In Vitro and In Vivo Characterization of a Novel Antibody-Like Single-Chain TCR Human IgG1 Fusion Protein
http://www.100md.com
免疫学杂志 2005年第7期
Abstract
We have constructed a protein composed of a soluble single-chain TCR genetically linked to the constant domain of an IgG1 H chain. The Ag recognition portion of the protein binds to an unmutated peptide derived from human p53 (aa 264–272) presented in the context of HLA-A2.1, whereas the IgG1 H chain provides effector functions. The protein is capable of forming dimers, specifically staining tumor cells and promoting target and effector cell conjugation. The protein also has potent antitumor effects in an in vivo tumor model and can mediate cell killing by Ab-dependent cellular cytotoxicity. Therefore, single-chain TCRs linked to IgG1 H chains behave like Abs but possess the ability to recognize Ags derived from intracellular targets. These fusion proteins represent a novel group of immunotherapeutics that have the potential to expand the range of tumors available for targeted therapies beyond those currently addressed by the conventional Ab-based approach.
Introduction
Targeted therapy for cancer has shown promise as a safe, effective alternative to highly toxic chemotherapeutic approaches. The clinical and commercial success of humanized mAbs Herceptin and Rituximab, which are used to treat metastatic breast cancer and B cell lymphoma, respectively, has prompted the development of a wide variety of Ab-based therapies for both solid and blood-borne tumors (1, 2, 3). The mechanisms of action of therapeutic Abs include direct antiproliferative and proapoptotic effects on their targets (4, 5). In addition, through their Fc domains, Abs may activate Ab-dependent cellular cytotoxicity (ADCC)3 (6, 7, 8) and phagocytosis (9), which may aid in tumor elimination.
Despite the promise of Ab-based immunotherapy, there are a number of potential problems with Ab-based tumor targeting. Abs only recognize Ags expressed on the cell surface; however, there are other tumor Ags that are expressed as intracellular proteins. Moreover, because many tumor-specific Ags are derived from an aberrant expression of cell type-specific proteins, many tumor Ags are expressed on only a small number of tumor types, potentially limiting the number of tumors that could be treated by any one-targeted therapy. Furthermore, tumor Ag expression is often heterogeneous, where the Ag may be expressed on the primary tumor but not on distant metastases of that tumor. Finally, tumor Ags shed from the surface of tumor cells can occupy the binding sites of antitumor-specific Abs, reducing the number of active molecules and resulting in decreased tumor cell death (10). Thus, it would be useful to be able to target a broader range of tumor Ags that may be derived from either intracellular or extracellular proteins that are not subject to the problems of Ag shedding.
The tumor suppressor protein, p53, is mutated and/or overexpressed in a wide variety of human malignancies (11, 12, 13), and its overexpression can correlate with poor prognosis (14, 15). Thus, p53 may be targeted as a general tumor Ag, but because it is an intracellular protein, it cannot be targeted by Ab-based therapies. Instead, p53 peptides presented in the context of HLA have been used as target Ags for T cell-based cancer therapy approaches, including tumor vaccination and ex vivo T cell stimulation and expansion (16, 17). Although these approaches have shown some promise in animal models and some early-stage clinical trials, both of these methods are time consuming, laborious, expensive, and lack standardization, making them unsuitable for treating large patient populations (16, 17). Thus, a therapeutic molecule that can recognize and bind to Ags derived from intracellular proteins such as p53 would significantly broaden the range of tumor Ags that can be targeted.
We have described the construction, expression, and characterization of a soluble three-domain mouse single-chain TCR (scTCR), which recognizes a wild-type human p53 peptide (aa 264–272) presented in the context of HLA-A2.1 (264scTCR) (18). As an IL-2 fusion protein, the 264scTCR retains its MHC-restricted, peptide-specific, Ag-binding properties, whereas the IL-2 portion binds to IL-2R and retains biological activity. Moreover, these studies demonstrate that this fusion protein mediates conjugation of target and effector cells, exhibits favorable pharmacokinetics in mice, binds to target tumor cells, and has antitumor effects (18).
In the studies reported here, we constructed a scTCR fusion protein that resembles an Ab. The Ag recognition portion of the fusion protein is composed of a 264scTCR, and a human IgG1 H chain provides the effector function domain. Fusion of the 264scTCR to the IgG1 H chain allows for dimer formation via the natural disulfide bonding in IgG1 molecules. We evaluate the ability of our fusion protein to bind specifically to cells in vitro as well as reduce metastasis in an in vivo experimental metastasis model. We also explore the possible mechanisms by which this protein might function. Our data indicate that this molecule exhibits many major characteristics of a mAb and that it is possible to create scTCR-based, Ab-like molecules that recognize Ags derived from intracellular proteins such as p53. These novel targeted therapies could prove to be potent immunotherapeutics.
Materials and Methods
Materials
A2.1 264 CTL clone no. 5 was derived by limiting dilution cloning (19) from a CTL line specific for the human p53 (aa 264–272) peptide generated in HLA-A2.1-transgenic mice (11). Expression vectors for producing peptide-loaded HLA-A2 tetramers were a generous gift from Dr. John Altman at Emory University (Atlanta, GA). CHO.K1 Chinese hamster ovary, T2 human lymphoblast, U937 human promyeloma cells, H57-597, BB7.2, and (BF1) 8A3.31 hybridoma cell lines were purchased from American Type Culture Collection. T2 human lymphoblast cells are positive for HLA-A2.1 but deficient in TAP 1 and 2 proteins, which allow them to display empty MHC molecules on their surface that can then be loaded with exogenous peptide (20). The H57-597 hybridoma produces a mAb that recognizes an epitope in the murine TCR- constant region, and the BB7.2 hybridoma produces a mAb that specifically recognizes an epitope on the 2 domain of HLA-A2. The (BF1) 8A3.31 hybridoma produces a mAb that recognizes an epitope in the human TCR- constant region. A highly metastatic subclone of the human melanoma cell line A375, A375-C15N, was used for in vivo metastasis studies and was maintained by Dr. John Francis at the Florida Hospital Cancer Research Institute (Orlando, FL) (21). Anti-TCR C mAb H57-597, anti-murine TCR V3 mAb, and FITC-labeled goat anti-mouse IgG were obtained from BD Pharmingen. Anti-IgG F(ab')2 and HRP-conjugated anti-IgG were purchased from Pierce and The Binding Site, respectively. All cell culture media and additives were purchased from Mediatech, and all cell culture materials were purchased from Nunc unless otherwise noted. All mice were purchased from Harlan. Biotinylated peptide/HLA-A2 reagents were generated as described previously (22).
Cell culture
All cell lines were maintained in complete culture medium comprising IMDM supplemented with 10% FBS plus 2 mM L-glutamine and grown at 37°C with 5% CO2. A375-C15N cells were maintained in RPMI 1640 with 10% heat-inactivated FBS, penicillin, and streptomycin (Invitrogen Life Technologies). For induction of FcR, U937 cells were cultured for 48 h in IMDM supplemented with 10% FBS, 2 mM L-glutamine, and 62.5 ng/ml human IFN- (R&D Systems).
Constructs
The TCR gene was cloned from the T cell clone A2.1 264 no. 5 as described previously (18). We designated the scTCR derived from this T cell clone 264scTCR. To generate the 264scTCR/IgG1 expression construct, the three-domain single-chain 264scTCR was amplified using the 264scTCR/IL-2 fusion protein construct as a template (18). The scTCR fragment was ligated into an Ab H chain expression vector, replacing the Ab variable region to yield a scTCR fused to a human IgG1 H chain region. A control fusion construct, 264scTCR/trIgG1, was created by truncating the IgG1 heavy domain before the hinge region, which prevented disulfide bonding. The CMV scTCR (CMVscTCR) was cloned from CTLs stimulated with HLA-A2-restricted CMV-pp65 peptides. The IgG1 fragment was amplified from 264scTCR/IgG1 DNA to create the CMVscTCR/IgG1 construct. For production of the fusion proteins in mammalian cells, CHO.K1 cells were electroporated using a Bio-Rad Gene Pulser, followed by limiting dilution cloning and selection in medium containing 1 mg/ml G418.
Protein purification
264scTCR/IgG1 and 264scTCR/trIgG1 were purified from cell culture supernatant fluid by immunoaffinity chromatography using the H57-597 mAb coupled to a Sepharose 4B column (Amersham Biosciences). CMVscTCR/IgG1 was purified from cell culture supernatant fluid by immunoaffinity chromatography using the (BF1) 8A3.31 mAb coupled to a Sepharose 4B column (Amersham Biosciences). The purified samples were concentrated and buffer-exchanged into PBS using an Ultrafree-15 centrifugal filter with a 30-kDa molecular mass cutoff membrane (Millipore). The TCR fusion protein samples were stored at 2°C to 8°C (short term) or at –80°C (long term) for biochemical and functional analyses.
SDS-PAGE
SDS-PAGE was performed under reducing or nonreducing conditions using 4–20% polyacrylamide gels (NOVEX) and the NOVEX EX-Cell II system. SDS-PAGE gels were stained with Coomassie blue.
ELISA
For determination of 264scTCR/IgG1 concentration from transfected CHO.K1 cells, Maxisorp 96-well plates (Nunc) were coated with anti-human IgG F(ab')2 (Pierce). Purified 264scTCR/IgG1 fusion protein was added and incubated for 1 h at room temperature. After washing, bound fusion protein was detected with HRP-conjugated sheep anti-human IgG1 (The Binding Site) and ABTS substrate (BioFX) and then quenched with 1% SDS (BioFX). Concentration of 264scTCR/IgG1 was determined by comparison to a known Ab standard curve. For demonstration of peptide/MHC binding, streptavidin plates were coated with p53 (aa 264–272)-loaded HLA-A2 monomers, and purified 264scTCR/IgG1 fusion protein was added and incubated for 1 h at room temperature. After washing, bound fusion protein was detected with HRP-conjugated sheep anti-human IgG1 (The Binding Site) and ABTS substrate (BioFX) and then quenched with 1% SDS (BioFX). For all ELISAs, absorbance was measured at 405 nm using a 96-well plate reader (Bio-Tek Instruments).
Cell staining with 264scTCR/IgG1
T2 cells pulsed with either p53 (aa 149–157) or p53 (aa 264–272) peptide were incubated with 8.0 pM 264scTCR/IgG1 or 264scTCR/trIgG1 fusion protein in 1% FBS in PBS for 30 min at room temperature. Bound fusion protein was detected with PE-conjugated anti-TCR C mAb. Samples were washed with 1% FBS in PBS before analysis on a FACScan flow cytometry instrument (BD Biosciences).
U937 cells were cultured for 48 h with 62.5 ng/ml human IFN- to increase expression of FcR on the cell surface. Cells (2 x 105/sample) were washed and incubated with or without 4 μg of 264scTCR/IgG1 fusion protein for 30 min at room temperature. After washing, cells were incubated with 4 μg of PE-conjugated, HLA-A2/p53 (aa 149–157)- or (aa 264–272)-loaded tetramers for 30 min at room temperature. After washing, the samples were analyzed by flow cytometry.
