WO2006029459A1

WO2006029459A1 - Antibodies specific for hepatocellular carcinoma and other carcinomas and uses thereof

Info

Publication number: WO2006029459A1
Application number: PCT/AU2005/001399
Authority: WO
Inventors: Vincent Batori; David S Wilson
Original assignee: Evogenix, Inc; Evogenix Ltd
Priority date: 2004-09-13
Filing date: 2005-09-13
Publication date: 2006-03-23
Also published as: CN101076542A

Abstract

Antibodies, humanized antibodies, super-humanizedTM antibodies, resurfaced antibodies, antibody fragments, derivatized antibodies, and conjugates of same with or without cytotoxic agents, which specifically bind to hepatocellular cancers and other cancers, and reduce the survival of such target cells, are provided. Said antibodies and fragments thereof may be used in the treatment of patients with tumors that express elevated levels of the antigen, such as hepatocellular carcinoma.

Description

ANTIBODIES SPECIFIC FOR HEPATOCELLULAR CARCINOMA AND OTHER CARCINOMAS AND USES THEREOF

FIELD OF THE INVENTION

The present invention relates to antibodies and antibody fragments that bind to CD147 and to their use in the treatment of cancer cells that express CD147. More particularly, the invention relates to non-human antibodies and antibody fragments that that have been mutated so as reduce their immunogenicity in humans.

BACKGROUND OF THE INVENTION

Primary liver cancer, hepatocellular carcinoma (HCC), is the fourth most common type of cancer worldwide, causing an estimated 1.3M deaths annually (Venook, 2000). This ranges from 17,000 cases per year in the United States to more than 250,000 cases per year in China. The number of cases in the United States is expected to increase dramatically in the next 10-20 years due to the hepatitis C epidemic that currently infects about 4 million Americans. Because HCC is often only detected in its late stages, it is one of the most lethal forms of cancer, with an average survival after onset of initial symptoms of only 3-4 months. New forms of therapy are therefore desperately needed.

Although surgical resection has been the treatment of first-choice, less than 10% of patients are suitable for curative resection due to factors such as metastasis and cirrhosis (for review, see Alsowmely & Hodgson, 2002). Some procedures have been shown to shrink tumors to allow them to be resectable, such as transarterial chemoembolisation, combined chemotherapy and radiation, hepatic artery ligation with infusion of a chemotherapeutic agent, radioimmunotherapy, etc. (Lau, 2002). Only a small proportion of patients respond significantly to these therapies.

HCC is one of the most chemotherapeutic drug-resistant types of tumors known (Alsowmely & Hodgson, 2002). A wide variety of agents have been investigated and are in use, including 5-fluorouracil, doxorubicin, mitoxantrone, cisplatin and etoposide. Systemic chemotherapy, however, has shown little clinical benefit and all of these therapies have been associated with significant toxicity. To date, there are no drugs specifically FDA-approved for HCC. Several HCC-specific antibodies conjugated to toxins or radioisotopes are under development for HCC therapy. These include the following:

The murine antibody called Hab 18 was generated by immunizing mice with human HCC tissue (Chen, 1992). After confirming the antibody's tumor-specificity, the antigen was identified as the cell surface protein CD 147 (also known as Hab 18 G, basigin, EMMPRIN, neurothelin, 5Al 1, CE9, HT7, M6, OX-47, or gp42) using expression cloning (Chen, et al., 1999). Other articles describing the antibody include Xing et al. (2003), Liu et al. (2003), Li et al. (2002), Lou et al. (2002), Yang et al. (2001), Bian et al. (2000), Qiu et al. (1998), Sui et al. (1998), Sui et al. (1996), Sui (1992) and Ji (1991).

One of the most advanced HCC-specific antibodies under development for HCC therapy is the ¹³¹I-HaM 8 antibody conjugate, a radio-iodinated mouse monoclonal antibody fragment that specifically binds to HCC cells. The murine Hab 18 antibody was processed to a F(ab)'₂ fragment by proteolytic methods, radiolabeled, and administered to mice bearing human HCC tumors (Lou et al, 2002; Bian et α/,,2000; Qiu et al, 1998). These studies demonstrated significant tumor uptake in this animal model and validated the approach of using radiolabeled versions of this antibody, particularly as F(ab)'₂ fragments, therapeutically in humans. Versions of the antibody fused to staphylococcal enterotoxin A were also prepared and shown to be effective in killing HCC cells (Yang et al, 2001).

CD147, the target of Habl8, is one of the most ubiquitously expressed cancer antigens known. Its functions are not fully understood, but its most relevant role in cancer appears to be in increasing expression of matrix metalloproteins (MMPs) that are involved in the degradation of basement membranes - a necessary step for tumor cells to break free and metastasize to new locations. The amino acid sequence of human CD147 is shown in SEQ ID NO: 70. CD147 is composed of 269 amino acids, has a 187 amino acid extracellular domain composed of two C2-type immunoglobulin loops with three Asn-linked oligosaccharides, a 24 amino acid transmembrane domain, a short single transmembrane domain and a 39 amino acid cytoplasmic domain. The first immunoglobulin domain is required for counter receptor activity, involved in MMP induction and oligomerization. The second immunoglobulin domain is required for association with caveolin-1 which leads to decreased self-association on the cell surface.

Degradation of basement membrane by matrix metalloproteinases is one of the most critical steps in various stages of tumor disease progression, including tumor angiogenesis, tumor growth, local invasion and subsequent distant metastasis (Nelson et al, 2000; Fang et al, 2000; Bergers et al, 2000). MMPs are a family of more than 25 metal-dependent endopeptidases that share a common modular domain structure. MMPs are overproduced in the tumor local environment. Collectively, these enzymes are capable of cleaving all of the extracellular matrix components of the parenchymal and vascular basement membranes that normally are mechanical barriers to tumor cell migration and invasion. High levels of MMPs are correlated with increased tumor invasion capacity both in vitro and in vivo (Gilles et al, 1994; Gilles et al, 1996). Imbalances in the production of MMPs and their endogenous inhibitors, tissue inhibitors of metalloproteinases (TIMPs), result in tumor angiogenesis and metastasis (Fang et al, 2000; Bergers et al, 2000; Lokeshwar et al, 1993; Powell et al, 1993; Wood et al, 1997; Sehgal et al, 1998; Kuniyasu et al, 2000)

Because of its ability to increase MMP expression, CD 147 has also been named extracellular matrix metalloproteinase inducer (EMMPRIN) (Guo et al, 1997; Lim et al, 1998). EMMPRIN was originally purified from the plasma membrane of cancer cells as a glycoprotein of M. Wt 58,000, with about half of its mass consisting of carbohydrate. It was designated tumor collagenase stimulating factor (TCSF) because of its ability to stimulate fibroblast synthesis of collagenase- 1 (MMP-I) (Biswas et al, 1995; Ellis et al, 1989). Subsequent research demonstrated that EMMPRTN also induced fibroblasts to synthesize MMP -2, MMP-3, as well as the membrane-type 1 MMP (MTl-MMP) and MT2-MMP, which function as endogenous activators of MMP-2 (Guo et al, 1997; Kataoka et al, 1993; Sameshima et al, 2000).

Therefore, EMMPRTN functions as an upstream modulator of MMP production in tumor local environment. EMMPRIN-positive tumor cells stimulate neighbouring fibroblast cells to express MMPs and therefore facilitate tumor invasion and metastasis. Several clinical studies have demonstrated high levels of EMMPRIN expression in tumor compartments as compared to peritumoral stromal tissues. These tumors include those originating in lung (Polette et al, 1997), breast (Polette et al, 1997), bladder (Muraoka et al, 1993; Javadpour & Guirguis 1992) and glioma (Sameshima et al, 2000a).

By transcriptome analysis of individual tumor cells isolated from the bone marrow of cancer patients and comparative genomic hybridization technique, EMMPRIN was shown to be the most frequently expressed protein in primary tumors and in micrometastatic cells (Klein et al, 2002), suggesting a central role in tumor progression and early metastasis.

The biological significance of increased expression of EMMPRIN in tumor cells was investigated by in vitro studies using recombinant EMMPRIN or native EMMPRIN purified from tumor cells. EMMPRIN has been shown to stimulate the expression of various MMPs produced by fibroblasts (Guo et al, 1997; Lim et al, 1998). The induction occurred at the transcription level and is at least in part mediated by a mitogen-activated protein kinase (MAPK) p38 kinase signaling pathway (Lim et al, 1998). The role of EMMPRIN in tumor growth and metastasis was directly illustrated using EMMPRIN-overexpressing human breast cancer cells. MDA MB 463 cells are normally slow-growing cells when implanted into nude mice. When these cells were transfected with EMMPRIN, however, they adopted a more aggressive growth pattern, with both accelerated growth rate and metastatic phenotypes (Zucker et al, 2001). CD147 has also been shown to have homotypic adhesion characteristics, and it is unclear how and to what extent this is related to its MMP -inducing activity.

CD 147 is also expressed on activated T and B cells and other accessory cells that stimulate them (Deeg et al, 2001). For this reason, an antibody approach targeting this antigen was suggested for reducing the immune response to foreign tissue engraftment. Acute graft-versus-host disease (GVHD) is a major complication of allogeneic hemopoietic stem cell transplantation (HSCT). GVHD is triggered by donor T cells that recognize recipient tissues as foreign. Resultant cell activation and cytokine release lead to the destruction of host tissue and GVHD. Despite in vivo pharmacologic prophylaxis of GVHD, acute GVHD requiring additional therapy develops in as many as 30% (with human leukocyte antigen [HLA] -identical donors) to 75% (with unrelated donors) of transplant recipients. The standard for initial therapy of acute GVHD is methylprednisolone. Steroid-resistant GVHD often develops in patients with initial diagnoses of more severe disease. The anti-CD 147 antibody ABX-CBL is a murine immunoglobulin M (IgM), described in PCT publication number WO 99/45031 (Abgenix Inc.). Activated T cells and B cells are depleted by ABX-CBL in vitro, whereas resting lymphocytes remain unaffected. ABX-CBL inhibited the in vitro mixed lymphocyte reaction by depleting monocytes, dendritic cells, and activated lymphocytes through a complement- dependent cytotoxic mechanism. In a published (Deeg et al, 2001) phase I/II trial, 27 patients with steroid-refractory acute GVHD received ABX-CBL at 0.01 (presumed no effect dose), 0.1, 0.2, or 0.3 mg/kg per day, and an additional 32 patients were given ABX-CBL at 0.2 or 0.15 mg/kg per day. All patients had undergone allogeneic transplantation and received GVHD prophylaxis. Among 51 patients which were able to be evaluated for efficacy, 26 (51%) responded, including 13 with complete responses (CR) and 13 with partial responses (PR). CR lasting 14 days or longer or PR lasting 7 days or longer occurred in 21 (41%; 8 CR, 13 PR) patients. Myalgias (muscle pain) were dose-limiting, resolved without sequelae and could be managed by pre-treatment with narcotics. At 6 months after the initiation of ABX-CBL therapy, 26 (44%) patients were still alive.

The Habl8 antibody is described in the US publication number 2005/0176933 (Chen Zhinan et al.). Other antibodies against CD147 are known as described in the tables below:

CD147-specifϊc MAbs newly assigned at Sixth International Workshop and Conference on Human Leukocyte Differentiation Antigens. Kobe, Japan, November 10-14, 1996.

Name (Workshop IDs) Source or Reference

AAA6 Walter Knapp, Vienna; Kasinrerk et al., 1992.

UM-8D6 David A. Fox, Michigan

HIl 97 De-Cheng Shen, Tianjin

HIM6 De-Cheng Shen, Tianjin

H84 ! Kimitaka Sagawa, Kurume Selection of other CD147-specific reference antibodies

There remains a need in the art for effective treatments for cancer and in particular primary liver cancers.

SUMMARY OF THE INVENTION

The present inventors have found that non-human antibodies or antibody fragments targeting CD 147 provide potentially useful agents for treating human cancers when the variable region of the antibody or antibody fragment has been mutated so as to reduce the immunogenicity to humans.

Thus, in a first aspect, the present invention provides a non-human antibody or antibody fragment comprising a variable region, which antibody or antibody fragment specifically binds to CD 147, wherein the variable region has been mutated so as to reduce the immunogenicity of the antibody or antibody fragment in humans.

The antibodies or antibody fragments of the invention provide a potentially useful new method for treating cancer, and in particular liver cancer, which avoid the undesirable side effects associated with existing antibody treatments. In a second aspect, the invention provides a conjugate comprising the antibody or antibody fragment of the first aspect linked to a cytotoxic agent.

In a third aspect, the invention provides a pharmaceutical composition comprising a pharmaceutically acceptable carrier and the antibody or antibody fragment of the first aspect or the conjugate of the second aspect. In a further aspect, the invention provides a method for inhibiting the growth of a cancer cell comprising contacting the cell with an antibody or antibody fragment, conjugate or pharmaceutical composition according to the first, second or third aspect of the invention, respectively.

In a further aspect, the invention provides a method for treating a patient having a cancer comprising administering to the patient an effective amount of an antibody or antibody fragment, conjugate or pharmaceutical composition according to the first, second or third aspect of the invention, respectively.

In a further aspect, the invention provides a method for diagnosing a subject suspect of having a cancer, the method comprising administering to the subject a conjugate comprising an antibody or antibody fragment according to the first aspect linked to a label capable of being detected within the subject.

In a further aspect, the invention provides an improved antibody or antibody fragment that specifically binds to CD147, prepared by: (a) providing a DNA encoding an antibody or fragment thereof comprising at least one sequence selected from the group consisting of SEQ ID NOS: 1 to 69; (b) introducing at least one nucleotide mutation, deletion or insertion into said DNA such that the amino acid sequence of said antibody or antibody fragment encoded by said DNA is changed; (c) expressing said antibody or antibody fragment; (d) screening said expressed antibody or antibody fragment for said improvement, whereby said improved antibody or antibody fragment is prepared.

In a further aspect, the invention provides an antibody or antibody fragment that binds to CD 147, wherein the heavy chain variable region is a mutated version of SEQ ID NO: 7 containing between 1 and 20 mutations in the heavy chain variable region that cause it to have a higher amino acid sequence identity to a natural human heavy chain than that of the heavy chain variable region described by SEQ ID NO: 7.

In a further aspect, the invention provides a humanized antibody heavy chain, said heavy chain comprising one or more CDRs derived from the murine antibody Habl8 heavy chain and comprising a framework sequence derived from a human antibody variable heavy chain having a CDRl of canonical structure type 1 and a CDR2 of length 19 according the antibody numbering system of Kabat.

In a further aspect, the invention provides a humanized antibody heavy chain, said heavy chain comprising one or more CDRs derived from the murine antibody Habl8 heavy chain and comprising a framework sequence derived from a human antibody variable heavy chain having a CDRl of canonical structure type 1 and a CDR2 of canonical structure type 3.

