IDp

MPIPa

b

c

d

e

f

a

ARRAA

KMDpp�

1

sbvTbdosec

MT

0h

Leukemia Research 37 (2013) 1576– 1582

Contents lists available at ScienceDirect

Leukemia Research

journa l h o me pag e: www.elsev ier .com/ locate / leukres

ncreased expression of phosphorylated NBS1, a key molecule of theNA damage response machinery, is an adverse prognostic factor inatients with de novo myelodysplastic syndromes

aria Kefalaa, Sotirios G. Papageorgioub, Christos K. Kontosc,anagiota Economopouloub, Athanasios Tsanasd, Vasiliki Pappab,oannis G. Panayiotidesa, Vassilios G. Gorgoulise, Eustratios Patsouris f,eriklis G. Foukasa,∗

2nd Department of Pathology, University of Athens Medical School, “Attikon” University Hospital, Chaidari, Greece2nd Department of Internal Medicine, Propaedeutic, Hematology Unit, University of Athens Medical School, “Attikon” University Hospital, Chaidari, GreeceDepartment of Biochemistry and Molecular Biology, University of Athens, Athens, GreeceSystems Analysis Modeling and Prediction Group, Oxford Centre for Industrial and Applied Mathematics, University of Oxford, UKMolecular Carcinogenesis Group, Department of Histology and Embryology, University of Athens Medical School, Athens, Greece1st Department of Pathology, University of Athens Medical School, Athens, Greece

r t i c l e i n f o

rticle history:eceived 18 June 2013eceived in revised form 22 August 2013ccepted 25 August 2013vailable online 5 September 2013

a b s t r a c t

The expression of activated forms of key proteins of the DNA damage response machinery (pNBS1,pATM and �H2AX) was assessed by means of immunohistochemistry in bone marrow biopsies of 74patients with de novo myelodysplastic syndromes (MDS) and compared with 15 cases of de novo acutemyeloid leukemia (AML) and 20 with reactive bone marrow histology. Expression levels were signifi-cantly increased in both MDS and AML, compared to controls, being higher in high-risk than in low-risk

eywords:yelodysplastic syndromesNA damage responseNBS1ATMH2AX

MDS. Increased pNBS1 and �H2AX expression possessed a significant negative prognostic impact foroverall survival in MDS patients, whereas pNBS1 was an independent marker of poor prognosis.

© 2013 Elsevier Ltd. All rights reserved.

. Introduction

Myelodysplastic syndromes (MDS) are clonal hematopoietictem cell disorders with great heterogeneity in terms of clinical,iological and prognostic features, which has been ascribed to theariety of genetic lesions contributing to MDS pathogenesis [1].hey are characterized by ineffective hematopoiesis, peripherallood cytopenias (anemia, neutropenia and/or thrombocytopenia),ysplasia in one or more myeloid cell lines and an increased riskf progression to acute myeloid leukemia (AML). Prognostic clas-

ification of various types of MDS concerning survival and risk ofvolution to AML is based on the percentage of bone marrow blasts,ytogenetic findings and number of cytopenias; based upon the

∗ Corresponding author at: 2nd Department of Pathology, University of Athensedical School, “Attikon” University Hospital, 1st Rimini St, 12462 Haidari, Greece.

el.: +30 6945750125.E-mail address: pfoukas@med.uoa.gr (P.G. Foukas).

145-2126/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.leukres.2013.08.018

above, the International Prognostic Scoring System (IPSS) [2] sub-divides MDS into four categories, namely low risk, intermediate-1risk (INT-1), intermediate-2 risk (INT-2) and high risk; these aregrouped into low-risk MDS (including low risk and INT-1) and high-risk MDS (including INT-2 and high risk MDS).

About 50% of MDS cases display aneuploidy (the state of acell containing an aberrant number of whole or parts of thechromosomes) [3,4] at the time of diagnosis, a finding indicativeof chromosomal instability (CIN) at early stages of hematopoi-etic stem cell clonal evolution. Although molecular mechanismsunderlying chromosomal instability remain largely unknown, for-mation of DNA double-strand breaks (DSBs) can lead to this typeof genomic instability [5]. Besides, formation of DNA DSBs acti-vates the DNA damage response (DDR) pathway which, accordingto the “oncogene-induced DNA damage model for cancer develop-

ment and progression” [6], represents an antitumor barrier raised inprecancerous lesions, by inducing p53-dependent cell cycle arrest,apoptosis and senescence, features that have been described inMDS [7,8].

Research 37 (2013) 1576– 1582 1577

Maoppplapsfcoa[

mccssttdst([

kiap

2

2

SDcantiw(hslcicAtHufifsTRs

2

GId

Table 1Clinical and biological characteristics of MDS patients.

