Chemotherapy-Induced CXC-Chemokine/CXC-ChemokineReceptor Signaling in Metastatic Prostate Cancer CellsConfers Resistance to Oxaliplatin through Potentiationof Nuclear Factor-�B Transcription and Evasion of Apoptosis

Catherine Wilson, Colin Purcell, Angela Seaton, Olabode Oladipo, Pamela J. Maxwell,Joe M. O’Sullivan, Richard H. Wilson, Patrick G. Johnston, and David J. J. WaughCentre for Cancer Research and Cell Biology, Queen’s University Belfast, Belfast, United Kingdom

Received July 22, 2008; accepted September 8, 2008

ABSTRACTConstitutive activation of nuclear factor (NF)-�B is linked withthe intrinsic resistance of androgen-independent prostate can-cer (AIPC) to cytotoxic chemotherapy. Interleukin-8 (CXCL8) isa transcriptional target of NF-�B whose expression is elevatedin AIPC. This study sought to determine the significance ofCXCL8 signaling in regulating the response of AIPC cells tooxaliplatin, a drug whose activity is reportedly sensitive to NF-�Bactivity. Administration of oxaliplatin to PC3 and DU145 cellsincreased NF-�B activity, promoting antiapoptotic gene tran-scription. In addition, oxaliplatin increased the transcription andsecretion of CXCL8 and the related CXC-chemokine CXCL1and increased the transcription and expression of CXC-chemo-kine receptors, especially CXC-chemokine receptor (CXCR) 2,which transduces the biological effects of CXCL8 and CXCL1.

Stimulation of AIPC cells with CXCL8 potentiated NF-�B acti-vation in AIPC cells, increasing the transcription and expressionof NF-�B-regulated antiapoptotic genes of the Bcl-2 and IAPfamilies. Coadministration of a CXCR2-selective antagonist,AZ10397767 (Bioorg Med Chem Lett 18:798–803, 2008), at-tenuated oxaliplatin-induced NF-�B activation, increased oxali-platin cytotoxicity, and potentiated oxaliplatin-induced apopto-sis in AIPC cells. Pharmacological inhibition of NF-�B or RNAinterference-mediated suppression of Bcl-2 and survivin wasalso shown to sensitize AIPC cells to oxaliplatin. Our resultsfurther support NF-�B activity as an important determinant ofcancer cell sensitivity to oxaliplatin and identify the induction ofautocrine CXCR2 signaling as a novel mode of resistance tothis drug.

Interleukin-8 (CXCL8) is a proinflammatory CXC-che-mokine whose expression is primarily regulated by theactivator protein-1 and NF-�B transcription factors (Bratet al., 2005). Overexpression of this chemokine has beendetected in the serum of patients with metastatic prostatecancer (CaP) (Veltri et al., 1999; McCarron et al., 2002),whereas colorimetric in situ hybridization has reported

elevated expression of CXCL8 in the tumor cells of andro-gen-independent CaP (AIPC) tissue (Uehara et al., 2005).Recently, we and others have demonstrated elevatedCXCL8 expression and CXCL8 receptor expression in can-cer cells of prostate biopsy tissue (Huang et al., 2005a;Murphy et al., 2005). Using immunohistochemistry, wedetermined that the intensity of CXCL8, CXCR1, andCXCR2 staining increased with stage of disease, reachinga maximal level in AIPC. The concurrent expression of theligand and its receptors suggests that CaP cells are subjectto a continuous autocrine CXCL8 signaling stimulus insitu. Significantly, increased expression of CXCL8 hasbeen correlated with the angiogenesis, tumorigenicity, andincidence of lymph node metastasis arising from orthotopicor xenograft implantation of human AIPC cells in athymicnude mice (Inoue et al., 2000; Kim et al., 2001), suggesting

This study was supported by the Ulster Cancer Foundation (to P.G.J. andD.J.J.W.), by the Prostate Research Campaign UK (to D.J.J.W.), by the North-ern Ireland Health, Social Services and Public Safety Research and Develop-ment Office (Grants RRG9.14 to D.J.J.W. and EAT/2546/03 to P.S., R.H.W.,and D.J.J.W.), by Cancer Research UK (Grant C212/A5720 to P.G.J., R.H.W.,and D.J.J.W.), and by the Association for International Cancer Research(Grant AICR-04-516).

Article, publication date, and citation information can be found athttp://jpet.aspetjournals.org.

doi:10.1124/jpet.108.143826.

ABBREVIATIONS: NF, nuclear factor; CaP, prostate cancer; AIPC, androgen-independent prostate cancer; CXCR, CXC-chemokine receptor;BAY 11-7082, 3-(4-methylphenylsulfonyl)-2-propenenitrile; SC-514, 5-(thien-3-yl)-3-aminothiophene-2-carboxamide; L-OHP, oxaliplatin; AZ767,AZ10397767; PARP, poly(ADP-ribose) polymerase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ELISA, enzyme-linked immunosorbentassay; PCR, polymerase chain reaction; qPCR, quantitative real-time PCR; siRNA, small interfering RNA; FACS, fluorescence-activated cellsorting; RI, R index; EMSA, electromobility shift assay; rCXCL, recombinant human-CXCL.

0022-3565/08/3273-746–759$20.00THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 327, No. 3Copyright © 2008 by The American Society for Pharmacology and Experimental Therapeutics 143826/3406733JPET 327:746–759, 2008 Printed in U.S.A.

746

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

that deregulation of this chemokine in patients may havefunctional significance with regard to disease progression.

Increasingly, monitoring CXCL8 expression in cancer pa-tients has been used as a prognostic marker in assessingpatient response to chemotherapy. In ovarian cancer, ele-vated serum CXCL8 levels identify those patients with a highresidual tumor burden after paclitaxel therapy (Uslu et al.,2005), whereas high levels of CXCL8 expression in the peri-toneal fluid and serum correlate with a poor initial responseto paclitaxel chemotherapy (Penson et al., 2000). Likewise,decreased CXCL8 serum levels have been described as anindicator of response to chemotherapy in stage IV melanoma(Brennecke et al., 2005) and non-small cell lung cancer pa-tients (Orditura et al., 2002), whereas reductions in intratu-moral CXCL8 expression has been reported in esophagealadenocarcinoma patients exhibiting a complete pathologicalresponse to chemotherapy (Abdel-Latif et al., 2005). Chemo-therapy agents have been shown to directly regulate CXCL8transcription in cancer cells. Paclitaxel increases CXCL8transcription and secretion in ovarian, breast, and lung can-cer cell lines (Collins et al., 2000; Uslu et al., 2005). Likewise,administration of doxorubicin (Adriamycin) and 5-fluoro-2�-deoxyuridine to breast cancer cells (De Larco et al., 2001), theaddition of fluorouracil to oral cancer cells (Tamatani et al.,2004), doxorubicin addition to small cell lung cancer cells(Shibakura et al., 2003), and dacarbazine administration tomelanoma cells (Lev et al., 2003) all result in increasedCXCL8 expression. However, the significance of this chemo-kine in modulating the response of cancer cells to chemother-apy is less well understood. Tumor necrosis factor-relatedapoptosis-inducing ligand-mediated increases in CXCL8 ex-pression attenuate cell death as a consequence of decreasedDR4 expression and reduced caspase-8 activation in ovariancarcinoma cells (Abdollahi et al., 2003). We have also dem-onstrated that CXCL8 signaling regulates the transcriptionof the native caspase-8-inhibitory protein, c-FLIP, to attenu-ate tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis in prostate cancer cells (Wilson et al.,2008). Conversely, the increased expression of CXCL8 wasnot shown to affect the sensitivity of osteosarcoma cells topaclitaxel (Duan et al., 2002).

The constitutive activity of NF-�B detected in AIPC celllines and tissue is proposed to underpin the poor response ofthis disease to chemotherapy (Fradet et al., 2004; Sweeneyet al., 2004). Because CXCL8 transcription is regulated byNF-�B (Brat et al., 2005), our study aimed to determinewhether this CXC-chemokine may contribute to the resis-tance of AIPC cells to cytotoxic chemotherapeutic agents. Wereport our findings that the expression of CXCL8, a relatedCXC-chemokine CXCL1, and their signaling competent re-ceptors CXCR1 and CXCR2 are each increased in AIPC cellsin response to administration of the platinum agent, oxali-platin, a chemotherapy agent whose activity has been shownto be modulated by NF-�B activity (Rakitina et al., 2003). Inaddition, we demonstrate that oxaliplatin-induced CXCR2signaling potentiates NF-�B activation and potentiates anti-apoptotic gene expression. Consistent with a role for CXCL8signaling in promoting cellular resistance to a chemicalstress, we further show that inhibition of CXCR2 signalingand its downstream effectors, NF-�B, Bcl-2, and survivin,increases the sensitivity of AIPC cells to undergo oxaliplatin-induced apoptosis. These findings support our recent charac-

terization of CXCL8 signaling in mediating the attenuatedresponse and reduced sensitivity of hypoxic AIPC cells toetoposide (Maxwell et al., 2007).

