Available online at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/yexcr

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2

http://dx.doi.org/10.10014-4827/& 2014 El

nCorresponding auSt. Joseph's Hospital,

E-mail address: d

Research Article

Prostate cancer stem-like cells proliferate slowly andresist etoposide-induced cytotoxicity via enhancingDNA damage response

Judy Yana,b,c, Damu Tanga,b,c,n

aDivision of Nephrology, Department of Medicine, McMaster University, Juravinski Innovation Tower, Room T3310,St. Joseph's Hospital, 50 Charlton Ave East, Hamilton, Ontario, Canada L8S 4L8bFather Sean O'Sullivan Research Institute, Hamilton, Ontario, Canada L8N 4A6cThe Hamilton Centre for Kidney Research (HCKR), St. Joseph's Hamilton Healthcare, Hamilton, Ontario, Canada L8N 4A6

a r t i c l e i n f o r m a t i o n

Article Chronology:

Received 11 April 2014Received in revised form20 June 2014Accepted 11 August 2014Available online 20 August 2014

Keywords:

Prostate cancer stem cellCell proliferationCell signalingDNA damage responseTumorigenesis

016/j.yexcr.2014.08.016sevier Inc. All rights reserv

thor at: Division of Nephr50 Charlton Ave East, Hamamut@mcmaster.ca (D. Ta

a b s t r a c t

Despite the development of chemoresistance as a major concern in prostate cancer therapy, theunderlying mechanisms remain elusive. In this report, we demonstrate that DU145-derived prostatecancer stem cells (PCSCs) progress slowly with more cells accumulating in the G1 phase in comparisonto DU145 non-PCSCs. Consistent with the important role of the AKT pathway in promoting G1progression, DU145 PCSCs were less sensitive to growth factor-induced activation of AKT in comparisonto non-PCSCs. In response to etoposide (one of the most commonly used chemotherapeutic drugs),DU145 PCSCs survived significantly better than non-PCSCs. In addition to etoposide, PCSCs demon-strated increased resistance to docetaxel, a taxane drug that is commonly used to treat castration-resistant prostate cancer. Etoposide produced elevated levels of γH2AX and triggered a robust G2/Marrest along with a coordinated reduction of the G1 population in PCSCs compared to non-PCSCs,

suggesting that elevated γH2AX plays a role in the resistance of PCSCs to etoposide-inducedcytotoxicity. We have generated xenograft tumors from DU145 PCSCs and non-PCSCs. Consistent withthe knowledge that PCSCs produce xenograft tumors with more advanced features, we were able todemonstrate that PCSC-derived xenograft tumors displayed higher levels of γH2AX and p-CHK1compared to non-PCSC-produced xenograft tumors. Collectively, our research suggests that theelevation of DNA damage response contributes to PCSC-associated resistance to genotoxic reagents.

& 2014 Elsevier Inc. All rights reserved.

Introduction

Prostate cancer (PC) is the most frequently diagnosed male cancerin the developed world [1]. The disease advances from prostatic

ed.

ology, Department of Medilton, Ontario, Canada L8Nng).

intra-epithelial neoplasia (PIN), to locally invasive carcinoma, tometastatic cancer and finally to castration-resistant prostatecancer (CRPC), which accounts for the majority of PC fatalities[2,3]. The reactivation of the androgen receptor (AR) axis in the

icine, McMaster University, Juravinski Innovation Tower, Room T3310,4A6. Fax: þ1 905 521 6181.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2 133

absence of androgen and the dysregulation of growth factorpathways have been extensively studied [4–6]. While localizedPC is commonly treated through radical prostatectomy and radio-therapy, advanced PC is treated with androgen ablation therapy.Chemotherapy is often the main salvage choice for patients withrecurrent PC after prostatectomy [7]. However, chemotherapy isineffective as resistance commonly develops and PC progressioncontinues [8].

While the exact mechanisms responsible for the developmentof chemoresistance are unclear, evidence supports that prostatecancer stem cells (PCSCs) contribute to chemotherapy resistance.Cancer stem cells (CSCs) possess functional traits of stem cells,including their expression of stem cell markers, their indefiniteself-renewal potential, ability for tumour initiation, differentiationinto a hierarchically organized tumour mass and a quiescent state[9–13]. More importantly, CSCs also possess an increased resis-tance to cytotoxic treatments [14–17]. PCSCs produce recurrenthormone refractory PCs [18] and the signatures of PCSCs areassociated with PC bone metastasis and poor prognosis [19,20].CD133þ, a PCSC marker, is associated with resistance to che-motherapy [21,22]. Despite the existence of indirect evidence for acritical role of PCSCs in the appearance of chemoresistance, directevidence supporting this notion has yet to be reported.

Chemotherapy primarily results in DNA damage. DNA lesionselicit the cellular DNA damage response (DDR) via a complexnetwork, leading to cell cycle arrest, senescence and apoptosis[23–25]. To investigate the contributions of PCSCs to the resis-tance of PC to DNA damage-induced cytotoxicity, we examinedthe impact of etoposide, a commonly used chemotherapeuticdrug on DU145 cell-derived PCSCs and the non-PCSC counter-parts. We report here that PCSCs proliferate substantially slowerthan non-PCSCs, exhibit significant resistance to etoposide-induced cytotoxicity, and were robustly sensitive to etoposide-elicited G2/M arrest. These observations collectively suggest thatthe kinetics of slow proliferation together with the enhancedsensitivity to DNA damage contribute to PCSC's resistance to DNAdamage-caused cytotoxicity.

Materials and methods

Cell lines and materials

DU145 was purchased from American Type Culture Collection(ATCC), and cultured in MEM supplemented with 10% FBS (SigmaAldrich, Oakville, ON) and 1% Penicillin–Streptomycin (Life Tech-nologies, Carlsbad, CA). Etoposide and propidium iodide werepurchased from Sigma Aldrich.

