NO316023B1 - Method for Preparation of a Therapeutically Active Antibody or Post-Therapeutically Active Fragment thereof selective for TAG-72 - Google Patents

Description

This invention relates to the area of immunoglobulin production and modifications of naturally occurring antibody amino acid sequences. Specifically, the invention relates to a method for the production of a therapeutically active antibody or a therapeutically active fragment thereof which are selective for TAG-72

Antibodies are specific immunoglobulin (Ig) polypeptides produced by the vertebrate immune system in response to infection by foreign proteins, glycoproteins, cells or other antigenic foreign substances The sequence of reactions that allow the organism to overcome invasion by foreign cells or for the system to get rid of foreign substances , is at least partially understood An important part of this process is the production of antibodies that bind specifically to a particular foreign substance The binding specificity of such polypeptides to a particular antigen is highly refined, and the bulk of specificities are able to be formed by the individual vertebrate is remarkable in its complexity and variability Millions of antigens are capable of eliciting antibody responses, each antibody being directed almost exclusively to the particular antigen that elicited it

Two main sources of vertebrate antibodies are currently used - in situ formation of mammalian B-lymphocytes, and formation in cell culture of B-cell hybrids Antibodies are formed in situ as a result of the differentiation of immature B-lymphocytes into plasma cells, which occurs in response to stimulation by specific antigens In the undifferentiated B cells, the parts of DNA that code for the different regions of the immunoglobulin chains are separated in the genomic DNA The sequences are assembled sequentially before expression A review of this process has been given by Gough, Trends in Biochem Sei 6,203(1981)

The resulting rearranged gene is capable of expression in the mature B-lymphocyte to produce the desired antibody. However, even when a particular mammal is exposed to only a single antigen, no uniform population of antibodies is formed. The in situ immune response to any one particular antigen is defined by the mosaic of responses to the various determinants present on the antigen Each subset of homologous antibodies is contributed by a single population of B cells - thus in situ formation of antibodies is "polyclonal"

This limited but inherent heterogeneity has been overcome in numerous special cases by the application of hybndoma technology to generate "monoclonal" antibodies in cell cultures at B-cell hybndomas (See Kohler and Milstein, C , Nature 256. 495-497(1975)

In this method, the relatively short-lived or lethal splenocytes or lymphocytes from a mammal injected with antigen are joined with an immortal tumor cell line, thus producing hybnd cells or "hybndomas" that are both immortal and capable of produce the genetically coded antibody in the B cell The hybrids thus formed are isolated into individual genetic strains by selection, dilution and regrowth, and each strain thus represents a single genetic line They therefore produce antibodies that are guaranteed to be homogeneous against a desired antigen These antibodies, which transmit their pure genetic parental inheritance, are called "monoclonal"

Monoclonal antibodies with monospecificity have greatly influenced immunology, and their applicability has already been demonstrated in such scientific fields as biology, pharmacology, chemistry, and others. Such monoclonal antibodies have gained widespread use not only as diagnostic reagents [(see, for example, Immunology for the 80's , Eds Voller, A , Bartlett, A , and Bidwell, D , MTP Press, Lancaster, (1981), but also therapy (see for example Ritz, J and Schlossman, S F , Blood 59,1-11, (1982)]

Monoclonal antibodies produced by hybndomas, while theoretically effective as discussed above and clearly preferred over polyclonal antibodies due to their specificity, suffer from an important drawback. In many applications, the use of monoclonal antibodies produced in non-human animals is severely limited way, where the monoclonal antibodies are to be used in humans Repeated injections of a "foreign" antibody into humans, such as a mouse antibody, can lead to harmful hypersensitivity reactions Such a non-human derived monoclonal antibody, when injected into humans, causes an anti-non-human antibody (ANHA) reaction For a discussion of a specific ANHA reaction caused by the use of mouse-d en verte antibodies, the human-anti-mouse antibody (HAMA) reaction, see Shawler et al, Journal of Immunology 135, 1530-1535(1985)

It is believed that the animal immunoglobulin which has human constant regions will form less ANHA reaction when injected into humans than the animal immunoglobulin which has non-human constant regions. Thus, monoclonal antibodies which have good binding affinities for selected antigens and which have human constant regions, thought to have great potential applicability in immunological diagnostics and therapy for human patients with cancer

Various attempts have so far been made to produce human-derived monoclonal antibodies using human hybridomas. For example, human-human hybridomas [Olsson, L et al, Proe Nati Acad Sei (USA), 77, 5429 (1980)] , human-mouse hybndomas [(Schlom, J , et al (ibid) 77. 6841 (1980)] and several other xenogeneic hybnd combinations have been prepared Human monoclonal antibodies have also been produced by transformation of lymphocytes using Epstein-Barr virus However can such hybridomas potentially contain pathogenic human viruses Alternatively, primary antibody-producing B cells have been immortalized in vitro by transformation with viral DNA Unfortunately yields of monoclonal antibodies from human hybridoma cell lines are relatively low (Vg/ml in humans compared to 100 //g/ ml in mouse hybndoms), and production costs are high

While human immunoglobulins are highly desirable in immunological diagnostics and therapy for human cancer patients, human cloning techniques have not yet reached the stage where human monoclonal antibodies with required antigen specificities can be easily obtained. In addition, for obvious ethical reasons, researchers cannot immunize human subjects with selected toxic or otherwise harmful antigens to form antibodies against the specific antigen This results in major restrictions on immunological diagnostics and therapy of human patients

Production of human-derived monoclonal antibodies is certainly possible,

but it remains ineffective in view of its low reproducibility and the other problems mentioned above [See also, Nature 300. 316-317 (1982)] Accordingly, most monoclonal antibodies are generated from non-human animals

A monoclonal antibody that reacts with high binding affinity to human tumor antigens, but is not recognized as a foreign substance by humans, is highly desirable. One method to overcome this difficulty is to artificially generate an antibody that is very similar to a human antibody and which is not recognized as a foreign substance in the human body, i.e. a chimeric or "humanized" antibody

Typically in chimeric antibodies, the variable region in both light and heavy chains mimics the variable regions in antibodies formed from a species of mammal, while the constant regions are homologous to the sequences in antibodies formed from humans. A clear advantage of such chimeric forms is that f e.g. the variable regions can be conveniently formed from currently known sources by using readily available hybridomas of B cells from non-human host organisms in combination with constant regions formed from e.g. human cell preparations. While the specificity of the variable region is not affected by its source, the constant region that is human is less likely to elicit any immune response from a human subject when the antibodies are injected than the constant region from a non-human source would

A known human tumor antigen is tumor-associated glycoprotein (TAG 72) TAG72 is associated with the surface of certain tumor cells of human origin, specifically the LS174T tumor cell line LS174T [American Type Culture Collection (here ATCC) No CL 188] is a variant of LS180 (ATCC no CL 187) colon adenocarcinoma-Unjen

The karyotype of LS174T is similar to the karyotype of LS 180 with a missing X chromosome in a majority of cells Data have been presented as described in Johnson, V G et al, Cancer Res 46 850-857 (1986), to characterize the TAG72 molecule as et mucin This conclusion is based on the following observations

(a) TAG72 has a high molecular weight (>1 x 10<6>) as shown by its exclusion from a Sepharose™ CL-4B column, (b) the density of TAG72 determined by equilibrium centrifugation in CsCl was 1.45 gm/ml, which indicated a heavily glycosylated glycoprotein (c) TAG72 shows a change in migration after the neuraminidase cleavage, indicating that it is a heavily sialylated molecule with abundant 0-glycoside-linked ohgosaccharides characteristic of mucins, (d) blood group antigens usually found on mucins are found on affinity -purified TAG72, and (e) Chondroitinase ABC cleavage had no effect on TAG72, thus showing that the TAG72 epitope is not expressed on a chondroitin sulfate proteoglycan

Numerous mouse monoclonal antibodies have been developed that have binding specificity for TAG72. One of these monoclonal antibodies, designated B72.3, is a mouse IgGI, produced by hybridoma B72.3 (ATCC no. HB-8108). B72.3 is a first-generation monoclonal antibody developed by to use a human breast carcinoma extract as an immunogen (see Colcher, D et al, Proe Nati Acad Sei (USA) 78, 3199-3203 (1981), and US Patents 4,522,918 and 4,612,282) As used herein, the term "first generation monoclonal antibody a monoclonal antibody produced by using a crude cell extract as an immunogen

Other monoclonal antibodies directed against TAG72 are designated "CC" (colon cancer) CC monoclonal antibodies are a family of second-generation mouse monoclonal antibodies As used herein, the term "second-generation monoclonal antibody" means a monoclonal antibody produced by using as an immunogen an antigen purified with a first generation monoclonal antibody CC monoclonal antibodies were prepared using TAG72 purified with B72.3 A discussion of the method for producing the CC antibodies is set forth in US Patent Application No. 7-073,685 (USPA 7-073,685), wherein the application was filed by Schlom et al on July 15, 1987 and is available to the public from the National Technical Information Service. No. HB 9459), CC83 (ATCC No. HB 9453), CC46 (ATCC No. HB 9458), CC92 (ATCC No. HB 9454), CC30 (ATCC No. HB 9457), CC11 (ATCC No. 9455), and CC 15 (ATCC No. HB 9460)

In the known field, no human antibody has been isolated that binds relatively strongly to TAG72 Consequently, suitable antibodies must be produced genetically

It is known that the effect of an lg molecule depends on its three-dimensional structure, which further depends on its primary amino acid sequence. Thus, changing the amino acid sequence of an lg can affect its activity in the opposite direction. Furthermore, a change in the DNA sequence that codes for lg can , affect the ability of the cell containing the DNA sequence to express, secrete or accumulate lg

USPA 7-073 685 describes that the CC antibodies can be changed to their chimeric form by substituting, for example, human constant regions (Fc) domains instead of mouse constant regions by recombinant DNA techniques known in the art. It is believed that proposals made in USPA 7 -073 685 did not lead to an actual attempt to express any chimeric Ig polypeptide chains, nor to produce Ig activity, nor to secrete and assemble Ig chains into the desired chimeric Igs

It is therefore not at all clear from the art that known recombinant DNA techniques will routinely produce a chimeric animal-human antibody from selected DNA sources that form functional chimeric antibodies that bind specifically to selected human carcinomas and that reduce the onset of ANHA -side effects when injected into humans

Surprisingly, the present invention is able to meet many of these above-mentioned needs and provides a method for producing the desired antibodies. For example, the present invention provides a method for producing a therapeutically active antibody, or a therapeutically active antibody fragment thereof , where the antibody or antibody fragment is able to selectively react with TAG-72 antigen and is obtained from one of the cell lines CH44-1 (ATCC HB 9884), CH44-2 (ATCC HB 9880), CH44-4 (ATCC HB 9877), CH88-1 (ATCC HB 9882), CH88-2 (ATCC HB 9881), CH88-3 (ATCC HB 9876), CH88-4 (ATCC HB 9874), CH 84-1 (ATCC HB 9883), CH84-2 (ATCC HB 9879), CH84-3 (ATCC HB 9878), or CH84-4 (ATCC HB 9875), characterized in that one of the aforementioned cell lines is grown in a suitable medium, after which the antibody or an antibody fragment thereof be isolated

The present invention provides a method for the production of new antibodies for use in in vivo diagnostic measurements and radioimmunoguided surgery

Accordingly, this invention relates to the production of an antibody or antibody fragment comprising a variable region having a heavy chain (VH), said VH being encoded by a DNA sequence that is effectively homologous to the Vh<g>tTAG germline gene (VHaTAG ), in which the variable region binds to TAG72 at least 25% more than the variable region in B72.3 binds to TAG72, and the binding affinities of the antibody and B72.3 are measured by the same technique

A DNA sequence is described which codes for at least part of an antibody heavy chain, said sequence comprising a DNA sequence segment which is effectively homologous to the VHaTAG kimlmje gene (VHaTAG), where the DNA sequence segment codes for i at least part of a heavy-chain variable region (VH)

Furthermore, a DNA sequence is described which comprises

(A) a sequence segment encoding a heavy chain, said sequence segment having

(1) a sequence subsegment that is effectively homologous to the VHaTAG germline gene (VHarTAG), wherein the DNA sequence segment encodes at least a portion of a Vh, and (2) a sequence subsegment that encodes at least a portion of a CH, and ( B) a sequence segment that codes for a light chain, said sequence segment having (1) a sequence subsegment that codes for at least part of an animal light chain variable region (VL), and (2) a sequence subsegment that codes for at least a part of a human light chain constant region (Cl),

where the antibody encoded by the DNA sequence binds to TAG72 at least 25% more than the variable region in B72.3 binds to TAG72, and the binding affinities of the antibody and B72.3 are measured by the same technique

The invention further includes the preparation of the aforementioned antibody alone or conjugated to an imaging marker or a therapeutic agent The invention also includes the preparation of a mixture comprising the aforementioned antibody in unconjugated or conjugated form in a pharmaceutically acceptable, non-toxic, sterile carrier

Description of the drawings

Figure 1 illustrates a basic immunoglobulin structure where the enzymatic cleavage sites are indicated Figure 2 illustrates the nucleotide sequences of VHoTAG VH, CC46 VH, CC49 VH, CC83 VH and CC92 VH Figure 3 illustrates the amino acid sequences of VHaTAG VH, CC46 VH, CC49 VH. CC83 VH and CC92 VH Figure 4a illustrates the nucleotide sequence and Figure 4b illustrates the corresponding amino acid sequence of CC49 VL Figure 5 illustrates the nucleotide sequence and Figure 5b illustrates the corresponding amino acid sequence of CC83 VL Figure 6a illustrates the nucleotide sequence and Figure 6b illustrates the corresponding amino acid sequence of CC92 VL

Figure 7 illustrates the nucleotide sequence of Hind III-Pst

The l fragment isolated from plasmid pGD1

Figure 8 illustrates the plasmid map of pBLUESCRIPT SK(-)

Figure 9 illustrates the plasmid map of pRL101

Figure 10 illustrates a residual function enzyme map of CC49

The L-chain genomic DNA insert in pRL101

Figure 11 illustrates the plasmid map of pRL200

Figure 12 illustrates a residue cleavage enzyme map of the CC83 L-chain genome DNA insert in pRL200 Figure 13 illustrates the nucleotide sequence of the Eco R1-Bam H1 fragment isolated from plasmid pNP9

Figure 14 illustrates the plasmid map of pHH49

Figure 15 illustrates the plasmid map of pHS83

Figure 16 shows the nucleotide sequence of CC49 VH, where the underlined segments show the sequences that are formed by using oligonucleotide polymers on mRNA Figure 17 shows the nucleotide sequence of CC83 VH, where the underlined segments show the sequences that are formed by using oligonucleotide primers on mRNA Figure 18 shows the amino acid sequence of CC49 VH, where the underlined segments show the sequences determined by protein sequencing Figure 19 shows the amino acid sequence of CC83 Vh, where the underlined segments show the sequences determined by protein sequencing Figure 20 shows the results from an SDS polyacrylamide gel, with the results from PNGase F treatment of CC83 antibody Figure 21 illustrates the restriction enzyme map of human y1, y2, y3 and yA

Figure 22 illustrates the plasmid map of pSV2gpt/R/B

Figure 23 illustrates the plasmid map of pSV2gpt-y 1-7.8

Figure 24 illustrates the plasmid map of pSV2gpt-y 1-2.3

Figure 25 illustrates the plasmid map of pSV2gpt-y2

Figure 26 illustrates the plasmid map of pSV2gpt-y3

Figure 27 illustrates the plasmid map of pSV2gpt-y4

Figure 28 illustrates the plasmid map of p49y1-7,8

Figure 29 illustrates the plasmid map of p49y1-2,3

Figure 30 illustrates the plasmid map of p49-γ2

Figure 31 illustrates the plasmid map of p49-y3

Figure 32 illustrates the plasmid map of p49-y4

Figure 33 illustrates the plasmid map of p83y1-7,8

Figure 34 illustrates the plasmid map of p83y 1-2,3

Figure 35 illustrates the plasmid map to p83-y2

Figure 36 illustrates the plasmid map of p83-y3

Figure 37 illustrates the plasmid map of p83-γ4

Figure 38 illustrates the entire reaction for the genetic engineering of barrier genes based on the method of Horton et al, Gene 77, 61 (1989) Figures 39A, 39B and 39C show the biological distribution and retention of CH44-1 throughout the body Figures 40A and 40B show the biological distribution and retention of CH-84-1 throughout the body

The immunoglobulin prepared according to this invention has been developed to meet the problems with mouse monoclonal antibodies described in the prior art. It is characterized in that it has a chimeric structure composed of a heavy chain variable region coded for by DNA formed from VhoTAG

Definitions

As used herein, "immunoglobulm" refers to a tetramer or aggregate thereof whether specific immunoreactive activity is a property or not "Antibodies" refers to such assemblies that have significant specific immunoreactive activity against an antigen, comprising light and heavy chains, with or without covalent bond between them, "non-specific immunoglobulin" ("NSI") means those immunoglobulins which have no known specificity towards an antigen

The basic immunoglobulin structural unit in vertebrate systems is relatively well understood [Edelman, G M , Ann N Y Acad Sei, 190, 5 (1971)] As seen in Figure 1, the units are composed of two identical light polypeptide chains of molecular weight approximately 23,000 daltons and two identical heavy chains with a molecular weight of 53,000-70,000 The four chains are united by disulfide bonds in a "Y" configuration where the light chains are juxtaposed with the heavy chains starting at the opening of the Y and continuing through the diversity region

Heavy chains are classified as y, fj, a, 6, or e, with some subclasses among them. The nature of this chain, as it has a long, constant region, determines the "class" of antibodies for IgA, IgD, IgE, IgG or IgM

Light chains are classified as either k (k) or X (lambda) Each heavy chain class can be bonded with either a k or X light chain Mainly the light and heavy chains are covalently bonded to each other, and the "tail" parts of the two heavy chains are bound to each other by covalent disulfide bonds when the immunoglobulins are formed either by hybndomes or by B cells However, if non-covalent connection between the chains can be carried out with the correct geometry, the collection of non-disulfide bonded chains will still be able to react with antigen

The amino acid sequences run from an N-terminus in the fork-shaped edges of Y to the C-terminus at the base of each chain. In the N-terminus is a variable region, and in the C-terminus is a constant region

The terms "constant" and "variable" are used functionally The variable regions in both light (Vl) and heavy (VH) chains determine binding recognition and specificity towards the antigen The constant region domains in light (Cl) and heavy (Ch) chains transmit important biological properties such as antibody chain association, secretion, transplacental mobility, and complement binding

The variable region is bound in each chain to the constant region by a bond that binds together in the V-gene sequence and the C-gene sequence The binding occurs at the genome level, combining nucleotide sequences via recombination sites The binding sequence is commonly known as a "J" sequence in the light chain gene, which codes for about 12 amino acids, and which is a combination of a "D" sequence and a "J" sequence in the heavy chain gene, which together codes for about 25 amino acids

"Chimeric antibody" for purposes of this invention refers to an antibody that has in its heavy chain a variable region amino acid sequence encoded by a nucleotide sequence derived from a mouse germline gene and a constant region amino acid sequence encoded by a nucleotide sequence derived from a human gene

However, the present invention is not intended to be narrowly limited to the mere substitution of human C-gene sequences encoding immunoglobulin constant regions for mouse C-gene sequences encoding immunoglobulin constant regions. Thus, the present invention is not limited to whether the fusion point is in the variable/constant limit or not

Through various techniques, it is now possible to produce altered chimaenic antibodies, composite chimaenic antibodies and fragmented chimaenic antibodies encoded by nucleotide sequences described herein

