**Abstract**

The chronic colonisation of the lower airways by bacterial pathogens is the leading cause of morbidity and mortality in patients with cystic fibrosis (CF). The use of novel culture-independent techniques such as next-generation sequencing (NGS) to analyse the lungs has allowed us to further understand the diversity, the complexity, the effects of acute exacerbations and the use of antibiotics on the bacterial communities. The understanding of the CF microbiome to airway disease remains a fascinating area of research; it presents new opportunities for disease management in CF and has the potential to explore the effects of cystic fibrosis transmembrane conductance regulator (CFTR) modulators. It also allows further appreciation regarding the roles played by anaerobic organisms within the CF airways. It is also of interest that a number of studies have demonstrated that the fluctuations of microbiome are not necessarily associated with the patient's clinical status. Despite the available evidence, there remain many challenges that must be overcome if microbiome profiling is going to influence future clinical practice. The effects of fungus and the emergence of nontuberculous mycobacteria in CF are also briefly discussed in this chapter.

**Keywords:** cystic fibrosis, microbiome, CFTR modulators, nontuberculous mycobacteria, *Aspergillus*

#### **1. Introduction**

Traditional culture techniques rely on growing bacteria on media in laboratory conditions often optimised for growth of specific organisms so that they can then subsequently be identified. In the last 20 years, novel techniques utilising nextgeneration sequencing (NGS) to identify bacteria have become available, enabling detection and description of bacterial communities without the need for conventional culture. These technologies have allowed a greater understanding of bacterial communities throughout the human body and have revealed functional roles in both health and disease.

A healthy human gut, for example, is home to a highly diverse community of bacteria, termed as microbiome, which has symbiotic functions including metabolism of otherwise indigestible compounds and defence against opportunistic pathogens [1, 2]. Furthermore, bacteria in the gut influence the stimulation and development of the innate mucosal immune system [3]. In addition to the roles in health, there has been significant interest in the relationship between microbiomes and diseases such as obesity, inflammatory bowel disease and diabetes mellitus [4–6].

Studies utilising culture-independent techniques to analyse the lungs have identified the presence of bacterial communities that are much more complex than the previously appreciated. The lungs were long considered to be an inherently sterile environment, in part due to the fact that conventional culture techniques often yielded negative results during health and it was only during disease that pathogens were detected. However, the advent of culture-independent techniques has demonstrated that multiple organisms comprise a community, termed the 'microbiome', in the lungs of patients, both healthy and diseased [7–9]. In this chapter, we discuss the techniques employed in 16S rRNA sequencing and the evidence these techniques have generated so far in relation to cystic fibrosis.

#### **2. 16S rRNA gene sequencing**

The 16S rRNA gene codes for a ribosomal subunit present in nearly all bacteria. The gene itself is approximately 1.5 kb long and consists of conserved regions, similar in nearly all microorganisms, and nine variable regions labelled, V1–V9, which are practically specific to each microorganism [10]. The identification of a specific DNA sequence that corresponds to the known variable region of 16S rRNA gene can allow discrimination of the presence and relative abundance of different microorganisms (**Figure 1**).

Once the samples have been processed, DNA is extracted, and the 16S rRNA gene is amplified using polymerase chain reaction (PCR). Next-generation sequencing allows elucidation of the precise gene sequences, and online reference databases can then be used to match each sequence to an organism and quantify its relative proportion within a multispecies population. However, it is important to note that sequencing of the 16S rRNA gene has limited resolution and often cannot distinguish species with similar gene sequences apart. Therefore, instead of distinct species, sequences are referenced against and assigned into operational taxonomic units (OTU) (see **Table 1**).

Given the large number of species identified even in healthy lungs, ecological theory and analyses are often employed to understand community dynamics [9]. According to ecological principles, the composition of the lung microbiome is determined by three factors:


3.Regional growth factors [11]

#### **Figure 1.**

*Schematic illustration of structure of the 16S rRNA gene. Orange regions—variable regions; blue region conserved region.*

**3**

the last decade.

Operational taxonomic unit (OTU)

*Glossary of terms and definitions.*

**Table 1.**

**3. CF respiratory microbiota in early life**

early estimation of the neonatal lower respiratory tract.

*The Pulmonary Microbiome in Cystic Fibrosis DOI: http://dx.doi.org/10.5772/intechopen.91765*

**Term Definition**

Microbiome The microorganisms in a particular environment

Richness A measure of the number of species in a community

relative abundance?

