Differences in cell-associated and cell-free microbial DNA in blood

In the absence of infection, blood has previously been understood to be free of microbes. However, with advances in sequencing technology this notion has been challenged, prompting new investigations into microbial DNA within the blood of both healthy and diseased individuals. To comprehensively survey microbial DNA in blood, we separated blood into fractions (plasma, red blood cells, and buffy coat) and assessed if the microbial-DNA is cell-free by the addition of DNase to a subset of each fraction. We measured 16S rRNA gene copy number with digital droplet PCR and identified the taxonomic origin of the microbial DNA with synthetic full-length 16S rRNA gene sequencing. As a use case, we examine microbial DNA from the blood of 5 men without prostate cancer (PC), 5 men with low-grade PC, and 5 men with high-grade PC. Our study demonstrates that the majority of microbial DNA is cell-free, indicating that it is not representative of proliferating microbes. Our analyses also revealed buffy coat had the lowest number of 16S rRNA gene copies yet highest number of genera of the fractions (median 23.3 copies/{micro}L and 10 genera) and thus may be a useful fraction to study moving forward. Additionally, microbial DNA in blood may have utility as a biomarker, as we detected disease-associated compositional differences in the plasma and buffy coat fractions. This study lays the groundwork for rigorously studying microbial DNA in blood, however larger studies are needed to confirm our disease-association findings. ImportanceThe concept of a "blood microbiome" has sparked debate in recent years, with questions about whether microbes truly exist in circulation. This study provides a crucial evaluation of the fractions of blood and their capacity to harbor microbial DNA, offering important context for prior and future research. By using DNase to differentiate between cell-associated and cell-free microbial DNA, we show that while microbial DNA is present in blood, it is sparse and heterogenous. These findings highlight the need for rigorous study design that carefully considers both positive and negative controls, as well as the specific blood fractions being examined.