Vascular Endothelial Growth Factor, a Novel and Highly Accurate Pancreatic Fluid Biomarker for Serous Pancreatic Cysts Michele T Yip-Schneider, PhD, Huangbing Wu, BS, Ryan P Dumas, MD, Brad A Hancock, Narasimhan Agaram, MD, Milan Radovich, PhD, C Max Schmidt, MD, PhD, FACS

BS,

Mucinous pancreatic cysts (intraductal papillary mucinous neoplasm and mucinous cystic neoplasm) have the potential to progress to invasive pancreatic adenocarcinoma, presenting an opportunity for early detection, prevention, and cure. Serous cystic neoplasms (SCN) have no malignant potential, but can mimic mucinous pancreatic cysts on imaging. Therefore, identification of biomarkers that can distinguish between cystic lesions is critically important. We hypothesize that vascular endothelial growth factor (VEGF)-A levels in pancreatic fluid correlate with pathologic diagnosis. STUDY DESIGN: Pancreatic cyst/duct fluid samples were prospectively collected from patients undergoing pancreatic resection and correlated with surgical pathology. VEGF levels were detected by ELISA. VEGF-A and VEGF receptor 2 expression in pancreatic tissue was localized by immunohistochemistry. Genetic alterations of the von Hippel-Lindau gene were determined by targeted next-generation sequencing. RESULTS: Eighty-seven patients met inclusion criteria for enrollment. Final pathologic diagnoses included pseudocyst (n ¼ 9), SCN (n ¼ 17), mucinous cystic neoplasm (n ¼ 24), low/ moderate grade intraductal papillary mucinous neoplasm (n ¼ 16), high-grade/invasive intraductal papillary mucinous neoplasm (n ¼ 10), and pancreatic ductal adenocarcinoma (n ¼ 11). VEGF-A was significantly upregulated in SCN cyst fluid compared with all other diagnoses (p < 0.0001). With a cut-off of 8,500 pg/mL, VEGF-A has 100% sensitivity and 97% specificity as an SCN biomarker. VEGF-A and VEGF receptor 2 are overexpressed in SCN cyst tissue. VEGF-C was also significantly elevated in SCN cyst fluid (p < 0.0001). With a cut-off set at 200 pg/mL, VEGF-C identifies SCN with 100% sensitivity and 90% specificity. The presence of a von Hippel-Lindau mutation in SCN cyst tissue correlates with elevated cyst fluid VEGF levels. CONCLUSIONS: This is the first report of a cyst fluid protein biomarker that can positively identify SCN. The ability to distinguish SCN from premalignant/malignant pancreatic cysts can spare the cost and risk of surveillance and surgical intervention in select patients. (J Am Coll Surg 2014; 218:608e619.  2014 by the American College of Surgeons)

BACKGROUND:

Pancreatic adenocarcinoma (pancreatic cancer) is the fourth leading cause of cancer-related deaths, with mortality nearly equal to incidence. Approximately 5% of patients survive 5 years from the time of diagnosis, and pancreatic cancer is anticipated to account for nearly 40,000 deaths in the United States this year alone.1 Surgery performed early in the course of disease can rarely cure patients with pancreatic cancer; other treatment modalities have been largely ineffective due to innate or acquired cancer cell resistance. Pancreatic cancer represents one of the deadliest cancers. Despite these dismal statistics, a substantial improvement in clinical outcomes is possible by identification and screening of patients at risk for pancreatic cancer developing.

Disclosure Information: Dr Schmidt is a paid advisor for Redpath Integrated Pathology Inc. and Asurgen, Inc. He is founder of B9, Inc and has a VEGF-A use patent pending. All other authors have nothing to disclose. This study received financial support from the Indiana Genomics Initiative of Indiana University (supported in part by Lilly Endowment Inc.). Presented at the Southern Surgical Association 125th Annual Meeting, Hot Springs, VA, December 2013. Received December 13, 2013; Accepted December 13, 2013. From the Departments of Surgery (Yip-Schneider, Wu, Dumas, Hancock, Radovich, Schmidt), Pathology (Agaram), Biochemistry/Molecular Biology (Schmidt), Walther Oncology Center (Schmidt), Indiana University School of Medicine, Indiana University Cancer Center (Schmidt), and Richard L Roudebush VA Medical Center (Schmidt), Indianapolis, IN. Correspondence address: Michele T Yip-Schneider, PhD, Department of Surgery, Indiana University School of Medicine, 980 W Walnut St, Bldg R3, Rm 541C, Indianapolis, IN 46202. email: [email protected] or [email protected]

ª 2014 by the American College of Surgeons Published by Elsevier Inc.

