M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

available at www.sciencedirect.com

ScienceDirect www.elsevier.com/locate/molonc

Combination therapy induces unfolded protein response and cytoskeletal rearrangement leading to mitochondrial apoptosis in prostate cancer Sandeep Kumara, Ajay K. Chaudharya, Rahul Kumara, Jordan O’Malleya, Anna Dubrovskab,c, Xinjiang Wanga, Neelu Yadava, David W. Goodricha, Dhyan Chandraa,* a

Department of Pharmacology and Therapeutics, Roswell Park Cancer Institute, Elm and Carlton Streets, Buffalo, NY 14263, USA b OncoRay-National Center for Radiation Research in Oncology, Medical Faculty and University Hospital Carl Gustav €t Dresden and Helmholtz-Zentrum Dresden-Rossendorf, Fetscherstrasse, Dresden, Carus, Technische Universita Germany c German Cancer Consortium (DKTK) Dresden and German Cancer Research Center (DKFZ), Heidelberg, Germany

A R T I C L E

I N F O

A B S T R A C T

Article history:

Development of therapeutic resistance is responsible for most prostate cancer (PCa) related

Received 13 January 2016

mortality. Resistance has been attributed to an acquired or selected cancer stem cell

Received in revised form

phenotype. Here we report the histone deacetylase inhibitor apicidin (APC) or ER stressor

13 March 2016

thapsigargin (TG) potentiate paclitaxel (TXL)-induced apoptosis in PCa cells and limit accu-

Accepted 23 March 2016

mulation of cancer stem cells. TXL-induced responses were modulated in the presence of

Available online 31 March 2016

TG with increased accumulation of cells at G1-phase, rearrangement of the cytoskeleton, and changes in cytokine release. Cytoskeletal rearrangement was associated with modula-

Keywords:

tion of the cytoplasmic and mitochondrial unfolded protein response leading to mitochon-

Prostate cancer

drial dysfunction and release of proapoptotic proteins from mitochondria. TXL in

Anticancer drugs

combination with APC or TG enhanced caspase activation. Importantly, TXL in combina-

Unfolded protein response

tion with TG induced caspase activation and apoptosis in X-ray resistant LNCaP cells.

Combination therapy

Increased release of transforming growth factor-beta (TGF-b) was observed while phos-

Apoptosis

phorylated b-catenin level was suppressed with TXL combination treatments. This was

Mitochondria

accompanied by a decrease in the CD44þCD133þ cancer stem cell-like population, suggesting treatment affects cancer stem cell properties. Taken together, combination treatment with TXL and either APC or TG induces efficient apoptosis in both proliferating and cancer stem cells, suggesting this therapeutic combination may overcome drug resistance and recurrence in PCa. ª 2016 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

* Corresponding author. Tel.: þ1 (716) 845 4882; fax: þ1 (716) 845 8857. E-mail address: [email protected] (D. Chandra). http://dx.doi.org/10.1016/j.molonc.2016.03.007 1574-7891/ª 2016 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

950

1.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

Introduction

Prostate cancer (PCa) is the most diagnosed and second leading cause of cancer related death among American men (Siegel et al., 2013). Worldwide, PCa is the second most commonly diagnosed cancer and fifth leading cause of cancer related death among men (Ferlay et al., 2015). Metastasis is responsible for nearly all PCa mortality. First line therapy for metastatic PCa is androgen deprivation therapy (ADT). While beneficial responses are observed in most cases, recurrence is inevitable. Taxane-based chemotherapy is commonly used to treat men recurring from ADT, but this is not curative as resistant disease inevitably develops. Mechanisms of resistance to ADT fall into one of three general categories: restored androgen receptor (AR) signaling, bypass of AR signaling through use of other nuclear hormone receptors, or trans differentiation to a phenotype completely independent of AR signaling (Watson et al., 2015). Disease recurring in heavily treated patients exhibits several characteristics including resistance to apoptosis and increased drug efflux and/or metabolism, characteristics inherent in cancer stem cells (CSCs) or cancer-initiating cells (CICs) (Cojoc et al., 2015a; Fulda, 2013, 2015; Liu and Tang, 2011). Paclitaxel (TXL) is a common prescription for the treatment of malignant epithelial cancers including PCa. TXL suppresses microtubule dynamics during mitosis thereby causing G2/M phase cell cycle arrest, growth inhibition and apoptosis (Yvon et al., 1999). The overexpression of multidrug transporters as well as hypoxia-inducing factor-1 in cancer cells diminishes the efficacy of TXL (Das et al., 2015; Statkiewicz et al., 2014). Other mechanisms underlying TXL resistance in PCa include changes in the kinetics of microtubule formation and elevated levels of antiapoptotic proteins like Bcl-2 (Murray et al., 2012; O’Neill et al., 2011). Combination therapy is one of the key approaches to overcome drug resistance (Al-Lazikani et al., 2012). For example, TXL has been used in combination with other anticancer drugs like butyrate, bevacizumab, and the Akt inhibitor MK2206 to treat different types of cancer (Hata et al., 2014; Molife et al., 2014; Rivkin et al., 2014). In advanced and progressive PCa, TXL in combination with estramustine or carboplatin showed increased antitumor activity (Kelly et al., 2001, 2003). Synergistic therapeutic efficacy of TXL was observed in combination with KML001 (sodium meta-arsenite) in treatment resistant PCa (Zhang et al., 2012). In advanced, hormone-refractory PCa, combinations of TXL, carboplatin, etoposide, and estramustine have shown enhanced antitumor activity in preclinical studies (Smith et al., 2003). Despite improved responses, no currently used single or combination therapy is curative in patients with metastatic PCa. In order to overcome therapeutic resistance in the treatment of PCa, several unconventional compounds like apicidin (APC) and thapsigargin (TG) have been evaluated as potential anticancer drugs. APC is a cyclic tetra-peptide, which causes histone deacetylases (HDAC) inhibition, increases accumulation of cells at G1 phase in a dose-dependent manner, and blocks cell migration and invasion of cancer cells (Ahn et al., 2009, 2012). TG is the active ingredient in several chemotherapeutic pro-drug formulations that induce ionositol-3-

phosphate (IP3)-independent intracellular calcium (Ca2þ) release and apoptosis by disrupting intracellular free Ca2þ levels (Dubois et al., 2013). TG also causes cancer cells to accumulate in G1 phase (Beaver and Waring, 1996). Considering TXL, APC, and TG function through distinct mechanisms of action (mitotic inhibitor, HDAC inhibitor, and ER stressor, respectively), combination therapy using these drugs may provide new options for overcoming therapeutic resistance in the treatment of PCa. We hypothesize that resistance of metastatic PCa cells to TXL can be inhibited by abrogation of TXL-induced G2/M arrest since dividing cells are more sensitive to death (Mitchison, 2012; Valeriote and van Putten, 1975). Here, we observe that TG reverses cell cycle arrest induced by TXL leading to alterations in the cytoskeleton and mitochondria. Combining TXL with TG or APC caused cell death mainly through the mitochondrial pathway of apoptosis. Furthermore, TXL in combination with TG or APC reduces the level of CSCs or CICs, that are typically chemoresistant. These findings suggest that these combinations may effectively target CSCs/CICs with potential implications for treating patients suffering from recurrent, therapeutically resistant PCa.

2.

Materials and methods

2.1.

Cells and reagents

Androgen-dependent (LNCaP) and androgen-independent cell lines (DU145 and PPC1) were procured from American Type Culture Collection (ATCC, USA). E006AA and its highly tumorigenic derivative E006AA-hT (Koochekpour et al., 2004, 2014) were kindly provided by Dr. Koochekpour at Roswell Park Cancer Institute. X-ray irradiated cells were generated by Dr. Anna Dubrovska and were culture as described earlier (Cojoc et al., 2015b). Cells were maintained in RPMI-1640 medium, DMEM, and McCoy’s 5A medium supplemented with 2 mM L-glutamine, 10% FBS (Atlanta Biologicals, USA), and 1% penicillin and streptomycin. All cells were cultured at 37  C in a humidified atmosphere in the presence of 5% CO2. All human cell lines were authenticated using the Short Tandem Repeat (STR) DNA profiling every 6 months. Antibodies for p53, p21, a-tubulin, E-cadherin were procured from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). Antibodies for cyclin B1, Bcl-xL, Bcl-2 and caspases were obtained from BD Biosciences (San Jose, CA, USA). Antibody for actin was purchased from MP Biomedicals, LLC. Antibodies for Bax and Bak were obtained from Upstate (Billerica, MA, USA). Heat shock protein 60 (Hsp60), Hsp70, Hsp90, Hsp10, and phospho-b catenin (T41/TS45) were obtained from MP Scientific and Cell Signaling, respectively.

2.2.

