What is nontuberculous mycobacterial pulmonary disease, who does it affect and how is it treated?

Nontuberculous mycobacteria (NTM) are aerobic, non-motile opportunistic pathogens. NTM comprises a phylogenetically diverse group of environmental rapid- and slow-growing species and subspecies within the genus Mycobacterium (M.), distinct from the M. tuberculosis complex and M. leprae [1]. Extrapulmonary NTM infections such as skin, soft-tissue and musculoskeletal infections occur, but pulmonary infection and disease predominates clinically [2]. Laboratory isolation of NTM can reflect transient infection and is not always indicative of disease, making it challenging to make the distinction. Guidelines for the diagnosis of NTM-PD (rather than infection) require clinical, radiographic and microbiological criteria all be met [2,3]. The significance of NTM-PD is increasing globally, as rates of NTM pulmonary isolation and disease continue to rise [4]. M. avium complex (MAC), slow-growing NTM, and M. abscessus, rapid-growing NTM, are the most common causative agents in NTM-PD [5–7], both contributing substantially to morbidity and mortality [8,9].

Clinically, NTM-PD is commonly broadly categorised into two distinct clinical phenotypes [3,10]. Fibrocavitary disease involves inflammatory cavitation, progressive scarring and surrounding fibrosis in the lung. It is often seen in middle-aged men with a history of smoking and excessive alcohol consumption, and underlying lung diseases such as past tuberculosis, chronic obstructive pulmonary disease (COPD) and emphysema [3]. Nodular bronchiectatic NTM-PD is commonly associated with the ‘Lady Windermere’ phenotype [11], characterised by distinct host morphological traits; typically peri- or post-menopausal women, with above average height and lean body habitus, high rates of thoracic deformities (scoliosis and pectus excavatum) and mitral valve prolapse, despite the absence of a shared immune or genetic profile [12–14].

Current NTM-PD guideline therapy entails prolonged multidrug regimens and is unacceptably long, usually at least 12–24 months [2]. The causative NTM species, mycobacterial susceptibility to available antibiotics, disease phenotype, patient characteristics and drug tolerability are just some of the many factors that influence the choice and success of NTM-PD therapy [2,15]. Clinical experience indicates that adverse effects from NTM-PD treatment are common, occurring in nearly 40% of patients, irrespective of NTM species [16]. Treatment success, referring to clinical cure, is low at only 40–60% [17,18]. Furthermore, rates of microbiological recurrence and chronically progressive NTM-PD are high, and often represent reinfection rather than relapse [18–20]. Given the substantial treatment burden, high incidence of adverse effects, and limited long-term efficacy, there is an urgent need to optimise our understanding of NTM-PD.

In healthy individuals, NTM are eliminated through coordinated innate and adaptive immune responses (Fig 1). Prior reviews have highlighted the complexity of these NTM–host immune-interactions [21–25]. However, it must be noted that current understanding of these interactions is commonly derived from models that do not reflect the specific host environments encountered in NTM-PD (such as Drosophila melanogaster [26] or Zebrafish embryo [27]), with knowledge gaps often filled by extrapolation from the more extensive understanding of host defence mechanisms against M. tuberculosis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Innate and adaptive immune response to NTM infection in the lung.

Alveolar macrophages are integral in the initiation of an immune response against inhaled NTM. This occurs in part through phagocytosis of NTM as well as the release of chemokines and cytokines by alveolar macrophages that recruit additional macrophages and other phagocytes to the site of infection. Macrophage activation by NTM is facilitated by pattern recognition receptors (PRRs), of which Toll-like receptor 2 (TLR2) plays a central role [28–31]. Macrophage recognition of NTM initiates phagosome maturation, lysosome acidification and antimicrobial defence mechanisms (e.g., generation of reactive oxygen species, expression of degradative enzymes), facilitating intracellular killing of mycobacteria [32]. Dendritic cells loaded with mycobacterial antigen that migrate to draining lymph nodes present mycobacterial antigens to naïve T cells, thereby facilitating T cell differentiation [33]. The associated cytokine milieu, such as via interleukin (IL)-12 release, shapes T cell differentiation into T helper 1 (Th1) cells. Infected macrophages also present mycobacterial antigens to Th1 cells. Th1 cells, alongside other innate and adaptive lymphocytes, coordinate mycobacteria-restricting immune responses, including interferon (IFN)-γ, tumour necrosing factor (TNF-α), and granulocyte macrophage colony-stimulating factor (GM-CSF)-mediated macrophage activation [34–37]. Fig 1 created with Biorender.com.