Generation of DU145 spheres

DU145 prostate cancer stem-like cells (spheres) were isolated andpropagated as we have previously published [26]. Briefly, DU145monolayer cells were individualized with TrypLE Express solution(Life Technologies, Carlsbad, CA) and subsequently resuspended ata density of 5000 cells/mL in serum-free (SF) media (DMEM/F12at a 3:1 mixture) (Life Technologies, Carlsbad, CA) supplementedwith 0.4% bovine serum albumin (BSA) (Bioshop Canada Inc.,Burlington, ON) and 0.2�B27 minus Vitamin A (Life Technologies,Carlsbad, CA) in T75 flasks. Typical spheres were formed in 10–12

days. All experiments were conducted with DU145 spheres thatwere subsequently individualized and grown in MEM with 10%FBS for up to four passages.

Western blot analysis

Cell lysates were prepared in a buffer containing 20 mM Tris (pH7.4), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100,25 mM sodium pyrophosphate, 1 mM NaF, 1 mM β-glyceropho-sphate, 0.1 mM sodium orthovanadate, 1 mM PMSF, 2 μg/mlleupeptin and 10 μg/ml aprotinin (Sigma Aldrich, Oakville, ON).A total of 50 μg of cell lysates were separated on SDS-PAGE geland transferred onto Amersham hybond ECL nitrocellulose mem-branes (Amersham, Baie d'Urfe, QC). Membranes were blockedwith 5% skim milk and then incubated with the indicatedantibodies at 4 1C overnight. Appropriate HRP-conjugatedsecondary antibodies were incubated for one hour at room tempera-ture. Signals were detected using an ECL Western Blotting Kit(Amersham, Baie d'Urfe, QC). The primary and secondary antibodiesand the concentrations used were: anti-γH2AX (1:1000, Cell Signal-ing, Danvers, MA), anti-H2AX (1:1000, Millipore, Mississauga, ON),anti-CHK1 Ser345 phosphorylation (1:300, Cell Signaling, Danvers,MA), anti-CHK1 (1:1000, Cell Signaling, Danvers, MA), anti-CHK2Thr68 phosphorylation (1:1000, Cell Signaling, Danvers, MA), anti-CHK2 (1:1000, Cell Signaling, Danvers, MA), anti-AKT Ser473 phos-phorylation (1:1000, Cell Signaling, Danvers, MA), anti-AKT (1:1000,Santa Cruz Biotechnology, Santa Cruz, CA), anti-GAPDH (1:5000, CellSignaling, Danvers, MA), anti-actin (1:1000, Santa Cruz Biotechnology,Santa Cruz, CA), anti-goat (1:3000, Santa Cruz Biotechnology, SantaCruz, CA), anti-mouse (1:3000, GE Healthcare, Mississauga, ON) andanti-rabbit (1:3000, GE Healthcare, Mississauga, ON).

Immunohistochemistry (IHC)

Xenograft tumors were deparaffinized in xylene, cleared in anethanol series, and heat-treated for 20 min in sodium citratebuffer (pH¼6.0) in a food steamer. Primary antibodies specific foranti-γH2AX (1:50, Cell Signaling, Danvers, MA) and anti-CHK1Ser345 phosphorylation (1:50, Cell Signaling, Danvers, MA) wereincubated with the sections overnight at 4 1C. Slides stainedwithout any primary antibody and non-specific rabbit IgG wereused as negative controls. Biotinylated secondary IgG and VectorABC reagent (Vector Laboratories, Burlingam, CA) were subse-quently added according to the manufacturer's instructions.Washes were performed with PBS. Chromogen reaction wascarried out with diaminobenzidine (Vector Laboratories, Burlin-gam, CA), and counterstained with hematoxylin (Sigma Aldrich,Oakville, ON).

Immunofluorescence (IF)

Cells were treated with etoposide for 2 h prior to staining.Immunofluorescence staining was carried out by fixing cells with4% paraformaldehyde for 20 min and permeabilized with 0.3%Triton X-100 (Sigma Aldrich, Oakville, ON) for 15 min. Anti-γH2AX(1:50, Cell Signaling, Danvers, MA) was added to the slides at 4 1Covernight. After washing, secondary antibody FITC Donkey IgG(1:200, Jackson ImmunoResearch Lab, West Grove, PA), wasapplied for 1 h at room temperature. The slide was subsequently

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2134

covered with VECTASHIELD mounting medium with DAPI (VectorLaboratories, Burlingam, CA). Images were taken with a fluores-cence microscope (Carl Zeiss, Axiovert 200).

Cell proliferation assay

A total of 1000 cells of DU145 and DU145 stem-like cells wereseeded in MEM supplemented with 10% FBS into a 96 well plateand incubated at 37 1C for 5 days. Proliferation was measuredusing the WST-1 cell proliferation assay kit (Millipore, Missis-sauga, ON) according to the manufacturer's instructions. Absor-bance readings were measured with a plate reader at 420 nm.

Colony formation

Cells were seeded at the indicated number of cells per well intriplicates and allowed to grow for 1 week. Media was changedevery 2 days. Cells were stained with crystal violet (0.5%). Emptywells were stained as background. Staining was dissolved in 33%acetic acid and absorbance read by a spectrometer at 550 nm.Specific staining was derived by the subtraction of the back-ground staining.

Cell survival assay

Cells were seeded in 6 well plates for 24 h. When they reach aconfluency of 80–90%, the cells were treated with DMSO (mocktreatment) and ETOP at the designated doses for 8 h or withdocetaxel at the designated doses for 6 h followed by culturing inuntreated media for 1 week. Surviving cells were stained withcrystal violet (0.5%). Empty wells were stained as background.Staining was dissolved in 33% acetic acid and absorbance read bya spectrometer at 550 nm. Specific staining was derived by thesubtraction of the background staining.

