The environmental microbiome, allergic disease and asthma

Abstract

The environmental microbiome represents the entirety of the microbes and their metabolites that we encounter in our environments. A growing body of evidence supports the role of the environmental microbiome in risk for and severity of allergic diseases and asthma. The environmental microbiome represents a ubiquitous, lifelong exposure to non-self antigens. During the critical window between birth and one year of life, interactions between our early immune system and the environmental microbiome have two consequences: our individual microbiome is populated by environmental microbes, and our immune system is trained regarding which antigens to tolerate. During this time, a diversity of exposures appears largely protective, dramatically decreasing the risk of developing allergic diseases and asthma. As we grow older, our interactions with the environmental microbiome change. While it continues to exert influence over the composition of the human microbiome, the environmental microbiome becomes increasingly a source for antigenic stimulation and infection. The same microbial exposure protective against disease development may exacerbate disease severity. While much has been learned about the importance of the environmental microbiome in allergic disease, much more remains to be understood about these complicated interactions between our environment, our microbiome, our immune system and disease.

Introduction

The prevalence of allergic diseases, including atopic dermatitis (AD), food allergy (FA), allergic rhinitis (AR) and asthma, appears to be increasing over the past century, suggesting that environmental factors—rather than changing genetics—are primarily responsible for this change (1, 2). Allergic diseases share a common underlying pathophysiology where the immune system reacts to environmental exposures that are commonly tolerated by others without allergic diseases. Strachan’s original hygiene hypothesis—demonstrating that risk of hay fever was predicted by the number of siblings and their birth order, postulated that the number of older siblings was a surrogate marker for increased microbial exposures. This represented one of the first steps in a growing understanding that changes in microbial exposures may explain part of the changing incidence of allergic diseases (3).

Research and theories regarding the importance of environmental microbial exposures in allergic disease development have continued to expand, progressing from the hygiene (3) and old friends (4) to the microbial diversity (5) and biodiversity (6) hypotheses. Largely, these theories highlight evidence suggesting that urbanization and other social changes have altered the way in which humans interact with the environment, resulting in altered exposure to environmental microbes and leading to increased prevalence of allergic diseases. While evidence continues to mount that diverse microbial exposures in early life appear beneficial, microbial exposures are not uniformly positive and the same microbial exposure plays a key role disease prevention may exacerbate severity after disease onset, such as with the case of microbial exposures and asthma (7–14). The idea of One Health represents a more all-encompassing theory stressing that the health of humans, animals, microbes and the environment is predicated on a complex interdependence, involving both symbiotic and pathogenic interactions (15).

We aim in this review to provide a framework outlining the two primary mechanisms by which the environmental microbiome impacts allergic diseases and asthma: through establishment and maintenance of the individual microbiome, and through exposure and/or infection of the host (Figure 1). We describe how exposure to a diverse environmental microbiome in early life helps establish a diverse human microbiome, crucially driving the maturing immune system toward tolerance and protecting against allergic disease development. We also highlight how diverse exposure in turn does not appear to be universally beneficial. We discuss the importance of infections as a risk factor for developing allergic diseases and examine the evidence suggesting that some microbial exposures may both protect against disease development and exacerbate disease severity depending on the timing of exposure. We summarize the evidence for these interactions as well as limitations in the field. We close by discussing how this knowledge impacts clinical practice today and future research in the field.

Figure 1.

Framework for understanding the interactions between the environmental microbiome, developing immune system and allergic disease.

Measuring microbes, exposures and communities

Defining the environmental microbiome

The environmental microbiome represents both the microbiota—the sum of the viruses, fungi, bacteria, archaea and lower eukaryotes—existing in an environment encountered by humans and the metabolites, structural molecules and surrounding environmental conditions created by the microbiota in that environment (16). The environmental microbiome is distinct from the human microbiome, which represents the microbiome specifically found on or in an individual host, for example inhabiting their skin, gut, or respiratory tract. Comparisons between individual microbiomes can be made using omnibus measures such as alpha and beta diversity, by evaluating differential abundance of microbial clades or microbial functions between samples, or by examining changes to microbial networks (17).