https://doi.org/10.1371/journal.ppat.1013603.g001

Susceptibility to NTM infection and NTM-PD is associated with a spectrum of immune dysfunction, ranging from overt primary (e.g., chronic granulomatous disease) or acquired immunodeficiency (e.g., chronic immunosuppressive drug exposure) [14,38–40]. However, in immunocompetent hosts, including individuals with the ‘Lady Windermere’ phenotype or underlying chronic lung disease, no single genetic or immunological defect has been identified to account for NTM susceptibility. Instead, we hypothesise that such susceptibility may arise from subtle or combined impairments in immune cell functions. Indeed, persistent NTM infection and/or NTM-PD are associated with reduced absolute counts of adaptive immune cells [41,42], reduced cytokine production (e.g., interleukin (IL)-1, IL-10, IL-18) [43], and/or immune cell exhaustion, suggesting that broader immune dysregulation could underlie susceptibility [12,43–47]. Toll-like receptor 2 (TLR2) is a pattern recognition receptor central to host recognition of MAC [28,30], and M. abscessus [29,31]. In peripheral blood monocytes from NTM-PD patients following MAC stimulation, TLR2 mRNA expression was reduced and cells showed diminished IL-12 and tumour necrosis factor (TNF)-α production [48]. Whether similar findings occur with M. abscessus stimulation remains to be determined. Thus, host susceptibility to NTM-PD appears to be underpinned by subtle and/or combined immune impairments, and the extent and combinations might vary between individuals. Detailed characterisation of immune responses in individuals with known NTM-PD susceptibility, prior to and during NTM infection, is needed to delineate the precise mechanisms underlying susceptibility and to better categorise susceptibility phenotypes in this multifactorial disease.

An important factor in addressing these knowledge gaps is that different NTM species appear to have unique mechanisms to subvert host defence [49–51]. For example, MAC replicates and persists intracellularly [49], through disrupting intracellular vesicle trafficking and impairing phagosome maturation [50]. In the context of M. abscessus infection, smooth and rough morphological variants show differential interactions with the immune system [52], where increased virulence of a M. abscessus rough variant was associated with phagosomal escape in macrophages, thereby triggering Type I interferon expression, cell death, and cell-to-cell spread [51]. In a comparison study of MAC-PD and M. abscessus-PD, multiparametric flow cytometry analysis of peripheral blood mononuclear cells indicated NTM-species-specific immune checkpoint marker ‘fingerprints’ in the host response: MAC-PD was associated with high T cell immunoglobulin and mucin-domain-containing protein 3 (TIM-3) expression by T cells; M. abscessus-PD was associated with high cytotoxic T-lymphocyte associated protein 4 (CTLA4) expression by CD4⁺ T cells (especially T-regulatory cells) [44]. Thus, characterisation of NTM-PD-associated host susceptibility will require consideration of the NTM species, strains, and morphotype.

The microbiome in health and disease

The microbiome is increasingly recognised as a dynamic and influential modulator of health and disease. Broadly, the term “microbiome” refers to the genomic composition and diversity of a microbial community (i.e., the microbiota) and how that might be affected by the physicochemical properties of its surrounding environment (biome) at the time of sampling [53]. The prevailing view is that the microbiota associated with an individual outnumbers the absolute count of human somatic and germ cells by a factor of ten [54]. Accordingly, the collective DNA content of the microbiota, known as the metagenome, is phylogenetically more variable and dynamic compared to the relatively stable human genome [55,56]. Microbiome composition is influenced by physicochemical factors (such as oxygen and moisture levels, pH, and nutritional ecology), host-derived factors (including immune responses and genetics), and environmental or lifestyle factors (birth-delivery mode, breast- or formula-feeding, dietary pattern, exercise and exposure to antibiotics/medication) [55,57,58]. As such, differences in the relative abundance of microbial species and community composition lead to dynamic and variable microbiotas at different anatomical sites, even among healthy individuals [59]. Measures of the taxonomic diversity and relative abundance of the microbiota recovered from different body sites are often reported; based principally on case-control and/or cohort studies [60,61]. Reductions in these measures are commonly referred to as a dysbiosis, which often, but not always, is associated with disease [61]. There may be patterns that characterise disease-associated dysbiosis with some shared microbiome changes across different diseases, whilst other changes may be disease-specific [60]. Some diseases are associated with the presence of pathogenic microbes in the microbiota whereas others are characterised by a relative reduction in microbes associated with health [60]. Indeed, although microbiome research generally suggests that susceptibility to and/or pathophysiology of various diseases may be associated with taxonomic variations in the microbiome, there is consensus that the concept of a ‘healthy’ microbiome lacks a universal definition and objective criteria [54,55,62].