In a further aspect, the invention provides a humanized antibody light chain, said light chain comprising one or more CDRs derived from the murine antibody Habl8 light chain and comprising a framework sequence derived from a human antibody variable light chain having a CDRl of canonical structure type 2, a CDR2 of canonical structure type 1 and a CDR3 of canonical structure type 1. In further aspects the invention provides polynucleotides encoding an antibody or antibody fragment according to various aspects of the invention, and vectors comprising the polynucleotides of the invention.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1: (i) Shows the Habl8 antibody heavy chain V regions aligned with human germline variable exons with canonical structures 1 (for CDRl) and 4 (for CDR2). (ii) Shows the Habl8 antibody CDR3 aligned with human JH sequences, (iii) Shows scoring of human germline variable exons with canonical structures 1 (for CDRl) and 4 or "4-like" (for CDR2) according to homology in the CDR regions to the Hablδ heavy chain V region, (iv) Shows three humanized Habl8 heavy chain variable region sequences (h3-72*01, h3-49*01 and h3-73*01) aligned with the Hablδ murine variable region heavy chain sequence. Figure 2: Shows the sequence of the heavy chain variable region of Habl8 aligned with human germline V exon sequences with canonical structures 1 (for CDRl) and 3 (for CDR2).

Figure 3: (i) Shows the Habl8 antibody light chain V regions aligned with human germline variable exons with canonical structures 2 (for CDRl), 1 (for CDR2) and 1 (for CDR3). (ii) Shows scoring of human germline variable exons with canonical structures 2 (for CDRl), 1 (for CDR2) and 1 (for CDR3) according to homology to the Habl8 antibody light chain V region, (iii) Shows the Hablδ variable region light chain CDR3/J region aligned with human germline JKappa (JK) regions, (iv) Shows two humanized Habl8 light chains (hi -6*01 and h3-l l*01) aligned to the Habl8 light chain variable region.

Figure 4: Shows vector construct for the expression of super-humanized scFv antibody in which: LacP = Lactose promoter; Ld = Shine-Dalgarno sequence + initiating ATG + leader sequence; Vheavy = heavy chain variable domain; Vlight = light chain variable domain; Fl = Flag tag = DYKDDDDK; 6H = HHHHHH; T = in frame stop codon(*); Amp resistance = gene encoding antibiotic resistance; OR =origin of replication

Figure 5: Shows an ELISA assay demonstrating specific binding by a humanized anti-CD 147 antibody based on Habl8. KEY TO SEQUENCE LISTINGS

SEQ ID NO: 1 : Amino acid sequence of CDR 1 of murine heavy chain of Habl 8

SEQ ID NO: 2: Amino acid sequence of CDR 2 of murine heavy chain Habl 8 SEQ ID NO : 3 : Amino acid sequence of CDR 3 of murine heavy chain Hab 18

SEQ ID NO: 4: Amino acid sequence of CDR 1 of murine light chain Hab 18

SEQ ID NO: 5: Amino acid sequence of CDR 2 of murine light chain Habl8

SEQ ID NO: 6: Amino acid sequence of CDR 3 of murine light chain Habl 8

SEQ ID NO: 7: Amino acid sequence of V region of murine heavy chain Habl8 SEQ ID NO: 8: Amino acid sequence of human germline variable region 3-15*01

SEQ ID NO: 9: Amino acid sequence of human germline variable region 3-15*03

SEQ ID NO: 10: Amino acid sequence of human germline variable region 3-49*01

SEQ ID NO: 11 : Amino acid sequence of human germline variable region 3-72*01

SEQ ID NO: 12: Amino acid sequence of human germline variable region 3-73*01 SEQ ID NO: 13: Amino acid sequence of Hablδ CDR3/JH region

SEQ ID NO: 14: Amino acid sequence of human germline JH region (JHl)

SEQ ID NO: 15: Amino acid sequence of human germline JH region (JH2)

SEQ ID NO: 16: Amino acid sequence of human germline JH region (JH3)

SEQ ID NO: 17: Amino acid sequence of human germline JH region (JH4) SEQ ID NO: 18: Amino acid sequence of human germline JH region (JH5)

SEQ ID NO: 19: Amino acid sequence of human germline JH region (JH6)

SEQ ID NO: 20: Amino acid sequence of humanized Hab 18 heavy chain variable region based on all 3 murine CDRS and human germline variable region sequences 3-72*01/JH2 SEQ ID NO: 21 : Amino acid sequence of humanized Habl 8 heavy chain variable region based on all 3 murine CDRS and human germline variable region sequences h3-49*01/JH2

SEQ ID NO: 22: Amino acid sequence of humanized Hab 18 heavy chain variable region based on all 3 murine CDRS and human germline variable region sequences h3 -73 * 01 /JH 1

SEQ ID NO: 23: Amino acid sequence of human germline variable region 1-2*01

SEQ ID NO: 24: Amino acid sequence of human germline variable region 1-2*02

SEQ ID NO: 25: Amino acid sequence of human germline variable region 1-3*01

SEQ ID NO: 26: Amino acid sequence of human germline variable region 1-8*01 SEQ ID NO: 27: Amino acid sequence of human germline variable region 1-46*01

SEQ ID NO: 28: Amino acid sequence of human germline variable region 1-58*01 SEQ ID NO: 29: Amino acid sequence of human germline variable region 3-07*01 SEQ ID NO: 30: Amino acid sequence of human germline variable region 3-09*01 SEQ ID NO: 31 : Amino acid sequence of human germline variable region 3-11*01 SEQ ID NO: 32: Amino acid sequence of human germline variable region 3-20*01 SEQ ID NO: 33: Amino acid sequence of human germline variable region 3-21*01 SEQ ID NO: 34: Amino acid sequence of human germline variable region 3-23*01 SEQ ID NO: 35: Amino acid sequence of human germline variable region 3-23*02 SEQ ID NO: 36: Amino acid sequence of human germline variable region 3-30*01 SEQ ID NO: 37: Amino acid sequence of human germline variable region 3-30*02 SEQ ID NO: 38: Amino acid sequence of human germline variable region 3-30*06 SEQ ID NO : 39 : Amino acid sequence of human germline variable region 3-30*18 SEQ ID NO: 40: Amino acid sequence of human germline variable region 3-30-3*01 SEQ ID NO: 41 : Amino acid sequence of human germline variable region 3-33*01 SEQ ID NO: 42: Amino acid sequence of human germline variable region 3-43*01 SEQ ID NO: 43 : Amino acid sequence of human germline variable region 3-48*01 SEQ ID NO: 44: Amino acid sequence of human germline variable region 3-48*03 SEQ ID NO: 45: Amino acid sequence of human germline variable region 3-64*01 SEQ ID NO: 46: Amino acid sequence of human germline variable region 3-74*01 SEQ ID NO: 47: Amino acid sequence of Light chain V region of murine Hab 18 SEQ ID NO: 48 : Amino acid sequence of germline variable region for human 1 -

6*01 SEQ ID NO: 49: Amino acid sequence of germline variable region for human 1-

9*01 SEQ ID NO: 50: Amino acid sequence of germline variable region for human 1- 12*01

SEQ ID NO: 51: Amino acid sequence of germline variable region for human ID-

16*01

SEQ ID NO: 52: Amino acid sequence of germline variable region for human ID- 16*02 SEQ ID NO: 53: Amino acid sequence of germline variable region for human 1-

17*01 SEQ ID NO: 54: Amino acid sequence of germline variable region for human 1-

33*01 SEQ ID NO: 55: Amino acid sequence of germline variable region for human 1- 39*01 SEQ ID NO: 56: Amino acid sequence of germline variable region for human ID-

43*01 SEQ ID NO: 57: Amino acid sequence of germline variable region for human 3-

11*01 SEQ ID NO: 58: Amino acid sequence of germline variable region for human 3-

11*02 SEQ ID NO: 59: Amino acid sequence of germline variable region for human 3-

15*01

SEQ ID NO: 60: Amino acid sequence of CDR3/J region of murine light chain Habl8

SEQ ID NO: 61 : Amino acid sequence of human germline JKappa region JKl SEQ ID NO: 62: Amino acid sequence of human germline JKappa region JK2 SEQ ID NO: 63 : Amino acid sequence of human germline JKappa region JK3 SEQ ID NO: 64: Amino acid sequence of human germline JKappa region JK4 SEQ ID NO: 65: Amino acid sequence of Human germline JKappa region JK5

SEQ ID NO: 66: Amino acid sequence of a humanized variable region light chain of

Habl8 based on human variable (kappa) region light chain sequences hl-6*01/JK3

SEQ ID NO: 67: Amino acid sequence of a humanized variable region light chain of Habl8 based on human variable (kappa) region light chain sequences h3-l l*01/JK3 SEQ ID NO: 68: Amino acid sequence of a super-humanized™ antibody scFv fragment, scFv-1, based on the heavy chain SEQ ID NO: 20 and the light chain SEQ ID NO: 66 SEQ ID NO: .69: Amino acid sequence of a super-humanized™ antibody scFv fragment, scFv-2, based on the heavy chain SEQ ID NO: 20 and the light chain SEQ ID NO: 67 SEQ ID NO : 70 : Amino acid sequence of CD 147.

DETAILED DESCRIPTION OF THE INVENTION

In the description that follows, reference is made to various publications that may assist one of ordinary skill in the art in understanding and practicing the invention to its fullest extent. Therefore, each reference cited in the description that follows is incorporated herein by reference in its entirety. Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is solely for the purpose of providing a context for the present invention. It is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application.

Definitions

"Mature antibody genes" are genetic sequences encoding immunoglobulins that are expressed, for example, in a lymphocyte such as a B cell, in a hybridoma or in any antibody-producing cell that has undergone a maturation process so that the particular immunoglobulin is expressed. The term includes mature genomic, cDNA or other nucleic acid sequence that encodes such mature genes, which have been isolated and/or recombinantly engineered for expression in other cell . types. Mature antibody genes have undergone various mutations and rearrangements that structurally distinguish them from antibody genes encoded in all cells other than lymphocytes. Mature antibody genes in humans, rodents, and many other mammals are formed by fusion of V and J gene segments in the case of antibody light chains and fusion of V, D, and J gene segments in the case of antibody heavy chains. Many mature antibody genes acquire point mutations subsequent to fusion, some of which increase the affinity of the antibody protein for a specific antigen.

"Germline antibody genes" or gene fragments are immunoglobulin sequences encoded by non-lymphoid cells that have not undergone the maturation process that leads to genetic rearrangement and mutation for expression of a particular immunoglobulin. One of the advantages provided by various embodiments of the present invention stems from the recognition that germline antibody genes are more likely than mature antibody genes to conserve essential amino acid sequence structures characteristic of individuals in the animal species, hence less likely to be recognized as foreign when used therapeutically in that species.

A "CDR" is the complementarity determining region within antibody variable sequences. There are three CDRs in each of the variable heavy and variable light sequences designated CDRl, CDR2 and CDR3, for each of the variable regions. The exact boundaries of these CDRs have been defined differently according to different systems, however, all have overlapping residues in what constitute the so called "hypervariable regions" within the variable sequences. The system described by Kabat et al. (1991) and Wu and Kabat (1970) not only provides an unambiguous residue numbering system applicable to any variable region of an antibody, but also provides precise residue boundaries defining the three CDRs. These CDRs may be referred to as Kabat CDRs. Chothia and coworkers (Chothia et al. (1987); Chothia et al. (1992); Tomlinson et al. (1995)) found that certain sub portions within Kabat CDRs adopt nearly identical peptide backbone conformations, despite having great diversity at the level of amino acid sequence. These sub portions were designated as Ll, L2 and L3 or Hl, H2 and H3 where the "L" and the "H" designates the light chain and the heavy chains regions, respectively. These regions may be referred to as Chothia CDRs, which have boundaries that overlap with Kabat CDRs. Table 1 illustrates the overlap of Chothia and Kabat CDRs according to the residue numbering system of Kabat.

TABLE 1 : Overlap of Chothia and Kabat CDRs according to the residue numbering system of Kabat

Other boundaries defining CDRs overlapping with the Kabat CDRs have been described by Padlan et al. (1995) or MacCallum et al. (1996). Still other CDR boundary definitions may not strictly follow one of the above systems, but will nonetheless overlap with the Kabat CDRs, although they may be shortened or lengthened in light of prediction or experimental findings that particular residues or groups of residues or even entire CDRs do not significantly impact antigen binding. The methods used herein may utilize CDRs defined according to any of these systems, although preferred embodiments use Kabat- or Chothia-defined CDRs.

"Framework" or "framework sequences" are the remaining sequences of a variable region minus the CDRs. Because the exact definition of a CDR sequence can be determined by different systems, the meaning of a framework sequences is subject to correspondingly different interpretations. To clarify the meaning used herein, a framework sequence means those sequences within the variable region of an antibody other than those defined to be CDR sequences, so that the exact sequence of a framework depends only on how the CDR is defined. For example, the CDRs used in the methods provided herein are usually a subset of what is considered a Kabat CDR, but in the case of CDRl of heavy chains for example, also includes residues that are classified as framework residues in the Kabat system.

"Canonical CDR structure types" are the structure types designated by Chothia (Chothia et al. (1987); Chothia et al. (1992); Tomlinson et al. (1995)). Chothia and coworkers found that critical portions of the CDRs of many antibodies adopt nearly identical peptide backbone conformations, despite great diversity at the level of amino acid sequence. Accordingly, Chothia defined for each CDR in each chain one or a few "canonical structures". Each canonical structure specifies primarily a set of peptide backbone torsion angles for a contiguous segment of amino acid residues forming a loop. The canonical CDR structure types defined by Chothia are listed in Table 2.

TABLE 2 : Canonical CDR structure types defined by Chothia

"Corresponding CDRs" refer relatively to the CDRs between two different variable sequences that correspond in position within the two different variable sequences. Thus, for example, a mouse light chain CDRl corresponds to a human light chain CDRl, and vice a versa, because each maps to a defined position in a Kabat numbering system, whether or not the actual boundary of the CDR is defined by Kabat, Chothia or some other system. Similarly, "corresponding" residues, sequences or amino acids refer relatively to the residue positions between two different peptide sequences mapped by the Kabat numbering system.

"Substantially the same" as applied to an amino acid sequence is defined herein as a sequence with at least about 90%, and more preferably at least about 95% sequence identity to another amino acid sequence, as determined by the FASTA search method in accordance with Pearson and Lipman (1988).

"Non-human antibodies" may be derived from any source. Preferably they are mammalian in origin. More preferably they are murine in origin. The non-human antibodies have a binding specificity for CD 147. An antibody is considered to have a binding specificity for CD 147 if it has an association constant of at least 10⁵M^"1 and at least a 10-fold lower affinity constant for most other proteins; more preferably, it should have an association constant of at least 10⁶M^"1 and at least a 10-fold lower affinity constant for most other proteins; still more preferably, it should an association constant of at least 10⁷M^"1 and at least a 100-fold lower affinity constant for most other proteins; still more preferably, it should an association constant of at least 10⁸M^"1 and at least a 100-fold lower affinity constant for most other proteins; still more preferably, it should an association constant of at least 10⁹M^"1 and at least a 1000-fold lower affinity constant for most other proteins. An "antibody fragment' or "fragment of antibody" preferably retains a binding specificity for CD 147 comparable with or the same as that of the antibody from which it is derived. Fragments include polypeptides with amino acid sequences substantially the same as the amino acid sequence of the variable regions of the antibodies. Such fragments may contain one or both Fab fragments or the F(ab')₂ fragment. Preferably the antibody fragments contain all six complementarity determining regions of the whole antibody, although fragments containing fewer than all of such regions, such as three, four or five CDRs, are also functional and encompassed by the present invention.

