Introduction and Overview

The advent of next-generation sequencing has amplified our ability to assess the microbiome of different body niches as well as their alterations in pathological conditions. While the oral cavity and the lung are often viewed as distinct clinical entities, they are part of a continuum. This continuity is reflected in their microbiomes. Emerging evidence suggests that dysbiosis of the oral cavity is at the very least associated with and may impact the progression of several lung pathologies. The intent of this review is to summarize the studies that show this relationship. Data is summarized in sections separated by pathology, following a summary of factors influencing the microbiome of the oral cavity. Table 1 provides an “at-a-glance” summary as well. We hope that this summary will be of use to the increasing number of investigators interested in this burgeoning field.

Table 1 Summary of studies evaluating oral microbiota alone or in comparison with the lung microbiota in healthy and diseased lungs
Full size table

Microbiome of the Oral Cavity

As a system open to the environment, the oral cavity has a unique microbiome as it consists of mucosal surfaces that shed constantly as well as non-shedding surfaces of teeth. The oral microbiome is composed of a diverse ecological community of commensal, symbiotic, and pathogenic microorganisms that share the human oral cavity. The ecological niche consists of 5 distinct areas: teeth, saliva, tongue, gingival sulcus and periodontal pocket, and the remaining oral mucosa [1, 2]. The abundant oral flora could not be fully identified previously because not all are cultivatable. But this changed with the advent of 16s rRNA polymerase chain reaction (PCR) gene sequencing. The Human Oral Microbiome Database (HOMD) has a repository of bacterial genome sequences of the oral cavity containing about 800 species [3]. More recently, the expanded HOMD (eHOMD) was developed to include microbiome databases of the human aerodigestive tract as well [4]. The microbiome of saliva comes from bacteria shed from biofilms on oral tissues and does not have its own indigenous microbiota [5]. The most common microbiota of the oral cavity are Streptococcus, Lactobacillus, and Prevotella [2]. In the periodontal pocket, a shift in the abundance of low-abundance species has led to the “dysbiosis hypothesis” theorized to be a cause of periodontitis [6]. P. gingivalis has been suspected to be one of the key bacterial species underlying periodontitis. The “keystone” pathogen hypothesis describes the effect of a low-abundance microbial pathogen such as P. gingivalis that exerts a disproportionately large effect on their communities [6, 7]. In the oral cavity, periodontitis is linked to P. gingivalis, which evolved to evade or circumvent the host immune system that triggers a destructive change in the normally homeostatic host-microbial interplay. In this manner, P. gingivalis acts as a keystone pathogen [6].

Factors Influencing the Oral Microbiota

The Lung HIV Microbiome Project studied the relationship between oral and lung microbiomes in HIV patients using bacterial 16S rRNA sequencing to compare the operational taxonomic units (OTUs) [8] between the two locations. Since HIV patients have impaired host defense, they are more vulnerable to infections, which may be reflected in their microbiota. Although the oral microbial populations were different in HIV-infected compared with HIV-uninfected patients, their bronchoalveolar lavage (BAL) microbial populations were not significantly different. CD4 cell counts did not correlate with the oral and lung microbiome on further analysis. Their lung microbiomes were mostly derived from their oral microbiome except for some unique bacterial such as Tropheryma whipplei. The use of antiretroviral therapy was associated with a reduced relative abundance of Tropheryma whipplei in the lungs of HIV-infected patients.

Ethnic background can determine salivary microbiota and can be altered with smoking [9, 10]. Using the same cohort of HIV-uninfected patients from the Lung HIV Microbiome Project, oral microbiota was found to be different between smokers and nonsmokers but lung microbiota was not significantly altered by smoking [11]. Unfortunately, the author did not provide an explanation for this lack of difference. A study into the temporal shift in the oral microbiome found that communities remained stable over time in healthy subjects but community diversity varied between individuals [12].