Culture-independent approach: progress and challenges

The development and widespread use of culture-independent techniques to identify microorganisms has allowed for a greater appreciation of the diversity and richness of environmental microbiota, yet these methods are not without limitations. Largely, direct microscopy and traditional culture are no longer widely used, as these techniques can be challenging to reproduce, and only a small subset of known microbes can be readily grown in vitro (18, 19). Many epidemiologic studies of microbial exposure have relied on measurement of microbial cell wall components such as endotoxin, using the Limulus Amebocyte Lysate (LAL) assay to quantify microbial load; these approaches do not identify the microbes present but rather seek to quantify the magnitude of exposure to a certain class of microbes (20). Culture-independent methods have largely replaced these techniques and have the added advantage of gaining insight into microbial load and species diversity in a sample (17). Some of these techniques look primarily at which microbiota are present. These include quantitative polymerase chain reaction (qPCR), which uses a prespecified marker gene limited to one species of microbe, and amplicon sequencing, which utilizes conserved marker gene sequences (often using 16S ribosomal RNA gene for bacteria or the internal transcribe spacer region for fungi) to visualize the diversity within a specific microbial kingdom. In contrast, metagenomics shotgun sequencing captures both predicted function of the microbial community and microbial identity at a deeper taxonomic resolution by sequencing all genes in a given sample. Metatranscriptomics goes further, capturing all actively transcribed genes and their products (Table 1). Metabolomics measures the small molecule products of cellular activities, contributing insight into not only what microbes are present but how they may be influencing the host. Although, current approaches do not distinguish between microbial-derived versus host-derived metabolites.

Table 1.

Culture-independent techniques to measure the environmental microbiome

However, keen attention to experimental design is essential when utilizing these sequencing techniques, as decisions made along the entire course of a study including sample collection, preservation, and storage decisions, nucleic acid extraction method, primer choice, library preparation, sequencing platform, and the bioinformatics pipeline can introduce variation that exceeds differences resulting from the biological variable of interest (17, 21–23). While studies performed on high biomass human samples such as stool have suggested that the variation introduced by choice of DNA extraction method is negligible, when evaluating more challenging samples with high host and low microbial nucleic acid content (for example built environment samples such as vacuumed dust), the choice of DNA extraction method introduces significantly more variability. This variability has the potential to overwhelm the signal from the biological factor of interest and makes between-study comparisons more challenging (24).

Microbiome sequencing generates high dimensional datasets with unique characteristics that require specialized statistical approaches to draw proper inferences, such as zero inflation, compositionality, and differences in sequencing depth between samples. Optimal methods for data analysis remain under some debate within the field (17, 25, 26). Most microbiome studies are designed to evaluate the relative rather than absolute abundance of microbial clades in a sample, as sequencing read count reflects laboratory technical factors rather than absolute microbial abundance (27). While absolute microbial abundance has been shown to vary between individuals and explain differences in disease states, techniques to measure this can be time-consuming and challenging (28). Measurement of absolute microbial abundance requires additional procedures at the time of sample processing such as the addition of known quantities of a microbe to the sample (“spike-in”) or low throughput time-intensive processes like flow cytometry (25, 29, 30). As a result, many microbiome studies are designed to evaluate the relative rather than absolute abundance of microbiota. Use of relative, rather than absolute abundance, however, requires statistical methods that deal with the compositionality of resulting datasets to avoid spurious correlations. A large number of analytic options exist for assessing diversity, differential abundance, and microbial network structure. Yet, different analytic choices applied to the same microbiome dataset can lead to widely different conclusions (26, 31). Adhering to common reporting guidelines, such as the STORMS checklist, will be critical for reproducibility of results moving forward (32). Finally, metagenomics and metatranscriptomics remain costly, meaning that most environment microbiome work has focused primarily on amplicon sequencing, which largely captures limited taxonomic information and lacks functional information (17, 33).

