SPARC Expression as a Prognostic Marker in Pancreatic Ductal Adenocarcinoma

Background: The aim is to investigate whether secreted protein acidic and rich in cysteine (SPARC) expression could be considered as an independent prognostic factor in pancreatic ductal adenocarcinoma (PDAC). Methods: The expression of SPARC in the tumor species was detected by immunochemistry and the effect of SPARC on the prognosis was analyzed by Kaplan-Meier and Cox proportional hazard models. Results: Positive SPARC expressions recognized in tumor and stroma were 31.33% and 64.66% respectively. The expression of SPARC in stroma was closely correlated with differentiation, vascular invasion and was an indicator for predicting poor outcome after surgery in Kaplan-Meier analysis. Univariate analysis and multivariate analysis showed the following factors to be significant: perineural invasion (HR = 2.098, 95% CI: 1.144 3.847) and stroma positive expression of SPARC (HR = 2.120, 95% CI: 1.307 3.440). Conclusions: High expression of SPARC in stroma was a promising pathological marker to predict the poor outcome in patients with pancreatic ductal adenocarcinoma.


Introduction
Pancreatic ductal adenocarcinoma (PDAC) is one of the leader causes of cancer-related mortality in the world [1].At the moment of diagnosis, most patients presented with locally advanced or metastatic status and were not considered as surgical candidate.Even for patients who were candidates for potentially curative resection, 5-year survival is poor, reported to be of 15% to 25% [2].Al-though the advancement in radiologic imaging, surgical treatment and novel drugs for adjuvant therapies have contributed to the outcomes improvement for those patients, significant progress hasn't been demonstrated.
It is crucial to detect the accurate staging and find the novel clinically useful prognostic factors for improving the prognosis.Some clinic parameters have been clearly considered as prognostic factors such as tumor size, degree of histological differentiation and resection margin status [3].But all of the above detections were post-surgery and could not be used as predictive factor preoperation.
Genomic alteration and the epigenetic changes have been validated involved in the development of tumor.It is well known that these changes are considered as an early event during the carcinogenesis and can predict the occurrence of tumor.The secreted protein acidic and rich in cysteine (SPARC), is a calcium binding glycoprotein that shows a high degree of interspecies sequence conservation and plays an important role in various cellular progression including tissue remodeling, wound repair, morphogenesis, cellular differentiation, cell pro-  [15] and promote invasion of glioma in vivo [16].Several studies have found that it could be considered as an independent prognostic factor in pancreatic carcinoma [17] [18].
The aim of this retrospective study was to delineate the clinicopathologic parameter and expression pattern of SPARC and the association between them in the patients diagnosed with pancreatic ductal adenocarcinoma.

Demographic features of 150 patients with PDAC
From the PDAC registration of the Department of General Surgery, Affiliated Hospital of Jiangsu University, PDAC patients recruited for the current study undergoing radical surgery between 2005 and 2015 were based on the availability of pathological samples for slides.Negative margin of the resection detected by microscopy was defined as curative resection.Surgical mortality defined as death of Open Journal of Gastroenterology patients within the first month was excluded the survival analysis.The tumors were staged according to the International Union Against Cancer tumor-node-metastasis (TNM) classification, 7th edition.The protocol of this study was reviewed and approved by the institutional ethics committee.Informed consent was obtained for every specimen examined.
Immunohistochemical study of SPARC Slides stained with hematoxylin/eosin from each patient were reviewed.Specimens from all PDAC patients had been fixed in formalin and embedded in paraffin.Formalin-fixed paraffin-embedded tissue sections (5 μm thick) were stained using the immunoperoxidase method with avidin-biotin complex as described previously [19].The primary antibody against SPARC (life science, H95031M) at 1:500 dilutions incubated with slides overnight at 4˚C.The slides were washed thrice in Tris-buffered saline supplement with Tween-20 for 5 min each before visualization with the Envision Plus Detection Kit (DAKO, Carpinteria, CA) followed the manufacturer's protocol.The absence of nonspecific staining was confirmed by the control staining omitting the primary antibody.
After washing thrice in Tris-buffered saline supplement with Tween-20 for 5 min each, the slides were mounted.Nuclei were counterstained with hematoxylin.Immunohistochemical results were judged by three pathologists who were unaware of the clinical data.
SPARC immunostaining was stratified as previous described [20]: tumor negative/stroma negative, tumor positive/stroma negative, tumor negative/stroma positive, and tumor positive/stroma positive.The staining results were judged by German semiquantitatives scoring system which considers both the stain intensity and the area extent and the definition of low and high expression as previous described [21].

Statistical analysis
Statistical analysis of correlation between the SPARC expression and the clinic parameter was performed using χ 2 test.The overall survival (OS) was estimated using Kaplan-Meier method.A univariable analysis of the clinicopathological variables was performed using the log-rank test.A multivariate regression was performed using the Cox proportional hazards model.All analyses were performed using the statistical software SPSS (version 16.0 for windows; SPSS Inc., Chicago, IL, USA).All p value derived from two-tailed test and p < 0.05 was considered statistically significant.

