Activation of Membrane Androgen Receptors in Colon Cancer Inhibits the Prosurvival Signals Akt/Bad In Vitro and In Vivo and Blocks Migration via Vinculin/Actin Signaling
Recently, we reported that membrane androgen receptors (mARs) are expressed in colon tumors triggering strong apoptotic responses. In the present study, we analyzed mAR-induced downstream effectors controlling cell survival and migration of Caco2 colon cancer cells. We show that long-term activation of mAR downregulated the activity of PI-3K and Akt and induced dephosphorylation/activation of the proapoptotic Bad (p-Bad). Moreover, treatment of APCMin/+ mice, which spontaneously develop intestinal tumors, with mAR-activating testosterone conjugates reduced the tumor incidence by 80% and significantly decreased the expression of p-Akt and p-Bad levels in tumor tissue. Furthermore, mAR activation strongly inhibited Caco2 cell migration. In accordance with these findings, vinculin, a protein controlling cell adhesion and actin reorganization, was effectively phosphorylated upon mAR activation. Phosphorylation inhibitors genistein and PP2 inhibited actin reorganization and restored motility. Moreover, silencing vinculin by appropriate siRNA’s, or blocking actin reorganization by cytochalasin B, restored the migration potential. From these results we conclude that mAR activation inhibits the prosurvival signals Akt/Bad in vitro and in vivo and blocks migration of colon cancer cells via regulation of vinculin signaling and actin reorganization, supporting the powerful tumoristatic effect of those receptors.
Recent studies established the expression of functional membrane androgen receptors (mARs) inducing rapid nongenomic androgen actions in tumors, including prostate (1,2), breast (3,4) and colon (5) as well as in other cell types such as macrophages and T cells (6,7), C6 (8) and vascular smooth muscle cells (9). The exact molecular identity of mAR remains unknown. It has been reported that signals emanating from this receptor, such as intracellular calcium or inositol 1,4,5-trisphosphate, are sensitive to pertussis-toxin inhibition (10,11), indicating that mAR may be a specific G-protein coupled receptor (GPCR) or a receptor in close association with a GPCR. Moreover, stimulation of mARs by membrane-impermeable testosterone conjugates triggered specific early signaling pathways and induced proapoptotic responses that could not be blocked by antiandrogens (1,5,12,13,14), implying that mAR effects are most likely different from those manifested upon activation of the intracellular androgen receptors (iARs).
The mAR-dependent nongenomic signaling was recently characterized in detail in prostate and breast cancer cells (15), and key prosurvival and proapoptotic gene products were identified that regulate the apoptotic response induced by mAR activation (16). According to these reports it was postulated that mAR might be a significant novel target for cancer treatment. Most recently, we reported that mARs (but not membranebound iARs) are expressed in colon tumors (5). Activation of mARs by testosterone-albumin conjugates triggered strong apoptotic responses and considerably reduced the colonic tumor incidence following chemical cancerogenesis in Balb/c mice (5). Moreover, mARs were predominantly expressed in colon tumor cells, whereas very low or even undetectable mARs were expressed in normal tissues (5). Although the proapoptotic responses were clearly dependent on mAR stimulation, downstream events regulating the expression and/or function of key prosurvival mediators in colon tumors remained undefined. In addition, the profound antitumorigenic mAR action has not been functionally correlated with changes in other key mechanisms such as cell motility and invasiveness.
The main goal of this work was to address the role of mAR activation toward major characteristics of tumor cells, namely cell survival and cell migration. Because mAR activation has been shown to promote strong apoptotic regression (2,5), we analyzed the functionality of the prosurvival PI-3K/Akt pathway. In addition, because this signaling pathway also plays a major role in the invasive potential of tumor cells, we further studied the migration potential upon mAR activation. Our findings indicate that prosurvival signaling prevails in colon tumor cells but is strongly downregulated upon mAR stimulation in vitro and in colon tumor tissues isolated from APCMin/+ mice following treatment with mAR agonists. Furthermore, mAR activation blocked migration and invasiveness of colon tumor cells, mainly recruiting the adhesionand actin cytoskeleton-regulator vinculin. These results provide novel mechanistic insights into the regulation of the proapoptotic and antimigratory mAR effects in colon tumors.