Cell cycle analysis

Cell cycle analysis was carried out using propidium iodide (PI) staining according to the methods described earlier (Fried et al., 1976). In brief, treated or untreated cells were harvested

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

and fixed with 2 ml of 70% ethanol and kept on ice for 30 min. Cells were centrifuged at 3200 rpm for 3 min and pellets collected. Pellets were washed with 1 ml of FCM buffer containing 1 PBS, 0.5% BSA, 0.04 g/L EDTA, 1 g/L sodium azide. Cells were stained with DNA staining dye in Krishan buffer (containing 0.1% sodium citrate, 0.02 mg/ml RNAse, 0.05 mg/ ml PI, 0.2% NP 40 and one drop 1 N HCl) and then analyzed via flow cytometry (LSR II, BD Biosciences). Flow cytometry results were analyzed by ModFitLT software.

2.3.

Immunolabeling and confocal microscopy

Cytoskeleton arrangement (a-tubulin) was studied using immunolabeling followed by confocal microscopy as described previously (Gogada et al., 2011). Briefly, control or treated cells (6000 cells/coverslip) were stained live with MitoTracker Orange (100 nM) and DAPI (1 mg/ml) for 15 min at 37  C in CO2 incubator. Cells were fixed with 4% formaldehyde containing 5% sucrose for 10 min at RT followed by permeabilization with 1% Triton X-100 in PBS for 10 min. Following washing and blocking with 10% goat serum containing 1% Triton X-100 in 1 PBS, primary antibody was applied for overnight at 4  C. Cells were incubated with AlexaFluor-488-conjugated secondary antibody for 2 h. ProLong gold antifade reagent (Molecular Probes and ThermoFisher Scientific, Grand Island, NY, USA) was used as mounting medium. Fluorescence images were acquired using a laser-scanning confocal system on an inverted microscope equipped with an oil-immersion lens (Carl Zeiss, Thornwood, NY).

2.4.

951

450 nm using a plate reader (BioTek Microplate Readers, Winooski, USA).

2.6. Whole cell lysate preparation, subcellular fractionation, and Western blotting Preparation of whole cell lysates, mitochondrial and cytosolic fractions was performed as previously described (Chandra et al., 2002, 2007). Western blotting to determine the levels of p53, p21, cyclin B1, E-cadherin, Bax, Bak, Bcl-2, Bcl-xL, Hsp10, Hsp60, Hsp70, Hsp90, phospho b-catenin, and actin in control and treated cells was performed as described previously (Gogada et al., 2011). In brief, proteins were separated on dodecyl sulfate-polyacrylamide gel electrophoresis (SDSPAGE) procured from (Bio-Rad, Hercules, CA, USA) and transferred on to nitrocellulose membrane (Millipore, Bedford, MA, USA). Membranes were blocked in 5% nonfat milk for 30 min and washed with PBS-T (1 PBS and 0.05% Tween 20) and further incubated with respective primary antibody. Anti-mouse or anti-rabbit horseradish peroxidaseconjugated antibodies (Amersham Pharmacia Biotech, Piscataway, NJ) were used as secondary antibody, respectively. Actin, lactate dehydrogenase (LDH), and translocase of outer mitochondrial membranes 20 kDa (TOM20) were analyzed as protein loading controls for whole cell lysates, cytosolic and mitochondrial fractions, respectively.

2.7. Quantification of apoptosis and caspase activity measurement

Cytoskeleton analysis

For the quantitative analysis of cytoskeletal changes, actin polymerization was determined as described previously (Hart et al., 2007). Briefly, cells were seeded and treated with drugs alone or in combinations. At the end of treatment, cells were fixed by 3.7% formaldehyde. Cells were permeabilized using permeabilization buffer (20 mM HEPES, 300 mM sucrose, 50 mM NaCl, 3 mM MgCl2 and 0.1% Triton X-100). Cells were stained with FITC-phalloidin (Sigma Aldrich, St. Louis, MO, USA), washed with PBS, analyzed by flow cytometry (LSRIIB, BD Bioscience, San Jose, USA). The percentage change in actin polymerization was calculated using the following formula.

Actin polymerizationð%Þ ¼

Images were captured without refreshing medium using an inverted microscope (Carl Zeiss, Thornwood, NY). The percentage nonviable cells were determined on the basis of 0.4% trypan blue staining. For quantification of apoptosis, cells were treated with various anticancer agents or with vehicle for various time periods followed by staining with annexinV-Alexafluor 488/PI Kit (ThermoFisher Scientific, Grand Island, NY, USA) according to the manufacturer’s instructions. The stained cells were analyzed by flow cytometry (LSR II, BD Biosciences) collecting 10,000 events. Graphs were plotted using Win List 3D software. Caspase-3 (DEVDase) and caspase-9 (LEHDase) activities were measured as previously described (Chandra et al., 2002).

ðMean of treated cells  Mean of control cellsÞ  100 Mean value of control cells

2.5. Enzyme-linked immunosorbent assay (ELISA) for cytokines The levels interleukin-8 (IL-8), interferon-gamma (IFN-g), and transforming growth factor-beta1 (TGF-b1) were measured in the culture supernatant and lysates of control, TXL, APC, TG, TXL þ TG, and TXL þ APC treated cells at 24 h and 48 h time points using commercially available ELISA kits (eBioscience, USA). The detection limit for IL-8, IFN-g and TGF-b1 were 2, 4 and 156.3 pg/ml, respectively. The plates were read at

2.8. Analysis of mitochondrial ROS (mtROS), mitochondrial membrane potential (mtMP), and mitochondrial mass (Mito mass) using flow cytometry Flow cytometry was used to measure mtROS, mtMP and Mito mass in control and treated groups as described previously (Yadav et al., 2014). The mtROS, Mito mass, and mtMP were estimated in the control and treated groups using MitoSox Red, MitoTracker Green, and MitoTracker Orange probes (ThermoFisher Scientific, Grand Island, NY, USA),

952

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

respectively. Briefly, control and treated cells were harvested by centrifugation at 1000 rpm for 5 min. Cells were resuspended in 1 ml of 1 PBS and centrifuged for 5 min at 2500 rpm. The pellets were stained with respective probes and incubated for 30 min in the dark at 37  C. Cells were harvested at 1500 rpm for 5 min and pellets were re-suspended in 1 PBS and 10,000 cells were analyzed by flow cytometry (LSR II, BD Biosciences). Data were analyzed with Win List 3D 7.1 software and presented as the fold change of geometric mean in comparison to untreated control.

2.9.

Real-time PCR analysis

Real time PCR was used to quantify the expressions of CHOP, C/ EBP, and Hsp70. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and actin were used as an internal reference gene to normalize expression of the apoptotic genes of interest. The primer sequences for multiple molecules are listed in Supplemental Table S1. Relative quantification of apoptosisrelated genes was analyzed by the comparative threshold cycle (Ct) method. For each sample, the Ct value of the apoptotic gene was normalized using the formula: DCt ¼ Ct (target gene)  Ct (GAPDH). To determine relative expression levels, the following formula was used: DDCt ¼ DCt (treated)  DCt (control). The value was used to plot the expression of apoptotic genes using the formula 2DDCt.

2.10.

Caspase-3 knock down using lentiviral shRNA

Lentiviral shRNA for silencing caspase-3 gene in LNCaP cells were obtained from the Roswell Park Cancer Institute shRNA Shared Resource Facility. The sequences for shRNA caspase3 were 50 -GGAAACATTCAGAAACTTGA-30 . LNCaP cells were plated in each well of a 12-well plate and infected the next day at a multiplicity of infection (MOI) of 3 with lentiviral particles designed to express shRNA targeting caspase-3. Empty lentiviral particles were used as a negative control. After 48 h, the cells were treated with puromycin for positive selection. 48 h later, cells were harvested for analysis.

2.11.

Gelatin zymography and total b-catenin analyses

Matrix metalloproteinase (MMPs) activity was analyzed using gelatin zymography as described previously (Issa et al., 2009). The gel was developed according to manufacturer’s instructions (Bio-Rad, PA, USA) and was stained using Coomassie Brilliant Blue R-250. Images were quantified by densitometry (Syngene Gel Doc System, Frederick, MD, USA). Intracellular total b-catenin expression was estimated using PE-tagged antibody (Cell Signaling Technology). Levels of fluorescence were captured by flow cytometry (LSR II, BD Bioscience) and data were analyzed by Win List 3D software. A total of 10,000 events were captured for each sample.

2.12.

