COVALENT DISULFIDE-LINKED DIABODIES AND USES THEREOF

Abstract

The present invention provides recombinant antibody fragments which include a variable domain which has been modified by the addition of a tail sequence to its C-terminal end. The tail sequence comprises a terminal cysteine residue and an amino acid spacer and does not substantially affect the fragment's target-binding affinity. The present invention also provides pharmaceutical compositions comprising the described antibody fragments and a pharmaceutically acceptable ⋅ carrier and methods of delivering an agent to cells of interest in a subject using the fragments as delivery vehicles. The invention further provides compositions comprising the described antibody fragments for the in vitro detection and measurement of target molecules which bind to the fragments and method of determining the presence or amount of such targets in a biological sample by contacting the sample with such compositions.

Description

CROSS-REFERENCE TO RELATED APPLICATION

The present application is related to and claims priority under 35 U.S.C. § 119(e) to U.S. provisional patent application No. 60/420,271 filed on Oct. 23, 2002. This application is incorporated herein by reference.

GOVERNMENT RIGHTS STATEMENT

The invention described herein was made with Government support under grant numbers P01 CA 43904 from the National Institutes of Health and DAMD17-00-1-0150 from the Department of Defense. Accordingly, the United States Government has certain rights in this invention.

FIELD OF THE INVENTION

The present invention relates to recombinant antibody fragments, nucleic acids encoding such recombinant antibody fragments, and methods of using such recombinant antibody fragments, particularly for in vivo delivery of agents to specific cells of interest.

BACKGROUND OF THE INVENTION

Throughout this application, various publications are referenced. Disclosures of these publications in their entireties are hereby incorporated by reference into this application to more fully describe the state of the art to which this invention pertains. Full bibliographic citations for the references may be found listed immediately preceding the claims.

Antibodies specific for tumor-associated antigens can provide effective vehicles for in vivo delivery of agents, such as radionuclides, for detection or therapy of cancer. The potential utility of cancer-targeting antibodies can be improved by protein engineering approaches, which can be used to modify characteristics such as affinity, immunogenicity, and pharmacokinetic properties. In particular, recombinant antibody fragments have been produced with favorable characteristics, including retention of high affinity for target antigen, rapid, high level accumulation in xenografts in murine models, and quick clearance from the circulation, resulting in high tumor:normal activity ratios. Furthermore, because antibody fragments do not persist in the circulation, they are less likely to be immunogenic than intact murine or even chimeric antibodies. With the advent of humanized and human antibodies, the issue of immunogenicity of recombinant antibodies is rapidly diminishing.

Recombinant fragments such as diabodies (e.g., 55 kDa dimers of single-chain Fv fragments, which self-assemble in a cross-paired fashion as described by Holliger, et al., 1993) or minibodies (e.g., 80 kDa scFv-C_H3 fusion proteins as described by Hu et al., 1996)) have shown promise as in vivo imaging agents in preclinical studies when radiolabeled with single-photon emitting radionuclides such as In-111 or 1-123, or positron emitters such as Cu-64 or 1-124 for positron emission tomography (Sundaresan et al., In press; Wu et al. 2000). Targeting and imaging of 1-123 radiolabeled single-chain Fv (scFv, 27 kDa) fragments has been demonstrated clinically, although the size and monovalency of scFv's may limit their utility (Begent, et al., 1996). Recent clinical imaging studies using I-123 radiolabled diabodies appear promising (Santimaria et al. 2003).

Most current antibody radiolabeling approaches involve conjugation to random sites on the surface of the protein. For example, standard radioiodination methods result in modification of random surface tyrosine residues. Many antibodies are highly susceptible to inactivation following iodination, presumably due to modification of key tyrosines in or near the binding site. (Nikula et al., 1995; Olafsen et al., 1996). Chemical modification of lysines located in or near the antigen-binding site could also potentially interfere with binding through sterical hindrance if a bulky group is added (Benhar et al., 1994; Olafsen et al., 1995). Alternative iodination approaches or radiometal labeling through conjugation of bifunctional chelates direct modifications to ϵ-amino groups of lysine residues, again randomly located on the surface of antibodies. The issue of inactivation following radiolabeling becomes more pressing as one moves to smaller and smaller antibody fragments, if equal reactivity is assumed, because the binding site(s) represent a larger proportion of the protein surface, and fewer “safe” sites for conjugation are available.

Site-specific radiolabeling approaches provide a means for both directing chemical modification to specific sites on a protein, located away from the binding site, and for controlling the stoichiometry of the reaction. Several strategies capitalize on naturally occurring moieties or structures on antibodies that can be targeted chemically. For example, the carbohydrate found on constant domains of immunoglobulins can be oxidized and conjugated with bifunctional chelates for radiometal labeling. (Rodwell, et al., 1993). In one instance, an unusual carbohydrate moiety occurring on a hypervariable loop of a kappa light chain was modified for site-specific chelation and radiometal labeling (Leung, et al., 1995). Others have exploited selective reduction of interchain disulfide bridges to enable modification using thiol-specific reagents. C-terminal cys residues on antibody Fab or Fab′ fragments have been used for direct labeling using ^99mTc (Behr, et al., 1995; Verhaar, et al., 1996). Novel approaches include the identification of a purine binding site in antibody Fv fragments, allowing specific photoaffinity labeling (Rajagopalan, et al., 1996).

More recently, genetic engineering approaches have been used to introduce specific sites for modification or radiolabeling of proteins and antibodies. Building on the above-mentioned work, glycosylation sites have been engineered into proteins to provide novel carbohydrate targets for chemical modification (Leung, et al., 1995; Qu, et al., 1998). The six-histidine tail commonly appended to recombinant proteins to provide a purification tag has been used in a novel 99mTc labeling method (Waibel, et al., 1999). Alternatively, a popular strategy has been to use site-directed mutagenesis to place cys residues on the surface of proteins to provide reactive sulfhydryl groups. This approach has been implemented by numerous groups to allow site-specific labeling of antibodies (Lyons, et al. 1990; Stimmel, et al., 2000) and other proteins (Haran, et al., 1992; Kreitman, et al., 1994).

Introduction of cys residues into engineered antibody fragments also has been used for stabilization or multimerization purposes. For example, introduction of strategically placed cys residues in the interface between the V_H and V_L domains of antibody Fv fragments has allowed covalent linkage and stabilization of these fragments (disulfide-stabilized Fv, or dsFv) (Glockshuber, et al., 1990; Webber, et al., 1995). Fitzgerald et al. described a disulfide bonded diabody in which cysteine residues were introduced into the V_L/V_H interface for stability and demonstrated its utility for fluorescent imaging of tumors (Fitzgerald, et al., 1997). Others have appended cys residues to the C-termini of single-chain Fv fragments (scFv, formed by fusing V_H and V_L with a synthetic peptide linker) to allow multimerization into scFv′₂ fragments (Adams, et al., 1993; Kipriyanov, et al., 1995).