Fragments also include single-chain antibody fragments, in particular known as single-chain variable region fragments (scFvs). These fragments contain at least one fragment of an antibody variable heavy-chain amino acid sequence (VH) tethered to at least one fragment of an antibody variable light-chain sequence (VL) with or without one or more interconnecting linkers. Such a linker may be a short, flexible peptide selected to assure that the proper three-dimensional folding of the (VL) and (VH) domains occurs once they are linked so as to maintain the target molecule binding- specificity of the whole antibody from which the single-chain antibody fragment is derived. Generally, the carboxyl terminus of the (VL) or (VH) sequence may be covalently linked by such a peptide linker to the amino acid terminus of a complementary (VL) and (VH) sequence. Single-chain antibody fragments may be generated by molecular cloning, antibody phage display or similar techniques. These proteins may be produced either in eukaryotic cells or prokaryotic cells, including the bacterium E.coli. scFv's can also be fused to other parts if antibody molecules. For example, scFv's can be attached, via a natural or artificial peptide linker, to the CH2- CH3 region of an IgG to form a divalent scFv-Fc construct.

The non-human antibody or antibody fragment comprises a variable region. At the very least, the non-human antibody or antibody fragment contains a variable region heavy or light chain sequence, preferably a heavy chain sequence. Preferably, the non- human antibodies and antibody fragments contain both variable heavy and light chain sequences. The non- human antibody or antibody fragment is sometimes referred to as the subject antibody or antibody fragment.

"Reducing the immunogenicity" of the antibody or antibody fragment in humans means that a molecule is created that has a reduced tendency to elicit an immune response when administered to a human (Hwang et al, 2005a); reduced tendency means that fewer patients display an immune response to the molecule than to the parent molecule, or that the degree of an immune response in certain patients is reduced compared to the parent molecule. Degree of immune response can be determined by measuring the titer of antibodies in the serum of patients, or the ability of the molecule to activate T cells in an in vitro assay.

The phrase "mutated so as to reduce the immunogenicity" is intended to encompass any type of modification or manipulation made to a non-human antibody variable region that results in a reduced tendency to elicit an immune response when administered to a human. The mutation may involve, for example, one or more point mutations in the framework region of a variable region, or it may involve replacement of one or more regions of a variable region framework sequence. It will be appreciated that it is not necessary to generate the original non-human variable region for subsequent mutation or manipulation. A mutated variable region may be generated de novo, for example, by synthetic means or by grafting CDR loops onto a selected human framework sequence. Examples of mutation techniques that can be used to reduce immunogenicity include deimmunizing, humanizing, super-humanizing™ and resurfacing of antibodies or antibody fragments

The present invention also encompasses "functional equivalents" of antibodies and antibody fragments which have been mutated in their variable regions so as to reduce immunogenicity to humans. Functional equivalents have binding characteristics that are comparable to those of the antibodies and antibody fragments and include, for example, chimerized and single chain antibodies. Methods of producing such functional equivalents are disclosed in PCT publication number WO 93/21319 (Cetus Oncology Corp), EP publication number 239,400 (Winter); PCT publication number WO 89/09622 (Protein Design Labs Inc), European publication number 338,745 (Celltech Ltd) and European publication number 332,424 (Hybritech Inc), which are incorporated in their respective entireties by reference.

Functional equivalents include polypeptides with amino acid sequences substantially the same as the amino acid sequence of the variable or hypervariable regions of the antibodies or antibody fragments of the invention.

Functional equivalents also include fragments of antibodies that have the same, or comparable binding characteristics to those of the whole antibody. Such fragments may contain one or both Fab fragments or the F(ab')₂ fragment. Preferably the antibody fragments contain all six complementarity determining regions of the whole antibody, although fragments containing fewer than all of such regions, such as three, four or five CDRs, are also functional. Further, the functional equivalents may be or may combine members of any one of the following immunoglobulin classes: IgG, IgM, IgA, IgD, or IgE, and the subclasses thereof.

Functional equivalents also encompass antibodies and antibody fragments which contain modifications which are essentially tantamount to conservative substitutions throughout their sequence which do not alter to any significant degree the binding specificity of the antibodies or antibody fragments.

Exemplary conservative substitutions.

Furthermore, if desired, non-naturally occurring amino acids or chemical amino acid analogues can be introduced as a substitution or addition into the polypeptide of the present invention. Such amino acids include, but are not limited to, the D-isomers of the common amino acids, 2,4-diaminobutyric acid, alpha-amino isobutyric acid, 4- aminobutyric acid, 2-aminobutyric acid, 6-amino hexanoic acid, 2-amino isobutyric acid, 3 -amino propionic acid, ornithine, norleucine, norvaline, hydroxyproline, sarcosine, citrulline,homocitruUine, cysteic acid, t-butylglycine, t-butylalanine, phenylglycine, cyclohexylalanine,fluoro-amino acids, designer amino acids such as (3- methyl) amino acids, C-alpha-methyl amino acids, N-alpha-methyl amino acids, and amino acid analogues in general.

Functional equivalents also include chemically modified derivatives of antibodies and antibody fragments which may provide advantages such as increasing stability and circulating time of the polypeptide. The chemical moieties for derivitization may be selected from water-soluble polymers such as polyethylene glycol, ethylene glycol/propylene glycol copolymers, carboxymethylcellulose, dextran, polyvinyl alcohol and the like.

Furthermore, the antibodies or antibody fragments may be differentially modified during or after synthesis, e.g. , by biotinylation, benzylation, glycosylation, acetylation, phosphorylation, amidation, derivatization by known protecting/blocking groups, proteolytic cleavage, etc. The antibodies or antibody fragments may be modified at random positions within the molecule, or at predetermined positions within the molecule and may include one, two, three or more attached chemical moieties. These modifications may, for example, serve to increase the stability and/or bioactivity of the binding moieties of the invention.

Diverse antibodies and antibody fragments, as well as antibody mimics may be readily produced by mutation, deletion and/or insertion within the variable and constant region sequences that flank a particular set of CDRs. Thus, for example, different classes of antibodies are possible for a given set of CDRs by substitution of different heavy chains, whereby, for example, IgGl -4, IgM, IgAl -2, IgD, IgE antibody types and isotypes may be produced. Similarly, artificial antibodies within the scope of the invention may be produced by embedding a given set of CDRs within an entirely synthetic framework such as a domain of a human fibronectin polypeptide sequence, as disclosed in US 6,818,418.

The term "variable" is used herein to describe certain portions of the variable domains that differ in sequence among antibodies and are used in the binding and specificity of each particular antibody for its antigen. However, the variability is not usually evenly distributed through the variable domains of the antibodies. It is typically concentrated in three segments called complementarity determining regions (CDRs) or hypervariable regions both in the light chain and the heavy chain variable domains. The more highly conserved portions of the variable domains are called the framework (FR). The variable domains of heavy and light chains each comprise four framework regions, largely adopting a beta-sheet configuration, connected by three CDRs, which form loops connecting, and in some cases forming part of the beta-sheet structure. The CDRs in each chain are held together in close proximity by the FR regions and, with the CDRs from the other chain, contribute to the formation of the antigen binding site of antibodies (Kabat et al, 1991). The constant domains are not generally involved directly in binding an antibody to an antigen, but exhibit various effector functions, such as participation of the antibody in antibody-dependent cellular toxicity.

Reducing immunogenecity

The advent of recombinant DNA methodology enabled structural engineering of antibody genes and production of modified antibody molecules with properties not obtainable by hybridoma technology. In the therapeutic arena, one aim of this methodology has been to reduce the immunogenicity in humans of rodent monoclonal antibodies by modifying their primary amino acid structure. Reduction of the immunogenicity of therapeutic antibodies is desirable because induction of an immune response can cause a spectrum of adverse effects in a patient, ranging from accelerated elimination of the therapeutic antibody, with consequent loss of efficacy, to fatal anaphylaxis at the most extreme.

One strategy to reduce immunogenicity of foreign monoclonal antibodies has been to replace the light and heavy chain constant domains of the monoclonal antibody with analogous domains of human origin, leaving the variable region domains of the foreign antibody intact. The variable region domains of the light and heavy chains are responsible for the interaction between the antibody and the antigen. The joining domains connecting variable domains to constant domains are situated in a region remote from the site of antigen-binding, therefore, the joining domains between variable and constant domains generally do not interfere with antigen-binding. Chimeric antibody molecules having mouse variable domains joined to human constant domains usually bind antigen with substantially the same affinity constant as the mouse antibody from which the chimeric was derived. Such chimeric antibodies are less immunogenic in humans than their fully murine counterparts. Nevertheless, antibodies that preserve entire murine variable domains tend to provoke immune responses in a substantial fraction of patients (Hwang et al, 2005a). For example, INFLIXIMAB™, a widely prescribed chimeric antibody that is considered safe, induced a human anti- chimeric antibody response in 7 out of 47 Crohns Disease patients (Rutgeerts et al., 1999). In the light of these undesirable antibody responses in humans, researchers have experimented with methods of obtaining variable domains with more human character. One category of methods, frequently referred to as "humanizing," aims to convert the variable domains of murine monoclonal antibodies to a more human form by recombinantly constructing an antibody variable domain having both mouse and human character. Humanizing strategies are based on several consensual concepts derived from antibody structure data. First, variable domains contain contiguous tracts of peptide sequence that are conserved within a species, but which differ between evolutionarily remote species, such as mice and humans. Second, other contiguous tracts are not conserved within a species, but differ even between antibody-producing cells within the same individual. Third, contacts between antibody and antigen occur principally through the non-conserved regions of the variable domain. Fourth, the molecular architecture of antibody variable domains is sufficiently similar across species that correspondent amino acid residue positions between species may be identified based on position alone, without experimental data. Humanization strategies share the premise that replacement of amino acid residues that are characteristic of murine sequences with residues found in the correspondent positions of human antibodies will reduce the immunogenicity in humans of the resulting antibody. Data supporting this concept has been published (Hwang et al., 2005a). However, replacement of sequences between species usually results in reduction of antibody-antigen binding. The art of humanization therefore lies in balancing replacement of the original murine sequence to reduce immunogenicity with the need for the humanized molecule to retain sufficient antigen binding to be therapeutically useful.

The present inventors have discovered and improved novel antibodies that specifically bind to CD 147 on the cell surface. They have successfully lowered the immunogenicity to humans of antibodies targeting CD 147 whilst maintaining effective affinity of the antibodies for CD 147.

Thus in a first aspect, the present invention provides a non-human antibody or antibody fragment comprising a variable region, which antibody or antibody fragment specifically binds to CD 147, wherein the variable region has been mutated so as to reduce the immunogenicity of the antibody or antibody fragment in humans.

In one approach, referred to in this application as "resurfacing", and as exemplified by U.S. Patent No. 5,869,619 (Studnicka) and by Padlan (1991), characteristically human residues may be substituted for murine variable domain residues that are determined or predicted (i) to play no significant chemical role in the interaction with antigens and (ii) to be positioned with side chains projecting into the solvent. Thus, exterior residues remote from the antigen binding site are humanized, while interior residues, antigen binding residues, and residues forming the interface between variable domains remain murine. One disadvantage of this approach is that rather extensive experimental data is required to determine whether a residue plays no significant chemical role in antigen binding or will be positioned in the solvent in a particular three dimensional antibody structure. The present invention includes the use of such methods, and the products thereof, for reducing the immunogenicity of non- human, e.g. murine, antibodies.

In another more general approach, exemplified by U.S. Patent No. 5,225,539 (Winter) and Jones et al. (1986), contiguous tracts of murine variable domain peptide sequence considered conserved may be replaced with the corresponding tracts from a human antibody. In this more general approach, all variable domain residues are humanized except for the non-conserved regions implicated in antigen-binding. To determine appropriate contiguous tracts for replacement, a classification of antibody variable domain sequences as developed by Wu and Kabat (1970) may be utilised as described in US 5,225,539 and Jones et al. (1986).

Wu and Kabat pioneered the alignment of antibody peptide sequences, and their contributions in this regard were several-fold. First, through study of sequence similarities between variable domains, they identified correspondent residues that to a greater or lesser extent were homologous across all antibodies in all vertebrate species, inasmuch as they adopted similar three-dimensional structures, played similar functional roles, interacted similarly with neighbouring residues, and existed in similar chemical environments. Second, they devised a peptide sequence numbering system in which homologous immunoglobulin residues were assigned the same position number. One skilled in the art can unambiguously assign what is now commonly called Kabat numbering to any variable domain sequence without reliance on any experimental data beyond the sequence itself. Third, for each Kabat-numbered sequence position, Kabat and Wu calculated variability, by which is meant the finding of few or many possible amino acids when variable domain sequences are aligned. They identified three contiguous regions of high variability embedded within four less variable contiguous regions. Other workers had previously noted variability approximately in these regions (hypervariable regions) and posited that the highly variable regions represented amino acid residues used for antigen binding. Kabat and Wu formally demarcated residues constituting these variable tracts, and designated these "complementarity determining regions" (CDRs), referring to chemical complementarity between antibody and antigen. A role in three-dimensional folding of the variable domain, but not in antigen recognition, was ascribed to the remaining less-variable regions, which are now termed "framework regions". Fourth, Kabat and Wu established a public database of antibody peptide and nucleic acid sequences, which continues to be maintained and is well known to those skilled in the art. The humanization method disclosed by U.S. 5,225,539 (Winter) and Jones et al.

(1986) using the Kabat classification results in a chimeric antibody comprising CDRs from one antibody and framework regions from another antibody that differs in species origin, specificity, subclass, or other characteristics. However, no particular sequences or properties were ascribed to the framework regions. Indeed, U.S. 5,225,539 (Winter) teaches that any set of frameworks could be combined with any set of CDRs. Framework sequences have since been recognized as being important for conferring the three dimensional structure of an antibody variable region necessary to retain sufficient antigen binding. Thus, the general humanizing methods described by U.S. 5,225,539 (Winter) and Jones et al. (1986) have the disadvantage of frequently leading to inactive antibodies because these references do not provide information needed to rationally select among the many possible human framework sequences, those most likely to support antigen binding required by a particular CDR region from a non-human antibody. Subsequent developments in the field have been refinements within the scope of U.S. 5,225,539 (Winter) to deal with loss of avidity for antigen observed with some humanized antibodies relative to the avidity of the corresponding mouse antibodies. Avidity is a quantitative measure of partitioning of an antibody, in the presence of antigen under conditions approximating chemical equilibrium, between free and antigen-bound forms. For reactions in solution not subject to multivalent binding effects, avidity is the same as affinity, a biochemical equilibrium constant.

