Adenovirally mediated p53 overexpression diversely influence the cell cycle of HEp2 and CAL 27 cell lines upon cisplatin and methotrexate treatment

J Cancer Res Clin Oncol (2009) 135:1747–1761

ORIGINAL PAPER

Adenovirally mediated p53 overexpression diversely inXuence the cell cycle of HEp-2 and CAL 27 cell lines upon cisplatin and methotrexate treatment

Sandra KraljeviT PaveliT · Marko MarjanoviT · Miroslav PozniT · Marijeta Kralj

Received: 5 September 2008 / Accepted: 2 June 2009 / Published online: 23 June 2009© Springer-Verlag 2009

AbstractPurpose p53 gene plays a crucial role in the response totherapy. Since it is inactivated in the majority of humancancers, it is strongly believed that the p53 mutationsconfer resistance to therapeutics. In this paper we analyzedthe inXuence of two mechanistically diverse antitumoragents—cisplatin and methotrexate on the proliferation andcell cycle of two head and neck squamous cancer cell linesHEp-2 (wild type p53 gene, but HPV 18/E6-inactivatedprotein) and CAL 27 (mutated p53 gene), along with theinXuence of adenovirally mediated p53 overexpression inmodulation of cisplatin and methoterexate eVects, wherebysubtoxic vector/compound concentrations were employed.Methods p53 gene was introduced into tumor cells usingadenoviral vector (AdCMV-p53). The cell cycle perturba-tions were measured by two parameter Xow cytometry. Theexpression of p53, p21WAF1/CIP1 and cyclin B1 proteins wasexamined using immunocytochemistry and western blotmethods.Results In CAL 27 cells overexpression of p53 com-pletely abrogated high S phase content observed in metho-trexate-treated cells into a G1 and slight G2 arrest, while it

sustained G2 arrest of the cells treated with cisplatin, alongwith the reduction of DNA synthesis and cyclin B1 expres-sion. On the other hand, in HEp-2 cell line p53 overexpres-sion slightly slowed down the progression through S phasein cells treated with methotrexate, decreased the cyclin B1expression only after 24 h, and failed to sustain the G2arrest after treatment with cisplatin alone. Instead, itincreased the population of S phase cells that were notactively synthesizing DNA, sustained cyclin B1 expressionand allowed the G2 cells to progress through mitosis.Conclusions This study demonstrates that adenovirallymediated p53 overexpression at sub-cytotoxic levelsenhanced the activity of low doses of cisplatin and metho-trexate in HEp-2 and CAL 27 cells through changes in thecell cycle. However, the mechanisms of these eVects diVerdepending on the genetic context and on the chemothera-peutics’ modality of action.

Keywords p53 · p21WAF1/CIP1 · Cell cycle · Cyclin B1 · HEp-2 cells · CAL 27 cells

Introduction

Since its discovery in 1979, the p53 gene was postulated tobe an important tumor suppressor whose inactivation isconsidered to be a critical step in tumorigenesis. Indeed,numerous publications have proved its fundamental role inthe modulation of the transformed state. The p53 proteinacts primarily as a multitarget transcription factor, whichmeans that it controls the expression of a wide range ofgenes with disparate functions. Additional cancer-relatedfunctions continue to be discovered, but thus far, its knownfunctions include cell cycle regulation, senescence, apopto-sis, repair of DNA damage caused by genotoxic agents,

S. KraljeviT PaveliT and M. MarjanoviT contributed equally to the preparation of this manuscript.

Electronic supplementary material The online version of this article (doi:10.1007/s00432-009-0621-5) contains supplementary material, which is available to authorized users.

S. KraljeviT PaveliT · M. MarjanoviT · M. PozniT · M. Kralj (&)Division of Molecular Medicine, Rudjer BonkoviT Institute, Bijenibka cesta 54, 10000 Zagreb, Croatiae-mail: [email protected]

123

1748 J Cancer Res Clin Oncol (2009) 135:1747–1761

angiogenesis, and regulation of oxidative stress (Foulkes2007).

The p53 gene functions as a “master regulator” of theapoptotic program, whereby it is capable of coordinatingthe process at multiple levels via diverse mechanisms. It isactivated in the cell response to stress, and plays a pivotalrole in the cellular response to a range of environmental andintracellular stresses including agents which cause DNAstrand breaks, ultraviolet radiation, hyper-proliferation andhypoxia. Since majority of antitumor drugs act as DNA-damaging agents, this gene plays a crucial role in theresponse to therapy. p53 acts as a node or hub for incomingstress signals transduced mostly in a transcription-depen-dent manner, but even transcriptionally independent pro-cesses have been documented as well (Meek 2004).Although the pro-apoptotic activities of p53 are well docu-mented and highly appreciated as one important aspect ofits multiple tumor suppressor functions, over the past fewyears relatively unexpected and unexplored side to thisstory emerged implicating p53 to be an active mediator ofpro-survival signaling pathways (Jänicke et al. 2008;Vousden 2006). Consequently, regardless of a plethora ofexperimental data published so far (more than 43,000 publi-cations on p53) we are still far from understanding themechanisms governing its versatile functions.

The p53 gene is one of the most commonly mutatedgenes in human cancer (Hollstein et al. 1996). Mutations ofp53 aVect its ability to suppress tumorigenesis mainlythrough the loss of wild-type p53 function, but also throughdominant negative eVect of mutant over wild-type p53 orgain of additional oncogenic properties. The presence ofp53 mutation has been associated with unfavorable progno-sis in a variety of tumor types, disease progression andoften with enhanced resistance to many anti-tumor agents(Wallace-Brodeur and Lowe 1999; Gallagher and Brown1999). Since the functions of p53 target genes are diverse,this may be the foundation by which p53 acts as a multi-functional protein and may begin to explain why loss ofp53 activity results in the development of human tumorsthat are associated with a range of phenotypes such as alack of diVerentiation, increased levels of angiogenesis andmetastasis (Slee et al. 2004). Consequently, complex p53biology has been studied to a great extent in order to recog-nize its exact role in diagnostic or prognostic beneWt and/orin improving cancer therapy. Recently, Ventura et al. pre-pared mice not expressing the p53 protein therefore sponta-neously developing tumors, by using a well-designedgenetic manipulation method. The p53 expression in thismice model has been restored upon tumors reached detect-able sizes. The restoration led to regression of autochtho-nous lymphomas and sarcomas in mice without aVectingnormal tissues. Additionally, it was shown that the mecha-nism responsible for tumor regression is dependent on the

tumor type, thus being apoptosis in lymphomas and cellularsenescence in sarcomas (Ventura et al. 2007). Taken alltogether, p53 tumor suppressor is an attractive targetfor pharmacological intervention (Blagosklonny 2002).Depending on the p53 status of cancer cells, diverse thera-peutic strategies could be exploited. Still, the response-mechanism depends equally on the p53 status, on thecellular genetic context which should be elucidated in indi-vidual tumors, as well as on type, strength and duration of astimulus known to be critical determinants of the cell fate.Oncologists might greatly beneWt from knowledge on p53role in cancer. First, the cancer treatment could be Wne-tuned according to the p53 status—e.g. patients bearing atumor with mutated p53 would receive one type of treat-ment, whereas patients without such a mutation wouldreceive a diVerent treatment. Second, a range of small mol-ecules could be speciWcally designed to directly target p53(Foulkes 2007). A third way is gene therapy employing thep53 gene as a drug. The latter is the only gene therapyapproach currently in various development stages (Kastan2007; Roth 2006) and approved by the State Food and DrugAdministration of China (SFDA) on 16 October 2003 forthe treatment of head and neck squamous cell carcinoma inChina (Peng 2005).

