WO2005112564A2 - Germline and sequence variants of humanized antibodies and methods of making and using them - Google Patents

Abstract

The present disclosure provides methods of designing humanized monoclonal antibodies having minimal immunogenicity and retained antigen binding affinity, by using human germline sequences as templates for each of the variable light chain and variable heavy chain regions (CDRs and frameworks). The present disclosure also provides humanized COL-1 antibodies generated by the disclosed methods. In several embodiments, methods are disclosed for the use of a humanized COL-1 antibody in the detection or treatment of a tumor in a subject. Also disclosed is a kit including the humanized COL-1 antibodies described herein.

Description

GERMLINE SEQUENCE VARIANTS OF HUMANIZED ANTIBODIES AND METHODS OF MAKING AND USING THEM

REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No.

60/562,781, filed April 15, 2004, and U.S. Provisional Application No. 60/580,839, filed June 18, 2004, both of which are incorporated herein by reference.

FIELD The present disclosure relates to methods of making humanized monoclonal antibodies that have minimal immunogenicity and that retain antigen binding affinity using human germline sequences as templates. The present disclosure also relates to humanized COL-1 antibodies that are generated using these methods. BACKGROUND The carcinoembyonic antigen (CEA) is a cell surface glycoprotein that is especially well characterized. CEA is a member of the immunoglobulin superfamily that includes normal fecal antigen, non-specific cross-reacting antigen, meconium antigen, and biliary glycoprotein. It is composed of seven domains linked to the cell membrane through a glycosylphosphatidylinositol anchor and has a molecular weight of 180 kDa. CEA is normally expressed in a variety of glandular epithelial tissues where it appears to be localized to the apical surface of the cells. It is over-expressed in 95% of gastrointestinal and pancreatic cancers and, as a result, is one of the most widely used human tumor markers for the diagnosis of human colon cancer. In addition, CEA is expressed in most non-small cell lung carcinomas, breast carcinoma and squamous cell carcinoma of the head and neck. Numerous monoclonal antibodies have been generated to detect various epitopes on CEA (Muraro et al, Cancer Res., 45:5769, 1985; Ohuchi et al, Cancer Res. 47:3565, 1987; Wilkinson et al, Proc. Natl. Acad. Sci. 98:10256, 2001). Of these monoclonal antibodies, COL-1 is of clinical importance because it has a high affinity for CEA. In addition, COL-1 reacts specifically with CEA and not with CEA-related antigens such as normal fecal antigen and non-specific cross-reacting antigen (Kuroki et al, Int. J. Cancer 44:208, 1989; Robbins et al, Int. J. Cancer 53:892, 1993). As a result of these properties, radiolabeled COL-1 has been used as a therapeutic agent in the treatment of patients with tumors that express CEA. Unfortunately, the murine origin of the antibody results in a human antimurine antibody (HAMA) response in these patients. The most widely used protocol in developing humanized antibodies involves the grafting of murine complementarity determining regions (CDRs) onto the frameworks of human antibodies. However, despite the numerous successes in developing humanized antibodies by CDR grafting, the murine CDRs in a humanized antibody could still evoke anti-variable (V) region responses in patients and as more humanized antibody is administered to a patient in multiple doses, the chance of an anti-V region response in the patient increases. Anti-V region responses are therefore the major impediment to the successful administration of humanized antibodies as a therapeutic treatment. Thus, there clearly exists a need to develop novel strategies to design antibodies with minimal immunogenicity and without compromising the antigen- binding affinity of the native antibody.

SUMMARY The present disclosure relates to methods of constructing humanized antibodies that retain antigen binding affinity and have reduced immunogenicity, compared to a parental antibody. The humanization of monoclonal antibodies (mAbs) by complementarity-determining region (CDR) grafting is a means to improve the clinical utility of xenogenic antibodies by reducing the human anti-murine antibody (HAMA) response elicited in patients. To minimize anti-V region responses, the antibody may be humanized by grafting onto the human templates only the specificity determining residues (SDRs), the residues that are essential for the surface complementarity of the antibody and its ligand. Generally, the humanization of an antibody, whether by CDR or SDR grafting, involves the use of a single human template for the entire variable light (VL) or variable heavy (VH) domain of an antibody. However, as disclosed herein, the homology between the human template sequences and the mAb to be humanized can be maximized by using a template from multiple human germline sequences conesponding to the different segments of the variable domain to reduce the immunogenicity of the humanized antibody. In one embodiment, the method involves, comparing at least one of a complementarity determining region (CDR)1 sequence, a CDR2 sequence, a CDR3 sequence, and a framework sequence, against a plurality of human germline sequences, wherein the CDRl, CDR2, CDR3, and framework sequences are all light chain sequences or are all heavy chain sequences from a non-human antibody that specifically binds an antigen. The method also involves selecting at least one of a first human germline sequence at least 50% homologous, but not identical, to the CDRl sequence, a second human germline sequence at least 50% homologous, but not identical, to the CDR2 sequence, a third human germline sequence at least 50%o homologous, but not identical, to the CDR3 sequence, and a fourth human germline sequence at least 50% homologous, but not identical, to the non-human framework sequence. The method further involves constructing a heavy chain antibody sequence or a light chain antibody sequence with at least one of the first, second, third, and fourth human germline sequences. The light chain sequence and/or the heavy chain sequence can be utilized to produce a minimally immunogenic humanized antibody with retained binding affinity for the antigen. In one example, the antigen is carcinoembryonic antigen (CEA). The present disclosure also relates to humanized COL-1 monoclonal antibodies that are minimally immunogenic and that retain CEA binding affinity, as compared to a parental humanized COL-1 antibody. The template for each light chain (L)-CDR, heavy chain (H)-CDR, variable light chain and heavy chain framework of the disclosed humanized COL-1 antibodies can be derived from human germline sequences that are homologous to the L-CDR, H-CDR, light chain and heavy chain sequences in the antibody to be humanized. More specifically, the disclosed humanized COL-1 antibodies contain a substitution of a murine non-ligand contact residue or a non- essential murine framework residue with a residue from the conesponding position of a homologous human gei line sequence. In one embodiment, the humanized antibodies contain CDRs and frameworks from murine COL-1 (mCOL-1), wherein a non-essential variable light chain framework residue is substituted with an amino acid from a conesponding Kabat position of a human gennline sequence. In another embodiment, a non-ligand contact residue of a mCOL-1 L-CDR is substituted with an amino acid from a conesponding Kabat position of a second human germline sequence. In yet another embodiment, a non-ligand contact residue in a mCOL-1 H-CDR is substituted with an amino acid from a conesponding Kabat position of a third human germline sequence. In one example of a humanized COL-1 monoclonal antibody, the light chain is an amino acid sequence as set forth in SEQ ID NO: 42, wherein a framework residue at position 9 of SEQ ID NO: 42 is substituted with an amino acid from a conesponding Kabat position of a first human germline sequence, and the heavy chain is an amino acid sequence as set forth in SEQ ID NO: 43, wherein an H-CDR2 residue at position 62 of SEQ ID NO: 43 is substituted with an amino acid from a conesponding Kabat position of a second human germlme sequence. Additional amino acid substitutions can be introduced into the humanized antibody sequence, for example (a) the light chain L- CDR1 residues at position 27 and at position 37 of SEQ ID NO: 42 are substituted with an amino acid from a conesponding Kabat position of third and fourth human germline sequences, respectively, (b) an L-CDR2 residue at position 57 of SEQ ID NO: 42 is substituted with an amino acid from a conesponding Kabat position of a fifth human gennline sequence, (c) an L-CDR3 residue at position 94 of SEQ ID NO: 42 is substituted with an amino acid from a conesponding Kabat position of a sixth human gennline sequence, or (d) L-CDR1 residues at position 27 and at position 37, the L- CDR2 residue at position 57, and the L-CDR3 residue at position 94 of SEQ ID NO: 42 are substituted with an amino acid from a conesponding Kabat position of third, fourth, fifth, and sixth human germline sequences, respectively. The humanized COL-1 antibody retains binding affinity for carcinoembryonic antigen (CEA) and has reduced immunogenicity, as compared to a parental humanized COL-1 antibody. Methods are also disclosed for the use of the humanized COL-1 monoclonal antibodies disclosed herein. A kit including the antibodies disclosed herein is also described.

The foregoing and other features and advantages will become more apparent from the following detailed description of several embodiments, which proceeds with reference to the accompanying figures.

BRIEF DESCRIPTION OF THE FIGURES FIG. 1 is a digital image of an analysis of the purified recombinant antibodies using the Agilent Bioanalyzer system under reducing conditions. Lane L, molecular weight markers; Lane 1, HuCOL-l_AbrCDR; Lane 2, SV1; Lane 3, SV2; Lane 4, SV3; Lane 5, SV4; Lane 6, SV5; Lane 7, HuCOL-l_SDR. Sizes of the molecular weight markers are given in the column at left. Three peaks are present in all lanes conesponding to the lower marker (6 kDa), system peak (9 kDa) and upper marker (210 kDa).

FIG. 2 is a graph of a competition radioimmunoassay (RIA) of specificity determining residue (SDR) variants of HuCOL-1. Increasing concentrations of monoclonal antibodies HuCOL-1 (open circle), HuCOL-l_AbrCDR (underscore), SV1 (closed square), SV2 (x), SV3 (closed circle), S V4 (open square), SV5 (closed diamond), HuCOL-l_SDR (open triangle) and human immunoglobulin G (HulgG) (plus 125 sign) were used to compete for the binding of I-labeled mCOL-1 to 200 ng of CEA coated in each well. FIGS. 3A-3C are plots of flow cytometric analyses of the binding of humanized COL-1 antibodies to cells expressing cell surface CEA. Binding profiles are of 1 μg of HuCOL-1 (FIG. 3A), HuCOL-l_AbrCDR (FIG. 3B) and HuCOL-l_SDR (FIG. 3C) monoclonal antibodies, to MC38 cells engineered to express CEA on their cell surface. Binding of an inelevant monoclonal antibody, human IgG (dashed line), is shown in each panel and represents less than 2% of the cell population.

FIGS. 4 A and 4B are graphs of binding sensorgrams showing the competition of HuCOL-1 (FIG. 4A) and HuCOL-l_SDR (FIG. 4B), for binding of the serum of patient MB to HuCOL-1 immobilized on the surface of a sensor chip. Different concentrations of the competitors were equilibrated with the MB serum before sample application.

FIGS. 5A and 5B are two plots showing sera reactivity, by surface plasmon resonance (SPR), of humanized COL-1 antibodies. Increasing concentrations of HuCOL-1 (closed square), HuCOL-l_AbrCDR (open circle), and HuCOL-l_SDR (closed triangle) monoclonal antibodies were used to compete with the anti-V region Abs to COL-1 present in the sera of patients MB (FIG. 5A) and EM (FIG. 5B) for binding to HuCOL-1 immobilized on a sensor chip. Percent binding of the sera to HuCOL-1 was calculated from the binding sensorgrams and plotted as a function of the competitor concentration.

FIGS. 6A and 6B are schematic representations of the humanization protocols for the mCOL-1. FIG. 6A shows the amino acid sequences of the variable light (V_L) regions of mCOL-1 (SEQ ID NO: 44), human antibody VJI'CL (SEQ ID NO: 31), HuCOL-1 derived from mCOL-1 and VJI'CL (SEQ ID NO: 46), and the light chain of HuCOL-1 variant HuCOL-l_AbrCDR (Variant; SEQ ID NO: 47). FIG. 6B shows the amino acid sequences of the variable heavey (VH) regions of mCOL-1 (SEQ ID NO: 45), human antibody MO30 (SEQ ID NO: 35), HuCOL-1 derived from mCOL-1 and MO30 (SEQ ID NO: 48), and the heavy chain of HuCOL-1 variant HuCOL-l_AbrCDR (Variant; SEQ ID NO: 49). Dashes indicate residues that are identical in mCOL-1, human and humanized antibodies. Asterisks mark frameworks residues that are deemed essential for maintaining the combining site structure of mCOL-1. Murine frameworks residues retained in the HuCOL-1 are shown in bold.

FIG. 7A-B are schematic diagrams of a method of designing humanized antibodies using different human germline sequences as templates for each of the variable light chain CDRs, variable heavy chain CDRs, variable light chain frameworks, and variable heavy chain frameworks onto which SDRs may be grafted.

SEQUENCE LISTING The nucleic and amino acid sequences listed in the accompanying sequence listing are shown using standard letter abbreviations for nucleotide bases, and three letter code for amino acids, as defined in 37 C.F.R. 1.822. Only one strand of each nucleic acid sequence is shown, but the complementary strand is understood as included by any reference to the displayed strand. In the accompanying sequence listing: SEQ ID NOs: 1-3 and 7 are the nucleic acid sequence of variable light chain 5' primers. SEQ ID NOs: 4-6 and 8 are the nucleic acid sequences of variable light chain 3' primers. SEQ ID NO: 9 is the amino acid sequence of the murine COL-1 L-CDR1. SEQ ID NO: 10 is the amino acid sequence of the murine COL-1 L-CDR2. SEQ ID NO: 11 is the amino acid sequence of the murine COL-1 L-CDR3. SEQ ID NO: 12 is the amino acid sequence of the human germline L-CDR1. SEQ ID NO: 13 is the amino acid sequence of the human gennline L-CDR2. SEQ ID NO: 14 is the amino acid sequence of the human germline L-CDR3. SEQ ID NO: 15 is the amino acid sequence of the murine COL-1 H-CDR1. SEQ ID NO: 16 is the amino acid sequence of the murine COL-1 H-CDR2. SEQ ID NO: 17 is the amino acid sequence of the murine COL-1 H-CDR3. SEQ ID NO: 18 is the amino acid sequence of the human germline H-CDR1. SEQ ID NO: 19 is the amino acid sequence of the human germline H-CDR2. SEQ ID NO: 20 is the amino acid sequence of FRl of the light chain of murine COL-1. SEQ ID NO: 21 is the amino acid sequence of FR2 of the light chain of murine COL-1. SEQ ID NO: 22 is the amino acid sequence of FR3 of the light chain of murine COL-1. SEQ ID NO: 23 is the amino acid sequence of FR4 of the light chain of murine

COL-1. SEQ ID NO: 24 is the amino acid sequence of FRl of the heavy chain of murine COL-1. SEQ ID NO: 25 is the amino acid sequence of FR2 of the heavy chain of murine COL-1. SEQ ID NO: 26 is the amino acid sequence of FR3 of the heavy chain of murine COL-1. SEQ ID NO: 27 is the amino acid sequence of FR4 of the heavy chain of murine COL-1.

SEQ ID NO: 28 is the amino acid sequence of FRl of a human germline light chain. SEQ ID NO: 29 is the amino acid sequence of FR2 of a human germline light chain. SEQ ID NO: 30 is the amino acid sequence of FR3 of a human germline light chain. SEQ ID NO: 31 is the amino acid sequence of the human VJI'CL light chain. SEQ ID NO: 32 is the amino acid sequence of FRl of a human germline heavy chain. SEQ ID NO: 33 is the amino acid sequence of FR2 of a human germline heavy chain. SEQ ID NO: 34 is the amino acid sequence of FR3 of a human germline heavy chain. SEQ ID NO: 35 is the amino acid sequence of the human MO30 heavy chain. SEQ ID NO: 36 is the amino acid sequence of the DPK22 human germline light chain sequence (encoded by the nucleic acid sequence in GenBank Accession No. X93639, herein incoφorated by reference). SEQ ID NO: 37 is the amino acid sequence of the DPK5 human germline light chain sequence (encoded by the nucleic acid sequence in GenBank Accession No. X93623, herein incorporated by reference). SEQ ID NO: 38 is the amino acid sequence of the DPK9 human germline light chain sequence (encoded by the nucleic acid sequence in GenBank Accession No. X93627/Jκ4, herein incorporated by reference). SEQ ID NO: 39 is the amino acid sequence of the DP-7 human germline heavy chain sequence (GenBank Accession No. Z12309, herein incorporated by reference). SEQ ID NO: 40 is the amino acid sequence of the DP-75 human gennline heavy chain sequence (GenBank Accession No. Z14071, herein incorporated by reference). SEQ ID NO: 41 is the amino acid sequence of the DPK24 human germline light chain sequence (encoded by the nucleic acid sequence in GenBank Accession No. X93640, herein incoφorated by reference). SEQ ID NO: 42 is the amino acid sequence of a humanized antibody light chain with mCOL-1 L-CDRs and HuCOL-l_AbrcDR light chain frameworks. SEQ ID NO: 43 is the amino acid sequence of a humanized antibody heavy chain with mCOL-1 H-CDRs and HuCOL-lAbrCDR heavy chain frameworks. SEQ ID NO: 44 is the amino acid sequence of the mCOL-1 light chain. SEQ ID NO: 45 is the amino acid sequence of the mCOL-1 heavy chain. SEQ ID NO: 46 is the amino acid sequence of the HuCOL-1 light chain. SEQ ID NO: 47 is the amino acid sequence of the HuCOL-l_Ab_rCDR light chain. SEQ ID NO: 48 is the amino acid sequence of the HuCOL-1 heavy chain. SEQ ID NO: 49 is the amino acid sequence of the HuCOL-l _brcDR heavy chain. SEQ ID NO: 50 is the amino acid sequence of the HuCOL-l_{A r}cDR L-CDR1. SEQ ID NO: 51 is the amino acid sequence of the HuCOL-l_AbrCDR H-CDR2. SEQ ID NO: 52 is the amino acid sequence of the HuCOL-l_AbrCDR light chain

FRl. SEQ ID NO: 53 is the amino acid sequence of the HuCOL-l_AbrcDR light chain FR2. SEQ ID NO: 54 is the amino acid sequence of the HuCOL-l_AbrcDR light chain FR3. SEQ ID NO: 55 is the amino acid sequence of the HuCOL-l _rcDR heavy chain FRl. SEQ ID NO: 56 is the amino acid sequence of the HuCOL-l_AbrcDR heavy chain FR2. SEQ ID NO: 57 is the amino acid sequence of the HuCOL-l_AbrCDR heavy chain

FR3. SEQ ID NO: 58 is the amino acid sequence of the HuCOL-l_AbrcDR heavy chain FR4. DETAILED DESCRIPTION

I. Abbreviations

Ab antibody

Ag antigen bp basepair

BSM bovine submaxillary mucin

C constant

CDR complementarity determining region

CEA carcinoembryonic antigen C_H constant heavy

CHO Chinese hamster ovary

C_L constant light

CMV cytomegalovirus

ELISA enzyme linked immunoassay

Fab fragment antigen binding

F(ab')₂ Fab with additional amino acids, including cysteines necessary for disulfide bonds

FACS fluorescence activated cell sorting

FITC fluorescein isothiocyanate

FR framework region

Fv fragment variable

H heavy

HAMA human antimurine antibody

HuCOL-1 humanized COL-1

HulgG human immunoglobulin G

IC₅₀ half maximal inhibition of binding

IFN interferon ig immunoglobulin

IL interleukin

Ka relative affinity constant

L light mAb monoclonal antibody mCOL-1 murine COL-1

PBS phosphate buffered saline

PCR polymerase chain reaction

PE Pseudomonas exotoxin

PDB protein data bank

RIA radioimmunoassay scFv single chain Fv

SDR specificity determining residue

SPR surface plasmon resonance

SV40 simian virus 40

TCR T cell receptor

TNF tumor necrosis factor

V variable

V_H variable heavy

V_L variable light

II. Terms Unless otherwise noted, technical terms are used according to conventional usage. Definitions of common terms in molecular biology may be found in Benjamin Lewin, Genes V, published by Oxford University Press, 1994 (ISBN 0- 19-854287-9);

Kendrew et al. (eds.), The Encyclopedia of Molecular Biology, published by Blackwell

Science Ltd., 1994 (ISBN 0-632-02182-9); and Robert A. Meyers (ed.), Molecular

Biology and Biotechnology: a Comprehensive Desk Reference, published by VCH

Publishers, Inc., 1995 (ISBN 1-56081-569-8). In order to facilitate review of the various embodiments of the invention, the following explanations of specific terms are provided:

Animal: Living multi-cellular vertebrate organisms, a category that includes, for example, mammals and birds. The term mammal includes both human and non- human mammals. Similarly, the term "subject" includes both human and veterinary subjects. Antibody: Immunoglobulin (Ig) molecules and immunologically active portions of Ig molecules, for instance, molecules that contain an antigen binding site which specifically binds an antigen. A naturally occurring antibody (for example, IgG) includes four polypeptide chains, two heavy (H) chains and two light (L) chains inter-connected by disulfide bonds and is produced by B cells or a hybridoma exposed to an antigen of interest. However, it has been shown that the antigen-binding function of an antibody can be performed by fragments of a naturally occurring antibody. Thus, these antigen-binding fragments are also intended to be designated by the term "antibody." Examples of binding fragments encompassed within the term antibody include (i) an Fab fragment consisting of the V_L, V_H, C_L and C_HI domains; (ii) an Fd fragment consisting of the V_H and C_HI domains; (iii) an Fv fragment consisting of the V_L and V_H domains of a single arm of an antibody, (iv) a dAb fragment (Ward et al. , Nature 341 :544, 1989) which consists of a VH domain; and (v) an F(ab')₂ fragment, a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region. Furthermore, although the two domains of the Fv fragment are coded for by separate genes, a synthetic linker can be made that enables them to be made as a single protein chain (known as single chain Fv (scFv); Bird et al, Science 242:423, 1988; and Huston et al. Proc. Natl. Acad. Sci. 85:5879, 1988) by recombinant methods. Such single chain antibodies, as well as dsFv, a disulfide stabilized Fv (Bera et al, J. Mol. Biol. 281:475, 1998), and dimeric Fvs (diabodies), that are generated by pairing different polypeptide chains (Holliger et al, Proc. Natl. Acad. Sci. 90:6444, 1993), are also included. In one embodiment, antibody fragments for use in this disclosure are those which are capable of cross-linking their target antigen, for example, bivalent fragments such as F(ab')₂ fragments. Alternatively, an antibody fragment which does not itself cross-link its target antigen (for example, a Fab fragment) can be used in conjunction with a secondary antibody which serves to cross-link the antibody fragment, thereby cross-linking the target antigen. Antibodies can be fragmented using conventional techniques and the fragments screened for utility in the same manner as described for whole antibodies. An antibody is further intended to include humanized monoclonal molecules that specifically bind the target antigen. "Specifically binds" refers to the ability of individual antibodies to specifically immunoreact with an antigen. This binding is a non-random binding reaction between an antibody molecule and the antigen. In one embodiment, the antigen is CEA. Binding specificity is typically determined from the reference point of the ability of the antibody to differentially bind the antigen of interest and an unrelated antigen, and therefore distinguish between two different antigens, particularly where the two antigens have unique epitopes. An antibody that specifically binds to a particular epitope is refened to as a "specific antibody." A variety of methods for linking effector molecules to antibodies are well known in the art. Detectable labels useful for such puφoses are also well known in the art, and include radioactive isotopes such as ³²P, fluorophores, chemiluminescent agents, and enzymes. Also encompassed in the disclosure are the chemical or biochemical modifications that incoφorate toxins in the antibody. In one embodiment, the toxin is chemically conjugated to the antibody, h another embodiment, a fusion protein is genetically engineered to include the antibody and the toxin. Specific, non- limiting examples of toxins are radioactive isotopes, chemotherapeutic agents, bacterial toxins, viral toxins, or venom proteins. The disclosure also includes chemical or genetically engineered modifications that link a cytokine to an antibody (such as by a covalent linkage). Specific, non-limiting examples of cytokines are interleukin (IL)-2, IL-4, IL-10, tumor necrosis factor (TNF)-alpha and interferon (IFN)-gamma. In one embodiment the antigen is CEA. Monoclonal and humanized immunoglobulins are encompassed by the disclosure. In one example, a murine monoclonal antibody that recognizes CEA is COL-1. In another example, an antibody that binds CEA is a humanized antibody, such as HuCOL- 1 (ATCC Accession Number PTA-4661). hi additional examples, a humanized antibody that binds CEA is HuCOL- I_SD_R, SV2, SV4, SV5, or HuCOL-l_AbrCDR (also refened to as HuCOL-l²⁴'²⁵'²⁷L/⁶¹H; ATCC Accession Number PTA-4644). The disclosure also includes synthetic and genetically engineered variants of these immunoglobulins. Antigen: Any molecule that can bind specifically with an antibody. An antigen is also a substance that antagonizes or stimulates the immune system to produce antibodies. Antigens are often foreign substances such as allergens, bacteria or viruses that invade the body. One specific, non-limiting example of an antigen is CEA. Carcinoembryonic antigen (CEA): A member of the immunoglobulin superfamily that includes normal fecal antigen, non-specific cross-reacting antigen, meconium antigen, and biliary glycoprotein. CEA is composed of seven domains linked to the cell membrane through a glycosylphosphatidylinositol anchor and has a molecular weight of 180 kDa (GenBank Accession Number A36319, herein incoφorated by reference). CEA is normally expressed in a variety of glandular epithelial tissues, where it appears to be localized to the apical surface of the cells, although it is also expressed in numerous carcinomas including gastrointestinal, colorectal, breast, ovarian and lung carcinomas (Robbins et al, Int'lJ. Cancer, 53:892- 897, 1993; Greiner et al, J. Clin. Oncol, 10:735-746, 1992; Ohuchi et al, Cancer Res. 47:3565-5780, 1985; Muraro et al, Cancer Res., 45:57695780, 1985). CEA is an especially well characterized human tumor antigen and is widely used for the diagnosis of human colon cancer. Monoclonal antibodies, designated COL-1 through COL- 15, have been generated to detect various epitopes on CEA (Muraro et al, Cancer Res., 45:5769-5780, 1985, herein incoφorated by reference), and in using these antibodies the differential expression of CEA has been determined (Muraro et al. , Cancer Res. , 45:5769-5780, 1985; Ohuchi et al, Cancer Res. 47:3565-3571, 1987; Wilkinson et al, Proc. Natl. Acad. Sci. 98:10256, 2001). Of these monoclonal antibodies, COL-1 is of clinical importance because it has a high affinity for CEA. cDNA (complementary DNA): A piece of DNA lacking internal, non-coding segments (introns) and regulatory sequences that determine transcription. cDNA is synthesized in the laboratory by reverse transcription from messenger RNA extracted from cells. Chimeric antibody: An antibody which includes sequences derived from two different antibodies, which typically are of different species. Most typically, chimeric antibodies include human and murine antibody domains. One example of a chimeric antibody is an antibody with human framework regions and murine CDRs. Complementarity Determining Region (CDR): Amino acid sequences which together define the binding affinity and specificity of the natural Fv region of a native Ig binding site. The light and heavy chains of an Ig each have three CDRs, designated light chain (L)-CDRl, L-CDR2, L-CDR3 and heavy chain (H)-CDRl, H-CDR2, H- CDR3, respectively. The CDRs of the light chain are bounded by the residues at positions 24 and 34 (L-CDRl), 50 and 56 (L-CDR2), 89 and 97 (L-CDR3); the CDRs of the heavy chain are bounded by the residues at positions 31 and 35b (H-CDR1), 50 and 65 (H-CDR2), 95 and 102 (H-CDR3), using the numbering convention delineated by Kabat et al. (Sequences of Proteins of Immunological Interest, 5^th Edition, Department of Health and Human Services, Public Health Service, National Institutes of Health, Bethesda, NIH Publication No. 91-3242, 1991, incoφorated herein by reference). Constant Region: The portion of the antibody molecule which confers effector functions. In the present disclosure, the variant antibodies include constant regions derived from human immunoglobulins. The heavy chain constant region can be selected from any of five isotypes: alpha, delta, epsilon, gamma or mu. Heavy chains of various subclasses (such as the IgG subclass of heavy chains) are responsible for different effector functions. Thus, by choosing the desired heavy chain constant region, humanized antibodies with the desired effector function can be produced. The light chain constant region can be of the kappa or lambda type. Corresponding Kabat position: A position of a residue in an amino acid sequence that follows the numbering convention delineated by Kabat et al. (Sequences of Proteins of Immunological Interest, 5^th Edition, Department of Health and Human Services, Public Health Service, National Institutes of Health, Bethesda, NIH Publication No. 91-3242, 1991). Also refened to herein as a conesponding position. DNA: Deoxyribonucleic acid. DNA is a long chain polymer which constitutes the genetic material of most living organisms (some viruses have genes composed of ribonucleic acid (RNA)). The repeating units in DNA polymers are four different nucleotides, each of which contains one of the four bases, adenine, guanine, cytosine and thymine bound to a deoxyribose sugar to which a phosphate group is attached. Triplets of nucleotides (refened to as codons) code for each amino acid in a polypeptide. The term codon is also used for the conesponding (and complementary) sequence of three nucleotides in the mRNA that is transcribed from the DNA. Effector Molecule: Therapeutic, diagnostic or detection moieties linked to an antibody, using any number of means known to those of skill in the art. Both covalent and noncovalent linkage means may be used. The procedure for linking an effector molecule to an antibody varies according to the chemical structure of the effector.

Polypeptides typically contain a variety of functional groups; for example, carboxylic acid (COOH), free amino (-NH₂) or sulfhydryl (-SH) groups, which are available for reaction with a suitable functional group on an antibody to result in the linkage of the effector molecule. Alternatively, the antibody is derivatized to expose or link additional reactive functional groups. The derivatization may involve linkage of any of a number of linker molecules such as those available from Pierce Chemical Company, Rockford, IL. The linker can be any molecule used to join the antibody to the effector molecule. The linker is capable of forming covalent bonds to both the antibody and to the effector molecule. Suitable linkers are well known to those of skill in the art and include, but are not limited to, straight or branched-chain carbon linkers, heterocyclic carbon linkers, or peptide linkers. Where the antibody and the effector molecule are polypeptides, the linkers may be joined to the constituent amino acids through their side groups (for example, through a disulfide linkage to cysteine) or to the alpha carbon amino and carboxyl groups of the terminal amino acids. An "immunoconjugate" is a covalent linkage of an effector molecule, such as a toxin, a chemical compound, or a detectable label, to an antibody. Examples of toxins include radioactive isotopes, chemotherapeutic drugs, bacterial toxins, viral toxins, and proteins contained in venom (for example, insect, spider, reptile, or amphibian venom). Specific, non-limiting examples of toxins include, but are not limited to, abrin, ricin, Pseudomonas exotoxin (such as PE35, PE37, PE38, andPE40), diphtheria toxin, anthrax toxin, botulinum toxin, or modified toxins thereof. For example, Pseudomonas exotoxin and diphtheria toxin are highly toxic compounds that typically bring about death through liver toxicity. Pseudomonas exotoxin and diphtheria toxin, however, can be modified into a form for use as an immunotoxin by removing the native targeting component of the toxin (for example, domain la of Pseudomonas exotoxin and the B chain of diphtheria toxin) and replacing it with a different targeting moiety, such as an antibody. Other toxic agents, that directly or indirectly inhibit cell growth or kill cells (cytotoxins), include chemotherapeutic drugs, cytokines, for example interleukin-2 or interferon, radioactive isotopes, viral toxins, or proteins contained within, for example, insect, reptile, spider, or amphibian venom. Specific, non-limiting examples of detectable labels include, but are not limited to, radioactive isotopes, enzyme substrates, co-factors, ligands, chemiluminescent agents, fluorescent agents, haptens, or enzymes. In one embodiment, an antibody is joined to an effector molecule. In another embodiment, an antibody joined to an effector molecule is further joined to a lipid or other molecule to a protein or peptide to increase its half-life in the antibody. The linkage can be, for example, either by chemical or recombinant means. In one embodiment, the linkage is chemical, wherein a reaction between the antibody moiety and the effector molecule has produced a covalent bond formed between the two molecules to form one molecule. A peptide linker (short peptide sequence) can optionally be included between the antibody and the effector molecule. In some circumstances, it is desirable to free the effector molecule from the antibody when the immunoconjugate has reached its target site. Therefore, in these circumstances, immunoconjugates will comprise linkages that are cleavable in the vicinity of the target site. Cleavage of the linker to release the effector molecule from the antibody may be prompted by enzymatic activity or conditions to which the immunoconjugate is subjected either inside the target cell or in the vicinity of the target site. When the target site is a tumor, a linker which is cleavable under conditions present at the tumor site (for example, when exposed to tumor-associated enzymes or acidic pH) may be used. In view of the large number of methods that have been reported for linking a variety of radiodiagnostic compounds, radiotherapeutic compounds, label (for example, enzymes or fluorescent molecules) drugs, toxins, and other agents to antibodies one skilled in the art will be able to determine a suitable method for linking a given agent to an antibody. Encode: A polynucleotide is said to "encode" a polypeptide if, in its native state or when manipulated by methods well known to those skilled in the art, it can be transcribed and/or translated to produce the mRNA for the polypeptide and/or a fragment thereof. The anti-sense strand is the complement of such a nucleic acid, and the encoding sequence can be deduced therefrom. Epitope: A site on an antigen recognized by an antibody, as determined by the specificity of the antibody amino acid sequence. Epitopes are also called antigenic determinants. Framework Region (FR): Amino acid sequences inteφosed between CDRs include variable light and variable heavy framework regions. The framework regions serve to hold the CDRs in an appropriate orientation for antigen binding. Essential framework residues are important for maintaining the combining site structure of an antibody. Non-essential framework residues can be substituted without compromising antigen binding. Germline sequence: Nucleic acid sequences encoding different domains of immunoglobulins (or T cell receptor) in their unreananged state rather than the reananged sequences (such as reananged nucleic acid sequences for production of immunoglobulins or T cell receptor molecules), and the amino acids encoded therein. Three separate loci encode, respectively, the Ig K light chain, the Ig λ light chain, and all the Ig heavy chain germline sequences. Each germline Ig locus is made up of at least three different types of gene segments, the variable (V), constant (C), and joining segments (J), that are separated from one another in the genome by large stretches of DNA that are never transcribed. The gennline organization of Ig (and T cell receptor) loci exists in all cell types of the body. However, gennline genes cannot be transcribed into mRNA that gives rise to functional antibodies. These are created only in developing B and T lymphocytes by reanangement of DNA that make the V, C, and J segments contiguous; the human V, D and J segments recombine during B cell development. Germline sequences of immunoglobulins do not include sequences wherein V-D-J recombination has occurred, and do not contain somatic hypermutations. A "human germline sequence" is a nucleic acid sequence encoding domains of immunoglobulins in their unreananged state or an amino acid sequence encoded thereby. The V regions from light and heavy chain germline sequences have been cloned (see, for example, Tomlinson et al. (J. Mol. Biol, 227:776-798, 1992, and Cox et al, Eur. J. Immunol, 24:827-836,.1994, which are incoφorated herein by reference). "Human germline sequence" refers to natural occuning and synthetic forms of the naturally occurring unreananged immunoglobulin sequences. HAMA (Human anti-murine antibody) response: An immune response in a human subject to the variable and constant regions of a murine antibody that has been administered to the patient. Repeated antibody administration may lead to an increased rate of clearance of the antibody from the patient's serum and may also elicit allergic reactions in the patient. Humanized antibody: A human antibody genetically engineered to include critical murine monoclonal antibody residues. A humanized antibody can include a human framework region and one or more CDRs from a non-human (such as a mouse, rat, or synthetic non-human) immunoglobulin. In one embodiment, the DNA encoding hypervariable loops of mouse monoclonal antibodies or variable regions selected in phage display libraries is inserted into the framework regions of human Ig genes. In another embodiment, murine residues important in antigen binding (ligand contact residues or specificity determining residues (SDRs), or essential framework residues) are inserted into the conesponding position of the variable region of human germline or Ig sequences. In yet another embodiment, a human residue, such as a human germline residue, is inserted into the corresponding position of a murine Ig sequence. Antibodies can be "customized" to have a desired binding affinity or to be minimally immunogenic in the humans treated with them. Humanized COL-1 antibodies: COL-1 antibodies humanized by grafting mCOL-1 (murine COL-1) CDRs onto the frameworks of the relevant human antibodies. The murine CDRs in the resultant humanized COL-1 antibody could evoke an anti- idiotypic response when administered in human subjects. COL-1 can be humamzed by grafting only a subset of the COL-1 CDR residues, for example those that are important for antigen binding (ligand contact residues or SDRs), onto the variable light and variable heavy framework regions of, for example, human germline antibody sequences. In one embodiment of a humanized COL-1 antibody, COL-1 CDR residues that are not involved in antigen binding (non-ligand contact residues or non-SDRs) are substituted with the conesponding residues of one or more human antibodies such as those from, for example, human germline antibody sequences. In another embodiment of a humanized COL-1 antibody, SDR and non-SDR residues are substituted with the cooesponding residues of one or more human germline antibody sequences. A specific humanized COL-1 monoclonal antibody, termed HuCOL- 1, has been deposited with ATCC as PTA-4661. In another embodiment, a humanized COL-1 antibody is HuCOL-l_AbrCDR (HuCOL-l²⁴'²⁵'²⁷L/⁶¹H; ATCC Accession Number PTA- 4644). For the puφoses of this disclosure, HuCOL-1 is refened to as the parental antibody. Other specific, non-limiting examples of a humamzed COL-1 monoclonal antibody are SV1, SV2, SV3, SV4, SV5, and HUCOL-1_SDR- For the puφoses of this disclosure, SV1, SV2, SV3, SV4, SV5, and HUCOL-I_SDR are also refened to as variant humanized COL-1 antibodies. Methods for making these antibodies and the amino acid sequence of the V_L and V_H chains of these antibodies are provided herein. IC₅₀ value: The concentration of a competitor antibody (for example, concentration of a humanized COL-1 antibody) required for half-maximal (50%) inhibition of binding of sera to another antibody (for example, the parental HuCOL- 1 or the HuCOL-l_Abr_CD_R). A higher IC50 value for a particular antibody indicates a decreased reactivity of that antibody to the serum, suggesting that the antibody with the higher ICso value has reduced irnmunogenicity in a subject. In one embodiment, a humanized COL-1 antibody has an IC₅₀ value that is greater than that of the parental HuCOL- 1 antibody, suggesting that the humanized COL-1 antibody has a decreased immunogenicity in a subject compared to the parental HuCOL- 1 antibody. Idiotype: The property of a group of antibodies or T cell receptors defined by their sharing a particular idiotope (an antigenic determinant on the variable region); for instance, antibodies that share a particular idiotope belong to the same idiotype. "Idiotype" may be used to describe the collection of idiotopes expressed by an Ig molecule. An "anti-idiotype" antibody can be prepared to a monoclonal antibody by methods known to those of skill in the art and may be used to prepare pharmaceutical compositions. Immune cell: Any cell involved in a host defense mechanism. These can include, for example, T cells, B cells, natural killer cells, neutrophils, mast cells, macrophages, antigen-presenting cells, basophils, eosinophils, and neutrophils. Immune response: A response of a cell of the immune system, such as a neuxrophil, a B cell, or a T cell, to a stimulus. In one embodiment, the response is specific for a particular antigen (an "antigen-specific response"). In another embodiment, the response is against an antibody, such as a HAMA response, including an anti-variable region response. Immunogenicity: A measure of the ability of a targeting protein or therapeutic moiety to elicit an immune response (humoral or cellular) when adniinistered to a subject. An antibody that generates a reduced, for example low, immune response when administered to a subject, such as a human subject, is "minimally immunogenic" or has "minimal immunogenicity." In one embodiment, a minimally immunogenic antibody is a murine antibody that is administered to a human subject without eliciting a HAMA response. In one example, a humanized COL-1 antibody, such as, but not limited to, SV2, SV4, SV5, or HuCOL- I_SDR, has minimal immunogenicity (or is minimally immunogenic) compared to the parental HuCOL-1. Various assays can be used to measure immunogenicity. In one embodiment, immunogenicity is measured in a competitive binding assay. In one specific, non- limiting example, immunogenicity is the ability of a variant humanized COL-1 antibody to prevent the parental HuCOL- 1 antibody from binding to COL-1 anti-idiotypic antibodies in a patient's serum. If a variant humanized COL-1 antibody competes with an equal molar amount of the parental HuCOL- 1 antibody (for instance, elicits greater than about 50% inhibition of parental HuCOL- 1 binding to anti-idiotypic antibodies in a patient's serum) then the variant humanized COL-1 antibody is immunogenic. If a variant humanized COL-1 antibody competes poorly with an equal molar or less amount of the parental HuCOL-1 antibody (for instance, elicits about 50% or less inhibition of parental HuCOL-1 binding to anti-idiotypic antibodies in a patient's serum) then the variant humanized COL-1 antibody is minimally immunogenic. In another embodiment, if a two-fold or greater molar concentration of a variant humanized COL-1 antibody is required to achieve about 50% inhibition of binding of the parental antibody to its cognate anti-idiotypic antibodies present in a subject's sera, then the variant antibody is minimally immunogenic. IC₅₀ is the concentration of the competitor antibody (for example, concentration of a variant humanized COL-1) required for half-maximal (50%) inhibition of binding of sera to HuCOL-1 or another variant humanized COL-1 antibody. Immunoreactivity: A measure of the ability of an Ig to recognize and bind to a specific antigen. Isolated: An biological component (such as a nucleic acid, peptide or protein) that has been substantially separated, produced apart from, or purified away from other biological components in the cell of the organism in which the component naturally occurs, for instance other chromosomal and extrachromosomal DNA and RNA, and proteins. Nucleic acids, peptides and proteins that have been "isolated" thus include nucleic acids and proteins purified by standard purification methods. The term also embraces nucleic acids, peptides and proteins prepared by recombinant DNA expression in a host cell as well as chemically synthesized nucleic acids. Label: A detectable compound or composition that is conjugated directly or indirectly to another molecule to facilitate detection of that molecule. Specific, non- limiting examples of labels include fluorescent tags, chemiluminescent tags, haptens, enzymatic linkages, and radioactive isotopes. Ligand contact residue or Specificity Determining Residue: A residue within a CDR that is involved in contact with a ligand or antigen. A ligand contact residue is also known as a specificity determining residue (SDR). A non-ligand contact residue is a residue in a CDR that does not contact a ligand. A non-ligand contact residue can also be a framework residue. Lymphocytes: A type of white blood cell that is involved in the immune defenses of the body. There are two main types of lymphocytes: B-cells and T-cells. Mammal: This term includes both human and non-human mammals. Similarly, the term "subject" includes both human and veterinary subjects. Monoclonal antibody: An antibody produced by a single clone of B- lymphocytes. Monoclonal antibodies are produced by methods known to those of skill in the art, for instance by making hybrid antibody-forming cells from a fusion of myeloma cells with immune spleen cells. Nucleic acid: A deoxyribonucleotide or ribonucleotide polymer in either single or double stranded form, and unless otherwise limited, encompasses known analogues of natural nucleotides that hybridize to nucleic acids in a manner similar to naturally occurring nucleotides. Oligonucleotide: A linear single-stranded polynucleotide sequence of up to about 200 nucleotide bases in length, for example a polymer of deoxyribonucleotides or ribonucleotides which is at least 6 nucleotides, for example at least 15, 50, 100 or even 200 nucleotides long. Operably linked: A first nucleic acid sequence is operably linked with a second nucleic acid sequence when the first nucleic acid sequence is placed in a functional relationship with the second nucleic acid sequence. For instance, a promoter is operably linked to a coding sequence if the promoter affects the transcription or expression of the coding sequence. Generally, operably linked DNA sequences are contiguous and, where necessary to join two protein coding regions, in the same reading frame. Phage display: A technique wherein DNA sequences are amplified and cloned into filamentous phage vector to create a library of fusion phages ("phage library") in which the phages display on their surface the proteins encoded by the foreign DNA. From the rescued phages, the individual phage clones are selected through interaction of the displayed protein with a ligand, and the specific phage is amplified by infection of bacteria. Antigen specific immunoglobulins can then be expressed and characterized for their antigen binding and sera reactivity (potential immunogenicity). Pharmaceutical agent: A chemical compound or composition capable of inducing a desired therapeutic or prophylactic effect when properly administered to a subject or a cell. "Incubating" includes a sufficient amount of time for a drug to interact with a cell. "Contacting" includes incubating a drug in solid or in liquid form with a cell. A "therapeutically effective amount" is a quantity of a specific substance sufficient to achieve a desired effect in a subject being treated. For instance, this can be the amount necessary to inhibit or suppress growth of a tumor or to decrease a sign or symptom of the tumor in the subject. In one embodiment, a therapeutically effective amount is the amount necessary to eliminate a tumor. When administered to a subject, a dosage will generally be used that will achieve target tissue concentrations (for example, in tumors) that has been shown to achieve a desired in vitro effect. Pharmaceutically acceptable carriers: The pharmaceutically acceptable carriers useful in this disclosure are conventional. Remington 's Pharmaceutical Sciences, by E. W. Martin, Mack Publishing Co., Easton, PA, 15^th Edition (1975), describes compositions and formulations suitable for pharmaceutical delivery of humanized COL-1 monoclonal antibodies disclosed herein. In general, the nature of the carrier will depend on the particular mode of administration employed. For instance, parenteral formulations usually comprise injectable fluids that include pharmaceutically and physiologically acceptable fluids such as water, physiological saline, balanced salt solutions, aqueous dextrose, glycerol or the like as a vehicle. For solid compositions (for example, powder, pill, tablet, or capsule forms), conventional non-toxic solid carriers can include, for example, pharmaceutical grades of mannitol, lactose, starch, or magnesium stearate. In addition to biologically-neutral carriers, pharmaceutical compositions to be administered can contain minor amounts of non-toxic auxiliary substances, such as wetting or emulsifying agents, preservatives, and pH buffering agents and the like, for example sodium acetate or sorbitan monolaurate. Polynucleotide: A single-stranded linear nucleotide sequence, including sequences of greater than 100 nucleotide bases in length. The term polynucleotide is specifically intended to cover naturally occurring polynucleotides, as well as those that are recombinantly or synthetically produced. A substantially purified polynucleotide, as used herein, refers to a polynucleotide that is substantially free of other polynucleotides or other materials with which it is naturally associated. In one embodiment, the polynucleotide is at least 50%, for example at least 80% free of other polynucleotides or other materials with which it is naturally associated. In another embodiment, the polynucleotide is at least 90% free of other polynucleotides or other materials with which it is naturally associated. In yet another embodiment, the polynucleotide is at least 95% free of other polynucleotides or other materials with which it is naturally associated. Polypeptide: A polymer in which the monomers are amino acid residues that are joined together through amide bonds. When the amino acids are alpha-amino acids, either the L-optical isomer or the D-optical isomer can be used, the L-isomers being prefened in nature. The term polypeptide or protein as used herein encompasses any amino acid sequence and includes, but may not be limited to, modified sequences such as glycoproteins. The term polypeptide is specifically intended to cover naturally occurring proteins, as well as those that are recombinantly or synthetically produced. Substantially purified polypeptide as used herein refers to a polypeptide that is substantially free of other proteins, lipids, carbohydrates or other materials with which it is naturally associated. In one embodiment, the polypeptide is at least 50%, for example at least 80% free of other proteins, lipids, carbohydrates or other materials with which it is naturally associated. In another embodiment, the polypeptide is at least 90% free of other proteins, lipids, carbohydrates or other materials with which it is naturally associated. In yet another embodiment, the polypeptide is at least 95% free of other proteins, lipids, carbohydrates or other materials with which it is naturally associated. Conservative amino acid substitution tables providing functionally similar amino acids are well known to one of ordinary skill in the art. The following six groups are examples of amino acids that are considered to be conservative substitutions for one another:

1) Alanine (A), Serine (S), Threonine (T); 2) Aspartic acid (D), Glutamic acid (E); 3) Asparagine (N), Glutamine (Q); 4) Arginine (R), Lysine (K); 5) Isoleucine (I), Leucine (L), Methionine (M), Valine (V); and 6) Phenylalanine (F), Tyrosine (Y), Tryptophan (W).

A non-conservative amino acid substitution can result from changes in: (a) the structure of the amino acid backbone in the area of the substitution; (b) the charge or hydrophobicity of the amino acid; or (c) the bulk of an amino acid side chain.

Substitutions generally expected to produce the greatest changes in protein properties are those in which: (a) a hydrophilic residue is substituted for (or by) a hydrophobic residue; (b) a proline is substituted for (or by) any other residue; (c) a residue having a bulky side chain, for example, phenylalanine, is substituted for (or by) one not having a side chain, for example, glycine; or (d) a residue having an electropositive side chain, for example, lysyl, arginyl, or histadyl, is substituted for (or by) an electronegative residue, for example, glutamyl or aspartyl. Variant amino acid sequences may, for example, be 80, 90 or even 95 or 98% identical to the native amino acid sequence. Programs and algorithms for determining percentage identity can be found at the NCBI website. Preventing or treating a disease: Preventing a disease refers to inhibiting completely or in part the development or progression of a disease, for example in a person who is known to have a predisposition to a disease, such as colorectal cancer, breast, ovarian, or prostate cancer. An example of a person with a known predisposition is someone with a history of cancer in the family, or who has been exposed to factors that predispose the subject to the development of a tumor. Treating a disease refers to a therapeutic intervention that inhibits, or suppressed the growth of a tumor, eliminates a tumor, ameliorates at least one sign or symptom of a disease or pathological condition, or interferes with a pathophysiological process, after the disease or pathological condition has begun to develop. Protein: A biological molecule encoded by a gene and comprised of amino acids. Recombinant: A recombinant nucleic acid is one that has a sequence that is not naturally occurring or was made artificially. Artificial combination is often accomplished by chemical synthesis or, more commonly, by the artificial manipulation of isolated segments of nucleic acids, for example, by genetic engineering techniques. Similarly, a recombinant protein is one encoded by a recombinant nucleic acid molecule. Relative binding affinity constant (Ka): Affinity of an antibody for an antigen can be expressed relative to the binding affinity of another antibody to the same antigen. In several embodiments, the relative affinity constant of a variant humanized COL-1 antibody is less than, similar to, or greater than, that of a murine COL-1 (mCOL-1), a parental humanized COL-1 antibody (HuCOL- 1), or HuCOl-l_AbrcD_R- In one embodiment, affinity is calculated by a modification of the Scatchard method described by Frankel et al, Mol. Immunol, 16:101, 1979. One of skill in the art can readily identify a statistical test that determines a statistically significant result for example, the Student's t-test, the Wilcoxon two sample test, or the Median test. In one embodiment, an antibody of interest retains antigen binding affinity, compared to a parental antibody, if the antibody binds the antigen and has a relative binding affinity constant about the same as the parental antibody. For example, if the parental antibody binds the antigen with an affinity constant of about 1.0 x 10^"8 M, the antibody can have a relative binding affinity constant of about 0.5 x 10^"8 M to about 5 x 10^"8 M, or about 1.5 x 10^"8 M. about 2 x 10^"8 M, about 2.5 x 10^"8 M, about 3 x 10^"8 M, about 3.5 x 10^" M or about 4 x 10^" M. In another example, if the parental antibody binds the antigen with an affinity constant of about 1.0 x 10^"7 M, the antibody can have a relative binding affinity constant of about 0.5 x 10^"7 M to about 5 x 10^"7 M, or about 1.5 x 10^"7 M, about 2 x 10^"7 M, about 2.5 x 10^"7 M, about 3 x 10^"7 M, about 3.5 x 10^"7 M or about 4 x lO^"7 M. In one embodiment, if the variant is a variant humanized COL-1 antibody, affinity can be compared to HuCOL- 1. In one example, the variant humanized COL-1 antibody binds CEA and has a relative binding affinity constant at least about 1.8 x 10^"8 M. In other embodiments, a variant humanized COL-1 antibody retains CEA binding affinity, compared to HuCOL- 1, if the relative binding affinity constant is at least about 2.0 x 10^"8, about 2.2 x 10^"8, about 2.4 x 10^"8, about 2.6 x 10^"8, about 2.8 x 10^"8, about 2.9 x 10^"8, about 3.0 x 10^"8, about 3.2 x 10^"8, about 3.5 x 10^"8, about 4.0 x 10^"8, about 4.5 x 10^"8, or about 5.0 10^"8 M. In another embodiment, a binding affinity is measured by an antigen/antibody dissociation rate. In yet another embodiment, a binding affinity is measured by a competition radioimmunoassay. In a further embodiment, a binding affinity is measured by flow cytometry as the number of gated cells labeled with the antibody, such as the HuCOL- 1 antibody. Selectively hybridize: Hybridization under moderately or highly stringent conditions that excludes non-related nucleotide sequences. In nucleic acid hybridization reactions, the conditions used to achieve a particular level of stringency, will vary depending on the nature of the nucleic acids being hybridized. For example, the length, degree of complementarity, nucleotide sequence composition (for example, GC versus AT content), and nucleic acid type (for example, RNA versus DNA) of the hybridizing regions of the nucleic acids can be considered in selecting hybridization conditions. An additional consideration is whether one of the nucleic acids is immobilized, for example, on a filter. A specific, non-limiting example of progressively higher stringency conditions is as follows: 2 x SSC/0.1% SDS at about room temperature (hybridization conditions); 0.2 x SSC/0.1%) SDS at about room temperature (low stringency conditions); 0.2 x

SSC/0.1% SDS at about 42°C (moderate stringency conditions); and 0.1 x SSC at about 68°C (high stringency conditions). One of skill in the art can readily determine variations on these conditions (for example, Molecular Cloning: A Laboratory Manual, 2nd ed., vol. 1-3, ed. Sambrook et al, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY,2001). Washing can be carried out using only one of these conditions, for example, high stringency conditions, or each of the conditions can be used, for example, for 10-15 minutes each, in the order listed above, repeating any or all of the steps listed. However, as mentioned above, optimal conditions will vary, depending on the particular hybridization reaction involved, and can be determined empirically. Sequence homology: The similarity between amino acid or nucleic acid sequences expressed in terms of the percentage identity between the sequences. The higher the percentage, the more homologous the two sequences are. Homologs of an antibody sequence will possess a significant degree of sequence identity when aligned using standard methods. Methods of alignment of sequences for comparison are well known in the art. Various programs and alignment algorithms are described in: Smith and Waterman, Adv. Appl. Math. 2:482, 1981; Needleman and Wunsch, J. Mol Biol. 48:443, 1970; Pearson and Lipman, Proc. Natl. Acad. Sci. U.S.A. 85:2444, 1988; Higgins and Shaφ, Gene 73:237, 1988; Higgins and Shaφ, CABIOS 5:151, 1989; Coφet et al, Nucleic Acids Research 16:10881, 1988; and Pearson and Lipman, Proc. Natl. Acad. Sci. U.S.A. 85:2444, 1988. Altschul et al, Nature Genet, 6:119, 1994 presents a detailed consideration of sequence alignment methods and identity calculations. The NCBI Basic Local Alignment Search Tool (BLAST) (Altschul et al. , J. Mol.

Biol. 215:403, 1990) is available from several sources, including the National Center for Biotechnology Information (NCBI, Bethesda, MD) and on the Internet, for use in connection with the sequence analysis programs blastp, blastn, blastx, tblastn, and tblastx. A description of how to determine sequence identity using this program is available on the NCBI website on the internet. Homologs of antibody sequences are typically characterized by possession of at least 50%, for example at least about 60%, 70%, 80%, 90%, 95%, 96%, 97%, 98% or 99%, sequence identity counted over the full length alignment with a reference sequence, for example a human germline sequence. The term "identical" in the context of an amino acid sequence or a nucleic acid sequence indicates that the sequence has 100% sequence identity to the reference sequence. Alignment can be performed using the NCBI Blast 2.0, gapped blastp set to default parameters. The comparison between the sequences is made over the full length alignment with the amino acid sequence given in this present disclosure, employing the Blast 2 sequences function using the default BLOSUM62 matrix set to default parameters, (gap existence cost of 11, and a per residue gap cost of

1). When aligning short peptides (fewer than around 30 amino acids), the alignment cand be performed using the Blast 2 sequences function, employing the PAM30 matrix set to default parameters (open gap 9, extension gap 1 penalties). Proteins with even greater similarity to the reference sequences will show increasing percentage identities when assessed by this method, such as at least about 50%), at least about 60%, at least about 70%, at least about 80%, at least about 90%, at least about 95%, at least about 96%, at least about 97%, at least about 98%, or at least about 99% sequence identity. When less than the entire sequence is being compared for sequence identity, homologues and, variants will typically possess at least 50% sequence identity over short windows of 10-20 amino acids, and may possess sequence identities of at least 50%, or at least 60%, 70%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, or 99% depending on their similarity to the reference sequence. Methods for determining sequence identity over such short windows are available at the NCBI website on the internet. One of skill in the art will appreciate that these sequence identity (homology) ranges are provided for guidance only; it is entirely possible that strongly significant homologues could be obtained that fall outside of the ranges provided. Therapeutically effective amount: A quantity of a specific substance sufficient to achieve a desired effect in a subject being treated. For instance, this can be the amount necessary to inhibit or suppress growth of a tumor. In one embodiment, a therapeutically effective amount is the amount necessary to reduce the growth of or eliminate a tumor, or to reduce a sign or symptom of the tumor. When administered to a subject, a dosage will generally be used that will achieve target tissue concentrations (for example, in tumors) that has been shown to achieve a desired in vitro effect. Treatment: Refers to both prophylactic inhibition of initial infection or disease, and therapeutic interventions to alter the natural course of an untreated infection or disease process, such as a tumor growth. Tumor: A neoplasm that may be either malignant or non-malignant. Tumors of the same tissue type are primary tumors originating in a particular organ (such as colorectal, breast, gastrointestinal, prostate, ovarian, bladder or lung). Tumors of the same tissue type may be divided into tumor of different sub-types (a classic example being bronchogenic carcinomas (lung tumors) which can be an adenocarcinoma, small cell, squamous cell, or large cell tumor). Breast cancers can be divided histologically into scinhous, infiltrative, papillary, ductal, medullary and lobular. In one embodiment, cells in a tumor express CEA. Variable region (also variable domain or V domain): The regions of both the light-chain and the heavy-chain on an Ig that contain antigen-binding sites. The regions are composed of polypeptide chains containing four relatively invariant "framework regions" (FRs) and three highly variant "hypervariable regions" (HVs). Because the HVs constitute the binding site for antigen(s) and determine specificity by forming a surface complementarity to the antigen, they are more commonly termed the "complementarity-determining regions," or CDRs, and are denoted CDRl, CDR2, and CDR3. Because both of the CDRs from the heavy- and light-chain domains contribute to the antigen-binding site, it is the three-dimensional combination of the heavy and the light chain that determines the final antigen specificity. Within the heavy- and light-chain, the framework regions sunound the CDRs. Proceeding from the N-terminus of a heavy or light chain, the order of regions is: FR1- CDRl -FR2-CDR2-FR3-CDR3-FR4. As used herein, the term "variable region" is intended to encompass a complete set of four framework regions and three complementarity-determining regions. Thus, a sequence encoding a "variable region" would provide the sequence of a complete set of four framework regions and three complementarity-determining regions. Variant Humanized COL-1: A humanized COL-1 antibody that has been further modified by introducing at least one amino acid substitution, and specifically binds CEA. A humanized COL-1 antibody can have at most 2, at most 3, at most 4, at most 5, at most 7, at most 10, at most 15, or more amino acid substitutions. Specific, non-limiting examples of variant humamzed COL-1 antibodies include SV1, SV2, SV3, S V4, SV5, and HuCOL- 1 _SDR.

Unless otherwise explained, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The singular terms "a," "an," and "the" include plural referents unless context clearly indicates otherwise. Similarly, the word "or" is intended to include "and" unless the context clearly indicates otherwise. It is further to be understood that all base sizes or amino acid sizes, and all molecular weight or molecular mass values, given for nucleic acids or polypeptides are approximate, and are provided for description. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incoφorated by reference in their entirety. In case of conflict, the present specification, including explanations of terms, will control. In addition, the materials, methods, and examples are illustrative only and not intended to be limiting.

Methods of Generating Germline Sequence Variants of Humanized Antibodies The humanization of monoclonal antibodies by CDR grafting onto human antibody templates (thereby utilizing the human framework region) is a known means to improve the clinical utility of xenogeneic (non-human) antibodies by reducing the human anti-murine antibody (HAMA) response elicited in patients, without compromising antigen binding affinity of the parental monoclonal antibody. However, CDR-grafted humanized antibodies may still evoke anti-variable (V) region response when administered to humans. To minimize anti-V region responses, the antibody can be humanized by grafting onto the human templates only the SDRs, the residues that are essential for the surface complementarity of the antibody and its ligand. Generally, humanized antibodies are generated using a single human antibody sequence as a template for either the V domain or the V_H domain of a xenogeneic antibody of interest, regardless of whether the humanized antibody is generated by CDR or SDR grafting. An alternative strategy to designing humanized antibodies is disclosed herein. The methods disclose herein utilize an analysis of the homology between the human template sequences and the xenogeneic monoclonal antibody to be humanized (see FIGS. 7A-7B). Generally, the homology is maximized by comparing the sequences of the segments (CDR and framework) of the variable domain of a single xenogeneic antibody of interest to a plurality of human germline sequences. The method can use a comparison of each CDR and the framework region. However, the method can include a comparison of a subset of the CDRs and/or the framework region. In one example, using this strategy, the human gennline sequence that is most homologous, but not identical to, each segment is identified. In this manner, humanized antibodies can be designed using as many as eight different human germline sequences as templates, one for each of eight different regions: the L-CDR1, L-CDR2, L-CDR3 (or conesponding SDRs), and the variable light chain framework of the variable light chain and the H-CDR1, H-CDR2, H-CDR3 (or conesponding SDRs) and variable heavy chain framework of the variable heavy chain. Fewer human germline sequences can be used if the same human germline sequence is most homologous to more than one variable domain segment from the same human light chain or human heavy chain. All of the L-CDR (1-CDRl , L-CDR2, L-CDR3), H-CDR (H-CDR1 , H-CDR2,