Materials and MethodsCell Culture. PC3 and DU145 cells were sourced and cultured as

described previously (MacManus et al., 2007; Maxwell et al., 2007).Chemicals. Chemicals were sourced from Sigma-Aldrich

(St. Louis, MO) unless otherwise stated. BAY 11-7082 and SC-514were purchased from Calbiochem (La Jolla, CA). Oxaliplatin (L-OHP)(Woynarowski et al., 2000) was obtained from the BridgewaterChemotherapy Suite, Belfast City Hospital. AZ10397767 (AZ767)was kindly provided by Dr. Simon T. Barry and Dr. David Blakey(AstraZeneca, Alderley Park, Cheshire, UK) (Walters et al., 2008).

Immunoblotting. Protein was prepared, resolved, and blotted asdescribed previously (Murphy et al., 2005; MacManus et al., 2007).Membranes were probed with monoclonal antibodies to anti-CXCL8(1:150 dilution) (Abcam plc, Cambridge, UK), anti-CXCR1 (1:500dilution) or anti-CXCR2 (1:250 dilution) (BioSource International,Camarillo, CA), anti-BCl2 (1:500 dilution) or anti-survivin (1:1000dilution) (all from Cell Signaling Technology Inc., Danvers, MA), andpoly(ADP-ribose) polymerase (PARP; 1:500 dilution) (eBioscience,San Diego, CA) overnight at 4°C. After washing in Tris-bufferedsaline/0.1% Tween 20, membranes were incubated with horseradishperoxidase-conjugated secondary antibodies (GE Healthcare, Chal-font St. Giles, UK). Specific staining was detected using chemilumi-nescence (SuperSignal, Pierce Chemical, Rockford, IL; or ECL Plus;GE Healthcare). Equal loading was assessed using a GAPDH mousemonoclonal primary antibody.

ELISA. Cells (1 � 105 cells/well) were incubated overnight at 37°Cin a humidified 5% CO2 atmosphere and replenished in serum-freeRPMI 1640 medium before treatment with oxaliplatin. Cell mediawere collected at indicated times, processed, and subjected to specificELISA. CXCL8 levels were measured using the Pelikine CompactCXCL8Elisa Kit (Sanquin Reagents, Amsterdam, The Netherlands),whereas CXCL1 levels were determined using the Quantikine kit(R&D Systems Europe Ltd, Abingdon, Oxfordshire, UK). The man-ufacturer’s instructions were employed in the application of eachELISA kit.

Electromobility Shift Assays. Nuclear extracts (8 �g of protein)were incubated with 35,000 cpm of a 22-bp oligonucleotide contain-ing the NF-�B consensus sequence, which had been previously end-labeled with [�-32P] ATP (10 mCi/mmol) by T4 polynucleotide kinase.Incubations were performed for 30 min at room temperature in thepresence of 2 �l of poly(dI�dC) and 100 mM Tris-HCl, pH 7.5, con-taining 10 mM EDTA, 50 mM dithiothreitol, and 40% (v/v) glycerol.The NF-�B complexes were resolved on 5% acrylamide gels, and themigration of NF-�B complexes was determined by detection of radio-activity on the gel by autoradiography.

Luciferase Reporter Assays. Cells (1 � 105 cells/well) wereseeded in six-well plates and incubated in RPMI 1640 medium at37°C for 48 h. Cells were then transfected using GeneJuice (Merck,Darmstadt, Germany) according to the manufacturer’s protocol in-corporating 2 �g of pGL3 basic vector (Promega, Madison, WI) or 2�g of an NF-�B-LUC-pGL3 plasmid (kindly provided by Dr. JamesPurcell, Queens University Belfast). Cells were also cotransfectedwith 0.2 �g of a Renilla luciferase plasmid as a transfection controlfor pGL3 and NF-�B. Transfected cells were incubated for a further24 h before drug addition. Either oxaliplatin or the CXC-chemokinewas added for the desired time, and the samples were analyzed byluciferase assay using the Promega Dual Luciferase assay kit (Pro-mega) according to the manufacturer’s protocol.

Quantitative Real-Time PCR Analysis. RNA was harvestedfrom cultured cells using RNAStat60 (Biogenesis, Poole, Dorset, UK)according to the manufacturer’s instructions, and cDNA was synthe-sized from 2 �g of total RNA by priming with random hexamers

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 747

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

(Invitrogen, Carlsbad, CA) and reverse transcribing with Moloneymurine leukemia virus reverse transcriptase (Invitrogen), as per themanufacturer’s protocol. Quantitative real-time PCR (qPCR) analy-sis was performed on the DNA Engine Opticon2 Continuous Fluo-rescence Detector (MJ Research, Watertown, MA) with product am-plification determined by SYBR Green 1 fluorescence detection(Finnzymes Diagnostics, Espoo, Finland). cDNA was synthesized asdescribed previously (Maxwell et al., 2007). The forward and reverseprimers used were: CXCL8 forward, ATg ACT TCC AAg CTg gCCgTg g; CXCL8 reverse, CAT AAT TTC TgT gTT ggC gCA gTg Tgg(Invitrogen); BCl2 forward, AAA ggA CCT gAT CAT Tgg gg; BCl2reverse, CAA CTC TTT TCC TCC CAC CA; survivin forward, TCTgCT TCA Agg AgC Tgg AA; survivin reverse, CgC ACT TTC TTCgCA gTT TC; CXCR1 forward, TgC ATC AgT gTg gAC CgT TA;CXCR1 reverse, TgT CAT TTC CCA ggA CCT CA; CXCR2 forward,TgC ATC AgT gTg gAC CgT TA; CXCR2 reverse, CCg CCA gTT TgCTgT ATT g; CXCL1 forward, CCC AAg AAC ATC CAA AgT gTC A;CXCL1 reverse, gTg gCT ATg ACT TCg gTT Tg; 18S forward, CATTCg TAT TgC gCC gCT A; and18S reverse, CgA Cgg TAT CTg ATCgTC (MWG Biotech, High Point, NC). The qPCR reaction consisted of1 �l of undiluted cDNA and 0.225 �M final concentration of forwardand reverse primers and 2� SYBR Green 1 master-mix (FinnzymesDiagnostics). Standard cycling procedures were employed with anannealing temperature of 55°C for all primer pairs tested. Specificamplicon formation with each primer pair was confirmed by meltcurve analysis. Gene expression was quantified relative to an 18Shousekeeping gene.

Cell Count Analysis. Cells (1 � 105 cells/well) were allowed toadhere overnight. Media were replaced with fresh RPMI 1640 me-dium and treated with a range of concentrations of oxaliplatin in theabsence or presence of AZ10397767 (20 nM) or BAY 11-7082 (1 �M).Cells were incubated at 37°C in a humidified 5% CO2 atmosphere for72 h, then trypsinized and counted in triplicate using a Coulter ZSeries particle count and size analyzer (Beckman Coulter, Fullerton,CA) at threshold values of TUPPER 21 and TLOWER 5. Cell numberswere normalized to control values, and statistical analyses of thedata were performed using GraphPad Prism 3.0 software (GraphPadSoftware Inc., San Diego, CA).

siRNA Strategy. Cells were seeded at 5 � 105 per P90 plate inOpti-MEM I (Invitrogen)-10% (v/v) fetal calf serum medium andallowed to grow to 50% confluence. siRNA oligonucleotide sequenceswere designed as follows: Bcl2 targeted, CAg CUg CAC CUg ACgCCC UUC; Bcl2 scrambled, CCU UCA gCC gCA UgC UgC CCA;survivin targeted, AAC gAg CCA gAC UUg gCC CA; and survivinscrambled, ACU ggC AAU CCA gCA gCC. siRNA transfections wereconducted by incubating cells in unsupplemented Opti-MEM I me-dium using Oligofectamine reagent (Invitrogen) for 4 h at 37°C, afterwhich cells were replenished in 20% (v/v) fetal calf serum-enrichedmedium. The concentration of oligonucleotides employed and theperiod of incubation for each experiment are described in the rele-vant textual description of results and the corresponding figurelegend.

FACS Analysis. Cells (5 � 104 cells) were seeded in a six-welltissue culture plate, allowed to adhere overnight, and then replen-ished in fresh RPMI 1640 medium. Cells were treated with increas-ing concentrations of oxaliplatin and incubated for a further 72 hbefore fixation in 100% ethanol. Cellular DNA profile was evaluatedby propidium iodide staining of cells using an EPICS XL Flow Cy-tometer (Beckman Coulter). In some experiments, oxaliplatin wasadministered in the presence of 20 nM AZ10397767 or 1 �M BAY11-7082.

Statistical Analysis. The nature of the interaction between ox-aliplatin and AZ10397767 was determined by calculating the R index(RI), using the methods recently described previously (Longley et al.,2004). Differences observed in comparing apoptosis levels observedin control and drug-treated groups were analyzed for statisticalsignificance using a two-tailed Student’s t test comparison (Prismversion 3.0 software).

ResultsOxaliplatin Promotes NF-�B Activation in AIPC Cells.

