Background
Breast cancer is the most common malignancy affecting women in the Western world. Systemic chemotherapy is indicated for women with metastatic breast cancer (MBC) [
1]. Several single-agent and combination chemotherapeutic options have been shown to be effective as first- or second-line therapy in the management of MBC with the taxanes and anthracyclines being the most active drugs [
2], however often with substantial side effects. An important area of cancer research is the search for new effective agents for targeted treatment with as small side effects as possible. Apoptosis, the well-regulated intrinsic suicide program that enables removal of unwanted cells is known to be disabled in many tumour cells and identifying agents that can overcome resistance to apoptosis could greatly improve current chemotherapy options [
3,
4].
Bacterial toxins such as AC-toxin, botulinum toxin, cholera toxin, and verotoxin have been suggested to be used as an approach to establish novel therapeutic agents against tumour malignancies either as independent anti-neoplastic agents or in combination treatment with chemo- or radiotherapy. Toxins effective against specific signalling pathways could reduce treatment side-effects to normal tissues and be an approach to generate specific anti-tumour agents [
5]. VT-1 is a member of the bacterial shigatoxin family expressed by some bacterial serotypes of
Escherichia coli, and
Shigella dysenteriae. Verotoxin-1 induces cytotoxicity, apoptosis, and cell cycle arrest in human tumour cells and has been suggested to be an anti-cancer agent candidate due to its low general toxicity and high specificity against tumours expressing its receptor, globotriasosylceramide (Gb3) [
6].
Gb3 (CD77) has been reported to be increased on the surface of several tumour cells lines originating from breast cancer, ovarian cancer, haematological malignancies, and astrocytoma tumours as well as normal epithelial cells [
7,
8]. The targeting of the toxin to a specific intracellular transport pathway is determined by the Gb3 isoform expressed on the cell surface and by the presence or absence of Gb3 in the lipid raft microdomains of the cell membrane [
9].
The purpose of this study was to investigate the potential of verotoxin-1 as a novel agent for enhancing apoptosis in breast cancer cells by examining the expression of the verotoxin-1 receptor Gb3 in breast cancer tissue followed by elucidation of the specific signalling pathways to cellular proliferation and apoptosis. We investigated the cytotoxicity and induction of apoptosis as well as the signal transduction mechanisms to apoptosis of Escherichia coli verotoxin-1 in human breast cancer cell lines. We demonstrate that verotoxin-1 has the potential to be an effective anticancer drug in Gb3-expressing breast cancer.
Methods
Breast cancer patients
Tumour specimens were collected in the northern healthcare region of Sweden between 1985 and 1989 from 25 unselected women with primary invasive breast carcinoma (11 node-negative and 14 node-positive patients) all with snap-frozen tumour tissue available. Node-negative patients were generally treated with a modified radical mastectomy or conservative surgery, followed by radiotherapy. All node-positive patients were treated with surgery and radiotherapy. Adjuvant systemic treatment was administered to all node-positive patients. Most postmenopausal patients also received adjuvant endocrine therapy. Prognostic and biological information were available for all patients regarding TNM-classification, histological type, oestrogen receptor (ER), progesterone receptor (PgR), and survival time. The median age was 64 years (range 34–83 years) and the median follow-up time was 63 months (last follow-up date was 30 June 1998). Samples was collected before informed consent became mandatory, the study was approved (Dnr 02-455) by the ethics committee of Umeå University.
Gb3 immunohistochemistry
Cryostat sections from biopsies of the breast cancer tumours were cut and post-fixed in acetone. Gb3 immunohistochemistry was performed with a monoclonal rat IgM antibody at a dilution of 1:40 using the Envision system (Dako, Denmark) and the staining evaluated by an experienced pathologist (TB).
Cell lines and cell culture conditions
Two human breast cancer cell lines T47D and MCF-7 were used. The cells were maintained under standard cell culture conditions, grown as monolayer culture in Eagle's MEM with Earl's salt supplemented by 10% foetal bovine serum and 200 μmol/L L-glutamine. They were incubated at 37°C in a humidified atmosphere containing 5% CO2.
Determination and inhibition of Gb3-expression of cultured cells
T47D and MCF-7 cells were screened for expression of Gb3 (CD77), the receptor for VT-1. Cellular expression of Gb3 was identified by a monoclonal rat IgM antibody on a FACSCalibur flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA). DL-threo-1-phenyl-2-palmitoylamino- 3-morpholino-1-propanol (PPMP) a chemical inhibitor of glucosylceramide synthesis was used to quench Gb3 expression. Cells were cultured with 2 μmol/L PPMP for 72 h prior to Gb3 expression analysis.