The lung microbiota: Could it be more stable in respiratory disease states than in health?

The lungs were historically believed to be sterile, but it is now established that the lungs host a resident microbiota distinct from the oropharynx [63,64]. Distal airways may have lower microbial richness and the right upper lobe is suggested to most closely resemble the oral cavity [64], therefore supporting the concept that the lung microbiota may be continually recolonised by microbes from adjacent sites and the environment through microaspiration and breathing [65,66]. Dickson and colleagues [67] propose that microbial composition in the airways is governed by a dynamic balance of three key factors; microbial immigration, microbial elimination and regional growth conditions. In the context of lung pathologies, it is hypothesised that alterations in the local airway environment, be it structural, immunological or metabolic, create conditions more conducive to bacterial proliferation, whereby microbial growth may surpass the airways’ capacity for elimination, this can result in persistent microbial colonisation and dysbiosis [67,68]. The relationship between the lung microbiota and respiratory diseases has been previously reviewed in the literature [69,70], and is relatively well-characterised in several chronic respiratory diseases that are associated with the development of NTM-PD. Shifts in microbial composition and the presence of certain taxa have been observed when comparing acute exacerbation states to clinically stable periods in COPD [71] and bronchiectasis [72]. These exacerbations are typically marked by reduced microbial diversity, associated with overgrowth of clinically significant pathogens such as Pseudomonas (P.) aeruginosa or Haemophilus influenzae [71]. A similar reduction in diversity is seen in cystic fibrosis patients chronically, but not intermittently, colonised by P. aeruginosa [73]. This pattern of dysbiosis, and the persistent dominance of certain taxa, may contribute to disease progression, supporting the hypothesis that the lung microbiota in respiratory disease is more stable and enduring than in health [65].

At the intersection of the lung microbiota and host susceptibility in NTM-PD

While environmental factors (e.g., temperature, rainfall) may influence the nature and degree of exposure to NTM [5], most individuals are able to control and eliminate NTM upon exposure. In cases where NTM infection can occur and progress to NTM-PD, underlying host susceptibility is multifactorial and multigenic [74,75]. At present, linking specific immune susceptibility factors to the microbiota in NTM-PD remains challenging, as we do not understand the varying combinations of subtle immune defects that may underlie susceptibility in different individuals. Broadly; however, immune dysfunction may contribute to pulmonary dysbiosis by impairing microbial clearance, allowing microbial expansion and promoting disease pathogenesis. In the following section, we examine the potential interplay between the lung microbiota, and two key susceptibility factors, impaired mucociliary clearance and microaspiration (see Fig 2), which are thought to predispose NTM infection and disease, and are also implicated in several chronic pulmonary conditions commonly associated with NTM-PD.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Model of host susceptibility factors driving lung and gut dysbiosis in NTM-PD.

The model is divided into two key categories of susceptible individuals: the ‘Lady Windermere’ phenotype and chronic lung disease. These categories are not mutually exclusive, and some individuals may exhibit features of both (or in fact neither). Common to both lies airway damage, impaired mucociliary clearance, and mucus stasis, which may promote niche colonisation by NTM or other pathogens, leading to recurrent infections and often requiring antibiotic use. This, in turn, may cause further dysbiosis in the lung and perturb the gut microbiota (see Fig 3). Fig 2 created with Biorender.com.

https://doi.org/10.1371/journal.ppat.1013603.g002

The gut–lung axis

Beyond direct anatomical connections such as in microaspiration, spatially distinct microbial communities do not operate as isolated populations. The lung environment and its associated microbiota are influenced by, and involved in complex interactions, including via the gut-lung axis. The gut-lung axis (GLA) describes the concept of the complex bidirectional microbiome–host interactions at each anatomical site, the gut and lungs, and the connectedness between these sites [65,66,68]. Microbial structural components and metabolic outputs from microbes possess immunomodulatory properties and can disperse via the bloodstream and lymphatic circulation, facilitating inter-niche communication that is fundamental in maintaining homeostasis and regulating host immune responses [110–112]. Fig 3 summarises the proposed roles of GLA in NTM-PD.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. A working model of the gut-lung axis (GLA) in nontuberculous mycobacterial pulmonary disease (NTM-PD).

