The Lung Microbiome in Lung Transplantation

Abstract

Culture-independent study of the lower respiratory tract after lung transplantation has enabled an understanding of the microbiome – that is, the collection of bacteria, fungi, and viruses, and their respective gene complement – in this niche. The lung has unique features as a microbial environment, with balanced entry from the upper respiratory tract, clearance, and local replication. There are many pressures impacting the microbiome after transplantation, including donor allograft factors, recipient host factors such as underlying disease and ongoing exposure to the microbe-rich upper respiratory tract, and transplantation-related immunosuppression, antimicrobials, and post-surgical changes. To date, we understand that the lung microbiome after transplant is dysbiotic; that is, it has higher biomass and altered composition compared to a healthy lung. Emerging data suggest that specific microbiome features may be linked to host responses, both immune and non-immune, and clinical outcomes such as chronic lung allograft dysfunction (CLAD), but many questions remain. The goal of this review is to put into context our burgeoning understanding of the lung microbiome in the post-lung transplant patient, the interactions between microbiome and host, the role the microbiome may play in post-transplant complications, and critical outstanding research questions.

Introduction

Lung transplantation is the only therapy for many end-stage lung diseases, but lung allografts have the shortest survival of all solid-organ transplants¹. Early challenges include primary graft dysfunction (PGD), acute cellular and antibody-mediated rejection (ACR, AMR), infections, and anastomotic complications, while the principal barrier to long-term allograft survival is chronic lung allograft dysfunction (CLAD), which leads to relentless graft failure^1,2. Several lung infections increase the risk of later CLAD and azithromycin slows progression^3,4 suggesting a role for host-microbiota interactions in CLAD pathogenesis.

The microbiome, defined as the collection of microbes and their genetic content in a distinct ecological location^5,6, considers entire communities and not just individual organisms, employs statistical analysis of community characteristics, and enables identification of microbiome “types.” While the microbe-rich gastrointestinal tract was the initial focus of such work⁷, emerging studies of the lung microbiome in lung transplantation have demonstrated marked abnormalities compared to health, termed “dysbiosis”^8–15. In the post-transplantation setting, however, the most relevant comparison is not to individuals without lung disease, but between post-transplant subjects with healthy allografts and those with dysfunction. This review addresses the unique post-transplantation lung microbiome, links to transplant outcomes, and host-microbiota cross-talk. In addition to the bacterial microbiome, which is the most-studied microbiome component, we address the less-studied communities of viruses and fungi.

Culture-Independent Sequencing

Microbial culture has been indispensable for understanding microbial pathogenesis but captures only a subset of the microbial landscape. Advances in sequencing technology (“next-generation” or “deep sequencing”) and computational methods have enabled a comprehensive, unbiased culture-independent approach¹⁶. Sequence-based approaches quantify the relative abundances of specific microbes in a sample and readily identify both potential pathogens and the vast commensal landscape. The most widely-used approach targets the bacterial 16S rRNA gene^17–20. Here polymerase chain reaction (PCR) amplifies bacterial DNA regions that, when sequenced, infer bacterial taxonomy^16,20. Targeted 16S rRNA gene analysis also captures archaea, which are scant in the lung and generally under-studied. Fungi (the “mycobiome”) are studied using analogous targeting of shared fungal regions, such as the 18S rRNA gene or internal transcribed spacer (ITS) sequence in the ribosomal gene locus.

In contrast, viruses have no such universally shared sequences so a complementary approach used is shotgun metagenomic sequencing, which sequences all the DNA (or RNA depending on sample preparation)^21,22. Thus, shotgun sequencing can profile viruses (the “virome”) as well as fungi and bacteria. Another advantage to shotgun sequencing is that it also captures microbial gene content, which assesses the genetic potential for microbial functions (e.g., metabolism, virulence factors, antimicrobial resistance). However, large amounts of host relative to microbial DNA in lung samples increases the cost and complexity of analysis²³; although with recent advances, shotgun metagenomics is enjoying broader use in lung microbiome research^23–26. Table 1 lists selected key elements in microbiome investigation relevant to this review.

Table 1.

Terms relevant to lung microbiome studies.