A375 melanoma cells were stained with either 264scTCR/IgG1 fusion protein or control fusion protein CMVscTCR/IgG1. A375 melanoma cells were cultured on coverslips for 24 h. The cells were fixed with 3.7% formaldehyde for 5 min and washed twice with washing buffer (0.5% BSA and 0.1% sodium azide in PBS). The cells were stained with 10 μg of 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein in 200 μl of PBS containing 5% normal goat serum for 45 min at 23°C. The cells were washed twice and incubated with 6 μg of FITC-conjugated F(ab')2 goat anti-human IgG Fc (Jackson ImmunoResearch Laboratories). The cells were washed three times with equilibration buffer (Molecular Probes). Coverslips were mounted on glass slides with anti-fade reagent in glycerol buffer (Molecular Probes) and sealed with nail polish. The slides were documented using a Nikon epifluorescence microscope (Nikon) with a SPOT RT camera and SPOT RT software v3.2 (Diagnostic Instruments).
For flow cytometric analysis, A375 cells were detached with 10 mM EDTA in PBS (pH 7.4) and washed twice with washing buffer. The cells were fixed with 3.7% formaldehyde for 5 min and washed twice. Cell staining was conducted using 4 μg of 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein in the presence or absence of 20 μg of HLA-A2.1/p53 (aa 264–272) tetramers for 45 min at 23°C. The cells were washed once and incubated with 6 μg of FITC-conjugated F(ab')2 goat anti-human IgG Fc as above. After washing twice, the cells were resuspended and analyzed on a FACScan flow cytometry instrument (BD Biosciences) with CellQuest software (BD Biosciences).
Affinity determinations
The affinity of the dimeric 264scTCR/IgG1 fusion protein for HLA-A2.1/p53 (aa 264–272) was compared with the affinity of the monomeric 264scTCR/trIgG1 fusion protein by surface plasmon resonance (Biacore). Approximately 1300 resonance units of biotinylated p53 (aa 264–272)/HLA-A2.1 complex were immobilized on a streptavidin sensor chip. Each surface was used one time only. The average affinity constants were calculated based on data from three different scTCR fusion protein concentrations (5, 10, and 20 μM) using the Biacore computer software Biaevaluation 3.1 (Biacore).
Cell conjugation
T2 cells pulsed with either p53 (aa 264–272) peptide or (aa 149–157) peptides were labeled with 12.6 μg/ml dihydroethidium (HE) (Molecular Probes), and U937 cells were labeled with 80 ng/ml calcein AM (Molecular Probes). After washing, the two populations of labeled cells were mixed together at a 1:1 ratio in the presence or absence of 2.5 μg of 264scTCR/IgG1 fusion protein for 20 min at room temperature. Cells were then analyzed on a FACScan flow cytometry instrument.
Pharmacokinetics in mice
For all experiments involving animals, principles of laboratory animal care were followed, as well as specific national laws where applicable. Female athymic nude mice were injected i.v. via the lateral tail vein with 250 μg of 264scTCR/IgG1 fusion protein diluted with PBS to a total volume of 100 μl. Serum was collected from one group of mice not injected with 264scTCR/IgG1 to establish background levels. Serum was collected by tail bleed from the injected groups at 15 and 30 min and 1, 2, 4, 8, and 24 h. Blood samples were centrifuged at 14,000 x g at 4°C for 10 min, and serum was collected and stored at –80°C until use. 264scTCR/IgG1 concentrations were determined by ELISA using anti-TCR V3 mAb for capture and HRP-conjugated sheep anti-human IgG1 Ab for detection.
Experimental metastasis assay
Female athymic nude mice were injected with 5.0 x 105 A375-C15N cells via the lateral tail vein on study day 0. The tumors were allowed to establish for 3 days and then the animals were injected i.v. with varying doses of either 264scTCR/IgG1 (0.1–30 mg/kg in 100 μl of total volume), PBS, or Stx-2 (an irrelevant Ab isotype-matched control at 30 mg/kg in 100 μl of total volume) on study days 3, 7, 10, 14, 17, 21, 28, and 35 posttumor cell injection. Forty-two days after tumor cell injection, all animals were sacrificed humanely, the lungs were removed and fixed in Bouin’s solution, and surface pulmonary tumor nodules were counted. Two observers counted tumor nodules on each lung, and the average counts were recorded. The Stx-2 Ab, a chimeric IgG1 Ab (mouse variable/human H and L chain constant region), was created in our lab and has been described previously (23). This Ab recognizes an epitope on Escherichia coli 0157 shigatoxin protein and was used for these studies as an irrelevant isotype control.
ADCC
T2 cells were loaded with p53 (aa 264–272) peptide and used as target cells. Target cells were labeled with 50 μg/ml calcein AM (Molecular Probes) for 30 min at 37°C and washed twice. Effector PBMCs were isolated from buffy coats (Community Blood Centers of South Florida) using the one-step gradient method with Histopaque-1077 (Sigma-Aldrich). To measure ADCC activity, 2 x 106 effector cells were incubated with 2 x 104 target cells in 0.25 ml of complete medium with either 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein at 20 μg/ml. The fluorescent intensity of culture supernatant fluid was measured after 4 h with a fluorescent plate reader (excitation = 485 nM; emission = 538 nM; cutoff = 530 nM). ADCC activity was calculated using the following formula: percentage of cytotoxicity = (fluorescent intensity of test sample – fluorescent intensity of target cells with medium)/(fluorescent intensity of target cells treated with 0.04% Triton X-100 – fluorescent intensity of target cells with medium) x 100. Data are reported as the mean of triplicate determinations ± SEM. Statistical significance was determined by paired t test analysis.
Results
Generation of TCR fusion protein constructs
We have constructed a fusion protein composed of a three-domain scTCR that recognizes an unmutated peptide from human p53 (aa 264–272) (18) fused to the constant domain of the human IgG1 H chain. The scTCR portion of this fusion protein comprises the TCR V region linked to the V/C regions via a flexible linker (G4S)4 (24). The C domain, which is directly linked to the V domain, is truncated at the amino acid residue just before the final cysteine, removing the transmembrane and cytoplasmic domains, to generate a soluble scTCR molecule. This scTCR is linked directly to the constant domain of the human IgG1 H chain (Fig. 1). The cysteine residues in the hinge region of the H chain are intact, which allows for appropriate disulfide bonding of the H chain regions and thus covalently linked dimer formation of the 264scTCR/IgG1 fusion protein.
FIGURE 1. The 264scTCR/IgG1 fusion protein. Schematic representation of the domain structure of the 264scTCR/IgG1 fusion protein.
Expression of TCR/IgG1 fusion protein in mammalian cells
To characterize the soluble 264scTCR/IgG1 fusion protein, CHO-K1 cells were stably transfected with the 264scTCR/IgG1 expression vector. To determine the concentration of secreted protein, the fusion protein was captured with an anti-IgG F(ab')2 and detected with an anti-IgG1 HRP conjugate. Concentrations of the 264scTCR/IgG1 fusion protein were estimated based on comparison to an IgG1 Ab standard. To determine whether secreted protein can bind peptide-loaded HLA, the fusion protein was captured with biotinylated HLA-A2/p53 (aa 264–272) monomers on streptavidin-coated plates and detected with anti-IgG1 HRP conjugate. Positive signals for ELISAs in either format indicate that the transfected cells secrete 264scTCR/IgG1 fusion protein that is assembled properly, folded, and secreted from the transfected cells. Moreover, the TCR portion of 264scTCR/IgG1 retains peptide-specific MHC binding ability in this fusion protein format.
The 264scTCR/IgG1 fusion protein was purified from transfected cell supernatants by immunoaffinity chromatography using an anti-murine TCR C mAb, H57-597. Purified protein was subjected to SDS-PAGE and Coomassie staining (Fig. 2). Under reducing conditions, the major protein band was 75–80 kDa, which is consistent with the calculated molecular mass for this protein based on its deduced amino acid sequence. Under nonreducing conditions, the molecular mass of the predominant protein band was 170–180 kDa. The apparent shift in molecular mass of the 264scTCR/IgG1 under reducing vs nonreducing conditions is to be expected because there are cysteine residues in the H chain regions that form disulfide bonds. Thus, as with an IgG1 Ab, the 264scTCR/IgG1 fusion protein forms dimers.
FIGURE 2. Production of 264scTCR/IgG1 fusion protein in transfected Chinese hamster ovary cells. The secreted fusion protein was purified by immunoaffinity chromatography and subjected to SDS-PAGE under either reducing (R) or nonreducing (N) conditions. SDS-PAGE gels was stained with Coomassie brilliant blue.
MHC/peptide binding of the TCR portion of 264scTCR/IgG1
The ability of the 264scTCR/IgG1 fusion protein to bind to peptide-loaded MHC on the surface of APCs was determined by flow cytometry. T2 lymphoblast cells were loaded with p53 (aa 264–272) or p53 (aa 149–157) control peptide and incubated with purified 264scTCR/IgG1 fusion protein. After several wash steps, the cells were incubated with PE-labeled anti-TCR mAbs to detect cell-bound fusion protein (Fig. 3). T2 cells loaded with p53 (aa 264–272) peptide stained positively with 264scTCR/IgG1, whereas cells loaded with p53 (aa 149–157) showed little or no staining, indicating that the 264scTCR/IgG1 could stain cells in a peptide-specific manner. We also attempted to stain peptide-loaded T2 cells with monomeric 264scTCR/trIgG1 fusion protein and found that no staining was observed on cells loaded with either p53 (aa 264–272) or p53 (aa 149–157) peptide. These results demonstrate that the bivalent nature of 264scTCR/IgG1 greatly increased its ability to stably bind MHC-peptide complexes on the surface of cells.
FIGURE 3. MHC/peptide-binding ability of the TCR portion of 264scTCR/IgG1 fusion protein. T2 cells were loaded with p53 (aa 264–272) peptide and stained with either 264scTCR/IgG1 (black thin line) or 264scTCR/trIgG1 fusion protein (gray thick line), followed by anti-TCR C mAb. T2 cells loaded with p53 (aa 149–157) control peptide were also stained with either 264scTCR/IgG1 (dotted black line) or 264scTCR/trIgG1 fusion protein (black thick line), followed by anti-TCR C mAb. All samples were analyzed by flow cytometry.
Binding affinity of 264scTCR/IgG1
To determine whether the intensity of staining observed with 264scTCR/IgG1 could be attributed to a high binding affinity, we performed surface plasmon resonance assays. The binding affinity (KD) of the dimeric 264scTCR/IgG1 with HLA-A2.1/p53 (aa 264–272) complexes was estimated to be 5.6 x 10–7 M, which is significantly higher than the monomeric 264scTCR/trIgG1 (KD of 2.2 x 10–6 M; Table I). The increased affinity of the 264scTCR/IgG1 fusion protein likely contributes to its ability to bind strongly to the correct peptide/MHC pair.