The endotoxin “paradox” and asthma

Research on the gram-negative bacterial cell membrane component lipopolysaccharide, also called endotoxin, represents an illustrative example highlighting both the growth of the field and the challenges that remain in measuring the impact of environmental microbial exposures on allergic diseases. Much work has been done investigating the link between bacterial and fungal markers and allergic diseases and asthma, including endotoxin and 1,3 beta-D-glucan. Yet, these studies, which have relied on assays quantifying microbial load by measuring microbial cell wall components such as endotoxin, have demonstrated conflicting associations with allergic diseases (20, 34, 35). One large study, utilizing three separate birth cohorts from Spain, Germany, and the Netherlands, examined early life endotoxin exposure in house dust and subsequent asthma diagnosis later in life found a protective effect of endotoxin in the Spanish cohort, a neutral effect in the German cohort and a harmful effect in the Dutch cohort (36). Recent work has shown that not all endotoxins are the same. Depending on the source bacterial species, endotoxin may elicit either a pro-inflammatory or anti-inflammatory host response (37). In this translational study, birth cohorts were established in Russia, Estonia, and Finland, where there is a known gradient of allergic diseases (37). Differences in the infant gut microbiome were observed by country, with a higher relative abundance of Bacteroides dorei in the Finnish infants and differences in microbial Lipid A (the bioactive component of endotoxin) biosynthesis (37). These investigators isolated a range of bacterial strains from the infants’ stool and purified endotoxin from each strain (37). Endotoxin from Escherichia coli elicited a robust inflammatory cytokine response from human peripheral blood mononuclear cells, while endotoxin from Bacteroides dorei inhibited the ability of Escherichia coli endotoxin to stimulate a cytokine response (37). Most epidemiological studies have characterized endotoxin exposure using the Limulus amebocyte lysate (LAL) assay, which does not approximate the biological response elicited in human-based assays such as incubation with either human whole blood (38) or human immune cells (39), and does not identify the source bacterial species for endotoxin. Assessment of endotoxin levels using the LAL assay are poorly correlated with microbiome attributes such as species diversity or richness (40). and cannot identify the source microbe, explaining the conflicting associations between endotoxin exposure and asthma. Understanding the best way to characterize microbial exposures will be critical for future efforts in reducing the incidence and morbidity due to allergic diseases (20).

The interplay between the environmental microbiome, the host microbiome and the immune system

Host microbiome is derived from the environmental microbiome

The controlled environments of mouse studies allow for insights into the importance of the environmental microbiome in populating the host microbiome with downstream effects on immune response. Although these studies have focused on using co-housing as a surrogate for the environmental microbiome rather than simultaneously measuring both the host and environmental microbiome, the control of diet as well as host genetic factors provide compelling proof of the important role environmental microbial exposures play in the host microbiome. In a murine study focused on the cecal, lung, and skin microbiomes, microbiome composition and diversity clustered strongly with environmental factors such as timing of shipment, lab of origin, and the cage in which mice were housed; inflammatory cytokine patterns were strongly correlated with lung microbiome composition (41). A study employing a randomized co-housing experiment of adult wild-type and transgenic Toll-like receptor (TLR) deficient mice demonstrated that a change in cage environment altered the gut microbiome of mice, and that co-housing had a large impact on gut microbial communities that was independent of host innate immunity based on TLR status (42). Further highlighting the importance of environmental rather than host factors in determining the composition of the gut microbiome, germ-free embryos of C57BL/6 mice, a common laboratory strain, were transferred into genetically distinct pseudopregnant wild mice (Mus musculus domesticus). The resulting “wildling” mice adopt the bacterial, fungal, and viral microbiome of their surrogate mothers (wild mice) rather than their biological mothers (C57BL/6), and their immune responses mirror those of the wild mouse (43). Taken together, these controlled murine experiments demonstrate that environmental microbial exposures have a large influence on the composition of the host microbiome and influences baseline immune tone.

Human studies of co-housing have likewise shown the strong influence of environmental exposures on the host microbiome. Large studies performed in adults have convincingly shown that it is environmental rather than host genetic factors that shape human gut microbiota. A large study of 1046 healthy adults showed that host genetics did not explain variability in the gut microbiome between individuals, and that genetically unrelated individuals living in the same household also had similarities in the composition of their gut microbiome (44). A separate study of co-housing which included the fecal, oral, and skin microbiome of both humans and pets showed that humans within the same household shared similar microbiomes both with each other and with their pets, and that dog ownership increased similarities between the skin microbiomes of adults in that house, likely due to increased direct and indirect contact between household members related to dog owner (45).

Besides co-housing, another surrogate for environmental microbial exposures in human studies is farm exposure. Growing up on a farm has been demonstrated to be protective against the development of asthma, AD, AR and allergic sensitization, with several potential mechanisms to explain this protective effect (46, 47). Acinetobacter lwoffii and Lactococcus lactis strains isolated from farm cowsheds induced maturation and Th1 polarization of human monocyte-derived dendritic cells (48). Intranasal treatment of mice with these strains prior to ovalbumin challenge reduced allergic sensitization, lung eosinophilic infiltration, and airway hyper-responsiveness, suggesting that exposure to specific farm-associated microbial strains is protective (48). Environmental microbes found in the indoor house dust of farm homes, like those of the Lachnospiraceae and Ruminococcaceae families, have been found in the gut microbiomes of children growing up on farms and appear to mediate the protective effect of farms for future development of allergic diseases (49, 50). Other studies have suggested that the protective farm-effect can be attributed to the consumption of unpasteurized milk (51, 52), with epigenetic reprogramming of host immune cells either directly or indirectly through microbial exposure (53).