Cell cycle distribution

Cell cycle analysis was carried out as previously described [27].Briefly, cells were individualized and washed 3 times with PBS.DNA content was stained using propidium iodide solution in thepresence of 100 mg/mL of RNase A (Sigma Aldrich, Oakville, ON)and 0.1% Triton X-100 for 1 h followed by detection of signalsusing a flow cytometer FC500 (Beckman Coulter, Mississauga,ON).

Formation of xenograft tumors

A total of 106 DU145 monolayer (non stem cells) and 104 DU145sphere (stem-like) cells were individualized and resuspended inMEM/Matrigel mixture (1:1 volume), followed by implantation of0.1 mL of this mixture subcutaneously (s.c.) into flanks of 8-week-old male NOD/SCID mice (The Jackson Laboratory, Bar Harbor,ME). Mice were inspected for tumour appearance, by observationand palpation, and tumour growth was measured weekly using acaliper. Tumour volumes were determined using the standardformula: L�W2�0.52, where L and W are the longest andshortest diameters, respectively. All animal work was carriedout according to experimental protocols approved by the McMas-ter University Animal Research Ethics Board.

Statistical analysis

Statistical analysis was performed using the student t-test andpo0.05 was considered statistically significant.

Results and Discussion

PCSCs proliferate at a significantly reduced rate comparedto non-PCSCs

It is a widely accepted concept that stem cells resist a variety oftoxicities largely owing to their non-proliferative status [11,28].However, whether quiescence remains an intrinsic feature forCSCs remains unclear. This is especially the case for PCSCs, as ithas been reported that DU145 PCSCs exhibited increased AKTactivity and proliferated more rapidly compared to DU145 non-PCSCs [29–31]. However, this can also be due to the unique mediarequirements of PCSCs for their isolation and maintenance. Whencultured under serum free medium, this condition supports PCSCbut not non-PCSC proliferation, indeed only DU145 PCSCs pro-liferate [26]. In our laboratory, we have recently isolated andcharacterized DU145 stem-like cells (for simplicity, we call itPCSCs) as suspension spheres with increased self-renewal andenhanced tumorigenicity. Approximately, 1.25% of DU145 cells areable to form spheres and 26% of the sphere cells are able togenerate secondary spheres [26], demonstrating an enrichment ofthe stem cell population in spheres. The PCSCs possess the abilityto regenerate the original heterogeneous population of cells andare highly tumorigenic than the non-PCSCs, with as little as 100cells capable of forming xenograft tumors. In addition, thesePCSCs express CD44, CD24, α2β1 integrin along with both basaland luminal markers [26]. It has been previously demonstratedthat DU145 stem-like cells proliferate as a monolayer in serum-containing mediumwith their ability to produce spheres in serumfree media (SFM) being fully maintained within the first 4passages [26]. However, the proliferation ability of PCSCs versusnon-PCSCs remains unknown under the same conditions whereboth are proliferating and their characteristics are also main-tained. By taking advantage of this knowledge, we examined theproliferation of non-PCSCs (DU145 monolayer cells) and that ofDU145 PCSCs side-by-side. In comparison to non-PCSC cells,DU145 PCSCs proliferated significantly slower (Fig. 1A). Whenseeded at comparable and low densities, DU145 PCSCs formeddramatically less colonies compared to non-PCSC DU145 cells(Fig. 1B). Furthermore, cell cycle analysis revealed more PCSCsaccumulated in the G1 phase with a concurrent reduction of Sphase cells in comparison to non-PCSCs (Fig. 1C). These observa-tions are consistent with the well demonstrated knowledge thatthe progression of G1 phase among cell division is regulated byextracellular signals [32–34]. Although we demonstrate a higherG1 phase in PCSCs compared to non-PCSCs, a number of proteinsinvolved in cell cycle remain unchanged (data not shown).As stem cells have a unique epigenetic pattern compared to theirrespective non stem cells [21,35,36], we speculate the possibilitythat multiple pathways are involved in ensuring a slow prolifera-tion rate of PCSCs. Conceptually, this property can be explored totarget PCSCs in future PC therapy. It is thus worthy of investigat-ing the underlying mechanisms that maintains PCSCs in aquiescent state.

Fig. 1 – DU145 PCSCs proliferate significantly slower than DU145 cells. (A) 1000 cells were seeded into 96 well plates. Cellproliferation was assayed daily for five days. Typical results from a single repeat are shown. (B) DU145 and DU145 PCSCs wereseeded at the indicated number of cells and cultured for one week. Cells were stained with crystal violet and quantified. Typicalimages after crystal violet staining are shown. (C) Cell cycle distribution was examined in DU145 and DU145 PCSCs. All experimentswere carried out three times. Means and standard deviations are presented. *: statistically significant (po0.05) in comparisonto DU145.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2 135

AKT is a key regulator protein that when activated by phos-phorylation is responsible for modifying a variety of eventsincluding cell growth, proliferation, apoptosis, motility, angiogen-esis and metastasis [37–40]. As G1 progression is regulated bygrowth factors via the AKT pathway [41–43], we examined theunderlying mechanism responsible for the slow G1 progressionobserved in PCSCs. We have examined the kinetics of growthfactor-induced activation of AKT. While serum dose-dependentlyinduces the activation of AKT in DU145 non-PCSCs, the magni-tudes of their activation were significantly attenuated in DU145PCSCs (Fig. 2). Taken together, as both types of cells are faced with

the presence of serum in situ these results demonstrate thatDU145 PCSCs are endowed with an intrinsic characteristic forslow proliferation.

