The Influence of Specific Mucin Content on the Microbiome of Middle Ear Effusions from Children with Chronic Otitis Media

Background: Otitis media (OM), an infection of the middle ear, occurs as a result of acute microbial pathogenesis and can progress to chronic OM (COM) following numerous episodes. The role of microbes in COM remains unclear. It has been reported that mucins are required for middle ear immune defense against pathogens. We aimed to characterize the microbiome of middle ear effusions (MEEs) from patients with COM in efforts to determine the relationship between varying microbial communities and the presence of secretory mucins MUC5B and MUC5AC. We hypothesized that MEEs with high mucin content would show a distinct microbiome largely devoid of typical acute OM pathogens. Methods: 42 MEEs from children ages 3 to 146 months with COM undergoing myringotomy at Children’s National Medical Center were recovered. Western blot analysis was performed using anti-MUC5B and anti-MUC5AC antibodies. DNA was extracted and sequenced using Illumina Nextseq500. Non-human sequences were classified using Kraken and Pathoscope, limiting classification to bacteria already identified within humans. Statistical analysis and comparison of the microbiome of MEEs to patient demographics, past medical history, and mucin content were performed using R. Results: 39 MEEs were positive for MUC5B, 29 for MUC5AC, and 28 for both. On average, 12.1% of sequenced DNA was identified as non-human of which Kraken classified 4.9%. 66% were bacterial, 7.8% archaeal, and 26% viral. After excluding an outlying purulent sample almost exclusively containing Pseudomonas, the five most prevalent bacterial genera included Mycoplasma, Methanococcus, Methanosarcina, Haemophilus, and Yersinia. After limiting classification using Pathoscope, we found that Actinobacillus, Cyanothece, Haemophilus, Achromobacter, and Stenotophomonas were the five most abundant genera. The overall proportion of Cyanothece within the microbiome of samples containing both mucins when compared to those containing just one was significantly increased (p= .041). Mucoid samples, compared to serous samples, showed an increase in the proportion of A. pleuropneumoniae (p= .044). Samples from children who experienced significant hearing loss contained a higher proportion of H. influenzae (p=.033). Conclusion: As one of the first to characterize the MEE microbiome, we saw a wide variety of bacteria, including some that have not yet been classified within humans. By limiting our sequences to those previously classified within the human body, we found significant variations in microbial communities between numerous variables including MUC5B and MUC5AC presence, hearing loss, and mucoid vs. serous effusions, suggesting an association between the MEE microbiome and COME pathogenesis