"Composite" immunoglobulins comprise polypeptide variable regions which have not so far been found associated with each other in nature. It is not essential whether any of the above mentioned are covalently or non-covalently assembled, as long as the assembly is capable of selectively reacting with a particular antigen or an antigenic family

"Altered antibodies" means antibodies in which the amino acid sequences, especially in the variable region, have been varied. Due to the relevance of recombinant DNA techniques in this invention, one need not be limited to the amino acid sequences of antibodies selected from natural sources, the amino acid sequences of the antibodies can be rearranged to achieve desired characteristics The possible variations are many and range from the change of only one or a few amino acids to the complete redesign of an antibody variable and/or constant region

Changes in the variable region will be made to improve the antigen-binding characteristics Changes in the constant region will mainly be made to improve cell process characteristics, such as complement fixation, interaction with membranes and other effector functions Changes can be made by standard recombinant techniques and also by oligonucleotide-directed mutagenesis techniques [Dalbadie-McFarland, et al Proe Nati Acad Sei (USA) 79, 6409 (1982)]

"Fragments" of immunoglobulins include segments of proteolytically cleaved or recombinantly produced portions of an antibody molecule capable of selectively reacting with a particular antigen or antigen family. Non-limiting examples of such proteolytic and/or recombinant fragments include "Fab", "F (ab')2", and "Fab"', and their proteolytic cleavage sites are shown in Figure 1, as well as "Fv" and "VH" By a "Fv" fragment it is meant that parts of light-chain variable and heavy-chain variable regions are linked together, mainly in their carboxy term Recombinant techniques for producing Fv fragments are set forth in WO 88/01649, WO 88/06630, WO 88/07085, WO 88/07086 and WO 88/09344 With a " VH fragment, it is intended that the variable region has at least a part of a heavy-chain variable region that is capable of being used as an antigen-binding functionality. The production and use of a light-chain variable region (VL) as an antigen-binding functionality has been demonstrated att in an article entitled "Development of biologically active peptides based on antibody structure" by Williams et al, Proe Nati Acad Sei (USA) 86, 5537-5541 (1989)

In this invention, "animal" is intended to include cattle, swine, rodents and primates, including humans and others

"Expression vector" is given a functional definition of any DNA sequence capable of effecting expression of a specified DNA code in a suitable host as included in this term. Currently such vectors are often in the form of plasmids, thus "plasmid " and "expression vector" are often used interchangeably. However, the invention is intended to include such other forms of expression vectors which serve equivalent functions and which from time to time become known in the art

By "transformation" is meant the introduction of DNA into a recipient host cell that changes the genotype and consequently results in a change in the recipient cell

"Host cells" refers to cells that have been recombinantly transformed with vectors constructed using recombinant DNA techniques. As defined herein, the antibody or modification thereof produced by a host cell, in light of this transformation

In descriptions of methods for the isolation of antibodies from recombinant hosts, the terms "cell" and "cell culture" are used interchangeably to denote the source of antibody unless clearly specified otherwise. In other words, recovery of antibody from the "cells" may mean either from centrifuged down, whole cells or from the cell culture containing both the medium and the suspended cells

Abbreviations

Nucleic acids, amino acids, peptides, protecting groups, active groups and similar parts, when abbreviated, are abbreviated according to IUPAC IUB (Commission on Biological Nomenclature) or the practice of the applicable fields The following are examples

Reagents

EDTA Ethylenediaminetetraacetic acid

SDS Natnumdodecyl sulfate

Nucleic acids

RNA Ribonucleic acid

DNA Deoxynbonucleic acid

Nitroqenholdiqe bases

Both DNA and RNA contain long chains of phosphoric acid, a sugar and nitrogenous bases DNA is a double-stranded helix, where the sugar is 2-deoxynbose, while RNA is single-stranded, while the sugar is D-nbose The four nitrogenous bases that characterize DNA nucleotides are bound together in complementary pairs by hydrogen bonds so they form the double helix in DNA adenine is bound to thymine, guanine is bound to cytosine In RNA, uracil is substituted for thymine in the arranged DNA pairs

Amino acids

Variable area

DNA that codes for the heavy chain consists of a Vh gene sequence, a Dh gene sequence and a JH gene sequence DNA that codes for the light chain consists of a V|_ gene sequence and a JL gene sequence

Vh gene sequence

The present invention is directed towards the production of selected chimeric antibodies in which the VH region is coded for by a DNA sequence formed from a germline gene that is specifically reactive against TAG72 (VHarTAG), the sequence of which is presented in Figure 2. The chimeric antibodies are selected on the basis of their ability to bind TAG72, namely where the variable region binds to TAG72 at least 25% more than the variable region in B72.3 binds to TAG72 Mainly the binding affinities of the chimeric antibody and B72.3 are measured by the same technique

Examples of techniques for measuring antibody binding affinity are presented in

following references Scatchard G , Annals of the N Y Acad of Sciences 5_1, 660

(1949), Steward, M W , and Petty, R E , Immunology 23, 881 (1972), Muraro, R , et al , Cancer Research 48, 4588 (1988), and Heyman.B , J of Immunol Methods 68, 193- 204 (1984)

One skilled in the art will understand that the nucleotide sequence of (and the amino acid sequences encoded by) VhøTAG that can be used in the present invention is intended to include effectively homologous nucleotide sequences and corresponding amino acid sequences "Effectively homologous" refers to identity or near identity of nucleotide or amino acid sequences Thus, in this description, it will be understood that a small degree of similarity may exist within homologous sequences and that all sequences showing at least 80% homology are considered equivalent

Homology is expressed by the fraction or percentage of "matching" bases (or amino acids) after two sequences (possibly of different lengths) have been aligned. The term alignment is used in the sense defined by Sankoff and Kruskal in Chapter 1 of their book, The Time Warps. Stnnq Edits. and Macromolecules The Theory and Practice of Sequence Companson, Addison-Wesley, Reading, MA,

(1983) Briefly, the two sequences are aligned by maximizing the number of matching bases (or amino acids) between the two sequences with the insertion of a minimal number of "blank" or "null" bases into each sequence to produce the maximum overlap

As is understood in the art, nucleotide mismatches can occur in the third or the uncertain base of the codon without causing amino acid substitutions in the final polypeptide sequence. Also, small nucleotide modifications (eg substitutions, insertions or deletions) in certain regions of the gene sequence can be tolerated and considered insignificant when any such modifications result in changes in amino acid sequence that do not change the functionality of the final product

It has been shown that chemically synthesized copies of whole or part of gene sequences can replace the corresponding regions in the natural gene without loss of gene function

Homologues of specific DNA sequences can be identified by those skilled in the art using a nucleic acid cross-hybridization test under conditions of stringency well understood in the art [as described in Nucleic Acid Hybridization, Hames and Higgens (eds ), IRL Press, Oxford, UK (1985)] Given two sequences, algorithms are available to calculate their homology eg Needleham and Wunsch, J Mol Biol ,48, 443-453 (1970), and Sankoff and Kruskal (cited above) pp 23-29 Also commercial services are available to perform such comparisons, eg Intelligenetics, Inc (Palo Alto, CA)

Dh and Jh gene sequences

The DH and JH gene segments exist in different types, although the type of D or J gene segment selected is not essential to the invention That is, DH and Jh can be formed from any animal Preferred animals include mice and humans Obviously human Dh and/or JH gene segments particularly preferred, but the invention is not limited in this way if a D or J gene segment from another animal species provides an important property, i.e. increased binding to TAG72

Examples of mouse DH and Jh sequences are set forth in "Organization, Stnjcture, and Assembly of Immunoglobulm Heavy Chain Diversity DNA Segments", Kurosawa and Tonegawa, J Exp Med 155, 201 (1982), and "Sequences of the Joining Region Genes for Immunoglobulm Heavy Chams and Their Role in Generation of Antibody Diversity", Gough and Bernard, Proe Nati Acad Sei (USA), 78, 509 (1981)

Examples of human DH and Jh sequences are set forth in an article entitled "Human Immunoglobulm D Segments encoded in Tandem Multigenic Families" by Siebenlist et al in Nature 294, 631 (1981), and examples of human JH sequences are set forth in "Structure of the Human Immunoglobulm // Locus Characterization of Embryonic and Rearranged J and D Genes" by Ravetch et al, Cell, 27, 583(1981)

\ A and J \ gene sequences

Essentially, any VL and JL gene sequence can be used, which encodes a portion of a VL that is complementary to the VH encoded by a nucleotide sequence that is effectively homologous to the VhctTAG By "complementary" is meant a VL that binds to the VH and which yields an antibody variable region having binding affinity of at least 25% greater than B72.3, as measured by any standard technique for measuring binding affinity constants

The type of VL and JL gene segment selected is not essential to the invention That is, VL and JL can be formed from any animal Preferred animals include mice and humans Obviously human VL and/or JL gene segments are particularly preferred, but the invention is not limited to this if a JL gene segment from another species gives an important property, i.e. increased binding to TAG72

Mouse J1 sequences are set forth in an article entitled "The nucleotide Sequence of a 5.5-kilobase DNA Segment Containing the Mouse Kappa Immunoglobulm J and C Region Genes" by Max, et al in J Biol Chem 256. 5116-5120 ( 1981) Human JL sequences are set forth in an article entitled "Evolution of Human Immunoglobulm K J Region Genes" by Heiter et al in The Journal of Biological Chemistry 357 (2), 1516-1522 (1982)

Formation of variable areas

Given the above description, it now becomes possible to form numerous specific embodiments of antibody variable regions within the scope of the present invention, i.e., which have effectively homologous VH sequences with VnaTAG and which bind to TAG72 at least 25% more than the vanable region in B72 ,3 binds to TAG72, and the binding affinities of the antibody and B72.3 are measured by the same technique Several possible techniques are presented below

Naturliq- produced variable areas

In response to an immunogen, TAG72, an immunized animal will increase selected antibody-producing B cells The variable region in antibodies produced by the B cells will be encoded by rearranged germline heavy and light chain DNA For example, the rearranged germline heavy chain will include V , D and J gene segments, including the leader sequence, as well as any mtrons that may be further removed The light chain encoding DNA will include the V and J gene segments, including the leader sequence, as well as any mtrons that may be further removed

The variability may be the result of somatic mutations occurring in a B cell during productive rearrangement of VhoTAG These somatic mutations are nucleotide changes that may or may not result in an amino acid change that alters the activity against TAG72 in the productively rearranged VH

Mapping techniques

Monoclonal or polyclonal antibodies can be screened to determine which of said antibodies selectively bind to TAG72. Such screening can be performed by any of a number of well-known methods, such as solid-phase radioimmunoassay, enzyme-linked immunosorbent assays, rosette assays, and blocking assays. The above-described methods is well known in the field

The nucleotide sequence for coding variable regions of antibodies produced from the productive rearrangement of VhøTAG has now been obtained. In addition to the nucleotide sequence of VHoTAG, Figure 2 also shows the nucleotide sequences that code for the heavy chain variable regions in CC46, CC49, CC83 and CC92 antibodies respectively Figure 3 shows the amino acid sequences of VhctTAG Vh, CC46 VH, CC49 VH, CC83 VH and CC92 VH, which correspond to the nucleotide sequences presented in Figure 2 A comparison of the nucleotide and amino acid sequences of VhctTAG VH, CC46 VH, CC49 VH, CC83 VH and CC92 VH shows a very striking feature , namely that the chains are very similar

The relative similarity of the DNA encoding the CC46 VH> CC49 VH, CC83 VH and CC92 Vh regions, especially in the 5' flanking segment, proves that these DNA sequences are formed from VHaTAG Somatic mutations occurring during productive rearrangement of the VH region gene to expressed in a B cell, gives rise to some nucleotide changes that may or may not result in a homologous amino acid change between two productively rearranged VHaTAG producing hyb domains

The nucleotide sequences and the corresponding amino acid sequences for CC49 VL are shown in Figures 4a and 4b respectively The nucleotide sequences and the corresponding amino acid sequences for CC83 VL are shown in Figures 5a and 5b respectively The nucleotide sequences and the corresponding amino acid sequences for CC92 VL are shown in Figures 6a and 6b respectively

Probe techniques

Other antibodies encoded by DNA formed from VhctTAG can be generated by using VhctTAG as a hybridization probe. Mainly, a probe made from DNA or RNA from VHaTAG or rearranged genes containing the recombined VHaTAG should be used by those skilled in the art to find homologous genes in unknown hybrid domains. Such homologous antibodies will have a DNA sequence whose mRNA hybridizes with the probe in all or part of the VhctTAG germline gene and its flanking region By "flanking" regions is meant to include the DNA sequences from the 5' end of VHaTAG to the 3'- the end of the gene in the upstream direction, and from the 3-end in VhctTAG to the 5' end of the gene in the downstream direction

Rationally synthesized variable areas

An even further approach is the rational synthesis of altered variable regions in the antibodies described here, as well as antibodies discovered via probing. Such an approach has several potential advantages. Namely, a researcher should not have to survey immunized host animals by first trying to select the antibodies that bind to TAG and then to select the antibodies that specifically have Vh regions encoded by DNA formed from VHaTAG

Mutagenic techniques

The VH and/or VL gene segments can be "changed" by mutagenesis Examples of techniques include the addition, deletion or non-conservative substitution of a limited number of different nucleotides or the conservative substitution of many nucleotides, provided that the correct reading frame is maintained

Substitutions, deletions, insertions or any subcombination can be combined to end up with a final construct Since there are 64 possible codon sequences but only 20 known amino acids, the genetic code is degenerate in the sense that different codons can give the same amino acid However, the code exactly for each amino acid, thus there is at least one codon for each amino acid, i.e. each codon gives a single amino acid and no other It will be clear that during translation the correct reading frame must be maintained in order to achieve the correct amino acid sequence in the finally produced polypeptide

Techniques for additions to predetermined amino acid sites having a known sequence are well known Examples of techniques include oligonucleotide-mediated, site-directed mutagenesis and polymerase chain reaction

Techniques for deletions at predetermined amino acid sites having a known sequence are well known. Examples of techniques include oligonucleotide-mediated site-directed mutagenesis and the polymerase chain reaction.

Techniques for substitutions at predetermined amino acid sites having a known sequence are well known

Examples of techniques include site-specific mutagenesis and the polymerase chain reaction technique

Oligonucleotide site-directed mutagenesis essentially involves hybridizing an oligonucleotide encoding a desired mutation with a single strand of DNA containing regions to be mutated and using the single strand as a template for extension of the oligonucleotide to produce a strand containing the mutation This technique is in various forms described by Zoller, MJ and Smith, M , Nuc Acids Res 10, 6487-6500 (1982), Norns, K , Norns, F , Christiansen,!, and Fm, N , Nuc Acids Res H, 5103-5112 ( 1983), Zoller.M J and Smith.M , DNA 3, 479-488 (1984), Kramer.W , Schughart.K and Fntz.W J , Nuc Acids Res 10,6475-6485(1982)

Polymerase chain reaction (PCR) mainly involves exponential amplification of DNA in vitro using sequence-specific oligonucleotides The oligonucleotides can incorporate sequence changes if desired The polymerase reaction technique is described in Mullis and Faloona, Meth Enz 155, 335-350

(1987) Examples of mutagenesis using PCR are described in Higuchi et al, Nucl Acids Res 16, 7351-7367 (1988), Ho et al, Genes 77, 51-59

(1989) and "Engineering Hybrid Restnction Genes Without the Use of Restnction Enzymes Gene Splicing by Overlap Extension", Horton et al, Gene 77, 61 (1989)

Alteration of the antibody-vanable regions may be of particular use in the therapeutic use of monoclonal antibodies Currently, when a chimeric antibody comprising a complete mouse-vanable region is injected into a human, the human body's immune system recognizes the mouse-vanable region, albeit less than a complete mouse antibody, as foreign and produces an immune reaction against it Thus, upon further injections of the mouse antibody or the chimeric antibody into humans, its effectiveness is considerably reduced by the action of the body's immune system against the foreign antibody Consequently, changes in the mouse VH and VL areas can reduce the human immune reaction to the altered antibody

Recombinant techniques

The antibodies can be constructed by recombinant techniques. In other words, because the nucleotide sequences for various VH and V[_ coding regions are now provided, one skilled in the art can in vitro produce a complete gene encoding the heavy and light chain variable regions

The engineered gene can be produced genetically in which selected Dg and JH gene segments are in functional combination with a selected Vh gene segment, i.e. the VHaTAG segment, or the VH gene segment in CC49 or CC83

For example, the engineered heavy chain coding DNA will include Dh and JH gene sequences contiguous with the 3' end of the germline VHcrTAG gene segment, thereby completing the CDR3 and frame (FR) 4 of the VH region A leader sequence may be present but may later is removed

Depending on the light chain used, it may also be necessary to provide a constructed light chain coding DNA. Thus, a DNA gene will comprise a VL gene segment in functional combination, e.g. contiguous with a JL gene segment, including the leader sequence which can further be removed, the JL gene segment will usually depending on whether the light chain is of the lambda or kappa system the J region sequence is contiguous with the end of the VL exon to complete FR4 in the VL region Such construction can be accomplished by the techniques used to construct the Vh gene

The engineered gene can be produced by conventional recombinant techniques, e.g. to make a gene insert in a plasmid ready for expression. The plasmids can then be expressed in host cells. Examples of recombinant biological techniques are presented below

When providing a fragment encoding either the light-chain or heavy-chain variable region, it will usually be desirable to include all or part of the intron downstream from the J region, particularly where the vanable region is formed from the host in which the spliced gene is to be expressed Where the intron is conserved, it will be necessary that there are functional splice acceptor and donor sequences t intron termini The intron between J and the constant region of the spliced gene may be primarily the intron sequence associated with (1) the constant region , (2) the J region, or (3) parts of each The latter may be for convenience where there is a suitable residue site in the introns from the two sources It may be necessary to provide adapters to join the intron to the constant region I in some cases all or part of the intron can be modified by deletion, nucleotide substitution(s) or insertion to make handling, expression or the like easier Advantageous is, an effective amount of the intron should be present to contain an enhancer that is functionally active with the naturally occurring promoter

Alternatively, it may be desirable that the spliced gene is free of the intron between the J gene and the C gene Thus the 3' terminus of the J gene will be adjacent to the 5' terminus of the C gene One can use an exonuclease and by using different digestion time, varying 3' termini can be provided which can then be used for binding to the constant region and selection can be made for a functional product in a number of ways, or by splicing with overlap extension using polymerase chain reaction technology, see Horton et al, supra In this case, an artificial promoter, which does not need to be functionally active with the enhancer, will mainly be used

In a preferred embodiment, the genes encoding the VH and VL regions can be changed by replacing at least parts of the complementary determining regions (CDRs) in the light or heavy chain vanable regions of the antibody with analogous parts of CDRs from an antibody with different specificity Example of a technique that replaces CDRs is described in European published patent application 0 239 400, by Gregory Winter, and in PCT application WO 88/09344 by Huston et al In an altered antibody according to the present invention, will only the CDRs in the antibody be foreign to a human body, and this should minimize side effects if used for human therapy However, human and mouse framework regions have characteristic features that separate human from mouse framework regions Thus, an antibody comprised of mouse CDRs in a human frame well not be anything more foreign to the body than a genuine human antibody