We have previously produced an anti-carcinoembryonic antigen (anti-CEA) diabody from the murine anti-CEA T84.66 antibody by joining V_L—eight amino acid linker—V_H. Tumor targeting, imaging, and biodistribution studies of a radiolabeled (at random sites on the protein) anti-CEA diabody demonstrated rapid tumor uptake, fast clearance from the circulation, and favorable properties for use as an imaging agent, when evaluated in nude mice bearing LS14T xenografts (Wu, et al., 1999; Yazaki et al., 2001b).

There remains a need in the art, however, for a stable, in vivo delivery vehicle that can be modified readily in specific locations without affecting the ability of the vehicle to specifically target cells of interest. There is also a continuing need for better in vitro detection methods. The invention provides a system for adding site-specific functional groups to antibody fragments that do not interfere with target binding by said fragment.

SUMMARY OF THE INVENTION

The present invention provides recombinant antibody fragments for use in in vivo delivery of agents for detection and treatment of diseases, primarily cancers. The present invention also provides recombinant antibody fragments for the in vitro detection of certain targets of interest. Preferred antibody fragments comprise at least two single chain polypeptide subunits, each subunit having a heavy-chain variable domain polypeptide sequence connected by a linker sequence to a light-chain variable domain polypeptide sequence. One of the variable domain polypeptide sequences in each subunit is modified at its C-terminal end by addition of a tail sequence. The tail sequence comprises a terminal cysteine residue and an amino acid spacer. The selection of the C-terminal end as the modification site for introduction of the tail sequence provides a location at the end of the variable region opposite the target combining site. This configuration avoids interference with antigen binding and thus, the addition of the tail sequence does not substantially affect the antibody fragment's target-binding affinity. A “target” in the context of the present invention is a molecule of interest that can bind to or complex with the antibody fragments of the present invention and includes any molecule against which an antibody can be isolated. Examples of a target include an antigen, an anti-antibody, a self antigen or a hapten. Accordingly, “target binding sites” in the context of the present invention are sites of antibody fragments that bind “targets” and include, but are not limited to, antigen binding sites.

The subunits assemble such that each heavy chain domain is bound to a light chain domain, thereby providing a specific target-binding site with each such light chain/heavy chain pairing. Moreover, in preferred embodiments of the invention, the addition of the tail sequence provides a disulfide covalent bond, or bridge, between the heavy chain variable regions or between the light chain variable regions, depending on to which variable region the tail sequence was added. Advantages of the bond include, but are not limited to, added stability and the presence of thiol groups in an internal, protected location, which then can be released when desired for site-specific chemical modification.

The present invention also provides pharmaceutical compositions comprising the described antibody fragments and a pharmaceutically acceptable carrier. The invention further provides methods of delivering an agent to cells of interest in a subject. Preferred delivery methods involve conjugating the agent to the recombinant antibody fragment and administering the conjugate to the subject under conditions permitting specific binding between the fragment and the cell of interest in the subject.

The present invention also provides in vitro diagnostic methods for detecting in a biological sample at least one target of interest. In such an diagnostic method complexes between at least one recombinant antibody fragment described herein and at least one target are detected.

Finally, it surprisingly has been found that the addition of a tail sequence having a terminal cysteine residue and an amino acid spacer provides advantages over known antibody fragment structures, including the formation of a stable disulfide bond and ease of site-specific chemical modification.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1a shows the design of anti-CEA cys-diabodies with a parental diabody with an 8 amino acid linker sequence (SEQ ID NO: 9) between the V domains and a C-terminus 5 amino acid sequence present in the V_H domain (SEQ ID NO: 10). (1) and (2) show cys-diabody variants made with the added amino acids highlighted and the cysteine residue underlined in each construct (SEQ ID NO: 2 and SEQ ID NO: 4). FIG. 1b is a schematic drawing of a non-covalent bound diabody.

FIG. 2 shows SDS-PAGE analysis of purified antibody fragment proteins in accordance with the invention.

FIG. 3 shows results of a competition ELISA between an embodiment of the present invention, a known antibody, and a known antibody fragment.

FIGS. 4a and 4b show traverse slices of serial microPET scans of a mouse bearing bilateral C6 (arrow) and LS174T (arrowhead) xenografts, injected with ⁶⁴Cu cys-diabody (57 μCi) and imaged at 4 hrs (4a) and 18 hrs (4b).

FIGS. 5a and 5b show the published crystal structure of the parental anti-CEA diabody (5a), and a model of the structure of -SGGC cys-diabody (5b) where the Fvs have been rotated to bring the C-termini close enough for disulfide bridge formation.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a recombinant antibody fragment comprising at least two single chain polypeptide subunits. Each subunit comprises a heavy-chain variable domain polypeptide sequence connected by a linker sequence to a light-chain variable domain polypeptide sequence. The linker sequence is preferably a glycine sequence typically having from about 5 to about 10 glycine residues, preferably 6, 7, 8 or 9 residues and most preferably 8 residues. One of the variable domain polypeptide sequences in each subunit is modified at its C-terminal end by addition of a tail sequence. See FIG. 1.a. The tail sequence comprises an amino acid spacer and a terminal cysteine residue. The spacer, positioned between the terminal cysteine residue and the end of the sequence to which the tail is added, is preferably between 1 and about 10 residues in length, more preferably 2 to 5 residues in length, most preferably 2 residues in length. However, depending on the distance of the C-termini, a longer tail sequence might be advantageous, e.g. a tail sequence having 11, 12 or more amino acid residues. In a preferred embodiment, the residues are glycine residues. The subunits assemble such that each heavy chain domain is bound to a light chain domain, wherein each such light chain/heavy chain pairing provides a target-binding site. Following addition of the tail sequence, each target-binding site retains binding affinity substantially similar or equivalent to that of a recombinant fragment without addition of the tail sequence, as described above and illustrated by FIG. 3. “Substantially similar binding affinity” in the context of the antibody fragments according to the present invention describes an in vitro binding affinity that a person skilled in the art would consider not inferior when compared to the in vitro binding affinity of the same antibody fragment without a tail sequence.

In a preferred embodiment, the tail sequence is added to the heavy chain variable domain of each subunit and provides a disulfide covalent bond between the heavy chain variable domains. Alternatively, the tail sequence is added to the light chain variable domain of each subunit and provides a disulfide covalent bond between the light chain variable domains.

In a preferred embodiment, at least one of the target binding sites of the recombinant antibody fragment specifically binds a carcinoembryonic antigen (CEA). A preferred diabody is an anti-CEA diabody. In a preferred embodiment, at least one light chain and at least one heavy chain of the fragment correspond to a light chain variable domain and a heavy chain variable domain of the murine anti-CEA T84.66 antibody. The tail sequence preferably is added to the heavy chain variable domain of each subunit and provides a disulfide covalent bond between the heavy chain variable domains. Alternatively, the tail sequence is added to the light chain variable domain of each subunit and provides a disulfide covalent bond between the light chain variable domains.