In our previous papers we showed that the both adenovi-rally mediated p53 and p21 overexpression inhibited thegrowth of several cell lines (Kralj et al. 2003; Kralj andPavelic 2003; Kraljevic Pavelic et al. 2008). However, p53was more potent in growth inhibition and apoptosis induc-tion of human tumor cells with non-functional p53. It wassupposed that this induction correlated with the basal p21expression level (Kralj et al. 2003). On the other hand, p21induced apoptosis in cells with non-functional p53 as well(HeLa, SW 620, HEp-2). Interestingly, we additionallydemonstrate that p21 can assume a dual role in apoptosis inthe same cell system and thus provided new grounds foroptimizing therapy in the treatment of cancer cells lackingp53 expression. In this paper, we analyzed the inXuence oftwo mechanistically diverse antitumor agents—cisplatinand methotrexate on the proliferation, cell cycle and celldeath of two head and neck cancer cell lines HEp-2 (wildtype p53 gene, but HPV 18/E6-inactivated protein) andCAL 27 (mutated p53 gene) and studied the inXuence ofp53 overexpression in modulation of cisplatin and methote-rexate eVects.

Materials and methods

Cell lines

Tumor cell lines HEp-2 (larynx carcinoma; HPV 18 inacti-vated p53 protein; ATCC number CCL-23), CAL 27

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1749

(tongue squamous carcinoma; mutated p53 gene; ATCCnumber CRL-2095) and Detroit 562 (pharynx carcinoma;mutated p53 gene; ATCC number CCL-138) were culturedas monolayers and maintained in the 10%-DMEM(Dulbecco’s modiWed Eagle’s medium—DMEM supple-mented with 10% fetal calf serum (FCS) (Gibco, USA),2 mM L-glutamine, 100 U/mL penicillin and 100 �g/mLstreptomycin) in a humidiWed atmosphere with 5% CO2 at37°C. HEK-293 cells (ATCC number CRL-1573) weremaintained in high glucose DMEM (HG-DMEM).

Adenoviruses and infection

Replication defective adenoviral recombinant Ad5CMV-p53 (Introgen Therapeutics, Inc., USA) was used for theintroduction of the p53 gene into cells. The vectors containthe CMV promotor, p53 cDNA, SV 40 polyadenylationsignal in a minigene cassette inserted into the E1-deletedregion of modiWed Ad5. As a control vector dl 312 wasused. Viral vectors were propagated and titrated inHEK-293 cells. Cells were harvested 36–40 h after infec-tion, pelleted, re-suspended in phosphate-buVered salineand lysed; cell debris was removed by subjecting the cellsto CsCl gradient puriWcation. Concentrated virus was dia-lyzed, aliquoted and stored at ¡80°C.

Immunocytochemistry

The cells were seeded on eight-well glass chamber slides(Nunc, USA) in 10%-DMEM at 1 £ 104 cells/well forHEp-2, 1.5 £ 104 cells/well for CAL-27 and 3 £ 104 cells/well for Detroit 562. Twenty-four or 72 h after the infectionwith AdCMV-p53 or Ad-dl 312 the cells were washed inthe phosphate buVer saline (PBS) and Wxed in methanolwith 1.5% hydrogen peroxide (H2O2) (Kemika, Croatia).The further procedure was assessed as described previously(Kraljevic Pavelic et al. 2008), using primary monoclonalantibodies against p53 (Calbiochem), at concentration2.5 �g/ml. The p53-positive cells (brownish colored) werecounted and the percentage was expressed as a number ofpositive infected cells compared to mock-infected cells. Atleast 100 cells were counted.

Antiproliferative assays

Antiproliferative eVect of cisplatin, methotrexate (Sigma)and AdCMV-p53 were tested. The cells were inoculatedonto standard 96-well microtitre plates on day 0 at concen-trations 1 £ 103 cells/well for HEp-2, 1.5 £ 103 cells/wellfor CAL-27 and 3 £ 103 cells/well for Detroit 562. Testagents in Wve 10-fold dilutions (10¡8–10¡4 mol/L), or viraldilutions [10–50 MOI—multiplicity of infection (MOI)]

were then added and incubated for a further 24, 48 and72 h. Working dilutions were freshly prepared on the day oftesting. After incubation, the cell growth rate was evaluatedby performing the MTT-assay, as described previously(Kraljevic Pavelic et al. 2008; Supek et al. 2008). Each testpoint was performed in quadruplicate in three individualexperiments. Each result is a mean value from three sepa-rate experiments.

Similarly, for determination and comparison of the ade-novectors eVect alone or in combination with cisplatin ormethotrexate, the modiWed MTT-test as described belowwas used. Viral dilutions at 10 MOI were prepared and thecells were either infected with adenoviruses, or mock-infected (10% DMEM). After 1.5 h of infection cisplatin(1 £ 10¡8 M–1 £ 10¡4 M and methotrexate dilutions(5 £ 10¡9–1 £ 10¡4 M) or 10% DMEM were then addedand incubated for 24, 48 and 72 h. After incubation, thecell growth rate was evaluated by performing the MTTassay as described above. The percentages of growth (PG)were calculated as ratios of mean absorbance values foreach treatment with its respective control multiplied with100.

Two-parameter Xow cytometry analysis: DNA content and BrdU incorporation

A total of 3 £ 105 cells/well were seeded in six-wellplates. The cells were infected with AdCMV-p53 at 10MOI and/or were treated with cisplatin and methotrexateat concentrations 0.5 �g/ml and 1 £ 10¡7 mol/L, respec-tively. After the desired length of time actively growingcells were pulsed for 45 min in the dark with 10 �M BrdUsolution in complete DMEM (Sigma). After incubationcells were detached from the culture plates, washed twicewith PBS and Wxed in 70% ethanol for 30 min at RT. Cellswere then treated with 2 N HCl with addition of 0.1% Tri-ton-X for 20 min. After washing with 0.1 M Na2B4O7 pH8.5 (Fluka), cells were resuspended in PBS with 0.5%Tween 20 (Sigma) and 0.5% BSA (Santa Cruz Biotechnol-ogy), counted and the number of 150,000 cells per 50 �Lwas adjusted for each test point. 10 �L of monoclonal anti-BrdU antibody (Becton Dickinson) was added to the eachreaction tube and incubated for 30 min in the dark at RT.After washing with PBS, cells were incubated for 45 minwith 10 �g/mL FITC labelled goat anti-mouse IgG anti-body (BD Pharmingen) in PBS with 0.5% Tween 20 and0.5% BSA. Cells were then washed and incubated for30 min at room temperature with a solution containing50 �g/mL Propidium-Iodide (Sigma) and 0.1 �g/�LRNase A (Sigma) and transferred to FACS tubes (BectonDickinson). The stained cells were then measured withBecton Dickinson FACSCalibur Xow cytometer (20,000counts were measured). Each test point was performed in

123

1750 J Cancer Res Clin Oncol (2009) 135:1747–1761

three individual experiments. The percentage of the cellsin each cell cycle phase was based on the obtained twoparameter plots (DNA content vs FITC-anti BrdU Xuores-cence) and determined using the FlowJo software (TreeStar Inc.).

Fluorescent microscopy of DAPI stained cells

HEp-2 and CAL 27 cells were seeded at the concentra-tion of 3 £ 105 cells/well in six-well plates, in whichsquare microscopy cover glass slides (20 mm) had beenpreviously placed. The cells were treated exactly as inthe cell cycle experiment to maintain the same condi-tions. After the desired length of time, cells were washedwith PBS and Wxed in 4% formaldehyde in PBS(Kemika, Croatia) for 15 min at RT. After Wxation cellswere permeabilized with 0.1% Triton-X in PBS (Sigma)for 10 min. After that, glass slides were washed4 £ 5 min with PBS and once in water, dried and posi-tioned onto microscopy glass slides using the Xuorescentmounting medium (Dako) containing 100 ng/mL DAPI(Sigma). Slides were kept in a dark humid container at4°C until they were analysed on the Olympus epiXuores-cent microscope.