Table I. Affinity determination with 264scTCR/IgG1 fusion proteina
Binding of the Fc domain of 264scTCR/IgG1 to FcR
To demonstrate the ability of the IgG1 Fc portion of the 264scTCR/IgG1 fusion protein to bind to FcR, U937 human promyeloma cells were stimulated with human IFN- to increase expression of the FcR on the cell surface. Stimulated as well as unstimulated cells were incubated with 264scTCR/IgG1 fusion protein, and bound fusion protein was detected with PE-conjugated HLA-A2.1/p53 (aa 264–272) or HLA-A2.1/p53 (aa 149–157) tetramers and analyzed by flow cytometry. When the 264scTCR/IgG1 fusion protein was used, cells stained positively following incubation with HLA-A2.1/p53 (aa 264–272) tetramers but not with HLA-A2.1/p53 (aa 149–157) tetramers. Moreover, positive staining with 264scTCR/IgG1 fusion protein was significantly more intense when the cells were stimulated with human IFN-, which is known to increase the expression of FcR on U937 cells (Fig. 4). Thus, these data indicate that the IgG1 portion of the 264scTCR/IgG1 fusion protein is capable of binding to FcRs.
FIGURE 4. Binding of the Fc domain of 264scTCR/IgG1 to FcR. rIFN--stimulated U937 cells were stained with 264scTCR/IgG1 fusion protein followed by HLA-A2/p53 (aa 264–272) (black thin line) or HLA-A2/p53 (aa 149–157) (black thick line) tetramers. Unstimulated U937 cells were also stained with fusion protein and HLA-A2/p53 (aa 264–272) (black dotted line) or HLA-A2/p53 (aa 149–157) (gray thick line) tetramers. All samples were analyzed by flow cytometry.
Conjugation of cells mediated by 264scTCR/IgG1 fusion protein
In order for the 264scTCR/IgG1 fusion protein to be useful as a therapeutic, it must be able to bring together target and effector cells through the TCR and Fc components, respectively. To demonstrate that the 264scTCR/IgG1 fusion protein can effectively conjugate cells, T2 cells were loaded with either p53 (aa 264–272) or p53 (aa 149–157) control peptide and then stained with HE. U937 effector cells were differentially stained with calcein AM. The two stained cell populations were mixed, incubated in the presence or absence of 264scTCR/IgG1, and analyzed by flow cytometry. When the two cell populations were incubated in the absence of the 264scTCR/IgG1 fusion protein (Fig. 5, A and B) or when the T2 cells were loaded with control peptide p53 (aa 149–157) and incubated with the U937 cells in the presence of 264scTCR/IgG1 fusion protein (Fig. 5C), the cells remained as two distinct populations on the flow cytometry histograms (Fig. 5, A–C, regions 1 and 3). However, when the T2 cells were loaded with p53 (aa 264–272) peptide and incubated with the U937 cells in the presence of the 264scTCR/IgG1 fusion protein (Fig. 5D), a double-stained population of cells appears (Fig. 5D, region 2, conjugated cells), indicating that T2 cells were conjugated to U937 cells via the 264scTCR/IgG1 fusion protein. These data demonstrate that the 264scTCR/IgG1 fusion proteins can mediate conjugation of effector cells to target cells.
FIGURE 5. Conjugation of cells mediated by 264scTCR/IgG1. T2 cells were loaded with either p53 (aa 264–272) (B and D) or p53 (aa 149–157) (A and C) peptides and then labeled with HE. U937 cells were labeled with calcein AM. Labeled cells were mixed and incubated in the presence (C and D) or absence (A and B) of 264scTCR/IgG1 fusion protein. Single-stained regions are marked 1 and 3, and the double-stained cell population is marked 2.
Tumor cell staining
To demonstrate the ability of 264TCR/IgG1 fusion protein to bind to tumor cells, the A375 human melanoma cell line, shown in our laboratory to be HLA-A2.1 and p53 positive (data not shown), was used. The 264TCR/IgG1 fusion protein positively stained the A375 cells (Fig. 6A), while staining with CMVscTCR/IgG1 was not detectable above background (Fig. 6B). Staining of A375 cells with 264scTCR/IgG1 was reduced by the addition of HLA-A2.1/p53 (aa 264–272) tetramers (Fig. 6C), indicating that 264scTCR/IgG1 specifically binds to tumor cells via the scTCR portion of the fusion protein.
FIGURE 6. Tumor cell staining with 264scTCR/IgG1 fusion protein. Formaldehyde-fixed A375 cells were stained with 264scTCR/IgG1 fusion protein (A) or CMVscTCR/IgG1 fusion protein (B). The images were captured at x400. Fixed A375 cells were stained with 264scTCR/IgG1 (black lines) or CMVscTCR/IgG1 (gray lines) fusion protein in the presence (thin or dotted lines) or absence (thick lines) of HLA-A2.1/p53 (aa 264–272) tetramers (C).
Serum half-life of 264scTCR/IgG1 in mice
The pharmacokinetics of the 264scTCR/IgG1 fusion protein was measured in nude mice. All mice were injected with a single i.v. dose of 264scTCR/IgG1 and showed no obvious signs of toxicity. Serum samples were collected at various time points, and the 264scTCR/IgG1 fusion protein concentration was determined using ELISA. In these assays, a maximum concentration of 8.8 μg/ml 264scTCR/IgG1 was detected with an apparent serum half-life of 4 h (Fig. 7). This is consistent with the apparent serum half-life of the 264scTCR/IL-2 fusion protein (18).
FIGURE 7. Serum half-life of 264scTCR/IgG1 fusion protein. Nude mice were injected with 264scTCR/IgG1 fusion protein, and serum samples were collected at 15 and 30 min and 1, 4, 8, and 24 h postinjection. Serum concentrations of 264scTCR/IgG1 were determined by ELISA.
Effect of 264scTCR/IgG1 on tumor growth in nude mice
To determine whether the 264scTCR/IgG1 fusion protein exhibits antitumor activity in vivo, an experimental metastasis assay was performed. Female athymic nude mice were injected with the highly metastatic A375 human melanoma subclone, A375-C15N, and tumors were allowed to establish for 3 days. Animals were then treated with varying doses of 264scTCR/IgG1, -Stx-2 (an irrelevant IgG1 isotype control Ab), or vehicle. Forty-two days after tumor cell injection, lung nodules were counted. 264scTCR/IgG1 reduced lung metastasis in a dose-dependent manner reaching maximum reduction of metastasis at the 1 mg/kg dose, whereas the isotype control failed to reduce metastasis (Fig. 8). These data indicate that the 264scTCR/IgG1 fusion protein exhibits potent anti-metastatic activities in vivo and has potential use as a cancer therapeutic in humans.
FIGURE 8. Antitumor effects of 264scTCR/IgG1. Female athymic nude mice were injected with highly metastatic A375-C15N cells and treated with 264scTCR/IgG1 at doses ranging from 0.1 to 30 or 30 mg/kg anti-Stx-2 (an irrelevant IgG1 isotype control Ab) or a dose volume equivalent of PBS. Forty-two days after tumor cell injection, the lungs were removed, lung nodules were counted, and the mean number of lung nodules was plotted.
ADCC
One possible mechanism of action for therapeutic Abs is ADCC, which is mediated by the FcR-binding portion of the IgG molecule. Because the 264scTCR/IgG1 fusion protein contains a FcR-binding domain, we investigated whether this fusion protein could mediate ADCC. T2 cells loaded with p53 (aa 264–272) peptide and labeled with calcein AM were used as target cells, and human PBMCs were used as effector cells. Target and effector cells were mixed together at a 1:100 ratio in the presence of 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein. Cell killing with 264scTCR/IgG1 ranged from 48 to 89%, while cell killing with CMVscTCR/IgG1 protein was significantly lower (p = 0.02) (Fig. 9). These data indicate that the 264scTCR/IgG1 protein is capable of mediating ADCC and that this mechanism may play a role in the antitumor activity of this fusion protein. Given that the population of FcR-positive effector cells in PBMCs used for the ADCC experiment is 15–20% of the population, the high level of killing seen in these assays suggests that there may be one or more other mechanisms of action used for cell killing mediated by 264scTCR/IgG1.
FIGURE 9. 264scTCR/IgG1 mediation of ADCC. T2 cells loaded with p53 (aa 264–272) peptide were labeled with calcein AM and incubated with human PBMCs (100:1 E:T ratio) in the presence of 264scTCR/IgG1 () or CMVscTCR/IgG1 ( ) fusion protein. After 4 h, culture supernatants were collected, and cell lysis was quantified.
Discussion
In these studies, we have described the construction and characterization of a novel 264scTCR/IgG1 fusion protein, which recognizes an unmutated peptide from human p53 (aa 264–272) presented in the context of HLA-A2.1. The IgG1 H chain domain confers FcR binding and allows for dimer formation, and the 264scTCR portion of the molecule confers peptide-specific MHC binding. This fusion protein is capable of specifically staining tumor cells and promotes target and effector cell conjugation. The 264scTCR/IgG1 fusion protein also has potent antitumor effects in an in vivo tumor model and can mediate cell killing by an ADCC mechanism of action. Thus, we have demonstrated that it is feasible to create scTCR fusion molecules, which are functionally equivalent to therapeutic mAbs but with the ability to recognize target Ags derived from intracellular proteins.
The development of TCR-based therapeutic agents has been precluded by the difficulty of producing stable, functionally active TCR molecules in sufficiently large quantities (25, 26, 27, 28, 29, 30, 31, 32). We have overcome this obstacle by constructing soluble, single-chain, three-domain TCR molecules comprising the TCR- variable domain linked to the TCR- variable and constant regions (18, 33). These constructs are active with respect to specific peptide/MHC binding, easily expressed in large quantities in mammalian cells, and readily purified for use in in vitro and in vivo assays (18). Expression and secretion of scTCR-based proteins in mammalian cells offers advantages over bacterial expression systems that require time-consuming, laborious refolding methods to produce TCR-/ heterodimers (28, 30, 32). In addition, solubly expressed scTCRs can be genetically fused with molecules exhibiting effector functions such as cytokines to create bifunctional-bispecific fusion proteins. For example, scTCR molecules linked to human IL-2 (18), GM-CSF, and IFN-2b (K. Card, unpublished observations) have been produced and shown to retain their respective bispecific functions. In another approach, Abs with specificity to peptide/MHC complexes have been generated and have been used to quantitate peptide/MHC complexes on the surface of APCs and deliver cytotoxic effector toxin to APCs in a peptide-dependent manner (34, 35).
TCRs are known to have low affinity for cognate peptide/MHC complexes due to negative selection in the thymus during the T cell maturation process (36). In addition, the density of a particular epitope displayed in the context of MHC on an APC surface is predictably low (34). These two factors raised significant doubts about whether it was feasible to use TCRs as targeting molecules. The tumor cell staining results in this study demonstrate that a dimeric form of the scTCR does in fact have sufficient affinity to function as an Ag recognition molecule in vitro. The antitumor effects of the 264scTCR/IgG1 fusion protein in our experimental metastasis model additionally demonstrate that the Ag-targeting capability of this type of scTCR molecule also functions in vivo. We are currently performing experiments to determine whether the density of the peptide/MHC complexes on the tumor cell surface directly influences the ability of scTCR fusion proteins to target tumor cells.
