New Opportunities for Managing Acute and Chronic Lung Infections

PERSPECTIVES

and their potential interactions with each OPINION other and airway commensals. We discuss the emergence of antibiotic multidrug- New opportunities for managing resistant (MDR) organisms and highlight the potential for the global spread of pathogens through circulation in cities, schools and acute and chronic lung infections hospitals. We then suggest novel diagnostic and therapeutic possibilities for mitigating William O. C. M. Cookson, Michael J. Cox and Miriam F. Moffatt these serious problems. We propose that Abstract | Lung diseases caused by microbial infections affect hundreds of millions the opportunities to improve patient care are substantial and that a coordinated and of children and adults throughout the world. In Western populations, the treatment of collaborative international Lung Microbiome lung infections is a primary driver of antibiotic resistance. Traditional therapeutic Project (to consider the collective genomes strategies have been based on the premise that the healthy lung is sterile and that of the microorganisms that reside in the infections grow in a pristine environment. As a consequence, rapid advances in our lung) together with systematic studies of understanding of the composition of the microbiota of the skin and bowel have not individual organisms (the microbiota) will be of inestimable value to global health. yet been matched by studies of the respiratory tree. The recognition that the lungs are as populated with microorganisms as other mucosal surfaces provides the The burden of lung infections opportunity to reconsider the mechanisms and management of lung infections. Overt respiratory infections are the leading Molecular analyses of the lung microbiota are revealing profound adverse responses cause of death in developing countries and to widespread antibiotic use, urbanization and globalization. This Opinion article are associated with more than 4 million 16 proposes how technologies and concepts flowing from the Human Microbiome deaths annually . In children under 5 years of age, pneumonia, which is usually Project can transform the diagnosis and treatment of common lung diseases. associated with a bacterial infection, accounts for 1.3 million deaths annually The airways and the lung are in a constant Until recently, standard medical teaching and 18% of all-cause mortality17. Moreover, state of challenge from environmental held that the healthy lung is sterile9 and that pneumonia kills far more adults than HIV microorganisms. A sedentary adult can pathogens invade an environment that is or malaria16. Globally, lower respiratory inhale more than 10,000 litres of air per normally kept free of microorganisms by the infections result in the loss of 103,000 day1, which washes over 40–80 m2 of lung host mucosa. However, the thoracic airways, disability-adjusted life years, which means surface2,3. In cities, this air may contain up which are the site of many infections, that lung infections are the largest cause of to 100,000 bacteria per litre4, a proportion are now known to support a distinctive disease burden to humanity. of which are able to penetrate into the community with a density of bacteria Acute infections can afflict healthy thoracic airways and beyond. By contrast, similar to that of the small bowel10,11. The adults but are more common in the the gut has a surface area of 30 m2 (REF. 5), consideration of how diverse commensal very young or very old or in those with and microbial ingression is limited by microbial communities resist exogenous underlying health problems. They usually gastric acid. Thus, it is not surprising that infection12–14 has had an enormous influence affect the lung parenchyma (pneumonia), lung infections commonly affect human on understanding and treating bowel the airways of the lung (bronchitis) or health. In the United Kingdom alone, diseases15, and it is timely now to reconsider both (bronchopneumonia) (FIG. 1). Less 16 million patients of the National Health received wisdom for the management of frequently, infections can spread to the Service are treated with antibiotics each year lung infections. Accessing the lower airways linings of the lung, causing pleurisy and for respiratory tract infections that include for routine microbial analyses typically empyema. CAP differs from hospital- 260,000 episodes of community-acquired involves invasive procedures such as acquired pneumonia (HAP) and pneumonia (CAP)6. Antibiotic prescription fibre-optic bronchoscopy. These practical ventilator-associated pneumonia (VAP), rates are higher in many other European difficulties, and the low priority given which occur in patients who are sick or countries than in the United Kingdom7. to non-HIV lung disease by the Human immunocompromised and result in high In the United States, acute lung infection Microbiome Project (HMP), mean that the morbidity, mortality and costs18. in the presence of underlying chronic technological and analytical advances of In contrast to acute infections, chronic obstructive pulmonary disease (COPD) is the HMP have been sparsely applied to the lung infections are typically a consequence the most common reason for administering airways and the lung. of underlying disease. For example, antimicrobial therapy to adults; this clinical In this Opinion article, we review the bronchiectasis describes dilations in the setting plays a substantial role in driving most common pulmonary infective agents, thoracic airways (FIG. 1), which can be antimicrobial resistance8. with an emphasis on bacterial pathogens caused by tuberculosis, scarring after

PERSPECTIVES

Tuberculosis Lobar pneumonia inflammation that causes wheezing and Chronic cavitating Alveolar consolidation: lung infections Gram-positive cocci (S. pneumoniae) shortness of breath. Notably, asthma with abundant affects 250–300 million individuals of all acid-fast bacilli ages, races and ethnicities globally, and (M. tuberculosis) its prevalence is increasing alarmingly in the developing world21. Urbanization, with the loss of the microbial environment Bronchopneumonia Healthy from traditional rural lifestyles, strongly Inﬂammation in airways airway 22 and parenchyma: increases susceptibility to disease . Gram-negative rods Exacerbations are often triggered by viral (H. inﬂuenzae) infections23. Asthmatic airways typically Commensal contain considerable numbers of bacterial bacteria pathogens24–28 (FIG. 1) and may also carry as yet unexplored fungi. However, a causal link between these organisms and disease is yet to be established. Bronchiectasis Abnormal dilated airways full of The microbiota of healthy lungs secretions and Pulmonary infections with pathogens polymicrobial infections take place in the presence of commensal lower-airway microbial communities (the microbiota; reviewed in REF. 29). In brief, the sequencing of variable regions of the 16S Mucus rRNA gene has shown that healthy airways contain in the region of 100 operational taxonomic units (OTUs; as defined by Empyema Airway disease sequence similarity), which are dominated Infection in the Chronic dominance of pathogens by the Firmicutes (primarily Streptococcus pleural space: in asthmatic and COPD airways: spp. and Veillonella spp. OTUs) but also (S. pneumoniae Gram-negative Protebacteria and M. tuberculosis) (Haemophilus spp., Neisseria spp. contain Bacteroidetes (mostly Prevotella and Moraxella spp.) spp.), Fusobacteria, Actinobacteria and 10,24–28 Figure 1 | Pulmonary infections associated with bacterial pathogens. The airways and lung can Proteobacteria . Many of these OTUs may represent facultative anaerobes that are be affected in a number of different ways by bacterial infection, as shownNature clockwise Reviews | Microbiologystarting from the top right: lobar pneumonia affects a whole lobe of the lung and may spill into the surrounding recalcitrant to routine culture. The limited pleura (two thin layers of tissue that protect the lungs and allow them to move during breathing). number of phyla and the consistency of the Lobar pneumonia is typically caused by Gram-positive short chains of Streptococcus pneumoniae. oropharyngeal and lower-airway microbiota Bronchopneumonia is centred in the bronchial tree and spreads into adjoining lung tissues; it is most between multiple individuals and across commonly caused by the Gram-negative rods of Haemophilus influenzae. Asthma and chronic studies10,30–32 indicate that the airways are obstructive pulmonary disease (COPD) are diseases of the small airways, characterized by over- not simply passive recipients of inhaled growth of Proteobacteria (such as H. influenzae and Neisseria spp.) and the loss of normal commensal microbiota and provide strong evidence for species. Empyema is an infection of the pleural space and is most commonly caused by S. pneumo- host selection. These microbial communities niae. The chronic lung diseases bronchiectasis and cystic fibrosis are associated with chronic resist- 30 ant infections with abnormal bronchial morphology; these infections are caused by H. influenzae, are present in infants , and airways seem S. pneumoniae and opportunistic pathogens such as Staphylococcus aureus (a Gram-positive coccus) likely to be colonized in a way similar to and Pseudomonas aeruginosa (a Gram-negative flagellate bacillus). Mycobacterium tuberculosis other body surfaces33 during the weeks causes chronic consolidation and cavitation with invasion into draining lymph nodes and eventual after birth. systematic spread. In health, the oropharynx, the thoracic airways and the fluids from bronchoalveolar lavage (BAL; fluid washings obtained acute infection, primary ciliary dyskinesia, cigarette smoking. Chronic lung diseases through a bronchoscope wedged into the immune deficiencies and cystic fibrosis such as COPD are all accompanied by subsegmental airways) contain similar (in populations of European descent). periods of acute worsening of respiratory microbiota10,11, and the extent to which the Bronchiectasis from all causes results symptoms, known as exacerbations. The lower airways are populated by aspiration in the progressive production of thick, aetiology of exacerbations encompasses from the upper respiratory tract is debated34. purulent secretions (sputum) and an host immune responses as well as complex However, in disease, the presence and advancing destruction of the airways and changes in the viral, fungal and bacterial abundance of pathogens in the lung are not lung parenchyma. communities19. In COPD, acute infective consistently reflected in the oropharynx10,34, Another underlying disorder that exacerbations cause airway inflammation which emphasizes that direct sampling of predisposes to infection is COPD, which and hasten disease progression20. the lower airways and BAL are desirable in is typified by inelastic constricted airways In other conditions, the relationship studies of the pulmonary microbiota34, with with or without loss of lung parenchyma with infection is more complex. Asthma strict controls for contamination during (emphysema) and is usually the result of is a syndrome of intermittent airway bronchoscopy11 and laboratory analyses35.