PCSCs display a significant survival advantage comparedto non-PCSCs in response to etoposide treatment

The property of slow proliferation or relative quiescence of PCSCswould suggest that PCSCs are resistant to chemotherapy-inducedcytotoxicity. One of the classic chemotherapeutic reagents isetoposide (ETOP) and ETOP is used in prostate cancer therapy

Fig. 2 – Decreased sensitivity to growth factor-induced activation of AKT in DU145 PCSCs. To examine the kinetics of AKT activation,DU145 and DU145 stem-like cells were serum starved over night followed by the addition of media with 10% FBS for the indicatedtimepoints. (A) Protein expression was examined for p-AKT, AKT and GAPDH. (B) Quantification of p-AKT was determined as a foldchange of expression relative to when no serum was added to the cells (0 min timepoint) following the normalization of p-AKT toAKT and GAPDH. Quantifications are from three independent experiments. Means and standard deviations are presented. n:statistically significant (po0.05) in comparison to DU145.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2136

[24,44]. The compound induces double-stranded DNA breaks(DSBs) by “poisoning” DNA topoisomerase II [24]. To examinethe cytotoxic effects of etoposide on PCSCs versus non-PCSCs, wetook advantage that DU145-derived PCSCs can be maintained asspheres in serum free medium (SFM) and that their stemnessmeasured by the cell's ability to produce subsequent spheres inSFM was not diminished even when cultured for more than 4passages as a monolayer in serum (10%) containing medium(SCM) [26]. Under identical culture conditions in SCM, DU145monolayer cells (non-PCSCs) did not survive 10 mM of etoposide(Fig. 3A and B), an average of 75% of the first passage of DU145sphere cell-derived re-monolayer cells, or DU145 PCSCs survived10 mM of etoposide treatment (Fig. 3A and B); interestingly,DU145 monolayer cells did not even survive as low as 3 mM and1 mM of etoposide (Supp Fig. 1). Although the percentage ofsurviving PCSCs declined following increasing doses of etoposide,PCSCs are substantially more resistant to etoposide in comparisonto DU145 non-PCSCs (Fig. 3A and B). Similar results were obtainedfor PCSCs that were maintained in serum containing mediumfrom less than 4 passages (data not shown), an observation that isconsistent with the knowledge that these cells are still PCSCs [26].To determine whether the surviving PCSCs are associated with

major defects that may prevent their proliferation and eventualsphere forming capabilities, we re-seeded surviving cells at lowdensities and these cells formed colonies in a density-dependentmanner (Fig. 3C). In addition, the surviving PCSCs were capable ofgenerating spheres and were able to be maintained demonstrat-ing their continued self-renewal after etoposide treatment(Fig. 3D). Therefore, at least for PCSCs that survived 5 mM of ETOPthese cells were fully recovered in terms of maintaining theirsphere forming and proliferation capacity. Docetaxel is a wellknown chemotherapeutic drug for a number of solid tumorsincluding prostate cancer. It is a semisynthetic taxane drug thatinhibits mitosis by binding and stabilizing microtubules to pre-vent disassembly [45]. In addition to etoposide, DU145 PCSCswere also more resistant to docetaxel induced cytotoxicitythen DU145 non-PCSCs (Supp Fig. 2). Taken together, theabove observations demonstrate that PCSCs significantly resistetoposide-induced cytotoxicity, and in part docetaxel-inducedcytotoxicity in comparison to their non-PCSC counterparts. As

slow proliferation has been shown to promote chemoresistance ina number of cancer cells [43], the observed decrease in prolifera-tion (Fig. 1) is a possible contributing factor to PCSCs' resistance togenotoxic agents-induced toxicity.

DU145 PCSCs display enhanced sensitivityto etoposide-induced DDR

After demonstrating that PCSCs proliferate slowly and are moreresistant to ETOP-induced cytotoxicity, we proceeded to examinethe underlying mechanisms responsible for PCSCs ability tosurvive etoposide treatment. Etoposide induces DSBs in cells,which elicit the DNA damage response (DDR). One aspect of DDRis to activate checkpoints to prevent cell proliferation by cell cyclearrest and apoptosis induction [46]. We thus reasoned that DU145PCSCs may undergo DDR differently compared to non-PCSCs inresponse to etoposide treatment. To determine this possibility, wetreated both DU145 PCSCs and non-PCSCs with different dosesfollowed by the examination of cell cycle distribution. In order todetermine the percentage of cells arresting at the G2/M phase, wetreated cells for 24 h to allow for cells to go through a round ofcell division. In order to detect the kinetics of G2/M arrest in bothDU145 PCSCs and non-PCSCs, a lower dosage of etoposide wasused. Surprisingly, at 0.4 mM of etoposide, a dose that increasedthe G2/M phase of cells from 6.7% to 19% for DU145 non-PCSCs,57% of PCSCs was arrested in the G2/M phase (Fig. 4, Table 1).While etoposide levels at 0.6 and 0.8 mM induced a 34% and 51%of non-PCSCs to arrest at the G2/M phase, respectively, the samedoses induced more than 70% of PCSCs to arrest in the G2/Mphase (Fig. 4, Table 1). Collectively, the above observationsdemonstrate a robustly enhanced sensitivity of DDR checkpointactivation for PCSCs in response to etoposide treatment.

To further examine PCSC-associated activation of DDR check-points, we have determined the kinetics of CHK1 and CHK2activation, two essential kinases involved in DDR checkpointactivation [46,47] in etoposide-treated PCSCs and non-PCSCs.Activation of CHK1 and CHK2 are commonly measured byphosphorylation of CHK1 at serine 345 (p-CHK1) and phosphor-ylation of CHK2 at threonine 68 (p-CHK2) [27,48]. Both non-PCSCs

Fig. 3 – DU145 stem-like cells exhibit an increase in resistance to etoposide treatment. (A) DU145 and DU145 stem-like cells weretreated with DMSO or etoposide at the indicated dosages for 8 h. Cells were then cultured in untreated media for one week. Typicalimages after crystal violet staining are shown. (B) Surviving cells were quantified and expressed as a percentage of the mock-treated cells. Means and standard deviations are presented. *: statistically significant (po0.05) in comparison to DU145 monolayercells. (C) DU145 stem-like cells were treated with 5 lM of etoposide for 8 h. After one week of culturing in untreated media, cellswere seeded at the indicated number of cells per well and allowed to grow for another one week. Cells were stained with crystalviolet and quantified. (D) Immediately after an 8-h etoposide treatment and one week after culturing in untreated media, DU145stem-like cells were examined for sphere forming capacity.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2 137

and PCSCs displayed a rapid kinetics of CHK2 activation and aslower CHK1 activation (Fig. 5A), observations that are consistentwith our understanding that DSBs primarily elicit CHK2 activation[46]. The kinetics and the levels of CHK2 activation remainscomparable in ETOP-treated non-PCSCs and PCSCs (Fig. 5A). Whilethe magnitudes of CHK1 activation appeared similar in both linesupon ETOP treatment, CHK1 activation reached the maximal levelmore rapidly in ETOP-treated PCSCs compared to non-PCSCs(Fig. 5B). Collectively, these observations support the elevatedsensitivity of checkpoint activation in PCSCs in response to ETOP-induced DDR.