Microbiome – the totality of microbes and their gene content in an environmental niche Mycobiome – the fungal microbiome Virome – the viral microbiome, including eukaryotic (host) viruses and bacteriophages 16S rRNA – gene commonly used for targeted sequence-based identification of bacteria and archaea 18S rRNA – gene commonly used for targeted sequence-based identification of fungi Internal Transcribed Spacer Sequence (ITS) – region in the rRNA locus commonly used for targeted sequence-based identification of fungi Taxon (or taxa, multiple) – A grouping of microbes at any level such as family, genus, or species Operational taxonomic unit (OTU) – closely-related sequences grouped together to represent a unit of bacteria (“taxon”), for example a cluster of bacterial 16S rRNA gene sequence sharing 97% sequence identity Richness – number of different species present in a community Evenness – evenness of distribution among species in a community Alpha diversity – characteristic of a community that incorporates both richness and evenness Beta diversity – compositional differences between two or more microbial communities Dysbiosis – deviation from the normal, optimal, or healthy microbiome in a particular niche; typically, this is operationally defined as increased absolute microbial abundance, altered alpha or beta diversity of communities, or altered relative taxonomic abundance in relation to the comparator samples Relative abundance – proportion of sequences aligning with a particular taxon within a sampled community, but does not account for differences in total microbial population present Shotgun sequencing – sequencing of the entire DNA content in a sample based on generation of random fragments Metagenomics – the study of genetic material recovered directly from an environmental niche without ex vivo culture or perturbation

The Normal Lung Microbiome

Composition and source of the lung bacterial microbiome in health.

Traditionally, the lung below the glottis was thought to be sterile, but recent work has dispelled this dogma even in health^27,28. This is perhaps not surprising, as the lung is contiguous with the microbe-rich upper respiratory tract (URT). However, the microbial biomass in the lung is orders of magnitude lower than the oropharynx. Therefore, stringent approaches are required to distinguish authentic signal from artifact derived from both oropharyngeal carryover during bronchoscopy and the background noise of microbial sequences from instruments and reagents^22,29. Applying such approaches, we now know the healthy lung microbiota largely overlaps those from the URT and glottic region²⁸. The most abundant taxa are from Firmicutes and Bacteroidetes phyla, with the most common genera Prevotella, Streptococcus, and Veillonella^{28,30,27,28,30,31}.

The near-identical composition of these two niches that are physically contiguous but ecologically distinct indicates that the primary source of lung bacteria is the URT^30,32–34. They are derived mainly by microaspiration known to occur in all healthy people²⁷, balanced by clearance through mucociliary and innate immune mechanisms^35–37. There are modest regional differences within the lung and slight divergence from upper respiratory communities, suggesting differential clearance and possibly limited local replication even in healthy people³⁰. Indeed, a proportion of these lung-derived bacteria are viable and transcriptionally active³⁸. Thus the healthy lung bacterial microbiome is low-biomass, transient, and dynamic, which differs from the gut and URT microbiomes and indicates distinct microbial ecology in the lung³⁹. This balance is disrupted in diseases such as cystic fibrosis (CF), bronchiectasis and post-transplantation where local growth may become a primary factor for lung microbial composition.

The lung virome and mycobiome.

These components of the lung microbiome are less studied. Apart from acute infection with community-acquired viruses, sequence-based analysis has identified Anelloviridae as the most common and abundant component of the lung virome, along with bacteriophage, and more variable detection of other eukaryotic viruses including Herpesviruses and Papillomaviruses^40,41. Anelloviridae are small, single-stranded DNA circular-genome viruses found in blood and many tissues, with nearly 100% prevalence in humans, but have not been associated with disease⁴².

The fungal constituents in health are also understudied and present unique challenges^{8,23,28,41,43–45}. In addition to URT and background contamination, many samples have very low fungal biomass and low number of fungal taxa present⁴⁶, fungal cell walls are difficult to lyse, there are biases in 18S and ITS primers, non-standardized nomenclature, and poor annotation of fungal databases⁴¹. The limited small studies in healthy lungs show low fungal biomass and inter-individual heterogeneity, with fungi from the Ascomycota phylum most common, followed less commonly by Basidiomycota. The main genera were Candida, Saccharomyces, Penicillium, Cladosporium, and occasionally other environmental taxa; Aspergillus and Pneumocystis have been inconsistently identified at low levels in health.

The lung bacterial microbiome and local immune tone.