H-CDR3), light chain framework, and heavy chain framework sequences can be compared to a plurality of human germline sequences. Alternatively, a subset of the L- CDR, H-CDR, light chain framework, and heavy chain framework sequences can be compared to the human germline sequences. For example, at least one of the L-CDR, H-CDR, light chain framework, or heavy chain framework sequences can be compared to a plurality of human germline sequences. This strategy is advantageous in reducing the immunogenicity of the humanized antibody since, by using this method, humanization of an antibody can be optimized independently for each segment of the variable domain of the antibody. In one embodiment, the method includes a comparison of at least one L-CDR sequence (L-CDR1, L-CDR2, and or L-CDR3 sequences, or conesponding SDRs) and/or a light chain framework sequence from a single xenogeneic (non-human) antibody with a plurality of human light chain germline sequences, in order to identify the portion of the light chain human germline sequences most homologous to the xenogeneic L-CDR sequence(s) and/or the light chain framework sequence. It is known that the CDRs of the light chain are bounded by the residues at positions 24 and 34 (L- CDR1), 50 and 56 (L-CDR2), 89 and 97 (L-CDR3), using the numbering convention delineated by Kabat et al. (Sequences of Proteins of Immunological Interest, 5^th Edition, Department of Health and Human Services, Public Health Service, National Institutes of Health, Bethesda, NIH Publication No. 91-3242, 1991). Thus, each of these sequences can be obtained from the antibody of interest and compared to a plurality of human germline sequences. A plurality of human germline sequences can include collections of at least 20, 50, 100, 150, 200, or more human germline sequences. Examples of collections of human gennline sequences include those listed in Tomlinson et al. (J. Mol. Biol, 227:776-798, 1992, and Cox et al, Eur. J. Immunol, 24:827-836, 1994, which are incoφorated herein by reference). Comparisons of homology can be made using any known method. For example, L-CDR1, L-CDR2, L-CDR3 sequences and or a light chain framework sequence of an antibody of interest can be compared to a plurality of human germline sequences and a human homolog selected that is at least 50%, for example at least about 60%, 70%, 80%, 90%, 95%, 96%, 97%, 98% or 99%, homologous to each of the L-CDR1, L-CDR2, L-CDR3 (or conesponding SDRs) and a light chain framework sequences from the antibody of interest. The selected homologous sequences generally are from different human germline sequences, but in certain examples, one, two, three or four of the homologous sequences can be from the same human germline sequence. The selected homologous sequences are not identical to the respective portion of the antibody of interest. Alignment can be perfooned using any known method, such as by using the NCBI Blast 2.0 program, gapped blastp set to default parameters (see above). When aligning short peptides (fewer than about 30 amino acids), the alignment can be performed using the Blast 2 sequences function, employing the PAM30 matrix set to default parameters (open gap 9, extension gap 1 penalties). The comparison between the sequences is generally made over the full length alignment with the selected amino acid sequence of interest (for example, the L- CDRl, L-CDR2, L-CDR3 or framework region). Additional methods are disclosed below. Thus, a portion of a first human germline light chain sequence that is most homologous, but not identical, to a first xenogeneic L-CDR sequence is identified. Subsequently, a portion of a second human germline light chain sequence that is most homologous, but not identical, to a second xenogeneic L-CDR sequence can be identified. A portion of a third human germline light chain sequence that is most homologous, but not identical, to a third xenogeneic L-CDR sequence can also be identified. In addition, a portion of a fourth human germline light chain sequence that is most homologous, but not identical, to the light chain framework sequence can be identified. The first, second, third, or fourth light chain human germline sequences can be different germline sequences, or two or more of the human germline sequences can be the same geonline sequence. In another embodiment, the method involves a comparison of at least one H- CDR sequence (H-CDRl , H-CDR2, H-CDR3 sequences or conesponding SDRs) or a heavy chain framework sequence from a single xenogeneic (non-human) antibody with a plurality of human heavy chain gennline sequences, in order to identify the heavy chain human gennline sequences most homologous to each xenogeneic H-CDR sequence and the heavy chain framework sequence. The CDRs of the heavy chain are bounded by the residues at positions 31 and 35b (H-CDRl), 50 and 65 (H-CDR2), 95 and 102 (H-CDR3), using the numbering convention delineated by Kabat et al. (supra). Thus, each of these sequences can be obtained from the antibody of interest and compared to a plurality of human germline sequences. For example, H-CDRl, H-CDR2, H-CDR3 sequences and a heavy chain framework sequence of an antibody of interest can be compared to a plurality of human germline sequences (see above) and a human homolog selected that is at least 50%, for example at least about 60%, 70%, 80%, 90%, 95%, 96%, 97%, 98% or 99%, homologous to each of the H-CDRl, H-CDR2, H-CDR3 (or conesponding SDR) and a heavy chain framework sequence from the antibody of interest. The selected homologous sequences generally are from different human gennline sequence, but one, two, three, or four of the homologous sequences can be from the same human germline sequence. The selected homologous sequences are not identical to the respective portion of the antibody of interested. Alignment against can be performed using any known method (see above). Thus, a first human germline heavy chain sequence that is homologous, but not identical, to a portion of a first xenogeneic H-CDR sequence is identified. Subsequently, a portion of a second human germline light chain sequence that is homologous, but not identical, to a second xenogeneic H-CDR sequence can be identified. A portion of a third human germline light chain sequence that is homologous, but not identical, to a third xenogeneic H-CDR sequence can also be identified. In addition, a portion of a fourth human germline light chain sequence that is most homologous, but not identical, to the heavy chain framework sequence can be identified. The first, second, third, or fourth heavy chain human germline sequences can be different germline sequences, or two or more of the human germline sequences can be the same germline sequence. In a further embodiment, the method involves a comparison of at least one of L- CDR sequences (L-CDRl, L-CDR2, L-CDR3) or conesponding SDRs and a light chain framework sequence from a single xenogeneic antibody, with a plurality of human light chain germline sequences, and a comparison of at least one of H-CDR sequences (H- CDR1, H-CDR2, H-CDR3) or conesponding SDRs and a heavy chain framework sequence from the same xenogeneic antibody, with a plurality of human heavy chain germline sequences, in order to identify the portion of the light chain human germline sequences most homologous to each xenogeneic L-CDR sequence and the light chain framework sequence, and the portion of the heavy chain human germline sequences most homologous to each xenogeneic H-CDR sequence and the heavy chain framework sequence. The heavy chain comparison and the light chain comparisons can be performed in any order. Once the portion of the human germline sequences are identified that are homologous but not identical to one or more portions (such as L-CDRl, L-CDR2, L- CDR3, light chain framework, H-CDRl, H-CDR2, H-CDR3, heavy chain framework) of the antibody of interest, one or more of these sequences is selected in order humanize the xenogeneic antibody of interest. For example, one, two, or three of the identified human germline light chain sequences homologous to the conesponding xenogeneic L- CDR sequences can be selected. Similarly, one, two, or three human germline heavy chain sequences homologous to the conesponding xenogeneic H-CDR sequences can be selected. In addition, one or both of the identified human gennline light chain and heavy chain framework sequences can be selected. Generally, if a human gennline sequence is not selected for each CDR and framework region, the conesponding xenogeneic (non-human) sequence of the antibody to be humanized is utilized. Thus, the humanized antibody can include any combination of xenogeneic and human gennline variable domain sequences. For example, if a human L-CDRl, L-CDR2, L-CDR2, human light chain framework and human heavy chain framework are selected, the H-CDRl, H-CDR2 and H-CDR3 from the xenogeneic antibody of interest are utilized. Similarly, if a human H-CDRl, H- CDR2, H-CDR2, human light chain framework and human heavy chain framework are selected, the L-CDRl, L-CDR2 and L-CDR3 from the xenogeneic antibody of interest are utilized. In another example, if a human L-CDRl, L-CDR2, H-CDR3, and heavy chain framework are selected, the L-CDR3 , H-CDRl , H-CDR2, and light chain framework from the xenogeneic antibody of interest are utilized. In one example, selected variable light chain sequences (L-CDRl, L-CDR2, L- CDR3, and the light chain framework) are assembled into a light chain antibody sequence. In another example, the selected variable heavy chain sequences (H-CDRl, H-CDR2, H-CDR3, and the heavy chain framework) are assembled into a heavy chain antibody sequence. A heavy chain sequence and a light chain sequence are utilized to produce an antibody that binds the antigen of interest. It should be noted that the methods disclosed herein can be used to produce a heavy chain antibody sequence, a light chain antibody sequence, or both the heavy chain and the light chain antibody sequence of an antibody that binds the antigen of interest. The methods described above use sequence comparisons based on the amino acid sequences of both a xenogeneic (non-human) antibody and human antibodies. However, the genetic code is well known. Thus, it will be clear to those of skill in the art that the method can be readily modified to utilize a comparison of nucleic acid sequences. Additional manipulations to the light chain and/or heavy chain sequences can be used to retain antigen binding or to further increase antigen binding of the antibody and/or to further reduce immunogenicity of the antibody in subjects, compared to a parental antibody. In one embodiment, xenogeneic (non-human) CDR and framework residues are identified that differ from the residues at the conesponding positions (according to the numbering convention delineated by Kabat et al. (supra)) in the most homologous human germline sequence. Identified residues that are further determined to not be involved in antigen binding (non-SDRs or non-essential framework residues) can be substituted with a different amino acid to further reduce immunogenicity of the humanized antibody in subjects. For example, residue at a position that conesponds to a non-SDR or non-essential framework residue can be substituted with the human germline residue at the conesponding Kabat position, or with a different amino acid. In some embodiments, residues at positions that conespond to SDRs or to essential framework residues are substituted with the human germline residue at the conesponding Kabat position, or with a different amino acid. In another embodiment, the essential xenogeneic residues (SDRs or essential framework residues) are grafted onto the conesponding positions (according to the numbering convention delineated by Kabat et al. (supra)) of a light chain sequence or a heavy chain sequence assembled from a combination of different human germline sequences, where each human germline sequence is most homologous to its conesponding xenogeneic CDR or framework region. The process of grafting murine SDRs onto a human template is commonly refened to as SDR-grafting. Without being bound by theory, grafting essential xenogeneic framework residues maintains the advantage that the structural features of the xenogeneic frameworks are preserved while the deviation between the xenogeneic framework sequence and the human germline sequence is minimized. The cunent general approach to humanization involves choosing the human antibody template which displays the closest overall sequence identity to the xenogeneic antibody to be humanized. This approach maximized the homology in the frameworks between the xenogeneic and human antibody sequences at the expense of the CDRs, because the highest homology is found in the frameworks. In contrast, the disclosed method to design humanized antibodies is such that the choice of the human template is neither exclusively framework-centered nor CDR-centered, thereby maximizing the homology between the human germline sequences and the xenogeneic monoclonal antibody to be humanized for each CDR and framework region (and reducing immunogenicity of the antibody as compared to a parental antibody), while retaining antigen binding activity. The methods disclosed herein of humanizing antibodies do not involve the use of phage display or CDR shuffling as described in U.S. Patent Nos. 5,565,332 and 6,225,447. Phage display is a technique in which phage are engineered to fuse a foreign peptide or protein with their capsid (surface) proteins and hence display it on their cell surfaces. The immobilized phage may then be used as a screen to identify what ligands bind to the expressed fusion protein exhibited (displayed) on the phage surface. CDR shuffling refers to artificially reananging CDRs by mixing different CDR sequences amplified by PCR, and assembling them into a single gene product. As noted above, when comparing human germline sequences and xenogeneic

(non-human) sequences, the term "homologous" in the context of an amino acid sequence or a nucleic acid sequence indicates that the sequence comprises at least 50% sequence identity, such as at least 60%, 70%, 80%, 85%, 90%, 95%, 96%, 97%, 98% or 99% sequence identity to the reference sequence over a comparison window of, for example a sequence encoding at least 10, 20, 40, 60, 80, or more amino acids. The term "identical" in the context of an amino acid sequence or a nucleic acid sequence indicates that the sequence has 100% sequence identity to the reference sequence. The term "most homologous" indicates that the sequence of interest shares the greatest homology with the selected germline sequence, as compared to the other germline sequences used in the comparison. In one embodiment, the germline sequence most homologous (or having highest homology) to a specified light chain or heavy chain CDR or framework region can be selected. The percentage of sequence identity is determined by comparing two optimally aligned sequences over a comparison window, wherein the portion of the sequence in the comparison window may comprise additions or deletions (for instance, gaps) as compared to the reference sequence (which does not comprise additions or deletions) for optimal alignment of the two sequences. The percentage is calculated by determining the number of positions at which the identical nucleic acid base or amino acid residue occurs in both sequences to yield the number of matched positions, dividing the number of matched positions by the total number of positions in the window of comparison and multiplying the result by 100 to yield the percentage of sequence identity. For sequence comparison of nucleic acid sequences, typically one sequence acts as a reference sequence, to which test sequences are compared. When using a sequence comparison algorithm, test and reference sequences are entered into a computer, subsequence coordinates are designated, if necessary, and sequence algorithm program parameters are designated. Default program parameters are used. Methods of alignment of sequences for comparison are well known in the art. Optimal alignment of sequences for comparison can be conducted, for example, by the local homology algorithm of Smith and Waterman, (Adv. Appl. Math. 2:482, 1981), by the homology alignment algorithm of Needleman and Wunsch, (J. Mol. Biol. 48:443, 1970), by the search for similarity method of Pearson and Lipman, (Proc. Nat 'I. Acad. Sci. USA 85:2444, 1988), by computerized implementations of these algorithms (GAP, BESTFIT, FASTA, and TFASTA in the Wisconsin Genetics Software Package,

Genetics Computer Group, 575 Science Dr., Madison, WI), or by manual alignment and visual inspection (see, for example, Current Protocols in Molecular Biology (Ausubel et al, eds 1995 supplement)). One example of a useful algorithm is PILEUP. PILEUP uses a simplification of the progressive alignment method of Feng and Doolittle (J. Mol. Evol 35:351-360, 1987). The method used is similar to the method described by Higgins and Shaφ (CABIOS 5:151-153, 1989). Using PILEUP, a reference sequence is compared to other test sequences to determine the percent sequence identity relationship using the following parameters: default gap weight (3.00), default gap length weight (0.10), and weighted end gaps. PILEUP can be obtained from the GCG sequence analysis software package, for example., version 7.0 (Devereaux et al, Nuc. Acids Res. 12:387-395, 1984). Another example of algorithms that are suitable for determining percent sequence identity and sequence similarity are the BLAST and the BLAST 2.0 algorithm, which are described in Altschul et al, J. Mol. Biol.215:403-410, 1990 and Altschul et al, Nucleic Acids Res. 25:3389-3402, 1977. Software for performing BLAST analyses is publicly available through the National Center for Biotechnology Information. The BLASTN program (for nucleotide sequences) uses as defaults a word length (W) of 11, alignments (B) of 50, expectation (E) of 10, M=5, N=-4, and a comparison of both strands. The BLASTP program (for amino acid sequences) uses as defaults a word length (W) of 3, and expectation (E) of 10, and the BLOSUM62 scoring matrix (see Henikoff and Henikoff, Proc. Natl. Acad. Sci. USA 89:10915, 1989). Using recombinant DNA techniques, the selected human germline and/or xenogeneic (non-human) sequences can be produced and assembled to form variable light and heavy chain sequences. Methods and plasmid vectors for producing the polynucleotides encoding humanized monoclonal antibodies and for expressing these polynucleotides in bacterial and eukaryotic cells are well known in the art, and specific methods are described in Sambrook et al. (In Molecular Cloning: A Laboratory

Manual, Ch. 17, CSHL, New York, 1989). DNA sequences can be manipulated with standard procedures such as restriction enzyme digestion, fill-in with DNA polymerase, deletion by exonuclease, extension by terminal deoxynucleotide transferase, ligation of synthetic or cloned DNA sequences, site-directed sequence-alteration via single- stranded bacteriophage intermediate or with the use of specific oligonucleotides in combination with polymerase chain reaction (PCR) or other in vitro amplification technique. A cDNA sequence (or portions derived from it) such as a cDNA encoding a humanized monoclonal antibody can be introduced into eukaryotic expression vectors by conventional techniques. These vectors are designed to permit the transcription of the cDNA in eukaryotic cells by providing regulatory sequences that initiate and enhance the transcription of the cDNA and ensure its proper splicing and polyadenylation. Vectors containing the promoter and enhancer regions of the cytomegalo virus (CMV), SV40, or long terminal repeat (LTR) of the Rous Sarcoma virus and polyadenylation and splicing signal from S V40 are readily available (Mulligan et al, Proc. Natl. Acad. Sci. USA 78:1078-2076, 1981; Gorman et al, Proc. Natl. Acad. Sci USA 78:6777-6781, 1982; Gonzales et al, Mol. Immunol, 40:337-349, 2003). The level of expression of the cDNA can be manipulated with this type of vector, either by using promoters that have different activities (for example, the baculovirus pAC373 can express cDNAs at high levels in S. frugiperda cells (Summers and Smith, In Genetically Altered Viruses and the Environment, Fields et al. (Eds.) 22:319-328, CSHL Press, Cold Spring Harbor, New York, 1985) or by using vectors that contain promoters amenable to modulation, for example, the glucocorticoid- responsive promoter from the mouse mammary tumor virus (Lee et al, Nature 294:228, 1982). The expression of the cDNA can be monitored in the recipient cells 24 to 72 hours after introduction (transient expression). In addition, some vectors contain selectable markers such as the gpt (Mulligan and Berg, Proc. Natl. Acad. Sci. USA 78:2072-2076, 1981) or neo (Southern and Berg, J. Mol. Appl. Genet. 1:327-341, 1982) bacterial genes. These selectable markers permit selection of transfected cells that exhibit stable, long-term expression of the vectors (and therefore the cDNA). The vectors can be maintained in the cells as episomal, freely replicating entities by using regulatory elements of viruses such as papilloma (Sarver et al, Mol. Cell Biol. 1:486, 1981) or Epstein-Ban (Sugden et al, Mol. Cell Biol 5:410, 1985). Alternatively, one can also produce cell lines that have integrated the vector into genomic DNA. Both of these types of cell lines produce the gene product on a continuous basis. One can also produce cell lines that have amplified the number of copies of the vector (and therefore of the cDNA as well) to create cell lines that can produce high levels of the gene product (Alt et al. , J. Biol. Chem. 253:1357, 1978). The transfer of DNA into eukaryotic, in particular human or other mammalian cells, is now a conventional technique. The vectors are introduced into the recipient cells as pure DNA (transfection) by, for example, precipitation with calcium phosphate (Graham and vander Eb, Virology 52:466, 1973) or strontium phosphate (Brash et al, Mol. Cell Biol. 7:2013, 1987), electroporation (Neumann et al, EMBOJ 1:841, 1982), lipofection (Feigner et al, Proc. Natl. Acad. Sci USA 84:7413, 1987), DEAE dex ran (McCufhan et al, J. Natl. Cancer Inst 41:351, 1968), microinjection (Mueller et al, Cell 15:579, 1978), protoplast fusion (Schafher, Proc. Natl. Acad. Sci. USA 77:2163-2167, 1980), or pellet guns (Klein et al, Nature 327:70, 1987). Alternatively, the cDNA, or fragments thereof, can be introduced by infection with virus vectors. Systems are developed that use, for example, retro viruses (Bernstein et al, Gen. Engr g 7:235, 1985), adenoviruses (Ahmad et al, J. Virol. 57:267, 1986), or Heφes virus (Spaete et al, Cell 30:295, 1982). Polynucleotides that encode proteins, such as humanized monoclonal antibodies, can also be delivered to target cells in vitro via non- infectious systems, for instance liposomes. Using the above techniques, the expression vectors containing a polynucleotide encoding a humanized monoclonal antibody (or the light or the heavy chain of an antibody), or fragments or variants or mutants thereof, can be introduced into human cells, mammalian cells from other species or non-mammalian cells as desired. The choice of cell is determined by the puφose of the treatment. For example, monkey COS cells (Gluzman, Cell 23:175-182, 1981) that produce high levels of the SV40 T antigen and permit the replication of vectors containing the SV40 origin of replication may be used. Similarly, Chinese hamster ovary (CHO), mouse NIH 3T3 fibroblasts or human fibroblasts or lymphoblasts can be used. Such recombinant humanized monoclonal antibodies can be made in large amounts and are easy to purify. Suitable methods are presented in Sambrook et al. (In Molecular Cloning: A Laboratory Manual, CSHL, New York, 1989) and are well known in the art. Humanized antibodies developed in this approach can be conjugated to effector molecules, such as radioactive molecules, toxins, drugs, fluorescent labels or other detectable materials for delivery to a target tissue (described in more detail below). An effector molecule can be coupled to an antibody by a linker. A "linker" is a molecule that joins two other molecules, either covalently, or through ionic, van der Waals or hydrogen bonds, for example, a nucleic acid molecule that hybridizes to one complementary sequence at the 5' end and to another complementary sequence at the 3' end, thus joining two non-complementary sequences. The humanized antibodies generated by the disclosed methods can then be tested for antigen reactivity of the antibody and for the level of production of the antibody by the transfected cells. Assays that can be used to test antigen reactivity and level of production of the humanized antibodies include the enzyme linked immunoassay (ELISA). Additional assays can be performed, such as competition radioimmunoassay, fluorescence activated cell sort (FACS) analysis, and competition assays in order to test binding activity of the xenogeneic monoclonal antibodies humanized by the disclosed methods. By identifying and selecting germline sequences most homologous to the xenogeneic sequences, the humanized monoclonal antibody is rendered minimally immunogenic (has reduced immunogenicity compared to a parental antibody), without compromising antigen binding activity, and would have a high rate of plasma clearance.

Thus, the antibodies are of use as therapeutic agents (such as for immunotherapy against tumors) or for canying out radioimmunoguided surgery. Minimally Immunogenic Germline Sequence Variants of Humanized COL-1 The following section describes humanized COL-1 antibodies. These antibodies were humamzed using the above method. Prior to the development of the disclosed method, reducing the immunogenicity of the COL-1 monoclonal antibody (humanization of the COL-1 monoclonal antibody) was undertaken by progressively reducing its murine content by genetic manipulation. For example, HuCOL- 1 was formed by grafting all six (three heavy chain and three light chain) monoclonal antibody COL-1 hypervariable regions (CDRs) onto the variable light (V_L) and variable heavy (V_H) frameworks of the human antibodies VJI'CL and MO30, respectively, while retaining only those murine framework residues that are required for the integrity of the antigen combining site structure. HuCOL- 1 was deposited with the ATCC as ATCC Accession Number PTA-4661. In another example, the HuCOL-1 _AbrcD_R antibody was formed by grafting partial or "abbreviated" CDRs onto the V_L and V_H frameworks of VJI'CL and MO30, respectively, while retaining murine framework residues that are determined to be required for the integrity of the antigen combining site structure. Specifically, an arginine at position 24 in the murine COL-1 L-CDRl was replaced with a lysine, an alanine at position 25 in the murine COL-1 L-CDRl was replaced with a serine, a lysine at position 27 in the murine COL-1 L-CDRl was replaced with a glutamine, and a proline at position 61 in the murine COL-1 H-CDR2 was replaced with a glutamine. HuCOL- brCDR, also refened to as HuCOL-l²⁴'²⁵'²⁷L/⁶¹H, and HuCOL-1 are described in De Pascalis et al, J Immunol, 169:3076-3084, 2002, U.S. Patent Application No. 60/408,703, and PCT Application No. PCT/US03/27976, all three of which are incoφorated herein by reference. These documents also disclose the amino acid sequences of COL-1 , MO30 and VJI'CL. Compared to mCOL-1, HuCOL-1 and HuCOL-l_AbrCDR showed decreased reactivity to the sera of patients who were earlier administered mCOL-1 in clinical trials (De Pascalis et al, J. Immunol, 169:3076-3084, 2002). However, HuCOL-1 had a relative binding affinity approximately 1.5-fold less than the mCOL-1 monoclonal antibody and HuCOL- 1 _AbrcDR had a relative binding affinity approximately 5 fold lower than the mCOL-1 monoclonal antibody. Thus, there was a need for an alternative method of generating humanized COL-1 antibodies exhibiting an even greater reduction in immunogenicity, but without compromising antigen binding affinity. The humanized COL-1 antibodies generated by the method disclosed herein bind CEA, include CDRs from mCOL-1 and framework regions from HuCOL- l_Abrc_DR, and contain one or more substitutions in the CDRs of the light and heavy chains, and in the framework regions of the light chain. As described above, in order to identify which residues are to be substituted, the sequences of each of the mCOL-1 CDRs and the HuCOL- 1 Abr_CD framework regions are compared to a plurality of human germline sequences, thereby identifying the most homologous human germline sequence for each CDR and framework region. Substitutions of specific residues can then be made at positions in the mCOL-1 CDRs and HuCOL- l_AbrcDR framework regions according to the numbering convention delineated by Kabat et al. (supra). The light chain and heavy chain CDRs of the mCOL-1 monoclonal antibody and

HuCOL- 1 humanized antibody have the following sequences: L-CDRl: RASKSVSASGYSYMH (SEQ ID NO: 9) L-CDR2: LASNLQS (SEQ ID NO: 10) L-CDR3: QHSRELPT (SEQ ID NO: 11) H-CDRl : DYYMH (SEQ ID NO: 15) H-CDR2: WIDPENGDTEYAPKFQG (SEQ ID NO: 16) H-CDR3: RGLSTMITTRWFFDV (SEQ ID NO: 17)

The light chain framework regions of the mCOL-1 monoclonal antibody have the following sequences: FRl: DIVLTQSPASLTVSLGLRATISC (SEQ ID NO: 20) FR2: WYQQRPGQPPKLLIY (SEQ ID NO: 21) FR3: GVPARFSGSGSGTDFTLNIHPVEEEDAATYYC (SEQ ID NO: 22) FR4: FGGGTKLEIK (SEQ ID NO: 23) The heavy chain framework regions of the mCOL-1 monoclonal antibody have the following sequences: FRl: EVQLQQSGAELVRSGASVKMSCTASGFNIK (SEQ ID NO: 24) FR2: WVKQRPEQGLEWIG (SEQ ID NO: 25) FR3: KATMTTDTSSNTAYLQLSSLTSEDTAVYYCNT (SEQ ID NO: 26) FR4: WGAGTTVAVSS (SEQ ID NO: 27)

The mCOL-1 monoclonal antibody variable light chain has the following sequence:

DIVLTQSPASLTVSLGLRATISCRASKSVSASGYSYMHWYQQRPGQPPK LLIYLASNLQSGVPARFSGSGSGTDFTLNIHPVEEEDAATYYCQHSRELPTFGGG TKLEIK (SEQ ID NO: 44).

