Germline Mutations in Pancreatic Adenocarcinoma

Background Pancreatic ductal adenocarcinoma has poor prognosis mainly highlighted by the low 5-year survival rate, which is estimated to be around 6%. Although pancreatic cancer incidence is approximately 0.5% in the general population, it has a high-mortality rate. It is being registered as the fourth leading cause of cancer-related deaths in the United States. Poor prognosis is mainly attributed to the lack of efficient treatment options.

With the majority of pancreatic adenocarcinoma cases being sporadic, familial clustering (defined by the presence of at least two first degree relatives with pancreatic cancer) is observed in ~20% of the cases, whereas a clear genetic cause is identified in approximately half of these cases. Pancreatic cancer is known to occur in a range of cancer predisposing syndromes, such as hereditary breast and ovarian cancer, Peutz-Jeghers, familial adenomatous polyposis, Lynch, Li-Fraumeni and familial atypical multiple mole syndromes, while there have been clear associations with loss-of-function mutations in non-syndromic cancer genes, such as PALB2 and ATM. In addition, individuals carrying mutations in PRSS1 gene that predisposes for hereditary pancreatitis are also considered to be at high-risk for pancreatic cancer. It is therefore clear that pancreatic cancer is not related with a single cancer predisposition gene, but with numerous known and probably yet to be discovered, cancer susceptibility genes.

On the other hand, the accurate fraction of genetic predisposition in pancreatic cancer hasn't been determined. Previous studies have evaluated the prevalence of loss-of-function mutations in known cancer predisposition genes in pancreatic cancer patients through the use of multi-gene panels. The mutation yield ranged from 3.8% to 9.7% in genes that were previously found relevant to pancreatic cancer predisposition.

Through these studies, it has been highlighted that family history and stage of pancreatic cancer or age at diagnosis is not always a predictor of an underlying genetic factor, while phenotypic heterogeneity is usually associated with mutation status. The search for novel predisposition genes in pancreatic cancer patients without family history is ongoing through candidate genes analysis or through whole exome sequencing.

Research methodology Clinicopathological characteristics

Genomic DNA from pancreatic cancer patients will be retrospectively and prospectively collected from the Hellenic Cooperative Oncology Group (HeCOG) Tumor Repository, through years 2000-2019. Written informed consent has already been obtained from all patients for the use and storage of their biologic material along with their approval for their participation in research studies. The study will be in agreement with the 1975 Helsinki statement, revised in 1983.

Clinicopathologic characteristics of patients with pancreatic cancer will be retrieved form their medical record. Diagnosis will be confirmed through pathology reports, which will also provide information about grade, proliferation rate, histological subtype and lymphovascular invasion. From the patient medical record, tumor size, nodal involvement as well as metastatic lesions (TNM staging) will be reported. Importantly, detailed family history will be collected. The collection of these information will be in compliance with the regulations of the Bioethics committees of participating hospitals. Clinicopathological characteristics will be correlated with the different genomic abnormalities identified by sequencing.

Germline DNA extraction

Blood samples will be collected from patients with pancreatic cancer. Germline DNA from white blood cells will be isolated using a commercially available blood DNA extraction kit. Germline DNA will then be subjected to mechanical fragmentation via sonication to reach a DNA size of approximately 200-250 bp for downstream next-generation sequencing (NGS) analysis.

Next-Generation Sequencing

Targeted capture enrichment NGS relies on in-solution hybridization of DNA libraries with TACS (Targeted Capture Enrichment Sequences) for the identification of germline mutations in cancer patients. First, TACS that target mutations associated with hereditary cancer (in a total of 62 cancer predisposing genes) are designed and prepared by PCR followed by biotinylation. Biotinylated TACS will then be immobilized on streptavidin coated magnetic beads for subsequent hybridisation with the DNA libraries. DNA libraries are prepared from patient derived-germline DNA that fulfill quality control criteria (DNA yield, DNA integrity and fragment size) according to established protocols. Library preparation involves the 3' end and 5' end adaptor ligation and each DNA library will be barcoded using unique oligonucleotide DNA sequences to allow subsequent discrimination after sequencing. Subsequent to barcoding, samples will undergo hybridization using custom made 5'-biotynylated TACS followed by PCR amplification. Finally, enriched DNA libraries will be normalised and pooled on a NGS platform prior to paired-end sequencing using manufacturer's protocols.

Bioinformatic and Data Analysis Sequence data from leucocyte (buffy coat) DNA will be demultiplexed and aligned to the human genome (hg19) using appropriate aligner algorithms (BWA) to generate a BAM file. A combination of variant calling algorithms will be used to call germline mutations. Local realignment will be performed for accurate detection of small insertions and deletions and appropriate algorithms to detect rearrangements. Bioinformatics tools for copy number alterations (CNAs) calling such as ASCAT-2, CNVkit or ANACONDA will be used for the assessment of focal gene amplification or deletion events. Read depth-based analysis will also be taken into consideration for assessment of gene-level CNAs. Appropriate variant annotators will be used to annotate the identified alterations to each gene. Germline variant databases such as ClinVar will be used for the classification of variants according to the ACMG guidelines.