In another preferred embodiment, the recombinant antibody fragment further comprises an agent, such as a diagnostic or therapeutic agent, conjugated to the cysteine residue of the tail sequence. This conjugation can be achieved readily by methods known in the art. The agent can be conjugated to the fragment via a thiol-specific bifunctional chelating agent or other suitable chelating agent. The agent can be, without limitation, a radionuclide label such as In-111 or 1-123, a positron emitter, such as Cu-64 or 1-124 or, in particular for the in vitro applications, a dye, such as a fluorescent dye for direct detection by colorimetric assays, for fluorescent detection including Fluorescence Activated Cell Sorting (FACS) or for Fluorescence Resonance Energy Transfer (FRET); a protein such as horseradish peroxidase or alkaline phosphatase, that generates a colored product with an appropriate substrate for ELISA type assays, or a luciferase that generates light upon addition of an appropriate substrate. Alternatively, the agent can be a cytotoxic agent such as a chemotherapeutic drug.

Another embodiment comprises an in vitro detection method in which an antibody fragment that has formed a complex with a target is detected via a labeled antibody or antibody fragment that is specific for a site of the fragment, attachment to which does not interfere with the fragment's ability to bind a target. Such a site is preferably the terminal LGGC or SGGC sequence.

The tail sequence for at least one of the variable domains preferably can comprise the sequence set forth in SEQ ID NO: 2. Alternatively, the tail sequence preferably can comprise the sequence set forth in SEQ ID NO: 4. The invention further provides nucleotide sequences encoding the antibody fragments described herein (SEQ ID NO: 5 and SEQ ID NO: 7).

The invention moreover provides pharmaceutical compositions which comprise a recombinant antibody fragment as described herein and a pharmaceutically acceptable carrier. These pharmaceutical compositions can be administered in accordance with the present invention as a bolus injection or infusion or by continuous infusion. Pharmaceutical carriers suitable for facilitating such means of administration are well known in the art. Examples of suitable pharmaceutical carriers are described in “Remington's Pharmaceutical Sciences” by E. W. Martin. Liquid formulations can be solutions or suspensions and can include vehicles such as suspending agents, solubilizers, surfactants, preservatives, and chelating agents.

The present invention also provides a method of delivering an agent to cells of interest in a subject which includes conjugating the agent to a recombinant antibody fragment and administering the conjugate, (or “conjugated recombinant fragment”) to the subject. As described above, the recombinant antibody fragment comprises at least two single chain polypeptide subunits in which each subunit comprises a heavy-chain variable domain polypeptide sequence connected by a linker sequence to a light-chain variable domain polypeptide sequence. One of the variable domain polypeptide sequences in each subunit is modified at its C-terminal end by addition of a tail sequence comprising a terminal cysteine residue and an amino acid spacer. The spacer, positioned between the terminal cysteine residue and the end of the sequence to which the tail is added, is preferably between 1 and about 10 residues in length, most preferably about 2 residues in length. In a preferred embodiment, the residues are glycine residues. The single chain subunits assemble such that each heavy chain domain associates with a light chain domain, each such light chain/heavy chain pairing thereby providing a target-binding site. At least one target-binding site specifically binds the cells of interest.

The conjugate is then administered to the subject under conditions permitting the conjugate to specifically bind targets on the cells of interest, which in turn thereby achieves delivery of the agent to cells of interest in the subject. Methods of in vivo targeting which are suitable for delivering conjugated antibody fragments are known in the art and described in the example provided herein. The subject is preferably a human patient. The amount to be administered to a human patient can be determined readily through methods known in the art, including those based on animal data. The cells of interest are preferably cancer cells and can be, without limitation, colon cancer cells, breast cancer cells, lung cancer cells, lymphoma cells, or cells from other human malignancies and other human diseases or conditions. The agent is preferably a detectable label. In a preferred embodiment of the invention, following administration of the conjugate to the subject, the detectable label can be detected to determine the location of the cells of interest in the subject. The detectable label can be a radioisotope, a thiol-specific label, an optical or fluorescent probe, or any other suitable label known in the art.

Alternatively, the agent can be a cytotoxic agent, such as a chemotherapeutic drug or radionuclide. Methods in accordance with the present invention also can be used to diagnose, without limitation, autoimmune, inflammatory, or angiogenic processes by selecting an appropriate antibody from which to derive the antibody fragment.

The invention also provides methods of detecting and quantitatively determining the concentration of a target in a biological fluid sample. In one embodiment the method comprises contacting a solid support with an excess of a certain type of antibody fragment which specifically forms a complex with a target, such as a tumor associated antigen, e.g., CEA, under conditions permitting the antibody fragments to attach to the surface of the solid support. The resulting solid support to which the antibody fragments are attached is then contacted with a biological fluid sample so that the target in the biological fluid binds to the antibody fragments and forms a target-antibody complex. The complex can be labeled with a detectable marker. Alternatively, either the target or the antibody fragments can be labeled before the formation the complex. For example, a detectable marker can be conjugated to the antibody fragments as described elsewhere herein. The complex then can be detected and quantitatively determined thereby detecting and quantitatively determining the concentration of the target in the biological fluid sample.

A biological fluid according to the present invention includes, but is not limited to tissue extract, urine, blood, serum, and phlegm. Further, the detectable marker includes but is not limited to an enzyme, biotin, a fluorophore, a chromophore, a heavy metal, a paramagnetic isotope, or a radioisotope.

Further, the invention provides a diagnostic kit comprising an antibody fragment that recognizes and binds an antibody or antibody fragment against a target; and a conjugate of a detectable label and a specific binding partner of the antibody or antibody fragment against the target. In accordance with the practice of the invention the label includes, but is not limited to, enzymes, radiolabels, chromophores and fluorescers.

In light of the preceding description, one of ordinary skill in the art can practice the invention to its fullest extent. The following example, therefore, is merely illustrative and should not be construed to limit in any way the invention as set forth in the claims which follow.

EXAMPLE

In the present example, cysteine residues were introduced into an anti-CEA diabody at different locations, in order to provide specific thiol groups for subsequent chemical modification. Modified proteins carrying an added C-terminal gly-gly-cys sequence were shown to exist exclusively as a disulfide-bonded dimer. This “cys-diabody” retained high binding to CEA and demonstrated tumor targeting and biodistribution properties identical to the non-covalent diabody. Furthermore, following reduction of the disulfide bond, the cys-diabody could be chemically modified using a thiol-specific bifunctional chelating agent, to allow labeling with radiometal. Thus, the disulfide-linked cys-diabody provides a covalently linked alternative to conventional diabodies, and, following reduction, generates specific thiol groups that are readily modified chemically. This format provides a useful platform for targeting a variety of agents to cells, such as for example, CEA-positive tumors.