Immunophenotypic characterization of CSCs

The surface expression of CD133, CD44, and CD24 was analyzed by immunophenotyping. Briefly, 1  106 cells were seeded in a 10 cm dish and incubated for 24 h in CO2 incubator at 37  C prior to treatment. Cells were treated with TXL, APC,

and TG alone or in combinations for 24 h. Cells were labeled with APC-conjugated anti-human CD133 (Miltenyi Biotec, San Diego, USA). FITC-conjugated mouse anti-human CD44 (BD Pharmingen, BD Biosciences, USA) and/or APC-conjugated CD24 in a buffer containing (1 PBS, 0.5% BSA and 2 mM EDTA) for 15 min at 4  C. Labeled cells were re-suspended in 1 PBS (pH 7.4), and analyzed by flow cytometer (LSR II, BD Bioscience). Unstained cells served as negative controls. A total of 100,000 events were captured for each sample. Estimation of ALDH bright (ALDHbr) cell populations was performed 24 h after treatment using the Aldefluor kit (Stem Cell Technology). Data were plotted using Win List 3D software.

2.13.

Statistical analysis

Results are presented as mean  standard deviation (SD) of data from at least three independent experiments. Significant differences between means were assessed via analysis of variance (ANOVA) using GraphPad Prism Version 6. A p < 0.05 value was deemed significant.

3.

Results

3.1. TG and APC modulate TXL-induced changes at the G1 and G2 phases of the cell cycle The anticancer activity of TXL associates with cell cycle arrest and apoptosis (Das et al., 2001), but effective doses for efficient anticancer effects are not well defined. We first investigated the effects of various concentrations of TXL on cell death and the cell cycle. TXL-induced cell death correlates with increasing concentrations ranging from 30 nM to 30 mM (Supplementary Figure S1A). Similarly, TXL induced a significant increase in the percentage of cells in G2/M phase (Supplementary Figures S1B and S2A). Proliferating cancer cells are more prone to undergo cell death compared to cell cycle arrested cells (Mitchison, 2012). We tested whether inhibition of TXL-induced cell cycle arrest enhances sensitivity to TXL investigating if changes in cell cycle phase distribution enhance cell death upon TXL treatment. Since ER stressors and HDAC inhibitors modulate cell cycle parameters (Bourougaa et al., 2010; Di Fazio et al., 2012), we treated LNCaP PCa cells with TXL in combination with either TG or APC and analyzed changes in cell cycle phase distribution using propidium iodide (PI) staining after 24 h drug exposure. We observed reversal of cell cycle changes when TXL was used in combination with TG or APC (Figure 1AeD; Supplementary Figure S2B). For example, reduction in G1 phase accumulation upon TXL treatment was inhibited when cells were treated with TXL þ TG or TXL þ APC (Figure 1A and B). TXL þ TG treatment abrogated increased G2/M phase accumulation observed upon treatment with TXL alone (Figure 1A and D). To understand the mechanisms underlying these changes, we determined the levels of key cell cycle related proteins upon treatment. We observed that tumor suppressor p53 and its target gene p21 were upregulated upon single agent treatment with each agent, whereas they were not significantly induced when used in combination (Figure 1E). It is interesting to note that TG alone

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

953

Figure 1 e TG and APC attenuate cell cycle arrest via reduced expression of p53, p21 and cyclin B1. LNCaP cells were treated with paclitaxel (TXL, 30 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (30 mM) D APC (1 mM), and TXL (30 mM) D TG (5 mM), respectively for a period of 24 h. Cell cycle phase distribution was measured by propidium iodide (PI) staining and flow cytometry. (A) Images demonstrating diploid (Dip) G1, S, and Dip G2-phase. Percentage changes as compared to control cells were plotted for Dip G1 (B), S (C), and Dip G2-phase (D). (E) Whole cell lysates (WCL) obtained from control and treated groups (for 24 h) were subjected to Western blot analysis of p53, p21 and cyclin B1. Actin serves as a loading control. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

did not induce p21 even though p53 was upregulated, suggesting p53 might be playing a role in cell death induction rather than promoting cell cycle arrest. In addition, TXLinduced cyclin B1 that plays a critical role in the G2 and M

phases of the cell cycle was reduced in the presence of TG and APC (Figure 1E). These findings suggest that TG and APC may restore the cancer cell cycle by attenuating TXLinduced cell cycle arrest.

954

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

3.2. Enhanced caspase activation and cell death in the presence of TXL plus TG or APC To understand whether TXL combinations induce caspase activation, we measured caspase-9 and -3 activities in response to single and combined agents in LNCaP cells. Caspase-9 activity was increased compared to controls in response to TXL plus TG and TXL plus APC (Figure 2A). A robust increase in caspase-3 activity was observed in response to both TXL plus TG and TXL plus APC supporting the idea that in the presence of TG or APC, TXL efficiently induces caspase activation (Figure 2B). To test whether caspase-3 is the main executioner caspase involved in response to TXL combinations, we treated caspase-3-silenced cells with drug alone or their combinations. Caspase-3-silencing inhibited DEVDase activity in response to single agent and TXL combinations (Figure 2C). We further investigated whether increased caspase-3 activation corresponds with cell death in response to single agents alone or TXL combinations. Annexin V/PI analysis demonstrated that TXL þ APC and TXL þ TG combinations significantly induced higher level of cell death compared to single agents alone (Figure 2D and Figure S3). Quantification of cell death using Trypan blue demonstrated a similar time-dependent increase in cell death upon combination treatment in Du145 cells (data not shown). A similar trend was observed in PPC-1, E006AA

and E006AA hT cells at 24 and 48 h following combined exposures to TXL with APC or TG (data not shown). Together, TXL in combination with APC or TG induces executioner caspase-3 activation that leads to increased cell death.

3.3. Combined exposure to TXL and TG or TXL and APC induces mtROS production, and alters mtMP and Mito mass leading to the release of proapoptotic proteins To understand the underlying mechanism of cell death, we first measured whether increase in cancer cell death corresponds with the generation of ROS, which can induce both cellular proliferation and death (Schieber and Chandel, 2014). MtROS levels were increased in response to TXL, TXL þ APC and TXL þ TG compared to control (Figure 3A). TXL induced mtMP and this response was significantly reduced in the presence of TG and APC. We also observed that APC or TG alone decreased mtMP by 3% and 40%, respectively in comparison to control (Figure 3B). Treatment of cancer cells with single agents or in combination reduced Mito mass to some extent relative to vehicle control (Figure 3C). Increased ROS production also associates with induction of mitochondrial apoptosis (Schieber and Chandel, 2014), therefore, we evaluated the levels of pro- and anti-apoptotic proteins that regulate mitochondrial apoptosis. We observed reduced expression of prosurvival proteins Bcl-2 and Bcl-xL

Figure 2 e TXL combinations induce caspase-3 dependent apoptosis. LNCaP cells were treated with paclitaxel (TXL, 1 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (1 mM) D APC (1 mM), and TXL (1 mM) D TG (5 mM), respectively, for a period of 24 h. Caspase-9 (A), and caspase-3/7 (B) activities were estimated using LEHD-AFC and DEVD-AFC as substrate, respectively, and presented as fold change compared to control. (C) Caspase-3 activity in response to drugs alone or their combinations using caspase-3 silenced cells or Mock shRNA (control shRNA) LNCaP cells. (D) LNCaP cells were treated with indicated drugs alone or in combination as described in panel A for 36 h and percent cell death was quantified by Annexin V/PI staining. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment. Casp-3, caspase-3.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

955

Figure 3 e TXL combination treatment increases ROS production, and modulates mtMP and Mito mass. LNCaP cells were treated with paclitaxel (TXL, 30 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (30 mM) D APC (1 mM), and TXL (30 mM) D TG (5 mM) for a period of 24 h. (A) Mitochondrial ROS (mtROS), (B) mitochondrial membrane potential (mtMP), (C) and mitochondrial mass (Mito mass) was measured by flow cytometry after live staining with MitoSox-Red, MitoTracker Orange, and MitoTracker Green, respectively. Data are presented as fold change compared to untreated control cells. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

upon exposure to TXL in combination with TG or APC (Figure 4A). The expression of Bak and Bax at mitochondria was not modulated in response to TXL plus APC treatment (Figure 4B). Unchanged expression of proapoptotic proteins and decreased expression of antiapoptotic effects encouraged us to analyze the release of Smac/DIABLO and cytochromec from mitochondria following exposure to drugs alone or TXL combinations at 24 h using purified cytosol and mitochondria. We observed higher levels of cytochrome c and Smac release upon TXL and TG combination compared to TG alone, whereas similar levels of cytochrome c and Smac release were observed with either APC alone or in APC and TXL combination. Smac and cytochrome c was highly depleted in the mitochondrial fraction, suggesting enhanced release from mitochondria in response to APC and TXL combination (Figure 4C and D). Conclusively, TXL with TG or APC induces ROS generation with subsequent release of proapoptotic proteins such as cytochrome c and Smac.