Materials and Methods
Cell Cultures and Wound Healing Assay
The Caco2 human colon cancer cell line and IEC06 nontransformed intestinal cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and were studied between passages 60 and 70. On the basis of previous titration experiments (5), we used a 10−7 mol/L testosterone-human serum albumin (HSA) concentration for mAR stimulation. For the wound healing assay, confluent cell cultures were scraped with a pipette tip across a 24-well plate. Following wounding, culture medium was replaced with fresh medium, and cells were exposed to 10−7 mol/L testosterone-HSA (Sigma, St. Louis, MO, USA) in the presence or absence of 10−6 mol/L cytochalasin B (Sigma) for the indicated time. It should be noted that the method for the wound healing assay requires that confluent cells are used, to provide the necessary cell layer to create wounding by scraping. On the other hand, subconfluent cells were used in Matrigel and Transwell assays (see below) to ensure correct evaluation of cell motility and invasiveness.
Matrigel and Transwell Assays
Matrigel assays were performed by using BD BioCoat™ BD Matrigel™ Invasion Chambers (BD Bioscience, San Jose, CA, USA). Matrigel was placed in each insert with 8.0-µm pore size in a 24-well plate. The chamber was allowed to polymerize at 37°C for 1 h. The inserts were then washed with serum-free DMEM, and 100 µL of complete cell culture medium with 1 × 105 cells was then seeded onto the insert. Five hundred microliters of complete cell culture medium with 10−7 mol/L testosterone-HSA or testosterone-bovine serum albumin (BSA) (Sigma) in the presence or absence of 10−6 mol/L cytochalasin B (Sigma) was added into the well below the insert. In control experiments, cells were preincubated with 10−7 mol/L testosterone-albumin conjugates (TAC) for 2 h. Then, TAC was washed out with complete cell culture medium and 500 were added into the well below the insert. After a 24-h incubation, the insert was wiped with a wet cotton swab. The lower surface was gently rinsed with phosphate-buffered saline (PBS), the cells were fixed and stained with 4,6-diamidino-2-phenylindole (DAPI) for 10 min, rinsed again with sterile water and allowed to dry. After the membranes were removed from the inserts, they were mounted with ProLong Gold antifade reagent (Invitrogen, Paisley, UK). To determine the total number of migrating cells, the slices were viewed and imaged under the microscope, and the number of cells/field in 10 random fields was counted. Experiments were performed in triplicates.
The Transwell assay was performed using Transwell inserts (BD Bioscience). The inserts were then washed with serum-free DMEM, and 100 µL of complete cell culture medium with 1 × 105 cells were seeded onto the insert. We then added 500 µL of complete cell culture medium with 10−7 mol/L testosterone-HSA or testosterone-BSA (Sigma) in the presence or absence of 10−6 mol/L cytochalasin B (Sigma), 10−6 mol/L anastrozole (Sigma), 10−6 mol/L flutamide (Sigma), 5 × 10−6 mol/L PP2 (Sigma) or 5 × 10−6 mol/L genistein (Sigma) into the well underside of the insert for 24 h at 37°C and 5% CO2. In control experiments, cells were preincubated with 10−7 mol/L TAC for 2 h. Then, TAC was washed out with complete cell culture medium and 500 µL was added into the well below the insert. After 24 h incubation, cells were fixed, stained with DAPI for 10 min and microscopically analyzed as described above.