Affinity plays an important role in tumor retention and penetration by macromolecules such as Abs. It has been shown that increasing the affinity of Abs increases tumor retention, but at a KD > 10–9 M, tumor retention is not further improved, and penetration is hindered significantly (37, 38). The 264scTCR was derived from a T cell line generated from a HLA-A2.1-transgenic mouse. TCRs derived from these mice are CD8 independent and are postulated to have high affinity for their cognate peptide/MHC complex (11). Our measurements indicate that the affinity of the dimeric 264scTCR/IgG1 fusion molecule for peptide/MHC is on the order of 5.6 x 10–7 M KD, which is exceptionally high for a TCR. Abs with this level of affinity show low levels of tumor retention but very efficient tumor penetration (37, 38). Whether the unusually high affinity of this scTCR for its cognate peptide/MHC is the key to successful tumor cell staining and antitumor activity is unknown but is receiving additional investigation in our laboratory. The affinity of scTCR molecules can also be increased substantially using a variety of in vitro enhancement methods in combination with mutagenesis (33, 39). For example, we have developed a phage display method wherein scTCRs are expressed as bacteriophage M13 gene VIII fusions (33). In vitro multimerization methods, such as birA-streptavidin conjugation or cross-linking with liposomes or dendrimers, may also increase the avidity of scTCR molecules. Such multimerization should increase the avidity of a scTCR-based therapeutic agent for its target without sacrificing Ag recognition specificity. Dimerization using the IgG H chain illustrates the feasibility of such an approach because the dimeric 264scTCR/IgG1 fusion protein has a significantly higher affinity for its cognate peptide/HLA complex than the monomeric 264scTCR/trIgG1 protein.
The serum half-life of the 264scTCR/IgG1 molecule is 4 h, which is consistent with that of our other scTCR fusion proteins (18). Although this apparent serum half-life is short compared with other full-length Ab molecules, it is significantly longer that that of Fab and single-chain variable fragments. Moreover, having a shorter half-life for a targeted IgG1-based therapeutic may help to improve on the safety profile of such a targeted therapeutic agent while maintaining the effector function of the FcR-binding portion of the molecule. It is possible that this fusion protein may be cleaved in vivo as has been shown for some other Ab-based fusion proteins (40). Analyzing the blood serum of animals given 264scTCR/IgG1 using ELISAs in other formats (different capture and detection, i.e., TCR/TCR, TCR/IgG, and IgG/IgG), as well as SDS-PAGE and immunoblots, may help determine whether this fusion protein is cleaved in vivo. In addition, it is possible that the fusion protein is taken up rapidly in tissues after administration. Biodistribution studies will elucidate whether the fusion protein is taken up in the tissues and, if so, what proportion is taken up compared with that remaining in circulation.
Mechanisms of action that have been described for therapeutic Abs include ADCC, direct blockage of critical survival signals, and induction of apoptotic signals leading to cell death (4, 5, 6, 7, 8, 9). Our data indicate that the 264scTCR/IgG1 fusion protein is capable of mounting a potent ADCC-mediated killing of targeted cells, although its effect on the unmanipulated tumor cells in vitro is minimal under similar experimental conditions (data not shown). This result does not eliminate the possibility that ADCC is a possible mechanism of action for the antitumor activity of the 264scTCR/IgG1 fusion protein in vivo. For instance, it has been shown that Ag processing and presentation by tumor cells can be altered by the localized inflammatory environment within the surrounding tissues (41). These changes in the tumor cells could make them favorable targets for 264scTCR/IgG1 fusion protein-mediated ADCC. However, the high level of killing in our in vitro ADCC assay and the high efficacy of 264scTCR/IgG1 against tumor metastasis in our in vivo metastasis model suggests that additional mechanisms might also be involved. 264scTCR/IgG1 does not induce apoptosis of APCs in vitro (data not shown). We believe that it might be possible that NK cells are also activated as a result of an altered interaction between NK inhibitory receptors and tumor cell-expressed HLA class I molecules when the 264scTCR/IgG1 fusion protein is present. Recently reported data demonstrates that a weak interaction between a NK cell inhibitory receptor subtype and its corresponding polymorphic HLA molecule are responsible for the rapid clearance of hepatitis C virus in certain individuals. The authors propose that when responding to hepatitis C virus infection, individuals homozygous for a weaker pair of HLA inhibitory receptors more effectively activate NK cells than individuals with other genotypes (42). Therefore, it is conceivable that the 264scTCR/IgG1 molecule blocks the inhibitory receptor interactions with HLA-A2 molecules and thereby stimulates NK cells via the release of the inhibitory signals.
In conclusion, we have demonstrated that the scTCR/IgG1 fusion protein is a viable alternative approach to more traditional Ab-targeted immunotherapeutics. The 264scTCR/IgG1 fusion protein exhibits Ab-like characteristics in that it can specifically bind to tumor cells in vitro and reduce tumor cell metastasis in vivo. TCRs are capable of recognizing intracellular proteins. Therefore, our soluble TCR-based reagents have the potential to expand the range of tumor or viral Ags for targeted therapies beyond those currently addressed by the conventional Ab-based approach.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Dr. Peter Rhode for his thoughtful review of this manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported in part by National Institutes of Health Grant 1R43CA88615-01.
2 Address correspondence and reprint requests to Dr. Hing C. Wong, Altor BioScience Corporation, 2810 North Commerce Parkway, Miramar, FL 33025. E-mail address: hingwong@altorbioscience.com
3 Abbreviations used in this paper: ADCC, Ab-dependent cellular cytotoxicity; scTCR, single-chain TCR; HE, dihydroethidium.
Received for publication October 29, 2004. Accepted for publication January 23, 2005.
References
Carter, P., L. Presta, C. M. Gorman, J. B. Ridgway, D. Henner, W. L. Wong, A. M. Rowland, C. Kotts, M. E. Carver, H. M. Shepard. 1992. Humanization of an anti-p185HER2 antibody for human cancer therapy. Proc. Natl. Acad. Sci. USA 89:4285.
Leget, G. A., M. S. Czuczman. 1998. Use of rituximab, the new FDA-approved antibody. Curr. Opin. Oncol. 10:548.
Pegram, M. D., A. Lipton, D. F. Hayes, B. L. Weber, J. M. Baselga, D. Tripathy, D. Baly, S. A. Baughman, T. Twaddell, J. A. Glaspy, D. J. Slamon. 1998. Phase II study of receptor-enhanced chemosensitivity using recombinant humanized anti-p185HER2/neu monoclonal antibody plus cisplatin in patients with HER2/neu-overexpressing metastatic breast cancer refractory to chemotherapy treatment. J. Clin. Oncol. 16:2659.
Shan, D., J. A. Ledbetter, O. W. Press. 1998. Apoptosis of malignant human B cells by ligation of CD20 with monoclonal antibodies. Blood 91:1644
Sliwkowski, M. X., J. A. Lofgren, G. D. Lewis, T. E. Hotaling, B. M. Fendly, J. A. Fox. 1999. Nonclinical studies addressing the mechanism of action of trastuzumab (Herceptin). Semin. Oncol. 26:60.
van de Winkel, J. G. J., P. J. Capel. 1993. Human IgG Fc receptor heterogeneity: molecular aspects and clinical implications. Immunol. Today 14:215.
Wallace, P. K., A. L. Howell, M. W. Fanger. 1994. Role of Fc receptors in cancer and infectious disease. J. Leukocyte Biol. 55:816.
Fanger, M. W., L. Shen, R. F. Graziano, P. M. Guyre. 1989. Cytotoxicity mediated by human Fc receptors for IgG. Immunol. Today 10:92.
Wallace, P. K., P. A. Kaufman, L. D. Lewis, T. Keler, A. L. Givan, J. L. Fisher, M. G. Waugh, A. E. Wahner, P. M. Guyre, M. W. Fanger, M. S. Ernstoff. 2001. Bispecific antibody-targeted phagocytosis of HER-2/neu expressing tumor cells by myeloid cells activated in vivo. J. Immunol. Methods 248:167.
Welch, W. R., J. M. Niloff, D. Anderson, A. Battaile, S. Emery, R. C. Knapp, R. C. Bast, Jr. 1990. Antigenic heterogeneity in human ovarian cancer. Gynecol. Oncol. 38:12.
Theobald, M., J. Biggs, D. Dittmer, A. J. Levine, L. A. Sherman. 1995. Targeting p53 as a general tumor antigen. Proc. Natl. Acad. Sci. USA 92:11993.
Hinds, P. W., C. A. Finlay, R. S. Quartin, S. J. Baker, E. R. Fearon, B. Vogelstein, A. J. Levine. 1990. Mutant p53 DNA clones from human colon carcinomas cooperate with ras in transforming primary rat cells: a comparison of the "hot spot" mutant phenotypes. Cell Growth Differ. 1:571.
Iggo, R., K. Gatter, J. Bartek, D. Lane, A. L. Harris. 1990. Increased expression of mutant forms of p53 oncogene in primary lung cancer. Lancet 335:675.
McLaughlin, R., D. O’Hanlon, T. McHale, C. E. Connolly, H. F. Given. 2001. Prognostic implications of p53 and bcl-2 expression in 108 women with stage two breast cancer. Ir. J. Med. Sci. 170:11
Temmim, L., H. Baker, F. Sinowatz. 2001. Immunohistochemical detection of p53 protein expression in breast cancer in young Kuwaiti women. Anticancer Res. 21:743.
Hilburger Ryan, M., S. I. Abrams. 2001. Characterization of CD8+ cytotoxic T lymphocyte/tumor cell interactions reflecting recognition of an endogenously expressed murine wild-type p53 determinant. Cancer Immunol. Immunother. 49:603
Kuball, J., M. Schuler, E. Antunes Ferreira, W. Herr, M. Neumann, L. Obenauer-Kutner, L. Westreich, C. Huber, T. Wolfel, M. Theobald. 2002. Generating p53-specific cytotoxic T lymphocytes by recombinant adenoviral vector-based vaccination in mice, but not man. Gene Ther. 9:833.
Card, K. F., S. A. Price-Schiavi, B. Liu, E. Thomson, E. Nieves, H. Belmont, J. Builes, J. A. Jiao, J. Hernandez, J. Weidanz, et al 2004. A soluble single-chain T-cell receptor IL-2 fusion protein retains MHC-restricted peptide specificity and IL-2 bioactivity. Cancer Immunol. Immunother. 53:345
Theobald, M., J. Biggs, J. Hernandez, J. Lustgarten, C. Labadie, L. A. Sherman. 1997. Tolerance to p53 by A2.1-restricted cytotoxic T lymphocytes. J. Exp. Med. 185:833.
