Background
Integrins are a family of glycosylated, heterodimeric transmembrane adhesion receptors that mediate cellular attachment to the extracellular matrix and to adjacent cells [
1]. They have a widespread distribution in cells and tissues, as they are involved in morphogenetic cell movements and migration, such as in gastrulation, neurulation and histogenesis [
2], as well as in inflammation, wound healing and thrombotic events [
3,
4]. Upon cellular interaction with extracellular matrix, integrin receptors function as bidirectional transducers of extra- and intracellular signals implicating in the regulation of immediate gene expression, cell proliferation, differentiation, survival/anoikis and angiogenesis [
5‐
8].
Furthermore, it is well established that integrins, due to aberrant adhesive events and cellular signals that alter gene expression and influence cell survival, contribute to various cancer stages, such as malignant transformation, tumor growth and progression, invasion and metastasis [
9‐
11]. In cancer growth, both quantitative and qualitative alterations in integrin expression have been observed. Some integrins are overexpressed or no longer expressed, while others become phosphorylated, affecting their cytoskeletal and extracellular ligand binding properties [
12].
Integrin alphavbeta3 – a vitronectin receptor – has been implicated in the pathophysiology and progression of several malignant tumors, such as melanoma [
13,
14], glioma [
15], ovarian [
16], prostate [
17] and breast cancer [
18,
19]. Specifically, nearly in all breast cancer tumors with a metastasis to bone, integrin alphavbeta3 was highly expressed [
20]. In vitro studies of breast cancer cultures have supported the positive correlation of the increased expression of alphavbeta3 with the ability of the cancer cells to adhere to extracellular matrix, to migrate, to regulate protease maturation [
21], as well as to interact and form aggregates with platelets, contributing to breast tumor cell adhesion to the subendothelial matrix under dynamic blood flow conditions [
22,
23]. Thus, the presence of integrin alphavbeta3 on breast cancer tumors signals the onset of widespread metastasis [
11,
24].
Integrins, being cell-substrate adhesion molecules, provide the physical and essential link between the actin cytoskeleton and the extracellular matrix during cell migration [
25,
26]. This connection is dynamically reorganized in response to mechanical, chemokine and growth factor signals, resulting in the continuous growth and reorganization of actin filaments in protruding organelles at the front of migrating cells, such as filopodia and lamellipodia, and in parallel, in the controlled retraction of adhesive contacts at the rear of these cells [
27‐
29]. The regions of the plasma membrane where integrins connect actin cytoskeleton to the extracellular matrix, through various structural and regulatory adaptor proteins, are specialized adhesive structures called
focal adhesions or
focal contacts [
30]. At focal adhesions, actin filaments are anchored in the form of bundles, termed stress fibers.
Ligand binding to integrins leads to integrin clustering and recruitment of actin filaments and signaling proteins to the cytoplasmic domain of integrins [
31]. Concerning the integrin alphavbeta3, the focal adhesion formation begins upon the presence of the extracellular ligand that triggers the activation and clustering of the integrin.
Given that the behavior of any given tumor is a complex interaction between tumor cells and the host environment, cell culture models are the first step toward characterizing molecules that influence tumor cell behavior in vivo [
24]. Established breast cancer cell lines have been widely used to provide meaningful data for evaluating the pathobiology of this kind of cancer, as the cultured cells are easy to handle and can be grown in almost infinite quantities with a high degree of homogeneity. However, there is a fundamental disadvantage: Cell lines are prone to genotypic and phenotypic alterations during their continual culture, resulting in a drift away from the phenotype of the originating tumor in vivo. On the contrary, primary cultures derived directly from tumors have the advantage that the cultured cells are directly isolated from the tumor site and their phenotype closely represents the pathobiology of the original tumor in vivo, as cancer cells are cultured for a finite length of time and they have little opportunity to undergo genotypic transformations [
32]. Although there is correlative evidence supporting the role of tumor-specific alphavbeta3 expression in breast cancer progression [
20], its precise contribution to primary tumor growth and metastasis is still unclear [
11].