PERSPECTIVES

Pulmonary infectious agents strains of S. pneumoniae are carried by between rhinoviruses and H. influenzae Pulmonary bacterial pathogens have been approximately 10% of children and cause augment rises in inflammatory markers and discovered almost entirely through microbial otitis media39. Their role in respiratory exacerbation severity in COPD49. culture. Historically, the most important disease has not been assessed. The carriage Finally, fungi form symbiotic as well lung pathogen has been Mycobacterium of S. pneumoniae and H. influenzae by a as pathogenic relationships with bacterial tuberculosis. Between the 17th and 19th substantial proportion of healthy children and eukaryotic communities. Some fungal centuries, it accounted for one-fifth of all indicates that colonization per se does not organisms can cause distinct syndromes of deaths in Europe and North America36. cause disease. The important factors that lung infection50. Furthermore, substantial It may lie latent for years before causing may cause the shift from colonization to fungal growths may be present in the a chronic cavitating lung infection that invasion and overt disease (BOX 1) have not microbiota of patients with cystic fibrosis produces highly infective sputum (FIG. 1). been extensively studied but may include and bronchiectasis51. Thus, important After a marked decline in prevalence, due concomitant viral infections, exposure interactions seem likely between fungal primarily to improvements in public health, to cigarette smoke and environmental and bacterial pathogens52. Invasive M. tuberculosis infections are now increasing pollution and progressive growth over the fungal disease associated with emergent in incidence, with MDR strains spreading in host microbiota40. organisms is increasing in incidence in many populations. Interestingly, different bacterial immunocompromised hosts53. In addition, Another important pathogen is pathogens are associated with the chronic fungal pathogens may be important in Streptococcus pneumoniae, the most infections that accompany cystic fibrosis and HAP but have been incompletely studied18. frequently detected bacterial cause of bronchiectasis. Pseudomonas aeruginosa, The difficulty in isolating fungi by culture acute pneumonia, which typically presents a facultative aerobic proteobacterium almost certainly means that detection with high fever, lobar consolidation and that causes opportunistic infection and is rates do not reflect the prevalence of pleurisy (FIG. 1). In addition, Haemophilus associated with antibiotic resistance, is an potentially pathogenic species in clinical influenzae infection is common in patients important pathogen, as are Burkholderia specimens. Fungi can cause disease by with smoking-related lung disease, causing cepacia, Staphylococcus aureus and varied mechanisms. Direct invasion of bronchial inflammation with or without nontuberculous mycobacteria (BOX 1). immunocompromised tissues may, at times, patchy infiltration into the surrounding Notably, the microbiota of those with be less important than overexuberant host lung (bronchopneumonia). H. influenzae is chronic lung diseases can contain abundant responses to the presence of the fungus. underdetected by standard clinical culture other potential pathogens; H. influenzae and For example, in allergic bronchopulmonary methods37. Other Proteobacteria that less S. pneumoniae most frequently dominate aspergillosis, colonization of the airways commonly cause acute lung infections the sputum and airway microbiota of both by Aspergillus fumigatus is accompanied by include Moraxella spp. and Neisseria spp. patients with cystic fibrosis and patients elevated blood levels of immunoglobulin E Unencapsulated non-typeable H. influenzae with bronchiectasis. Common pathogens and eosinophils. Treatment of the immune (NTHI) strains are the main drivers of isolated by culture from patients with HAP response with corticosteroids is currently recurrent airway infection in COPD, and VAP include Pseudomonas spp. and given priority over fungal eradication. together with Moraxella catarrhalis and other Gram-negative bacilli, staphylococci, S. pneumoniae38. Asthmatic airways appear streptococci and Haemophilus spp.41. Mixed Current therapeutic strategies to be similarly dominated by Proteobacteria or polymicrobial infections occur in patients Although the aim of this article is to (which may be Neisseria spp. rather than who are immunocompromised or suffering suggest improvements and new directions H. influenzae) and Firmicutes (potentially from chronic lung infections, posing further to the management of lung infections, S. pneumoniae)24–28, although asthma- therapeutic challenges42. it is worthwhile first to summarize associated OTUs have yet to be defined by In addition to bacterial pulmonary contemporary diagnosis and treatment of isolated culture. disease, viral respiratory infections are these illnesses, with an emphasis on the The microorganisms described above common43, and their ability to spread problems that are faced. have the ability to adhere to and invade rapidly and cause epidemics has a major other airway mucosal surfaces. In addition, global impact. Although viral infections Current diagnosis. Bacteriological S. pneumoniae and capsulated forms of are not the subject of this review, it should diagnosis for lung infections in most H. influenzae and Neisseria meningitidis be recognized that they do not act alone. clinical laboratories depends on isolating are the principal causes of bacterial H. influenzae was first discovered in 1892 pathogens from patient samples by meningitis, and S. pneumoniae, NTHI and during an influenza epidemic44, and bacterial microbial culture. The technology for M. catarrhalis also cause chronic infections pneumonia was the principal cause of death culture on agar plates was invented in 1880 of the middle ear (otitis media). All these in the 1918–1919 Spanish flu pandemic45. in Berlin54. Plate culture can be effective in organisms are commonly carried in the nose Viral infections often trigger exacerbations isolating pathogens, followed by testing for and oropharynx by healthy children (BOX 1), of asthma and COPD, and multiple microbial sensitivities and identification and extensive vaccination programmes are interactions can occur between viruses and by molecular techniques such as mass changing the prevalence of their capsulated common bacterial pathogens during acute spectrometry55. However, microbial culture forms. However, it is unencapsulated NTHI infection46. HIV infection, which supresses has many subjective elements, beginning strains (which are difficult to vaccinate the adaptive immune system, increases the with pathogen differentiation from against) that drive recurrent airway risk of pneumococcal pneumonia 20‑fold47. commensals by colony morphology. Failure infection in COPD38, and colonization with Mechanistically, influenza infection to identify causal organisms occurs in 50% these strains is also associated with wheeze promotes adherence and invasion by of patients who have been hospitalized with in infants24. In addition, unencapsulated S. pneumoniae48. In addition, interactions pneumonia in intensive care units56, with

PERSPECTIVES

similar low microbial diagnostic rates in presence of polymicrobial infections is often pathogens’ that pose the greatest threat to pneumonia following immunosuppressive beyond current diagnostic capabilities, which human health; of these, four affect the lungs therapy or bone marrow transplants57,58. can make it difficult to make rational clinical (P. aeruginosa, S. pneumoniae, H. influenzae The administration of antibiotics before decisions about antimicrobial therapy. and S. aureus). specimen collection is an omnipresent Antibiotics for respiratory infections problem. Furthermore, the use of specialized Current treatments. Although it is are usually given orally, with β‑lactams culture media or anaerobic culture can universally accepted that widespread use often the drugs of first choice. Although the stretch the capacity of routine clinical of antimicrobial agents is causing a global upper airway microbiota appears somewhat laboratories, and protocols may be erratically crisis in antibiotic resistance59, it has not resilient and returns to the preantibiotic guided by clinical suspicion. been sufficiently recognized that antibiotic state quickly, in the gut, substantial selection Deconvoluting the complex interactions prescription for lung infections is a primary for antibiotic-resistant organisms can be between the microbiota and different contributor to this problem6,8. In 2017, the detected many months after treatment60. pathogens in acute exacerbations or in the WHO listed 12 antibiotic-resistant ‘priority Because early treatment for lung infections

