Figure. Heatmap visualization of the family-level microbiota found in each anatomic site. A, Comparison of the microbiota composition in the adenoid, tonsil, and middle ear specimens from the chronic serous otitis media case shows that the adenoid microbiota is dominated by multiple bacterial groups, including Pseudomonadaceae, Streptococcaceae, and Fusobacteriaceae, among others, in contrast to the single bacterial group–dominated microbiota in the middle ear and the tonsil. As shown by the heatmap color legend, red and purple represent the most abundant bacterial types, whereas yellow and gray signify the more rare types. The number by the color legend corresponds with the number of sequences from a given bacterial group in each sample. B, Visualization with hierarchal clustering results using the family-level microbiota suggests that while the adenoid microbiota is composed of bacterial families found in both tonsil and middle ear, the tonsil and the adenoid microbiota are more closely related, with the middle ear microbiota as an out group. This finding likely reflects the higher proximity and exposure of the tonsil to the adenoid and the enclosed nature of the middle ear.
Liu CM, Cosetti MK, Aziz M, Buchhagen JL, Contente-Cuomo TL, Price LB, Keim PS, Lalwani AK. The Otologic MicrobiomeA Study of the Bacterial Microbiota in a Pediatric Patient With Chronic Serous Otitis Media Using 16SrRNA Gene-Based Pyrosequencing. Arch Otolaryngol Head Neck Surg. 2011;137(7):664-668. doi:10.1001/archoto.2011.116
Author Affiliations: Division of Pathogen Genomics, Translational Genomics Research Institute, Flagstaff, Arizona (Drs Liu, Price, and Keim and Mss Aziz, Buchhagen, and Contente-Cuomo); Center for Microbial Genetics and Genomics, Northern Arizona University, Flagstaff (Drs Liu and Keim); and Departments of Otolaryngology (Drs Cosetti and Lalwani) and Physiology and Neuroscience and Pediatrics (Dr Lalwani), New York University School of Medicine, New York.
Objective To characterize bacterial microbiota in middle ear, adenoid, and tonsil specimens using 16SrRNA gene-based pyrosequencing analysis.
Design Cross-sectional study of bacterial microbiota in middle ear, adenoid, and tonsil specimens from a pediatric patient with chronic serous otitis media. Middle ear, adenoid, and tonsil specimens from a pediatric patient were collected and underwent cell lysis and DNA isolation. Pyrosequencing was performed on the 454 Life Sciences GS FLX platform (Roche Diagnostics Corp, Branford, Connecticut). Pyrosequencing data were processed, quality-checked, and taxonomically classified to generate an abundance-based matrix. Ecological analyses were performed.
Setting Academic, tertiary referral center.
Main Outcome Measures Comparative microbiome analysis.
Results We detected a total of 17 unique bacterial families, with 9, 9, and 12 bacterial families from the middle ear, tonsil, and adenoid specimens, respectively. Pseudomonadaceae dominated the middle ear microbiota at 82.7% relative abundance, whereas Streptococcaceae dominated the tonsil microbiota at 69.2%. Multiple bacteria, including Pseudomonadaceae, Streptococcaceae, Fusobacteriaceae, and Pasteurellaceae, dominated the adenoid microbiota. Overlap between the middle ear and the tonsil microbiota was minimal. In contrast, the adenoid microbiota encompassed bacteria detected from middle ear and tonsil.
Conclusions Bacterial community analysis using pyrosequencing analysis revealed diverse, previously unknown bacterial communities in a set of pediatric middle ear, tonsil, and adenoid specimens. Our findings suggest that the adenoid may be a source site for both the middle ear and tonsil microbiota. An ecological framework is appropriate in comparative analysis of microbiota from nonsterile body sites.