In order to allow site-specific radiolabeling using thiol-specific reagents, the present example describes mutant anti-CEA diabodies engineered by substitution or addition of unique cys residues. Two variants, namely a variant having the C-terminal sequence-LGGC, and a variant having the C-terminal sequence-SGGC, were found to exist as a stable disulfide-linked dimer. The former demonstrated equivalent antigen binding in vitro and tumor targeting in vivo compared to the parental diabody shown in FIG. 1. a., and had the added advantage of allowing site-specific chemical modification following reduction of the interchain disulfide bridge. SEQ ID NO: 5 represents the nucleotide sequence of the VTVS-SGGC Cys diabody used in the examples. SEQ ID NO: 6 represents the respective amino acid sequence. SEQ ID NO: 7 represents the nucleotide sequence of the VTVS-LGGC Cys diabody used in the examples. SEQ ID NO: 8 represents the respective amino acid sequence.

Materials and Methods

Design of Cysteine-Modified Diabodies (Cys Diabodies)

Two variants of anti-CEA diabodies (FIG. 1, I & II; Table 1) were constructed by PCR-based mutagenesis (QuickChange Site Directed Mutagenesis Kit, Strategene, LaJolla, Calif.) of the pEE12 expression vector (Lonza Biologics, Slough, UK) containing the original anti-CEA diabody constructed with an 8 amino acid glycine-serine linker (GS8) (Wu, et al., 1999). The two constructs contained a cysteine at the C-terminus of the V_H and two glycines inserted in front of the cysteine as a spacer. One retained the original C-terminal sequence of the V_H with GGC appended (VTVS-SGGC) and in the other, serine 113 was exchanged to a leucine (VTVS-LGGC).

Mammalian Expression, Selection and Purification

1×10⁷ NSO cells (provided by Lonza Biologics)(Galfe and Milstein, 1981) were transfected with 40 μg of linearized vector DNA by electroporation and selected in glutamine-deficient media as described (Yazaki, et al., 2001a). Clones were screened for expression by ELISA, in which the desired protein was captured either by protein L or by a recombinant CEA fragment, N-A3 (You, et al., 1998) and detected using alkaline phosphatase-conjugated goat anti-mouse Fab antibodies (Sigma, St. Louis, Mo.). Supernatants also were examined by Western blot for size analysis, using the alkaline phosphatase-conjugated goat anti-mouse Fab antibodies. The best producing clones were expanded. Cys-diabodies were purified from cell culture supernatants, using a BioCad 700E chromatography system (Applied Biosystems, Foster City, Calif.) as described (Yazaki et al. 2001a). Briefly, the supernatants were treated with 5% AG1®-X8 (Bio-Rad Labs, Hercules, Calif.) overnight to remove phenol red and cell debris and then dialyzed versus 50 mM Tris-HCl, pH 7.4. Treated supernatant was loaded onto an anion exchange chromatography column (Source 15Q, Amersham Pharmacia Biotech AB, Uppsala, Sweden), and proteins were eluted with a NaCl gradient to 0.2 M in the presence of 50 mM HEPES, pH 7.4. Eluted fractions, containing the desired protein, were subsequently loaded onto a Ceramic Hydroxyapatite (Bio-Rad Laboratories, Hercules, Calif.) column and eluted with a KPi gradient to 0.15 M in the presence of 50 mM MES, pH 6.5. Fractions containing pure proteins were pooled and concentrated by Centriprep 10 (Amicon Inc., Beverly, Mass.). Elution was monitored by absorption at 280 nm. The concentration of purified protein per ml was determined by OD₂₈₀, but also by applying a small sample on protein L using known amounts of parental diabody and later cys-diabody standards quantitated by amino acid composition analysis (Wu, et al., Immunotechnology, 1996).

Characterization of Purified Cys-Diabodies

Purified proteins were analyzed by SDS-PAGE pre-cast 4-20% polyacrylamide Ready gels (Bio-Rad Laboratories under non-reducing and reducing (1 mM DTT) conditions and stained using MicrowaveBlue™ (Protiga Inc., Frederick, Md.). Samples were also subjected to size-exclusion HPLC on Superdex 75 (Amersham Biosciences). Retention time was compared to a standard of parental diabody. Binding to CEA was initially assessed by ELISA as described above. Competition/Scatchard was also carried out in ELISA microtiter plates wells coated with N-A3, using a fixed concentration (1 nM) of biotinylated chimeric T84.66 antibody, and increasing concentration of non-biotinylated competitors (0.01-100 nM). Displacement was monitored with alkaline phosphatase-conjugated streptavidin (1:5000 dilution) (Jackson ImmunoResearch Labs, West Grove, Pa.) and color was developed with Phosphatase substrate tablets (Sigma, St. Louis, Mo.) dissolved in diethanolamine buffer, pH 9.8. All experiments were carried out in triplicate.

Radioiodination

70 μg of purified cys-diabody was radiolabeled with 140 μCi Na¹³¹I (Perkin Elmer Life Sciences, Inc., Boston, Mass.) in 0.1 phosphate buffer at pH 7.5, using 1.5 ml polypropylene tubes coated with 10 μg Iodogen (Pierce, Rockford, Ill.). Following a 5-7 min. incubation at room temperature, the sample was purified by HPLC on Superdex 75. Peak fractions were selected and diluted in normal saline/1% human serum albumin to prepare doses for injection. The labeling efficiency was 85%. Immunoreactivity and valency were determined by incubation of radiolabeled protein with a 20-fold excess of CEA at 37° C. for 15 min., followed by HPLC size-exclusion chromatography on a calibrated Superose 6 column (Amersham Biosciences).

Biodistribution in Tumor-Bearing Mice

7-8 week-old female athymic mice were injected subcutaneously in the flank with 10⁶ LS174T human colon carcinoma cells (ATCC #CL-188). At 7 days post inoculation, mice bearing LS174T xenografts were injected with 1 μg of ¹³¹I-labeled cys-diabody (specific activity, 1.7 μCi/μg) via the tail vein. Groups of five mice were sacrificed and dissected at 0, 2, 4, 6, and 24 h post injection. Major organs were weighed and counted in a gamma scintillation counter. Radiouptakes in organs were corrected for decay and expressed as percentage of injected dose per gram of tissue (% ID/g) and as percentage of injected dose per organ (% ID/organ). Tumor masses ranged from an average of 0.580 mg (0 h group) to 1.058 mg (24 h group). Biodistribution data are summarized as means and corresponding standard errors (sem). Animal blood curves were calculated using ADAPT II software (D' Argenio and Schumitzky, 1979) to estimate two rate constants (k_i) and associated amplitudes (A_i).