Box 1 | The mobility and evolution of the lung microbiota Respiratory pathogens such as Streptococcus virulent multidrug-resistant (MDR) Burkholderia causing serious invasive infections with a pneumoniae, Haemophilus influenzae, Neisseria cepacia was discovered in 1990 (REF. 127), before mortality approaching 60%134. C. auris infections spp. and Moraxella spp. are all commonly detection of transmission of MDR strains of were first reported in Japan, followed by reports present in healthy individuals yet lead to disease P. aeruginosa128 and Mycobacterium abscessus129. from South Korea, India, Kuwait, South Africa, in only some. They require human contact to Alarmingly, many M. abscessus infections are Pakistan and the United Kingdom and then spread, and the extent of the contact differs caused by dominant clones that have spread South America and the United States134. in many environments (see the figure). These internationally130. Knowledge of these risks to These studies show how cosmopolitan MDR bacteria may be carried in the nose and patients with cystic fibrosis has led to aggressive strains spread nationally and internationally in a nasopharynx but cause disease in the lower measures to control transmission of MDR strains highly connected world, giving new meaning to airways and lung parenchyma. Factors that can in hospitals and clinics131. the ‘everything is everywhere’ hypothesis135. trigger the switch between colonization and The spotlight has now turned to emerging The next decade promises to yield much more infection include the number of bacteria in fungal pathogens in cystic fibrosis132, information about virulent strains of more vulnerable sites, synergism and nutrient exemplified by Scedosporium spp.133. In patients mundane pathogens, such as H. influenzae. It is competition with commensals122, intercurrent without cystic fibrosis hospitalized for a long almost certain that measures to control the viral infections, cigarette smoking and air time with lines or tubes entering their body, spread of organisms from all clinical pollution. By contrast, bacteria that cause a recently emergent MDR Candida auris is circumstances will need to be greatly changed. opportunistic infections of damaged lungs (for Healthy diverse example, in patients with cystic fibrosis), such as commensal bacteria Pseudomonas aeruginosa and nontuberculous Airborne mycobacteria, follow the exemplar of many Pathogen Pathogen pathogens enteropathogenic bacteria by having an colonization dominance environmental reservoir123 where they can persist before infecting susceptible hosts. The potential for viruses to mutate and spread pandemically is universally feared. Less attention has been given to transmission of Pathogen bacterial and fungal pathogens, although Airway Mucus invasion international travel, high population densities epithelial and selective pressures for pathogenicity cells provide fertile grounds for such events. For example, whole-genome sequencing of Mycobacterium tuberculosis isolates shows that this exclusively human pathogen emerged as a chronic infection 70,000 years ago, following Inﬂammation modern humans out of Africa and evolving traits of virulence and transmissibility to take advantage of progressive human overcrowding124. M. tuberculosis sequencing can now trace where modern outbreaks arise and how they spread125. SCHOOL The movement of pathogens within hospitals was first shown by whole-genome sequencing Rural populations Cities Schools and nurseries Hospitals and clinics during an outbreak of lung infections with • Rich microbial • Large pathogen pool • Naive immunity • Intense mixing 126 environment • Much mixing • Selection for • Strong selection carbapenem-resistant Klebsiella pneumoniae . • Small pathogen pool • Selection for AMR transmission for virulence Children with cystic fibrosis are highly prone • Low mixing • More virulent strains • Strong selection to chronic lung infections and are thus in circulation for AMR subject to multiple antibiotic administrations, Intensity of mixing often within hospitals and their associated clinics. Patient-to-patient transmission of AMR, antimicrobial resistance. Nature Reviews | Microbiology

PERSPECTIVES

improves outcomes, most patients are given Illness Problem Mitigation antibiotics without a microbial diagnosis, and British Thoracic Society guidelines Acute community- • Microbial diagnosis not sought acquired pneumonia • Broad-spectrum antibiotics Rapid PCR urge empirical antibiotic use in patients diagnostics treated for CAP in primary care61 (FIG. 2). In addition to patients’ toxic and allergic Existing disease • Recurrent infections Selective responses to antibiotics, antibiotic damage to (COPD) in the • Microbial diagnosis not sought inhaled community • AMR enhanced antibiotics resident microbial populations considerably • Commensal depletion increases susceptibility to bowel infections62. Similar processes are possible in the • Previous antibiotics Vaccination lung, and the balance between the beneficial • No microbial diagnosis in 50% of patients Acute hospital- • AMR and virulence gained from and detrimental effects of antibiotics could acquired pneumonia environment Commensal profitably be investigated in COPD, cystic • Immunosuppression and restoration fibrosis and bronchiectasis. concomitant disease in recurrent • Commensal depletion The therapeutic strategy in patients disease with pneumonia of unknown aetiology is management diﬃculty Progressive to escalate the complexity and intensity • Complex pathogen load Enhanced Chronic infection and inaccurate microbial diagnosis microbial of the antibiotic regimen, often with managed in • Chronic prophylactic antibiotic use culture with aminoglycosides, with the associated risks hospital clinics • Transmission of AMR and virulence by whole-genome emergent pathogens sequencing of selecting for and disseminating MDR • Commensal depletion (FIG. 2). The microbiological causes of | HAP and VAP vary by location, and MDR Figure 2 Current therapeutic strategies for lung infections. The flowNature chart Reviews details | problems Microbiology with to these organisms is almost universal. the current diagnosis and treatment strategies for various lung infections, from simple community- Broad-spectrum antibiotic therapy for acquired pneumonia to the therapy of chronic bronchial infections. Potential generic improvements HAP and VAP is given at the onset of to disease management are shown on the right. AMR, antimicrobial resistance; COPD, chronic obstructive pulmonary disease. disease41. De‑escalation of antibiotics is recommended once culture data are available41, but this is unlikely to prevent samples is the prospect of improved interpretation than one accounting for selection for antimicrobial resistance accuracy of clinical diagnostics (FIG. 2). For 2%). Similar diagnostic possibilities (AMR) in the gut60. example, 16S rRNA gene sequence analysis exist for fungal pathogens by use of, for Antibiotic prophylaxis is widely used of DNA isolated from clinical specimens example, high-throughput sequencing of in the management of bronchiectasis, as can better identify poorly described, rarely the amplified internal transcribed spacer 2 exemplified by children with cystic fibrosis63. isolated or phenotypically aberrant strains. (ITS2) region70. Diagnostic fungal cultures The isolation of S. aureus and P. aeruginosa It can be routinely used for identification are technically difficult and currently take by microbial culture is related to worse of mycobacteria (which are difficult to 2–4 weeks of incubation71, so there are outcomes, but antibiotic prophylaxis against identify by culture-based methods) and can considerable possible speed and accuracy staphylococci has no clear benefit and lead to the recognition of novel pathogens benefits in switching to sequence-based may increase carriage of P. aeruginosa64. and bacteria that are resistant to standard methods of detection72. Attempting an early, targeted eradication clinical culture67–69 (BOX 2). By and large, Problems with diagnosis by the 16S of P. aeruginosa, similarly, has no definite 16S rRNA gene sequencing from bacterial rRNA gene and ITS2 sequencing derive beneficial effect on mortality or morbidity65. isolates has been envisaged for microbial from the accuracy of sequences in databases identification67–69. However, 16S rRNA gene as well as those generated by the diagnostic Evolving therapies analysis of the whole bacterial community laboratory and may result in an uncertain Many of the mutations that confer AMR from biological specimens such as sputum definition of species when based only on come with a fitness cost to the bacteria and samples collected from throat swabs sequence information instead of on the that carry them66, which offers the may provide direct information of the traditional phenotypic classification of hope that finishing a course of antibiotics presence of pathogens without the need for cultured microbial isolates67,68. In addition, may eventually be followed by reversion of culture (BOX 2). multiple copies of the 16S rRNA gene may the microbiota towards its pre-antibiotic In contrast to open-ended sequence be present and show sequence variation state. Continued efforts to prescribe fewer analysis of 16S rRNA gene amplicons, in some bacterial species69. Limited 16S antibiotics are essential, but other innovative specific quantitative PCR (qPCR) assays for rRNA gene sequence variation makes approaches to therapy are now possible. common pathogens are already possible and discrimination difficult between different They are summarized in FIGS 2,3 and may deliver results in a fraction of the time Streptococcus spp., including S. pneumoniae, discussed in more detail below. of standard clinical cultures. Additionally, which may be overcome by sequencing other qPCR with general primers can measure polymorphic loci73. Improved diagnostics. Modern sequencing the overall number of all bacterial genomes, In the clinical context, rigorous control is technology offers a variety of alternative which can be combined with species- necessary to prevent microbial nucleic acid diagnostic approaches to traditional specific assays to interpret the relative contamination of swabs, containers of human culture-based methods (BOX 2). An dominance of a particular pathogen in the fomites, PCR kits, sequencing kits and immediate benefit of sequence-based microbiota (as an organism accounting for reagents35. Mixtures of bacteria that have a detection of the microbiota in clinical 50% of the microbiota carries a different different resistance to lysis74 can be overcome