While moving to a farm environment may not be a practical clinical recommendation, studies suggest that a farm-like environmental microbiome could be created. A farm-like microbiome was also shown to correlate with families who wore their shoes in the home, supporting the notion that soil-derived microbial exposure may represent part of this effect (49). Mouse models support this, showing that exposure to soil rather than bedding mediates a protective effect in murine asthma models (54). Living close to or spending time in green spaces with high environmental microbial diversity can influence both the host microbiome and risk of allergic disease (55). While studies have shown mixed associations between pet ownership and atopy, animal exposure may provide similar protective farm-like influences on the host microbiome and allergic disease (56–58).

There are a few studies that have conducted simultaneous evaluation of environmental and host microbial communities, and demonstrate convincing evidence that there is transfer of environmental microbes to the host even in adulthood (49, 50, 56, 59–61). There is a critical window at birth and early in infancy when the host microbiome is more sensitive to environmental microbial exposures, with downstream effects on immune maturation. Strain-level analyses convincingly show that there is vertical transmission of maternal microbes to the infant, particularly in infants born via vaginal delivery as compared to infants born via Caeserean section (62). Beyond mode of delivery, diet, and early life antibiotic administration (60, 61, 63), recent studies have shown that even environmental factors such as daycare attendance, a source of environmental microbes, independently influences the maturation of the early life gut microbiome (64). The early-life microbiome (birth through early childhood) is considered highly dynamic whereas relative stability is a hallmark of the adult microbiome (14). Yet, even later in life there is evidence that our microbiome continues to be influenced by the microbiome in the environment, particularly in the context of new microbial exposures associated with animals (40, 45, 65–67). Home studies show that people share microbes with their pets (45, 65). The upper respiratory tracts of livestock workers are colonized by strain-specific methicillin-resistant Staphylococcus aureus (MRSA) present in their work environment (66, 67). In a pre-post study of animal care technicians where the built environment and human microbiomes were concurrently measured, a single eight hour work shift cleaning mouse cages led to transfer of environmental microbes to the nasal and skin microbiomes of the workers (40).

Both murine studies and human studies where surrogate indicators such as co-housing or farm exposure for the environmental microbiome, or direct measurement of the environmental microbiome, have shown that there is colonization of the human microbiome by environmental microbes. While the host microbiome is more susceptible to environmental microbial colonization during a critical window early in life, even in adulthood the host microbiome is influenced by environmental microbial exposures with downstream implications for the development or exacerbation of allergic diseases.

Environmental microbiome as source of pathogens

So far, we have outlined how a healthy environmental microbiome—such as that found on a farm—populates and develops our own microbiome and trains our early life immune system what to tolerate. Yet, some might argue, the larger purpose of the immune system is host defense, protecting us from harmful microbial invaders. While perhaps not often thought about in this regard, pathogens that result in infections also represent exposures from the environmental microbiome. Which microbes we encounter, even down to the specific strains of viruses and bacteria in our environment, can impact risk for and severity of allergic diseases.

The role of the host microbiome in immune training

As reviewed above, the human microbiome is derived from our environmental microbiome. Extensive, recent reviews have highlighted the importance of the human microbiome in diseases along the atopic march (atopic dermatitis (68–71), food allergy (72–78), allergic rhinitis (79) and asthma (80–88)). A large part of this risk is attributable to the role of the human microbiome in training the immune system and engender tolerance (89, 90), and through persistent epigenetic modifications on host immune cells (53, 91–93).

In contrast with the more rigid and reactive adult immune system, in early life the immune system is more plastic with a predilection toward tolerance. Interactions between the host immune system, environmental microbiome and host microbiome during the first year of life are essential to training the immune system to balance tolerance with immune response (89, 90). Studies from germ-free mice have identified the importance of the microbiome in developing organ specific lymphoid tissue and a balanced immune response (90). Germ-free mice exhibit Th2 skewing, impaired organ-associated invariant natural killer T (iNKT) cell function, hyper-IgE response, and a decreased regulatory T-cell (Treg) abundance in the colon, lung and skin (90). This “allergic phenotype” in germ-free mice can be reversed when they are exposed to commensal microbes during early life, highlighting the crucial role the host microbiome plays in establishing immune tolerance (90).