Cell viability assays
A fluorometric method utilising fluorescein diacetate (FDA) was used to quantify cell viability and determine VT-1 sensitivity of T47D and MCF-7 cells
in vitro [
10]. Cells (1 × 10
4) were cultured and incubated with VT-1 (0.1 – 5 μg/L) followed by fluorescence determination as described earlier [
11]. The same procedure was made with cells pre-treated with 2 μmol/L PPMP for 72 h prior to VT-1 exposure, in which case 2 μmol/L PPMP was present in the medium during the whole VT-1 incubation.
A colorimetric assay (Cytotoxicity detection kit, LDH Roche Applied Science, Mannheim, Germany) was used as described earlier [
12] to measure lactate dehydrogenase (LDH) release following cell membrane damages not seen in apoptotic cells [
13] to quantify necrosis induction by VT-1 (0.1 – 5 μg/L) after 6 h.
Cell cycle analysis with propidium iodide staining measured by flow cytometry
Propidium iodide (PI) is a fluorescence dye that binds specifically to double-stranded nucleic acids [
14,
15]. In the flow cytometry assay employed, PI fluorescence is indicative of the DNA content of the cells. Cells in the G2/M phase are preparing to divide and they contain double amount of DNA (4
n) compared to cells in the G1 phase that have not yet replicated their DNA (2
n DNA content). Upon completion of 72 h incubation with/without 0.01, 0.1 or 1 μg/L VT-1, the cell cultures were washed with PBS and treated with 0.1% trypsin at 37°C. The cell suspension was collected, washed once with PBS (200 g, 5 min), and re-suspended in PBS containing 70% cold absolute ethanol, for fixation and permeabilisation of the cell membrane (1 h, 4°C). After fixation the cells were washed twice with PBS (900 g, 5 min), and the cells were treated with 40 μg/mL RNase in PBS (final volume 100 μL), for 15 min at 37°C. Finally, the PI staining solution (50 μg/mL PI and 3.8 mmol/L sodium citrate in PBS) was added to the cells, followed by 3 h incubation in the dark, at 4°C. The cell cycle analysis was performed by a Fluorescence Activated Cell Sorter (FACSCalibur, Becton Dickinson, San Jose CA USA), and PI fluorescence (designated as FL-2 Height in the histogram plots) was measured at 488 nm. Ten thousand cells were analyzed in each experiment. The percentage of cells in the G2/M phase of the cell cycle was then determined. The same procedure was used on cells pre-treated with 2 μmol/L PPMP for 72 h before VT-1 treatment.
DNA fragmentation analysis
TUNEL (TdT-mediated dUTP nick end labelling) staining detecting nuclear DNA fragmentation was used as a marker for late stage apoptosis using Roche's In situ cell death detection kit, TMR red (Roche Applied Science, Mannheim, Germany). T47D and MCF-7 cells were cultured to about 80% confluence. Medium was thereafter changed to fresh medium containing 0–5 μg/L VT-1 and incubation continued for 72 h. Cells were thereafter harvested with trypsin and any floating cells collected by centrifugation. Followed by TUNEL-stained according to the manufactures instructions and TUNEL-staining was determined with flow cytometry. The same procedure was used on cells pre-treated with 2 μmol/L PPMP for 72 h before VT-1 treatment.
Caspase activity determination
Fluorometric assays were used to detect caspase -3, -8, and -9 enzyme activities. T47D and MCF-7 cells were treated with 0–5 μg/L VT-1 for 24 h. Thereafter caspase-enzyme activity was measured with fluorometric assays (R&D Systems Inc., Minneapolis, MN, USA) as earlier described [
12]. The same procedure was used on cells pre-treated with 2 μmol/L PPMP for 24 h before VT-1 treatment.
SDS-PAGE gel electrophoresis and immunoblotting
VT-1 influence on specific proteins involved in apoptosis signal transduction was investigated through Western blotting. Cells were exposed to 0.1, 1, 5 μg/L VT-1, and 5 μg/L VT-1 with 2 μmol/L PPMP for 24 h and then lysed in lysis buffer. Gel electrophoresis and immunoblotting was preformed as earlier described [
11] using primary antibodies against AKT, p-Akt
ser473, p-Akt
tyr308, Bad, p-Bad
ser112, p-Bad
ser136, Bax, Bcl-2, p-Bcl-2
ser70, Bcl-X
L, SAPK/JNK, p-SAPK/JNK, P44/42, p-P44/42 (ERK1/2), MKK 3/6, p-MKK 3/6 and Monoclonal β-Actin antibody for detection of actin as loading control.