Oxaliplatin induces cytotoxicity through formation of adductswith cellular DNA (Woynarowski et al., 2000). Previously pub-lished studies suggest that the sensitivity of cancer cell lines tooxaliplatin is modulated by the level of constitutive NF-�Bactivity (Rakitina et al., 2003). An electromobility shift assay(EMSA) and luciferase reporter assays were each conductedto determine whether oxaliplatin promoted NF-�B activationin AIPC cells. Initial studies conducted in the PC3 cell lineconfirmed that administration of 1 �M oxaliplatin resulted ina rapid increase in the DNA binding activity of NF-�B (Fig.1a). Subsequently, luciferase reporter assays confirmed anincreased level of NF-�B transcriptional activity in two AIPCcell lines, PC3 and DU145 cells (Fig. 1b). Furthermore, usingreal-time PCR analysis, we observed increased mRNA tran-script levels for two representative NF-�B target genes, Bcl-2and survivin, after exposure of PC3 and DU145 cells to 1 �Moxaliplatin (Fig. 1c, left and right, respectively). In addition,oxaliplatin was also shown to increase Bcl-2 and survivinprotein expression in PC3 (Fig. 1c, left) and DU145 (Fig. 1c,right) cells.

Oxaliplatin Potentiates CXCL8 and CXC-ChemokineReceptor Expression in PC3 Cells. Transcription of theCXCL8 gene is regulated by NF-�B. Using a qPCR protocol,oxaliplatin was shown to induce a rapid and sustained in-crease in the CXCL8 mRNA transcript level in PC3 andDU145 cells (Fig. 2a, left). Furthermore, oxaliplatin treat-ment induced the expression of the related CXC-chemokine,CXCL1, in each of these CaP cell lines (Fig. 2a, right). Thisinduction of CXCL8 and CXCL1 by oxaliplatin was attenu-ated by addition of the NF-�B inhibitor, BAY 11-7082 (Fig.2b), suggesting that NF-�B activation is the predominantfactor underpinning oxaliplatin-induced transcriptional reg-ulation of these CXC-chemokines. Immunoblotting also con-firmed a time-dependent increase in the intracellular CXCL8protein level in response to 1 �M oxaliplatin (Fig. 2c),whereas ELISA-based analysis of PC3 or DU145 cell culturemedium revealed an increased level of CXCL8 and CXCL1secretion from each of these CaP cell lines after exposure to1 �M oxaliplatin for 3 h (Fig. 2d).

It is significant that the administration of oxaliplatin alsoincreased the expression of CXCR1 and CXCR2, the recep-tors that mediate the biological activity of both CXCL8 andCXCL1. The increase in the mRNA transcript level detectedfor CXCR2 was more sustained and greater in magnitudethan that detected for CXCR1 in both the PC3 and DU145cells in response to 1 �M oxaliplatin, respectively (Fig. 2e).Immunoblotting also confirmed increased CXCR1 and CXCR2receptor expression in the PC3 and DU145 cells after additionof oxaliplatin, especially that of the CXCR2 receptor (Fig. 2f). Asa consequence, the potentiation of CXCL8 and CXCL1 secretionand the increase in CXCR2 expression suggest that exposure tooxaliplatin increases the magnitude of the autocrine CXC-che-mokine signaling stimulus that AIPC cells are subject to and,furthermore, promotes a selective increase in CXCR2-mediatedsignaling.

CXC-Chemokine Signaling Potentiates NF-�B Tran-scription in AIPC Cells. Further experiments were con-ducted to characterize the effect of CXCL8 signaling uponestablished cell survival pathways, focusing on whether

748 Wilson et al.

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

NF-�B was a downstream effector of CXCL8 signaling inthese AIPC cells. PC3 cells were stimulated with 3 nM re-combinant human-CXCL8 (rCXCL8), and an EMSA wasused to determine whether this stimulus increased the DNAbinding activity of NF-�B. Increased complexing of NF-�Bwith DNA was detected within 20 min of adding rCXCL8,reaching a peak level of interaction at 40 min (Fig. 3a, left).Analysis over a longer time course also confirmed that NF-�Bcomplexing with DNA was detected within 1 h of stimulationwith rCXCL8 and identified a secondary increase in the DNAbinding of this transcription factor 4 h after addition ofrCXCL8 (Fig. 3a, right). The secondary increase in NF-�Bcomplexing was lower in intensity than the initial responseto exogenous rCXCL8. We also examined the receptor selec-tivity underpinning the CXCL8-induced activation of NF-�Band the CXCL8-induced potentiation of Bcl-2 expression.Blockade of CXCR2 receptor signaling was effected using asmall-molecule CXCR2 receptor antagonist, AZ10397767,which when administered at a concentration (20 nM) selec-tively antagonizes CXCR2 activation but has no activity onthe CXCR1 receptor in vitro (Walters et al., 2008). In EMSAanalysis, CXCL8-promoted DNA binding of NF-�B was at-tenuated in the presence of AZ10397767 (Fig. 3b). Luciferasereporter assays were also conducted to demonstrate that CXC-

chemokine signaling not only potentiated nuclear translocationand DNA binding of NF-�B but also induced the transcriptionalactivity of this factor in AIPC cells. Stimulation with either 3nM CXCL8 (Fig. 3c, left) or 3 nM CXCL1 (Fig. 3c, right) wasobserved to increase NF-�B-promoted luciferase activity in bothPC3 and DU145 cells. Consistent with the EMSA analysis, theCXC-chemokine-promoted NF-�B transcriptional activationwas attenuated by coadministration of the CXCR2 antagonist,AZ10397767.

CXCL8 Signaling Potentiates Antiapoptotic GeneTranscription in AIPC Cells. qPCR analysis was con-ducted to determine whether CXCL8 signaling increased thetranscriptional of established, NF-�B-regulated antiapop-totic genes. Bcl-2 mRNA transcript levels were shown toincrease by a factor in excess of 3- and 2.5-fold within 1 h ofadding 3 nM rCXCL8 to PC3 and DU145 cells, respectively(Fig. 4a, left). A sustained increase in the transcription of theBcl-2 gene was evident in both AIPC cells at further timepoints out to 12 h after stimulation. CXCL8 signaling alsoinduced statistically significant increases in survivin mRNAtranscript levels in PC3 and DU145 cells (Fig. 4a, right).Immunoblotting further confirmed that CXCL8 signalingincreased the expression of these representative NF-�B-regulated, antiapoptotic genes in both PC3 (Fig. 4b, left) and

Fig. 1. Oxaliplatin potentiates NF-�B-pro-moted transcription of antiapoptotic genesin AIPC cells. a, EMSA blot demonstratingthe promotion of NF-�B binding to its con-sensus binding site in PC3 cells in responseto the addition of 1 �M L-OHP. b, bar graphillustrating the magnitude and time-depen-dent induction of NF-�B-induced luciferaseactivity in PC3 and DU145 cells after stim-ulation with 1 �M L-OHP. Data shown arethe mean � S.E.M. values, calculated fromfour independent experiments. c, bar graphsillustrating the time-dependent increase inthe mRNA transcript levels detected forBcl-2 (left) and survivin (right) in PC3 cellsin response to addition of 1 �M L-OHP. Allvalues were normalized to the levels de-tected in unstimulated cells. Data shownare the mean � S.E.M. values, calculatedfrom four and three independent experi-ments, respectively. d, immunoblots demon-strating an increase in Bcl-2 and survivinexpression in PC3 (left) and DU145 (right)cells in response to the addition of 1 �ML-OHP. Equal protein loading was confirmedby reprobing the membrane for GAPDHexpression.

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 749

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

DU145 (Fig. 4b, right) cells. Expression of Bcl-2 was observedto increase in both cell lines at time points out to 4 h afterstimulation, with a secondary increase in expression detected10 h after stimulation. This may reflect the secondary in-creases in NF-�B/DNA binding and the secondary rise intranscript levels detected for this gene in each of the AIPC

cells after stimulation with rCXCL8. In contrast, CXCL8induced a sustained increase in the expression of survivin outto 6 h after stimulation in both AIPC cells (Fig. 4b). Similarto the effect of CXCL8, CXCL1 also induced increases in theexpression of Bcl-2 and survivin in these AIPC cells, al-though the induction of antiapoptotic protein expression in

Fig. 2. Oxaliplatin potentiates CXC-chemokine expression in AIPC cells. a, bar graphs illustrating the time-dependent increase in mRNA transcriptlevels detected for CXCL8 (left) and CXCL1 (right) in PC3 and DU145 cells over a 12-h time course, in response to stimulation with 1 �M L-OHP. Datashown are the mean � S.E.M. values, calculated from five independent experiments. b, bar graph illustrating the effect of the NF-�B inhibitor, BAY11-7082, on L-OHP-induced CXCL8 mRNA levels in PC3 and DU145 cells, measured 24 h after treatment with 1 �M L-OHP. Data shown are themean � S.E.M. values, calculated from four independent experiments. c, immunoblot illustrating the time-dependent increase in intracellular CXCL8expression after the addition of 1 �M L-OHP to PC3 cells. d, bar graph illustrating the time-dependent increase in the secretion of CXCL8 (left) andCXCL1 (right) from PC3 and DU145 cells after stimulation with 1 �M L-OHP. Data shown are the mean � S.E.M. values, calculated from threeindependent experiments. e, bar graph illustrating the time-dependent increase in mRNA transcript levels detected for CXCR1 (left) and CXCR2(right) in PC3 and DU145 cells in response to stimulation with 1 �M L-OHP. Data shown are the mean � S.E.M. values, calculated from fourindependent experiments. f, immunoblots illustrating the effect of L-OHP stimulation upon the expression of CXCR1 and CXCR2 receptors in PC3 (left)and DU145 (right) cells. All immunoblots shown are representative of three independent experiments, and equal protein loading was confirmed byreprobing the membranes for GAPDH expression. All values shown for qPCR analysis were normalized to the respective levels detected inunstimulated cells. Statistically significant differences were determined using Student’s t test: �, p � 0.05; ��, p � 0.01; and ���, p � 0.001.