Conjugation and Radiometal Labeling of Cys-Diabodies

The VTVS-LGGC cys-diabody was reduced and conjugated with a novel bifunctional chelating agent comprised of the macrocyclic chelate DOTA (1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid), a tetrapeptide linker, and hexanevinylsulfone group for chemical attachment to thiol groups. This compound, DOTA-glycylleucylglycyl-(ϵ-amino-bis-1,6 hexanevinylsufone)lysine, abbreviated DOTA-GLGK-HVS, is described in greater detail elsewhere (Li, et al., 2002). VTVS-LGGC cys-diabody (2 mg in 0.5 mL PBS) was reduced by treatment with 20 μL of 20 μM tris-(carboxyethyl)phosphine (TCEP) (Pierce) in PBS for 2 h at 37° under Ar and centrifuged through a Sephadex G25 spin column. DOTA-GLGK-HVS, (58 μL of 20 mM in PBS) was added and the solution rotated at 10 rpm for 4 h at 25° C. The conjugate was dialyzed against 0.25 M NH4OAc, pH 7.0. Extent of modification was evaluated by isoelectric focusing gels (Li, et al., 2002).

Radiolabeling of Cys-Diabody Conjugates with Copper-64

Copper-64 (copper chloride in 0.1 N HCL; radionuclide purity >99%) was produced in a cyclotron from enriched ⁶⁴Ni targets, at the Mallinckrodt Institute of Radiology, Washington University Medical Center (McCarthy, et al., 1997). DOTA-GLGK-HVS-conjugated cys-diabody (200 μg) was incubated with 7.3 mCi of ⁶⁴Cu in 0.1 M NH₄ citrate, pH 5.5 for 1 h at 43° C. The reaction was terminated by addition of EDTA to 1 mM. Labeled protein was purified by size exclusion HPLC on a TSK2000 column (30 cm×7.5 mm I.D.; Toso-Haas; Montgomeryville, Pa.). The radiolabeling efficiency was 56% and the specific activity was 1.7 μCi/μg.

MicroPET Imaging

CEA-positive (LS174T) and CEA-negative (C6 rat glioma) xenografts were established in nude mice by subcutaneous injection of 1-2×10⁶ cells subcutaneously into the shoulder area, 10-14 days prior to imaging. Mice were imaged using the dedicated small animal microPET scanner developed at the Crump Institute for Biological Imaging (UCLA) (Chatziioannou, et al., 1999). Mice were injected in the tail vein with 57 μCi of ⁶⁴Cu-diabody. After the appropriate time had elapsed, mice were anesthetized with a 4:1 mixture of ketamine (80 mg/kg body weight) and xylazine (10 mg/kg body weight, injected intraperitoneally), placed in a prone position, and imaged using the microPET scanner with the long axis of the mouse parallel to the long axis of the scanner. Acquisition time was 56 min. (8 min. per bed position; 7 bed positions), and images were reconstructed using MAP reconstruction algorithm (Qi, et al., 1998).

Results

Expression, Purification, and Characterization of Cys-Diabodies

The cys-diabodies were expressed in the mouse myeloma cell line, NSO. The expression levels for the VTVS-SGGC and VTVS-LGGC constructs were between 5-20 μg/mL as determined by ELISA. Cultures were expanded in T flasks and supernatants collected. The VTVS-SGGC and VTVS-LGGC cys-diabodies were purified essentially as described (Yazaki, et al., 2001).

Analysis of the purified proteins on SDS-PAGE demonstrated that the two-step purification scheme yielded VTVS-SGGC and VTVS-LGGC diabodies that were >95% pure (FIG. 2). Under non-reducing conditions both proteins migrated as single species in the range of 55-60 kDa, with the VTVS-LGGC version exhibiting slightly lower mobility (FIG. 2, lanes 1 and 2). Under reducing conditions both proteins demonstrated the presence of the expected 25 kDa monomer (FIG. 2, lanes 4 and 5). Thus, when purified under native conditions, both of these cys-diabodies existed as essentially pure disulfide-bonded homodimers. Incorporation of a serine (polar) versus a leucine (nonpolar) residue made no difference at position 113. One of ordinary skill in the art would recognize that most other amino acid substitutions at this position would be suitable with the probable exception of cys, (which would introduce another thiol) and proline (which would restrict flexibility).

Size exclusion chromatography demonstrated that the covalently linked cys-diabody was slightly smaller than the regular diabody, as it eluted at 20.42 min. (average of 5 experiments) as opposed to 20.13 min. (average of 4 experiments) (not shown).

The binding-activity of the cys-diabody to CEA was initially demonstrated by ELISA. Affinity was measured by competition ELISA in the presence of competitors at different concentrations. As shown in FIG. 3, by competition assay the affinities of the cys-diabody and parental diabody were essentially the same as that of the parental intact murine T84.66 monoclonal antibody.

The immunoreactivity and valency of the cys-diabody were analyzed following radioiodination by solution-phase incubation in the presence of excess CEA. Size-exclusion HPLC analysis demonstrated that 90% of the cys-diabody shifted to high molecular weight complexes indicated by two peaks, suggesting that the cys-diabody was bound to one and two CEA molecules (not shown).

In Vivo Biodistribution and Targeting

The ¹³¹I-labeled cys-diabody was assessed for its ability to target tumor in athymic mice bearing xenografts of LS174T human colon carcinoma cells. As can be seen in Table 2 the accumulation of the ¹³¹I-labeled cys-diabody reached 9.32% ID/g at 2 h and this level of localization was maintained at 4 and 6 hours post injection. Blood clearance was rapid and nearly complete by 18 h (0.55% ID/g), with the half life in the beta phase being 2.68 hrs., essentially the same as that observed with the non-covalently bound diabody (2.89 hrs.) (Yazaki et al., 2001 b). Activities in other normal organs (liver, spleen, lung, kidney) fell rapidly as well and were below 1% ID/g by 18 h. These biodistribution results were essentially identical to those observed for the parental anti-CEA diabody (Wu, et al., 1999).

In Vivo Imaging by microPET

Cys-diabody was conjugated with the macrocyclic chelate DOTA using a novel peptide-hexanevinylsufone derivative described in detail elsewhere (Li, et al., 2002). This allowed efficient radiolabeling with ⁶⁴Cu, a positron-emitting radionuclide with a 12.7 h half-life, well-matched to the targeting and clearance kinetics observed for diabodies in murine systems in vivo. MicroPET imaging studies were conducted on athymic mice bearing LS174T xenografts (CEA-positive human colorectal carcinoma) or C6 xenografts (CEA-negative rat glioma). Specific targeting to the CEA-positive xenograft was observed at 4 and 18 h post injection, with little evidence of activity in the CEA-negative tumor. The results are shown in FIGS. 4a and 4b. However, this particular combination of protein/chelate/radionuclide resulted in elevated liver activity (19.4% ID/g at 4 hrs) in addition to kidney activity (55.1% ID/g at 4 hrs) (Li et al. 2002).