In the present work, for the first time, the correlation of integrin alphavbeta3 expression with actin cytoskeleton distribution was studied in primary cell cultures of human breast cancer biopsies using immunofluorescent localization of integrin alphavbeta3 combined with phalloidin-rhodamine fluorescence for F-actin visualization. The relative integrin alphavbeta3 fluorescence intensities at the periphery and the main body of the cancer cells were semi-quantitatively assessed with the aid of a computerized image analysis system. A qualitative evaluation of integrin alphavbeta3 localization in correlation with the distribution of F-actin bundles in breast cancer cells, was also undertaken ultrastructurally, while quantification of integrin and actin expression was evaluated by Western immunoblotting. For comparison, the same procedure was carried out in primary cultures of breast cells developed from biopsies of normal tissues adjacent to the tumor. The results showed a shift of integrin alphavbeta3 immunofluorescence from the main cell body – particularly at the perinuclear area, near the ventral cell surface – in the primary normal breast epithelial cells to the marginal area in the primary breast cancer epithelial cells. In parallel, in the latter, F-actin cytoskeleton appeared well-formed, consisting of numerous and thicker stress fibers. At the ultrastructural level, increased integrin alphavbeta3 immunogold localization was observed in epithelial breast cancer cells over the area of stress fibers at the basal cell surface. These findings were also verified with Western immunoblotting by the higher expression of integrin beta3 subunit and actin in primary breast cancer cells. Given the above, the primary breast cancer cell cultures developed without any enzymatic treatment – to retain the integrin alphavbeta3 status intact – consist a model system that can be successfully used for the study of breast tumor cell behaviour, concerning the adhesion capacity and the migrating potential of these cells.
Discussion
In the present study, for the first time, integrin alphavbeta3 localization was correlated to the F-actin cytoskeleton formation in primary cell cultures of human breast cancer biopsies in comparison to normal tissue. By means of immunofluorescence, a shift of integrin alphavbeta3 localization was observed from the main cell body – particularly around the nucleus at the ventral cell surface of primary normal breast epithelial cells – to the marginal area of primary breast cancer epithelial cells. In parallel, in the latter, F-actin cytoskeleton appeared well-formed, consisting of numerous thicker stress fibers, compared to the F-actin cytoskeleton in normal epithelial cells. Furthermore, by means of electron microscopy, increased integrin alphavbeta3 immunogold localization was observed in epithelial breast cancer cells over the area of stress fibers at the basal cell surface, in comparison to normal breast epithelial cells. These findings were also verified by the higher expression of integrin beta3 subunit in primary breast cancer cells, as shown by Western immunoblotting.
Many preclinical studies have been carried out using established breast cancer cell lines, due to their easy manipulation, the infinite cell quantities and the high degree of homogeneity [
32,
33]. However, cell lines frequently have undergone multiple genotypic and phenotypic changes, resulting in a drift away from the biological behavior of the primary tumor of origin. On the contrary, short-term primary cell cultures of epithelial breast cancer cells, despite the culturing difficulties, consist an experimental model system with a greater biological and clinical relevance [
32,
34]. In the present study, primary cell cultures of breast cancer and normal tissues were developed using the explant technique, in which small fragments of minced tissue were allowed to attach onto substrate without any previous trypsinization, in order to preserve the integrin alphavbeta3 status of the cells [
35]. From the adhered tissue fragments, a mixed cell population consisted of epithelial cells and stromal fibroblasts, were grown out. Although the presence of fibroblasts is considered by many investigators as "contamination" due to their high rate of proliferation, resulting in outgrowing the cancer epithelial cells, the culture conditions which were used in this study, allowed the controlled growth of a low number of fibroblast-like cells. Actually, their presence in the developed primary breast cancer cell cultures was considered essential, resembling more closely the in vivo environment of the tumor tissue, since tumor epithelial cells within the cancerous mammary gland are in direct contact with the highly activated collagenous tumor stroma, communicating with the fibroblasts and interacting with the extracellular matrix, where integrins are implicated [
36,
37]. Furthermore, according to recent studies, fibroblasts, except their fundamental functions, such as deposition of extracellular matrix, regulation of epithelial differentiation, regulation of inflammation [
38], have also a prominent role in defining the rate and extent of cancer progression [
37,
39]. Hence, the model system of primary breast cancer cell culture that was developed, in the present study, retains the most closely resembling environment to the tumor tissues.