The nucleotide sequences corresponding to the VH amino acid sequences in VHaTAG, CC46, CC49, CC83 and CC92, as well as in the CC49, CC83 and CC92 VL gene segments, are given Accordingly, it is assumed that the CDRs from the antibodies prepared according to the present invention can be grafted mn in the framework regions of the human antibody Mainly the CDR regions from a human VH or VL region can be replaced by CDRs from the VH or VL regions in antibodies according to the present invention Examples of human antibodies from which the framework parts can be used include human plasmacytoma NEWN, [Jones et al, "Replacing the complementarity-determining regions in a human antibody with those from a mouse", Nature 321, 522-525 (1986)], publicly available from Dr Greg Winter, and various other human VH and VL genes available from Dr Terrence Rabbitts, both researchers who are from the Medical Research Council, 20 Park Crescent, London W1N4AL

The determination as to what constitutes a CDR and what constitutes a framework region can be made on the basis of the amino acid sequences of a selected lg as indicated in Kabat et al, Sequences of Proteins of Immunological Interest. 4 ed

(1987), US Dept of Health and Human Services, NI H

The four framework regions broadly adopt a fi-sheet conformation and the CDRs form loops that connect and in some cases form part of the fi-sheet structure

Furthermore, not all amino acid residues in the loop regions are accessible to solvent, and in one case the amino acid residues in the framework regions are involved in antigen binding [Amit, A G , Manuzza, R A , Phillips, S E V , and Poljak, R J , Science 233, 747-753, (1986) ]

It is also known that the variable regions in the two parts of an antigen-binding site are held in correct orientation by meter chain, non-covalent interactions. These may involve amino acid residues in the CDRs

Thus, to transfer the antigen-binding capacity of one variable region to another, it may not always be necessary to replace all the CDRs with the complete CDRs from the donor variable region. It may be necessary to transfer only those residues that are necessary for the antigen binding site, and this may involve the transfer of framework region residues as well as CDR residues

It is thus clear that simply replacing one or more CDRs with complementary CDRs need not always result in a functionally altered antibody However, given the explanations set forth in European Published Patent Application No. 0 239 400, it will be well within the competence of professionals in the field, either by performing routine experimentation, or by trial and error testing to obtain a functionally altered antibody

Preferably, the variable regions of both the heavy and light chains are altered by at least partial CDR replacement and, if necessary, by partial framework region replacement and sequence alteration. Although the CDRs may be generated from an antibody of the same class or even subclass of the antibody from which the framework regions are formed, it is envisaged that the CDRs will be formed from an antibody of a different class and preferably from an antibody of a different species

Composite variable areas

Essentially, the V gene that codes for V is the same V gene that codes for VL naturally combined with VH by choice For example, the VH gene that codes for the VL regions in CC49 and CC83 is advantageously used when using the V gene which encode Vh in CC49 and CC83 respectively

Surprisingly, because the VH regions of the antibodies produced according to the present invention are encoded by VH genes formed from VHaTAG, composite antibodies can be advantageously formed. In other words, the VH region of an antibody produced according to the present invention can invention, is suitably combined with the VL region of another antibody prepared according to the present invention. Although the amino acid sequences of the CC49 and CC83 heavy chains are superficially close, it would be expected that a change of a few or even an amino acid will drastically affect the binding function of the antibody, i.e. the resulting antibodies are mainly assumed to be a non-specific immunoglobulin (NSI), i.e. - lacking antibody character, (see European published patent application no. 0 125 023)

Quite surprisingly, it has now been found that an antibody having the required VH according to this invention does not need to be recombined only with a VL from the same naturally occurring animal antibody F e.g., as set forth in the examples, it is possible to produce a chimeric antibody having a heavy chain from a VH from CC83 and a light chain from a VL from CC49, while the composite antibody thus formed has a binding specificity 25% greater than the binding affinity of B72.3 to TAG72

Constant areas

Tunqkiede (Ch) area

The Cn regions can be of different human isotypes, ie IgG (ie, IgGi, lgG2, lgG3 and lgG4), IgA, IgD, IgM, as well as the different subtypes of the individual groups

For a discussion of the human y1, see Ellison et al, "The nucleotide sequence of a human immunoglobulin C-gamma-1 gene", Nucl Acid Res 10, 4071-4079 (1982), Takahashi et al, "Structure of human immunoglobulin gamma genes Implications for evolution of a gene family", Cell 29, 671-679 (1982) For a discussion of the human gamma 2 (y2), see Krawmkel et al, "Companion of the hinge-coding segments in human immunoglobulin gamma heavy genes and the linkage of the gamma 2 and gamma 4 subclass genes, EMBO J1,403-407 (1982), Ellison et al, "Linkage and sequence homology of two human immunoglobulin gamma heavy cham constant region genes, Proe Nat Acad Sei ( USA) 79, 1984-1988 (1982), Takahashi et al, mfra For a discussion of human gamma 3 (y3), see Krawmkel et al, mfra and Takahashi et al, mfra For a discussion of human gamma 4 (y4), see Ellison et al "Nucleotide sequence of a human immunoglobulin C-gamma-4 gene, DNA i, 11-18 (1981), Krawmkel et al mfra, and Takahasi et al, mfra

For a discussion of the human mu, see Rabbitts et al, Human Immunogobuhn Heavy Chain Genes Evolutionary Comparisons of C/ j, Cd, and Cy genes and Associated Switch Sequences", Nucl Acid Res 9,4509-4524

For a discussion of the human alpha, see Flanagan et al, "Mechanisms of Divergence and Convergence of the Human Immunoglobulm alpha 1 and alpha 2 Constant Region Gene Sequences", Cell 36,681-688 (1984)

For a discussion of the human delta, see White et al, "Human Immunoglobulm D Genomic Sequences of the Delta Heavy Chain", Science 228.

733-737(1985)

For a discussion of the human epsilon, see Max et al, "Duplication and Deletion in the Human Immunoglobulm e Genes", Cell 29, 691-699 (1982)

Lettkiede (CO area

The CL region can be human kappa (k) or human lambda (X)

For a discussion of the human k, see "Cloned Human and Mouse Kappa Immunoglobulm Constant and J Region Genes Conserve Homology in Functional Segments", Heiter et al, Cell 22, 197-207, November (1980)

For a discussion of the human X, see "Processed Genes A Dispersed Human Immunoglobulm Gene Beanng Evidence of RNA-Type Processing", Hollis et al, Nature 296, 321-325 (1982)

The CH and/or Ci_ gene segments can be "changed" by mutagenesis Examples of techniques include addition, deletion or non-conservative substitution of a limited number of different nucleotides or the conservative substitution of many nucleotides, provided that the correct reading frame is maintained In addition, the entire areas in the protein are changed, e.g. by substituting CH2 with CH3 This substitution can be done at DNA level by insertion, deletion or substitution of entire sequence exons

Construction of antibodies

Immunization beds

The first technique for producing antibodies, either monoclonal or polyclonal, having Vh regions encoded by DNA formed from VHaTAG, is to immunize a host animal with purified TAG72 Examples of methods for immunizing a host animal with TAG72 are set forth in US patents 4 522,918 and 4,612,282, using a human mammary gland extract as an immunogen, and US Patent Application 7-073,685 (which is in the public domain), using TAG72 purified with B72.3 as an immunogen

Thereafter, monoclonal or polyclonal antibodies produced from the immunoassay are mapped to determine which of said antibodies selectively bind to TAG72. Such mapping can be performed by any number of a variety of well-known methods, such as solid phase radioimmunoassay, enzyme-linked immunosorbent assays, rosette assays and blocking assays The methods described above are well known in the art

Synthesis of amino acid sequences

Immunoglobulins produced according to the present invention can be synthesized from their amino acids Suitable techniques are the Mernfield solid phase method as described in J Amer Chem Soc 85, 2149-2154 (1963) This solid phase method for synthesizing sequences of amino acids is also described on page 1 -4 in a book by Stewart and Young, Solid Phase Peptide Synthesis (W H Freeman and Co , San Fransisco, 1969)

Construction of DNA

DNA that codes for Vh and Vi_

DNA encoding the antibody heavy and light chains can be obtained from a variety of sources known to those of ordinary skill in the art, eg genomic DNA, cDNA, synthetic DNA or a combination of these

Cells encoding the desired sequence can be isolated, and genomic DNA fragmented by one or more restriction enzymes Genomic DNA may or may not include naturally occurring mtrons The resulting fragments can then be cloned and mapped using a heavy chain J region (Jh) probe for the presence of the DNA sequence that codes for the polypeptide sequence of interest DNA fragments isolated by preparative agarose gel electrophoresis are ligated Recombinant plaques from the collections are mapped with a mouse-Jn probe

DNA can also be obtained from a cDNA pool Messenger RNA encoding heavy or light chain is isolated from a suitable source, either mature B cells or a hybridoma culture, by using standard techniques for RNA isolation, and by using oligo-dT cellulose chromatography to segregate poly-A mRNA Poly-A mRNA can be further fractionated to obtain sequences of sufficient size to encode the amino acid sequences of the light or heavy chain of the desired antibody as necessary

A cDNA pool is then prepared from the mixture from mRNA using a suitable primer, preferably a nucleic acid sequence characteristic of the desired cDNA. Such a primer can be synthesized based on the amino acid sequence of the antibody. Alternatively, cDNA from unfractionated poly-A mRNA from a cell line that produces the desired antibody or poly-dT is also used. The resulting cDNA is optionally size-fractionated on a polyacrylamide gel and then extended with e.g. dC residues for base part matching with pBR322 or another suitable cloning vector that has been cleaved with a suitable residue removal enzyme such as Pst I , and extended with dG residues Alternative ways of creating cloning vectors containing cDNA by using other tails and a different cloning vector residue can for example also be used, but the above is a standard and a preferred choice A suitable host cell strain, typically Eschenchia coli (E coli), the base-pair adapted cloning vectors are transformed, and the successful transformants are identified using, for example, ampicillin or tetracycline resistance or other phenotypic characteristics located on the cloning vector plasmid

Successful transformants are picked up and transferred to microtiter plates or other support for further growth and preservation Nitrocellulose filter extracts of these adult cultures are then probed with suitable nucleotide sequences containing bases known to be complementary to desired sequences in the cDNA Several types of probe can be used, preferably synthetic single -stranded DNA sequences labeled by kinased with γ-<32>P ATP Cells fixed to the nitrocellulose filter, lysed, DNA denatured, and then fixed before reaction with kinase-treated probe Clones that hybridize successfully are detected by contact with a photographic plate , then plasmids from the adult colonies are isolated and sequenced using methods known in the art to verify that the desired parts of the gene are present

The desired gene fragments are excised and aligned to ensure appropriate reading frame with the control segments when inserted into appropriate expression vectors. Typically, nucleotides are added to the 5' end to include an initiation signal and an appropriately positioned restriction endonuclease site

Because the inventors have provided the nucleotide sequences in VHcrTAG, DNA can also be synthesized synthetically, for example by using an Applied Biosystems™ Model 380A DNA Synthesizer, and constructed by standard techniques

Finally, an example of a technique for using a combination of the above-mentioned techniques is overlapping extension splicing using polymerase chain reaction technology, see Horton et al, supra. Essentially, a synthetically synthesized primer, which has a so-called "wagging tail" can " is inserted together with a selected sequence, e.g. genomic DNA. The sequences are then amplified and spliced together

DNA that codes for CH and Ci

The DNA fragment encoding the amino acid sequence of the human constant region can be obtained by mapping the chromosomal DNA in cells producing human immunoglobulin

Vectors

The desired DNA fragment can be placed in a biologically functional expression carrier which can contain suitable control sequences not present in the selected DNA fragment By "biologically functional" it is meant that the expression carrier ensures replication and/or expression in a suitable host, either by maintaining as an extrachromosomal element, or by integration mn into the host genome A large number of vectors are available or can be readily prepared, and are well known to those skilled in the art

A number of plasmids have been described, such as those described in European published patent applications 0036776, 0048970 and 0051873, which already contain a promoter in reading frame with the gene and compatible with the proposed host cell

The vectors and methods described herein are suitable for use across a wide range of microorganisms, either prokaryotic or eukaryotic, which are amenable to transformation. The plasmid will be able to replicate in the microorganism, particularly a bacterium

Mainly, plasmid vectors containing the appropriate promoters, which can be used by the microbial organism for expression of its own protein, also contain control sequences, nbosome binding sites, and transcription termination sites. Mainly, the replicon and control sequences generated from species compatible with the host cell are used in connection with these hosts

Smaller or larger SV40 fragments can also be used, provided that the approximate 250 base pair (bp) sequence extending from the Hind III site towards the Pvu II site located at the origin of replication of the virus is included. Furthermore, it is also possible and often desirable to use promoter or control sequences normally associated with the desired gene sequence, provided that such control sequences are compatible with the host cell systems

Finally, the plasmid should preferably have a gene, a marker gene, which is capable of conferring a phenotypic characteristic enabling the selection of host cells containing the expression vector. Particularly useful is a gene conferring survival selection. Survival selection can be achieved by conferring resistance to a weight-inhibiting substance. or by imparting a growth factor ability to a bacterium that lacks such ability

Mainly prokaryotes are preferred For example, pBR322 is a plasmid formed from an E coli species, see [Bolivar, et al, Gene 2, 95 (1977)] particularly useful pBR322 contains genes for ampicillin and tetracychn resistance and thus provides an easy way to identify transformed cells on

While these prokaryotes are the most commonly used, other microbial strains that may be used include E coli strains such as E coli B, E coli K12 strain 294 (ATCC No. 31466) and E coli X1776 (ATCC No. 31537), E coli W3110 (F "' r prototropic, ATTC No. 27325), bacilh such as Baallus subtitus, and other enterobactenaceae such as Salmonella typhimunum or Serratia marcescens, and various Pseudbmonas species may be used These examples are intended to be illustrative only

In addition to prokaryotes, eukaryotic microbes can also be used Saccharomyces ceævisiae, or common baker's yeast is the most commonly used among eukaryotic microorganisms, although a number of other strains are commonly available

For expression in Saccharomyces, plasmid YRp7, for example [Stinchcomb, et al, Nature 282, 39 (1979), Kingsman et al, Gene 7, 141 (1979), Tschemper, et al, Gene 10, 157 (1980)] commonly used This plasmid already contains the trpl gene which provides a selection marker for a mutant strain of yeast cell lacking the ability to grow in tryptophan, for example ATCC No. 44076 or PEP4-1 [Jones, Genetics 85,12 (1977)] Presence of trpl the lesion as a characteristic of the yeast host cell genome, then provides an efficient environment for the detection of transformation by growth in the absence of tryptophan

Any plasmid vector containing a yeast cell-compatible promoter, origin of replication and termination sequence is suitable for use in yeast cells. Suitable promoter sequences in yeast cell vectors include the promoters for 3-phosphoglycerate kinase [Hitzeman, et al, J Biol Chem 225, 2073 (1980)] or other glycolytic enzymes [ Hess, et al, J Adv Enzyme Reg 7, 149 (1968) Holland et al, Biochemistry 17, 4900 (1978)]

For use in mammalian cells, the control functions on the expression vectors are often provided by viral material. For example, commonly used promoters are formed from polyoma, Adenovirus 2, and most often Simian Virus 40 (SV40). The early and late promoters for SV40 virus are particularly useful because both are easily obtained from the virus as a fragment that also contains the SV40 virus origin of replication [Fiers, et al, Nature 273, 113 (1978)]

For example, pSV2neo contains a gene for both ampicillin resistance and neomycin resistance, which is under the control of an SV40 promoter. Thus, pSV2neo provides an easy way to identify cells transformed with genes for both the animal-vanable region and the human constant region

Preparation of chimeric DNA

Genes that code for the heavy chain or the light chain will be constructed by joining the 5' end of a DNA fragment that codes for the constant region to the 3' end of a DNA fragment that codes for the variable region the DNA sequence that codes for the antibody amino acid sequence, can be obtained in conjunction with the promoter and replication site from genomic DNA To the extent that the host cells recognize the transcriptional regulatory and translational initiation signals associated with the heterologous genes, the region 5' and 3' of the variable region coding sequence can be retained with the variable region coding sequence and used for transcription and translation initiation regulation The non-coding region 3' to the constant region may be conserved for its transcriptional initiation regulatory sequences, such as termination and polyadenylation When referring to 5' or 3' for a double strand, it is meant to mean the direction of transcription, 5' being upstream from 3'

The intron sequence between the variable region of each respective strand can be joined to the corresponding human constant DNA fragment at any suitable restriction site. When providing a fragment encoding the variable region, it will usually be desirable to include part of the intron in downstream from the J region Where the intron is conserved, it will be necessary for there to be functional splice acceptor and donor sequences in the mtron termini The adjacent non-coding region 5' to the variable region will normally include the sequences involved in the initiation of transcription and . translation, such as the TATA box, cap sequence and CAAT sequence Usually the 5' non-coding sequence does not exceed about 1-2 kilobases (kb)

An enhancer sequence should exist between the J region and the constant region The enhancer used can be the enhancer of either (1) the animal V region or (2) the human constant region

By preserving the 3" region naturally adjacent to the DNA sequence encoding the constant region, the transcription termination signals can be provided for the gene. Where the transcription termination signals are not satisfactorily functional in the expression host cell, a 3' region functional in the host cell can be substituted. Appropriately, the non- the 3' coding region is obtained from a non-coding adjacent 3' region in a constant region from the expression host. fragments This area can then be used as a building block in the production of the gene

Production of expression carriers

Construction of suitable expression vectors containing the desired coding and control sequences can be produced as follows. Termini of the vectors and the DNA fragments can then be religated to form the desired expression vectors. The methods used are not dependent on the DNA source or the intended host.

DNA fragments encoding the light chain and heavy chain can be inserted into separate expression carriers or into the same vector Preferably, the joined genes encoding the light and heavy chimeric chains are assembled into two different expression vectors that can be used to cotransform a receiver cell, either simultaneously or sequentially

The methods for inserting the DNA fragments containing the chimeric genes into expression vectors include the use of restriction endonucleases "Restriction endonucleases" (or "residue enzymes") are hydrolytic enzymes capable of catalyzing site-specific cleavage of DNA molecules The locus of restriction endonuclease action has been determined by the existence of a specific nucleotide sequence Such a sequence is called a recognition site for the restriction endonuclease Many restriction endonucleases from a number of bacterial species have been isolated and characterized using the nucleotide sequence of their recognition sites Some restriction endonucleases hydrolyze the phosphodiester bonds on both strands at the same point, producing blunt ends Others catalyze hydrolysis of bonds separated by a few nucleotides from each other, producing fne single-stranded regions at each end of the cleaved molecule. Such single-stranded ends are self-complementary, thus cohesive and can use ndes to rejoin the hydrolyzed DNA Examples of resection enzymes include Aat II, Barn Hl, Eco RI, Hind III, Nde I, Spe I, Xba I, Sac I, Bgl II, Pst I, Sal I and Pvu II

In addition, the expression vector can have an insert polylinker that has a number of unique restriction sites. Upon digestion of the expression vector with the appropriate restriction enzymes, the polylinker will be cleaved so that at least one DNA fragment containing the gene can be inserted. Where the polylinker allows separable termini, DNA- the fragment is inserted in a single orientation, where the termini are the same, insertion of the DNA fragment will result in plasmids that have two different orientations

Cleavage is carried out by treating the plasmid with restriction enzyme(s) Mainly about 10 //g of plasmid or DN A fragment with about 10 units of enzyme in about 100 µl buffer solution is used Endonuclease digestion will normally be carried out at temperatures varying from 37°C to 65° C, at a pH of from

7 to 9

(Appropriate buffers and substrate amounts for particular retraction enzymes are specified by the manufacturers) The time for the reaction will be from 1 to 18 hours

It may be useful to prevent religation of the cleaved vector by pretreatment with alkaline phosphatase Specific conditions are prescribed by the manufacturer

After the restriction enzyme digestion is complete, the protein is removed by extraction with phenol and chloroform. The nucleic acid is removed from the aqueous fraction (containing about 0.3M sodium acetate) by precipitation with about 2.5 volumes of ethanol.