750 Wilson et al.

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

response to this chemokine was more transient in the DU145cells than PC3 cells (Fig. 4c).

To confirm the mechanism regulating CXC-chemokine-induced antiapoptotic gene expression, PC3 and DU145cells were stimulated with 3 nM rCXCL8 in the absenceand presence of the CXCR2 antagonist, AZ10397767, or theNF-�B inhibitor, BAY 11-7082. AZ10397767 was shown toreverse the CXCL8-promoted increase in Bcl-2 and survivinmRNA transcript levels to basal or sub-basal levels in boththe PC3 and DU145 cells (Fig. 4d). Likewise, administrationof BAY 11-7082 attenuated the CXCL8-promoted transcrip-tion of the Bcl-2 and survivin genes in each of these AIPCcell lines (Fig. 4d). Reduced expression of Bcl-2 and survivinprotein was also observed in immunoblotting experimentsafter inhibition of CXCL8-signaling using AZ10397767 orBAY 11-7082 (data not shown). In accordance, our data in-dicate that the CXCR2 receptor primarily couples CXCL8to NF-�B activation and that CXCR2/NF-�B signaling un-derpins CXC-chemokine-promoted regulation of antiapop-totic protein expression in these cells.

Inhibition of CXCR2 Signaling Attenuates Oxaliplatin-Induced Activation of NF-�B. Because CXCR2-mediatedsignaling promotes NF-�B activation in PC3 cells, we deter-mined whether oxaliplatin-induced CXC-chemokine signalingmay reinforce and prolong oxaliplatin-induced NF-�B acti-vation. Oxaliplatin was added to PC3 cells in the absence orpresence of the CXCR2 antagonist, AZ10397767. EMSA analy-sis confirmed that coadministration of AZ10397767 attenuatedoxaliplatin-induced complexing of NF-�B with the radiola-beled oligonucleotide, with marked decreases observed 1 and2 h after addition of the two drugs (Fig. 5a). Coadministra-tion of AZ10397767 was also shown to inhibit oxaliplatin-induced NF-�B transcriptional activity, using a further lucif-erase reporter assay (Fig. 5b). Furthermore, qPCR analysisconfirmed that the presence of AZ10397767 attenuated theoxaliplatin-induced increases in mRNA transcript levelsfor each of the CXC-chemokines (CXCL8 and CXCL1) and

antiapoptotic genes (Bcl-2 and survivin) in the PC3 (Fig. 5c,left) and DU145 (Fig. 5c, right) cells. Coadministration ofAZ10397767 also attenuated oxaliplatin-induced increases inBcl-2 and survivin protein expression in AIPC cells (data notshown).

Inhibition of CXCR2 Signaling Potentiates the Cyto-toxicity and Induction of Apoptosis by Oxaliplatin inAIPC Cells. Cell count analysis was used to determine thesensitivity of AIPC cells to a 72-h continuous exposure tooxaliplatin. PC3 and DU145 cells were largely insensitive tooxaliplatin at concentrations up to 1 �M. However, reducedcell viability was detected above this concentration. Interest-ingly, the coadministration of AZ10397767 rendered bothPC3 and DU145 cells more sensitive to oxaliplatin-inducedcytotoxicity, with marked decreases in cell number detectedin cell populations treated with lower concentrations of thisplatinum drug (Fig. 6a). Analysis of cell count data conductedon either PC3 or DU145 cells determined that a two-site curvefit was statistically superior for oxaliplatin-induced cytotoxicityin the presence of AZ10397767 as compared with the one-sitemodel, which was preferential for modeling the effects ofoxaliplatin alone. The potentiating effect of AZ10397767 onoxaliplatin-induced cytotoxicity in PC3 cells was evident atconcentrations in the 10 to 50 nM range but not at higherconcentrations of oxaliplatin (i.e., greater than 10 �M). Theinteraction between oxaliplatin and AZ10397767 was syn-ergistic in the PC3 and DU145 cell lines, respectively, atoxaliplatin concentrations ranging from 0.1 to 1 �M (RI �1.5 and RI � 1.3, respectively). In addition, coadministra-tion of AZ10397767 corresponded to calculated increasesof 10- and 19-fold in the potency of oxaliplatin in PC3 cells(oxaliplatin and AZ10397767, IC30 0.2 � 0.005 �M ver-sus oxaliplatin alone, IC30 1.9 � 0.01 �M; n 6, p �0.01) and DU145 cells (IC25 0.021 � 0.02 versus 0.4 �0.09 �M; n 3), respectively.

Addition of AZ10397767 by itself failed to induce apoptosisin either PC3 or DU145 cells. However, coadministration of

Fig. 3. Characterization of CXC-chemokine-induced NF-�B activa-tion in AIPC cells. a, EMSA blotsdemonstrating the promotion ofNF-�B binding to its consensusbinding site in PC3 cells in re-sponse to short-term (left) or long-term (right) stimulation with 3 nMrCXCL8. b, EMSA blot illustratingthe effect of coadministration of 20nM AZ767, a CXCR2 receptor an-tagonist, upon the time-dependentDNA binding of NF-�B induced bystimulation of PC3 cells with 3 nMCXCL8. c, bar graph illustratingthe magnitude of NF-�B-inducedluciferase activity in PC3 andDU145 cells, 24 h after stimula-tion with 3 nM CXCL8 (left) or 3nM CXCL1 (right), in the absenceand presence of the CXCR2 antag-onist AZ767 (20 nM). Data shownare the mean � S.E.M. values, cal-culated from five and three inde-pendent experiments, respectively.Statistically significant differenceswere determined using Student’s ttest: �, p � 0.05; ��, p � 0.01.

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 751

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

Fig. 4. Characterization of CXC-chemokine-induced antiapoptotic protein expression in AIPC cells. a, bar graphs illustrating the effect of 3 nMrCXCL8-induced signaling upon the mRNA transcript levels detected for Bcl-2 (left) and survivin (right) in PC3 and DU145 cells. All values werenormalized to the levels detected in unstimulated cells. Data shown are the mean � S.E.M. values, calculated from three independent experiments.b, immunoblots demonstrating the time-dependent increases in Bcl-2 and survivin expression in PC3 (left) and DU145 (right) cells after stimulationwith 3 nM rCXCL8. c, immunoblots demonstrating the time-dependent increases in Bcl-2 and survivin expression in PC3 (left) and DU145 (right) cellsafter stimulation with 3 nM rCXCL1. All immunoblots shown in the figure are representative of two to three experiments. d, bar graph illustratingthe effect of the CXCR1 antagonist, AZ10397767 (AZ767), or the NF-�B antagonist BAY 11-7082 on rCXCL8-induced increases in the mRNA transcriptlevels detected for Bcl-2 (left) and survivin (right) in PC3 and DU145 cells. Data shown are the mean � S.E.M. values, calculated from five independentexperiments. Statistically significant differences in mRNA transcript level were determined using Student’s t test: �, p � 0.05; ��, p � 0.01; and ���,p � 0.001.

752 Wilson et al.

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

AZ10397767 with 0.1 or 1 �M oxaliplatin resulted in a markedincrease in the sub-G0/G1 cell population in either cell line (Fig.6b, left and right). In PC3 and DU145 cells, respectively, thelevel of apoptosis induced by oxaliplatin increased from 5.48 �0.69 to 10.55 � 2.5% (p � 0.05; n 4) and 4.99 � 0.49 to 9.13 �1.74% (p � 0.05; n 3) in the presence of the CXCR2 antago-nist. AZ10397767-mediated potentiation of oxaliplatin-inducedapoptosis was also studied by immunoblotting for cleavage ofthe caspase-3 and caspase-7 substrate PARP (Fig. 6c, left andright). Although treatment with 1 �M oxaliplatin reduced full-length PARP expression in PC3 cells relative to untreated cells,the concurrent administration of AZ10397767 with 1 �M oxali-platin effected a more significant reduction in PARP expression.Likewise, in the DU145 cells, AZ10397767 had the ability toenhance oxaliplatin-induced cleavage of PARP. In accordance,these cell-based and molecular experiments indicate that inhi-bition of CXCR2 receptor signaling potentiates oxaliplatin-in-duced apoptosis in AIPC cells.