608

ISSN 1072-7515/13/$36.00 http://dx.doi.org/10.1016/j.jamcollsurg.2013.12.019

Vol. 218, No. 4, April 2014

Yip-Schneider et al

Abbreviations and Acronyms

¼ guanine nucleotide binding protein a stimulating activity polypeptide 1 IPMN ¼ intraductal papillary mucinous neoplasm MCN ¼ mucinous cystic neoplasm SCN ¼ serous cystic neoplasm VEGF ¼ vascular endothelial growth factor VEGFR-2 ¼ vascular endothelial growth factor receptor 2 VHL ¼ von Hippel-Lindau

GNAS

Although many cases of pancreatic cancer occur sporadically, 2 groups of patients have been identified that are at risk of pancreatic cancer. One group is patients with familial or hereditary pancreatic cancer,2 and the other is patients with cystic lesions of the pancreas.3 Cystic lesions of the pancreas are diagnosed in increasing numbers due to the use of high-resolution imaging and increased awareness of pancreatic cyst symptoms4; in fact, >2% of American adults have pancreatic cysts. Cystic lesions include mucinous (intraductal papillary mucinous neoplasm [IPMN] and mucinous cystic neoplasm [MCN]) and serous (serous cystic neoplasm [SCN]) types.3 Cystic lesions of the pancreas exhibit variable malignant potential. Mucinous cystic lesions, IPMN and MCN, have considerable potential to progress to invasive pancreatic adenocarcinoma. Because these patients are at increased risk of pancreatic cancer developing, screening, risk stratification, and pancreatic cyst removal in select patients at highest risk (ie, where risk of cancer significantly outweighs the risk of surgery) will promote early detection and prevention of pancreatic cancer. Serous cystic neoplasm, unlike IPMN and MCN, is a nonmucinous, benign cyst of the pancreas that has little or no risk of malignant transformation; however, SCN can mimic IPMN and MCN on imaging and symptom presentation. Radiologists, when blinded to pathologic diagnosis, are at best approximately 50% accurate in differentiation. In addition, the overall accuracy of cytology in identifying mucinous lesions is only approximately 50%.5,6 Cyst fluid cytology analysis can often be difficult to perform due to insufficient cells or contamination. In addition, there are no known protein biomarkers to date that can accurately distinguish SCN from mucinous cystic lesions7; patients with SCNs that are not accurately distinguished from mucinous cystic lesions might undergo surgical intervention and its attendant risks unnecessarily. Pancreatic surgery has high morbidity (40% to 50%) and potential mortality (1% to 5%), even in high-volume pancreatic surgery centers. To guide the management of mucinous cystic lesions, the International Consensus Guidelines

Benign Pancreatic Cyst Biomarker

609

were established in 2006 and included parameters such as cyst size, intramural nodules, cyst fluid cytology, and symptoms as indications for surgical resection. In a retrospective study of 147 patients, these guidelines were reported to have a sensitivity of 100%, but a specificity of only 23%dall patients with malignancy were identified as surgical candidates, but 85% of the patients who underwent surgical resection had benign disease; other studies have confirmed these findings.8-12 To accurately identify groups of patients at low or high risk of pancreatic cancer developing, improvements must be made in the ability to distinguish between serous and mucinous cysts.13 Commercial biomarkers, such as pancreatic cyst fluid CEA level and K-Ras mutation status, exist that can help diagnose mucinous cysts. Mucinous cysts typically show elevated CEA levels (>192 ng/mL) with a sensitivity and specificity of 75% and 84%, respectively.5 K-Ras mutations are common, although not universal, in mucinous cysts, so molecular analysis for K-Ras mutations can also identify some mucinous cysts.14 Although these tests can identify a portion of precancerous mucinous cysts, none can accurately identify serous cysts or distinguish benign serous cysts from all other cysts. We report here that vascular endothelial growth factor (VEGF)-A, a critical regulator of vascular growth and function, is considerably elevated in the pancreatic cyst fluid from patients with SCN and can differentiate SCN, a completely benign pancreas cyst, from other pancreas cysts with 100% sensitivity and 97% specificity.