3.4. TXL-induced cytoskeletal and mitochondria rearrangements are altered by APC and TG treatment TXL binds microtubules and abnormally stabilizes the dynamics of the cytoskeleton thereby causing cell cycle arrest and apoptosis (Arnal and Wade, 1995; Yvon et al., 1999). To understand whether APC and TG modulate TXL-induced cytoskeletal changes, we elucidated the structural distribution of

a-tubulin after 24 h exposures to drugs alone or in combination. Confocal microscopy analysis of a-tubulin demonstrated a well-organized filamentous structure with a portion of atubulin co-localized with mitochondria in control untreated LNCaP cells (Figure 5A). TXL treatment caused uneven atubulin distribution in cells, and loss of co-localization with mitochondria (Figure 5A). This suggests that TXL treatment leads to dissociation of microtubule from mitochondria. APC and TG as single agent did not modulate microtubule structure or association of mitochondria with a-tubulin. TXL in combination with TG or APC showed circumferential or unevenly localized a-tubulin accompanied by reduction in Mito mass (Figure 5A). For quantitative analysis of cytoskeletal changes, untreated and treated cells were stained with FITCphalloidin, which selectively label F-actin, followed by flow cytometry analysis. We observed significant downregulation of actin polymerization in both combination treatments (Figure 5B). Alteration of cytoskeletal organization is closely associated with E-cadherin loss (Chen et al., 2014), thus we also evaluated E-cadherin expression in response to these treatments. Fulllength E-cadherin expression was elevated upon exposure to TXL or TG alone in comparison to control (Figure 5C). TXLmediated E-cadherin upregulation was attenuated in the presence of TG or APC, but was still higher than that of untreated cells. Disruption of tight junctions and induction of proinflammatory cytokines including IL-8 and IFN-g associates

Figure 4 e Release of apoptogenic proteins from mitochondria upon exposure to TXL combinations associates with decreased antiapoptotic proteins and increased proapoptotic proteins. LNCaP cells were treated with paclitaxel (TXL, 1 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (1 mM) D APC (1 mM), and TXL (1 mM) D TG (5 mM), respectively, for a period of 24 h. (A) Equal amounts of protein were subjected to Western blotting for Bcl-2 and Bcl-xL. Actin serves as a loading control. (B) LNCaP cells were treated with various drugs for 24 h. Mitochondrial fractions were isolated using differential centrifugation. Western blot analysis was carried out for Bak and Bax. Actin serves as loading control. (C and D) LNCaP cells were treated with various drugs for 24 h. Cytosolic and mitochondrial fractions were isolated using differential centrifugation. Western blot analysis was carried out for Smac and cytochrome c (Cyt. c). LDH and TOM20 serve as markers of cytosolic and mitochondrial fractions, respectively. Actin serves as a loading control.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

with survival and invasion of cancer cells (Brysse et al., 2012; Feigin and Muthuswamy, 2009; Naydenov et al., 2013). Combination treatment at 24 h reduced expression of CXCL8 compared to treatment with TXL alone (Figure 5D). Increased release of IL-8 was observed in response to TG or APC

957

exposure alone or in combination with TXL, compared to controls (Figure 5E). We observed significant release of IFN-g in response to TXL þ TG and TXL þ APC combinations (Figure 5F). Since IL-8 and IFN-g function as prosurvival proteins, increased release of these cytokines may modulate

Figure 5 e TXL combinations induce cytoskeletal rearrangements leading to increased release of IL-8 and IFN-g. LNCaP cells were treated with paclitaxel (TXL, 1 or 30 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (1 or 30 mM) D APC (1 mM), and TXL (1 or 30 mM) D TG (5 mM), respectively for a period of 24 h. (A) Equal numbers of treated or untreated LNCaP cells were stained live with MitoTracker Orange and DAPI. Finally, cells were immunostained for a-tubulin and images were captured by confocal microscopy. (B) LNCaP cells treated with drugs alone or in combination were analyzed for actin polymerization using FITC-tagged phalloidin and flow cytometry. (C) E-cadherin (E-cad) expression was analyzed by Western blotting. Actin and Ponceau staining serve as loading controls. (D) Equal amounts of RNA were subjected to real time PCR analysis for the CXCL-8 gene. GAPDH serves as an internal control. (E) IL-8 release in the culture supernatant was quantified using ELISA at 24 and 48 h. The concentration of IL-8 (pg/mL) in culture supernatants was determined by comparing the OD to an IL-8 standard curve. (F) IFN-g release into the culture supernatant was quantified as in E. Data were normalized by cell number. Bar represents 5 mm. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

958

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

survival proteins including heat shock proteins or Hsps (Singh and Lokeshwar, 2009; Wilson et al., 2012). Together, TXL in combination with APC or TG induces cytoskeletal rearrangements causing changes in cellular structure and function.

3.5. APC enhances Hsp70 accumulation whereas TG reduces Hsp70 level in mitochondria Cellular stress in response to anticancer agents induces an unfolded protein response (UPR) which helps maintain cellular structure and function (Harrington et al., 2015). UPR is regulated by cytosolic, endoplasmic reticulum (ER), and mitochondria proteins through a mechanism requiring heat shock proteins (Hsps) including Hsp60, Hsp60 cofactor Hsp10, Hsp70, and Hsp90 (Jovaisaite et al., 2014). Expression of these Hsps associates with transcription factors like C/EBP and the homologous protein CHOP (Jovaisaite et al., 2014; Zhao et al., 2002). Real-time PCR analysis demonstrated significantly increased expression of C/EBP and CHOP upon treatment with TG alone and in combination with TXL (Figure 6AeB). We observed upregulation of Hsp70 in response to both APC alone and APC in combination with TXL, at the transcriptional and protein levels (Figure 6C and D). Hsp10 protein level was reduced in response to APC alone and TG alone, but TG or APC did not significantly modulate total cellular level of Hsp10 in combination with TXL (Figure 6D). These drug treatments did not significantly modulate accumulation of Hsp60 and Hsp90 proteins. APC treatment, either alone or in combination with TXL caused increased accumulation of Hsp70 while TG appeared to reduce it. Subcellular analysis demonstrates Hsp60 was downregulated whereas Hsp90 was unchanged by TG alone or in combination treatments (Figure 6E). Hsp60 cofactor Hsp10 showed reduced mitochondrial accumulation in response to single agents or their combinations with TXL (Figure 6F). Interestingly, drug treatment also appeared to affect the subcellular localization of Hsp70. TG alone and TXLþTG reduced accumulation of Hsp70 in mitochondria whereas APC alone and its combination with TXL enhanced Hsp70 accumulation in mitochondria (Figure 6F). These findings suggest that TG or APC in combination with TXL alter mitochondrial UPR (UPRmt) and cytosolic UPR (UPRcyto) through modulation of Hsp10, Hsp60, Hsp70, and Hsp90.

3.6. TXL in the presence of APC or TG modulates matrix metalloproteinase (MMP), TGF-b, and b-catenin phosphorylation

member of TGF-b family protein, regulates MMPs through p38 MAPK and ERK1/2 in highly invasive cancer cells (Gomes et al., 2012). Thus we measured TGF-b1 level in cellular lysates and release in the cell culture supernatant. Decreased levels of TGF-b1 were found in cancer cells treated with TXL þ TG and TXL þ APC (Figure 7B). Secretion of TGF-b1 was timedependent (Figure 7C and D). TGF-b1 is a modulator of b-catenin (Gomes et al., 2012), and higher levels of total b-catenin were evident in all treatment groups with the highest levels observed in TXL þ TG treated cells (Figure 7E). Phosphorylated b-catenin at residues T41/S45 is predominantly nuclear and continues active transcription in Wnt-stimulated cells (Gerlach et al., 2014; Maher et al., 2010). Indeed, diminished expression of T41/S45 phosphorylated b-catenin was evident in TG alone, TXL plus TG, and TXL plus APC treated cells (Figure 7F). Taken together, expression of active MMPs, TGFb1, and T41/S45 phosphorylated b-catenin indicate an adverse growth environment for CSCs.

3.7. Modulation of CSCs cell populations indicates potential efficacy of combination therapy Increased b-catenin/Wnt signaling contributes to self-renewal of CSCs (Mao et al., 2014). Thus we measured the levels of CSC populations using CD133þCD44þ and CD44þCD24 markers. We observed reduced CD133þCD44þ CSCs in response to TXL plus TG as well as TXL plus APC treatment (Figure 8A) compared to TXL alone. Combination treatments also decreased the abundance of CD133þCD44 and CD133CD44þ cells (Figure 8A). Enrichment of CD44þCD24 CSCs was observed in TXL alone and TXLþAPC treatment (Figure 8B), whereas TXL þ TG combination showed reduced enrichment, similar to C133þCD44þ cells. We also observed elevated numbers of ALDH bright (ALDHbr) cells in response to TXL, APC, and TXL plus APC combination at early stages of exposure, while decreased levels were evident in TXL plus TG (Figure 8C). To further support our hypothesis that TXL in combination with TG induces apoptosis in radiation-resistant cells that possess CSCs characteristics (Cojoc et al., 2015b), we observed elevated caspase-3 and -9 activities in response to TXL combinations (Figure 9A and B; and data not shown). Increased apoptosis in radiation-resistant cells upon treatment with TXL combinations establish that caspase activation in CSCs induces apoptosis (Figure 9B and C). Thus, therapeutic efficacy of TXL could be enhanced in the presence of TG.

