RESEARCH ARTICLE

Deregulation of base excision repair gene expressionand enhanced proliferation in head and neck squamouscell carcinoma

Ishrat Mahjabeen & Kashif Ali & Xiaofeng Zhou &

Mahmood Akhtar Kayani

Received: 10 December 2013 /Accepted: 24 February 2014 /Published online: 13 March 2014# International Society of Oncology and BioMarkers (ISOBM) 2014

Abstract Defects in the DNA damage repair pathway con-tribute to cancer. The major pathway for oxidative DNAdamage repair is base excision repair (BER). Although BERpathway genes (OGG1, APEX1 and XRCC1) have beeninvestigated in a number of cancers, our knowledge on theprognostic significance of these genes and their role in headand neck squamous cell carcinoma is limited. Protein levels ofOGG1, APEX1 and XRCC1 and a proliferation marker, Ki-67, were examined by immunohistochemical analysis, in acohort of 50 HNSCC patients. Significant downregulation ofOGG1 (p<0.04) and XRCC1 (p<0.05) was observed in poor-ly differentiated HNSCC compared to mod–well-differentiat-ed cases. Significant upregulation of APEX1 (p<0.05) andKi-67 (p<0.05) was observed in poorly differentiatedHNSCC compared to mod-well-differentiated cases. Signifi-cant correlation was observed between XRCC1 and OGG1(r=0.33, p<0.02). Inverse correlations were observed be-tween OGG1 and Ki-67 (r=−0.377, p<0.005), betweenAPEX1 and XRCC1 (r=−0.435, p<0.002) and betweenOGG1 and APEX1 (r=−0.34, p<0.02) in HNSCC. To con-firm our observations, we examined BER pathway genes anda proliferation marker, Ki-67, expression at the mRNA levelon 50 head and neck squamous cell carcinoma (HNSCC) and

50 normal control samples by quantitative real-time polymer-ase chain reaction. Significant downregulation was observedin case of OGG1 (p<0.04) and XRCC1 (p<0.02), whilesignificant upregulation was observed in case of APEX1(p<0.01) and Ki-67 (p<0.03) in HNSCC tissue samples com-pared to controls. Our data suggested that deregulation of baseexcision repair pathway genes, such as OGG1, APEX1 andXRCC1, combined with overexpression of Ki-67, a markerfor excessive proliferation, may contribute to progression ofHNSCC in Pakistani population.

Keywords HNSCC . BER pathway gene . Proliferationmarker . Immunohistochemistry

Introduction

DNA repair systems play a critical role in protecting thehuman genome from damage caused by carcinogens presentin the environment [1]. The major pathway for oxidative DNAdamage repair is the base excision repair (BER) pathway [2].Key genes such as 8-hydroxyguanine DNA glycosylase(OGG1), apurininc/apyrimidinic endonuclease 1 (APEX1)and X-ray repair cross-complementing 1 (XRCC1) are re-quired for viability and efficient repair of DNA damage inthe BER pathway [3].

The OGG1 gene is located on human chromosome 3p26.2,a region that frequently shows loss of heterozygosity (LOH) inseveral human cancers [4]. The human OGG1 interacts phys-ically and functionally with XRCC1, and the DNAglycosylase activity of OGG1 increases two- to threefold[5]. Substantial evidence indicates an important role for X-ray repair cross-complementing 1 (XRCC1) in single-strandbreak repair (SSBR) and base excision repair pathway (BER)[6]. Apparently, devoid of any enzymatic activity, this proteinis thought to act as a scaffolding protein for other repair factors

I. Mahjabeen :K. Ali :M. A. Kayani (*)Cancer Genetics Lab, Department of Biosciences, COMSATSInstitute of Information and Technology, Park Road Chakshazad,Islamabad, Pakistane-mail: mkayani@comsats.edu.pk

I. Mahjabeen :X. ZhouCenter for Molecular Biology of Oral Diseases, College of Dentistry,University of Illinois at Chicago, Chicago, IL, USA

X. ZhouDepartment of Periodontics, College of Dentistry, Graduate College,UIC Cancer Center, University of Illinois at Chicago, Chicago, IL,USA

Tumor Biol. (2014) 35:5971–5983DOI 10.1007/s13277-014-1792-5

[7]. Apurininc/apyrimidinic endonuclease 1 (APEX1) pro-teins also have important functions in the BER pathway [8].APEX1 resides on 14q11.2–14q12 and processes the AP sitesor single-strand breaks (SSBs), remaining after the damagedbase has been excised by DNA glycosylases, which is con-sidered the rate-limiting step in the base excision repair path-way (BER) [9]. APEX1 has also been shown to interact withXRCC1 [10], and APEX1 overexpression can compensate forXRCC1-deficient cells in the repair of DNA SSBs induced byoxidative DNA damage, both in vivo and in whole-cell ex-tracts [11].

In addition to the DNA repair system, cell prolifera-tive activity is another reliable prognostic factor in headand neck tumours and may be helpful in differentialdiagnosis of head and neck cancer (HNC). Studies in-dicated that the Ki-67 labelling index of more than10 % is diagnostic of HNC. The prognostic significanceof Ki-67 has been reported in various tumours, includ-ing laryngeal carcinoma [12], salivary gland adenoidcystic carcinoma [13], mucoepidermoid carcinoma [14],hepatocellular carcinoma [15], breast carcinoma [16]and lung carcinoma [17]. Very few studies are availableregarding the role of DNA repair gene expression andtheir association with proliferation and head and neckcarcinogenesis [12, 18].

The previous studies by same research group havefound novel and known mutations in OGG1 [19, 20],APEX1 [21] and XRCC1 genes [22] in germline screen-ing. Similarly at RNA screening, significant variationsin expression have been reported in OGG1, XRCC1[23] and APEX1 [21] genes, and a significant correla-tion was observed between BER pathway genes andproliferation marker Ki-67. Therefore, in continuationof the previous germline screening and somatic screen-ing at the mRNA level of these genes in head and neckcancer patients, the current study was designed to in-vestigate BER pathway (OGG1, APEX1 and XRCC1)protein expression and its relationship to excessive pro-liferation (as measured by the proliferation marker Ki-67). In addition to this, we also determined whetherexpression of BER pathway proteins and Ki-67 is dif-ferent in primary tumours of early versus advanced headand neck squamous cell carcinoma (HNSCC).

Materials and methods

Samples

Tumours along with adjacent normal tissues (controls)were collected after informed consent from 50 HNSCCpatients, immediately after surgery from Military Hospi-tal (MH, Rawalpindi) and Allied Hospital (Faisalabad).