METHODS Patient samples Patients signed informed consent for collection of pancreatic fluid at the time of routine endoscopy (endoscopic ultrasonography or ERCP) and/or operation for the Indiana University Pancreatic Tissue-Fluid Bank. Fluid specimens were placed immediately on ice after procurement and aliquoted for storage at 80 C. For this study, samples from 87 patients collected between 2003 and 2012, including pseudocyst (n ¼ 9), SCN (n ¼ 17), MCN (n ¼ 24), IPMN (low/moderate grade, n ¼ 16 or high-grade/ invasive, n ¼ 10), and pancreatic ductal adenocarcinoma (n ¼ 11), were pathologically confirmed. Intraductal papillary mucinous neoplasm dysplasia was determined according to the WHO criteria. Enzyme-linked immunosorbent assay Pancreatic fluid samples (150 uL) were analyzed for VEGF-A or VEGF-C by Quantikine ELISA (R&D Systems) according to manufacturer’s protocol. Fluid

Yip-Schneider et al

J Am Coll Surg

Benign Pancreatic Cyst Biomarker

A

B 600

500

Serum VEGF-A (pg/ml)

610

400

300

200

100

0

C

Pseudo

(n=4

SCN

MCN

IPMN

16

11

11)

Vol. 218, No. 4, April 2014

Yip-Schneider et al

samples were also assayed using the VEGF165 ELISA (Siemens Healthcare Diagnostics Inc.), which does not cross react with VEGF121. Immunohistochemistry Pancreatic tissue sections were stained with antibodies specific for VEGF (n ¼ 147) (Santa Cruz Biotechnology Inc) and vascular endothelial growth factor receptor 2 (VEGFR-2) (KDR) (Cell Signaling). Briefly, slides were deparaffinized and hydrated in running water. For VEGF, slides were placed in antigen retrieval citrate buffer (pH 6.0; Dako North America) in a pressure cooker for 15 minutes; for VEGFR-2, slides were placed in antigen retrieval citrate buffer (pH 9.0) in a pressure cooker for 15 minutes. All slides were then placed in 3% H2O2 for 10 minutes, incubated with appropriate primary and secondary antibodies, and then counterstained. Intensity of staining was graded by a certified pathologist blinded to the sample identity. Western blotting Representative tissues were lysed in radioimmunoprecipitation assay buffer (phosphate-buffered saline, 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 1 mmol/L phenylmethylsulfonyl fluoride, 10 mg/mL aprotinin, and 1 mmol/L Na3VO4), and the supernatants were obtained. Lysates (10 mg protein) were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis on 4% to 20% gradient gels (Invitrogen) and transferred to Immobilon P membranes (Millipore). The blots were probed with VEGFR-2 (Cell Signaling Technology Inc.) or extracellular signalregulated kinase (Santa Cruz Biotechnology) antibodies according to the manufacturer’s protocol. Recombinant VEGFR-2 protein (GST fusion protein; Cell Signaling Technology) was run as a positive control. Protein expression was detected with enhanced chemiluminescence (Perkin-Elmer Life Sciences). Molecular genetic analysis of von Hippel-Lindau The DNA was extracted from 2 SCN samples and their matched adjacent normal tissue using the Qiagen AllPrep

Benign Pancreatic Cyst Biomarker

611

DNA-RNA-miRNA Universal Kit (Qiagen). A custom multiplex polymerase chain reaction capturing all coding exons and significant portions of introns and untranslated regions was designed using the Ion Torrent Ampliseq Designer (Life Technologies). Barcoded libraries were prepared using the Ion Ampliseq Library Kit, followed by sequencing using the Ion Proton Sequencer and the Ion Proton PIv2 chip. Sequencing data was mapped to the human genome (hg19) and analyzed for somatic mutations using the Ion Torrent Suite 4.0 software and the Ion Reporter 4.0 software. Copy number for VHL was determined using a predesigned VHL copy number variation quantitative polymerase chain reaction assay from Life Technologies, with quantitative polymerase chain reaction performed on an ABI 7900HT Real-Time PCR system, and analysis using the Applied Biosystems CopyCaller Software v2.0. Statistical analysis Statistical significance was determined by analysis of variance with Tukey’s post test for multiple comparisons; statistical significance was set at p < 0.05.