The mCOL-1 monoclonal antibody variable heavy chain has the following sequence:

EVQLQQSGAELVRSGASVKMSCTASGFNIKDYYMHWVKQRPEQGLE WIGWIDPENGDTEYAPKFQGKATMTTDTSSNTAYLQLSSLTSEDTAVYYCNTR GLSTMITTRWFFDVWGAGTTVAVSS (SEQ ID NO: 45)

The light chain and heavy chain CDRs of the HuCOL- 1 _Abrc_DR antibody have the following sequences: L-CDRl : KSSQSVSASGYSYMH (SEQ ID NO: 50) L-CDR2: LASNLQS (SEQ ID NO: 10) L-CDR3: QHSRELPT (SEQ ID NO: 11) H-CDRl: DYYMH (SEQ ID NO: 15) H-CDR2: WIDPENGDTEYAQKFQG (SEQ ID NO: 51) H-CDR3: RGLSTMITTRWFFDV (SEQ ID NO: 17)

The light chain framework regions of the HuCOL- l_AbrcDR antibody have the following sequences: FRl : DIVLTQSPASLAVSLGERATINC (SEQ ID NO: 52) FR2: WYQQKPGQPPKLLIY (SEQ ID NO: 53) FR3: GVPARFSGSGSGTDFTLTISSVQAEDVATYYC (SEQ ID NO: 54) FR4: FGGGTKLEIK (SEQ ID NO: 23) The heavy chain framework regions of the HuCOL- 1 _AbrCD antibody have the following sequences: FRl: EVQLVQSGAEVVKPGASVKMSCKASGFNIK (SEQ ID NO: 55) FR2: WVKQAPGQGLEWIG (SEQ ID NO: 56) FR3: KATMTTDTSTSTAYLELSSLRSEDTAVYYCNT (SEQ ID NO: 57) FR4: WGAGTLVTVSS (SEQ ID NO: 58)

In one embodiment, the humanized COL-1 antibodies disclosed herein include a substitution of a non-ligand contact residue (a non-SDR residue) with a residue from the conesponding position, according to the numbering convention delineated by Kabat et al. (supra), of a homologous germline sequence. The non-ligand contact residue substituted can be in the mCOL-1 L-CDRl, L-CDR2, or L-CDR3. For example, the non-ligand contact residue can be at position 24, 25, 26, 27, 27a, 27b, 27c, or 33 of mCOL-1 L-CDRl, at position 51, or 54 of mCOL-1 L-CDR2, or at position 90 of mCOL-1 L-CDR3. In one example, the residue at position 27 of L-CDRl is substituted with a residue at the conesponding position, according to the numbering convention of Kabat et al. (supra), in the L-CDRl of the DPK22 germline sequence (SEQ ID NO: 12, Table I). Thus, the humanized COL-1 antibody has a glutamine at position 27 in the mCOL-1 L-CDRl. In another example, the residue at position 33 of L-CDRl is substituted with a residue at the conesponding position in the L-CDRl of the DPK22 gennline sequence. Thus, the humanized COL-1 antibody has a leucine at position 33 of the mCOL-1 L-CDRl. In a further example, the residue at position 27 of L-CDRl and the residue at position 33 of L-CDR2 are substituted with residues at the conesponding positions from two different human germline sequences. In yet another example, the residue at position 90 of L-CDR3 is substituted with a residue at the conesponding position in the L-CDR3 of the DPK9 gennline sequence (SEQ ID NO: 14, Table I). Thus, the humanized COL-1 antibody has a glutamine at position 90 of the mCOL-1 L- CDR3. In another embodiment, the non-ligand contact residue substituted can be in

H-CDRl, H-CDR2, or H-CDR3. For example, the non-ligand contact residue can be at position 34 of H-CDRl, position 59, 60, 61, 62, 63, 64, or 65 of H-CDR2, or position 102 of H-CDR3 from the mCOL-1 monoclonal antibody. In one example, the residue at position 61 of H-CDR2 is substituted with a residue at the cooesponding position, according to the numbering convention of Kabat et al. (supra), in the H-CDR2 of the DP-75 germline sequence (SEQ ID NO: 19, Table I). Thus, the humanized COL-1 antibody has a glutamine at position 61 of H-CDR2. In yet another embodiment, the humanized COL-1 antibody includes a substitution of a non-essential light chain or heavy chain framework residue with a residue from the conesponding position, according to the numbering convention of Kabat et al. (supra), of a homologous germline sequence. In one example, at least one of the non-essential residues in the light chain (see Fig. 6) is substituted with a residue at the conesponding position in the framework of the DPK24 gennline sequence (SEQ ID NOs: 28, 29, 30). In another example, at least one of the non-essential residues in the heavy chain (see Fig. 6) is substituted with a residue at the conesponding position in the framework of the DP-75 germline sequence (SEQ ID NOs: 32, 33, 34). In another embodiment, an SDR residue or an essential framework residue, is substituted. In one specific, non-limiting example, the residue at position 53 of L- CDR2 is substituted with a residue at the conesponding position in the L-CDR2 of the DPK5 germline sequence (SEQ ID NO: 13, Table I). Thus, the humanized COL-1 antibody has a serine at position 53 of the mCOL-1 L-CDR2. As described above, the light chain framework of HuCOL- l_AbrcDR was derived from the VJI'CL variable light chain and the heavy chain framework of HuCOL- l_AbrCDR 5 was derived from the MO30 variable heavy chain. Thus, a humanized COL-1 antibody light chain having HuCOL- l_AbrCDR light chain framework regions and mCOL-1 L- CDRs has the following sequence:

DΓVLTQSPASLAVSLGERATINCRASKSVSASGYSYMHWΎQQKPGQPP o KLLIYLASNLQSGVPARFSGSGSGTDFTLTISSVQAEDVATYYCQHSRELPT FGGGTKLEIK (SEQ ID NO: 42)

A humanized COL-1 heavy chain having HuCOL- lAbr_CDR heavy chain framework regions and mCOL-1 H-CDRs has the following sequence:5 EVQLVQSGAEVVKPGASVKMSCKASGFNIKDYYMHWVKQAPGQGL EWIGWIDPENGDTEYAPKFQGKATMTTDTSTSTAYLELSSLRSEDTAVYYCNT RGLSTMITTRWFFDVWGAGTLVTVSS (SEQ ID NO: 43) 0 In one example, the residue at position 27 of SEQ ID NO: 42 is substituted with the residue at position 27 of the amino acid sequence as set forth in SEQ ID NO: 36. In another example, the residue at position 37 of SEQ ID NO: 42 is substituted with the residue at position 34 of the amino acid sequence as set forth in SEQ ID NO: 36. In other examples, the residue at position 9 of SEQ ID NO: 42 is substituted with the5 residue at position 9 of the amino acid sequence as set forth in SEQ ID NO: 41, the residue at position 57 of SEQ ID NO: 42 is substituted with the residue at position 53 of the amino acid sequence as set forth in SEQ ID NO: 37, the residue at position 94 of SEQ ID NO: 42 is substituted with the residue at position 90 of the amino acid sequence as set forth in SEQ ID NO: 38, or the residue at position 62 of SEQ ID NO: 43 is substituted with the residue at position 62 of the amino acid sequence as set forth in SEQ ID NO: 40. In one specific, non-limiting example, the residue at position 27 of SEQ ID NO: 42 is substituted with a glutamine. In other specific, non-limiting examples, the residue at position 37 of SEQ ID NO: 42 is substituted with a leucine, the residue at position 57 of SEQ ID NO: 42 is substituted with a serine, or the residue at position 94 of SEQ ID NO: 42 is substituted with a glutamine. In yet another specific, non-limiting example, the residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid. In a further specific, non-limiting example, the residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine. The germline antibody sequences described above have the following sequences:

DPK24 (SEQ ID NO: 41 ; encoded by GenBank Accession No. X96340, herein incoφorated by reference)

DIVMTQSPDSLAVSLGERATINCKSSQSVLYSSNNKNYLAWYQQKPGQPPKLLI YWASTRESGVPDRFSGSGSGTDFTLTISSLQAEDVAVYYCQQYYSTPP

DPK22 (SEQ ID NO: 36; encoded by GenBank Accession No. X93639, herein incoφorated by reference)

EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSR ATGIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQYGSSPP

DPK5 (SEQ ID NO: 37; encoded by GenBank Accession No. X93623, herein incoφorated by reference)

DIQMTQSPSSVSASVGDRVTITCRASQGISSWLAWYQQKPGKAPKLLIYAASSL QSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQANSFPP DPK 9 (SEQ ID NO: 38; encoded by GenBank Accession No. X93627Jκ4, herein incoφorated by reference)

DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQ SGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSYSTPP

DP-7 (SEQ ID NO: 39; encoded by GenBank Accession No. Z12309, herein incoφorated by reference)

QVQLVQSGAEVKKPGASVKVSCKASGYTFTSYYMHWVRQAPGQGLEWMGII NPSGGSTSYAQKFQGRVTMTRDTSTSTVYMELSSLRSEDTAVYYCAR

DP-75 (SEQ ID NO: 40; encoded by GenBank Accession No. Z14071, herein incoφorated by reference)

QVQLVQSGAEVKKPGASVKVSCKASGYTFTGYYMHWVRQAPGQGLEWMGW INPNSGGTNYAQKFQGRVTMTRDTSISTAYMELSRLRSDDTAVYYCAR

The DPK24 light chain frameworks have the following sequences: FRl: DTVMTQSPDSLAVSLGERATINC (SEQ ID NO: 28) FR2: WYQQKPGQPPKLLIY (SEQ ID NO: 29) FR3: GVPDRFSGSGSGTDFTLTISSLQAEDVAVYYC (SEQ ID NO: 30)

The DP-75 heavy chain frameworks have the following sequences: FRl: QVQLVQSGAEVKKPGASVKVSCKASGYTFT (SEQ ID NO: 32) FR2: WVRQAPGQGLEWMG (SEQ ID NO: 33) FR3: RVTMTRDTSISTAYMELSRLRSDDTAVYYCAR (SEQ ID NO: 34)

Humanized COL-1 antibodies with mCOL-1 CDRs and HuCOL- lAbr_CDR framework regions can include combinations of the CDR and framework substitutions disclosed herein. In one embodiment, the humanized COL-1 antibody can include a substitution of the residue at position 61 of H-CDR2 and at position 9 of the light chain framework. In one example, the humanized COL-1 antibody includes substitutions of the residues at positions 27 and 33 of L-CDRl, at position 61 of H-CDR2, and at position 9 of the light chain framework region. In one specific, non-limiting example of a humanized COL-1 antibody, the residue at position 27 of SEQ ID NO: 42 is substituted with a glutamine, the residue at position 37 of SEQ ID NO: 42 is substituted with a leucine, the residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid, and the residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine. This humanized COL-1 antibody is termed SV2 (Table II). In another example, the humanized COL-1 antibody includes a substitution of the residue at position 53 of L-CDR2, in addition to the residues at positions 61 of H- CDR2 and at position 9 of the light chain framework. In one, non-limiting example of a humanized COL-1 antibody, the residue at position 57 of SEQ ID NO: 42 is substituted with a serine, the residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid, and the residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine. This humanized COL-1 antibody is termed SV4 (Table II). In yet another example, the humanized COL-1 antibody includes substitutions of the residue at position 90 of L-CDR3, in addition the residues at position 61 of H- CDR2 and position 9 of the light chain framework. In one, non-limiting example of a humanized COL-1 antibody, the residue at position 94 of SEQ ID NO: 42 is substituted with a glutamine, the residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid, and the residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine. This humanized COL-1 antibody is termed SV5 (Table II). In a further example, the humanized COL-1 antibody includes substitutions of the residues at positions 27 and 33 of L-CDRl, positions 53 of L-CDR2, positions 90 of L-CDR3, position 61 of H-CDR2, and position 9 of the light chain framework. In a specific, non-limiting example of a humanized COL-1 antibody, the residue at position 27 of SEQ ID NO: 42 is substituted with a glutamine, the residue at position 37 of SEQ ID NO: 42 is substituted with a leucine, the residue at position 57 of SEQ ID NO: 42 is substituted with a serine, the residue at position 94 of SEQ ID NO: 42 is substituted with a glutamine, the residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid, and the residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine. This humanized COL-1 antibody is termed HUCOL-I_SDR (Table II). The humanized COL-1 antibodies disclosed herein, such as SV2, SV4, SV5 and HuCOL- 1 _SDR, contain a reduced murine content, and consequently, reduced immunogenicity, when compared to HuCOL- 1 and HuCOL- l_AbrcD_R- Nonetheless, the humanized COL-1 antibodies of the invention have a binding affinity that is similar or is increased as compared to a parental antibody, such as HuCOL- 1. Thus, the humanized monoclonal antibodies disclosed herein bind CEA with a retained binding affinity, as compared to the HuCOL- 1 antibody. In one embodiment, the humanized COL-1 antibody retains a binding affinity for CEA that is at least about 1.8 x 10^"8 M. In other embodiments, the humanized COL-1 antibody retains a binding affinity for CEA that is at least about 2.0, x 10^"8,about 2.2, x 10^"8, about 2.4 x 10^"8, about 2.6 x 10^"8, about 2.8 x 10^"8, about 2.9 x 10^"8, about 3.0 x 10^"8, about 3.2 x 10^"8, about 3.5 x 10^"8, about 4.0 x 10^"8, about 4.5 x 10^"8, or about 5.0 x 10^"8 M. In one embodiment, the humanized