Anderson, K. S., J. Alexander, M. Wei, P. Cresswell. 1993. Intracellular transport of class I MHC molecules in antigen processing mutant cell lines. J. Immunol. 151:3407.
van Golen, K. L., S. Risin, A. Staroselsky, D. Berger, M. A. Tainsky, S. Pathak, J. E. Price. 1996. Predominance of the metastatic phenotype in hybrids formed by fusion of mouse and human melanoma clones. Clin. Exp. Metastasis 14:95.(Luis A. Mosquera, Kimberl)
We have constructed a protein composed of a soluble single-chain TCR genetically linked to the constant domain of an IgG1 H chain. The Ag recognition portion of the protein binds to an unmutated peptide derived from human p53 (aa 264–272) presented in the context of HLA-A2.1, whereas the IgG1 H chain provides effector functions. The protein is capable of forming dimers, specifically staining tumor cells and promoting target and effector cell conjugation. The protein also has potent antitumor effects in an in vivo tumor model and can mediate cell killing by Ab-dependent cellular cytotoxicity. Therefore, single-chain TCRs linked to IgG1 H chains behave like Abs but possess the ability to recognize Ags derived from intracellular targets. These fusion proteins represent a novel group of immunotherapeutics that have the potential to expand the range of tumors available for targeted therapies beyond those currently addressed by the conventional Ab-based approach.
Introduction
Targeted therapy for cancer has shown promise as a safe, effective alternative to highly toxic chemotherapeutic approaches. The clinical and commercial success of humanized mAbs Herceptin and Rituximab, which are used to treat metastatic breast cancer and B cell lymphoma, respectively, has prompted the development of a wide variety of Ab-based therapies for both solid and blood-borne tumors (1, 2, 3). The mechanisms of action of therapeutic Abs include direct antiproliferative and proapoptotic effects on their targets (4, 5). In addition, through their Fc domains, Abs may activate Ab-dependent cellular cytotoxicity (ADCC)3 (6, 7, 8) and phagocytosis (9), which may aid in tumor elimination.
Despite the promise of Ab-based immunotherapy, there are a number of potential problems with Ab-based tumor targeting. Abs only recognize Ags expressed on the cell surface; however, there are other tumor Ags that are expressed as intracellular proteins. Moreover, because many tumor-specific Ags are derived from an aberrant expression of cell type-specific proteins, many tumor Ags are expressed on only a small number of tumor types, potentially limiting the number of tumors that could be treated by any one-targeted therapy. Furthermore, tumor Ag expression is often heterogeneous, where the Ag may be expressed on the primary tumor but not on distant metastases of that tumor. Finally, tumor Ags shed from the surface of tumor cells can occupy the binding sites of antitumor-specific Abs, reducing the number of active molecules and resulting in decreased tumor cell death (10). Thus, it would be useful to be able to target a broader range of tumor Ags that may be derived from either intracellular or extracellular proteins that are not subject to the problems of Ag shedding.
The tumor suppressor protein, p53, is mutated and/or overexpressed in a wide variety of human malignancies (11, 12, 13), and its overexpression can correlate with poor prognosis (14, 15). Thus, p53 may be targeted as a general tumor Ag, but because it is an intracellular protein, it cannot be targeted by Ab-based therapies. Instead, p53 peptides presented in the context of HLA have been used as target Ags for T cell-based cancer therapy approaches, including tumor vaccination and ex vivo T cell stimulation and expansion (16, 17). Although these approaches have shown some promise in animal models and some early-stage clinical trials, both of these methods are time consuming, laborious, expensive, and lack standardization, making them unsuitable for treating large patient populations (16, 17). Thus, a therapeutic molecule that can recognize and bind to Ags derived from intracellular proteins such as p53 would significantly broaden the range of tumor Ags that can be targeted.
We have described the construction, expression, and characterization of a soluble three-domain mouse single-chain TCR (scTCR), which recognizes a wild-type human p53 peptide (aa 264–272) presented in the context of HLA-A2.1 (264scTCR) (18). As an IL-2 fusion protein, the 264scTCR retains its MHC-restricted, peptide-specific, Ag-binding properties, whereas the IL-2 portion binds to IL-2R and retains biological activity. Moreover, these studies demonstrate that this fusion protein mediates conjugation of target and effector cells, exhibits favorable pharmacokinetics in mice, binds to target tumor cells, and has antitumor effects (18).
In the studies reported here, we constructed a scTCR fusion protein that resembles an Ab. The Ag recognition portion of the fusion protein is composed of a 264scTCR, and a human IgG1 H chain provides the effector function domain. Fusion of the 264scTCR to the IgG1 H chain allows for dimer formation via the natural disulfide bonding in IgG1 molecules. We evaluate the ability of our fusion protein to bind specifically to cells in vitro as well as reduce metastasis in an in vivo experimental metastasis model. We also explore the possible mechanisms by which this protein might function. Our data indicate that this molecule exhibits many major characteristics of a mAb and that it is possible to create scTCR-based, Ab-like molecules that recognize Ags derived from intracellular proteins such as p53. These novel targeted therapies could prove to be potent immunotherapeutics.
Materials and Methods
Materials
A2.1 264 CTL clone no. 5 was derived by limiting dilution cloning (19) from a CTL line specific for the human p53 (aa 264–272) peptide generated in HLA-A2.1-transgenic mice (11). Expression vectors for producing peptide-loaded HLA-A2 tetramers were a generous gift from Dr. John Altman at Emory University (Atlanta, GA). CHO.K1 Chinese hamster ovary, T2 human lymphoblast, U937 human promyeloma cells, H57-597, BB7.2, and (BF1) 8A3.31 hybridoma cell lines were purchased from American Type Culture Collection. T2 human lymphoblast cells are positive for HLA-A2.1 but deficient in TAP 1 and 2 proteins, which allow them to display empty MHC molecules on their surface that can then be loaded with exogenous peptide (20). The H57-597 hybridoma produces a mAb that recognizes an epitope in the murine TCR- constant region, and the BB7.2 hybridoma produces a mAb that specifically recognizes an epitope on the 2 domain of HLA-A2. The (BF1) 8A3.31 hybridoma produces a mAb that recognizes an epitope in the human TCR- constant region. A highly metastatic subclone of the human melanoma cell line A375, A375-C15N, was used for in vivo metastasis studies and was maintained by Dr. John Francis at the Florida Hospital Cancer Research Institute (Orlando, FL) (21). Anti-TCR C mAb H57-597, anti-murine TCR V3 mAb, and FITC-labeled goat anti-mouse IgG were obtained from BD Pharmingen. Anti-IgG F(ab')2 and HRP-conjugated anti-IgG were purchased from Pierce and The Binding Site, respectively. All cell culture media and additives were purchased from Mediatech, and all cell culture materials were purchased from Nunc unless otherwise noted. All mice were purchased from Harlan. Biotinylated peptide/HLA-A2 reagents were generated as described previously (22).
Cell culture
All cell lines were maintained in complete culture medium comprising IMDM supplemented with 10% FBS plus 2 mM L-glutamine and grown at 37°C with 5% CO2. A375-C15N cells were maintained in RPMI 1640 with 10% heat-inactivated FBS, penicillin, and streptomycin (Invitrogen Life Technologies). For induction of FcR, U937 cells were cultured for 48 h in IMDM supplemented with 10% FBS, 2 mM L-glutamine, and 62.5 ng/ml human IFN- (R&D Systems).
Constructs
The TCR gene was cloned from the T cell clone A2.1 264 no. 5 as described previously (18). We designated the scTCR derived from this T cell clone 264scTCR. To generate the 264scTCR/IgG1 expression construct, the three-domain single-chain 264scTCR was amplified using the 264scTCR/IL-2 fusion protein construct as a template (18). The scTCR fragment was ligated into an Ab H chain expression vector, replacing the Ab variable region to yield a scTCR fused to a human IgG1 H chain region. A control fusion construct, 264scTCR/trIgG1, was created by truncating the IgG1 heavy domain before the hinge region, which prevented disulfide bonding. The CMV scTCR (CMVscTCR) was cloned from CTLs stimulated with HLA-A2-restricted CMV-pp65 peptides. The IgG1 fragment was amplified from 264scTCR/IgG1 DNA to create the CMVscTCR/IgG1 construct. For production of the fusion proteins in mammalian cells, CHO.K1 cells were electroporated using a Bio-Rad Gene Pulser, followed by limiting dilution cloning and selection in medium containing 1 mg/ml G418.
Protein purification
264scTCR/IgG1 and 264scTCR/trIgG1 were purified from cell culture supernatant fluid by immunoaffinity chromatography using the H57-597 mAb coupled to a Sepharose 4B column (Amersham Biosciences). CMVscTCR/IgG1 was purified from cell culture supernatant fluid by immunoaffinity chromatography using the (BF1) 8A3.31 mAb coupled to a Sepharose 4B column (Amersham Biosciences). The purified samples were concentrated and buffer-exchanged into PBS using an Ultrafree-15 centrifugal filter with a 30-kDa molecular mass cutoff membrane (Millipore). The TCR fusion protein samples were stored at 2°C to 8°C (short term) or at –80°C (long term) for biochemical and functional analyses.
SDS-PAGE
SDS-PAGE was performed under reducing or nonreducing conditions using 4–20% polyacrylamide gels (NOVEX) and the NOVEX EX-Cell II system. SDS-PAGE gels were stained with Coomassie blue.
ELISA
For determination of 264scTCR/IgG1 concentration from transfected CHO.K1 cells, Maxisorp 96-well plates (Nunc) were coated with anti-human IgG F(ab')2 (Pierce). Purified 264scTCR/IgG1 fusion protein was added and incubated for 1 h at room temperature. After washing, bound fusion protein was detected with HRP-conjugated sheep anti-human IgG1 (The Binding Site) and ABTS substrate (BioFX) and then quenched with 1% SDS (BioFX). Concentration of 264scTCR/IgG1 was determined by comparison to a known Ab standard curve. For demonstration of peptide/MHC binding, streptavidin plates were coated with p53 (aa 264–272)-loaded HLA-A2 monomers, and purified 264scTCR/IgG1 fusion protein was added and incubated for 1 h at room temperature. After washing, bound fusion protein was detected with HRP-conjugated sheep anti-human IgG1 (The Binding Site) and ABTS substrate (BioFX) and then quenched with 1% SDS (BioFX). For all ELISAs, absorbance was measured at 405 nm using a 96-well plate reader (Bio-Tek Instruments).
Cell staining with 264scTCR/IgG1
T2 cells pulsed with either p53 (aa 149–157) or p53 (aa 264–272) peptide were incubated with 8.0 pM 264scTCR/IgG1 or 264scTCR/trIgG1 fusion protein in 1% FBS in PBS for 30 min at room temperature. Bound fusion protein was detected with PE-conjugated anti-TCR C mAb. Samples were washed with 1% FBS in PBS before analysis on a FACScan flow cytometry instrument (BD Biosciences).
U937 cells were cultured for 48 h with 62.5 ng/ml human IFN- to increase expression of FcR on the cell surface. Cells (2 x 105/sample) were washed and incubated with or without 4 μg of 264scTCR/IgG1 fusion protein for 30 min at room temperature. After washing, cells were incubated with 4 μg of PE-conjugated, HLA-A2/p53 (aa 149–157)- or (aa 264–272)-loaded tetramers for 30 min at room temperature. After washing, the samples were analyzed by flow cytometry.
