Background
We previously described the isolation and characterization of a large panel of fully human monoclonal antibodies from patients with breast cancer [
1]. Many of these antibodies are highly sensitive and specific for breast cancer and some also demonstrate high sensitivity and specificity for non-autologous malignancies of different types. The antigen target of two of these antibodies, 27.F7 and 27.B1 is the protein GIPC1, which is a member of a family of PDZ-domain conserved proteins.
GIPC1 is a carboxy-terminal GAIP interacting protein and together they are components of a G-protein-coupled signaling complex thought to be involved in vesicular trafficking. The PDZ domain of the GIPC family proteins interacts with C terminal regions of FZD3, IGF1 receptor, TrkA, TGF-β RIII, integrin α6A, 5T4 and RGS19 [
2]. Thus GIPC1, like other PDZ domain-containing proteins, may function to cluster signaling molecules and membrane receptors in specific membrane microdomains [
3]. Because RGS19 is a member of the RGS family that regulates heterotrimeric G-protein signaling, the GIPC1 family of proteins might function as scaffolds linking heterotrimeric G-proteins to receptor tyrosine kinases.
It is also known that GIPC1 not only interacts with TGF-β type III receptor (TGF-β RIII) [
4], but also induces its increased expression on the cell surface, leading to an enhanced responsiveness to TGFβ. Down-regulation of GIPC1 mRNA in tumors might promote cellular proliferation through interference of TGFβ signaling [
5]. On the other hand, Awan et.al. suggested a metastatic role for GIPC1 protein demonstrating its close interaction with 5T4 protein, which has a great impact on the actin cytoskeleton and cell migration [
6]. In addition, GIPC1 was shown to interact with alpha-actinin-1 [
7], which is important for stabilizing actin bundles. It was also shown to be involved with cell adhesion through its close link with E-cadherin in epithelial cells [
8]. Therefore, GIPC1 might play key roles in carcinogenesis and embryogenesis through modulation of growth factor signaling
In our previous study antibodies 27.F7 and 27.B1 were studied using immunofluorescence on breast cancer specimens. They were highly specific for breast cancer and did not stain normal breast tissue. To provide a more in depth analysis of these antibodies and to further clarify the association of GIPC1 with different types of breast cancer, we carefully studied 27.F7 and 27.B1 antibodies via immunohistochemical analysis of breast cancer tissue. In addition, we determined that these antibodies stained the ovarian cancer cell line SKOV-3 quite strongly and as a result we also performed a similar analysis on serous carcinoma of the ovary, the most common and aggressive type of ovarian malignancy.
Breast cancer claims the lives of many women yearly. Recently, there have been improvements in breast cancer treatment and survival, which has rested to a large extent on detection and treatment of early stage disease [
9‐
11]. Although mammography has been quite successful in detection of breast cancer, there are still many women that die as a result of identification of the malignancy at a late stage [
10‐
13]. Ovarian cancer, on the other hand, in particular epithelial carcinoma of the ovary is the leading cause of death from gynecologic cancers in the United States, and is the fifth leading cause of cancer death among U.S. women. It usually occurs in women over the age of 35, with most affected women being above the age of 50 [
14].
Approximately 5% to 10% of ovarian cancers are familial and in most families affected with breast and ovarian cancer a genetic linkage has been established with the
BRCA1 locus.
BRCA2, is also responsible for some instances of inherited ovarian and breast cancer [
9]. Although treatment modalities for ovarian cancer are lacking, survival in most cases seems to depend on early detection and treatment [
14]. One of the principal reasons for such a high mortality is the lack of effective and reliable methods for early diagnosis of the disease [
15]. Because ovarian cancer is often asymptomatic in its early stages, most patients have widespread disease at the time of diagnosis. Partly as a result of this, yearly mortality in ovarian cancer is approximately 65% of the incidence rate [
16]. In addition, most diagnostic tests including manual examination and transvaginal ultrasound cannot reliably predict early onset of malignancy [
17]. For early cancer screening physicians have relied, for the most part, on a variety of cancer markers such as CA-125. Many of these, however, are not reliable, as evidenced by the fallibility of CA-125 in predicting ovarian cancer, since many women with low values have malignancy [
12,
18].
As early detection of ovarian cancer is a key factor for long-term survival, it is an imperative that new markers for early onset ovarian cancer be developed. Since we identified GIPC1 as a new tumor-associated antigen that is linked to breast cancer and that stains ovarian cancer cells in vitro, we reasoned that it might be useful as a new marker for these malignancies. Therefore, we extended our studies of GIPC1 in breast and ovarian cancer to determine the utility of this cancer-associated antigen as a marker for malignancy. In particular, a study of GIPC1 protein expression in malignant cells of ovarian tumors might provide new avenues for the diagnosis, prognosis and treatment of ovarian cancer, and might constitute the basis for further study of the role that this protein might play in malignant disease.