Alpha diversity Within-sample diversity Beta diversity Between-sample diversity

The lung microbiome in healthy individuals is dictated largely by immigration

16S rRNA gene A gene which codes for a ribosomal subunit. Present in all prokaryotes and has

Evenness A measure of similarity of the relative abundance for each species in a

Diversity A measure of variety in a community. Combination of richness and evenness

variable regions, which differ slightly between bacterial species

Group of strains/species with similar 16S rRNA gene sequences

community That is, does one species dominate, or do all species have similar

Understanding the development of the CF respiratory microbiota in early life has attracted interest in order to appreciate the driving factors behind the distinct microbiota seen later in life and also to identify potential opportunities for intervention. Neonates and infants cannot expectorate sputum independently, and bronchoalveolar lavage (BAL) is only used sparingly; hence, upper respiratory tract samples are often used as surrogates. The imperfection of this approach was recently highlighted where large differences in concordance between BAL samples and upper respiratory tract (URT) samples were observed in some taxa [12]. Nevertheless, concordance was high for some important taxa such as *Moraxella* and *Staphylococcus,* and in the absence of less invasive techniques, URT sampling enables

The composition of CF nasopharyngeal microbiota diverges from that of non-CF infants as early as the first few months of life [13, 14]. Newborn healthy infants appear to have nasopharyngeal microbiota dominated by *Moraxella* spp*., Corynebacterium spp.* and *Haemophilus* spp., a community structure that persists for at least the first 6 months of life. Conversely the CF nasopharynx microbiome is initially dominated by *Staphylococcus aureus* before a gradual increase in *Streptococcus* spp*.* and *Moraxellaceae* at 3 months of age [13]. Despite the increased *S. aureus* seen in CF, there were no decreases in measures of richness or diversity indicating that changes are due

to differing microenvironment rather than interspecies competition [15].

and elimination and hence generally consists predominantly of those Gramnegative anaerobes also resident in the oral flora such as *Prevotella* and *Veillonella* spp. [11]. However, in disease the regional growth conditions are altered, and niches for other species to thrive are created. In CF, for example, viscous secretions, altered pH, nutrient availability and architectural disturbance may all help select for a community of altered composition to that of healthy individuals. How this community changes over time, in response to the intensive antibiotic treatment that people with CF are exposed to, has been the subject of much interest in

*The Pulmonary Microbiome in Cystic Fibrosis DOI: http://dx.doi.org/10.5772/intechopen.91765*


#### **Table 1.**

*Cystic Fibrosis - Heterogeneity and Personalized Treatment*

niques have generated so far in relation to cystic fibrosis.

**2. 16S rRNA gene sequencing**

microorganisms (**Figure 1**).

units (OTU) (see **Table 1**).

determined by three factors:

3.Regional growth factors [11]

1.Immigration of organisms into the lung

2.Elimination of microbes from the airways

Studies utilising culture-independent techniques to analyse the lungs have identified the presence of bacterial communities that are much more complex than the previously appreciated. The lungs were long considered to be an inherently sterile environment, in part due to the fact that conventional culture techniques often yielded negative results during health and it was only during disease that pathogens were detected. However, the advent of culture-independent techniques has demonstrated that multiple organisms comprise a community, termed the 'microbiome', in the lungs of patients, both healthy and diseased [7–9]. In this chapter, we discuss the techniques employed in 16S rRNA sequencing and the evidence these tech-

The 16S rRNA gene codes for a ribosomal subunit present in nearly all bacteria.

Once the samples have been processed, DNA is extracted, and the 16S rRNA

Given the large number of species identified even in healthy lungs, ecological theory and analyses are often employed to understand community dynamics [9]. According to ecological principles, the composition of the lung microbiome is

*Schematic illustration of structure of the 16S rRNA gene. Orange regions—variable regions; blue region—*

gene is amplified using polymerase chain reaction (PCR). Next-generation sequencing allows elucidation of the precise gene sequences, and online reference databases can then be used to match each sequence to an organism and quantify its relative proportion within a multispecies population. However, it is important to note that sequencing of the 16S rRNA gene has limited resolution and often cannot distinguish species with similar gene sequences apart. Therefore, instead of distinct species, sequences are referenced against and assigned into operational taxonomic