A375 melanoma cells were stained with either 264scTCR/IgG1 fusion protein or control fusion protein CMVscTCR/IgG1. A375 melanoma cells were cultured on coverslips for 24 h. The cells were fixed with 3.7% formaldehyde for 5 min and washed twice with washing buffer (0.5% BSA and 0.1% sodium azide in PBS). The cells were stained with 10 μg of 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein in 200 μl of PBS containing 5% normal goat serum for 45 min at 23°C. The cells were washed twice and incubated with 6 μg of FITC-conjugated F(ab')2 goat anti-human IgG Fc (Jackson ImmunoResearch Laboratories). The cells were washed three times with equilibration buffer (Molecular Probes). Coverslips were mounted on glass slides with anti-fade reagent in glycerol buffer (Molecular Probes) and sealed with nail polish. The slides were documented using a Nikon epifluorescence microscope (Nikon) with a SPOT RT camera and SPOT RT software v3.2 (Diagnostic Instruments).
For flow cytometric analysis, A375 cells were detached with 10 mM EDTA in PBS (pH 7.4) and washed twice with washing buffer. The cells were fixed with 3.7% formaldehyde for 5 min and washed twice. Cell staining was conducted using 4 μg of 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein in the presence or absence of 20 μg of HLA-A2.1/p53 (aa 264–272) tetramers for 45 min at 23°C. The cells were washed once and incubated with 6 μg of FITC-conjugated F(ab')2 goat anti-human IgG Fc as above. After washing twice, the cells were resuspended and analyzed on a FACScan flow cytometry instrument (BD Biosciences) with CellQuest software (BD Biosciences).
Affinity determinations
The affinity of the dimeric 264scTCR/IgG1 fusion protein for HLA-A2.1/p53 (aa 264–272) was compared with the affinity of the monomeric 264scTCR/trIgG1 fusion protein by surface plasmon resonance (Biacore). Approximately 1300 resonance units of biotinylated p53 (aa 264–272)/HLA-A2.1 complex were immobilized on a streptavidin sensor chip. Each surface was used one time only. The average affinity constants were calculated based on data from three different scTCR fusion protein concentrations (5, 10, and 20 μM) using the Biacore computer software Biaevaluation 3.1 (Biacore).
Cell conjugation
T2 cells pulsed with either p53 (aa 264–272) peptide or (aa 149–157) peptides were labeled with 12.6 μg/ml dihydroethidium (HE) (Molecular Probes), and U937 cells were labeled with 80 ng/ml calcein AM (Molecular Probes). After washing, the two populations of labeled cells were mixed together at a 1:1 ratio in the presence or absence of 2.5 μg of 264scTCR/IgG1 fusion protein for 20 min at room temperature. Cells were then analyzed on a FACScan flow cytometry instrument.
Pharmacokinetics in mice
For all experiments involving animals, principles of laboratory animal care were followed, as well as specific national laws where applicable. Female athymic nude mice were injected i.v. via the lateral tail vein with 250 μg of 264scTCR/IgG1 fusion protein diluted with PBS to a total volume of 100 μl. Serum was collected from one group of mice not injected with 264scTCR/IgG1 to establish background levels. Serum was collected by tail bleed from the injected groups at 15 and 30 min and 1, 2, 4, 8, and 24 h. Blood samples were centrifuged at 14,000 x g at 4°C for 10 min, and serum was collected and stored at –80°C until use. 264scTCR/IgG1 concentrations were determined by ELISA using anti-TCR V3 mAb for capture and HRP-conjugated sheep anti-human IgG1 Ab for detection.
Experimental metastasis assay
Female athymic nude mice were injected with 5.0 x 105 A375-C15N cells via the lateral tail vein on study day 0. The tumors were allowed to establish for 3 days and then the animals were injected i.v. with varying doses of either 264scTCR/IgG1 (0.1–30 mg/kg in 100 μl of total volume), PBS, or Stx-2 (an irrelevant Ab isotype-matched control at 30 mg/kg in 100 μl of total volume) on study days 3, 7, 10, 14, 17, 21, 28, and 35 posttumor cell injection. Forty-two days after tumor cell injection, all animals were sacrificed humanely, the lungs were removed and fixed in Bouin’s solution, and surface pulmonary tumor nodules were counted. Two observers counted tumor nodules on each lung, and the average counts were recorded. The Stx-2 Ab, a chimeric IgG1 Ab (mouse variable/human H and L chain constant region), was created in our lab and has been described previously (23). This Ab recognizes an epitope on Escherichia coli 0157 shigatoxin protein and was used for these studies as an irrelevant isotype control.
ADCC
T2 cells were loaded with p53 (aa 264–272) peptide and used as target cells. Target cells were labeled with 50 μg/ml calcein AM (Molecular Probes) for 30 min at 37°C and washed twice. Effector PBMCs were isolated from buffy coats (Community Blood Centers of South Florida) using the one-step gradient method with Histopaque-1077 (Sigma-Aldrich). To measure ADCC activity, 2 x 106 effector cells were incubated with 2 x 104 target cells in 0.25 ml of complete medium with either 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein at 20 μg/ml. The fluorescent intensity of culture supernatant fluid was measured after 4 h with a fluorescent plate reader (excitation = 485 nM; emission = 538 nM; cutoff = 530 nM). ADCC activity was calculated using the following formula: percentage of cytotoxicity = (fluorescent intensity of test sample – fluorescent intensity of target cells with medium)/(fluorescent intensity of target cells treated with 0.04% Triton X-100 – fluorescent intensity of target cells with medium) x 100. Data are reported as the mean of triplicate determinations ± SEM. Statistical significance was determined by paired t test analysis.
Results
Generation of TCR fusion protein constructs
We have constructed a fusion protein composed of a three-domain scTCR that recognizes an unmutated peptide from human p53 (aa 264–272) (18) fused to the constant domain of the human IgG1 H chain. The scTCR portion of this fusion protein comprises the TCR V region linked to the V/C regions via a flexible linker (G4S)4 (24). The C domain, which is directly linked to the V domain, is truncated at the amino acid residue just before the final cysteine, removing the transmembrane and cytoplasmic domains, to generate a soluble scTCR molecule. This scTCR is linked directly to the constant domain of the human IgG1 H chain (Fig. 1). The cysteine residues in the hinge region of the H chain are intact, which allows for appropriate disulfide bonding of the H chain regions and thus covalently linked dimer formation of the 264scTCR/IgG1 fusion protein.
FIGURE 1. The 264scTCR/IgG1 fusion protein. Schematic representation of the domain structure of the 264scTCR/IgG1 fusion protein.
Expression of TCR/IgG1 fusion protein in mammalian cells
To characterize the soluble 264scTCR/IgG1 fusion protein, CHO-K1 cells were stably transfected with the 264scTCR/IgG1 expression vector. To determine the concentration of secreted protein, the fusion protein was captured with an anti-IgG F(ab')2 and detected with an anti-IgG1 HRP conjugate. Concentrations of the 264scTCR/IgG1 fusion protein were estimated based on comparison to an IgG1 Ab standard. To determine whether secreted protein can bind peptide-loaded HLA, the fusion protein was captured with biotinylated HLA-A2/p53 (aa 264–272) monomers on streptavidin-coated plates and detected with anti-IgG1 HRP conjugate. Positive signals for ELISAs in either format indicate that the transfected cells secrete 264scTCR/IgG1 fusion protein that is assembled properly, folded, and secreted from the transfected cells. Moreover, the TCR portion of 264scTCR/IgG1 retains peptide-specific MHC binding ability in this fusion protein format.
The 264scTCR/IgG1 fusion protein was purified from transfected cell supernatants by immunoaffinity chromatography using an anti-murine TCR C mAb, H57-597. Purified protein was subjected to SDS-PAGE and Coomassie staining (Fig. 2). Under reducing conditions, the major protein band was 75–80 kDa, which is consistent with the calculated molecular mass for this protein based on its deduced amino acid sequence. Under nonreducing conditions, the molecular mass of the predominant protein band was 170–180 kDa. The apparent shift in molecular mass of the 264scTCR/IgG1 under reducing vs nonreducing conditions is to be expected because there are cysteine residues in the H chain regions that form disulfide bonds. Thus, as with an IgG1 Ab, the 264scTCR/IgG1 fusion protein forms dimers.
FIGURE 2. Production of 264scTCR/IgG1 fusion protein in transfected Chinese hamster ovary cells. The secreted fusion protein was purified by immunoaffinity chromatography and subjected to SDS-PAGE under either reducing (R) or nonreducing (N) conditions. SDS-PAGE gels was stained with Coomassie brilliant blue.
MHC/peptide binding of the TCR portion of 264scTCR/IgG1
The ability of the 264scTCR/IgG1 fusion protein to bind to peptide-loaded MHC on the surface of APCs was determined by flow cytometry. T2 lymphoblast cells were loaded with p53 (aa 264–272) or p53 (aa 149–157) control peptide and incubated with purified 264scTCR/IgG1 fusion protein. After several wash steps, the cells were incubated with PE-labeled anti-TCR mAbs to detect cell-bound fusion protein (Fig. 3). T2 cells loaded with p53 (aa 264–272) peptide stained positively with 264scTCR/IgG1, whereas cells loaded with p53 (aa 149–157) showed little or no staining, indicating that the 264scTCR/IgG1 could stain cells in a peptide-specific manner. We also attempted to stain peptide-loaded T2 cells with monomeric 264scTCR/trIgG1 fusion protein and found that no staining was observed on cells loaded with either p53 (aa 264–272) or p53 (aa 149–157) peptide. These results demonstrate that the bivalent nature of 264scTCR/IgG1 greatly increased its ability to stably bind MHC-peptide complexes on the surface of cells.
FIGURE 3. MHC/peptide-binding ability of the TCR portion of 264scTCR/IgG1 fusion protein. T2 cells were loaded with p53 (aa 264–272) peptide and stained with either 264scTCR/IgG1 (black thin line) or 264scTCR/trIgG1 fusion protein (gray thick line), followed by anti-TCR C mAb. T2 cells loaded with p53 (aa 149–157) control peptide were also stained with either 264scTCR/IgG1 (dotted black line) or 264scTCR/trIgG1 fusion protein (black thick line), followed by anti-TCR C mAb. All samples were analyzed by flow cytometry.
Binding affinity of 264scTCR/IgG1
To determine whether the intensity of staining observed with 264scTCR/IgG1 could be attributed to a high binding affinity, we performed surface plasmon resonance assays. The binding affinity (KD) of the dimeric 264scTCR/IgG1 with HLA-A2.1/p53 (aa 264–272) complexes was estimated to be 5.6 x 10–7 M, which is significantly higher than the monomeric 264scTCR/trIgG1 (KD of 2.2 x 10–6 M; Table I). The increased affinity of the 264scTCR/IgG1 fusion protein likely contributes to its ability to bind strongly to the correct peptide/MHC pair.