Middle ear effusion in otitis media was at one time considered sterile. Senturia et al1 first cultured bacteria from middle ear effusions in 1958. Since then, the presence of Streptococcus pneumoniae, Haemophilus influenza, and Moraxella catarrhalis with otitis media, including acute otitis media, otitis media with effusion, and chronic serous otitis media, has become accepted. In children, pathogens may migrate to the middle ear from adenoid tissue of the nasopharynx through a short, patent eustachian tube or insufflated by unintentional valsalva maneuvers.2 The potential role of the nasopharyngeal lymphoid tissue as a reservoir of both middle ear and tonsil disease is supported by a wide range of clinical and microbiological studies.2- 15
Yet, the relationship between the adenoid microbiota—the collective bacterial community that lives in and on the adenoids—and middle ear health and disease remains to be elucidated. In healthy children, the adenoid microbiota consist of diverse bacterial species.4,6,7,15- 17 Our ability to effectively study these polymicrobial anatomic sites has been limited by laboratory techniques and the classic “one-organism, one-disease” infectious disease framework. Using 16SrRNA gene-based pyrosequencing, which is an advanced molecular-based technique that can characterize complex microbial communities with high sensitivity,18 combined with an ecological infectious disease framework—a framework in which bacterial community features such as co-occurrence patterns or interactions between diverse bacterial types are considered in the study of pathogenesis—we can better assess the polymicrobial middle ear and adenotonsillar microbiota in health and disease.
To our knowledge, there has been no published study on the adenoid, tonsil, or middle ear microbiota using an open molecular approach. In this report, 16SrRNA gene-based pyrosequencing analysis is applied to remnant otologic, adenoid, and tonsil surgical specimens of a single pediatric patient with chronic serous otitis media, adenotonsillar hypertrophy, and obstructive sleep apnea. We hypothesize that the microbiota of the middle ear, adenoid, and tonsil will have a high degree of microbial diversity and will be significantly correlated.
The patient was an 8-year-old boy with a 3-year history of hearing loss and right-sided tonal tinnitus. Additional symptoms included speech delay and nasal congestion that was unresponsive to oral antihistamines and topical nasal steroid sprays, as well as nightly snoring, mouth breathing, and witnessed apenic episodes during sleep. He had not taken antibiotics in the previous 3 months and had not had an upper respiratory tract infection within the previous 8 weeks. Pediatric immunizations were up to date. Findings from the patient's otologic examination demonstrated bilaterally intact and retracted tympanic membranes with a greenish-yellow, right-sided serous otitis media, and shallow pars flaccida retraction pockets without squamous debris were noted bilaterally. A Weber examination with a 512-Hz tuning fork lateralized to the right ear, but bilateral Rinne testing was unreliable. Findings from a nasal examination demonstrated clear rhinorrhea bilaterally and oral cavity inspection showed symmetric, tonsillar hypertrophy greater than level 4.
Results from a behavioral audiogram demonstrated a mild right conductive hearing loss with a speech reception threshold (SRT) of 25 dB and a word discrimination score of 100%. In the left ear, the SRT was 10 dB with 96% word discrimination. The right ear yielded a flat tympanogram (type B), while a type Ad tympanogram was found in the left. A polysomnogram demonstrated an apnea-hypopnea index (AHI) of 1.7 and a respiratory distress index (RDI) of 2.3.
A diagnosis of conductive hearing loss, chronic serous otitis media, and obstructive sleep apnea was made, and the patient underwent uncomplicated bilateral myringotomy and tube insertion and adenotonsillectomy.
Preoperative written informed consent was obtained according to a New York University, New York, internal review board–approved protocol (No. 08-174). The study was also approved at the Translational Genomics Research Institute (TGen) by the Western Institutional Review Board through an amendment (pkeim 08-012) of the TGen Umbrella protocol pkeim08-018. The ear canal was irrigated with alcohol. Aspirate from the middle ear was obtained intraoperatively with the aid of binocular microscopy, and deep lymphoid tissues from within the adenoids and tonsils were collected using a sterile technique. Samples were transferred under sterile conditions from the operative instrument to a 15-mL conical tubes containing 200 μL of AllProtect storage media (Qiagen, Valencia, California). The collected specimen was immediately frozen on dry ice in the operating room, then transferred to −70°C storage within 2 hours and stored at −70°C until processing.
