Herg K+ Channel-Dependent Apoptosis and Cell Cycle Arrest in Human Glioblastoma Cells

Glioblastoma (GB) is associated with poor patient survival owing to uncontrolled tumor proliferation and resistance to apoptosis. Human ether-a-go-go-related gene K+ channels (hERG; Kv11.1, KCNH2) are expressed in multiple cancer cells including GB and control cell proliferation and death. We hypothesized that pharmacological targeting of hERG protein would inhibit tumor growth by inducing apoptosis of GB cells. The small molecule hERG ligand doxazosin induced concentration-dependent apoptosis of human LNT-229 (EC50 = 35 µM) and U87MG (EC50 = 29 µM) GB cells, accompanied by cell cycle arrest in the G0/G1 phase. Apoptosis was associated with 64% reduction of hERG protein. HERG suppression via siRNA-mediated knock down mimicked pro-apoptotic effects of doxazosin. Antagonism of doxazosin binding by the non-apoptotic hERG ligand terazosin resulted in rescue of protein expression and in increased survival of GB cells. At the molecular level doxazosin-dependent apoptosis was characterized by activation of pro-apoptotic factors (phospho-erythropoietin-producing human hepatocellular carcinoma receptor tyrosine kinase A2, phospho-p38 mitogen-activated protein kinase, growth arrest and DNA damage inducible gene 153, cleaved caspases 9, 7, and 3), and by inactivation of anti-apoptotic poly-ADP-ribose-polymerase, respectively. In summary, this work identifies doxazosin as small molecule compound that promotes apoptosis and exerts anti-proliferative effects in human GB cells. Suppression of hERG protein is a crucial molecular event in GB cell apoptosis. Doxazosin and future derivatives are proposed as novel options for more effective GB treatment.


Introduction
Glioblastoma (GB) is the most common malignant primary brain tumor in adults. Current treatment is based on maximal safe surgical resection, followed by chemo-and radiotherapy when feasible [1]. However, outcome is poor despite optimal therapy with a mean survival rate of 1 year following diagnosis, which is due to uncontrolled tumor proliferation, infiltrative growth, angiogenesis, and resistance to apoptosis and medical treatment [2,3]. Thus, effective therapy of GB still remains an unmet medical need.
Given that GB cells express hERG channels and that hERG suppression by doxazosin induces apoptosis, we hypothesized that pharmacological targeting of hERG channels would induce apoptosis of GB cells. To explore hERG-associated GB cell apoptosis and underlying molecular pathways, human glioblastoma cell lines (LNT-229 and U87MG [17,18]) and the hERG inhibitor doxazosin were employed. Doxazosin triggered apoptosis and caused cell cycle arrest of GB cells. Suppression of hERG protein expression via siRNA-mediated knock down mimicked pro-apoptotic effects of doxazosin. HERG receptor binding competition of doxazosin and the small molecule compound terazosin that had no apparent effect on cell viability attenuated doxazosin-induced apoptosis of GB cells. In summary, a hERGdependent pro-apoptotic pathway is revealed in human glioblastoma cells, providing a novel therapeutic opportunity for future treatment of GB.

TUNEL Staining
Apoptosis was detected by terminal deoxyribonucleotide transferase-mediated dUTP nick end labeling (TUNEL) as described [19][20][21]. Following exposure to doxazosin for 24 h, cells grown in 12-well tissue culture dishes were fixed and TUNEL reaction mixture (Roche Applied Science, Mannheim, Germany) was added to the sections according to the manufacturer's instructions, followed by incubation at 37uC for 60 min. After removal of the TUNEL reagent slides were rinsed with phosphate-buffered saline (PBS), and TUNEL-positive cells were evaluated using a fluorescence microscope (IX 50; Olympus, Hamburg, Germany).

