Background
In patients with cancer, the body mounts an immune response following the onset of malignant disease since the new cells are recognized as non-self. It is composed of both immune cells that mediate innate, non-specific immunity, and adaptive, antigen-specific immunity [
1‐
3]. Tumor cell proteins can elicit an immune response for various reasons; aberrant gene expression (e.g. cancer-testis antigens) [
4‐
10], overexpression (neu/Her2) [
11,
12], aberrant processing (mucin) [
13,
14] and mutation events (p53) [
11,
15]. Although it is evident that a natural humoral response to cancer exists, tumor-associated antigens (TAAs) are generally notoriously bad immunogens. This is likely due to systemic tolerance to the autoantigens and, as a result, the natural humoral immune response against tumor antigens fails to reach high antibody titers and is not effective [
16].
During the last decade, the search for TAAs that can be targeted by the immune system, and as such are "immunovisible", has been the focus of much research in cancer immunology. In addition, the isolation and production of fully human monoclonal antibodies (fhMAb) to such antigens has also made significant advances over the past few years [
16‐
18]. The potential utility of these antibodies to identify TAAs, to discriminate between neoplastic and normal tissues and potentially act as anti-cancer therapeutics has been the impetus for this work.
To identify tumor-associated antigens, one of the more fruitful approaches has been to employ naturally occurring anti-cancer antibodies that arise in cancer patients. To this end, serological expression technology (SEREX) has facilitated the identification of novel TAAs by screening patients' whole sera on cDNA expression libraries that were prepared from autologous tumors or human cancer cell lines [
19‐
23]. This technology has led to the creation of a database of protein antigens that are associated with and specific to a variety of cancers. However, the native immune response to these antigens is not identified or captured by this methodology. Therefore, although proteins that are associated specifically with cancer can be pinpointed, the antibodies that can effectively target these antigens remain mostly unidentified.
To overcome this limitation we designed and implemented an alternate strategy that relies on a unique trioma fusion partner cell line, MFP-2, which we developed [
24]. MFP-2 can efficiently fuse with both peripheral blood and lymph node lymphocytes. Following fusion, surviving hybridoma clones are stable for prolonged periods and many produce significant quantities of human monoclonal antibodies. We employed this unique fusion partner cell line to develop a panel of native autologous fully human monoclonal antibodies (fhMAb) that were culled from the natural repertoire arising in breast cancer patients [
25]. These fhMAbs reacted specifically with breast cancer cells and malignant tissues. They are useful not only for identification of the target antigens, but also for immunodiagnostic procedures [
26] and eventually for immunotherapy of breast cancer, since they can be produced on an industrial scale.
We identified the protein targets of two of the anti-breast cancer autoantibodies that we isolated, and determined that they target the protein GIPC1. Using our fhMAbs that target GIPC1, we studied its expression in human breast tissue and in cultured cells. We determined that this protein is specifically up-regulated in malignant breast epithelial tissue/cells and in breast cancer cell lines and is not detected in normal breast epithelia or in live primary fibroblast cell lines. Therefore, GIPC1 is a novel breast cancer-associated antigen that may play a role in the initiation and/or progression of breast cancer.
Methods
Cell culture
All human cancer and normal cell lines were purchased from ATCC. Human breast cancer cell lines MCF-7 and SK-BR-3 and primary human fibroblasts are among those used in this study. SK-BR-3 were cultured in McCoy's 5a medium supplemented with L-glutamine and 10% FBS. MCF-7 was grown in MEM medium supplemented with L-glutamine, non-essential aminoacids, 10% FBS and 0.01 mg/ml bovine insulin. Fibroblasts were cultured in DMEM, supplemented with L-glutamine and 10% FBS. Other normal and neoplastic cell lines were cultured according to the conditions recommended by the ATCC. Hybridoma clones were produced and cultured according to previously described techniques [
25].
Antibody characterization
The isotype of human Abs was determined by ELISA using murine anti-human isotype-specific MAbs to μ-, γ-, κ- and λ-chains (Sigma, USA) and goat anti-mouse Ig (25 μg/mL) conjugated to peroxidase and absorbed with human Ig.
