Nuclear Survivin and Its Relationship to DNA Damage Repair Genes in Non-Small Cell Lung Cancer Investigated Using Tissue Array

Purpose To investigate the predictive role and association of nuclear survivin and the DNA double-strand breaks repair genes in non-small cell lung cancer (NSCLC): DNA-dependent protein kinase catalytic subunit (DNA-PKcs), Ku heterodimeric regulatory complex 70-KD subunit (Ku70) and ataxia-telangiectasia mutated (ATM). Methods The protein expression of nuclear survivin, DNA-PKcs, Ku70 and ATM were investigated using immunohistochemistry in tumors from 256 patients with surgically resected NSCLC. Furthermore, we analyzed the correlation between the expression of nuclear survivin, DNA-PKcs, Ku70 and ATM. Univariate and multivariate analyses were performed to determine the prognostic factors that inuenced the overall survival and disease-free survival of NSCLC. Results The expression of nuclear survivin, DNA-PKcs, Ku70 and ATM was significantly higher in tumor tissues than in normal tissues. By dichotomizing the specimens as expressing low or high levels of nuclear survivin, nuclear survivin correlated significantly with the pathologic stage (P = 0.009) and lymph node status (P = 0.004). The nuclear survivin levels were an independent prognostic factor for both the overall survival and the disease-free survival in univariate and multivariate analyses. Patients with low Ku70 and DNA-PKcs expression had a greater benefit from radiotherapy than patients with high expression of Ku70 (P = 0.012) and DNA-PKcs (P = 0.02). Nuclear survivin expression positively correlated with DNA-PKcs (P<0.001) and Ku70 expression (P<0.001). Conclusions Nuclear survivin may be a prognostic factor for overall survival in patients with resected stage I-IIIA NSCLC. DNA-PKcs and Ku70 could predict the effect of radiotherapy in patients with NSCLC. Nuclear survivin may also stimulates DNA double-strand breaks repair by its interaction with DNA-PKcs and Ku70.


Introduction
Non-small-cell lung cancer (NSCLC) is the leading cause of cancer-related death worldwide, with an increasing incidence and a poor prognosis. As the clinico-pathologic characteristics can not accurately predict the patient outcome, understanding the biology of NSCLC at the molecular level will help in the development of new and effective treatment modalities and prediction of the prognosis.
Evasion of apoptotic cell death is critical for tumor growth and is reported to be a hallmark of cancer cells [1] that develop resistance to anticancer treatments. Thus, targeting the apoptotic pathways may represent a promising strategy to counteract resistance and sensitize the cancer cells to anticancer modalities, including radiotherapy [2]. Among these anti-apoptotic factors, survivin, a 16.5 KD protein of 142 amino acid residues that is the smallest member of the inhibitor of apoptosis protein (IAP) family, deserves much attention due to its universal over-expression in human tumors, and its prominent role in the regulation of a variety of cellular networks, including those that control tumor cell proliferation and adaption to an unfavorable environment [3]. Survivin exist in distinct subcellular pools including the nucleus, the cytoplasm and the mitochondria [4][5][6]. The subcellular distribution of survivin plays a distinct role in the ability of this molecule to regulate cell division and survival. The localization of survivin in the cytoplasm is considered to be cytoprotective due to the anti-apoptotic activities of this molecule [7]. The nuclear localization of survivin is associated with cell division, as survivin is a subunit of the chromosomal passenger complex (CPC) [8].
In mammalian cells, DNA double-strand breaks (DSBs) are mainly repaired via homologous recombination (HR) or nonhomologous end joining (NHEJ) [9]. A key element of the NHEJ pathway is the DNA-dependent protein kinase (DNA-PK), which consists of a 465-KD catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs), and a heterodimeric regulatory complex, Ku,which includes a 70-KD subunit (Ku70) and an 86-KD subunit (Ku80) [10]. Current models suggest that the heterodimeric Ku proteins rapidly bind to double stranded DNA ends and recruit DNA protein kinase (DNA-PK), generating a DNA-PK holoenzyme complex. The ATM protein has been identified as the principal activator and master controller of the cellular response to DSBs. The ATM kinase seems to be the primary activator and master controller of the cellular response to DNA DSBs and phosphorylates key players of the DNA damage response network, including cell cycle arrest apoptosis and DNA repair [11].
Recently it was shown that survivin is able to provoke both suppression of apoptosis and stimulation of DNA repair. The underlying mechanisms seem to be multifaceted and involve caspase-dependent and caspase-independent pathways. Some reports confirmed a higher incidence of DNA-damage in colorectal-and non-small cell lung cancer cell lines after treatment with survivin siRNA or the transcriptional repressor YM155 based on the detection of phospho-histone c-H2AX detection as a marker of radiation-induced DSBs [12,13]. Co-immunoprecipita-tion analyses revealed an interaction between survivin, Ku70 and DNA-PKcs in nuclear foci [14]. These results suggested that the nuclear accumulation of survivin was linked to DNA-DSB repair.
The aim of this study was to examine the potential correlation and prognostic value of nuclear survivin, DNA-PKcs, Ku70 and ATM using immunohistochemistry. To our knowledge, this is the first study.