Inhibition of NF-�B Signaling Increases Sensitivityof AIPC Cells to Oxaliplatin. Further cell count assays wereconducted to determine whether the suppression of drug-induced NF-�B activity by AZ10397767 may explain its abil-ity to enhance oxaliplatin cytotoxicity in AIPC cells. In these

experiments, the activity of NF-�B was perturbed using wellcharacterized pharmacological inhibitors of this transcrip-tion factor. Administration of BAY 11-7082 at a final concen-tration of 1 �M to PC3 and DU145 cells was shown to beeffective in enhancing the cytotoxicity of oxaliplatin in eachof these AIPC cells (Fig. 7a). As observed with AZ10397767,noniterative nonlinear regression analysis of the cell countdata predicted a two-site model to be more appropriate inmodeling the data resulting from the coadministration ofoxaliplatin and BAY 11-7082. The combination of oxaliplatinwith BAY 11-7082 increased the calculated IC25 value from0.73 � 0.63 to 0.008 � 0.01 �M (p � 0.05; n 3) in PC3 cells.In DU145 cells, inhibition of NF-�B activity using BAY 11-7082 increased the calculated IC30 from 0.287 � 0.456 to0.015 � 0.04 �M (p � 0.05; n 3). Although oxaliplatin wasineffective in inducing apoptosis in either cell line, treatmentwith 1 �M BAY 11-7082 did induce apoptosis in PC3 (Fig. 7b,left) and DU145 (Fig. 7b, right) cells. Furthermore, the inhi-bition of NF-�B activity using BAY 11-7082 further enhancedthe apoptosis induced by exposure of PC3 or DU145 cells to0.1 and/or 1 �M oxaliplatin (Fig. 7b). These results wereverified by use of a second inhibitor of NF-�B, the I�K-2inhibitor, SC-514, which was administered to PC3 cells at a

Fig. 5. Inhibition of CXCR2 signaling attenuates oxaliplatin-induced NF-�B activation. a, EMSA blot illustrating the L-OHP-stimulated binding ofNF-�B to an oligonucleotide containing the consensus nucleotide sequence for this transcription factor, in the absence and presence of the CXCR2antagonist, AZ10397767 (AZ767). In these experiments, cells were treated with 1 �M L-OHP, and AZ10397767 was coadministered where indicatedat a concentration of 20 nM. b, bar graph illustrating the magnitude and time-dependent induction of NF-�B-induced luciferase activity in PC3 andDU145 cells after a 24-h stimulation with 1 �M L-OHP, in the absence and presence of the CXCR2 antagonist, AZ767. Data shown are the mean �S.E.M. values, calculated from four independent experiments. c, bar graph illustrating the time-dependent alterations in mRNA transcript levelsdetected for CXCL8, CXCL1, Bcl-2, and survivin in PC3 (left) and DU145 (right) cells after a 24-h stimulation with 1 �M oxaliplatin in the absenceor presence of AZ767. Data shown are the mean � S.E.M. values, calculated from five and four independent experiments, respectively. Statisticallysignificant differences in mRNA transcript level were determined using Student’s t test: �, p � 0.05; ��, p � 0.01; and ���, p � 0.001.

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 753

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

final concentration of 2 �M. SC-514 inhibited oxaliplatin (1�M)-induced NF-�B binding in EMSA analysis, sensitizedPC3 cells to sub-micromolar concentrations of oxaliplatin incell count analysis, and, when coadministered with 0.1 �Moxaliplatin, SC-514 increased the accumulation of apoptoticcells detected by FACS analysis in the sub-G0/G1 peak rela-tive to that observed when treated with 0.1 �M oxaliplatinalone (data not shown).

Suppression of Bcl-2 and Survivin Sensitizes PC3Cells to Oxaliplatin. The functional significance of thedownstream transcriptional targets of NF-�B, Bcl-2 and sur-vivin, in underpinning the resistance of AIPC cells to oxali-platin was studied using gene-targeted RNA interference-based approaches to selectively suppress the expression ofthese genes in AIPC cells. PC3 cells were initially transfectedusing 50 nM of a single gene-specific oligonucleotide (oligo)against either Bcl-2 (Bcl2-T) or survivin (Sur-T) or with ascrambled oligonucleotide (Sc), used at an identical concen-tration. At this concentration, transfection of PC3 cells withthe Bcl2-T oligo was shown to decrease but not abrogate Bcl-2

expression in these cells 72 h after transfection. In contrast,transfection with Sur-T oligonucleotide was shown to bemarkedly more effective in depleting survivin expression inthe PC3 cells (Fig. 8a, top and bottom).

The effect of transiently reducing antiapoptotic gene ex-pression upon the induction of apoptosis in PC3 cells in theabsence and presence of oxaliplatin was determined usingFACS analysis. Initially, PC3 cells were treated with 50 nMBcl2-T oligo, Sur-T oligo, or Sc oligo for 48 h, after which cellswere exposed to increasing concentrations of oxaliplatin for afurther 72 h. In control experiments, transfection of PC3 cellsusing concentrations up to 200 nM Sc oligo did not increasethe level of apoptosis over that observed in control PC3 cells(data not shown).

Transfection with 50 nM Bcl2-T alone increased the mag-nitude of the sub-G0/G1 fraction of the cell population (8.53 �0.57%, p � 0.01, n 4) relative to both mock-transfectedcontrol (4.68 � 0.18%) and scrambled oligo-transfected(4.75 � 0.18%) (Fig. 7c) cells, suggesting that even subopti-mal reductions in the expression of this protein initiates

Fig. 6. Inhibition of CXCR2 signaling potenti-ates oxaliplatin cytotoxicity and oxaliplatin-in-duced apoptosis in PC3 cells. a, graph illustrat-ing the effect of increasing concentrations ofL-OHP upon the viability of PC3 cell populations(left) and DU145 (right) cell populations in theabsence or presence of 20 nM AZ767, calculatedover the course of 72h. b, bar graph illustratingthe percentage of apoptotic cells detected in PC3(left) or DU145 (right) cell populations aftertreatment with 1 �M L-OHP in the absence orpresence of coadministering AZ767 (20 nM). Val-ues shown represent the percentage of cells de-tected in the sub-G0/G1 fraction as determined byFACS analysis. c, immunoblots illustrating lossof full-length PARP expression in PC3 cells (left)and PARP cleavage in DU145 cells (right) inresponse to addition of 1 �M oxaliplatin in theabsence and presence of AZ767. Equal proteinloading was confirmed by reprobing the mem-brane for GAPDH expression. Statistically sig-nificant differences in cell viability or apoptoticcell populations were determined using Stu-dent’s t test: �, p � 0.05; and ��, p � 0.01.

754 Wilson et al.

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

spontaneous apoptosis in PC3 cells. In contrast, abrogation ofsurvivin expression using the Sur-T oligo only effected amodest increase in the apoptotic fraction (7.55 � 0.4% of thecell population; p � 0.01), indicating that PC3 cells are notcritically dependent upon survivin for viability. Althoughtreatment with 1 �M oxaliplatin alone was largely ineffectivein increasing the level of apoptosis in Sc oligo-transfectedPC3 cells, oxaliplatin did increase the apoptotic cell popula-tion to 10.58 � 0.72 and 8.3 � 0.4% in 50 nM Bcl2-T-trans-fected PC3 cells (p � 0.001; n 4) and 50 nM Sur-T oligo-transfected PC3 cells (p � 0.05, n 4), respectively (Fig. 8b).The relative importance of these antiapoptotic proteins inmodulating PC3 cell viability and the sensitivity of thesecells to oxaliplatin was reaffirmed using PARP cleavage as amarker of apoptosis induction. In Sc oligo-transfected cells,treatment with 1 �M oxaliplatin resulted in a modest de-crease in PARP expression. In contrast, transfection of PC3cells with 50 nM Bcl2-T had a pronounced effect in promotingPARP degradation in both the absence and presence of 1 �Moxaliplatin, shown by the reduced expression of the full-length protein. Suppression of survivin alone had no effect onPARP expression but did potentiate the degradation of PARPobserved when cells were cotreated with 1 �M oxaliplatin(Fig. 8c). Therefore, decreasing Bcl-2 expression in PC3 cellsappears to be of major importance in sensitizing cells tooxaliplatin, whereas suppressing survivin expression con-tributes to an enhanced response.

To determine the effect of increasing the effectiveness oftargeting Bcl-2 expression, PC3 cells were treated withincreasing concentrations of the Bcl2-T oligo. Transfectionwith 200 nM Bcl2-T oligo was shown to deplete endogenousBcl-2 expression in these cells in immunoblots conducted

on protein lysates extracted from transfected cells (Fig.8d). Suppression of Bcl-2 expression alone was observed tocoincide with a spontaneous loss of full-length PARP ex-pression and the emergence of PARP cleavage, indicativeof apoptosis. Furthermore, use of this Bcl2-T oligo at thishigher concentration induced a more marked increase inthe apoptotic cell fraction detected by FACS analysis, in-creasing the level to 20.4 � 2.9% (p � 0.001) (Fig. 8e). Afurther increase in the apoptotic cell fraction was observedafter the addition of 1 �M oxaliplatin to 200 nM Bcl2-T-transfected PC3 cells. However, because of the high level ofspontaneous apoptosis resulting from treatment with theBcl2-T oligo at this higher concentration, there was nostatistically significant enhancement when 200 nM Bcl2-Twas combined with 1 �M oxaliplatin.

