Introduction

Lung cancer is the leading cause for cancer-related deaths worldwide [1]. Non-small cell lung cancer (NSCLC) accounts for approximately 85 % of all lung cancers, and it consists of mainly two subtypes: adenocarcinoma and squamous cell carcinoma [2]. Despite recent advances in targeted therapy, the prognosis for this neoplasm remains poor. There are a number of factors known to affect prognosis such as TNM stage, surgical resectability of the tumor, and pathologic features including tumor size, histologic subtype, lymphovascular invasion, and tumor differentiation. Among these, the most precise and important prognostic factor is the TNM stage [3]. However, the prognosis of the patients with the same stage of lung cancer is variable. Therefore, it is of great importance to identify novel biomarkers to predict the prognosis of lung cancer patients.

Numerous studies have evaluated clinicopathologic implications of various molecules in NSCLC and many of these molecules are being investigated as potential therapeutic targets or as biomarkers to identify subgroups of lung carcinoma with different prognoses. The apoptosis pathways are also widely investigated in NSCLC [4]. Positive regulators of apoptosis are often lost or inactivated, while inhibitor proteins of apoptosis are often upregulated in cancer cells. Defects in apoptosis are implicated in both tumorigenesis and drug resistance, and these defects may also be responsible for some cases of chemotherapy failure in lung cancer [5].

Survivin is a member of the inhibitor of apoptosis protein family which is selectively and abundantly expressed in embryonic and fetal tissues, but is undetectable or found at very low levels in normal differentiated tissues [6]. Survivin is overexpressed in most human cancers, including lung, colon, breast, brain, stomach, esophagus, pancreas, liver, and prostate cancers, as well as hematopoietic malignancies; thus, it plays an important role in the development and progression of neoplasm [6–9]. In lung cancer, the expression of survivin significantly increased in the atypical adenomatous hyperplasia lesions to bronchioloalveolar carcinomas [10, 11]. In addition, survivin mRNA levels were higher in advanced tumor stages, thus suggesting that survivin expression tends to be associated with tumor progression in NSCLC [12].

Survivin was initially described as inhibitors of apoptosis by binding and inactivating caspases [13], but recent data have revealed broader and more complex roles of this protein. High expression of survivin in lung cancer is associated with poor clinical outcome, more aggressive clinicopathologic features, and resistance to both conventional chemotherapy and radiation therapy [8, 14, 15]. Survivin has attracted considerable attention through its deregulation by overexpression in various cancer cells, representing an interesting target for cancer therapy including NSCLC [4, 8, 16]. The pharmacological survivin inhibitors LY2181308 and YM155 have demonstrated acceptable toxicity and evidence of therapeutic efficacy as single agents in early phase clinical trials in patients with advanced prostate cancer, large-cell non-Hodgkin’s lymphoma, and NSCLC [17].

Several studies investigated the prognostic value of survivin expression in NSCLC patients, however, data from the previous studies on the clinicopathologic implication of survivin in NSCLC are inconsistent [8, 12, 18]. Most studies analyzed expression of survivin in NSCLC without consideration of histologic subtypes, patient treatment data, or molecular alteration status.

Over the past decade, it has become evident that subsets of NSCLC can be further defined at the molecular level by “driver” mutations that occur in multiple oncogenes, including EGFR, KRAS, and ALK [19]. These discoveries in the molecular pathogenesis of lung cancer have resulted in the successful application of targeted therapeutic strategies. Defects in apoptotic signaling pathways can be induced by deregulated oncogenes in human cancer and it has been known that tyrosine kinase inhibitors induce apoptosis in NSCLC cells. In addition, overexpression of survivin has been found to attenuate antitumor effect of gefitinib in human tumor xenografts and NSCLC cells [20]. A recent study reported that forced expression of survivin attenuated TAE684 anaplastic lymphoma kinase (ALK)-induced apoptosis in EML4-ALK-positive lung cancer cell line [21]. It will therefore be of interest to determine whether increased survivin expression is clinically significant as a negative predictive marker in lung cancer patients with molecular changes. The relationship between survivin expression and molecular alterations in lung cancer has not been well studied.

Taken together, in this study, we aimed to assess the prevalence and clinical significance of survivin expression in NSCLC patients.