Descriptions of cleavage enzyme cleavage methods can be found in the following articles Greene et al, Methods in Molecular Biol<q>v. Vol 9, ed Wickner, R B , Marcel Dekker, Inc , New York, "DNA Replication and Biosyn-thesis" Mertz and Davis, Proe Nat Acad Sei, (USA), 69, 3370 (1972)

Size separation of the cleaved fragments by agarose gel electrophoresis is easily performed to follow the course of the reaction Once cleavage has proceeded to the desired extent, the endonuclease can be inactivated by heating above 65°C for approximately 10 minutes or organic extraction

The desired fragment is then purified from the cleavage. Suitable purification techniques include gel electrophoresis or sucrose gradient centrifugation.

The plasmid carrier and foreign DNA fragments are then ligated with DNA ligase to recycle This process is referred to as base pairing and DNA ligation

A suitable buffered medium containing the DNA fragments, DNA hgas and suitable cofactors is used The temperature used will be between 25°C to 4°C When the DNA segments are hydrogen-bonded, the DNA ligase will be able to introduce a covalent bond between the two segments The time used for the base part matching will vary with the temperature used, the nature of the salt solution as well as the nature of the sticky ends or cohesive termini Mainly the time for ligation can be from 5 to 18 hours See Maniatis T , Molecular Cloning, Cold Spring Harbor, supra

Host cells

The expression carrier constructs can then be used to transform a suitable host cell Suitable host cells include cells derived from unicellular as well as multicellular organisms

The chimeric immunoglobulin genes can be expressed in non-lymphoid cells such as bacteria or yeast cells

Various unicellular microorganisms can be transformed, such as bacteria. That is, the unicellular organisms that are capable of being grown in cultures or fermentation. Since bacteria are mainly the most suitable organisms to work with, bacteria will be referred to hereafter as examples of the other unicellular organisms. Bacteria susceptible to transformation include members of the Enterobactenaceae, such as strains of Eschenchia coli, Salmonella, Bacillaceae, such as Bacillus subtilis, Pneumococcus, Streptococcus, and Haemophilus mfluenzae

When expressed in bacteria, the immunoglobulin heavy chains and light chains become part of inclusion bodies. The chains must then be isolated, purified and then assembled into functional immunoglobulin molecules

In addition to prokaryotes, eukaryotic microbes such as yeast cell cultures can also be used Saccharomyces cerevisae or common baker's yeast are the most commonly used among eukaryotic microorganisms, although a number of other strains are commonly available Presence of the trpl lesion as a characteristic of the yeast host cell genome provides an efficient environment for detection of transformation by growth in the absence of tryptophan

In addition to microorganisms, cultures of cells derived from multicellular organisms can also be used as hosts. In principle, any such cell culture is easy to work with, whether from a vertebrate or invertebrate culture, provided the cell line is one that at least originally produced antibodies Propagation of cells from vertebrates in culture has become a routine procedure in recent years [Tissue Culture, Academic Press, Kruse and Patterson, publishers, (1973)] Examples of such useful host cell lines are Sp2/0, VERO and HeLa cells, Chinese hamster ovary (CHO ) cell lines, and W138, BHK, COS-7 and MDCK cell lines

The preferred recipient cell line is a plasmacytoma cell or such as B-lymphocytes or hybndoma cells. Plasmacytoma cells can synthesize, assemble and secrete immunoglobulins coded for by transformed immunoglobulin genes. Furthermore, they have the mechanism for glycosylation of the immunoglobulin Sp2/0 is a preferred recipient cell because it is an immunoglobulin non-producing plasmacytoma cell The cell only produces immunoglobulin coded for by the transformed immunoglobulin genes Plasmacytoma cells can be grown in culture or in mouse pentoneum, where secreted immunoglobulin can be obtained from ascites fluid

Transformation of host cells

Transformation of host cells is performed as follows. The expression carrier is linearized, and DNA is inserted into host cells for production of the antibody. Examples of methods for inserting DNA into host cells include electroporation, protoplast fusion, calcium phosphate precipitation, or other conventional techniques using dextran sulfate and PEG.

If cells without massive cell wall barriers are used as host cells, transformation can be performed by the calcium phosphate precipitation method as described by Graham and Van der Eb, Virology, 52,546 (1978)

If prokaryotic cells or cells containing substantial cell wall structures are used, the preferred method of transformation is calcium treatment using calcium chloride as described by Cohen, F N et al, Proe Nati Acad Sei (USA) 69, 2110 (1972)

The host cells can be transformed either via co-transformation or targeted transformation

For co-transformation, the genes encoding the light chain and heavy chain can be used to transform separate cell cultures, either from the same or from different species, separate light and heavy chain plasmids can be used to co-transform a single cell culture, or finally, a single expression plasmid containing both genes and capable of expressing the genes for both light and heavy chains can be transformed mn in a single cell culture

In the targeted transformation technique, the host cells are transformed with genes encoding the light chain and the cells containing the light chain marker are selected The light chain is found by using cytostaining or possibly by detecting the light chain in the supernatant if it has been secreted Cells selected to have the light chain, is transformed with the heavy chain construct, and resulting cells that additionally contain the heavy chain marker are selected

It is known that some immortalized lymphoid cell lines, such as plasmacytoma cell lines, in their normal state secrete isolated lg light or heavy chains. Accordingly, if such a cell line is transformed with the vector containing the chimeric heavy or light chain according to the present invention, it may not be necessary to transform the cell line or another cell line with the second lg chain, provided that the normally secreted chain is complementary to the variable region of the lg chain encoded by the vector initially used to transform the cell line

Selection and expression of transformed host cells

Essentially, after transformation of the host cells, the cells can be cultured for approximately 48 hours to allow expression of marker genes. The cells are then placed in a selective medium, where untransformed cells are killed, leaving only cells transformed with the DNA constructs

Heavy and light chains or parts thereof can be produced in isolation from each other and antibodies and fragments thereof can be obtained Such preparations require the application of techniques to reassemble isolated chains

The ability of the method according to the invention to produce heavy and light chains or parts thereof, in isolation from each other, gives the opportunity to obtain unique accumulations of immunoglobulins, Fab regions and univalent antibodies It is possible to recombine the heavy and light chains in vitro, interrupted by cleavage of only the interchain disulfides, and regaining antibody activity even without restoration of the interchain disulfides [see Edelman, G M , et al, Proe Nati Acad Sei (USA) 50, 753 (1963)]

The transformed cells are cultured under conditions appropriate for production of the light chains and/or heavy chains, and measured for heavy and/or light chain protein synthesis. Examples of measurement techniques include enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (RIA), or fluorescence-activated cell sorter -analysis (FACS), immunohistochemistry and the like

The binding affinity of monoclonal antibodies for TAG72 is determined by methods well known in the art (see Heyman, B et al, J Immunol Methods 68, 193-204 (1984) and as described in detail in the examples provided hereinafter)

Selected positive cultures are subcloned to isolate pure transformed colonies A suitable technique for obtaining subclones is via the limiting dilution method described by McKeara in Monoclonal Antibodies, Plenum Press, NY (1980)

Hybndomas producing such chimeric antibodies can be cultured using known methods. The transformed cells can secrete large amounts of the light chains and/or heavy chains by culture in vitro, such as by the hollow fiber system, spinner culture, static culture, or in vivo, such as ascites production

The chimeric antibodies can be produced in large quantities by injecting a hybnbom into the pentoneal cavity of pnstanut-exposed mice, and after an appropriate time (about 1-2 weeks) harvesting ascites fluid from the mice, which gives a very high titer of homogeneous monoclonal antibody, and isolating the monoclonal antibodies therefrom by methods well known in the art [see Stramignoni, P et al,

Intl J Cancer 3J[, 543-552 (1983)] Hybndomas are cultured in vivo, as tumors in animals, from which serum or ascites fluid can provide up to about 50mg/ml of monoclonal antibodies. Typically, injection (preferably intrapentoneal) of about 10 <6> to 107 histologically compatible hybridoma cells in mice or rats result in tumor formation after a few weeks The antibodies can then be collected and produced by well-known methods (See mainly Immunoloqical Methods, Vol I & II, ed. Lefkovits, I and Perms, B (1979 & 1981) Academis Press, New York, N Y , and Handbook of Experimental Immunology. ed Weir, D , (1978) Blackwell Scientific Publications, St Louis MO

The antibodies can then be stored in various buffer solutions such as phosphate-buffered saline (PBS), which provides an essentially stable antibody solution for further use

The chimeric antibodies according to the present invention can be fragmented by using known protease enzymes, e.g. papain and pepsin, to obtain highly immunoreactive F(ab')2, F(ab') and Fab fragments. In addition, active fragments can of lg formed by proteolysis (about 50,000 MW) are split into their fully reduced heavy chain and light chain components and sufficiently efficiently reconstituted to yield an active antibody [Haber, E , Proe Nati Acad Sei (USA) 52, 1099 (1064), Whitney.P L , et al, Proe Nati Acad Sei (USA) 52, 524 (1965)] The reactivity of the resulting F(ab') 2 , F(ab') and Fab fragments is determined by methods as described above for the complete monoclonal antibody molecule

Areas of use for the antibodies

The antibodies produced by the method according to the present invention, as well as immunoreactive fragments or recombinants thereof, provide unique advantages for use in a number of cancer treatments. In addition to their ability to bind specifically to malignant cells, and to localize tumors, have the antibodies constantly vanable areas that do not bind demonstrably to normal cells such as fibroblasts, endothelial cells or epithelial cells in the main organs

Specifically, the antibodies, the immunoreactive fragments or recombinants thereof, are applicable to, but not limited to, the following types of cancer treatment (1) in vivo diagnostic measurements conjugated to an imaging marker, for in situ detection of carcinoma lesions as further described below, (2 ) in vivo therapy, by using the antibodies produced according to the method of the present invention, alone or conjugated to a therapeutic agent such as a radionuclide, toxin, effector cells, other antibodies or via a complement mechanism as described below, and (3 ) radioimmunoguided surgery, as described below

Furthermore, a pharmaceutical composition comprising the antibodies produced by the method according to the present invention in a pharmaceutically acceptable, non-toxic, sterile carrier such as physiological saline, non-toxic buffers and the like is also now possible

Injectable mixtures prepared by the method according to the present invention can be either in suspension or solution form. In solution form, the complex (or, when desired, the separate components) is dissolved in a pharmaceutically acceptable carrier. Such carriers include a suitable solvent, preservatives such as benzyl alcohol, if it is needed, and buffers Applicable solvents include, for example, water, aqueous alcohols, glycols and phosphonate or carbonate esters Such aqueous solutions do not contain more than 50% of the organic solvent by volume

Injectable suspensions such as mixtures prepared by the method according to the present invention require a liquid suspension medium, with or without adjuvants that carry The suspension medium can be, for example, aqueous polyvinylpyrrolidone, inert oils such as vegetable oils or highly refined mineral oils or aqueous carboxymethylcellulose Suitable physiologically acceptable adjuvants, if necessary to preserve the complex in suspension, can be chosen from thickeners such as carboxymethyl cellulose, polyvinylpyrrolidone, gelatin and alginates. Many surfactants are also usable as suspending agents, e.g. lecithin, alkylphenol, polyethylene oxide, additive products, naphthalene sulphonates, alkylbenzene sulphonates and the polyoxy-ethylene sorbitan esters Many substances that provide hydrophobicity, density and surface tension in the liquid suspension medium can help in the production of the injectable suspensions in individual cases F e.g. are silicone antifoam agents, sorb itol and sugar all applicable suspending agents

Because cancer cells are heterogeneous, and consequently a single monospecific chimeric antibody may not be able to recognize all cells expressing different epitopes in a tumor

Thus, it may be desirable to administer several different chimeric antibodies produced by the method according to the present invention. The sequential use of these different antibodies should substantially reduce the anti-ideotypic reactions in human patients when compared to the simultaneous use of a single antibody F eg CH92, CH88 and CH44 can be administered sequentially to a patient Since these antibodies have different light chains and indeed different CDR3 regions, anti-idiotypic reactions should be minimized

In vivo diagnostic measurements

In vivo diagnostic measurements of human tumors or metastases thereof using the antibodies, immunoreactive fragments or recombinants thereof, conjugated to a marker, administered to a patient and then detecting the presence of the imaging marker in the patient by exposing the patient to a suitable detection agent

Administration and detection of the antibody-imaging marker conjugate, as well as methods of conjugating the antibody to the imaging marker, are carried out by methods that can be easily recognized or determined, as described, for example, in Goldenber, DM et al, New England J Med, 298, 1384-1388 ( 1978), Goldenberg, D M et al, J Amer Med Assoc 280, 630-635 (1983), Goldenberg, DM et al. Gastroenterol 84,524-532 (1983), Siccardi. G G et al, Cancer Res 46, 4817^822 ( 1986), Epeneetos, AA et al, Cancer 55, 984-987 (1985), Philben,

V J et al, Cancer 57, 571-576 (1986), Chiou, R et al, Cancer Inst 76, 849-855

(1986), Colcher, E et al, Cancer Res , 43, 736-742 (1983), Colcher, E et al, Laboratory Research Methods in Biology and Medicine Immunodiaqnostics, New York, Alan R Liss, pp 215-258 (1983) , Keenan, AM et al, J Nucl Med 25, 1197-1203 (1984), Colcher, D et al, Cancer Res 47, 1185-1189(1987), Estaban, J M et al, Intl J Cancer 39, 50-59 (1987), Martin.D T et al, Curr Surg 41., 193-194

(1984), Martin, E W Jr et al, Hybndoma 5, S97-S108 (1986), Martin. D T et al, Am J Surg 150, 672-675 (1985), Meares et al , Anal Biochem 142, 68-78 (1984), and Krejcarek et al, Biochem and Biophys Res Comm 77. 581-585 (1977)

The dosage will vary depending on the age and weight of the patient. Mainly, the dosage should be performed to visualize or detect tumor sites, separate from normal tissues. Preferably, a single dosage will be between 0.1-200 mg of an antibody-marker conjugate per patient.

Examples of imaging markers that can be conjugated to the antibody are well known to those skilled in the art, and include substances that can be detected by diagnostic imaging using a gamma scanner or handheld gamma probe or Positron Emission Tomography or the like as described in the references mentioned above, and substances that can be detected by nuclear magnetic resonance imaging using a nuclear magnetic resonance spectrometer or the like, as described in the references mentioned above

Suitable examples of substances that can be detected by using a gamma scanner or the like include, for example, radioisotopes such as 125|, 1311, 123|, 111m, 105Rh, 153Sm, 67Cu, 67Ga, 166Ho, 177Lu, 186Re, 188Re , and 99mTc 125,, 123,, 153sm and 99nriTc are preferred because of their low energy and suitability for detection over a long range

An example of a substance that can be detected by using a nuclear magnetic resonance spectrometer or similar is gadolinium (Gd)

In vivo cancer treatment

In this method, the antibody-therapeutic agent conjugate can be delivered to the carcinoma site to thereby directly expose the carcinoma tissue to the therapeutic agent

The antibodies produced by the method according to the present invention, immunoreactive fragments or recombinants thereof, can be administered in a pharmaceutically effective amount for the in vivo treatment of human carcinomas or metastases thereof. A "pharmaceutically effective amount of the antibody" an immunoreactive fragment or a recombinant of this, conjugated or unconjugated to a therapeutic agent, means that the amount of said antibodies in the pharmaceutical mixture should be sufficient to achieve effective binding with the antigens, against which said antibodies have specific affinity The pharmaceutical mixture can be administered in a single or multiple dosage

Methods for the preparation and administration of conjugates of the antibody, immunoreactive fragments or recombinants thereof and a therapeutic agent are well known or can be easily determined by those skilled in the art. Furthermore, suitable dosages will depend on the age and weight of the patient and the therapeutic agent used and are well known or readily determined by those skilled in the art. Representative methods are described in the references cited below

Examples of the antibody-therapeutic agent conjugates that can be used in therapy include the following (1) antibodies bound to radionuclides, such as 1311, 90y, 105Rh, 47sc. 67cu, 212B,,211At, 67Ga, 125i. 186Re, 188Re, 177Lu, 99mTc, 153Sm, 123, and 111ln as described for example in Goldenberg, DM et al, Cancer Res 4J., 4354-4360 (1981), Carrasquillo.J A et al, Cancer Treat Rep 68, 317-328 ( 1984), Zalcberg. J R et al, J Nat Cancer Inst 72,697-704 (1984), Jones, DH et al, Int J Cancer 35, 715-720 (1985), Lange. P H et al, Surgery 98, 143-150 (1985), Kaltovich.F A et al,

J Nucl Med 27, 897 (1986), Order, S E , et al, Int J Radiother Oncol Biol Phys 8, 259-261 (1982), Courtenay-Luck, N et al, Lancet 1, 1441-1443 (1984) and Erninger, Ds et al, Cancer Treat Rep 66, 289-297 (1982),

(2) antibodies bound to drugs or biological response modifiers such as methotrexate, adrimyacin, and lymphokines such as interferon, as described in, for example, Chabner.B et al, Cancer, Principles and Practice of Oncoloqv. Philadelphia, PA, J B Lippincott Co Vol 1, p 290 328 (1985), Oldham. R K et al, Cancer, Principles and Practice of Oncoloqv. Philadelphia, PA, J B Lippincott Co , Vol 2, pp 2223-2245 (1985), Deguchi.T et al, Cancer Res 46, 3751-3755 (1986), Deguchi.T et al ,Fed Proc_44,1684 (1985), Embleton. M J et al, Br J Cancer 49, 559-565 (1984 and Pimm. M V et al, Cancer Immunol Immunother 12, 125-135

(1982), (3) antibodies bound to toxins as described in eg Uhr, J W et al, Monoclonal Antibodies and Cancer. Academic Press, Inc, pp 85-98 (1983), Vitetta, ES et al, Biotechnology and Bio Frontiers. ed P H Abelson, p 644-650 (1983) and Vitetta.E S et al, Sei, 219, 644-650 (1983), (4) heterofunctional antibodies, e.g. antibodies bound to or combined with another antibody so that the complex binds to both the carcinoma and the effector cells, eg killer cells such as T cells, as described eg in Perez, P et al, J Exper Med 163,166-178 (1986) and Lau. MA et al

Proe Nati Acad Sei (USA) 82, 8648-8652 (1985), and

(5) natural, i.e. non-conjugated or non-complexed antibodies, as described in e.g. Herlyn.D et al, Proe Nati Acad Sei (USA) 79, 4761-4765

(1982), Schulz. G et al, Proc Nati Acad Sci (USA) 80, 5407-5411 (1983), Capone. P M et al, Proc Nati Acad Sci (USA) 80, 7328-7332 (1983), Sears. H F , et al, Cancer Res 45, 5910-5913 (1985), Nepom GT et al, Proe Nati Acad Sei (USA) 81, 2864-2867 (1984), Koprowski, H et al, Proe , Natl Acad Sei (USA ) 81., 216-219 (1984), and Houghton, A N et al, Proe Nati Acad Sei (USA) 82, 1242-1246 (1985)

The methods for combining the antibody or antibody fragment into a desired therapeutic agent as described above are conventional and well known in the field, e.g. the methods given in the references above