The presence of tumour cells in the collected tissueswas rectified by examination of frozen sections follow-ing haematoxylin and eosin stain (HE stain) by a con-sultant pathologist. All individuals were personallyinterviewed using the specifically designed questionnaire.Clinical or pathologic staging and identification of theanatomical site of the lesions were based on the Inter-national Union against Cancer TNM classification ofmalignant tumours (2003). Clinical characterization ofthe patients is summarized in Table I. The distributionof the 50 tumours included 24 (48 %) cases of squa-mous cell carcinoma of the larynx, 12 (24 %) cases ofthe pharynx and 14 (28 %) cases of squamous cellcarcinoma of the oral cavity (Table 1).This study wasconducted with prior approval from ethical committeesof both CIIT and collaborating hospitals.

Immunohistochemistry and scoring

Immunohistochemical analysis was carried out using theDAB chromogen for staining. Three and half micron-thicksections were cut from a representative paraffin block andplaced on glass slides. The sections were first de-paraffinized in xylene and then rehydrated in a descendingseries of ethanol concentrations, as described previously [24].

Table 1 Demographic and clinical characterization of the cohort

Variable HNSCC Cases

Age Median(range) 60 (20-80)

Gender Male: n (%) 27 (54)

Female: n (%) 23 (46)

Area of cancer Larynx: n (%) 24 (48)

Pharynx: n (%) 12 (24)

Oral cavity: n (%) 14(28)

Survival Disease free: n (%) 31 (62)

Dead: n (%) 19 (38)

Clinical stage Stage I: n (%) 05 (10)

Stage II: n (%) 09 (18)

Stage III: n(%) 26 (52)

Stage IV: n (%) 10 (20)

T stage Stage I: n (%) 05 (10)

Stage II: n (%) 11 (22)

Stage III: n (%) 28 (56)

Stage IV: n (%) 06(12)

N stage Stage 0: n (%) 16(32)

Stage I: n (%) 28 (56)

Stage II: n (%) 06 (12)

Grade Poor: n (%) 05 (10)

Mod: n (%) 15 (20)

Well: n (%) 30 (60)

The M stage data for HNSCC patients are not available

5972 Tumor Biol. (2014) 35:5971–5983

Sections were then blocked with 10 % normal serum for15 min at 37 °C followed by incubation with rabbit anti-OGG1 (Novus Biologicals), mouse anti-APEX1 (Novus Bio-logicals), mouse anti-XRCC1 (Novus Biologicals) or mouseanti-Ki-67 (Abcam) at a dilution of 1:1,000 for 1 h at roomtemperature. After washing three times in PBS, the sectionswere incubated with single stain boost immunohistochemistry(IHC) detection (Cell Signaling Technology) for 30 min atroom temperature. After additional washing in PBS, the sec-tions were incubated with two drops of Reagent 2 conjugatefor 10 min at room temperature. This was followed by rinsingin PBS solution three times. One drop of reagent A, two dropsof reagent B and one drop of reagent C were added in 1 ml ofdistilled water. It was mixed and protected from light and usedwithin 1 h. Two drops of DAB chromogen were added toslides and left for 10–15 min.

The slides were washed in distilled water and counterstained with haematoxylin for approximately 10 min. Slideswere rinsed in distilled water, coverslipped using aqueousmounting medium and allowed to dry at room temperature.Negative controls were prepared using the same procedureexcept that the primary antibodies were replaced with PBS.

The relative intensities of the completed immunohisto-chemical reactions were evaluated using light microscopy bythree independent trained observers, unaware of the clinicaldata. Tumour cells were counted randomly in ten high powerfields for measurement of immunoreactivity. The followingformula was used for the evaluation of immunoreactivity:

Immunoreactive score ¼ intensity score � propotion scrore:

The intensity score was categorized as 0 at negative inten-sity, 1 at weak intensity, 2 at moderate intensity or 3 at strongstaining intensity. The propotion score was categorized as 0 atno positive cell, 1 at ≤10 % of positive cells, 2 at 11–50 % ofpositive cells, 3 at 51–80 % of positive cells or 4 at >80 %distribution of positive cells. This immunoreactivity (rangingfrom 0 to 12) was divided into low immunoreactivity (totalimmunoreactive score of 0–4) and high immunoreactivity(total immunoreactive score of >4).

RNA extraction and real-time PCR

RNA isolation was carried out from tumour samples of a studygroup by using the standard Trizol reagent method [21] andstored at −80 °C. Reverse transcription PCR (RT–PCR) wascarried out using SuperScript First-Strand Synthesis System(Invitrogen, USA). The SuperScript III First-Strand SynthesisSystem for RT–PCR was optimized to synthesize first-strandcDNA from purified poly(A)+or total RNA. It provides highcDNAyields, sensitivity and specificity.

For quantitative PCR (qPCR), primers specific for geneOGG1, APEX1, XRCC1, Ki-67 and β actin (internal control)were purchased from OriGene (USA). Each qPCR was per-formed in a 20-μl reaction mixture containing ~1 μl of prod-uct from RT reaction, 10 μl of 2× Sybr Green, 1 μl of eachprimer and 7-μl RNase free water. qPCRwas performed usingReal-Time PCR system (Applied Biosystems Step one plus)under standard conditions. The relative mRNA levels ofOGG1, APEX1, XRCC1 and β actin were computed usingthe 2−ΔΔCt analysis method [23].

Statistical analysis

One-way ANOVA and chi-square test were used to assess theassociation of OGG1, APEX1, XRCC1 and Ki-67 expressionwith clinical and histopathological parameters (e.g. TNM andgrade). Spearman correlation coefficient was used to assesscorrelations among the gene–gene expression and gene–clin-ical and histopathological parameters. Statistical significancewas established at a value of p<0.005. Asterisks (*), (**) and(***) indicate p<0.05, p<0.01 and p<0.001, respectively.Statistical analyses were performed using GraphPad Prism5software and SPSS software package.

Results

Immunohistochemical analysis of BER pathway genesand Ki-67

Immunohistochemical analysis was used to assess the expres-sion of selected BER pathway genes (i.e. OGG1, APEX1 andXRCC1) and the proliferation marker Ki-67. Fifty tumoursamples of HNSCC were used in this analysis. Immunostain-ing for OGG1, APEX1 and XRCC1 was observed in nuclearand cytoplasmic regions. Immunostaining for Ki-67 was ob-served in the nuclear region only. The immunostaining foreach protein was also determined as positive or negative by acutoff value determined as follow: OGG1, APEX1 andXRCC1 staining was interpreted as positive when >10 % ofthe tumour cells showed distinct staining. Ki-67 staining wasinterpreted as positive when >25 % showed distinct nuclearstaining [25].

Positive control and negative control are used to test theprocedure and to check the specificity of primary antibody andsecondary antibody, respectively. Ductal carcinoma of breastwas used as positive control for OGG1, APEX1 and XRCC1.Tonsil tissue was used as positive control for Ki-67. Fornegative control in place of primary antibody, phosphatebuffer saline (PBS) was applied on a duplicate slide and thentreated with secondary antibody and so on.