Western blot analysis

Treated and untreated cells were lysed with the RIPA buVer(Tris–HCl (pH = 7.5) 10 mM, 1% Nonidet P40, 150 mMNaCl, 0.1% SDS and protease inhibitor cocktail (Roche).40 �g of total proteins were resolved on 9 or 12% SDS-polyacrylamide gel depending on the studied protein atconstant 100 V and subsequently transferred to nitrocellu-lose membrane (BIO-RAD) at constant 200 mA usingMini-PROTEAN Cell (BIO-RAD). Membranes wereblocked for 1 h at room temperature with 4% non-fat drymilk in TBST (50 mM Tris base, 150 mM NaCl, 0.1%Tween 20, pH 7.5). Subsequently, membranes were incu-bated overnight at 4°C in 3% non-fat dry milk in TBSTsupplemented with primary antibodies against p53 (Calbio-chem, 1.4 �g/ml); p21 (Santa Cruz Biotechnology, 2.5 �g/ml),cyclin B1 (Santa Cruz Biotechnology; 1.2 �g/ml). Themembranes were then washed with TBST and incubated for1 h at room temperature in TBST containing a secondaryantibody linked to horseradish peroxidase. The signal wasdetected by the Western Lightening ChemiluminiscenceReagent Plus kit (PerkinElmer) and visualized on the Ver-saDoc Imaging System 4000 (BIO-RAD). Signal intensitiesof the particular bands were measured and analysed by theQuantity One software (BIO-RAD). Anti-�-tubulin (Sigma,monoclonal anti-�-tubulin mouse IgG, diluted 1:1,000) wasused as a loading control.

Results

Antiproliferative eVect of methotrexate, cisplatin and AdCMV-p53 on head and neck squamous cell carcinoma cell lines

The gene transfer eYciency was determined by immunocy-tochemical detection of p53 protein expression in cells. Theendogenous expression of the mutated p53 protein wasdetected in CAL 27 cells (62% of total CAL 27 cell num-ber) and Detroit 562 cells (50% of total cell number) (datanot shown). The p53 protein expression levels increased inall cell lines upon AdCMV-p53 treatment, depending onthe MOI. In HEp-2 cells, no endogenous p53 proteinexpression was detected, so increased p53 expression isattributed solely to the adenovirally mediated wt-p53 genetransduction (Table 1).

The responses of HEp-2, CAL 27 and Detroit 562 cellsto the individual therapeutic agents [cisplatin (cis), metot-rexate (MTX) and AdCMV-p53] or the combination of thechemotherapeutics and adenovirally mediated wt-p53 geneexpression were analyzed by determining the relative via-bility of the treated cells (Figs. 1–4).

AdCMV-p53 induced a time- and a dose-dependent anti-proliferative eVect on HEp-2 and CAL 27 cells, whileDetroit 562 cells were almost completely resistant to thetreatment (Fig. 1). After 3 days of treatment AdCMV-p53caused a marked growth inhibitory eVect on HEp-2 at 40and 50 MOI (30–75%) and CAL-27 at all vector concentra-tions (70–100%), while after 6 days its eVect was cytotoxicto both HEp-2 and CAL-27 cells whereby the cell growthwas completely inhibited. However, at low MOI ofAdCMV-p53 signiWcant variations in the percentages ofviable HEp-2 cells were observed. Detroit 562 cells werealmost completely resistant to the AdCMV-p53 treatmentthat exerted a growth inhibitory eVect only after the 6 daystreatment at all vector concentrations.

Morphological changes pointing to apoptosis (chromatincondensation) were observed both in HEp-2 (Fig. 2)and CAL 27 (data not shown) cells after the 24 h- and72 h- treatments with AdCMV-p53 at higher MOIs.

Table 1 Percentage of HEp-2 cells expressing wt-p53 after 24- and72 h-treatments with Ad-p53

Treatment (HEp-2) Percentage

24 h 72 h

Control 0 0

Ad-p53 10MOI 47 84

Ad-p53 20MOI 65 90

Ad-p53 30MOI 72 100

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1751

As discussed previously (Kraljevic Pavelic et al. 2008),cisplatin exerted a strong and similar antiproliferative eVecton all tested cell lines (IC50 values for the 72-h treatmentperiod were about 3 £ 10¡6 M, which corresponds to0.5 �g/mL), while methotrexate diVerentially inhibited thegrowth of all tested cell lines, being somewhat less activeon Detroit 562 and slightly cytotoxic to HEp-2 cells(Fig. 3).

In all further experiments we used relatively low concen-trations of chemotherapeutics (corresponding to approxi-mate IC50 values for the 72-h treatment period of HEp-2cells) as well as a viral particle number that was suYcientto induce the expression of p53 protein in cca 50% of HEp-2cells. Such experimental procedure enabled a better correla-tion of the obtained results to those acquired previously inour study dealing with the inXuence of p21 gene overex-pression (Kraljevic Pavelic et al. 2008). Besides, we werethus able to check the eYcacy of the combined treatment ofcells using low doses of both chemotherapeutics and p53-bearing vectors.

The results of treatments with AdCMV-p53, chemother-apeutics and combination treatments are shown in Fig. 4. Amore pronounced eVect on the growth of all tested cell lineswas detected for the combined treatment with cisplatin(0.25–5 �g/mL) and AdCMV-p53 at the lower vector con-centration (MOI 20) in comparison to cells treated with cis-platin alone (25–60% for HEp-2, 24–64% for CAL 27 and22–44% for Detroit 562).

The antiproliferative eVect of methotrexate (1 £ 10¡4–5 £ 10¡9 M) was similarly, but less eYciently enhanced byAdCMV-p53 on all three cell lines (12–30%). These resultsconWrm that the overexpression of p53 gene sensitize cellswith diverse genetic backgrounds and non-functional p53to mechanistically diVerent chemotherapeutics. Contrary top53, the p21 gene overexpression induced opposite eVectsin HEp-2 cells, enhancing cell death induced by cisplatinand attenuating the methotrexate activity (Kraljevic Pavelicet al. 2008).

The inXuence of AdCMV-p53, cisplatin, methotrexate and their combinations on the cell cycle of HEp-2 and CAL 27 cells

In order to establish the mechanisms underlying previ-ously observed eVects (enhancement of antiproliferativeeVects of cisplatin or methotrexate), we monitored cellcycle perturbations upon diVerent treatments using Xowcytometry. For that purpose we analyzed HEp-2 (Figs. 5a,b, 6a, b) and CAL 27 (Figs. 5c, d, 6c, d) cells treated with:cisplatin at 0.5 �g/ml, methotrexate at 1 £ 10¡7 M andcells treated with the combination of cytostatic com-pounds and AdCMV-p53 at 10 MOI. We did not performadditional experiments with the Detroit 562 cells sincethese cells were almost completely resistant to theAdCMV-p53 treatment and less sensitive to chemothera-peutics treatment. We used vector concentrations that

Fig. 1 The eVect of AdCMV-p53 on the growth of HEp-2, CAL 27and Detroit 562 cell lines after 3 (a) and 6 days (b) of treatments. Theresults were obtained by MTT-assay and are shown as percentages of

growth (PG) for each treatment point § standard deviation. Ad-dl 312(empty vector) was used as a control. The statistically signiWcant diVer-ences are marked with the sign asterisk (P > 0.05)

123

1752 J Cancer Res Clin Oncol (2009) 135:1747–1761

infected about 50% of cells (positive to immunocyto-chemical detection of p53). Treatment of HEp-2 cells withAdCMV-p53 at 10 MOI caused a slight G1 arrest at bothtime points compared to control. Interestingly, whenHEp-2 cells were infected with higher vector number(MOI 20) a substantial increase in the fraction of apopto-tic cells after the 24 h-treatment could be detected (datanot shown) and raised considerably during the next 48 hof treatment (up to 60% of apoptotic cells), pointing to adose-dependent eVect of p53 overexpression—lower

vector concentrations induce solely a cell cycle arrestwhile higher vector concentrations strongly induce apop-tosis. Therefore, the enhanced cytotoxic eVect of chemo-therapeutics in combination with p53-bearing adenoviralvectors at MOI 20 and higher could be ascribed to pro-nounced apoptosis. However, we focused our research tothe diVerences in the cell cycle/death changes between therespective treatment regimes when no cytotoxicity wasobserved.