PCSCs possess significantly high levels of γH2AX

Although CHK1 may be activated with a faster kinetic in PCSCscompared to non-PCSCs in response to etoposide treatment(Fig. 5B), the difference does not seem to match the dynamicsdisplayed in PCSCs' G2/M arrest induced by ETOP (Fig. 4, Table 1).Etoposide is known to cause toxicity to cells by producing double-stranded breaks, and thereby leading to the production and therecruitment of γH2AX to the damaged DNA site; this recruitmentinduces the assembly of repair complexes to repair DSBs[46,49,50]. Considering the elevated G2/M arrest (Fig. 4, Table 1)

Fig. 4 – DU145 stem-like cells exhibit an increased sensitivity to etoposide-induced G2/M arrest. DU145 and DU145 stem-like cellswere treated with DMSO or etoposide at the indicated doses for 24 h. Cells were collected and examined for DNA content.Representative cell cycle images are shown.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2138

and the substantiated resistance to etoposide, we reasoned thatthe G2/M arrest may coordinate with DSB repair to allow PCSCs tosurvive etoposide treatment. To test this possibility, we examinedγH2AX in both DU145 PCSCs and DU145 non-PCSCs with andwithout etoposide treatment. While etoposide induced the pro-duction of γH2AX and the formation of γH2AX foci in DU145 non-PCSCs (Fig. 6A and B) indicative of the production of DSBs [51],etoposide elicited significantly higher levels of γH2AX (Fig. 6A andB) in PCSCs. This suggests that DU145 PCSCs undergo moredynamic DSB repair compared to the non-PCSC counterparts.We have previously shown that as little as 100 PCSCs can

efficiently generate tumors in NOD/SCID mice [26]. In addition,tumors generated from PCSCs displayed more characteristics thatare associated with advanced prostate cancer compared to non-PCSCs, as PCSC derived xenografts contained more CD44þ cells[26]. This is consistent with the report that CD44þ cells are moretumorgenic and metastatic than CD44� cells [52]. To takeadvantage of this knowledge and further investigate PCSCs-associated elevation of γH2AX, we produced xenograft tumorsusing non-PCSCs and PCSCs and determined the levels of γH2AX,a typical marker of DDR in xenograft tumors. When injectedsubcutaneously, PCSCs formed tumors and grew at the same rateas non-PCSCs despite a 100 fold decrease in the number of cellsinjected (Fig. 7A). The difference in the starting cell numbers is toensure the tumors grew at a relatively same rate and tumorvolumes are comparable during the time of sacrifice. However,this further demonstrates the advanced nature of PCSCs com-pared to non-PCSCs. In comparison to DU145 non-PCSC-producedxenograft tumors, those generated by DU145 PCSCs containedincreased levels of γH2AX (Fig. 7B and C), observations thatsupport the apparent high basal level of γH2AX in DU145 PCSCs

compared to the non-PCSC counterparts (Fig. 6). Consistent withγH2AX being indicative of double stranded breaks, we detected asubstantially higher level of CHK1 phosphorylation (p-CHK1),indicative of CHK1 activation in xenografts derived from PCSCs(Fig. 7B). As chemoresistant tumors are in a more advanced stagecompared to the tumors that are sensitive to chemotherapytreatments [3,53,54], along with the knowledge that DU145 PCSCsproduced tumors with characteristics that are associated withadvanced prostate cancer [26], it is thus tempting to speculatethat high levels of γH2AX and p-CHK1 could potentially play arole in chemoresistance in more advanced tumors. The highlysensitive nature of PCSCs to etoposide-induced G2/M arrest seemsinconsistent with their resistance to etoposide. While DDR iscertainly utilized for genotoxic reagents to inhibit tumour growthor cause tumour regression, DDR is also a protective mechanismto prevent catastrophes caused by the inability for cells to arrestto allow for the repair of DNA lesions [55–57]. Based on thisknowledge, it is conceivable that being unable to effectively arrestcell cycle progression in the presence of etoposide-induced DSBsis a major cause for sensitivity of DU145 non-PCSCs to etoposide-induced cytotoxicity and that the robust cell cycle arrest on theother hand protected PCSCs from etoposide-elicited lethal effects.This interpretation is consistent with the observed elevation ofγH2AX in PCSCs in comparison to the non-PCSCs, indicative ofPCSC-associated increases in DSB repair. This indication is con-sistent with the demonstrated enhancement of DNA damagerepair in CSCs [58]. Taken together, we demonstrate here thatDU145 PCSCs possess an increased survival advantage comparedto DU145 non-PCSCs to etoposide-induced cytotoxicity, likely dueto a coordinated elevation of checkpoint activation and DSBrepair.

Table 1 – Cell cycle on DU145 and DU145 PCSCs treated with ETOP.

ETOP (lM)

DMSO 0.4 0.6 0.8

G1 phaseDU145 45.076.0 45.875.7 32.678.0 21.879.3DU145 stem-like 57.872.8n 14.470.8n 4.870.6n 1.170.66n

S phaseDU145 48.275.3 34.771.0 33.171.8 27.176.0DU145 stem-like 36.771.7n 28.178.3 24.575.6 17.374.8

G2/M phaseDU145 6.771.8 19.474.8 34.179.9 50.9715.1DU145 stem-like 5.371.2 57.479.1n 70.575.8n 81.475.0n

Cell cycle distribution is presented as mean7SD.n po0.05 compared to DU145.