No. patients 74Sex (male/female) 46/28Median (range)

Age (years) 75 (41–87)Hemoglobin (g/dL) 9.8 (5.7–14.1)WBCa (×103/mL) 4195 (1100–30,360)ANCb (×103/mL) 1735 (80–7425)Platelets (×103/mL) 131,500 (7000–440,000)Blast count (%) 6 (0–20)Overall survival (months) 24 (1–148)Karyotype (67/74 patients)

Normal 41 (61%)Abnormal 26 (39%)

Cytogenetic risk (67/74 patients)Low 47 (70%)Intermediate 11 (16.5%)High 9 (13.5%)

WHOc classification (74/74 patients)RA 9 (12.2%)RARS 3 (4.1%)RCMD 23 (31.1%)RAEB-1 23 (31.1%)RAEB-2 16 (21.6%)

IPSSd risk (70/74 patients)Low 20 (28.6%)Intermediate-1 27 (38.6%)Intermediate-2 16 (22.9%)High 7 (10.0%)

a White blood cells.b Absolute neutrophil count.c World Health Organization classification (RA, refractory anemia; RARS, refrac-

tory anemia with ringed sideroblasts; RCMD, refractory cytopenia with multilineage

risk), according to their IPSS score. For categorization of pNBS1, �H2AX, pATM, and

M. Kefala et al. / Leukemia

An early event in response to DNA DSBs is the recruitment ofRN complex proteins (MRE11, RAD50 and NBS1), which serves

s the sensor of DNA breakage and leads to the phosphorylationf NBS1 [9,10]. Phosphorylated NBS1 (pNBS1) activates throughhosphorylation the key signal transducer of the DDR pathway,rotein kinase ATM, which in turn phosphorylates many substrateroteins, including histone protein H2AX, whose rapid phosphory-

ation leads to �H2AX foci formation at nascent DSB sites [9]. Theforementioned events finally induce the activation of tumor sup-ressor protein p53, which leads to cell cycle arrest, apoptosis orenescence, constituting an evolutionary pressure for the transitionrom precancerous lesion to full blown malignancy [6]. These pro-esses can be monitored by means of either immunofluorescencer immunohistochemistry, using primary antibodies that detect thectivated, phosphorylated forms of NBS1, ATM and H2AX and p5311,12].

Following publication of the aforementioned cancer develop-ent model, many studies performed mainly on solid tumors

onfirm the initial prediction, i.e. the activation of DDR in pre-ancerous lesions [11–14]. Concerning neoplastic hematopoietictem cell disorders, two studies performed on bone marrowpecimens from MDS and AML patients detected activation ofhe DDR machinery in preleukemic lesions, i.e. MDS, in contrasto normal bone marrow samples; their results were contra-ictory concerning the expression of these molecules in AMLamples [15,16]. Moreover, in a recent study we have shownhat proteins involved in the Non Homologous End JoiningNHEJ) mechanism of DSBs repair are expressed in MDS cases17].

The aim of the present study was to examine the expression ofey proteins of the DDR machinery, i.e. pNBS1, pATM and �H2AX,n bone marrow cells of adult de novo MDS and to determine theirssociation with clinical and biological parameters, as well as withrognosis.

. Materials and methods

.1. Study population

The current study included 74 cases of newly diagnosed adult de novo MDS.amples were obtained by bone marrow biopsy before the onset of any treatment.iagnosis was issued according to the updated (2008) WHO classification (1). Theirlinical and biological characteristics are summarized in Table 1. Patients with lownd INT-1 score according to the IPSS (47/70 patients) who were asymptomatic didot receive any treatment or received only growth factors (Epo and GM-CSF) andransfusions of red blood cells (RBC) and/or platelets (PLT) as needed. The vast major-ty of patients with INT-2 and high risk IPSS scores (19/23 patients) were treated

ith 5-azacytidine (Vidaza). Four patients were treated with AML chemotherapyAracytine and Idarubicin; 7 + 3), while one of them received additional allogeneicematopoietic stem cell transplantation. In addition, 20 cases of bone marrow biop-ies from age-matched patients (11 men, 9 women) in the process of non-Hodgkinymphoma staging (7 cases of low-grade B-cell lymphomas, 10 of diffuse large B-ell lymphomas and 3 of T-cell lymphomas), without any evidences of lymphomanfiltration, served as controls, as well as 15 cases of de novo AML, were added foromparison purposes; hence, the total study population consisted of 109 cases.ll patients represented in this study underwent trephine biopsy in the Hema-

ology Unit of the 2nd Department of Internal Medicine in “Attikon” Universityospital between December 2003 and September 2010. All of them had a follow-p period of minimum one year (median 24, range 1–148). All specimens werexed in aceto-zinc formalin solution for 24 h followed by decalcification in Decal®

or 3 h and were then embedded into paraffin. Hematoxylin-and-Eosin-stainedlides, as well as appropriate immunostains were used for diagnostic purposes.he study was approved by the “Attikon” University Hospital Institutional Ethicseview Board; all patients had given informed consent for participation in thetudy.