As expected, cisplatin markedly increased the number ofcells in the S phase of HEp-2 cells after 24 h, showing thereduced rate of S phase progression. These cells accumu-lated in G2 phase during the next 24 h. Methotrexate treat-ment also slowed down the S phase progression (themajority of cells were in the early S phase, as documentedfrom the BrdU incorporation plots Fig. 6a) throughout thewhole experimental period (Figs. 5, 6a, b). Similar cellcycle changes were observed in CAL 27 cells, except thatAdCMV-p53 induced strong G1 and G2 arrest throughoutthe whole experimental period, with the reduction of Sphase fraction (Fig. 5c, d).

A combined treatment with AdCMV-p53 and cisplatininduced an increase of HEp-2 cells in the S in compari-son to cells treated with cisplatin alone and a concomi-tant decrease of cells in the G1 phase after 48 h. TheBrdU incorporation experiments showed that in the com-bined treatment there was an increase of cells in S phasethat did not incorporate BrdU (16%), i.e. were notactively synthesizing DNA (Fig. 6b). Also should benoted that the proportion of non BrdU incorporating cellswas of DNA content close to G1 (2n) that was not thecase with the cisplatin treatment alone where most of thecells seemed to accumulate in the early G2 phase(Fig. 6b). In the combined treatment the cells shiftedfrom the G1 phase but were not actively synthesizingDNA (probably aimed to die), while the rest of delayedG2 cells progressed to mitosis (as documented by themicroscopy data, Fig. 10).

On the other hand, in CAL 27 cells overexpression ofp53 gene induces an additional G1 arrest compared to cis-platin treatment alone, accompanied with the pronouncedreduction of S phase cells compared to both control and cis-platin-treated cells (Figs. 5, 6c, d).

A combined treatment with AdCMV-p53 and metho-trexate showed similar eVect on the CAL 27 cell cycle asthe treatment of the p53 overexpression alone (a strongreduction of S phase, compared to methotrexate treatmentalone, while in HEp-2 cells AdCMV-p53 slightly enhancesthe G1 phase after 24 h (Fig. 6a). On the other hand, after48 h the cell cycle was not signiWcantly inXuenced byAdCMV-p53 and was similar to the methotrexate treat-ment, whereby the majority of cells were in the earlyS phase (Fig. 6b).

Fig. 2 p53 expression in HEp-2 cells 24 (a) and 72 h (b) after infec-tion with Ad-CMVp53 at 10, 20, 30 MOI, along with control, unin-fected cells. Blue arrows point to apoptotic cells

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1753

p53, p21, cyclin B1 protein expression changes in HEp-2 and CAL 27 cells treated with cisplatin, methotrexate, AdCMV-p53 and their combinations

Further on, we examined the changes in p21 and p53 pro-tein levels in HEp-2 and CAL 27 cells treated withAdCMV-p53, cisplatin, methotrexate and their respectivecombinations. The p53 protein expression in HEp-2 cellswas detectable only in treatments with AdCMV-p53(Fig. 7; Supplementary Figures 1, 2), whereby initially thecombined treatments with both chemotherapeutics wereless marked in comparison to the treatment with AdCMV-p53 alone. However, it gradually rose to equal expression(48 h), while after 72 h in the combined MTX/AdCMV-p53treatment it had several-fold higher level in comparison to

the treatment with AdCMV-p53 alone, or in the combina-tion with cisplatin (Supplementary Figure 1). On the otherhand, the mutated p53 protein expression was apparent incontrol and treated CAL 27 cells. Similarly to HEp-2 cells,neither cisplatin, nor methotrexate induced the p53 expres-sion, while the adenovirus-mediated expression was readilydetected with no diVerences between the treatment combi-nations (Fig. 7b).

The p21 protein expression was observed in treatmentsof HEp-2 cells with AdCMV-p53 and in the combinedtreatments with AdCMV-p53 and chemotherapeutics onlyafter 24 h, whereby it was signiWcantly lower in the com-bined treatment with metotrexate (2-fold) and cisplatin(1.6-fold) than in the AdCMV-p53 treatment alone, corre-lating with the equally lower exogenous p53 expression

Fig. 3 Dose response curves obtained by MTT-test after the 3-daystreatment of HEp-2, CAL 27 and Detroit 562 cells with cisplatin andmethotrexate at concentrations 1 £ 10¡4 M–1 £ 10¡8 M. The results

are shown as percentages of viable cells [PG percentage of growth(%)] for each tested drug concentration § standard deviations

Fig. 4 The eVects of treatments with AdCMV-p53 at 20 MOI, cis-platin (0.25–5 �g/mL) (a), methotrexate (1 £ 10¡4–5 £ 10¡9 M) (b)and their combinations on the growth of HEp-2, CAL 27 and Detroit

562 cell lines after the 3 days-treatment. The results were obtained byMTT-assay and are shown as percentages of growth (%) for each treat-ment point § standard deviation

123

1754 J Cancer Res Clin Oncol (2009) 135:1747–1761

(Fig. 8a). After 48 h, the p21 protein expression wasdetected only in cells treated with AdCMV-p53, althoughp53 was still highly expressed in the combined treatments.On the contrary, p21 protein expression was strongly andequally expressed in CAL 27 cells in all AdCMV-p53-treatments after 24 h (Fig. 8b), while it gradually dimin-ished during further 24 h.

Since we detected a rather marked G2 phase delay aftertreatment with both cisplatin and in some cases withAdCMV-p53 as well, and p53 gene is involved in G1, aswell as G2 phase arrest, whereby it regulates a G2 check-point through cyclin B1, we monitored changes in cyclinB1 expression. Methotrexate induced non-signiWcant (24 h)or only slight changes (48 h) in cyclin B1 expression ineach cell line; actually the only signiWcant upregulation ofcyclin B1 protein expression was documented in HEp-2treated with MTX after 48 h (2-fold increase), while after24 h there was no signiWcant cyclin B1 increase (1.01-fold).Similarly, in CAL 27 cells the expression of cyclin B1 wasraised by 1.2-fold (24 h) and 1.5-fold (48 h) compared tocontrol, respectively (Fig. 9a, b; Supplementary Figure 2).The combination treatment with MTX/AdCMV-p53decreased the cyclin B1 protein level in HEp-2 cells

(Fig. 9a, b; Supplementary Figure 2) 2-fold comparing tocontrol cells, MTX and AdCMV-p53 treated cells at 24 h.However, after 48 h in MTX/AdCMV-p53 treated cells thecyclin B1 level was comparable to both the control andMTX-only treated cells, while it is fourfold lower com-pared to AdCMV-p53-only treated cells. In CAL 27 cells,AdCMV-p53 downregulates cyclin B1 expression strongly(2-fold compared to control cells, and 2.6-fold compared toMTX-treated cells). In contrast, cisplatin induced a 2-foldincrease in cyclin B1 level in both HEp-2 and CAL 27 after24 h, whose expression levels sustained for next 24 h beingaccentuated in HEp-2 cells in comparison to control cells.Interestingly, unlike in CAL 27 cells, overexpression ofp53 gene did not diminish levels of cyclin B1 in HEp-2cells, either alone or in combination with cisplatin, espe-cially after 48 h, which should be correlated to the pro-nounced S/G2 accumulation (combined treatment ofAdCMV-p53 and cisplatin) and probably normal transitionof cells through G2 and M in both treatments (being inaccordance with the mitotic cells detected by the micros-copy) (Figs. 5, 6, 10). p53 overexpression in CAL 27 cellsstrongly decreased cyclin B1 levels at both time pointswhich is a direct consequence of p53 mediated G1 arrest

Fig. 5 The eVects of AdCMV-p53 at 10 MOI, cisplatin at 0.5 �g/ml,methotrexate at 1 £ 10¡7 M and their combinations on the cell cycledistribution of HEp-2 and CAL 27 cells after the 24- and 48 h-treat-

ments. The graphs represent the number of cells in respective cell cyclephase (G1, S and G2/M) obtained by Xow cytometry. One representa-tive experiment is shown

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1755

and a decrease of S phase-cells that eventually are aimed tomitosis.