Fig. 5 – Faster kinetics for CHK1 activation but not CHK2 in PCSCs compared to non-PCSCs. DU145 and DU145 stem-like cells weretreated with DMSO or etoposide at the designated doses for 2 h. (A) Protein expression was examined for p-CHK1, CHK1, p-CHK2,CHK2 and GAPDH. (B) Quantification of p-CHK1 was based on the percentage of activation relative to 100 lM followingnormalization with CHK1 and GAPDH. Quantifications are from three independent experiments. Means and standard deviationsare presented. *: statistically significant (po0.05) in comparison to DU145.

Fig. 6 – PCSCs exhibit higher γH2AX levels compared to non-PCSCs. DU145 and DU145 stem-like cells were treated with DMSO oretoposide at the designated doses for 2 h. (A) Protein expression was examined and quantified for γH2AX and H2AX.Quantifications are from three independent experiments. γH2AX expression was normalized to H2AX expression. Means andstandard deviations are presented. *: statistically significant (po0.05) in comparison to DU145. (B) Immunofluorescence of γH2AXfoci in DU145 and DU145 stem-like cells. Nuclei were counterstained with DAPI. Scale bar represents 20 lm.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2 139

Fig. 7 – Increased γH2AX levels in PCSCs derived xenograft tumors. (A) DU145 (106 cells) and DU145 stem-like cells (104 cells) wereinjected subcutaneously into the flanks of male NOD/SCID mice and tumors were measured each week for 8 weeks. (B) Xenografttumors derived from DU145 (n¼3) and DU145 stem-like cells (n¼3) were stained for γH2AX and CHK1 activation (p-CHK1). Typicalimages with magnified regions (inset) are shown. (C) The number of cells staining positive for γH2AX, along with the total numberof cells were counted from five random fields for each xenograft tumors and expressed as a percentage of cells positive for γH2AX.Means and standard deviations are presented. *: statistically significant (po0.05) in comparison to DU145.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2140

Conclusion

It is becoming increasingly clear that CSCs are critical to cancerprogression and are responsible for failure to a variety totherapies. Our research demonstrated that PCSCs contribute tothe resistance to DNA damage-induced cytotoxicity. Potentiallyattributed to their slow proliferation, DU145 PCSCs survivedsignificantly better than non-PCSCs and triggered a robust G2/Marrest with elevated levels of γH2AX. As PCSCs generate tumorsmore aggressively than non-PCSCs, indicative of an advancedfeature of PC, higher levels of γH2AX and p-CHK1 were alsodetected in PCSC-derived xenografts. Our research thus providesthe first evidence that PCSCs and most likely other types of CSCsenhance their survival of DNA lesions in a coordinated effortthrough sensitization of checkpoint activation and through eleva-tion of lesion repairs.

Conflict interest

The authors declare no conflict of interest.

Acknowledgments

This work was supported by a CIHR Grant (MOP - 84381) toD. Tang.

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.yexcr.2014.08.016.

r e f e r e n c e s

[1] A. Jemal, F. Bray, M.M. Center, J. Ferlay, E. Ward, D. Forman,Global cancer statistics, CA: Cancer J. Clin. 61 (2011) 69–90.

[2] C. Moon, J.C. Park, Y.K. Chae, J.H. Yun, S. Kim, Current status ofexperimental therapeutics for prostate cancer, Cancer Lett. 266(2008) 116–134.

[3] F.G. el-Kamar, M.L. Grossbard, P.S. Kozuch, Metastatic pancreaticcancer: emerging strategies in chemotherapy and palliative care,Oncologist 8 (2003) 18–34.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2 141

[4] Y. Chen, C.L. Sawyers, H.I. Scher, Targeting the androgen receptorpathway in prostate cancer, Curr. Opin. Pharmacol. 8 (2008)440–448.

[5] D. Djakiew, Dysregulated expression of growth factors and theirreceptors in the development of prostate cancer, Prostate 42(2000) 150–160.

[6] C. Fiorentini, S. Bodei, F. Bedussi, M. Fragni, S.A. Bonini,C. Simeone, D. Zani, A. Berruti, C. Missale, M. Memo, P. Spano,S. Sigala, GPNMB/OA protein increases the invasiveness of humanmetastatic prostate cancer cell lines DU145 and PC3 throughMMP-2 and MMP-9 activity, Exp. Cell Res. 323 (2014) 100–111.

[7] T.E. Bjerklund Johansen, U. Witzsch, D. Greene, Salvage treatmentin prostate cancer: a clinical approach, Expert Rev. AnticancerTher. 13 (2013) 613–623.

[8] J.E. Damber, G. Aus, Prostate cancer, Lancet 371 (2008)1710–1721.

[9] I. Baccelli, A. Trumpp, The evolving concept of cancer andmetastasis stem cells, J. Cell Biol. 198 (2012) 281–293.

[10] A. Chauchereau, N. Al Nakouzi, C. Gaudin, S. Le Moulec,D. Compagno, N. Auger, J. Benard, P. Opolon, F. Rozet, P. Validire,G. Fromont, K. Fizazi, Stemness markers characterize IGR-CaP1,a new cell line derived from primary epithelial prostate cancer,Exp. Cell Res. 317 (2011) 262–275.

[11] M. Dean, T. Fojo, S. Bates, Tumour stem cells and drug resistance,Nat. Rev. Cancer 5 (2005) 275–284.

[12] R. Hellsten, M. Johansson, A. Dahlman, O. Sterner, A. Bjartell,Galiellalactone inhibits stem cell-like ALDH-positive prostatecancer cells, PLoS One 6 (2011) e22118.

