The authors have declared that no competing interests exist.
Conceived and designed the experiments: JAPCJ JSP KN RJS LMD PS JGM PJW. Performed the experiments: JAPCJ. Analyzed the data: JAPCJ JSP KN RJS LMD PS JGM. Contributed reagents/materials/analysis tools: JAPCJ JSP KN PJW. Wrote the paper: JAPCJ PS JGM PJW.
There is increasing evidence to suggest that the sinus microbiome plays a role in the pathogenesis of chronic rhinosinusitis (CRS). However, the concentration of these microorganisms within the sinuses is still unknown. We show that flow cytometry can be used to enumerate bacteria and virus-like particles (VLPs) in sinus flush samples of CRS patients. This was achieved through trialling 5 sample preparation techniques for flow cytometry. We found high concentrations of bacteria and VLPs in these samples. Untreated samples produced the highest average bacterial and VLP counts with 3.3 ± 0.74 x 107 bacteria ml-1 and 2.4 ± 1.23 x 109 VLP ml-1 of sinus flush (n = 9). These counts were significantly higher than most of the treated samples (p < 0.05). Results showed 103 and 104 times inter-patient variation for bacteria and VLP concentrations. This wide variation suggests that diagnosis and treatment need to be personalised and that utilising flow cytometry is useful and efficient for this. This study is the first to enumerate bacterial and VLP populations in the maxillary sinus of CRS patients. The relevance of enumeration is that with increasing antimicrobial resistance, antibiotics are becoming less effective at treating bacterial infections of the sinuses, so alternative therapies are needed. Phage therapy has been proposed as one such alternative, but for dosing, the abundance of bacteria is required. Knowledge of whether phages are normally present in the sinuses will assist in gauging the safety of applying phage therapy to sinuses. Our finding, that large numbers of VLP are frequently present in sinuses, indicates that phage therapy may represent a minimally disruptive intervention towards the nasal microbiome. We propose that flow cytometry can be used as a tool to assess microbial biomass dynamics in sinuses and other anatomical locations where infection can cause disease.
Chronic rhinosinusitis (CRS) is a common disease amongst the human population, and there is increasing evidence to show that microorganisms are involved in the inflammation of the sinus mucosal layer leading to exacerbation of the disease [
Flow cytometry has been used as a method for enumerating heterotrophic bacteria and virus-like particles (VLPs) in environmental samples for decades [
Maxillary sinus flush fluid samples were obtained from nine patients diagnosed with CRS in accordance to criteria defined by the Chronic Rhinosinusitis Task Force [
Immediately after opening the maxillary sinus, approximately 5 ml of sterile saline was used to flush the sinus and re-collected in sterile specimen containers. Volume of flush fluid collected ranged from approximately 2 to 4 ml. Once samples were collected, they were transported on ice ready for immediate fixation with glutaraldehyde (0.5% final concentration) on ice in the dark, then snap freezing in liquid nitrogen and storage at -80°C until analysis [
Five sample preparation techniques for flow cytometry were investigated. Fixed sinus flush fluid samples were thawed before each treatment was applied.
Sputasol was made using 0.02 μm filtered MilliQ water according to the manufacturer’s instructions (Oxoid). Equal volumes of Sputasol and fixed sinus sample were mixed together then incubated at 37°C for 15 minutes.
Methanol, 0.2 μm filtered, was added to fixed sinus samples to a final concentration of 20% [
Potassium citrate tribasic solution (1M, Sigma) was added to the fixed sample to a 1% final concentration [
Sodium pyrophosphate solution was added to 100 μl of fixed sample to a final concentration of 10 mM [
Fixed samples were diluted in 0.2μm filtered TE buffer (10 mM Tris, 1 mM EDTA, pH 7.4) without pre-treatment [
Bacterial and VLP populations present in sinus flush fluid were identified and enumerated using a BD ACCURI C6 flow cytometer (Becton Dickinson). Samples using each extraction technique were run in triplicate for each patient. Samples were diluted (1:100) in 0.2 μm filtered TE buffer, stained with the DNA-binding dye SYBR-I Green (1:20,000 final dilution; Molecular Probes) then incubated at 80°C in the dark for 10 minutes [
Samples were analysed using an Accuri C6 flow cytometer (Accuri C6) and BD ACCURI CFlow software. All samples were run for 2 minutes at machine fluidics setting of fast, with the threshold set to FL-1 (green fluorescence). As a control, 1 μm diameter fluorescent beads (Molecular Probes) were used. Beads were added to each sample to a final concentration of 105 beads ml-1 [
Flow cytometry data was analysed using FlowJo software (Tree Star, Inc.). VLP and bacterial populations were categorised based on variations in side scatter, a representation of cell size, and SYBR Green fluorescence, an indication of nucleic acid content [
Rank abundance plots were generated for bacterial and VLP concentrations using all method triplicates and their averages to distinguish between any patient groupings formed on abundance. Comparisons between bacterial and VLP abundances for each treatment method were made using the statistical analysis program SPSS version 22 (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp.) using the Wilcoxon signed-rank test. Statistical significance between treatments was considered when p < 0.05.
Cytograms showed discrete bacterial and VLP populations present within the sinus fluid of CRS patients (
Representative cytogram shows the VLP and bacterial populations in a patient’s untreated sinus wash.
Error represents standard error of the mean.