Fluorescence microscopy of DAPI stained nuclei of HEp-2 and CAL 27 cells treated with cisplatin, methotrexate, AdCMV-p53 and their combinations

We performed microscopy experiments using DAPI underthe same experimental conditions correlating with the pre-vious results, we could clearly detect mitotic HEp-2 cellstreated with AdCMV-p53, cisplatin, or their combination(Fig. 10), while no or only scarce mitotic cells could bedetected in the cells treated with methotrexate, or its combi-nation with AdCMV-p53. On the other hand, in CAL 27

cells substantial number of mitotic cells could be detectedonly in control cell population (Fig. 10). Besides, a fewcells with condensed and/or fragmented nuclei were readilydistinguished in both cell lines upon the combined treat-ments with AdCMV-p53 and both citostatics, but mostly inthe AdCMV-p53/cisplatin-treated HEp-2 cells. Such anuclear appearance should point to the programmed celldeath mechanisms activated in these cells.

Discussion

Anticancer agents that target DNA are among the mosteVective in clinical use and have proved to signiWcantly

Fig. 6 Two parameter Xow cytometry histograms showing DNAcontent (horizontal axis) and the level of BrdU incorporation (verticalaxis) for each treatment of HEp-2 (24 and 48 h after treatment, a, b,respectively) and CAL 27 cells (24 and 48 h after treatment, c, d,

respectively). Density of the cells in the histogram is depicted bydiVerent color with blue indicating low cell density; green and yellow/red indicate medium and high cell density, respectively. One represen-tative experiment is shown

123

1756 J Cancer Res Clin Oncol (2009) 135:1747–1761

increase the survival of patients with cancer when used incombination with drugs that have diVerent mechanisms ofactions. Despite their eVectiveness and widespread use asanticancer drugs, some key limitations hamper their use inclinical practice. Indeed, many patients with diagnosed can-cer do not respond to such therapeutical approach or evenmore discouraging, do develop drug resistance. The molec-ular mechanisms underlying drug resistance to DNA-dam-aging agents are believed to involve modulation of cellularlevels of drug and drug target, defects in apoptosis, or proW-cient DNA repair (Zhou and Bartek 2004). Clearly, diverseoutcomes exist for defects occurring in the above men-tioned cellular processes might occur. Therefore, the suc-cess of targeted therapy will depend to a large extent on themolecular Wngerprinting of individual tumors (Evan andVousden 2001). For all these reasons, cellular responses tovarious DNA-damaging agents, as well as the signal-trans-duction pathways that mediate these responses have beenextensively studied.

Although current anticancer agents do not directly inter-act with p53, the mechanism of action for many of these

agents involves accumulation of wt p53. Diverse multipleand overlapping mechanisms include DNA intercalation,inhibition of topoisomerase I and II and DNA crosslinking,whereby DNA strand breaks appear to be the most impor-tant lesion triggering elevation of p53 protein levels(Blagosklonny 2002). Since the p53 gene is inactivated inthe majority of human cancers, it is strongly believed thatthe p53 mutations confer resistance to therapeutics. Indeed,many studies have examined the role of p53 in therapeuticresponses. Still, while the majority of the data support arole for mutated p53 in reducing chemosensitvity, the otherindicate either no correlation or even increased sensitivityto chemotherapeutic agents (Bunz et al. 1999). Based onvarious results, it might be deduced that although mutatedp53 gene cause diVerent response patterns to therapeuticagents depending on the type of drug and genetic back-ground, p53 mutant cells are still generally resistant to mostDNA-damaging drugs since p53 aVects the access of drugsto intracellular targets, drug target interactions, and thedownstream response to cellular injury. Thus an overallhypothesis involving the status of p53 in cancer chemotherapy

Fig. 7 The p53 protein expres-sion in HEp-2 (a) and CAL 27 (b) cells 24 and 48 h after cell treatment. cont Untreated cells, MTX methotrexate at 1 £ 10¡7 M, cis cisplatin at 0.5 �g/ml, p53 AdCMV-p53 at 10 MOI; as well as their respec-tive combinations. �-tubulin expression was used as the loading control

Fig. 8 The p21 protein expres-sion in HEp-2 (a) and CAL 27 (b) cells 24 and 48 h after cell treatment. cont Untreated cells, MTX methotrexate at 1 £ 10¡7 M, cis cisplatin at 0.5 �g/ml, p53 AdCMV-p53 at 10 MOI; as well as their respec-tive combinations. �-tubulin expression was used as the loading control

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1757

response prediction seems plausible. Indeed, numerousrecent studies have been performed to establish clinical evi-dence that the p53 genotype serve as a predictive markerfor response to therapy as well as to deWne proWles ofpatients who actually beneWt from a certain type of therapy(Hait and Yang 2006; Kandioler et al. 2008; Kupryjanczyket al. 2008). These studies demonstrate the usefulness ofindividualized treatments which are based on the p53 statusor employ the gene replacement therapy of the mutated p53gene with its wild form. This gene therapy approach is anattractive strategy to cancer treatment in such tumor cellsthat do not respond well to conventional therapy. The mostconventional approach to the p53 gene replacement therapyis based on use of adenoviral vectors for the introduction ofwt p53 gene into cells (Horowitz 1999). Indeed, numerousstudies have conWrmed that reintroduction of the wildtype p53 suppresses tumor cell growth, induces apoptosis

(Nielsen and Maneval 1998; Kralj et al. 2003) and/orincreases sensitivity to conventional anti-tumor agents irre-spective of the p53 status (Horio et al. 2000; Inoue et al. 2000;Ganjavi et al. 2006).

In this study we demonstrated an increased sensitivity ofsquamous head and neck cancer cell lines HEp-2 and CAL27 bearing non-functional p53 protein to two, mechanisti-cally diverse, DNA-active drugs cisplatin and methotrex-ate. Cisplatin directly binds to DNA resulting in formationof intra- and inter-strand adducts as well as in cross-linkingDNA to proteins. This alkylating ability of cisplatin is pri-marily responsible for its cytotoxicity and anticancer activ-ity, whereby cisplatin-DNA adducts structurally block themovement of DNA replication and transcription machineryalong DNA, which results in inhibition of DNA replicationand transcription (Sun et al. 2002; Kraljevic Pavelic et al.2008). On the other hand, methotrexate mechanism of

Fig. 9 The cyclin B1 protein expression in HEp-2 (a) and CAL 27 (b) cells 24 and 48 h after cell treatment. cont Un-treated cells, MTX methotrexate at 1 £ 10¡7 M, cis cisplatin at 0.5 �g/ml, p53 AdCMV-p53 at 10 MOI; as well as their respec-tive combinations. �-tubulin expression was used as the loading control

Fig. 10 Fluorescent microscopy of DAPI stained nuclei of HEp-2 andCAL 27 cells 48 h after treatment. cont Untreated cells, MTX metho-trexate at 1 £ 10¡7 M, cis cisplatin at 0.5 �g/ml, p53 AdCMV-p53 at

10 MOI; as well as their respective combinations. White arrows pointto cells in the process of mitosis, while green arrows show apoptoticcells

123

1758 J Cancer Res Clin Oncol (2009) 135:1747–1761

action is linked to its antimetabolic activity and block of theDNA synthesis. Methotrexate inhibits various enzymaticactivities associated with both purine and pyrimidine syn-thesis including thymidylate synthetase and dihydrofolatereductase (Hattangadi et al. 2004).