In the healthy lung, aspirated bacteria encounter efficient mechanical and innate immune defenses which oppose colonization^37,47. Nevertheless, despite the low biomass and dynamic nature of the lung bacterial microbiome, it impacts the local immune tone. In healthy humans, the presence of bacterial sequences in the lung that match those in the oropharynx correlates with higher basal Th17 lymphocytes when compared to people with little authentic lung bacterial content⁴⁸. Thus, entry of URT-derived microbiota may regulate tonic immune profiles in the lung.

At the same time, it is well-established that the gut microbiome, with exponentially higher biomass, sets systemic immune tone, and it may also impact the lung immune milieu through metabolites that enter the systemic circulation, the so-called gut-lung axis^49–51. In mice, both local lung and gut microbiota influenced lung immune profiles, but local lung microbial diversity and composition more strongly influenced pulmonary immune tone (referred to as the lung-lung axis)⁵². Thus, both the lung-lung and gut-lung axes contribute to the lung immune setpoint.

Establishment of the Post-Transplantation Lung Microbiome

Studies of the post-transplantation lung microbiome^8–15,53,54 have uniformly found higher microbial biomass and dysbiosis⁵⁵ characterized by lower alpha-diversity (a measure incorporating both the number of species and their evenness) and altered composition compared to healthy subjects. Dysbiosis likely results from donor-dependent, recipient-specific, and transplantation-related factors.

Pre-transplantation lung disease influences the post-transplantation microbiome.

Patients with CF or other suppurative lung diseases have distinct lung microbiomes^8,10,56,57, which can impact the post-transplant microbiome. Patients with CF have upper respiratory mucosal and sinus pathology with persistent colonization, and despite bilateral transplantation, these microbes often repopulate the lung allograft^56–58. The post-transplant bacterial microbiome in CF tends to be distinct from non-suppurative lung disease and, similar to pre-transplant profiles, often enriched in Proteobacteria, particularly Pseudomonas, though this is not universal^8–10,57.

Non-suppurative lung conditions requiring transplantation have microbiota that deviate from non-diseased healthy individuals, including emphysema, interstitial pulmonary fibrosis, and other interstitial diseases^59–62. Although less dramatic and lower microbial biomass than that seen in CF, pre-transplant lung dysbiosis may also impact the post-transplant microbiome, especially in cases of single lung transplant⁵⁴.

As described above, the URT is the source of the lung microbiome in health, and microaspiration is especially frequent post-transplantation^63,64. The URT is markedly dysbiotic in individuals with advanced lung disease both before and after transplantation⁶⁵. Whether this results from the underlying lung disease, its treatment, or both, is unclear. Nevertheless, the markedly dysbiotic oropharyngeal microbiome that provides a source of lung microbes likely contributes to lung microbiota post-transplantation.

Donor factors may impact the post-transplant microbiome.

Donor lungs are selected based on the absence of significant pathology, but post-transplant donor-derived infections are well-known⁶⁶. Donors are mechanical ventilated, and in this setting the lung microbiome is dysbiotic^67–70. While not well-studied, donor microbiota persistence in the allograft may be a factor shaping the post-transplant lung microbiome¹⁰. This may be altered further by the use of peri-transplantation antibiotics, but the degree to which this happens is unknown. A striking example of donor dysbiosis involves the lung virome, which is dominated by Anelloviridae. Anelloviridae sequences in donor lungs were elevated 100-fold compared to healthy people⁷¹, and these elevated Anelloviridae persist post-transplantation (and disseminate from the allograft to recipient circulation)⁷².

Transplant-related factors impact the lung microbiome.

Multiple surgical factors alter post-transplant lung microbial defense⁷³. Mainstem bronchial anastomoses and airway ischemia both may impair the mucociliary elevator, transplantation results in a lymphatic system that is not surgically reconnected, and vagal denervation leads to loss of afferent stimulation and dysfunctional cough reflex, impairing clearance.

Immunosuppression is required for allograft tolerance but increases the risk of both community-acquired and opportunistic infections⁷⁴. It seems logical that immune suppression would impact the lung microbiome through modulation of immune clearance but this has not been studied directly and is a key area for investigation⁷⁵.