Radioimmunosuppressed surgery

Antibodies, immunoreactive fragments or recombinants thereof are important for radioimmunoguided surgery (RIGS) In RIGS, an intraoperative therapy, tumors are located and excised An antibody labeled with an imaging marker is injected into the patient, and bound antibody located by a handheld gamma detection probe (GDP) and excised An example of GDP is Neoprobe™, commercially available from Neoprobe™ Corporation, Tampa, Fl See Martin et al, "Radioimmunoguided surgery a new approach to the intraoperative detection of tumor usmg antibody B72.3" Amer J Surg 156, 386-392 (1988), Martin et la "Radioimmunoguided surgery intraoperative use of antibody 17-1A in colorectal cancer", Hybridoma 5, 597-S108 (1986)

Administration and detection of the antibody-imaging marker conjugate, as well as procedures for conjugating the antibody to the imaging marker, are carried out by methods that can be easily known or easily determined by a person skilled in the art, as described, for example, above

The dosage will vary depending on the age and weight of the patient, but mainly a time dosage of 0.1 to 200 mg of antibody-marker conjugate per patient is sufficient

The following examples are for illustration only of the construction and expression of chimeric DNA sequences encoding the antibodies produced by the method of this invention All temperatures not otherwise stated are in °C All percentages not otherwise stated are by weight

EXAMPLES

Exploitation of mouse constant regions

CC antibodies were generated from mice, and are significantly less able to perform the effector functions possessed by the human constant regions

Accordingly, in the following examples, selected antibodies are "humanized" by genetically removing the constant regions of the heavy and light chains and replacing them with their human equivalents

The mouse light chain constant region genes were replaced with the human kappa (k) gene, and the mouse heavy chain genes were replaced with each of the four

human gamma isotypes (y1, y2, y3 and y4) Each of these four gamma isotypes has unique biological properties For a general review see "The Human IgG subclasses", Hamilton, R G (1989) Docnr CB0051-289 Calbiochem Corporation

Production of heavy and light keed variable range

Isolation of CC49 easy kiede

CC49 thymoma cells secrete an antibody that has an IgGi isotype heavy chain and a kappa light chain

Total DNA from CC49 hybndoma cells, Galb/C mouse kidney cells and NSI plasmacytoma cells was isolated according to the methods set forth in Cell, 24, 353-356(1981)

Essentially, approximately 10-20 µg of the extracted DNA from each cell line was digested completely with 80 units of Barn Hl, Eco RI, Hind III, Spe I, Xba I, Sac I, Bgl II and Pst 11 50-100 microliters of a reaction mix containing the appropriate reaction buffer at 37°C overnight

Then, the total extracted DNA from each cell line was subjected to the Southern Hybridization technique, developed by E M Southern, [J Mol Biol 98, 503-517

(1975)] The DNA fragments were fractionated on the basis of their size by electrophoresis on a 0.8% agarose gel. The double-stranded DNA fragments were modified into single-stranded DNA fragments in an alkaline solution, and then a nitrocellulose filter was placed in near contact with the gel to transfer the modified DNA segments onto filters in the presence of a solution with a high salt concentration

Hybridization was performed by using as probe a randomly released < 32 P >-labelled L chain

More specifically, the probe was a 1.71 kilo base pair (kbp) Hind Ill-Pst I fragment containing the coding exons for the mouse Jl regions (J1-J5) and was isolated from plasmid pGD1 A nucleotide sequence of the probe fragment is given in Figure 7 This plasmid is described in "Site Directed Cleavage of Immunoglobulm Gene Segments by Lymphoid Cell Extracts", Agostaro et al, Can J Biochem Cell Biol 63, 969-976

(1985) The plasmid was provided by Nobumichi Hozumi and John Roder, Mt Sinai Research Institute, Toronto, Ontario, Canada

To radiolabel the probe, alpha < 32P > dCTP was obtained from Amersham, Arlmgton Heights, IL, USA, and the random release test kit was obtained from Pharmacia, Piscataway, NJ, USA

The signals in Southern transfers were visualized by autoradiography using Kodak X-OMAT™ AR film No obviously rearranged bands were observed Thus, relative to the standards, no unique band was detected on the autoradiogram of CC49 DNA digested with Hind III However, did not reveal from the Southern data that the rearranged band for the L chain was masked by a band that migrated in CC49 Hind III cleaved DNA parallel to the band resulting from a Hind III cleavage of mouse kidney cell DNA (representing germline DNA) This was indeed shown to be the case

Preparation of plasmid containing Mouse V\ genes

LSAMBDA-ZAP™, a lambda-based insert cloning vector capable of self-cutting, was purchased from Stratagene Co , La Jolla, CA, USA LAMBDA-ZAP™ is described on pages 20-21 of the 11987 Stratagene Catalog The cohesive (cos) ends of LAMBDA-ZAP™ was ligated overnight following the manufacturer's procedure

Twenty micrograms of the ligated LAMBDA-ZAP™ was digested with 5 microliters (15 units) of Spe I, purchased from New England Biolabs, Inc. The total volume of digestion was 100 microliters. After 55 minutes of digestion, 6 more units of Spe I were added After 70 minutes, the reaction was stopped by phenol extraction and ethanol precipitation carried out as via Stratagene's method

Cleavage with Spe I residue removal enzyme results in the production of "sticky ends" at both termini. These sticky ends were modified with T4 DNA polymerase to form half-filled Spe I sticky ends, e.g. 5'ACT/3TCATG. To perform the half-filling reaction, The DNA pellet obtained in the above ethanol precipitation dissolved in 8 microliters of water To this was added 2 microliters of 10 millimolar dTTP, 2 microliters of 10 millimolar dCTP, 2 microliters of Stratagene's 10X hgas buffer, 4 microliters of reionized, distilled water, and 2 microliters of a Klenow fragment from Bethesda Research Laboratories (BRL) The reaction was carried out at room temperature for 30 minutes The reaction was stopped by inactivating DNA polymerase at 65°C 110 minutes

160 micrograms of total CC49 hybndoma DNA (containing the mouse lett chain promoter and Log VJ exons) was digested completely with Hind III Fragments between approximately 1 kb and two approximately 20 kb were excised from 0.8% agarose gels DNA was purified using GENECLEAN™ , which is commercially available from BIO 101 (La Jolla, CA, USA)

The total CC49 hybndom DNA Hind III cleaved fragments were half filled, hk LAMBDA-ZAP™'s Spel fragments except that dATP and dGTP were used The half filled Hind III cleaved fragments produced STXGCTRøGAA sticky ends, which are compatible with the Spe I half filled LAMBDA -ZAP™ fragment above

After phenol extraction and ethanol precipitation according to the descriptions of Maniatis, the total CC49 hybndom Hind III modified and LAMBDA-ZAP™ Spe I modified DNA fragments were ligated using T4 DNA hgas The ligation reaction was set up using a 6.1 microliter ligation mix which contained the following approximately 0.2 microgram of the total CC49 hybndom Hind III modified DNA in a 3 microliter solution, approximately 1 microgram of LAMBDA-ZAP™ Spe I modified DNA in a 1 microliter solution, 0.6 microliter of Stratagene's 10X hgase buffer, 0, 5 microliters of 10 millimolar ATP and 1 microliter of Stratagenes ligase. This was incubated overnight in a water bath and the temperature lowered in small steps from about 18°C to about 4°C. This ligase eliminated both the Hind III and Spe I sites

A genomic pool of ligated mixture was made according to Stratgene's procedure Briefly, 2 microliters of the ligated mixture produced above were used in Stratagene's Gigapck Gold packaging system, following the manufacturer's guidelines, 15,150 mm plates having a density of 50,000 plaques per plate, was screened, as per the manufacturer's guidelines, for positive clones by hybridization to the nitrocellulose filter, obtained from Schleicher-Schuell, Keene, NH, USA The < 32P > arbitrarily labeled probe generated from pGD1, described above, was used for hybridization Two positive clones were obtained

Each clone was plaque-purified, and recombinant plasmids (phagmids) of LAMBDA-ZAP™ containing the CC49 L chain-vanable region were obtained using Stratagene's automated dilution procedure. The vector portion of the resulting recombinant plasmid is named pBLUESCRIPT SK(-) and consists of 2964 bp as described in Stratagene's catalog 1987 A plasmid map of pBLUESCRIPT SK(-) is shown in Figure 8

DNA from the two positive clones was partially sequenced, and both were identical. One of the clones, called pRL 101, was used for further studies

Restriction mapping of CC49 light k| oath

pRL101 was 7.61 kb, and the size of the DNA insert was determined by restriction enzyme mapping to be 4.65 kb A plasmid map of pRL101 is shown in Figure 9 A residual cleavage enzyme map of the CC49 L chain genome DNA insert in pRL101 is shown in Figure 10

Isolation of CC83 light-keyed variable range

The procedures used to isolate the CC83 light chain were essentially those used to isolate the CC49 light chain, with the following exceptions

Genomic context containing 7 x 10<5> plaque was mapped by using as probe < 32P > arbitrarily labeled 1.71 Hind Ill-Pst I fragment generated from pGD1 as described above A positive clone was obtained The positive clone was named pRL200

Restriction map leqqinq of CC83 easy kiede

pRL200 was 7.44 kb, and the size of the DNA insert was determined by restriction enzyme mapping to be 4.48 kb A plasmid map of pRL200 is shown in Figure 11 A restriction enzyme map of the CC83 L chain genome The DNA insert in pRL200 is shown in Figure 12

Isolation of CC49 tunqkeyed variable range

The procedures used to isolate the CC49 heavy chain were essentially those used to isolate the CC49 light chain, including mapping the same CC49 Hind III modified DNA

The hybridization probe used to map the assembly was generated from pNP9, which contains a 1.98 kbp Eco R1-Bam H1 fragment containing the coding exons for JH3 and Jh4 in the CC49 immunoglobulin heavy chain. The nucleotide sequence of the probe fragment is given in Figure 13 The plasmid was provided by Dr Nobumichi Hozume and Dr John Roder, Mt Sinai Research Institute, Toronto, Ontario, Canada

A genomic sequence containing 9.5 x 10<5> plaques was mapped, from which a positive clone was obtained, the positive clone was named pHH49

Restriction map analysis of CC49 heavy kiede

pHH49 was approximately 7.0 kg, and the size of the DNA insert was determined by restriction enzyme mapping to be approximately 4.0 kb A plasmid map of pHH49 is shown in Figure 14

Insulation of CC83 heavy kied eva ria belt area

The procedures used to isolate the CC83 heavy chain were essentially those used to isolate the CC49 heavy chain, with the following exceptions

Approximately 13 micrograms of ligated LAMBDA-ZAP™ vector DNA was digested with 12 units of Spe I, purchased from New England Biolabs, Inc, in a total volume of 100 microliters of an appropriate buffer LAMBDA-ZAP™ was digested at 37°C for one hour The reaction mixture was phenol extracted and ethanol precipitated as per Stratagene's method The Spe I cleaved LAMBDA-ZAP™ was dephosphorylated according to the method set forth in Maniatis with the exception that a 40-fold excess of calf intestinal alkaline phosphatase (Boehnnger Mannheim, Indianapolis, IN, USA) was used

DNA from CC83 was completely digested with Spe I Fragments between approximately 3 kb and 40 kb were isolated from a 0.8% agarose gel slice electroeluent as described by Maniatis, and ligated with the dephosphorylated Spe-I cut LAMBDA-ZAP™ vector

A genome sequence containing 5 x 10<5> plaques was mapped using the probe generated from nNP9, the sequence of which is given in Figure 13. A positive clone was obtained. The positive clone was named pHS83

Restriction mapping of CC83 heavy chain

pHS83 was 7.95 kb, and the size of the DNA insert was determined by restriction enzyme mapping to be approximately 5 kb. A plasmid map of pHS83 is shown in Figure 15

Sequencing of CC46. CC49. CC83 and CC92 mRNA

Total RNA from approximately 1 x 10<7> CC49 cells frozen at -70°C was extracted essentially as reported by Maniatis, with the following exceptions

Four molar guanidium isocyanate and 2.5 molar sodium citrate, pH 7.0 and a SW40Ti rotor centrifuged at 31,000 rpm were used

A total amount of 2.7 mg of CC49 was isolated After centrifugation, poly A+ mRNA was purified from approximately 1.68 mg of RNA by ogo(dT) cellulose chromatography using Type 3 ogo(dT) cellulose obtained from Collaborative Research, Inc , Bedford, MA, USA The procedure was as described by AViv and Leder, Proe Nat'1 Acad Sei (USA) 69, 1408 (1972) A total amount of 50.24 micrograms of poly A+ mRNA was obtained from 1.68 milligrams of mRNA

A total amount of 3.82 mg of CC83 was isolated from approximately 1 x 10<7> cells A total amount of 54.6 //g of poly A+ mRNA was isolated from 1.91 milligrams of total RNA

A total amount of 0.814 mg of CC92 RNA was isolated from approximately 2.6 x 10<8 >cells A total amount of 41.88 micrograms of poly A+ RNA was isolated from 0.814 mg of total RNA

A total amount of 1.7 mg of CC46 RNA was isolated from approximately 2.89 x 10<8 >cells A total amount of 68.88 micrograms of poly A+ RNA was isolated from 1.7 mg of total RNA

Synthetic oligonucleotide primers were synthesized using an Applied Biosystems' (Applied Biosystems (ABI), Foster City, CA, USA) Model 380A DNA synthesizer, by phosphoramadite-based chemistry as described by ABI The oligonucleotides were purified, as described by the manufacturer, after electrophoresis on a 20% polyacrylamide gel containing 7M urea

Oligonucleotide concentrations were determined electrophotometrically at an optical density of 260 nm, where 1 OD 260 nm unit is hk 33 fjglml single-stranded DNA

The following oligonucleotide primers were made for mRNA sequencing (1) For CC49, CC83 and CC92 light chains, KL(-), a 22-mer

5-GGAAGATGGATACAGTTGGTGC-3'

complementary to the coding sequence in the 5' of the constant region of mouse immunoglobulin kappa chains, is used to determine the mRNA sequence in the 3' of the light chain variable region

In addition, for CC49 light chain, 49FR1(-), a 17-mer

5-GGAAGATGGATACAGTTGGTGC-3'

used to determine the remaining sequence

In addition, for CC83 light chain, J4(-), a 24-mer

5'-CCAACTTTGTCCCCGAGCCGAACG-3'

and also 83L CDR2(-), a 17-mer

5-CAGGACTCCAGTGTGC-3'

used to determine the remaining sequence

In addition, for the CC92 light chain, J5(-) was

5'-CGTTTCAGCTCCAGCTTGGTCCC-3'

used to determine the remaining sequence

For the CC46, CC49, CC83 and CC92 y*\ heavy chains, CH1(-), a 24-mer

5'-ATGGAGTTTAGTTTGGGCAGCAGAT-3'

complementary to the coding sequence at the 5" end of the mouse

y1 heavy chain constant region CH1(-) 24-mer is used to determine the mMRA sequence at the 3'end in the heavy chain vanable regions

In addition, for the CC49 heavy chain, the JH4(-)-20mer became

5-GGTGACTGAGGTTCCTTGAC-3'

used to determine the remaining sequence

In addition, for CC83 heavy chain, JH2(-)-16-mer

5'-CTGAGGAGACTGTGAG-3'

used to determine the remaining sequence

In addition, for the CC92 heavy chain and the B72.3 heavy chain, the B72.3/CC92 HC-20 became more

5'-CCTTGAACTTCTCATTGTAC-3'

used to determine the remaining sequence

The following procedures were performed as described by Jan Gelliebter in BRL FOCUS9.1 (1987)

The oligonucleotide primers were end-labeled as follows 100 ng of oligonucleotide was combined in 50 mM Tris HCL (pH 8), 10 mM MgCl 2 , 5 mM dithiothreitol and 1 mM spermidine, 100 µCi (γ-<32>P) ATP (Amersham, 5000 Ci/mMol) and 7 units of T4 polynucleotide kinase in a volume of 13 µL The reaction was allowed to proceed at 37°C for 30 min, then heated for 5 min at 65°C to inactivate the kinase, and then 7 µL of water was added to to make the concentration 5 ng///l The labeled monomers were stored at -20°C until needed

Separate samples, each containing approximately 13 micrograms of poly(A)+ mRNA of CC49, CC83, CC92, or CC46, respectively, were resuspended in 110 µl base pair matching buffer [10 mM Tris HCl (pH 8.3), and 250 mM CKI]

A 5 ng sample of end-labeled oligonucleotide primers was added to each mRNA sample, heated to 80°C for 3 minutes and base-paired for 145 minutes at 61°C, for KL(-) and 65°C for the CH1 (-) oligonucleotides AMV reverse transcriptase ( Boehnnger Mannheim) was used at a level of 6 units for this mRNA sequencing reaction The remainder of the sequencing was performed as set forth in BRL FOCUS9, 1 (1987)

Initial sequence data showed that the heavy and light chains were rearranged as follows CC49 kappa light chain used a J5, CC49 y1 heavy chain used a JH4 The CC83 light chain used a J4, the CC83 gamma 1 used a JH2 The CC46 kappa light chain used a J2, the CC46 heavy chain used a JH3 The CC92 light chain used a J5, the CC92 gamma 1 used a JH2 Figure 16 shows the nucleotide sequence of CC49 VH, and the underlined segments show the sequences generated by applying oligonucleotide primers to mRNA Figure 17 shows the nucleotide sequence of CC83 VH, and the underlined segments show the sequences formed by using oligonucleotide primers on mRNA The entire nucleotide sequences of CC46 VH and CC92 VH, shown in Figure 2, were formed by using oligonucleotide primers on mRNA Figure 4a shows the nucleotide sequence of CC49 VL, and the underlined segments shows the sequences formed by applying oligonucleotide primers to mRNA Figure 5a shows the nucleotide sequence of CC83 VL, and the underlined se gments show the sequences generated by applying oligonucleotide primers to mRNA

The entire nucleotide sequence of CC92 VL, shown in Figure 6, was generated by applying oligonucleotide primers to mRNA

Protein sequence

Purified mouse CC49 and CC83 immunoglobulin molecules were sent to Dr George Tarr at the University of Michigan Protein Sequencing equipment for NH2-terminal amino acid sequence analysis Dr Tarr used the Edman degradation method, as modified by Tarr, G E , in "Manual Edman Sequencing System", Microcharacterization of Polypeptides A Practical Manual 91 John E Shively, ed , Humana Press, Inc , Clifton, N J , pp 155-194 (1986)] l briefly Dr Tarr reduced and alkylated the immunoglobulin molecules The light and heavy chains of the immunoglobulin molecules were separated by reverse phase HPLC

Figure 4b shows the amino acid sequence of CC49 VL, and the results of the amino acid sequence determination for the first 24 amino acids in the mature CC49

VL is underlined Figure 5b shows the amino acid sequence of CC83 VL, and the results of the amino acid sequence determination of the first 51 amino acids in the mature CC83 VL is underlined ASN-20 could not be determined in the CC83 light chain, due to the presence of N-linked carbohydrate residues in this position, which is shown in the PNGase F experiment below The sequence Asn-lle-Thr corresponds to the corresponding sequence Asn-X-The/Ser for carbohydrate attachment to Asn

Since the heavy chains of immunoglobulins CC49 and CC83 are blocked at the N-terminus and unavailable for amino acid sequence determination, the natural glycopeptide was treated with cyanogen bromide (CNBr) to cleave at the methionine residues. The cleavage resulted in fragments that were purified by reverse phase HPLC N- terminal amino acid sequencing was performed on the CNBr fragments

The results of the amino acid determination of one of the CC49 Vh CNBr peptide fragments are indicated as underlined residues in Figure 18 The results of the amino acid determination of one of the CC83 VH CNBr peptide fragments are indicated as underlined residues in Figure 19 As with CC49, all other peptide sequences correspond to CNBr fragments formed from the constant region in museum 1