Materials and methods

Patients

From May 2003 to December 2009, the archival formalin-fixed, paraffin-embedded tissue blocks of 373 consecutive pulmonary specimens from NSCLC patients who underwent surgical resection at our university hospital were acquired. The tumors consisted of 225 (60.3 %) adenocarcinomas and 148 (39.7 %) squamous cell carcinomas. The patients given diagnoses other than adenocarcinoma or squamous cell carcinoma were excluded due to their small number. None of the patients had received neoadjuvant chemotherapy or irradiation preoperatively. Clinicopathologic information was obtained from medical records and pathology reports. The period of follow-up ranged from 0 to 84 months.

Tissue microarray

Representative core tissue sections (2 mm in diameter) were taken from paraffin blocks and arranged in new tissue microarray (TMA) blocks using a trephine apparatus (SuperBioChips Laboratories, Seoul, Korea). In cases with variable histologic features, the most representative area was selected for TMA construction. Six cores were sampled and included in the TMA block from each patient.

Immunohistochemistry

Four-micrometer sections from the tissue microarray blocks were deparaffinized in xylene and rehydrated in graded alcohol. Immunohistochemistry (IHC) was performed using the automated staining systems, Ventana BenchMark XT and NexES received from Ventana Medical Systems (Tucson, AZ). The primary antibody of survivin was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The slides were examined independently by two observers (PL.S. and J.H.C.) blinded to both clinical and pathologic data. Expression of survivin was quantified based on the extent of staining (by percentage of positive tumor cells: 0–100 %) and the intensity of staining (graded on a scale of 0–3 as follows: 0 = no staining, 1 = weak staining, 2 = moderate staining, 3 = strong staining; Fig. 1). A semiquantitative score was obtained by multiplying the grades of extent and intensity of staining. The median value of all the scores was chosen as the cut-off value for dividing the expression of proteins into high and low [22].

Fig. 1
figure 1

Immunohistochemical expression of survivin in lung adenocarcinoma tissue (a) and squamous cell carcinoma tissue (b)

Full size image

Molecular testing

Polymerase chain reaction and a direct DNA sequencing method were used for detecting both EGFR and KRAS mutations, as previously described [23]. The presence of ALK translocation was evaluated by fluorescence in situ hybridization [24].

Statistical analysis

The analyses were performed using the software package Statistical Package for Social Sciences, version 18.0, for Windows (SPSS, Chicago, IL). The chi-square test or Fisher’s exact test was used to compare the expression of proteins according to each clinicopathologic parameter. Correlations between each protein expression were analyzed using Spearman correlation. Kaplan–Meier analysis was performed for survival curves, and statistical significance was assessed using the log-rank test. To evaluate whether a biomarker is an independent prognostic factor of overall survival, multivariate analysis using the Cox proportional hazard regression model was performed. All p values were based on the two-sided statistical analysis and p value less than 0.05 was considered to indicate statistical significance.

Results

Patient characteristics

The clinicopathologic features of the patients were summarized in Table 1. Patients consisted of 258 (69.2 %) men and 115 (30.8 %) women with a median age of 65 years (range, 21–84), with 136 (36.5 %) nonsmokers and 237 (63.5 %) smokers. Two hundred fifty-one (67.3 %) of 373 patients received adjuvant platinum-based chemotherapy. Pathologic tumor stage of each tumor was determined according to the seventh edition of the TNM classification system. The specimens were classified into 225 cases of adenocarcinoma (60.3 %) and 148 cases of squamous cell carcinoma (39.7 %). The patients with adenocarcinoma were younger than the patients with squamous cell carcinoma (p = 0.001), and the tumor size of adenocarcinoma was smaller than that of squamous cell carcinoma (p = 0.009). Among the squamous cell carcinoma patients, 92.6 % (137/148) had a history of smoking, but only 44.4 % (100/225) of the adenocarcinoma patients were smokers (p < 0.001). There were statistical differences in pleural invasion and lymph node metastasis between the two histological subtypes (p = 0.005 and p = 0.031, respectively).