Treatment | Bacteria ml-1 (±SE) | VLP ml-1 (±SE) |
---|---|---|
Untreated | 3.3 x 107 (7.4 x 106) | 2.4 x 109 (1.2 x 109) |
Sodium pyrophosphate | 2.9 x 107 (6.7 x 106) | 2.0 x 109 (9.9 x 108) |
Sputasol | 2.2 x 107 (6.5 x 106) | 1.8 x 109 (9.1 x 108) |
Methanol | 2.0 x 107 (4.3 x 106) | 2.2 x 109 (9.9 x 108) |
Potassium citrate | 1.9 x 107 (4.6 x 106) | 2.2 x 109 (1.1 x 109) |
For patient samples, the mean bacterial abundance for untreated samples was the highest of all treatments, with 3.3 ± 0.74 x 107 cells ml-1 (n = 27;
VLP mean abundance for the untreated method was the highest for all patient samples with 2.4 ± 1.2 x 109 cells ml-1 (n = 27;
Rank abundance was used to identify possible groupings among the patient’s bacterial abundance. Breaks in the plot suggest 3 groupings of patients with bacterial abundances classified as high, greater than 107 cells ml-1, medium, between 105 to 106 cells ml-1, and low, less than 105 cells ml-1 (
Three clear groups of patients with high, medium and low bacterial abundances are apparent. Differences in treatments used on the samples can be seen not to influence steps of bacterial abundance for patients.
Data points follow a logarithmic trend achieved by steps of power laws for each observed group. High medium and low bacterial concentration groups fit the power law equations y = 2E+08x-0.84 (R2 = 0.84), y = 5E+10x-3.57 (R2 = 0.98) and y = 6E+09x-3.49 (R2 = 1) respectively.
A rank abundance plot for VLP abundance was generated using the method triplicates of untreated, sodium pyrophosphate, potassium citrate, Sputasol and methanol samples (
Highest abundances are rare while lower abundances are common. Differences in treatments used on the patient samples can be seen not to influence the high range of VLP origination levels.
Data points follow a power law.
Prior to sinus surgery, each patient completed a questionnaire regarding basic clinical information and provided a severity score from 0, being no problem, to 5, being a problem as bad as it can be, for CRS symptoms for the past two weeks. Based on the rank abundance plots (Figs
This is the first study to use flow cytometry to enumerate bacteria and VLPs within the maxillary sinus of CRS patients. We present a number of snapshot enumerations, using flow cytometry, of the microbial composition of sinuses of CRS patients. We tested a number of sample preparation techniques for bacterial and VLP enumeration that are used for environmental samples, in particular, techniques used for disruption of coral mucus, for microscopy [
Although the primary focus of this study was to enumerate the bacteria and VLPs in the sinus fluid, there was concern surrounding the presence of small fragments of mucus or biofilms within the samples prepared for flow cytometry. The mucus may have caused the bacteria and VLPs to clump together resulting in an overestimation on particle size, shape and DNA content. This is a similar concern in regards to analyzing bacteria and VLPs in coral mucus using microscopy [
Common chemicals used in environmental sample preparation include potassium citrate, sodium pyrophosphate and methanol. Sodium pyrophosphate and potassium citrate are commonly used in environmental microbiology for desorbing viral particles from soil and marine sediment [
Our results show that the untreated and sodium pyrophosphate treatment methods yielded significantly higher bacterial abundances than all other methods tested (p < 0.05). For VLP enumeration, no treatment (as in the untreated samples) was the optimal method. Although there was no significant difference between untreated and sodium pyrophosphate treated samples (p < 0.05), sodium pyrophosphate did not yield significantly higher VLP abundances than methanol and potassium citrate (p < 0.05). This result contrasts to previous microscopy studies which found potassium citrate better for viral enumeration in coral mucus [
Previous studies have shown that the human sinus is colonized with an array of microbes [
Large variations were observed between patients bacterial and VLP concentrations, which is not uncommon with human microbial flora studies [
For the bacterial rank abundance, the three orders of magnitude range among the 9 patients may reflect temporal variation or that the bacterial abundances are defined by processes or a variable that was not measured. The skew of the distribution towards the lower concentrations in the VLP rank abundance is consistent with the highly episodic nature of viral infections, particularly in bacteriophage where large burst sizes can quickly reduce the concentrations of particular bacterial species, leaving high bacteriophage concentrations at least temporarily [
Knowledge of the abundance of microorganisms in CRS will further our understanding of the disease as the presence of certain bacterial species does not always imply infection. The aim of his research was to produce a snapshot enumeration of the sinus microbes of 9 patients known to suffer from CRS, and to determine if they had similar abundances of bacteria and VLPs. Within the group of 9 patients sampled, there was 3 orders of magnitude difference in abundance for bacterial populations and almost 4 orders of magnitude difference for VLPs. This suggests that not all CRS patients are infected at the same level of bacteria and VLPs. Knowledge of the differences in bacterial abundances may facilitate the development of personalised treatment options.
This work indicates the potential for future studies in other microbial disease related health conditions. We propose that flow cytometry has potential as a tool to monitor microbial dynamics in patients and in future may assist in determining appropriate dosages required when treating microbial related health conditions.
(TIF)
Replicates (Rep) for each method are shown.
(PDF)
Replicates (Rep) for each method are shown.
(PDF)
We would like to thank all the staff at the Flow Cytometry Unit in the Flinders Medical Centre for their technical support throughout the duration of this study. We would also like to extend our gratitude to the tissue registrars in Professor Wormald’s research group for conducting the patient surveys and their assistance in sample collection.