Abstract and Introduction
Abstract
This study was undertaken to determine whether recently identified proteins could be translated to clinical practice as markers to distinguish pancreatic adenocarcinoma from chronic pancreatitis on fine-needle aspirate (FNA) samples. Resected pancreatic tissue sections (n = 40) and FNA samples (n = 65) were stained for clusterin-β, MUC4, survivin, and mesothelin. For each biomarker, the staining patterns in adenocarcinoma and in reactive ductal epithelium were evaluated and compared. clusterin-β stained reactive ductal epithelium significantly more frequently than pancreatic adenocarcinoma (P < .001). In comparison, MUC4 and mesothelin were expressed more frequently in pancreatic adenocarcinoma on tissue sections. Positive staining for MUC4 (91% vs 0%; P < .001) and mesothelin (62% vs 0%; P = .01) and absence of staining for clusterin-β (90% vs 7%; P < .001) were noted significantly more frequently in adenocarcinoma cells than in reactive cells in FNA samples. clusterin-β and MUC4 can help distinguish reactive ductal epithelial cells from the cells of pancreatic adenocarcinoma in FNA samples.
Introduction
Pancreatic carcinoma is the fourth leading cause of cancer-associated deaths in the United States. The majority of these tumors occur at advanced stages, and, therefore, resection may not be an option. Our study and studies by others show that resection of small (early) pancreatic tumors (<3.0 cm) and tumors of low histologic or cytologic grade and stage are correlated with improved survival. It is, therefore, important to identify these tumors in their early stages.
In a patient clinically suspected to have a pancreatic tumor, imaging by conventional computed tomography scanning has become one of the standard modalities to indicate the size and extent of the tumor. Recently, it has been documented that endoscopic ultrasound (EUS) is a more sensitive modality than imaging by conventional computed tomography scan for detecting small pancreatic tumors and for detecting invasion of vessels, which may help to assess resectability of pancreatic carcinomas. Imaging modalities alone, however, can neither differentiate benign from malignant lesions nor determine the type of neoplasm. The National Comprehensive Cancer Network consensus group, formed of experts treating pancreatic cancer, also has recognized that tissue diagnosis should be obtained before instituting definitive therapy. In cases in which imaging studies suggest an unresectable pancreatic tumor, a diagnosis of adenocarcinoma on cytologic assessment of fine-needle aspirate (FNA) becomes important before starting protocol-associated therapy.
A preoperative tissue diagnosis can be obtained by using image-guided FNA and needle biopsies. In recent years, EUS-guided FNA has been used increasingly to obtain preoperative diagnosis and for staging pancreatic carcinoma. This application of EUS-FNA is becoming common since it has emerged as a very sensitive and specific modality in the diagnosis of pancreatic adenocarcinoma. Furthermore, we also have demonstrated that the sensitivity and specificity of detecting pancreatic carcinoma using EUS-FNA does not vary whether the lesion is small or large. Although many studies suggest improved specificity, experts agree that interpreting FNA samples from the pancreas is inherently difficult and that interpretation is even more challenging when samples are scant and bloody, as often noted on EUS-FNA. In our experience and that of others, nondiagnostic samples (inadequate or equivocal diagnoses) from pancreas and other organs when using EUS-FNA range from 11% to 30%. In such a scenario, a marker that can serve as an adjunct in separating the cells of pancreatic adenocarcinoma from the cells of reactive ductal epithelium would be very useful.
Recently, the use of high-throughput technologies on multiple pancreatic cell lines has identified potential genes and proteins that could have roles in pancreatic carcinogenesis. It would be important to determine the clinical usefulness of such differentially regulated proteins, especially if they could be used as markers to separate the cells of reactive ductal epithelium from the cells of pancreatic adenocarcinoma. FNA procedures guided by EUS imaging are increasingly being performed throughout major centers to obtain preoperative diagnostic and staging information; thus, performing biomarkers on the limited samples obtained using EUS-FNA could become extremely important for multidisciplinary patient management teams if they were useful in aiding the diagnosis of pancreatic adenocarcinoma.
Clusterin/apolipoprotein J is a secreted heterodimeric glycoprotein of 70 to 80 kd. Clusterin (apolipoprotein J) also serves as a heat shock protein and has chameleon-like activity that influences many basic cell functions, including cell remodeling, differentiation, apoptosis, and cell proliferation. As a result, it has been proposed as a candidate for the development of antisense therapies for certain tumors. Clusterin has been reported to be overexpressed in anaplastic lymphoma and various carcinomas. In experimental studies, overexpression of clusterin was associated with regeneration and development of pancreas and pancreatitis.
MUC4 is a transmembrane apomucin that has been reported to be overexpressed in pancreatic cancer cells. The messenger RNA for MUC4 also has been identified as overexpressed in more than 75% of pancreatic adenocarcinomas but not in chronic pancreatitis. Expression of MUC4 also has been reported to be a prognostic indicator for pancreatic adenocarcinoma. Decreased expression of MUC4 induced in a pancreatic cell line xenograft via plasmid-transfected anti-MUC4 led to decreased expression of MUC4, reduced clonogenicity, decreased cell proliferation, and reduced tumor volume. These studies suggest that MUC4 expression has an important role in the development and progression of pancreatic adenocarcinoma. Its usefulness as a diagnostic marker in limited samples remains unexplored.
Survivin, described in 1997, is an inhibitor of apoptosis. Overexpression of survivin in human pancreatic tissues is postulated to be an early molecular event in pancreatic carcinogenesis, and its expression has been correlated with prognosis; however, to our knowledge, there has been no study to characterize its usefulness as a marker to aid in preoperative diagnosis using FNA samples.
Mesothelin is a differentiation antigen expressed on the cell membrane of normal mesothelial cells. Recent gene expression data and serial analysis of gene expression also demonstrated that mesothelin could be expressed not only by mesotheliomas and ovarian epithelium but also by pancreatic cancer cells. Tissue expression studies also have demonstrated that mesothelin could be expressed in 85% to 100% of pancreatic cancer cells.
To our knowledge, however, except for mesothelin, there is not a single study in the literature that shows that the aforementioned candidate markers could be translated to diagnostic markers for limited cytology samples. The present study was undertaken to determine whether the recently identified proteins could be translated to clinical practice as diagnostic markers to distinguish pancreatic adenocarcinoma from chronic pancreatitis on FNA samples.