SEM indicates scanning electron microscopy.
Circles represent means, and error bars, standard deviations. CFU indicates colony-forming units.
Circles represent means, and error bars, standard deviations. CFU indicates colony-forming units.
Circles represent means, and error bars, standard deviations. S aureus total (planktonic + biofilm) counts were significantly higher without oxacillin compared with all the other treatments with oxacillin (P < .001). CFU indicates colony-forming units.
A, Ethylene oxide treatment. B, Ciprofloxacin treatment at 0 µg/mL. C, Ciprofloxacin treatment at 10 µg/mL. D, Ciprofloxacin treatment at 3000 µg/mL.
Esin L, Antonelli PJ, Ojano-Dirain C. Effect of Haemophilus influenzae Exposure on Staphylococcus aureus Tympanostomy Tube Attachment and Biofilm Formation. JAMA Otolaryngol Head Neck Surg. 2015;141(2):148-153. doi:10.1001/jamaoto.2014.3208
Tympanostomy tube (TT) biofilm formation may lead to sequelae.
To determine whether the acute pathogen Haemophilus influenzae promotes TT attachment and biofilm formation by the chronic pathogen Staphylococcus aureus.
Design and Setting
Controlled, in vitro microbiological study at an academic research laboratory using TTs treated with 0 (untreated), 10, or 3000 µg/mL ciprofloxacin or ethylene oxide and TTs with and without prior H influenzae exposure.
Fluoroplastic TTs (18 per treatment) were cultured with H influenzae. The TTs were gas-sterilized or exposed to 0, 10, or 3000 µg/mL ciprofloxacin. One-third of the TTs from each treatment group underwent H influenzae counts or scanning electron microscopy (SEM). Another one-third were used for an S aureus attachment assay. The remainder, as well as TTs not exposed to H influenzae, were cultured with S aureus and then treated with oxacillin to kill planktonic S aureus. S aureus counts and SEM were performed.
Main Outcomes and Measures
Attachment and biofilm formation of S aureus on TTs assessed by quantitative bacterial counts and SEM.
Mean (SD) H influenzae counts were lower on TTs treated with 3 mg/mL than with 10 µg/mL ciprofloxacin (2.06 × 103 [1.00 × 103] vs 4.21 × 105 [1.67 × 105]; P < .001). Mean (SD) S aureus attachment was higher on TTs with untreated preexisting H influenzae (8.88 × 105 [3.20 × 105]; P < .001) and lower on TTs with prior exposure to H influenzae treated with 10 (3.43 × 104 [2.10 × 104]; P = .006) or 3000 µg/mL ciprofloxacin (6.41 × 102 [3.59 × 102]; P < .001). S aureus biofilm formation was similar across groups, except TTs with prior exposure to H influenzae treated with ciprofloxacin 3 mg/mL, which had significantly less (9.30 × 101 [3.51 × 102]; P < .001).
Conclusions and Relevance
Exposure to live H influenzae may promote S aureus attachment on TTs. Treatment of H influenzae on TTs with ciprofloxacin, 3 mg/mL, as found in ototopical therapy, may reduce subsequent S aureus attachment and biofilm formation.
Tympanostomy tubes (TTs) are placed in more than 600 000 children in the United States each year.1 Use of TTs may be complicated by posttympanostomy tube otorrhea (PTTO) and occlusion, which have reported rates as high as 83% and 74%, respectively.2,3 Biofilm formation on the TT surface has been implicated in the development of and complications in the treatment of these sequelae.4- 6
Posttympanostomy tube otorrhea in young children (ie, younger than 3-6 years) typically involves pathogens that cause acute otitis media, including Streptococcus pneumonia and Haemophilus influenzae.7- 9 In contrast, PTTO in older children more commonly involves pathogens found in chronic suppurative otitis media, including Staphylococcus aureus and Pseudomonas aeruginosa.7- 9H influenzae and the chronic pathogens are known to be capable of biofilm formation.10 The transition from acute pathogens to chronic pathogens is concerning because chronic pathogens, and disease, are more difficult to eradicate. Although this transition may simply represent attrition of more susceptible microbes and replacement by more resistant microbes,11 there is evidence that the presence of an early colonizer may promote later colonization by other species.12
We have previously found that the presence of H influenzae does not promote P aeruginosa biofilm formation in vitro.13 The present study was conducted to determine whether H influenzae promotes TT attachment and biofilm formation by S aureus.