PERSPECTIVES

in health and disease. Sputum, lung Pathogen biopsy samples and brushings, unlike inhalation or stool samples, are problematic for proliferation microbial metagenomics because the Community ratio of human to microbial DNA may diversity and health be higher than 99:1. Strategies to enrich for microbial DNA before making sequencing libraries are technologically possible82 but not yet effective in ‘Magic bullets’ ‘The bludgeon’ practice. Promisingly, next-generation Vaccination sequencing of RNA from BAL samples has Narrow- proved effective in diagnosing viral and spectrum bacterial infections83. antibiotics Phage Community Broad-spectrum Improved antibiotic regimens. The depletion antibiotics therapy emergence of AMR in pathogens shares many factors in common with the emergence of drug-resistant malignant cells during chemotherapy for cancer. The management of AMR may be improved by strategies that are similar to those Community Community Antimicrobial used for cancer chemotherapy, where the preservation restoration resistance quantification of resistance mechanisms has been followed by specific countertherapies84. This approach has been exemplified by the Figure 3 | New therapeutic strategies for lung infections. Healthy airways contain a diverse mixture development of β‑lactamase inhibitors to Nature Reviews | Microbiology of commensals. Lung infections with pathogens (red) are typically treated with systemic broad- counter penicillin resistance. spectrum antibiotics (the bludgeon) that may deplete healthy commensals (green) and leave Extraordinarily, the experimental and antibiotic-resistant organisms in the bowel as well as the lung. New therapeutic approaches (magic bullets) include vaccination, inhaled narrow-spectrum antibiotics with primary effects on particular theoretical basis that underlies antibiotic pathogens and phage therapy. Commensal community restoration may replenish a depleted dosage regimens is limited and the results 85 microbiota and improve resistance to subsequent reinfections. contentious . High-dose antimicrobial chemotherapy has long been considered the best means of controlling drug resistance, by standardized methods, such as bead may actively support the presence of some but recent studies have shown that this beating (high-energy mixing with grinding commensals76, and the addition of mucins can encourage the appearance of resistant beads), with well-understood performances to culture media may provide essential strains at the same time as it releases for important organisms. An important nutrients for colony growth. them from ecological competition86. consideration is that DNA sequences will Whole-genome sequencing of bacterial The effectiveness of the common use of not distinguish between live and quiescent isolates has numerous applications to the combinations of antibiotics has also been microorganisms or extracellular DNA from problem of antibiotic resistance, including questioned, and it has been suggested that dead microorganisms. the development of novel antibiotics in many circumstances, such combinations Other than pathogens, most of the and diagnostic tests, the stewardship of will actually increase pathogen load87. pulmonary microbiota have not been isolated available antibiotics to prevent AMR and These findings, and the rising tide of or characterized in any detail29. Studies the elucidation of factors that generate and AMR, indicate a need for an experimental of the bowel microbiota have shown how perpetuate resistance77,78. Proof-of-principle re‑examination of antibiotic regimens in improved culture followed by whole-genome studies for the value of whole-genome common clinical states. To help this process, sequencing can permit sophisticated analyses sequencing have come from investigations elements important in antibiotic resistance of human microbial communities in health of M. tuberculosis79,80 and P. aeruginosa are increasingly accessible through genomic and in disease75. The diagnostic utility of infections81. This systematic approach is sequencing of pathogen isolates and also the sequencing of strain-specific loci or easily extended to other pathogens such from partial-genome metagenomic sequence of metagenomes will depend on future as S. pneumoniae and Haemophilus spp. data88, which allow a better anticipation systematic culture and genomic sequencing isolated from patients with lung infections. of the response to antibiotics. Microbial of all the lung microbiota. The perception A publicly accessible culture collection for sequencing can also be used to make that human indigenous bacteria are largely the lung microbiota will encourage the decisions about undesirable consequences of recalcitrant to culture is inaccurate75, and development of model systems to investigate antibiotic treatment of (as yet unidentified) simple measures may improve isolation interactions between hosts, pathogens and keystone commensal organisms as well as of organisms from the lung. Anaerobic commensals with straightforward replication about AMR emergence. culture, for example, has not routinely been of important results. used for pulmonary specimens, despite the Metagenomic shotgun sequencing Inhaled antibiotics. The administration of high numbers of anaerobic OTUs detected of bowel samples has delivered a mass of oral antibiotics can result in long-lasting by 16S RNA gene sequencing. Mucins information about microbial processes adverse alteration of gut microbial

PERSPECTIVES

60 communities , which suggests that Box 2 | Complex clinical uses of microbial community analyses benefits will follow from reducing or avoiding exposure of gut organisms to DNA and RNA sequence analysis of specimens from the lung may be of direct clinical benefit to the antimicrobials (FIG. 3). Inhalations have management of lung infections. However, there are obstacles yet to be overcome. been used medically for two millennia, and • Simple PCR assays can be used to quantify known pathogens. However, not enough is yet known a wide range of aerosolized medications about many of the unencapsulated Proteobacteria, including non-typeable Haemophilus influenzae and Neisseria spp. that dominate in chronic obstructive pulmonary disease and asthma. The are now in use or in development for 89 systematic isolation and sequencing of these organisms is desirable for the successful many diseases . The pulmonary delivery development of rapid diagnostic assays. of antibiotics (often as powders) for lung • In contrast to culture techniques, a simple presence or absence of a particular bacterial infections has shown success, with reduced operational taxonomic unit may not be sufficient to guide therapy. Absolute counts (by dosing required and reductions in side quantitative PCR or digital PCR) will be more helpful. Criteria are needed to differentiate 90 effects . Aerosolized (inhaled) antibiotics between asymptomatic carriage of common pathogens and the presence of disease. may be more effective than their systemic Concomitant measure of the total bacterial load can be used to estimate the degree of counterparts in treating pneumonia91. When pathogen dominance. used to treat persistent airway infection in • Sequence analysis also allows consideration of synergistic effects between members of a microbial cystic fibrosis, inhaled antibiotics appear community, avoiding the misleading assumption of ‘one pathogen, one disease’ (REF. 136). more effective than oral or parenteral • It will be desirable to develop measures of the health of a patient’s microbiome, identify optimal therapy92. It has been suggested that inhaled patterns of diversity and resilience and establish the presence or absence of keystone organisms antibiotics do not promote AMR, and a associated with health. These parameters will permit public health promotion of infection meta-analysis of studies in cystic fibrosis resistance in otherwise susceptible individuals and environments. did not observe significant development of • The effects of antibiotics will be assessed by their impact on whole microbial communities so that antibiotic-resistant organisms90,91. antimicrobial actions may be balanced between commensals and pathogens. A limited range of antibiotics is currently • Mixed infections are common, particularly in the presence of underlying lung disease. The marketed for inhalation and approved identification of all pathogens in an infection and a knowledge of pathogen synergies will for use in cystic fibrosis by the FDA and improve treatments. the European Medicines Agency93. These • Metagenomic transcriptomic analysis (RNA) of bacterial gene expression may identify targets include tobramycin, colistin and nebulized for therapy in particular infections. Metagenomic analysis (DNA) of lung samples is currently aztreonam93. Although not approved for severely hampered by high concentrations of human DNA. Single-cell transcriptomics of bacteria other diseases, injectable formulations purified from samples may be helpful. of gentamicin, tobramycin, amikacin, • Antibiotic sensitivities can currently be predicted by sequence analysis of bacterial isolates. ceftazidime and amphotericin are currently The application of this technique to complex microbial mixtures may give a measure of general community resistance. Pathogen-specific signatures may be discerned by unique sequences nebulized ‘off-label’ to manage non-cystic within the community. Alternatively, single-cell sequencing may identify the antimicrobial fibrosis bronchiectasis, drug-resistant resistance profile of individual organisms. nontuberculous mycobacterial infections, • Epidemiological surveys of the transmission and community carriage of common respiratory VAP and airway infections occurring pathogens, such as non-typeable H. influenzae, will underpin strategies to reduce transmission 93 after transplant . and prevent colonization. The advantages of inhaled therapy encourage a search, beyond the treatment of chronic infections in cystic fibrosis93, increasing AMR95. Phages can be modified and inflammatory bowel disease105. There is for novel antibiotics with chemical genetically to induce desirable biological good reason to consider similar possibilities characteristics that encourage retention in and therapeutic properties96. Importantly, in the lungs (FIG. 3). the lung (FIG. 3). A more selective antibiotic phage infection leads to only transient Disordered microbial communities specificity (narrow spectrum) towards changes in the microbial communities that have a loss of commensal diversity major pathogens such as H. influenzae is that harbour targeted pathogens97. are a feature of asthma10,106, which may also desirable (FIG. 2). Patients with COPD Despite these advantages, phage therapy be a reasonable pulmonary target for who suffer frequent lower respiratory tract is not proving easy to develop within the interventional trials. Community restoration infections, which contribute enormously to stringent regulatory requirements for new might also be of benefit after long-term AMR in the community8, would be an ideal treatments98. Nevertheless, it has shown antibiotic use in children with cystic group for the comparison of inhaled and encouraging results for the treatment of fibrosis or primary ciliary dyskinesia or in systemic antibiotic therapies. P. aeruginosa infections in patients with adults with COPD. However, in contrast to cystic fibrosis99,100 and has major potential faecal microbiota transplantation, healthy Phage therapy. In 1917, bacteriophages for targeted treatment of more common individuals do not spontaneously provide were discovered to cause epizootic pathogens, such as H. influenzae (FIG. 3). airway microbiota in a form suitable for infections of bacteria and were used for transplantation; in addition, little is known antibacterial therapy and prophylaxis94. Restoring bacterial communities. about the characteristics of lower-airway However, a poor understanding of their Manipulation of the microbiota confers commensals, and pulmonary pathogens biology and the advent of antibiotics colonization resistance to intestinal commonly colonize the airways without curtailed research after 1945 (REF. 94). The infections101–103, and faecal microbiota causing disease. early literature suggests that phage therapy transplantation is markedly effective in A structured approach to identifying is very effective94, and there has been a treating the severe illnesses associated with bowel protective organisms has been modern renewal of interest in response to persistent Clostridium difficile infection104 exemplified for C. difficile infections102.