Table 1 Clinicopathologic characteristics according to the histologic subtypes of NSCLCs
Full size table

Immunohistochemical expression of survivin

In most cases, positive expression of survivin was detected in cytoplasm, and a few cases showed both nucleus and cytoplasm staining (Fig. 1), in these cases, staining of both components was assessed and combined. A total of 52.9 % (119/225) of adenocarcinoma and 54.1 % (80/148) of squamous cell carcinoma were considered as high survivin expression. The expression of survivin was not significantly different between adenocarcinoma and squamous cell carcinoma (p = 0.825). There was a significant correlation between the high expression of survivin and the presence of venous invasion (p = 0.001), lymph node metastasis (p = 0.014) and the presence of recurrence (p = 0.010) in adenocarcinoma. There were no significant correlations between survivin expression and clinicopathologic characteristics of the squamous cell carcinoma patients (Table 2).

Table 2 Immunohistochemistry status in NSCLCs according to clinicopathologic characteristics
Full size table

Correlations between survivin expression and molecular characteristics

To investigate the association of survivin expression and molecular characteristics of lung cancer, we analyzed survivin expression according to the major driver mutations including EGFR and KRAS mutations and ALK translocation in 146 adenocarcinomas. EGFR mutations at exons 18–21 were investigated and 86 (58.9 %) patients harbored mutated EGFR status. A KRAS mutation study of codons 12, 13, and 61 was performed and 7 (4.8 %) patients were found to have KRAS mutation. Fifteen (10.3 %) patients were found to have ALK translocation. The expression of survivin was not significantly different among different molecular subgroups (Table 3).

Table 3 Immunohistochemistry status in ADC according to molecular characteristics
Full size table

Survival analysis

In univariate analysis, we first analyzed the associations between the conventional clinicopathologic factors and patient outcomes in adenocarcinoma and squamous cell carcinoma, separately. In adenocarcinoma, age >65 years, smoking, tumor size >3 cm, the presence of pleura invasion, venous invasion, lymphatic invasion, lymph node metastasis, and more advanced pathologic stage were associated with a significantly worse disease-free survival (DFS) or/and overall survival (OS) (Table 4). In squamous cell carcinoma, clinicopathologic factors including the presence of pleura invasion, venous invasion, lymphatic invasion, and more advanced pathologic stage were associated with a significantly worse DFS or/and OS (Table 4).

Table 4 Survival analysis by Kaplan–Meier estimates in NSCLCs according to clinicopathologic and IHC parameters
Full size table

Then, we analyzed the prognostic significance of survivin. In adenocarcinoma, high expression of survivin was associated with significantly shorter DFS (42.2 vs. 58.8 months; p = 0.001) and shorter OS (60.8 vs. 71.5 months; p = 0.009) (Fig. 2). In squamous cell carcinoma, the expression of survivin was not associated with the prognosis of the patients (Fig. 2). Multivariate Cox analysis confirmed that survivin was an independent poor prognostic factor in adenocarcinoma (DFS: HR, 1.687; 95 % CI, 1.123–2.532; p = 0.012; OS: HR, 1.965; 95 % CI, 1.108–3.486; p = 0.021) (Table 5 and 6).

Fig. 2
figure 2

Kaplan–Meier survival estimate graphs of disease-free survival (DFS) and overall survival (OS) according to the expression of survivin in patients with lung adenocarcinoma (a, b) and squamous cell carcinoma (c, d)

Full size image
Table 5 Multivariable Cox proportional hazard analysis of survival in lung ADCs
Full size table
Table 6 Multivariable Cox proportional hazard analysis of survival in lung SCCs
Full size table

In addition, we analyzed the prognostic significance of survivin within the different molecular subgroups. High expressions of survivin were associated with significantly shorter DFS and/or OS in EGFR-mutated group (DFS; 37.16 vs. 50.20 months; p = 0.027, OS; 60.26 vs. 65.75 months; p = 0.096) (Fig. 3), KRAS-negative group (DFS; 38.98 vs. 50.88 months; p = 0.006, OS; 59.77 vs. 64.39 months; p = 0.048) and ALK-negative group (DFS; 37.61 vs. 48.25 months; p = 0.014, OS; 59.35 vs. 63.73 months; p = 0.090) (Table 7). There was no significant difference between the high and low expressions of survivin in EGFR-negative, KRAS-mutated and ALK-rearranged group (p > 0.05).