As this study involved neither human nor animal subjects, it was exempt from institutional review. Fluoroplastic TTs (Medtronic) were exposed to samples of human plasma (donated by a local hospital) to promote biofilm formation on TTs and to simulate conditions during TT placement.14 The study design is shown in Figure 1. The TTs were cultured with H influenzae in microtiter plates for 7 days, then exposed to ethylene oxide gas sterilization (yielding complete eradication of viable H influenzae without loss of residual cellular debris), ciprofloxacin (10 or 3000 µg/mL, akin to in vivo exposure with systemic and topical administration, respectively) for 24 hours, or no treatment (control). One-third of the TTs from each treatment group underwent analysis (ie, bacterial counts or scanning electron microscopy [SEM]), without exposure to S aureus. Another one-third were used for a 2-hour S aureus attachment assay. The remainder, as well as fresh TTs not exposed to H influenzae, were cultured with S aureus for 2 days and then treated with 1 mg/mL oxacillin sodium to kill planktonic organisms to assess subsequent S aureus biofilm formation. A group of TTs cultured with S aureus was not exposed to oxacillin to represent total (planktonic and biofilm) bacterial load.
Bacterial growth was quantified by means of bacterial counts (16 samples per treatment) and SEM (2 samples per treatment). A total of 252 TTs were used, divided accordingly among the 14 treatment groups (Table). Because of the large numbers of samples involved, the study was run in 2 batches, with equal distribution of samples.
H influenzae strain 62094 and S aureus strain 29213 were isolated aseptically from quad-streaked chocolate agar and tryptic soy agar plates, respectively. Bacterial colonies were examined for correct morphology. The culture medium for H influenzae was brain-heart infusion (Becton Dickinson) supplemented with 2 µg/mL nicotinamide adenine dinucleotide (Sigma) and 10 µg/mL hemin (Remel). The TTs were dipped in 200 μL of log-phase H influenzae for 5 minutes and then provided with 200 μL fresh culture medium. The TTs were cultured with H influenzae in 96-well microtiter plates at 37°C for 7 days because our earlier study10 revealed mature H influenzae biofilms after 4 days. Nutrient broth was replaced daily. After 7 days, TTs were treated with either ethylene oxide gas sterilization, ciprofloxacin (10 or 3000 µg/mL), or no antimicrobial agent (control). One-third of the TTs were used to quantify bacterial growth using quantitative counts and SEM. No bacterial counts were conducted on 7-day H influenzae after gas sterilization because preliminary studies consistently showed no bacterial growth after ethylene oxide sterilization. The TTs treated with ethylene oxide gas sterilization were ensured adequate ventilation prior to use in S aureus attachment and biofilm assays.
An aliquot of S aureus overnight culture was grown to logarithmic phase, and a working solution, 500 μL of log-phase S aureus and 19.5 mL of tryptic soy broth (MP Biomedicals), was prepared. One-third of the TTs with preexisting H influenzae biofilm that were either gas-sterilized or treated with 0, 10, or 3000 μg/mL ciprofloxacin were transferred individually in borosilicate culture tubes and inoculated with 250 μL of the S aureus working solution. Fresh TTs not exposed to H influenzae (S aureus control) were also exposed to human plasma, then inoculated with 250 μL of S aureus working solution. The culture tubes containing the TTs were transferred to a water bath set at 37°C with constant agitation at 100 rpm.
After 2 hours, TTs were removed from the water bath and were washed 3 times with 500 µL of 1-M phosphate-buffered saline (PBS). The PBS washes were aspirated using a sterile glass Pasteur pipette (Fisher Scientific). Two TTs were processed for SEM and 16 TTs were transferred to 15-mL flip-top conical tubes (Thermo Fisher Scientific) containing 2 mL of PBS with 5 ppm Tween-80 (Fisher Chemical) to assess S aureus attachment. Samples from the TTs with prior H influenzae exposure were plated on both tryptic soy agar and chocolate agar plates to determine S aureus and H influenzae colony counts, respectively. The SEM and bacterial count procedures are described in Assessment of Biofilm Formation by Means of Quantitative Bacterial Counts and Scanning Electron Microscopy.
The remaining one-third of TTs with preexisting H influenzae with or without gas sterilization or ciprofloxacin treatment, to be exposed to S aureus, were placed in 96-well microtiter plates containing 200 μL log-phase S aureus in brain-heart infusion supplemented with 2 µg/mL nicotinamide adenine dinucleotide and hemin to allow both H influenzae and S aureus to grow. The plates were maintained in the incubator at 37°C for 2 days because mature S aureus biofilms are formed after 2 days.7 Nutrient broth was replaced daily. After 2 days, oxacillin (1 mg/mL, Sigma) was added to the broth to eradicate planktonic organisms.