Tumor Biol. (2014) 35:5971–5983 5973

Differences in expression profile in early and advancedHNSCC

The expression profiles of each investigated markers in well-differentiated, moderately and poorly differentiated HNSCCcases were compared. As illustrated in Fig. 1, OGG1 expres-sion was lower in HNSCC tumours compared to positivecontrol tissue samples. This protein level was observed lower(p<0.04) in poorly differentiated tumours compared to well-differentiated and moderately differentiated tumours. The ex-pression profile of APEX1 was observed higher in HNSCCtumours compared to positive control tissue samples, and thisupregulation was observed higher in poorly differentiatedtumours (p<0.05) compared to well-differentiated and mod-erately differentiated HNSCC cases (Fig. 2). As shown in

Fig. 3, the XRCC1 protein level is significantly downregulat-ed in HNSCC tumour samples compared to positive controltissue samples, and this downregulation is significantly lower(p<0.05) in poorly differentiated tumours when compared tomoderately or well-differentiated tumours. In case of Ki-67,the protein level was overexpressed (p<0.05) in HNSCCtumour tissues when compared with positive control tissuesamples. This overexpression was observed higher in poorlydifferentiated tumours compared to well- or moderately dif-ferentiated HNSCC cases (Fig. 4).

Protein levels of BER pathway genes and Ki-67

As illustrated in Fig. 5, among 50 HNSCC samples that weexamined, 78 % cases exhibited downregulated OGG1

Fig. 1 Immunohistochemistryanalysis of OGG1 expression inHNSCC tissue samples.Immunohistopathologicalanalysis was performed toexamine the OGG1 expression ona negative control, b positivecontrol, c well-differentiatedprimary, d moderatelydifferentiated, e poorlydifferentiated or undifferentiatedtumour and f lymph nodemetastasis tissue samples. Fornegative control in place ofprimary antibody, phosphatebuffer saline (PBS) was used, andfor positive control, ductalcarcinoma of breast was used.Representative images a, b, c, d, e(×20) and f (×40) were shown

5974 Tumor Biol. (2014) 35:5971–5983

expression, 10% cases exhibited upregulated OGG1 level and2 % cases showed no results. The relative intensities of OGG1protein levels showed that 44 % cases showed weak immu-noreactive intensity, 36 % cases showed moderate immuno-reactive intensity and 20 % cases showed strong immunore-active intensity. Weak immunoreactive intensity of OGG1was prevalent in significantly higher (p<0.04) number ofHNSCC tumour samples when compared to moderate orstrong immunoreactive intensities (Table 2). In case of theAPEX1 protein level we examined, 32 % cases exhibiteddownregulated APEX1 expression, 65 % cases exhibited up-regulated APEX1 level and 3 % cases showed no results(Fig. 5). Relative immunoreactive intensities of APEX1

protein levels exhibited that 18 % cases indicated weak im-munoreactive intensity, 37 % cases showed moderate immu-noreactive intensity and 45 % cases showed strong immuno-reactive intensity. Strong immunoreactive intensity of APEX1is significantly higher (p<0.05) in HNSCC tumour sampleswhen compared to weak and moderate immunoreactive inten-sities (Table 2).

The expressional analysis of the XRCC1 gene exhibitedthat 64 % cases showed downregulation, 30 % showed up-regulation and 6 % cases showed no change (Fig. 5). Semi-quantitative analysis revealed that 48 % cases showed weakimmunoreactive intensity, 32 % cases showed moderate im-munoreactive intensity and 20 % cases showed strong

Fig. 2 Immunohistochemistryanalysis of APEX1 expression inHNSCC tissue samples.Immunohistopathologicalanalysis was performed toexamine the APEX1 expressionon a negative control, b positivecontrol, c well-differentiatedprimary, d moderatelydifferentiated, e poorlydifferentiated or undifferentiatedtumour and f lymph nodemetastasis tissue samples. Fornegative control in place ofprimary antibody, phosphatebuffer saline (PBS) was used, andfor positive control, ductalcarcinoma of breast was used.Representative images a, b, e, f(×40) and c, d (×20) were shown

Tumor Biol. (2014) 35:5971–5983 5975

immunoreactive intensity. Weak immunoreactive intensitywas prevalent in significantly higher (p<0.05) number ofHNSCC tumour samples compared to moderate or strongimmunoreactive intensity (Table 2). Among 50 HNC samplesthat we examined, 19 % cases exhibited downregulated Ki-67expression, 75 % cases exhibited upregulated Ki-67 level and6 % cases showed no results (Fig. 5). Relative immunoreac-tive intensities of Ki-67 protein levels showed that 21 % casesshowed weak immunoreactive intensity, 30 % cases showedmoderate immunoreactive intensity and 49 % cases showedstrong immunoreactive intensity. Strong immunoreactive in-tensity of Ki-67 is significantly higher (p<0.05) in HNSCCtumour samples when compared to weak and moderate im-munoreactive intensities (Table 2). The impact of clinicalfactors on the expression level of investigated proteins wasevaluated in Table 2. C stage, pT stage, pN stage, grade andsurvival were evaluated according to weak, moderate andstrong immunoreactive intensity.

Correlations among BER pathway genes, Ki-67 proteinexpression and clinicopathological features in HNSCC

When gene–gene relationship was explored, a statisticallysignificant correlation between XRCC1 expression andOGG1 expression was observed (r=0.33, p<0.02). Inversecorrelations were observed between OGG1 and Ki-67 (r=−0.377, p<0.005), between APEX1 and XRCC1(r=−0.435,p<0.002) and between OGG1 and APEX1 (r=−0.34,p<0.02) in HNSCC cases. An apparent correlation and anapparent inverse correlation were also observed betweenAPEX1 and Ki-67 (r=0.071, p=0.62) and between XRCC1and Ki-67 (r=−0.069, p=0.63), respectively. However, thesecorrelations were not statistically significant (Table 3).