.2. Cytogenetic analysis

Cytogenetic analysis was performed by direct culture of bone marrow cells and-banding technique. Chromosomal aberrations were described according to the

nternational System of Human Cytogenetic Nomenclature (ISCN 1995) recommen-ations [18].

dysplasia; RAEB, refractory anemia with excess blasts).d International prognostic scoring system.

2.3. Immunohistochemistry

Immunostaining was performed on 3-�m deparaffinized sections. Primary anti-bodies against p95-NBS1 (phospho S432) (rabbit monoclonal, clone EPR2470Y[abcam], 1:100 dilution), ATM (phospho S1981) (rabbit monoclonal, clone EP1890Y[abcam], 1:100 dilution), �H2AX (phospho S139) (mouse monoclonal, clone 9F3[abcam], 1:500 dilution) and p53 (mouse monoclonal, clone DO-7 [DAKO], 1:50dilution) were used. For pATM, �H2AX and p53, slides were immersed in ahigh pH target retrieval solution (K8004 [DAKO]), boiled in a microwave at650 W for 20 min, and subsequently cooled at room temperature for 20 min.Endogenous peroxidase activity was blocked by means of the peroxidase-blockingreagent S2001. Immunohistochemistry was performed using the DAKO AutostainerPlus device with EnVisionTM detection and visualization kit (K5007 [DAKO]).For pNBS1 immunostaining, BondTM Automated Immunohistochemistry Systemwas used. In sixteen randomly chosen cases (5 reactive, 4 RA, 1 RARS, 2RAEB-1, 1 RAEB-2 and 3 AML) double immunostaining was performed usingprimary antibodies against CD34 (mouse monoclonal, clone QBEnd-10 [DAKO],1:50 dilution) and �H2AX. Double staining was performed with EnVisionTM

G/2 Doublestain System, Rabbit/Mouse (DAB+/Permanent Red), Code K5361. Allsections were lightly counterstained with Hematoxylin for 25 sec prior to mount-ing.

2.4. Scoring of immunohistochemical staining

Cell counting was done blindly with an 60× objective [Nikon eclipse E400microscope] by a trained (MK) and a qualified (PGF) pathologist. At least1000 marrow cells per case were assessed and the percentage of positivenuclei calculated. Double staining was scored by counting the percentage of�H2AX positive cells in at least 100 CD34 positive cells, with blastic morphol-ogy.

2.5. Statistical analysis

MDS patients were divided into two subgroups (low/INT-1 risk and INT-2/high

p53 percentage of immunopositive cells, the X-tile algorithm [19] was used to gen-erate optimal cutpoints: 35.8% (46th percentile) for pNBS1, 66.3% (72nd percentile)for pATM, 66.1% (75th percentile) for �H2AX and 5.0% (76th percentile) for p53.Finally, extensive statistical analysis was performed (Online Supplementary Designand Methods).

1578 M. Kefala et al. / Leukemia Resea

Table 2Frequency of the detected numerical and structural abnormalities in MDS patients.

Karyotype n (%) MDS subtype

Normal 41 (61.0) 6RA, 3 RARS, 15RCMD, 11 RAEB-1,6 RAEB-2

Abnormal 26 (39.0)Del (5q) 5 (19.0) 1 RA, 2 RAEB-1, 2

RAEB-2Del (20q) 3 (11.5) 1 RCMD, 1 RAEB-1,

1 RAEB-2−Y 3 (11.5) 1 RA, 2 RCMDTrisomy 8 3 (11.5) 1 RCMD, 1 RAEB-1,

1 RAEB-2Del (11q) 2 (8.0) 1 RCMD, 1 RAEB-2Inv (9) 2 (8.0) 2 RAEB-1Inv (3) 1 (4.0) 1 RCMDAbnormalities of

chromosome 77 (27.0) 1 RCMD, 5 RAEB-1,

1 RAEB-2Complex karyotype 3 (11.5) 3 RAEB-2

MDS, myelodysplastic syndromes, RA, refractory anemia; RARS, refractory anemiawR

3

3

si

3

cit

Fs

ith ringed sideroblasts; RCMD, refractory cytopenia with multilineage dysplasia;AEB, refractory anemia with excess blasts.

. Results

.1. Cytogenetic analysis

Cytogenetic analysis was successful in 67 out of 74 patients. Ashown in Table 2, structural or numerical abnormalities were foundn 26 patients (39%).