Patients are given antimicrobials post-transplant for both prophylactic and therapeutic indications^74,76. Antimicrobials profoundly alter the gut bacterial microbiome composition, decrease diversity, and increase the reservoir of antibiotic resistance genes^77,78. Emerging findings suggest parallel effects may impact the lung⁷⁹. In mice, ceftriaxone reduced the relative abundance of lung Proteobacteria (Gram-negative bacteria) and increased Firmicutes (Gram-positives)⁵². Chronic azithromycin is used to reduce exacerbations in obstructive lung diseases, and in lung transplantation for CLAD prevention. Azithromycin may involve both immunomodulatory and antibacterial mechanisms, but whether it works through modifying the lung microbiome is unknown and discordant effects been reported^53,80,81. In post-transplant patients, azithromycin did not influence lung microbiome total bacterial burden, alpha-diversity, or high-abundance taxa, though differences in low abundance taxa such as reduced Enterobacteriaceae and Pseudomonas were seen at 1–2 years⁵³.

The Lung Bacterial Microbiome After Transplantation

Microbiome dysbiosis post-transplantation.

Multiple studies have found elevated bacterial burden compared to healthy controls, ranging from 15-fold to 1000-fold higher^8,9,15,53,54. Despite substantial methodological differences, studies indicate that a subset of post-transplant patients maintains a lung microbiome composition predominantly of oropharyngeal-type flora while most have a marked decrease in alpha-diversity accompanied an outgrowth of specific taxa^8,14,15,53. Quantitative comparison of lung versus URT communities shows outgrowth in the lung of commensals and potential pathogens, such as Pseudomonas, Enterobacteriaceae, Staphylococcus, as well as anaerobes⁸. In longitudinal studies of BAL from routine care, URT flora usually dominated in the first year after transplant^14,15. The most abundant phylum was Bacteroidetes, followed by Firmicutes, then Proteobacteria, and Actinobacteria^14,15. A smaller proportion of samples exhibited outgrowth of Corynebacterium and potential pathogens such as Pseudomonas, Staphylococcus, and Stenotrophomonas. In CF, two main types of post-transplant lung bacterial communities have been described: one typical of oral bacteria (particularly Streptococcus and Veillonella), and one dominated by Pseudomonas¹⁰.

Enrichment of Proteobacteria following transplantation is common, notably Pseudomonas^{9,10,57,58,82}. A study that carried out a more granular analysis revealed two species within the genus, P. aeruginosa and P. flourescens⁹. A higher relative abundance of P. aeruginosa was associated with lower diversity, more bacterial DNA, greater neutrophilia, and having symptoms at time of sampling. These culture-independent data add nuance to the culture-based associations between Pseudomonas aeruginosa and bronchiolitis obliterans syndrome (BOS; the most common phenotype of CLAD)^11,82–85.

Thus, across studies, the post-transplant lung microbiome is higher biomass and deviates from healthy individuals, but there is a high degree of patient-specificity. While this finding precludes broad generalizations about any homogenous “post-transplant lung microbiome” composition, it is this very heterogeneity that suggests distinct outcomes may be linked to microbiome profiles. Furthermore, the term “dysbiosis” is generally used to connote divergence from health, but essentially all post-transplant patients’ lung microbiomes differ from non-transplant healthy people. Therefore, the key question is whether there is an optimal post-transplantation lung microbiome versus communities that associate with elevated risk of post-transplantation complications. Similarly, it is important to distinguish microbiomes characterized during “routine” surveillance bronchoscopy, versus sampling done for-cause such as symptoms or allograft dysfunction.

Spatial and temporal heterogeneity of the lung microbiome.

While the lung microbiome is often conceptualized as a homogenous entity, there is under-appreciated spatial heterogeneity³⁰. In transplant patients, differences have been reported between proximal and distal samples^12,86 and between native lungs and allografts in after single lung transplant for interstitial lung disease (ILD) (in individuals with lung function decline)⁵⁴. This could result from regional differences in the entry, clearance, or local replication.

There are complex dynamics at the time of transplant and in the first post-transplant year⁶⁶. Early on, donor microbiota persists in some subjects post-transplant, while in others, it is rapidly replaced¹⁰. In CF, Pseudomonas strains derived from recipient populated the allograft within days^58,57. Far less is known about how pre-transplant native lung communities impact the post-transplant microbiome in non-CF patients⁶⁶.