Patients
The study was approved by the Ethical Review Committee of Harbin Medical University, Harbin, China. All patients provided written informed consent to participate in the study. This study was retrospective. Surgical resection specimens were obtained from 256 lung cancer patients at stage IA to IIIB between December 2004 and December 2006. No patients received preoperative chemotherapy and radiotherapy. There were 176 males and 80 females with a mean age of 57.7 yrs. The tumors were staged according to the revised version of lung cancer TNM staging published in 1997 [15]. Histopathological diagnosis was

Construction of Tissue Microarrays
The lung cancer tissue microarray (TMAs) was constructed as follows. Briefly, a tissue arraying instrument was used to create holes in a receptive paraffin block and to acquire tissue cores from the donor tissue block using a thin-walled needle with an inner diameter of 2 mm that was held in an X-Y precision guide. The core samples were retrieved from the selected region in the donor and extruded directly into the receptive block at defined array coordinates. A solid steel wire that was closely fit in the tube, was used to transfer the tissue cores into the receptive block. After the construction of the array block, all the tissue blocks were cut with a microtome to 4 mm and affixed to the slide. Blocks from 256 patients were arrayed as triplicate spots of 2 mm diameter on slides.

Immunohistochemistry
The tissue sections were deparaffinized in xylene and rehydrated with graded alcohol concentrations using standard procedures. The sections were subsequently submerged in citrate buffer (pH 6.0) and autoclaved at 121uC for 5 min to retrieve the antigenicity. After washing in phosphate-buffered saline (PBS,0.1 M,PH 7.4,3 times for 5 min), endogenous peroxidase was blocked by incubation in 3% hydrogen peroxide for 15 min at room temperature. Then, the samples were incubated with a survivin antibody (71G4B7,Cell Signaling Technology, Beverly, MA) diluted at 1:400, a DNA-PKcs antibody (MS-423-P0,Neomarkers, Fremont, CA), a Ku70 antibody (EPR4027,Epitomics, CA) and an ATM antibody (Y170,Epitomics, CA) diluted at 1:200 in PBS containing 0.5% BSA overnight at 4uC in a moist chamber and washed with buffer to remove unbound antibodies. The sections were incubated with the biotinylated secondary antibody followed by peroxidase-conjugated streptavidin for 30 min. 3-39diaminobenzidine tetrahydrochloride (Dako,Germany) was added to visualize the reaction. After rinsing in deionized water and counterstaining with commercial hematoxylin, the slides were dehydrated and mounted. Appropriate tissue sections as positive controls for each primary antibody were simultaneously stained to be used as positive controls. Mouse IgG1 or rabbit IgG (Cell Signaling Technology, Beverly, MA) were used as negative controls at the same dilutions as the corresponding primary antibodies.