.2. Expression of proteins involved in DDR machinery

In the present study, we examined the expression of proteinomponents of the DDR machinery i.e. pNBS1, �H2AX and pATMn 74 MDS patients. Expression of all molecules was localized inhe nucleus. Any cytoplasmic staining was not further considered

ig. 1. Immunohistochemistry of reactive bone marrow, low-risk MDS, high-risk MDS and

how differences in mean values of percentage of immunopositive cells. Error bars repres

rch 37 (2013) 1576– 1582

in this analysis. The percentage of immunopositive cells for thethree DDR-involved proteins was significantly increased in MDSand AML compared to controls [for pNBS1, 38.7 (range: 0.0–85.8)in MDS patients, 80.7 (range: 65.5–95) in AML patients and 8.8(range: 1.2–36.8) in normal controls (P < 0.001), for �H2AX, 32.2(range: 1.2–93.4) in MDS patients, 86.8 (range: 35.1–96.8) in AMLpatients and 16.1 (range: 0.0–70.3) in controls (P < 0.001) and forpATM, 38.3 (range: 4.9–85) in MDS, 79.7 (range: 0.0–100) in AMLand 19.3 (range: 8.4–34.7) in controls (P < 0.001)] (Online Supple-mentary Table S1). In addition, we found that the expression ofthe proteins that participate in the DDR machinery differed sig-nificantly among the subtypes of MDS and was higher in high riskMDS (RAEB-1 and RAEB-2) compared to low risk MDS (RA, RARS,and RCMD), categorized according to WHO classification (2008),and even more elevated in AML patients (P < 0.001 for all examinedDDR-involved proteins) (Fig. 1; Online Supplementary Table S2).

Finally, in MDS patients the expression level of any one of pNBS1,�H2AX and pATM proteins was positively correlated to the expres-sion levels of the rest two, while the expression of �H2AX and pATMalso correlated with p53 expression (Online Supplementary TableS3).

3.3. Relationships between pNBS1, �H2AX and pATM expression,and MDS patients’ clinicopathological variables

Positive correlations were observed between pNBS1, �H2AXand pATM expression status and blast count (rs = 0.462, P < 0.001,rs = 0.274, P = 0.018 and rs = 0.342, P = 0.003 respectively). For�H2AX, this finding was also confirmed in double-stained slides,whereby the median percentage value of CD34+ blasts immu-nostained for �H2AX was 5-fold higher in high risk MDS andmore than 9-fold higher in AML cases compared to controls. No

difference was detected between controls and low risk MDS.Additionally, a significant negative correlation was found betweenpNBS1 expression status and hemoglobin as well as white bloodcount (WBC), suggesting higher pNBS1 expression in patients with

AML representative cases for pNBS1, �H2AX and pATM. Box-and-whiskers diagramsent standard errors of the mean.

Resea

aP�ePs(sc(2p(aS

optgpwmp(

3p

icpA0vwt1aiooTpnMm

3

2TmO(ptKp

DCTa

M. Kefala et al. / Leukemia

nemia (rs = −0.357, P = 0.002) and leucopenia (rs = −0.245, = 0.035). Finally, a negative correlation was observed betweenH2AX expression and platelet count, implying higher �H2AXxpression in patients with thrombocytopenia (rs = −0.241,

= 0.038) (Online Supplementary Table S4). In addition, expressiontatus of each protein under study was categorized into two groupspositive vs. negative), as described in the “Design and Methods”ection. Thus, among the 74 MDS cases examined, 40 (54.1%) werelassified as pNBS1-positive and 34 (45.9%) as pNBS1-negative, 1824.3%) as �H2AX-positive and 56 (75.7%) as �H2AX-negative and0 (27.0%) as pATM-positive and 54 (73.0%) as pATM-negative.NBS1-positive patients had more frequently abnormal karyotypeP = 0.011), intermediate or high cytogenetic risk profile (P = 0.041)nd belonged to a more advanced IPSS category (P = 0.006) (Onlineupplementary Table S5).

Regarding pATM expression, pATM-positive patients had moreften abnormal karyotype compared to the pATM-negative MDSopulation (P = 0.046), while �H2AX-positive patients belongedo an advanced IPSS category compared to the negative sub-roup of patients, (P = 0.007). When the percentage of eachrotein expression was analyzed as a continuous variable, itas shown that the expression of all three analyzed DDRachinery molecules increased gradually among, normal controls,

atients with low/INT-1 and INT-2/high MDS, and AML patientsFig. 1).

.4. p53 Expression and its correlations with DDR-involvedroteins and clinicopathological characteristics of MDS patients

We also examined the expression status of p53 protein andts association with DDR-involved proteins, as well as with clini-opathological characteristics of MDS patients. The percentage of53-immunopositive cells was significantly increased in MDS andML compared to normal controls [median % positivity: 1.6 (range:.0–72.8) in MDS patients vs. 45 (range: 9.3–95.0) in AML patientss. 0.0 (range: 0.0–1.6) in normal controls, P < 0.001]. In addition,e found that the expression of p53 differed significantly among

he different subtypes of MDS, being higher in high risk MDS (RAEB- and RAEB-2) compared to low-risk MDS (RA, RARS, and RCMD),ccording to WHO (2008) classification, and even more elevatedn AML (P < 0.001) (Online Supplementary Tables S1 and S2). More-ver, p53 expression was positively correlated with the expressionf �H2AX (P = 0.016) and pATM (P = 0.007) (Online Supplementaryable S3). A significant association was also observed between the53% positivity and the karyotype category, as MDS patients withormal karyotype were less frequently p53-positive compared toDS patients with abnormal karyotype (P = 0.021) (Online Supple-entary Table S6).

