Background
Chronic rhinosinusitis is a disease cluster with a significant societal burden, and despite extensive research efforts, has an unknown pathophysiology. There is emerging evidence that microorganisms play an important role in the exacerbation and perpetuation of mucosal inflammation. However, the microbial biodiversity in disease and controls is not well defined. Furthermore, the importance of microorganism abundance, and potential relationship to disease manifestation is unknown. To further our understanding of the role of microorganisms in CRS, it is important to comprehensively characterize the resident microbial community in healthy and diseased tissue and examine the specific host immunological responses to these organisms. Thus, we sought to characterize the microbial populations in CRS and controls to establish a basis for further species directed research into this heterogeneous disease. Only through such a systematic approach can we determine the importance, or otherwise, of these microorganisms in the disease phenotypes. Furthermore, comparative microbiome studies will provide important information for the selection of antimicrobial therapies, and enable the determination of the effectiveness of such treatments.
Increasingly, we are discovering that complex polymicrobial communities exist, especially at the host-environment interface such as mucosal surfaces, and the majority of these bacterial species are refractory to culture [
1]. Many of these organisms are found to be residing in complex communities known as biofilms; communities of organisms surrounded by a self-produced exopolysaccharide matrix, irreversibly attached to a live or inert surface [
2]. There have been recent advances in our understanding of biofilms in CRS, and their importance regarding disease evolution following treatment [
3], increased postoperative infection [
3], and altered host immune mechanisms [
4]. Many biofilm organisms are resistant to culture [
5], and their detection requires specialized techniques [
6]. Phenotypic differences which occur between planktonic and biofilm based organisms, may contribute to the relative incapacity of biofilm associated organisms to grow on nutrient media [
7,
8].
Traditional culture-dependant techniques have been the mainstay of microbial diagnostics in CRS. However, the reliance of cultivation on nutrient media often results in ‘enrichment bias’ with detection of a narrow range of microbes which is not representative of the actual diversity present, particularly in environmental samples [
9]. Competition between organisms during enrichment often results in dominance of one or two organisms with the fastest growth rates [
9]. Selective media techniques use nutrient restriction to enhance or restrict growth of organisms based on inherent microbial characteristics for identification. In a complex microbial community such as the diseased sinonasal mucosa, the identification of every organism using this method would be exhaustive. Additionally, many organisms may not thrive on nutrient media once the advantages of biofilm structures, and symbiotic relationships are lost.
The objective of this study was to determine the relationship between disease manifestation, and the biodiversity and abundance of mucosal microorganisms in CRS patients and controls using broad-based molecular diagnostics and conventional culture. Additionally, we sought to determine the specificity of contemporary biofilm detection methods by comparing them with molecular detection techniques in the same patients. Whilst our understanding of the importance of these communities in CRS, and their capacity to benefit or harm the host is in it’s infancy, the investigation is fundamental to furthering our understanding of this disease.
Discussion
This study compares the microbial bioburden in CRS patients, with healthy controls using three markedly different, but complementary detection techniques. We have shown that conventional laboratory culture has a tendency to polarize the detected microbes, selecting for abundant, rapidly growing aerobic organisms with favourable growth characteristics, such as S. aureus and S. epidermidis. Biosensor molecular detection allows a more comprehensive analysis of the microbial community and has the advantage of not requiring a priori knowledge of the flora. S. aureus was the most commonly detected organism in CRS patients, and was relatively more abundant in CRS patients compared to controls. It was detected with high sensitivity with FISH compared to the biosensor, but at lower levels by conventional culture. We have shown that molecular quantification can provide additional information with respect to microbial abundance, and comparisons between health and disease may assist our understanding of the role of these organisms in pathogenesis. Additionally we have shown that the detection of organisms using FISH and conventional culture is significantly dependant on microbial abundance as measured using molecular quantification. Fungi were present uncommonly, in a select group of nasal polyp patients.
Staphylococcus aureus was the most prevalent organism in CRS patients using molecular detection, followed by
Staphylococcus epidermidis and
Propionibacterium acnes. When the microbial density was compared for these three organisms between CRS and control patients,
S. aureus was present at approximately 10 times higher genomes per sample in CRS patients, whereas
S. epidermidis and
P. acnes were found in similar abundance between both patient groups. The increased bioburden of
S. aureus in the disease group is of particular interest, as it is emerging as a prominent disease modifying organism in CRS and it’s presence in patients has important clinical implications [
14,
15]. The capacity of
S. aureus to exist within dense mucosal biofilms has been documented [
11,
16], which may explain the variable detection from clinical specimens using traditional culture techniques.
