Characterizing the ‘normal’ vs ‘abnormal’ esophageal luminal flora
Metagenomic sequencing of the human population revealed that the gastroesophageal (GE) microbiome is predominated roughly in order of prevalence by six major phyla: Firmicutes, Bacteroidetes, Actinobacteria, Proteobacteria, Fusobacteria, and Saccharibacteria. Typically, Bacteroidetes and Firmicutes often predominate-primarily in response to abundance of either Bacteroides or Clostridium spp.. Here, phylum-level classification is an oversimplification that does not account for the diversity that exists in a relatively simple microbiome like that found in the distal esophagus. Considerable variation of both the identity and relative abundance of specific bacteria exists, especially when the taxa are characterized with greater resolution by elucidation of taxa to species or strain-level. In 2009, one group identified two distinct types of GE microbiomes. Comparing the results from individuals with GERD to healthy controls, they characterized the control population, which is predominated by gram-positive organisms, Streptococcus spp., as a type I microbiome. The type II microbiome was largely associated with pathological states, such as GERD and BE, and demonstrated a relative increase in abundance of gram-negative anaerobes.
Further work delineated the taxa and observed that three distinct clusters of esophageal microbiotas were predominant in biopsies of human esophageal tissue. Each is characterized by their relative abundance of Streptococcus and Prevotella spp.. Cluster 1 is intermediate, with an approximately equal proportion of both genera with an increased presence of Haemophilus and Rothia spp. Cluster 2 consists predominantly of Streptococcus spp. Cluster 3 is primarily represented by Prevotella spp..
In addition to inter-individual variation, the composition of luminal microbiota also varies in the esophagus from the mouth to the stomach both in health and disease. Specifically, the commensal flora of the proximal, mid-, and distal esophagus varies both in makeup, and relative abundance. In a study of 12 patients under routine surveillance for BE, the proximal esophagus was more similar to the oropharynx in that it had higher concentrations of gram-positive organisms than the distal esophagus. Streptococcus spp. were found throughout the entirety of the esophagus, increasing in relative abundance from the proximal to mid-esophagus and markedly decreasing thereafter in the distal esophagus. Gram-negative organisms included Prevotella and Delftia spp., which overall were more concentrated in the distal esophagus. This is not surprising, because the lipopolysaccharide (LPS) “shell” around gram-negative organisms hardens them to variation in pH, bile salt concentration, proteases, and to some extent, temperature[8-15]. This is why most enteric pathogens are gram negative—they can survive the selection and potentially adverse effects related to proximity to gastric contents. It would be thereby expected that the microbiome becomes much less diverse and increasingly enriched in gram-negative species. Diversity is greatest in the region nearest to the source (oronasal cavity) and with mildest conditions, e.g., saliva/mucous, luminal physiological pH, and moderate physiological temperatures[16,17]. It would be expected to see a gradient toward facultative and obligate anaerobes (probably from the subgingival space) in the esophageal lumen distally towards the stomach. Notably, many bacteria also have increased abundance because of the protective nature of sporulation (e.g., Clostridia spp.) within the harsh surrounding environment[18,19]. Accordingly, the underlying GE pathology appears to be associated with alterations in the composition of this gram positive/gram negative continuum and balance.
The biomic differences for esophageal disease is notable. Patients with BE had overall higher levels of Streptococcus spp. in tissue biopsies throughout the entirety of the tract compared to those without the disease. While this appears to be in contrast with the gram-positive/gram-negative imbalance discussed previously, this may be an indirect effect of persistent local irritation caused by immunogenic gram-negative species that facilitates bacterial proliferation and the infiltration of underlying tissue with gram-positive bacteria—of which streptococci are a major part. Whether this is a cause of BE metaplasia or a consequence is unknown. Furthermore, the sharper decrease in overall abundance of Streptococcus spp. from mid- to distal-esophagus is greater in individuals with BE compared to those without metaplasia, which suggests that the overall effect on relative composition of Streptococcus spp. is negative despite an increased tissue prevalence.
These findings suggest that an increase in relative abundance of gram-negative flora in the distal esophagus leads to a local inflammatory response which negatively impacts the barrier function. This ultimately leads to tissue proliferation of all flora, including commensal Streptococcus spp. The proximal-distal variation in specific flora changes from healthy to diseased states, in conjunction with the previous GE microbiome subtyping, illustrate the pathological potential of dysbiosis. The emerging data may offer insight into new treatment paradigms focused on microbial alteration which could supplement, and even possibly replace, the need for acid suppression in these disease states.