Interestingly, there is extensive change over time in bacterial composition at lower (more granular) taxonomic levels, but more stability at higher taxonomic levels. In a longitudinal study, only ~10% of operational taxonomic units (OTUs, the molecular surrogate for bacterial species) found at the final time point were present initially¹¹. However, 46% of subjects exhibiting consistent phylum-level profiles over time. Furthermore, of those that had transitions in phylum-level profiles, about half went from dysbiosis (one phylum constituting >70% of the community) to no dysbiosis.

The bacterial microbiome and transplantation outcomes.

BOS, the more common obstructive form of CLAD, has had variable associations with the lung microbiome. An early study found that in CF, the presence of Pseudomonas species post-transplant appeared protective against BOS, particularly if the patient had Pseudomonas (identified by culture) pre-transplant¹⁰. In contrast, enrichment with oral taxa increased BOS incidence. This did not hold for non-CF subjects. Furthermore, decreased diversity over time within an individual increased BOS risk. These findings raised the hypothesis that the stability of microbiome pre-/post-transplant, rather than specific taxa, impacts BOS risk.

However, subsequent studies reported different microbiome correlations with BOS. A small, uncontrolled study found enrichment of Actinomyces and Xanthomonadaceae in BOS¹¹. Adding further complexity, both time- and location-specificity also matter. A study using shotgun metagenomic sequencing from routine bronchoscopy samples before BOS onset found Actinobacteria-dominance (Propionibacterium and Corynebacterium) at three months post-transplant had lower rates of BOS, less inflammation in BAL, and lower rates of ACR¹³. However, this association did not hold for later time points, suggesting a time-dependent association with outcomes. In an examination of topographical features of the post-transplant lung microbiome, microbial and inflammatory profiles differed in proximal (early BAL return) and distal airways (late BAL return), and BOS was linked to specific taxa from three phyla (Proteobacteria, Bacteroidetes, Firmicutes) in distal but not proximal specimens¹².

A recent study prospectively analyzed the microbiome from BAL from routine bronchoscopy one year after transplantation and looked at the predictors of subsequent CLAD⁸⁷. They found that bacterial biomass was a strong predictor of CLAD development or death within 500 days of the sample collection. For the highest tertile of biomass there was a hazard ratio (HR) of 5.98 (10.56 in multivariable modeling) for the composite outcome and when analyzed continuously for each log10 increase in bacterial biomass there was a HR of 1.89 (2.49 in multivariable model). In their compositional cluster analysis they found subjects who had “pharyngeal” or “inflammation-associated” taxa (such as Pseudomonas) had the highest risk of CLAD or death. No specific taxon was independently associated with CLAD. Additionally, immunosuppression neither significantly impacted bacterial microbiome composition nor mediated the effect between the microbiome and outcome. Overall, this study that found that early microbiome ‘type’ was linked to subsequent adverse outcomes.

In summary, emerging data suggest a complex relationship between the lung bacterial microbiome and BOS. Unraveling this complexity must account for temporal, spatial, and patient-specific factors. In addition, cross-kingdom interactions with the virome and mycobiome, as well as an integration with the multifaceted host response after lung transplantation needs future study.

The Lung Virome and Mycobiome after Transplantation

Eukaryotic viruses.

Herpesviruses, particularly cytomegalovirus (CMV) and herpes simplex virus (HSV), are important pathogens after lung transplantation, and are targets of prophylactic antiviral regimens⁷⁴. Community-acquired respiratory viruses (CARV’s), such as influenza, enteroviruses, and rhinoviruses, lead to acute morbidity and increased risk of later CLAD^3,88,89. These viruses are typically identified by targeted PCR methods. While there is a burgeoning literature probing the human virome through sequence-based methods^{40,71,72,89,90}, only a few focused on the lung and even fewer in lung transplantation.

The most abundant eukaryotic viruses in lung transplant BAL are of the Anelloviridae family--small circular DNA viruses (such as Torque Teno Virus; TTV)^{40,71,72,89,90}. Despite their prevalence and abundance, Anelloviridae effects on the host are not well understood. However, they are under immune control, as blood levels increase in organ transplantation and AIDS⁹¹. Anelloviridae levels typically drop before and during episodes of ACR in solid organ transplantation, which has led to interest in potentially using Anelloviridae as a functional measure of therapeutic immunosuppression^40,71,89,92.