In case of the relationship among clinicopathological pa-rameters, statistically significant correlations were observedbetween gender and area (r=0.319, p<0.02), between pT andC stage (r=0.942, p<0.000), between pT and PN (r=0.98,

Fig. 3 Immunohistochemistryanalysis of XRCC1 expression inHNSCC tissue samples.Immunohistopathologicalanalysis was performed toexamine the APEX1 expressionon a negative control, b positivecontrol, c well-differentiatedprimary, d moderatelydifferentiated, e poorlydifferentiated or undifferentiatedtumour and f lymph nodemetastasis tissue samples. Fornegative control in place ofprimary antibody, phosphatebuffer saline (PBS) was used, andfor positive control, ductalcarcinoma of breast was used.Representative images a, b, c, d(×20) and e, f (×40) were shown

5976 Tumor Biol. (2014) 35:5971–5983

p<0.000), between pTand grade (r=0.291, p<0.04), betweenC stage and pN (r=0.929, p<0.000) and between C stage andgrade (r=0.289, p<0.04). Inverse correlation was observedbetween age and pT (r=−0.366, p<0.009), between age and Cstage (r=−0.362, p<0.01), between age and pN (r=−0.336,p<0.02) and between age and area (r=−0.284, p<0.05)(Table 3).

mRNA level of BER pathway genes and Ki-67

Quantitative real-time PCRwas used for expressional analysisof selected BER pathway genes, i.e. OGG1, APEX1 andXRCC1 along with the proliferation marker Ki-67. Fifty tu-mour samples of HNSCC along with controls were used forreal-time PCR of respective genes. Among selected fourgenes, OGG1 expression was significantly downregulated(p<0.04) in HNSCC samples than in normal tissue samples,

and a significant upregulation (p<0.02) of APEX1 was ob-served in HNSCC compared to normal tissue samples. Whilein case of XRCC1 and Ki-67, significant downregulation(p<0.01) and upregulation (p<0.03) were observed inHNSCC compared to normal tissue samples, respectively(Fig. 6).

Correlations between the mRNA level of BER pathway genes(OGG1, APEX1, XRCC1) and proliferation marker (Ki-67)expression

When the gene–gene relationship was explored, we observeda positive significant Spearman correlation between XRCC1vs OGG1 (r=0.55**, p<0.0001) and a negative significantcorrelation between XRCC1 vs Ki-67 (r=−0.377**, p<0.007)in HNC cases. A negative non-significant correlation wasobserved between APEX1 vs XRCC1 (r=−0.133, p=0.36),

Fig. 4 Immunohistochemistryanalysis of Ki-67 expression inHNSCC tissue samples.Immunohistopathologicalanalysis was performed toexamine the Ki-67 expression ona negative control, b positivecontrol, c well-differentiatedprimary, d moderatelydifferentiated, e poorlydifferentiated or undifferentiatedtumour and f lymph nodemetastasis tissue samples. Fornegative control in place ofprimary antibody, phosphatebuffer saline (PBS) was used, andfor positive control, tonsil tissuewas used for Ki-67.Representative images a, b, c, d, e(×20) and f (×40) were shown

Tumor Biol. (2014) 35:5971–5983 5977

APEX1 vs OGG1 (r=−0.029, p=0.84) and OGG1 vs Ki-67(r=−0.192, p=0.18). A positive correlation was observedbetween APEX1 vs Ki-67 (r=0.089, p=0.35) (Table 4).

Discussion

In this study, we have examined the expression profile ofOGG1, APEX1 and XRCC1 genes, which are involved inthe repair of damage from DNA through the base excisionrepair pathway and then correlated the protein level of thesegenes with the expression profile of the proliferation markerKi-67. Our study reports that the OGG1 protein level isdownregulated in HNSCC tumour samples. The OGG1 pro-tein level is significantly downregulated in poorly differenti-ated tumours compared to well- or moderately differentiatedtumours. Many earlier studies have reported similar results indifferent cancers [26–29].While contradictory results ofOGG1 expression in different cancers have also been ob-served in different studies [30–32]. In the current study, local-ization of OGG1 protein is cytoplasmic as well as nuclearwhich is conflicting with several previous reports in whichOGG1 immunohistochemical expression was assessed [29,30] although the disease in question was different. The

cause(s) of significantly reduced expression of OGG1 inHNSCC is unknown. However, mutations observed in theOGG1 gene [19, 20] and reduced expression in the OGG1mRNA level [23] have been shown to be associated withimpaired BER activity, ultimately leading to accumulation ofmore 8-oxoG in HNSCC. This accumulation of 8-oxoG mayincrease the mutational load and help to derive tumour pro-gression [28].

The expression profile of APEX1 in this study by IHCanalysis, demonstrated upregulation of APEX1 in HNSCCsamples. A significant increase in the protein level was ob-served in poorly differentiated tumours compared to well- ormoderately differentiated tumours. This upregulation ofAPEX1 has been consistently observed in different types ofcancers [33–39]. Both nuclear and cytoplasmic staining ofHNSCC tumour tissue were observed and ranged from strongto weak staining in the current study. Similar trend in theAPEX1 protein level has also been observed in differentcancers such as cervical cancer [40], epithelial ovarian cancer[41] and sinonasal SCC [42]. Previous studies have observedthat nuclear and cytoplasmic expression of APEX1 may berelated to tumour invasiveness, prognosis and malignant be-haviour of different cancers [33, 42]. The cause(s) of signifi-cantly higher protein level of APEX1 in HNC is unknown.

Fig 5 Changed expressionallevels of BER pathway gene(OGG1, APEX1 and XRCC1)and proliferation marker (Ki-67)in HNSCC tissue samples

5978 Tumor Biol. (2014) 35:5971–5983

However, several studies have reported that SNPs/mutationsin APEX1 are an important factor found to be linked withchange in protein levels and increased susceptibility to carci-nogenesis [43, 44]. Recent in vitro studies have reported thatgenetic changes in APEX1 translate into change in APEX1protein levels and DNA repair capabilities in different cancers[33, 41]. In addition to current data, we have previouslyreported that APEX1 mutations and overexpression of

APEX1 mRNA levels might be associated with increased riskof HNSCC in Pakistani population [21].

We also observed downregulation of the XRCC1 proteinlevel in HNSCC. This downregulation was higher in poorlydifferentiated tumours compared to well- or moderately differ-entiated tumours. Similar observation have been made in lungcarcinoma [45], in gastric carcinoma [46, 47] and in medullo-blastoma [48], whereas conflicting results for XRCC1 protein

Table 2 Protein expression of BER pathway genes and Ki-67 in HNSCC†

Variables OGG1 APEX1 XRCC1 Ki-67

Weak Moderate Strong Weak Moderate Strong Weak Moderate Strong Weak Moderate Strong

C-stage

I-II 09(64) 03(22) 02(14) 03(21) 06(43) 05(36) 06(43) 05(36) 03(21) 03(21) 05(36) 06(43)

III-IV 13(36) 15(42) 08(22) 06(16) 12(24) 18(50) 18(50) 11(31) 07(90) 08(22) 10(28) 18(50)

pT-stage

T1-T2 08(50) 04(25) 04(25) 02(12.5) 10(62.5) 04(25) 06(37.5) 06(37.5) 04(25) 05(31) 03(19) 08(50)

T3-T4 14(41) 14(41) 06(18) 07(21) 09(26) 18(53) 18(53) 10(29) 06(18) 06(18) 12(35) 16(47)

pN-stage

N0 02(13) 08(50) 06(38) 02(13) 04(25) 10(62) 05(31) 04(25) 07(44) 03(19) 04(25) 09(56)