Integrins provide the physical link between the actin cytoskeleton and the extracellular matrix [
25] contributing to the formation of the focal adhesion sites [
40]. Specifically, integrin alphavbeta3 expression provides broad adhesive capacity of the cells, mediating the cell spreading and migration [
24,
41]. It behaves as an allosteric switch between a freely diffusing non-activated plasma membrane protein and an activated bivalent linker protein, binding simultaneously to extracellular ligands and cytoplasmic adaptor proteins [
40]. Upon extracellular ligand binding, integrin alphavbeta3 is activated, triggering its clustering, which is required for the formation of focal adhesion sites [
40,
42].
In the present study, the integrin alphavbeta3 immunofluorescence in the primary epithelial cells of the normal breast tissues showed a slightly diffused pattern at the ventral surface of the cells, representing the non-clustered non-activated state of the integrin [
40], freely diffused in the plasma membrane. In addition, integrin alphavbeta3 was evenly localized around the nucleus, at the main cell body of many cells. According to Cluzel et al. [
40], integrin beta3 subunit clusters were observed underneath the main cell body of the mouse B16F1 melanoma cells with confocal microscope at the level of the ventral cell surface representing de novo-activated integrins, which are precursors of focal adhesions. Concerning the integrin alphavbeta3 trafficking via a "long-loop" of recycling, after its internalization, it proceeds to early endosomes, then to the perinuclear recycling compartment and finally returns to the plasma membrane [
43]. Therefore, in the present study, the integrin alphavbeta3 immunofluorescence at the perinuclear area of the primary normal epithelial cells may represent either de novo accumulations of this integrin or integrins transferred to the basal cell surface after their recycling through the perinuclear compartment. Taking into account that the primary normal breast epithelial cells are not characterized by high motility, this perinuclear localization may indicate a pool of non-recruited heterodimers of integrin alphavbeta3 in the basal cell surface.
In the primary epithelial cells of the breast cancer tissues, the integrin alphavbeta3 immunofluorescence was not observed at the ventral cell surface around the nucleus, suggesting that its recycling was completed via a "short-loop", without passing through the perinuclear recycling compartment, but returning – after its internalization and proceeding to early endosomes – to the plasma membrane [
43], in response to the elevated requirements of integrin alphavbeta3 availability due to the higher cellular motility. On the contrary, integrin alphavbeta3 immunofluorescence was mostly observed as bright aggregations at the cell periphery and particularly at the leading edge of the advancing lamellipodium, contributing to the adherence of the cancer cells to the substrate, forming focal adhesions and mediating their spreading and migration [
24,
26]. The occasional presence of integrin alphavbeta3 immunofluorescence at the rest of the main cell body in the primary breast cancer epithelial cells, as shown by the frequency distribution histogram of the immunofluorescence intensity values, may be, either, due to non-clustered non-activated integrins, where a diffused pattern is observed, or due to de novo-formed clusters of integrin, where immunofluorescent aggregations are displayed.