Although the increased sensitivity to cisplatin by adeno-virus-mediated p53 overexpression has already beenreported (Inoue et al. 2000; Ganjavi et al. 2006; Weinribet al. 2001), the involvement of p53 in cisplatin-mediatedapoptosis, as well as cisplatin-induced cell death is highlydose-dependant and still remains controversial (Yip et al.2006). The combination therapy might be clinically inter-esting if low doses of chemotherapeutics are used whichultimately leads to good therapeutic outcomes along withlow toxicity. Therefore, our study was intended to revealwhether and how low doses of either cisplatin (0.5 �g/ml)or methotrexate (1 £ 10¡7 M) and p53 overexpression(about 50% of p53-overexpressing cells) inhibit the growthof tumor cell lines. We previously reported that p21WAF1/CIP1

overexpression signiWcantly increased the sensitivity to lowdoses of cisplatin, while it attenuated the methotrexate-mediated apoptosis in the HEp-2 tumor cell line (KraljevicPavelic et al. 2008). This data was in accordance with theversatile roles of p21 protein. For example, p21 mainlybinds to cyclin/CDK2 complexes and negatively modulatesprogression through G1 checkpoint of the cell cycle(Brugarolas et al. 1999), especially in the p53-dependentmanner in a response to DNA damage after �-irradiation orexposure to DNA damaging agent such as doxorubicin(el-Deiry et al. 1993; Waldman et al. 1995). On the otherhand, under special conditions p21 can stabilize and pro-mote formation of cyclin/CDK complexes and its nuclearlocalization, with two contrast functions of the proteinoccurring in a concentration dependent manner (LaBaeret al. 1997). Additionally it can block cell cycle in S and G2phase prior to cytokinesis (Bunz et al. 1998; Radhakrishnanet al. 2004), block or promote apoptosis (Gartel and Tyner2002) and have a negative eVect on autophagy (Fujiwaraet al. 2008).

Unlike the p21 gene, p53-overexpression sensitizestumor cells to both chemotherapeutics, although the eVectis more accentuated to cisplatin. Neither cisplatin, normethotrexate induced p53 or p21 expression in HEp-2 orCAL 27 compared to non treated cells, which was partiallyexpected because these cell lines bear non-functional p53protein, which is degraded by the HPV 18-E6 protein inHEp-2 and mutated in CAL 27. Contrary to our observa-tions, Yip and coworkers observed a marked increase in theexpression of p21, without p53 activation after treatmentwith cisplatin at 6 �g/ml in CAL 27 cells (Yip et al. 2006).Also, in our study and again contrary to the Yip et al.results, we were not able to detect basal p21 protein expres-sion in CAL 27 cells by western blotting; although we

managed to detect it by immunocytochemistry. Further-more, Yip et al. detected raised p53 protein levels inCCL23 cells, in spite of p53-inactivation by human papillo-mavirus type 18-E6 protein. They speculate that cisplatinmight inactivate E6 protein in CCL23, or induce p21 via ap53-independent pathway in CAL 27 cells. We believe thatthe dose of cisplatin is crucial for the observed divergenteVects, since we administered an order-of-magnitude lowerdose that could not induce such eVects (6 vs. 0.5 �g/ml).The p21 protein was readily expressed after adenovirus-mediated p53 overexpression, which conWrms its transcrip-tional activity. No diVerence in p21 (as well as p53) proteinexpression was detected in all AdCMV-p53-treated CAL27 cells, while p21 was somewhat less expressed in bothcombination-treatments than in AdCMV-p53 treatmentalone in HEp-2 cells after 24 h. The p53-expression gradu-ally and strongly increased with the time passing in bothcombinations, but mostly in the combination of AdCMV-p53 with methotrexate in HEp-2 pointing to a fact that bothchemotherapeutics, but mostly methotrexate initiallyrepressed the exogenous p53 expression in HEp-2 cells,while afterwards both combined treatments, but againmostly methotrexate, led to a possible synergistic eVect.Interestingly, despite the high p53 protein expressionobserved in both combined treatments after 48 h, no p21protein was detected that could be the main reason for theobserved failure of HEp-2 cells to arrest the cell cycle. TheeVects of cisplatin and methotrexate on the cell cycleperturbation were expected—cisplatin caused an initialS phase delay which ended in a G2 phase accumulation,while methotrexate induced a strong accumulation of G1/Scell population in both cell lines, primarily in HEp-2 cells,being in accordance with the previous reports demonstrat-ing the early S-phase arrest after the treatment of variouscell lines with 0.1 �M methotrexate (Linke et al. 1996).

A similar eVect of cisplatin was already described forsome of p53-defective cells (Lee et al. 1999; Sun et al.2002). The accumulation of cells in S and G2 phases wasconcomitant with the cyclin B1 expression in both HEp-2and CAL 27 cells, which was in agreement with previousstudies (Dan and Yamori 2001). Cyclin B1 is a welldescribed G2 cyclin and its accumulation begins in the Sphase, reaches the maximal level at mitosis, being rapidlydegraded after the metaphase/anaphase transition (Piaggioet al. 1995; Pines and Hunter 1989; Clute and Pines 1999;Chang et al. 2003). Cyclin B1-Cdk1 is a master mitotic reg-ulator being therefore an ultimate target toward which mostantiproliferative signals generated during S and G2 phaseconverge (Charrier-Savournin et al. 2004).

In general, cells with diVerent p53 status respond diVer-ently to DNA damage, whereby p53wt cells execute a longterm cell cycle arrest after treatment, whereas the p53-deW-cient cells undergo a transient cell cycle arrest followed by

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1759

apoptosis or mitotic catastrophe. The reason for such diver-gent responses is that the p53mut cells usually do not main-tain the G2 arrest after DNA damage, but initiate prematuremitosis that leads to cell death (mitotic catastrophe),accompanied by the emergence of aberrant mitoses and dis-tinct nuclear forms of micronucleation and apoptoticnuclear condensation (Bhonde et al. 2006).

In CAL 27 cells p53 overexpression strongly reduced thecyclin B1 accumulation, pointing to the re-activation ofG2/M checkpoint, resulting in a sustained G2 arrest withalso G1 arrest, i.e. the cells behaved like p53wt. Indeed, itwas shown that p53 inhibits the transcription of B type ofcyclins and that p53 regulates a G2 checkpoint throughcyclin B1 (Krause et al. 2000; Innocente et al. 1999).Besides, it was documented that not only p53, but p21expression is essential for maintaining the G2 checkpoint(Bunz et al. 1998; Maeda et al. 2002). Moreover, Charrier-Savournin et al. (2004) showed that p21 exerts a critical rolein maintaining the permanent G2 arrest by nuclear seques-tration and by blocking activation of cyclin B1-Cdk1. In ourexperiments p21 protein is still expressed in CAL 27 cellstreated with cisplatin/AdCMV-p53 even after 48 h, contraryto HEp-2 cells where p21 is still expressed only in the Ad-p53 treated cells. We propose that the expression of p21 inCAL 27 cells even after 48 h together with a decrease ofcyclin B1 ensures the G2 arrest, along with G1 arrest andreduction of cells in S phase. These results are also in goodconcordance with and complement the recent results shownby Rakitina et al. 2007 that point to interventions to cellcycle (e.g. G1/S transition inhibition) in cells with impairedp53 function (by introduction of p21 gene, or inhibition ofcdk 1 and/or 2), could potentiate the activity of cytotoxicdrugs, such as oxaliplatin (Rakitina et al. 2007).

Contrary to this, HEp-2 cells treated with cisplatin andAdCMV-p53 did not respond by diminishing the cyclin B1levels. Additionally, cyclin B1 expression after 48 h inAdCMV-p53 treatment alone was associated with the G2/Mtransition (as documented by the microscopy data) andaccumulation of cells in the G1 phase. Therefore, the p53overexpression (achieved by infecting the cells at MOI 10)in the HPV-E6/E7 infected cells failed to maintain G2arrest in cells suVering DNA damage in the late S phase(such as cisplatin-treated cells in our study) for a prolongedincubation period (48 h), while it is still slightly able toslow down the progression through S phase in the cellstreated with methotrexate, not allowing their progressionthrough mitosis. The most probable factors for this obser-vation are: (a) a lack of p21 protein expression, which wasshown to be required for both the G1 and G2 phase arrestinduction and the G2 arrest maintenance (Charrier-Savour-nin et al. 2004; Bhonde et al. 2006), and (b) the expressionof human papilloma virus (HPV) E6 and E7 proteins inHEp-2 cells that degrade both p53 and pRb proteins,

respectively. Namely, Charrier-Savournin et al. (2004)nicely demonstrated that even in the presence of high p21levels, HPV18-E7-inactivated pRb protein compromisesthe irreversible cell cycle exit. Therefore, unlike CAL 27,HEp-2 cells progressed through M phase despite higher p53and/or p21 levels induced upon AdCMV-p53 treatments.Besides, Funk et al. (1997) showed that E7 protein inacti-vates both CDK and PCNA-dependant inhibitory functionsof p21 through binding to sequences in the carboxy-termi-nal end of p21. These and other data imply E7 protein candisrupt normal cell cycle control. In the end it seems thatthese cells treated with AdCMV-p53 or in combinationwith cisplatin enter an aberrant form of mitosis and die afterprolonged time period (as documented by microscopy datashowing dead/apoptotic cells), especially in the treatmentswhere the amount of cyclin B1 was still high enough toforce cell into mitosis. This should be more closely investi-gated. On the other hand, signiWcantly higher p53 expres-sion (higher MOIs used for adenovirus infection) issuYcient to overcome the E6/7-induced p53/pRb inhibitionand is able to induce early apoptosis. Nevertheless, in bothcases p53 overexpression substantiates the cisplatin-induced cell cycle changes and leads to more pronouncedeVects.