A multitude of other studies have recently reported CRS microbiological data using a variety of techniques, including molecular diagnostics and FISH. Comparison with these studies is of interest to further our understanding of this disease. In the current study,
Haemophilus influenzae was detected at relatively low levels (13%) in CRS patients, and was not detected in controls. This is in agreement with Stephenson
et al., who detected
H. influenzae in 17% of CRS patients [
17], but contrasts with earlier FISH based studies which proposed it as the dominant organism in CRS [
18,
19].
Pseudomonas aeruginosa was also found infrequently in the current study (8%), confirming the findings of Stephenson
et al. These results contrast with a recent molecular study which found
P. aeruginosa to be the dominant organism in CRS patients, however control patients were not assessed [
20]. The disparity in organism profiles between studies may reflect regional variation, patterns of antimicrobial use, methodological differences, or disease severity patterns.
The disease burden in the current study was relatively high as measured by the radiological severity and the rate of non-primary surgery, which were at least twice as high as others [
17,
20], reflecting the tertiary nature of the practice. Some microbial biofilms have been associated with a more severe disease process than others [
16], and the microbial community in this study may reflect some bias towards the more severe end of the CRS spectrum. However, this is arguably the population which is most resistant to current treatment paradigms, and therefore of utmost importance to investigate.
The high prevalence of anaerobes detected in CRS patients is in agreement with previous molecular studies [
17,
20], however we also found high rates in controls, with similar abundance, casting doubt on a direct pathogenic role. There was poor agreement between the Ibis biosensor and conventional culture data for anaerobes, which reflects the paucity of culture-based studies reporting anaerobic species in CRS. These different detection methodologies seem to have comparable efficacy at detecting abundant, fast growing organisms. The pathologic importance of these species remains to be determined however.
Through the use of molecular detection we are beginning to understand the natural flora of non-diseased sinuses. Contrary to previous reports of sterility [
21,
22], we have shown that all healthy sinuses in this study are associated with a microbial community, and many of these organisms are also found in diseased sinuses. Nevertheless, we have shown that the abundance of organisms is significantly greater in CRS patients compared to controls - a phenomenon which requires further investigation with greater numbers of patients.
The cultivation of microorganisms using traditional culture has many pitfalls which may explain it’s limited utility in describing polymicrobial communities such as that within diseased sinuses. To be detected, organisms must grow on media after being removed from the native mucosal surface, with significant environmental changes in temperature, pH, nutrient sources, and without the complex dynamics of polymicrobial communities and host immune systems. Molecular methods such as the Ibis biosensor offer great potential for analysis of microbial diversity in CRS, which is untempered by the limitations of conventional culture. Molecular methods are sensitive and accurate and provide a more complete view of the bacterial communities present, and even provide a molecular antibiogram to guide treatment decisions. An important caveat however, is the capacity of molecular techniques to detect the genetic material of non-viable microorganisms. This must be taken into consideration when comparing between these, and culture-based techniques. Access to molecular detection instrumentation and expertise may also limit it’s utility in the clinical setting, but this is rapidly improving.
The formation of biofilms also impacts on the culture rates from patients. Organisms that form these immobilized consortia often undergo phenotypic transformations with reduced metabolic activity, which impacts the capacity to grow on selective media [
2]. Using the model organism
S. aureus, we have shown significantly increased detection rates using both FISH and molecular detection compared to cultivation techniques. Similar to previous studies [
11,
23,
24], biofilm was not detected in the control patients, despite detection of
S. aureus in 2/6 controls by the Ibis biosensor.
In situ hybridisation techniques have previously been reported to have reduced sensitivity in detecting low copy number nucleic acids of scarce bacteria [
1]. In an attempt to counter this, we have employed protein nucleic acid (PNA) FISH which has higher affinity and stability for microbial DNA sequences than conventional FISH probes [
25]. Despite this however, some of the CRS patients, and both controls with
S. aureus detected at low DNA copy number by Ibis, were not detected by FISH. This limitation not with standing, we have shown a good correlation between FISH and biosensor detection of
S. aureus in CRS patients, with high sensitivity and specificity, suggesting it has good clinical and research applicability.