Each TT was aseptically transferred into individual conical tubes with 2 mL of PBS with 5 ppm Tween-80. The conical tubes were placed into a water bath and sonicated (Branson 2510, Branson Ultrasonics) for a total of 7.5 minutes, with serial 1.5-minute sonication exposures separated by a 1-minute rest. After sonication, tubes were vortexed (setting 8) for 30 seconds, serially diluted, and spread plated in triplicate on chocolate or tryptic soy agar plates for H influenzae and S aureus, respectively. Specimens with prior exposure to H influenzae were plated on both chocolate and tryptic soy agar plates to check for the presence of H influenzae after subsequent S aureus culture. Plates were incubated for 18 to 24 hours at 37°C, and colonies were counted manually. Data were expressed as colony-forming units (CFUs) per milliliter.
Representative TT samples not processed for bacterial counts were fixed in 1 mL Trumps fixative (1% glutaraldehyde, 4% formaldehyde in PBS) and stored at 4°C until processed. Specimens were washed 3 times with PBS for 10 minutes, then fixed for 1 hour in 1% osmium tetroxide in PBS (Electron Microscopy Sciences). Specimens were washed once with PBS and 3 times with deionized water for 10 minutes each. Specimens were dehydrated in ethanol series for 10 minutes each (25%, 50%, 75%, 95%, 100%) and then hexamethyldisilazane (Electron Microscopy Sciences) for 5 minutes. Specimens were allowed to air dry overnight. The TTs were then cut in half to allow visualization of their inner and outer surfaces. Specimens were sputter-coated with gold/palladium with argon gas (Desk II sputter coater, Denton Vacuum USA) for 45 seconds and stored under vacuum until imaged (Phenom scanning electron microscope, FEI Company).
Mean CFUs were analyzed using a 2-tailed t test (comparison of 2 treatment groups) and 1-way analysis of variance (comparison of 3 or more groups) followed by Student t test or Dunnett multiple-comparison test for comparisons of means. P ≤ .05 was considered significant. All statistical analyses were carried out using JMP, version 10.0 (SAS Institute Inc).
High levels of total (planktonic and biofilm) H influenzae were observed on fluoroplastic TTs following H influenzae exposure, without ciprofloxacin treatment (Figure 2). Tympanostomy tubes treated with the ototopical dose (3 mg/mL) and the systemic dose (10 µg /mL) had lower H influenzae counts compared with TTs not treated with ciprofloxacin (P < .001). Viable H influenzae persisted on TTs treated with 3 mg/mL ciprofloxacin, albeit more than 2 logs lower than on TTs treated with 10 µg/mL ciprofloxacin, consistent with functional criteria for biofilm formation.
Viable H influenzae also remained on TTs after the 2-hour S aureus attachment assay, with the counts following the same trend as the H influenzae counts after ciprofloxacin exposure, prior to the attachment assay (ie, H influenzae counts decreased with increasing ciprofloxacin concentration; P < .001). Among TTs exposed in the 2-hour S aureus attachment assay, TTs with untreated preexisting H influenzae had higher S aureus attachment (P < .001) (Figure 3) compared with TTs without preexisting H influenzae. In contrast, S aureus attachment on ethylene oxide–sterilized TTs was not different than on TTs without preexisting H influenzae (P = .69). The TTs with preexisting H influenzae that were treated with 10 µg/mL and 3 mg/mL ciprofloxacin had approximately 1.5 and 3.5 log lower S aureus counts, respectively, compared with TTs without prior H influenzae exposure (P = .006 and < .001).
After 2 days of S aureus exposure, followed by oxacillin treatment, viable H influenzae was isolated from TTs with prior exposure to H influenzae that had been treated with 0 or 10 µg/mL ciprofloxacin, whereas H influenzae–exposed TTs treated with 3 mg/mL ciprofloxacin yielded none (P < .001; data not shown). All TTs treated with oxacillin had lower S aureus counts compared with TTs without prior H influenzae exposure that had not been exposed to oxacillin (P < .001; Figure 4), consistent with elimination of planktonic organisms. S aureus biofilm formation, as measured by viable S aureus CFUs after oxacillin exposure, was not different among treatment groups, except for TTs treated with 3 mg/mL ciprofloxacin, which had significantly lower S aureus counts (P < .001). The minimum inhibitory concentrations of ciprofloxacin and oxacillin against H influenzae and S aureus, respectively, were 0.06 and 0.244 µg/mL. Thus, the antibiotic concentrations used to eradicate planktonic bacteria for each bacterial pathogen far exceeded the observed minimum inhibitory concentrations for the specific antibiotic.
Scanning electron microscopy showed that, qualitatively, the number of H influenzae microbes seen after 7-day culture was inversely dependent on the ciprofloxacin concentration, similar to our previous study.13 Ethylene oxide–sterilized TTs had abundant cellular and extracellular material and lysed bacteria, consistent with H influenzae biofilm. Scanning electron micrographs yielded similar findings after the 2-hour S aureus attachment assay, with TTs treated with 10 and 3000 µg/mL ciprofloxacin containing fewer bacteria than ethylene oxide–treated TTs and TTs without H influenzae exposure. After a 2-day S aureus exposure, H influenzae-exposed TTs treated with 3000 µg/mL ciprofloxacin had fewer bacteria than those that received the other treatments (Figure 5).
