Jing Lv, Lei Guo, Ji-Jun Liu, He-Ping Zhao, Jun Zhang, Ji-Han Wang

Abstrac t The incidence of esophageal adenocarcinoma (ΕAC) has increased in recent decades, and its 5-year survival rate is less than 20%. As a well-established precursor, patients with Barrett's esophagus (BΕ) have a persistent risk of progression to ΕAC. Many researchers have already identified some factors that may contribute to the development of BΕ and ΕAC, and the identified risks include gastroesophageal reflux (GΕR), male sex, older age, central obesity,tobacco smoking, Helicobacter pylori (H. pylori) eradication, and the administration of proton pump inhibitors (PPIs) and antibiotics. The human gut harbors trillions of microorganisms, the majority of which are bacteria. These microorganisms benefit the human host in many ways, such as helping in digestion, assisting in the synthesis of certain vitamins, promoting the development of the gastrointestinal immune system, regulating metabolism and preventing invasion by specific pathogens. In contrast, microbial dysbiosis may play important roles in various diseases, such as inflammation and cancers. The composition of the microbiota located in the normal esophagus is relatively conserved without distinct microbial preferences in the upper, middle and lower esophagus. Six major phyla constitute the esophageal microbiota, including Firmicutes,Bacteroides, Actinobacteria, Proteobacteria, Fusobacteria and TM 7, similar to the oral microbiota. Streptococcus dominates the esophageal microbiota. However, the microbiota varies in different esophageal diseases compared to that in the healthy esophagus. The type I microbiota, which is primarily composed of gram-positive bacteria, is closely associated with the normal esophagus, while type II microbiota has enriched gram-negative bacteria and is mainly associated with the abnormal esophagus. These increased gram-negative anaerobes/microaerophiles include Veillonella, Prevotella, Haemophilus, Neisseria, Granulicatella and Fusobacterium, many of which are associated with BΕ. The microbial diversity in the esophagus is decreased in ΕAC patients, and Lactobacillus fermentum is enriched compared to that in controls and BΕ patients. Furthermore, the microbiota may be associated with BΕ and ΕAC by interacting with their risk factors, including central obesity, GΕR, H. pylori, administration of PPIs and antibiotics. Therefore, a large gap in research must be bridged to elucidate the associations among these factors. Some studies have already proposed several potential mechanisms by which the microbiota participates in human carcinogenesis by complicated interactions with the human host immune system and signaling pathways. The activation of the LPS-TLR4-NF-κB pathway may contribute to inflammation and malignant transformation. This exciting field of gastrointestinal microbiota allows us to unravel the mystery of carcinogenesis from another perspective. Further studies are needed to explore whether the microbiota changes before or after disease onset, to improve our understanding of the pathogenesis, and to find novel targets for prevention, diagnosis and therapy, which could offer more cost-effective and relatively safe choices.

Key words: Barrett's esophagus; esophageal adenocarcinoma; microorganisms; esophageal microbiota; alteration; dysbiosis manuscript

INTRODUCTION

Εsophageal cancer is one of the most common malignancies in the world, and the incidence of esophageal adenocarcinoma (ΕAC) has markedly increased in recent decades, as it accounts for almost half of all esophageal cancers[1,2]. The 5-year survival rate is less than 20%[3]because most ΕAC patients are first diagnosed in the advanced stages, which are not curable[4]. As a well-established risk factor for ΕAC, Barrett's esophagus (BΕ) confers a persistent risk of progression to ΕAC[5], increasing a patient's risk more than 30 times than the general population[6,7]. In the first global definition,namely, the Montreal d efinition, BΕ is d efined as the replacement of normal squamous epithelial lining w ith metaplastic columnar epithelium[8]. Notably, the incidence of ΕAC progression from BΕ varies among studies. Some studies have indicated that more than 85% of newly diagnosed ΕAC patients have no history of BΕ[9,10], whereas other investigations[6,7]have reported that almost half of ΕAC patients have progressed from BΕ. Nevertheless, BΕ is the only well-recognized precursor of ΕAC, and the underlying mechanisms of pathogenesis and carcinogenesis need to be elucidated[11]. Many researchers have already shown that some factors may contribute to the development of BΕ and ΕAC, and the identified risks include, but are not limited to, gastroesophageal reflux (GΕR), male sex, older age, central obesity, tobacco smoking, Helicobacter pylori (H. pylori) erad ication w ith antibiotics and acid suppression therapy[3,12-14]. However, preventive strategies are lacking.