Reagents
Antibodies were obtained from the following: rat monoclonal IgM for Gb3 expression were from Immunotech, Marseille, France and all other antibodies were from Cell Signalling Technology Inc. Danvers, MA, USA. Eagle's MEM in Earl's salt was from Gibco Ltd, Paisley, Scotland, UK, and foetal bovine serum and L-glutamine were from Biochrom KG, Berlin, Germany. Trypsin. DL-threo-1-phenyl-2-palmitoylamino- 3-morpholino-1-propanol (PPMP), fluorescein diacetate (FDA), verotoxin-1, propidium iodide (PI), and RNase were from Sigma-Aldrich, St. Louis, MO, USA. Lysis buffer were from R&D Systems Inc. Minneapolis, MN, USA. The bicinchoninic acid (BCA) protein assay kit was from Pierce Biotechnology Inc. Rockford, IL, USA. NuPAGE antioxidant and reducing agent were from Invitrogen, Carlsbad, CA, USA. Tris-HCl SDS-PAGE criterion precast gel and Immune-Blot PVDF membranes were from Bio-Rad, Hercules, CA, USA. ECL Advance Western blotting detection system were from Amersham Biosciences, Buckinghamshire, UK.
Statistics
Statistical significance was tested with one-way ANOVA. The level of significance for rejecting the null hypothesis of zero treatment effect was taken to be p = 0.05.
Discussion
The use of toxins targeted against specific tumour cells or signal transduction pathways could reduce cancer treatment side effects and toxicity to normal tissues and thus be an approach to generate specific antitumour effects [
16]. Our results demonstrate that verotoxin-1- (VT-1) induced cytotoxicity involves induction of apoptosis in breast cancer cells expressing the toxin receptor Gb3. Gb3 expressing T47D cells were highly sensitive to VT-1
in vitro, In contrast, the MCF-7 cells lacking Gb3 were completely resistant to VT-1 treatment. Moreover, using PPMP, that inhibits glucosylceramide synthesis and thus makes exposed cells unable to synthesize Gb3 [
17] rendered T47D cells resistant or partly resistant to VT-1 depending on toxin concentration.
We also demonstrated that VT-1 at higher concentrations induces the intrinsic pathway to apoptosis, probably through activation of JNK. This leads to disruption of the mitochondrial membrane potential, activation of caspase-9 and -3, ultimately leading to DNA fragmentation and cell death. At higher VT-1 concentrations the mode of cell death was still through Gb3-dependent apoptosis. This was indicated by complete PPMP-blockable VT-1-induced MAP kinase protein expression, caspase activities and DNA fragmentation.
Most chemotherapeutic drugs exert their action by inducing apoptosis and the development of resistance to chemotherapy appears to be mediated by alterations in the sensitivity towards apoptosis [
18,
19]. Progress of tumour growth and resistance to treatment is associated with decreased rate of apoptosis [
20]. Many cytotoxic agents have a relatively narrow therapeutic window and in many patients the risk of side effects are higher than the probability of obtaining a clinical benefit. Targeted toxins against specific pathways could reduce the toxicity to non-tumour cells and be an approach to generate specific anti-tumour agents.
Several reports have demonstrated that VT-1 induces apoptosis in solid tumour cell lines such as astrocytoma, renal cell carcinoma, and colon cancer [
21‐
23]. VT-1 has also been shown to induce apoptosis and rapid elimination of mice tumour xenografts in Gb3-expressing human renal cell carcinoma, colon carcinoma, bladder carcinoma, glioblastoma, malignant meningiomas [
23‐
27].
