Benchmarking the Intratumoral Microbiome in Pancreatic Ductal Adenocarcinoma: A Longitudinal Assessment of Contamination Sources and Decontamination Strategies

Résumé fourni par la source

Abstract Pancreatic ductal adenocarcinoma (PDAC) harbors a distinct intratumoral microbiome. Yet rigorous characterization of its composition is hampered by pervasive environmental and procedural contamination. Sources of contamination have not been thoroughly explored, and the methods of decontamination have not been sufficiently evaluated in a benchmarking manner. We systematically collected >300 negative control (NCT) samples comprising paraffin from formalin-fixed paraffin-embedded (FFPE) samples, lysis buffer and sterile water over a period of four years processed by different laboratory persons (LP). All samples were sequenced using full-length 16S rRNA gene sequencing with Oxford-Nanopore Technologies. We benchmarked four decontamination methods (restrictive filtering, decontam, SCRuB, and the Nejman et al.-derived (Nj) pipeline) against fresh-frozen tumor samples (FF) from LSL-Kras G12D/+ ; LSL-Trp53 R172H/+ ; Pdx-1-Cre (KPC) mice, using the abovementioned contamination assessment to calculate a composite score for the assessment. Further, we validated those methods via technical replicates. Microbial profiles of NCT samples were significantly determined by control type, LP, year and season reflecting complex batch effects. The 15 most abundant contaminants spanned well-characterized environmental taxa and human commensals from the oral cavity. The LP processing samples left a significant microbial trace highly contributing to the batch effect. Decontamination benchmarking demonstrated that the Nj method consistently outperformed alternatives in both composite score and inter-replicate concordance. Application of Nj to fresh frozen PDAC samples substantially reduced contaminant burden while preserving putative tumor-associated signals in FF but not FFPE samples. Our results support the adoption of the Nj decontamination approach for future intratumoral microbiome studies in fresh frozen tumor samples. Importance Tumors are increasingly recognized to harbor their own bacterial communities. For pancreatic ductal adenocarcinoma (PDAC)—one of the deadliest cancers—it has previously been shown that the intratumoral microbiome influences cancer proliferation and treatment response. But tumor tissue contains very little bacterial DNA, so trace contamination from reagents, lab surfaces, and the people processing samples can easily be mistaken for real biological signal. Using one of the largest longitudinal collections of negative control samples, together with tumor samples from a mouse model of pancreatic cancer, we show that the person handling a sample can introduce more bacterial noise than the sample’s environment, and we identify a decontamination strategy that reliably distinguishes genuine tumor-associated bacteria from background contamination in fresh-frozen tissue, but not in formalin-preserved tissue. This provides practical guidance for researchers designing low-biomass microbiome studies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.