[13] J.E. Visvader, G.J. Lindeman, Cancer stem cells in solid tumours:accumulating evidence and unresolved questions, Nat. Rev.Cancer 8 (2008) 755–768.

[14] S. Bao, Q. Wu, R.E. McLendon, Y. Hao, Q. Shi, A.B. Hjelmeland,M.W. Dewhirst, D.D. Bigner, J.N. Rich, Glioma stem cells promoteradioresistance by preferential activation of the DNA damageresponse, Nature 444 (2006) 756–760.

[15] P.C. Hermann, S.L. Huber, T. Herrler, A. Aicher, J.W. Ellwart, M.Guba, C.J. Bruns, C. Heeschen, Distinct populations of cancer stemcells determine tumor growth and metastatic activity in humanpancreatic cancer, Cell Stem Cell 1 (2007) 313–323.

[16] M.F. Shi, J. Jiao, W.G. Lu, F. Ye, D. Ma, Q.G. Dong, X. Xie,Identification of cancer stem cell-like cells from human epithelialovarian carcinoma cell line, Cell. Mol. Life Sci. 67 (2010)3915–3925.

[17] M.H. Wright, A.M. Calcagno, C.D. Salcido, M.D. Carlson,S.V. Ambudkar, L. Varticovski, Brca1 breast tumors containdistinct CD44þ/CD24� and CD133þ cells with cancer stem cellcharacteristics, Breast Cancer Res. 10 (2008) R10.

[18] J. Qin, X. Liu, B. Laffin, X. Chen, G. Choy, C.R. Jeter, T. Calhoun-Davis, H. Li, G.S. Palapattu, S. Pang, K. Lin, J. Huang, I. Ivanov,W. Li, M.V. Suraneni, D.G. Tang, The PSA(-/lo) prostate cancer cellpopulation harbors self-renewing long-term tumor-propagatingcells that resist castration, Cell Stem Cell 10 (2012) 556–569.

[19] M. Colombel, C.L. Eaton, F. Hamdy, E. Ricci, G. van der Pluijm,M. Cecchini, F. Mege-Lechevallier, P. Clezardin, G. Thalmann,Increased expression of putative cancer stem cell markers inprimary prostate cancer is associated with progression of bonemetastases, Prostate 72 (2012) 713–720.

[20] Z.A. Wang, A. Mitrofanova, S.K. Bergren, C. Abate-Shen,R.D. Cardiff, A. Califano, M.M. Shen, Lineage analysis of basalepithelial cells reveals their unexpected plasticity and supports acell-of-origin model for prostate cancer heterogeneity, Nat.Cell Biol. 15 (2013) 274–283.

[21] T. Baba, P.A. Convery, N. Matsumura, R.S. Whitaker, E. Kondoh,T. Perry, Z. Huang, R.C. Bentley, S. Mori, S. Fujii, J.R. Marks,A. Berchuck, S.K. Murphy, Epigenetic regulation of CD133 andtumorigenicity of CD133þ ovarian cancer cells, Oncogene 28(2009) 209–218.

[22] X. Yu, Y. Lin, X. Yan, Q. Tian, L. Li, E.H. Lin, CD133, stem cells, andcancer stem cells: myth or reality? Curr. Colorectal Cancer Rep. 7(2011) 253–259.

[23] S. Bhattacharya, R.M. Ray, L.R. Johnson, Role of polyamines inp53-dependent apoptosis of intestinal epithelial cells, Cell Signal.21 (2009) 509–522.

[24] C.J. Lord, A. Ashworth, The DNA damage response and cancertherapy, Nature 481 (2012) 287–294.

[25] N. Pabla, Z. Ma, M.A. McIlhatton, R. Fishel, Z. Dong, hMSH2recruits ATR to DNA damage sites for activation during DNAdamage-induced apoptosis, J. Biol. Chem. 286 (2011) 10411–10418.

[26] A.P. Rybak, L. He, A. Kapoor, J.C. Cutz, D. Tang, Characterization ofsphere-propagating cells with stem-like properties from DU145prostate cancer cells, Biochim. Biophys. Acta 1813 (2011)683–694.

[27] F. Wei, Y. Xie, L. Tao, D. Tang, Both ERK1 and ERK2 kinasespromote G2/M arrest in etoposide-treated MCF7 cells by facil-itating ATM activation, Cell Signal. 22 (2010) 1783–1789.

[28] P. Dalerba, R.W. Cho, M.F. Clarke, Cancer stem cells: models andconcepts, Annu. Rev. Med. 58 (2007) 267–284.

[29] A. Dubrovska, S. Kim, R.J. Salamone, J.R. Walker, S.M. Maira,C. Garcia-Echeverria, P.G. Schultz, V.A. Reddy, The role of PTEN/Akt/PI3K signaling in the maintenance and viability of prostatecancer stem-like cell populations, Proc. Natl. Acad. Sci. USA 106(2009) 268–273.

[30] D. Ponti, A. Costa, N. Zaffaroni, G. Pratesi, G. Petrangolini,D. Coradini, S. Pilotti, M.A. Pierotti, M.G. Daidone, Isolation andin vitro propagation of tumorigenic breast cancer cells withstem/progenitor cell properties, Cancer Res. 65 (2005)5506–5511.

[31] S.K. Singh, I.D. Clarke, M. Terasaki, V.E. Bonn, C. Hawkins,J. Squire, P.B. Dirks, Identification of a cancer stem cell in humanbrain tumors, Cancer Res. 63 (2003) 5821–5828.

[32] J. Boonstra, Progression through the G1-phase of the on-goingcell cycle, J. Cell. Biochem. 90 (2003) 244–252.

[33] E. Hulleman, J. Boonstra, Regulation of G1 phase progression bygrowth factors and the extracellular matrix, Cell. Mol. Life Sci. 58(2001) 80–93.

[34] S.M. Jones, A. Kazlauskas, Growth factor-dependent signaling andcell cycle progression, FEBS Lett. 490 (2001) 110–116.

