Abstract Introduction The penile prosthesis is the gold standard surgical treatment for erectile dysfunction refractory to medical treatment, but a subset of individuals require device revision for mechanical failure, pain, erosion or infection. Clinical cultures have been used to demonstrate that both infected and non-infected explanted devices harbor bacteria. We hypothesized that penile prostheses generate complex adherent biological structures called biofilms that consistently harbor microbes, even in the absence of infection, and that their composition would differ by device type and infection status. Objective Using shotgun metagenomics, we sought to determine the biofilm composition of indwelling penile prostheses and compare them by device type and infection status. Methods Approval for this study was obtained from our institutional IRB, and informed consent was provided by all participants who underwent removal and replacement of a penile prosthesis for any indication. Prior to access of the device, the dartos fascia was swabbed as a control. Upon access of the device, instruments and gloves from the surgical team were changed to minimize contamination from the surrounding tissue. The device was explanted, and each component was analyzed in parallel. Device components were subjected to sonication, DNA extraction, and shotgun metagenomics. The curated reads were then compared to annotated databases of known bacterial, viral, fungal, and protozoal species. Metabolic pathway genes were identified and compared to KEGG pathways to generate a functional metabolic analysis. Results 19 devices (16 Boston Scientific; 3 Coloplast) were included in the study. Indications for removal were infection (n=4), pain (n=2), mechanical failure (n=11), and impending erosion (n=2). Of non-host derived reads, approximately 65% were prokaryotic, 15% viral, 15% protozoan, and no fungal in origin (Figure, left). Beta-diversity of bacteria (p=0.046, Figure, right), virus (p=0.04), non-human genes (p=0.02), and gene pathways differed by device type (p=0.04). Interestingly, none of these differed by infection status (p>0.05). Sulfur metabolism was commonly enriched throughout devices, indicative of anaerobic growth conditions. Sphingobacterium hotanense, and the flavin mononucleotide metabolism pathway, were enriched in the context of infection. After exclusion of the 4 devices explanted for infection, there were 16 differentially abundant prokaryotic taxa by device type including 4 Pseudomonas species. Biofilm and antibiotic resistance genes were commonly represented independent of device type or infection status. Conclusions Both non-infected and infected devices consistently harbored microbiota in the form of biofilms, which were rich in prokaryotic, viral, and protozoal taxa. Device type was a greater driver of variability than infection status. Our data suggest that different device types harbor variable affinity for microbiota. Mechanistic studies are needed to understand whether the identified microbiota might protect against or contribute to the risk of infection, pain or erosion. Disclosure No
Read full abstract