Methotrexate treatment was slowing down the S phaseprogression in both cell lines, along with only slight or nosigniWcant cyclin B1 induction. The combination with p53overexpression slightly accentuated the G1 arrest andstrongly decreased cyclin B1 accumulation in CAL 27 (24and 48 h) and HEp-2 cells only after 24 h. Interestingly, inthe combination treatment of HEp-2 cells with AdCMV-p53 and methotrexate the p53 expression gradually roseeven above the protein level detected in the AdCMV-p53treatment alone indicating that as time passed, methotrexatecould not repress the exogenous p53 expression, and aswell might have initiated transcription of endogenous pro-tein as a response to the chemotherapy. This should alsocontribute to the observed reduced proliferation rate in themethotrexate-treated cells (as opposed to the complete lackof arrest in the cisplatin-treated cells). The mechanismunderlying this phenomenon and a possibility of a synergis-tic eVect should be additionally investigated.

In conclusion, this study demonstrates that adenovirallymediated p53 overexpression at sub-cytotoxic levelsenhanced the activity of two mechanistically diVerent che-motherapeutics cisplatin and methotrexate at low doses insquamous head and neck tumor cell lines through changesin the cell cycle. However, the mechanisms of these eVectsdiVer depending on the cell line, i.e. genetic context. InCAL 27 cells, bearing mutated p53 gene, overexpression ofp53 completely abrogated high S phase content observed inmethotrexate treatment into a G1 and slight G2 arrest,while it sustained G2 arrest of the cells treated with

123

1760 J Cancer Res Clin Oncol (2009) 135:1747–1761

cisplatin, along with the reduction of DNA synthesis.Therefore, the low levels of p53 protein expression in thiscell line was able to achieve suYcient eVect, which shouldbe compared to the p53wt cells, which were able to activateand/or maintain both G1 and G2 arrest following DNAdamage. On the other hand, p53 overexpression in HPV 18E6/7 infected cell line HEp-2, which therefore has nonfunc-tional p53 and pRb protein, had diVerent eVect concerningthe chemotherapeutic used; (a) in the combination withmethotrexate, it slightly delayed the progression through Sphase (without diminishing the percentage of S phasecells), not allowing the cells to proceed to mitosis; (b)failed to sustain the initial S/G2 delay after treatment withcisplatin, but instead it enhanced S/G2/M progression withan increase of the cells in S phase that were not activelysynthesizing DNA. Our assumption is that it probablyforced the cells to proceed to an aberrant form of mitosisthat was driven by high level of cyclin B1.

Acknowledgments Support for this study by the Ministry ofScience, Education and Sport of Croatia is gratefully acknowledged(Projects 098-0982464-2393, 098-0982464-2514). The authors areparticularly thankful to Dr. Anamaria BrozoviT for exceptionally use-ful discussions and experimental help and Dr. Ksenija Zahradka forproviding us with the BrdU chemical and anti-BrdU antibody.

References

Bhonde MR, Hanski M-L, Budczies J, Cao M, Gillissen B, Moorthy D,Simonetta F, Scherübl H, Truss M, Hagemeier C, Mewes H-W,Daniel PT, Zeitz M, Hanski C (2006) DNA damage-inducedexpression of p53 suppresses mitotic checkpoint kinase hMps1:the lack of this suppression in p53MUT cells contributes to apopto-sis. J Biol Chem 281:8675–8685. doi:10.1074/jbc.M511333200

Blagosklonny MV (2002) p53: an ubiquitous target of anticancerdrugs. Int J Cancer 98:161–166. doi:10.1002/ijc.10158

Brugarolas J, Moberg K, Boyd SD, Taya Y, Jacks T, Lees JA (1999)Inhibition of cyclin-dependent kinase 2 by p21 is necessary forretinoblastoma protein-mediated G1 arrest after gamma-irradia-tion. Proc Natl Acad Sci USA 96:1002–1007. doi:10.1073/pnas.96.3.1002

Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP,Sedivy JM, Kinzler KW, Vogelstein B (1998) Requirement forp53 and p21 to sustain G2 arrest after DNA damage. Science282:1497–1501. doi:10.1126/science.282.5393.1497

Bunz F, Hwang PM, Torrance C, Waldmann T, Zhang Y, Dillehay L,Williams J, Lengauer C, Kinzler KW, Vogelstein B (1999) Dis-ruption of p53 in human cancer cells aster s the responses to ther-apeutic agents. J Clin Invest 104:263–269. doi:10.1172/JCI6863

Chang DC, Xu N, Luo KQ (2003) Degradation of cyclin B is requiredfor the onset of anaphase in mammalian cells. J Biol Chem278:37865–37873. doi:10.1074/jbc.M306376200

Charrier-Savournin FB, Chateau M-T, Gire V, Sedivy J, Piette J, DuliTV (2004) p21-mediated nuclear retention of cyclin B1-Cdk1 inresponse to genotoxic stress. Mol Biol Cell 15:3965–3976.doi:10.1091/mbc.E03-12-0871

Clute P, Pines J (1999) Temporal and spatial control of cyclin B1destruction in metaphase. Nat Cell Biol 1:82–87. doi:10.1038/10049

Dan S, Yamori T (2001) Repression of cyclin B1 expression after treat-ment with adriamycin, but not cisplatin in human lung cancerA549 cells. Biochem Biophys Res Commun 280:861–867.doi:10.1006/bbrc.2000.4231

el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, TrentJM, Lin D, Mercer WE, Kinzler KW, Vogelstein B (1993) WAF1,a potential mediator of p53 tumor suppression. Cell 75:817–825

Evan GI, Vousden KH (2001) Proliferation, cell cycle and apoptosis incancer. Nature 411:342–348

Foulkes WD (2007) p53—master and commander. N Engl J Med357:2539–2541

Fujiwara K, Daido S, Yamamoto A, Kobayashi R, Yokoyama T, AokiH, Iwado E, Shinojima N, Kondo Y, Kondo S (2008) Pivotal roleof the cyclin-dependent kinase inhibitor p21WAF1/CIP1 in apop-tosis and autophagy. J Biol Chem 283:388–397

Funk JO, Waga S, Harry JB, Espling E, Stillman B, Galloway DA(1997) Inhibition of CDK activity and PCNA-dependent DNAreplication by p21 is blocked by interaction with the HPV-16 E7oncoprotein. Genes Dev 11:2090–2100

Gallagher WM, Brown R (1999) p53-oriented cancer therapies: cur-rent progress. Ann Oncol 10:139–150

Ganjavi H, Gee M, Narendran A, Parkinson N, Krishnamoorthy M,Freedman MH, Malkin D (2006) Adenovirus-mediated p53 genetherapy in osteosarcoma cell lines: sensitization to cisplatin anddoxorubicin. Cancer Gene Ther 13:415–419

Gartel AL, Tyner AL (2002) The role of the cyclin-dependent kinaseinhibitor p21 in apoptosis. Mol Cancer Ther 1:639–649

Hait WN, Yang J-M (2006) The individualization of cancer therapy:the unexpected role of p53. Trans Am Clin Climatol Assoc117:85–101