Factors affecting the composition of the microbiota
The GE microbiome is shaped by the oral cavity, oropharynx, and stomach due to migration of oral bacteria to the esophagus and reflux of gastric microbiota. Recognizably, this varies considerably from person to person, even in the apparently healthy population. In addition to anatomic location, factors that have been noted to alter the EM composition include age, diet, proton pump inhibitor (PPI) use, oral hygiene, and smoking. Studies of these factors have helped provide a framework for understanding the GE microbiota.
Contrary to the philosophy that, for example, the colon is a discrete microbiome that stands alone, we view the whole of the gastrointestinal tract as a contiguous system separated by “gates” imposed by selection pressure driven by factors related to function (pH, osmolarity, proteases, indigenous flora, etc.). That is, a series of discrete “neighborhoods” connected by means of a tube and the assumption that the oronasal cavity if the sole source of inoculum. Prior to weaning, infants do not have established “gates,” and an evolving colonization with what will ultimately become the adult microbiome takes place until the age of approximately three. Thereafter, barring unnatural perturbation, the gastrointestinal neighborhoods are established, stable, and at equilibrium with the host. Unfortunately, humans are pioneers of the unnatural, and a number of behaviors, many now considered essential, systematically undermine the balance. Recognizably, there may be sequential changes associated with age, medication exposures, diet, hygiene, sleep efficiency, and environmental exposures.
Age: Age has an effect on the GE microbiome, although the full significance has not yet been determined[22-25]. During early life, the human colonic microbiome varies greatly. Analysis of the microbial composition of the human colonic microbiome in patients ranging from newborns to 80 years old, and across three distinct geographic locations (United States, Venezuela, and Malawi) has found that the phylogenetic composition fluctuates dramatically during the first 3 years of life before stabilizing into a more stable adult-like composition, regardless of geographic location. Conceivably, a similar dynamic microbial shift exists for the esophagus given the same multifactorial environmental factors in early life, based on mode of delivery (vaginal birth or cesarean section),the type of dietary feeding (breast or formula feeding), as well as the timing of adult food introduction[27,28].
With aging, humans seem to have a less dramatic, but still notable shift in the GE microbiome. Evaluation of the EM of adults of ages 30 years to 60 years, using 16S rRNA-, 18S rRNA-amplicon sequencing, and shotgun sequencing, has found age to be a significant factor driving microbiome composition. Notably, they indicated a positive correlation with age and the relative abundance of Firmicutes such as some Streptococcus spp., including Streptococcus parasanguinis with increasing age. Furthermore, increasing age was inversely correlated with prevalence of Bacteroidetes including Prevotella melaninogenica. To better place this in the context of our current understanding of the GE microbiome and the previously demonstrated community clusters (Streptococcus predominant, Prevotella predominant, and intermediate predominant), this study showed that regardless of disease state, with increased age, there is a more robust microbiome composition and a higher number of gram-positive (Streptococcus parasanguinis) species and a lower number of gram-negative (Prevotella melaninogenica) species. Thus, age may contribute to the different esophageal microbial community types. Despite this, gram-negative proliferation is associated with progression of esophageal disease at all ages. It may be that age may affect and predict a ‘baseline’ microbiome that is incrementally altered by microbial imbalance.
Notably, there is a degenerative effect of aging on esophageal motor function which may play a role in the differences seen in the GE microbiome of the elderly population as esophageal function naturally deteriorates after the age of 40. The presence of GERD has a significant impact on esophageal contraction wave amplitude but not on peristalsis. Accordingly and hypothetically, the mechanistic and functional changes of the esophagus influence the microbiota as a direct or indirect consequence of the aging process.
Diet: Dietary factors influence the colonic microbiome both as an infant (breast vs formula feeding) and as an adult (affecting the colonic microbiome with short-term macronutrient changes)[31-33]. With specific focus on the GE microbiome, dietary intake has been associated with the development of esophageal diseases such as BE, EAC, and esophageal squamous cell carcinoma (ESCC)[34,35]. In particular, consumption of leafy and cruciferous vegetables, as well as raw fruits is associated with decreased risk of BE and EAC, while red meat intake is associated with increased risk.
In early life, breastfeeding, formula feeding and the introduction of solid foods, play a large role in development of the gastrointestinal microbiome. While more specific investigation is needed to evaluate the specific effects on the EM, it is likely that similarities exist in the progression due to the same factors. Human breast milk is predominantly composed of the microbes Corynebacterium, Ralstonia, Staphylococcus, Streptococcus, Serratia, Pseudomonas, Propionibacterium, Sphingomonas, and Bradyrhizobiaceae in addition to milk oligosaccharides[37-39].