RESULTS VEGF-A levels in patient pancreatic fluid Pancreatic fluid (cyst or duct) samples from 87 patients, including pseudocyst (n ¼ 9), SCN (n ¼ 17), MCN (n ¼ 24), IPMN (low/moderate grade, n ¼ 16 or highgrade/invasive n ¼ 10), and pancreatic adenocarcinoma (n ¼ 11) were collected and pathologically confirmed. The IPMN dysplasia, was determined according to the WHO criteria. Pancreatic fluid samples were subsequently analyzed for VEGF-A by ELISA and correlated with surgical pathologic diagnosis. Levels of VEGF-A were significantly elevated in pancreatic cyst fluid obtained from patients with SCN compared with all other groups tested (p < 0.0001, Fig. 1A). For SCN, the level of VEGF-A ranged from 9,780 pg/mL to 267,530 pg/mL, with a mean of 81,870  20,630 pg/mL. Means for the other groups were as follows: pseudocyst 1,360  500 pg/mL; MCN

Figure 1. Vascular endothelial growth factor A (VEGF-A) levels in pancreatic fluid and serum. (A) The level of VEGF-A was determined by ELISA, as shown in the scatter plot (mean, horizontal black line). Cyst or duct fluid was obtained from patients with confirmed pathological diagnosis of pseudocyst, serous cystic neoplasms (SCN), mucinous cystic neoplasm (MCN), intraductal papillary mucinous neoplasm (IPMN) (low/moderate grade [lo/mod] or high grade/invasive [hi/ inv]), or pancreatic adenocarcinoma (PDA). The red dotted line is the cut-off (8,500 pg/mL) giving 100% sensitivity and 97% specificity. *p < 0.0001 for SCN vs other groups. (B) The level of VEGF165 was determined by ELISA, as shown in the plot with the corresponding level of total VEGF for each fluid specimen analyzed. (C) VEGF-A levels in patient serum were determined by ELISA as shown in the scatter plot (mean, black line).

612

Table 1. Sample no.

Method of Fluid Procurement and VEGF-A Levels Diagnosis

P10

SCN

P128

SCN

P50

Pseudocyst

P53

MCN

P119

IPMN hi/inv

Procedure Fluid type VEGF-A, pg/mL

EUS OR EUS OR EUS OR EUS OR ERCP OR

Cyst Cyst Cyst Cyst Duct

26,780 30,190 9,280 10,130 140 1,680 2,790 2,460 2,590 3,160

EUS, endoscopic ultrasound; hi/inv, high-grade/invasive; IPMN, intraductal papillary mucinous neoplasm; MCN, mucinous cystic neoplasm; OR, operation; SCN, serous cystic neoplasm; VEGF-A, vascular endothelial growth factor A.

2,940  740 pg/mL; IPMN low/moderate 2,240  460 pg/mL; IPMN high grade/invasive 2,810  770 pg/mL; and pancreatic adenocarcinoma 1,950  330 pg/mL. With a cut-off set at 8,500 pg/mL, VEGF-A provides 100% sensitivity and 97% specificity as a biomarker for benign SCN lesions; the receiver operating characteristic curve area is >0.99. Two MCN fluid samples had VEGF levels >8,500 pg/mL, at 9,870 and 15,230 pg/ mL. The highest level of VEGF-A detected in any of the other lesion groups was 6,880 pg/mL (IPMN high grade/invasive). This ELISA can detect the presence of all VEGF-A splice isoforms. When a VEGF165-specific ELISA was performed, much lower levels of VEGF were detected (approximately 4- to 10-fold lower) in the SCN fluids, suggesting that the predominant isoform present is not VEGF165 (Fig. 1B). Table 2.

The method of fluid procurement (routine endoscopy [endoscopic ultrasonography or ERCP] vs operation) did not affect the level of VEGF-A detected (Table 1). Multiple freeze-thaws of the cyst fluid also did not affect VEGF-A levels (data not shown). In addition, cyst fluid levels of VEGF-A in SCN patients did not correlate with patient age, sex, or cyst parameters, such as size, location, or whether micro-/macrocystic (Table 2). VEGF-A levels in plasma, serum, bile, and urine The level of VEGF-A was also determined in other fluids collected including plasma, serum, bile, and urine. There was no significant difference in serum VEGF-A levels (p ¼ 0.70, Fig. 1C) between any of the groups tested. In addition, for the patient with the highest level of cyst fluid VEGF-A (267,530 pg/mL), VEGF-A was not elevated in the patient’s serum (200 pg/mL), plasma (0 pg/mL), or urine (0 pg/mL). In a subset of patients for which bile and urine fluids were available for analysis, VEGF-A levels were relatively low (

Vascular endothelial growth factor, a novel and highly accurate pancreatic fluid biomarker for serous pancreatic cysts.

Mucinous pancreatic cysts (intraductal papillary mucinous neoplasm and mucinous cystic neoplasm) have the potential to progress to invasive pancreatic...
1MB Sizes 0 Downloads 0 Views