The gut and lung have specialised microbiotas that are interconnected via the bloodstream and lymphatic system, via the GLA [65,68]. We propose gut microbiota dysbiosis as a potential initiating factor in the cycle of NTM-PD; however, whether it represents a cause or consequence of disease remains unclear. Dysbiosis contributes to changes in migrating immune cells, microbial immunomodulatory components and immune mediators, together representing a level of systemic immune dysregulation that affects the lung microbiota via the GLA [110–112]. Microaspiration and reflux are also implicated in the GLA and respiratory disease pathophysiology with a degree of direct seeding of gut microbes into the airways [99,104]. Lung microbiota dysbiosis may then cause lung injury and inflammation, contributing to poor clearance of NTM infection, allowing establishment of disease. Antibiotic treatment for NTM-PD then causes further dysbiosis of both the gut and lung microbiotas [61,113,114], so the cycle continues. Fig 3 created with Biorender.com.

https://doi.org/10.1371/journal.ppat.1013603.g003

The gut microbiome, particularly within the distal (ileocolonic) regions of the gastrointestinal tract, has been extensively reviewed for its contributions to the development and regulation of peripheral immune functions, including in the lung [65,115,116]. Emerging evidence suggests that while the gut microbiota can be protective against respiratory infections [112], gut dysbiosis can impair immune tolerance and contribute to inflammation. Gut dysbiosis has been associated with manifestations and exacerbations in various respiratory disease contexts [117,118]. For example, in a murine model, reduced gut microbial diversity increased mortality from respiratory viral infection by disrupting regulatory T cell balance and elevating pro-inflammatory cytokines [119]. In COPD, faecal microbiota profiles were characterised by a Prevotella-dominated enterotype with reduced levels of short-chain fatty acids (SCFAs); faecal microbiota transplantation from these patients into mice induced ‘COPD-like’ lung inflammation and accelerated disease progression [120]. SCFAs such as acetate, propionate and butyrate, are produced in the gut as by-products of bacterial anaerobic fermentation of dietary fibre [121]. Their diverse physiological roles and implications for human health and disease were recently reviewed by Mukhopadhya and colleagues [122], including their regulation of local gut immune responses and their systemic immunomodulatory effects [110,111]. Altered SCFA metabolism has been identified in pulmonary diseases, with reduced abundance of SCFA-producing bacteria and/or decreased production of SCFAs reported in patients with bronchiectasis [123], active tuberculosis (TB) [124], and COPD [125]. To our knowledge, assessments on SCFA-focussed analyses in NTM-PD have not yet been reported. Such analyses, paired with immune profiling in gut and lung tissue microenvironments, would be highly valuable in determining whether altered SCFA metabolism contributes to NTM-PD and how it may influence the immune response to NTM.

One metabolite which has been examined in the context of NTM-PD is L-arginine, an amino acid obtained through dietary intake, but also synthesised endogenously, as well as produced by gut microbes. L-arginine is proposed to have roles within multiple metabolic pathways and to shape immune functions [126–128]. Kim and colleagues [129] reported decreased L-arginine in serum of M. abscessus-PD patients as well as in M. abscessus-infected mice. Oral administration of L-arginine to mice infected with M. abscessus was associated with remodelling of the gut microbiota, and a reduction in pulmonary NTM burden and amelioration of lung histopathology. Transcriptomic profiling of the lung tissues revealed elevated expression of interferon (IFN)-γ and IL-12 in these M. abscessus-infected mice treated with L-arginine, suggesting enhanced Th1-mediated pulmonary immune responses. This example highlights how gut microbial alterations via microbe-derived metabolites might influence immune responses and disease severity in the lungs. However, further research is needed to establish whether there is a causal relationship between L-arginine-induced changes in the microbiota and enhanced pulmonary defence, especially as current evidence is derived from murine-infected mice and may not fully translate to humans. Additionally, L-arginine may also exert effects through other mechanisms, such as increasing nitric oxide synthase activity [130], which is known to have antimicrobial properties against mycobacteria [131,132]. To define the specific roles of microbe-derived metabolites and better understand the mechanisms contributing to the GLA in NTM-PD, more detailed characterisation of the microbiota and their functional capabilities in this condition is required.

Challenges in the lung microbiota in NTM-PD: the need for disease stratification and methodological consistency

Variations in the lung microbiota are inconsistent across studies examining NTM infection and NTM-PD. Of note, there has been a transition in the field from culture-dependent to culture-independent methods of microbiota analysis. As culture-independent methods of surveying the microbiota are less affected by the biases introduced by culturing, we focussed our assessment on studies based on culture-independent analyses (details of studies outlined in Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary table of culture-independent studies characterising the impact of NTM infection or NTM-PD on the lung microbiota.