SPARC staining of human PDAC specimens in relation to clinic parameter features
In the IHC analysis, we can see the main staining of SPARC was located in the stroma and weak in tumor cells as previous results [20].In stroma, SPARC expression was mostly obviously seen in the cytoplasm of peritumoral fibroblasts while the staining was weak or even absent in the stroma distant from the cancer pression of SPARC.Within total 150 subjects, 30 (20%) cases were classified as tumor negative/stroma negative, 23 (15.3%) cases as tumor positive/stroma negative, 73 (48.7%) cases as tumor negative/stroma positive and 24 (16%) cases as tumor positive/stroma positive (Figure 1).Parameters association was compared for all the 150 patients according to their stromal SPARC status (Table 1

Discussion
In the present study, we demonstrated that the positive SPARC expression in pancreatic ductal adenocarcinoma mainly located in stroma while compared with the tumor cells, and were 64.66% (97/150) and 31.3%(47/150), respectively.
The high SPARC expression in stroma was associated with vascular invasion and differentiation in PDAC.Infante et al. reported this association in PDAC in the head of pancreases under Whipple procedure [20].Previous research approves that through mediating MMPs activity, SPARC could enhance the invasive ability of prostate and breast cancer cells [22] [23].Also in MCF7 breast cancer cells, increased motility and invasion induced by transfecting with c-Jun proto-oncogene could inhibit by the suppression of SPARC expression [24].Shi et al. demonstrates that in glioma, SPARC delays tumor growth but promotes in- vasion in vivo [25].It also finds that SPARC endows the ability of mediate growth factors involved in angiogenesis, such as platelet-derived growth factor (PDGF), fibroblast growth factor (FGF)-2, and vascular endothelial growth factor (VEGF) [26].
In the survival analysis, we demonstrated that high SPARC expression in the stroma is associated with poor prognosis in primary PDAC samples while the status in tumor cells was not.No matter the SPARC expression status in the tumor cells, the positive status in stroma was associated with the worst prognosis; notwithstanding median survival was both of 11 months.Meanwhile, the median survival time in tumor (+)/stroma (−) was slightly shorter than tumor (−)/stroma (−) although there was no statistic significance.The results consisted with previous research, but the underlying mechanism is unknown.Some studies show that transplanted tumors grow faster in xenograft tumor model in SPARC −/− mice [27] and the growth of pancreatic cancer cells is suppressed by exogenous SPARC protein in vitro [28].These results suggest that a SPARC originated from peritumoral fibroblasts may play a role in inhibiting the growth of pancreatic cancer cells in vivo, in which SPARC expression losing frequently dues to the promoter methylation.In breast cancer, in cases with strong SPARC stromal staining, there is a trend towards increased disease-specific survival [29].
In the immunohistochemical analysis, high SPARC expression in stroma correlated with poor prognosis.Moreover, based on the multivariate analysis of OS, stroma high SPARC expression translated into an independently unfavorable prognostic factor.The involvement of SPARC in tumorigenesis of pancreatic cancer has been mentioned in several previous investigations.Early studies demonstrated that SPARC is a frequent target for aberrant methylation in pancreatic cancer [28] and exogenous SPARC protein has growth-suppressive activity on pancreatic cancer cells [30].Pancreatic tumors grown orthotopically in Sparc-null (Sparc −/− ) mice are more metastatic than tumors grown inwild-type (Sparc +/+ ) littermates [31].Some other studies found that the expression of SPARC by peritumoral fibroblasts portends a poorer prognosis for patients with pancreatic cancer [28].Therefore, it suspects that SPARC may be a protective mechanism and acts as an indicator of invasive disease.Some reports have also revealed that high SPARC expression is related to poor prognosis in the following cancers: melanoma, gliomas [6] [16] [22].The results of the present study were in agreement with these previous reports.Therefore, SPARC was expected to serve as a prognostic-predictive marker of PDAC.
We demonstrated that SPARC were over expressed in a large proportion of PADC tissues in stroma.In addition, high expressions of SPARC were correlated with the disease progression and poor clinical outcome of PDAC.Furthermore, SPARC was proved to be an independent prognostic factor for overall survival in patients with PDAC.Therefore, targeted inhibition of SPARC might be a new strategy for therapy of PDAC.Certainly, further strong supports from basic investigations are still needed.

Figure 1 .
Figure 1.The expression of SPARC detected by immunohistochemistry in pancreatic ductal adenocarcinomas samples.(a) SPARC expression: cancer negative, stromal expression negative; (b) SPARC expression: cancer positive, stromal expression negative; (c) SPARC expression: cancer negative, stromal expression positive; and (d) SPARC expression: cancer positive, stromal expression positive.

Figure 2 .
Figure 2. Kaplan-Meier survival curves stratified by tumor and stromal SPARC status of patients with pancreatic ductal adenocarcinoma after pancreaticoduodenectomy.

Figure 3 .
Figure 3.Comparison of Os rate of 111 PDAC patients who had undergone radical surgery in relation to (a) SPARC expression in stroma and (b) perineural invasion result.Abbreviations: OS, overall survival; PDAC,pancreatic ductla adenocarcinoma; SPARC, secreted protein acidic and rich in cysteine.

Table 2 .
Univariate analysis of prognostic factors for overall survival (n = 111).

Table 3 .
Multivariate analysis of prognostic factors for overall survival (n = 111).