Determination of N-bonded carbohydrate on CC83 L chain

This experiment was done to verify that there is an N-linked carbohydrate attached to the CC83 light chain, presumably at Asn-20 (see Figure 5b) The enzyme glycopeptidase F (PNGase F), which is isolated from the culture filtrate of Flavobactenum meningosepticum [Tarentino, AL et al, Biochemistry 24, 4665-4671 (1985)], will cleave high mannose and/or biantennary complex sugars N-linked to Asn to form a free carbohydrate structure and an ASP residue from the ASN to which it was attached The difference in molecular weight between the glycosylated and non-glycosylated form of the same peptide can be determined by SDS-PAGE

Twelve microgram reactions with and without PNGase F (Boehringer Mannheim, Indianapolis, IN, USA) for the purified mouse antibodies CC49, CC83 and CC11 F(ab')2 (a positive control) were performed in a final aqueous reaction volume of 40 microliters Four microliters 10x buffer (1M potassium phosphate, 0.1M dinatnum EDTA pH 7.4) was added to each reaction mixture. To the tubes labeled "with PNGase F", 7.5 µL of PNGase F was also added and all tubes were incubated at 37°C 11 hours 40 microliters of Laemmli 2X sample dilution buffer containing fimercaptoethanol was added to the reaction tubes. A 10% SDS polyacrylamide gel was electrophoresed, the gel was stained with Coomassie blue R-250 and destained. Figure 20 shows the results As shown in lane 2, a new band (<*>) appears in the PNGase F treated CC83 sample but not in the untreated CC83 sample (lane 3) The new band is about 2000-3000 molecular weight less than the natural light chain band, which represents the removal of an N-bond carbohydrate part The only matching glycosylation site for the CC83 light chain is at ASN 20, so by interference it is assumed that this is the actual site of glycosylation, and why it did not show up in the N-terminal sequence analysis of the CC83 light chains as ASN Den CC49 light chain does not change mobility when treated with PNGase F (lane 6), but a new band is observed for the heavy chain fragment from CC11 F(ab')2 (lane 4<*>) which serves as positive control mRNA sequence data for CC11 heavy chain indicate a corresponding glycosylation site in the V region (data not shown) The standards (lane 1) are bovine serum albumin (BSA), MW 68,000 and soybean trypsin inhibitor (STI), MW 21,500

DNA sequence

Plasmid DNA was sequenced directly using the Sequenase DNA sequencing test kit, obtained from United States Biochemical (USB), Cleveland, OH, USA USB's protocol was followed to sequence double-stranded DNA DNA from each variable region was sequenced using JH or JL oligo determined from the mRNA sequence information to be specific for each productively rearranged heavy chain or light chain gene

After the initial sequences were determined, the sequence was further extended using additional primers. The additional primers were synthesized using information gathered from the sequences formed earlier

Using the above technique, the DNA sequences for the entire heavy chain variable region exons and light chain vanable region exons for CC49 and CC83 were obtained. The DNA sequence was collected and analyzed using Hitachf's DNA sequence analysis computer program DNASIS™

The following oligonucleotide primers were designed for DNA sequencing

(1) for both light chains, C« intron (-)

5'-GAAAACCTGTGTCTTACAC 3'

(2) For the CC49 light chain, CC49 FREE(+)

The complete nucleotide sequences for CC49 VL and CC83 VL are shown in Figures 4a and 5a, respectively. Subsequently, the sequencing of each heavy chain was expanded with the following sequences CC49/83 HC/5'(+)

The complete nucleotide sequences for CC49 VH and CC83 VH are shown in Figure 2

Comparisons were made between the characterized mRNA sequence and the characterized DNA sequence, and between the characterized amino acid sequence and the amino acid sequence predicted from the DNA sequence. Based on these comparisons, the plasmid clones were identified as containing the correct DNA sequence to encode the CC49 and CC83 heavy and light chain variable ranges

The predicted amino acid sequences from the nucleotide sequences in the heavy chain variable regions in CC49 and CC83, as shown in Figure 2, show extensive sequence similarity throughout the framework regions and the hypervariable regions 1 and 2 Hyper-variable region 3 is quite different between the two due to the recombination of VH region with different D and JH sequences, namely that the CC49 y1 heavy chain used a Jh4, and the CC83 y1 used a JH2

The extensive DNA sequence homology 5' to the coding regions of the CC49 and CC83 heavy chain variable region genes shows that the two heavy chain variable region genes were formed from the same germline exons

Isolation of VHaTAG, the germline precursor gene of the heavy gene of CC46. CC49. CC83 and CC92

The procedure used to isolate the germline precursor gene of the heavy chain variable regions in CC 46, CC49, CC83 and CC92 was essentially that used to isolate the CC49 heavy chain variable region with the exception that DNA was used to form the LMABDA-ZAP™ assembly came from an irrelevant hydndom cell line, (ie a cell line that produces antibodies that do not bind satisfactorily to TAG72) A genome sequence containing approximately 900,000 plaques was mapped, from which a positive clone was isolated The positive clone was named pVHaTAG pVHoTAG was approximately 5 .2 kb, and the size of the DNA insert was determined by restriction enzyme mapping to be approximately 2.2 kb

DNA sequence of VhøTAG

The following oligonucleotide primers were used to determine the DNA sequence of VhgtTAG

The complete nucleotide sequence of VoTAG is shown in Figure 2

Isolation of human heavy constant genes

Plasmid constructs containing the various heavy-chain human constant regions (py1, py2, py3 and py4) were provided by Dr lian R Kirsch of the National Cancer Institute, Betseda, Maryland

Restriction enzyme mapping was performed on these genes to confirm their identity Restriction maps for the human constant regions are described in Figure 21

Kimensk easy kiede

Mouse CC49 V area

The Hind III site in the CC49 light chain genomic DNA located in the mouse intron region between J5 and Ck (see Max, Edward E et al, J Biol Chem 256, 5116

(1981), was lost in the cloning procedure where the half-filled in Hind III site was ligated to half-filled in Spe I sites in the LamBDA-ZAP vector. Plasmid pRL101 (Figure 9) carried this modification. The intron Hind III site was regenerated as described in the steps below to enable a Hind III-Bam H1 human germline kappa light chain DNA fragment (see Hieter, P et al, J Bio Chem 257, 1516 (1982) to be ligated directly to the mouse vaneable region. All steps were performed using standard molecular biology techniques known to professionals and that can be found a handbook such as Manatis

A 1.69 kb Barn Hl-Pst I fragment is isolated from pRL101, described supra A 2.96 kb Barn Hl-Pst I fragment is isolated from pBluescnpt SK(-) (purchased from Stratagene), described supra The two fragments are then ligated and pRL103 below is isolated

Plasmid pGB1 (described supra), was digested with Pst I and Hind III cleavage enzymes to yield the required 1.03 kb intron-containing fragment, and pRL103 was also digested with Pst I and Hind III cleavage enzymes to remove the small fragment of DNA. in the polylinker

The resulting fragments were ligated with T4 DNA ligase to produce a 5.68 kg plasmid, named pRL 104 A partial residue map of pGD1 and pRL 104 is shown below

Human Ck area

Plasmid phum Ck was obtained from Dr John Roder, Mt Sinai Research Institute, Toronto, Ontario, Canada The plasmid is generated from pBR322, with a 12 kb Barn H1 fragment containing the human Ck exon inserted into this pBR322 is described on page 171 of Stratagenes The 1987 catalog 12 kb Barn Hl fragment restriction map is shown below [from Heiter, P et al, J Viol Chem 257, 1516(1982)]

Plasmid phum Ck was cleaved with Hind III and Barn HI cleavage enzymes to give a 5.0 kb fragment containing the human Ck exon pRL104 was cleaved with Fsp I and Hind III cleavage enzymes to give a 4.2 kb fragment containing the mouse light-chain variable exons from CC49

The two resulting fragments were joined with T4 DNA ligase to produce a 9.2 kb fragment from the mixture of resulting fragments. This mixture was digested with Barn I to give a 7.7 kb Barn Hl CC49 L chain chimeric construct with Barn I sticky ends, containing both the mouse vanavable region exons and the human constant region (k) exon These constructs use the human enhancer sequences and the mouse promoter sequences

The chimeric Barn Hl fragment containing both the mouse light chain variable region exons (L and VJ) and the human constant region kappa (k) exon was ligated into the Barn I site with plasmid pSV2neo (5.6 kb), a pBR322 denvert plasmid containing the selectable marker gene neo (obtained from ATCC) The presence of the active neogene makes a cell resistant to growth inhibition by Geneticin, a neimycin-like drug also called G418

The chimeric Gam H1 fragment was inserted into pSV2neo in both orientations as shown below. Both transcriptional entries in the chimeric light chain gene, relative to the neo gene, were constructed. Plasmid pSV2neo was cloned into the Barn H1 site, dephosphorylated (according to the procedure presented in Aniatis) using calf intestinal alkaline phosphatase (to prevent self-ligation) and ligated with chimeric CC49 L chain Barn Hl fragments from above

The transcription orientations of the neo gene and the CC49 chimeric light chain are indicated by arrows in pRL150 and pRL105 The portions generated from pSV2neo are indicated These plasmids were purified on a large scale from preparative scale (1.0 L) by fermentation of E coli clones that replicated each of the plasmids. The purified plasmids were used to introduce the chimeric CC49 light chain into SP2/0 plasmacytoma cells as discussed below

Mouse CC83\A region and human Ck region

The Hind III site in pRL200 that was lost in the cloning process of the CC83 light chain was restored for the same reason as for the CC49 light chain chimeric construct. The restoration was performed as follows. Plasmid pRL200 was cloned into a unique Nhe I site, and both its sticky ends were converted to blunt ends by complementing with dNTPs is DNA polymerase I A Barn Hl phosphorylated linker (purchased from New England Biolabs) was ligated to the filled-in site The new plasmid is called pRL201 and is shown below

The 2 5 kb Barn Hl-Pst I fragment from pRL201 containing the CC83 light chain-vanable region genomic DNA was appropriately ligated to the 4 kb Barn Hl-Pst I vector fragment from pRL104 that was described previously in the CC49 light-chain constructs and which already had the Hind lll-carrying intron fragment The new plasmid is named pRL202 and is shown below The approximately 5.05 kb Fsp l-Hind III fragment from pRL202 was isolated and ligated with the human cK-containing 5.0 kb Hind lll-Bam Hl fragment already described for the CC49 light-chain chimeric construct The generation of the CC83 light-chain vector was performed from this point in an identical manner as for the CC49 light chain The resulting 8.5 kb Barn Hl CC83 light-chain chimeric construct was also ligated into pSV2neo-Bam H1 (phosphatase treated), and plasmids with both possible orientations of the insert were obtained as shown in the diagram below

The transcription orientations of the neo gene and the CC83 chimeric light chain are indicated by arrows in pRL203 and pRL 230. These plasmids were purified on a large scale from preparative scale (1.0 L) fermentation in a commercial incubator of E coli clones replicating each of the plasmids The purified plasmids were used to introduce the chimeric CC83 light chain into SpP2/0 plasmacytoma cells, as discussed later

All four of the chimeric light chain plasmid constructs (pRL105, pRL150, pRL203 and pRL 230) can be ligated by digesting with the restriction enzyme Aat II The Aat II site in the plasmids is in a region that is not essential for the expression of the chimeric light chain gene or the selectable marker gene , neo

Chimenske heavy kieders

Human Gamma constant gene exons

The plasmid vector used to carry out the chimeric heavy chain constructions is designated pSV2gpt, set forth in Mulligan and Berg, "Selection of animal cells that express the E coli gene coding for xanthine guanine phosphonbosyltransferase", Proe Nati Acad Sei (USA) 78,(4 ), 2072-2076 (1982) pSV2gpt is a pBR322 derived plasmid containing the selectable marker gene, guanine phosphonbosyl transferase (gpt), which can be used for selective growth in media containing myciphenolic acid To prepare pSVgpt as a recipient for the human C-> 1, Cy2, Cy3, Cy4 exons, it was digested with Eco RI and Barn Hl. The digested DNA was fractionated on a 4% polyacrylamide gel, and the 4.5 kb vector fragment was recovered from the gel by electroelution as described in Maniatis. This linearized plasmid was designated pSV2gpt/R/B, a plasmid map is shown in Figure 22 It is possible to receive fragments with Eco R1-Bam H1 ends

The 5' Hind III sites, present on the human Igd constant region fragments, were converted to Eco RI sites for directed cloning into the Eco RI site in pSV2-gpt For y1, y2, y3 and y4, the Eco RI site in vector pBR322 was used

Cy1

The fragment containing the human Cy1 exons was obtained by cleaving and ligating py1 with Hind III followed by filling in the Hind III sticky ends using all four dNTPs and the Klenow fragment from DNA polymerase to blunt the Hind III ends An Eco RI linker was ligated to the blunt ends to replace the Hind III site with an Eco RI site This construct was then digested with Eco RI and Barn I to release a 7.8 kb fragment containing the Cy1 exons This fragment was named Cy1 -7.8 kb

The fragment was ligated into the Eco Rl-Bam Hl sites of pSV2-gpt/R/B. The design of this vector (pSV2-gpt-y1-7,2) enables us to insert a mouse heavy chain variable region gene (with Eco ends) into the Eco RI site of the human IgG heavy chain vectors More specifically, 125 ng of the human Cy1-7.8 kg fragment was ligated to 100 ng of the linearized pSV2-gpt-R/B vector at 10 µl using 400 units of T4 DNA hgase (obtained from New England Biolabs) Frozen competent E coli DH1 cells from Invitrogen (San Diego, CA) were transformed by a ligation reaction according to Invitrogen's procedure The resulting plasmid was designated pSV2gpty1-7,8 A plasmid map of pSV2gpty 1-7,8 is shown in Figure 23

In addition, another shorter fragment containing the Cy1 exons was generated. of 1.9 kb (total = 14.2 kb) prompted the need to reduce the large size of the 7.8 kb Cy1 Eco Rl-Bam Hl fragment The coding region of 7.8 kb Cy1 occupies only the first 1/3 of 5" end of the fragment

Size reduction was performed by converting a Pvu II site in the downstream direction to a Barn Hl site by butt-end addition of a Barn Hl linker The Hind III site in py-1 was converted to an Eco RI site by cleavage of py-1 with Hind III, filling in the 3'-end to form a butt-end and addition of Eco RI linkers that above the Pvu II site 2.3 kb downstream was converted to a Barn Hl site by subsequent cleavage with Pvu II and ligation of Barn Hl linkers directly to the blunt Pvu II ends This construct was then digested with Eco RI and Barn Hl to release a 2.3 kb fragment containing the Cy1 exons The shortened EcoRI-BamHI fragment (2.3 kb) still contains the y1 exons and 3' polyadenylation sequence This reduces the total vector size by 5.5 kb, making the overall construct more manageable (total = 8.7 kb)

Approximately 200 ng of the human Cy1 2s.3 kb fragment was ligated to 100 ng of the linearized plasmid pSV2pgt/R/B vector in a volume of 10 µl using 400 units of T4 DNA ligase (New England Biolabs) Frozen competent coli cells, obtained from Invitrogen, were transformed with the ligation reaction according to Invitrogen's procedure. The resulting plasmid was designated pSV2pgty1-2.3 A plasmid map of pSV2pgty1-2.3 is shown in Figure 24

DNA fragments containing the other three human IgG constant region exons were also isolated. The Cy2 exons were recovered from plasmid py2 as a 4.0 kb Eco Rl-Bam Hl fragment. The Cy3 exons were recovered from plasmid py3 as an 8.0 kb Eco Rl -Bam Hl fragment The Cy4 exons were recovered from plasmid Cy4 as a 7.6 kb Eco Rl-Bam Hl fragment The fragments were ligated separately into ipSV2gpt/R/B as described for Cy1-7,8 and Cy1-2,3 Plasmid map of the resulting plasmids are shown in Figure 25, pSV2gpt-y2, Figure 26, pSV2gpt-y3, and

Figure 27, pSV2gpt-y4

Heavy-duty Chimen constructions

The complete heavy chain variable region human y1 constant region chimeric constructs were generated by inserting a fragment containing the mouse heavy chain variable region exons into the plasmids containing the human y1 constant region exons described as follows

Eco RI fragments containing the mouse heavy chain variable region genes from CC49 and CC83 hybdenoma cells were then ligated into each of the y1-y4-containing pSV2-gpt vectors (pSV2gpt-y1, pSV2gpt-y2, pSV2gpt-y3, pSV2gpt-y4). as follows

CC49

A fragment containing the heavy chain variable region exons encoding the CC49 heavy chain variable region was prepared by digesting 14 µg of pHH49 with 50 units of Eco RI (obtained from BRL) at 37°C for 2 hours. The digested fraction was fractionated on a 4% polyacrylamide gel, and the 1.9 kb Eco RI fragment containing the heavy chain variable region exons for CC49 was recovered by electroelution as described by Maniatis This fragment was designated f49R

A fragment containing the 7.8 kb sequence encoding y1 was prepared as follows

Approximately 50 µg of the vector pSV2gpt y1-7.8 was digested with Eco RI. The resulting fragment was dephosphorylated (to prevent cell ligation) using calf intestinal alkaline phosphatase as described by Maniatis. The fragment was purified from the 0.8% agarose gel by electroelution. This vector was designated pSV2gpty1-7.8/R

The Eco RI site is located 245 bp upstream from the transcription initiation sites, and contains the promoter and the necessary tissue-specific sequences for efficient expression The intron regions 3' of the variable region genes contain the mouse heavy chain enhancer sequences that are absent on the human IgG heavy chain vectors Therefore, the heavy chain chimeric vectors use both mouse promoter and enhancer sequences

Approximately 325 ng of annealed pVS2gpty1-7.8/R was ligated with 188 ng of f49R in a volume of 10 / l with 1 unit of T4 DNA hgase (BRL). Frozen competent E coli AG-1 cells from Stratagene were transformed with the ligation reaction in according to their method The resulting plasmid was designated p49y1-7.8 Figure 28 illustrates the plasmid map for p49y1-7.8

Approximately 50 µl of the vector pSV2gpty1-2.3 was digested as for SV2gpty1-7.8 with Eco RI. The resulting fragment was dephosphorylated using calf intestinal alkaline phosphatase as described by Maniatis. The fragment was purified from a 0.8% agarose gel by electroelution This cloned plasmid was designated pSV2gpty1-2.3/R

Approximately 300 ng of the ligated plasmid pSV2pgty1-2.3/R was ligated with 188 ng of f49R in a volume of 10 µl with 1 unit of T4 DNA hgase (BRL) Frozen competent E coli AG-1 cells from Stratagene ( La Jolla, CA) was transformed with the ligation reaction according to their method The resulting plasmid was designated p49y1 -2.3 Figure 29 illustrates a plasmid map for p49y1 -2.3

Plasmids pSV2gpt-y2, pSV2gpt-y3 and pSV2gpt-y4 were digested separately with Eco RI to produce the linear plasmid vectors pSV2gpt-y2/R, pSV2gpt-y3/R and pSV2gpt-y4/R, respectively. Each of these three linear plasmid vectors was ligated separately with f49R Plasmid maps of the resulting plasmids are shown in Figure 30, p49-y2, Figure 31, p49-y3, and Figure 32, p49-y4

CC83

Chimeric constructs containing the CC83 heavy chain variable region were generated similarly to the chimeric constructs for CC49 A fragment containing the heavy chain variable region exons encoding the CC83 heavy chain region was prepared by cleaving 19 µg of pHS83 with 50 units Eco RI (obtained from BRL) at 37°C for 2 h. The cleaved fraction was fractionated on a 4% polyacrylamide gel, and the 2.9 kb Eco RI fragment containing the heavy chain variable region exons from CC83 was recovered by electroelution as described in Maniatis This fragment was designated f83R