The gut microbiome specifically plays a role in helping facilitate oral tolerance, down-regulating reactions to foreign stimuli and modulating risk of atopy. Human studies have demonstrated that in people with gut microbiome communities associated with allergic disease, there is also a concomitantly increased CD4+ and decreased Treg cell abundance (89). Gut accumulation of Bacteroides, Clostridium, Lactobacillus and Veillonella species, in contrast, upregulates Treg cell function (14). Beyond the role in training the immune response, another mechanism by which environmental microbiome modulates immune activity is through epigenetic changes (53). Exogenous Lactobacillus reuteri administration to pregnant mothers altered DNA methylation in CD4+ cell populations in their children (93). Short chain fatty acids such as acetate and butyrate produced by gut microbes results in altered histone acetylization thus regulating Treg and mast cells through epigenetic modifications (91, 92). Thus mechanisms by which the environmental microbiome influences allergic disease is through changes in the gut microbiome, or through epigenetic modification of immune cell populations.

The environmental microbiome and allergic disease

Having defined the environmental microbiome, explored how it is measured, and examined its influence on the human microbiome, and mechanisms by which it influences tolerance, we next discuss the evidence on the influence of the environmental microbiome on individual allergic diseases.

Asthma

The evidence supporting the protective effect of growing up on a farm against asthma development remains one of the best studied examples of the influence of environmental microbial exposures on allergic disease. Growing up on a farm has been demonstrated to be protective against the development of asthma, AD, AR and allergic sensitization (46). This association has now been directly linked to the farm environmental microbiome. Increasing environmental microbiome diversity—as measured in home dust samples in the first year of life—correlates with degree of protection from allergic diseases (10, 49, 50, 52, 94–96).

The protective effect of the farm environment is mediated through multiple mechanisms. A key role involves appropriate gut microbiome diversity and maturation in early life (49, 50, 52). Gut and nasal colonization by environmental microbes found in the indoor house dust of farm homes, like those of the Lachnospiraceae and Ruminococcaceae families, appears to be strongly associated with the protective farm effect (49, 50). Both the timing at which these exposures occur and the relative abundance of different microbes in the microbiome community appear important for this asthma-protective effect (50). Exposure to the farm environmental microbiome guides infant gut microbiome development on a course toward appropriate tolerance, whereas developing too quickly or too slowly appears to confer risk. Both diversity of microbial exposure and appropriate maturation of the microbiome appear key (50).

Even in instances when people did not grow up on farms, if their environmental microbiome was farm-like, similar protection from asthma has been observed (49). This farm-like indoor microbiome contains more rumen-associated and fewer human-associated microbes (49). While human-specific microbes can colonize us (i.e. Streptococcaceae family and Staphylococcus genus), they also have the potential to infect us. In contrast, human disease from bovine-specific microbes, such as those of the Methanobrevibacter genus, are less likely to infect humans, so our immune system can upregulate tolerance to these types of species. Consistent with the critical window hypothesis, it the farm-like effect is strongest in the first year of life but persists through childhood (49).

While numerous studies show increased bacterial richness and diversity on farms, more mixed data exists regarding fungal diversity. Increased fungal diversity on farms has not been consistently associated with change in asthma risk, whereas high fungal diversity in house dust was associated with decreased risk of asthma in more urban cohorts (49, 52, 97).

Farms, and a farm-like environmental microbiome, appear protective against asthma, whereas developing in an urban area, and obtaining an urban microbiome, conversely appears to confer risk. Infants growing up in urban environments develop a more homogeneous urbanized gut and airway microbiome, with increased relative abundance of Veillonella, Rothia and Haemophilus and with decreased Bacteroidetes (98). The microbial community structure of urban house dust samples is significantly different from that derived from rural house dust (95). In neighboring geographic regions, urbanization was associated with significantly decreased Acinetobacter diversity in the nasal microbiome, which was also strongly associated with increased asthma prevalence (99). Having an urban gut microbiome during the critical window of early life was also associated with increased asthma risk (98). Analogously, cohort studies of US urban children show that both high allergen exposure and high environmental bacterial diversity in house dust at one year are strongly associated with lower risk of atopy and lower risk of asthma (10, 96). In particular, presence of some bacterial in the Bacteroidetes and Firmicute phyla in house dust—Prevotellaceae, Lachnospiraceae and Ruminococcaceae families—are associated with this protective effect (10).

Growing evidence suggest that farm versus urban gut microbiome signatures are also associated with discordant immune responses. Blood leukocytes from children with a farm-like microbiome were shown to have less inflammatory cytokine response to bacterial cell wall antigens, supporting the notion that farm environment microbial exposures may engender tolerance (49). Gut microbes produce short-chain fatty acids (such as acetate, butyrate and propionate) which have been demonstrated to modulate epigenetic changes in genes responsible for balancing immune responsiveness and Treg activity (53, 100). Abundance of gut bacteria known to produce butyrate, prevalent in the guts of farm infants, has been associated with decreased risk of asthma (50, 101). In contrast, markers of a Th2 immune response (C-C motif chemokine ligands 11, 13 and 17) were higher in children with an urban microbiome profile (98).