XTT Cell Viability Assay
Cell viability was quantified using an assay that utilizes the ability of live cells to reduce 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide salt (XTT) to produce a colored formazan compound. Cells grown in 96-well tissue culture dishes were transferred into drug-free media after 24 h drug application. XTT (125 mM; AppliChem, Darmstadt, Germany) was then added as reported [19], and cells were maintained at 37uC and 5% CO 2 for 2 hours in accordance with the manufacturer's instructions. Adsorption was determined at 450 nm using a spectrophotometer (PHOmo, Anthos Mikrosysteme, Krefeld, Germany) and normalized to control measurements obtained from corresponding cells cultured in drug-free medium.

Annexin V-FITC Apoptosis Assay
The annexin V-fluorescein isothiocyanate (FITC) assay was employed to quantify apoptosis at an early stage. Annexin V binds to phophatidylserine (PS) that is translocated to the outer leaflet of the plasma membrane during apoptosis. In addition, propidium iodide (PI) staining was applied as marker indicating compromised plasma membranes of late apoptotic LNT-229 cells. Following experimental treatment cells were rinsed with PBS, harvested using accutase (PromoCell, Heidelberg, Germany), and stained for 10 min at room temperature in the dark according to the manufacturer's instructions (Annexin V-FITC Detection Kit; PromoCell). Cell fluorescence was detected by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, NJ, USA) and analyzed using CellQuest software (Becton Dickinson).

Fluorocytometric Cell Cycle Analysis
Different phases of the cell cycle were distinguished by flow cytometry [22]. The assay is based on stoichiometric binding of propidium iodide to increasing amounts of DNA in cell cycle phases G0/G1, S, and G2/M. After doxazosin treatment, LNT-229 cells were trypsinized, fixed and permeabilized using ethanol, rinsed in phosphate-buffered saline (PBS), and treated with RNase A to remove RNAs. DNA was then quantitatively stained with propidium iodide for 1 h at room temperature protected from light. Fluorescence was analyzed using a FACSCanto flow cytometer (Becton Dickinson) and FlowJo software (Treestar, Ashland, OR, USA). DNA histogram data were fit with the Dean-Jett-Fox model.

Small Interfering RNA Knock Down
Anti-human ether-a-go-go-related gene (hERG; Kv11.1) siRNA (sc-42498; Santa Cruz) was used to knock down hERG protein expression. Scrambled siRNA (sc-37007; Santa Cruz) served as control. LNT-229 cells were allowed to settle in 6-well tissue culture dishes and antibiotic-free DMEM for 24 hours. SiRNAs (10 mmol/L stock solutions) were then transfected according to the manufacturer's recommendation. After 9 hours, DMEM supplemented with 20% fetal bovine serum (FBS) and 2% penicillin/ streptomycin was added. Following incubation for 24 hours doxazosin was added and cell viability was quantified 24 h later using the XTT assay.

Drugs
Doxazosin, desipramine and terazosin (Sigma-Aldrich, St. Louis, MO, USA) were prepared as 10 mM stock solutions in dimethyl sulfoxide (DMSO) and stored at 220uC.

Statistics
Data are presented as mean 6 standard error of the mean (SEM) of n experiments. Statistical differences of continuous variables were determined with Origin 6 software (OriginLab, Northampton, MA, USA) using unpaired Student's t tests (twosided tests). Statistical analyses were carried out prior to normalization of data. P,0.05 was considered statistically significant.

The Small Molecule Compound Doxazosin Induces Apoptosis of Human LNT-229 Glioblastoma Cells
Apoptosis of LNT-229 cells was analyzed in situ by TUNEL (terminal deoxyribonucleotide transferase-mediated dUTP nick end labeling) fluorescence, assessing DNA damage and fragmentation as characteristic apoptotic features. Compared to baseline conditions ( Figures 1A and 1E), increased apoptosis rates were detected following administration of the hERG inhibitor and a 1adrenoceptor antagonist, doxazosin (10-50 mM), for 24 h ( Figures 1B to 1D, 1F). Application of 10 mM doxazosin achieved significant hERG current reduction in vitro in human embryonic kidney (HEK) cells that were either stably (.95% inhibition) or transiently transfected with hERG cDNA (,90% inhibition) [14,15]. Quantification of GB cell death using a XTT-based cell viability assay revealed a half-maximal effective doxazosin concentration of 35.365.2 mM (n = 4-6 independent assays; Figure 1G). 50 mM doxazosin reduced cell viability to 14.1610.8% (n = 6; p = 0.003). The treatment duration required to achieve half-maximal pro-apoptotic effects of 10, 20, and 30 mM doxazosin was 48.861.2 h, 26.262.8 h, and 29.661.0 h, respectively (n = 3-6 assays; Figure 1H).