Relationship Between Oral and Lung Microbiota

A study of microaspiration in healthy subjects using BAL samples showed that lower airway samples that are enriched with oral taxa (Prevotella or Veillonella) are associated with increased numbers of lymphocytes and neutrophils [13]. These taxa are also associated with a Th 17 lung inflammation phenotype [14]. Oral-derived microbiota such as Prevotella spp. is responsible for the regulation of pulmonary inflammatory responses shown by IL-17A in a mouse model [15]. In healthy subjects, as expected, the microbial biomass decreases from the oral cavity to the lungs but studies showed conflicting findings of whether oral and lung microbiome are similar or different. The topological continuity theory asserts that the respiratory tract from the nasal and oral cavities to the upper and lower airways are contiguous and the microbiota is indistinguishable between them [16]. This theory is supported by Charlson et al. who showed that the oral and lung microbiome are homogenous as the lungs contain bacterial sequences largely indistinguishable from the upper respiratory flora [17]. The countervailing theory is the island biogeography theory that describes the different human anatomic locations as different “islands” of habitation that differ in time and location [16]. A study by Bassis et al. showed that the oral and lung microbiome are found to be different in only half of the patients. Microaspiration was suggested as the cause of the similarity in the remaining half of the patients [18]. Risk of contamination by oral bacteria of the bronchoscope during sampling of BAL may also explain the differences. Since the studies evaluating oral and lung microbiota use bronchoscopy and BAL, evaluation of potential contamination was made by comparing the bacterial sequence and load of the serial BAL. Both authors were confident that carry over of oral contaminant was not an issue. Yu et al. overcame the potential oral contamination by comparing oral microbiota with lung microbiota (of normal lung tissue) obtained via surgical resection for lung cancer, which showed that the lung microbiome is different from the oral microbiome [19••]. Although the cohort of patients in the Yu et al. study had lung cancer, their non-malignant tissues were similar between patients and had greater phylogenetic diversity compared with malignant tissue which had low phylogenetic diversity and also showed different microbiota in different cancer histological diagnoses (adenocarcinoma vs squamous cell carcinoma).

Dysbiosis in Acute Respiratory Tract Infections

Poor oral hygiene has been linked to respiratory tract infection [20, 21]. Dental plaque, which is a biofilm on tooth surfaces, has been identified as a reservoir of bacteria causing pneumonia [22,23,24]. Accumulation of dental plaque increased with ICU stay duration, which also increased the likelihood of colonization by aerobic pathogens (gram-negative rods first) [24]. This then led to nosocomial infections. Critical illness may allow more rapid dental plaque formation since these patients tend to have diminished salivation (xerostomia) and salivary pH [25]. Immigration of food-associated bacteria is reduced in critical illness when catabolic starvation state predominates due to reduced nutritional supply to commensal bacteria [26, 27]. Regular oral intake involves the ingestion of hard and fibrous food. This and the movement of tongue and cheeks during speech are absent in patients with critical illness especially if the patient is intubated [28]. Xerostomia due to stress of critical illness is worsened by lines and tubes traversing the oral airways which leaves the mouth open as well as medication that dries up secretions. The natural distribution of salivary immune factors such as IgA and lactoferrin is compromised in the setting of xerostomia [29]. The predominant bacterial species shift from gram-positive to gram-negative in the critical illness state [27, 30]. The alteration in carbohydrates in buccal cells during critical illness has been shown to promote adherence of pathogenic bacteria to epithelial cells [31]. Benign Prevotella spp. and Veillonella spp. population [32] are displaced by potentially pathogenic bacteria such as P. aeruginosa and K. pneumonia [30, 33]. Colonization of the oral cavity was found to be greater in patients with teeth or wearing dentures (73%) compared with that in edentulous patients (37.5%) [34]. This suggests that non-shedding surfaces (teeth and dentures) favor bacterial colonization greater than shedding surfaces (mucosa). This was confirmed by more rigorous studies of respiratory pathogens isolated from the oral cavity of patients with pneumonia which matched the strains found in the lung via bronchoalveolar lavage [35, 36]. Using bacterial floral analysis of 16s rRNA gene among patients with pneumonia with aspiration risk, oral streptococci were the most common bacterial phylotypes detected [37]. Another 16s rRNA gene PCR amplification study showed that 88% of ventilator-associated pneumonia (VAP) patients had overlapping pathogens in the oral cavity and the lungs and identification of new putative uncultivatable and unreported species in 56% of patients [38]. The dorsum of the tongue was suggested as a potential reservoir of bacteria for VAP.