Beyond early life, the indoor microbiome influences the severity of asthma. However, while exposure to diverse microbes in early life is protective against asthma (9, 102), studies in established asthmatics show the opposite; high environmental microbial diversity is associated with more asthma symptoms (11, 12). In patients with atopic sensitization, lower fungal diversity of house dust was associated with more severe asthma symptoms (11). In school-aged children, increasing house dust fungal and bacterial diversity has been associated with increased asthma symptoms (11). Low house dust bacterial diversity is associated with increased rates of atopy in adults, while no association was found with bacterial diversity and asthma (103). Some studies have demonstrated that high levels of dust fungi are associated with asthma (104, 105). Specifically, the presence of the fungal genus Volutella appeared to be associated with worse asthma severity (11).

Analogous to the home indoor microbiome, the school environment represents an important—less studied—location of exposure for childhood asthma symptoms. Paralleling what is seen in home-based studies of school age children with asthma, increasing bacterial diversity of classroom dust is associated with increased asthma severity (106). In university students, an indoor microbiome containing species that have been shown to be protective early in life, such as Ruminococcus, was associated with increased asthma symptoms (107). In children sensitized to fungi, their presence in schools can greatly exacerbate asthma symptoms (108). Built environment factors, including age of the school-building, humidity and presence of air-conditioning systems, appear to shape the indoor dust microbiome including both bacterial and fungal changes, and impact severity of asthma symptoms (109, 110).

Perhaps more than any other allergic disease, interactions with the environmental virome influence the development and severity of asthma. Infection with rhinovirus (RV) or respiratory syncytial virus (RSV) before the age of three has been associated with dramatic increases in the risk of asthma, up to ten-fold by the age of six (111–115). Infection with these viruses is exceedingly common in childhood; yet, timing of infection and specific viral virulence factors confer differential risk of asthma development. Infection within the first three months of life is riskier than infections later in childhood for subsequent development of asthma in childhood (116). Severity and strain of RSV and RV appear important as severe but not mild infection increases risk and RSV Long strain—as opposed to A2 or line 19 strains—induces heightened reactivity of the airway (111, 117, 118). Further, RV-B, compared to RV-A and RV-C, appears more mild and less likely to cause wheezing related illness (111, 119). While it remains less clear whether the altered Th2 immune response precedes RSV or RV infection or is a result of it, evidence suggests that especially early in life, RSV or RV infection precedes asthma and may drive a push toward Th2 skewing (111, 116, 120). However, viruses do not appear to uniformly drive Th2 skewing. In mouse models of asthma, influenza infection protected against asthma development mediated by both CD4⁺ and CD8⁺ memory effector T cells (7).

Those with asthma are at risk of viral illness as the disease phenotype drives an imbalanced immune response, favoring a Th2 over Th1 driven response and resulting in inadequate viral clearance (111). Viral illnesses represent the largest driver of asthma exacerbations (121). More than half of acute asthma exacerbations in adults and children are associated with viral infection (122, 123). RV is the most commonly identified virus associated with wheezing in those greater than one year old, with RV-A and RV-C causing more severe exacerbations (111, 121).

Beyond their intrinsic risk, the severity of viral infections is modified by characteristics of the upper airway microbiome. Colonization with Haemophilus, Streptococcus and Moraxella is associated with more severe respiratory symptoms during viral illnesses in asthma whereas high levels of Corynebacterium, Staphylococcus and Dolosigranulum are associated with protection (115, 124, 125). Early colonization with these same bacteria in the upper airway at one month—Haemophilus, Streptococcus and Moraxella—was shown to increase probability of pneumonia and bronchiolitis within the first three years of life (126). Infants with this bacterial colonization pattern were also shown to exhibit heightened inflammatory immune responses (126).

Atopic dermatitis

Much like in other allergic diseases, dysbiosis appears to precede development of AD, especially the overabundance of Staphylococcus aureus (SA) in the skin microbiome (127–129). Yet, only a few studies have directly shown that increased environmental SA exposure or burden is directly related to colonization or AD risk (95, 130). SA content of bedroom dust is associated with both presence and severity of AD (130). Diversity of exposure also appears to influence risk of AD as decreased house dust bacterial diversity in rural homes was associated with AD, with Clostridiales, Ruminococcaceae, and Bacteroidales all shown to be individually associated with decreased likelihood of AD (95). Studies have demonstrated that various farm exposures are correlated with less AD, yet the mechanism remains unclear (46). Whereas an urban gut microbiome at one week of age increased risk for AD risk this association was lost at one month and one year (98).