.5. Prognostic role of DDR-involved proteins in MDS

Follow-up information was available for 70 out of 74 patients;6 deaths were recorded, all of which due to causes related to MDS.he median overall survival (OS) was 24 months (range: 1–148onths). Kaplan–Meier analysis revealed significantly reducedS of pNBS1-positive compared to pNBS1-negative MDS patients

P = 0.002), of �H2AX-positive compared to �H2AX-negative MDSatients (P = 0.005) and reduced OS of p53-positive comparedo p53-negative MDS patients (P = 0.024). On the other hand,aplan–Meier OS curves did not differ significantly between pATM-ositive and pATM-negative MDS patients (Fig. 2).

To evaluate the prognostic effect of the positive expression of

DR machinery proteins on the OS in MDS patients, we performedox regression analysis. As demonstrated in Online Supplementaryable S7, univariate Cox regression analysis revealed that pNBS1nd �H2AX positivities carry a significant negative prognostic

rch 37 (2013) 1576– 1582 1579

impact for OS in MDS patients [hazard ratio (HR) = 3.65, 95%confidence interval (CI) = 1.52–8.73, P = 0.004 and HR = 2.98, 95%CI = 1.32–6.75, P = 0.009, respectively]. Interestingly, the multivari-ate Cox regression models, adjusted for IPSS risk and patients’ age,revealed an independent negative prognostic significance of pNBS1positivity for OS (HR = 2.79, 95% CI = 1.15–6.81, P = 0.024) (OnlineSupplementary Table S7).

4. Discussion

In this retrospective study we assessed the expression of pNBS1,pATM and �H2AX in a series of MDS as well as de novo AML patients.To the best of our knowledge, no studies dealing with the expres-sion of all three key molecules of DNA-damage response in MDSpatients have so far been published. Our data show that the phos-phorylation events of these proteins correlate among each other,indicating the participation in a common activated pathway inMDS cases. Moreover, the expression of all molecules significantlycorrelates with blast count and is increased in MDS cases com-pared to controls, with a statistical significant further increase inAML cases compared to MDS cases. Accordingly, in MDS patients,the expression of all 3 molecules is positively correlated with thetype of MDS, being higher as the prognostic IPSS category worsens.The baseline positive cells counted in tissue sections from casesserved as controls may represent an overestimation, supposing thatlymphomas induce DNA damage in distant proliferative tissues inhumans, in analogy to what has been shown in a mouse tumormodel [20].

DDR activation in MDS has also been suggested in recent studies,through the detection of pATM and/or �H2AX [15,16,21]. Fur-thermore, although in one of them [16], pATM but not �H2AXexpression was significantly increased in AML samples comparedto MDS, correlating with the blast percentage, in the second [15],no further increase of pATM, pCHK2 and �H2AX was detected inAML cases. These discrepancies might be related to the AML pop-ulation under investigation, this including de novo AML cases inthe first study (n = 5) compared to MDS-related AML in the sec-ond (n = 10). It is well known that the activation of DNA damageresponse is an early event during neoplastic transformation raisinga barrier to progression, through increased apoptosis or senescenceof transformed cells [6]. Increased apoptosis (which may under-lie ineffective hematopoiesis [7,22,23]), as well as senescence havebeen described in MDS patients [8]. Furthermore, even in low-riskMDS subtypes, DNA double-strand breaks have been detected bycomet assay [24], suggesting a causative role for the activation ofDDR in MDS. The fact that pNBS1, pATM and �H2AX are overex-pressed in MDS, even in low grade, is an indicator of activation ofDNA damage response pathway early in preleukemic stages.