4. Increased death in bulk proliferating cancer cells prompted us to evaluate effects of TXL combinations on CSCs, which play an important role in the acquisition of epithelialemesenchymal transition (EMT) phenotype (Scheel and Weinberg, 2012). E-cadherin is an important mediator of EMT and regulates the invasive function of MMPs (Lu et al., 2014; Nawrocki-Raby et al., 2003). Thus, we characterized the release of MMP-2, and MMP-9 upon exposure to TXL, APC or TG alone and in combination. Gelatin zymographic analysis demonstrated MMP-9 and MMP-2 activity were upregulated upon TG alone or TGþTXL treatment, but not by other treatments (Figure 7A, and data not shown). TGF-b1, a major

Discussion

The development of resistance to current PCa therapy, including TXL, associates with expression of multidrug resistance genes, reduced cellular proliferation, the presence of resistant CSCs populations, and cell cycle arrest (Fulda, 2013; Mitchison, 2012; Murray et al., 2012; O’Neill et al., 2011; Rycaj and Tang, 2014). Since TXL induces both cell cycle arrest and preferentially kills fast dividing cells (Mitchison, 2012), restoration of cell cycle progression via attenuation of TXLinduced cell cycle arrest may lead to enhanced efficacy of TXL therapy for patients with PCa. Our study, for the first time, demonstrates that TG alters G2/M-phase arrest induced

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

959

Figure 6 e TXL combinations induce both cytosolic and mitochondrial unfolded protein response. LNCaP cells were treated with paclitaxel (TXL, 1 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (1 mM) D APC (1 mM), and TXL (1 mM) D TG (5 mM), respectively for a period of 24 h. Equal amounts of RNA were subjected to real time PCR analysis for C/EBP (A), CHOP (B), and Hsp70 (C). (D) Equal amounts of proteins were used for Western blot analysis of heat shock proteins. Actin serves as a loading control. (E and F) LNCaP cells were treated with the indicated drugs for 24 h. Cytosolic and mitochondrial fractions were isolated using differential centrifugation. Western blot analysis was carried out for Hsp10, Hsp60, Hsp70, and Hsp90. LDH and TOM20 serve as markers for cytosolic and mitochondrial fractions, respectively. Actin serves as loading control. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

by TXL in PCa cells. In addition, CSCs populations are altered by TXL þ TG and TXL þ APC, suggesting CSCs are sensitive to the combination treatment. These combination therapies may provide a means to overcome TXL resistance in the treatment of PCa. How does TG attenuate TXL-induced cell cycle arrest? MtROS induce phosphoinositide 3-kinase (PI3K) and Akt

signaling, and stabilize hypoxia-inducing factors (Connor et al., 2005; Pelicano et al., 2006; Sullivan and Chandel, 2014; Waypa et al., 2010). Collectively, these changes may promote cellular proliferation. The observed increased production of mtROS upon treatment with TXL combinations may thus support the proliferative potential of cancer cells. Indeed, our findings suggest that TXL-induced cell cycle arrest was

960

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

Figure 7 e TXL combinations modulate stemness via decreased expression of MMPs and phospho b-catenin, and increased release of TGF-b1. LNCaP cells were treated with paclitaxel (TXL, 30 mM or 1 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (30 mM or 1 mM) D APC (1 mM), and TXL (30 mM or 1 mM) D TG (5 mM), respectively for a period of 24 h. Equal amounts of culture supernatant were subjected to gelatin zymographic analysis for determination of total MMPs. (A) Representative images of pro-MMP-9 and t-MMP-2 (total MMP) are shown. (B) Total TGF-b1 level in cytosol upon exposure 1 mM dose of TXL alone and TXL combinations or release of TGF-b1 in supernatant at 24 h (C) and 48 h (D) after treatment using 30 mM dose of TXL and its combinations were quantified using ELISA. The concentration of TGF-b1 in culture supernatants was determined by comparing to a TGF-b1 standard curve. The levels of TGF-b1 were normalized to protein concentration or number of cells. (E) Flow cytometry analysis of total b-catenin expression in different treatment groups (1 mM of TXL or combinations). Data were analyzed by Win List software. (F) Western blot analysis of phospho b-catenin (1 mM of TXL or combinations) was performed using equal amounts of protein. Actin serves as a loading control. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

severely attenuated in the presence of TG or reduced in the presence of APC. Since cell cycle associated proteins, including p53/p21/cyclin B, regulate various phases of the cell cycle (Eymin and Gazzeri, 2010), inhibition of TXLinduced expression of p53/p21/cyclin B by TG or APC is consistent with the notion that TXL þ TG or TXL þ APC combinations could restore PCa cells to an actively cycling state which is more sensitive to apoptosis.

Inhibition of the antioxidant system can cause increased levels of mtROS (Neumann et al., 2003; Sullivan and Chandel, 2014) leading to oxidative stress. Therefore, enhanced levels of mtROS upon attenuation of the antioxidant system lead to severe oxidative stress and cell death. Indeed, increased mtROS concomitantly associates with efficient release of apoptogenic proteins such as cytochrome c and Smac leading to caspase-dependent apoptosis in response to TXL combinations. Low levels of caspase

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

961

Figure 8 e TXL combinations decrease CSCs. LNCaP cells were treated with paclitaxel (TXL, 1 mM), apicidin (APC, 1 mM), thapsigargin (TG, 5 mM), TXL (1 mM) D APC (1 mM), and TXL (1 mM) D TG (5 mM), respectively for a period of 24 h. Immunophenotyping of CD24, CD44, and CD133 cell populations were carried out using CD24-APC, CD44-FITC, and CD133-APC tagged antibodies, respectively, and flow cytometry. Data were analyzed using Win List software (A). Representative dot plot showing CD44LCD133D, CD44DCD133L, and CD44DCD133D populations in control and treated groups. Quantification of CSC-like cells is also plotted relative to controls. (B) Representative dot plot showing CD44LCD24D, CD44DCD24L, and CD44DCD24D populations in control and treated groups. Quantification of CSCs is also plotted relative to controls. (C) ALDHbr cell populations were measured using flow cytometry at 24 h. Data were analyzed by Win List software. Data are presented in fold change compared to controls. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

962

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

Figure 9 e Effect of combinational therapy on X-ray irradiated LNCaP cells. X-ray resistant LNCaP cells were treated with paclitaxel (TXL, 1 mM), thapsigargin (TG, 5 mM), and TXL (1 mM) D TG (5 mM), respectively for a period of 24 h and 48 h. Images were taken (A) and equal amounts of proteins were subjected to caspase-3 activity measurements (B). Annexin/PI staining at 24 h or 48 h (C) was measured by flow cytometry and data analyzed by Win List software. Early and late apoptosis are presented as fold change compared to respective controls. Data are mean ± SD (n [ 3). *p < 0.05 compared with untreated cells, and $p < 0.05 compared to TXL alone treatment.

activation due to limited mitochondrial permeabilization and in some cases caspase-dependent apoptosis also induces cellular proliferation and tumorigenesis (Ichim et al., 2015; Ryoo and Bergmann, 2012). It is interesting to note that TXL combination with TG or APC induces IFN-g, which is known to enhance mitochondrial apoptotic signaling (Barton et al., 2005). For execution of mitochondrial apoptosis, TXL combinations induce efficient cytochrome c and Smac release, which was partially facilitated by the reduced level of Bcl-2 and Bcl-xL. Proapoptotic activities of Bax and Bak are inhibited by antiapoptotic proteins including Bcl-xL (Czabotar et al., 2014). Therefore, unchanged levels of proapoptotic channel proteins (e.g. Bax and Bak) along with reduced Bcl2 and BclxL will induce efficient permeabilization of the outer mitochondrial membrane leading to the release apoptogenic proteins (cytochrome c and Smac) causing enhanced apoptosis. Increased cell death in either androgen-independent or -dependent PCa cells indicates the potential efficacy of combinatorial therapy irrespective of PCa disease stage. These results indicate different mechanisms of cell death induced by these two TXL combinations. Cytosolic, ER, and mitochondrial UPR are regulated by heat shock proteins including Hsp60, Hsp70, and Hsp90 (Jovaisaite et al., 2014). Mitochondrial rearrangement as well as modulation of mtROS, mtMP, and Mito mass in response to TXL combinations indicates possible alterations of UPRmt, while upregulated Hsp70 indicates altered UPRcyto and UPRmt. Since CHOP and C/EBP are required for both ER UPR and UPRmt (Horibe and Hoogenraad, 2007; Marciniak et al., 2004),