Methods
Cell culture and antibodies
Human breast cancer cells MCF-7, ovarian cancer cells SKOV3 and human primary fibroblasts were obtained from the ATCC and maintained according to the supplier's instructions. Hybridoma cell line, secreting 27.F7 and 27.B1 antibodies, were produced as a result of fusion between the MFP-2 fusion partner cell line and human lymphocytes derived from lymph node lymphocytes of breast cancer patients [
1]. They were maintained in 24-well plates in RPMI 1640 media (Gibco), supplemented with 10% Fetal Calf Serum, L-glutamine, non-essential amino acids, pyruvate, vitamins and hypoxanthine and thymidine (HT).
Human anti-GIPC1 monoclonal antibodies 27.B1 and 27.F7 are both IgM,κ, (Rudchenko et al., manuscript in press BMC cancer), and were produced by growing hybridoma cells and used as culture media supernatant with defined antibody concentration. For some experiments antibodies were further purifed by gel filtration chromatography on a 1 meter × 2.5 cm column packed with Sephacryl S300 in sodium phosphate buffer (pH 7.8), 300 mM NaCl at a flow rate of 6 mL/minute. Normal goat serum used for blocking was purchased from Sigma (G9023). Monovalent Fab Fragment Goat Anti-Human IgM+IgG, used for secondary blocking of light chains, was purchased from Jackson Immunoresearch Laboratories, Inc. (109-006044). ELISA capturing antibody, for determination of 27.B1 and 27.F7 antibody concentrations, goat anti-human IgM, Fc-specific, was purchased from Jackson Immunoresearch Laboratories (109007043). Peroxidase labeled goat anti-human IgM for ELISA was purchased from Sigma (A-8400). Secondary goat anti-Human kappa light chain FITC conjugated was purchased from Enco Scientific Services, (2060-02), and biotinylated goat anti-human kappa light chain, affinity purified, was purchased from Vector Laboratories, Inc. (BA-3060, P1216).
Human tissues
The paraffin-embedded samples of biopsies of normal ovaries, ovarian and breast tumor tissues were obtained from the archives of the Institute of Pathology of Soroka Medical Center, Israel. The research was approved by the Institutional Ethics Committee. Classification of tumor samples according to clinical staging, differentiation state and tumor type was performed by the experienced pathologists from The Institute of Pathology, Soroka Medical Center.
Cellular ELISA (cELISA)
Preparation of fixed target cells for cELISA antibody screening
Upon reaching 80–90% confluency, the target cells were detached from the plastic dishes by incubation with 0.02% ethylenediaminetetraacetic acid (EDTA) and 0.25% trypsin for 5 min at 37°C. After a few washes with phosphate-buffered saline (PBS) the cells were exposed to 4% formaldehyde for 15 min at room temperature and then stored at 4°C for up to 4 weeks.
cELISA on fixed cells
Millipore suction plates were blocked with 0.3% dry milk in PBS for 1 h at room temperature. Fixed cells were applied to each well at 5 × 103–104 cells per well. Before application the fixed cells were washed and resuspended in 0.1% Tween-20/PBS (to perforate cell membranes) and incubated for 30 min at 37°C. All the following steps were performed in the presence of 0.3% milk in PBS. After applying cells to the wells and washing them at least twice by applying vacuum to the plate, antibodies were applied to the wells and incubated with the cells for 2 h at room temperature. Following removal of the antibodies and subsequent washes, secondary antibodies conjugated with HRP were applied. Conjugates used were Goat anti-Human Ig Kappa (27.F7 and 27.B1 are IgM κ). After 30 min of incubation followed by several washings, the orthophenylendiamine (OPD) substrate was added and the color intensity at 492 nm was recorded.
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). Equal amounts of protein were separated on 10% SDS polyacrylamide gels, transferred to a nitrocellulose membrane and 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.
Reactivity of 27.F7 and 27.B1 monoclonal antibodies against SK-BR-3 cell line by flow cytometry
Purified 27.B1 and 27.F7 antibodies were tested against the representative tumor cell line to determine the reactivity by flow cytometry. Briefly, SK-BR-3 cells (0.9 × 106/300 μL) were incubated with purified antibodies (27.B1 and 27.F7) or control human myeloma IgM at 50 μg/mL for 2 hours on ice. After incubation, the cells were washed with PBS-5% FBS and incubated with biotin-conjugated anti-human-IgM (Pierce cat #31778, diluted 1:100) for 1 hour on ice. The cells were washed with PBS-5% FBS, followed by incubation with Streptavidin-Cy-Chrome (Pharmingen cat# 13038A, diluted 1:120) for 30 minutes on ice. Finally, the cells were washed and resuspended in 0.5 mL of buffer containing propidium iodide (Molecular Probes cat# P-1304) at 0.6 μg/mL. Tumor cell binding was determined using a FACSCalibur. Antibodies were considered positive if antibody-treated tumor cells exhibited a positive shift in fluorescence of 30% or more of the cell population as compared to the negative control.