N1-N2 20(59) 10(29) 04(12) 07(21) 15(44) 12(35) 19(56) 12(35) 03(9) 08(23) 11(33) 15(44)

Grade

Poor-mod 12(75) 02(12.5) 02(12.5) 03(19) 08(50) 05(31 06(38) 05(31) 05(31) 04(25) 06(37.5) 06(37.5)

Well 10(29) 16(47) 08(24) 06(18) 11(32) 17(50) 18(53) 11(32) 05(15) 06(19) 09(26) 18(53)

Survival

Disease free 11(35) 12(39) 08(26) 08(35) 12(39) 11(26) 23(35) 05(39) 03(26) 03(10) 07(23) 21(67)

Dead 11(58) 06(32) 02(10) 01(5) 06(32) 12(63) 01(5) 11(58) 07(37) 07(37) 08(42) 04(21)

†The expression levels of selected BER pathway genes and Ki-67 levels were measured by immunohistochemistry (IHC) on study cohort. The caseswere categorized based on IHC staining as weak, moderate and strong

Table 3 Correlations among BER pathway genes (OGG1, APEX1 and XRCC1), proliferation marker (Ki-67) expression and clinicopathologicalparameters of the HNSCC cases†

Gender Age pT stage C stage pN stage Grade Area OGG1 APEX1 XRCC1 Ki-67

Gender -0.194 0.017 0.124 0.038 -0.008 0.319* -0.316* 0.410** -0.263 0.170

Age -0.366** -0.362** -0.366** -0.242 -0.284* -0.167 -0.163 0.016 0.058

pT stage 0.942** 0.987** 0.291* 0.162 0.052 0.163 0.020 -0.096

C stage 0.929** 0.289* 0.180 0.003 0.120 -0.026 - 0.093

pN stage 0.251 0.143 0.094 0.183 0.14 -0.062

Grade 0.123 0.003 0.064 -0.032 -0.006

Area -0.322* -0.279 0.374** -0.044

OGG1 -0.30* 0.330* -0.377**

APEX1 -0.435** 0.071

XRCC1 -0.069

Ki-67

† Spearman correlation coefficients

The expression levels of OGG1, APEX1, XRCC1 and Ki-67 genes for HNSCC cases were determined based on IHC analysis

*p<0.05. The p values were computed using one-way ANOVA and χ2 test

Tumor Biol. (2014) 35:5971–5983 5979

level have been reported in some other studies [49–53]. In thecurrent study, localization of XRCC1 protein was cytoplasmicas well as nuclear, similar to staining localization observed inlung cancer [45] which is contradictory with several previousstudies in which XRCC1 immunohistochemical expression hasbeen assessed specifically nuclear [48, 51, 52, 54]. In previousstudies, we reported that combination of silent mutation(Pro206Pro, Gln632Gln) and missense mutation (Tyr576Asnand Arg399Gln) in XRCC1 gene [22], decreased XRCC1mRNA level [23] and decreased XRCC1 protein level mayinfluence the HNC in Pakistani population, but the exact mech-anism remains unknown. Nevertheless, many studies havereported that nuclear and cytoplasmic expression of BER genesmay be related to tumour invasiveness, prognosis and malig-nant behaviour of different cancers [33, 42].

In the present study, we also examined the expressionprofile of Ki-67, tumour proliferation marker, in 50 HNSCCtumour samples by IHC. Upregulation of Ki-67 was observedin HNSCC tumour samples compared to positive controls.

This upregulation was higher in poorly differentiated tumoursamples compared to well- or moderately differentiated tu-mour samples. Upregulation of Ki-67 has been consistentlyobserved in different types of cancers [55–59]. Nuclear stain-ing of the malignant epithelium was seen and ranged fromstrong to weak staining intensity in this study. Similar stainingtrend of Ki-67 has been reported in TSCC [56], in non-Hodgkin’s lymphoma [57], in breast carcinoma [58] and inmyoepithelial carcinoma [60]. The main reason for the higherprotein level of Ki-67 in HNSCC is unknown. However,previous studies have reported that Ki-67 upregulation isrelated to excessive proliferation, recurrence, metastasis andmore aggressiveness of tumours in different cancers [23, 60].

Our data also demonstrated that the expression of OGG1 iscorrelated with XRCC1 and inversely correlated with APEX1and Ki-67. APEX1 is correlated with Ki-67 and is inverselycorrelated with XRCC1. An apparent inverse correlation be-tween XRCC1 and Ki-67 was also observed, but this correla-tion is not statistically significant. Similar trends in correlation

Fig. 6 mRNA expressions of BER pathway genes (OGG1, APEX1 and XRCC1) and proliferation marker (Ki-67) in HNSCC tissue samples comparedwith controls

Table 4 Correlations between BER pathway genes (OGG1, APEX1 and XRCC1), proliferation marker (Ki-67) expression and clinicopathologicalcharacteristics of primary HNC†

Age pT stage C stage pN stage Grade Area Survival OGG1 APEX1 XRCC1 Ki-67

Age -0.366** -0.362** -0.366** -0.242 -0.284* -0.191 -0.01 -0.204 0.035 0.112

pT stage 0.942** 0.987** 0.291* 0.162 0.064 0.066 0.311* 0.031 -0.092

C stage 0.929** 0.289* 0.180 0.126 0.059 0.287* -0.007 -0.027

pN stage 0.251 0.143 0.076 0.074 0.304* 0.019 -0.083

Grade 0.123 0.106 0.045 0.142 -0.172 -0.015

Survival 0.024 -0.112 0.041 -0.188 0.062

OGG1 -0.045 -0.113 -0.057 0.031

APEX1 -0.133 0.554*** -0.192

XRCC1 -0.029 0.089

Ki-67 0.377**

† Spearman correlation coefficients.

The expression levels of OGG1, APEX1, XRCC1 and Ki-67 for patient cohort were based on the relative mRNA level

*p<0.05. The p values were computed using one-way ANOVA and χ2 test

5980 Tumor Biol. (2014) 35:5971–5983

between BER pathway gene and Ki-67 at mRNA level havealready been reported in HNC in Pakistani population [23].