U.S. Patent No. 5,693,761 (Queen et al) discloses a refinement on U.S. 5,225,539 (Winter) for humanizing antibodies, and is based on the premise that ascribes avidity loss to problems in the structural motifs in the humanized framework which, because of steric or other chemical incompatibility, interfere with the folding of the CDRs into the binding-capable conformation found in the mouse antibody. To address this problem, U.S. 5,693,761 (Queen et al) teaches using human framework sequences closely homologous in linear peptide sequence to framework sequences of the mouse antibody to be humanized. Accordingly, the methods of U.S. 5,693,761 (Queen et al.) focus on comparing framework sequences between species. Typically, all available human variable domain sequences are compared to a particular mouse sequence and the percentage identity between correspondent framework residues is calculated. The human variable domain with the highest percentage is selected to provide the framework sequences for the humanizing project. U.S. 5,693,761 (Queen et al.) also teaches that it is important to retain in the humanized framework certain amino acid residues from the mouse framework critical for supporting the CDRs in a binding- capable conformation. Potential criticality is assessed from molecular models. Candidate residues for retention are typically those adjacent in linear sequence to a CDR or physically within 6 angstroms of any CDR residue. The present invention includes the use of such methods taught by U.S. 5,693,761 (Queen et al), and the products thereof, for the humanization of the Habl8 antibody.

In other approaches, criticality of particular framework amino acid residues is determined experimentally once a low-avidity humanized construct is obtained, by reversion of single residues to the mouse sequence and assaying antigen binding as described by Riechmann et al. (1988). Another example approach for identifying criticality of amino acids in framework sequences is disclosed by U.S. Patent Nos. 5,821,337 (Carter et al) and 5,859,205 (Adair et al). These references disclose specific Kabat residue positions in the framework, which, in a humanized antibody may require substitution with the correspondent mouse amino acid to preserve avidity. The present invention includes the use of such methods, and the products thereof, for the humanization of the Habl8 antibody.

One of the disadvantages of the refinements described in U.S. 5,693,761 (Queen et al), and the approaches of Riechmann et al. (1988), U.S. 5,821,337 (Carter et al.) and U.S. 5,859,205 (Adair et al) is that a very large number of human framework sequences are required for comparison, and/or the guidelines for preserving critical amino acid residues are not completely sufficient to predict functionality. Accordingly, the resulting frameworks constructed, which are part human and part mouse, still frequently exhibit human immunogenicity or lowered antigen binding, thereby requiring numerous iterations in framework construction to obtain a suitable framework for therapeutic uses.

In the resurfacing technology, molecular modeling, statistical analysis and mutagenesis are combined to adjust the non-CDR surfaces of variable regions to resemble the surfaces of known antibodies of the target host. Strategies and methods for the resurfacing of antibodies, and other methods for reducing immunogenicity of antibodies within a different host, are disclosed in U.S. Patent No. 5,639,641 (Pedersen et al), which is hereby incorporated in its entirety by reference.

The deimmunisation of antibodies or antibody fragments is achieved by the identification and elimination of potentially immunogenic murine T and B-cell epitopes from the non-human, e.g. mouse, antibody or antibody fragment. Removal of T cell epitopes may be achieved by identification of such epitopes from the variable regions of the antibody. For example, amino acid sequences of the variable region may be analyzed for the presence of MHC class II-binding motifs by a 3 -dimensional "peptide threading" method. Removal of at least one or all of the B cell epitopes from the variable region may be achieved by the "veneering" of surface residues where this will not interfere with antibody recognition.

Any of the above-described methods may be employed to mutate the variable region of an antibody or antibody fragment that specifically binds to CD 147 so as to reduce the immunogenicity of the antibody or antibody fragment in humans. In a preferred embodiment, immunogenicity is reduced by a CDR-grafting method. CDR-grafting methodologies provide a prescription for arriving at an appropriate human framework sequence for humanizing a subject non-human antibody. The preferred CDR-grafting methods of the present invention are the so-called super- humanization™ methods based on those described in detail in US publication number US 2003/0039649 (Foote) and PCT publication number WO 04/006955 (Foote) and further explained by Tan et al. (2002) and by Hwang et al. (2005b). U.S. Publication No. 2003/0039649 (Foote) designates the methods as SUPER-HUMANIZING ANTIBODIES™ and the antibodies made thereby as SUPER-HUMANIZED ANTIBODIES™. In contrast to previous CDR-grafting methods, where the choice of the humanized framework sequence was based on comparing the human frameworks to the subject (murine) frameworks, the methods for SUPER-HUMANIZING ANTIBODIES™ are based on the human antibody providing the humanized framework based on similarity of its CDRs to those of the subject non-human antibody, without regard to comparing the framework sequences between the two antibodies.

The similarity to the subject CDRs of candidate human antibody sequences is assessed for each domain at two levels. Primarily, identical three-dimensional conformations of CDR peptide backbones are sought. Experimentally determined atomic coordinates of the subject CDRs are seldom available, hence three-dimensional similarity is approximated by determining Chothia canonical structure types of the subject CDRs and excluding from further consideration candidates possessing different canonical structures. Secondarily, residue-to-residue homology between subject CDRs and the remaining human candidate CDRs is considered, and the candidate with the highest homology is chosen.

Choosing highest homology is based on various criteria used to rank candidate human variable regions having the same canonical structure as the subject the non- human variable regions. The criterion for ranking members of the selected set may be by amino acid sequence identity or amino acid homology or both. Amino acid identity is simply a score of position by position matches of amino acid residues. Similarity by amino acid homology is position by position similarity with respect to residue structure and character. Homology may be scored, for example, according to the tables and procedures described by Henikoff and Henikoff (1992) or by the BLOSUM series described by Henikoff and Henikoff (1996).

The steps of the methods are as follows:

(i) Determine the peptide sequences of the heavy and light chain variable domains of the subject antibody that binds to CD147. These can be determined by any of several methods, such as DNA sequencing of the respective genes after conventional cDNA cloning; DNA sequencing of cloning products that have been amplified by the polymerase chain reaction from reverse transcripts or

DNA of the subject hybridoma line; or peptide sequencing of a purified antibody protein. (ii) Apply the Kabat numbering system (Kabat et al, 1991) to the heavy and light chain sequences of the subject non-human CD147-bmding antibody, (iii) Determine canonical structure types for each of the CDRs of the subject non- human antibody that binds to CD 147. This determination is made from examination of the peptide sequence in light of the guidelines discussed in Chothia and Lesk (1987), Chothia et al., (1992), Tomlinson et al. (1995), MacCallum et al. (1996) and Al-Lazikani et al. (1997). The salient features of canonical structure determination for each of the CDRs are as follows.

For heavy chain CDRl, three canonical structure types are currently known.

Assignment of a new sequence is straightforward because each canonical structure type has a different number of residues. As described in Al-Lazikani et al. (1997), when

Kabat numbering is assigned to the sequence, the numbering for residues 31-35 will be as follows for the respective canonical structures.

For heavy chain CDR2, four canonical structure types are currently known. Several have unique numbers of residues, and are easily distinguished from their unique Kabat numbering of positions 52-56. Canonical structure types 2 and 3 for heavy chain CDR2 have equal numbers of residues, hence must be distinguished by clues within their sequence, as discussed by Chothia et al. (1992). Canonical structure type 2 has Pro or Ser at position 52a and GIy or Ser at position 55, with no restriction at the other positions. Canonical structure type 3 has GIy, Ser, Asn, or Asp at position 54, with no restriction at the other positions. These criteria are sufficient to resolve the correct assignment in most cases. Additionally framework residue 71 is commonly Ala, VaI, Leu, He, or Thr for canonical structure type 2 and commonly Arg for canonical structure type 3.

Heavy chain CDR3 is the most diverse of all the CDRs. It is generated by genetic processes, somewhat of a random nature, unique to lymphocytes. Consequently, canonical structures for CDR3 have been difficult to predict. In any case, human germline V gene segments do not encode any part of CDR3; because the V gene segments end at Kabat position 94, whereas positions 95 to 102 encode CDR3. For these reasons, canonical structures of CDR3 are not considered for choosing candidate human sequences.

For light chain CDRl, six canonical structure types are currently known for CDRl in kappa chains. Each canonical structure type has a different number of residues, hence assignment of a canonical structure type to a new sequence is apparent from the Kabat numbering of residue positions 27-31. For light chain CDR2, only a single canonical structure type is known for CDR2 in kappa chains, hence, barring exceptional subject antibody sequences, assignment is automatic.

For light chain CDR3, up to six canonical structure types have been described for CDR3 in kappa chains, but three of these are rare. The three common ones can be distinguished by their length, reflected in Kabat numbering of residue positions 91-97.

After identifying the canonical CDR structure types of the subject non-human antibody, human genes of the same chain type (heavy or light) that have the same combination of canonical structure types as the subject antibody are identified to form a candidate set of human sequences. In preferred embodiments, only the peptide sequences of human germline immunoglobulin V_H and V_k gene fragments are considered for comparison. Most of these gene fragments have been discovered and have already been assigned to a canonical structure type (Chothia et al. (1992); Tomlinson et al. (1995)). Additional V gene fragments not disclosed by these references are provided in US publication number 2003/0039649 (Foote). For the heavy chain, conformity of CDRl and CDR2 to the mouse canonical structure types is assessed, and genes that do not conform are excluded. For the light chain, conformity of CDRl and CDR2 of each human sequence to the canonical structure types of the subject antibody is first assessed. The potential of residues 89-95 of a candidate V_k gene to form a CDR3 of the same canonical structure type as the subject antibody is assessed, by positing a fusion of the gene with a J region and applying criteria for CDR3 canonical CDR structure type determination to the fused sequence, and non conforming sequences are excluded.

In another embodiment, appropriate when a variable domain of the subject antibody is of a canonical structure type not available in the human genome, human germline V genes that have three-dimensionally similar, but not identical, canonical structure types are considered for comparison. Such a circumstance often occurs with kappa chain CDRl in murine antibodies. All 6 possible canonical structure types have been observed at this CDR in murine antibodies, whereas the human genome encodes only canonical types 2, 3, 4 and 6. In these circumstances, a canonical CDR structure type having length of amino acid residues within two of the length of the amino acid residues of the subject non-human sequence may selected for the comparison. For example, where a type 1 canonical structure is found in the subject antibody, human Vk sequences with canonical structure type 2 should be used for comparison. Where a type 5 canonical structure is found in the murine antibody, human Vk sequences with either canonical structure type 3 or 4 should be used for comparison. In another embodiment, mature, rearranged human antibody sequences can be considered for the sequence comparison. Such consideration might be warranted under a variety of circumstances, including but not limited to instances where the mature human sequence (1) is very close to germline; (2) is known not to be immunogenic in humans; or (3) contains a canonical structure type identical to that of the subject antibody, but not found in the human germline.

In preferred embodiments, for each of the candidate V genes with matching canonical structure types, residue-to-residue sequence identity and/or homology with the subject sequence is also evaluated to rank the candidate human sequences. In a specific embodiment, the residues evaluated are as follows:

In preferred embodiments, residue-to-residue homology is first scored by the number of identical amino acid residues between the subject and the candidate human sequences. The human sequence used for subsequent construction of a converted antibody is chosen from among the 25 percent of candidates with the highest score. In other embodiments, appropriate when several candidate sequences have similar identity scores, similarity between non-identical amino acid residues may be additionally considered. Aliphatic-with-aliphatic, aromatic-with-aromatic, or polar-with-polar matches between subject and object residues are added to the scores. In another embodiment, quantitative evaluation of sequence homology may be performed using amino acid substitution matrices such as the BLOSUM62 matrix of Henikoff and Henikoff (1992).

A suitable sequence for the framework region C-terminal to the CDR3 sequence is selected from the set of known human germline J segments. A preferred J peptide sequence is selected by evaluating residue to residue homology for each J segment for sequence positions for which CDR3 and J overlap, using the scoring criteria specified for the evaluation of candidate V genes as mentioned above. The J gene segment peptide sequence used for subsequent construction of a converted antibody is chosen from among the 25 percent of candidates with the highest score. CDR3 of the heavy chain, which is part of the JH region thereof, does not have a limited number of three- dimensional structures that can be predicted from its sequence, however, any JH region may be used for constructing humanized heavy chain variable regions according to this method.

In all embodiments, the humanized molecule comprises at least one variable region, preferably a heavy chain variable region. Preferably the molecule also contains a second light chain variable region. Preferably, the humanized variable region contains at least one, at least two, or at least three CDR regions from the subject non- human antibody.

In one embodiment, where the antibody or antibody fragment comprises a heavy chain variable region and a light chain variable region, only the CDR3 of the heavy chain variable region is derived from or substantially the same as the subject non- human antibody CDR3, while the remaining 5 CDRs are derived from the human antibody sequences. In a more preferred embodiment, at least two of the 6 possible CDRs (one of which is heavy chain variable region CDR3) are derived from or substantially the same as those from the corresponding CDRs of the subject non-human antibody. In yet a more preferred embodiment, at least 3 of the 6 possible CDRs (one of which is heavy chain variable region CDR3) are derived from or substantially the same as those from the corresponding CDRs of the subject non-human antibody. Preferably, at least 4 or 5 of the 6 possible CDRs (one of which is heavy chain variable region CDR3) are derived from or substantially the same as those from the corresponding CDRs of the subject non-human antibody. In one embodiment, all 6 of the possible CDRs are derived from or substantially the same as those from the corresponding CDRs of the subject non-human antibody

In other embodiments, when only a heavy chain variable region is present, only the CDR3 of the heavy chain variable region is derived from or substantially the same as the subject non-human antibody CDR3, whilst the remaining 2 CDRs are derived from the human antibody sequence. In other embodiments, one or both of the other two CDRs (CDRl and/or CDR2) are also derived from the subject non-human antibody.

In all cases, however, it is preferred that the humanized antibody molecule contains no more than 10 amino acid residue in the framework sequence that differ from those in the framework sequence of the candidate human variable region.

Preferably, the antibody or antibody fragment framework sequences derived from a candidate human antibody have at least 65, more preferably at least 75, more preferably at least 80, yet more preferably at least 85, yet more preferably at least 90% and yet more preferably at least 95% sequence identity with the native candidate human framework sequence.

In a further preferred embodiment, a humanized antibody of the present invention has no more than a 100-fold reduction in affinity for CD 147 compared with a subject murine antibody raised against CD 147 (preferably Habl8), more preferably no more than 25-fold reduction in affinity, more preferably no more than 5-fold reduction in affinity, and more preferably about equal affinity for CD 147 as a subject murine antibody raised against CD147.

In a further preferred embodiment, a humanized antibody of the present invention has an affinity for CD147 of at least Kd InM, 1OnM, 10OnM, or IuM.

In another embodiment, appropriate when increased affinity of a humanized antibody is desired, residues within the CDRs of a converted antibody may be additionally substituted with other amino acids. Typically, no more than four amino acid residues in a CDR are changed, and most typically no more than two residues in the CDR will be changed, except for heavy chain CDR 2, where as many as 10 residues may be changed. Similarly, in certain embodiments, some of the amino acids in the framework sequences may be changed. Preferably, no more than 10 amino acid residues are changed The humanized antibody sequence is then physically assembled by methods of gene synthesis and recombinant protein expression known by those skilled in the art. The final form of the humanized sequences having the chimeric variable chains made by the methods disclosed herein may take many forms. Most typically, the chimeric antibodies will be made by construction a nucleic acid sequence encoding the chimeric variable chains, which are recombinantly expressed in a suitable cell type. Most typically, these variable regions will be linked to the constant regions of human immunoglobulin genes such that, when expressed, full-size immunoglobulins will be produced. In many cases, full-size IgG will be the preferred format. In other cases, IgG, IgM, IgA, IgD, or IgE may be preferred.