The impact of breastfeeding on the infant gastrointestinal microbiome was highlighted in two studies that found formula-fed infants to have a lower proportion of Bifidobacterium and Lactobacillus spp. and a higher proportion of Clostridiales and Proteobacteria when compared with breast-fed infants. Furthermore, formula fed infants have lower microbial diversity after the first year of life when compared to breast-fed infants. Several other epidemiologic studies have suggested breastfeeding to have a protective role against asthma, autism, and type 1 diabetes, while also showing a lower association of inflammatory and autoimmune diseases[37,42]. As stated earlier, the phylogenetic composition fluctuates dramatically during the first 3 years of life before evolving into a more mature and stable adult-like configuration. This shift is likely to allow infants to be better equipped to handle processing of a more robust diet.
In adults, the GE microbiome and the relationship to diet is still under investigation. Our focus here will be on diet and its relationship to esophageal disease as a foundation for possible future studies into the GE microbiome role. There are several difficulties, particularly with confounding and study-design issues, when correlating dietary factors in adults with esophageal disease. Thus, most of the existing literature on diet and BE or EAC is based on case-control studies in which minor to moderate inverse associations have been reported with a diet low in fruits and vegetables (green, leafy, and cruciferous vegetables). It has been theorized that fruits and vegetables, which are high in antioxidants, phytosterols, and other substances, may inhibit carcinogenesis by free-radical reduction or by blocking the formation of N-nitroso compounds in the alimentary and respiratory tract[43,44]. Other case-control studies have shown an association with a diet high in red and processed meats and an increased risk of esophageal cancers, likely due to processed meats being a major source of nitrites and nitrosamines[35,45]. Given the potential for multiple interactions between specific macronutrients, other studies have turned to looking at diet-regimens for easier study design. They found that the Mediterranean diet is inversely associated with both BE and EAC, whereas the “Western diet”, high in meat consumption and low in fruits and vegetables, appears to increase the risk of these diseases[46,47]. Although this relationship between dietary intake of fruits, vegetables, as well as red and processed meats, has been more recently implicated, further evaluation is needed in regard to the interaction of these diets and the GE microbiome composition.
A study specifically looking at the relationship of diet with the GE microbiome evaluated patients with high overall fiber vs high fat intake and found that dietary fiber, but not fat intake was associated with a distinct EM. In particular, increasing fiber intake was significantly associated with increasing relative abundance of Firmicutes, including Streptococcus spp., and decreasing relative abundance of gram-negative bacteria overall. Low fiber intake was associated with increased relative abundance of several gram-negative flora, including Prevotella, Neisseria, and Eikenella spp.. These findings offer a potential dietary therapeutic option for prevention or slowing the progression of esophageal disease by decreasing exposure to a higher abundance of gram-negative influence, and thereby reducing the induction of a gram-negative-LPS induced inflammatory cascade.
PPIs: PPIs are the therapeutic first-line treatment for many esophageal disorders such as GERD, erosive esophagitis, and BE. The main mechanistic action of PPIs is to lower acid production at the level of the stomach by inhibiting the hydrogen-potassium ATPase pump, a transmembrane protein responsible for releasing hydrochloric acid into the stomach lumen. PPIs inhibit acid secretion by binding within this domain, promoting a higher gastric pH, and thus increasing the pH of the refluxate.
The use of PPIs has been demonstrated to alter both GE and colonic microbiomes, although the full extent is yet unknown. The clearest defined role is reduction of gastric acid, thereby allowing survival of orally ingested organisms to populate the more distal esophagus. This pH-related microflora change may allow propagation of bacterial species that would otherwise not flourish under more acidic conditions. For example, a significant increase in oral microbiome species such as Rothia dentocariosa, Rothia mucilaginosa, Scardovia spp., and Actinomyces spp. in the gut microbiome has been noted following PPI use[50-52].
In the distal esophagus, the effect of PPIs may be more likely to be due to microbial related inflammatory changes, whereas previously attributed to direct acid contact mucosal injury. A study of patients with non-erosive reflux disease (NERD), erosive GERD, and BE compared PPI use vs no use within each respective group and found no change in α diversity or β diversity between PPI and non-PPI users of each group was reported, but composition of specific bacteria taxa at the phylum level was noted. In particular, PPI use was associated with an increase in Firmicutes and Proteobacteria in BE, and a decrease in Bacteroidetes in NERD and reflux esophagitis (RE). In another study, biopsies taken before and after 8 wk of PPI treatment (lansoprazole 30 mg twice daily) revealed a significant decrease in the gram-negative Comamonadaceae spp. and increased gram-positive Clostridia (Clostridiaceae and Lachnospiraceae spp.) and Actinomycetales (Micrococcaceae and Actinomycetaceae spp.).