The laboratory methods and bioinformatics and statistical analyses can be found in detail elsewhere.19 Briefly, DNA was isolated and purified from each clinical specimen using an enzymatic and mechanical extraction protocol. To fully assess the diverse bacteria in each sample, the V3-V4 regions of the conserved bacterial 16SrRNA gene were amplified using bar-coded fusion polymerase chain reaction primers and sequenced on the 454 Life Sciences GS FLX platform (Roche Diagnostics Corp, Branford, Connecticut). The resultant sequences were assigned to its source sample using the barcode. Because the lowest number of sequences per sample was 214, we normalized the number of sequences per sample to 214 by random sampling without replacement. Sequences from the normalized data set were analyzed to determine its bacterial source based on taxonomic designation (eg, genus). The resultant data set was used to compare the richness (ie, the total number of unique bacterial types found), diversity (ie, the collective measurement of both abundance and number of unique bacterial types found), and composition the adenoid, tonsil, and middle ear microbiota.
DNA was isolated from the collected samples and analyzed using 16SrRNA gene-based pyrosequencing analysis to characterize the bacterial microbiota. We obtained a total of 1042 sequences from the 3 samples. During taxonomic classification using the subset data, we were able to classify more than 90% of the sequences at a 95% bootstrap confidence level at the phyla, class, order, and family level; however, this rate decreased to below 90% at the genus level (Table 1) and led us to perform subsequent comparative microbiota analysis at the family level.
The microbiota of each anatomic site had a distinct profile. Among the 3 samples, a total of 17 bacterial families were detected, with 9, 9, and 12 bacterial families from the middle ear, tonsil, and adenoid samples, respectively. The Pseudomonadaceae bacterial family was found in all specimens and dominated the middle ear microbiota at a 82.7% relative abundance rate. Streptococcaceae dominated the tonsil microbiota at a relative abundance level of 69.2%. The adenoid microbiota was dominated by multiple bacteria, including Pseudomonadaceae, Streptococcaceae, Fusobacteriaceae, and Pasteurellaceae.
Overall, adenoid tissue demonstrated the most diverse bacterial profile that correlated separately with the microbiota of the tonsil and the middle ear. The Shannon diversity index was greatest in the adenoid at 1.84, followed by 0.87 in the tonsil and 0.68 in the middle ear. The heatmap visualization showed that the multiorganism-dominated adenoid microbiota overlapped with both the middle ear and the tonsil (Figure). Hierarchal clustering further indicated that the adenoid microbiota was more closely related to the bacterial composition of the tonsil than the middle ear (Figure).
Analysis of microbial composition by genus again found that Pseudomonas species was the only genus found in all 3 sites, constituting 97.1% of the Pseudomonadaceae sequences (Table 2). Anaerobe identification was most common in the adenoid, with Fusobacterium species and Prevotella species found at greater abundance in adenoid tissue than either tonsil or middle ear specimens. Fusobacterium species was also uniquely abundant in the adenoid.
In this study, we demonstrate that 16SrRNA gene-based pyrosequencing analysis is a powerful molecular technique to characterize complex microbial communities in the middle ear, adenoid, and tonsil. Using 16SrRNA gene-based pyrosequencing, we found that the middle ear, adenoid, and tonsil microbiota harbor previously unknown levels of microbial diversity. The adenoid microbiota was the most complex, encompassing bacteria found in both the tonsil and the middle ear supporting the hypothesis that nasopharyngeal adenoid tissue may serve as a bacterial reservoir for both middle ear and tonsillar diseases.
The complex bacterial communities identified encompassed a wide range of bacterial types, including many that have not been previously reported in culture-based studies. For example, in addition to Hemophillus species and Streptococcus species , known to be associated with otitis media and adenotonsillar disease, additional previously unreported bacterial families, including Comamonadaceae, Oxalobacteraceae, Clostridiales family XI, and Xanthomonadaceae, were detected. These are fastidious and notoriously difficult to culture by traditional methods because they require special media not routinely used for bacterial culture. An open, nontargeted molecular-based method is thus ideal for characterizing body sites colonized by these organisms. In addition, this method also allows the identification of rare organisms within a polymicrobial community that would otherwise be masked by dominant bacteria in a traditional culture-based assay. Because many of the bacterial types we identified were not previously reported in adenotonsillar and middle ear tissues, their clinical roles are, as yet, undefined. Future studies of their role in pathogenesis are thus needed. However, it is worth noting that the sites from this study had relatively lower richness and diversity than most other human body sites.