In the lung, altered peri-transplantation dynamics of Anelloviridae has been associated with primary graft dysfunction (PGD)⁷¹. As noted earlier, donor lungs have elevated Anelloviridae levels before procurement. Perhaps counterintuitively, subjects with PGD had smaller increases in Anelloviridae levels across the peri-transplant period. This may reflect immune mechanisms that both inhibit viral replication and injure tissue, or early tissue injury itself limiting viral replication. Interestingly, donor Anelloviridae variants were transplanted along with the allograft and disseminated systemically in the recipient, while variants present in recipient blood at time of transplant often migrated into the new lung⁷². Lung Anelloviridae levels remain elevated after transplant and correlate with the degree of bacterial dysbiosis, indicating inter-kingdom covariation^40,72. The consequences of persistent Anelloviridae elevations and allograft/recipient virome intermixing remain to be determined.

In addition to the DNA viral landscape, RNA viruses are of great interest⁸⁸, and some reports have described a high frequency of persistent RNA CARVs post-transplant, including influenza A and B, parainfluenza, and rhinovirus⁸⁹. No association between viruses and ACR was identified, though BOS was not investigated.

Bacteriophages.

Bacteriophages (or phages) are abundant in the lungs of transplant recipients, particularly phages that infect bacterial inhabitants of the human respiratory tract⁷¹. Phages can influence bacterial pathogenesis in the lung, particularly in CF⁹³. Also, emerging data suggest that phages may directly interact with human cells and modulate innate immune function⁹⁴. However, what role phages play post-transplantation is unknown. Phages have also garnered interest as tools to fight multi-drug resistant bacterial infections such as Pseudomonas and Mycobacterium species, including in lung transplant recipients^95,96. Therefore, understanding the lung bacteriophage landscape may contribute to future therapies in lung transplantation.

Novel viruses in the respiratory tract.

A limitation in interrogating the virome by deep-sequencing is the incomplete annotation of viral databases. Indeed, a sizable portion of sequences in viral metagenomic studies cannot be annotated, termed viral “dark matter.” Sometimes low-quality or partial sequence alignments are found to unexpected non-human viruses, leaving the significance uncertain. Recently, sequences were identified in lung transplant specimens that partially aligned with a porcine circovirus⁹⁷. Through sequential contig-building and alignment, those sequences were determined to be a novel viral family termed Redondoviridae (named for their small, circular DNA genome)⁹⁷. This newly-identified virus was found in human BAL and oropharyngeal samples but rarely from other sites or species, and appears to be the second most common DNA virus in such specimens (after Anelloviridae). Future studies are needed to determine what role these (and potentially additional) newly-discovered members of the human virome have in lung pathogenesis, including post-transplantation.

The lung mycobiome.

Based on fungal culture, Aspergillus and Candida are commonly identified in lung samples after transplantation^8,46. Aspergillus is associated with early anastomotic complications, as it grows in ischemic anastomotic tissue⁹⁸. Several studies have linked Aspergillus infection or colonization to later CLAD^99,100, although recent work found no association¹⁰¹. Invasive Candida infections are common after lung transplantation, usually involving bloodstream, pleural space, or surgical site^74,102.

Sequence-based investigation of fungal microbiome (mycobiome) can employ targeted amplification or shotgun sequencing. The first post-transplantation metagenomic analysis found fungi were highly variable between patients in absolute quantity and composition^8,46. Candida was the most prevalent taxa, and when abundant in the lung it was also abundant in the oropharynx. This contrasted with healthy controls, where Candida was prevalent in oral samples but scarce in lung. Aspergillus was seen in BAL, sometimes at high levels, while other subjects had Cryptococcus and environmental fungi. Additionally, there was significant inter-kingdom covariation between Candida and Rothia, Streptococcus, and Veillonella⁴⁶.

Deep-sequencing was recently applied to microbial biofilms on endobronchial stents used to treat airway complications in a predominantly post-lung transplant cohort¹⁰³. Candida spp. were the most common fungal taxa in biofilms, followed by Aspergillus, which was particularly prevalent in stents placed to treat anastomotic dehiscence. Fungi generally considered environmental taxa (Scedosporium, Stereum, Sarcinomyces) were often seen at high absolute abundances. There was also evidence for fungal-bacterial co-variation and a link between stent material and biofilm type.