Discussion

The present example provides the design, production and evaluation of a novel antibody format, a covalently-linked (disulfide-bonded) diabody which is referred to as the cys-diabody. Two constructs were made. The initial intent was to introduce cysteine residues into the anti-CEA diabody in order to provide specific sites for chemical modification including conjugation and radiolabeling. In course of these experiments, it was unexpectedly discovered that addition of the sequence GGC to the end of the protein resulted in a diabody in which the C-termini of the V_L-V_H subunits came together and formed a disulfide bond. Two slightly different versions of the cys-diabody (with C-terminal sequences of LGGC or SGGC) resulted in essentially 100% formation of the disulfide linkage. This protein provides various improvements over a standard diabody, including, but without limitation, covalent linkage for greater potential stability, and the feasibility of site-specific modification following reduction of the disulfide bond and generation of free reactive thiols. Additionally, the introduced cysteine residues are essentially “protected” through the internal linkage and prevented from forming random disulfide bonds with small molecules (such as glutathione) or other sulfhydryl-containing proteins present in the cell. As a result, the cys-diabody can be obtained in higher amounts and with greater purity than might be expected for proteins containing engineered cysteine residues that are unpaired and accessible to random chemical modification and fortuitous disulfide formation.

The crystal structure of the parental T84.66/GS8 diabody has recently been solved (Carmichael, et al., 2003). In the crystal, the Fv units of the diabody assumed a very compact, twisted structure, with the binding sites oriented in a skewed orientation at a tight 70° angle. The C-termini of the heavy chain variable regions (where we have appended the cys residues in the present example) are about 60 Å apart. However, the structure that was solved is likely to represent one of many conformations that the parental diabody can adopt. The fact that the cys-diabody forms with such high efficiency implies that the parental diabody is in fact quite flexible, and that the Fv domains can swivel such that the C-termini are juxtaposed. FIG. 5a shows the crystal structure of the diabody and FIG. 5b shows a model where the Fv domains have been rotated to bring the C termini close enough for disulfide bridge formation (the disulfide bridge is the light-colored loop protruding from the model in the central bottom part of the Figure). Size exclusion HPLC analysis confirms the notion that the native, non-covalent diabody is quite flexible; it elutes at an earlier retention time suggesting a larger Stokes radius. By contrast, the covalently-linked cys-diabody elutes at a later retention time, implying a more compact, more highly constrained molecule. Taken together, these results suggest that native diabody has a more open and flexible structure.

TABLE 1
Cys-diabody constructs
VTVS-SGGC: 5′-GTC ACC GTC TCC TCA GGT GGA TGT-3′ (SEQ ID NO: 1)
           Val Thr Val Ser Ser Gly Gly Cys (SEQ ID NO: 2)
VTVS-LGGC: 5′-GTC ACC GTC TCC TTA GGT GGA TGT-3′ (SEQ ID NO: 3)
           Val Thr Val Ser Leu Gly Gly Cys (SEQ ID NO: 4)

TABLE 2
Biodistribution of 131I-labeled T84.66 cys-diabody in
athymic mice bearing LS174T xenograftsa
Organ (% ID/g)	0	2	4	6	18	24
Tumor	2.49 ± 0.33	9.32 ± 1.73	10.02 ± 0.69	9.15 ± 0.77	6.43 ± 1.05	4.79 ± 1.28
Blood	36.34 ± 2.70	5.87 ± 0.63	4.04 ± 0.53	3.28 ± 0.27	0.55 ± 0.13	0.36 ± 0.09
Liver	8.40 ± 0.61	3.16 ± 0.81	2.66 ± 0.64	1.67 ± 0.27	0.64 ± 0.24	0.32 ± 0.06
Spleen	5.67 ± 0.80	2.54 ± 0.39	1.57 ± 0.20	1.30 ± 0.14	0.34 ± 0.11	0.19 ± 0.04
Kidney	18.58 ± 1.22	4.03 ± 0.57	2.99 ± 0.51	2.09 ± 0.25	0.47 ± 0.09	0.37 ± 0.04
Lung	9.90 ± 0.69	3.40 ± 0.48	2.50 ± 0.39	1.94 ± 0.17	0.46 ± 0.10	0.27 ± 0.07
Ratiosb
T:B	0.07	1.59	2.48	2.79	11.69	13.31
T:kidney	0.13	2.31	3.35	4.38	13.68	12.95
T:Liver	0.30	2.95	3.77	5.48	10.05	14.97
Tumor
Tumor weight	0.58 ± 0.28	0.7 ± 0.20	0.69 ± 0.09	0.63 ± 0.19	1.02 ± 0.54	1.06 ± 0.53
Groups of five mice were analyzed at each time point. Tumor and normal organ uptake are expressed as percent injected dose per gram (% ID/g). Table values are the means with corresponding standard errors of the means (s.e.m.). The ratios presented are the averages of the T:B, Tumor:kidney and Tumor:Liver ratios for the individual mice.
aTime given in hours
bRatios were determined for each individual mouse, and then averages were calculated