Hattangadi DK, DeMasters GA, Walker TD, Jones KR, Di X,Newsham IF, Gewirtz DA (2004) InXuence of p53 and caspase 3activity on cell death and senescence in response to methotrexatein the breast tumor cell. Biochem Pharmacol 68:1699–1708

Hollstein M, Shomer B, Greenblatt M, Soussi T, Hovig E, MontesanoR, Harris CC (1996) Somatic point mutations in the p53 gene ofhuman tumors and cell lines: updated compilation. Nucleic AcidRes 24:141–146

Horio Y, Hasegawa Y, Sekido Y, Takahashi M, Roth JA, Shimokata K(2000) Synergistic eVects of adenovirus expressing wild-type p53on chemosensitivity of non-small cell lung cancer cells. CancerGene Therapy 7:537–544

Horowitz J-A (1999) Adenovirus-mediated p53 gene therapy: over-view of preclinical studies and potential clinical applications.Curr Opin Mol Ther 1:500–509

Innocente SA, Abrahamson JLA, Cogswell JP, Lee JM (1999) p53 reg-ulates a G2 checkpoint through cyclin B1. Proc Natl Acad Sci96:2147–2152

Inoue A, Narumi K, Matsubara N, Sugawara S, Saijo Y, Satoh K,Nukiwa T (2000) Administration of wild-type 53 adenoviral vec-tor synergistically enhances the cytotoxicity of anti-cancer drugsin human lung cancer cells irrespective of the status of p53 gene.Cancer Lett 157:105–112

Jänicke RU, Sohn D, Schulze-OsthoV K (2008) The dark side of atumor suppressor: anti-apoptotic p53. Cell Death DiVer 15:959–976

Kandioler D, Stamatis G, Eberhardt W, Kappel S, Zöchbauer-Müller S,Kührer I, Mittlböck M, Zwrtek R, Aigner C, Bichler C, Tichy V,Hudec M, Bachleitner T, End A, Müller MR, Roth E, Klepetko W(2008) Growing clinical evidence for the interaction of the p53genotype and response to induction chemotherapy in advanced non-small cell lung cancer. J Thorac Cardiovasc Surg 135:1036–1041

Kastan MB (2007) Wild-type p53: tumors can’t stand it. Cell 128:837–840

Kralj M, Pavelic J (2003) p21WAF1/CIP1 is more eVective than p53 ingrowth suppression of mouse renal carcinoma cell line Renca invitro and in vivo. J Cancer Res Clin Oncol 129:463–471

123

J Cancer Res Clin Oncol (2009) 135:1747–1761 1761

Kralj M, Husnjak K, Körbler T, Pavelic J (2003) Endogenous p21WAF1/CIP1

status predicts the response of human tumor cells to wild-type p53and p21WAF1/CIP1 overexpression. Cancer Gene Ther 10:457–467

Kraljevic Pavelic S, Cacev T, Kralj M (2008) A dual role of p21waf1/cip1

gene in apoptosis of HEp-2 treated with cisplatin or methotrexate.Cancer Gene Ther 15:576–590

Krause K, Wasner M, Reinhard W, Haugwith U, Lange-ze Dohna C,Mossner J, Engeland K (2000) The tumor suppressor protein p53can repress transcription of cyclin B. Nucl Acid Res 28:4410–4418

Kupryjanczyk J, Kraszewska E, Ziolkowska-Seta I, Madry R, TimorekA, Markowska J, Stelmachow J, Bidzinski M, Polish Ovarian Can-cer Study Group (POCSG) (2008) TP53 status and taxane-plati-num versus platinum-based therapy in ovarian cancer patients: anon-randomized retrospective study. BMC Cancer 8:27

LaBaer J, Garrett MD, Stevenson LF, Slingerland JM, Sandhu C, ChouHS, Fattaey A, Harlow E (1997) New functional activities for thep21 family of CDK inhibitors. Genes Dev 11:847–862

Lee SI, Brown MK, Eastman A (1999) Comparison of the eYcacy of7-hydroxystaurosporine (UCN-01) and other staurosporine ana-logs to abrogate cisplatin-induced cell cycle arrest in humanbreast cancer cell lines. Biochem Pharmacol 58:1713–1721

Linke SP, Clarkin KC, Di Leonardo A, Tsou A, Wahl GM (1996) Areversible, p53-dependent G0/G1 cell cycle arrest induced byribonucleotide deplenon in the absence of detectable DNA dam-age. Genes Dev 10:934–947

Maeda T, Chong MT, Espino RA, Chua PP, Cao JQ, Chomey EG,Luong L, Tron VA (2002) Role of p21Waf-1 in regulating theG1 and G2/M checkpoints in ultraviolet-irradiated keratinocytes.J Invest Dermatol 119:513–521

Meek DW (2004) The p53 response to DNA damage. DNA Repair3:1049–1056

Nielsen LL, Maneval DC (1998) p53 tumor suppressor gene therapyfor cancer. Cancer Gene Ther 5:52–63

Peng Z (2005) Current status of gendicine in China: recombinanthuman Ad-p53 agent for treatment of cancers. Hum Gene Ther16:1016–1027

Piaggio G, Farina A, Perrotti D, Manni I, Fuschi P, Sacchi A, GaetanoC (1995) Structure and growth-dependent regulation of thehuman cyclin B1 promoter. Exp Cell Res 216:396–402

Pines J, Hunter T (1989) Isolation of a human cyclin cDNA: evidencefor cyclin mRNA and protein regulation in the cell cycle and forinteraction with p34cdc2. Cell 58:833–846

Radhakrishnan SK, Feliciano CS, Najmabadi F, Haegebarth A, KandelES, Tyner AL, Gartel AL (2004) Constitutive expression ofE2F-1 leads to p21-dependent cell cycle arrest in S phase of thecell cycle. Oncogene 23:4173–4176

Rakitina TV, Vasilevskaya IA, O’Dwyer PJ (2007) Inhibition of G1/Stransition potentiates oxaliplatin-induced cell death in colon can-cer cells lines. Biochem Pharmacol 73:1715–1726

Roth RA (2006) Adenovirus p53 gene therapy. Expert Opin Biol Ther6:55–61

Slee EA, O’Connor DJ, Lu X (2004) To die or not to die: how does p53decide? Oncogene 23:2809–2818

Sun D, Urrabaz R, Buzello C, Nguyen M (2002) EVects of cisplatin onexpression of DNA ligases in MiaPaCa human pancreatic cancercells. Biochem Biophys Res Commun 298:537–544

Supek F, Kralj M, MarjanoviT M, Kuman L, Kmuc T, KrizmaniT I,LiniT B (2008) Atypical cytostatic mechanism of N-1-sulfonylcy-tosine derivatives determined by in vitro screening and computa-tional analysis. Invest New Drugs 26(2):97–110

Ventura A, Kirsch DG, McLaughlin ME, Tuveson DA, Grimm J,Lintault L, Newman J, Reczek EE, Weissleder R, Jacks T (2007)Restoration of p53 function leads to tumour regression in vivo.Nature 445:661–665

Vousden KH (2006) Outcomes of p53 activation-spoilt for choice.J Cell Sci 119:5015–5020

Waldman T, Kinzler KW, Vogelstein B (1995) p21 is necessary for thep53-mediated G1 arrest in human cancer cells. Cancer Res55:5187–5190

Wallace-Brodeur RR, Lowe SW (1999) Clinical implications of p53mutations. Cell Mol Life Sci 55:64–75

Weinrib L, Li J-H, Donovan J, Huang D, Liu F-F (2001) Cisplatinchemotherapy plus adenoviral p53 gene therapy in EBV-positiveand -negative nasopharyngeal carcinoma. Cancer Gene Ther8:352–360

Yip HT, Chopra R, Chakrabarti R, Veena MS, Ramamurthy B, SrivatsanES, Wang MB (2006) Cisplatin-induced growth arrest of headand neck cancer cells correlates with increased expression of p16and p53. Arch Otolarynog Head Neck Surg 132:317–326

Zhou B-B, Bartek J (2004) Targeting the checkpoint kinases: chemo-sensitization versus chemoprotection. Nat Rev Cancer 4:1–10

123