Concurrent Suppression of Bcl-2 and Survivin Po-tentiates L-OHP Cytotoxicity in PC3 Cells. Bcl-2 primar-ily antagonizes Bax/Bak-induced mitochondrial membranepermeabilization, whereas the survivin/IAP gene family di-rectly inhibits the activation of downstream caspases, includ-ing caspase-9 (Reed, 2001). To determine whether oxaliplatinand CXCL8-promoted increases in Bcl-2 and survivin expres-sion describe two complementary mechanisms of resistanceto oxaliplatin, we determined the effect of concurrently sup-pressing Bcl-2 and survivin expression on oxaliplatin-in-duced apoptosis. The sensitivity of Bcl2-T/Sur-T cotrans-fected cells to undergo apoptosis was compared with theeffect of transfecting PC3 cells with equivalent concentra-tions of either the Bcl2-T or Sur-T oligos alone or with a Scoligo (Fig. 8b). All transfections were conducted using a finaloligonucleotide concentration of 50 nM, given our observationof high levels of spontaneous apoptosis when antiapoptotic

Fig. 7. Inhibition of the NF-�B signalingpathway sensitizes AIPC cells to oxalip-latin. a, graphs illustrating the effectof coadministering the NF-�B inhibitorBAY 11-7082 at a concentration of 1 �Mupon the cytotoxicity of L-OHP in PC3(left) and DU145 (right) cells. Cells wereexposed to both agents for 72 h. Datashown are the mean � S.E.M. of fourindependent experiments. b, bar graphspresenting the level of apoptosis detectedby FACS analysis in untreated, L-OHP-treated, BAY-11–702-treated, and L-OHP/BAY 11-7082-treated PC3 (left) and DU145(right) cells. Data shown are the mean �S.E.M. of four independent experiments.Statistically significant differences in theapoptotic cell populations were determinedusing Student’s t test: �, p � 0.05; and ��,p � 0.01.

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 755

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

gene oligos were used at increased concentrations. Cotrans-fection with 50 nM Bcl2-T and 50 nM Sur-T oligos was shownto induce a more pronounced accumulation of cells in thesub-G0/G1 peak (15.0 � 1.3%; n 4) relative to transfectionwith Bcl2-T (8.53 � 0.58%, p � 0.05) or Sur-T (7.55 � 0.41%)alone. In the context of sensitizing PC3 cells to oxaliplatin,the combined suppression of Bcl-2 and survivin potentiatedthe level of apoptosis detected in PC3 cells in response tooxaliplatin (sub-G0/G1 fraction of 18.77 � 0.89%) in Bcl2-T/Sur-T cotransfected cells relative to that observed in Bcl2-T

(10.58 � 0.72%, p � 0.05, n 4) or Sur-T transfected (8.3 �0.39%, p � 0.01, n 4) cells.

DiscussionDeregulation of NF-�B has been proposed as a major factor

contributing to the resistance of AIPC to chemotherapy. Nu-clear localization of this transcription factor, indicative ofconstitutive activity, has been detected in biopsy tissues ofAIPC (Fradet et al., 2004; Sweeney et al., 2004). Further

Fig. 8. Attenuation of antiapoptotic gene expression induces spontaneous apoptosis and increases oxaliplatin-induced apoptosis in PC3 cells. a,immunoblots demonstrating the relative suppression of Bcl-2 or survivin in PC3 cells 72 h after transfection with either the Bcl2-T oligonucleotide orSur-T oligonucleotide, respectively, at a final concentration of 50 nM. The effect of the gene-targeted oligonucleotides is shown relative to that observedafter transfection of PC3 cells with a scrambled oligonucleotide (Sc) at 50 nM. b, graph illustrating the percentage of apoptotic cells detected as thesub-G0/G1 fraction by FACS analysis in PC3 cell populations after transfection with the Bcl2-T or Sur-T oligonucleotides at a concentration of 50 nM,singly or in combination, in the absence or presence of 1 �M L-OHP. The effect of transfection cells with the scrambled oligonucleotide at a similarconcentration is shown for comparison. Values shown represent the mean � S.E.M. percentage, determined in four independent experiments. c,immunoblots illustrating the effect of suppressing Bcl-2 or survivin expression using 50 nM Bcl2-T or Sur-T, respectively, upon PARP cleavage in PC3cells in the absence and presence of 1 �M L-OHP. The response is shown relative to the effect of transfecting PC3 cells with 50 nM Sc oligo. Equalprotein loading was confirmed by reprobing the membrane for GAPDH expression. d, immunoblots reporting the expression of full-length PARP,cleaved PARP, and Bcl-2 protein expression after the transfection of PC3 cells with 200 nM Bcl2-T oligo, in the absence and presence of 1 �Moxaliplatin. Equal protein loading was confirmed by reprobing the membrane for GAPDH expression. e, bar graph illustrating the percentage ofapoptotic cells detected as the sub-G0/G1 fraction by FACS analysis in PC3 cell populations after transfection with the Bcl2-T oligonucleotide at aconcentration of 200 nM, in the absence or presence of 1 �M L-OHP. The effect of transfection cells with the scrambled oligonucleotide at a similarconcentration (200 nM) is shown for comparison. Values shown represent the mean � S.E.M. percentage, determined in four independent experiments.Statistical analysis of data presented in (b) or (e) was performed using a Student’s t test, comparing the effect of the combined oligonucleotidetransfection (Combo) with that of Sc oligo-treated cells (directly above column) or with individual gene-targeted transfections as indicated. �, p � 0.05;��, p � 0.01; ���, p � 0.001.

756 Wilson et al.

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

studies conducted in cell lines have shown that inhibition ofNF-�B renders CaP cells more susceptible to apoptosis, me-diated in part by the ability of this transcription factor toregulate antiapoptotic gene expression (Huang et al., 2005;Li et al., 2005). The sensitivity of colorectal cancer cells tooxaliplatin-induced death is adversely affected by elevatedNF-�B activity (Rakitina et al., 2003), suggesting that thedysregulation of this transcription factor in AIPC may un-derpin the low sensitivity of this disease to oxaliplatin (Drozet al., 2003).

We have shown that administration of oxaliplatin furtherpotentiates NF-�B activity in AIPC cells, increasing tran-scriptional regulation and expression of antiapoptotic genes.Furthermore, we have shown that oxaliplatin induced a dy-namic process of increased CXCL8 gene transcription, intra-cellular protein expression, and ultimately secretion fromPC3 cells that is predominantly mediated downstream ofNF-�B activation. Oxaliplatin was also shown to increase thetranscription and expression of the CXCL8 receptors, CXCR1and CXCR2, in AIPC cells. To our knowledge, this is the firstreport demonstrating an effect of chemotherapy drugs uponthe expression of these CXC-chemokine receptors in cancercells. Furthermore, these observations are consistent withour recent demonstration that hypoxia-induced transcriptionof the CXCR1 and CXCR2 receptors is mediated in part byincreased NF-�B transcriptional activity (Maxwell et al.,2007). As a consequence, our demonstration of increasedligand and receptor expression suggests that exposure ofAIPC cells to oxaliplatin increases the magnitude of theautocrine CXCL8 signaling stimulus received by these cells.

To determine whether drug-induced or constitutive CXCL8signaling may alter the sensitivity of the cells to the chemo-therapy drug, we characterized the effect of stimulatingAIPC cells with rCXCL8, focusing on the NF-�B cell survivalpathway. Administration of pharmacologically relevant con-centrations of rCXCL8 induced NF-�B activation, enhancingthe transcription and expression of the antiapoptotic genes,Bcl-2 and survivin, in AIPC cells. These responses were at-tenuated by coadministration of the CXCR2 receptor antag-onist, AZ10397767, at a concentration that selectively blocksCXCR2 activation. Furthermore, pharmacological inhibitionof NF-�B attenuated the rCXCL8-induced increase in Bcl-2expression, suggesting that the CXCR2 receptor primarily,but not exclusively, couples CXCL8 to the activation ofNF-�B and transcriptional regulation of these antiapoptoticgenes in AIPC cells.

The selective increase in CXCR2 expression suggests thatsignaling through this receptor may predominate overCXCR1 in dictating the response of AIPC cells to the admin-istration of oxaliplatin. Although CXCR2 binds and is acti-vated by CXCL8, this receptor also binds several relatedCXC-chemokines, including the growth-related oncogenes(CXCL1, CXCL2, CXCL3), whose expression is also regulatedat the level of transcription by NF-�B (Brat et al., 2005). Asa consequence, we have shown that oxaliplatin treatmentalso induces expression of CXCL1 in AIPC cells. As observedfor CXCL8, stimulation of these AIPC cells with CXCL1potentiates a CXCR2-dependent activation of NF-�B and in-creases antiapoptotic protein expression. Irrespective of theligand activating the receptor, the significance of CXCR2-induced signaling in modulating the sensitivity of AIPC cellsto oxaliplatin was illustrated by the ability of the CXCR2

antagonist, AZ10397767, to attenuate oxaliplatin-inducedNF-�B activation and reduce oxaliplatin-induced transcrip-tion of the CXCL8, CXCL1, Bcl-2, and survivin genes. Coad-ministration of AZ10397767 also markedly increased thesensitivity of AIPC cells to oxaliplatin and enhanced oxali-platin-induced apoptosis.