https://doi.org/10.1371/journal.ppat.1013603.t001

To date, there is no clear consensus on whether NTM infection or NTM-PD is consistently associated with changes in microbial diversity. Some studies report no change in diversity [133–136], others observe a decrease [137–140], and few report an increase [95,141,142]. One hypothesis is that microbial diversity remains stable during transient infection, but declines with disease progression. This pattern mirrors observations in chronic P. aeruginosa infection in cystic fibrosis, where diversity declines only once chronic infection is established [73]. However, testing this hypothesis in NTM-PD is challenging due to the difficulty in distinguishing transient NTM infection from true NTM-PD in some individuals. Sulaiman and colleagues [133] found no difference in diversity between NTM-positive and NTM-negative bronchiectasis samples. However, all participants in this study exhibited respiratory symptoms and imaging abnormalities suggestive of NTM-PD, making it difficult to determine whether they were transiently infected or had established disease. Alternatively, the similar diversity observed between groups could reflect a shared underlying lung pathology, namely bronchiectasis. No difference in diversity was also observed in a study by Song and colleagues [135], which compared stable NTM-PD and progressive NTM-PD cases (those requiring treatment). This similarity could reflect that both groups had established disease, and that difference might only emerge when compared to transient infection—a group not included in the study. Future research would benefit from longitudinal studies that can track changes in the lung microbiota from transient infection through to established NTM-PD and progressive NTM-PD. One such study involving cystic fibrosis patients with NTM infection or NTM-PD, although not reporting specific diversity metrics, identified distinct differences in microbial network clustering between transient NTM infection and persistent NTM-PD [143]. The networks suggested that microbial communities in persistent NTM-PD were more interconnected and less modular, indicating reduced community stability in disease. Although not all individuals with NTM-PD follow a trajectory from transient infection to progressive disease, such analyses may offer valuable insights into the relationship between lung microbiota dynamics and disease progression.

In studies reporting decreased diversity, overgrowth of specific taxa is often observed. For instance, in a bronchiectasis cohort, NTM-positive individuals exhibited significantly reduced diversity compared to NTM-negative counterparts, with overrepresentation of Mycobacterium, Pseudomonas, Veillonella, Actinomyces, and Fusobacterium [138]. Similarly, decreased diversity in an NTM-PD cohort compared to controls was observed alongside overgrowth of Pseudomonas and Rhodococcus [139]. These findings align with observations in other chronic lung diseases, where reduced microbial diversity is seen in established disease and often driven by the dominance of specific taxa [71,73,144]. Of the studies reporting increased microbial diversity associated with NTM-PD, one was a pilot study significantly limited by sample size [141], the other provides a paired analysis of disease-involved segments compared to non-involved segments in resected lung specimens from individuals with NTM-PD [95]. It is possible that in advanced NTM-PD (such as cases requiring surgical resection) microbial diversity increases as the disease progresses, potentially due to invasion by other bacterial species or immune exhaustion. To better understand these dynamics, further research is needed to quantify changes in microbial diversity across the spectrum of NTM-PD severity.

Specific microbial taxa have been proposed as potential biomarkers for distinct disease characteristics in NTM-PD; however, substantial variability between studies limits the identification of consistent markers and their clinical applicability. Nevertheless, some patterns may be beginning to emerge. Anaerobic genera such as Veillonella [137,138,143,145], Fusobacterium [138,145], and Prevotella [145] are frequently reported at higher relative abundance in individuals with NTM infection or NTM-PD compared to healthy controls or those with NTM-negative bronchiectasis. These same taxa are among the most common in health and other chronic respiratory diseases [63,146], suggesting their presence is not NTM-PD specific. Whether the increased abundance of these anaerobic microbes could relate to changes in the airway microenvironment in individuals with NTM-PD, such as the development of hypoxic niches driven by mucus stasis [90], could be a future research question. The potential for sputum microbiota biomarkers to distinguish between disease progression and stability in NTM-PD has also been suggested, but findings remain inconsistent and unvalidated. For instance, Prevotella nanceiensis was found in one study as a marker of NTM-PD stability [135], while another study constructed a predictive model for NTM-PD progression using seven genera (Burkholderia, Pseudomonas, Sphingomonas, Candidatus, Saccharibacteria, Phocaeicola, Pelomonas and Phascolarctobacterium) [140]. Among these genera, Pseudomonas is frequently reported, albeit inconsistently. Pseudomonas abundance was increased in a study in NTM infection [138] and a study in NTM-PD [139], yet another study found it absent in all MAC-PD samples, with authors proposing a possible mutual exclusivity between Pseudomonas and MAC in the lung environment [134]. While such biomarkers may hold promise for future disease monitoring, the variability across studies currently limits the development of robust predictive models or biomarkers for disease progression.