In parallel, in primary breast cancer epithelial cells the well-formed F-actin cytoskeleton – in comparison with that in normal counterparts – consisted by numerous thicker stress fibers, reveals the dynamic reorganization of actin filaments, reciprocating the increased integrin alphavbeta3 clustering and the formation of focal adhesion structures [
30,
40]. Stress fibers were observed to be oriented towards the boundary between cells, where integrin alphavbeta3 is located, but mostly towards integrin aggregations at the marginal region of the cells, as well as at the leading edge of the advancing lamellipodia, terminating at the focal adhesion structures [
44,
45]. Furthermore, the orientation of F-actin stress fibers to the integrin alphavbeta3 aggregations at the cell periphery may indicate the involvement of stress fibers in the translocation of this integrin also to the focal adhesion sites, as it was proved for integrin beta1 by Kawakami et al. [
44].
The immunogold ultrastructural localization of the integrin alphavbeta3 in correlation with the actin cytoskeleton distribution is performed for the first time in primary breast cancer cell cultures. At the focal adhesion sites at the ventral surface of the cells, an increased integrin alphavbeta3 immunogold labeling was observed over a wider area of well-formed stress fibers arranged in parallel, compared with that of normal cells. The increased integrin alphavbeta3 immunogold labeling, combined with a well-formed actin cytoskeleton, suggests the involvement of integrin alphavbeta3 to better cellular adhesion as part of the enhanced cellular motility process [
24].
The increased integrin alphavbeta3 immunolabeling at the ultrastructural level was quantified in primary breast cancer cells vs. normal ones, performing Western immunoblotting for the integrin beta3 subunit. Integrin beta3 subunit was chosen as it forms heterodimers uniquely with the v and the platelet-specific IIb alpha-chains [
46]. In breast epithelial cells integrin beta3 subunit pairs exclusively with alphav to form the heterodimer alphavbeta3. Thus, the quantitation of beta3 subunit by Western immunoblotting in primary cell cultures of breast biopsies, indirectly equalizes with the quantitation of integrin alphavbeta3 expression. In primary cells of normal breast tissues, using the optimal protein load (40 μg), integrin beta3 was expressed at low no-detectable levels, which is in agreement with the no-detectable levels of integrin alphavbeta3 in MCF-7 low metastatic cell line [
24]. On the contrary, in primary cells of breast cancer tissues, integrin beta3 was expressed at various levels, although the samples of breast cancer biopsies originated from the same histological type and grade, revealing the heterogeneity which characterizes the breast cancer. In one of the samples, integrin beta3 was expressed almost at the same high level, as in the highly metastatic MDA-MB-435 cell line, which was used as a positive control. On the other hand, in Western immunoblotting, the higher expression levels of actin in primary breast cancer cells compared to that in normal breast cells is consistent with the well-formed actin cytoskeleton observed in the former by immunofluorescence, suggesting responsiveness of actin cytoskeleton to the higher motility demands which are determined by the malignant potential of the breast cancer cells.
Methods
Primary cell cultures of breast tissues
Breast tissue biopsies after the operation, were immediately placed into sterilized culture medium consisting of RPMI 1640 supplemented with 10% Foetal Calf Serum (FCS), 1 mM L-glutamine, 0.5 mgr/ml gentamycin, and transported to the tissue culture facility. Small tissue fragments (2–3 mm3) of each biopsy were put on sterilized round glass cover slips, 13 mm in diameter (one tissue fragment/coverslip), which were already placed in plastic flasks 25 cm2. The cells were allowed to migrate out from tissue fragments and to grow up adhering to the glass coverslips, in the culture medium, at 37°C, 5% CO2 in incubator. The cell growth was complete in about 2–3 weeks.
Scanning electron microscopy (SEM)
For SEM, cells – attached on the glass coverslips – were fixed in 2.5% glutaraldehyde in 0.05 M cacodylate-Na buffer, pH 7.4, for 30 min at 37°C and then rinsed in buffer. Dehydration was accomplished by serial immersion of the specimens for 3 min each in 25%, 50%, 70%, 95% ethanol and for 5 min in 100% ethanol twice, at room temperature (RT). The specimens were then infiltrated gradually in a mixture of amyl acetate diluted in 100% ethanol (1:2, 1:1, 2:1), 5 min each at RT and finally in 100% amyl acetate 3 times for 5 min. The specimens were then covered with one drop of hexamethyldisilazane and were dried overnight. Before the observation, they were rendered conductive by sputtering them with gold before being observed by the scanning electron microscope JEOL JSM 4500 operated at 3 kV. Digitized images were saved as high resolution *BMP files and were printed at 1000 dpi on a Xerox HP Laserjet 4250 n high resolution laser printer.