Traditional culture techniques are also known to be less sensitive for identifying bacteria residing in biofilms.13,18 Biofilms are intricate, 3-dimensional aggregates of bacteria that are highly resistant to both immune-mediated killing and antimicrobial agents and have been detected in a wide range of otorhinolaryngologic infections.14,20 In human tissue, biofilms comprise a community of sessile organisms embedded in an adherent matrix of extracellular polymeric substances. These characteristics both enhance survival of the bacteria and decrease sensitivity of culture-based detection. Biofilm formation of multiple mucosal pathogens, notably Pseudomonas species, has been implicated in a variety of otorhinolaryngologic diseases, such as otitis media with effusion, chronic serous otitis media, and recurrent adenotonsillitis.21,22 Although this study does not provide direct evidence of biofilm formation in the tissue specimens, it does report increased detection of bacteria genera associated with known biofilm-forming capabilities, such as Pseudomonas species.
The molecular method used in this study is highly sensitive for identification of bacteria within human tissues. However, detection of the bacterial 16SrRNA does not necessarily equate with bacterial viability and activity; these would be better supported by detection of bacterial transcripts or proteins in tissues. Reagent contaminants is also an important challenge in studying small-size surgical specimens because the bacterial DNA carried through reagent synthesis can overwhelm the true bacterial content of the specimens. Contamination during specimen collection may also be a problem. For example, Pseudomonas species is known to colonize the external auditory canal, and thus, middle ear specimens are theoretically susceptible to unintentional contamination by this organism; we have worked to minimize the potential for cross-contamination by collecting the specimens in the operating room using sterile techniques. Finally, because this study is restricted to a sample from a single patient, the finding cannot be generalized until reproduced in a larger study population.
In conclusion, pyrosequencing analysis of the bacterial 16SrRNA gene revealed diverse bacterial communities in a set of pediatric middle ear, tonsil, and adenoid specimens. Our results indicate that there is much greater microbial diversity present at these anatomic sites than previously detected. Among the 3 sites, the adenoid microbiota was the most complex and demonstrated the greatest overlap with the tonsil and the middle ear, suggesting that nasopharyngeal tissue may serve as a bacterial reservoir for both middle ear and tonsillar disease.
Correspondence: Anil K. Lalwani, MD, Department of Otolaryngology, New York University School of Medicine, 540 First Ave, Skirball 7Q, New York, NY 10016 (Anil.Lalwani@nyumc.org).
Submitted for Publication: October 26, 2010; final revision received March 29, 2011; accepted April 22, 2011.
Author Contributions: Drs Keim and Lalwani contributed equally to the study. All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Liu, Price, Keim, and Lalwani. Acquisition of data: Liu, Cosetti, Aziz, Buchhagen, Contente-Cuomo, and Lalwani. Analysis and interpretation of data: Liu, Cosetti, Aziz, Price, Keim, and Lalwani. Drafting of the manuscript: Liu, Cosetti, Price, and Lalwani. Critical revision of the manuscript for important intellectual content: Cosetti, Aziz, Buchhagen, Contente-Cuomo, Price, Keim, and Lalwani. Statistical analysis: Liu. Obtained funding: Liu. Administrative, technical, and material support: Liu, Contente-Cuomo, Keim, and Lalwani. Study supervision: Liu, Keim, and Lalwani.
Financial Disclosure: None reported.
Funding/Support: Funding for this research was provided by the TGen Sylvia-Chase Postdoctoral Fellowship to Dr Liu and by TGen Foundation funding awarded to Dr Price.
Role of the Sponsor: The funders did not influence the design and conduct of the study; the collection, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript.
Previous Presentation: This study was presented as a poster at the American Society of Pediatric Otolaryngology; April 30–May 2, 2010; Las Vegas, Nevada.