The gene itself is approximately 1.5 kb long and consists of conserved regions, similar in nearly all microorganisms, and nine variable regions labelled, V1–V9, which are practically specific to each microorganism [10]. The identification of a specific DNA sequence that corresponds to the known variable region of 16S rRNA gene can allow discrimination of the presence and relative abundance of different

**2**

**Figure 1.**

*conserved region.*

*Glossary of terms and definitions.*

The lung microbiome in healthy individuals is dictated largely by immigration and elimination and hence generally consists predominantly of those Gramnegative anaerobes also resident in the oral flora such as *Prevotella* and *Veillonella* spp. [11]. However, in disease the regional growth conditions are altered, and niches for other species to thrive are created. In CF, for example, viscous secretions, altered pH, nutrient availability and architectural disturbance may all help select for a community of altered composition to that of healthy individuals. How this community changes over time, in response to the intensive antibiotic treatment that people with CF are exposed to, has been the subject of much interest in the last decade.

### **3. CF respiratory microbiota in early life**

Understanding the development of the CF respiratory microbiota in early life has attracted interest in order to appreciate the driving factors behind the distinct microbiota seen later in life and also to identify potential opportunities for intervention. Neonates and infants cannot expectorate sputum independently, and bronchoalveolar lavage (BAL) is only used sparingly; hence, upper respiratory tract samples are often used as surrogates. The imperfection of this approach was recently highlighted where large differences in concordance between BAL samples and upper respiratory tract (URT) samples were observed in some taxa [12]. Nevertheless, concordance was high for some important taxa such as *Moraxella* and *Staphylococcus,* and in the absence of less invasive techniques, URT sampling enables early estimation of the neonatal lower respiratory tract.

The composition of CF nasopharyngeal microbiota diverges from that of non-CF infants as early as the first few months of life [13, 14]. Newborn healthy infants appear to have nasopharyngeal microbiota dominated by *Moraxella* spp*., Corynebacterium spp.* and *Haemophilus* spp., a community structure that persists for at least the first 6 months of life. Conversely the CF nasopharynx microbiome is initially dominated by *Staphylococcus aureus* before a gradual increase in *Streptococcus* spp*.* and *Moraxellaceae* at 3 months of age [13]. Despite the increased *S. aureus* seen in CF, there were no decreases in measures of richness or diversity indicating that changes are due to differing microenvironment rather than interspecies competition [15].

The divergence observed in the first few months of life usually precedes antibiotic administration and demonstrates that CF itself is associated with compositional changes in the microbiota, but as CF infants grow older, exposure to antibiotics, either via acute treatment for respiratory illnesses or prophylaxis against classic CF pathogens, becomes inevitable. Mika et al. investigated the relationship between antibiotics and the nasal microbiota by prospectively following 30 newborn infants with CF with fortnightly sampling for the first 12 months of life [14]. Antibiotic administration was associated with an increase in the Shannon diversity measure (a measure of the richness and evenness of a community), but this was judged to be most likely secondary to an increase in transient colonisers. Interestingly, antibiotic therapy was staphylococcal directed, but decreases in *Staphylococcus* OTUs were not seen. Instead significant reductions in *Moraxellaceae* were observed, and when oligotyping was used to observe changes in *Staphylococcus* at the species level, mild reductions in *S. aureus* were offset by increases in *S. epidermidis*, leading the authors to suggest that *S. epidermidis* may act as a reservoir of resistance. These findings were supported by Prevaes et al. who conducted a similar study in a slightly older population (mean age 2 years old) [13]. Antibiotic treatment was observed to be associated with reductions in *Moraxellaceae* and *S. aureus* OTUs, with increases in other staphylococcal OTUs, although more specific oligotyping was not performed.