Approximately 300 ng of the ligated plasmid pSV2gpty1-7.8/R, obtained as above, was ligated with 270 ng of f83R in a volume of 10 µl with one unit of T4 DNA igase (obtained from BRL) Frozen competent E coli AG -1 cells, obtained from Stratagene, were transformed with the ligation reaction according to Stratagene's procedure The resulting plasmid was designated p83y1-7.8 Figure 33 illustrates the plasmid map of p83y1-7.8

Approximately 90 ng of linearized plasmid pSV2gpty1-2,3/R, obtained as above, was ligated with 270 ng of f83R in a volume of 10//I with 1 unit of T4 DNA-ltgase (BRL) Frozen competent E co//-AG- 1 cells from Stratagene were transformed with the ligation reaction according to their method The resulting plasmid was designated p83y1,2,3 Figure 34 illustrates the plasmid map of p83y1 -2,3

Plasmids pSV2gpt-y2, pSV2gpt-y3 and pSV2gpt-y4 were cleaved separately as above for pSV2gpt-y2/R, pSV2gpt-y3/R and pSV2gpt-y4/R with Eco RI to produce the linear plasmid vectors pSV2gpt-y2/R, respectively , pSV2gpt-y3/R and pSV2gpt-y4/R Each of these 3 linear plasmid vectors was ligated separately with f83R Plasmid maps for the resulting plasmids are shown in Figure 35, p83-y2, Figure 36, p83-y3 and Figure 37, p83- y4

All ten of the circular plasmid constructs (p49y1-7.8, p49y1-2.3, p83y1-7.8, p83y1-2.3. p49y2, p83-y2, p49-y3, p49-y4 and p83-y4) were so linearized for transformation by cleavage with restriction enzyme Nde I The Nde I site in the plasmids is in a region that is not essential for expression of the chimeric immunoglobulin gene or the selectable marker gene since the plasmids need to be in a linear form before transformation into a recipient cell to enhance selective integration of DNA into host cell genome DNA use of ammopterin and thymidine (to further block the guanine pathway) found to be unnecessary

Generation of clones producing chimeric 44 antibody

CH44-1

49K-13-13 cells were used as targets for chimeric heavy chain constructs. The cells were transformed with 20 fg of chimeric heavy chain DNA vector (p49y1-7,8 or p49y1-2,3) obtained by Nde I cleavage. Transformation by electroporation was performed as above for chimen light chains

However, selection after 48 hours was performed by replacing the Geneticin-containing medium with medium containing Geneticin and 0.3/<y>g/ml mycophenolic acid, 250 //g/ml xanthine and 10//g/ml hypoxanthine

Transformed cells grew into macroscopically visible colonies in 14 days. At that time, 50 µl of supernatant was removed and measured by ELISA methods for binding to TAG and expression of human IgG constant region. Wells containing cells with positive TAG binding were expanded into 24-well plates with fresh drug selection medium and allowed to grow for 3-7 days

Subcloning was performed as follows Viable cell numbers were determined and the cells were plated into two 96-well plates One plate received 50 live cells and the other received 250 live cells The non-subcloned cells were expanded into 6-well plates until the cell density was sufficient to allow storage in liquid nitrogen in the event re-subcloning would be necessary

After subcloning, the clones that exerted the highest chimeric antibody production were selected for chimeric antibody production in bioreactors

CH44-2

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH44-1 were repeated with the exception that 20 fg P49-y2 was used as the chimeric heavy chain vector

CH44-3

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH44-1 were repeated with the exception that 20 fg p49-γ3 was used as the heavy chain vector

CH44-4

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells in the CH44-1 construct were repeated with the exception that p49-y4 was used as the heavy chain vector

Generation of clones producing chimeric 88 antibody

CH88-1

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells of the construct into CH44-1 were repeated with the following exceptions

83K-26-5, 83K-34-10 and 83K-42-2 cells showing chimeric CC83 light chain production were transformed as described in the transformation of CH44-1, except that 20 fjg (p83y1-7.8 or p83y1-2,3, the pSV2gpt vector containing the CC83 heavy chain gene was used as the heavy chain vector

CH88-2

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells in the construction of CH88-1 were repeated with the exception that 20 fg p83-y2 was used as the heavy chain vector

CH88-3

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH88-1 were repeated with the exception that 20 fg p83-y3 was used as the heavy chain vector

CH88-4

The procedures used to sequentially transform the Sp2/0 plasma cytocells into the CH88-1 construct were repeated except that 20 fg p83-γ4 was used as the heavy chain vector

The linearized DNA was ethanol precipitated and further dissolved in 110-20 microliters of PBS. After 15 minutes on ice, electroporation was performed using a Gen Pulser electroporation apparatus with added capacitance increaser (BioRad) at 0.2 kilovolts and 960 µF. The time constant ( r ) was mainly about 26 milliseconds

After transformation, cells were allowed to recover on ice for 115 min to allow relaxation of agitated membranes. Afterwards, cells were suspended in 14 ml of RPMI 1640 medium containing 5% fetal calf serum (RPMI+) and transferred to a 96 or 24 well plate. To minimize the likelihood for more than one drug-resistant cell per well, the cells were also diluted 10 times in medium (RPMI+) and plated on another 96-well (or 24-well) plate The cell suspension was incubated at 37°C and 5% CO2 atmosphere

After 48 hours (to allow expression of drug resistance), the medium was removed and replaced with medium containing 1 mg/ml Geneticin

After 7-10 days, Geneticin-resistant clones were sub-cultured, and the cells mapped for chimeric light chains by cytostaining

Cytostaining

Small sample amounts of cells were pelleted on a glass plate using a cytospin-2 centrifuge (Shandon, Inc ) After air drying, the cells were fixed in acetic acid/ethanol (5 parts acetic acid/95 parts ethanol) After rinsing 3 times with PBS (without CA< +2> and Mg<+2>), the plates were placed in a humid chamber (100% RH) and stained for 20 minutes with 20 //I goat anti-human Kappa-FITC, a fluorescent dye-conjugated antibody specific for human kappa light chains The conjugated antibody was diluted 1 3 with 1% BSA in PBS After washing overnight with PBS, the plates were mounted with fluoromont-G, histological mounting medium (obtained from Southern Biotech) under a coverslip The plates were observed with an Olympus model BH-2 microscope equipped with an epi-illumination UV to-heftnmg

Based on the intensity of fluorescence, the constructs with the orientation of the light chain in the opposite direction of transcription relative to the direction of transcription of the neo gene in the vector were found to give the highest L-chain expression. Therefore, pRL 105 was the preferred CC49L chain construct and pRL230 was the favored CC83L chain construct. As a result of these experiments, the following chimeric light chain-containing cell lines (derived from Sp2/0) were used for the targeted transformations

For the CC49 chimaenic L chain, a cell line (49K-13-13) was obtained that expressed the chimaenic light chain formed from CC49. This cell line was used for all further targeted transformations with chimaenic heavy chain vectors for constructs using the chimaenic CC49 light chain chain

For the CC83 chimenic light chain, 3 cell lines (83K-26-5, 83K-34-10 and 83K-42-2) were obtained that expressed the chimenic light chain formed from CC83 One cell line (83K-26-5) stained more strongly than the others and had localized areas of cytoplasmic immunofluorescence All three cell lines were compared for their relative ability to produce high levels of chimeric antibody after transformation with the chimeric CC83 g1 heavy chain vector Several clones expressing chimeric antibodies were generated from electroporation of 83K- 34-10 target than from each of the other two chimeric light chain target cell lines Therefore, the 83K-34-10 light chain cell line was used as a target for further electroporations with chimeric heavy chain vectors for constructs containing the CC83 light chain variable region

Generation of gpt-resistant clones carrying CC49 and CC83 chimeric H-linked constructs

Prior to transformation with pSV2-gpt vectors containing chimeric heavy chain constructs, drug selection was established for inhibition of growth of untransformed Sp2/0 plasmacytoma cells [obtained from the American Type Culture Collection (ATCC)] Conditions for drug selection of cells transformed with pSV2-gpt vectors were more difficult to create. The E coli gpt gene, which encodes the enzyme guanosine phosphonbosyl transferase, transfers the ability to use xanthine and hypoxanthine as substrates for the biosynthesis of guanine when the guanine metabolism pathway in mammals is inhibited by mycophenolic acid (MPA).

Published values for the concentrations of PMA that allow growth of other lymphoid cell lines transformed with pSV2-gpt vectors were found to be almost two orders of magnitude too high to allow growth of Sp2/0 cells transformed with pSV2-gpt vectors in our tissue culture environment. concentration of 0.1 //g/ml of MPA found to be optimal for selection of gpt resistance In addition,

Verification of construction

Since the EcoRI fragments can be ligated in either orientation, the correct orientation was determined by digestion with Nco I. In the constructs presented above, correct ligation for plasmid construction is confirmed by performing restriction enzyme site mapping on the plasmid. individual starting fragments Due to experience with the transcription orientation in the light chain vectors, the heavy chain vectors were constructed only in the opposite transcription orientation relative to the gpt gene

Transformation of plasmids into mouse plasma cvtoma cells

When both the light chain and heavy chain chimene genes were transformed into the same cell, tetrameric (H2L2) immunoglobulins are obtained. Synthesis and secretion of these "chimene" antibody proteins was accomplished by introducing the chimene (mouse V human C region) genes into mouse plasma cvtoma cells

(Sp2/0)

Transformation was achieved by electroporation [Sahagan, B G et al,

J Immunology 137,1066 (1986)]

Expression of chimeric (mouse V human C region) genes in transformed mouse plasma cvtoma cells (Sp2/0) is achieved by using two different techniques In one way different ratios of light chain genes and heavy chain genes can be introduced together This is referred to as cotransformation Alternatively stable clones carrying the chimeric light chain gene, is obtained and further used in another way referred to as targeted transformation In each method, the aim is to obtain clones containing genes for both the H chain and the L chain that produce intact H2L2 immunoglobulin mentioned above

A_ Cotransformations

Cotransformation involves transformation of cells with both drug resistance markers at the same time and further selection with one or both drugs Cotransformation of heavy chain and light chain vectors (in ratios of 1 1 and 1 10 respectively) was originally performed using only neo-selection Neo was obtained -resistant cell lines that expressed the first chimeric IgG1 antibodies with demonstrable TAG72 binding activity. Cotransformation was performed in accordance with the methods set forth in Corneha Gorman, "High Efficiency Gene Transfer into Mammalian Cells", DNA Clonin<q>, Vol II, D M Glover ed IRL Press , Oxford, England (1985)

B_ Targeted transformations

Constructs containing light and heavy chimeric immunoglobulin genes were sequentially transformed into Sp2/0 mouse plasma cvtoma cells Targeted transformation involves transformation and selection with a vector containing a first drug resistance gene (ie Geneticin for the chimeric light chain gene vector), followed by transformation and selection with a vector containing another drug resistance gene (ie mycophenolic acid for the chimeric heavy chain gene vector)

Neo-selection

Before transformation with pSV2-neo vectors containing chimeric light chain constructs, drug selection conditions for inhibition of growth of non-transformed Sp2/0 plasmacytoma cells [obtained from the American Type Culture Collection (ATCC)] were established by titration of the neomycin-like drug, Geneticin (GIBCO) Published values for concentrations of Geneticin used for drug selection ranged from 100-1000 //g/ml Concentrations above 400 //g/ml were found to prevent growth of Sp2/o cells in our tissue culture media

Construction of light-weight-containing cells

Sp2/0 mouse plasma cvtoma cells were initially transformed with light chain-containing pSV2 neovectors as follows Cells were grown in RPMI

1640 medium with 5% fetal calf serum Cells were washed in PBS and suspended to a concentration of 1 x 10<7> live cells/ml PBS 0.8 ml cells were transferred to an electroporation cuvette (on ice) containing 20 //g light chain -containing pSV2 neovector (pRL105 and pRL150 for CC49 chimeric light chain and pRL 203 and pRL230 for CC83 chimeric L chain) annealed with Aat II residue junction endonuclease Aatll was inactivated by heating the samples to 65°C for 110 minutes

The linearized DNA was ethanol precipitated and further dissolved in 110-20 microliters of PBS. After 15 minutes on ice, electroporation was performed using a Gen Pulser electroporation apparatus with added capacitance enhancer (BioRad) at 0.2 kilovolts and 960 jjF. The time constant (r) was essentially approx. 26 milliseconds

After transformation, cells were allowed to recover on ice for 115 min to allow relaxation of agitated membranes. Afterwards, cells were suspended in 14 ml of RPMI 1640 medium containing 5% fetal calf serum (RPMI+) and transferred to a 96 or 24 well plate. To minimize the likelihood for more than one drug-resistant cell per well, the cells were also diluted 10 times in medium (RPMI+) and plated on another 96-well (or 24-well) plate The cell suspension was incubated at 37°C and 5% CO2 atmosphere

After 48 hours (to allow expression of drug resistance), the medium was removed and replaced with medium containing 1 mg/ml Geneticin

After 7-10 days, Geneticin-resistant clones were sub-cultured, and the cells mapped for chimeric light chains by cytostaining

Cytostaining

Small sample amounts of cells were pelleted on a glass plate using a cytospin-2 centrifuge (Shandon, Inc ) After air drying, the cells were fixed in acetic acid/ethanol (5 parts acetic acid/95 parts ethanol) After rinsing 3 times with PBS (without CA+ 2 and Mg<+2>), the plates were placed in a humid chamber (100% RH) and stained for 120 minutes with 20 µl of goat anti-human Kappa-FITC, a fluorescent dye-conjugated antibody specific for human kappa light chains The conjugated antibody was diluted 1 3 with 1% BSA in PBS After washing overnight with PBS, the plates were mounted with fluoromont-G, histological mounting medium (obtained from Southern Biotech) under a coverslip The plates were observed with an Olympus model BH -2 microscopes equipped with an epi-illumination UV for— adhesion

Based on the intensity of fluorescence, the constructs with the orientation of the light chain in the opposite direction of transcription relative to the direction of transcription of the neo gene in the vector were found to give the highest L-chain expression. Therefore, pRL 105 was the preferred CC49L chain construct and pRL230 was the favored CC83L chain construct. As a result of these experiments, the following chimeric light chain-containing cell lines (derived from Sp2/0) were used for the targeted transformations

For the CC49 chimaenic L chain, a cell line (49K-13-13) was obtained that expressed the chimaenic light chain formed from CC49. This cell line was used for all further targeted transformations with chimaenic heavy chain vectors for constructs using the chimaenic CC49 light chain

For the CC83 chimenic light chain, 3 cell lines (83K-26-5, 83K-34-10 and 83K-42-2) were obtained that expressed the chimenic light chain formed from CC83 One cell line (83K-26-5) stained more strongly than the others and had localized areas of cytoplasmic immunofluorescence All three cell lines were compared for their relative ability to produce high levels of chimeric antibody after transformation with the chimeric CC83 g1 heavy chain vector Several clones expressing chimeric antibodies were generated from electroporation of 83K- 34-10 target than from each of the other two chimeric light chain target cell lines Therefore, the 83K-34-10 light chain cell line was used as a target for further electroporations with chimeric heavy chain vectors for constructs containing the CC83 light chain variable region

Generation of qpt-resistant clones carrying CC49 and CC83 chimeric H-gated constructs

Before transformation with pSV2-gpt vectors containing chimeric heavy chain constructs, drug selection was established for inhibition of growth of untransformed Sp2/0 plasmacytoma cells [obtained from the American Type Culture Collection (ATCC)] Conditions for drug selection of cells transformed with pSV2-gpt vectors were more difficult to create. The E coli gpt gene, which encodes the enzyme guanosine phosphonbosyl transferase, transfers the ability to use xanthine and hypoxanthine as substrates for the biosynthesis of guanine when the guanine metabolism pathway in mammals is inhibited by mycophenolic acid (MPA).

Published values for the concentrations of PMA that allow growth of other lymphoid cell lines transformed with pSV2-gpt vectors were found to be nearly two orders of magnitude too high to allow growth of Sp2/0 cells transformed with pSV2-gpt vectors in our tissue culture environment. concentration of 0.1 //g/ml of MPA found to be optimal for selection of gpt resistance l in addition, the use of aminoptene and thymidine (to further block the guanine pathway) was found to be unnecessary

Generation of clones producing chimeric 44 antibody

CH44-1

49K-13-13 cells were used as targets for chimeric heavy chain constructs. The cells were transformed with 20 µg of chimeric heavy chain DNA vector (p49y1-7.8 or p49y1-2.3) linearized by Nde I cleavage. Transformation by electroporation was performed as above for Chimen light chains

However, selection after 48 hours was performed by replacing the Genetcin-containing medium with medium containing Geneticin and 0.3 µg/ml mycophenolic acid, 250 µg/ml xanthine and 10 µg/ml hypoxanthine.