The invention also includes functional equivalents of the antibodies described in this specification as described previously.

Preferred non-human antibodies or antibody fragments which bind to CD 147 and which can be mutated so as to reduce immunogenicity to humans are murine antibodies or antibody fragments. A preferred murine antibody is the Habl8 antibody which is described by Xing et al. (2003), Liu et al. (2003), Li et al, (2002), Lou et al. (2002), Yang et al. (2001), Bian et al. (2000), Qui et α/.(1998), Sui (1992), Sui et al. (1996), Sui et al. (1998), Chen (1992) and Ji (1991).

The sequences of the variable regions for both the heavy and light chain of the Habl8 antibody are described in US publication number 2005/0176933 (Chen Zhinan et al.). There is an error in the amino acid sequence published in the aforementioned patent application. The correct amino acid sequences, as derived from the DNA sequences from the same application, are shown in Figure 1 and in SEQ ID NOs: 7 and 8. The variable region of murine heavy chain of the Habl8 antibody is shown in

SEQ ID NO: 7. The variable region of murine light chain of the Hablδ antibody is shown in SEQ ID NO: 47. The CDR sequences of the Habl8 antibody are shown in SEQ ID NOs:l-6: SEQ ID NO: 1 is the CDR 1 of murine heavy chain; SEQ ID NO: 2 is the CDR 2 of murine heavy chain; SEQ ID NO: 3 is the CDR 3 of murine heavy chain; SEQ ID NO: 4 is the CDR 1 of murine light chain; SEQ ID NO: 5 is the CDR 2 of murine light chain; SEQ ID NO: 6 is the CDR 3 of murine light chain.

Antibody or antibody fragments of the present invention may also be generated by changing the sequences of the heavy and light chain genes in the CDRl, CDR2, CDR3, or framework regions, using methods such as oligonucleotide-mediated site- directed mutagenesis, cassette mutagenesis, error-prone PCR, DNA shuffling, mutator- strains of E. coli, (Vaughan et al, 1998; Adey et al., 1996) or RNA-directed RNA polymerases such as the one from bacteriophage Q-beta, as described in PCT publication number WO 99/58661 (Diatech Pty Ltd). These methods of changing the sequence of the primary antibody have resulted in improved affinities of the secondary antibodies (Gram et al, 1992; Boder et al, 2000; Davies and Riechmann, 1996; Thompson et al, 1996; Short et al, 2002; Furukawa et al, 2001).

By a similar directed strategy of changing one or more amino acid residues of the antibody, the antibody sequences described in this invention can be used to develop anti-CD147 antibodies with improved functions.

In a preferred embodiment, the non-human antibody or antibody fragment of the present invention comprises a heavy chain in which one or more mutations have been made at Kabat residue numbers H3, H5, H18, H19, H23, H37, H40, H41, H42, H49, H73, H76, H77, H78, H82b, H83, H84, H88, H89, H93 and H105. Preferably, one or more mutations have been made at Kabat residue numbers H3, H5, H18, H23, H40, H42, H49, H77, H78, H82a, H84, H88 and H89, and more preferably at Kabat residue numbers H3, H5, Hl 8, H23, H42, H77, H78, H88 and H89.

In a preferred embodiment, the non-human antibody or antibody fragment comprises a heavy chain having a sequence as shown in SEQ ID NO: 7 in which one or more of the following residues have been mutated: 3, 5, 18, 19, 23, 37, 40, 41, 42, 49, 76, 79, 80, 81, 87, 89, 90, 94, 95, 99, or 109. Preferred mutations at these positions are as follows:

It will be appreciated that any combination of one or more or all of these mutations is encompassed by the present invention

In another preferred embodiment, the non-human antibody or antibody fragment of the present invention comprises a light chain in which one or more mutations have been made at Kabat residue numbers Ll, L3, L4, L7, L9, LlO, L12, L13, L15, L17,

L19, L21, L22, L42, L43, L49, L58, L60, L63, L67, L73, L77, L78, L79, L80, L83,

L85, L87, LlOO, Ll 04 and Ll 05. Preferably, one or more mutations have been made at

Kabat residue numbers Ll, L7, L9, LlO, L12, L13, L15, L43, L49, L60, L63, L67, L73, L77, L78, L80, L83 and L87, and more preferably at Kabat residue numbers L7, L9,

LlO, L12, L15, L49, L63, L67, L73, L77, L83 and L87.

In a preferred embodiment, the non-human antibody or antibody fragment comprises a light chain having a sequence as shown in SEQ ID NO: 47 in which one or more of the following residues have been mutated: 1, 3, 4, 7, 9, 10, 12, 13, 15, 17, 19, 21, 22, 42, 43, 49, 58, 60, 63, 67, 73, 77, 78, 79, 80, 83, 85, 87, 100, 104 or 105.

Preferred mutations at these positions are as follows:

It will be appreciated that any combination of one or more or all of these mutations is encompassed by the present invention.

It will be appreciated that any combination of the one or more mutations to a heavy chain may be combined with any combination of the one or more mutations to a light chain. etal

Preferably, the variant antibody or antibody fragment framework in which one or more amino acid residues have been mutated have at least 65, more preferably at least 70, more preferably at least 75, yet more preferably at least 80, yet more preferably at least 85% and yet more preferably at least 90% sequence identity with the native subject non-human framework sequence.

In a preferred embodiment, the antibody is a super-humanized™ Hablδ antibody. Habl8 antibodies may be super-humanized™ as follows.

First, canonical structures for the CDR' s of the heavy and light chains are determined by examining their sequences according to the method taught in US publication number US 2003039649 (Foote) and PCT publication number WO 04/006955 (Foote) and as further explained by by Tan (2002) and by Hwang et ah (2005b) and in the present application.

The heavy chain CDRl is of canonical structure 1 and the heavy chain CDR2 is of canonical structure 4-like. As there are three additional residues inserted after residue 52, heavy chain CDR2 most closely resembles canonical structure type 4. However, this canonical structure generally has a tyrosine at position 55, and a serine or a lysine residue at position 54. In the case of Habl8, the histidine in position 55 and asparagine in position 54 consequently classify this CDR as an irregular canonical structure most closely related to canonical structure 4 (hereafter referred to as "4-like"). The light chain CDRl is of canonical structure 2, the light chain CDR2 of canonical structure 1 and the light chain CDR3 of canonical structure 1

For the heavy chain, alignments are preferably made of human germline V region exons (Hwang et ah, 2005b) that have the canonical structures 1 and 4 at CDRl and CDR2, respectively, and the CDR' s of these may be aligned against the VH region of the Hablδ antibody, as shown in Figure 1. Any of these germline sequences (as set out in SEQ ID NOs: 8 to 12) could supply the framework regions for making a humanized heavy chain. There are a limited number of human geπnline variable regions, however, with this set of canonical structures. Therefore, the CDR' s of Hablδ were also aligned with those human germline V sequences that have canonical structures 1 and 3 at CDRl and CDR2, respectively, since canonical structure 3 is similar to canonical structure 4 (Figure 2). Any of these germline sequences (as set out in SEQ ID NOs:23 to 46) may also supply appropriate framework regions for making humanized heavy chains.

In a preferred embodiment, the human germline variable region sequence designated 3-72*01 (set out in SEQ ID NO: 11) is selected as an appropriate framework for making a humanized heavy chain. Thus in a preferred embodiment, the antibody or antibody fragment of the present invention comprises a humanized Hablδ antibody heavy chain comprising an amino acid sequence according to SEQ ID NO: 20.

The Habl8 light chain variable regions are preferably aligned against human germline Vkappa region sequences with the same canonical structures at the 3 CDRs, as shown in Figure 3. While any of these sequences (as set out in SEQ ID NOs: 48 to 59) could serve as framework sequences to humanize the Hablδ light chain, preferably the human germline variable region sequence designated 1-6*01 (SEQ ID NO: 48) or 3-11*01 (SEQ ID NO: 57) is selected as an appropriate framework for making a humanized light chain. Thus, in a preferred embodiment, the antibody or antibody fragment of the present invention comprises a humanised Habl8 antibody light chain comprising an amino acid sequence according to SEQ ID NO: 66 or 67.

The scope of the present invention is not limited to antibodies and fragments comprising manipulated Habl8 antibodies or fragments thereof. All non-human antibodies and fragments that specifically bind to CD 147 and which have been manipulated so as to reduce their immunogenicity in humans fall within the scope of the present invention. Thus, antibodies and antibody fragments which differ from those described herein and which use different CDR and/or different framework regions are also included in the present invention. However, the knowledge of the amino acid and nucleic acid sequences for the

Hablδ antibody and its humanized variants, which are described herein, can be used to develop other antibodies which also bind to the CD 147 and thereby directly or indirectly cause cell death or reduced metastasis. Several studies have surveyed the effects of introducing one or more amino acid changes at various positions in the sequence of an antibody, based on the knowledge of the primary antibody sequence, on its properties such as binding and level of expression (Yang et al, 1995; Rader et al, 1998; Vaughan et al, 1998).

Thus, in a further embodiment, the antibody or antibody fragment employs a framework sequence derived from a human antibody variable heavy chain having a CDRl of canonical structure type 1 and a CDR2 of canonical structure type 4. Examples of suitable human germline VH region genes having such CDR canonical structure types are given in Figure 1.

Alternatively, the antibody or antibody fragment employs a framework sequence derived from a human antibody variable heavy chain having a CDRl of canonical structure type 1 and CDR2 of canonical structure type 3. Examples of suitable human germline VH region genes having such CDR canonical structure types are given in

Figure 2.

In another embodiment, the antibody or antibody fragment employs a framework sequence derived from a human antibody variable light chain having a CDRl of canonical structure type 2, a CDR2 of canonical structure type 1 and a CDR3 of canonical structure 1. Examples of suitable human germline Vkappa region exons having such canonical structures types are given in Figure 3.

The conjugates of the present invention comprise the antibody, fragments, and their analogs as disclosed herein, linked to a cytotoxic agent. Preferred cytotoxic agents are maytansinoids, taxanes, and analogs of CC-1065. The conjugates can be prepared by in vitro methods. In order to link the cytotoxic agent to the antibody, a linking group is used. Suitable linking groups are well known in the art and include disulfide groups, thioether groups, acid labile groups, photolabile groups, peptidase labile groups and esterase labile groups. Preferred linking groups are disulfide groups and thioether groups. For example, conjugates can be constructed using a disulfide exchange reaction or by forming a thioether bond between the antibody and the cytotoxic agent.

Maytansinoids and maytansinoid analogs are among the preferred cytotoxic agents. Examples of suitable maytansinoids include maytansinol and maytansinol analogs. Suitable maytansinoids are disclosed in U.S. Patent Nos. 4,424,219; 4,256,746; 4,294,757; 4,307,016; 4,313,946; 4,315,929; 4,331,598; 4,361,650;

4,362,663; 4,364,866; 4,450,254; 4,322,348; 4,371,533; 6,333,410; 5,475,092;

5,585,499; and 5,846,545.

Taxanes are also preferred cytotoxic agents. Taxanes suitable for use in the present invention are disclosed in U.S. Patent Nos. 6,372,738 and 6,340,701. CC-1065 and its analogs are also preferred cytotoxic drugs for use in the present invention. CC-1065 and its analogs are disclosed in U.S. Patent Nos. 6,372,738;

6,340,701; 5,846,545; 5,843,937 and 5,585,499.

An attractive candidate for the preparation of such cytotoxic conjugates is CC-

1065, which is a potent anti-tumor antibiotic isolated from the culture broth of Streptomyces zelensis. CC-1065 is about 1000-fold more potent in vitro than are commonly used anti-cancer drugs, such as doxorubicin, methotrexate and vincristine

(Bhuyan et al, 1982).

Cytotoxic drugs such as methotrexate, daunorubicin, doxorubicin, vincristine, vinblastine, melphalan, mitomycin C, chlorambucil, and calicheamicin are also suitable for the preparation of conjugates of the present invention, and the drug molecules can also be linked to the antibody molecules through an intermediary carrier molecule such as serum albumin.

For diagnostic applications, the antibodies of the present invention typically will be labeled with a detectable moiety. The detectable moiety can be any one which is capable of producing, either directly or indirectly, a detectable signal. For example, the detectable moiety may be a radioisotope, such as ³H, ¹⁴C, ³²P, ³⁵S, or ¹³¹I; a fluorescent or chemiluminescent compound, such as fluorescein isothiocyanate, rhodamine, or luciferin; or an enzyme, such as alkaline phosphatase, beta-galactosidase or horseradish peroxidase. Radioisotopes may also be used for therapeutic purposes by targeting a tumor for destruction by ionizing radiation. In particular ¹³¹I and ⁹⁰Y have been used extensively to arm antibodies and increase their destruction of targeted cancer cells. The murine Habl8 antibody, in a F(ab)'2 format, has been administered to animals and patients as a I-conjugated antibody (see BACKGROUND section).

Any method known in the art for conjugating the antibody to the detectable moiety may be employed, including those methods described by Hunter et ah, (1962), David et al, (1974), Pain and Surolia (1981) and Nygren (1982).

The antibodies of the present invention can be employed in any known assay method, such as competitive binding assays, direct and indirect sandwich assays, and immunoprecipitation assays (Zola, 1987). The antibodies of the invention also are useful for in vivo imaging, wherein an antibody labeled with a detectable moiety such as a radio-opaque agent or radioisotope is administered to a subject, preferably into the bloodstream, and the presence and location of the labeled antibody in the host is assayed. This imaging technique is useful in the staging and treatment of malignancies. The antibody may be labeled with any moiety that is detectable in a host, whether by nuclear magnetic resonance, radiology, or other detection means known in the art.

For therapeutic applications, the antibodies or conjugates of the invention are administered to a subject, in a pharmaceutically acceptable dosage form. They can be administered intravenously as a bolus or by continuous infusion over a period of time, by intramuscular, subcutaneous, intra-articular, intrasynovial, intrathecal, oral, topical, or inhalation routes. The antibody may also be administered by intratumoral, peritumoral, intralesional, or perilesional routes, to exert local as well as systemic therapeutic effects. For the treatment of hepatocellular carcinoma, the antibody may be delivered via the hepatic artery. Suitable pharmaceutically acceptable carriers, diluents, and excipients are well known and can be determined by those of skill in the art as the clinical situation warrants. Examples of suitable carriers, diluents and/or excipients include: (1) Dulbecco's phosphate buffered saline, pH about 7.4, containing about 1 mg/ml to 25 mg/ml human serum albumin, (2) 0.9% saline (0.9% w/v NaCl), and (3) 5% (w/v) dextrose. The addition of polyvinyl pyrrolidone may be used in the case of radiolabeled antibodies so as to protect them against radiolysis. The method of the present invention can be practiced in vitro, in vivo, or ex vivo.