Among precancerous disorders, MDS are unique in that in theusual diagnostic workflow, information concerning the degree ofcell differentiation (assessed by means of complete blood count),as well as the presence of cytogenetic abnormalities (through clas-sical karyotype analysis) is essential in the evaluation of everysuspected case. MDS is, accordingly, an ideal human model to inves-tigate possible correlations between DDR activation and degreeof cellular differentiation or ploidy status. Interestingly, increasedexpression of pNBS1 was found in patients with lower levels ofhemoglobin (P = 0.002) and WBC (P = 0.035), whereas increasedexpression of �H2AX was related to thrombocytopenia (P = 0.038).In previous studies, evidence of inhibition of differentiation afterDNA damage has been observed in skeletal myogenesis [25,26]

and pancreatic organogenesis [27]. It is considered that the dif-ferentiation checkpoint is a barrier that inhibits the expression ofdifferentiation-specific genes in precursor cells after DNA dam-age, preventing the generation of differentiated cells bearing

1580 M. Kefala et al. / Leukemia Research 37 (2013) 1576– 1582

pNBS

uiils

ii(tmpf[ohs

Fig. 2. Overall survival of MDS patients in relation to

nresolved lesions [28,29]. In accordance to these ex vivo stud-es, our results may suggest that the activation of DDR machineryn MDS cases might interfere with differentiation of hematopoieticineages, providing a second potential mechanism beyond apopto-is for the observed peripheral cytopenias.

Another interesting finding of the present work was the pos-tive correlation between the expression of molecules involvedn the DDR machinery, namely pNBS1 and pATM and aneuploidyP = 0.011 and 0.046, respectively). It is well known that aneuploidy,he numerical and structural alteration of chromosomes, is a com-

on characteristic of tumor cells [30]. Overall, about 50% of MDSatients possess an abnormal karyotype, highest frequencies beingound in patients who have refractory anemia with excess of blasts

3]. The role of aneuploidy in the process of carcinogenesis remainsbscure, as both oncogenic as well as tumor suppressor propertiesave been attributed to cells with an aberrant number of chromo-omes [31]. Higher expression of DDR molecules in aneuploid MDS

1 (A), �H2AX (B), pATM (C) and p53 (D) expression.

cases raises the possibility that even in early stages of carcino-genesis, aneuploidy is proportional to genetic instability [32,33]probably through oncogene induced DNA DSBs [6]. On the otherhand, higher expression of DDR molecules might be the result of dif-ferential gene expression and stress response in aneuploid cells, ashas been recently described in yeast and mammalian cells [34] andthis could explain the positive correlation of pNBS1 expression withintermediate and high cytogenetic risk cases, in which a wide vari-ety of different karyotypes are included. Interestingly, in all threeMDS cases with trisomy 8, high expression of pNBS1 was found,suggesting a correlation between MYC oncogene overexpressionand its transcriptional target, NBS1 [35].

Our results suggest that overexpression of the activated form of

key molecules of DDR machinery might have an adverse effect onprognosis in MDS patients. In fact, univariate analysis (n = 70) indi-cates that increased expression of pNBS1 and �H2AX is associatedwith poor prognosis (P = 0.004 and 0.009, respectively), whereas

Resea

mierappvhwco

aaomstfao

C

R

G7idm

A

a

ddc

m

l

m

A

f2

R

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[[

M. Kefala et al. / Leukemia

ultivariate analysis (n = 67) confirmed this association only forncreased expression of pNBS1 (P = 0.024). Moreover, pATM over-xpression was associated in univariate analysis with a trend foreduced overall survival (P = 0.082). As suggested earlier for headnd neck squamous cell carcinomas [36], NBS1 may act as an onco-rotein, participating in the MDS-related development of AML,ossibly through the activation of PI3K/AKT, a pathway that is acti-ated in high risk MDS [37]. In accordance, NBS1 overexpressionas been associated with worse prognosis in uveal melanomas [38],hereas NBS1 defect is negatively related to survival in breast can-

er patients, indicating a tumor suppressor function in the contextf that disease [39].

In conclusion, our results show that key molecules of DNA dam-ge response pathway are expressed in their activated form in MDSnd AML. Further studies are needed to better elucidate the rolef this pathway in MDS and AML pathogenesis. Activated NBS1ay be a potential independent prognostic factor by means of

tandard immunostaining during initial MDS diagnosis, thus iden-ifying patients with poor prognosis; these high risk patients mayurther benefit from therapeutic targeting of phosphorylated NBS1,s has been shown ex vivo for BCR/ABL+ leukemic cells [40], anption that needs to be supported by additional studies.

onflict of interest statement

Nothing to report.

ole of the funding source

This work was supported by the Special Account for Researchrants, National and Kapodistrian University of Athens (grant0/4/9960 to Periklis G. Foukas). The funding source had no

nvolvement in the study design, collection and interpretation ofata, writing of the manuscript and the decision to submit theanuscript for publication.

cknowledgements

We thank Konstantinos Savvatakis for his excellent technicalssistance.

Contributions. MK, SGP and PGF provided the conception andesign of the study, acquisition, analysis and interpretation of data;rafting the article and revised it critically for important intellectualontent.

