Comprehensive Analysis of Bacterial Flora in Postoperative Maxillary Cyst Fluid by 16S rRNA Gene and Culture Methods

Intracystic fluid was aseptically collected from 11 patients with postoperative maxillary cyst (POMC), and DNA was extracted from the POMC fluid. Bacterial species were identified by sequencing after cloning of approximately 580 bp of the 16S rRNA gene. Identification of pathogenic bacteria was also performed by culture methods. The phylogenetic identity was determined by sequencing 517–596 bp in each of the 1139 16S rRNA gene clones. A total of 1114 clones were classified while the remaining 25 clones were unclassified. A total of 103 bacterial species belonging to 42 genera were identified in POMC fluid samples by 16S rRNA gene analysis. Species of Prevotella (91%), Neisseria (73%), Fusobacterium (73%), Porphyromonas (73%), and Propionibacterium (73%) were found to be highly prevalent in all patients. Streptococcus mitis (64%), Fusobacterium nucleatum (55%), Propionibacterium acnes (55%), Staphylococcus capitis (55%), and Streptococcus salivarius (55%) were detected in more than 6 of the 11 patients. The results obtained by the culture method were different from those obtained by 16S rRNA gene analysis, but both approaches may be necessary for the identification of pathogens, especially of bacteria that are difficult to detect by culture methods, and the development of rational treatments for patients with POMC.


Introduction
Three theories, that is, the retention cyst, cleft cyst, and closed cavity theories, have been proposed for the formation of postoperative maxillary cysts (POMCs) [1][2][3], which are a long-term delayed complication arising from years to decades after radical operations such as the Caldwell-Luc operation for maxillary sinusitis [4,5]. This complication is generally known as POMC, although other terms are also found in the literature, such as postoperative buccal cyst, postoperative wangenzyste, mucocele, postoperative paranasal cyst, surgical ciliated cyst, and respiratory implantation cyst [6]. These cysts are considered rare in the West, whereas cysts in the maxillooral region constitute 20% of reported cysts in Japan [6]. They grow painlessly in the maxillary sinus where they are often perceived through swelling or pain in the maxillary mucobuccal fold and buccal region. Other symptoms noted with significant cyst growth include nasal obstruction, rhinorrhea, dysosmia, exophthalmos, and ocular displacement. Bacterial infections in the oral cavity may be closely related to cyst growth and subsequent symptoms [4]. Over 600 bacterial species have been detected in the oral cavity [7], and the majority of these are viable but not yet cultured (VBNC) bacteria [8]. Thus, it is often extremely difficult to identify pathogens in the oral cavity for this type of cyst as well as for other types. Only a few studies have reported the presence of bacteria in the pus of POMCs, but they are unclear [4,9]. We investigated the bacterial species involved in cyst infection by comprehensively analyzing the bacterial flora in POMC fluid by 16S rRNA gene analysis.

Subject Population and Characteristics.
Study patients included 7 males and 4 females with POMC (mean age: 58.8 years; range: 33-87 years). All patients had a history of surgery for sinusitis (14-50 years prior; mean: 34.6 years prior). Cyst fluid was obtained by aspiration, and most fluid was dark brown in color. The fluid was serous in 4 patients and viscous in 7 patients. In all patients, signs of inflammation were evident in the affected region, such as swelling, redness, pain, fever, and/or nasal obstruction (Table 1). Antibiotics were administered prior to the collection of cyst fluid in 9 patients, but it was unclear in 1 patient.

Collection of Samples.
After gargling with gluconic acid, chlorhexidine (ConCool F; WelTec, Japan), the most swollen maxillary mucobuccal area was disinfected with povidoneiodine solution before direct intracystic puncture using an 18-G needle (Terumo Corporation, Tokyo, Japan) and aspiration. Samples were immediately transferred into sterile tubes containing CO 2 (Kenkiporter; Terumo Clinical Supply Co., Ltd., Tokyo, Japan) to ensure the preservation of anaerobes. One tube was used for the identification of pathogens by bacterial culture testing, while the other was subjected to DNA extraction within 1 h of collection. C for 7 min. The PCR products (approximately 580 bp) were determined by gel electrophoresis using a 2.0% agarose gel. After confirming the product size, the product was purified and added to 25 μL of TE buffer using a Centricon YM-100 concentrator (Millipore, Bedford, MA, USA).