Immunohistochemical Staining Evaluation
Evaluation of the staining for survivin, DNA-PKcs, Ku70 and ATM expression was conducted with bright-field light microscopy independently by two experienced pathologists independently. Nuclear survivin, DNA-PKcs,Ku70 and ATM protein expression levels were classified semiquantitatively by combining the proportion and intensity of the positively stained tumor cells. The percentage of positively stained tumor cells was scored as follows: 0 (no positive tumor cells), 1 (1-25% positive tumor cells), 2 (26-50% positive tumor cells), 3 (51-75% positive tumor cells), and 4 (76-100% positive tumor cells). Staining intensity was scored as follows: 0(no staining); 1(weak staining); 2(moderate staining) and 3(strong staining). The staining intensity score multiplied by the percentage of positive staining was used to define the expression levels. The mean score among the three tissue sections was entered for statistical analyses. Subsequently, the median value of all scores was used as the cut-off point for the classification of the expression of the 4 proteins, thus lung cancer patients were classified into two groups: low expression and high expression groups. Furthermore, cases with discrepancies were reviewed simultaneously by the original two pathologists and a senior pathologist until a consensus was reached.

Statistical Analysis
The associations between the IHC parameters, as well as the IHC/clinical-pathological parameters were analyzed using the Chi-squared test, the Fisher exact test and the Pearson rank correlation. Overall survival (OS), disease-free survival (DFS) and local recurrence-free survival (LRS) were obtained according to the Kaplan-Meier method. The persistence of the disease and the occurrence of relapses were the end points for DFS, the occurrence of local recurrence was the end point for LRS, and death from any cause was considered an event for the calculation of OS. The comparisons between each prognostic variable were

Correlation with Prognostic Factors
The correlations between patients expressing nuclear survivin, DNA-PKcs, Ku70 and ATM at different levels and their clinicopathological characteristics were presented in Table 1. Nuclear survivin correlated significantly with the pathologic stage and lymph node status. Nuclear survivin expression levels in patients with stage III tumors were markedly higher than those in patients with stage I and stage II tumors (P = 0.009). Nuclear survivin expression in patients with N2 and N3 disease was higher than those in patients with N0 and N1 disease (P = 0.004). There was no statistically significant difference between DNA-PKcs, Ku70 or ATM expression and the clinicopathologic variables.

Expression of Nuclear Survivin and DNA DSB Repair Genes in Relation to Prognosis
The median follow-up time was 64 months. The median OS of the patients was 4765.8 months, and the median DFS time was 2564.5 months. Univariate survival analysis revealed that the tumor histology, differentiation grade, pathologic stage, T status, N status, nuclear survivin expression, DNA-PKcs expression and Ku70 expression significantly influenced the OS and DFS ( Table 2). Multivariate analysis was performed according to the results of the univariate analysis. The results indicated that the tumor differentiation grade, pathologic stage and nuclear survivin expression were the independent prognostic factors ( Table 3). The calculated relative risks of death and recurrence was 1.68 (P = 0.047) and 1.83 (P = 0.041) for patients with high survivin expression, 1.67 (P = 0.001) and 1.59 (P = 0.003) for patients with a high pathologic stage and 0.63 (P = 0.015) and 0.51 (P = 0.002) for patients with low differentiation, respectively.
Kaplan-Meier analysis revealed that patients with nuclear survivin high expression had poorer OS (44.8 vs. 55.7 months; P = 0.001) ( Fig. 2A) and DFS (35.7 vs. 49.7 months; P = 0.001) (Fig. 2B) compared with patients with lower levels of nuclear survivin. Patients with a moderate or strong differentiation grade had a better survival than patients with low differentiation grade (OS, P = 0.007; DFS, P = 0.008). Moreover, higher pathologic stage (OS, P,0.001; DFS, P,0.001) were significantly associated with poor OS and DFS. Patients with high DNA-PKcs or Ku70 expression had a significantly worse OS and DFS than those with lower expression but they were not independent prognostic factors.