Similarly to asthma, AD leads to a Th2 skewed immune system altering responses to viral pathogens making infections more likely, which in turn can lead more severe viral infections in those with AD (131). Herpesvirus infections, leading to eczema herpeticum, and severe symptoms with Cocksackie virus infections are common in AD (131). Whereas infection with Molluscum contagiosum virus is often benign in young children, children with AD are both more prone to infection and experience disease flares when infected (132). Geographic differences in microbial exposures are highlighted by Cowpox infections causing AD exacerbations in Europe, a virus that does not typically exist outside of Europe and Asia (132, 133).

Additionally, bacterial skin and soft tissue infections, most often from Staphylococcus and Streptococcus species, are both more likely and more severe in AD, contributing significantly to AD morbidity (131). Even which specific strains and virulence factors of SA one is exposed to appear important in AD. SA superantigens increase risk for skin and soft tissue infections in AD. SA isolates from those with AD have been shown to produce greater amounts of toxic shock syndrome toxin-1 (TSST-1), staphylococcal enterotoxin (SE) and exfoliative toxin (134). While SA is relatively ubiquitous in our environmental microbiome, SA virulence factors associated with skin infections show large geographic differences, highlighting the importance of these interactions (134).

Allergic rhinitis

Environmental microbiome research has less often explored AR as the primary outcome, often preferring allergic sensitization, yet the existent data suggests that early, diverse exposure prevents AR development. As with other allergic diseases, farm living appears protective against AR (46, 135). However, studies looking at microbiome diversity and association with allergic sensitization have shown mixed results (10, 98, 136). Studies that have explored AR directly as the outcome suggest that, like other allergic disease processes, diversity and richness of the indoor microbiome are protective against AR (137, 138).

Research suggests that while the environmental microbiome does not seem to prevent sensitization, it does appear to protect against AR development. The reasons for this apparent paradox between allergic sensitization and AR outcomes are not entirely clear. However, it is known that some with allergic sensitization do not go on to develop allergic disease, and both allergic sensitization and AR continue to increase in prevalence throughout childhood (139, 140). The mechanism by which the environmental microbiome modulates AR are not known, but colonization of the host nasal microbiome with subsequent epigenetic modification of nasal epithelial cells may be one potential pathway; a recent study showed that the association between nasal microbial richness and AR diagnosis was largely mediated through epigenetic modifications of the nasal epithelium (141). Yet further mechanistic understanding—whether mediated by the presence of certain microbes, the interplay between early life exposures and immune system training, or other currently unknown factors remains to be seen.

Even less data exists regarding the impact of the environmental microbiome on AR symptomatology. Presence of indoor mold has long been associated with AR, but this has most relied largely on home inspection (visible mold) rather than microbiology techniques (142). Recent culture-independent studies show mixed results with regard to the association between indoor microbial diversity and allergic rhinitis diagnosis (103, 143). One study showed that Firmicute abundance was associated with well controlled hay fever in adults living on farms in the US (103). Another using shotgun metagenomic profiling revealed AR symptoms were associated with several bacteria that are more likely to be found in urban schools in China (143). More research focusing on AR as the outcome utilizing modern culture-independent techniques to measure exposure will be important to our understanding of the environmental microbiome’s impact on AR severity.

Food allergy

Despite the vast research into the importance of the gut microbiome in FA, the role of the environmental microbiome in FA is less well established (73). Associations between microbial exposures and FA have more often been observed via surrogate measures of diverse microbial exposures. But, the microbiological underpinnings of these findings are poorly understood. Exposure to older siblings and pets have been shown protective against FA, whereas Cesarean section delivery seems to confer risk (144, 145). Breast feeding may be beneficial in those already with AD or FA as it has been shown to restore the gut microbiome closer to that of non-atopic infants (146). While the farm effect has been well studied in other allergic diseases, it has been less robustly analyzed with respect to FA (147). Studies assessing the link between FA and diversity of exposure, as measured by endotoxin level, have generated mixed results (35). A cohort study of Old Order Mennonite communities in the US demonstrated that a more traditional lifestyle was associated with decreased FA risk (148). The urban microbiome signature was not shown to increase risk of food sensitization by six years of age in a Danish cohort (98). Yet, germ-free mice appear to offer more direct evidence of the importance of a diverse environmental microbiome as gut colonization with a high-diversity microbiome in early life protected against exaggerated-IgE immune response, whereas colonization with a low-diversity microbiome was similar to no microbiome colonization (149). Given what is known about the importance of the gastrointestinal tract and the gut microbiome in promoting tolerance, and evidence that environmental microbes populate the gut microbiome, the lack of direct associations between the environmental microbiome and FA in human studies may reflect the paucity of studies in this area.