Immunocytochemistry and immunohistochemistry
Cells were plated on ethanol pre-treated cover slips (Fisher, USA) and placed in 6-well plates (Falcon, USA) in culture medium. After 24 hours the cover slips with attached cells were repeatedly washed in PBS and fixed in ethanol. Following fixation and repeated washes, cover slips were incubated with the primary and secondary antibodies according to standard protocols, stained with propidium iodide 1 μg/ml and analyzed by confocal fluorescent microscopy using a Zeiss Axiovert 100 TV microscope and Zeiss software. For immunohistochemistry randomly selected 5 μm thick sections of paraffin embedded breast cancer tissue were used. Endogenous peroxidase activity was blocked by incubation of slides in 3% H2O2 in methanol. Following washing, tissue slides were blocked with 5% normal goat serum in PBS. Monovalent Fab fragments of goat anti-human IgM+IgG (Jackson Immunoresearch Laboratories, Inc.), in blocking solution, was then applied for secondary blocking. After 3 washes in PBS, the human monoclonal antibody was applied at an approximate concentration of 5 ug/ml. The slides were then washed and incubated with a second FITC conjugated antibody to human κ-light chains (Sigma, USA) and propidium iodide at 1 μg/ml. Following a few washes, mounting medium (Biomeda, USA) and cover slips were applied and sections were analyzed by standard fluorescent microscopy.
Western blotting
Cells were lysed with freshly prepared ice cold lysis buffer [20 mM Tris-HCl, pH 7.6, 420 mM NaCl, 0.25% NP40, 2 mM phenylmethylsulfonyl fluoride, 1 ug/ml leupeptin, 250 U/ml Trasylol (aprotinin)] and stored at -80°C or used immediately. Protein concentration was determined with the BioRad Protein Detection Reagent (BioRad). Tissue samples were mechanically homogenized on ice, spun down at 3000 g for 30 min at 4°C and the lower (non lipid) phase was used for further analysis.
Equal amounts of protein were separated on 10% SDS polyacrylamide gels and either Coomassie blue or silver stained according to established techniques [
27]. Following electrophoresis, the proteins in the gel were transferred to a nitrocellulose membrane using a variation of the methods of Towbin [
28] and Burnette [
29], and following blocking, probed with relevant primary and HRP-conjugated secondary antibody. Membranes were processed using an enhanced chemiluminescence kit (ECL, Amersham), and visualized on Kodak BioMax MR-1 film. The immunoblotting with recombinant GIPC1 protein was carried out as previously described [
26].
Binding of 125I-labeled monoclonal antibody to SK-BR-3 cells and Scatchard analysis
Antibody 27.F7 was labeled with Na-
125I (specific activity 17.4 mCi/mg) (New England Nuclear, MA, USA) using Iodogen as previously described [
30]. The resulting specific activity of
125I-27.F7 was (100 mCi/mmol). SK-BR-3 cells were grown in 24-well plates supplemented with DMEM media with 10% FCS and used in these experiments at subconfluent phase at a density of 2 × 10
5 cells per well. Cell monolayers were suspended with trypsin, cooled to 4°C by placing them on ice and washed twice with PBS containing 1% BSA (Sigma, USA). The cells (50,000 cells per sample) were blocked with 1% BSA-PBS for 1 hour at 4°C followed by incubation with
125I-27.F7 (approx. 10
5 cpm per sample) in the presence of increasing concentrations of cold unlabeled 27.F7 (ranging 0.1 – 200 ng/ml) for 1 hour at 4°C. After incubation the samples were applied to Millipore filters using Millipore 96-well membrane plates (Millipore, USA). The wells were broken off and counted individually in a Cobra γ-counter (Hewlett Packard, USA). Maximum binding, B
max, was determined by incubating varying numbers of cells (ranging from 1.25 × 10
4 to 32 × 10
5 cells) with radiolabeled 27.F7 antibody. Nonspecific binding of the tracer was determined in the presence of an excessive amount of unlabeled antibody (500 μg/ml) and was generally less than 5% of maximum binding. All experimental measurements of K
a and the number of antigen targets per cell were done in triplicate. Analysis of the data was performed according to previously described methods [
31‐
33].
Antigen identification
RNA was purified from SK-BR-3 cells according to standard protocols. Preparation of mRNA was performed by oligo dT chromatography with a kit from Stratagene (La Jolla, CA). A lambda phage expression library was prepared in lambda ZAP (Stratagene, La Jolla, CA) and plaque lifts were screened with human monoclonal 27.B1 and 27.F7 according to previously described methods [
21,
34]. Phage plaques that were positive on the first screened were picked and two rounds of plaque purification was performed to ensure that they were true positives. Positive lambda phage clones were autoexcised according to the protocol provided by Stratagene, grown as plasmids according to standard protocols and sequenced to identify the cDNA inserts.
Northern blot analysis
Total cellular RNA was isolated by the Guanidinium/Phenol extraction method and Northern blotting was performed as previously described [
35,
36]. Briefly, 15 μg of RNA is denatured and electrophoresed in a 1.2% Agarose gel along with 3.5% formaldehyde, transferred to a nylon membrane and hybridized sequentially with
32P-labeled cDNA probes. The GIPC1 cDNA fragment that we isolated using SEREX technology was used for the gene specific probe and a cDNA fragment of the GAPDH gene was used as an internal control to normalize expression. Following hybridization, the filters were washed and exposed for autoradiography.