[35] T.H. Cheung, T.A. Rando, Molecular regulation of stem cellquiescence, Nat. Rev. Mol. Cell Biol. 14 (2013) 329–340.

[36] J. Lee, M.J. Son, K. Woolard, N.M. Donin, A. Li, C.H. Cheng,S. Kotliarova, Y. Kotliarov, J. Walling, S. Ahn, M. Kim, M. Totonchy,T. Cusack, C. Ene, H. Ma, Q. Su, J.C. Zenklusen, W. Zhang, D. Maric,H.A. Fine, Epigenetic-mediated dysfunction of the bone mor-phogenetic protein pathway inhibits differentiation ofglioblastoma-initiating cells, Cancer Cell 13 (2008) 69–80.

[37] S. Brader, S.A. Eccles, Phosphoinositide 3-kinase signallingpathways in tumor progression, invasion and angiogenesis,Tumori 90 (2004) 2–8.

[38] S.J. Grille, A. Bellacosa, J. Upson, A.J. Klein-Szanto, F. van Roy,W. Lee-Kwon, M. Donowitz, P.N. Tsichlis, L. Larue, The proteinkinase Akt induces epithelial mesenchymal transition andpromotes enhanced motility and invasiveness of squamous cellcarcinoma lines, Cancer Res. 63 (2003) 2172–2178.

[39] A.K. Nalla, N. Estes, J. Patel, J.S. Rao, N-cadherin mediatesangiogenesis by regulating monocyte chemoattractant protein-1expression via PI3K/Akt signaling in prostate cancer cells, Exp.Cell Res. 317 (2011) 2512–2521.

[40] A. Toker, M. Yoeli-Lerner, Akt signaling and cancer: surviving butnot moving on, Cancer Res. 66 (2006) 3963–3966.

[41] N. Gao, Z. Zhang, B.H. Jiang, X. Shi, Role of PI3K/AKT/mTORsignaling in the cell cycle progression of human prostate cancer,Biochem. Biophys. Res. Commun. 310 (2003) 1124–1132.

[42] J. Liang, J.M. Slingerland, Multiple roles of the PI3K/PKB (Akt)pathway in cell cycle progression, Cell Cycle 2 (2003) 339–345.

E X P E R I M E N T A L C E L L R E S E A R C H 3 2 8 ( 2 0 1 4 ) 1 3 2 – 1 4 2142

[43] G.T. McDonald, R. Sullivan, G.C. Pare, C.H. Graham, Inhibition ofphosphatidylinositol 3-kinase promotes tumor cell resistance tochemotherapeutic agents via a mechanism involving delay in cellcycle progression, Exp. Cell Res. 316 (2010) 3197–3206.

[44] J.M. Kamradt, K.J. Pienta, Etoposide in prostate cancer, ExpertOpin. Pharmacother. 1 (2000) 271–275.

[45] B.A. Hadaschik, M.E. Gleave, Therapeutic options for hormone-refractory prostate cancer in 2007, Urol. Oncol. 25 (2007)413–419.

[46] B.B. Zhou, S.J. Elledge, The DNA damage response:putting checkpoints in perspective, Nature 408 (2000)433–439.

[47] R.T. Abraham, Cell cycle checkpoint signaling through the ATMand ATR kinases, Genes Dev. 15 (2001) 2177–2196.

[48] F. Wei, Y. Xie, L. He, L. Tao, D. Tang, ERK1 and ERK2 kinasesactivate hydroxyurea-induced S-phase checkpoint in MCF7 cellsby mediating ATR activation, Cell Signal. 23 (2011) 259–268.

[49] M.T. Mok, B.R. Henderson, A comparison of BRCA1 nuclearlocalization with 14 DNA damage response proteins anddomains: identification of specific differences between BRCA1and 53BP1 at DNA damage-induced foci, Cell Signal. 22 (2009)47–56.

[50] G.S. Stewart, B. Wang, C.R. Bignell, A.M. Taylor, S.J. Elledge, MDC1is a mediator of the mammalian DNA damage checkpoint, Nature421 (2003) 961–966.

[51] C. Redon, D. Pilch, E. Rogakou, O. Sedelnikova, K. Newrock,W. Bonner, Histone H2A variants H2AX and H2AZ, Curr. Opin.Genet. Dev. 12 (2002) 162–169.

[52] L. Patrawala, T. Calhoun, R. Schneider-Broussard, H. Li, B. Bhatia,S. Tang, J.G. Reilly, D. Chandra, J. Zhou, K. Claypool, L. Coghlan,D.G. Tang, Highly purified CD44þ prostate cancer cells fromxenograft human tumors are enriched in tumorigenic andmetastatic progenitor cells, Oncogene 25 (2006) 1696–1708.

[53] J. Crown, J. O’Shaughnessy, G. Gullo, Emerging targeted therapiesin triple-negative breast cancer, Ann. Oncol. 23 (Suppl 6) (2012)Svi56–Svi65.

[54] J.H. Schiller, D. Harrington, C.P. Belani, C. Langer, A. Sandler,J. Krook, J. Zhu, D.H. Johnson, Comparison of four chemotherapyregimens for advanced non-small-cell lung cancer, N. Engl. J.Med. 346 (2002) 92–98.

[55] S. Gasser, D. Raulet, The DNA damage response, immunity andcancer, Semin. Cancer Biol. 16 (2006) 344–347.

[56] J.W. Harper, S.J. Elledge, The DNA damage response: ten yearsafter, Mol. Cell 28 (2007) 739–745.

[57] A. Sancar, L.A. Lindsey-Boltz, K. Unsal-Kacmaz, S. Linn, Molecularmechanisms of mammalian DNA repair and the DNA damagecheckpoints, Annu. Rev. Biochem. 73 (2004) 39–85.

[58] M. Maugeri-Sacca, M. Bartucci, R. De Maria, DNA damage repairpathways in cancer stem cells, Mol. Cancer Ther. 11 (2012)1627–1636.