In other therapeutic treatments, the antibodies, antibody fragments or conjugates of the invention are co-administered with one or more additional therapeutic agents. Suitable therapeutic agents include, but are not limited to, cytotoxic or cytostatic agents. Cancer therapeutic agents are those agents that seek to kill or limit the growth of cancer cells while doing minimal damage to the host. Thus, such agents may exploit any difference in cancer cell properties (e.g. metabolism, vascularization or cell-surface antigen presentation) from healthy host cells. Differences in tumor morphology are potential sites for intervention: for example, the second therapeutic can be an antibody such as an anti-VEGF antibody that is useful in retarding the vascularization of the interior of a solid tumor, thereby slowing its growth rate. Other therapeutic agents include, but are not limited to, adjuncts such as granisetron HCL, androgen inhibitors such as leuprolide acetate, antibiotics such as doxorubicin, antiestrogens such as tamoxifen, antimetabolites such as interferon alpha-2a, cytotoxic agents such as taxol, enzyme inhibitors such as ras farnesyl-transferase inhibitor, immunomodulators such as aldesleukin, and nitrogen mustard derivatives such as melphalan HCl, and the like.

When present in an aqueous dosage form, rather than being lyophilized, the antibody typically will be formulated at a concentration of about 0.1 mg/ml to 100 mg/ml, although wide variation outside of these ranges is permitted. For the treatment of disease, the appropriate dosage of antibody or conjugate will depend on the type of disease to be treated, as defined above, the severity and course of the disease, whether the antibodies are administered for preventive or therapeutic purposes, the course of previous therapy, the patient's clinical history and response to the antibody, and the discretion of the attending physician. The antibody is suitably administered to the patient at one time or over a series of treatments.

Depending on the type and severity of the disease, about 0.015 to 15 mg of antibody/kg of patient weight is an initial candidate dosage for administration to the patient, whether, for example, by one or more separate administrations, or by continuous infusion. For repeated administrations over several days or longer, depending on the condition, the treatment is repeated until a desired suppression of disease symptoms occurs. However, other dosage regimens may be useful and are not excluded.

The invention will now be described by reference to the following examples, which are illustrative only and are not intended to limit the present invention.

EXAMPLES

EXAMPLE 1 : Humanization of murine antibody Habl 8

The murine Habl 8 antibody was super-humanized jTM by determining the canonical structures for the CDR' s of the heavy and light chains in accordance with the method taught in US publication number US 2003/039649 (Foote) and PCT publication number WO 04/006955 (Foote) and as further explained by by Tan et al. (2002) and by Hwang et al. (2005b) and in the present application. First, canonical structures for the CDR' s of the heavy and light chains were determined to be the following by examining their sequences according to the method taught in the US patent application serial number 10/194975 and US patent number 6,881,557 and as further explained by Tan et al (2002. J. Immunol. 169(2): 1119-25), by Hwang et al (2005b) and in this application (see above). For sequences shown below, the Kabat numbering for each residue is shown above the one letter amino acid code for each residue.

Heavy CDRl: canonical structure 1 (SEQ ID NO: 1):

Since there are no insertions after residue 35, CDRl has canonical structure type 1.

Heavy CDR2: canonical structure 4-like (SEQ ID NO: 2):

As there are three additional residues inserted after residue 52, heavy chain CDR2 most closely resembles canonical structure type 4. However, this canonical structure W 2

41

generally has a tyrosine at position 55, and a serine or a lysine residue at position 54. In the case of Hablδ, the histidine in position 55 and asparagine in position 54 consequently classify this CDR as an irregular canonical structure most closely related to canonical structure 4 (hereafter referred to as "4-like).

Light CDRl: canonical structure 2 (SEQ ID NO: 4):

Because there are no insertions or deletions between residues 27 and 31, CDRl has 0 canonical structure type 2.

Light CDR2: canonical structure 1 (SEQ ID NO: 5):

5 This is not an exceptional sequence; its canonical structure is of type 1.

Light CDR3: canonical structure 1 (SEQ ID NO: 6)

0 Because of the length, the glutamine at position 90 and the Proline at position 95, this sequence is consistent with canonical structure type 1.

The sequence of the V region of the heavy chain of Habl8 is shown in SEQ ID NO: 7. Human germline VH region sequences (Hwang et al, 2005b) that have the 5 canonical structures 1 at CDRl and structures 4 or 4-like (same number of residues in CDR2 but having sequences lacking the typical "signature" residues at positions 54 or 55 or both) at CDR2, respectively, were selected and are set out in SEQ ID NOs 8 to 12. These sequences were aligned against the VH region of the Habl8 antibody (SEQ ID NO: 7), as shown in Figure 1. The Habl8 variable region heavy chain CDR3/JH region sequence is set out in SEQ ID NO: 13. Human germline JH region sequences (i.e. JHl - JH6) are set out in SEQ ID NOs: 14 to 19 respectively. Any of the germline sequences shown in SEQ ID NOs 8 to 12, in combination with any of the six possible JH sequences (SEQ ID NOs: 14 to 19) could supply the framework regions for making a humanized heavy chain.

Super-humanized™ antibody variable regions were designed by combining sequences as follows: the heavy chain variable domain consisted of the heavy chain Kabat CDR sequences from Habl8 (SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3) and framework regions identical to either the germline sequence 3-73*01 (SEQ ID NO: 12), 3-49*01 (SEQ ID NO: 10) or 3-72*01 (SEQ ID NO: 11). An example of a humanized heavy chain variable region based on the murine Habl8 CDRs and the 3- 72*01 (SEQ ID NO: 11) and JH2 (SEQ ID NO: 15) germline sequences consists of the sequence shown in SEQ ID NO: 20. Another example of a humanized heavy chain variable region based on the murine Habl8 CDRs and the 3-49*01 (SEQ ID NO: 10) and JH2 (SEQ ID NO: 15) germline sequences consists of the sequence shown in SEQ ID NO: 21. Yet another example of a humanized heavy chain variable region based on the murine Habl8 CDRs and the 3-73*01 (SEQ ID NO: 12) and JHl (SEQ ID NO: 14) germline sequences consists of the sequence shown in SEQ ID NO: 22. Because there are a limited number of human germline variable regions with this set of canonical structures, we also aligned the CDR' s of Habl8 with those human germline V exons that have canonical structures 1 and 3 at CDRl and CDR2, respectively, since canonical structure 3 is similar to canonical structure 4 (see Figure 2). Any of these germline sequences (as shown in SEQ ID NOs: 23 to 46), in combination with any of the six possible JH sequences (as shown in SEQ ID NOs 14 to 19), may supply appropriate framework regions for making humanized heavy chains.

Super-humanized™ antibody light chain variable region sequences were designed by combining sequences as follows: the Kabat-defϊned CDR' s (SEQ ID NO: 4, SEQ ID NO: 5 and SEQ ID NO: 6) were derived from the murine Hablδ antibody variable region light chains; the frameworks were derived from any combination of the human germline Vkappa sequences with the same canonical structures as the murine Hablδ light chain variable region, i.e. canonical structures 2, 1 and 1 at CDRl, CDR2 and CDR3, respectively, with any of the five possible Jkappa segments.

The V region sequence of the Habl8 light chain is shown in Figure 3 and SEQ ID NO: 47. Selected human germline Vkappa sequences with the same canonical structures as the murine Hablδ light chain variable region are also shown in Figure 3 and are set out in SEQ ID NOs: 48 to 59.

In order to choose the most preferred human light chains sequences to work with, we aligned the selected human germline Vkappa sequences with the CDR' s of the murine Habl 8 kappa light chain variable region as shown in Figure 3. These were then ranked by considering either the total number of identical (i.e. same as in corresponding position in the Habl 8 CDRs) amino acids in the three CDRs or the total number of identical amino acids in the regions of the three CDRs that are likely to be in contact with antigen (see Figure 3 for exact list of amino acid positions). Using these criteria, two preferred examples were chosen, namely 1-6*01 (SEQ ID NO: 48) and 3-11*01 (SEQ ID NO: 57). The sequence 1-6*01 (SEQ ID NO: 48) has the highest sequence identity with the murine Habl 8 in the regions of the three CDRs that are likely to be involved in contacting the antigen.

The sequence of the Habl 8 variable light chain CDR3/J region is set out in SEQ ID NO: 60. The sequences of human germline JKappa regions (JKl -JK5) are set out in SEQ ID NOS 61 to 65 respectively. JK3 (SEQ ID NO: 63) was chosen as a preferred embodiment due to its identity with SEQ ID NO: 60 at Kabat residues 96 and 97. By combining sequence 1-6*01 (SEQ ID NO: 48) with the JK3 sequence (SEQ ID NO: 63) and the murine Habl 8 CDRs (as shown in SEQ ID NOs: 4 to 6), the humanized Vkappa (light) chain represented by SEQ ID NO: 66 was generated.

As the precise binding interface for the Habl 8 antibody to its target antigen is not known, we also considered the sequence 3-11*01 (SEQ ID NO: 57), which has the highest sequence identity with the murine Habl 8 in the three CDRs (all CDR residues considered). By combining this with the JK3 sequence (SEQ ID NO: 63) and the murine Hab 18 CDRs (as shown in SEQ ID NOs: 4 to 6), the humanized Vkappa (light) chain represented by SEQ ID NO: 67 was generated.

EXAMPLE 2: Binding affinities of humanized Hab 18 scFv fragments

To assess the success of the super-humanization™ procedure in transferring binding affinity, the following super-humanized™ single chain variable region fragments (scFvs) were constructed by gene synthesis (Stemmer et al, 1995. Gene 16:49-53): scFV - 1 (SEQ ID NO: 68): incorporating the humanized heavy chain sequence as shown in SEQ ID NO: 20 and the humanized light chain sequence as shown in SEQ ID NO: 66.

scFV - 2 (SEQ ID NO: 69): incorporating the humanized heavy chain sequence as shown in SEQ ID NO: 20 and the humanized light chain sequence as shown in SEQ ID NO: 67.

A cassette consisting of a super-humanized™ heavy chain and a super- humanized™ light chain, connected through a 15 residue linking region featuring a (G₄S)₃ motif (Huston et α/.,1988) , was assembled using Ncol, BamHI and Notl restriction sites at beginning of the heavy chain, within the linker region and at the end of the light chain, respectively (see Figure 4). Complete scFv constructs were inserted into a bacterial expression vector based on pUC18, featuring a signal peptide at the N- terminus (which is cleaved off in the periplasm) and a hexahistidine tag at the C- terminus, allowing for purification on a Ni²⁺-chelating resin. A FLAG tag (Slootstra et al., 1997), consisting of the sequence DYKDDDDK, was also included at the C- terminal end (see Figure 5), to aid in the detection or purification using commercial antibodies against the FLAG epitope. The expression constructs were transformed into E coli strain HB2151 and the DNA sequence was confirmed. Protein expression in 2xYT media supplemented with 0.05% glucose was performed at 30°C, induced in late log phase by addition of 0.5 mM IPTG and then grown overnight. After separation of the cells from the media by centrifugation, the periplasmic fraction was released by osmotic shock by resuspending and incubating the cells on ice in 20% sucrose solution buffered in 30 mM Tris (pH 8) supplemented with ImM EDTA, then separating the osmotic fraction by centrifugation. Western blotting confirmed the presence of monomelic scFv protein. Roughly equivalent amounts of expressed scFv in the periplasmic fraction was observed for each construct. Protein was purified further using Ni-NTA superflow beads (Qiagen, Doncaster) according to manufacturers instructions, which resulted in >90% pure protein as assessed by SDS PAGE.

Binding to the target of Habl8, the cell surface molecule CD 147 was assessed by ELISA: 100ng/well recombinant human CD 147 (provided by AmProx, Carlsbad, CA, expressed in E coli) was immobilized on plates in 0.1M sodium carbonate (pH 9.5) overnight at 4°C. Wells were then washed and subsequently blocked with 4% skim milk by incubating for 2 hours at room temperature, washed with phosphate buffered saline (pH 7.4, PBS) supplemented with 0.05% Tween-20 (Sigma, St Louis) and then incubated with serial dilutions of crude protein periplasmic fraction of super- humanized™ scFv in PBS for 2 hours while rocking at room temperature, after which the wells were washed again with PBS plus 0.05% Tween-20. Binding was detected with an anti-FLAG-horseradish peroxidase conjugate (Sigma, St. Louis MO), with which the wells were incubated for one hour at room temperature while rocking according to manufacturers specifications. Binding was visualized using SureBlue peroxidase substrate (KPL, Gaithersburg). The enzymatic reaction was stopped after 15 minutes by adding 0.1M HCl and peroxidase product was measured by spectrophotometric absorption at 450nm. Averaged data of duplicate measurements are compared to a positive control, the parental mouse antibody that was expressed recombinantly as a Fab fragment (see Figure 5). The mouse Fab fragment was expressed and periplasmic fraction extracted from E. coli using similar conditions as for the scFv constructs. The construct scFv-1 (SEQ ID NO: 68) appears to retain approximately 30%-50% of the binding affinity seen for the mouse Fab fragment: scFv fragments generally have lower affinities that the corresponding Fab fragments, so the somewhat reduced degree of binding of this humanized scFv (SEQ ID NO: 68) relative to the murine Fab may be due to this fact, and/or to slightly reduced affinity of the humanized sequences relative to the murine sequence. The humanized light chain appears to be important also, as illustrated by the difference in binding of scFv-1 (SEQ ID NO: 68) versus scFv-2 (SEQ ID NO: 69) where the same heavy chain is present.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

REFERENCES

Adey, N. B. et al, (1996). In "Phage Display of Peptides and Proteins", Chapter 16, pp. 277-291, Eds. Kay, B. K. et al., Academic Press.

Al-Lazikani B., Lesk, A.M., Chothia, C. (1997). Standard conformations for the canonical structures of immunoglobulins. J. MoI Biol, 273 (4): 927-48.

Alsowmely, A.M. & Hodgson, H.J.F. (2002). Non-surgical treatment of hepatocellular carcinoma. Aliment Pharmacol. Ther., 16:1-15.

Bergers, G., Brekken, R., McMahon, G., Vu, TH., Itoh, T., Tamaki, K., Tanzawa, K., Thorpe, P., Itohara, S., Werb, Z. & Hanahan, D. (2000). Matrix metalloproteinase-9 triggers the angiogenic switch during carcinogenesis. Nat. Cell. Biol., 2:737-744.

Bernhard, M.I., Foon, K.A., Oeltmann, T.N., Key, M.E., Hwang, K.M., Clarke, G.C., Christensen, W.L., Hoyer, L.C., Hanna, M.G. & Oldham, R.K. (1983). Guinea pig line 10 hepatocarcinoma model: characterization of monoclonal antibody and in vivo effect of unconjugated antibody and antibody conjugated to diphtheria toxin A chain. Cancer Res., 43:4420-4428.

Bhuyan, B.K., Newell, K.A., Crampton, S.L. & Von Hoff, D.D. (1982). CC-1065 (NSC-218223), a most potent antitumor agent: Kinetics of inhibition of growth, DNA synthesis and cell survival. Cancer Res., 42:3532-3537.