Oxaliplatin cytotoxicity in AIPC cells was also increased bydirect targeting of NF-�B activation or its downstream tran-scriptional targets. Pharmacological inhibition of NF-�B us-ing the BAY 11-7082 or SC-514 compounds sensitized AIPCcells to low but not high concentrations of oxaliplatin,whereas RNA interference-mediated suppression of Bcl-2and survivin was shown to increase the level of oxaliplatin-induced apoptosis detected in PC3 cells. Interestingly, reduc-tion of Bcl-2 and survivin expression alone was shown toinduce apoptosis in PC3 cells, suggesting that this AIPC cellline may have a marked dependence on the expression ofthese antiapoptotic proteins for maintenance of cell viability.The concurrent suppression of both Bcl-2 and survivin wasalso shown to result in a significantly greater induction ofapoptosis in PC3 cells in response to oxaliplatin. Therefore,we propose that the potentiation of CXCL8- and CXCL1-induced CXCR2 signaling underpins a novel and contribut-ing mode of resistance to oxaliplatin that is mediated, at leastin part, by sustaining NF-�B transcription and reinforcingthe expression of antiapoptotic proteins in AIPC cells (Fig. 8).

Prior studies have shown that the combination of oxalipla-tin with the heat shock protein-90 inhibitor 17-N-allylamino-17-demethoxygeldanamycin results in a more effective sup-pression of cellular NF-�B activation and promotion ofanticancer activity (Rakitina et al., 2003). Therefore, theability of AZ10397767 to also suppress NF-�B activity andreduce antiapoptotic protein expression in the presence ofoxaliplatin is a likely explanation of its capacity to increasethe cytotoxicity of this drug in AIPC cells. Our observationsthat coadministration of either the CXCR2 antagonist or theNF-�B inhibitor BAY 11-7082 sensitizes the AIPC cells to lowrather than high concentrations of oxaliplatin suggests thatCXCL8/CXCL1-induced NF-�B-signaling in AIPC cells con-fers a surmountable resistance to a specific mode of oxalipla-tin-induced cytotoxicity. Further detailed studies will be re-quired to establish the precise mechanism (mitochondrial ordeath receptor mediated) by which CXCR2 and NF-�B sig-naling impact on oxaliplatin-induced apoptosis in AIPC cells.Prior studies have shown that oxaliplatin induces apoptosisin colorectal cancer cells through the activation of caspase-8and Bid cleavage, suggesting a role for the extrinsic apoptosispathway (Griffiths et al., 2004; Longley et al., 2006). Ourcurrent observations that CXCL8 signaling regulates Bcl-2gene family expression together with our recently publisheddata demonstrating that this chemokine promotes a NF-�B-driven expression of the endogenous caspase-8 inhibitorc-FLIP in CaP cells (Wilson et al., 2008) suggest that CXCL8signaling may attenuate the capacity of oxaliplatin to initiatemitochondrial-driven apoptosis in these cells by altering theratio of proapoptotic cleaved Bid to the mitochondrial stabi-lizing effect of increased Bcl-2 expression. Furthermore, ourobservation that CXCL8/CXCL1 signaling increases survivinexpression in AIPC cells suggests a further mechanismdownstream of the mitochondria by which AIPC cells maywithstand the apoptosis-inducing stress experienced afterthe administration of oxaliplatin.

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 757

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

Several final issues warrant mention. First, oxaliplatin iscurrently employed in the treatment of metastatic colorectalcancer. Experiments conducted in our laboratory have char-acterized a similar induction of CXCL8 signaling in colorectalcancer cells and demonstrated that the inhibition of chemo-kine signaling also has a similar effect in sensitizing colorec-tal cancer cells to oxaliplatin (C. Purcell and D. Waugh,unpublished observations). Therefore, these ongoing studiespoint to a clinical relevance of CXC-chemokine signaling indetermining the response of malignant cells to oxaliplatin inthis disease setting. Furthermore, in conducting experimentsin the parental p53 wild-type HCT116 and the matched p53-deficient HCT116 cells, we will be able to undertake a directinvestigation of how p53 status influences the magnitude ofthe CXC-chemokine-promoted antiapoptotic response in can-cer cells. This has not been possible in the current studiesgiven the p53-deficient and p53 mutant status in the PC3and DU145 cell lines, respectively. Finally, we have alsoconducted experiments to determine whether other platinumagents induce a similar response in prostate cancer cells.Administration of cisplatin was also shown to induce NF-�Bactivity and increase CXC-chemokine expression in AIPCcells. However, inhibition of CXCR2 signaling did not in-crease the sensitivity of these cells to cisplatin (C. Wilson andD. Waugh, unpublished observations). We suspect that thisreflects the inability of this pharmacological intervention toovercome the primary mode of resistance, i.e., that of mis-match repair enzyme deficiency in AIPC cells (Chen et al.,2001) that drives resistance to cisplatin but not oxaliplatinadducts (Zdraveski et al., 2002; Chaney et al., 2004).

In summary, we indicate a role for NF-�B signaling andtwo of its transcriptional targets, Bcl-2 and survivin, in at-tenuating oxaliplatin cytotoxicity in AIPC cells. Further-more, we report the significance of oxaliplatin-induced CXC-chemokine signaling in reinforcing NF-�B transcriptionalactivity and, consequently, potentiating the resistance ofthese cells to this platinum drug. In accordance, direct tar-geting of NF-�B, its transcriptional targets, or inactivatingCXCR2-mediated signaling may be appropriate strategies toincrease the effectiveness of using oxaliplatin to treat cancer(Fig. 9). The demonstration that CXCL8 signaling potenti-ates antiapoptotic protein expression and attenuates chemo-therapy-induced apoptosis further links CXCL8 signaling toanother described “hallmark” of cancer cells (Hanahan andWeinberg, 2000), that of evading apoptosis, in addition to its

proposed roles in promoting angiogenesis, invasion and me-tastasis.

Acknowledgments

We thank Simon T. Barry and David Blakey (AstraZeneca, Ald-erley Park, Cheshire, UK) for provision of AZ10397767 and StephenMcNiece and the Dispensing Pharmacy of the Bridgewater Suite,Belfast City Hospital for consultation and for provision of the che-motherapy drugs used in the study.

ReferencesAbdel-Latif MM, O’Riordan JM, Ravi N, Kelleher D, and Reynolds JV (2005) Acti-

vated nuclear factor-kappa b and cytokine profiles in the esophagus parallel tumorregression following neoadjuvant chemoradiotherapy. Dis Esophagus 18:246–252.

Abdollahi T, Robertson NM, Abdollahi A, and Litwack G (2003) Identification ofinterleukin 8 as an inhibitor of tumor necrosis factor-related apoptosis-inducingligand-induced apoptosis in the ovarian carcinoma cell line OVCAR3. Cancer Res63:4521–4526.

Brat DJ, Bellail AC, and Van Meir EG (2005) The role of interleukin-8 and itsreceptors in gliomagenesis and tumoral angiogenesis. Neuro Oncol 7:122–133.

Brennecke S, Deichmann M, Naeher H, and Kurzen H (2005) Decline in angiogenicfactors, such as interleukin-8, indicates response to chemotherapy of metastaticmelanoma. Melanoma Res 15:515–522.

Chaney SG, Campbell SL, Temple B, Bassett E, Wu Y, and Faldu M (2004) Proteininterations with platinum-DNA adducts: from structure to function. J Inorg Bio-chem 98:1551–1559.

Chen Y, Wang J, Fraig MM, Metcalf J, Turner WR, Bissada NK, Watson DK, andSchweinfest CW (2001) Defects of DNA mismatch repair in human prostate can-cer. Cancer Res 61:4112–4121.

Collins TS, Lee LF, and Ting JP (2000) Paclitaxel up-regulates interleukin-8 syn-thesis in human lung carcinoma through an NF-kappaB- and AP-1-dependentmechanism. Cancer Immunol Immunother 49:78–84.

De Larco JE, Wuertz BR, Manivel JC, and Furcht LT (2001) Progression andenhancement of metastatic potential after exposure of tumor cells to chemother-apeutic agents. Cancer Res 61:2857–2861.

Droz JP, Muracciole X, Mottet N, Ould Kaci M, Vannetzel JM, Albin N, Culine S,Rodier JM, Misset JL, Mackenzie S, et al. (2003) Phase II study of oxaliplatinversus oxaliplatin combined with infusional 5-fluorouracil in hormone refractorymetastatic prostate cancer patients. Ann Oncol 14:1291–1298.

Duan Z, Lamendola DE, Penson RT, Kronish KM, and Seiden MV (2002) Overex-pression of IL-6 but not IL-8 increases paclitaxel resistance of U-2OS humanosteosarcoma cells. Cytokine 17:234–242.

Fradet V, Lessard L, Begin LR, Karakiewicz P, Masson AM, and Saad F (2004)Nuclear factor-kappaB nuclear localization is predictive of biochemical recurrencein patients with positive margin prostate cancer. Clin Cancer Res 10:8460–8464.

Griffiths GJ, Koh MY, Brunton VG, Cawthorne C, Reeves NA, Greaves M, Tilby MJ,Pearson DG, Ottley CJ, Workman P, et al. (2004) Expression of kinase-defectivemutants of c-Src in human metastatic colon cancer cells decreases Bcl-xL andincreases oxaliplatin- and Fas-induced apoptosis. J Biol Chem 279:46113–46121.