Doxazosin Causes Cell Cycle Arrest in the G0/G1 Phase
In addition to apoptosis, we examined the effects of doxazosin on cell cycle regulation of LNT-229 cells. Cell cycle phases were distinguished by flow cytometry after doxazosin treatment for 24 h ( Figures 2D to 2F). The fraction of cells in the G0/G1 phase was enhanced by doxazosin from 28.262.6% (solvent control; Figure 2D) to 43.761.1% (20 mM doxazosin; p = 0.008; Figure 2E) and 64.861.8% (30 mM doxazosin; p = 0.023; Figure 2F) (n = 3 assays). In contrast, GB cells in the G2/M phase were reduced from 23.161.6% to 8.161.4% (p = 0.016) and 11.461.8% (p = 0.012) under these conditions (n = 3), reflecting cell cycle arrest and anti-proliferative effects of the drug.

Doxazosin Triggers Apoptosis of U87MG Glioblastoma Cells
Pro-apoptotic effects of doxazosin were investigated in U87MG glioblastoma cells to exclude LNT-229-specific cellular effects. GB cell death ( Figures 3A and 3B) was quantified using the XTT assay following doxazosin application (24 h), yielding a half-maximal effective concentration of 28.961.3 mM (n = 4-6 independent assays; Figure 3C).

Desipramine Induces Death of LNT-229 Cells Similar to Doxazosin
To further assess pro-apoptotic effects of hERG antagonism, we applied the antidepressant drug desipramine that has previously been shown to reduce hERG currents and prevent hERG protein trafficking to the plasma membrane [11]. LNT-229 cells were treated with 30 mM desipramine for 24 h (Figure 4). This concentration achieved ,70% reduction of hERG protein and ,98% decrease of hERG currents in human embryonic kidney cells [11]. Desipramine reduced cell viability quantified using the XTT assay to 42.568.1% compared to control cells (n = 3; p = 0.008).

Modulation of Doxazosin-induced Apoptosis by Terazosin, a Small Molecule hERG Inhibitor
The mechanistic role of small molecule-induced hERG protein suppression in apoptosis of GB cells was tested using terazosin, an a 1 -adrenoceptor antagonist that reduces hERG currents in culture cells [14]. Terazosin was applied to LNT-229 cells for 24 h. In contrast to doxazosin (Figure 1), terazosin at concentrations between 10 mM to 100 mM did not significantly affect cell viability quantified by XTT assay (n = 5 to 6 cells; Figure 9A) despite ,90% blockade of hERG current at 100 mM terazosin [14]. Relative cell viability yielded 96.663.1% (n = 6; p = 0.86) after application of 100 mM terazosin. This observation supports a role for hERG protein expression in GB cell apoptosis independent of hERG ion channel function. Consistent with this notion, terazosin (100 mM; 24 h) exhibited weak effects on hERG protein levels in LNT-229 cells (220.462.8%; n = 3; p = 0.03), while doxazosin (30 mM; 24 h) suppressed hERG expression by 48.261.7% (n = 3; p = 0.002; Figure 9B). However, application of 30 mM doxazosin +100 mM terazosin did not result in additive reduction of hERG protein. An unexpected finding was revealed: simultaneous application of 100 mM terazosin attenuated doxazosin-induced hERG suppression to 24.5611.4% (n = 3; p = 0.07; Figure 9B). This observation suggests that hERG receptor occupancy interferes with binding of terazosin. Antagonism of doxazosin binding by terazosin was associated with restoration of protein expression and increased survival of GB cells. LNT-229 cell viability was increased by co-application of 50 mM and 100 mM terazosin to 53.364.8% (n = 3; p = 0.008) and 66.766.9% (n = 4; p = 0.02), compared to application of 30 mM doxazosin for 24 h (41.061.5%; n = 4; p = 0.002; Figure 9C).