Bian, HJ., Chen, Z-N. & Deng, J.L. (2000). Direct technetium-99m labeling of anti- hepatoma monoclonal antibody fragment: a radioimmunoconjugate for hepatocellular carcinoma imaging. World J Gastroenterol., 6(3):348-352.

Biswas, C, Zhang, Y., DeCastro, R., Guo, H., Nakamura, T., Kataoka, H. & Nabeshima, K. (1995). The human tumor cell-derived collagenase stimulatory factor (renamed EMMPRIN) is a member of the immunoglobulin superfamily. Cancer Res., 55:434.439 Boder, E.T., Midelfort, K.S. & Wittrap, K.D. (2000). Directed evolution of antibody fragments with monovalent femtomolar antigen-binding affinity. Proc. Natl. Acad. Sd. U. S. A., 91:10101-10105.

Chen Z-N. (1992). Significance and application of anti-malignant hepatoma MAb HAb 18 in radioimmunal diagnosis of human hepatocellular carcinoma. Zhonghua Zhong Liu Za Zhi., 14(1):9-12.

Chen, Z-N., Yang, Z., Mi, L., Jiang, J.L., and Guo, X.N. (1999). J. Cell. MoI. Immunol, 15:34

Chothia, C. & Lesk, A.M. (1987). Canonical structure types for the hypervariable regions of immunoglobulins. J. MoL Biol., 96:901-917;

Chothia, C, Lesk, A.M., Gherardi, E., Tomlinson, I.M., Walter, G., Marks, J.D., Llewelyn, M.B. & Winter, G. (1992). Structural repertoire of the human VH segments. J. MoI. Biol, 227:799-817.

David, G.S. & Reisfeld, R.A. (1974). Protein iodination with solid state lactoperoxidase. Biochemistry, 13:1014.

Davies, J. & Riechmann, L. (1996). Affinity improvement of single antibody VH domains: residues in all three hypervariable regions affect antigen binding. Immunotechnology, 2(3):169-179.

Deeg, H.J., Blazar, B.R., Bolwell, BJ., Long, G.D., Schuening, F., Cunningham, J., Rifkin, R.M., Abhyankar, S., Briggs, A.D., Burt, R., Lipani, J., Roskos, L.K., White, J.M., Havrilla, N., Schwab, G. & Heslop, H.E. (2001). Treatment of steroid-refractory acute graft- versus-host disease with anti-CD 147 monoclonal antibody ABX-CBL. Blood, 98:2052-2058.

Ellis, S.M., Nabeshima, K. & Biswas, C. (1989). Monoclonal antibody preparation and purification of a tumor cell collagenase-stimulatory factor. Cancer Res., 49:3385-3391. Fang, J., Shing, Y., Wiederschain, D., Yan, L., Butterfield, C, Jackson, G., Harper, J., Tamvakopoulos, G. and Moses. M.A. (2000). Matrix metalloproteinase-2 is required for the switch to the angiogenic phenotype in a tumor model. PNAS 97(8):3884-3889.

Fuhrer, J.P., Xie, H., Murphy, M.I. Jr., Ye, J.N. & Yao, Z. (1991). Characterization of a membrane-associated glycoprotein (gp43) on human hepatocellular carcinomas by a monoclonal antibody. Cancer Res., 51:2158-2163.

Furukawa, K., Shirai, H., Azuma, T., Nakamura, H. (2001). A Role of the Third Complementarity-determining Region in the Affinity Maturation of an Antibody. J. Biol. Chem., 276:27622-27628.

Gilles, C, Polette, M., Piette, J., Birembaut, P. & Foidart, J.M. (1994). Epithelial-to- mesenchymal transition in HPV-33-transfected cervical keratinocytes is associated with increased invasiveness and expression of gelatinase A. Int. J. Cancer, 59:661—666.

Gilles, C, Polette, M., Piette, J., Munaut, C, Thompson, E. W. & Birembaut, P. (1996). High level of MT-MMP expression is associated with invasiveness of cervical cancer cells. Int. J. Cancer, 65:209-213.

Gram, H., Marconi, L.A., Barbas, C.F.I., Collet, T.A., Lerner, R.A., & Kang, A.S. (1992). In vitro selection and affinity maturation of antibodies from a naive combinatorial immunoglobulin library. Proc. Natl. Acad. Sd. USA, 89:3576-3580.

Guo, H., Zucker, S., Gordon, M.K., Toole, B.P. & Biswas, C. (1997). Stimulation of Matrix Metalloproteinase Production by Recombinant Extracellular Matrix Metalloproteinase Inducer from Transfected Chinese Hamster Ovary Cells. J. Biol. Chem., 272:24-27

Henikoff, S. & Henikoff, J.G. (1992). Amino acid substitution matrices from protein blocks. Proc. Natl. Acad. Sd. USA, 89:10915-10919.

Henikoff, S. & Henikoff, J.G. (1996). Using substitution probabilities to improve position-specific scoring matrices. CABIOS, 12:135-143. J.G. Henikoff, J.G. & Henikoff, S. (1996). Blocks Database and its Applications. Meth. Enzymolog, 26:88-105.

Hunter & Greenwood. (1962). Nature, 144:945.

Huston, J.S., Levinson, D., Mudgett-Hunter, M., Tai, M., Novotny, J., Margolies, M.N., Ridge, RJ, Bruccoleri, R.E., Haber, E., Crea, R.and Oppermann, H. (1988). Protein Engineering of Antibody Binding Sites: Recovery of Specific Activity in an Anti-Digoxin Single-Chain Fv Analogue Produced in Escherichia coli. PNAS, 85: 5879-5883.

Hwang, K.M., Foon, K.A., Cheung, P.H., Pearson, J.W. & Oldham, P.K. (1984). Selective antitumor effect on LlO hepatocarcinoma cells of a potent immunoconjugate composed of the A chain of abrin and a monoclonal antibody to a hepatoma-assaciated antigen. Cancer Res., 44:4578-4586.

Hwang, WYK, & Foote, J. (2005a) Immunogenicity of engineered antibodies. Methods 36:3-10.

Hwang, W. Y., Almagro, J.C., Buss, T.N., Tan, P. & Foote, J. (2005b). Use of human germline genes in a CDR homology-based approach to antibody humanization. Methods, 36:35-42.

Javadpour, N. & Guirguis, R. (1992). Tumor collagenase-stimulating factor and tumor autocrine motility factor as tumor markers in bladder cancer - an update. Eur. Urol., 21 Suppl. 1:1-4

Ji, Y. (1991). Internalization of monoclonal antibodies against human hepatoma in the target cells. Zhonghua YiXue Za ZhL, 71(7):369-72, 26. Jones, P.T., Dear, P.H., Foote, J., Neuberger, M.S. & Winter, G. (1986). Replacing the complementarity determining regions in a human antibody with those from a mouse. Nature, 321:522-525.

Kabat, E.A., Wu, T.T., Reid-Miller, M., Perry, H.M., Gottesmann, K.S. & Foeller, C. (1991). Sequences of Proteins of Immunological Interest. Fifth Edition, Bethesda: U.S. Dept. of Health and Human Services, PHS, NIH. Kasinrerk, W., Fiebiger, E., Stefanova, L, Baumruker, T., Knapp, W. & Stockinger., H. (1992). Human leukocyte activation antigen M6, a member of the Ig superfamily, is the species homologue of rat OX-47, mouse basigin, and chicken HT7 molecule. J. Immunol, 149:847-854.

Kataoka, H., DeCastro, R., Zucker, S. & Biswas, C. (1993). Tumor cell-derived collagenase-stimulatory factor increases expression of interstitial collagenase, stromelysin, and 72-kDa gelatinase. Cancer Res., 53 :3154-3158.

Klein, C.A., Seidl, S., Petat-Dutter, K, Offner, S., Geigl, J.B., Schmidt-Kittler, O., Wendler, N., Passlick, B., Huber, R.M., Schlimok, G., Baeuerle, P.A. & Riethmϋller, G. (2002). Combined transcriptome and genome analysis of single micrometastatic cells. Nature Biotechnology, 20(4):387-392.

Kuniyasu, H., Troncoso, P., Johnston, D., Bucana, CD., Tahara, E., Fidler, IJ. & Pettaway, CA. (2000). Relative Expression of Type IV Collagenase, E-cadherin, and Vascular Endothelial Growth Factor/Vascular Permeability Factor in Prostatectomy Specimens Distinguishes Organ-confined from Pathologically Advanced Prostate Cancers. CHn. Cancer Res., 6:2295-2308.

Lau W. Y. (2002). Clinical Review Management of hepatocellular carcinoma. J. R. Coll Surg. Edinb,. 47:389-399.

Li, Y., Cai, M., Wang, Z., Liu, X. & Liang, Z. (2002). Preparation and in vitro killing effect of adriamycin-loaded immunonanosphere against hepatoma led by F (ab')2 Fragment of monoclonal antibodies. Hua Xi Yi Ke Da Xue Xue Bao., 33(1):8-11.

Lim, M., Martinez, T., Jablons, D., Cameron, R., Guo, H., Toole, B., Li, J.D. & Basbaum, C (1998). Tumor-Derived EMMPRIN (Extracellular Matrix Metalloproteinase Inducer) Stimulates Collagenase Transcription Through MAPK p38. FEBS Lett., 441:88-92.

Liu X., Shang, B. & Liu, X. (2001). Inhibition of the growth of hepatoma and hepatic metastasis by pingyangmycin conjugated with Fab' fragment of monoclonal antibody. Zhonghua YiXue Za Zhi, 81:201-201. Liu, X., Wang, X., Wei, D., Cai, M. & Li, G. (2003). Internalization of antibody- targeted immunonanoparticles into human hepatoma cells and its reversal effect on MDR. Sichuan Da Xm Xw Bao YiXue Ban., 34(3):431-4.

Lokeshwar, B.L., Selzer, M.G., Block NX. & Gunja-Smith, Z. (1993). Secretion of matrix metalloproteinases and their inhibitors (tissue inhibitor of metalloproteinases) by human prostate in explant cultures: reduced tissue inhibitor of metalloproteinase secretion by malignant tissues. Cancer Res., 53:4493-4498.

Lou, C., Chen, Z-N., Bian, HJ., Li, J. & Zhou, S.B. (2002) Pharmacokinetics of radio- immunotherapeutic agent of direct labeling mAb 188Re-HAbI 8. World J Gastroenterol, 8(l):69-73.

MacCallum, R.M., Martin, A.C.R. & Thornton, J.M. (1996). Antibody-antigen interactions: contact analysis and binding site topography. J. MoI. Biol., 262, 732-745.

Muraoka, K., Nabeshima, K., Murayama, T., Biswas, C. & Koono, M. (1993). Enhanced expression of a tumor-cell-derived collagenase-stimulatory factor in urothelial carcinoma: its usefulness as a tumor marker for bladder cancers. Int. J. Cancer, 55:19-26.

Nelson, A.R., Fingleton, B., Rothenberg, MX. & Matrisian, L.M. (2000). Matrix metalloproteinases: biologic activity and clinical implications. J. Clin. Oncol., 18:1135-1139.

Nygren, H. (1982).. Conjugation of horseradish peroxidase to Fab fragments with different homobifunctional and heterobifunctional cross-linking reagents. A comparative study. J. Histochem. Cytochem., 30:407-412.

Padlan, E.A. (1991). A possible procedure for reducing the immunogenicity of antibody domains while preserving their ligand-binding properties. MoI. Immunol., 28:489-498.

Padlan, E.A., Abergel, C. & Tipper, J.P. (1995). Identification of specificity- determining residues in antibodies. FASEB J., 9:133-139. Pain, D. & Surolia, A. (1981). Preparation of protein A-peroxidase monoconjugate using a heterobifunctional reagent, and its use in enzyme immunoassays. J. Immunol. Metk, 40:219-220.

Pearson, W.R. & Lipman, DJ. (1988). Improved tools for biological sequence comparison. Proc. Natl. Acad. Sd. USA, 85:2444-2448.

Polette, M., Gilles, C, Marchand, V., Lorenzato, M., Toole, B., Tournier, J-M., Zucker, S. & Birembaut, P. (1997). Tumor Collagenase Stimulatory Factor (TCSF) Expression and Localization in Human Lung and Breast Cancers. J. Histochem. Cytochem., 45:703-710.

Powell, W.C., Knox, J.D., Navre, M., Grogan, T.M., Kittelson, J., Nagle, R.B.& Bowden, G.T. (1993). Expression of the metalloproteinase matrilysin in DU-145 cells increases their invasive potential in severe combined immunodeficient mice. Cancer Res., 53:417-422.

Qiu, K., Wang, B.C., Chen, Z-N., Fang, P., Liu, CG., Wan, W.X. & Liu, Y.F. (1998). 99mTc-labeled HAb 18 McAb Fab fragment for radioimmunoimaging in nude mice bearing human hepatocellular carcinoma. World J Gastroenterol., 4(2):117-120.

Rader, C, Cheresh, D. A. & Barbas, C.F., III. (1998). A phage display approach for rapid antibody humanization: designed combinatorial V gene libraries. Proc. Natl. Acad. ScL U.S.A., 95: 8910-8915.

Riechmann, L., Clark, M., Waldmann, H. & Winter, G. (1988). Reshaping human antibodies for therapy. Nature, 332:323-327.

Rutgeerts, P., D'Haens, G., Targan, S. (1999). Efficacy and safety of retreatment with anti-tumor necrosis factor antibody (INFLIXIMAB) to maintain remission in Crohn's disease. Gastroenterology, 117, 761-769.

Sameshima, T., Nabeshima, K., Toole, B.P., Yokogami, K., Okada, Y., Goya, T., Koono, M. & Wakisaka, S. (2000). Glioma cell extracellular matrix metalloproteinase inducer (EMMPRIN) (CD 147) stimulates production of membrane-type matrix metalloproteinases and activated gelatinase A in co-cultures with brain-derived fibroblasts. Cancer Lett, 157:177-184.

Sameshima, T., Nabeshima, K., Toole, B. P., Yokogami, K., Okada, Y., Goya, T., Koono, M., & Wakisaka, S. (2000a). Expression of emmprin (CD147), a cell surface inducer of matrix metalloproteinases, in normal human brain and gliomas. Int. J. Cancer, 88:21-27.

Sehgal, G., Hua, J., Bernhard, EJ., Sehgal. L, Thompson, T.C. & Muschel, R.M. (1998). Requirement for matrix metalloproteinase-9 (gelatinase B) expression in metastasis by murine prostate carcinoma. Am. J. Pathol., 152:591-596.

Short, M.K., Krykbaev, R.A., Jeffrey, P.D. & Margolies, M.N. (2002). Complementary Combining Site Contact Residue Mutations of the Anti-digoxin Fab 26-10 Permit High Affinity Wild-type Binding. J. Biol. Chem., 277(19):16365-16370.

Slootstra, J.W., Kuperus. D., Pluckthun, A. and Meloen, R.H. (1997). Identification of new tag sequences with differential and selective recognition properties for the anti- FLAG monoclonal antibodies Ml, M2 and M5. MoI Divers. 2(3): 156-64.

