HPB
Matthew K. Iyer, MD/PhD
Fellow, Complex General Surgical Oncology
Duke University Medical Center
DURHAM, North Carolina, United States
Disclosure: I do not have any relevant financial / non-financial relationships with any proprietary interests.
Detection and treatment of intraductal papillary mucinous neoplasms (IPMN) harboring high grade dysplasia before progression to pancreatic ductal adenocarcinoma (PDAC) remains an important challenge. Molecular assays show promise in guiding the management of IPMN. Here, we employ spatial transcriptomics to discover genes that differentiate high-grade dysplasia (HGD) from low-grade dysplasia (LGD) and nominate a set of candidate cyst fluid biomarkers.
Methods:
Spatial transcriptome profiling was conducted using the NanoString GeoMx Digital Spatial Profiler. Regions of HGD and LGD within IPMN tissue slides were designated by a pancreatic pathologist. Segmental profiling of epithelial cells stained with an Anti-pan-Cytokeratin (PanCK) antibody was performed using oligonucleotide probes from the Cancer Transcriptome Atlas kit. Photocleaved barcodes were sequenced using an Illumina NextSeq machine. Probe count data was analyzed using the R project software, including the DESeq2 package for normalization and differential expression (DE) analysis. We obtained a catalogue of secreted proteins and a list of genes associated with unfavorable prognosis in PDAC from The Human Protein Atlas (HPA) online resource.
Results:
We profiled 64 regions of dysplastic cyst epithelium from 8 patients (4 HGD, 4 LGD per patient). Four regions that did not meet quality control filters were excluded. Differential expression (DE) analysis yielded a set of dysregulated genes for each patient. The union of the individual DE gene sets comprised 226 genes. Of these, 83 genes (36%) were DE in at least 2 of the 8 specimens (see Figure). We nominated 8 cyst fluid biomarkers according to two criteria: 1) evidence of a secreted protein product and 2) evidence that high expression was associated with unfavorable prognosis in PDAC. These included CCL20, COL17A1, DKK1, IL1RN, LAMB3, LAMC2, SFN, and SPP1.
Conclusions:
Spatial transcriptomics of IPMN revealed DE genes specific to regions of high-grade dysplasia. Several of these genes produce secreted proteins associated with unfavorable prognosis in PDAC. Future studies will validate these candidate biomarkers in pancreatic cyst fluid.