Immunoprecipitation and Western Blotting
Cells were incubated with 10−7 mol/L testosterone-HSA, free testosterone or free HSA for the indicated time, washed twice with ice-cold PBS and suspended in 500 µL ice-cold lysis buffer (50 mmol/L Tris/HCl, 1% TritonX-100 pH 7.4, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS], 0.15% NaCl, 1 mmol/L EDTA, 1 mmol/L sodium orthovanadate) containing protease inhibitor cocktail (Sigma). The protein concentration was determined using the Bradford assay (Bio-Rad, Munich, Germany). Sixty micrograms of protein were solubilized in sample buffer at 95°C for 5 min and resolved by 10% SDS-polyacrylamide gel electrophoresis. For immunoblotting, proteins were electrotransferred onto a polyvinylidene difluoride membrane and blocked with 5% nonfat milk in Tris-buffered saline-0.10% Tween 20 at room temperature for 1 h. Then, the membrane was incubated with phospho-Akt (Thr308), phospho-PI-3K p85 (Tyr458)/p55 (Tyr199) (1:1000; Cell Signaling, Danvers, MA, USA), or phospho-Bad (Ser136) (1:100, Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C overnight. After washing (in PBS with 0.05% Tween [PBST]) and subsequent blocking, the blot was incubated with secondary antirabbit antibody (1:2000; Cell Signaling) or antimouse antibody (1:5000, GE Healthcare, Piscataway, NJ, USA) for 1 h at room temperature. After washing, antibody binding was detected with the enhanced chemiluminescence detection reagent (Amersham, Freiburg, Germany). For controls the blots were stripped in stripping buffer (Carl Roth, Karlsruhe, Germany) at 56°C for 30 min. After washing with PBST, blots were blocked with Tris-buffered saline with Tween + 5% milk for 1 h at room temperature. Then, they were incubated with anti-Akt or anti-Bad (1:100; Santa Cruz Biotechnology) antibodies at 4°C overnight. After washing with PBST and incubation with antirabbit antibody (1:2000; Cell Signaling), antibody binding was detected and quantified with Quantity One Software (Bio-Rad).
For immunoprecipitation, equal amounts of protein (500 µg) in the presence or absence of 50 µmol/L genistein were subjected to immunoprecipitation with a monoclonal antiphosphotyrosine antibody (7 µg/500 µg total protein, Santa Cruz Biotechnology). After incubation with the antibody for 1 h at 4°C, 50 of the homogeneous protein A suspension was added into the mixture and incubated overnight at 4°C on a rocking platform. After three washing steps, samples were resuspended in SDS sample buffer, subjected to SDS electrophoresis and transferred to nitrocellulose membrane. Proteins were incubated with vinculin monoclonal antibody (1:100; Santa Cruz Biotechnology) followed by the appropriate antimouse antibody (1:5000; GE Healthcare). Detection of protein bands was performed with an enhanced chemiluminescence kit. Bands were quantified with Quantity One Software (Bio-Rad).
Immunofluorescence Analysis and Confocal Laser Scanning Microscopy
For testosterone-HSA-fluorescein isothiocyanate (FITC) staining, 5-µm-thick frozen tissue sections from the adenomatous polyposis coli (APC) mouse tumors were fixed with 4% paraformaldehyde (PFA) for 15 min and incubated with 5% BSA/1 × PBS/0.3% Triton for 1 h at room temperature. After two washes with PBS with 1.5% fetal bovine serum, specimens were exposed to testosterone-HSA-FITC (10−7 mol/L; Sigma) for 1 h at room temperature. Nuclei were stained with DRAQ-5 dye (1:1000; Biostatus, Leicestershire, UK) for 10 min at room temperature.
To quantify the expression of phosphorylated Akt and Bad, 5-µm-thick frozen tissue sections from the APC mice colon tumors (at least three tumors of similar size from each animal) were fixed with 4% PFA for 15 min at room temperature. After washing twice with PBS the slides were incubated with 5% normal goat serum/1 × PBS/0.3% Triton for 1 h at room temperature. Then, the specimens were exposed overnight at 4°C to phospho-Akt (Thr308) (1:800; Cell Signaling) or phospho-Bad (Ser136) (1:100, Santa Cruz Biotechnology). The slides were rinsed three times with PBS and incubated for 1.5 h at room temperature with secondary FITC goat antirabbit antibody (1:500; Invitrogen). After three washing steps the nuclei were stained for 10 min at room temperature with DRAQ-5 dye (1:1000; Biostatus).
To determine the phosphorylation of vinculin, cells were cultured on glass cover slips with testosterone-HSA or control without testosterone-HSA for different time periods, which are indicated in the figure legends. After washing twice with PBS, cells were incubated with 4% PFA for 15 min and then incubated with 5% normal goat serum/1 × PBS/0.3% Triton for 1 h at room temperature. Then, the cells were exposed to antivinculin antibodies (1:400; Gene Tex, Irvine, CA, USA) at 4°C overnight. The cells were rinsed three times with PBS and incubated with secondary FITC goat antirabbit antibody (1:500; Invitrogen) or goat antimouse antibody (1:500; Invitrogen) for 1.5 h at room temperature. For F-actin staining, cells were incubated with rhodamine-phalloidin (1:100; Molecular Probes, Eugene, OR, USA) for 40 min in the dark. After three washing steps the nuclei were stained with DRAQ-5 dye (1:1000; Biostatus) for 10 min at room temperature. All the slides and cover slips were mounted with ProLong Gold antifade reagent (Invitrogen). Images were taken on a Zeiss LSM 5 EXCITER Confocal Laser Scanning Microscope (Carl Zeiss MicroImaging, Jena, Germany) with a water-immersion Plan-Neofluar 40x/1.3 NA DIC. Images were analyzed with the instrument’s software.