PERSPECTIVES

Similar strategies for the airway microbiota of Neisseria spp. and Moraxella spp. as well of and minimizing AMR. A systematic may identify the minimum number of as of Haemophilus spp. and Streptococcus understanding of the healthy pulmonary commensal organisms that might provide spp.10,28,30,118,119. In addition, S. pneumoniae microbiota is a necessary platform from protection against colonization with has a high reported carriage rate (45%) which to build new therapies that may pathogenic Proteobacteria and Streptococcus in children with asthma108. Should these include novel inhaled antibiotics, phages, spp. or could be used to replace a microbial bacteria be shown definitively to be causing vaccination and techniques to restore community with particularly high levels of airway inflammation, the success of healthy microbial communities. AMR. The organisms to be used may vary for vaccination against such organisms leads William O. C. M. Cookson, Michael J. Cox and different diseases and different individuals. to the intriguing possibility of eventual Miriam F. Moffatt are at the Asmarley Centre for The dose of bacteria as well as the species to vaccination to protect against childhood Genomic Medicine, National Heart and Lung Institute, be introduced is an important consideration. asthma. Inherent in a vaccination strategy Imperial College London, Dovehouse Street, London SW3 6LY, UK. In an analogy of restoring a garden, therapy would be the recognition that many of the with antimicrobials to remove pathogens organisms associated with asthma and Correspondence to W.O.C.M.C. (weeds) before treatment may be helpful. COPD are likely to be unencapsulated [email protected] As with other novel therapies, an ethical strains. This further emphasizes the need for doi:10.1038/nrmicro.2017.122 framework will need to be established a concerted attempt to isolate and sequence Published online 24 Oct 2017 around early-stage clinical trials to examine the lower-airway microbiota. 1. Adams, W. C. California Air Resources Board contract no. A033‑205: Measurement of breathing rate and safety before therapies can be explored. Vaccines against P. aeruginosa are not volume in routinely performed daily activities. effective in preventing infections in patients California Air Resources Board https://www.arb.ca.gov/ 120 research/apr/past/a033-205.pdf (1993). Vaccination. The potential for vaccination with cystic fibrosis , perhaps suggesting 2. Weibel, E. R. & Gomez, D. M. Architecture of the to have an impact on common lung that vaccinations have less value in patients human lung. Use of quantitative methods establishes fundamental relations between size and infections (FIG. 3) is exemplified by invasive with infections that flourish in luminal number of lung structures. Science 137, 577–585 infections with S. pneumoniae107. Although secretions, which may be at a distance from (1962). 3. Hasleton, P. S. The internal surface area of the adult vaccination in the first years of life does not mucosal defence mechanisms. human lung. J. Anat. 112, 391–400 (1972). confer long-lasting immunity108, prevention A better understanding of the population 4. Bowers, R. M. et al. Sources of bacteria in outdoor air across cities in the midwestern United States. of invasive disease by S. pneumoniae distributions of potentially pathogenic Appl. Environ. Microbiol. 77, 6350–6356 (2011). with AMR is achievable109. The rate of Proteobacteria as well as H. influenzae and 5. Helander, H. F. & Fandriks, L. Surface area of the digestive tract — revisited. Scand. J. Gastroenterol. antibiotic-resistant, invasive pneumococcal S. pneumoniae in the healthy population, 49, 681–689 (2014). infections has decreased in young children as well as in patients with asthma and 6. Guest, J. F. & Morris, A. Community-acquired pneumonia: the annual cost to the National Health and older adults with the widespread use patients with COPD, is important to help Service in the UK. Eur. Respir. J. 10, 1530–1534 of pneumococcal conjugate vaccines110, develop rational vaccination strategies. (1997). 7. Adriaenssens, N. et al. European Surveillance of and vaccination is effective in preventing Effective vaccination is likely to be followed Antimicrobial Consumption (ESAC): outpatient pneumonia in children and adults by the evolution of population serotypes, antibiotic use in Europe. J. Antimicrob. Chemother. 66 111 (Suppl 6.), vi3–vi12 (2011). following splenectomy . which would impart a need for continuous 8. Murphy, T. F. Vaccines for nontypeable Haemophilus Respiratory infections caused by NTHI vaccine development121. However, the global influenzae: the future is now. Clin. Vaccine Immunol. 38 22, 459–466 (2015). result in enormous global morbidity . numbers of children and adults at risk 9. Dickson, R. P., Erb-Downward, J. R. & Huffnagle, G. B. Effective H. influenzae vaccines have suggest that the effort would be worthwhile. The role of the bacterial microbiome in lung disease. Expert Rev. Respir. Med. 7, 245–257 been directed against capsular antigens (2013). and are given routinely in childhood for Conclusions 10. Hilty, M. et al. Disordered microbial communities in asthmatic airways. PLoS ONE 5, e8578 (2010). meningitis and otitis media in most Western Lung infections have enormous effects on 11. Charlson, E. S. et al. Topographical continuity of societies. However, the unencapsulated human health, but current treatments are bacterial populations in the healthy human respiratory tract. Am. J. Respir. Crit. Care Med. 184, 957–963 NTHI strains that commonly cause lung imperfect and fuel a global engine for AMR. (2011). diseases are not influenced by current Genome sequencing is fundamentally 12. Brook, I. Bacterial interference. Crit. Rev. Microbiol. 112–114 25, 155–172 (1999). vaccines . A vaccine against NTHI changing our understanding of the airway 13. Reid, G., Howard, J. & Gan, B. S. Can bacterial would prevent disease and reduce antibiotic microbiome in health and disease, not interference prevent infection? Trends Microbiol. 9, 8 424–428 (2001). use , most clearly in the context of COPD least by revealing the profound pressures 14. Falagas, M. E., Rafailidis, P. I. & Makris, G. C. Bacterial exacerbations115,116 but also in CAP, cystic that drive the evolution and spread of interference for the prevention and treatment of infections. Int. J. Antimicrob. Agents 31, 518–522 fibrosis and chronic bronchitis. An ideal pathogenic microorganisms. Genomic (2008). vaccine antigen should be conserved studies of opportunistic pathogens 15. Pamer, E. G. Resurrecting the intestinal microbiota to combat antibiotic-resistant pathogens. Science 352, among strains, have abundant epitopes on that accompany cystic fibrosis show 535–538 (2016). the bacterial surface and induce protective how easily hyper-resistant and virulent 16. Ferkol, T. & Schraufnagel, D. The global burden of 8 respiratory disease. Ann. Am. Thorac Soc. 11, immune responses . A conjugate vaccine clones can disseminate from clinic to 404–406 (2014). against protein D (a highly conserved clinic across the world. These studies 17. Walker, C. L. et al. Global burden of childhood pneumonia and diarrhoea. Lancet 381, 1405–1416 and stable protein in the outer membrane provide an exemplar for investigation of (2013). of encapsulated and unencapsulated the epidemiological suspicion that the 18. Park, D. R. The microbiology of ventilator-associated pneumonia. Respir. Care 50, 742–765 (2005). H. influenzae strains), although only partially evolution and transmission of apparently 19. Goss, C. H. & Burns, J. L. Exacerbations in cystic effective, has demonstrated the feasibility of mundane pathogens may be common. fibrosis. 1: Epidemiology and pathogenesis. Thorax 8,114,117 62, 360–367 (2007). vaccination for NTHI . Sequence-based techniques are ready 20. Donaldson, G. C., Seemungal, T. A., Bhowmik, A. & Characterization of the asthmatic airway to transform the microbial diagnosis Wedzicha, J. A. Relationship between exacerbation frequency and lung function decline in chronic microbiome is at an early stage, but there of common lung infections and are the obstructive pulmonary disease. Thorax 57, 847–852 is a consensus that there is an overgrowth first priority for improving management (2002).