Table I. Affinity determination with 264scTCR/IgG1 fusion proteina
Binding of the Fc domain of 264scTCR/IgG1 to FcR
To demonstrate the ability of the IgG1 Fc portion of the 264scTCR/IgG1 fusion protein to bind to FcR, U937 human promyeloma cells were stimulated with human IFN- to increase expression of the FcR on the cell surface. Stimulated as well as unstimulated cells were incubated with 264scTCR/IgG1 fusion protein, and bound fusion protein was detected with PE-conjugated HLA-A2.1/p53 (aa 264–272) or HLA-A2.1/p53 (aa 149–157) tetramers and analyzed by flow cytometry. When the 264scTCR/IgG1 fusion protein was used, cells stained positively following incubation with HLA-A2.1/p53 (aa 264–272) tetramers but not with HLA-A2.1/p53 (aa 149–157) tetramers. Moreover, positive staining with 264scTCR/IgG1 fusion protein was significantly more intense when the cells were stimulated with human IFN-, which is known to increase the expression of FcR on U937 cells (Fig. 4). Thus, these data indicate that the IgG1 portion of the 264scTCR/IgG1 fusion protein is capable of binding to FcRs.
FIGURE 4. Binding of the Fc domain of 264scTCR/IgG1 to FcR. rIFN--stimulated U937 cells were stained with 264scTCR/IgG1 fusion protein followed by HLA-A2/p53 (aa 264–272) (black thin line) or HLA-A2/p53 (aa 149–157) (black thick line) tetramers. Unstimulated U937 cells were also stained with fusion protein and HLA-A2/p53 (aa 264–272) (black dotted line) or HLA-A2/p53 (aa 149–157) (gray thick line) tetramers. All samples were analyzed by flow cytometry.
Conjugation of cells mediated by 264scTCR/IgG1 fusion protein
In order for the 264scTCR/IgG1 fusion protein to be useful as a therapeutic, it must be able to bring together target and effector cells through the TCR and Fc components, respectively. To demonstrate that the 264scTCR/IgG1 fusion protein can effectively conjugate cells, T2 cells were loaded with either p53 (aa 264–272) or p53 (aa 149–157) control peptide and then stained with HE. U937 effector cells were differentially stained with calcein AM. The two stained cell populations were mixed, incubated in the presence or absence of 264scTCR/IgG1, and analyzed by flow cytometry. When the two cell populations were incubated in the absence of the 264scTCR/IgG1 fusion protein (Fig. 5, A and B) or when the T2 cells were loaded with control peptide p53 (aa 149–157) and incubated with the U937 cells in the presence of 264scTCR/IgG1 fusion protein (Fig. 5C), the cells remained as two distinct populations on the flow cytometry histograms (Fig. 5, A–C, regions 1 and 3). However, when the T2 cells were loaded with p53 (aa 264–272) peptide and incubated with the U937 cells in the presence of the 264scTCR/IgG1 fusion protein (Fig. 5D), a double-stained population of cells appears (Fig. 5D, region 2, conjugated cells), indicating that T2 cells were conjugated to U937 cells via the 264scTCR/IgG1 fusion protein. These data demonstrate that the 264scTCR/IgG1 fusion proteins can mediate conjugation of effector cells to target cells.
FIGURE 5. Conjugation of cells mediated by 264scTCR/IgG1. T2 cells were loaded with either p53 (aa 264–272) (B and D) or p53 (aa 149–157) (A and C) peptides and then labeled with HE. U937 cells were labeled with calcein AM. Labeled cells were mixed and incubated in the presence (C and D) or absence (A and B) of 264scTCR/IgG1 fusion protein. Single-stained regions are marked 1 and 3, and the double-stained cell population is marked 2.
Tumor cell staining
To demonstrate the ability of 264TCR/IgG1 fusion protein to bind to tumor cells, the A375 human melanoma cell line, shown in our laboratory to be HLA-A2.1 and p53 positive (data not shown), was used. The 264TCR/IgG1 fusion protein positively stained the A375 cells (Fig. 6A), while staining with CMVscTCR/IgG1 was not detectable above background (Fig. 6B). Staining of A375 cells with 264scTCR/IgG1 was reduced by the addition of HLA-A2.1/p53 (aa 264–272) tetramers (Fig. 6C), indicating that 264scTCR/IgG1 specifically binds to tumor cells via the scTCR portion of the fusion protein.
FIGURE 6. Tumor cell staining with 264scTCR/IgG1 fusion protein. Formaldehyde-fixed A375 cells were stained with 264scTCR/IgG1 fusion protein (A) or CMVscTCR/IgG1 fusion protein (B). The images were captured at x400. Fixed A375 cells were stained with 264scTCR/IgG1 (black lines) or CMVscTCR/IgG1 (gray lines) fusion protein in the presence (thin or dotted lines) or absence (thick lines) of HLA-A2.1/p53 (aa 264–272) tetramers (C).
Serum half-life of 264scTCR/IgG1 in mice
The pharmacokinetics of the 264scTCR/IgG1 fusion protein was measured in nude mice. All mice were injected with a single i.v. dose of 264scTCR/IgG1 and showed no obvious signs of toxicity. Serum samples were collected at various time points, and the 264scTCR/IgG1 fusion protein concentration was determined using ELISA. In these assays, a maximum concentration of 8.8 μg/ml 264scTCR/IgG1 was detected with an apparent serum half-life of 4 h (Fig. 7). This is consistent with the apparent serum half-life of the 264scTCR/IL-2 fusion protein (18).
FIGURE 7. Serum half-life of 264scTCR/IgG1 fusion protein. Nude mice were injected with 264scTCR/IgG1 fusion protein, and serum samples were collected at 15 and 30 min and 1, 4, 8, and 24 h postinjection. Serum concentrations of 264scTCR/IgG1 were determined by ELISA.
Effect of 264scTCR/IgG1 on tumor growth in nude mice
To determine whether the 264scTCR/IgG1 fusion protein exhibits antitumor activity in vivo, an experimental metastasis assay was performed. Female athymic nude mice were injected with the highly metastatic A375 human melanoma subclone, A375-C15N, and tumors were allowed to establish for 3 days. Animals were then treated with varying doses of 264scTCR/IgG1, -Stx-2 (an irrelevant IgG1 isotype control Ab), or vehicle. Forty-two days after tumor cell injection, lung nodules were counted. 264scTCR/IgG1 reduced lung metastasis in a dose-dependent manner reaching maximum reduction of metastasis at the 1 mg/kg dose, whereas the isotype control failed to reduce metastasis (Fig. 8). These data indicate that the 264scTCR/IgG1 fusion protein exhibits potent anti-metastatic activities in vivo and has potential use as a cancer therapeutic in humans.
FIGURE 8. Antitumor effects of 264scTCR/IgG1. Female athymic nude mice were injected with highly metastatic A375-C15N cells and treated with 264scTCR/IgG1 at doses ranging from 0.1 to 30 or 30 mg/kg anti-Stx-2 (an irrelevant IgG1 isotype control Ab) or a dose volume equivalent of PBS. Forty-two days after tumor cell injection, the lungs were removed, lung nodules were counted, and the mean number of lung nodules was plotted.
ADCC
One possible mechanism of action for therapeutic Abs is ADCC, which is mediated by the FcR-binding portion of the IgG molecule. Because the 264scTCR/IgG1 fusion protein contains a FcR-binding domain, we investigated whether this fusion protein could mediate ADCC. T2 cells loaded with p53 (aa 264–272) peptide and labeled with calcein AM were used as target cells, and human PBMCs were used as effector cells. Target and effector cells were mixed together at a 1:100 ratio in the presence of 264scTCR/IgG1 or CMVscTCR/IgG1 fusion protein. Cell killing with 264scTCR/IgG1 ranged from 48 to 89%, while cell killing with CMVscTCR/IgG1 protein was significantly lower (p = 0.02) (Fig. 9). These data indicate that the 264scTCR/IgG1 protein is capable of mediating ADCC and that this mechanism may play a role in the antitumor activity of this fusion protein. Given that the population of FcR-positive effector cells in PBMCs used for the ADCC experiment is 15–20% of the population, the high level of killing seen in these assays suggests that there may be one or more other mechanisms of action used for cell killing mediated by 264scTCR/IgG1.
FIGURE 9. 264scTCR/IgG1 mediation of ADCC. T2 cells loaded with p53 (aa 264–272) peptide were labeled with calcein AM and incubated with human PBMCs (100:1 E:T ratio) in the presence of 264scTCR/IgG1 () or CMVscTCR/IgG1 ( ) fusion protein. After 4 h, culture supernatants were collected, and cell lysis was quantified.
Discussion
In these studies, we have described the construction and characterization of a novel 264scTCR/IgG1 fusion protein, which recognizes an unmutated peptide from human p53 (aa 264–272) presented in the context of HLA-A2.1. The IgG1 H chain domain confers FcR binding and allows for dimer formation, and the 264scTCR portion of the molecule confers peptide-specific MHC binding. This fusion protein is capable of specifically staining tumor cells and promotes target and effector cell conjugation. The 264scTCR/IgG1 fusion protein also has potent antitumor effects in an in vivo tumor model and can mediate cell killing by an ADCC mechanism of action. Thus, we have demonstrated that it is feasible to create scTCR fusion molecules, which are functionally equivalent to therapeutic mAbs but with the ability to recognize target Ags derived from intracellular proteins.
The development of TCR-based therapeutic agents has been precluded by the difficulty of producing stable, functionally active TCR molecules in sufficiently large quantities (25, 26, 27, 28, 29, 30, 31, 32). We have overcome this obstacle by constructing soluble, single-chain, three-domain TCR molecules comprising the TCR- variable domain linked to the TCR- variable and constant regions (18, 33). These constructs are active with respect to specific peptide/MHC binding, easily expressed in large quantities in mammalian cells, and readily purified for use in in vitro and in vivo assays (18). Expression and secretion of scTCR-based proteins in mammalian cells offers advantages over bacterial expression systems that require time-consuming, laborious refolding methods to produce TCR-/ heterodimers (28, 30, 32). In addition, solubly expressed scTCRs can be genetically fused with molecules exhibiting effector functions such as cytokines to create bifunctional-bispecific fusion proteins. For example, scTCR molecules linked to human IL-2 (18), GM-CSF, and IFN-2b (K. Card, unpublished observations) have been produced and shown to retain their respective bispecific functions. In another approach, Abs with specificity to peptide/MHC complexes have been generated and have been used to quantitate peptide/MHC complexes on the surface of APCs and deliver cytotoxic effector toxin to APCs in a peptide-dependent manner (34, 35).
TCRs are known to have low affinity for cognate peptide/MHC complexes due to negative selection in the thymus during the T cell maturation process (36). In addition, the density of a particular epitope displayed in the context of MHC on an APC surface is predictably low (34). These two factors raised significant doubts about whether it was feasible to use TCRs as targeting molecules. The tumor cell staining results in this study demonstrate that a dimeric form of the scTCR does in fact have sufficient affinity to function as an Ag recognition molecule in vitro. The antitumor effects of the 264scTCR/IgG1 fusion protein in our experimental metastasis model additionally demonstrate that the Ag-targeting capability of this type of scTCR molecule also functions in vivo. We are currently performing experiments to determine whether the density of the peptide/MHC complexes on the tumor cell surface directly influences the ability of scTCR fusion proteins to target tumor cells.