As babies grow older, sampling from the lower airways becomes more common, and comparisons between the lower and upper airways become feasible. Given the close proximity and interrelated spaces of the nose, throat and lungs, it could be expected that they all may share similar community structures; however, the reality is that the nasal community appears different from that of the throat and lung, which are much more closely aligned. Boutin et al. found differences in the community structure of the nasal cavity compared to throat and sputum samples; in that diversity, richness and evenness were significantly higher in nasal samples, and up to 21 of the 76 most abundant nasal OTUs were not present in the throat or sputum samples [16]. Interestingly, the authors also found that subjects could be broadly defined into one of two ecotypes based on the presence or absence of *Pseudomonas*, and the similarities between throat and sputum samples began to diminish once *Pseudomonas* was present. Muhlebach et al. supported these findings in a recently published study, which for the first time included routine sequential BAL sampling in young children as part of the large AREST-CF cohort in Australia and the USA. They showed that lower airway cultures mirrored that of the oral cavity until approximately age 2, when increasing predominance of known CF pathogens was observed and communities diverged [17]. This has a number of clinical implications in that, firstly, throat swabs can provide adequate representation of the lower airways in very young children and, secondly, prevention or delay of this transition point by manipulation of the microbiota could theoretically be a strategy to improve outcomes later in life.

#### **4. Progressive loss of diversity**

Once the lungs are colonised with CF pathogens, a pattern of progressively uneven community structures ensues. Cox et al. examined biobanked sputum samples from a cohort of 63 clinically stable people with CF of ages ranging from 9 months to 72 years [7]. This cross-sectional approach identified the loss of community richness, evenness and diversity as age increased. *Pseudomonas* and *Burkholderia* OTUs began to progressively dominate in older subjects, and the changes in community structure were inversely associated with pulmonary function. In a similar study design with 269 patients, Coburn et al. also found sample

**5**

*The Pulmonary Microbiome in Cystic Fibrosis DOI: http://dx.doi.org/10.5772/intechopen.91765*

richness and diversity [21].

clinical outcomes [24].

predictor for decreased diversity [29].

which notably increased after the age of 25 [18].

diversity inversely correlated with age and disease stage. Progressive loss of diversity was particularly correlated with *Pseudomonas* and *Burkholderia* abundance,

Zhao et al. were the first to confirm these findings longitudinally when they followed up six patients over a 9-year period with serial sputum collections. It was observed that the three patients with what they termed as more 'progressive' disease had significant decreases in community diversity over the course of several years. Decreasing lung function and increasing age were also associated with decreasing community diversity [19]. This study was soon followed by Fodor et al. who focussed more on changes in the microbiota associated with acute changes in clinical status but did observe a strong correlation between low species richness and poor lung function [20]. Stokell et al. followed up a single patient up to over 3 years and observed increasing total bacterial load as well as diminishing community

Contrastingly, Whelan et al. recently published a study of six patients who submitted thrice-weekly sputum samples for a year [22]. No overall changes in community structure were observed over the course of the year, and the authors concluded that the respiratory microbiome is unique to each patient and the previously reported associations between community structure and clinical parameters may be true on a cohort/population level but not at an individual level. There is some merit in this argument, but it is also worth noting that the six patients in the study appeared relatively stable with a median of only one exacerbation in the 12-month study period. It is also therefore a possibility that the follow-up period was not long enough to capture the more indolent changes likely to be present in those patients [23]. A much longer study period was adopted by Acosta et al. [24] who analysed samples from matched patients with biobanked sputum samples in three historic cohorts spanning nearly 20 years at a single centre. Across all the three cohorts, the core microbiome constituents were preserved, but the proportion of *Pseudomonas*-dominated communities was reduced, and overall diversity increased in the more recent cohorts. Community structure improved gradually from the most historic cohort to the most recent, and these changes appeared to correlate with the generally improving clinical status of people with CF, confirming the previously described observed association between community structure and

The association reported in most studies between community structure and clinical outcomes has inevitably led to the question of whether a less diverse or even rich microbiome is simply a marker of increased pulmonary disease or is itself a driver in disease pathogenesis [25]. If the latter were true, efforts to promote a more diverse community could have the potential to slow pulmonary disease progression. An Italian group has led efforts to find patterns or signatures in the microbiome that may predispose patients to accelerated lung function decline; however, no causal association has been elucidated [26–28]. Instead, Zhao et al. found that the relationship between age, lung function and community diversity disappeared once controlled for antibiotic use, thus suggesting antibiotic therapy is the predominant driver of reducing community diversity [19]. The same group later developed a statistical approach to more precisely correct the antibiotic exposure when examining relationships between microbiota and clinical outcomes. The approach was applied to 478 sputum samples and confirmed that antibiotic use was an independent

Accurately recording antibiotic use is troublesome in longitudinal studies due to the frequent episodic use of antibiotics in CF which is often self-directed by patients themselves, due to the widespread use of long-term antibiotics for which compliance may be heterogeneous and also due to the retrospective nature of a number