PERSPECTIVES

21. Beasley, R. & The International Study of Asthma and 47. Scott, J. A. et al. Aetiology, outcome, and risk factors 73. Hanage, W. P. et al. Using multilocus sequence data Allergies in Childhood (ISAAC) Steering Committee. for mortality among adults with acute pneumonia in to define the pneumococcus. J. Bacteriol. 187, Worldwide variation in prevalence of symptoms of Kenya. Lancet 355, 1225–1230 (2000). 6223–6230 (2005). asthma, allergic rhinoconjunctivitis, and atopic 48. McCullers, J. A. Insights into the interaction between 74. Vebø, H. C., Karlsson, M. K., Avershina, E., Finnby, L. eczema: ISAAC. Lancet 351, 1225–1232 (1998). influenza virus and pneumococcus. Clin. Microbiol. & Rudi, K. Bead-beating artefacts in the 22. Cookson, W. The immunogenetics of asthma and Rev. 19, 571–582 (2006). Bacteroidetes to Firmicutes ratio of the human stool eczema: a new focus on the epithelium. Nat. Rev. 49. Wilkinson, T. M. et al. Effect of interactions between metagenome. J. Microbiol. Methods 129, 78–80 Immunol. 4, 978–988 (2004). lower airway bacterial and rhinoviral infection in (2016). 23. Johnston, S. et al. Community study of role of viral exacerbations of COPD. Chest 129, 317–324 75. Browne, H. P. et al. Culturing of ‘unculturable’ human infections in exacerbations of asthma in 9–11 year old (2006). microbiota reveals novel taxa and extensive children. BMJ 310, 1225–1229 (1995). 50. Wheat, L. J., Goldman, M. & Sarosi, G. State‑of‑the-art sporulation. Nature 533, 543–546 (2016). 24. Bisgaard, H. et al. Childhood asthma after bacterial review of pulmonary fungal infections. Semin. Respir. 76. Linden, S. K., Sutton, P., Karlsson, N. G., Korolik, V. colonization of the airway in neonates. N. Engl. J. Med. Infect. 17, 158–181 (2002). & McGuckin, M. A. Mucins in the mucosal barrier to 357, 1487–1495 (2007). 51. Pihet, M. et al. Occurrence and relevance of infection. Mucosal Immunol. 1, 183–197 (2008). 25. Green, B. J. et al. Potentially pathogenic airway filamentous fungi in respiratory secretions of patients 77. Koser, C. U., Ellington, M. J. & Peacock, S. J. bacteria and neutrophilic inflammation in treatment with cystic fibrosis — a review. Med. Mycol. 47, Whole-genome sequencing to control antimicrobial resistant severe asthma. PLoS ONE 9, e100645 387–397 (2009). resistance. Trends Genet. 30, 401–407 (2014). (2014). 52. Kastman, E. K. et al. Biotic interactions shape the 78. Punina, N. V., Makridakis, N. M., Remnev, M. A. 26. Huang, Y. J. et al. Airway microbiota and bronchial ecological distributions of Staphylococcus species. & Topunov, A. F. Whole-genome sequencing targets hyperresponsiveness in patients with suboptimally mBio 7, e01157‑16 (2016). drug-resistant bacterial infections. Hum. Genom. 9, 19 controlled asthma. J. Allergy Clin. Immunol. 127, 53. Enoch, D. A., Ludlam, H. A. & Brown, N. M. Invasive (2015). 372–381.e3 (2011). fungal infections: a review of epidemiology and 79. Walker, T. M. et al. Whole-genome sequencing for 27. Goleva, E. et al. The effects of airway microbiome on management options. J. Med. Microbiol. 55, prediction of Mycobacterium tuberculosis drug corticosteroid responsiveness in asthma. Am. 809–818 (2006). susceptibility and resistance: a retrospective cohort J. Respir. Crit. Care Med. 188, 1193–1201 (2013). 54. Hesse, W. Walther and Angelina Hesse – early study. Lancet Infect. Dis. 15, 1193–1202 (2015). 28. Huang, Y. J. & Boushey, H. A. The microbiome in contributors to bacteriology. ASM News 58, 425–428 80. Bradley, P. et al. Rapid antibiotic-resistance asthma. J. Allergy Clin. Immunol. 135, 25–30 (2015). (1992). predictions from genome sequence data for 29. Man, W. H., de Steenhuijsen Piters, W. A. & 55. Croxatto, A., Prod’hom, G. & Greub, G. Applications of Staphylococcus aureus and Mycobacterium Bogaert, D. The microbiota of the respiratory tract: MALDI-TOF mass spectrometry in clinical diagnostic tuberculosis. Nat. Commun. 6, 10063 (2015). gatekeeper to respiratory health. Nat. Rev. Microbiol. microbiology. FEMS Microbiol. Rev. 36, 380–407 81. Winstanley, C. et al. Newly introduced genomic 15, 259–270 (2017). (2012). prophage islands are critical determinants of in vivo 30. Cardenas, P. A. et al. Upper airways microbiota in 56. Waters, B. & Muscedere, J. A. 2015 update on competitiveness in the Liverpool Epidemic Strain of antibiotic-naive wheezing and healthy infants from the ventilator-associated pneumonia: new insights on its Pseudomonas aeruginosa. Genome Res. 19, 12–23 tropics of rural Ecuador. PLoS ONE 7, e46803 (2012). prevention, diagnosis, and treatment. Curr. Infect. Dis. (2009). 31. Stearns, J. C. et al. Culture and molecular-based Rep. 17, 496 (2015). 82. Feehery, G. R. et al. A method for selectively enriching profiles show shifts in bacterial communities of the 57. Joos, L. et al. Pulmonary infections diagnosed by BAL: microbial DNA from contaminating vertebrate host upper respiratory tract that occur with age. ISME J. 9, a 12‑year experience in 1066 immunocompromised DNA. PLoS ONE 8, e76096 (2013). 1246–1259 (2015). patients. Respir. Med. 101, 93–97 (2007). 83. Fischer, N. et al. Evaluation of unbiased next-generation 32. Molyneaux, P. L. et al. The role of bacteria in the 58. Boyton, R. J. Infectious lung complications in patients sequencing of RNA (RNA-seq) as a diagnostic method pathogenesis and progression of idiopathic pulmonary with HIV/AIDS. Curr. Opin. Pulm. Med. 11, 203–207 in influenza virus-positive respiratory samples. fibrosis. Am. J. Respir. Crit. Care Med. 190, 906–913 (2005). J. Clin. Microbiol. 53, 2238–2250 (2015). (2014). 59. Bush, K. et al. Tackling antibiotic resistance. Nat. Rev. 84. Coley, H. M. Mechanisms and strategies to overcome 33. Twigg, H. L. et al. Use of bronchoalveolar lavage to Microbiol. 9, 894–896 (2011). chemotherapy resistance in metastatic breast cancer. assess the respiratory microbiome: signal in the noise. 60. Zaura, E. et al. Same exposure but two radically Cancer Treat. Rev. 34, 378–390 (2008). Lancet Respir. Med. 1, 354–356 (2013). different responses to antibiotics: resilience of the 85. Kupferschmidt, K. Resistance fighters. Science 352, 34. Dickson, R. P. et al. Spatial variation in the healthy salivary microbiome versus long-term microbial shifts 758–761 (2016). human lung microbiome and the adapted island in feces. mBio 6, e01693‑15 (2015). 