upregulation of C/EBP and CHOP by TG alone and TXL plus TG suggest elevation of both cellular UPR and UPRmt. TG induces ER stress due to decreased ER calcium levels and loss of chaperone activity leading to the accumulation of unfolded proteins causing ER UPR (Davenport et al., 2007). In contrast, APC-induced UPR is mediated via Hsp70 expression, which was elevated by APC alone or TXLþAPC treatments. Thus both TXL combinations induce UPR but by different mechanisms. UPR causes cytoskeletal rearrangement and may affect cell cycle arrest (Brewer et al., 1999; Sharon-Friling et al., 2006) and TXL binds with tubulin leading to cytoskeleton rearrangement in cells (Yang et al., 2009). We observed significant change in the cytoskeletal structure during combination therapy. Furthermore, tubulin is an inherent component of mitochondrial membranes and specifically interacts with the voltage-dependent anion channel (Carre et al., 2002), which indicates a correlation between cytoskeletal rearrangement and mitochondrial dysfunction. Cytoskeleton and cadherin family proteins are major components of cell adhesion and are important regulators of cellular proliferation and apoptosis (Hannigan et al., 2005). In the present study, reduced expression of TXL-induced E-cadherin in the combination treatments indicates E-cadherin is a potential target for TXL in combination with either APC or TG. E-cadherin is an important mediator of MMP functions in tumor cells (Nawrocki-Raby et al., 2003), thus expression of E-cadherin shows inverse relation with the expression of MMP-2 and MMP-9. Increased release of TGF b1 upon exposure to TXL combinations indicates the tumor suppressive effects

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

of TXL in the presence of APC and TG. Similar to our results, mifepristone and tamoxifen induced secretion of TGF-b1 has been reported to reduce survival and proliferation of LNCaPC4 cells (Liang et al., 2002). Down-regulated phosphorylation of b-catenin further validates our hypothesis that TXL plus APC or TXL plus TG may reduce CSCs population. Indeed, our study demonstrated decreased enrichment of CD133þCD44þ CSCs, thus indicating the potential efficacy of TXL combinations. Although CD133, CD44, and ALDH1, are considered as markers for CSCs (Okudela et al., 2012), CD44þCD133þ may also be considered a potentially CSC marker for prostate CSCs (Collins et al., 2005). CD44þCD24 cells were relatively rare in a PCa cell line, making up only 0.04% of the cells (Hurt et al., 2008). Further, decreased levels of CD44þCD24 cell population in TXL plus TG, while enriched in TXL plus APC suggest different mechanisms of action by the two combinations. Similar to CD44þCD24, TXL plus APC treatment demonstrate enrichment of ALDHbr cell population at early time period. These findings also suggest that death of non-CSCs may have contributed to enrichment of CSCs populations. Together, our findings highlight several seminal points that have long-term implications for the treatment of metastatic PCa. Mechanistically, TXL combinations trigger increased ROS production and mitochondrial stress causing UPRmt. UPRmt increases in cell proliferation followed by mitochondrial apoptotic cell death. Alteration of TXL-induced cell cycle arrest will enhance the therapeutic response of bulk cancer cells (i.e., non-CSCs). On the other hand, reduced enrichment of CSCs supports the notion that TXL combinations promote removal of these cancer-initiating cells. Removal of bulk cancer cells and CSCs may lead to enhanced therapeutic responses and potential cures. Indeed, TXL combination induces caspase-dependent apoptotic cell death in radiation-resistant CSCs as shown previously (Cojoc et al., 2015b). It is also interesting to note that TXL combinations target CD133þCD44þ CSCs, which may represent early stem cells or tumor progenitor cells in PCa (Collins et al., 2005; Tang, 2012). Therefore, removal of both early CSCs and nonCSCs may facilitate in designing novel strategy to enhance efficacy of current anticancer agent TXL leading to cure of metastatic PCa cells.

Conflict of interest Authors declare no potential conflict of interest.

Acknowledgments This work was supported in part by the National Cancer Institute of the National Institutes of Health under Award Number R01CA160685, the American Cancer Society Research Scholar Grant RSG-12-214-01 e CCG, and U.S. Department of Defense under Award Number W81XWH-14-1-0013 to DC; and the National Cancer Institute Center Support Grant P30 CA016056 to the Roswell Park Cancer Institute. We apologize to those

963

colleagues whose publications could not be cited due to space constraints.

Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.molonc.2016.03.007. R E F E R E N C E S

Ahn, M.Y., Kang, D.O., Na, Y.J., Yoon, S., Choi, W.S., Kang, K.W., Chung, H.Y., Jung, J.H., Min do, S., Kim, H.S., 2012. Histone deacetylase inhibitor, apicidin, inhibits human ovarian cancer cell migration via class II histone deacetylase 4 silencing. Cancer Lett. 325, 189e199. Ahn, M.Y., Lee, J., Na, Y.J., Choi, W.S., Lee, B.M., Kang, K.W., Kim, H.S., 2009. Mechanism of apicidin-induced cell cycle arrest and apoptosis in Ishikawa human endometrial cancer cells. Chem. Biol. Interact. 179, 169e177. Al-Lazikani, B., Banerji, U., Workman, P., 2012. Combinatorial drug therapy for cancer in the post-genomic era. Nat. Biotechnol. 30, 679e692. Arnal, I., Wade, R.H., 1995. How does taxol stabilize microtubules? Curr. Biol. e CB 5, 900e908. Barton, C., Davies, D., Balkwill, F., Burke, F., 2005. Involvement of both intrinsic and extrinsic pathways in IFN-gamma-induced apoptosis that are enhanced with cisplatin. Eur. J. Cancer 41, 1474e1486. Beaver, J.P., Waring, P., 1996. Thapsigargin induces mitochondrial dysfunction and apoptosis in the mastocytoma P815 cell line and in mouse thymocytes. Cell Death Differ. 3, 415e424. Bourougaa, K., Naski, N., Boularan, C., Mlynarczyk, C., Candeias, M.M., Marullo, S., Fahraeus, R., 2010. Endoplasmic reticulum stress induces G2 cell-cycle arrest via mRNA translation of the p53 isoform p53/47. Mol. Cell 38, 78e88. Brewer, J.W., Hendershot, L.M., Sherr, C.J., Diehl, J.A., 1999. Mammalian unfolded protein response inhibits cyclin D1 translation and cell-cycle progression. Proc. Natl. Acad. Sci. U. S. A. 96, 8505e8510. Brysse, A., Mestdagt, M., Polette, M., Luczka, E., Hunziker, W., Noel, A., Birembaut, P., Foidart, J.M., Gilles, C., 2012. Regulation of CXCL8/IL-8 expression by zonula occludens-1 in human breast cancer cells. Mol. Cancer Res. e MCR 10, 121e132. Carre, M., Andre, N., Carles, G., Borghi, H., Brichese, L., Briand, C., Braguer, D., 2002. Tubulin is an inherent component of mitochondrial membranes that interacts with the voltagedependent anion channel. J. Biol. Chem. 277, 33664e33669. Chandra, D., Choy, G., Tang, D.G., 2007. Cytosolic accumulation of HSP60 during apoptosis with or without apparent mitochondrial release: evidence that its pro-apoptotic or prosurvival functions involve differential interactions with caspase-3. J. Biol. Chem. 282, 31289e31301. Chandra, D., Liu, J.W., Tang, D.G., 2002. Early mitochondrial activation and cytochrome c up-regulation during apoptosis. J. Biol. Chem. 277, 50842e50854. Chen, A., Beetham, H., Black, M.A., Priya, R., Telford, B.J., Guest, J., Wiggins, G.A., Godwin, T.D., Yap, A.S., Guilford, P.J., 2014. Ecadherin loss alters cytoskeletal organization and adhesion in non-malignant breast cells but is insufficient to induce an epithelial-mesenchymal transition. BMC Cancer 14, 552. Cojoc, M., Mabert, K., Muders, M.H., Dubrovska, A., 2015a. A role for cancer stem cells in therapy resistance: cellular and molecular mechanisms. Semin. Cancer Biol. 31, 16e27.