Clone Library Construction and Nucleotide Sequence
Determination [10]. PCR products were cloned using a TOPO TA Cloning kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. Transformation was performed using competent E. coli TOP10 cells provided by the manufacturer. A total of 96-288 white colonies were randomly selected from each clone library for sequencing analysis. To prepare a template for sequencing analysis, a partial fragment was amplified from the cloning vector (pCR II) containing an inserted PCR product using M13 forward, M13 reverse, and AmpliTaq Gold DNA polymerase. The reaction mixtures were incubated in a thermocycler at 95 • C for 2 min, followed by 20 cycles at 95 • C for 30 s, 60 • C for 30 s, and 72 • C for 30 s, with a final extension at 72 • C for 3 min. After the primers and dNTP were eliminated from the PCR mixture using an ExoSAP-IT Kit (USB, Cleveland, OH, USA) according to the manufacturer's instructions, a 1 μL aliquot was used as a template for the sequencing reaction. The sequencing reactions used the M13 reverse primer and a BigDye Terminator Cycle Sequencing kit v3.1 (Applied Biosystems, Carlsbad, CA, USA). The nucleic acid sequences were determined using a 3130xl Genetic Analyzer (Applied Biosystems). Homology searches were performed using the basic local alignment search tool (BLAST) with an in-house software system and database, which contained type strains (5,878 species) from the RDP-II website (http://rdp.cme.msu.edu/) and DNA Data Bank of Japan (http://www.ddbj.nig.ac.jp/). A level of 98% sequence identity was used as the cutoff for the identification of a species taxon. Phylogenetic trees were constructed using the neighbor-joining method of Saitou and Nei [11].

Bacterial
Culture. Pathogens were isolated from intracystic fluid at the central and clinical laboratories in Saga University Hospital. Intracystic fluid from each patient was inoculated onto 6 different types of agar plates, that is, sheep blood agar, sheep blood agar supplemented with phenylethyl alcohol, bromothymol blue agar, chocolate agar, prereduced Brucella HK (RS) agar, and prereduced Brucella HK-PV agar (Kyokuto Pharmaceutical Industrial Co., Ltd., Tokyo, Japan). The plates were cultured aerobically or anaerobically at 37 • C for 2-7 days. Based on the abundance and morphology of colonies, 2-5 colonies per sample were isolated individually on the agar plates, as described above, and reincubated aerobically or anaerobically at 37 • C for 2-7 days. The isolates were identified using VITEK2 (Biomerieux Japan Co., Ltd., Tokyo, Japan), WalkAway (Siemens Healthcare Diagnostics K.K., Tokyo, Japan), or the RapID-ANA II system (Remel Inc., Lenexa, KS, USA) according to the manufacturer's instructions.