Discussion and conclusion
Our studies have determined that GIPC1 is a novel breast cancer associated antigen that is up-regulated in breast cancer cell lines and in malignant tissue from all breast cancer patients tested in this study. In our previous studies [
25], we determined that the reactivity of the human monoclonal antibodies 27.B1 and 27.F7 are specific to breast cancer tissue and cells. Our present study demonstrated that these antibodies identify the unique antigen target GIPC1 (GAIP interacting protein, C domain), a PDZ domain containing protein [
38,
39].
GIPC1 was first described as a PDZ domain protein that binds to the C terminus of RGS-GAIP (for regulators of G signaling – Gα
i3 interacting protein) [
38]. GAIP itself is considered to be a Gα
i3 regulator, which acts as GTP-ase activating protein switching Gα
i3 into an inactive mode [
40‐
45]. Although the functional pathway of Gα
i3 is apparently vesicular trafficking, with GAIP serving as its regulator, the physiological relevance of the interaction between GIPC1 and GAIP is been actively investigated [
46‐
53].
With respect to breast tissue, the recognition by 27.B1 and 27.F7 of GIPC1 is strictly limited to neoplastic cells. Furthermore, the subcellular localization of the bound antibodies in the cell membrane and cytoplasm is consistent with what has been previously described for GIPC1 localization [
46]. Taken together, our findings suggest that the up-regulation of GIPC1 is cancer cell specific. Although 27.B1 does not detect any protein by FACS or Western blot analysis in normal cells and tissues it is likely below the level of detection since GIPC1 is likely involved in many diverse cellular processes [
47,
49‐
55].
GIPC1 plays a role in mediating the assemblage of molecules involved in signaling transduction pathways [
53]. As such, these molecules are involved in protein-protein interactions and likely modulate the activity of their targets [
53]. Proteins containing PDZ domains, and the interactions that they mediate, may be involved in a wide variety of signal transduction cascades including interaction with receptors, adhesion molecules, ion channels, gap junctions, cytoskeleton proteins and other vital proteins, such as Fas [
48‐
54,
56]. Moreover, GIPC1 appears to be a highly conserved protein. In rodents it regulates distribution of M-Sem-F, a neuronal membrane-associated protein and binds to a glucose transporter protein, GLUT1 [
57,
58]. It is tempting to speculate on the role of over-expressed GIPC1 in binding to a glucose transporter protein with the subsequent influx of glucose supporting growth of tumor cells. Of course, other proteins may be linked to GIPC1 function in cancer cells; this is currently under investigation in our laboratory.
Research of cancer-associated antigens is an extremely important pursuit. Identification of these antigens can provide insight into the cause of a malignancy, identify targets for immunotherapy and immunodiagnostics [
26] as well as lead to the development of new cancer vaccines. Previous studies on natural monoclonal autoantibodies from cancer patients did not describe the target antigens [
59,
60]. Our studies demonstrate that by combining the "immunoprospecting" of cancer autoantibodies and SEREX technology discovery of target antigens for monoclonal cancer autoantibodies can be accomplished. In conclusion, our studies revealed that GIPC1 is a novel cancer-associated antigen; its role in carcinogenesis, however, needs further clarification. It also needs to be clarified whether GIPC1 is a specific breast cancer-associated antigen or it is overexpressed in other malignant diseases as well.
Acknowledgements
We would like to thank Marina Tashker and Ekaterina Hahiashvili for excellent technical assistance.
This project was supported by American Society Grant and the Department of Medicine at Columbia College of Physicians and Surgeons (SR, GK, AE, IT); the grant form Ludwig Cancer Institute (MS and LO); The Morningside Foundation (GC) and Research Development Grant from BGU of Beer Sheva, Israel (VY, CL and LL).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
SR carried immunochemical and biochemical studies and organized all the data for the manuscript; MS carried the molecular biology experiments and performed SEREX for identification of the antigen; GK developed hybridoma clones producing specific monoclonal antibody; VY performed flow cytometry and Western blot studies for confirmation of the antigen identity; CL did cell culture work related to cloning and selection of antibody-producing clones; AE, LO and GLC provided an expert clinical information on breast cancer and contributed to the interpretation of data and consideration of potential applications; LL and IT are senior co-investigators who conceived the study, developed its design participating in its coordination and drafting the manuscript. All authors read and approved the final manuscript.