COL-1 antibody retains a binding affinity if it has a significantly lower antigen/antibody dissociation rate compared to that of the HuCOL- 1 antibody. In another embodiment, the humanized COL-1 antibody retains a binding affinity if less antibody is required for a 50%) inhibition of the binding of ¹²⁵I-labeled mCOL-1 to CEA compared to the HuCOL-1 and HuCOL-lAbrCDR antibodies, h yet another embodiment, the humamzed COL-1 antibody retains a binding affinity when the number of cells labeled with humanized COL-1 antibody is significantly greater than the number of cells labeled by the HuCOL- 1 and HuCOL- l_AbrCDR antibodies, as measured by flow cytometry. Immunogenicity of humanized COL-1 antibodies can be measured in a competitive binding assay as the ability of a humanized COL-1 antibody to prevent a mCOL-1, HuCOL-1, or HuCOL-l_Abrc_DR antibody from binding to anti-idiotypic antibodies in a human subject's serum. In one embodiment, the humamzed COL-1 antibody generates a reduced, for example low, immune response and is minimally immunogenic in a subject. In one embodiment, at least about two-fold higher molar concentration of the humanized COL-1 antibody, than that of the parental antibody, for example the HuCOL- 1 antibody, is required to elicit 50% inhibition of HuCOL- 1 antibody binding to its cognate anti-idiotypic antibodies in a subject's sera. In other embodiments, at least about three-fold, at least about five-fold, at least about ten-fold, at least about twenty-fold, at least about twenty-five-fold, at least about thirty-fold, at least about thirty-five-fold, at least about forty-fold, at least about fifty-fold, at least about seventy-fold, or at least about one hundred-fold higher molar concentration of the humanized COL-1 antibody, than that of the HuCOL-1 antibody, is required to elicit 50% inhibition of HuCOL- 1 antibody binding to its cognate anti-idiotypic antibodies in a subject's sera. In one embodiment, the humanized COL-1 antibody has a CH2 domain deletion (Slavin-Chiorini et al, Int. J. Cancer, 53:97-103, 1993; Slavin-Chiorini et al, Cancer Research, 55:5957s-5967s, 1995; Slavin-Chiorini et al, Cancer Biother. Radiopharm., 12:305-316, 1997, incoφorated herein by reference). The generation and characterization of CH2 domain deleted antibodies is described in Mueller et al, Proc. Natl. Acad. Sci. USA., 87:5702-5705, 1990. In one specific embodiment, a humanized COL-1 antibody with a CH2 domain deletion is cleared more quickly from the plasma compared to the murine COL-1 monoclonal antibody or a parental antibody, for example HuCOL- 1. In other specific embodiments, a humanized COL-1 antibody with a CH2 domain deletion has reduced immunogenicity compared to the murine COL-1 antibody or the HuCOL- 1 antibody. In yet other embodiments, a humanized COL-1 antibody with a CH2 domain deletion has reduced immunogenicity compared to the murine COL-1 antibody or the HuCOL- 1 antibody, and retains CEA binding affinity. Effector molecules, for example, therapeutic, diagnostic, or detection moieties, can be linked to a humanized COL-1 antibody that specifically binds CEA, using any number of means known to those of skill in the art. Thus, a humanized COL-1 antibody with an amino acid substitution can have any one of a number of different types of effector molecules linked to it. In addition, the antibody can be linked to an effector molecule by a covalent or non-covalent means. In one embodiment, the antibody is linked to a detectable label. In some embodiments, the antibody is linked to a radioactive isotope, an enzyme substrate, a chemotherapeutic drug, a co-factor, a ligand, a chemiluminescent agent, a fluorescent agent, a hapten, or an enzyme. In other embodiments, the antibody is linked to a cytokine. Specific, non-limiting examples of cytokines are IL-2, IL-4, IL-10, TNF-alpha and IFN-gamma. In yet other embodiments, the antibody is linked to a cytotoxin, such as a bacterially-expressed toxin, a virally- expressed toxin, or a venom protein, to yield immunotoxins. Specific, non-limiting examples of cytotoxins include ricin, abrin, Pseudomonas exotoxin (PE), diphtheria toxin and subunits thereof, as well as botulinum toxins A through F. These toxins are readily available from commercial sources (for example, Sigma Chemical Company, St. Louis, MO). Diphtheria toxin is isolated from Corynebacterium diphtheriae. Ricin is the lectin RCA60 from Ricinus communis (Castor bean). The term also references toxic variants thereof. For example, see, U.S. Patent No. 5,079,163 and U.S. Patent No. 4,689,401. Ricinus communis agglutinin (RCA) occurs in two forms designated RCA₆o and RCA₁₂₀ according to their molecular weights of approximately 65 and 120 kDa respectively (Nicholson & Blaustein, J. Biochim. Biophys. Ada 266:543, 1972). The A chain is responsible for inactivating protein synthesis and killing cells. The B chain binds ricin to cell-surface galactose residues and facilitates transport of the A chain into the cytosol (Olsnes et al, Nature 249:627-631 , 1974 and U.S. Patent No. 3,060, 165). Abrin includes toxic lectins from Abrus precatorius. The toxic principles, abrin a, b, c, and d, have a molecular weight of from about 63 and 67 kDa and are composed of two disulfide-linked polypeptide chains A and B. The A chain inhibits protein synthesis; the B-chain (abrin-b) binds to D-galactose residues (see, Funatsu, et al, Agr. Biol. Chem. 52:1095, 1988; and Olsnes, Methods Enzymol. 50:330-335, 1978). In several embodiments, the toxin is Pseudomonas exotoxin (PE). The term "Pseudomonas exotoxin" as used herein refers to a full-length native (naturally occurring) PE or a PE that has been modified. Such modifications may include, but are not limited to, elimination of domain la, various amino acid deletions in domains lb, II and III, single amino acid substitutions and the addition of one or more sequences at the carboxyl terminus such as KDEL and REDL (see Siegall et al, J. Biol. Chem.264: 14256, 1989). In a prefened embodiment, the cytotoxic fragment of PE retains at least 50%, preferably 75%, more preferably at least 90%, and most preferably 95% of the cytotoxicity of native PE. In a most prefened embodiment, the cytotoxic fragment is more toxic than native PE. Native Pseudomonas exotoxin A is an extremely active monomeric protein (molecular weight 66 kD), secreted by Pseudomonas aeruginosa, which inhibits protein synthesis in eukaryotic cells. The native PE sequence is provided as SEQ ID NO: 1 of U.S. Patent No. 5,602,095, incoφorated herein by reference. The method of action is inactivation of the ADP-ribosylation of elongation factor 2 (EF-2). The exotoxin contains three structural domains that act in concert to cause cytotoxicity. Domain la (amino acids 1-252) mediates cell binding. Domain II (amino acids 253-364) is responsible for translocation into the cytosol and domain III (amino acids 400-613) mediates ADP ribosylation of elongation factor 2. The function of domain lb (amino acids 365-399) remains undefined, although a large part of it, amino acids 365-380, can be deleted without loss of cytotoxicity. See Siegall et al, J. Biol. Chem. 264: 14256- 14261, 1989, incoφorated by reference herein. PE employed includes the native sequence, cytotoxic fragments of the native sequence, and conservatively modified variants of native PE and its cytotoxic fragments. Cytotoxic fragments of PE include those which are cytotoxic with or without subsequent proteolytic or other processing in the target cell (for example, as a protein or pre-protein). Cytotoxic fragments of PE include PE40, PE38, PE37, and PE35. PE40 is a truncated derivative of PE as previously described in the art. See, Pai et al, Proc. Nat'lAcad. Sci. USA 88:3358-62, 1991; and Kondo et al, J. Biol. Chem. 263:9470-9475, 1988. PE35 is a 35 kD carboxyl-terminal fragment of PE composed of a Met at position 280 followed by amino acids 281-364 and 381-613 of native PE. PE37, another truncated derivative of PE, is described in U.S. Patent No. 5,821,238. PE38 is a truncated PE pro-protein composed of amino acids 253-364 and 381-613 which is activated to its cytotoxic form upon processing within a cell (see U.S. Patent No. 5,608,039, incoφorated herein by reference). In a particularly prefened embodiment, PE38 is the toxic moiety of the immunotoxin, however, other cytotoxic fragments, such as PE35, PE37, and PE40, are contemplated and are disclosed in U.S. Patent No. 5,602,095; U.S. Patent No. 5,821,238; and U.S. Patent No. 4,892,827, each of which is incoφorated herein by reference. Polynucleotides encoding the V and or the V_H of humanized antibodies that bind CEA, such as SV2, SV4, SV5 and HuCOL- 1_SDR are also provided. These polynucleotides include DNA, cDNA and RNA sequences which encode the humanized antibody. It is understood that all polynucleotides encoding these antibodies are also included herein, as long as they encode a polypeptide with the recognized activity, such as the binding to CEA. The polynucleotides of this disclosure include sequences that are degenerate as a result of the genetic code. There are 20 natural amino acids, most of which are specified by more than one codon. Therefore, all degenerate nucleotide sequences are included as long as the amino acid sequence of the antibody encoded by the nucleotide sequence is functionally unchanged. Primers, such as polymerase chain reaction (PCR) primers can readily be prepared that hybridize to a specific V_H or V_L, or a component thereof. In one embodiment, the primers include at least ten, at least 15, 16, 17, 18, 18, or 20 consecutive nucleotides of a nucleic acid encoding the V_H or V of interest. Also included are fragments of the above-described nucleic acid sequences that are at least 15 bases in length, which is sufficient to permit the fragment to selectively hybridize to DNA that encodes the antibody of interest under physiological conditions. The term "selectively hybridize" refers to hybridization under moderately or highly stringent conditions, which excludes non-related nucleotide sequences. A nucleic acid encoding a V_L and or V_H of a humanized COL-1 antibody that specifically binds CEA can be cloned or amplified by in vitro methods, such as the polymerase chain reaction (PCR), the ligase chain reaction (LCR), the transcription- based amplification system (TAS), the self-sustained sequence replication system (3SR) and the Qβ replicase amplification system (QB). For example, a polynucleotide encoding the protein can be isolated by PCR of cDNA using primers based on the DNA sequence of the molecule. A wide variety of cloning and in vitro amplification methodologies are well known to persons skilled in the art. PCR methods are described in, for example, U.S. Patent No. 4,683,195; Mullis et al., Cold Spring Harbor Symp. Quant. Biol. 51:263, 1987; and Erlich, ed.,PCR Technology, (Stockton Press, NY, 1989). Polynucleotides also can be isolated by screening genomic or cDNA libraries with probes selected from the sequences of the desired polynucleotide under stringent hybridization conditions. The polynucleotides include a recombinant DNA which is incoφorated into a vector; into an autonomously replicating plasmid or virus; or into the genomic DNA of a prokaryote or eukaryote, or which exists as a separate molecule (for example, a cDNA) independent of other sequences. The nucleotides of the invention can be ribonucleotides, deoxyribonucleotides, or modified forms of either nucleotide. The term includes single and double forms of DNA. DNA sequences encoding a V_L and/or V_H of a humanized antibody that specifically binds CEA can be expressed in vitro by DNA transfer into a suitable host cell. The cell may be prokaryotic or eukaryotic. The term also includes any progeny of the subject host cell. It is understood that all progeny may not be identical to the parental cell since there may be mutations that occur during replication. Methods of stable transfer, meaning that the foreign DNA is continuously maintained in the host, are known in the art. A polynucleotide sequence encoding a V and/or V_H of a humanized COL-1 antibody that specifically binds CEA can be operatively linked to expression control sequences. An expression control sequence operatively linked to a coding sequence is ligated such that expression of the coding sequence is achieved under conditions compatible with the expression control sequences. The expression control sequences include, but are not limited to, appropriate promoters, enhancers, transcription terminators, a start codon (for instance, ATG) in front of a protein-encoding gene, splicing signal for introns, maintenance of the conect reading frame of that gene to permit proper translation of mRNA, and stop codons. Thus, an scFv can be produced. The polynucleotide sequences encoding a V_L and/or V_H of a humanized COL-1 antibody that specifically binds CEA can be inserted into an expression vector including, but not limited to, a plasmid, virus or other vehicle that can be manipulated to allow insertion or incoφoration of sequences and can be expressed in either prokaryotes or eukaryotes. Hosts can include microbial, yeast, insect and mammalian organisms. Methods of expressing DNA sequences having eukaryotic or viral sequences in prokaryotes are well known in the art. Biologically functional viral and plasmid DNA vectors capable of expression and replication in a host are known in the art. Transformation of a host cell with recombinant DNA may be carried out by conventional techniques as are well known to those skilled in the art. Where the host is prokaryotic, such as E. coli, competent cells which are capable of DNA uptake can be prepared from cells harvested after exponential growth phase and subsequently treated by the CaCl₂ method using procedures well known in the art. Alternatively, MgCl₂ or RbCl can be used. Transformation can also be performed after forming a protoplast of the host cell if desired, or by electroporation. When the host is a eukaryote, such methods of transfection of DNA as calcium phosphate coprecipitation, conventional mechanical procedures such as microinjection, electroporation, insertion of a plasmid encased in liposomes, or virus vectors may be used. Eukaryotic cells can also be cotransformed with the polynucleotide sequence of interest, and a second foreign DNA molecule encoding a selectable phenotype, such as the heφes simplex thymidine kinase gene. Another method is to use a eukaryotic viral vector, such as simian virus 40 (S V40) or bovine papilloma virus, to transiently infect or transform eukaryotic cells and express the protein (see for example, Eukaryotic Viral Vectors, Cold Spring Harbor Laboratory, Gluzman ed., 1982). Isolation and purification of recombinantly-expressed polypeptides may be carried out by conventional means including preparative chromatography and immunological separations. Pharmaceutical Compositions and Therapeutic Methods Pharmaceutical compositions are disclosed herein that include a humanized COL-1 monoclonal antibody, such as SV2, SV4, SV5 or HUCOL-I_SDR that can be foonulated with an appropriate solid or liquid carrier, depending upon the particular mode of administration chosen. In addition, a humanized COL-1 monoclonal antibody linked to an effector molecule (for instance, a toxin, a chemotherapeutic drug, or a detectable label) can be prepared in pharmaceutical compositions. The pharmaceutically acceptable carriers and excipients useful in this disclosure are conventional. For instance, parenteral formulations usually comprise injectable fluids that are pharmaceutically and physiologically acceptable fluid vehicles such as water, physiological saline, other balanced salt solutions, aqueous dextrose, glycerol or the like. Excipients that can be included are, for instance, other proteins, such as human serum albumin or plasma preparations. If desired, the pharmaceutical composition to be administered can also contain minor amounts of non-toxic auxiliary substances, such as wetting or emulsifying agents, preservatives, and pH buffering agents and the like, for example sodium acetate or sorbitan monolaurate. The dosage form of the pharmaceutical composition will be determined by the mode of administration chosen. For instance, in addition to injectable fluids, topical, inhalation, oral and suppository formulations can be employed. Topical preparations can include eye drops, ointments, sprays and the like. Inhalation preparations can be liquid (for example, solutions or suspensions) and include mists, sprays and the like. Oral formulations can be liquid (for example, syrups, solutions or suspensions), or solid (for example, powders, pills, tablets, or capsules). Suppository preparations can also be solid, gel, or in a suspension form. For solid compositions, conventional non-toxic solid carriers can include pharmaceutical grades of mannitol, lactose, starch, or magnesium stearate. Actual methods of preparing such dosage forms are known, or will be apparent, to those skilled in the art. The pharmaceutical compositions that include a humamzed COL-1 monoclonal antibody, such as SV2, SV4, SV5 or HuCOL- I_SDR, can be formulated in unit dosage form, suitable for individual administration of precise dosages. In addition, the phannaceutical compositions may be administered as an immunoprophylactic in a single dose schedule or as an immunotherapy in a multiple dose schedule. A multiple dose schedule is one in which a primary course of treatment may be with more than one separate dose, for instance 1-10 doses, followed by other doses given at subsequent time intervals as needed to maintain or reinforce the action of the compositions. Treatment can involve daily or multi-daily doses of compound(s) over a period of a few days to months, or even years. Thus, the dosage regime will also, at least in part, be determined based on the particular needs of the subject to be treated and will be dependent upon the judgment of the administering practitioner. In one specific, non-limiting example, a unit dosage can be about 0.1 to about 10 mg per patient per day. Dosages from about 0.1 up to about 100 mg per patient per day may be used, particularly if the agent is administered to a secluded site and not into the circulatory or lymph system, such as into a body cavity, into a lumen of an organ, or directly into a tumor. In one embodiment, about 10 mCi of a radiolabeled humanized COL-1 monoclonal antibody is administered to a subject. In other embodiments, about 15 mCi, about 20 mCi, about 50 mCi, about 75 mCi or about 100 mCi of a radiolabeled humanized COL-1 monoclonal antibody is administered to a subject. The amount of active compound(s) administered will be dependent on the subject being treated, the severity of the affliction, and the manner of administration, and is best left to the judgment of the prescribing clinician. Within these bounds, the formulation to be administered will contain a quantity of the active component(s) in amounts effective to achieve the desired effect in the subject being treated. The compounds of this disclosure can be administered to humans on whose tissues they are effective in various manners such as topically, orally, intravenously, intramuscularly, intraperitoneally, intranasally, intradermally, intrathecally, subcutaneously, via inhalation or via suppository. The particular mode of administration and the dosage regimen will be selected by the attending clinician, taking into account the particulars of the case (for example, the subject, the disease, the disease state involved, and whether the treatment is prophylactic). In one embodiment, a therapeutically effective amount of a humanized COL-1 antibody, such as SV2, SV4, SV5 or HuCOL- I_SDR, is the amount of humanized COL-1 antibody necessary to inhibit further growth of a CEA-expressing tumor or suppress the growth of a CEA-expressing tumor, without eliciting a HAMA response in the patient receiving the treatment. In other embodiments, a therapeutically effective amount of humanized COL-1 antibody is the amount of humanized COL-1 antibody necessary to eliminate or reduce the size of a CEA-expressing tumor, without eliciting a HAMA response. Specific, non-limiting examples of CEA-expressing tumors are adenocarcinoma, colorectal, gastric, pancreatic, breast, lung, and ovarian tumors. In yet another embodiment, a therapeutically effective amount of humanized COL-1 antibody is an amount of humanized COL-1 antibody that is effective at reducing a sign or a symptom of the tumor and induces a minimal HAMA response. A therapeutically effective amount of a humanized COL-1 monoclonal antibody, such as SV2, SV4, SV5 or HuCOL- I_SDR, can be administered in a single dose, or in several doses, for example daily, during a course of treatment. In one embodiment, treatment continues until a therapeutic result is achieved. However, the effective amount of humanized COL-1 antibody will be dependent on the subject being treated, the severity and type of the affliction, and the manner of administration of the therapeutic(s). Controlled release parenteral formulations of a humanized COL-1 monoclonal antibody can be made as implants, oily injections, or as particulate systems. For a broad overview of protein delivery systems (see Banga, A.J., Therapeutic Peptides and Proteins: Formulation, Processing, and Delivery Systems, Technomic Publishing Company, Inc., Lancaster, PA, 1995). Particulate systems include microspheres, microparticles, microcapsules, nanocapsules, nanospheres, and nanoparticles. Microcapsules contain the therapeutic protein as a central core. In microspheres the therapeutic is dispersed throughout the particle. Particles, microspheres, and microcapsules smaller than about 1 μm are generally refened to as nanoparticles, nanospheres, and nanocapsules, respectively. Capillaries have a diameter of approximately 5 μm so that only nanoparticles are administered intravenously. Microparticles are typically around 100 μm in diameter and are administered subcutaneously or intramuscularly (see Kreuter, J., Colloidal Drug Delivery Systems, J. Kreuter, ed., Marcel Dekker, Inc., New York, NY, pp. 219-342, 1994; Tice & Tabibi, Treatise on Controlled Drug Delivery, A. Kydonieus, ed., Marcel Dekker, Inc. New York, NY, pp. 315-339, 1992). Polymers can be used for ion-controlled release. Various degradable and nondegradable polymeric matrices for use in controlled drug delivery are known in the art (Langer, R., Accounts Chem. Res. 26:537, 1993). For example, the block copolymer, polaxamer 407 exists as a viscous yet mobile liquid at low temperatures but forms a semisolid gel at body temperature. It has shown to be an effective vehicle for formulation and sustained delivery of recombinant IL-2 and urease (Johnston et al, Pharm. Res. 9:425, 1992; and Pec et al, J. Parent. Sci. Tech. 44:58, 1990). Alternatively, hydroxyapatite has been used as a microcarrier for controlled release of proteins (Ijntema et al, Int. J. Pharm. 112:215, 1994). In yet another aspect, liposomes are used for controlled release as well as drug targeting of the lipid-capsulated drug (Betageri et al. , Liposorne Drug Delivery Systems, Technomic Publishing Co., Inc., Lancaster, PA, 1993). Numerous additional systems for controlled delivery of therapeutic proteins are known (for example, U.S. Patent No. 5,055,303; U.S. Patent No. 5,188,837; U.S. Patent No. 4,235,871; U.S. Patent No. 4,501,728; U.S. Patent No. 4,837,028; U.S. Patent No. 4,957,735 and U.S. Patent No. 5,019,369; U.S. Patent No. 5,055,303; U.S. Patent No. 5,514,670; U.S. Patent No. 5,413,797; U.S. Patent No. 5,268,164; U.S. Patent No. 5,004,697; U.S. Patent No. 4,902,505; U.S. Patent No. 5,506,206, U.S. Patent No. 5,271,961; U.S. Patent No. 5,254,342 and U.S. Patent No. 5,534,496). Site-specific administration of the disclosed compounds can be used, for instance by applying the humanized COL-1 antibody to a pre-cancerous region, a region of tissue from which a tumor has been removed, or a region suspected of being prone to tumor development. In some embodiments, sustained intra-tumoral (or near-tumoral) release of the pharmaceutical preparation that includes a therapeutically effective amount of humanized COL-1 antibody may be beneficial. The present disclosure also includes therapeutic uses of humanized COL-1 monoclonal antibodies, such as SV2, SV4, SV5 or HuCOL- I_SDR, that are non- covalently or covalently linked to effector molecules. In one specific embodiment, the humanized COL-1 monoclonal antibody is covalently linked to an effector molecule that is toxic to a tumor or cell expressing CEA. In one specific, non-limiting example, the effector molecule is a cytotoxin. In other specific, non-limiting examples, the effector molecule is a detectable label, a radioactive isotope, a chemotherapeutic drug, a bacterially-expressed toxin, a virally-expressed toxin, a venom protein, or a cytokine. Humanized COL-1 antibodies linked to effector molecules can be prepared in pharmaceutical compositions. In one specific, non-limiting example, a pharmaceutical composition for intravenous administration, such as an immunotoxin, would be about 0.1 to 10 mgper patient per day. Dosages from 0.1 up to about 100 mg per patient per day may be used, particularly if the agent is administered to a secluded site and not into the circulatory or lymph system, such as into a body cavity or into a lumen of an organ. Actual methods for preparing aαjninistrable compositions will be known or apparent to those skilled in the art and are described in more detail in such publications as Remingtons Pharmaceutical Sciences, 19^th Ed., Mack Publishing Company, Easton, PA (1995). Humanized COL-1 monoclonal antibodies covalently linked to an effector molecule have a variety of uses. For example, a humanized COL-1 antibody such as SV2, SV4, SV5 or HuCOL- I_SDR can be linked to a radioactive isotope and used in immunotherapy. A humanized COL-1 antibody covalently linked to a radioactive isotope is of use to localize a tumor in radioimmunoguided surgery, such that the tumor can be surgically removed. The present disclosure also includes combinations of a humanized COL-1 monoclonal antibody, such as SV2, SV4, SV5 or HuCOL- I_SDR, with one or more other agents useful in the treatment of tumors. For example, the compounds of this disclosure can be administered in combination with effective doses of immunostimulants, anti- cancer agents, anti-inflammatory agents, anti-infectives, and/or vaccines. The term "administration in combination" or "co-administration" refers to both concurrent and sequential administration of the active agents. A subject that has a tumor, or is predisposed to the development of a tumor, will be a candidate for treatment using the therapeutic methods disclosed herein.

Diagnostic Methods and Kits A method is provided herein for the in vivo or in vitro detection of CEA- expressing tumors or cells. An in vivo detection method can localize any tumor or cell that expresses CEA in a subject. In one embodiment, a humanized COL-1 antibody such as SV2, SV4, SV5 or HUCOL-I_SDR is administered to the subject for a sufficient amount of time for the antibody to localize to the tumor or cell in the subject and to form an immune complex with CEA. The immune complex is then detected. In one specific, non-limiting example detection of an immune complex is performed by immunoscintography. Other specific, non-limiting examples of immune complex detection include radiolocalization, radioimaging, or fluorescence imaging. In one example, the antibody is directly linked to an effector molecule that is a detectable label. Specific, non-limiting examples of detectable labels include a radioactive isotope, an enzyme substrate, a co-factor, a ligand, a chemiluminescent agent, a fluorescent agent, a hapten, or an enzyme. In another example, a humanized COL-1 antibody and a secondary antibody are administered to the subject for a sufficient amount of time for the humanized COL-1 antibody to form an immune complex with CEA on a tumor or cell, and for the secondary antibody to form an immune complex with the humanized COL-1 antibody. In one embodiment, the humanized COL-1 antibody is complexed with the secondary antibody prior to their administration to the subject, hi one specific, non-limiting embodiment, the secondary antibody is linked to a detectable label. In one embodiment, the immune complex, which includes CEA, the humamzed COL-1 antibody, and the secondary antibody linked to a detectable label, is detected as described above. A method of detecting tumors in a subject includes the administration of a humanized COL-1 antibody such as SV2, SV4, SV5 or HuCOL- I_SDR, complexed to an effector molecule, such as a radioactive isotope. After a sufficient amount of time has elapsed to allow for the administered radiolabeled antibody to localize to the tumor, the tumor is detected. In one specific, non-limiting example, a radiolabeled immune complex is detected using a hand held gamma detection probe. Primary tumors, metastasized tumors or cells expressing CEA can be detected. For example, a humanized COL-1 antibody complexed to an effector molecule, such as a radioactive isotope, is administered to a subject prior to surgery or treatment. In one specific embodiment, the detection step is performed prior to surgery to localize the tumor. In another embodiment, the detection step is performed during surgery, for example to detect the location of the tumor prior to removing it, as in radioimmunoguided surgery. A humanized COL-1 antibody complexed to an effector molecule, such as a radioactive isotope, can also be administered to a subject following surgery or treatment, to determine the effectiveness of the treatment, such as to ensure the complete removal of the tumor, or to detect a recurrence of the tumor. In vitro detection methods are provided herein. These methods can be used to screen any biological sample to assess for the presence of a tumor or cell that expresses CEA. A biological sample can be obtained from a mammal, such as a human, suspected of having a tumor expressing CEA. In one embodiment the subject has a colorectal tumor. In other embodiments, the subject has a gastric tumor, a pancreatic tumor, a breast tumor, a lung tumor, an adenocarcinoma, or an ovarian tumor. Other biological samples that can be detected by the in vitro detection method include samples of cultured cells that express CEA. Such samples include, but are not limited to, tissue from biopsies, autopsies, and pathology specimens. Biological samples also include sections of tissues, such as frozen sections taken for histological pmposes. Biological samples further include body fluids, such as blood, serum, saliva, or urine. A kit is provided herein for detecting a CEA-expressing tumor or cell. Kits for detecting a CEA-expressing tumor or cell will typically include a humanized COL-1 antibody that specifically binds CEA, such as one or more of SV2, SV4, SV5 or HuCOL- I_SDR- An antibody fragment, such as an Fv fragment can be included in the kit. The antibody can also be provided as an immunoconjugate. Thus, in several examples, the antibody is conjugated to a detectable label, such as a radioactive isotope, an enzyme substrate, a co-factor, a ligand, a fluorescent agent, a hapten, an enzyme, or a chemiluminescent agent. The kit can further include instructional materials disclosing means of use of an antibody that specifically binds CEA, such as SV2, SV4, SV5 or HuCOL- I_SDR, or a fragment thereof. The instructional materials can be written, in an electronic form (for example, computer diskette or compact disk) or may be visual (for example, video files). The kits can also include additional components to facilitate the particular application for which the kit is designed. Thus, for example, the kit can additionally contain a means of detecting a label (for example, enzyme substrates for enzymatic labels, filter sets to detect fluorescent labels, appropriate secondary labels such as a secondary antibody, or the like). In one example, the kit contains a secondary antibody that is conjugated to a detectable label. Kits can additionally include buffers and other reagents, such as an antigen (for example, purified CEA) routinely used for the practice of a particular method, or of use in the preparation of a suitable control. Such kits and appropriate contents are well known to those of skill in the art. The invention is illustrated by the following non-limiting Examples.

