Abstract
Background: Secreted protein acidic and rich in cysteine (SPARC) is a protein which governs diverse cellular functions and matrix remodeling systems. It was reported that SPARC expression of peritumoral fibroblasts was correlated with poor prognosis in resectable pancreatic cancer patients in immunohistochemical analysis. However, the significance of SPARC mRNA expression in pancreatic adenocarcinoma remains unclear. Materials and Methods: SPARC mRNA expression was evaluated in 104 formalin-fixed paraffin-embedded tissue samples of pancreatic adenocarcinoma patients with macro-dissection technique. Results: High SPARC mRNA expression was associated with a poorer prognosis than low SPARC. In univariate Cox's proportion hazard model, depth of tumor invasion, lymph node metastasis, histopathological tumor grade, lymphatic invasion, vascular invasion, surgical margin, no chemotherapy and high SPARC expression were significant prognostic factors. In multivariate analysis, histopathological tumor grade, surgical margin, no chemotherapy and high SPARC expression were independent significant prognostic factors. Conclusion: SPARC mRNA expression is a prognostic marker for pancreatic adenocarcinoma patients.
Pancreatic cancer is a deadly disease which ranks the fourth leading cause of cancer death in the United States (1). The vast majority of patients with pancreatic cancer experience a poor clinical course because of the limited efficiency of currently available therapeutic approaches, including surgical resection, chemotherapy, and radiotherapy (2).
Secreted protein acidic and rich in cysteine (SPARC), also known as osteonectin, is a matricellular glycoprotein involved in diverse biological processes including tissue remodeling, wound repair, morphogenesis, cellular differentiation, cell proliferation, cell migration, and angiogenesis (3-6). Increased SPARC expression has been described in many tumor types, including colonic, esophageal (7), breast (8), pancreatic (9) and melanoma. On the contrary, SPARC expression is frequently lost in pancreatic cancer cells through promoter methylation (9). Interestingly, SPARC is often overexpressed in stromal fibroblasts immediately adjacent to pancreatic cancer cells, suggesting its role in the tumor-stroma interactions (9). SPARC expression in fibroblasts from noncancerous pancreatic tissue was augmented by co-culture with pancreatic cancer cells (9). Furthermore, stromal SPARC expression was shown to be an independent prognostic factor in a large series of patients with pancreatic cancer in immunohistochemical analysis (10). These finding altogether suggest that SPARC plays a critical role in pancreatic cancer progression.
Previous studies have showed SPARC is a prognostic marker in various tumors. It seems, however, that the role of SPARC in prognosis is complex and cell type-specific. In colorectal carcinoma, lacking SPARC expression was poor prognostic factor (11) and high SPARC expression is a marker of chemosensitivity (12). On the other hand, high SPARC expression was associated with poor prognosis and survival in tongue carcinoma (13), breast (8, 14), head and neck (15), esophageal (7), bladder (16) and gastric cancer (17), and melanoma (18). In pancreatic ductal adenocarcinoma, SPARC expression in cancer cells is not significantly associated with prognosis. However, high SPARC expression in peritumoral fibroblasts is a marker of poor prognosis in immunohistological analysis (10).
Despite the progress in the past decade, there is little information on the role of SPARC in pancreatic ductal adenocarcinoma. In particular, there is little information on the relationship between SPARC mRNA expression and long-term clinical outcome. In this study, we analyzed total RNA extracted from formalin-fixed paraffin-embedded tissue samples using macrodisssection to identify SPARC mRNA expression.
Materials and Methods
Patients and study design. Our study series consisted of 104 patients who underwent pancreatic resection for pancreatic cancer at the Department of Surgery and Oncology, Kyushu University Hospital (Fukuoka, Japan) from 1992 through 2007. The patients were comprised of 66 men and 38 women with a median age of 66 years (range, 36-86 years). Survival was measured from the time of pancreatic resection and death was the endpoint. Prognosis was examined in October 2008 and follow-up data of 104 cases were available. The median observation time for overall survival was 20 months, ranging from 1 to 101 months. Sixty-four patients died during the follow-up; the other patients were alive and censored.