Immunocytofluorescence
Microscope slides were treated with 1 M hydrochloric acid (HCl) to obtain proper cell adhesion, rinsed thoroughly with DI water and autoclaved. The SKOV-3, MCF-7 and SKBR-3 cell lines were trypsinized using a 0.25% trypsin solution, transferred to the slides and incubated overnight in growth media. After the incubation, immunofluorescent staining of the cells was performed, using primary 27.F7 and 27.B1 human antibodies for GIPC1 detection. SKOV-3, SKBR-3, MCF-7 cells were stained according to the following immunofluorescence staining protocol. Briefly, slides were washed with PBS and blocked using blocking solution (5% normal goat serum) (Sigma, G9023) in PBS. After a series of washings (3 times in PBS, 5 minutes each), the slides were covered with 2 micrograms/ml of primary human monoclonal antibody 27.F7 or 27.B1 (IgM, k), for 1 hour at room temperature in a humid chamber. After the incubation, the slides were washed in PBS, blocked as above, and incubated with a secondary anti-human kappa light chain FITC conjugated antibodies (Enco Scientific Services, 2060-02) for 30 minutes. The slides were washed several times, drained, mounted with mounting medium (Biomedia, M01) and coverslips were applied. Cell staining was then examined by fluorescence microscopy.
Antigen blocking
GIPC1 protein was expressed in bacteria with a 6× histidine tag, purified in denatured form by NTA resin chromatography (Qiagen), and refolded as previously described [
19]. GIPC1 protein (15 micrograms/mL final concentration in blocking solution) was preincubated with 27.F7/27.B1 antibody (1.5 micrograms/mL final concentration in blocking solution) for 1 hr at RT. Samples of breast malignant tissue and an ovarian malignant tumor, which were previously found positive for GIPC1 staining, were utilized for this analysis. Normal ovary and normal breast tissue served as negative controls. The staining procedure was performed according to the immunohistochemistry protocol outlined below, except that instead of applying primary antibody alone a mixture of antibody preincubated with GIPC1 protein was applied to the slides during staining. The results were examined by light microscopy.
Immunohistochemistry
Five μm sections were obtained from paraffin blocks. 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. The slides were washed 3 times in PBS and incubated with primary human monoclonal antibody 27.F7 or 27.B1 in blocking solution. The slides were then washed and incubated with a secondary antibody. The next steps in the staining procedure were performed using VECTASTAIN® ABC KIT (Standard) (Vector Laboratories Inc.) according to the recommended protocol. The slides were then incubated with the Diaminobenzidine peroxidase substrate (DAB) (Sigma FAST™ 3,3 Diaminobenzidine tablet sets). Mayer's Hematoxylin was applied for nuclear counterstaining. The slides were subsequently washed several times, drained, dehydrated, and mounted with Eukitt® quick-hardening mounting medium and covered with coverslips. For negative controls, the primary antibody was replaced by phosphate-buffered saline in each set of staining.
Immunohistochemical analysis
The sections were reviewed by two pathologists (R.S.L and B.D). Extent and intensity of staining were evaluated. Extent assessed roughly how much of the pertinent area in the tissue was stained and was scored as percentages: 0%, 10%, 20%, 30%, etc. up to 100% immunoreactive epithelial cells. Intensity assessed the strength of staining and was scored as 1+, weak; 2+, moderate; and 3+, strong staining. Only moderate and strong staining, exceeding the background staining, observed in ≥ 10% of the section was considered positive. Such staining was not observed in the non-neoplastic tissue in tumor sections.
Statistical analysis
Statistical analysis using Fisher one-tail and two-tail tests for two independent samples was performed, testing the significance of the difference between 27.B1 and 27.F7 antibodies in their ability to detect positive GIPC1 cases in different malignant and benign tumors of the ovary and breast. This test was also used to make comparisons between different tumors with respect to the frequency of positive cases detected by these antibodies. The P value indicates significance at a value of < 0.05.
Discussion
In this pilot study we determined that the reactivity of fully human monoclonal antibodies (fhMAbs) 27.B1 and 27.F7, derived from lymph node B-cells of a breast cancer patient and targeting the PDZ domain-containing protein known as GIPC1, are specific to breast and ovarian cancer. Through exploration of the interaction of 27.F7 and 27.B1 autoantibodies with breast cancer cell lines SKBR-3, MCF-7 and ovarian cancer cell line SKOV-3, we found a very strong signal in staining of breast, ovarian and pancreas cancer cell lines, but not in normal controls. The results of these experiments reveal that these two anti-GIPC1 human monoclonal antibodies bind to cancer cells specifically but are not limited to binding a single malignant cell type. Instead they have broader cancer reactivity than we previously thought. Moreover, flow cytometry has demonstrated that the reactivity of 27.F7 antibody with the SKBR-3 cell surface is significantly stronger than that of 27.B1. Furthermore, we have also demonstrated that these antibodies not only bind to the surface of cancer cells, but are internalized, in contrast to control immunoglobulin. As such, these antibodies may impact directly on GIPC1 intracellularly, and therefore may affect a variety of different cancer-associated signaling pathways. Therefore, these antibodies may be useful not only diagnostically but also therapeutically in the future.