We purport that the variability in findings and absence of a distinct microbial profile compared to healthy lungs reflects methodological heterogeneity rather than the absence of a disease-associated microbiota in NTM-PD. Most studies have included varying proportions of different species within their NTM-PD cohorts [95,133,135,137–139,142,145], primarily MAC and M. abscessus, but also some with smaller percentages with less commonly seen pathogens in NTM-PD such as M. xenopi and M. lentiflavum (specific percentages for each study are detailed in Table 1). Few studies have focussed on a single species, specifically MAC-PD [134,141], or M. abscessus NTM-PD [136]. This lack of species-specific focus is problematic, as data suggests differential taxa enrichment in lung tissue microbiota depending on NTM causative species (MAC vs M. abscessus) [95], and even potential differences in sputum diversity and taxonomy based on subspecies of M. abscessus (massiliense versus abscessus) [136]. Future studies should, at a minimum, ensure species-level consistency within NTM-PD groups, with consideration given to subspecies-level distinctions as well. Similarly, there may be differences in microbiota composition associated with radiographic phenotype (fibrocavitary versus nodular bronchiectatic) [95]. However, most studies have neither stratified participants by phenotype nor consistently reported it. Incorporating radiographic phenotype into future analyses may help clarify microbiota-disease associations and improve the interpretability.

Inconsistencies in the types of respiratory samples used to assess the lung microbiota in NTM-PD significantly hinder cross-study comparisons. Sputum samples are widely used in clinical and research settings due to their ease of collection and capacity to capture a broad range of microbial genera present in the lungs. Comparative analyses have shown that bronchoscopy does not yield additional genera beyond those identified in sputum samples [147]. Nevertheless, sputum represents a surrogate measure of the lower respiratory tract, as it comprises a composite of mucus from the lower respiratory tract and aqueous secretions from the oral cavity acquired during transit through the upper airway. This passage introduces contamination from oropharyngeal sources [147,148], which can confound interpretations of microbial composition. Studies comparing sputum with BALF and protected specimen brushings [147], and comparing sputum with direct lung sampling acquired at the time of transplantation [148], both in cystic fibrosis patients, reported significant differences in both microbial diversity and relative abundance, underscoring the limitations of sputum as a proxy for the lung microbiota. Similarly, a comparison of diversity and bacterial load in a subset of bronchiectasis patients, both NTM-positive and NTM-negative, suggested sputum microbial composition more closely resembled that of the oral cavity than BALF samples [133]. Other sampling techniques may be less affected by contamination. Intraindividual inter-site analysis in healthy individuals suggests BALF is not confounded by contamination from the oral microbiota [63]. Protected sampling methods, particularly protected specimen brushings (PSB), have also been shown to reduce oral contamination, potentially yielding even greater distinction from oral microbiota than other bronchoscopic techniques [149]. In NTM-PD, there were discrepancies when comparing protected specimen brushing and bronchial washing samples, with lower diversity in protected specimen brushings [139]. The lung microbiota is likely more accurately characterised using bronchoscopic samples than sputum samples, however, the clinical risks, invasiveness, and time-consuming nature of bronchoscopy limit its routine use. Nevertheless, to characterise the lung microbiota in NTM-PD, future studies require greater methodological consistency to allow for better comparison across the literature.

The gut microbiota in NTM-PD: many questions to be answered

Gut microbiota dysbiosis has been associated with pulmonary disorders, such as COPD [117], bronchiectasis [123], and cystic fibrosis [150], but specific studies of gut-associated dysbiosis in NTM-PD are limited. To date, three studies have reported on the faecal microbiota in NTM-PD patients [details in Table 2] [136,137,151].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary table of studies characterising the faecal microbiota in NTM-PD.

https://doi.org/10.1371/journal.ppat.1013603.t002

Gut microbiota profiles may differ between individuals with NTM-PD (who are treatment naïve) when compared to healthy controls. Choi and colleagues [137] and Lin and colleagues [151] reported significantly decreased diversity measures in faecal microbial profiles of individuals with NTM-PD compared to healthy controls. Both studies included relatively small sample numbers and were unable to stratify by NTM-PD phenotype, causative NTM species, or disease severity. Broadly, the observed decrease in diversity in NTM-PD, corresponds with the decrease in diversity seen in other infectious [124,152] and non-infectious lung pathologies [123]. Changes in taxonomy; however, were not consistent across the two NTM-PD studies despite both using linear discriminant analysis effect size (LEfSe) to compare taxa abundances between groups and to determine statistically significant biomarkers [137,151]. Both studies identified that members of the genus Prevotella were altered in NTM-PD patients [137,151]. However, Choi and colleagues [137] reported that Prevotella fusca, P. marshii, and P. nanceiensis were predominant in NTM-PD subjects, whereas Lin and colleagues [151] saw reductions in the relative abundance of Prevotella_9 in NTM-PD patients. Species of the genus Prevotella synthesise SCFAs [153], which might offer potential focus for future studies to integrate microbial taxonomy and their metabolic capabilities in NTM-PD. The variable findings related to Prevotella may arise from inconsistent definitions of a ’healthy’ control population, which are rarely reported in microbiota studies. We propose that the most appropriate matched control would be individuals with similar underlying host phenotype and lung disease, but without NTM infection, recent antibiotic exposure, or other significant co-morbidities that may be affecting the microbiome. Identifying such controls is challenging, so studies may benefit from multiple comparison groups. Greater confidence in characterising NTM-PD-associated gut dysbiosis at the genus or species level will also require larger patient and control cohorts, and more detailed consideration of clinical parameters including, but not limited to, NTM causative species, disease severity and concomitant diseases or infections.