Integrin alphavbeta3 immunofluorescence combined with F-actin phalloidin-rhodamine fluorescence
When cell culture was completed, the cells – attached onto the glass coverslips – were fixed in 3% paraformaldehyde in cytoskeleton buffer, pH 6.8, for 15 min at room temperature (RT). Each one of the coverslips was placed in a well of multiwell Petri dish and washed three times (5 min each) with cytoskeleton buffer. Cells were then incubated in primary monoclonal mouse anti-human integrin alphavbeta3 antibody LM609 (60 μg/ml) (MAB1976, Chemicon) diluted in cytoskeleton buffer, pH 6.8, containing 5% normal goat serum (NGS), overnight at 4°C, without any previous detergent permeabilization. After three washes (5 min each) with cytoskeleton buffer, cells were incubated in the secondary goat anti-mouse IgG FITC-conjugated antibody (1:300) (AP308F, Chemicon) diluted in cytoskeleton buffer, pH 6.8, containing 1% bovine serum albumin (BSA) (A7638, Sigma), for 1 h at RT. Subsequently, cells were washed three times with cytoskeleton buffer (5 min each) and permeabilized with 0.5% (vol/vol) Triton X-100 for 3 min at RT, for the visualization of filamentous actin with the application of the rhodamine-conjugated phalloidin method. After three washes with cytoskeleton buffer (5 min each), cells were incubated with rhodamine-conjugated phalloidin (3 ng/ml) (P1951, Sigma), for 1 h at 37°C. Finally, the cells were washed three times with cytoskeleton buffer and three times more with distilled water. Each glass coverslip was placed upside-down and mounted on a slide in a drop of aqua-poly mount (18606, Polysciences) antifading agent. Cells were observed with a Zeiss Axiovert S-100 inverted photomicroscope equipped with epifluorescence optics and using a 63× oil immersion objective. Photographs were recorded on Kodak 400ASA color film. The negatives were scanned in an Agfa Duoscan T2500 high resolution scanner and when required prints were obtained using a HP color Laserjet 4700 n printer.
Image analysis morphometry
For the assessment of the relative integrin alphavbeta3 fluorescence intensities with image analysis morphometry, at the periphery and the main body of the epithelial breast cancer cells and of the epithelial breast cells of normal tissues, approximately 35 micrographs of cells of each category were taken randomly using the 63× oil immersion objective, with the same identical settings, on a Zeiss Axiovert S-100. The micrograph negatives were scanned with an Agfa DuoScan T2500 scanner at an optical resolution of 1000 ppi (pixels per inch). The digitized images were converted automatically into positive grayscale images (256 grey levels), stored as tiff files and calibrated before use.
The image analysis was carried out using the Image-Pro Plus, version 3.0, software for Windows (Media Cybernetics, Silver Spring, MD, USA) on a Pentium IV, 2.4 GHz, workstation. The estimation of the relative integrin alphavbeta3 densities was made by measuring the integrin fluorescence intensities per defined cellular areas at the cell periphery and the main cellular body with specially developed macros, as follows: a) calibration of the intensity curve per image b) interactive drawing of the cell periphery as it is defined by the outer limit of the cell surface and 1–5 μm inwards, as this distance corresponds to the usual breadth of a lamellipodium [
29], (c) interactive drawing of the main body of the cell as it is defined by the rest cellular area, excluding the nuclear area. All adjustments, including programming and calibrations, were made by a certified (factory approved) expert programmer.