Pro-apoptotic and Anti-proliferative Functions of Doxazosin in Human Glioblastoma Cells
We identified and characterized doxazosin, an a 1 -adrenoceptor antagonist clinically used to treat hypertension and benign prostatic hyperplasia, as a novel small molecule trigger of apoptosis in GB cells. Doxazosin induced apoptosis of human GB cells in time-and concentration-dependent fashion. Apoptotic cell death was confirmed using TUNEL-and annexin V-based assays to demonstrate DNA fragmentation and phosphatidylserine translocation to the outer leaflet of the plasma membrane as characteristic apoptotic features. In addition, concentration-dependent activation of specific pro-apoptotic signaling (i.e., cleavage of PARP and of caspases 3, 7, and 9) was observed in the presence of doxazosin. Suppression of mature hERG protein by doxazosin was identified as crucial molecular event in GB cell apoptosis. Proapoptotic effects were similarly observed with desipramine, an antidepressant drug that induces reduction of hERG protein [11].
Doxazosin treatment resulted in activation of the pro-apoptotic receptor tyrosine kinase EphA2 and associated degradation of EphA2 that is characteristic to activated receptor tyrosine kinases, consistent with data obtained from prostate cancer (PC3) and breast cancer (MDA-MB-231) cells [27]. Doxazosin stimulation also activated pro-apoptotic factors p38 MAPK and GADD153. Furthermore, initiator caspase 9 and effector caspases 7 and 3 were activated in response to doxazosin treatment. PARP inactivation reflected loss of DNA repair activity. Finally, phagocytosis of dead cells was indicated by increased expression of the lower migrating form of LC3 protein. The correlation of doxazosin application with established pro-apoptotic signaling reported in this work requires mechanistic validation in future studies.
Pro-apoptotic effects were accompanied by GB cell cycle arrest in the G0/G1 phase that shows anti-proliferative actions of doxazosin. Similar effects were observed in GH3 and AtT-20 pituitary tumor cells and attributed to attenuated nuclear factor-kB signaling [28]. Moreover, hERG protein reduction through siRNA knock down decreased proliferation in small cell lung  The Mechanistic Role of hERG K + Channel Protein A pivotal role of hERG protein expression for apoptosis of GB cells is suggested by reduced hERG expression following doxazosin treatment. Similar reduction of cell viability observed  with desipramine further supports mechanistic significance of hERG protein levels in GB cell apoptosis. Furthermore, specific siRNA-mediated knock down of hERG expression triggered apoptosis, confirming the mechanistic role of small moleculeinduced hERG protein suppression. Finally, antagonism of doxazosin-induced hERG protein suppression and glioblastoma cell death by co-treatment of terazosin and doxazosin indicates the functional relevance of hERG protein in GB cell apoptosis. In contrast to doxazosin, terazosin did not induce apoptosis despite similar degrees of acute hERG K + channel blockade in vitro (doxazosin, .95%; terazosin, ,90%; [14]). Mechanistic differentiation of acute blockade and apoptosis indicates a role for hERG protein expression independent of hERG function. Doxazosin and other hERG K + channel ligands impair protein expression by selective disruption of hERG trafficking into the cell surface membrane [10,11,15]. This mechanism of action occurs mechanistically independent of acute hERG current blockade, a wellestablished class III antiarrhythmic mode of action and proarrhythmic property of several non-antiarrhythmic compounds [30]. The presence of independent drug-channel interaction sites for inhibition of hERG forward trafficking and acute blockade is supported by the observation that certain compounds induce current block with little or no trafficking inhibition (and vice versa) [10,20]. Moreover, elimination of a common drug binding site formed by aromatic residues in the hERG channel pore [31] abolished acute current block without affecting protein trafficking inhibition (reviewed in [10]). Here, protection from GB cell apoptosis by terazosin that exhibits low intrinsic trafficking inhibition may be readily explained by pharmacological ''rescue'' of hERG surface protein expression through competitive binding of doxazosin and terazosin at a common trafficking receptor site. Treatment with terazosin prevented doxazosin binding and preserved hERG protein expression. However, an allosteric interaction between separate doxazosin and terazosin binding sites cannot be excluded, and the precise molecular mechanism underlying the suppression of mature hERG protein in doxazosinassociated apoptosis remains to be elucidated.
In addition to its interaction with hERG channels, doxazosin may directly bind and activate EphA2 [27]. We detected high apoptosis rates following specific siRNA knock down of hERG protein that were increased further by ,16% through additional application of doxazosin. This difference may be partially mediated through direct activation of EphA2 signaling by doxazosin. Furthermore, additional reduction of hERG viability in the presence of anti-HERG siRNA may be caused by doxazosin-induced suppression of residual hERG protein (,42%).
In summary, there are multiple lines of evidence that indicate and confirm the mechanistic significance of hERG channels in GB cell apoptosis: 1. Previous data revealed that doxazosin specifically induced apoptosis in hERG-expressing cells [15]. 2. Concentration-dependent hERG protein reduction was observed in apoptotic cells exposed to doxazosin ( Figure 5A). 3. The antidepressant drug desipramine that blocks hERG protein trafficking to the plasma membrane [11] exhibited pro-apoptotic effects similar to doxazosin (Figure 4). 4. Anti-hERG siRNA reduced hERG protein expression and induced apoptosis similar to doxazosin (Figure 8). 5. Terazosin did not reduce hERG protein expression ( Figure 9B) and did not show pro-apoptotic effects ( Figure 9A) despite structural similarities between terazosin and doxazosin.
6. In the presence of doxazosin, terazosin rescued hERG protein expression and protected from apoptosis ( Figure 9C).