Gut microbiota profiles may also be related to disease severity measures or disease exacerbation. Decreased gut microbiota diversity and richness has been seen in acute exacerbations of COPD compared to stable disease [144], and in acute exacerbation of bronchiectasis compared to stable bronchiectatic disease [123]. In NTM-PD, Lin and colleagues [151] found that 11 bacterial genera had significantly different abundance compared to a healthy control groups. These 11 genera were correlated with NTM-PD disease severity measures, including the presence of a cavity, grading of sputum acid-fast smear and radiographic scores [151]. The strongest correlation was observed between greater disease severity and reduced levels of the Prevotella_9 genus [151]. Prevotella again emerges as a possible genus of interest, with potential relevance to NTM-PD severity. These findings warrant validation in larger, well-characterised NTM-PD cohorts. Further research should explore whether Prevotella, or other taxa, could serve as biomarkers of NTM-PD progression, and whether associated microbial shifts reflect changes in host metabolism, such as altered SCFA synthesis.

Antibiotic treatment and the gut microbiota in NTM-PD

Antibiotic treatment causes significant disruption of the gut microbiota, affecting both composition and abundance to varying degrees [61,114]. Antibiotics have different modes of action, and may selectively deplete specific subpopulations of microbes [154–156]. Even antibiotics from the same class can induce different and specific effects on the microbiota [157]. Specific drug, route, spectrum and concurrent disease [113,157], as well as interactions with other non-antibiotic medication and polypharmacy [158], are just some of the factors that make the impact of antibiotic treatment highly individualised and complex.

Suggested treatment for nodular-bronchiectatic MAC-PD, is a 3-drug regimen including a macrolide (e.g., azithromycin, ethambutol and rifampicin), with intravenous amikacin commonly prescribed in addition to the 3-drugs for cavitary, severe or refractory disease [2,15]. Treatment for M. abscessus-PD is more complex and variable than that for MAC-PD. The optimal drugs, regimens and duration for treatment of M. abscessus-PD are not pre-defined, and rather antibiotic choices are informed by mutational or inducible macrolide resistance patterns [2]. Given the substantial antibiotic burden, we hypothesise that NTM-PD therapy induces significant gut dysbiosis.

Inference of the impact of commonly used antibiotics in disease contexts other than NTM-PD may provide first insights. However, of the drugs used in NTM-PD treatment, to date, there is only information on azithromycin used independently. Azithromycin exposure has been associated with a significant decrease in gut microbial diversity in other disease contexts [155,157,159], and has been negatively correlated with presence of Actinobacteria [155,159] and Bifidobacterium [157,159]. These changes in microbial diversity and taxonomy are consistent with the broad-spectrum activity of azithromycin. In the context of the GLA and its role in NTM-PD, it is important to consider that antibiotic-associated microbial dysbiosis has been linked to downstream metabolic and immunological changes that may cause or exacerbate disease [155,157,160]. Metabolically, azithromycin exposure in healthy individuals resulted in Bifidobacterium depletion which was associated with increases in C-peptide, leading to reduced microbiota-associated glucose metabolism and reduced SCFA biosynthesis [157]. The same study identified azithromycin exposure was correlated with a significant decrease in serum pro-inflammatory cytokines, specifically monocyte chemoattractant protein (MCP)-1 and IL-5 in healthy individuals [157]. Azithromycin has also been reported to attenuate airway inflammation and reduce inflammatory markers in an allergic asthma mouse model [161], as well as to exert anti-inflammatory effects in the lungs of humans with asthma [162]. While these studies suggest possible azithromycin-induced perturbations, the extent, nature, and downstream inflammatory effects of these changes must be independently and thoroughly investigated in NTM-PD, as the impact of antibiotic perturbations may differ across disease states. For example, in bronchiectasis patients treated with erythromycin (which has a similar mechanism of action to azithromycin), significant alterations in microbiota composition were only observed in individuals who did not have an airway infection dominated by P. aeruginosa [113]. We hypothesise that similar variability could exist between NTM species, highlighting the need for future studies to examine both drug and species-specific effects on the microbiota. Further research is needed to clarify how azithromycin and other antibiotics used in NTM-PD, both individually and as multi-drug therapy, affect the gut microbiota. The impact of antibiotic treatment on the lung microbiota remains underexplored, both in respiratory diseases and NTM-PD, despite evidence suggesting microbial disruption [113]. Investigating antibiotic effects on both the gut and lung microbiotas, and their potential cross-talk via the GLA, may offer valuable insights into NTM-PD pathophysiology.