Small Interfering RNA Experiments
Caco2 cells were grown in DMEM medium containing 10% fetal calf serum under standard culture conditions (37°C, 5% CO2). We then seeded 4 × 104 cells in 24-well plates and cultivated them with fresh culture medium for 8 h. The cells were subsequently transfected with validated small interfering RNA (siRNA) for vinculin (ID# s14764; Ambion, Darmstadt, Germany) or with a negative control siRNA by using an siPORT Amine (Ambion) transfection agent according to the manufacturer’s protocol. The efficiency of silencing was checked by Western blot 72 h after transfection. Upon silencing, 37.6% of the vinculin protein was still detectable in cells treated with siRNA for Vinculin compared with cells treated with a negative control siRNA.
In Vivo Experiments
In vivo animal experiments were carried out in 2- to 6-month-old age-matched mice of both sexes with mutated apc resulting in spontaneous colon tumor development (APCMin/+) obtained from the Jackson Laboratory (Bar Harbor, ME, USA). The animals were housed under controlled environmental conditions (22°–24°C, 50–70% humidity and a 12-h light/dark cycle). Throughout the study the mice had free access to standard pelleted food (C1000; Altromin, Lage, Germany) and tap water. All animal experiments were conducted according to the German law for the care and welfare of animals and were approved by local authorities.
The animals were divided into two groups. Group A (n = 6) received 2× 5 mg/kg subcutaneous TAC injections 3 times per wk for 8 wks. In the control group B (n = 4) similar doses of normal saline were given. At the end of 8 wks all animals were anesthetized with ether and killed. After death, the entire colorectum from the colorectal junction to the anal verge was examined. Then, the colon was opened longitudinally, washed with PBS, and divided into three portions (proximal, middle and distal). Tumors were counted with a dissecting microscope at 3× magnification. After inspection the colon was fixed in a 40% g/L formaldehyde buffer solution (pH 7.4).
Terminal Deoxynucleotidyl Transferase dUTP Nick-End Labeling Assay
The colonic cancer tissues (at least three independent tumors from each animal) were cut into 8-µm frozen sections and subsequently fixed in 4% PFA for 30 min at room temperature. After rinsing with PBS the samples were permeabilized in a solution of 0.1% Triton X-100 in sodium citrate for 2 min. Samples, washed with PBS, were then incubated in the terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) reaction mix for 1 h at 37°C, according to the manufacturer’s instructions (Roche, Mannheim, Germany). Nuclei were stained with DRAQ5™ (Biostatus). Sections were analyzed with a confocal laser scanning microscope (Carl Zeiss).
APOPercentage Apoptosis Assay
Caco2 cells were cultured in 96-well plates for the APOPercentage apoptosis assay (Biocolor, Belfast, Ireland). In the presence or absence of 10−6 mol/L anastrozole (Sigma), the cells were stimulated or not with 10−7 mol/L TAC for 24 h in serum-containing medium. Untreated cells cultured in serum-free medium were used as a positive control for the apoptotic response.
Data are provided as means ± SEM, n represents the number of independent experiments. Data were tested for significance by using the unpaired Student t test or ANOVA as appropriate. Differences were considered statistically significant when P values were < 0.05. Statistical analysis was performed with GraphPad InStat version 3.00 for Windows 95 (GraphPad Software, San Diego, CA; USA, https://doi.org/www.graphpad.com).