REFERENCES

• Adams, G. P., J. E. McCartney, M.-S. Tai, H. Oppermann, J. S. Huston, W. F. Stafford, M. A. Bookman, I. Fand, L. L. Houston and L. W. Weiner (1993). “Highly specific in vivo tumor targeting by monovalent and divalent forms of 741F8 anti-c-erbB-2 single-chain Fv.” Canc. Res. 53: 4026-4034.
• Begent, R. H. J., M. J. Verhaar, K. A. Chester, J. L. Casey, A. J. Green, M. P. Napier, L. D. Hope-Stone, N. Cushen, P. A. Keep, C. J. Johnson, R. E. Hawkins, A. J. W. Hilson and L. Robson (1996). “Clinical evidence of efficient tumor targeting based on single-chain Fv antibody selected from a combinatorial library.” Nature Med. 2: 979-984.
• Behr, T., W. Becker, E. Hannappel, D. M. Goldenberg and F. Wolf (1995). “Targeting of liver metastases of colorectal cancer with IgG, F(ab′)2, and Fab′ anti-carcinoembryonic antigen antibodies labeled with 99mTc: the role of metabolism and kinetics.” Cancer Research (Suppl.) 55: 5777s-5785s.
• Benhar, I., Brinkmann, U., Webber, K. O., and Pastan, I. “Mutations in the CDR loops of a recombinant immunotoxin that reduce its sensitivity to chemical derivatization.” Bioconjug Chem 5 (1994), 321-6.
• Carmichael, J. A., Power, B. E., Garrett, T. P., Yazaki, P. J., Shively, J. E., Raubischek, A. A., Wu, A. M., and Hudson, P. J. “The crystal structure of an anti-CEA scFv diabody assembled from T84.66 scFvs in V(L)-to-V(H) orientation: implications for diabody flexibility.” (2003) J Mol Biol 326, 341-51.
• Chatziioannou, A. F., S. R. Cherry, Y. Shao, R. W. Silverman, K. Meadors, T. H. Farquhar, M. Pedarsani and M. E. Phelps (1999). “Performance evaluation of microPET: A high-resolution lutetium oxyorthosilicate PET scanner for animal imaging.” J. Nucl. Med. 40: 1164-1175.
• D'Argenio, D. Z., and Schumitzky, A. “A program package for simulation and parameter estimation in pharmacokinetic systems.” (1979) Comput Programs Biomed 9, 115-34.
• Fitzgerald K, Holliger P, Winter G. (1997 October) “Improved tumour targeting by disulphide stabilized diabodies expressed in Pichia pastoris.” Protein Eng. 10(10):1221-5.
• Galfre, G., and Milstein, C. “Preparation of monoclonal antibodies: strategies and procedures.” (1981) Methods Enzymol 73, 3-46.
• Glockshuber, R., M. Malia, I. Pfitzinger and A. Pluckthun (1990). “A comparison of strategies to stabilize immunoglobulin Fv-fragments.” Biochemistry 29: 1362-1367.
• Haran, G., E. Haas, B. K. Szpikowska and M. T. Mas (1992). “Domain motions in phosphoglycerate kinase: Determination of interdomain distance distributions by site-specific labeling and time-resolved fluorescence energy transfer.” Proc. Natl. Acad. Sci. USA 89: 11764-11768.
• Holliger, P., T. Prospero and G. Winter (1993). “‘Diabodies’: Small bivalent and bispecific antibody fragments.” Proc. Natl. Acad. Sci. USA 90: 6444-6448.
• Hu, S. Z., L. Shively, A. Raubitschek, M. Sherman, L. E. Williams, J. Y. C. Wong, J. E. Shively and A. M. Wu (1996). “Minibody: A novel engineered anti-carcinoembryonic antigen antibody fragment (single-chain Fv-CH3) which exhibits rapid, high-level targeting of xenografts.” Cancer Res. 56: 3055-3061.
• Kipriyanov, S. M., S. Dubel, F. Breitling, R. E. Kontermann, S. Heymann and M. Little (1995). “Bacterial expression and refolding of single-chain Fv fragments with C-terminal cysteines.” Cell Biophys 26(3): 187-204.
• Kreitman, R. J., R. K. Puri, P. Leland, B. Lee and I. Pastan (1994). “Site-specific conjugation to interleukin 4 containing mutated cysteine residues produces interleukin 4-toxin conjugates with improved binding and activity.” Biochemistry 33(38): 11637-44.
• Leung, S., M. J. Losman, S. V. Govindan, G. L. Griffiths, D. M. Goldenberg and H. J. Hansen (1995). “Engineering a unique glycosylation site for site-specific conjugation of haptens to antibody fragments.” J. Immunol. 154: 5919-5926.
• Li, L., T. Olafsen, A. L. Anderson, A. Wu, A. A. Raubitschek and J. E. Shively (2002). “Reduction of kidney uptake in radiometal labeled peptide linkers conjugated to recombinant antibody fragments. Site-specific conjugation of DOTA-peptides to a Cys-diabody.” Bioconjug Chem 13(5): 985-95.
• Lyons, A., D. J. King, R. J. Owens, G. T. Yarranton, A. Millican, N. R. Whittle and J. R. Adair (1990). “Site-specific attachment to recombinant antibodies via introduced surface cysteine residues.” Protein Engineering 3: 703-708.
• McCarthy, D. W., R. E. Shefer, R. E. Klinkowstein, L. A. Bass, W. H. Margeneau, C. S. Cutler, C. J. Anderson and M. J. Welch (1997). “Efficient production of high specific activity 64Cu using a biomedical cyclotron.” Nucl. Med. Biol. 24: 35-43.
• Nikula, T. K., Bocchia, M., Curcio, M. J., Sgouros, G., Ma, Y., Finn, R. D., and Scheinberg, D. A. “Impact of high tyrosine fraction in complementarity determining regions: measured and predicted effects of radioiodination on IgG immunoreactivity.” (1995) Mol Immunol 32, 865-72.
• Olafsen, T., Bruland, O. S., Zalutsky, M. R., and Sandlie, I. “Cloning and sequencing of V genes from anti-osteosarcoma monoclonal antibodies TP-1 and Tp-3: location of lysine residues and implications for radiolabeling.” (1995) Nucl Med Biol 22, 765-71.
• Olafsen, T., Bruland, 0. S., Zalutsky, M. R., and Sandlie, I. “Abundant tyrosine resides in the antigen binding site in anti-osteosarcoma monoclonal antibodies TP-1 and TP-3: Application to radiolabeling.” (1996) Acta Oncol 35, 297-301.
• Qi, J., R. M. Leahy, S. R. Cherry, A. Chatziioannou and T. H. Farquhar (1998). “High-resolution 3D Bayesian image reconstruction using the microPET small-animal scanner.” Phys. in Med. and Biol. 43: 1001-1013.
• Qu, Z. X., R. M. Sharkey, H. J. Hansen, L. B. Shih, S. V. Govindan, J. Shen, D. M. Goldenberg and S. O. Leung (1998). “Carbohydrates engineered at antibody constant domains can be used for site-specific conjugation of drugs and chelates.” J. Immunol. Methods 213: 131-144.
• Rajagopalan, K., G. Pavlinkova, S. Levy, P. R. Pokkuluri, M. Schiffer, B. E. Haley and H. Kohler (1996). “Novel unconventional binding site in the variable region of immunoglobulins.” Proc. Natl. Acad. Sci. USA 93: 6019-6024.
• Rodwell, J. D., V. L. Alvarez, C. Lee, A. D. Lopes, J. W. F. Goers, H. D. King, H. J. Powsner and T. J. McKearn (1993). “Site-specific covalent modification of monoclonal antibodies: In vitro and in vivo evaluations.” Proc. Natl. Acad. Sci. USA 83: 2632-2636.
• Santimaria, M., Moscatelli, G., Viale, G. L., Giovannoni, L., Neri, G., Viti, F., Leprini, A., Borsi, L., Castellani, P., Zardi, L., Neri, D., and Riva, P., “Immunoscintigraphic Detection of the ED-B Domain of Fibronectin, a Marker of Angiogenesis, in Patients with Cancer.” (2003) Clin Cancer Res 9, 571-9.
• Stimmel, J. B., B. M. Merrill, L. F. Kuyper, C. P. Moxham, J. T. Hutchins, M. E. Fling and F. C. Kull, Jr. (2000). “Site-specific conjugation on serine right-arrow cysteine variant monoclonal antibodies.” J Biol Chem 275(39): 30445-50.
• Sundaresan, G., Yazaki, P. J., Shively, J. E., Finn, R. D., Larson, S. M., Raubitschek, A. A., Williams, L. E., Chatziioannou, A. F., Gambhir, S. S., and Wu, A. M. “Iodine-124 labeled engineered anti-CEA minibodies and diabodies allow high-contrast, antigen-specific microPET imaging of xenografts in athymic mice.” (In press) J Nucl Med.
• Verhaar, M. J., P. A. Keep, R. E. Hawkins, L. Robson, J. L. Casey, B. Pedley, J. A. Boden, R. H. J. Begent and K. A. Chester (1996). “Technetium-99m radiolabeling using a phage-derived single-chain Fv with a C-terminal cysteine.” J. Nucl. Med. 37: 868-872.
• Waibel, R., R. Alberto, J. Willuda, R. Finnern, R. Schibli, A. Stichelberger, A. Egli, U. Abram, J.-P. Mach, A. Pluckthun and P. A. Schubiger (1999). “Stable one-step technetium-99m labeling of His-tagged recombinant proteins with a novel Tc(I)-carbonyl complex.” Nat. Biotechnol. 17(897-901).
• Webber, K. O., Y. Reiter, U. Brinkmann, R. Kreitman and I. Pastan (1995). “Preparation and characterization of a disulfide-stabilized Fv fragment of the anti-Tac antibody: Comparison with its single-chain analog.” Mol. Immunol. 32: 249-258.
• Wu, A. M., Chen, W., Raubitschek, A., Williams, L. E., Neumaier, M., Fischer, R., Hu, S. Z., Odom-Maryon, T., Wong, J. Y., and Shively, J. E. “Tumor localization of anti-CEA single-chain Fvs: improved targeting by non-covalent dimers.” (1996) Immunotechnology 2, 21-36.
• Wu, A. M., L. E. Williams, L. Zieran, A. Padma, M. A. Sherman, G. G. Bebb, T. Odom-Maryon, J. Y. C. Wong, J. E. Shively and A. A. Raubitschek (1999). “Anti-carcinoembryonic antigen (CEA) diabody for rapid tumor targeting and imaging.” Tumor Targeting 4: 47-58.
• Wu, A. M., Yazaki, P. J., Tsai, S., Nguyen, K., Anderson, A. L., McCarthy, D. W., Welch, M. J., Shively, J. E., Williams, L. E., Raubitschek, A. A., Wong, J. Y., Toyokuni, T., Phelps, M. E., and Gambhir, S. S. “High-resolution microPET imaging of carcinoembryonic antigen-positive xenografts by using a copper-64-labeled engineered antibody fragment.” (2000) Proc Natl Acad Sci USA 97, 8495-500.
• Yazaki, P. J., L. Shively, C. Clark, C.-W. Cheung, W. Le, B. Szpikowska, J. E. Shively, A. A. Raubitschek and A. M. and Wu (2001a). “Mammalian expression and hollow fiber bioreactor production of recombinant anti-CEA diabody and minibody for clinical applications.” J. Immunol. Methods 253: 195-208.
• Yazaki, P. J., A. M. Wu, S. W. Tsai, L. E. Williams, D. N. Ikler, J. Y. Wong, J. E. Shively and A. A. Raubitschek (2001b). “Tumor targeting of radiometal labeled anti-CEA recombinant T84.66 diabody and t84.66 minibody: comparison to radioiodinated fragments.” Bioconjug Chem 12(2): 220-8.
• You, Y. H., L. J. Hefta, P. J. Yazaki, A. M. Wu and J. E. Shively (1998). “Expression, purification, and characterization of a two domain carcinoembryonic antigen minigene (N-A3) in Pichia Pastoris. The essential role of the N-domain.” Anticancer Res. 18: 3193-3202.