The prevalence of fungi in CRS and healthy control patients has long been debated [
26]. Many studies have found a predominance of fungi in both patient groups when the nasal cavity is sampled [
27‐
29]. It is possible that these studies are detecting inhaled environmental fungi, which is trapped within nasal mucus en-route to the oropharynx. When sinus mucosa is specifically analysed, studies suggest a lesser prevalence in CRS patients and absence in controls [
30,
31]. The characterization of fungal species in CRS is far less advanced than for bacteria, and whether or not certain fungal species are more prevalent or important in CRS is still unknown. Therefore we employed a pan-fungal FISH probe to detect all species, in conjunction with culture and molecular detection. We found fungi in a small proportion of the sinus mucosa of CRS patients using all three detection methods, and an absence in control patients. The three methods showed similar sensitivity. Fungi may be playing a role in this small subset of CRS patients, all of whom had nasal polyposis. This study refutes the theory that current culture methods are insensitive and missing a large proportion of patients in whom fungi are playing a central role.
The importance of understanding the complex polymicrobial communities in the sinuses is highlighted by the concept of dysbiosis, where organisms interact in positive (mutualistic) or negative ways to alter the local community, and interaction with the host. There is evidence to suggest that microbial diversity is important for health [
32], and a reduced diversity with increased abundance is associated with chronic inflammation and poor healing [
33]. There is also literature to suggest that host genetics and immunity strongly influence the composition of the mucosal microflora [
34]. Microbial communities inhabiting mucosal surfaces such as the gastrointestinal tract can result in a significant mutualism including local immune homeostasis [
35,
36], and protection from pathogens through processes such as nutrient consumption, occupation of attachment sites, and secretion of antimicrobial substances [
37].
Propionibacterium acnes, which was found in more than 80% of control patients in the current study, has been shown to produce bacteriocins which have antibacterial and antifungal activity which may be protective against pathogens [
38]. Competition between microbes on mucosal surfaces can result in selection of virulence factors which can be detrimental to the host [
39]. In an elegant model of polymicrobial interactions, Sibley
et al., have shown that avirulent organisms can enhance the pathogenicity of other organisms, highlighting the importance of comprehensive community analysis to investigate disease [
40].
It is possible that disruption of mutualistic relationships, through shifts in mucosal-associated microbial composition, could contribute to the onset, progression, or recalcitrance of CRS. For example, the change in microbial dynamics during a viral upper respiratory tract infection or acute bacterial exacerbation of CRS. The often protracted use of antibiotics in our patients may also have detrimental sequelae on the microbial balance in CRS which requires further substantiation. There is poor understanding of such mechanisms in CRS, but they are an active focus of research at present. The first step in this journey is to greatly improve our knowledge of the mucosal microbial communities in CRS patients and controls.
Future directions for research should examine larger populations of CRS patients to characterize the microbiome of the different CRS phenotypes in comparison with healthy controls. Other molecular techniques, such as 16S rRNA gene pyrosequencing and quantitative real-time PCR will provide greater resolution for more detailed microbiome studies. Furthermore, longitudinal molecular studies evaluating the effect of antibiotics, endoscopic sinus surgery, and topical treatments on microbial diversity and abundance in CRS patients would be invaluable. Investigating the relationships within the microbial communities, and their interactions with the host immune system in health and disease, will ultimately lead to a greater understanding of the pathogenesis of chronic rhinosinusitis.
Competing interests
PJW receives royalties from Medtronic and is a consultant for Neilmed. GDE is a consultant for Stryker and Medtronic. SB, AF, EC, LWT, RM-K, HP & FZH have no competing interests. Funding: nil received.
Authors’ contributions
SB - Project design, tissue collection & analysis, fungal and bacterial biofilm preparation & imaging, image analysis, statistical analysis, manuscript preparation. AF - Project design, assistance with biofilm analysis, manuscript editing. EC - Assistance with biofilm analysis, manuscript editing. LWT - Project co-supervision, manuscript editing. RM-K - Ibis sample preparation & analysis. HP - Project co-supervision, manuscript editing. FZH - Ibis analysis & interpretation. GDE - Ibis analysis & interpretation, manuscript editing. PJW - Principal project supervision, surgical procedures, manuscript editing. All authors read and approved the final manuscript.