To confirm our observation and further elucidate the role ofBER pathway genes (OGG1, APEX1 and XRCC1) and tu-mour marker (Ki-67) in HNC, the expression of these geneswas observed by quantitative real-time PCR in the same studycohort. Our study reported that the OGG1 mRNA level isdownregulated in HNSCC samples compared to normal con-trol tissue samples. In the current study, reduced expression ofOGG1 is likely to exhibit a reduced DNA repair activity,ultimately resulting in accumulation of more 8-oxoG inHNC. These observations are in agreement with previousstudies showing that downregulation in OGG1 protein andmRNA expression was correlated with a manifold increase in8-OHdG levels [26, 61]. The second important gene of theBER pathway, APEX1, is upregulated in HNSCC tumoursamples compared to normal control tissue samples in thepresent study. Similar results have also been observed earlierin pancreatic carcinoma [62] and ovarian, gastro-esophagealand pancreatico-biliary carcinomas [63]. Earlier studies havedemonstrated that different oxidative agents promote an in-crease of APEX1mRNA and protein expression and ultimate-ly result in carcinogenesis [34, 64]. The third gene of the BERpathway, XRCC1, mRNA level is downregulated in HNCtumour tissue samples compared to control tissue samples inthe present study. Various studies have been reported that thelow expressional level of this gene leads to impaired BER andSSB repair, greater mutational load and therefore increasedcancer risk [47, 65]. In case of proliferation marker Ki-67,significantly higher mRNA expression was observed in HNCtumours than in control tissue samples in the present study.Many earlier studies have reported similar overexpression ofKi-67 in different cancers [66–68]. Upregulation of Ki-67might be associated with growth arrest or growth stimulationand deregulation of apoptotic or proliferative process [69].Since uncontrolled proliferation is a common feature of tu-mour cells, therapeutic agents that target Ki-67 may be usefultools in cancer treatment. In conclusion, increased Ki-67expression is associated with pathologic and clinical variablesindicative of aggressive disease. For gene to gene interaction,we observed a statistically significant positive correlationbetween OGG1 and XRCC1 mRNA expression. A signifi-cantly negative correlation was also observed betweenXRCC1 and Ki-67 mRNA expression. APEX1 is negativelycorrelated with OGG1 and XRCC1 and positively correlatedwith Ki-67 mRNA expression, but this correlation is statisti-cally non-significant.

In summary, our results confirmed that downregulation ofBER pathway genes, such as OGG1 and XRCC1, and over-expression of APEX1 are associated with enhanced prolifer-ation (as measured by the proliferationmarker Ki-67) and maycontribute to the initiation and progression of HNSCC inPakistani population. This deregulation of BER pathway

genes and enhanced proliferation are more pronounced inpoorly differentiated HNSCC cases. Taken together, our find-ings suggested that BER pathway genes may be useful as abiomarker for the molecular classification and prognosis ofHNSCC. Additional studies will be needed to fully explorethe molecular mechanisms that contribute to the deregulationof BER pathway genes and Ki-67 in HNSCC, and theirpotential as a biomarker and a novel therapeutic target forHNSCC patients.

Acknowledgments This study was supported by grants from theHigher Education Commission (HEC) of Pakistan, the COMSATS Insti-tute of Information Technology (CIIT), Islamabad, and the NIH PHSgrant (CA139596).

Conflicts of interests None

References

1. Spitz M, Wei Q, Dong Q, Amos CI, Wu X. Genetic susceptibility tolung cancer: the role of DNA damage and repair. Cancer EpidemiolBiomarkers Prev. 2003;12:689–98.

2. Tudek B. Base excision repair modulation as a risk factor for humancancers. Mol Asp Med. 2007;28:258–75.

3. Mitra S, Izumi T, Boldogh I, Bhakat KK, Hill JW, Hazra TK.Choreography of oxidative damage repair in mammalian genomes.Free Radic Biol Med. 2002;33:15–28.

4. Kohno T, Shinmura K, Tosaka M, Tani M, Kim SR, Sugimura H,et al. Genetic polymorphisms and altered splicing of the hOGG1gene, that is involved in the repair of 8-hydroxyguanine in damagedDNA. Oncogene. 1998;16:3219–25.

5. Campalans A, Marsin S, Nakabeppu Y, O'connor TR, Boiteux S,Radicella JP. XRCC1 interactions with multiple DNA glycosylases: amodel for its recruitment to base excision repair. DNA Repair.2005;4:826–35.

6. Fan CY, Liu KL, Huang HY, Barnes EL, Swalsky PA, Bakker A,et al. XRCC1 co-localizes and physically interacts with PCNA.Nucleic Acids Res. 2004;32:2193–201.

7. Kumar A, Pant MC, Singh SH, Khandelwal S. Reduced expression ofDNA repair genes (XRCC1, XPD, and OGG1) in squamous cell carci-noma of head and neck in North India. Tumor Biol. 2012;33(1):111–9.

8. Fortini P, Pascucci B, Parlanti E, D’Errico M, Simonelli V, DogliottiE. The base excision repair: mechanisms and its relevance for cancersusceptibility. Biochimie. 2003;85:1053–71.

9. Parsons JL, Dianova I, Dianov GL. APE1 is the major 3’-phosphoglycolate activity in human cell extracts. Nucleic AcidsRes. 2004;32:3531–6.

10. Vidal AE, Boiteux S, Hickson ID, Radicella JP. XRCC1 coordinatesthe initial and late stages of DNA abasic site repair through protein–protein interactions. EMBO J. 2001;20:6530–9.

11. SossouM, Flohr-Beckhaus C, Schulz I, Daboussi F, Epe B, RadicellaJP. APE1 overexpression in XRCC1-deficient cells complements thedefective repair of oxidative single strand breaks but increases geno-mic instability. Nucleic Acids Res. 2005;33(1):298–306.

12. AshrafMJ, MaghbulM, Azarpira N, Khademi B. Expression of Ki67and P53 in primary squamous cell carcinoma of the larynx. Indian JPathol Microbiol. 2010;53:661–5.

13. Tang QL, Fan S, Li HG, Chen WL, Shen XM, Yuan XP, et al.Expression of Cyr61 in primary salivary adenoid cystic carcinomaand its relation to Ki-67 and prognosis. Oral Oncol. 2011;47:365–70.

Tumor Biol. (2014) 35:5971–5983 5981

14. Okabe M, Inagaki H, Murase T, Inoue M, Nagai N, Eimoto T.Prognostic significance of p27 and Ki-67 expression inmucoepidermoid carcinoma of the intraoral minor salivary gland.Mod Pathol. 2001;14:1008–14.

15. Schmilovitz-Weiss H, Tobar A, Halpern M, Levy I, Shabtai E, Ben-Ari Z. Tissue expression of squamous cellular carcinoma antigen andKi67 in hepatocellular carcinoma-correlation with prognosis: a his-torical prospective study. Diagn Pathol. 2011;6:121.

16. Faratian D, Munro A, Twelves C, Bartlett JM. Membranous andcytoplasmic staining of Ki67 is associated with HER2 and ER statusin invasive breast carcinoma. Histopathology. 2009;54:254–7.

17. Han B, Lin S, Yu LJ, Wang RZ, Wang YY. Correlation of 18F-FDGPET activity with expressions of survivin, Ki67, and CD34 in non-small-cell lung cancer. Nucl Med Commun. 2009;30:831–7.

18. Fischer CA, Jung M, Zlobec I, Green E, Storck C, Tornillo L, et al.Co-overexpression of p21 and Ki-67 in head and neck squamous cellcarcinoma relative to a significantly poor prognosis. Head Neck.2011;33:267–73.