The values of the relative integrin alphavbeta3 fluorescence intensities per cell were automatically recorded in a sheet of Excel. Their statistical assessment between normal and cancer cells was performed using the "frequency distribution" test. The values were divided into 8 classes. The distribution of frequency of values which corresponds to the cell periphery and the main cellular area is displayed comparatively in each histogram concerning the normal and cancer epithelial cells.
Transmission electron microscopy
Tissue preparation
For electron microscopy, cells allowed to migrate from biopsy tissue pieces and to adhere on the plain plastic bottom of the culture flask for 2–3 weeks. They were then fixed in 2.5% glutaraldehyde made up in phosphate buffer saline (PBS) 0.01 M, pH 7.4, for 15 min at RT, post-fixed in 1% aqueous OsO
4 for 30 min at 4°C, isolated according to Kouloukoussa et al. [
48] and finally embedded in a mixture of epoxy resins. Ultrathin (80–90 nm thick) sections were cut with a Diatome diamond knife on an Leica Ultracut R ultramicrotome and were collected onto 200 mesh uncoated nickel grids.
Electron immunocytochemistry
Immunogold labeling of integrin alphavbeta3 was carried out as described previously [
49], modified for single labeling. Briefly, all incubations were carried out by floating the grids on drops of solutions at RT, except for the primary antibody step, using the multiwell Terasaki plates as humid chambers. Before use, all solutions were micro-filtered with millipore filters (0.2 μm pore size) fitted to a 5 ml syringe. Grid washing was carried out by floating them on drops of the washing solutions with simultaneous gentle agitation on a mechanical agitator. Before the immunolabeling procedure, the nickel grids with the osmicated epoxy-resin sections were first incubated on drops of 8% H
2O
2, in a humid chamber for 8 min and then on drops of saturated aqueous solution of NaIO
4 [
50], pH 3.8 for 30 min at RT, in order to remove osmium from the superficial regions of the thin sections to be labeled. Between these two incubations, thorough rinsing with distilled water before immunolabeling was carried out.
Immunogold labeling of integrin alphavbeta3
Grids were first incubated for 4 min (four changes, 1 min each) on drops of 0.05 M Tris/HCl buffer, pH 7.4, and then on blocking buffer [0.05 M Tris/HCl, pH 7.4+0.2% BSA+0.1% coldwater fish gelatin+5% NGS+0.1% Tween-20] for 30 min. Grids were then incubated in the primary monoclonal mouse anti-human integrin alphavbeta3 (10 μg/ml) antibody diluted in 0.05 M Tris/HCl buffer, pH 7.4, containing 1% BSA, overnight at 4°C. Control sections were incubated in the absence of primary antibody. The grids were rinsed, with agitation, in five changes of 0.05 M Tris/HCl, pH 7.4, for 2 min each time, and then they were washed, with agitation, consecutively in five changes (1 min each) of 0.05 M Tris/HCl, pH 7.4 containing 0.1% Tween-20 (solution I), three changes (1 min each) of 0.05 M Tris/HCl, pH 7.2 containing 0.2% BSA and 0.1% Tween-20 (solution II), one change for 5 min of 0.05 M Tris/HCl, pH 8.2 containing 1% BSA and 0.1% Tween-20 (solution III). The grids were incubated for 1 h at RT in the secondary colloidal gold antibody [goat anti-mouse IgG conjugated to 15 nm gold particles (1:20) (25133, EMS)] diluted in solution III and they were washed with agitation in 3 changes (1 min each) of solution II, five changes (1 min each) of solution I and five changes (1 min each) of distilled water. Finally, the grids were dried and counterstained with saturated ethanolic uranyl acetate and lead citrate. Sections were viewed in a Zeiss EM 900 electron microscope, at 80 kV accelerating voltage, with an objective aperture of 30 μm. The electron micrographs were recorded on Kodak Electron Image film SO-163 (gelatin plates, 8 cm × 10 cm, Code No: 1679257). The negatives were scanned with an Agfa DuoScan T2500 scanner using optical resolution of 1000 ppi and stored as tiff files. When required, the digitized images were printed at 1200 dpi on a HP laserjet 4250 n high resolution laser printer.