Several distinct pathogeneses of oral microbiota linked to pneumonia have been described [39, 40]. Oral pathogens may cause pneumonia via (1) aspiration of oral pathogens; (2) modification of lung mucosal surfaces by aspirated periodontal disease-associated enzymes and cytokines allowing adhesion and colonization by pathogens; (3) destruction of salivary pellicles on pathogenic bacteria by periodontal disease–associated enzymes; (4) airborne translocation; and (5) systemic bacteremia from periodontal infections. Mucosal alteration of epithelial cells in the form of loss of fibronectin, which functions to promote bacterial adhesion, occurs in P. aeruginosa colonization. Due to the action of proteases, the loss of fibronectin leads to unmasking of mucosal surfaces for respiratory pathogen adhesins. Oral bacteria can also destroy salivary pellicles through the action of sialidase on sialic acid residuals. This decreases the ability of mucins in the saliva to clear pathogens such as H. influenza. Cytokine release due to oral bacteria from the gingival crevice can stimulate the respiratory cells to produce other cytokines that recruit inflammatory cells that release hydrolytic enzymes. These enzymes can damage the respiratory epithelium leading to increase susceptibility to colonization of pathogens. In support of the relevance of these findings, improving oral hygiene and cleansing in at risk patients (ICU patients and elderly) have been advocated to prevent the occurrence of aspiration pneumonia, hospital-acquired pneumonia (HAP), and VAP. The use of chlorhexidine mouthwash reduces the risk of developing VAP but did not show mortality benefit [41, 42]. However, all results examining this issue do not show consensus in the conclusions. A more recent meta-analysis on the use of chlorohexidine mouthwash found an increase in mortality [43] although the reason was unclear. Clearly, further work needs to be done in this area.

Dysbiosis in Cystic Fibrosis

Cystic fibrosis (CF) lung disease is characterized by chronic colonization of bacteria such as Haemophilus influenzae, Staphylococcus aureus, and Pseudomonas aeruginosa. The oral-lung axis is thought to be bidirectional in CF. A case-controlled study showed genetic relatedness between subgingival plaque and lung P. aeruginosa [44••]. The oral cavity is suggested as a potential reservoir of P. aeruginosa allowing for initial colonization and subsequent recolonization of P. aeruginosa in CF patients. The ascent of pulmonary bacteria can also replete the oral bacterial reservoir. The dorsum of the tongue was found to be the most common location for P. aeruginosa colonization in the oral cavity [45]. Although normal healthy lungs have coordinated mucociliary clearance to limit bacteria migration into the lungs, CF patients are particularly vulnerable to bacterial migration into the lungs due to dysfunctional cilia caused by dense mucus [46] and impaired alveolar macrophages and autophagy [47, 48]. The use of sputum to reflect the lung microbiome was studied by Hogan et al. who showed that the most abundant pathogen in sputum reflects the predominant taxa identified from protected brushing of the lung [49]. Nonetheless, the diversity of lung microbiome is lower in the lung compared with sputum especially in advance CF [49,50,51]. Toothbrushes were found to be a potential reservoir of CF-associated bacteria such as P. aeruginosa, S. aureus, S. maltophilia, A. xylosoxidans, and S. marcescens [52]. CF bacteria on the toothbrush’s bristles can be introduced from the environment into the patient or be a source of re-inoculation to the oral cavity of the host.

An interesting mechanistic basis for the influence of the oral microbiome on the lung microbiome was proposed in cystic fibrosis patients. Oral metabolites can travel passively to the lungs and affect the lung microbiome in CF patients through cross-feeding. 2,3-Butanedione is a byproduct of alternative fermentation pathway which produces a neutral pH and avoids the lethal acidification of low pH fermentation. The 2,3-butanedione gas likely produced by oral Streptococcus spp. is volatile and easily travels through the airways into the lungs which acts as a substrate for phenazine production by P. aeruginosa in CF lungs [53]. Phenazines are redox-active pigments that can serve as alternative electron acceptor for metabolism in hypoxic biofilm subregions [54]. Through proteome analysis 2,3-butanedione production was linked to biofilm production and increases the virulence factor of P. aeruginosa [55].

Dysbiosis in COPD

Tooth loss, periodontal disease, poor dental care, and lack of oral health knowledge were found to be associated with higher risk of having COPD [56,57,58]. Worse dental hygiene was associated with more respiratory symptoms in COPD and greater number of teeth has a positive correlation with more respiratory symptoms and sputum production [59]. As described in the pneumonia section above, the presence of non-shedding surfaces such as teeth and dentures favors bacterial colonization [34]. Periodontitis and COPD are hypothesized to be linked because both have similar pathophysiology in terms of elevated circulating inflammatory cytokines and mediators such as C-reactive protein, interleukin-8, tumor necrosis factor-α, and matrix metalloproteinase [60, 61]. Periodontitis can act as an inflammatory reservoir [62]. The cytokines from a local inflammatory response of periodontitis can spill into the systemic circulation with subsequent inflammatory damage to distal organ such as the lung. Neutrophilic inflammation is characteristic of both COPD and periodontitis [60]. Nevertheless, the causal relationship between periodontal disease and COPD can be due to the confounding effect of smoking [63]. Smoking is a major risk factor for periodontal disease [64] and is the main cause of COPD.