Stein, R., Wang, Z.F., Sharkey, R.M., Klein, K.M. & Golderberg, D.M. (1991). A new murine monoclonal antibody against human hepatoma. Hybridoma, 10:255-267.

Stemmer, W.P.C., Crameri, A., Ha, K.D., Brennan, T.M. & Heyneker, H.L. (1995). Single-step assembly of a gene and entire plasmid from large numbers of oligodeoxyribonucleotides. Gene, 164(1): 49-53.

Sui, Y. (1992). Targeting therapy with adriamycin-monoclonal antibody (HAbI 8) conjugate in nude mice with human hepatocellular carcinoma xenografts. Zhonghua Yi Xwe Za Zfø., 72(7):397-400, 445.

Sui, Y., Sun, Z. & Chen Z-N. (1996). The internalization of immunotargeting drugs for hepatocellar carcinoma and its significance. Zhonghua YiXue Za Zhi., 76(11):845-7. Sui, Y., He, F. & Chen Z-N. (1998). Radioimmunoimaging of hepatocellular carcinoma with 1311 labeled antihepatoma monoclonal antibody Fab fragment. Zhonghua YiXue Za ZhL, 78(7):537-9.

Tan, P., Mitchell, D.A., Buss, T.N., Holmes, M.A., Anasetti, C. & Foote, J. (2002). "Superhumanized" Antibodies: Reduction of Immunogenic Potential by Complementarity-Determining Region Grafting with Human Germline Sequences: Application to an Anti-CD28. J. Immunol., 169:1119-1125.

Thompson, J., Pope, T., Tung, J.S., et al. (1996). Affinity maturation of a high-affinity human monoclonal antibody against the third hypervariable loop of human immunodeficiency virus: use of phage display to improve affinity and broaden strain reactivity. J. MoI. Biol, 256:77-88.

Tomlinson, I.M., Cox, J.P.L., Gherardi, E., Lesk, A.M. & Chothia, C. (1995). The structural repertoire of the human Vk domain. EMBO J., 14: 4628-4638.

Vaughan, T. J., Osbourn, J.K. & Tempest, P.R. (1998). Human antibodies by design. Nature Biotechnology, 16:535-539.

Venook, A.P. (2000). Hepatocellular carcinoma. Curr. Treat. Options Oncol., 1:407- 15.

Wang, Q-c, Ying, W-b., Xie, H., Zhang, Z-c, Yang, Z-h. & Ling, L-q. (1991). Trichosanthin-monoclonal antibody conjugate specifically cytotoxic to human hepatoma cells in vitro. Cancer Res., 51 :3353-3355

Wood, M., Fudge, K., Mohler, J.L., Frost, A.R., Garcia, F., Wang, M. & Stearns, M.E. (1997). In situ hybridization studies of metalloproteinases 2 and 9 and TIMP-I and TIMP-2 expression in human prostate cancer. Clin. Exp. Metastasis, 15:246-258.

Wu, T.T. & Kabat, E. A. (1970). An analysis of the sequences of the variable regions of Bence Jones proteins and myeloma light chains and their implications for antibody complementarity. J. Exp. Med., 132:211-250. Xing JL., Wang, YQ., Yang, XM., Song, F. & Chen, Z-N. (2003). Cloning and expression of Fab gene of mAb HAb 18 against human hepatocellular carcinoma. Xi Bao Yu Fen Zi Mian YiXue Za ZhL, 19(6):570-3.

Yang, W.P., Green, K., Pinz-Sweeney, S., Briones, A.T., Burton, D.R. & Barbas, C.F.I. (1995). CDR walking mutagenesis for the affinity maturation of a potent human anti- HIV-I antibody into the picomolar range. J. MoI. Biol, 254:392-403.

Yang, LJ., Sui, Y.F. & Chen, Z-N. (2001). Preparation and activity of conjugate of monoclonal antibody Habl8 against hepatoma F(ab')(2) fragment and staphylococcal enterotoxin A. World J. Gastroenterol, 7(2) :216-221.

Zeng, Z-C, Tang, Z-Y., Liu, K-D., Lu, J-Z., Xie, H. & Yao, Z. (1998). Improved long- term survival for unresectable hepatocellular carcinoma (HCC) with a combination of surgery and intrahepatic arterial infusion of ¹³¹I-anti-HCC mAb. Phase I/II clinical trials. J. Cancer Res. Clin. Oncol, 124:275-280.

Zola (1987). Monoclonal Antibodies: A Manual of Techniques, pρ.147-158, CRC Press, Inc.

Zucker, S., Hymowitz, M., Rollo, E.E., Mann, R., Conner, C.E., Cao, J., Foda, H.D., Tompkins, D.C. & Toole. B.P. (2001). Tumorigenic Potential of Extracellular Matrix Metalloproteinase Inducer. Am. J. Pathol, 158:1921-1928.

Claims

CLAIMS:

1. A non-human antibody or antibody fragment comprising a variable region, which antibody or antibody fragment specifically binds to CD 147, wherein the variable region of the subject antibody or antibody fragment has been mutated so as to reduce the immunogenicity of the antibody or antibody fragment in humans.

2. The antibody or antibody fragment according to claim 1, wherein the antibody or antibody fragment is deimmunized.

3. The antibody or antibody fragment according to claim 1 or claim 2, wherein the antibody or antibody fragment is humanized.

4. The antibody or antibody fragment according to claim 3, wherein the antibody or antibody fragment is super-humanized™.

5. The antibody or antibody fragment according to any preceding claim, wherein the antibody or antibody fragment is resurfaced.

6. The antibody or antibody fragment according to any preceding claim, wherein the variable region comprises a heavy chain variable region and only the CDR3 of the heavy chain variable region is derived from or substantially the same as the subject non-human antibody or antibody fragment.

7. The antibody or antibody fragment according to claim 6, wherein the CDR3 of the heavy chain is derived from or has a sequence substantially identical to SEQ ID NO: 3.

8. The antibody or antibody fragment according to claim 6 or claim 7, wherein the only variable region present is a heavy chain variable region.

9. The antibody or antibody fragment according to claim 9, wherein one or both of the CDRl and CDR2 are derived from or are substantially the same as the subject non-human antibody or antibody fragment.

10. The antibody or antibody fragment according to any one of claims 1 to 7, wherein the antibody or antibody fragment comprises a heavy chain variable region and a light chain variable region.

11. The antibody or antibody fragment according to claim 10, wherein at least two of the six CDRs of the variable regions are derived from or substantially the same as those from the corresponding CDRs of the subject antibody or antibody fragment and wherein at least one of the CDRs is the heavy chain variable region CDR3.

12. The antibody or antibody fragment according to claim 11, wherein at least three, four, five or all six two of the CDRs of the variable regions are derived from or substantially the same as those from the corresponding CDRs of the subject antibody or antibody fragment.

13. The antibody or antibody fragment according to claim 10, comprising at least one, two, three or four CDR(s) selected from the group consisting of a heavy chain CDRl of SEQ ID NO: 1; a heavy chain CDR2 of SEQ ID NO: 2, a light chain CDRl of SEQ ID NO: 4, a light chain CDR2 of SEQ ID NO: 5 and a light chain CDR3 of SEQ ID NO: 6.

14. The antibody or antibody fragment of claim 10, comprising a heavy chain CDRl of SEQ ID NO: 1; a heavy chain CDR2 of SEQ ID NO: 2, a light chain CDRl of SEQ ID NO: 4, a light chain CDR2 of SEQ ID NO: 5 and a light chain CDR3 of SEQ ID NO: 6.

15. The antibody or antibody fragment according to any one of claims 3, wherein the variable region comprises a heavy chain amino acid sequence according to SEQ ID NO: 20.

16. The antibody or antibody fragment according to any one of claims 3, wherein the antibody or antibody fragment comprises a light chain amino acid sequence according to SEQ ID NO: 66 or to SEQ ID NO: 67.

17. The antibody or antibody fragment according to claim 3, wherein the variable region comprises a heavy chain amino acid sequence having at least 90% sequence identity to the amino acid sequence represented by SEQ ID NO: 20.

18. The antibody or antibody fragment according any preceding claim, wherein the antibody or antibody fragment in which the variable region has been mutated contains no more than 10 amino acid residues in the framework sequence that differ from those in the framework sequence of the candidate human variable region.

19. The antibody or antibody fragment according any preceding claim, wherein the antibody or antibody fragment in which the variable region has been mutated has no more than a 100-fold reduction in affinity for CD 147 compared with the subject murine antibody raised against CD 147.

20. The antibody or antibody fragment according any preceding claim, wherein the antibody or antibody fragment in which the variable region has been mutated has an affinity for CD 147 of at least Kd InM, 1OnM, 10OnM, or lμM.

21. The antibody or antibody fragment according to claim 3, wherein the antibody or antibody fragment in which the variable region has been mutated comprises a framework sequence derived from or substantially the same as a framework sequence of a human antibody variable heavy chain having a CDRl of canonical structure type 1 and a CDR2 of canonical structure type 4.

22. The antibody or antibody fragment according to claim 3, wherein the antibody or antibody fragment in which the variable region has been mutated comprises a framework sequence derived from or substantially the same as a framework sequence of a human antibody variable light chain having a CDRl of canonical structure type 2, a CDR2 of canonical structure type 1 and a CDR3 of canonical structure 1.

23. A pharmaceutical composition comprising the antibody or antibody fragment of any one of claims 1 to 22 and a pharmaceutically acceptable carrier.

24. A conjugate comprising the antibody or antibody fragment according to any one of claims 1 to 22 linked to a cytotoxic agent.

25. A conjugate of claim 24, wherein the said cytotoxic agent is a radionuclide.

26. A conjugate of claim 25, wherein the said radionuclide is selected from ¹³¹I and

90γ

27. A pharmaceutical composition comprising the conjugate of any one of claims 24 to 26 and a pharmaceutically acceptable carrier.

28. A method for inhibiting the growth of a cancer cell comprising contacting said cell with an antibody or antibody fragment of any one of claims 1 to 22.

29. A method for treating a patient having a cancer comprising administering to said patient an effective amount of the antibody or antibody fragment of any one of claims 1 to 22.

30. The method according to claim 29 further comprising administering to said patient another therapeutic agent.

31. The method according to claim 30, wherein said therapeutic agent is a cytotoxic agent.

32. The method for diagnosing a subject suspected of having a cancer, said method comprising: administering to said subject a conjugate comprising an antibody or antibody fragment according to any one of claims 1 to 22 linked to a detectable label; and detecting the distribution of said labelled conjugate within said subject.

33. The method according to any one of claims 28 to 32, wherein said cancer is a carcinoma.

34. The method of claim 33, wherein said carcinoma is hepatocellular carcinoma.

35. An improved antibody or antibody fragment that specifically binds to CD 147, prepared by:

(a) providing a DNA encoding an antibody or fragment thereof comprising at least one sequence selected from the group consisting of SEQ ID NOS: 1 to 69;

(b) introducing at least one nucleotide mutation, deletion or insertion into said DNA such that the amino acid sequence of said antibody or antibody fragment encoded by said DNA is changed;

(c) expressing said antibody or antibody fragment; and (d) screening said expressed antibody or antibody fragment for said improvement, whereby said improved antibody or antibody fragment is prepared.

36. The antibody or antibody fragment according to claim 35, wherein said improvement is an increased affinity for CD 147.

37. The antibody or antibody fragment according to claim 35 or claim 36, wherein said at least one nucleotide mutation, deletion or insertion is made by a method selected from the group consisting of oligonucleotide-mediated site-directed mutagenesis, cassette mutagenesis, degenerate oligonucleotide-primed PCR, error-prone PCR, DNA shuffling and the use of mutator-strains of bacteria.

38. The antibody or antibody fragment according to any one of claims 35 to 37, wherein at least one of the mutations is in a CDR.

39. A polynucleotide comprising a nucleic acid sequence encoding the antibody or antibody fragment according to any one of claims 1-22.

40. A vector comprising the polynucleotide of claim 39.

41. The vector of claim 40, wherein said vector is an expression vector capable of expressing said antibody or antibody fragment.

42. An antibody or antibody fragment that binds to CD147 comprising a heavy chain variable region, wherein the heavy chain variable region is a mutated version of SEQ ID NO: 7 containing between 1 and 20 mutations in the heavy chain variable region that cause it to have a higher amino acid sequence identity to a natural human heavy chain than that of the heavy chain variable region described by SEQ ID NO: 7.

43. An antibody or antibody fragment according to claim 42 further comprising a light chain variable region, wherein the light chain variable region has one, two or three CDRs chosen from SEQ ID NOs: 4-6.

44. An antibody or antibody fragment according to claim 42 further comprising a light chain variable region wherein the light chain comprises a mutated version of a light chain selected from SEQ ID NO: 50, wherein the said mutated version contains between 1 and 20 mutations in the light chain variable region that cause it to have a higher amino acid sequence identity to a natural human light chain.

45. An antibody that binds to an epitope of CD 147 overlapping with that of Habl8 and which has a reduced number of non-human amino acids compared to Habl8.

46. An antibody according to claim 45 that has reduced immunogenicity compared to Hablδ.

47. A humanized antibody heavy chain, said heavy chain comprising one or more CDRs derived from the murine antibody Habl8 heavy chain and comprising a framework sequence derived from a human antibody variable heavy chain having a CDRl of canonical structure type 1 and a CDR2 of length 19 according the antibody numbering system of Kabat.

48. A humanized antibody heavy chain according to claim 47 wherein the human antibody variable heavy chain has s sequence substantially the same as any one of SEQ ID NOs: 8 to l2.

49. A humanized antibody heavy chain, said heavy chain comprising one or more CDRs derived from the murine antibody Hablδ heavy chain and comprising a framework sequence derived from a human antibody variable heavy chain having a CDRl of canonical structure type 1 and a CDR2 of canonical structure type 3.

50. A humanized antibody heavy chain according to claim 49, wherein the human antibody variable heavy chain has a sequence substantially the same as any one of SEQ ID NOs: 23 to 46.

51. A humanized antibody heavy chain of any one of claims 47 to 50 wherein the heavy chain comprises at least one CDR selected from the group consisting of a CDRl of SEQ ID NO: 1; a CDR2 of SEQ ID NO: 2 and a CDR3 of SEQ ID NO: 3.

52. A humanized antibody light chain, said light chain comprising one or more CDRs derived from the murine antibody Habl8 light chain and comprising a framework sequence derived from a human antibody variable light chain having a CDRl of canonical structure type 2, a CDR2 of canonical structure type 1 and a CDR3 of canonical structure type 1.

53. A humanized antibody light chain according to claim 52, wherein the human antibody variable light chain has a sequence substantially the same as any one of SEQ ID NOs: 48 to 59.

54. A humanized antibody light chain according to claim 52 or claim 53 wherein the heavy chain comprises at least one CDR selected from the group consisting of a CDRl of SEQ ID NO: 4; a CDR2 of SEQ ID NO: 5 and a CDR3 of SEQ ID NO: 6.

55. A humanized antibody comprising a heavy chain of any one of claims 47 to 51 and/or a light chain of any one of claims 52 to 54.