Dynamics of intestinal and intratumoral microbiome signatures in genetically engineered mice and human pancreatic ductal adenocarcinoma

Background: Recently, emerging evidence has revealed a prominent role of the microbiome in pancreatic ductal adenocarcinoma (PDAC). While to date most observations were made in PDAC patients, mouse models of PDAC still require a precise characterization of their disease-related microbiome in order to understand whether these models could be used for mechanistic and interventional preclinical studies. Methods: To investigate the fecal and tumoral microbiome of LSL-KrasG12D/+;LSL-Trp53R172H/+;Pdx-1-Cre (KPC) mice and control (CTRL) animals, Oxford Nanopore sequencing was applied. Feces were collected from 6-weeks-old, 12-weeks-old, and tumor-bearing KPC mice and age-matched CTRLs. After DNA extraction, metagenomic sequencing was performed. DNA from KPC tumor and healthy pancreas samples was subject to 16S rRNA gene sequencing. Additionally, the 16S rRNA gene and lipopolysaccharide (LPS) as bacterial marker components were detected in KPC tumor tissue over time by fluorescent-in situ-hybridization (FISH) and immunohistochemistry (IHC). Results: Fecal samples from KPC mice have a significantly distinguishable microbiome compared to age-matched healthy CTRLs over time regarding beta diversity (p = 0.001, R2 = 0.2-0.25 for Bray-Curtis). Recently published human PDAC classifiers robustly predict disease status from feces of KPC mice achieving area under the receiver operating characteristic (AUROC) values of up to 0.822. The results remain significant even after factoring in confounding variables sex and age. Furthermore, KPC tumors harbor significantly more bacterial components than healthy pancreas quantified by FISH and IHC. Notably, also the microbial composition differs significantly between KPC tumors and healthy pancreas tissue (p = 0.042 for Bray-Curtis). Microbiota found in human PDAC samples are significantly more abundant in KPC tumors as compared to healthy pancreas (p-value < 0.001). Conclusion: KPC fecal samples show similarities with the microbial composition of stool samples from human PDAC patients. A comprehensive characterization of both fecal and tumoral microbiome in the KPC mouse model demonstrates its potential to serve as a predictive biomarker in the future.