All supplementary materials are available online at https://doi.org/www.molmed.org.
mAR Stimulation Inhibits Akt Activity and Induces Bad Dephosphorylation in Caco2 but Not in IEC06 Cells
p-Akt and p-Bad Are Downregulated in Colon Tumor Tissues Treated by Testosterone-HSA
mAR Activation Inhibits Cell Motility in Colon Cancer Cells
mAR Activation Triggers Vinculin Phosphorylation
Vinculin Is Necessary for Actin Reorganization and Migration of mAR Stimulated Caco2 Cells
The present paper discloses the novel finding that prosurvival signals are effectively downregulated in mAR-expressing colon tumors following stimulation by TAC in vitro and in vivo. First, phosphorylated kinase Akt, which is constitutively upregulated in colon tumors but not in nontransformed cells (Figure 1A, B; Supplemental Figure 1; and data not shown), was significantly downregulated upon long-term mAR activation by TAC. In line with this result, PI-3K, an upstream regulator of Akt, was dephosphorylated upon long-term TAC treatment (Figure 1D), indicating that the prosurvival PI-3K/Akt signaling is downregulated in mAR-activated Caco2 cells. In addition, the proapoptotic Bad protein was efficiently dephosphorylated and thus activated by TAC following similar kinetics (Figure 1C). These results suggest that, in contrast to nontransformed intestinal IEC06 cells, mAR-expressing colon tumor cells express activated prosurvival signals that may protect them from apoptotic cell death. mAR activation downregulated the activity of these signals via dephosphorylation, a finding consistent with the strong apoptotic regression upon mAR stimulation reported recently (5). These results were confirmed by in vivo experiments in TAC-treated APCMin/+ mice. This mouse model carries defective APC, which results in the spontaneous development of gastrointestinal tumors (25). Indeed, the colon tumor incidence observed in the treated 2- to 6-month-old age-matched animals was significantly reduced by 80% (Figure 2B), corroborating previously reported data obtained in a chemically induced colon tumor model in Balb/c mice (5). Interestingly, histological immunostaining analysis revealed that both p-Akt and p-Bad were effectively downregulated in colon tumor specimens isolated from TAC-treated animals (Figure 2D). These findings collectively provide novel mechanistic evidence, pointing to p-Akt/p-Bad signaling, which may control the mAR-induced antitumorigenic effects in vitro and in vivo. It should be noted that APCMin/+ mutant mice of both sexes were used in this study to avoid conflicting data regarding the sex-related prevalence of colon tumors of APCMin/+ mutant mice. It has been pointed out that the enhanced susceptibility of male mice to intestinal tumor growth (26,27) results instead from the classical androgen receptor, because mAR antitumorigenic effects seem to be independent of classical androgen-receptor signaling (1,5,12,13).
In vitro downregulation of the PI-3K/Akt signaling upon mAR activation has also been reported in DU145 prostate cancer cells (16). In these cells, PI-3K was constitutively activated (13), whereas long-term mAR stimulation by specific agonists induced dephosphorylation of both PI-3K and its downstream effector Akt (16). Taken together, the results observed in prostate and colon tumors indicate that mAR expression is associated with active prosurvival pathways and thus protects cells from apoptotic regression. Activation of these receptors triggers specific signaling to restrict this prosurvival machinery. This assumption is fully supported by the in vitro and in vivo results presented here, which support the argument that the in vitro findings reported previously are not simply a side effect of mAR activation.