Claims

1-47. (canceled)

48. A diabody comprising:

a first single-chain polypeptide subunit that comprises: a first heavy chain variable domain polypeptide connected by a first linker sequence to a first light chain variable region polypeptide, and a first tail sequence that comprises a first amino acid spacer a first cysteine residue, wherein the first amino acid spacer comprises from 1 to about 10 amino acid residues, and wherein the first cysteine residue is at the C terminus of the first single-chain polypeptide subunit; and

a second single-chain polypeptide subunit that comprises: a second heavy chain variable domain polypeptide connected by a second linker sequence to a second light chain variable region polypeptide, and a second tail sequence that comprises a second amino acid spacer a second cysteine residue, wherein the second amino acid spacer comprises from 1 to about 10 amino acid residues, and wherein the second cysteine residue is at the C terminus of the second single-chain polypeptide subunit,

wherein the first cysteine residue forms a disulfide bond with the second cysteine residue.

49. The diabody of claim 48, wherein the first heavy chain variable domain polypeptide and the second light chain variable region polypeptide together form a first target binding site, and wherein the second heavy chain variable domain polypeptide and the first light chain variable region polypeptide together form a second target binding site.

50. The diabody of claim 49, wherein the first target binding site and the second target binding site bind the same target.

51. The diabody of claim 49, wherein the first target binding site and the second target binding site bind different targets.

52. The diabody of claim 48, wherein the first amino acid spacer comprises from about 5 to about 10 amino acid residues, and wherein the second amino acid spacer comprises from about 5 to about 10 amino acid residues.

53. The diabody of claim 48, wherein the first amino acid spacer comprises from 2 to 5 amino acid residues, and wherein the second amino acid spacer comprises from 2 to 5 amino acid residues.

54. The diabody of claim 48, wherein the first amino acid spacer comprises 6, 7, 8, or 9 amino acid residues, and wherein the second amino acid spacer comprises 6, 7, 8, or 9 amino acid residues.

55. The diabody of claim 48, wherein the first amino acid spacer comprises 2 amino acid residues, and wherein the second amino acid spacer comprises 2 amino acid residues.

56. The diabody of claim 48, wherein the amino acid residues of the first amino acid spacer are glycine residues.

57. The diabody of claim 56, wherein the amino acid residues of the second amino acid spacer are glycine residues.

58. The diabody of claim 48, wherein the first linker sequence comprises from about 5 to about 10 amino acid residues, and wherein the second linker sequence comprises from about 5 to about 10 amino acid residues.

59. The diabody of claim 58, wherein the first linker sequence comprises 8 amino acid residues, and wherein the second linker sequence comprises 8 amino acid residues.

60. The diabody of claim 58, wherein the residues of the first linker sequence are glycine residues and wherein the residues of the second linker sequence are glycine residues.

61. The diabody of claim 49, the first target binding site and/or the second target binding site specifically binds a carcinoembryonic antigen (CEA).

62. The diabody of claim 61, wherein the first heavy chain variable domain polypeptide corresponds to a heavy chain variable domain polypeptide sequence of the murine anti-CEA T84.66 antibody, and wherein the second light variable domain polypeptide corresponds to a light chain variable domain polypeptide sequence of the murine anti-CEA T84.66 antibody.

63. The diabody of claim 62, wherein the second heavy chain variable domain polypeptide corresponds to a heavy chain variable domain polypeptide sequence of the murine anti-CEA T84.66 antibody, and wherein the first light variable domain polypeptide corresponds to a light chain variable domain polypeptide sequence of the murine anti-CEA T84.66 antibody.