All resected specimens were fixed in formalin and embedded in paraffin for pathological diagnosis. All tissues adjacent to the specimens were evaluated histologically according to the criteria of the World Health Organization (19). Two pathologists were in agreement with regard to the pathological features of all cases and the diagnoses were confirmed. The stage of tumors was assessed according to the UICC classification (20). The clinicopathological characteristics of the tumor collection are noted in Table I. The study was approved by the Ethics Committee of Kyushu University and conducted according to the Ethical Guidelines for Human Genome/Gene Research enacted by the Japanese Government and the Helsinki Declaration.
Isolation of total RNA. Total RNA was isolated from tumor tissue by macrodissection using a High Pure RNA Isolation kit (Roche Diagnostics, Mannheim, Germany) and RNeasy Mini Kit (Qiagen, Tokyo, Japan) according to the manufacturer's instructions; RNA concentrations were measured with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Rockland, DE, USA) at 260 and 280 nm (A260/280). RNA integrity was assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Palo Alto, CA, USA).
Quantitative reverse transcription-polymerase chain reaction (qRT-PCR). Specific primers for SPARC and β-actin were used. One-step quantitative real-time RT-PCR (qRT-PCR) with gene-specific priming was performed with a QuantiTect SYBR Green RT-PCR Kit (Qiagen) with a Chromo4 Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA), as described previously (21). We designed specific primers for SPARC (forward primer, 5′-gagaaggtgtgcagcaatga-3′ and reverse primer, 5′-aagtggcaggaa gagtcgaa-3′), β-actin (forward, 5′-tgagcgcggctacagctt-3′; reverse, 5′-tccttaatgtcacgcacgattt-3′) and screened a database with BLAST (DNA data bank of Japan, http://blast.ddbj.nig.ac.jp/top-j.html) to confirm the specificity of these primers. The level of expression of the mRNA for each gene was calculated on a standard curve constructed from values for total RNA from the Capan-1 pancreatic cancer cell line for β-actin and pancreatic fibroblasts cell line for SPARC. Expression of SPARC was normalized to that of β-actin. The cut-off value of SPARC expression was determined by partition analysis to separate group of high and low mRNA expression of SPARC.
Statistical analyses. SPARC expression was split into high and low level groups using recursive descent partition analysis, as described by Hoffmann et al. (22). The significance of differences between groups of continuous data was identified by Student's t-test. Categorical variables were assessed by χ2 or Fisher's exact probability test. Speaman's rank correlation coefficient test was carried out for testing the association between ordinal variables. Univarate analysis of overall survival and survival curves were performed by Kaplan-Meier method, and Cox proportional hazards model was used for multivariate analysis. All statistical analyses were performed with JMP software (SAS Institute, Inc, Cary, NC, USA). A value of p<0.05 was considered statistically significant.
Association between SPARC expression and clinicopathological variables.
Kaplan-Meier survival curves for patients who had undergone surgical resection of pancreatic ductal adenocarcinoma, stratified according to the expression of SPARC mRNA in the macrodissection bulk tissue. Low SPARC <4.3, High SPARC >4.3; log-rank test, p<0.0001.
Results
Association between SPARC mRNA expression and clinicopathologic features in pancreatic ductal adenocarcinoma. The associations between the mRNA expression of SPARC and clinicopathological features are presented in Table I. The median value of SPARC mRNA expression was 0.60 (range, 0.05-23.37), and the cut-off value of 4.3 was determined by partition analysis to separate group of high and low mRNA expression of SPARC. High SPARC mRNA expression was significantly associated with chemotherapy (p=0.0062, Table I). However, we found no significant association between high SPARC mRNA expression and age, gender, depth of tumor invasion, lymph node metastasis, stage, histopathologic tumor grade, lymphatic invasion, vascular invasion or surgical margin (Table I).
Survival analysis. Overall survival was defined as the period from the day of surgery until death of the patient. Death from a cause other than cancer relapse or survival until the end of observation period was considered a censoring event. Follow-up data were available for 104 patients. The median follow up time was 20 months and the survival rate was 64.9% and 40.8% at 1 and 2 years. The 5-year overall survival rate of the 104 patients was 20.24%. As shown in Figure 1, patients with a low SPARC mRNA level had a significantly better 5-year survival rate (22.48%) than those with a high SPARC mRNA level (0.00%; p<0.0001).