The most current understanding of bacterial acute exacerbation of COPD (AECOPD) comes from study of the lung microbiota, which showed that acquisition of new strain of Haemophilus influenzae, Moraxella catarrhalis, or Streptococcus pneumonia is strongly associated with AECOPD [65, 66]. A study comparing the oral and sputum microbiota (representing the lung microbiota) showed that frequent exacerbators (≥ 1 exacerbations per year) had a lower alpha diversity in their sputum microbiota than sputum microbiota of infrequent exacerbators [67]. Nonetheless, oral wash samples did not differ between frequent and infrequent exacerbators. PERMANOVA analyses found clustering of microbiota based on oral hygiene status, COPD severity, anatomic site, inhaler corticosteroid use, and smoking. Investigation of the potential use of oropharyngeal swab as a surrogate for sputum in AECOPD found that oropharyngeal swab and sputum had similar microbiota composition but oropharyngeal samples had higher diversity [68]. With the use of deep sequencing, the investigation Wang et al. on the lung microbiome in COPD showed clustering of phyla to Proteobacteria, Firmicutes, and Bacteroidetes subgroups [69]. During AECOPD, there was overall reduction in microbial α diversity and increase in relative abundance of Proteobacteria and decrease in Firmicutes. Wang et al. also showed that corticosteroid treatment decreased microbial α diversity with an increase of Proteobacteria over Firmicutes. On the other hand, treatment with antibiotics created an opposite trend.

Dysbiosis in Lung Cancer

Although an association of oral microbiome and cancer risk has been found in pancreatic cancers [70], our understanding of the impact of the oral microbiome on lung cancer pathophysiology remains limited. Yan et al. were the first to demonstrate the association between salivary microbiota with lung cancer [71]. Using 16s sequencing, levels of Capnocytophaga and Veillonella were significantly higher in patients with squamous cell cancer and adenocarcinoma suggesting their levels as potential use as biomarkers for disease detection or classification. The AUC of ROC of Veillonella for squamous cell cancer and adenocarcinoma were 0.81 and 0.68, respectively, and the AUC of Capnocytophaga for squamous cell cancer and adenocarcinoma was 0.79 and 0.81, respectively. In a different study, excluding the effect of smoking, salivary microbiome among 75 non-smoking female patients with lung cancer compared with 172 matched healthy control found decreased microbial diversity and occurrence of dysbiosis in the lung cancer group [72]. Bacterial genera Blastomonas and Sphingomonas were found to be significantly higher in the oral microbiota of the lung cancer group while Acinetobacter and Streptococcus were higher in the control group. The study also found a positive correlation between immunocytochemistry markers TTF-1 and CK 7 with Enterobacteriaceae, and Napsin A with genus Blastomonas. A study of salivary dysbiosis showed that genera Veillonella and Streptococcus were strongly increased in NSCLC compared with controls [73]. The UniFrac distance was significantly different between the groups on principal coordinates analysis. It also showed cross links among salivary microbiota dysbiosis, systemic inflammatory markers, and predicted Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.

Dysbiosis in Lung Transplant

Oral pharyngeal microbiota in patients undergoing lung transplant showed severe dysbiosis in taxonomic composition and respiration phenotypes with reduced richness and diversity and increased facultative and reduced aerobic bacteria in the pre-transplant stage in the setting of their end-stage lung disease [74]. In 6 weeks to 3 months post-transplant, the richness and diversity were intermediate between healthy and pre-transplant patients. By 6 months, the post-transplant patients’ oral pharyngeal microbiota resembled that of pre-transplant patients. All post-transplant patients were on antimicrobial and immunosuppressive therapy, which may have affected the patients’ microbiota. However, analysis of pre-transplant patients on these agents suggested that the dysbiosis is not driven by antimicrobial nor immunosuppressive therapy.

Conclusion

With the improvement in PCR sequencing tools available to investigators, we have a greater understanding of the oral microbiome diversity and its systemic effect especially on the lungs. The diversity of microbiota decreases as we descend from the oral cavity to the lungs. Dysbiosis of the oral microbiota is linked to oral infections and a number of lung diseases especially pneumonia, CF, COPD, and lung cancer. The oral cavity was found to be a reservoir of bacteria causing disease in the acute conditions (aspiration pneumonia, HAP, and VAP) and chronic disease (CF). In lung cancer, changes in oral microbiota can be a biomarker of disease. Future studies should explore the use of oral microbiota dysbiosis as biomarker of disease and the manipulation of oral microbiota therapeutically to change lung disease progression.