86. Day, T. & Read, A. F. Does high-dose antimicrobial model of lung biogeography. Ann. Am. Thorac Soc. 12, 61. Lim, W. S. et al. BTS guidelines for the management chemotherapy prevent the evolution of resistance? 821–830 (2015). of community acquired pneumonia in adults: update PLoS Comput. Biol. 12, e1004689 (2016). 35. Salter, S. J. et al. Reagent and laboratory 2009. Thorax 64 (Suppl. 3), iii1–iii55 (2009). 87. Pena-Miller, R. et al. When the most potent contamination can critically impact sequence-based 62. Freter, R. In vivo and in vitro antagonism of intestinal combination of antibiotics selects for the greatest microbiome analyses. BMC Biol. 12, 87 (2014). bacteria against Shigella flexneri: II. The inhibitory bacterial load: the smile-frown transition. PLoS Biol. 36. Wilson, L. G. Commentary: Medicine, population, and mechanism. J. Infect. Dis. 110, 38–46 (1962). 11, e1001540 (2013). tuberculosis. Int. J. Epidemiol. 34, 521–524 (2005). 63. Mogayzel, P. J. Jr. et al. Cystic Fibrosis Foundation 88. McArthur, A. G. et al. The comprehensive antibiotic 37. Kennedy, W. A. et al. Incidence of bacterial meningitis pulmonary guideline. Pharmacologic approaches to resistance database. Antimicrob. Agents Chemother. in Asia using enhanced CSF testing: polymerase chain prevention and eradication of initial Pseudomonas 57, 3348–3357 (2013). reaction, latex agglutination and culture. aeruginosa infection. Ann. Am. Thorac Soc. 11, 89. Rubin, B. K. Pediatric aerosol therapy: new devices Epidemiol. Infect. 135, 1217–1226 (2007). 1640–1650 (2014). and new drugs. Respir. Care 56, 1411–1423 38. Sethi, S. & Murphy, T. F. Infection in the pathogenesis 64. Smyth, A. R. & Walters, S. Prophylactic anti- (2011). and course of chronic obstructive pulmonary disease. staphylococcal antibiotics for cystic fibrosis. Cochrane 90. Traini, D. & Young, P. M. Delivery of antibiotics to the N. Engl. J. Med. 359, 2355–2365 (2008). Database Syst. Rev. 11, CD001912 (2014). respiratory tract: an update. Expert Opin. Drug Deliv. 39. Keller, L. E., Robinson, D. A. & McDaniel, L. S. 65. Langton Hewer, S. C. & Smyth, A. R. Antibiotic 6, 897–905 (2009). Nonencapsulated Streptococcus pneumoniae: strategies for eradicating Pseudomonas aeruginosa in 91. Palmer, L. B. & Smaldone, G. C. Reduction of bacterial emergence and pathogenesis. mBio 7, e01792 (2016). people with cystic fibrosis. Cochrane Database Syst. resistance with inhaled antibiotics in the intensive care 40. Siegel, S. J. & Weiser, J. N. Mechanisms of bacterial Rev. 11, CD004197 (2014). unit. Am. J. Respir. Crit. Care Med. 189, 1225–1233 colonization of the respiratory tract. Annu. Rev. 66. Melnyk, A. H., Wong, A. & Kassen, R. The fitness costs (2014). Microbiol. 69, 425–444 (2015). of antibiotic resistance mutations. Evol. Appl. 8, 92. Ryan, G., Singh, M. & Dwan, K. Inhaled antibiotics for 41. American Thoracic Society & Infectious Diseases 273–283 (2015). long-term therapy in cystic fibrosis. Cochrane Society of America. Guidelines for the management 67. Clarridge, J. E. III. Impact of 16S rRNA gene sequence Database Syst. Rev. 3, CD001021 (2011). of adults with hospital-acquired, ventilator-associated, analysis for identification of bacteria on clinical 93. Quon, B. S., Goss, C. H. & Ramsey, B. W. Inhaled and healthcare-associated pneumonia. Am. J. Respir. microbiology and infectious diseases. Clin. Microbiol. antibiotics for lower airway infections. Ann. Am. Crit. Care Med. 171, 388–416 (2005). Rev. 17, 840–862 (2004). Thorac Soc. 11, 425–434 (2014). 42. Holcombe, L. J., O’Gara, F. & Morrissey, J. P. 68. Woo, P. C., Lau, S. K., Teng, J. L., Tse, H. & Yuen, K. Y. 94. Summers, W. C. Bacteriophage therapy. Annu. Rev. Implications of interspecies signaling for virulence of Then and now: use of 16S rDNA gene sequencing for Microbiol. 55, 437–451 (2001). bacterial and fungal pathogens. Future Microbiol. 6, bacterial identification and discovery of novel bacteria 95. Young, R. & Gill, J. J. Phage therapy redux — what is 799–817 (2011). in clinical microbiology laboratories. Clin. Microbiol. to be done? Science 350, 1163–1164 (2015). 43. Lessler, J. et al. Incubation periods of acute Infect. 14, 908–934 (2008). 96. Nobrega, F. L., Costa, A. R., Kluskens, L. D. & respiratory viral infections: a systematic review. 69. Janda, J. M. & Abbott, S. L. 16S rRNA gene Azeredo, J. Revisiting phage therapy: new applications Lancet Infect. Dis. 9, 291–300 (2009). sequencing for bacterial identification in the diagnostic for old resources. Trends Microbiol. 23, 185–191 44. Kuhnert, P. & Christensen, H. Pasteurellaceae: laboratory: pluses, perils, and pitfalls. J. Clin. Microbiol. (2015). Biology, Genomics and Molecular Aspects (Caister 45, 2761–2764 (2007). 97. Reyes, A., Wu, M., McNulty, N. P., Rohwer, F. L. & Academic Press, 2008). 70. Koljalg, U. et al. Towards a unified paradigm for Gordon, J. I. Gnotobiotic mouse model of phage- 45. Morens, D. M., Taubenberger, J. K. & Fauci, A. S. sequence-based identification of fungi. Mol. Ecol. 22, bacterial host dynamics in the human gut. Predominant role of bacterial pneumonia as a cause 5271–5277 (2013). Proc. Natl Acad. Sci. USA 110, 20236–20241 of death in pandemic influenza: implications for 71. Bosshard, P. P. Incubation of fungal cultures: how long (2013). pandemic influenza preparedness. J. Infect. Dis. 198, is long enough? Mycoses 54, e539–e545 (2011). 98. Servick, K. Beleaguered phage therapy trial presses 962–970 (2008). 72. Balajee, S. A. et al. Sequence-based identification of on. Science 352, 1506 (2016). 46. Bosch, A. A., Biesbroek, G., Trzcinski, K., Aspergillus, Fusarium, and Mucorales species in the 99. Alemayehu, D. et al. Bacteriophages phiMR299‑2 and Sanders, E. A. & Bogaert, D. Viral and bacterial clinical mycology laboratory: where are we and where phiNH‑4 can eliminate Pseudomonas aeruginosa in interactions in the upper respiratory tract. should we go from here? J. Clin. Microbiol. 47, the murine lung and on cystic fibrosis lung airway cells. PLoS Pathog. 9, e1003057 (2013). 877–884 (2009). mBio 3, e00029‑12 (2012).