These results provided the impetus for our present immunohistochemical studies, and further examination of 27.B1 and 27.F7 human monoclonal antibodies can provide a base for their potential future applications in immunohistochemical research, cancer screening and diagnostics.
We also studied the interactions of 27.F7 and 27.B1 antibodies with human breast and ovarian tumor tissue under the working hypothesis that these two human monoclonal antibodies target different GIPC1 epitopes. The immunohistochemical studies of 27.F7 and 27.B1 have also enabled a comparison of GIPC1 levels in different types of breast and ovarian tumors. Our results clearly demonstrate elevated levels of GIPC1 in malignant, but not benign breast tumors. Therefore, we suppose that this protein is cancer-associated, and might play a role in a cancer-associated process. Another potential explanation that might be suggested is that malignant processes taking place in the proliferating cells require an increased amount of GIPC1 protein, enabling its detection in immunohistochemical staining.
Analysis of ovarian tumors with 27.F7 and 27.B1 antibodies demonstrated differential staining between them, which was statistically significant for borderline ovarian tumors, displaying exclusive binding of 27.B1, but not 27.F7 antibody.
Based on all of our immunohistochemical findings we hypothesize that the antibodies target different epitopes. This is also supported by the fact that they display differential binding upon examination of invasive malignant tumors, and also with respect to borderline ovarian tumors. Since our immunohistochemical results were obtained in a pilot study, this finding requires further investigation on a larger sampling to clarify if this might be used as cancer diagnostic and prognostic tool.
Benign, borderline and malignant conditions are different in terms of their gross morphology and fine structure and are expected to be associated with different cellular processes, perhaps leading to different protein distribution and compartmentalization. The conformation of a cellular protein might also change according to its function at a given moment and in a given tissue implying different epitope accessibility, which might explain in part the differential staining obtained by 27.B1 and 27.F7 antibodies. Nonetheless, the reason for such a difference between the sensitivities of these antibodies to the GIPC1 protein in the tissues that were examined still remains unclear. Recently, it has been shown that GIPC1 can form multimers by binding to itself and that the PDZ domain is involved in the GIPC-GIPC interaction. Furthermore, it was shown that whereas the bulk of cytosolic GIPC1 was present as monomer, GIPC1 homotrimer was readily detectable in the membrane fraction [
21]. These results support our findings and help explain the differential tissue staining obtained by 27.B1 and 27.F7 antibodies.
The results obtained from the examination of benign ovarian cystadenoma, borderline ovarian tumor and invasive ovarian serous carcinoma demonstrate a direct correlation between tumor malignancy and staining. It can be clearly seen that the highest number of GIPC1 positive cases was found for epithelial serous carcinoma – a malignant tumor of the ovary. The behaviour of borderline ovarian tumors is uncertain; they usually behave in a benign fashion, but they have a potential for recurrences in the form of peritoneal implants or even as a metastatic disease. With this in mind, along with other results, we developed a few hypotheses regarding the possible cause for significantly elevated levels of GIPC1 detected in malignant and borderline ovarian tumors. We propose that GIPC1 overproduction is not a direct cause of cancer, as mentioned above, but is rather a byproduct of the cellular response to the transformed state. In addition, it could be evidence for an attempt by the cell to balance and regulate itself.
Regardless of the role that GIPC1 might play in carcinogenesis, it is clearly a novel cancer-associated antigen, suggesting that further research should be conducted in order to define its role in cell transformation and cancer development. To further support the correlation between GIPC1 expression and clinical pathologies, a full-scale study is being performed. Our study, however, forms the basis for further broad-based research of the differential interaction of the fully human 27.B1 and 27.F7 autoantibodies with GIPC1 antigen in different malignant and benign tumors, which will likely be useful for research and diagnostics. Further investigations of the role these antibodies play in the progression of cancer, and the role of GIPC in cancer, is therefore warranted.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
VY and AT performed many of the experiments and prepared much of this manuscript. SR, TA and HK performed immunocytochemcial and immunohistochemcial experiments. RS–L and BD are pathologists at Soroka Hospital who performed immunohistochemical analysis and chose tissue blocks for these studies. AR and BP provided patient specimens and helped prepare the manuscipt. IT and LL initiated these studies, designed most of the experiments and wrote this manuscript.