Univariate analysis of prognostic factors in Cox's proportional hazards model.
Multivariate analysis of prognostic factors in Cox's proportion hazards model.
Univariate analysis of prognostic factors. The univarate analysis of prognostic factors of the patients with pancreatic ductal adenocarcinoma is summarized in Table II. Depth of tumor invasion (p=0.0001), lymph node metastasis (p=0.0003), histopathological tumor grade (p=0.0243), lymphatic invasion (p=0.0097), vascular invasion(p=0.0014), surgical margins (p=<0.0001), no chemotherapy (p=0.0219) and high SPARC expression (p=0.0017) were positively correlated with a poor prognosis.
Multivariate Cox proportional hazards regression analysis. All factors that were significant for predicting overall survival by univariate analysis were included in the multivariate Cox regression analysis. Multivariate analysis identified high SPARC expression as a significant independent prognostic factor for poor survival (hazard ratio (HR) 2.92 (95% confidence interval (CI) 1.63-5.50), p=0.0161) (Table III). Histopathological tumor grade (HR 2.99 (95% CI 1.63-5.04), p=0.0005), surgical margins (HR 3.22 (95% CI 1.84-5.68), p=<0.0001) and no chemotherapy (HR 2.16 (95% CI 1.22-3.78), p=0.0092) were also significant for poor survival, while depth of tumor invasion, lymph node metastasis, lymphatic invasion and vascular invasion were not.
Discussion
It was reported that high SPARC expression is a poor prognostic marker for many malignant tumors in immunohistochemical analysis. In particular, immunohistochemical analysis showed prognosis for ampullary cancer (23) and pancreatic cancer (10) patients is correlated to SPARC expression in stroma. In pancreatic ductal adenocarcinoma, the SPARC gene is down-regulated by aberrant hypermethylation(9). In contrast, peritumoral fibroblasts generally express SPARC. Microdissection was not performed in this study, but we suppose high mRNA expression of SPARC from formalin-fixed paraffin-embedded tissue may represent stromal SPARC expression in pancreatic ductal adenocarcinoma.
It has been suggested that SPARC has growth-inhibitory properties in many, but not all, tumor types. In previous reports, addition of exogenous SPARC protein inhibited the proliferation of pancreatic cancer cells (9, 24), suggesting a growth-inhibitory effect of SPARC. Meanwhile SPARC is also known to influence the invasive and metastatic capacities of certain tumors (25-29). SPARC has been suggested to play a role in the invasive and metastatic phenotype of pancreatic cancer because SPARC was initially identified as a pancreatic cancer invasion-specific gene through serial analysis of gene expression (30). However, there is only limited information regarding the effects of SPARC on the invasive behavior of pancreatic cancer cells. Addition of exogenous SPARC to Colo-357 pancreatic cancer cells increased their invasion, while antisense inhibition of SPARC in Panc-1 cells reduced their invasion in vitro (24). Similarly, exogenous SPARC increased the invasion of Panc-1 cells both in monoculture and in co-culture with pancreatic stellate cells (31). To confirm any influence of exogenous SPARC from peritumoral fibroblast in pancreatic cancer stromal interaction, we tried an invasion assay of pancreatic cancer cells co-cultured with fibroblasts overexpressing SPARC and transfected with Si-SPARC, but we found no significant difference (data not shown).
In pancreatic cancer, fine-needle aspiration biopsy is often used to confirm diagnosis. At that time, samples can be obtained but they are small and not always suitable for immunohistochemical analysis. However, mRNA can be analyzed in small samples. If fine-needle aspiration biopsy is carried out, it may provide significant information in tailor-made medicine.
Our study demonstrates that the further investigation of SPARC is warranted due to its status as a potential prognostic and therapeutic agent.
In conclusion, the present study is the first report showing that in mRNA extracted from formalin-fixed paraffin-embedded tissues using macrodissection, a high level of SPARC expression is a significant independent prognostic marker in pancreatic ductal adenocarcinoma. These findings suggest that assessment of SPARC mRNA expression could contribute to determining therapy of pancreatic ductal adenocarcinoma patients.
Footnotes
- Received September 3, 2009.
- Revision received February 16, 2010.
- Accepted February 18, 2010.
- Copyright© 2010 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