It has long been postulated that inadequately treated acute otitis media predisposes the patient to chronic suppurative otitis media by way of a tympanic membrane perforation or TT.15 This view is supported by the relatively low rate of isolation of S aureus and P aeruginosa from middle ears with otitis media and intact tympanic membranes16 and a relatively high rate in those with otitis media and nonintact tympanic membranes.9,17 With the exception of limited data from animal models,18 there have been few experimental data to support or refute this.
In the present study, we sought to elucidate the impact of TT exposure to H influenzae on attachment and biofilm formation by S aureus. We observed higher S aureus counts on TTs that had been exposed to viable H influenzae than on TTs that had been exposed to H influenzae killed by ethylene oxide or ciprofloxacin. This suggests an active, synergistic process in which H influenzae promotes the attachment of S aureus, such as cell-to-cell signaling, which is a key initial step in biofilm formation.19 This did not, however, lead to higher levels of S aureus biofilm when measured after 2 days of exposure. The reason for this may be that S aureus forms a mature biofilm on TT materials (ie, microbial counts plateau) by 2 days.10 Whereas this was the basis for our choosing this time frame, it may well have been too late to observe an effect. Furthermore, if the same rapid biofilm development occurs clinically, interventions aimed at preventing subsequent S aureus biofilm formation may have a very narrow window for initiation.
In our prior study evaluating the influence of H influenzae and P aeruginosa biofilm development, we similarly observed no impact of H influenzae exposure on P aeruginosa biofilm formation; however, our cell counts were made after 4 days of P aeruginosa exposure.13 We did not measure the effect of H influenzae exposure on P aeruginosa attachment. Because P aeruginosa biofilm cell counts also plateau after 2 days, it is possible that we missed any early effects.
The literature on a possible synergistic link between H influenzae exposure and S aureus growth is mixed. Introduction of the pneumococcal vaccine has been met with an increase in S aureus nasopharyngeal carriage,20H influenzae otitis media,21 and S aureus acute otitis media with spontaneous tympanic membrane perforation in vaccinated children.22 However, in the nasopharynx, there appears to be a negative association between S aureus and H influenzae,23,24 or if any effect, a positive influence of S aureus on H influenzae growth (ie, the converse to what was studied here).25
The second important observation from the present study was the influence of the TT sterilization method on subsequent S aureus biofilm formation. Although samples were effectively washed following H influenzae and ciprofloxacin exposure, use of high-dose ciprofloxacin, equivalent to ototopical therapy, demonstrated a residual inhibitory effect on S aureus counts. We have previously observed this phenomenon with ototopical agents, most notably with ciprofloxacin, on subsequent biofilm development.13,26 This lends additional credibility to the relative value of ototopical therapy over systemic antimicrobial therapy for PTTO.27
Exposure to live H influenzae may promote S aureus attachment on TTs. Treatment of H influenzae on TTs with high-dose ciprofloxacin 3 mg/mL, as found in ototopical therapy, may reduce subsequent S aureus attachment and biofilm formation. Further investigation is needed to determine the clinical significance of these findings.
Submitted for Publication: April 1, 2014; final revision received June 3, 2014; accepted July 4, 2014.
Corresponding Author: Carolyn Ojano-Dirain, PhD, Department of Otolaryngology, University of Florida College of Medicine, 1345 Center Dr, PO Box 100264, Gainesville, FL 32610-0264 (firstname.lastname@example.org).
Published Online: December 18, 2014. doi:10.1001/jamaoto.2014.3208.
Author Contributions: Dr Ojano-Dirain had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Antonelli, Ojano-Dirain.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Esin, Ojano-Dirain.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Ojano-Dirain.
Administrative, technical, or material support: All authors.
Study supervision: Esin, Ojano-Dirain.
Conflict of Interest Disclosures: Dr Antonelli has previously received financial support from Medtronic ENT (grant support and consulting fees) and Alcon Laboratories (grant support and honoraria for speaking engagements and advisory board service). No other disclosures are reported.
Funding/Support: Financial support for this study was provided by the University of Florida. Tympanostomy tubes were provided by Medtronic ENT.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentation: This study was presented at the Triological Society Combined Sections Meeting; January 11, 2014; Miami, Florida.
Additional Contributions:H influenzae strain 62094 and S aureus strain 29213 were donated by David Stroman, PhD, Alcon Research, Ltd.