#### *The Pulmonary Microbiome in Cystic Fibrosis DOI: http://dx.doi.org/10.5772/intechopen.91765*

*Cystic Fibrosis - Heterogeneity and Personalized Treatment*

The divergence observed in the first few months of life usually precedes antibiotic administration and demonstrates that CF itself is associated with compositional changes in the microbiota, but as CF infants grow older, exposure to antibiotics, either via acute treatment for respiratory illnesses or prophylaxis against classic CF pathogens, becomes inevitable. Mika et al. investigated the relationship between antibiotics and the nasal microbiota by prospectively following 30 newborn infants with CF with fortnightly sampling for the first 12 months of life [14]. Antibiotic administration was associated with an increase in the Shannon diversity measure (a measure of the richness and evenness of a community), but this was judged to be most likely secondary to an increase in transient colonisers. Interestingly, antibiotic therapy was staphylococcal directed, but decreases in *Staphylococcus* OTUs were not seen. Instead significant reductions in *Moraxellaceae* were observed, and when oligotyping was used to observe changes in *Staphylococcus* at the species level, mild reductions in *S. aureus* were offset by increases in *S. epidermidis*, leading the authors to suggest that *S. epidermidis* may act as a reservoir of resistance. These findings were supported by Prevaes et al. who conducted a similar study in a slightly older population (mean age 2 years old) [13]. Antibiotic treatment was observed to be associated with reductions in *Moraxellaceae* and *S. aureus* OTUs, with increases in other staphylococcal OTUs, although more specific oligotyping was not performed. As babies grow older, sampling from the lower airways becomes more common, and comparisons between the lower and upper airways become feasible. Given the close proximity and interrelated spaces of the nose, throat and lungs, it could be expected that they all may share similar community structures; however, the reality is that the nasal community appears different from that of the throat and lung, which are much more closely aligned. Boutin et al. found differences in the community structure of the nasal cavity compared to throat and sputum samples; in that diversity, richness and evenness were significantly higher in nasal samples, and up to 21 of the 76 most abundant nasal OTUs were not present in the throat or sputum samples [16]. Interestingly, the authors also found that subjects could be broadly defined into one of two ecotypes based on the presence or absence of *Pseudomonas*, and the similarities between throat and sputum samples began to diminish once *Pseudomonas* was present. Muhlebach et al. supported these findings in a recently published study, which for the first time included routine sequential BAL sampling in young children as part of the large AREST-CF cohort in Australia and the USA. They showed that lower airway cultures mirrored that of the oral cavity until approximately age 2, when increasing predominance of known CF pathogens was observed and communities diverged [17]. This has a number of clinical implications in that, firstly, throat swabs can provide adequate representation of the lower airways in very young children and, secondly, prevention or delay of this transition point by manipulation of the microbiota could theoretically be a strategy to improve

**4**

outcomes later in life.

**4. Progressive loss of diversity**

Once the lungs are colonised with CF pathogens, a pattern of progressively uneven community structures ensues. Cox et al. examined biobanked sputum samples from a cohort of 63 clinically stable people with CF of ages ranging from 9 months to 72 years [7]. This cross-sectional approach identified the loss of community richness, evenness and diversity as age increased. *Pseudomonas* and *Burkholderia* OTUs began to progressively dominate in older subjects, and the changes in community structure were inversely associated with pulmonary function. In a similar study design with 269 patients, Coburn et al. also found sample

diversity inversely correlated with age and disease stage. Progressive loss of diversity was particularly correlated with *Pseudomonas* and *Burkholderia* abundance, which notably increased after the age of 25 [18].

Zhao et al. were the first to confirm these findings longitudinally when they followed up six patients over a 9-year period with serial sputum collections. It was observed that the three patients with what they termed as more 'progressive' disease had significant decreases in community diversity over the course of several years. Decreasing lung function and increasing age were also associated with decreasing community diversity [19]. This study was soon followed by Fodor et al. who focussed more on changes in the microbiota associated with acute changes in clinical status but did observe a strong correlation between low species richness and poor lung function [20]. Stokell et al. followed up a single patient up to over 3 years and observed increasing total bacterial load as well as diminishing community richness and diversity [21].