964

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

Cojoc, M., Peitzsch, C., Kurth, I., Trautmann, F., KunzSchughart, L.A., Telegeev, G.D., Stakhovsky, E.A., Walker, J.R., Simin, K., Lyle, S., Fuessel, S., Erdmann, K., Wirth, M.P., Krause, M., Baumann, M., Dubrovska, A., 2015b. Aldehyde dehydrogenase is regulated by beta-catenin/TCF and promotes radioresistance in prostate cancer progenitor cells. Cancer Res. 75, 1482e1494. Collins, A.T., Berry, P.A., Hyde, C., Stower, M.J., Maitland, N.J., 2005. Prospective identification of tumorigenic prostate cancer stem cells. Cancer Res. 65, 10946e10951. Connor, K.M., Subbaram, S., Regan, K.J., Nelson, K.K., Mazurkiewicz, J.E., Bartholomew, P.J., Aplin, A.E., Tai, Y.T., Aguirre-Ghiso, J., Flores, S.C., Melendez, J.A., 2005. Mitochondrial H2O2 regulates the angiogenic phenotype via PTEN oxidation. J. Biol. Chem. 280, 16916e16924. Czabotar, P.E., Lessene, G., Strasser, A., Adams, J.M., 2014. Control of apoptosis by the BCL-2 protein family: implications for physiology and therapy. Nat. Rev. Mol. Cell Biol. 15, 49e63. Das, G.C., Holiday, D., Gallardo, R., Haas, C., 2001. Taxol-induced cell cycle arrest and apoptosis: dose-response relationship in lung cancer cells of different wild-type p53 status and under isogenic condition. Cancer Lett. 165, 147e153. Das, V., Stepankova, J., Hajduch, M., Miller, J.H., 2015. Role of tumor hypoxia in acquisition of resistance to microtubulestabilizing drugs. Biochim. Biophys. Acta 1855, 172e182. Davenport, E.L., Moore, H.E., Dunlop, A.S., Sharp, S.Y., Workman, P., Morgan, G.J., Davies, F.E., 2007. Heat shock protein inhibition is associated with activation of the unfolded protein response pathway in myeloma plasma cells. Blood 110, 2641e2649. Di Fazio, P., Ocker, M., Montalbano, R., 2012. New drugs, old fashioned ways: ER stress induced cell death. Curr. Pharm. Biotechnol. 13, 2228e2234. Dubois, C., Vanden Abeele, F., Sehgal, P., Olesen, C., Junker, S., Christensen, S.B., Prevarskaya, N., Moller, J.V., 2013. Differential effects of thapsigargin analogues on apoptosis of prostate cancer cells: complex regulation by intracellular calcium. FEBS J. 280, 5430e5440. Eymin, B., Gazzeri, S., 2010. Role of cell cycle regulators in lung carcinogenesis. Cell Adh. Migr. 4, 114e123. Feigin, M.E., Muthuswamy, S.K., 2009. Polarity proteins regulate mammalian cell-cell junctions and cancer pathogenesis. Curr. Opin. Cell Biol. 21, 694e700. Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M., Parkin, D.M., Forman, D., Bray, F., 2015. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int. J. Cancer 136, E359eE386. Fried, J., Perez, A.G., Clarkson, B.D., 1976. Flow cytofluorometric analysis of cell cycle distributions using propidium iodide. Properties of the method and mathematical analysis of the data. J. Cell Biol. 71, 172e181. Fulda, S., 2013. Regulation of apoptosis pathways in cancer stem cells. Cancer Lett. 338, 168e173. Fulda, S., 2015. Targeting apoptosis for anticancer therapy. Semin. Cancer Biol. 31, 84e88. Gerlach, J.P., Emmink, B.L., Nojima, H., Kranenburg, O., Maurice, M.M., 2014. Wnt signalling induces accumulation of phosphorylated beta-catenin in two distinct cytosolic complexes. Open Biol. 4, 140120. Gogada, R., Prabhu, V., Amadori, M., Scott, R., Hashmi, S., Chandra, D., 2011. Resveratrol induces p53-independent, Xlinked inhibitor of apoptosis protein (XIAP)-mediated Bax protein oligomerization on mitochondria to initiate cytochrome c release and caspase activation. J. Biol. Chem. 286, 28749e28760. Gomes, L.R., Terra, L.F., Wailemann, R.A., Labriola, L., Sogayar, M.C., 2012. TGF-beta1 modulates the homeostasis

between MMPs and MMP inhibitors through p38 MAPK and ERK1/2 in highly invasive breast cancer cells. BMC Cancer 12, 26. Hannigan, G., Troussard, A.A., Dedhar, S., 2005. Integrin-linked kinase: a cancer therapeutic target unique among its ILK. Nat. Rev. Cancer 5, 51e63. Harrington, P.E., Biswas, K., Malwitz, D., Tasker, A.S., Mohr, C., Andrews, K.L., Dellamaggiore, K., Kendall, R., Beckmann, H., Jaeckel, P., Materna-Reichelt, S., Allen, J.R., Lipford, J.R., 2015. Unfolded protein response in cancer: IRE1alpha inhibition by selective kinase ligands does not impair tumor cell viability. ACS Med. Chem. Lett. 6, 68e72. Hart, G., Flaishon, L., Shachar, I., 2007. IL-12 and IL-18 downregulate B cell migration in an Ly49D-dependent manner. Eur. J. Immunol. 37, 1996e2007. Hata, A., Katakami, N., Tanaka, K., Takeshita, J., Matsumoto, T., Monden, K., Nagata, K., Masago, K., Kaji, R., Fujita, S., Tachikawa, R., Otsuka, K., Otsuka, K., Tomii, K., 2014. Bevacizumab plus weekly paclitaxel with or without carboplatin for previously-treated non-squamous non-small cell lung cancer. Anticancer Res. 34, 275e281. Horibe, T., Hoogenraad, N.J., 2007. The chop gene contains an element for the positive regulation of the mitochondrial unfolded protein response. PLoS One 2, e835. Hurt, E.M., Kawasaki, B.T., Klarmann, G.J., Thomas, S.B., Farrar, W.L., 2008. CD44þ CD24(-) prostate cells are early cancer progenitor/stem cells that provide a model for patients with poor prognosis. Br. J. Cancer 98, 756e765. Ichim, G., Lopez, J., Ahmed, S.U., Muthalagu, N., Giampazolias, E., Delgado, M.E., Haller, M., Riley, J.S., Mason, S.M., Athineos, D., Parsons, M.J., van de Kooij, B., Bouchier-Hayes, L., Chalmers, A.J., Rooswinkel, R.W., Oberst, A., Blyth, K., Rehm, M., Murphy, D.J., Tait, S.W., 2015. Limited mitochondrial permeabilization causes DNA damage and genomic instability in the absence of cell death. Mol. Cell 57, 860e872. Issa, A., Le, T.X., Shoushtari, A.N., Shields, J.D., Swartz, M.A., 2009. Vascular endothelial growth factor-C and C-C chemokine receptor 7 in tumor cell-lymphatic cross-talk promote invasive phenotype. Cancer Res. 69, 349e357. Jovaisaite, V., Mouchiroud, L., Auwerx, J., 2014. The mitochondrial unfolded protein response, a conserved stress response pathway with implications in health and disease. J. Exp. Biol. 217, 137e143. Kelly, W.K., Curley, T., Slovin, S., Heller, G., McCaffrey, J., Bajorin, D., Ciolino, A., Regan, K., Schwartz, M., Kantoff, P., George, D., Oh, W., Smith, M., Kaufman, D., Small, E.J., Schwartz, L., Larson, S., Tong, W., Scher, H., 2001. Paclitaxel, estramustine phosphate, and carboplatin in patients with advanced prostate cancer. J. Clin. Oncol. e Off. J. Am. Soc. Clin. Oncol. 19, 44e53. Kelly, W.K., Zhu, A.X., Scher, H., Curley, T., Fallon, M., Slovin, S., Schwartz, L., Larson, S., Tong, W., Hartley-Asp, B., Pellizzoni, C., Rocchetti, M., 2003. Dose escalation study of intravenous estramustine phosphate in combination with Paclitaxel and Carboplatin in patients with advanced prostate cancer. Clin. Cancer Res. e Off. J. Am. Assoc. Cancer Res. 9, 2098e2107. Koochekpour, S., Maresh, G.A., Katner, A., Parker-Johnson, K., Lee, T.J., Hebert, F.E., Kao, Y.S., Skinner, J., Rayford, W., 2004. Establishment and characterization of a primary androgenresponsive African-American prostate cancer cell line, E006AA. Prostate 60, 141e152. Koochekpour, S., Willard, S.S., Shourideh, M., Ali, S., Liu, C., Azabdaftari, G., Saleem, M., Attwood, K., 2014. Establishment and characterization of a highly tumorigenic African American prostate cancer cell line, E006AA-hT. Int. J. Biol. Sci. 10, 834e845.