19. Mahjabeen I, Baig RM, Masood N, Sabir M, Malik FA, Kayani MA.OGG1 Gene sequence variation in head and neck cancer patients inPakistan. Asian Pac J Cancer Prev. 2011;12:2779–83.

20. Mahjabeen I, Baig RM, Masood N, Sabir M, Inayat U, Malik FA,et al. Novel mutations of OGG1 base excision repair pathway gene inlaryngeal cancer patients. Fam Cancer. 2012;11(4):587–93.

21. Mahjabeen I, Baig RM, Sabir M, Kayani MA. Genetic and expres-sional variations of APEX1 are associated with increased risk of headand neck cancer. Mutagenesis. 2013;28(2):213–8.

22. Mahjabeen I, Baig RM, Masood N, Sabir M, Inayat U, Malik FA,et al. Genetic variations in XRCC1 gene in sporadic head and neckcancer (HNC) patients. Pathol Oncol Res. 2012;19(20):183–8.

23. Mahjabeen I, Chen Z, Zhou X, Kayani MA. Decreased mRNAexpression levels of base excision repair (BER) pathway genes isassociated with enhanced Ki-67 expression in HNSCC. Med Oncol.2012;29(5):3620–5.

24. Wang YX, Sun YE, Li XH, Wang ZB, Tong XY, Liu YL.Comparative study on molecular staging of lymph nodes in non-small cell lung cancer patients. Ai Zheng. 2009;28(3):318–22.

25. Al-Moundhri MS, Nirmala V, Al-Hadabi I, Al-Mawaly K, Burney I,Al-Nabhani M, et al. The prognostic significance of p53, p27 kip1,p21 waf1, HER-2/neu, and Ki67 proteins expression in gastric can-cer: a clinicopathological and immunohistochemical study of 121Arab patients. J Surg Oncol. 2005;91(4):243–52.

26. Mambo E, Chatterjee A, Souza-Pinto NC, Mayard S, Hogue BA,Hoque M, et al. Oxidized guanine lesions and hOgg1 activity in lungcancer. Oncogene. 2005;24:4496–508.

27. Kunisada M, Sakumi K, Tominaga Y, Budiyanto A, Ueda M,Ichihashi M, et al. 8-Oxoguanine formation induced by chronicUVB exposure makes Ogg1 knockout mice susceptible to skincarcinogenesis. Cancer Res. 2005;65:6006–10.

28. Huang XX, Scolyer RA, Abubakar A, Halliday GM. Human 8-oxoguanine-DNA glycosylase-1 is downregulated in human basalcell carcinoma. Mol Genet Metab. 2012;106:127–30.

29. Karihtala P, Kauppila S, Puistola U, Jukkola-Vuorinen A. Absence ofthe DNA repair enzyme human 8-oxoguanine glycosylase is associ-ated with an aggressive breast cancer phenotype. Br J Cancer.2012;106:344–7.

30. Salim EI, Morimura K, Menesi A, El-Lity M, Fukushima S,Wanibuchi H. Elevated oxidative stress and DNA damage and repairlevels in urinary bladder carcinomas associated with schistosomiasis.Int J Cancer. 2008;123:601–8.

31. Ku YP, Jin M, Kim KH, Ahn YJ, Yoon SP, You HJ, et al.Immunolocalization of 8-OHdG and OGG1 in pancreatic islets ofstreptozotocin-induced diabetic rats. Acta Histochem. 2009;111:138–44.

32. Preston TJ, Henderson JT, McCallum GP, Wells PG. Base excisionrepair of reactive oxygen species-initiated 7, 8- dihydro-8-oxo-2′-

deoxyguanosine inhibits the cytotoxicity of platinum anticancerdrugs. Mol Cancer Ther. 2009;8(7):2015–26.

33. Moore DH,Michael H, Tritt R, Parsons SH, KelleyMR. Alternationsin the expression of the DNA repair/ redox enzyme APE/ref-1 inepithelial ovarian cancers. Clin Cancer Res. 2000;6:602–9.

34. Kelley MR, Cheng L, Foster R, Tritt R, Jiang J, Broshears J, et al.Elevated and altered expression of the multifunctional DNA baseexcision repair and redox enzyme Ape1/ref-1 in prostate cancer. ClinCancer Res. 2001;7:824–30.

35. Bobola MS, Finn LS, Ellenbogen RG, Geyer JR, Berger MS, BragaJM, et al. Apurinic/apyrimidinic endonuclease activity isassociatedwith response to radiation and chemotherapy in medullo-blastoma and primitive neuroectodermal tumors. Clin Cancer Res.2005;11(20):7405–14.

36. Yang S, Irani K, Heffron SE, Jurnak F, Meyskens FLJ. Alterations inthe expression of the apurinic/apyrimidinic endonuclease-1/redoxfactor-1 (APE/ref-1) in human melanoma and identification of thetherapeutic potential of resveratrol as an APE/ref-1 inhibitor. MolCancer Ther. 2005;4(12):1923–35.

37. Raffoul JJ, Banerjee S, Singh-Gupta V, Knoll ZV, Fite A,Zhang H, et al. Down-regulation of apurinic/apyrimidinic en-donuclease 1/redox factor-1 expression by soy isoflavonesenhances prostate cancer radiotherapy in vitro and in vivo.Cancer Res. 2007;67(5):2141–9.

38. Jiang L, Liu X, Chen Z, Jin Y, Heidbreder CE, Kolokythas A, et al.MicroRNA-7 targets insulin-like growth factor 1 receptor (IGF1R) intongue squamous cell carcinoma cells. Biochem J. 2010;432:199–205.

39. Fishel ML, Jiang Y, Rajeshkumar NV, Scandura G, Sinn AL, He Y,et al. Impact of APE1/ref-1 redox inhibition on pancreatic tumorgrowth. Mol Cancer Ther. 2011;10(9):1698–708.

40. Schindl M, Oberhuber G, Pichlbauer EG, Obermair A, Birner P,Kelley MR. DNA repair-redox enzyme apurinic endonuclease incervical cancer: evaluation of redox control of HIF-1α and prognos-tic significance. Int J Oncol. 2001;19:799–802.

41. Freitas S, Moore DH, Michael H, Kelley MR. Studies of apurinic/apyrimidinic endonuclease/ref-1 expression in epithelial ovarian can-cer: correlations with tumor progression and platinum resistance. ClinCancer Res. 2003;9:4689–94.

42. Lee JW, Jin J, Rha KS, Kim YM. Expression pattern of apurinic/apyrimidinic endonuclease in sinonasal squamous cell carcinoma.Otolaryngol Head Neck Surg. 2012;147(4):788–95.

43. Ford BN, Ruttan CC, Kyle VL, Brackley ME, Glickman BW.Identification of single nucleotide polymorphisms in human DNArepair genes. Carcinogenesis. 2000;21:1977–81.