Contrastingly, Whelan et al. recently published a study of six patients who submitted thrice-weekly sputum samples for a year [22]. No overall changes in community structure were observed over the course of the year, and the authors concluded that the respiratory microbiome is unique to each patient and the previously reported associations between community structure and clinical parameters may be true on a cohort/population level but not at an individual level. There is some merit in this argument, but it is also worth noting that the six patients in the study appeared relatively stable with a median of only one exacerbation in the 12-month study period. It is also therefore a possibility that the follow-up period was not long enough to capture the more indolent changes likely to be present in those patients [23]. A much longer study period was adopted by Acosta et al. [24] who analysed samples from matched patients with biobanked sputum samples in three historic cohorts spanning nearly 20 years at a single centre. Across all the three cohorts, the core microbiome constituents were preserved, but the proportion of *Pseudomonas*-dominated communities was reduced, and overall diversity increased in the more recent cohorts. Community structure improved gradually from the most historic cohort to the most recent, and these changes appeared to correlate with the generally improving clinical status of people with CF, confirming the previously described observed association between community structure and clinical outcomes [24].

The association reported in most studies between community structure and clinical outcomes has inevitably led to the question of whether a less diverse or even rich microbiome is simply a marker of increased pulmonary disease or is itself a driver in disease pathogenesis [25]. If the latter were true, efforts to promote a more diverse community could have the potential to slow pulmonary disease progression. An Italian group has led efforts to find patterns or signatures in the microbiome that may predispose patients to accelerated lung function decline; however, no causal association has been elucidated [26–28]. Instead, Zhao et al. found that the relationship between age, lung function and community diversity disappeared once controlled for antibiotic use, thus suggesting antibiotic therapy is the predominant driver of reducing community diversity [19]. The same group later developed a statistical approach to more precisely correct the antibiotic exposure when examining relationships between microbiota and clinical outcomes. The approach was applied to 478 sputum samples and confirmed that antibiotic use was an independent predictor for decreased diversity [29].

Accurately recording antibiotic use is troublesome in longitudinal studies due to the frequent episodic use of antibiotics in CF which is often self-directed by patients themselves, due to the widespread use of long-term antibiotics for which compliance may be heterogeneous and also due to the retrospective nature of a number

of CF microbiome studies [30]. However, Pittman et al. were able to prospectively perform bronchoscopy and record antibiotic exposure of 32 subjects as part of the AREST-CF study. In that study, community diversity was much lower in the BAL of those patients receiving antibiotics [31].

Thus it appears likely that the strong association between community structure and degree of lung disease is related to the inevitable prolonged and aggressive use of antibiotics in CF rather than direct pathogenesis from a less diverse microbiome.

### **5. Community changes with acute pulmonary exacerbations**

Despite the importance of exacerbations on long-term outcomes of people with CF, the pathophysiology of these events remains undefined [32, 33]. Clinically, exacerbations are frequent and are characterised by rapid changes in symptoms such as an increase in sputum volume or purulence, shortness of breath and fatigue. The precise mechanisms underlying these important events remain elusive, and studies looking for answers using culture-independent techniques have not found consistent answers. For example, one may expect to find evidence of increases in known pathogens at the time of exacerbations, yet there is no consistent evidence of this [23]. In fact, a number of studies have found the CF microbiota to be extremely stable over time and resilient to change at exacerbation and following subsequent treatment [19, 20, 34, 35].

However, when the community structure as a whole is considered, a number of larger studies have found reduced diversity or richness at the times of exacerbation compared to clinical stability. Coburn et al. found small decreases in Shannon diversity in exacerbation samples compared to their baseline study of 269 people with CF [18]. Similarly, Filkins et al. found that samples taken during exacerbations had significantly lower diversity than samples taken when patients were stable [36]. Perhaps most convincingly, Li et al. collated data from 18 previous studies to analyse over 700 sputum samples and found that there were significant reductions in community richness at exacerbation [37].

Whilst increases in *P. aeruginosa* at the time of exacerbation have not been seen consistently, they have been observed in some cases. Carmody et al. followed up four patients for 3 months with daily sputum sampling and observed daily stability between exacerbations but increased *P. aeruginosa* abundance at the time of exacerbation in some patients and increases in *Prevotella* in others [38]. These findings help introduce two new concepts: firstly, the potential for exacerbations to appear similar phenotypically but have different underlying aetiology with only some being due to changes that can be observed in the microbiota and, secondly, that previously overlooked anaerobes may play a pathogenic role.