Results
The phylogenetic identity of 1139 16S rRNA gene clones was determined by sequencing 517-596 bp for each clone. A total of 1114 clones were classified while the remaining 25 clones were unclassified ( Table 2). The 25 clones showed less than 85% homology to nucleotide sequences present in the database. Overall, 41 genera and 64 species were detected. The genera and species are listed in Figures 1  and 2, respectively, in the order of decreasing prevalence. Prevotella (91%), Neisseria (73%), Fusobacterium (73%), and Porphyromonas (73%) had the highest prevalence in all patients ( Figure 1). S. mitis (64%), Fusobacterium nucleatum (55%), P. acnes (55%), S. capitis (55%), and S. salivarius (55%) were detected in more than 6 of the 11 patients ( Figure 2). Figure 3(a) shows 70 of the most prominent taxa among the 105 taxa/species detected. The taxa detected fell into 5 bacterial phyla, that is, Firmicutes, Fusobacteria, Actinobacteria, Proteobacteria, and Bacteroidetes. Spirochaetes such as Treponema spp. were not detected in any of the samples by 16S rRNA gene analysis. Differences in the bacterial profiles of each patient are shown in Figure 3(b) and Table 2. The mean numbers of taxa/species detected in purulent and serous fluids were 28 (range: 13-41) and 12 (range: 3-23), respectively. As shown in Table 2, 5 of the 11 samples were negative for bacterial growth, regardless of whether inflammation was observed. Four of the five negative samples were collected from patients who had received antibiotics. The mean numbers of taxa/species detected by 16S rRNA gene analysis in POMC fluids of patients who had received macrolides and β-lactams were 27 (range: 6-39) and 23 (range: 17-29), respectively (Tables 1 and 2). Although S. sanguinis was identified in case 3 by culture method, this species was not detected by 16S rRNA gene analysis. Klebsiella pneumoniae, Streptococcus. anginosus, Peptoniphilus asaccharolyticus, and Finegoldia magna were identified in case 5, but they were not detected by 16S rRNA gene analysis. Prevotella melaninogenica is not the same sample as P. denticola even by biochemical methods, although both are Prevotella species. Streptococci viridans were also identified by the culture method for cases 10 and 11, but they were not detected by 16S rRNA gene analysis. Only 3 taxa/species were found in case 10 who had received no antibiotics prior to sample collection, that is, F. nucleatum (54%), Porphyromonas gingivalis (29%), and Porphyromonas spp. (17%) ( Table 2). In case 8, Enterococcus faecalis was the most prevalent bacteria detected by both the culture method and 16S rRNA gene analysis, comprising 35% of the total population (Table 2). Only in case 8, the species identified by the culture method matched that detected as the most prevalent bacteria species by 16S rRNA gene analysis.