Clinical Implications and Future Directions
Current treatment of GB is associated with low patient survival rates owing to uncontrolled proliferation and resistance to apoptosis or cytotoxic treatment [2,3]. This work identifies doxazosin as small molecule compound that triggers apoptosis and exerts anti-proliferative effects in human GB cells. Thus, we propose that the Food and Drug Administration (FDA)-and European Medicines Agency (EMA)-approved drug doxazosin may be re-purposed for treatment of human glioblastoma in the future following in vivo concept validation. In patients treated with doxazosin for hypertension or benign prostatic hyperplasia, plasma levels between 42 nM and 244 nM have been reported [32][33][34] with 98.3% protein binding [35]. These data indicate that apoptosis of GB cells will not be induced using doses that are currently applied for non-malignant diseases. Rather, increased doxazosin doses would be required for the use as anticancer drug that have not yet been determined in vivo.
More widespread expression of hERG protein in tumors (e.g. colorectal or endometrial cancer) and cancerous cells (e.g. acute myologenous or chronic lymphocytic leukemia) [12] suggests potential significance of hERG-based anticancer therapy beyond glioblastoma that warrants further exploration, carefully considering the significance of the individual cellular context for regulation of cell death and proliferation. A systematic in vivo investigation of anticancer properties and potential cardiac side effects of small molecule-induced hERG protein suppression is required. Specifically, significant apoptosis of cardiac myocytes and marked inhibition of repolarizing hERG channels have been observed at similar concentrations compared to induction of GB cell apoptosis [13][14][15]. Thus, proarrhythmic and cardiotoxic risks of hERG inhibitors and associated cardiac I Kr current reduction/ QTc interval prolongation or apoptosis of cardiac myocytes require careful pre-clinical evaluation and clinical monitoring when applying doxazosin and derivatives in clinical oncology [6,8,10,11,16,19,20,36].

Conclusions
The small molecule compound doxazosin induces apoptosis in GB cells and exhibits anti-proliferative function. HERG potassium channels, previously recognized to regulate cardiac action potential repolarization, modulate GB cell apoptosis. The data imply that doxazosin-related reduction of hERG protein expression may be linked to increased apoptosis rates. This novel finding of apoptosis regulation in GB cells may provide new options for anticancer therapies. Particularly, doxazosin and doxazosin-based derivatives may be useful as novel treatment for GB due to their pro-apoptotic and anti-proliferative properties.