However, there are challenges to mycobiome analysis including issues of quantitative variability between samples, low absolute fungal biomass in many patients, sequence contamination, biases in ITS primers and database limitations⁴⁶, and further studies are need to address the lung mycobiome and to relationship to clinical outcomes.

Mechanisms of Lung Microbiome-Host Interactions and Injury

Both microbial pathogen-associated molecular patterns (PAMPs) sensed by host pathogen-recognition receptors (PRRs) and microbial metabolites mediate microbiome-host interactions³⁷, and it is essential to understand these interactions after transplant (Figure 2).

Figure 2. Host-microbiome interactions after lung transplantation.

Emerging evidence supports cross-talk between the lung microbiome and host responses after transplantation. While data is still evolving, the lung microbiome composition has been implicated in regulating basal immune tone and host response to resident microbiota and pathogens through cellular components like myeloid-derived suppressor cells (MDSC) and regulatory T cells (Treg), cytokines and chemokines, and other pathways (acetylated proline-glycine-proline (Ac-PGP); cyclooxygenase (COX)). Distinct microbiome profiles have also been linked to remodeling and fibrosis pathways that may contribute to post-transplant injury via multiple cytokines, growth factors, metalloproteinases, and collagen and matrix proteins. Additional abbreviations: Platelet-derived growth factor receptor-D, PDGFR-D; interleukin, IL; tumor necrosis factor-alpha, TNF-a; interferon gamma-induced protein-10, IP-10; matrix metalloproteinase, MMP; insulin-like growth factor-1, IGF-1

Microbiome modulation of inflammatory responses.

When a potential respiratory pathogen is identified by culture following transplantation, it is challenging to distinguish infection from colonization. One study of individuals suspected to have infection found the BAL microbiome profiles combined with cytokine signatures added useful context to culture data, and suggested that the microbiome modulated inflammatory responses¹⁰⁴. Specifically, clinically-diagnosed pneumonia had low alpha-diversity, enrichment of potential pathogens, and elevated inflammatory cytokines. In contrast, in colonization (culture positive without clinical signs of infection) there were more oral commensals and an adaptive immune profile characterized by regulatory T-cells (T_reg), decreased IL-17_A, and increased IL-2 and IP-10. The authors postulated that commensal anaerobes activate T_regs, which constrain the inflammatory response and limit progression to pneumonia. Another report linked the relative abundance of Pseudomonas in allografts with neutrophilic inflammation and levels of the neutrophil chemoattractant CXCL8 (IL-8), whereas the presence of oral taxa (Streptococcus, Prevotella, Veillonella, Actinomyces) negatively correlated with CXCL8, as did allograft alpha-diversity⁵³. These findings suggest that diverse communities comprised of predominantly URT taxa are less likely to promote neutrophilic inflammation and pneumonia. The post-transplant lung microbiome has also been linked to myeloid-derived suppressor cells (MDSCs) in BAL, which are bone marrow cells that home to the lung and modulate T-cell responses¹². Multiple MDSC subtypes were found in the post-transplant lung, including pro-inflammatory and immunosuppressive phenotypes, and the MDSC subtype correlated with microbiome composition.

Additional insight comes from comparison of allografts versus native lungs of single lung transplant recipients with underlying ILD⁵⁴. Compared to native lungs with ILD, allografts had higher microbial biomass and increased Acinetobacter and Pseudomonas, along with inflammatory cytokines and elevated acetylated proline-glycine-proline (Ac-PGP), a collagen breakdown product that increases neutrophilic inflammation and vascular permeability. Metabolome profiles correlated in particular with allograft Proteobacteria, implicating microbial-derived small molecules as potential drivers of inflammation. Together, these findings suggest that the lung allograft microbiome regulates local immune responses^48,52,105.

Remodeling and fibrosis.

