UCSF

Proteolytic enzymes are central regulators of many of the hallmarks of cancer development and progression. Understanding how proteolytic activity evolves over the course of tumor development has the potential to provide mechanistic insights into disease progression and enable personalized treatment regimens. In this work, we sought to characterize protease activity in pancreatic ductal adenocarcinoma (PDAC) and its precursor lesions. In order to study proteolysis in a global and unbiased manner, we applied a technology developed in our lab, termed Multiplex Substrate Profiling by Mass Spectrometry (MSP-MS). This technology utilizes a physicochemical diverse peptide library and mass spectrometry to determine protease specificity from various biological samples.

We first used MSP-MS to determine that a PDAC cell line secretes the lysosomal, aspartyl protease, cathepsin E. Secreted cathepsin E primarily exists as a zymogen and displays minimal activity at neutral pH. However, we were able to use MSP-MS to identify two substrates that were cleaved by the cathepsin E zymogen at pH 6.5, suggesting that this enzyme could be playing a proteolytic role in cancer progression. We also used our MSP-MS assay to assess global proteolytic activity in fluid from pancreatic cysts. Mucinous pancreatic cysts are precursor lesions to PDAC and should be resected if high-grade dysplasia or invasive cancer is suspected (HGD/IC). Unfortunately, clinicians are often unable to determine the stage preoperatively or differentiate mucinous from nonmucinous cysts, which are fully benign lesions of the pancreas. Through MSP-MS coupled with proteomic analysis, we identified three acid-activated proteases with increased activity in mucinous pancreatic cysts: cathepsin E, gastricsin, and tripeptidyl peptidyl peptidase 1 (TPP1). We developed a simple, fluorescence-based assay for analysis of all three proteases and applied this to a cohort of 110 cyst fluid samples. Analysis of gastricsin activity was 95% accurate for differentiating mucinous from nonmucinous cysts, significantly outperforming the most commonly applied biomarker, CEA. Combined analysis of gastricsin with CEA and cathepsin E improved accuracy to 99%. TPP1 activity demonstrated 89% sensitivity and 40% specificity for distinguishing mucinous cysts with HGD/IC from those with only LGD. This performance compares favorably to commonly assessed clinical and radiographic features. Taken together, our results demonstrate that protease activity is differentially regulated in PDAC and activity analysis can serve as an accurate diagnostic tool for helping clinicians identify cysts with the highest potential for malignant transformation.