EXAMPLES EXAMPLE 1 Methods used to generate germline-derived humanized COL-1 antibodies Unlike other humanized antibodies that have been developed using a single human antibody as the template for the V_L or V_H domains of the humanized antibody, a humanized COL-1 antibody is disclosed herein that was developed using a human template for the V_L domain that contains several different human germline sequences. SDR grafting of critical murine residues onto the human template derived from multiple human germline sequences generated HuCOL-1 variants that have a binding activity comparable to the parental HuCOL- 1 and display reduced immunogenicity compared to the parental HuCOL- 1. This example describes the methods used to generate these germline-derived HuCOL- 1 antibodies.

Synthetic oligonucleotides Oligonucleotide primers listed below were used for site-specific mutagenesis of the V_L domain of HuCOL- l_AbrCDR- They were supplied by Gene Probe Technologies (Gaithersburg, MD). The mutagenic bases are underlined, the positions of the residue changes are parenthetically enclosed, and the sequences recognized by restriction endonucleases are in bold.

V primers: 5' V_L (24,25): 5'-CACTTTGGCTGGCCCTGCAGTTGATGG-3' (SEQ ID NO: 1) 5' V_L (27d): 5^!-ATAGCCAGAGGAACTGACACT-3' (SEQ ID NO: 2) 5' V_L (33): 5'-GTACCAGTGCAGATAACTATAGC-3 ' (SEQ ID NO: 3) 3 ' V_L (50): 5 '-CTCATTTACGCCGCATCCAGC-3 ' (SEQ ID NO: 4) 3' V_L (53): 5'-CTTGCATCCAGCCTGCAATCTG-3' (SEQ ID NO: 5)

3' V_L (90): S'-TACTGTCAGCAGAGTAGGGAG-S' (SEQ ID NO: 6) The sequences of the 20- to 21- bp-long end primers used for DNA amplification were as follows: 5 ' V_L: 5 '-GCAAGCTTCCACCATGGATA-3 ' (SEQ ID NO: 7)

3' V_L: 5'-TGCAGCCGCGGTACGTTTGAT-3' (SEQ ID NO: 8)

The 5' V end primer carries the Hindϊll restriction endonuclease site followed by the Kozak consensus sequence and a sequence encoding the N-terminus of the signal peptide. The 3' V_L end primer contains a unique SβcII site located 10 bp downstream from the start of the human K C region.

DNA amplification and mutagenesis All PCR amplifications were carried out in EasyStart™ 50 tubes (Molecular BioProducts, San Diego, CA) in a final volume of 50 μl of PCR buffer (20 mM Tris- HC1, pH 8.4 and 50 mM KC1) containing 2 mM MgCl₂, 200 μM of dNTPs, 2.5 units of the high-fidelity copy PfuTurbo DNA polymerase (Stratagene Cloning Systems, La Jolla, CA), 0.1% Triton X-100, 25 pmoles each of the appropriate 5' and 3' primers, and 100 ng of DNA template. Initial denaturation at 95°C for 7 minutes was followed by 30 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 30 seconds and extension at 72°C for 1 minute. It was followed by a final primer extension step at 72 °C for 10 minutes. The V_L genes of the SDR variants of HuCOL- 1 were generated using primer-induced mutagenesis by a dual step PCR as described by Landt et al. (Gene 96:125-128, 1990).

Expression vectors and generation of expression constructs Prior to cloning into expression vectors, the synthesized V_L genes were sequenced using the ABI PRISM dRhodamine terminator cycle sequencing kit (Perkin Elmer Applied Biosystems, Foster City, CA). The mammalian expression vector pDCM-dhfr (Gonzales et al, Mol Immunol, 40:337-349, 2003) was used for the co-expression of the Ig L and H chains in Chinese hamster ovary (CHO) cells. pDCM- dhfr contains two distinct cloning sites, each downstream from an enhancer-promoter complex of the immediate early genes of human cytomegalo virus (CMV). The plasmid also includes a dhfr expression unit driven by an enhancer-deficient simian virus 40 (SV40) early promoter (Ryu et al, Hum. Antibodies Hybridomas, 7:113-122, 1996) and a neomycin resistance gene for drug selection of transfectants. To generate a pDCM-dhfr expression construct of the genes encoding the L and

H chains of the HuCOL- l_AbrCDR antibody, a construct containing the L chain was first generated. An approximately 400 bp DNA fragment encoding the V_L of HuCOL- l_AbrCD was prepared by digestion of the pre-existing construct pBScHuCOL-l_Abrc_DRVL (De Pascalis et al, J Immunol, 169:3076-3084, 2002) with HindlU and Sαell. This fragment was then swapped with the HindlU and SαcII fragment of the pre-existing construct of HuCC49V10 V_L in pDCMdhfr (Gonzales et al, Mol Immunol, 40:337- 349, 2003). To assemble the H chain of HuCOL- l_AbrCDR, the EcoRVApal fragment encoding the V_H of the HuCC49V10 in pDCMdhfr (Gonzales et al, Mol Immunol, 40:337-349, 2003) was replaced with the EcoKUApal fragment of the COL-1 V_H released from a pre-existing construct of pBScHuCOL-1 (De Pascalis et al, J Immunol, 169:3076-3084, 2002). The resulting construct was then treated with EcoRUNotl and the released DNA fragment was then inserted into the pDCM-dhfr construct containing the DNA encoding the HuCOL- l_Abrc_DR L chain. To generate expression constructs of the SDR variants of HuCOL-1, the mutated V region sequences were exchanged with the V_L of HuCOL-1 _AbrCDR in the pDCM-dhfr L chain construct through the HindUVSacU site, prior to insertion of the H chain gene into the vector.

Mammalian cell culture and production of recombinant Abs To develop transfectants expressing HuCOL- l_AbrCDR and the SDR variants, CHOdhfr cells were transfected with the pDCM-dhfr derived expression construct using liposome-mediated DNA transfer (Lipofectamine Plus, Invitrogen, Carlsbad, CA) according to the guidelines of the manufacturer. Following transfection, cells were incubated at 37°C in DMEM/F12 medium overnight, and were then trypsinized and seeded in 96-well plates at 2 x 10⁴ cells per well in selection medium (alpha MEM, 10% dialyzed fetal bovine serum, 550 μg/mL G418). After 2 weeks of selection, the culture supematants of the stable transfectants were monitored by ELISA assay and Western blotting.

ELISA Enzyme linked immunoassay (ELISA) was carried out by coating 96-well polyvinyl microtiter plates with Fcγ-fragment-specific goat anti -human IgG (100 ng/well) (Jackson ImmunoResearch Laboratories, West Grove, CA) or with CEA (100 ng/well) (Research Diagnostic Inc., Flanders, NJ) to test for the production of Ig by the transfected mammalian cells and to assess the antigen reactivity of the purified antibodies, respectively. To detect reactivity of the samples to the ELISA plates, the SureBlue™ detection reagent was used (KPL, Gaithersburg, MD) according to the manufacturer's instructions. The details of the assay procedure have been reported previously (De Pascalis et al, J Immunol, 169:3076-3084, 2002, Gonzales et al, Mol Immunol, 40:337-349, 2003).

Purification of recombinant Abs The highest antibody-producing clones were grown in CHO-S-SFM II serum- free medium (Invitrogen, Carlsbad, CA) supplemented with 550 μg mL G418. Cell culture supematants were collected and centrifuged to remove cellular debris.

Antibodies were purified from the supematants using a protein A column (Invitrogen, Carlsbad, CA) as detailed previously (Gonzales et al, Mol Immunol, 40:337-349, 2003). The purified proteins were concentrated using Centricon 30 (Amicon, Beverly, MA) and buffer-exchanged in PBS (pH 7.4). The protein concentration was determined using a Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA) based on the method developed by Bradford (Bradford, Anal Biochem 72:248-254, 1976). The purity of the antibody preparation was evaluated using the Agilent 2100 Bioanalyzer system (Agilent Technologies, Waldronn, Germany), under reducing and non-reducing conditions, using the Protein 200 LabChip kit (Agilent Technologies).

Competition RIA The relative antigen-binding affinity of HuCOL- 1, HuCOL- l_AbrCDR and the HuCOL- 1 SDR variants were determined using a competition RIA as described previously (De Pascalis et al., J Immunol, 169:3076-3084, 2002). Twenty-five μl of serial dilutions of the antibodies to be tested and HulgG (negative control), prepared in 1% BSA in PBS, were added to microtiter plates containing 200 ng of CEA saturated with 5% BSA in PBS. ¹²⁵I-labeled mCOL-1 (100,000 cp in 25 μl) was then added to each well. After an overnight incubation at 4°C, the plates were washed and counted in a γ-scintillation counter. The relative affinity constants were calculated from the linear parts of the curve by a modification of the Scatchard method (Frankel and Gerhard, Mol Immunol 16:101-106, 1979).

Flow cytometric (FACS) analysis To evaluate the ability of the HuCOL- 1 SDR variants to bind to cell-surface

CEA, FACS analysis was carried out as described previously (De Pascalis et al, J Immunol, 169:3076-3084, 2002). Retrovirally transduced MC38 cells expressing CEA (Robbins et al, Cancer Res., 51:3657, 1991) were resuspended in cold Ca - and Mg - free Dulbecco's phosphate-buffered saline (PBS) and incubated with the antibodies for 30 minutes on ice. A human IgG was used as an isotype control. After one washing cycle, the cell suspension was stained with fluorescein isothiocyanate (FITC)- conjugated mouse anti-human antibody (Pharmingen, San Diego, CA) for 30 minutes on ice. A second washing cycle was performed and then the samples were analyzed with a FACScan (Becton Dickinson, Mountain View, CA) using CellQuest for Macintosh. Data from analysis of 10,000 cells were obtained.

Immunoadsorption of patient serum Stored patients' sera from a phase I clinical trial that involved the administration of I-mCOL-1 to gastrointestinal carcinomas patients (Yu et al, J Clin Oncol,

14:1798, 1996) were used to assess the sera reactivity of the HuCOL-1 SDR variants. Several sera showed the presence of anti-V region antibodies to murine monoclonal antibody COL-1 (De Pascalis et al, J Immunol, 169:3076-3084, 2002). However, the sera contained circulating antigen and anti-murine Fc antibodies, which could interfere with the binding of the HuCOL- 1 variants to the mCOL-1 anti-V region Abs. As detailed previously (De Pascalis et al, J Immunol, 169:3076-3084, 2002), the circulating CEA and anti-murine Fc antibodies were removed from the sera by sequential pre-adsoφtion with purified mCOL-6 and mCOL-4 mAbs, two murine Abs that react with epitopes of CEA different from the one recognized by mCOL-1 (Kuroki et al. , Int J Cancer 44:208, 1989). mCOL-4 is of the same isotype as mCOL- 1.

Sera reactivity The reactivity of COL-1 variants to anti-V region antibodies was determined using a recently developed SPR-based competition assay (Gonzales et al, J Immunol Methods 268: 197-210, 2002). Competition experiments were performed at 25°C using a CM5 sensor chip containing HuCOL- 1 in flow cell 1 and rabbit gamma globulin (Bio- Rad, Hercules, CA), as a reference, in flow cell 2. Typically, HuCOL- 1 or its variants, diluted at different concentrations in running buffer (10 mM HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA and 0.005%) Tween-20), were used to compete with the HuCOL- 1 immobilized on the sensor chip for binding to serum anti-V region antibodies. Patient's serum with or without the competitor (HuCOL- 1 or its variants) was applied across the sensor surfaces using a recently developed sample application technique (Abrantes et al, Anal Chem 73:2828-2835, 2001) at a unidirectional flow of 1 μl/min. After the binding was measured for 600 seconds, the unbound samples were washed from the surfaces with running buffer at a flow rate of 100 μl/min, and the surfaces were regenerated with 10 mM glycine (pH 2.0). The percent binding at each antibody concentration was calculated as follows: % binding = [slope of the signal obtained with competitor (serum + antibodies) / slope of the signal obtained without competitor (serum only)] x 100. The IC₅o for each antibody, the concentration required for 50% inhibition of the binding of the serum to immobilized HuCOL- 1, was calculated.

Example 2 Design and generation of the genes encoding the HuCOL-1 SDR variants This example describes the generation of nucleic acid constructs encoding

HuCOL-1 SDR variants. To develop a humanized COL-1 antibody, the potential SDRs and non-SDRs within the COL-1 CDRs, as identified in Table I, were evaluated using the Protein Data Bank database (Bernstein et al, J. Mol. Biol. 112-535, 1977; Abola et al, In Crystallographic Databases-Information Content, Software Systems, Scientific

Applications, F.C. Allen, G. Bergerhoff, and R. Sievers, eds., Data Commission of the International Union of Crystallography, Bonn, pp. 107-132, 1987). It should be noted that the database does not always help in reaching a definite conclusion regarding the designation of some CDR residues as ligand contact residues (SDRs). Even if the information provided by the database is compelling to designate some residues SDRs, amino acid substitutions at these positions may reveal that such residues are not, in fact, SDRs. Also in Table I are the sequences of the CDRs of COL-1 and the human germline V_L and V_H sequences used as templates for SDR grafting. The human germline CDR sequence that shows the best homology to the conesponding murine CDR was chosen as the human template for that particular CDR. The numbering convention of Kabat et al. (Sequence of proteins of immunological interests, 5th Ed. U.S. Department of Health and Human Services, National Institutes of Health, Bethesda, MD, NIH Publication no. 91-3242, 1991) is used.

" Numbering convention of Kabat et al. [ref. 30] is used in this table. ⁶ Residue positions shown in bold denote SDRs. "Human germline V_κ sequences from Cox et al. (Eur. J. Immunol 24:827-836, 1994) were used as templates for the light chain CDRs. 'Αmino acid residues in italics indicate differences between COL-1 and the human template at the corresponding positions. " Human germline V_H sequences from Tomlinson et al. (J. Mol. Biol., 227:776-798, 1992) were used as templates for the heavy chain CDRs. 10 In LCDR1, the amino acid differences between COL-1 and the germline sequence X93639 are located at positions 27, 27d, 28, 29, 30, 33 and 34. Among these residue positions, only 27 and 33 are designated non-SDRs. A variant (SV2) containing substitutions at the two positions (K27Q, M33L) was generated. Another variant (SV1) 15 was generated containing a change at position 27d, in addition to the changes at 27 and 33. In LCDR2, there are two differences between the COL-1 sequence and the human germline X93623, located at positions 50 and 53. Both of these positions are designated as SDRs. To test whether the residue at position 53 was necessary for antigen binding, a variant (SV4) containing a human substimtion at this position (N53S) was produced. Another variant (SV3) containing a human residue at position 50 (L50A), in addition to position 53, was also produced. In LCDR3, the residues at positions 90, 92, 93 and 94 are different between COL-1 and the germline sequence X93627/Jκ4. A variant (SV5) containing a mutation to change the murine residue at position 90 with its human counteφart (H90Q) was generated. HCDR3 was not considered in designing variants because each residue position in this CDR is highly likely to be an SDR. In addition, most segments of HCDR3 are not encoded in the V_H germline sequences (Tomlinson et al, J. Mol. Biol, 227:776-798, 1992) and consequently, no human template was chosen for this CDR. In HCDRl, only position 31 varies between the COL-1 and the human germline Z12309 sequences. No substitution was made at this position, hi HCDR2, the amino acid differences between the COL-1 sequence and the human germline Z14071 sequence are located at positions 52, 53, 54, 56, 58 and 61. No human substitutions were made at positions 52, 53, 54, 56, and 58. The residue at position 61, on the other hand, is a non-SDR. Thus, all the light chain SDR variants contain this human substitution (P61Q). Table II summarizes the human substitutions made in the COL-1 variants compared with the murine sequence. Table II. Amino acid substitution variants of COL-1 Amount of Ab K_v (relative

Variant CDR Positions Substituted" required for 50% affinity constant)

Designation Inhibition (ng) x 10^s M^"1 Light Chain Heavy Chain

HuCOL-1 None none 24.7 2.60

HuCOL- l_Abtc_DR 24, 25, 27 61 31.3 1.16

SV1 27, 27d, 33 61 285.5 0.27

SV2 27, 33 61 25.2 2.18

SV3 50, 53 61 688.8 0.16

SV4 53 61 29.1 1.84

SV5 90 61 21.9 2.85

HuCOL-l_SDR 27, 33, 53, 90 61 20.5 2.98

Numbering convention of Kabat et al. is used in this table. Genes encoding the V_L domains of the variants were generated by primer- induced mutagenesis, as described in Example 1, using the pBScHuCOL-l_AbrCDRVL construct (De Pascalis et al, J Immunol, 169:3076-3084, 2002) as the template. The most homologous human germline sequence to the frameworks of COL-1 is X93640 (SEQ ID NO: 41), which was used as the template for the V_L frameworks in this study. In developing HuCOL- 1 and HuCOL- l_AbrCDR, the template used for the VL frameworks was the reananged human antibody VJI'CL (De Pascalis et al, J Immunol, 169:3076- 3084, 2002), which differed from X93640 only by a residue at position 9 in the frameworks. In COL-1 and VJI'CL, the residue at position 9 is an alanine, while it is an aspartic acid in X93640. The residue in this position is unlikely to be crucial in maintaining the overall structure of the antibody combining site as it is not involved in CDR contact or in the V_L V_H interaction. Therefore, a single substitution in the V frameworks was first incoφorated at position 9 (A9D) before mutations in the CDRs were incoφorated. The V_L regions of the SDR variants were generated by primer- induced mutation, inserted into the pDCM-dhfr expression casette, and subsequently sequenced, as described in Example 1. Example 3 Expression of COL-1 SDR variants and characterization of the purified monoclonal antibodies The expression constructs of the genes encoding the H and L chains of the HuCOL- l_Abrc_DR antibody and the SDR variants were introduced into CHOdhfr- cells grown in 96-well plates. The supematants harvested from the G418 resistant clones were assayed for Ig production by ELISA as described in Example 1. Most of the transfectants were positive for Ig production. The supematants of the twelve highest- producing clones for each construct, based on ELISA results, were further analyzed for Ig production through Western blotting. Based on the results from Western blots, the highest producing clone for each construct was cultured and grown for antibody production and purification. The purity of the antibody preparations was verified using the Agilent 2100 Bioanalyzer system. Under reducing conditions (FIG. 1), all antibodies yielded two protein bands of approximately 24-27 kDa and 60 kDa. These molecular masses are in conformity with those of the Ig L and H chains, respectively.

Example 4 Relative CEA-binding of humanized Abs derived from COL-1 An ELISA was carried out to assess the CEA reactivity of the SDR variants. S VI and SV3 showed a significantly reduced reactivity to the CEA antigen, while S V2, SV4 and SV5 showed reactivity comparable to HuCOL- l_AblCDR, which was used as a positive control. Based on the ELISA results, a final SDR-grafted variant (HuCOL- I_SDR) was generated by combining, in one construct, all the human substitutions present in the SV2, SV4 and SV5 variants. A competition radioimmunoassay (RIA) was performed to determine the relative CEA-binding affinity of HuCOL-1, HuCOL- l_AbrCDR, and the different SDR variants, including the final HuCOL-l_SDR variant. Serial dilutions of unlabeled antibodies were used to compete with the binding of I-mCOL-1 (FIG. 2) to CEA. Most of the COL-1-derived recombinant antibodies, except SVl and SV3, were able to completely inhibit the binding of I-mCOL-1 to CEA. The competition profiles of SV2, SV5 and HUCOL-1_SDR were comparable to that of the CDR-grafted HuCOL-1.

The amount of antibody required for 50% inhibition of the binding of I-mCOL-1 to CEA was 24.7, 25.2, 21.9, and 20.5 ng for HuCOL-1, SV2, SV5, and HUCOL-1_SDR, respectively (Table III). These results, calculated from the linear parts of the graph shown in FIG. 2, are in conformity with the relative affinity constants (Ka) of 2.60 x 10⁸, 2.18 x 10⁸, 2.85 x 10⁸ and 2.98 x 10⁸ M^"1 for HuCOL-1, SV2, SV5, and HuCOL- I_SDR, respectively. The competition profiles of antibodies HuCOL-l_AbrCDR and SV4 were shifted slightly to the right as compared with HuCOL- 1, cooesponding to 31.3 and 29 A ng, respectively, of antibody required for 50% inhibition of I-mCOL-1 binding to CEA. The shift to the right of the competition profiles conesponds to a decrease in the CEA-binding affinities of HuCOL- l_{A rCDR} and SV4, calculated to be 1.16 x 10⁸M^"1 and 1.84 x 10 M^" , respectively. The variants SVl and SV3 showed substantial decreases in their CEA-binding capacity, requiring 285.5 and 688.8 ng, respectively, of 125 antibody to inhibit the I-mCOL-1 binding to CEA by 50%. This translates to binding affinities of 0.27 x 10⁸ M^"1 and 0.16 x 10⁸ M^"1 for SVl and SV3, respectively, showing an approximately 10- and 20-fold decrease in binding reactivity compared with HuCOL-1.