CLAD pathogenesis involves remodeling and fibrosis. Studies from the SysCLAD Consortium identified links between the post-transplant bacterial microbiome and host-response gene expression profiles^14,15. BAL samples were characterized as pro-inflammatory (elevated TNF and cyclooxygenase-2 (COX2)), pro-remodeling (elevated platelet-derived growth factor-D (PDGF-D) and tissue inhibitors of metalloproteinases-1/matrix metalloproteinase-12 (TIMP1/MMP-12) ratio), or intermediate profiles¹⁵. The pro-inflammatory profile correlated with increased Firmicutes and Proteobacteria, peaked 3–4 months post-transplant, and had low macrophage abundance and elevated neutrophils. In contrast, the remodeling profile had increased Bacteroidetes, peaked about 12 months, and had more macrophages and fewer neutrophils. In vitro stimulation of monocytic cell lines with bacteria from these phyla recapitulated the inflammatory versus remodeling profiles. Another study extended those findings to address a fibroproliferative phenotype¹⁴. It identified an “anabolic profile” in 20% of post-transplant BAL samples characterized by increased host thrombospondin-1 and PDGF-D, which correlated with Prevotella, Streptococcus, Veillonella, and Neisseria. Conversely, a “catabolic profile” seen in 12% of subjects showed upregulated MMPs, collagen-V1 A-2, fibronectin-1, insulin-like growth factor 1 (IGF1), along with reduced microbial diversity, Staphylococcus, Corynebacterium, Stenotrophomonas, and Haemophilus, and culture-positive infections. In co-culture of macrophage and fibroblast cell lines, stimulation with Staphylococcus or Pseudomonas led to increased MMPs, consistent with the catabolic profile, whereas Prevotella or Streptococcus led to more collagen and fibronectin deposition, like the anabolic profile. These reports link distinct microbiome profiles to host fibroproliferation responses relevant to CLAD and provide tractable in vitro model systems to dissect host-microbe interactions.

Conclusions and future directions

The post-transplantation lung microbiome is dysbiotic, with markedly higher bacterial, fungal and viral abundance and altered composition compared to health. Emerging data suggest that the composition of bacterial communities is associated with adverse outcomes such as CLAD, although inconsistencies between reports limit conclusions, and the mycobiome and virome remain under-studied. Critical issues in understanding the microbiome in relation to outcomes include the need to account for patient-specificity, spatial heterogeneity, time-dependence, indications for sampling, and methodological differences. While data so far do not provide an answer about what might be an optimal or “healthy” post-transplantation lung microbiome, studies have begun to suggest links between microbiome composition and inflammation, fibrosis, and remodeling mechanisms relevant to both ACR and CLAD. As an emerging field, these promising early findings need future validation in high-quality clinical studies to establish links to outcomes as well as basic investigations to dissect host-microbiome mechanisms, which offer the potential for novel diagnostics and therapeutics after lung transplantation.

Future work is needed to examine the generalizability of the microbiome “types” seen in previous studies and determine what microbiome-related metrics might be operationalized as biomarkers for adverse outcome risk stratification or to guide therapeutic decisions. A second key area is to better understand the mechanisms of cross-talk between microbiome and host by which microbial constituents impact allograft outcomes. A third area needing additional research is the understudied virome and mycobiome.

Finally, a key frontier is targeting the respiratory tract microbiome for therapeutic purposes after lung transplantation. Perhaps the first step would be understanding how the microbiome influences or perhaps mediates responses to treatment with macrolides, such as reduced neutrophilic inflammation, and whether distinct microbiome profiles may be responsible for the clinical heterogeneity in response^4,109. Beyond that, therapeutic targeting of the microbiome might take the form of supplementing communities (a “probiotic” approach)¹⁰⁶, selectively reducing pro-inflammatory taxa (such as in phage therapy)^94,95, or using microbially-mediated small molecules^107,108 that have beneficial host-effects.

Figure 1. Factors contributing to lung dysbiosis following transplantation.

(1) The upper respiratory tract microbiome is the source of lung bacteria in health and is markedly dysbiotic both in people with advanced lung disease awaiting transplantation and patients post-transplant. (2) Microbial entry to the lung occurs via microaspiration, which is increased in post-transplant subjects. (3) Recipients’ microbiota seed the allograft from large airway colonization, particularly in cystic fibrosis and related conditions, or via dysbiosis in the native lung in single lung transplantation. (4) The donor lung may bring an endogenous microbiome along with it. (5) Mechanical clearance is disrupted in transplant recipients through mucociliary dysfunction, anastomosis site barriers, disrupted afferent loop of the cough reflex, lymphatic disruption. (6) Immune clearance is impeded by immunosuppressive treatment and disrupted lymphatics. (7) Local replication in the lung is increased compared with healthy people, with greater divergence (beta-diversity) between URT and lung microbiome communities. (8) The pressure exerted by frequent or continuous antimicrobial treatment may alter the microbiome in ways that remain to be determined.