Affinity plays an important role in tumor retention and penetration by macromolecules such as Abs. It has been shown that increasing the affinity of Abs increases tumor retention, but at a KD > 10–9 M, tumor retention is not further improved, and penetration is hindered significantly (37, 38). The 264scTCR was derived from a T cell line generated from a HLA-A2.1-transgenic mouse. TCRs derived from these mice are CD8 independent and are postulated to have high affinity for their cognate peptide/MHC complex (11). Our measurements indicate that the affinity of the dimeric 264scTCR/IgG1 fusion molecule for peptide/MHC is on the order of 5.6 x 10–7 M KD, which is exceptionally high for a TCR. Abs with this level of affinity show low levels of tumor retention but very efficient tumor penetration (37, 38). Whether the unusually high affinity of this scTCR for its cognate peptide/MHC is the key to successful tumor cell staining and antitumor activity is unknown but is receiving additional investigation in our laboratory. The affinity of scTCR molecules can also be increased substantially using a variety of in vitro enhancement methods in combination with mutagenesis (33, 39). For example, we have developed a phage display method wherein scTCRs are expressed as bacteriophage M13 gene VIII fusions (33). In vitro multimerization methods, such as birA-streptavidin conjugation or cross-linking with liposomes or dendrimers, may also increase the avidity of scTCR molecules. Such multimerization should increase the avidity of a scTCR-based therapeutic agent for its target without sacrificing Ag recognition specificity. Dimerization using the IgG H chain illustrates the feasibility of such an approach because the dimeric 264scTCR/IgG1 fusion protein has a significantly higher affinity for its cognate peptide/HLA complex than the monomeric 264scTCR/trIgG1 protein.
The serum half-life of the 264scTCR/IgG1 molecule is 4 h, which is consistent with that of our other scTCR fusion proteins (18). Although this apparent serum half-life is short compared with other full-length Ab molecules, it is significantly longer that that of Fab and single-chain variable fragments. Moreover, having a shorter half-life for a targeted IgG1-based therapeutic may help to improve on the safety profile of such a targeted therapeutic agent while maintaining the effector function of the FcR-binding portion of the molecule. It is possible that this fusion protein may be cleaved in vivo as has been shown for some other Ab-based fusion proteins (40). Analyzing the blood serum of animals given 264scTCR/IgG1 using ELISAs in other formats (different capture and detection, i.e., TCR/TCR, TCR/IgG, and IgG/IgG), as well as SDS-PAGE and immunoblots, may help determine whether this fusion protein is cleaved in vivo. In addition, it is possible that the fusion protein is taken up rapidly in tissues after administration. Biodistribution studies will elucidate whether the fusion protein is taken up in the tissues and, if so, what proportion is taken up compared with that remaining in circulation.
Mechanisms of action that have been described for therapeutic Abs include ADCC, direct blockage of critical survival signals, and induction of apoptotic signals leading to cell death (4, 5, 6, 7, 8, 9). Our data indicate that the 264scTCR/IgG1 fusion protein is capable of mounting a potent ADCC-mediated killing of targeted cells, although its effect on the unmanipulated tumor cells in vitro is minimal under similar experimental conditions (data not shown). This result does not eliminate the possibility that ADCC is a possible mechanism of action for the antitumor activity of the 264scTCR/IgG1 fusion protein in vivo. For instance, it has been shown that Ag processing and presentation by tumor cells can be altered by the localized inflammatory environment within the surrounding tissues (41). These changes in the tumor cells could make them favorable targets for 264scTCR/IgG1 fusion protein-mediated ADCC. However, the high level of killing in our in vitro ADCC assay and the high efficacy of 264scTCR/IgG1 against tumor metastasis in our in vivo metastasis model suggests that additional mechanisms might also be involved. 264scTCR/IgG1 does not induce apoptosis of APCs in vitro (data not shown). We believe that it might be possible that NK cells are also activated as a result of an altered interaction between NK inhibitory receptors and tumor cell-expressed HLA class I molecules when the 264scTCR/IgG1 fusion protein is present. Recently reported data demonstrates that a weak interaction between a NK cell inhibitory receptor subtype and its corresponding polymorphic HLA molecule are responsible for the rapid clearance of hepatitis C virus in certain individuals. The authors propose that when responding to hepatitis C virus infection, individuals homozygous for a weaker pair of HLA inhibitory receptors more effectively activate NK cells than individuals with other genotypes (42). Therefore, it is conceivable that the 264scTCR/IgG1 molecule blocks the inhibitory receptor interactions with HLA-A2 molecules and thereby stimulates NK cells via the release of the inhibitory signals.
In conclusion, we have demonstrated that the scTCR/IgG1 fusion protein is a viable alternative approach to more traditional Ab-targeted immunotherapeutics. The 264scTCR/IgG1 fusion protein exhibits Ab-like characteristics in that it can specifically bind to tumor cells in vitro and reduce tumor cell metastasis in vivo. TCRs are capable of recognizing intracellular proteins. Therefore, our soluble TCR-based reagents have the potential to expand the range of tumor or viral Ags for targeted therapies beyond those currently addressed by the conventional Ab-based approach.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Dr. Peter Rhode for his thoughtful review of this manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported in part by National Institutes of Health Grant 1R43CA88615-01.
2 Address correspondence and reprint requests to Dr. Hing C. Wong, Altor BioScience Corporation, 2810 North Commerce Parkway, Miramar, FL 33025. E-mail address: hingwong@altorbioscience.com
3 Abbreviations used in this paper: ADCC, Ab-dependent cellular cytotoxicity; scTCR, single-chain TCR; HE, dihydroethidium.
Received for publication October 29, 2004. Accepted for publication January 23, 2005.
References
Carter, P., L. Presta, C. M. Gorman, J. B. Ridgway, D. Henner, W. L. Wong, A. M. Rowland, C. Kotts, M. E. Carver, H. M. Shepard. 1992. Humanization of an anti-p185HER2 antibody for human cancer therapy. Proc. Natl. Acad. Sci. USA 89:4285.
Leget, G. A., M. S. Czuczman. 1998. Use of rituximab, the new FDA-approved antibody. Curr. Opin. Oncol. 10:548.
Pegram, M. D., A. Lipton, D. F. Hayes, B. L. Weber, J. M. Baselga, D. Tripathy, D. Baly, S. A. Baughman, T. Twaddell, J. A. Glaspy, D. J. Slamon. 1998. Phase II study of receptor-enhanced chemosensitivity using recombinant humanized anti-p185HER2/neu monoclonal antibody plus cisplatin in patients with HER2/neu-overexpressing metastatic breast cancer refractory to chemotherapy treatment. J. Clin. Oncol. 16:2659.
Shan, D., J. A. Ledbetter, O. W. Press. 1998. Apoptosis of malignant human B cells by ligation of CD20 with monoclonal antibodies. Blood 91:1644
Sliwkowski, M. X., J. A. Lofgren, G. D. Lewis, T. E. Hotaling, B. M. Fendly, J. A. Fox. 1999. Nonclinical studies addressing the mechanism of action of trastuzumab (Herceptin). Semin. Oncol. 26:60.
van de Winkel, J. G. J., P. J. Capel. 1993. Human IgG Fc receptor heterogeneity: molecular aspects and clinical implications. Immunol. Today 14:215.
Wallace, P. K., A. L. Howell, M. W. Fanger. 1994. Role of Fc receptors in cancer and infectious disease. J. Leukocyte Biol. 55:816.
Fanger, M. W., L. Shen, R. F. Graziano, P. M. Guyre. 1989. Cytotoxicity mediated by human Fc receptors for IgG. Immunol. Today 10:92.
Wallace, P. K., P. A. Kaufman, L. D. Lewis, T. Keler, A. L. Givan, J. L. Fisher, M. G. Waugh, A. E. Wahner, P. M. Guyre, M. W. Fanger, M. S. Ernstoff. 2001. Bispecific antibody-targeted phagocytosis of HER-2/neu expressing tumor cells by myeloid cells activated in vivo. J. Immunol. Methods 248:167.
Welch, W. R., J. M. Niloff, D. Anderson, A. Battaile, S. Emery, R. C. Knapp, R. C. Bast, Jr. 1990. Antigenic heterogeneity in human ovarian cancer. Gynecol. Oncol. 38:12.
Theobald, M., J. Biggs, D. Dittmer, A. J. Levine, L. A. Sherman. 1995. Targeting p53 as a general tumor antigen. Proc. Natl. Acad. Sci. USA 92:11993.
Hinds, P. W., C. A. Finlay, R. S. Quartin, S. J. Baker, E. R. Fearon, B. Vogelstein, A. J. Levine. 1990. Mutant p53 DNA clones from human colon carcinomas cooperate with ras in transforming primary rat cells: a comparison of the "hot spot" mutant phenotypes. Cell Growth Differ. 1:571.
Iggo, R., K. Gatter, J. Bartek, D. Lane, A. L. Harris. 1990. Increased expression of mutant forms of p53 oncogene in primary lung cancer. Lancet 335:675.
McLaughlin, R., D. O’Hanlon, T. McHale, C. E. Connolly, H. F. Given. 2001. Prognostic implications of p53 and bcl-2 expression in 108 women with stage two breast cancer. Ir. J. Med. Sci. 170:11
Temmim, L., H. Baker, F. Sinowatz. 2001. Immunohistochemical detection of p53 protein expression in breast cancer in young Kuwaiti women. Anticancer Res. 21:743.
Hilburger Ryan, M., S. I. Abrams. 2001. Characterization of CD8+ cytotoxic T lymphocyte/tumor cell interactions reflecting recognition of an endogenously expressed murine wild-type p53 determinant. Cancer Immunol. Immunother. 49:603
Kuball, J., M. Schuler, E. Antunes Ferreira, W. Herr, M. Neumann, L. Obenauer-Kutner, L. Westreich, C. Huber, T. Wolfel, M. Theobald. 2002. Generating p53-specific cytotoxic T lymphocytes by recombinant adenoviral vector-based vaccination in mice, but not man. Gene Ther. 9:833.
Card, K. F., S. A. Price-Schiavi, B. Liu, E. Thomson, E. Nieves, H. Belmont, J. Builes, J. A. Jiao, J. Hernandez, J. Weidanz, et al 2004. A soluble single-chain T-cell receptor IL-2 fusion protein retains MHC-restricted peptide specificity and IL-2 bioactivity. Cancer Immunol. Immunother. 53:345
Theobald, M., J. Biggs, J. Hernandez, J. Lustgarten, C. Labadie, L. A. Sherman. 1997. Tolerance to p53 by A2.1-restricted cytotoxic T lymphocytes. J. Exp. Med. 185:833.
Anderson, K. S., J. Alexander, M. Wei, P. Cresswell. 1993. Intracellular transport of class I MHC molecules in antigen processing mutant cell lines. J. Immunol. 151:3407.
van Golen, K. L., S. Risin, A. Staroselsky, D. Berger, M. A. Tainsky, S. Pathak, J. E. Price. 1996. Predominance of the metastatic phenotype in hybrids formed by fusion of mouse and human melanoma clones. Clin. Exp. Metastasis 14:95.(Luis A. Mosquera, Kimberl)