Hanahan D and Weinberg RA (2000) The hallmarks of cancer. Cell 100:57–70.Huang J, Yao JL, Zhang L, Bourne PA, Quinn AM, di Sant’Agnese PA, and Reeder

JE (2005a) Differential expression of interleukin-8 and its receptors in the neu-roendocrine and non-neuroendocrine compartments of prostate cancer. Am JPathol 166:1807–1815.

Huang YT, Pan SL, Guh JH, Chang YL, Lee FY, Kuo SC, and Teng CM (2005b) YC-1suppresses constitutive nuclear factor-kappaB activation and induces apoptosis inhuman prostate cancer cells. Mol Cancer Ther 4:1628–1635.

Inoue K, Slaton JW, Eve BY, Kim SJ, Perrotte P, Balbay MD, Yano S, Bar-Eli M,Radinsky R, Pettaway CA, et al. (2000) Interleukin 8 expression regulates tumor-igenicity and metastases in androgen-independent prostate cancer. Clin CancerRes 6:2104–2119.

Fig. 9. Schematic representation depict-ing the importance of NF-�B in oxalipla-tin resistance. a, oxaliplatin promotesactivation of NF-�B in AIPC cells, result-ing in increased transcription/expressionof antiapoptotic protein expression,CXCL8, CXCL1, and CXCR2. The result-ant autocrine/paracrine signaling rein-forces NF-�B activation, promoting a sur-vival phenotype in the cells. b, inhibitionof CXCR2 signaling, NF-�B activation,Bcl-2, and/or survivin expression atten-uates the survival phenotype and poten-tiates oxaliplatin-induced apoptosis inAIPC cells.

758 Wilson et al.

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from

Kim SJ, Uehara H, Karashima T, Mccarty M, Shih N, and Fidler IJ (2001) Expres-sion of interleukin-8 correlates with angiogenesis, tumorigenicity, and metastasisof human prostate cancer cells implanted orthotopically in nude mice. Neoplasia3:33–42.

Lev DC, Ruiz M, Mills L, McGary EC, Price JE, and Bar-Eli M (2003) Dacarbazinecauses transcriptional up-regulation of interleukin 8 and vascular endothelialgrowth factor in melanoma cells: a possible escape mechanism from chemotherapy.Mol Cancer Ther 2:753–763.

Li Y, Ahmed F, Ali S, Philip PA, Kucuk O, and Sarkar FH (2005) Inactivation ofnuclear factor kappaB by soy isoflavone genistein contributes to increased apopto-sis induced by chemotherapeutic agents in human cancer cells. Cancer Res 65:6934–6942.

Longley DB, Allen WL, McDermott U, Wilson TR, Latif T, Boyer J, Lynch M, andJohnston PG (2004) The roles of thymidylate synthase and p53 in regulatingFas-mediated apoptosis in response to antimetabolites. Clin Cancer Res 10:3562–3571.

Longley DB, Wilson TR, McEwan M, Allen WL, McDermott U, Galligan L, andJohnston PG (2006) c-FLIP inhibits chemotherapy-induced colorectal cancer celldeath. Oncogene 25:838–848.

MacManus CF, Pettigrew J, Seaton A, Wilson C, Maxwell PJ, Berlingeri S, PurcellC, McGurk M, Johnston PG, and Waugh DJJ (2007) Interleukin-8 signalingpromotes translational regulation of cyclin D in androgen-independent prostatecancer cells. Mol Cancer Res 5:737–748.

Maxwell PJ, Gallagher R, Seaton A, Wilson C, Scullin P, Pettigrew J, Stratford IJ,Williams KJ, Johnston PG, and Waugh DJJ (2007) HIF-1 and NF-�B-mediatedupregulation of CXCR1 and CXCR2 expression promotes cell survival in hypoxicprostate cancer cells. Oncogene 26:7333–7345.

McCarron SL, Edwards S, Evans PR, Gibbs R, Dearnaley DP, Dowe A, Southgate C,Easton DF, Eeles RA, and Howell WM (2002) Influence of cytokine gene polymor-phisms on the development of prostate cancer. Cancer Res 62:3369–3372.

Murphy C, McGurk M, Pettigrew J, Santinelli A, Mazzucchelli R, Johnston PG,Montironi R, and Waugh DJ (2005) Non-apical and cytoplasmic expression ofinterleukin-8, CXCR1 and CXCR2 correlates with cell proliferation and microves-sel density in prostate cancer. Clin Cancer Res 11:4117–4127.

Orditura M, De Vita F, Catalano G, Infusino S, Lieto E, Martinelli E, Morgillo F,Castellano P, Pignatelli C, and Galizia G (2002) Elevated serum levels of inter-leukin-8 in advanced non-small cell lung cancer patients: relationship with prog-nosis. J Interferon Cytokine Res 22:1129–1135.

Penson RT, Kronish K, Duan Z, Feller AJ, Stark P, Cook SE, Duska LR, Fuller AF,Goodman AK, Nikrui N, et al. (2000) Cytokines IL-1beta, IL-2, IL-6, IL-8, MCP-1,GM-CSF and TNFalpha in patients with epithelial ovarian cancer and theirrelationship to treatment with paclitaxel. Int J Gynecol Cancer 10:33–41.

Rakitina TV, Vasilevskaya IA, and O’Dwyer PJ (2003) Additive interaction of oxali-platin and 17-allylamino-17-demethoxygeldanamycin in colon cancer cell lines

results from inhibition of nuclear factor kappaB signaling. Cancer Res 63:8600–8605.

Reed JC (2001) The survivin saga goes in vivo. J Clin Invest 108:965–969.Shibakura M, Niiya K, Kiguchi T, Kitajima I, Niiya M, Asaumi N, Huh NH, Nakata

Y, Harada M, and Tanimoto M (2003) Induction of IL-8 and monoclyte chemoat-tractant protein-1 by doxorubicin in human small cell lung carcinoma cells. Int JCancer 103:380–386.

Sweeney C, Li L, Shanmugam R, Bhat-Nakshatri P, Jayaprakasan V, Baldridge LA,Gardner T, Smith M, Nakshatri H, and Cheng L (2004) Nuclear factor-kappaB isconstitutively activated in prostate cancer in vitro and is overexpressed in pros-tatic intraepithelial neoplasia and adenocarcinoma of the prostate. Clin CancerRes 10:5501–5507.

Tamatani T, Azuma M, Ashida Y, Motegi K, Takashima R, Harada K, Kawaguchi S,and Sato M (2004) Enhanced radiosensitization and chemosensitization in NF-kappaB-suppressed human oral cancer cells via the inhibition of gamma-irradiation- and 5-FU-induced production of IL-6 and IL-8. Int J Cancer 108:912–921.

Uehara H, Troncoso P, Johnston D, Bucana CD, Dinney C, Dong Z, Fidler IJ, andPettaway CA (2005) Expression of interleukin-8 gene in radical prostatectomyspecimens is associated with advanced pathologic stage. Prostate 64:40–49.

Uslu R, Sanli UA, Dikmen Y, Karabulut B, Ozsaran A, Sezgin C, Muezzinoglu GG,Omay SB, and Goker E (2005) Predictive value of serum interleukin-8 levels inovarian cancer patients treated with paclitaxel-containing regimens. Int J GynecolCancer 15:240–245.

Veltri RW, Miller MC, Zhao G, Ng A, Marley GM, Wright GL Jr, Vessella RL, andRalph D (1999) Interleukin-8 serum levels in patients with benign prostatichyperplasia and prostate cancer. Urology 53:139–147.

Walters I, Austin C, Austin R, Bonnert R, Cage P, Christie M, Ebden M, Gardiner S,Grahames C, Hill S, et al. (2008) Evaluation of a series of bicyclic CXCR2 antag-onists. Bioorg Med Chem Lett 18:798–803.

Wilson C, Wilson T, Johnston PG, Longley DB, and Waugh DJJ (2008) Interleukin-8signaling attenuates TRAIL and chemotherapy-induced apoptosis through tran-scriptional regulation of c-FLIP in prostate cancer cells. Mol Cancer Ther 7:2649–2661.

Woynarowski JM, Faivre S, Herzig MC, Arnett B, Chapman WG, Trevino AV,Raymond E, Chaney SG, Vaisman A, Varchenko M, et al. (2000) Oxaliplatin-induced damage of cellular DNA. Mol Pharmacol 58:920–927.

Zdraveski ZZ, Mello JA, Farinelli CK, Essigmann JM, and Marinus MG (2002) MutSpreferentially recognizes cisplatin- over oxaliplatin-modified DNA. J Biol Chem277:1255–1260.

Address correspondence to: Dr. David Waugh, Centre for Cancer Researchand Cell Biology, Queens University Belfast, 97 Lisburn Road, Belfast BT97BL, UK. E-mail: d.waugh@qub.ac.uk

CXC-Chemokine Signaling Attenuates Platinum Cytotoxicity 759

at ASPE

T Journals on O

ctober 21, 2018jpet.aspetjournals.org

Dow

nloaded from