Survivin expression and its clinical significance in pancreatic cancer

Survivin has recently been identified as a novel inhibitor of apoptosis (IAP). It blocks common downstream elements of both the mitochondrial pathway and the death receptor pathway, by directly inhibiting terminal effector caspase-3, caspase-7, and caspase-9 activity[1, 2]. Thus, it inhibits apoptosis pathway differently from bcl-2, which blocks mitochondrial cytochrome c release into the cytosol, resulting in the inhibition of mitochondrial apoptotic pathway. Survivin is expressed during embryonic and fetal development but is undetectable in terminally differentiated normal adult tissue. However, it is re-expressed in transformed cell lines and several human cancer cells at a frequency of 34–100 % [3, 4]. As a prognostic factor, survivin expression is significantly associated with poor clinical outcome in cancers, such as neuroblastoma, colorectal cancer, breast cancer, lung cancer and esophageal cancer [5‐10].

Survivin expression in pancreatic cancer has not been widely studied. Satoh et al reported that survivin was expressed in malignant portion of intraductal papillary-mucinous tumor (IPMT) but not in chronic pancreatitis, and thus proposed that survivin expression may be upregulated at an early stage of tumorigenesis by reducing cancer cell apoptosis [11]. In another study, Asanuma et al reported that survivin acts as an inducible radioresistance factor in pancreatic cancer cell line [12].

In the present study, we investigated survivin expression in pancreatic cancer and its association with clinical outcome.

Survivin, an IAP, has been studied as a prognostic marker in various cancers. Adida et al reported that survivin expression in neuroblastoma correlated with unfavorable histology, aggressive and disseminated disease [5]. In colorectal cancer and breast cancer, patients with survivin-positive tumors had a decreased apoptotic index and worse survival rates than those with survivin-negative tumors [7, 8]. In addition, survivin expression correlated with poor prognosis in esophageal cancer and non-small cell lung cancer [9, 10].

The clinical significance of survivin expression in pancreatic cancer is not well understood. Previous studies reported that survivin was expressed at an incidence of 77%–88% [11, 13]. In our study, survivin was expressed in 93.9%, and was highly expressed in most cases (59.2%). In pathological parameters, neural invasion was more common in survivin positive group. Neural invasion is a specific metastatic route in pancreatic cancer and perineural invasion is known to be a prognostic marker for the recurrence of pancreatic cancer after surgery [14]. However, venous invasion was more common in survivin negative group. Venous invasion in pancreatic cancer can be a risk factor for liver metastasis [15]. We did not explain this discrepancy, but we suggested that venous invasion may be an independent prognostic factor irrelevant to survivin expression. Recent study reported that survivin expression has different clinical significance according to expression pattern [16]. Although we did not distinguish the expression pattern, it may be another explanation for this discrepancy. It is important to predict recurrence or metastatic pattern after curative surgery because of determining adequate adjuvant treatment modality in pancreatic cancer, so further study may be warranted to investigate the association between survivin and pathological parameters.

We did not find any other association between survivin expression and clinical course due to the limitation of patients enrolled in this study. Most patients in our study were at an early stage, making it difficult to compare survivin expression according to various stage. Secondly, the vast majority of pancreatic cancer tissue in the present study expressed survivin, meaning survivin-negative group was very small. For the same reason, we could not determine that survivin expression was a prognosis factor. Recently, Kami et al reported that survivin expression may be a prognostic factor in pancreatic cancer. In that study, 63.8% of patients showed survivin expression, resulting in a good proportion of survivin-negative patients for comparison [17]. We also investigated whether survivin expression was associated with poor clinical outcome, including recurrence and survival, but no significant correlation was observed. However, It is not determined that survivin may not be associated with poor clinical outcome because of limitation of our study. It may be that further study involving great number of patients is required in order to determine the clinical significance of survivin expression in pancreatic cancer.

It is not certain whether survivin expression is a predictive marker for anticancer therapy such as radiotherapy or chemotherapy in pancreatic cancer. In a study using a pancreatic cancer cell line, Asanuma et al reported that survivin may be a radioresistant factor [12], and Kato et al suggested that survivin expression may be a predictive marker for chemotherapy as well as prognostic marker in esophageal cancer in their previous study [9]. We examined the survivin expression in fourteen patients treated with chemotherapy to evaluate the potential of survivin as a predictive marker for chemotherapy. Although the findings were statistically not significant, there was a suggestion that non-responder group showed high survivin expression. On the basis of this finding, we hypothesize that survivin is a potential marker for predicting responsiveness to chemotherapy.

The p53 mutation is frequently seen in pancreatic cancer, and its expression correlates with overall survival after pancreatectomy [18]. Previous study found survivin and p53 may play important roles in the transition from adenoma to carcinoma in situ in IPMT, and that survivin expression was higher in p53-positive pancreatic cancers [18, 19]. In the present study, while 34.6% of tumors expressed p53, there was no correlation between p53 and survivin expression. Bcl-2 family proteins also play important roles in regulation of apoptosis. Previous work showed Bax expression was associated with a better prognosis and was found to be a predictive marker for adjuvant chemotherapy [20]. In the present study, while 84.6% of tumors expressed Bax, there was no correlation between its expression and survivin expression or clinico-pathological parameters. The present data did not indicate that p53 or Bax protein correlated with survivin expression in pancreatic cancer. It is not certain why this discrepancy exists, although all these proteins may block apoptotic pathway. Whether p53 or Bax expression in pancreatic cancer is related to prognosis is still under debated. In addition, some authors suggest that survivin may be independent prognostic marker not associated with p53 or Bax expression [21, 22]. Therefore, we hypothesized that tumor cell may have other pathway to evade apoptosis after Bax or p53-involved steps in pancreatic cancer. It appears further large scale studies are required to investigate any correlation between survivin and cell cycle regulatory protein expression in pancreatic cancer.

In summary, survivin was expressed at high levels in the vast majority of pancreatic cancer. While the data did not indicate any correlation between survivin expression and clinical outcome, they may be interpreted as suggesting survivin may be a predictive marker in pancreatic cancer for anticancer therapy in pancreatic cancer. We believe further prospective studies are warranted to determine whether survivin is a prognostic or predictive marker in pancreatic cancer.