Transformed cells grew into macroscopically visible colonies in 14 days. At that time, 50 µL of supernatant was removed and measured by ELISA methods for binding to TAG and expression of human IgG constant region. Wells containing cells with positive TAG binding were expanded into 24-well plates with fresh drug selection medium and allowed to grow for 3-7 days

Subcloning was performed as follows Viable cell numbers were determined and the cells were plated into two 96-well plates One plate received 50 live cells and the other received 250 live cells The non-subcloned cells were expanded into 6-well plates until the cell density was sufficient to allow storage in liquid nitrogen in the event re-subcloning would be necessary

After subcloning, the clones that exerted the highest chimeric antibody production were selected for chimeric antibody production in bioreactors

CH44-2

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH44-1 were repeated with the exception that 20 µg of P49-y2 was used as the chimeric heavy chain vector

CH44-3

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH44-1 were repeated with the exception that 20 fg p49-γ3 was used as the heavy chain vector

CH44-4

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH44-1 were repeated with the exception that 20 µg of p49-y4 was used as the heavy chain vector

Generation of clones producing chimeric 88 antibody

CH88-1

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells of the construct into CH44-1 were repeated with the following exceptions

83K-26-5, 83K-34-10 and 83K-42-2 cells showing chimeric CC83 light chain production were transformed as described in the transformation of CH44-1, except that 20 fjg (p83y1-7.8 etler p83y1-2,3, the pSV2gpt vector containing the CC83 heavy chain gene, was used as the heavy chain vector

CH88-2

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells in the construction of CH88-1 were repeated with the exception that 20 fg p83-y2 was used as the heavy chain vector

CH88-3

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the construct in CH88-1 were repeated with the exception that 20 fg p83-y3 was used as the heavy chain vector

CH88-4

The procedures used to sequentially transform the Sp2/0 plasma cytocells into the CH88-1 construct were repeated except that 20 fg p83-γ4 was used as the heavy chain vector

Generation of clones producing chimeric 84 antibody

Due to the high degree of sequence similarity between the heavy chain variable regions of CC49 and CC83, chimeric antibodies were generated whose light and heavy chains were formed from different parents by mixed targeted transformations. To form both "mixed" combinations, the chimeric heavy chain y1 isotype vectors in CC49 and CC83 electroporated into the chimeric light chain targets 83K34-10 and 49K-13-13, respectively. The resulting cell lines were designated CH48-1 and CH84-1, where the first numerical designation represents the heavy chain and light chain parents respectively For example, CH48-1 represents the y1 isotype with the heavy chain formed from CC49 and the light chain formed from CC83

The CH48-1 composite antibody did not bind to TAG72. This was not due to the inability to make genuine chimeric antibody, since most drug-resistant cell lines produced chimeric IgG (as determined by ELISA analysis using goat anti-human IgG trap with goat anti-human IgG alkaline phosphatase as probe) If any binding affinity was present, it was significantly less than that observed for the first-generation antibody B72.3, which had about an order of magnitude less affinity for TAG72 than both CC49 and CC83

Surprisingly, CH84-1 bound to TAG72 with affinity hk both parents This novel antibody was newly generated in our laboratory and has not yet been demonstrated in nature

Competition studies were performed to determine the specificity of this new mixed antibody, CH84-1. It should be noted that both CC49 and CC83 exert some degree of competitive recognition for the TAG72 antigen. It was found that CH84-1 competed more with CC49 for binding to TAG72 than it did with CC83 This would indicate that the specificity for binding to TAG72 lies in the light chain

Human γ2, -3, and -4 isotypes were also generated with this mixed antibody, producing CH84-2, CH84-3, CH84-4 clones

CH84-1

The procedure used to sequentially transform the Sp2/0 plasmacytoma cells at the concentration in HC44-1 was repeated with the following exceptions

49K-13-13 cells, which showed production of CH44 light chain by cytostaining, were then transformed as described in that transformed into CH44-1, with the exception that 20 fg of p83y1-2.3, the pSV2gpt vector containing the CH83 heavy chain gene, were substituted for p49y1-2,3, the pSV2gpt vector containing the CH44 heavy chain gene

CH84-2

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells into the CH84-1 construct were repeated except that 20 µg of p83-y2 was substituted for p83y-2,3

CH84-3

The procedures used to sequentially transform the Sp2/0 plasmacytoma cells in the construct into CH84-1 were repeated with the exception that 20 fg p83-y3 was substituted for p83y1-2,3

CH84-4

The procedures used to sequentially transform

The Sp2/0 plasmacytoma cells in the construct in CH84-1 were repeated with the exception that 20 fg p83-y4 was substituted for p83y1-2,3

Purification of recombinant antibodies

Cells expressing the chimene antibodies were removed by centrifugation from the culture medium, and the medium was filtered through a 0.2 µm filter. Chimene antibodies were purified in two steps from culture supernatants. In the first step of the purification, a protein A affinity cartridge (Nygene Corporation, Yonkers, NY) according to the manufacturer's specifications. Up to 1.0 L of culture supernatant was passed through a 1 mg capacity cartridge at 5 ml/min. The cartridge was washed with phosphate-buffered saline (PBS) to remove traces of albumin. antibody was recovered by elution with 0.5M sodium nitrate buffer, pH 3.0 pH of the fractions containing the chimeric antibody was immediately adjusted to neutrality with a 1M solution of Trizma base Final purification was obtained from this solution after concentration on an Amicon centncon 30 unit, by gel filtration using a Pharmacia Superose 12 HR 16/5 column as described by the manufacturer (Pharmacia, Piscataway, NJ)

EXAMPLE Formation of an immunoglobulin containing the mouse vVjaTAq germline variable region

The following examples are set forth to provide the skilled artisan with a reproducible technique for producing an antibody having a VH region encoded by a DNA sequence formed from VHaTAG

Components for a VHaTAG tungkie dough that can be expressed

A mouse-human chimeric antibody can be generated that contains the mouse VHaTAG germline heavy chain variable region, a light chain variable region complementary to the VnaTAG VH, such as either the CC49 or CC83 mouse light chain vanable region, and human constant regions

The 2.2 kb Hind Ili germline DNA fragment containing the VHaTAG VH exon sequence is used as a template to obtain a functionally rearranged VnaTAg variable region The mouse genomic J-C/j mtron region is used as a source for the mouse heavy chain enhancer sequences This latter region is obtained from plasmid pNP9 (see example above of "Isolation of CC49 Heavy Chain Variable Region") Figure 38 shows the entire reaction for the preparation of the hybrid gene based on the method of Horton et al, (1989), supra Four oligonucleotides (oligos) are designed to be used in a enzymatic amplification and modification of target DNA Oligo 1 is base-paired to the 5' end of VHaTAG which spans the EcoRI site which is 249bp 5' in relation to the ATG initiation codon Oligo 2 is base-paired to sequences complementary to the 3'-end of the VhoTAG exon and also contains sequences that codes for a D-segment D-segment the sequences in Oligo 2 basepartils do not match with any VHaTAG sequences Oligo 3 contains sequences co complementary to the 5'-end in the mouse genome J-C/j region and incorporates sequences that code for the D-segment (same as in Oligo 2) and the J-segment Oligo 4 basepartial fits to the 3'-end in the J-C/j region and contains sequences complementary to the EcopRI site located 1219 pb 3' in relation to JH4 The sequences for these oligos follow

In this example, the D sequence SP2.3 is taken from the published sequence of Kurosawa and Tonegawa J Exp Med , 155 201 (1982) The D sequence is shown in bold type in oligos 2 and 3 Any other characterized mouse or human D- segment can be used by substituting their sequence in these positions in Oligo 2 and 3

The J segment in Oligo 3 is underlined It is mouse Jh4 taken from the published sequence of Gough and Bernard Proe Nati Acad Sei (USA), 78 509

(1981) The inclusion of any other mouse or human J segment can be done by substitution of their sequences for that of Jh4 in Oligo 3

In Oligo 1 and 4, the EcoRI sites (GAATTC) are shown in italics

Collection of intact VH aTAG genes

Two separate DNA amplification reactions are performed using the components described above DNA amplification reaction #1 copies the VnaTAG sequence and adds a D segment to its 3' end DNA amplification reaction #2 copies the mouse intron sequences containing the heavy chain enhancer sequences and adds D and J- the segments coded for in oligo 3 The amplified products from reactions 1 and 2 are gel purified, combined, and oligos 1 and 4 are added to start reaction #3 In reaction 3, base match the products from reactions 1 and 2 over their common D sequences Further DNA -amplification from oligos 1 and 4 gives the product shown at the bottom of Figure 38 This fragment is digested with EcoRI and gel purified The modified VnaTAG fragment is ligated into the EcoRI site in pSV2gpty 1(2,3) as described in the example above, "heavy chain chimeric constructions" The entire VHaTAG-D-J enhancer containing fragment is completely sequenced to ensure that no mutations have been introduced during DNA amplification the reactions The other three heavy chain y isotypes can be formed by ligating the same modified VhoTAG fragment into the other three y containing pSV2gpt vectors (pSV2gpt-y2, pSV2gpt-y3, pSV2gpt-y4)

Expression of the modified VhctTAG gene

The modified VhoTAG gene containing plasmids can be linearized with Ndel and introduced via electroporation into the chimeric CC49 or CC83 light chain expressing cell lines see example above, "C Targeted Transformations" The transformed cells are selected for growth in the presence of Geniticin and mycophenolic acid as described above in "C Targeted transformations" Presence of expressed antibody is measured by the TAG72 ELISA (see section in RESULTS, Enzyme-linked immunoassays (ELISA) The expressed antibody from these cells will contain human lg y1, k constant regions with the CC49 or CC83 light chain vanable region and a heavy chain functional region from the modified VhoTAG germline VH exons

Four examples of modified VHcrTAG heavy chain vane site constructs having a vanadate of D and J segments are shown below,

The sequence of the human D sequence D1 is obtained from Siebenlist et al, Nature, 294,63 (1981) The sequence of the human Jh1 is obtained from Ravetch et al, Cell 27, 583 (1981)

The formation of VHaTAG #i is described with oligos 1 to 4 in the diagram above To form VhoTAG #m to - iv the corresponding D and J segments need to be changed in oligos 2 and 3 The following oligos align these changes Substitution of these oligos in reaction #1 and reaction #2 will result in the formation of VGaTAG #n to - iv

Results

A_ Chimen antibody - Producing cell hnie

Simultaneous detection of heavy and light chains was performed using two probe antibodies 1) Goat anti-human kappa labeled with the fluorescent dye FITC and, 2) Goat anti-human IgG labeled with the fluorescent dye TRITC

Cell lines that are positive reactions for both heavy and light chains were further tested for associated chimeric immuno-globuhn production and biological activity, namely binding to TAG72

Enzyme-linked immunoassays (ELISA)

To select a transformed cell producing a chimeric monoclonal antibody, the ELISA technique was used The clones containing the heavy chain and light chain drug selection constructs were selected by their weight in selective culture medium The following cell lines were tested (1) CH44-1 a cell line which had CC49VH, CC49 VL, and constant region IgGi, (2) CH44-2 a cell line that had CC49 VH, CC49 VL, and constant region lgG2, (3) CH44-4 a cell line that had CC49 VH, CC49 VL, and cash area lgG4,

(4) CH88-1 a cell line that had VH, CC83 VL, constant region IgGi,

(5) CH88-2 a cell line that had CC83 VH, CC83 VL and constant region lgG2, (6) CH88-3 a cell line that had CC83 VH, CC83 VL, and constant region lgG3, (7) CH88-4 a cell line that had CC83 VH, CC83 VL, and constant region lgG4, (8) CH84-1 a cell line that had CC83 VH, CC49 VL, and constant region IgGi, (9) CH84-2 a cell line that had CC83 VH, CC49 VL, and constant region lgG2, (10) CH84-3 a cell line that had CC83 VH, CC49 VL, and constant region lgG3, and (11) CH84-4 a cell line that had CC83 VH, CC49 VL, and constant region lgG4

Supernatants and these cultures were subjected to ELISA. The presence of chimeric anti-TAG72 antibody was measured directly by reaction with an excess of goat anti-human IgG antibody labeled with an enzyme such as alkaline phosphatase, after allowing the chimeric anti-TAG72 antibody to bind. onto microtiter wells coated with antigen (TAG72) Anti-TAG72 activity was determined as a criterion for successful recombination

After growth 114 days, 50 μl of supernatant was removed from the wells of the subcloned cells and remeasured for TAG binding by ELISA. Samples of supernatants (50 μl) from drug-resistant cell lines were added to wells of Nunc Immulon 96-well plates previously had been coated with TAG antigen

(1 50 dilution) After washing to remove unbound material, the wells were incubated with goat anti-human IgG antibodies conjugated with alkaline phosphatase (GAHIgG-AP) as a probe to detect the human constant regions in the chimeric antibodies that had bound to TAG the antigen immobilized on the plate. One more wash to remove unbound probe (GAHIgG-AP) followed by the addition of a chromogenic alkaline phosphatase substrate allowed color to develop in the wells that had TAG binding associated with human constant regions (ie, chimeric anti- TAG72 antibodies) Absorbance readings at 405 nm indicate the relative amount of chimeric antibody produced by the drug-resistant cell lines

CH44-1

Anti-TAG72 activity was used as the criterion for successful recombination Wells of microtiter plates were coated with TAG by incubating 50 µl of a 1 75 dilution of purified TAG72 [Muraro, R , et al, Cancer Research 48, 4588^596 (1988 )] 118 hours at room temperature The wells were then washed 4 times with phosphate buffered saline (PBS), and then blocked with BSA by incubating 50 μl of 0.5% BSA in PBS for 2 hours at 37°C, followed by washing 4 times with PBS These plates are stable if kept moist at 4°C 50 microliters of sample was then applied to each well A blank containing fresh medium is used as a control All samples were incubated either in the plate for 90 minutes at 37°C or overnight at 4°C in a closed container

The plates were then washed 4 times with PBS, and goat anti-human IgG alkaline phosphate (Southern Biotech Assoc ) was applied to each well by adding 50 microliters of a 1,250 dilution. The solution was incubated at 37°C for 90 minutes. Color development was checked after washing the plates 4 times with PBS to remove the probe

The substrate was incubated in a 200 µl solution of substrate p-nitrophenyl phosphate (Kirkegaard & Perry) in ethanolamine-buffered saline for 16 minutes at room temperature for color development. The optical density at 450 nm in each well was read using a Dynatech microplate reader (Dynatech Inc )

The Sp2/0 colonies with supernatant wells that had TAG72-binding chimeric antibody activity were subcloned by limiting dilution Individual subclones were selected on the basis of relatively high production of chimeric antibody

CH44-2

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH44-2

CH44-3

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH44-3

CH44-4

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH44-4

CH88-1

The TAG-ELISA procedure used with CH44-1 was repeated with an exception that the antibody was CH88-1

CH88-2

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH88-2

CH88-3

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH88-3

CH88-4

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH88-4

CH84-1

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was C84-1

CH84-2

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH84-2

CH84-3

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH84-3

CH84-4

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH84-4

CH48-1

The TAG-ELISA procedure used with CH44-1 was repeated with the exception that the antibody was CH48-1

B In vivo Carcinom targeting

The chimene monoclonal antibodies used in animal studies and shown in Tables 1-4 below were labeled with Na<125>l using iodine (Pierce Chemical, Rockford, IL) More specifically, from about 0.5-2 mg of purified chimene monoclonal antibodies adjusted to 0.5 ml of 0.1 M sodium phosphate buffer (pH 7.2) and then added to a 12 cm x 75 cm glass tube coated with 50 µg of iodine followed by the addition of from 0.1-0.5 mCi Na< 125>l (New England Nuclear, Boston, MA) After a 2 minute incubation at room temperature, the protein was removed from the insoluble gel, and the unincorporated <125>l was separated from the antibody by gel filtration through a 10 ml Sephadex™ G column -25 using PBS as a buffer the dilution procedure gave labeled IgG chimeric antibody with a specific activity of 0.05 to 0.2 /jCi//yg

Female athymic mice (nu/nu) on a CD1 background were obtained from Charles River at approximately 4 weeks of age. Nine days later, mice were inoculated subcutaneously (0.1 ml/mouse) with LS174T cells (1 x 10<6> cells/ animals)

Athymic mice bearing carcinomas of 70 to 400 mg in weight, approximately 12 to 13 days after inoculation of the cells, were given injections intravenously of from 0.5 to 2.0 fjC\ (10-50 fjg protein) in PBS of the chimen monoclonal antibodies that had been soldered as described above Groups of five mice euthanized at varying times by exsanguination, carcinoma and normal tissues were excised and weighed, and cpm was measured in a gamma counter cpm/mg in each tissue was then determined and compared to which was found in the carcmomet

The results for CH44-1 are shown in Tables 1-2 and Figures 39A, 39B and 39C The results for CH84-1 are shown in Tables 3-4 and Figures 40A and 40B

Percentage dose per gram of 125l-labelled antibody

As shown in Table 1, at approximately 122 hours after injection, the percent injected dose to tumor for CC44-1 was 44.16 percent CH44-1 was therefore effective in targeting the human tumor in situ This shows that the chimeric monoclonal antibodies produced by the method according to the present invention, were effective for in vivo carcinoma targeting and thus are applicable for in vivo treatment of cancer

Percentage inserted dose per organ of 12<5>l labeled antibody

As shown in Table 2, at 122 hours after injection, the percent injected dose in tumor for CH44-1 was 4.55% CH44-1 was therefore effective in targeting the human tumor in situ This shows that the chimeric monoclonal antibodies produced by the method according to the present invention were effective for in vivo carcinoma targeting and thus were applicable for in vivo treatment of cancer

Percentage inserted dose per gram of 125l-labeled antibody

As shown in Table 3, at approximately 118 hours after injection, the percent injected dose to tumor for CH84-1 was 30.27% CH84-1 was therefore effective in targeting the human tumor in situ This shows that the chimeric monoclonal antibodies produced by the method according to the present invention, were effective for in vivo carcinoma targeting and thus were applicable for in vivo treatment of cancer

Percentage inserted dose per organ 125l-labeled antibody

As shown in Table 4, at approximately 118 hours after injection, the percentage of injected dose to tumor for CH84-1 was 7.02% CH84-1 was therefore effective in targeting the human tumor in situ This shows that the chimeric monoclonal antibodies produced by the method according to the present invention, were effective for in vivo carcinoma targeting and thus were applicable for in vivo treatment of cancer

Deposition of cell lms producing chimeric antibodies

Eleven illustrative cell lines secreting chimeric antibodies, all having kappa light chains, made by the above-mentioned examples, were deposited with the American Type Culture Collection (ATCC) on October 19, 1988. Specifically, the following cell lines have been deposited.

(1) CH44-1 a cell line having CC49VH, CC49 VL, and constant region lgG1 (ATCC No. 9884), (2) CH88-2 a cell line having CC83VH, CC83VL and constant region lgG2 (ATCC No. HB9880), (3 ) CH44-4 a cell line having CC49 VH, CC49 VL, and constant region IgG4 (ATCC No. 9877), (4) CH88-1 a cell line having CC83 VH, CC83 VL, and constant region IgG-i (ATCC No. 9882 ), (5) CH44-2 a cell line having CC49 VH, CC49 VL, and constant region lgG2 (ATCC no. 9881), (6) CH88-3 a cell line having CH83 VH, CC83 VL, and constant region lgG3 (ATCC No. 9876), (7) CH88-4 a cell line having CC83 VH, CC83 VL, and constant region lgG4 (ATCC No. 9874), (8) CH84-1 a cell line having CC83 VH, CC49 VL, and constant region Igd (ATCC No. 9883), (9) CH84-2 a cell line having CC83 VH, CC49 VL, and constant region lgG2 (ATCC No. 9879), (10) CH84-3 a cell line having CC83 VH > CC49 VL, and constant region lgG3 (ATCC No. 9878), and (11) CH84-4 a cell line having CC83 VH, CC49 VL, and constant region lgG4

(ATCC no. 9875)

Claims (8)

1 Method for the production of a therapeutically active antibody, or a therapeutically active antibody fragment thereof, where the antibody or antibody fragment is capable of selectively reacting with TAG-72 antigen and is obtained from one of the cell lines CH44-1 (ATCC HB 9884), CH44-2 (ATCC HB 9880), CH44-4 (ATCC HB 9877), CH88-1 (ATCC HB 9882), CH88-2 (ATCC HB 9881), CH88-3 (ATCC HB 9876), CH88 -4 (ATCC HB 9874), CH 84-1 (ATCC HB 9883), CH84-2 (ATCC HB 9879), CH84-3 (ATCC HB 9878), or CH84-4 (ATCC HB 9875), characterized in that a of the mentioned cell lines are cultured in a suitable medium, after which the antibody or an antibody fragment thereof is isolated 2 Method for producing a therapeutic antibody according to claim 1, further comprising producing an antibody conjugate or antibody fragment conjugate, where the antibody or antibody fragment is contacted with a therapeutic agent, characterized in that a corresponding method is used 3 Method according to claim 2, where the antibody or antibody fragment according to claim 1 is conjugated to a therapeutic agent, characterized in that a corresponding method is used 4 Method according to claim 3, where the therapeutic agent is a radionuclide, drug or biological response modifier, toxin or other antibody, characterized in that a corresponding method is used 5 Method according to claim 4, where the radionuclide is 1311, ^ Y, 105Rn,47 Sc, 67Cu, 212Bl, 211At, 6<7>G3i 125|, 186Re> 188Re, 177Lu> 99mTc, 153Sm, 123| or H1|n, characterized in that a corresponding method is used 6 Method according to claim 4, where the drug or the biological response modifier is methotrexate, adriamycin or lymphokine, characterized in that a corresponding method is used 7 Method for producing a therapeutically active antibody or therapeutically active antibody fragment thereof according to claim 1, further comprising contacting the antibody or antibody fragment thereof with a pharmaceutically acceptable, non-toxic solid carrier, characterized in that a corresponding method is used 8 Method according to any one of claims 2-6, further comprising contacting the antibody conjugate or antibody-fragment conjugate with a pharmaceutically acceptable, non-toxic sterile carrier, characterized in that a corresponding method is used