Relationship between Nuclear Survivin and DNA Damage Repair Genes
We investigated the possible relationship between nuclear survivin and DNA damage repair genes among the 256 patients. Among the three relevant genes, DNA-PKcs and Ku70 expression were significantly positively related to nuclear survivin expression ( Table 4

Prediction of Treatment Outcome
Postoperative radiotherapy could significantly reduce the risk of local recurrence. In our study, 92 patients received postoperative radiotherapy. Multivariate analysis indicated that the tumor differentiation grade, DNA-PKcs and Ku70 expression were the independent prognostic factors ( Table 5). The calculated relative risks of local recurrence was 2.025(P = 0.007) for patients with high DNA-PKcs expression, 2.161(P = 0.029) for patients with high Ku70 expression and 0.522 (P = 0.008) for patients with low differentiation grade, respectively. In patients with low expression of Ku70 and DNA-PKcs, the local recurrence-free survival (LRS) was significantly higher compared to patients with high expression of Ku70 (P = 0.012; Figure. 3A) and DNA-PKcs (P = 0.02; Figure. 3B). The greatest benefit of radiotherapy was observed in patients whose tumors had low expression of Ku70 and DNA-PKcs.

Discussion
In the current study, we determined the expression of nuclear survivin and 3 key components of the DNA DSB repair pathways in patients with NSCLC using TMAs. Compared with normal lung tissue, the proportion of cells with positively stained nuclei and the staining intensity were higher in tumor tissues than in normal lung tissue. Reduced expression of messenger RNA for DNA-PKcs and ATM in normal lung tissue has also been shown in a previous study [16]. Survivin expression was only observed in tumor samples, not in the surrounding normal tissues, which is in accordance with the results of previous studies [17,18].
The prognostic predictive value of survivin expression and that of nuclear staining versus cytoplasmic staining in several malignancies, including NSCLC, have been investigated in a number of studies with conflicting results. Similarly, the results of studies investigating the prognostic predictive value of the nuclear staining of survivin are also conflicting. These controversies could be explained, in IHC studies, on the basis of using antibodies of different specificities or concentrations, and employing different cut-off points and different approaches for storing and processing tissues. Our study showed that the nuclear expression of survivin is an independent negative prognostic factor for survival in surgically resected NSCLC patients. Consistent with our results, other retrospective series have identified nuclear rather than cytoplasmic survivin expression as an independent negative prognostic factor for survival in NSCLC [19][20][21][22]. Even in NSCLC, nuclear survivin expression was shown to be correlated with a better outcome in a recent series of advanced (stage IIIA-IV) inoperable patients who underwent chemotherapy [17]. However, a meta-analysis of survivin conducted by Jiang Fan [23] using individual patient data found conflicting results. Data from three studies were combined to find that the level of nuclear survivin did not have impact on OS of NSCLC patients (RR1.58, 95% CI 0.87-2.85, P = 0.13). But the studies only included three studies and the number of the patients were small, the negative result may be due to the small number of patients.
In our study, nuclear survivin expression positively correlated with the pathologic stage and the lymph node status. The current results are consistent with those that reported that survivin was a marker of lymph node metastasis [18,24,25]. The link between survivin expression and the potential for lymph node metastasis could be explained as follows. Survivin could be an apoptosis inhibitor: the proportion of cancerous cells in a tissue, which would otherwise be removed by apoptosis, increase with continued growth and increased potential for invasion and metastasis.