PERSPECTIVES

100. Debarbieux, L. et al. Bacteriophages can treat and 113. Lerman, S. J., Kucera, J. C. & Brunken, J. M. 128. Cheng, K. et al. Spread of β-lactam-resistant prevent Pseudomonas aeruginosa lung infections. Nasopharyngeal carriage of antibiotic-resistant Pseudomonas aeruginosa in a cystic fibrosis clinic. J. Infect. Dis. 201, 1096–1104 (2010). Haemophilus influenzae in healthy children. Pediatrics Lancet 348, 639–642 (1996). 101. He, X., McLean, J. S., Guo, L., Lux, R. & Shi, W. 64, 287–291 (1979). 129. Bryant, J. M. et al. Whole-genome sequencing to The social structure of microbial community involved 114. Murphy, T. F. et al. Nontypeable Haemophilus identify transmission of Mycobacterium abscessus in colonization resistance. ISME J. 8, 564–574 influenzae as a pathogen in children. Pediatr. Infect. between patients with cystic fibrosis: a retrospective (2014). Dis. J. 28, 43–48 (2009). cohort study. Lancet 381, 1551–1560 (2013). 102. Lawley, T. D. et al. Targeted restoration of the 115. Dagan, R. & Klugman, K. P. Impact of conjugate 130. Bryant, J. M. et al. Emergence and spread of a human- intestinal microbiota with a simple, defined pneumococcal vaccines on antibiotic resistance. transmissible multidrug-resistant nontuberculous bacteriotherapy resolves relapsing Clostridium Lancet Infect. Dis. 8, 785–795 (2008). mycobacterium. Science 354, 751–757 (2016). difficile disease in mice. PLoS Pathog. 8, e1002995 116. Wilby, K. J. & Werry, D. A review of the effect of 131. Saiman, L. et al. Infection prevention and control (2012). immunization programs on antimicrobial utilization. guideline for cystic fibrosis: 2013 update. Infect. 103. Adamu, B. O. & Lawley, T. D. Bacteriotherapy for the Vaccine 30, 6509–6514 (2012). Control Hosp. Epidemiol. 35 (Suppl. 1), S1–S67 treatment of intestinal dysbiosis caused by Clostridium 117. Leach, A. J. et al. Reduced middle ear infection with (2014). difficile infection. Curr. Opin. Microbiol. 16, 596–601 non-typeable Haemophilus influenzae, but not 132. Chotirmall, S. H. & McElvaney, N. G. Fungi in the (2013). Streptococcus pneumoniae, after transition to cystic fibrosis lung: bystanders or pathogens? 104. Gough, E., Shaikh, H. & Manges, A. R. Systematic 10‑valent pneumococcal non-typeable H. influenzae Int. J. Biochem. Cell Biol. 52, 161–173 (2014). review of intestinal microbiota transplantation protein D conjugate vaccine. BMC Pediatr. 15, 162 133. Parize, P. et al. Impact of Scedosporium apiospermum (fecal bacteriotherapy) for recurrent Clostridium (2015). complex seroprevalence in patients with cystic fibrosis. difficile infection. Clin. Infect. Dis. 53, 994–1002 118. Simpson, J. L. et al. Airway dysbiosis: Haemophilus J. Cyst. Fibros. 13, 667–673 (2014). (2011). influenzae and Tropheryma in poorly controlled 134. Larkin, E. et al. The emerging pathogen Candida 105. Colman, R. J. & Rubin, D. T. Fecal microbiota asthma. Eur. Respir. J. 47, 792–800 (2015). auris: growth phenotype, virulence factors, activity of transplantation as therapy for inflammatory bowel 119. Zhang, Q. et al. Airway microbiota in severe asthma antifungals, and effect of SCY‑078, a novel glucan disease: a systematic review and meta-analysis. and relationship to asthma severity and phenotypes. synthesis inhibitor, on growth morphology and biofilm J. Crohns Colitis 8, 1569–1581 (2014). PLoS ONE 11, e0152724 (2016). formation. Antimicrob. Agents Chemother. 61, 106. Dickson, R. P., Martinez, F. J. & Huffnagle, G. B. 120. Johansen, H. K. & Gotzsche, P. C. Vaccines for e02396‑16 (2017). The role of the microbiome in exacerbations of preventing infection with Pseudomonas aeruginosa 135. de Wit, R. & Bouvier, T. ‘Everything is everywhere, but, chronic lung diseases. Lancet 384, 691–702 in cystic fibrosis. Cochrane Database Syst. Rev. 8, the environment selects’; what did Baas Becking and (2014). CD001399 (2015). Beijerinck really say? Environ. Microbiol. 8, 755–758 107. Conklin, L. et al. Systematic review of the effect of 121. Maiden, M. C. Population genomics: diversity and (2006). pneumococcal conjugate vaccine dosing schedules on virulence in the Neisseria. Curr. Opin. Microbiol. 11, 136. Byrd, A. L. & Segre, J. A. Infectious disease. Adapting vaccine-type invasive pneumococcal disease among 467–471 (2008). Koch’s postulates. Science 351, 224–226 (2016). young children. Pediatr. Infect. Dis. J. 33 (Suppl. 2), 122. Rao, D., Webb, J. S. & Kjelleberg, S. Microbial S109–S118 (2014). colonization and competition on the marine alga Ulva Acknowledgements 108. Esposito, S. et al. Streptococcus pneumoniae australis. Appl. Environ. Microbiol. 72, 5547–5555 W.O.C.M.C., M.J.C. and M.F.M. receive funding from the colonisation in children and adolescents with asthma: (2006). Wellcome Trust and from the Asmarley Trust. W.O.C.M.C. and impact of the heptavalent pneumococcal conjugate 123. Ley, R. E., Peterson, D. A. & Gordon, J. I. Ecological M.F.M. are joint Wellcome Senior Investigators. W.O.C.M.C. vaccine and evaluation of potential effect of thirteen- and evolutionary forces shaping microbial diversity in also received funding for microbiome studies with a Senior valent pneumococcal conjugate vaccine. BMC Infect. the human intestine. Cell 124, 837–848 (2006). Investigator award from the National Institute for Health Dis. 16, 12 (2016). 124. Comas, I. et al. Out‑of‑Africa migration and Neolithic Research, United Kingdom. 109. Hampton, L. M. et al. Prevention of antibiotic- coexpansion of Mycobacterium tuberculosis with nonsusceptible Streptococcus pneumoniae with modern humans. Nat. Genet. 45, 1176–1182 Author contributions conjugate vaccines. J. Infect. Dis. 205, 401–411 (2013). W.O.C.M.C. carried out research, wrote the article and con‑ (2012). 125. Walker, T. M. et al. Whole-genome sequencing to tributed to discussions, review and editing of the manuscript. 110. Kyaw, M. H. et al. Effect of introduction of the delineate Mycobacterium tuberculosis outbreaks: M.J.C. and M.F.M. carried out research and contributed to pneumococcal conjugate vaccine on drug-resistant a retrospective observational study. Lancet Infect. Dis. discussions, review and editing of the manuscript. The Streptococcus pneumoniae. N. Engl. J. Med. 354, 13, 137–146 (2013). authors gratefully acknowledge the considerable input of two 1455–1463 (2006). 126. Snitkin, E. S. et al. Tracking a hospital outbreak of anonymous reviewers. 111. Ammann, A. J. et al. Polyvalent pneumococcal- carbapenem-resistant Klebsiella pneumoniae with polysaccharide immunization of patients with sickle- whole-genome sequencing. Sci. Transl Med. 4, Competing interests statement cell anemia and patients with splenectomy. N. Engl. 148ra116 (2012). The authors declare no competing interests. J. Med. 297, 897–900 (1977). 127. LiPuma, J. J., Dasen, S. E., Nielson, D. W., Stern, R. C. 112. Bou, R. et al. Prevalence of Haemophilus influenzae & Stull, T. L. Person-to‑person transmission of Publisher’s note pharyngeal carriers in the school population of Pseudomonas cepacia between patients with cystic Springer Nature remains neutral with regard to jurisdictional Catalonia. Eur. J. Epidemiol. 16, 521–526 (2000). fibrosis. Lancet 336, 1094–1096 (1990). claims in published maps and institutional affiliations.

ERRATUM Inserting proteins into the bacterial cytoplasmic membrane using the Sec and YidC translocases Kun Xie and Ross E. Dalbey Nature Reviews Microbiology 6, 234–244 (2008) This article was published with an incorrect DOI. The correct DOI is 10.1038/nrmicro3595. This has now been corrected in the online version. We apologize to the authors and to readers for any confusion caused.

CORRIGENDUM Archaea and the origin of eukaryotes Laura Eme, Anja Spang, Jonathan Lombard, Courtney W. Stairs and Thijs J. G. Ettema Nature Reviews Microbiology 15, 711–723 (2017) On pages 714–715 of this article, in the first paragraph of the section What do we currently know about LECA?, the sentence “Phylogenomic and comparative genomic analyses have led to the hypothesis that LECA, estimated to have lived ~1–1.9 million years ago54, already was a fully fledged eukaryote and possessed a large number of features that are uniquely found in modern eukaryotes55,56.” should have read “Phylogenomic and comparative genomic analyses have led to the hypothesis that LECA, estimated to have lived ~1–1.9 billion years ago54, already was a fully fledged eukaryote and possessed a large number of features that are uniquely found in modern eukaryotes55,56.” This has been corrected in the online version of the article. The authors apologize to the readers for any misunderstanding caused.