Discussion
Although more than 600 species of bacteria are believed to be present in the oral cavity [7], the majority of these bacteria are VBNC [8], that is, bacteria not yet detected by currently available culture methods. Therefore, we conducted a comprehensive survey of the bacterial flora in POMC fluid by 16S rRNA gene analysis and culture methods to better determine the actual abundance of bacteria, including VBNC bacteria.
Infections originating in the oral cavity are complex because of the presence of a large numbers of bacterial species [12,13], which makes it difficult to determine the bacteria involved in most oral infections. As previously stated, there are 3 current theories regarding the formation of POMCs, and the relationship between the formation of POMCs, their progression, and bacteria is not well understood. Very few studies have identified the bacteria involved in POMC infections [4,9]. Bacterial infections may be closely linked to cyst growth and the progression of its symptoms because the maxillary sinus floor is adjacent to the oral cavity, and bacteria reach the cyst via the teeth and periodontal pockets despite the cyst cavity being isolated by the cyst wall. In this study, we aseptically collected intracystic POMC fluid and compared the accuracy of identification of bacteria by conventional culture methods and 16S rRNA gene analysis.
As shown in Figure 3(a), a high bacterial diversity was observed in the POMC fluid, and the bacterial flora in this fluid is shown for each patient (Figure 3(b), Table 2). All POMC fluids contained bacteria, and purulent fluids contained higher numbers of bacteria species/taxa than serous fluids (Tables 1 and 2), suggesting that oral bacterial infections of POMC may be involved in inflammatory symptoms, such as cheek swelling and pain caused by POMC, and symptom progression. Prevotella, Neisseria, Fusobacterium, Porphyromonas, and Propionibacterium were detected in more than 70% of all patients (Figure 1), while S. mitis, F. nucleatum, P. acnes, S. capitis, and S. salivarius were detected in more than 50% of all patients ( Figure 2). P. acnes was the dominant bacterium in 6 of 11 patients (Figure 3(b)). It is present on the skin, in the oral cavity, and the intestinal mucosa where it is known to be the causative agent of a wide range of diseases and opportunistic infections, including alveolar abscesses, maxillary sinusitis, osteomyelitis, meningitis, noma, pericarditis, sepsis, hepatitis, granuloma, acne, and orbital abscesses [14][15][16]. S. mitis, which was detected in 7 patients (Figure 3(b)), is an indigenous oral bacterium that is primarily present in the buccal mucosa and on tooth surfaces. It is a potential causative agent of endocarditis [17,18]. P. denticola, which was detected in 6 patients (Figure 3(b)), is an indigenous oral bacterium and commonly detected in subgingival biofilm of periodontitis patients [19], F. nucleatum, which was detected in 6 patients (Figure 3(b)), is associated with periodontitis, and according to Kuriyama et al. [20,21] it is often present in an abscess along with the anginosus group of Streptococci. Other bacterial species were detected, including E. faecalis, S. capitis, P. gingivalis, S. haemolyticus, S. parasanguinis, S. salivarius, and S. sinensis, but they were present at a lesser extent in multiple cases compared with the species mentioned above. E. faecalis was the dominant bacteria in case 8 (Table 2), and it is normally part of the indigenous bacterial flora in the human intestinal tract, although it can cause urinary tract infections, endocarditis, meningitis, and opportunistic infections [22]. E. faecalis is often found  in the teeth [23] during root canal treatment, where it is associated more with the asymptomatic initial period of root canal bacterial infections than with symptomatic root canal bacterial infections. It is significantly more common in cases of failed root canal treatment than of early root canal bacterial infections [24].
There are very few reports on the potential origin of the microbiota in POMC fluid. In case 8, the potential origin might be an endodontic source because only E. faecalis was detected by both the culture method and 16S rRNA gene analysis. Case 10 received no antibiotics, and F. nucleatum and P. gingivalis, which are closely related to periodontitis, were detected by 16S rRNA gene analysis, suggesting that the origin might be a periodontal source. However, Treponema spp. are often detected in periodontal diseases, and none were found in any of the samples. Further large-scale studies are required to understand the potential origin of the microbiota in POMC fluid.
are often observed during the first culture examination of oral clinical samples. Several colonies are selected and used for subsequent species identification tests. The results obtained by the culture method were different from those obtained by 16S rRNA gene analysis. There are several possible reasons for this result. It is most probable that the number of bacteria identified by conventional culture methods differs from the actual number present because the actual causative agents might not be detected in a culture medium as a result of their VBNC status or competition with other bacterial species for survival, thereby leading to a negative result. Long-term clinical administration of macrolide antibiotics is often followed in patients with maxillary sinusitis [25]. Macrolide antibiotics had been administered by general practitioners prior to definite POMC diagnosis in 6 of the 9 patients in our study. Thus, it is possible that some bacteria were eliminated by the antibiotics used. In addition, DNA extraction is not limited solely to living individuals because all environmentally present 16S rRNA genes are amplified by universal primers in community structure analysis. Therefore, it is possible that the remnants of DNA of dead bacterial cells were also amplified. We could not identify the possible causative bacteria or the susceptibilities of antimicrobial agents to the possible causative bacteria by 16S rRNA gene analysis; therefore, the conventional culture method was necessary to identify species and develop rational treatments for patients with odontogenic infections. In addition, odontogenic infections are often mixed, and the true pathogen may not be the dominant organisms in the pus where it may be hidden by larger numbers of opportunistic organisms. The numbers of low-prevalence taxa/species are shown in Table 2 as other spp. In addition, some of the lowprevalence taxa/species are also shown in Figures 1 and 2. However, the significance of the low-prevalence taxa/species for ecology or etiology remains unclear in this study.

Conclusions
We identified the bacterial flora present in POMC fluid by 16S rRNA gene analysis and conventional culture methods. Bacterial fell into 41 genera and 64 species were detected in the 11 patients with POMC bacterial infections, which may be closely involved in both inflammatory symptoms, such as cheek swelling and pain caused by POMC, and symptom progression. The results obtained by the culture method were different from those obtained by 16S rRNA gene analysis, but both approaches may be necessary for the identification of pathogens, especially of bacteria that are difficult to detect by culture methods, and the development of rational treatments for patients with POMC. Further large-scale studies are required to understand the potential origin of the microbiota in POMC fluid.