Western immunoblotting of beta3 integrin subunit
Breast cancer, as well as normal cells from primary cultures were harvested by culture flasks by brief treatment with 0.05% Trypsin-0.53 mM EDTA (GIBCO-Invitrogen), pelleted with centrifugation and lysed in modified RIPA buffer (150 mM NaCl, 50 mM Tris-HCl pH 8.0, 0.1% SDS, 1% sodium deoxycholate, 1% NP-40, 1% Triton) containing 10 μM PMSF and 1:100 Protease inhibitors cocktail (SERVA). For homogenization, cells were left on ice for 1 h while pippeting 4–5 times every 10 min. Cell debris was removed by centrifugation at 15,000 g for 20 min and supernatant was stored. Protein concentration was determined with the BCA method (PIERCE).
After several trials, the optimum amount of loaded protein was determined at 40 μg. Therefore, equal amounts of protein (40 μg) were loaded and resolved on 10% SDS-PAGE, after addition of 4× Laemli buffer (200 mM Tris-HCl pH 6.8, 8% SDS, 40% Glycerol, 20% 2-mercaptoethanol, 0.8% Bromophenol blue). In each gel, 10 μL of prestained Molecular Weight Marker (PageRuler, Fermentas) was loaded. Gels were electroblotted on nitrocellulose membranes (0.45 μm pore size, Amersham) at 20 V overnight and membranes were separated in two parts according to the bands of the Molecular Weight Marker. Membranes were then blocked with 5% nonfat dry milk in TBS, containing 0.05% Tween-20, 0.01% azide, for 2 h at RT. The parts of the membranes containing proteins from 75 to 175 kDa were incubated overnight at 4°C with primary antibody rabbit anti-human integrin beta3 (1:2,000) (AB1932, Chemicon) and the parts of the membranes containing proteins from 20 to 75 kDa were incubated overnight at 4°C with the mouse anti-human actin antibody (1:2,000) (MAB1501, Chemicon). The next day the membranes were washed with TBS, containing 0.05% Tween-20 and 0.01% azide (4 × 15 min). The parts of the membranes incubated with anti-human beta3 were subsequently incubated with secondary antibody goat anti-rabbit IgG-HRP conjugated (1:5,000) (AP307P, Chemicon) and the parts of the membrane incubated with anti-actin were subsequently incubated with rabbit anti-mouse IgG-HRP conjugated (1:3,000) (P0161, Dako), for 1 h at RT. After washing with TBS containing 0.05% Tween-20 and 0.01% azide (4 × 10 min), peroxidase activity was visualized with ECL Plus reagent (Amersham).
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
SH is the principal investigator employed by a post-doctoral fellowship. She made all necessary lab work, cell culturing, morphometry, electron microscopy and photography. MK is a cell culture specialist who coordinated primary culture establishment, growth and maintenance. KA worked on the fluorescent experiments and photography. YD was responsible for cell isolation and Western immunoblotting. LDA is a resident pathologist who examined tissue biopsies. NG is a pathologist, responsible for tissue biopsy delivery and grading. DV is a pathologist responsible for handling tissue biopsies and isolation of cancer lesions. IVB is a cell culture expert who acted as a consultant. VAM is a human genetics specialist who worked on Western immunoblottings and electrophoreses. SDV is a senior surgeon of the hospital responsible for routine surgeries. EZK was involved in designing and performing western immunoblotting experiments. CK is Head of the Department and pathologist who participated in experiment design and helped to draft the manuscript. EM worked as an electron microscopy specialist for the examination of the samples under the electron microscope and coordinated the research group.