The first concept is supported by Whelan et al. who found in longitudinal sampling of six patients that some but not all exacerbations were associated with changes in the microbiota [22]. Attempts to identify different types of exacerbations in COPD have identified four distinct aetiological clusters, bacterial, viral, eosinophilic predominant and 'paucinflammatory', and even though these clusters may not be mirrored in CF, it is plausible that not all exacerbation clusters would be associated with changes apparent in either individual taxa or overall bacterial community structure [39].

Changes in the metabolic activity of specific taxa or the community as a whole triggering an exacerbation could be another explanation for an apparent lack of change in the community structure seen in some studies. The metabolites lactate and putrescine were found by Twomey et al. to increase during exacerbation in the absence of clear changes in the community structure [40]. Quinn et al. used

**7**

*The Pulmonary Microbiome in Cystic Fibrosis DOI: http://dx.doi.org/10.5772/intechopen.91765*

pathogenesis of some exacerbations [43, 44].

abundant OTUs were all reduced [20].

community structure.

**exacerbations**

infections.

the ecological functional networking to identify the non-mevalonate pathway of isoprenoid synthesis as a 'keystone' pathway in CF infections. Intriguingly fosmidomycin, an antimalarial agent, is known to be effective at targeting this pathway [41]. The second concept to emerge from the study of Carmody et al. relates to the changes in *Prevotella* abundance at the time of exacerbation and raises the prospect that species not considered conventional CF pathogens may play a role in exacerbations [42]. Anaerobic species are easily overlooked in conventional selective culturing due to the requirement for anoxic culture yet are identified frequently in culture-independent analyses of the CF lower airways. In addition to *Gemella*, both *Prevotella* and *Streptococcus anginosus (milleri)* have been found to have associations with clinical stability [36, 38, 42]. Anaerobes have been shown to have the potential to modulate *P. aeruginosa* gene expression in the polymicrobial setting; hence, even if they are not directly pathogenic, they may still play a contributory role to the

To summarise, the aetiologies underpinning the transition from a stable state to an acute exacerbation are not well understood. It is likely that there multiple aetiological clusters but only some of which may be associated with changes in

**6. Community changes associated with treatment for acute pulmonary** 

Traditional dogma would dictate that intensive, targeted antimicrobial therapy with dual anti-pseudomonal agents will result in significant reductions in abundance of *P. aeruginosa*; however, in a similar vein to the findings from studies of the onset of exacerbations, microbiota responses to treatment for acute pulmonary exacerbations in CF have not aligned with this conventional understanding of

One of the predominant themes that has emerged from studies of the respiratory microbiota response to acute antibiotics is that *P. aeruginosa* is not impacted to the same degree as other members of the community. For example, Daniels et al. studied 12 adult CF subjects across the cycle of an exacerbation and treatment and found that following initiation of anti-pseudomonal antimicrobials, the relative abundance of *P. aeruginosa* actually increased, alongside a reduction in the total number of species detected [45]. Cuthbertson et al. also found no evidence of reduced *P. aeruginosa* abundance in a study of 12 CF patients receiving treatment for pulmonary exacerbations [45]. Instead, reductions in *Streptococcus sanguinis*, *Prevotella* and *Porphyromonas* OTUs were observed [35]. Similarly, Li et al., in their analysis of over 700 sputum samples, found that antibiotic treatment had no effect on *Pseudomonas* abundance but did have significant effects on *Gemella*, *Staphylococcus*, *Actinomyces, Moraxellaceae* and *Fusobacterium* [37]. Further, Fodor et al. again found that dominant taxa such as *Pseudomonas* and *Burkholderia* were unchanged when compared at the beginning and end of an exacerbation but the relative abundances of *Gemella, Streptococcus* and a small number of other less

In contrast, two studies have found reductions in *P. aeruginosa* following treatment. Firstly, Zemanick et al. investigated the association between inflammation and changes to the airway microbiota during treatment for exacerbations and found that although bacterial load did not change, the relative abundance of *P. aeruginosa* was observed to decrease and that these changes correlated with improved lung function [46]. A reduction in *P. aeruginosa* abundance following treatment was also reported by Smith et al., who noted rapid decreases in *P. aeruginosa* abundance and