These studies offer evidence that PPI use may have effects beyond that of acid suppression. This supports a possible mechanistic role for PPIs altering the GE microbiome, favoring gram-positive bacteria that prefer environments with higher pH. This effect would reduce induction of the Toll-like-receptor (TLR)/inflammatory cascade by gram-negative LPS producing bacteria.
Although their association with GERD is unknown, acid-producing bacteria are found in the esophagus and oral cavity. The use of PPIs may directly target the proton pumps (P-type ATPase enzymes) of these bacteria (notably Streptococcus pneumoniae and Helicobacter pylori). Further studies are warranted to determine if these bacteria are a causal factor of GERD by directly producing acid, which are in turn inhibited by PPIs. In addition, PPI use may indirectly change the natural bacterial flora in non-gastric tissues that express H+/K+-ATPases by shutting down proton pumps[55,56].
PPIs may also reduce inflammation apart from direct acid suppression. In esophageal squamous epithelial cells, omeprazole has been shown to inhibit interleukin (IL)-8 expression by blocking the nuclear translocation of a nuclear factor-kappa beta (NF-kB) subunit and the binding of AP-1 subunits to the IL-8 promoter. IL-8 is an inflammatory mediator that has been implicated in the GERD, BE, and EAC pathways. Increased expression of LPS from gram-negative bacteria, and subsequent activation of the TLR-4-NF-κB pathway are associated with expression of downstream mediators such as IL-8 and cyclooxygenase (COX)-2. The levels of both are directly correlated with transition from metaplasia to dysplasia in BE. Thus, if PPI therapy has an effect on IL-8 expression by blocking NF-kb and AP-1, there may be a role for therapeutic use outside of direct acid suppression.
Oral hygiene: Oral hygiene is thought to play a vital role in the GE microbiome. Bacteria found in the oral cavity can migrate distally via deglutition. The role of oral microbiota in colonizing the esophagus, and becoming part of the commensal GE microbiota, remains uncertain.
Maintained oral hygiene is associated with a higher proportion of gram-positive cocci and rods, mostly comprised of Streptococcus spp., which contrasts with those with poor oral hygiene showing shifts to a higher proportion of anaerobic gram-negative bacteria such as Prevotella spp.. The oral microbiome shift to a more gram-negative dominant flora may have distal effects of LPS-inducing TLRs and activation of an inflammatory cascade in the esophagus as described previously. It is also unclear whether antibiotic mouthwashes damage an otherwise healthy microbiome. Further research is needed to better define the relationship between oral hygiene and the GE microbiome.
One recent population-based, case-control study reported that poor oral health was associated with an increased risk of ESCC. More specifically, they found that tooth loss was associated with a moderately significant increased risk of esophageal cancer and that brushing once per day or less was associated with an 80% increased risk of developing ESCC in this population. They propose that tooth brushing influences the balance of microorganisms by directly removing plaque, food residue, and carcinogenic products of tobacco and alcohol. Accordingly, this affects the levels of inflammation and/or production of the carcinogenic by-products of nitrosamines and acetaldehyde. While this is currently theoretical, given the proximity of the oral cavity to the distal esophagus, it is reasonable that oral hygiene would have downstream effects on the GE microbiome and esophageal disease as well as perhaps on intestinal microbial mediated disease as well.
Smoking: Up to 20% of United States adults use a tobacco product, and tobacco’s effect on the GE microbiome is uncertain. Esophageal balloon procured cytology found that current smoking was associated with an increase in both α and β diversity of the esophagus. It is suspected that the increase is due to smoking-related immunosuppression, permitting novel bacteria to colonize the upper gastrointestinal tract. The study also found two anaerobic bacteria, Dialister invisus and Megasphaera micronuciformis, are more commonly detected in current smokers. Increased α and β diversity after smoking exposure may also be a result of biofilm formation. There is some evidence that cigarette smoking induces staphylococcal biofilm formation in an oxidant-dependent manner by increasing fibronectin binding protein-A. This leads to increased binding of staphylococci to fibronectin and increased adherence to human cells.
Smoking exposure can affect a wide range of human physiologic processes by inducing a proinflammatory state, increasing cytokines such as tumor necrosis factor (TNF)-α, IL-1, IL-6, IL-8 all while decreasing anti-inflammatory cytokines such as IL-10. Further investigation into the EM and its relationship to smoking and the development of disease are needed.