Akt has been also shown to play a major role in the invasive potential of colon cancer cells in response to a variety of stimuli, for example, heregulin (28), PAK1 (29) and Sprouty-2 (30). Moreover, inhibition of Akt-dependent pathways has been linked to reduced cell motility in colon cancer (21), whereas mAR stimulation downregulates p-Akt (16 and this work), and mAR-dependent activation has been shown to block cell motility and invasion in prostate cancer cells (1). Thus, we sought to determine an effect of mAR activation on cell motility in colon tumor cells. The present observations further reveal that mAR stimulation modulates specific molecular targets controlling cell motility. Activation of mAR by two distinct nonpermeable TAC (testosterone-HSA and testosterone-BSA) markedly inhibited cell motility as documented by different assays (Figure 3). The possibility that testosterone conjugates may be converted to estrogen and influence the overall interpretation of our results was also considered. However, previous binding studies in colon cancer (5) as well as in prostate cancer cells (2) have clearly indicated that estrogen (and progesterone) displaced radiolabeled testosterone with significantly lower affinity (104- to 102-fold). These findings indicate that even if such a conversion takes place, it can not influence the mAR-induced effects described so far, because of the high androgen selectivity of these membrane receptors. In addition, control experiments showed that neither migration nor apoptotic responses were influenced by the aromatase inhibitor anastrozole (Figure 4), further supporting the androgen specificity of the mAR-induced effects. Finally, because flutamide did not influence the motility effects (Figure 4A) nor the apoptotic responses in colon cancer (5), it is believed that the mAR pool mediating the observed effects is most likely unrelated to a membrane-associated form of the classical, intracellular AR that may be present in the plasma membrane of colon tumor cells. Although the existence of such a form of membrane tethered intracellular AR has not been reported in colon cells, experimental data in prostate cells places iARs on the cell membrane (31). In that case, however, and in sharp contrast to what we have observed in our assays in colon tumors, membrane-associated iARs induced cell proliferation (instead of apoptosis), which was efficiently blocked by antiandrogens and antiestrogens (31). In conclusion, our data support the existence of an active, non-AR/ER-related membrane receptor bearing anticancer action in colon cancer cells. This conclusion is in-line with recent findings demonstrating that iARs could not be detected in membrane preparations of Caco2 cells (5).
Inhibition of migration is usually correlated with impaired expression/activation of adhesion molecules and reorganization of focal contact structures, including the actin cytoskeleton (32). Because actin reorganization is a major effect of mAR activation in tumor cells (2,3,5,13), we focused on the molecular mechanism underlying the mAR-induced inhibition of cellular motility in Caco2 cells (Figure 3). Our results demonstrate that vinculin is a main target of mAR activation that may regulate cell motility. Phosphorylation of vinculin (Figure 5A) was an early and persistent event leading to significant morphological changes of Caco2 cells. It was clearly correlated with actin restructuring, as indicated by the visualization of newly organized actin filaments emanating from the vinculin spots on the cell-adhesion contacts (Figure 5B). Vinculin silencing or inhibition of vinculin phosphorylation by specific inhibitors largely reversed actin reorganization and the inhibition of migration. These findings imply that vinculin phosphorylation/activaton upon mAR stimulation regulates cell adhesion and inhibits the migration potential of Caco2 tumor cells. This conclusion is in line with several reports in the literature. Thus, vinculin was shown to be important in regulating adhesion dynamics and cell migration (33), and it was postulated that vinculin may connect early adhesion sites to the actin-driven protrusive machinery (23). According to this model vinculin stabilizes focal adhesions and thereby suppresses cell migration, an effect that is relieved by modifications of inositol phospholipids (34). Although the precise role of vinculin in focal adhesions remains to be elucidated, recent experimental evidence suggests that vinculin overexpression reduces cell migration, whereas vinculin downregulation enhances cell motility (34). This hypothesis is in-line with the results presented in our study, showing inhibition of cell motility ahead of vinculin activation in mAR-stimulated Caco2 cells. Notably, this effect was efficiently reversed by silencing of endogenous vinculin by use of siRNAs. Finally, we noted that the effects of TAC on cell invasion were manifested even after short treatment of cells with this compound (Figure 8). Moreover, these effects were still silenced by siRNAs against vinculin (Figure 8). These results clearly indicate that the inhibitory signals on invasion are activated early upon mAR stimulation, are dependent on vinculin and are present in cells well before they commit to the mAR-induced apoptotic program.
In conclusion, our results provide novel mechanistic insights into mAR-induced antitumorigenic actions. The long-term activity of the prosurvival regulators PI-3K, Akt and Bad is effectively suppressed and the specific molecular target for cell adhesion, vinculin, is modulated upon mAR activation. Because this molecule may adapt signaling pathways involved in apoptosis, cell survival and motility (35,36), we hypothesize that it represents a key signaling effector regulating the mAR-dependent antitumorigenic effects observed in our reported studies. Further experiments are now needed to address the molecular identity of mAR and to evaluate the potential role of these signaling targets for the development of novel antitumorigenic strategies based on specific mAR activation.
The authors declare that they have no competing interests as defined by Molecular Medicine, or other interests that might be perceived to influence the results and discussion reported in this paper.
This work was supported by grants from Deutsche Forschungsgemeinschaft (GRK 1302; SFB773; Mercator program) and the Greek Ministry of Health (KESY program).