#### *The Pulmonary Microbiome in Cystic Fibrosis DOI: http://dx.doi.org/10.5772/intechopen.91765*

*Cystic Fibrosis - Heterogeneity and Personalized Treatment*

those patients receiving antibiotics [31].

treatment [19, 20, 34, 35].

community structure [39].

in community richness at exacerbation [37].

previously overlooked anaerobes may play a pathogenic role.

of CF microbiome studies [30]. However, Pittman et al. were able to prospectively perform bronchoscopy and record antibiotic exposure of 32 subjects as part of the AREST-CF study. In that study, community diversity was much lower in the BAL of

Thus it appears likely that the strong association between community structure and degree of lung disease is related to the inevitable prolonged and aggressive use of antibiotics in CF rather than direct pathogenesis from a less diverse microbiome.

Despite the importance of exacerbations on long-term outcomes of people with CF, the pathophysiology of these events remains undefined [32, 33]. Clinically, exacerbations are frequent and are characterised by rapid changes in symptoms such as an increase in sputum volume or purulence, shortness of breath and fatigue. The precise mechanisms underlying these important events remain elusive, and studies looking for answers using culture-independent techniques have not found consistent answers. For example, one may expect to find evidence of increases in known pathogens at the time of exacerbations, yet there is no consistent evidence of this [23]. In fact, a number of studies have found the CF microbiota to be extremely stable over time and resilient to change at exacerbation and following subsequent

However, when the community structure as a whole is considered, a number of larger studies have found reduced diversity or richness at the times of exacerbation compared to clinical stability. Coburn et al. found small decreases in Shannon diversity in exacerbation samples compared to their baseline study of 269 people with CF [18]. Similarly, Filkins et al. found that samples taken during exacerbations had significantly lower diversity than samples taken when patients were stable [36]. Perhaps most convincingly, Li et al. collated data from 18 previous studies to analyse over 700 sputum samples and found that there were significant reductions

Whilst increases in *P. aeruginosa* at the time of exacerbation have not been seen consistently, they have been observed in some cases. Carmody et al. followed up four patients for 3 months with daily sputum sampling and observed daily stability between exacerbations but increased *P. aeruginosa* abundance at the time of exacerbation in some patients and increases in *Prevotella* in others [38]. These findings help introduce two new concepts: firstly, the potential for exacerbations to appear similar phenotypically but have different underlying aetiology with only some being due to changes that can be observed in the microbiota and, secondly, that

The first concept is supported by Whelan et al. who found in longitudinal sampling of six patients that some but not all exacerbations were associated with changes in the microbiota [22]. Attempts to identify different types of exacerbations in COPD have identified four distinct aetiological clusters, bacterial, viral, eosinophilic predominant and 'paucinflammatory', and even though these clusters may not be mirrored in CF, it is plausible that not all exacerbation clusters would be associated with changes apparent in either individual taxa or overall bacterial

Changes in the metabolic activity of specific taxa or the community as a whole triggering an exacerbation could be another explanation for an apparent lack of change in the community structure seen in some studies. The metabolites lactate and putrescine were found by Twomey et al. to increase during exacerbation in the absence of clear changes in the community structure [40]. Quinn et al. used

**5. Community changes with acute pulmonary exacerbations**

**6**

the ecological functional networking to identify the non-mevalonate pathway of isoprenoid synthesis as a 'keystone' pathway in CF infections. Intriguingly fosmidomycin, an antimalarial agent, is known to be effective at targeting this pathway [41].

The second concept to emerge from the study of Carmody et al. relates to the changes in *Prevotella* abundance at the time of exacerbation and raises the prospect that species not considered conventional CF pathogens may play a role in exacerbations [42]. Anaerobic species are easily overlooked in conventional selective culturing due to the requirement for anoxic culture yet are identified frequently in culture-independent analyses of the CF lower airways. In addition to *Gemella*, both *Prevotella* and *Streptococcus anginosus (milleri)* have been found to have associations with clinical stability [36, 38, 42]. Anaerobes have been shown to have the potential to modulate *P. aeruginosa* gene expression in the polymicrobial setting; hence, even if they are not directly pathogenic, they may still play a contributory role to the pathogenesis of some exacerbations [43, 44].

To summarise, the aetiologies underpinning the transition from a stable state to an acute exacerbation are not well understood. It is likely that there multiple aetiological clusters but only some of which may be associated with changes in community structure.