44. Hu JJ, Smith TR, Miller MS, Mohrenweiser HW, Golden A, CaseLD. Amino acid substitution variants of APE1 and XRCC1 genesassociated with ionizing radiation sensitivity. Carcinogenesis.2001;22:917–22.

45. Liu W, AO L, Cui Z, Zhou Z, Zhou Y, Yuan X, et al. Molecularanalysis of DNA repair gene methylation and protein expressionduring chemical-induced rat lung carcinogenesis. Biochem BiophysRes Commun. 2011;408:595–601.

46. Wang P, Tang JT, Peng YS, Chen XY, Zhang YJ, Fang JY. XRCC1downregulated through promoter hypermethylation is involved inhuman gastric carcinogenesis. J Dig Dis. 2010;11:343–51.

47. Wang S, Wu X, Chen Y, Zhang J, Ding J, Zhou Y, et al. Prognosticand predictive role of JWA andXRCC1 expressions in gastric cancer.Clin Cancer Res. 2012;18(10):2987–96.

48. Chetty C, Dontula R, Gujrati M, Dinh DH, Lakka SS. Blockade ofSOX4 mediated DNA repair by SPARC enhances radioresponse inmedulloblastoma. Cancer Lett. 2012;323:188–98.

49. Sak SC, Harnden P, Johnston CF, Paul AB, Kiltie AE. APE1 andXRCC1 protein expression levels predict cancer-specific survivalfollowing radical radiotherapy in bladder cancer. Clin Cancer Res.2005;11:6205.

5982 Tumor Biol. (2014) 35:5971–5983

50. Cheng X, Lu W, Ye F, Wan X, Xie X. The association of XRCC1gene single nucleotide polymorphisms with response to neoadjuvantchemotherapy in locally advancedcervical carcinoma. J Exp ClinCancer Res. 2009;28:91.

51. Kang CH, Jang BG, Kim DW, Chung DH, Kim YT, Jheon S, et al.The prognostic significance of ERCC1, BRCA1, XRCC1, andbetaIII-tubulin expression in patients with non-small cell lung cancertreated by platinum- and taxane-based neoadjuvant chemotherapyand surgical resection. Lung Cancer. 2010;68:478–83.

52. Ang MK, Patel MR, Yin XY, Sundaram S, Fritchie K, Zhao N, et al.High XRCC1 protein expression is associated with poorer survival inpatients with head and neck squamous cell carcinoma. Clin CancerRes. 2011;17(20):6542–52.

53. Rybarova S, Vecanova J, Hodorova I, Mihalik J, Cizmarikova M,Mojzis J, et al. Association between polymorphisms of XRCC1, p53and MDR1 genes, the expression of their protein products andprognostic significance in human breast cancer. Med Sci Monit.2011;17:BR354–63.

54. Fujimura M, Morita-Fujimura Y, Noshita N, Yoshimoto T, Chan PH.Reduction of the DNA base excision repair protein, XRCC1, maycontribute to DNA fragmentation after cold injury-induced braintrauma in mice. Brain Res. 2000;869:105–11.

55. Rodrigues RB, Motta Rda R, Machado SM, Cambruzzi E, ZettlerEW, Zettler CG, et al. Prognostic value of the immunohistochemistrycorrelation of Ki-67 and p53 in squamous cell carcinomas of thelarynx. Braz J Otorhinolaryngol. 2008;74:855–9.

56. Kim SJ, Shin HJ, Jung K, Baek S, Shin BK, Choi J, et al. Prognosticvalue of carbonic anhydrase IX and Ki-67 expression in squamouscell carcinoma of the tongue. Jpn J Clin Oncol. 2007;37:812–9.

57. Szczuraszek K, Mazur G, Jelen M, Dziegiel P, Surowiak P, Zabel M.Prognostic significance of Ki-67 antigen expression in non-Hodgkin’s lymphomas. Anticancer Res. 2008;28:1113–8.

58. Cheang MCU, Chia SK, Voduc D, Gao D, Leung S, Snider J, et al.Ki67 index, HER2 status, and prognosis of patients with luminal Bbreast cancer. JNCI. 2009;101(10):736–50.

59. Sym S, Hong J, Cho E, LeeW, ChungM, Ha S, Park Y, Park J, Lee J,Shin D (2011) Prognostic impact of immunohistochemical

expression of Ki-67 in patients with advanced gastric cancer whounderwent curative Resection. J Clin Oncol 29

60. Jiang Y, Cheng B, Ge M, Zhang G. The prognostic significance ofp63 and Ki-67 expression in myoepithelial carcinoma. Head NeckOncol. 2012;4:9.

61. Habib SL. Insight into mechanism of oxidative DNA damage inangiomyolipomas from TSC patients. Mol Cancer. 2009;8:13.

62. Jiang Y, Zhou S, Sandusky GE, Kelley MR, Fishel ML.Reduced expression of DNA repair and redox signaling pro-tein APE1/ref-1 impairs human pancreatic cancer cell survival,proliferation, and cell cycle progression. Cancer Investig.2010;228:885–95.

63. Al-Attar A, Gossage L, Fareed KR, Shehata M, MohammedM, Zaitoun AM, et al. Human apurinic/apyrimidinic endonu-clease (APE1) is a prognostic factor in ovarian, gastro-oesophageal and pancreatico-biliary cancers. Br J Cancer.2010;102:704–9.

64. Grosch S, Kaina B. Transcriptional activation of apurinic/apyrimidinic endonuclease (Ape, ref-1) by oxidative stressrequires CREB. Biochem Biophys Res Commun. 1999;261:859–63.

65. Vaezi A, Feldman CH, Niedernhofer LJ. ERCC1 and XRCC1 asbiomarkers for lung and head and neck cancer. Pharmgenomics PersMed. 2011;4:47–63.

66. Krskova L, Kalinova M, Brızova H, Mrhalova M, Sumerauer D,Kodet R. Molecular and immunohistochemical analyses of BCL2,KI-67, and cyclin D1 expression in synovial sarcoma. Cancer GenetCytogenet. 2009;193:1–8.

67. Brouwer-Visser J, Cossio MJ, Chao SK, Huang GS. Effect of IGF2overexpression on the tumorigenicity of human ovarian carcinomacells. Cancer Res. 2012;72:1.

68. Ma J, Li J, Li H, Xiao X, Shen L, Fang L. Downregulation ofpancreatic-duodenal homeobox 1 expression in breast cancer pa-tients: a mechanism of proliferation and apoptosis in cancer. MolMed Rep. 2012;6:983–8.

69. Urruticoechea A, Smith IE, DowsettM. Proliferationmarker Ki-67 inearly breast cancer. J Clin Oncol. 2005;23:7212–20.

Tumor Biol. (2014) 35:5971–5983 5983