Example 5 Flow cytometric analysis Flow cytometric analysis was used to measure the binding of HuCOL- 1, HuCOL- l_AbrcDR and HuCOL- 1_SDR to the surface-expressed CEA of the retro virally transduced tumor cell line, MC38 (Robbins et al, Cancer Res., 51:3657, 1991). No significant differences were found in the mean fluorescence intensity, or in the percentage of cells that were reactive with the three humanized Abs (FIG. 3). The percentages of gated cells, calculated after exclusion of inelevant binding, for HuCOL- 1, HuCOL-l_AbrCDR and HuCOL- 1_SD were 23.2, 27.4 and 31.1, respectively, while the mean fluorescence intensities were between 24 and 25 when 1 μg of each antibody was used.

Example 6 Reactivity of Humanized COL-1 SDR variants to patients' sera A measure of the immunogenicity of a variant antibody is its in vitro reactivity to the sera of patients who were administered the parental antibody in a clinical trial. To assess the potential immunogenicity of HuCOL- 1, HuCOL- l_AbrCDR and HuCOL- l_SDRin patients, the antibodies were characterized for their reactivity to sera from gastrointestinal carcinoma patients who were administered I-mCOL-1 in a phase I clinical trial (Yu et al, J. Clin. Oncol, 14:1798, 1996). As described in Example 1, any circulating CEA and anti-murine Fc antibodies were removed from the sera by immunoadsoφtion with mCOL-6 and mCOL-4, two murine anti-CEA antibodies of IgG_! and IgG_2a isotypes, respectively. Specific binding profiles of immobilized HuCOL-1 to the sera of patients MB and EM showed that the pre-adsorbed sera contained antibodies against the variable regions of mCOL-1. Sera reactivity of the humanized antibodies was determined by their ability to compete with HuCOL- 1 immobilized on a sensor chip for binding to the mCOL-1 anti- V region antibodies present in the patients' sera. FIG. 4 shows two sets of sensorgrams depicting the binding of immobilized HuCOL- 1 to the serum of patient MB prior to, and after, equilibration with increasing concentrations of competitors HuCOL-1 (FIG. 4A) and HuCOL- 1_SDR (FIG. 4B). The binding of the patient's serum alone (without the competitor) was measured both at the start and the end of each set of binding measurements to show the stability of the HuCOL- 1 surface activity as a function of time. Therefore, the decrease in the magnitude of the binding signal with increasing concentration of the competitor is attributable exclusively to binding inhibition by the competitor. Sensorgrams showing the inhibition by HuCOL- I_AOI-_CDR of the binding of MB serum to the HuCOL-1 surface and the competition of all three antibodies with patient EM's serum were also generated. FIG. 5 shows the competition profiles generated by the three humanized antibodies when they were used to compete with the HuCOL- 1 immobilized on the sensor chip for binding to the anti-V region antibodies to COL-1 present in the sera of patients MB (FIG. 4A) and EM (FIG. 4B). The competition profiles of the antibodies for both patients' sera follow the same pattern, with HuCOL- 1 as the most reactive and HuCOL- 1_SDR as the least reactive. The IC₅₀ values, the concentrations of the competitor antibody required for half-maximal inhibition of the binding of HuCOL- 1 to the patient's serum, were calculated from the competition curves and are given in Table III. A higher IC₅₀ value indicates a decreased reactivity to the serum, suggesting a potentially reduced immunogenicity of the antibodies in patients. Compared with those of HuCOL- 1 for both patients' sera, the IC₅₀ values of HuCOL- 1 A_DICDR are 1.5-fold higher while those of HuCOL- 1 _SDR are approximately 2- to 3.5-fold higher. In addition, the slopes of the competition curves of HUCOL-I_SDR are quite different from that of the HuCOL-1 curves. Compared with HuCOL-1, a 7- and 25-fold increase in the concentration of HUCOL-1_SDR is required to achieve 70% inhibition of the binding of HuCOL- 1 immobilized on a sensor chip to the anti-V region antibodies present in the sera of patients MB and EM, respectively. Table III. Reactivity of the humanized Abs with patients' sera^α Patient MB Patient EM Competitor Antibody (nM) (nM) HuCOL-1 4.9 1.9 HuCOL- l_AbrCDR 6.9 2.9 HuCOL-l_SDR 8.4 6.8

Competitor antibody concentrations required for the half-maximal inhibition of the binding of patients' sera to immobilized HuCOL-1 were calculated. EXAMPLE 7 Humanized COL-1 Antibody Testing in Patients

Patients and Sample Collection Patients with recunent colorectal cancer are assessed to determine the maximum tolerated dose of intravenously administered ¹⁷⁷Lutetium radiolabeled HuCOL- 1_SDR (Mulligan, (1995) Clin. Cancer Res. 1:1447-1454). Colorectal cancer patients are given a test dose of 0.1 mg (intravenous bolus) of HuCOL- 1_SDR and are observed for 30 minutes prior to administration of the ¹⁷⁷Lu-labeled HuCOL-l_SDR. The radiolabeled antibody is given as an intravenous infusion over the course of a one hour time interval. Blood samples are collected prior to and at the end of the infusion, as well as 0.5, 1 and 2 hours following the completion of the infusion. In addition, blood samples are collected daily over the subsequent 7 days. Patients return for a follow-up examination at 3, 6 or 8 weeks. Blood samples are again collected during these visits. Sera are separated and stored at -20°C.

Determination of Patient Humoral Response The sera from the patients are evaluated for the presence of human anti-murine antibodies (HAMA) in response to radiolabeled HuCOL-l_SDR using the SPR-based assay described in Example 6, above. The sera are pre-absorbed with an mCOL-4 monoclonal antibody that recognizes an epitope of CEA which is different from the epitope recognized by the humanized COL-1 monoclonal antibody. Pre-absoφtion using the COL-4 antibody removes circulating CEA and anti-murine Fc antibodies from the sera. To monitor the sera-reactivity of the anti-variable antibodies in the pre- absorbed sera, HuCOL- 1_SDR is coated on the surface of flow cell 1 and a reference protein (HuIgG2a, bovine serum albumin, or rabbit gamma globulin) is immobilized on the surface of flow cell 2. A small, known volume of a patient serum sample us applied to each flow cell using the recently developed sample application technique previously described (Abrantes et al, Anal Chem. 73:2828, 2001). Sensograms to flow cell 1 and flow cell 2 are generated and the response difference between the two cells is plotted for each serum sample, thus providing a measure of the anti-variable region response against HuCOL- 1_SDR in each particular serum sample. Results indicate that the patients' sera have a minimal anti-variable region response against the HuCOL- 1_SDR antibody.

This disclosure provides humanized COL-1 monoclonal antibody. The disclosure further provides methods of diagnosing and treating tumors using these humanized COL-1 antibodies. It will be apparent that the precise details of the methods described may be varied or modified without departing from the spirit of the described disclosure. We claim all such modifications and variations that fall within the scope and spirit of the claims below.

Claims

We claim: 1. A humanized COL-1 monoclonal antibody, comprising: a light chain comprising an amino acid sequence as set forth in SEQ ID NO: 42, wherein a framework residue at position 9 of SEQ ID NO: 42 is substituted with an amino acid from a conesponding Kabat position of a first human gennline sequence, a heavy chain comprising an amino acid sequence as set forth in SEQ ID NO: 43, wherein a heavy chain Complementarity Determining Region (H-CDR)2 residue at position 62 of SEQ ID NO: 43 is substituted with an amino acid from a conesponding Kabat position of a second human germline sequence; and wherein the antibody comprises additional substitutions selected from the group consisting of (a) light chain Complementarity Determining Region (L-CDR) 1 residues at position 27 and at position 37 of SEQ ID NO: 42 are substituted with an amino acid from a conesponding Kabat position of third and fourth human germline sequences, respectively, (b) an L-CDR2 residue at position 57 of SEQ ID NO: 42 is substituted with an amino acid from a corresponding Kabat position of a fifth human germline sequence (c) an L-CDR3 residue at position 94 of SEQ ID NO: 42 is substituted with an amino acid from a conesponding Kabat position of a sixth human germline sequence, and (d) L-CDRl residues at position 27 and at position 37, the L-CDR2 residue at position 57, and the L-CDR3 residue at position 94 of SEQ ID NO: 42 are substituted with an amino acid from a conesponding Kabat position of third, fourth, fifth, and sixth human germline sequences, respectively, wherein the humanized COL-1 antibody retains binding affinity for carcinoembryonic antigen (CEA) and has reduced immunogenicity, as compared to a parental humanized COL-1 antibody.

2. The humanized antibody of claim 1, wherein L-CDRl residues at position 27 and at position 37 of SEQ ID NO: 42 are substituted with an amino acid from a conesponding Kabat position of third and fourth human germline sequences, respectively.

3. The humanized antibody of claim 1 , wherein the L-CDR2 residue at position 57 of SEQ ID NO: 42 is substituted with an amino acid from a conesponding Kabat position of a sixth human gennline sequence.

4. The humanized antibody of claim 1, wherein the L-CDR3 residue at position 94 of SEQ ID NO: 42 is substimted with an amino acid from a conesponding

Kabat position of a sixth human gennline sequence.

5. The humanized antibody of claim 1, wherein L-CDRl residues at position 27 and at position 37, the LCDR2 residue at position 57, the L-CDR3 residue at position 94 of SEQ ID NO: 42 are substituted with an amino acid from a conesponding Kabat position of third, fourth, fifth, and sixth human germline sequences, respectively.

6. The humanized antibody of claim 1 , wherein the residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid.

7. The humanized antibody of claim 1 , wherein the residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine.

8. The humamzed antibody of claim 2, wherein a residue at position 27 of SEQ ID NO: 42 is substituted with a glutamine and the residue at position 37 of SEQ ID NO: 42 is substituted with a leucine.

9. The humanized antibody of claim 3, wherein the residue at position 57 of

SEQ ID NO: 42 is substituted with a serine.

10. The humanized antibody of claim 4, wherein the residue at position 94 of SEQ ID NO: 42 is substituted with a glutamine.

11. The humanized antibody of claim 5, wherein the L-CDRl residue at position 27 of SEQ ID NO: 42 is substituted with a glutamine and the L-CDRl residue at position 37 of SEQ ID NO: 42 is substituted with a leucine, the L-CDR2 residue at position 57 of SEQ ID NO: 42 is substimted with a serine, and the L-CDR3 residue at position 94 of SEQ ID NO: 42 is substituted with a glutamine.

12. The antibody of claim 1 , further comprising a detectable label.

13. The antibody of claim 12, wherein the detectable label comprises a fluorescent tag, a chemiluminescent tag, a hapten, an enzymatic molecule, or a radioisotope.

14. The humanized antibody of claim 1, further comprising an effector molecule.

15. The humanized antibody of claim 14, wherein the effector molecule is a toxin.

16. A composition, comprising a functional fragment of the humanized antibody of claim 1, wherein the functional fragment specifically binds CEA.

17. The composition of claim 16, wherein the fragment comprises an Fv, an Fab, or an F(ab')₂.

18. A pharmaceutical composition, comprising a therapeutically effective amount of the humanized antibody of claim 1 in a pharmaceutically acceptable carrier.

19. A method for treating a subject with a tumor that expresses CEA, comprising: administering a therapeutically effective amount of the humanized antibody of any one of claims 1-15 to the subject, thereby treating the subject with the rumor.

20. A method for detecting a cell that expresses CEA in a subject, comprising: contacting a sample from the subject with the antibody of claim 1, and detecting the presence of a complex of the antibody with CEA, thereby detecting the cell expressing CEA.

21. The method of claim 20, wherein the subject has a tumor.

22. The method of claim 20, wherein the antibody is labeled.

23. The method of claim 20, wherein the sample is a biopsy specimen, autopsy specimen, pathology specimen, or a biological fluid.

24. A method for in vivo diagnosis of a tumor in a subject, comprising: administering to a mammal a diagnostically effective amount of the antibody of claim 12; allowing sufficient time for the antibody to become specifically localized to at least one tumor cell; and detecting the labeled antibody in vivo at a site where the antibody has become localized, thereby diagnosing the tumor.

25. A kit, comprising: a container comprising the antibody of claim 1.

26. The kit of claim 25, further comprising a container containing an antigen, a container containing a secondary antibody conjugated to a detectable label, instructions for using the kit, or any combination thereof.

27. An isolated nucleic acid encoding the antibody of claim 1.

28. A vector comprising a promoter operably linked to the nucleic acid of claim 26.

29. A host cell, transformed with the vector of claim 28.

30. The host cell of claim 29, wherein the cell is a eukaryotic cell.

31. A humanized COL-1 monoclonal antibody, comprising: a light chain comprising an amino acid sequence as set forth in SEQ ID NO: 42 and a heavy chain comprising an amino acid sequence as set forth in SEQ ID NO: 43, wherein a residue at position 9 of SEQ ID NO: 42 is substituted with an aspartic acid, a residue at position 27 of SEQ ID NO: 42 is substituted with a glutamine, a residue at position 37 of SEQ ID NO: 42 is substituted with a leucine, a residue at position 57 of SEQ ID NO: 42 is substituted with a serine, a residue at position 94 of SEQ ID NO: 42 is substituted with a glutamine, a residue at position 62 of SEQ ID NO: 43 is substituted with a glutamine, and wherein the humanized COL-1 antibody retains binding affinity for carcinoembryonic antigen (CEA) and has reduced immunogenicity, as compared to a parental humanized COL-1 antibody.

32. A method of constructing a humanized antibody, comprising: comparing at least one of (a) a complementarity determining region (CDR)1 sequence, (b) a CDR2 sequence, (c) a CDR3 sequence, and (d) a framework sequence, from a non-human antibody that specifically binds an antigen, against a plurality of human germline sequences; selecting at least one of (a) a first human germline sequence at least 50% homologous, but not identical, to the CDRl sequence, (b) a second human germline sequence at least 50% homologous, but not identical, to the CDR2 sequence, (c) a third human germline sequence at least 50% homologous, but not identical, to the CDR3 sequence, and (d) a fourth human germline sequence at least 50% homologous, but not identical, to the non-human framework sequence; and constructing a heavy chain antibody sequence or a light chain antibody sequence comprising at least one of the first, second, third, and fourth human germline sequences, wherein the light chain antibody sequence or the heavy chain antibody sequence can be utilized to produce a humanized antibody with retained binding affinity for the antigen and reduced immunogencitiy compared to a parental antibody.

33. The method of claim 32, wherein the CDRl sequence, the CDR2 sequence, the CDR3 sequence, and the framework sequence are light chain sequences, and wherein the method further comprises comparing at least one of (a) a heavy chain complementarity determining region (H-CDR) 1 sequence, (b) a H-CDR2 sequence, (c) a H-CDR3 sequence, and (d) a heavy chain framework sequence, from a non-human antibody that specifically binds the antigen, against a plurality of human germline sequences; selecting at least one of (e) a fifth human germline sequence at least 50%> homologous, but not identical, to the H-CDRl sequence, (f) a sixth human germline sequence at least 50% homologous, but not identical, to the H-CDR2 sequence, (g) a seventh human germline sequence at least 50% homologous, but not identical, to the H- CDR3 sequence, and (h) an eighth human gennline sequence at least 50% homologous, but not identical, to the heavy chain non-human framework sequence; and constructing a heavy chain antibody sequence with the selected human germline sequences, wherein the light chain antibody sequence and the heavy chain antibody sequence can be utilized to produce a minimally immunogenic humanized antibody with retained binding affinity for the antigen.

34. The method of claim 32, wherein the CDRl sequence, the CDR2 sequence, the CDR3 sequence, and the framework sequence are heavy chain sequences, and wherein the method further comprises comparing at least one of (a) a light chain complementarity determining region (L-CDR) 1 sequence, (b) a L-CDR2 sequence, (c) a L-CDR3 sequence, and (d) a light chain framework sequence, from a non-human antibody that specifically binds the antigen, against a plurality of human germline sequences; selecting at least one of (e) a fifth human germline sequence at least 50% homologous, but not identical, to the L-CDRl sequence, (f) a sixth human gennline sequence at least 50% homologous, but not identical, to the L-CDR2 sequence, (g) a seventh human germline sequence at least 50% homologous, but not identical, to the L- CDR3 sequence, and (f) an eighth human gennline sequence at least 50% homologous, but not identical, to the light chain non-human framework sequence; and constructing a light chain antibody sequence with the selected human germline sequences, wherein the light chain antibody sequence and the heavy chain antibody sequence can be utilized to produce a minimally immunogenic humanized antibody with retained binding affinity for the antigen.

35. The method of claim 32, wherein the method comprises comparing the framework sequence from the non-human antibody that specifically binds the antigen, against a plurality of human framework germline sequences; selecting a human framework gennline sequence at least 50% homologous, but not identical, to the non-human framework sequence; and identifying a non-human framework residue involved in antigen binding that is different from a human residue at a conesponding Kabat position in the selected human framework germline sequence, and substituting the human residue with the non- human residue involved in antigen binding in the humanized antibody.

36. The method of claim 32, wherein the method comprises comparing at least one of the (a) CDRl sequence, (b) CDR2 sequence, and (c) CDR3 sequence, from the non-human antibody that specifically binds the antigen, against a plurality of human germline sequences; selecting a human germline sequence at least 50% homologous, but not identical, to the at least one of (a) the non-human CDRl sequence, (b) CDR2 sequence, and (d) CDR3 sequence; and identifying a non-human CDR residue involved in antigen binding that is different from a human residue at a conesponding Kabat position in the selected human germline sequence, and substituting the human residue with the non-human residue involved in antigen binding in the humanized antibody.

37. The method of claim 32, wherein the method comprises comparing at least one of the (a) CDRl sequence, (b) CDR2 sequence, (c) CDR3 sequence, and (d) framework sequence, from the non-human antibody that specifically binds the antigen, against a plurality of human germline sequences; selecting a human germline sequence at least 50% homologous, but not identical, to the at least one of (a) the non-human CDRl sequence, (b) CDR2 sequence, (c) CDR3 sequence, and (d) framework sequence; and identifying in the at least one of the (a) CDRl sequence, (b) CDR2 sequence, (c) CDR3 sequence, and (d) framework sequence, from the non-human antibody, a CDR or framework residue not involved in antigen binding, and substituting a residue at a cooesponding Kabat position in the selected human germline sequence with a different amino acid.

38. The method of claim 32, comprising comparing (a) a complementarity determining region (CDR)1 sequence, (b) a CDR2 sequence, (c) a CDR3 sequence, and (d) a framework sequence from a non- human antibody that specifically binds an antigen, against a plurality of human germline sequences selecting (a) a first human germline sequence at least 50% homologous, but not identical, to the CDRl sequence, (b) a second human germline sequence at least 50% homologous, but not identical, to the CDR2 sequence, (c) a third human gennline sequence at least 50% homologous, but not identical, to the CDR3 sequence, and (d) a fourth human germline sequence at least 50%> homologous, but not identical, to the non- human framework sequence; and constructing a heavy chain antibody sequence or a light chain antibody sequence comprising all of the first, second, third, and fourth human germline sequences, wherein the light chain antibody sequence or the heavy chain antibody sequence can be utilized to produce a humanized antibody with retained binding affinity for the antigen and reduced immunogencitiy compared to a parental antibody

39. The method of claim 38, wherein the CDRl sequence, the CDR2 sequence, the CDR3 sequence, and the framework sequence are light chain sequences, and wherein the method further comprises comparing (a) a heavy chain complementarity determining region (H- CDR)1 sequence, (b) a H-CDR2 sequence, (c) a H-CDR3 sequence, and (d) a heavy chain framework sequence, from a non-human antibody that specifically binds the antigen, against a plurality of human germline sequences; selecting (e) a fifth human germline sequence at least 50%> homologous, but not identical, to the H-CDRl sequence, (f) a sixth human germline sequence at least 50% homologous, but not identical, to the H-CDR2 sequence, (g) a seventh human germline sequence at least 50% homologous, but not identical, to the H-CDR3 sequence, and (h) an eighth human germline sequence at least 50% homologous, but not identical, to the heavy chain non-human framework sequence; and constructing a heavy chain antibody sequence with all of the selected human germline sequences, wherein the light chain antibody sequence and the heavy chain antibody sequence can be utilized to produce a minimally immunogenic humanized antibody with retained binding affinity for the antigen.

40. The method of claim 38, wherein the CDRl sequence, the CDR2 sequence, the CDR3 sequence, and the framework sequence are heavy chain sequences, and wherein the method further comprises comparing (a) a light chain complementarity determining region (L-

CDR)1 sequence, (b) a L-CDR2 sequence, (c) a L-CDR3 sequence, and (d) a light chain framework sequence, from a non-human antibody that specifically binds the antigen, against a plurality of human germline sequences; selecting (e) a fifth human germline sequence at least 50% homologous, but not identical, to the L-CDRl sequence, (f) a sixth human germline sequence at least 50% homologous, but not identical, to the L-CDR2 sequence, (g) a seventh human germline sequence at least 50% homologous, but not identical, to the L-CDR3 sequence, and (h) an eighth human germline sequence at least 50% homologous, but not identical, to the light chain non-human framework sequence; and constructing a light chain antibody sequence with all of the selected human germline sequences, wherein the light chain antibody sequence and the heavy chain antibody sequence can be utilized to produce a minimally immunogenic humamzed antibody with retained bmding affinity for the antigen.