Fig. 3
figure 3

Kaplan–Meier survival estimate graphs of disease-free survival (DFS) (a) and overall survival (OS) (b) according to the expression of survivin in EGFR-mutated lung adenocarcinoma

Full size image
Table 7 Survival analysis by Kaplan–Meier estimates in ADCs according to molecular characteristics
Full size table

Discussion

In this study, we evaluated the immunohistochemical expression of survivin in 225 adenocarcinomas and 148 squamous cell carcinomas, separately, and investigated its association with the clinicopathologic parameters, molecular features, and prognosis of the patients. We demonstrated that high expression of survivin was positively associated with clinicopathologic parameters which have been implicated in unfavorable prognosis, and multivariate analysis revealed high expression of survivin was an independent prognostic factor in adenocarcinoma. In squamous cell carcinoma, there was no significant correlation between the expression of survivin and outcome of patients.

Numerous studies have discussed the prognostic value of survivin expression in NSCLC patients, but the results are inconsistent [8, 18]. Some studies have demonstrated the association of survivin with poor prognosis in NSCLC [18, 25] while other studies documented that survivin expression had no prognostic implications in NSCLC patients [12]. Vischioni et al. demonstrated that nuclear expression of survivin predicted longer overall and relapse-free survival in NSCLC patients [26]. A study by Rosato et al. also showed that higher level of survivin mRNA was correlated with a better prognosis in NSCLC [27]. Our results demonstrated that survivin high expression had a poor prognostic value in patients with lung adenocarcinoma but not in those with squamous cell carcinoma. The different prognostic role of survivin in adenocarcinoma and squamous cell carcinoma shows the biological differences between these two major histologic types of NSCLC. Most of the studies analyzed the whole NSCLC without consideration of histologic subtypes and only a few studies performed the analysis in adenocarcinoma and squamous cell carcinoma separately, but the case number is limited in all of these studies [8, 25]. Different histologic subtypes of NSCLC show accumulated evidences of different molecular events during oncogenesis [28]. It is important that these biomarkers should be validated on an independent series of cases according to histologic types. A meta-analysis showed that significant heterogeneity was present in different studies, and histological type might be a major cause of heterogeneity [25]. Recently, specific genetic abnormalities have been identified in NSCLC, which shows distinct features according to the histologic types. Activating mutations of EGFR, KRAS, and ALK are mainly identified in adenocarcinomas, whereas p53 mutation appears to be more frequent in squamous cell carcinoma [29]. Because these oncogenic dysregulation can influence apoptosis signaling pathways, the prognostic role of survivin expression can be different between adenocarcinoma and squamous cell carcinoma.

Research on anticancer targeted therapy has opened a new era of personalized medicine; targeted therapies have improved and will continue to improve the outcome of lung cancer patients. The induction of apoptosis has been considered as the major mechanism for drug-mediated anticancer treatment. However, there is few data regarding the effect of survivin in different molecular subgroup of lung cancer. To the best of our knowledge, this is the first study analyzing the survivin expression according to the major driver mutations in lung cancer. We analyzed the association between survivin expression and molecular characteristics (EGFR, KRAS, and ALK) to examine whether the expression of survivin is different among different molecular subgroups. Although there was no statistically significant difference in the expression of survivin among different molecular subtypes, survival analysis within the different molecular subgroups showed that high expression of survivin was associated with significantly shorter DFS in EGFR-mutated group. Tumors that have higher survivin expression showed worse DFS than those with lower survivin expression in EGFR-mutated group, suggesting that high expression of survivin plays an inhibitory role in the antitumor effect of chemotherapeutic agents including EGFR-tyrosine kinase inhibitors in lung cancer patients with EGFR mutation. A study found that the stable overexpression of survivin attenuated gefitinib-induced apoptosis in NSCLC cells positive for EGFR mutation and also inhibited the antitumor effect of gefitinib in human tumor xenografts [20]. Whether the overexpression of survivin can be a target of the molecular therapy in NSCLC needs to be further clarified.

In conclusion, our results revealed a significant association of survivin expression with various clinicopathologic characteristics, particularly those related with aggressive features of lung adenocarcinoma and survival analysis has shown that survivin is an independent, negative prognostic factor in adenocarcinoma, but not in squamous cell carcinoma. The different prognostic roles played by survivin in adenocarcinoma and squamous cell carcinoma highlights the biological differences between these two histologic types.