In this study, we observed that high expression of DNA-PKcs and Ku70 was associated with poor overall survival in patients with NSCLC but it did not reach statistical significance. DNA double-strand breaks (DSBs) are regarded as the most lethal of all DNA lesions. Prompt and efficient repair of DNA DSBs is critical for maintaining genomic integrity. Defects in the DSB repair pathway may cause genetic alterations, chromosome instability, and, ultimately, malignant transformation. Conversely, tumor cells with increased DNA DSB repair capacity would be more likely to survive and proliferate, leading to the poor prognosis in cancer patients. The prognostic value of these genes were investigated and the results were conflicting. In the study conducted by Jinliang et al in which RT-PCR was used to measure DNA-PKcs and ATM expression in lung cancer, patients with high tumor:normal (T/N) expression ratios of ATM or DNA-PKcs had a notably shorter median survival than patients with low ratios [16]. However, there are also some inconsistent results in terms of the prognostic value of these genes. In a comprehensive review, Wynand reported that cells that exhibit a defect in their DNA repair pathways are liable to suffer from DNA damage-induced apoptosis [26]. For example, the survival of patients with high expression of DNA-PKcs or Ku86 was significantly better when compared with the survival of patients with low expression of DNA-PKcs or Ku86 [27]. These inconsistencies may be the result of many factors, including different cancer types, different histology and tumor characteristics, or methodological differences, such as the technique used to probe gene expression or the antibody reacting against a specific epitope.
The molecular mechanism underlying the differential localization in tumor cells is still unclear. The subcellular distribution of survivin plays a distinct role in the ability of the molecule to regulate cell division. Recent data further indicate that the interaction of the leucine-rich nuclear export signal (NES) with the nuclear export receptor chromosome region maintenance protein 1 homolog (Crm1) is critically involved in survivin's intracellular localization and cancer-related functions [28,29]. DNA damage can stimulate a rapid discharge of the mitochondrial pool of survivin into the cytosol, which preserves the viability of tumor cells during a protracted G2 block by antagonizing DNA damage-induced apoptosis [30]. Using Pearson correlation coefficient analysis, the correlation between the expression levels of nuclear survivin and those of DNA-PKcs or Ku70 was observed in our study. Survivin is a member of the inhibitor of apoptosis (IAP) family, and it directly inhibits the activity of caspases, notably caspase-3. Recently, it was shown that survivin also stimulates DSB repair by its interaction with DNA-PK [14,31]. Therefore, we speculate that DNA damage stimulates a rapid discharge of survivin to the nucleus, which physically interacts with factors of the DNA DSB repair proteins DNA-PKcs and Ku70. Thus survivin stimulates DNA double-strand repair capacities by regulating DNA-PKcs and Ku70 activity and, thereby inhibit DNA damage-induced apoptosis. However, more studies are required to clarify this open question.
The novel finding in our study is that the protein expression levels of DSB repair proteins can be of value for the prediction of treatment outcome. Patients with low Ku70 and DNA-PKcs expression had a greater benefit from radiotherapy than patients with high expression of Ku70 and DNA-PKcs because the latter group had a high rate of recurrence after treatment with radiotherapy. Our interpretation is that Ku70 and DNA-PKcs are necessary for DNA damage recognition and signaling and that loss of these proteins would lead to impaired DNA repair, increased radiosensitivity, and improved local tumor control. Karin et al. reported that low expression of Ku70/Ku80 predicted a good effect of radiotherapy in early breast cancer [32]. Beskow also found that radioresistant cervical cancer showed an increased frequency of DNA-PKcs, Ku70 and Ku86 positive cells [33].  Bouchaert et al. reported that positive DNA-PKcs nuclear staining was closely associated with biochemical recurrence and suggested that DNA-PKcs could be a predictive marker of recurrence after radiotherapy in prostate cancer [34]. However, there are also results that are inconsistent with our findings [27,32,35]. It remains controversial whether DNA-PKcs or Ku70 expression could be a prognostic factor for the response of patients with lung cancer to radiotherapy.
In conclusion, we observed that nuclear survivin was significantly associated with poor survival in patients with NSCLC and that nuclear survivin was strongly related to pathologic stage and lymph node metastasis. Nuclear survivin was positively correlated with DNA-PKcs and Ku70 and thus may stimulate DNA doublestrand breaks repair by its interaction with DNA-PKcs and Ku70.
The current results indicate that nuclear survivin may be useful for predicting the clinical outcomes of patients with NSCLC.