M O L E C U L A R O N C O L O G Y 1 0 ( 2 0 1 6 ) 9 4 9 e9 6 5

Liang, Y., Eid, M.A., El Etreby, F., Lewis, R.W., Kumar, M.V., 2002. Mifepristone-induced secretion of transforming growth factor beta1-induced apoptosis in prostate cancer cells. Int. J. Oncol. 21, 1259e1267. Liu, C., Tang, D.G., 2011. MicroRNA regulation of cancer stem cells. Cancer Res. 71, 5950e5954. Lu, M., Marsters, S., Ye, X., Luis, E., Gonzalez, L., Ashkenazi, A., 2014. E-cadherin couples death receptors to the cytoskeleton to regulate apoptosis. Mol. Cell 54, 987e998. Maher, M.T., Mo, R., Flozak, A.S., Peled, O.N., Gottardi, C.J., 2010. Beta-catenin phosphorylated at serine 45 is spatially uncoupled from beta-catenin phosphorylated in the GSK3 domain: implications for signaling. PLoS One 5, e10184. Mao, J., Fan, S., Ma, W., Fan, P., Wang, B., Zhang, J., Wang, H., Tang, B., Zhang, Q., Yu, X., Wang, L., Song, B., Li, L., 2014. Roles of Wnt/beta-catenin signaling in the gastric cancer stem cells proliferation and salinomycin treatment. Cell Death Dis. 5, e1039. Marciniak, S.J., Yun, C.Y., Oyadomari, S., Novoa, I., Zhang, Y., Jungreis, R., Nagata, K., Harding, H.P., Ron, D., 2004. CHOP induces death by promoting protein synthesis and oxidation in the stressed endoplasmic reticulum. Genes Dev. 18, 3066e3077. Mitchison, T.J., 2012. The proliferation rate paradox in antimitotic chemotherapy. Mol. Biol. Cell 23, 1e6. Molife, L.R., Yan, L., Vitfell-Rasmussen, J., Zernhelt, A.M., Sullivan, D.M., Cassier, P.A., Chen, E., Biondo, A., Tetteh, E., Siu, L.L., Patnaik, A., Papadopoulos, K.P., de Bono, J.S., Tolcher, A.W., Minton, S., 2014. Phase 1 trial of the oral AKT inhibitor MK-2206 plus carboplatin/paclitaxel, docetaxel, or erlotinib in patients with advanced solid tumors. J. Hematol. Oncol. 7, 1. Murray, S., Briasoulis, E., Linardou, H., Bafaloukos, D., Papadimitriou, C., 2012. Taxane resistance in breast cancer: mechanisms, predictive biomarkers and circumvention strategies. Cancer Treat. Rev. 38, 890e903. Nawrocki-Raby, B., Gilles, C., Polette, M., Martinella-Catusse, C., Bonnet, N., Puchelle, E., Foidart, J.M., Van Roy, F., Birembaut, P., 2003. E-Cadherin mediates MMP downregulation in highly invasive bronchial tumor cells. Am. J. Pathol. 163, 653e661. Naydenov, N.G., Baranwal, S., Khan, S., Feygin, A., Gupta, P., Ivanov, A.I., 2013. Novel mechanism of cytokine-induced disruption of epithelial barriers: Janus kinase and protein kinase D-dependent downregulation of junction protein expression. Tissue Barriers 1, e25231. Neumann, C.A., Krause, D.S., Carman, C.V., Das, S., Dubey, D.P., Abraham, J.L., Bronson, R.T., Fujiwara, Y., Orkin, S.H., Van Etten, R.A., 2003. Essential role for the peroxiredoxin Prdx1 in erythrocyte antioxidant defence and tumour suppression. Nature 424, 561e565. O’Neill, A.J., Prencipe, M., Dowling, C., Fan, Y., Mulrane, L., Gallagher, W.M., O’Connor, D., O’Connor, R., Devery, A., Corcoran, C., Rani, S., O’Driscoll, L., Fitzpatrick, J.M., Watson, R.W., 2011. Characterisation and manipulation of docetaxel resistant prostate cancer cell lines. Mol. Cancer 10, 126. Okudela, K., Woo, T., Mitsui, H., Tajiri, M., Masuda, M., Ohashi, K., 2012. Expression of the potential cancer stem cell markers, CD133, CD44, ALDH1, and beta-catenin, in primary lung adenocarcinoma e their prognostic significance. Pathol. Int. 62, 792e801. Pelicano, H., Xu, R.H., Du, M., Feng, L., Sasaki, R., Carew, J.S., Hu, Y., Ramdas, L., Hu, L., Keating, M.J., Zhang, W., Plunkett, W., Huang, P., 2006. Mitochondrial respiration defects in cancer cells cause activation of Akt survival pathway through a redox-mediated mechanism. J. Cell Biol. 175, 913e923.

965

Rivkin, I., Cohen, K., Bod, T., Argov, M., Margalit, R., 2014. Cancer cell sensitization and improved treatment efficacy by combined sodium butyrate and paclitaxel formulations is cancer-type specific. Int. J. Pharm. 461, 437e447. Rycaj, K., Tang, D.G., 2014. Cancer stem cells and radioresistance. Int. J. Radiat. Biol. 90, 615e621. Ryoo, H.D., Bergmann, A., 2012. The role of apoptosis-induced proliferation for regeneration and cancer. Cold Spring Harbor Perspect. Biol. 4, a008797. Scheel, C., Weinberg, R.A., 2012. Cancer stem cells and epithelialmesenchymal transition: concepts and molecular links. Semin. Cancer Biol. 22, 396e403. Schieber, M., Chandel, N.S., 2014. ROS function in redox signaling and oxidative stress. Curr. Biol. e CB 24, R453eR462. Sharon-Friling, R., Goodhouse, J., Colberg-Poley, A.M., Shenk, T., 2006. Human cytomegalovirus pUL37x1 induces the release of endoplasmic reticulum calcium stores. Proc. Natl. Acad. Sci. U. S. A. 103, 19117e19122. Siegel, R., Naishadham, D., Jemal, A., 2013. Cancer statistics, 2013. CA Cancer J. Clin. 63, 11e30. Singh, R.K., Lokeshwar, B.L., 2009. Depletion of intrinsic expression of Interleukin-8 in prostate cancer cells causes cell cycle arrest, spontaneous apoptosis and increases the efficacy of chemotherapeutic drugs. Mol. Cancer 8, 57. Smith, D.C., Chay, C.H., Dunn, R.L., Fardig, J., Esper, P., Olson, K., Pienta, K.J., 2003. Phase II trial of paclitaxel, estramustine, etoposide, and carboplatin in the treatment of patients with hormone-refractory prostate carcinoma. Cancer 98, 269e276. Statkiewicz, M., Maryan, N., Lipiec, A., Grecka, E., Grygorowicz, M.A., Omiotek, M., Gorska, A., Mikula, M., Malecki, M., 2014. The role of the SHH gene in prostate cancer cell resistance to paclitaxel. Prostate 74, 1142e1152. Sullivan, L.B., Chandel, N.S., 2014. Mitochondrial reactive oxygen species and cancer. Cancer Metab. 2, 17. Tang, D.G., 2012. Understanding cancer stem cell heterogeneity and plasticity. Cell Res. 22, 457e472. Valeriote, F., van Putten, L., 1975. Proliferation-dependent cytotoxicity of anticancer agents: a review. Cancer Res. 35, 2619e2630. Watson, P.A., Arora, V.K., Sawyers, C.L., 2015. Emerging mechanisms of resistance to androgen receptor inhibitors in prostate cancer. Nat. Rev. Cancer 15, 701e711. Waypa, G.B., Marks, J.D., Guzy, R., Mungai, P.T., Schriewer, J., Dokic, D., Schumacker, P.T., 2010. Hypoxia triggers subcellular compartmental redox signaling in vascular smooth muscle cells. Circ. Res. 106, 526e535. Wilson, C., Maxwell, P.J., Longley, D.B., Wilson, R.H., Johnston, P.G., Waugh, D.J., 2012. Constitutive and treatmentinduced CXCL8-signalling selectively modulates the efficacy of anti-metabolite therapeutics in metastatic prostate cancer. PLoS One 7, e36545. Yadav, N., Pliss, A., Kuzmin, A., Rapali, P., Sun, L., Prasad, P., Chandra, D., 2014. Transformations of the macromolecular landscape at mitochondria during DNA-damage-induced apoptotic cell death. Cell Death Dis. 5, e1453. Yang, Y., Alcaraz, A.A., Snyder, J.P., 2009. The tubulin-bound conformation of paclitaxel: T-taxol vs “PTX-NY”. J. Nat. Prod. 72, 422e429. Yvon, A.M., Wadsworth, P., Jordan, M.A., 1999. Taxol suppresses dynamics of individual microtubules in living human tumor cells. Mol. Biol. Cell 10, 947e959. Zhang, B., Suer, S., Livak, F., Adediran, S., Vemula, A., Khan, M.A., Ning, Y., Hussain, A., 2012. Telomere and microtubule targeting in treatment-sensitive and treatment-resistant human prostate cancer cells. Mol. Pharmacol. 82, 310e321. Zhao, Q., Wang, J., Levichkin, I.V., Stasinopoulos, S., Ryan, M.T., Hoogenraad, N.J., 2002. A mitochondrial specific stress response in mammalian cells. EMBO J. 21, 4411e4419.