In Conclusion

The environmental microbiome represents a ubiquitous, lifelong exposure to non-self antigens. During the critical window between birth and one year of life, the interactions between our tolerance-prone early immune system and the environmental microbiome have two consequences: our individual microbiome is populated by environmental microbes and the environmental microbiome participates in setting immune tone either directly through environmental microbial exposure, or indirectly through the host microbiome. During this time period, a diversity of microbial exposures is largely protective for the future risk of developing allergic diseases. Atopy risk is altered through various mechanism including antigenic stimulation, metabolite production, and epigenetic changes to immune populations. As we grow older, the environmental microbiome exerts a smaller influence on the host microbiome, and our interactions with the environmental microbiome change. The environmental microbiome becomes increasingly a source for antigenic stimulation and infection. The same microbial exposures that play a critical role in allergic disease prevention early in life are often also associated with exacerbation once disease is established. This process has importance to both the development of allergic diseases in addition to the severity of disease and frequency of exacerbations. While much has been learned about the importance of the environmental microbiome in allergic disease, much more remains to be understood about these complicated interactions between our environment, our microbiome, our immune systems and allergic disease.

In practice today

We have an ever-expanding understanding of the critical role environmental microbes play in altering our human microbiome and predisposing us to allergic diseases and asthma. The importance of the early childhood period for microbiome development and maturation in protection against allergy seems clear. But actual recommendations, such as living on a farm or owning a dog, remain both understudied in a clinical trial setting and are impractical as interventions. Given that some bacterial species appear beneficial to human microbiome development, attempting to use these as therapeutics in the form of prebiotics, probiotics, or synbiotics seems commonsensical and has been used successfully for treatment of some non-allergic diseases (150, 151). Despite numerous studies on use of pre- and probiotics to treat or prevent allergic diseases, recent reviews highlight the challenges of interpreting this vast literature (152, 153). Largely results have been conflicting, with significant heterogeneity with respect to microbe used as an intervention, mechanism, dose, and timing of administration, and outcome of interest (152, 153). Intervention trials focused on changing the environmental rather than host microbiome will be critical to move the field forward. A small pilot study of an integrated pest-management intervention in elementary school classrooms showed that the intervention changed the classroom microbiome (106). This suggests that environmental interventions can influence the environmental microbiome and should be further studied.

Likewise, timing and severity of early life infection with RV and RSV clearly confer risk for asthma, whereas the role of SA colonization and infection in AD is compelling. Yet, nearly all children are exposed to these microbes in early life, and it remains unclear how to counsel families on their avoidance. Palivizumab, a monoclonal antibody directed against RSV given prophylactically to high-risk children, has been demonstrated to reduce childhood wheezing but does not appear to prevent progression to asthma (154, 155). While much has been learned about the exposures to the environmental microbiome that protect against development of allergic diseases and asthma, significant work needs to be done to understand how we can translate this knowledge into clinical therapeutics or interventions.

Future Directions

The body of literature demonstrating the important contribution of the environmental microbiome to the development of and morbidity associated with allergic disease continues to grow. Future culture-independent studies will need to follow best practices for reproducible research in order to enhance study-to-study comparability and generalizability of results in the field. The environmental microbiome has been demonstrated to vary significantly across geographies and climates (156). Generalizability of environmental microbiome research is additionally limited by the reality that it has largely focused on populations with high burden of allergic disease skewing toward primarily white, affluent populations in industrialized nations, leaving us largely without the insights to be gained from a more equitable distribution of microbiome research (157, 158). Research focused on the environmental microbiome and allergic diseases should focus on representative sampling of populations from both high and low resource settings. Finally, further defining the mechanisms by which the environmental microbiome influences allergic diseases and identification of key interventions that influence the environmental microbiome are needed to translate research into improved outcomes for patients with allergic diseases.

Abbreviations:

AD

Atopic dermatitis; eczema

AR

Allergic rhinitis; allergic rhinoconjunctivitis

AS

Allergic sensitization

FA

Food allergy

qPCR

Quantitative polymerase chain reaction

RV

Rhinovirus

RSV

Respiratory syncytial virus

SA, MRSA, MSSA

(methicillin sensitive/resistant) Staphylococcus aureus

Treg

Regulatory T-cell