Furthermore, 80% of all primary and metastatic breast cancer biopsies have been shown to express Gb3 [
8]. A recent study demonstrated expression of Gb3 in metastatic colon cancer, whereas expression was virtually absent in normal colonic epithelial cells [
26]. Elevated expression of Gb3 has also been demonstrated in drug-resistant ovarian tumours compared to the primary tumour [
28]. Gb3 has also shown to co-localize with the multi drug resistance protein P-glycoprotein (Pgp/MDR1) and have thus been suggested as a possible target for multi drug resistant cells [
29]. We recently demonstrated Gb3 in vascular endothelial- and tumour cells by immunostaining of Gb3 expression in glioma cryostat sections. The present results of the immunostaining showed expression of Gb3 in endothelial cells but also in tumour cells. This suggests that VT-1 exposure could be effective directly against breast cancer tumour cells but also affect tumour vascularity as suggested also for gliomas [
30]. The present demonstration of Gb3 expression in breast cancer tissue indicates that it is possible to use VT-1 also in the treatment of breast cancer. Further investigations should, however, be conducted to ensure that the cytotoxic effect of VT-1 does not lead to severe adverse effects due to targeting of normal Gb3-expressing cells outside of the tumor. The most important target expressing high levels of Gb3 would be the microvasculature of the kidney. It is also the main target of E. Coli VT-1-induced haemolytic uraemic syndrome [
31] Other non-tumour target cells of concern for adverse effects would be tonsil germinal-center B lymphocytes which regularly express Gb3 [
32]. Finding safe modes of administration are of critical importance for use of VT-1 as an anti-neoplastic agent.
We demonstrated that VT-1 induced a dose-dependent decrease of cell viability in Gb3-expressing T47D cells, but not in MCF-7 cells which lacked Gb3 expression. Expression of the membrane receptor Gb3 is obviously required for VT-1 internalization and transport of the toxin to the endoplasmic reticulum [
33]. When T47D cells were incubated with PPMP, that inhibits glucosylceramide synthesis and thus Gb3 expression [
17], prior to VT-1 exposure the sensitivity to VT-1 cytotoxicity at lower toxin concentrations was lost.
We then studied if VT-1-induced cytotoxicity was due to induction of apoptosis. Massive VT-1-induced DNA-fragmentation as well as apoptosis-related morphological changes was detected in the Gb3-expressing cell line T47D after 72 h incubation with high concentrations of VT-1. Pre-treatment with PPMP for 72 h prior to VT-1 exposure completely abolished VT-1-incuded DNA fragmentation. Apoptosis is most often executed by the activation of caspase -3. Activation of caspase-3 can be achieved either by caspase-8, the most proximal caspase in the receptor signalling pathway or by caspase-9 in the mitochondrial pathway to apoptosis. We therefore examined the activation of the three caspases after 24 h VT-1 exposure. Caspases-3 and -9, but not caspase-8 was found to be activated in both cell lines indicating that the intrinsic pathway to apoptosis was activated by VT-1. There was no VT-1-induced cellular activation of caspase-3,-8, or -9 after pre-incubation with PPMP. The mode of VT-1 cytotoxicity at low concentrations is still unclear. Nevertheless, cytotoxicity is apparently dependent on the cellular expression of Gb3 but not due to apoptosis, necrosis or cell cycle arrest as noted by lack of caspase activity or DNA fragmentation, cell lysis, and changes in cell cycle distribution, respectively.
Protein phosphorylation is the major cellular mechanism used to regulate protein functions, among them controlling cell growth, death, differentiation and apoptosis. The phosphorylation state of the key components of MAPK signalling pathways [
34] controlling Bax and mitochondrial function related to cell survival and apoptosis was therefore investigated. VT-1 induced JNK and MKK3/6 phosphorylation in T47D cells, suggesting that survival signal pathways were overruled by VT-1-induced JNK and p38 activation leading to mitochondrial depolarization, caspase-9 activation and apoptosis. Similar increases of JNK phosphorylation leading to apoptosis have been demonstrated to be essential for
Pseudomonas aeruginosa ExoS-induced apoptosis [
35] and for verotoxin-1 in glioma cell lines [
11].
Despite clinical improvements there are still major problems associated with breast cancer treatment. Chemotherapy and radiotherapy are associated with many, sometimes severe side effects. Therefore, targeted therapies that effectively kill cancer cells, without affecting normal tissues is a major objective in clinical cancer research. The high specificity and the ability of verotoxin-1 to selectively induce breast cancer cell death indicate that verotoxin-1 may be used as a potential anti-neoplastic agent for treatment of Gb3-positive breast cancers.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
DJ, PBM and AJ conceived and designed the study. EK and JM carried out most of the assays. IL collected the clinical data and TB evaluated the immunohistochemical staining. The manuscript was prepared by DJ, PBM and AJ. All authors read, critically advised, and approved the manuscript.