To our knowledge, there is one study to date that documents the impact of antibiotic treatment on the gut microbiota in NTM-PD; a study in 27 M. abscessus-PD patients with faecal microbiota analysed at baseline, 2 weeks and 6 months of antibiotic treatment (25 of 27 patients on a treatment regimen of azithromycin + amikacin + imipenem/cefoxitin ± clofazimine) [136]. This study provides a clinically relevant perspective by comparing microbiota signatures of M. abcessus-PD patients who are treatment-‘responders’ (defined as negative mycobacterial sputum culture at 2 weeks of treatment) to ‘non-responders’ (persistent positive mycobacterial culture at 2 weeks) [136]. The ‘responders’ group showed a significant decrease from pre-treatment baseline faecal microbiota diversity by 2 weeks of antibiotic treatment, and were found to have increased Enteroccocus abundance at week 2 [136]. ‘Non-responders’ displayed relative stability in diversity from pre-treatment to 2 weeks and continuing stability at 6 months of antibiotic treatment [136]. Diversity measures at 2 weeks and 6 months were significantly lower in the ‘responders’ group than ‘non-responders’ group [136]. ‘Non-responders’ were also found to have increased abundance of Eubacterium hallii group at baseline compared to ‘responders’ [136]. Given the long duration of NTM-PD treatment [2,3,163], determining treatment-responsiveness at 2 weeks is relatively early. However, 14 of the 15 treatment-responders at 2 weeks remained culture negative at 6 months and only four patients displayed a different outcome at 6 months compared to 2 weeks [136]. As the antibiotic regimens in the two patient groups were not significantly different, it was hypothesised that non-responders may have a host-dependent mechanism underlying their treatment unresponsiveness and relatively stable gut microbiota diversity [136]. These observations encourage future studies to assess whether relative stability of microbial diversity following commencement of treatment could be an early indicator for treatment non-responsiveness, or whether taxonomic biomarkers, such as Enteroccocus abundance or Eubacterium hallii group may predict a patient’s favourable or unsuccessful antibiotic treatment outcome. Such early indicators would hold great potential for early stratification of patients to pursue alternative treatment options. Given the gut microbiota’s potential influence on antibiotic pharmacokinetics, the cyclical relationship between the microbiota and treatment responsiveness may benefit from therapeutic drug monitoring alongside microbiota composition analysis in the future. Differences between M. abscessus subspecies (massiliense or abscessus) [136] and extension to other NTM-PD causative organisms [137,151] should also be considered, to assess potential for universal, early biomarkers of NTM treatment success or failure.

Recovery dynamics following antibiotic treatment—beyond the immediate perturbation—are crucial, as prior antibiotic exposure may influence susceptibility to future infections [154]. Initial and long-term dysbiosis post-treatment cessation depends on the degree of perturbation, the breadth of spectrum of antibiotics, microbial resistance, and the capacity of the microbiota to re-establish equilibrium following perturbation [164]. Reduction in microbial diversity can favour re-colonisation with antibiotic-resistant pathogens [160]. Recovery rates following antibiotic treatment in other diseases are highly variable, from as little as a few weeks, to persistent perturbation up to 14 months post cessation of antibiotic treatment [165,166]. In NTM-PD, studies to date have not addressed long-term gut microbiota composition following treatment. This research will be complicated by the high rates of co-infection with respiratory pathogens in individuals with NTM-PD [167,168]. Antibiotic treatment of these co-infections may cause dysbiosis, a factor that warrants consideration in both research and clinical settings. We hypothesise that the high rates of reinfection and recurrence of disease in NTM-PD may relate, at least in part, to the dysbiosis associated with long-term antibiotic use. Understanding the impact of both NTM-PD-specific antibiotic therapy and treatment for respiratory co-infections on the microbiome is essential, including how the microbiome recovers post-treatment and how these recovery dynamics may influence the risk of relapse or reinfection.