MK and PGF performed experiments.CKK carried out the statistical analysis and drafted the

anuscript.PE and AT supplied the analysis of data.VP, IGP, VGG, EP were responsible critically for important intel-

ectual content of the article.All authors have read and approved the submitted form of the

anuscript.

ppendix A. Supplementary data

Supplementary data associated with this article can beound, in the online version, at http://dx.doi.org/10.1016/j.leukres.013.08.018.

eferences

[1] Swerdlow S. WHO classification of tumours of haematopoietic and lymphoidtissues. 4th ed; 2008. p. 88–93. Lyon.

[2] Greenberg P, Cox C, LeBeau MM, Fenaux P, Morel P, Sanz G, et al. Internationalscoring system for evaluating prognosis in myelodysplastic syndromes. Blood1997;89:2079–88.

[

[

rch 37 (2013) 1576– 1582 1581

[3] Pozdnyakova O, Miron PM, Tang G, Walter O, Raza A, Woda B, et al. Cytoge-netic abnormalities in a series of 1,029 patients with primary myelodysplasticsyndromes: a report from the US with a focus on some undefined single chro-mosomal abnormalities. Cancer 2008;113:3331–40.

[4] Fenaux P. Chromosome and molecular abnormalities in myelodysplastic syn-dromes. Int J Hematol 2001;73:429–37.

[5] Jeggo PA, Lobrich M. DNA double-strand breaks: their cellular and clinicalimpact. Oncogene 2007;26:7717–9.

[6] Halazonetis TD, Gorgoulis VG, Bartek J. An oncogene-induced DNA damagemodel for cancer development. Science 2008;319:1352–5.

[7] Parker JE, Mufti GJ, Rasool F, Mijovic A, Devereux S, Pagliuca A. The roleof apoptosis, proliferation, and the Bcl-2-related proteins in the myelodys-plastic syndromes and acute myeloid leukemia secondary to MDS. Blood2000;96:3932–8.

[8] Wang YY, Cen JN, He J, Shen HJ, Liu DD, Yao L, et al. Accelerated cellular senes-cence in myelodysplastic syndrome. Exp Hematol 2009;37:1310–7.

[9] Uziel T, Lerenthal Y, Moyal L, Andegeko Y, Mittelman L, Shiloh Y. Require-ment of the MRN complex for ATM activation by DNA damage. EMBO J2003;22:5612–21.

10] Economopoulou P, Pappa V, Papageorgiou S, Dervenoulas J, Economopoulos T.Abnormalities of DNA repair mechanisms in common hematological malignan-cies. Leuk Lymphoma 2011;52:567–82.

11] Bartkova J, Horejsi Z, Koed K, Kramer A, Tort F, Zieger K, et al. DNA damageresponse as a candidate anti-cancer barrier in early human tumorigenesis.Nature 2005;434:864–70.

12] Gorgoulis VG, Vassiliou LV, Karakaidos P, Zacharatos P, Kotsinas A, LiloglouT, et al. Activation of the DNA damage checkpoint and genomic instability inhuman precancerous lesions. Nature 2005;434:907–13.

13] Oka K, Tanaka T, Enoki T, Yoshimura K, Ohshima M, Kubo M, et al. DNA damagesignaling is activated during cancer progression in human colorectal carcinoma.Cancer Biol Ther 2010;9:246–52.

14] Koorstra JB, Hong SM, Shi C, Meeker AK, Ryu JK, Offerhaus GJ, et al. Widespreadactivation of the DNA damage response in human pancreatic intraepithelialneoplasia. Mod Pathol 2009;22:1439–45.

15] Horibe S, Takagi M, Unno J, Nagasawa M, Morio T, Arai A, et al. DNA damagecheck points prevent leukemic transformation in myelodysplastic syndrome.Leukemia 2007;21:2195–8.

16] Boehrer S, Ades L, Tajeddine N, Hofmann WK, Kriener S, Bug G, et al. Suppressionof the DNA damage response in acute myeloid leukemia versus myelodysplasticsyndrome. Oncogene 2009;28:2205–18.

17] Economopoulou P, Pappa V, Kontsioti F, Papageorgiou S, Foukas P, Liakata E,et al. Expression analysis of proteins involved in the non homologous endjoining DNA repair mechanism, in the bone marrow of adult de novo myelodys-plastic syndromes. Ann Hematol 2010;89:233–9.

18] Schreck R. Curr Protoc Hum Genet. Appendix 4: Appendix 4B.19] Camp RL, Dolled-Filhart M, Rimm DL. X-tile: a new bio-informatics tool for

biomarker assessment and outcome-based cut-point optimization. Clin CancerRes 2004;10:7252–9.

20] Redon CE, Dickey JS, Nakamura AJ, Kareva IG, Naf D, Nowsheen S, et al. Tumorsinduce complex DNA damage in distant proliferative tissues in vivo. Proc NatlAcad Sci USA 2010;107:17992–7.

21] Head DR, Jacobberger JW, Mosse C, Jagasia M, Dupont W, Goodman S,et al. Innovative analyses support a role for DNA damage and an aber-rant cell cycle in myelodysplastic syndrome pathogenesis. Bone Marrow Res2011;2011:950934.

22] Delia D, Aiello A, Soligo D, Fontanella E, Melani C, Pezzella F, et al. bcl-2 Proto-oncogene expression in normal and neoplastic human myeloid cells. Blood1992;79:1291–8.

23] Tsoplou P, Kouraklis-Symeonidis A, Thanopoulou E, Zikos P, Orphanos V, Zoum-bos NC. Apoptosis in patients with myelodysplastic syndromes: differentialinvolvement of marrow cells in ‘good’ versus ‘poor’ prognosis patients andcorrelation with apoptosis-related genes. Leukemia 1999;13:1554–63.

24] Novotna B, Neuwirtova R, Siskova M, Bagryantseva Y. DNA instability inlow-risk myelodysplastic syndromes: refractory anemia with or without ringsideroblasts. Hum Mol Genet 2008;17:2144–9.

25] Ito H, Miller SC, Billingham ME, Akimoto H, Torti SV, Wade R, et al. Doxorubicinselectively inhibits muscle gene expression in cardiac muscle cells in vivo andin vitro. Proc Natl Acad Sci USA 1990;87:4275–9.

26] Kurabayashi M, Jeyaseelan R, Kedes L. Antineoplastic agent doxorubicin inhibitsmyogenic differentiation of C2 myoblasts. J Biol Chem 1993;268:5524–9.

27] Lebrun P, Montminy MR, Van Obberghen E. Regulation of the pancreatic duo-denal homeobox-1 protein by DNA-dependent protein kinase. J Biol Chem2005;280:38203–10.

28] Puri PL, Bhakta K, Wood LD, Costanzo A, Zhu J, Wang JY. A myogenic differen-tiation checkpoint activated by genotoxic stress. Nat Genet 2002;32:585–93.

29] Simonatto M, Latella L, Puri PL. DNA damage and cellular differentiation: morequestions than responses. J Cell Physiol 2007;213:642–8.

30] Mitelman Database of Chromosome Aberrations in Cancer.31] Weaver BA, Silk AD, Montagna C, Verdier-Pinard P, Cleveland DW. Aneu-

ploidy acts both oncogenically and as a tumor suppressor. Cancer Cell 2007;11:25–36.

32] Duesberg P, Rausch C, Rasnick D, Hehlmann R. Genetic instability of cancercells is proportional to their degree of aneuploidy. Proc Natl Acad Sci USA1998;95:13692–7.

33] Maeck L, Haase D, Schoch C, Hiddemann W, Alves F. Genetic instability inmyelodysplastic syndrome: detection of microsatellite instability and loss

1 Resea

[

[

[

[

[

[

582 M. Kefala et al. / Leukemia

of heterozygosity in bone marrow samples with karyotype alterations. Br JHaematol 2000;109:842–6.

34] Sheltzer JM, Torres EM, Dunham MJ, Amon A. Transcriptional consequences ofaneuploidy. Proc Natl Acad Sci USA 2012;109:12644–9.

35] Chiang YC, Teng SC, Su YN, Hsieh FJ, Wu KJ. c-Myc directly regulates the tran-scription of the NBS1 gene involved in DNA double-strand break repair. J BiolChem 2003;278:19286–91.

36] Yang MH, Chiang WC, Chou TY, Chang SY, Chen PM, Teng SC, et al.

Increased NBS1 expression is a marker of aggressive head and neck cancerand overexpression of NBS1 contributes to transformation. Clin Cancer Res2006;12:507–15.

37] Follo MY, Mongiorgi S, Bosi C, Cappellini A, Finelli C, Chiarini F, et al. TheAkt/mammalian target of rapamycin signal transduction pathway is activated

[

rch 37 (2013) 1576– 1582

in high-risk myelodysplastic syndromes and influences cell survival and pro-liferation. Cancer Res 2007;67:4287–94.

38] Ehlers JP, Harbour JW. NBS1 expression as a prognostic marker in uvealmelanoma. Clin Cancer Res 2005;11:1849–53.

39] Bartkova J, Tommiska J, Oplustilova L, Aaltonen K, Tamminen A, Heikkinen T,et al. Aberrations of the MRE11-RAD50-NBS1 DNA damage sensor complex inhuman breast cancer: MRE11 as a candidate familial cancer-predisposing gene.Mol Oncol 2008;2:296–316.

40] Rink L, Slupianek A, Stoklosa T, Nieborowska-Skorska M, Urbanska K, Sefer-ynska I, et al. Enhanced phosphorylation of Nbs1, a member of DNArepair/checkpoint complex Mre11-RAD50-Nbs1, can be targeted to increasethe efficacy of imatinib mesylate against BCR/ABL-positive leukemia cells.Blood 2007;110:651–60.