The human gut harbors trillions of microorganisms[15-18], and the composition of the microbial communities that inhabit the mouth, esophagus, stomach and intestine are d iverse and host-sp ecific. The majority of microorganisms are bacteria and are estimated to comp rise ～1014bacterial cells, w hich is ten times more than the total number of human cells[19]. These microorganisms benefit the human host in many w ays, such as help ing in d igestion, assisting in the synthesis of certain vitamins,p romoting the d evelop ment of the gastrointestinal immune system, regulating metabolism and preventing invasion by specific pathogens[15,16,19,20]. On the other hand,microbial dysbiosis may lead to tissue d amage and play significant roles in various diseases, includ ing inflammatory d isorders and cancers[21-25]. Dysbiosis refers to an abnormal cond ition of the microbial ecosystem in a host[26]. Therefore, equilibrium must be achieved and maintained to support the interactions of the human host and microbiota. Before the 1990s, researchers mainly focused on the role of certain microorganisms by using protocols largely limited to culture-dependent methods, but cultivation is not suitable for d efining a complicated microbial community and may induce bias as well[19]. Since the development of next-generation sequencing, which is ind ep end ent of cultivation, the sensitivity of research techniques has been dramatically improved, and microbiota exploration has begun[20].

How ever, little is know n about the relationship betw een the microbiota and the pathogenesis of BΕ and ΕAC. Here, w e review the features of microbial communities in BΕ and ΕAC p atients, w hich may provid e some evid ence of the relationship s between altered esophageal microbiota and BΕ/ΕAC.

ESOPHAGEAL MICROBIOTA

The colon has the largest microbiota in the body[20], whereas the esophagus has far fewer microbes. Although bacterial communities have been observed with high interand intra-individual variations, an overlapping community has been identified between sites[27,28]. Previous studies[20,29-31]have suggested that Firmicutes (Clostridium,Ruminococcus, Eubacterium, Peptostreptococcus, Peptococcus, Lactobacillus-L), Bacteroidetes,Proteobacteria (Enterobacteriaceae) and Actinobacteria (Bifidobacterium-BF) phyla constitute the majority of the human gut microbiota. The composition of the microbiota located in the normal esophagus is relatively conserved, and the estimated resident microbes are mainly composed of more than 100 species, most of which have alread y been identified[32,33]. The dominant microbes that colonize the normal esophagus are Streptococcus[32,33], and Dong L et al.[28]reported that no distinct microbial preference exists in the upper, middle and lower esophagus. However, the microbiota varies in different esophageal diseases compared to the healthy esophagus[34].

In 2004, Pei Z et al.[32]found that six major phyla constituted the esophageal microbiota, namely, Firmicutes, Bacteroides, Actinobacteria, Proteobacteria, Fusobacteria and TM 7, which are comparable to the oral microbiota. The genus Streptococcus dominates the esophageal microbiota[32]. Since then, more studies have emerged, and a classification for the esophageal microbiota was proposed[33]. In 2009, Yang L et al.[33]reported that type I microbiota, which is mainly composed of gram-positive bacteria,is closely associated with the normal esophagus and is dominated by the Firmicutes phylum. Consistent w ith previous studies, Streptococcus was the most dominant genus, and its relative abundance was higher. The type II microbiota is enriched in gram-negative bacteria (more than 50%) and is mainly associated with the abnormal esophagus. The relative abundances of 24 other genera are increased in the type II microbiota, many of w hich are relevant to BΕ. These increased gram-negative anaerobes/microaerophiles include Veillonella, Prevotella, Haemophilus, Neisseria,Granulicatella and Fusobacterium. Moreover, one study[19]conducted in Japan showed that the numbers of bacteria were similar in control, esophagitis and BΕ groups,despite the changes in the relative abundance of taxa. In patients with esophagitis or BΕ, the microbial diversity changed, and the abundance of Streptococcus species was red uced[33]. Gram-negativ e anaerobes/microaerop hiles occup ied greater proportions[33], such as Veillonella, Prevotella, Fusobacterium and Neisseria[19,33]. This shift from a gram-positive aerobic microbiota to a gram-negative anaerobic microbiota may be influenced by microenvironmental changes and related to abnormal disease states[33]. These consistent observations suggested that the altered microbiota is reliable in BΕ and could be further studied. Of note, Macfarlane S et al.[35]found that Campylobacter colonized the esophagus of the majority of BΕ patients and could not be identified in the control group. Moreover, Amir I et al.[36]strongly suggested that the family Enterobacteriaceae (mainly the genus Escherichia) is associated with esophageal abnormalities, such as esophagitis and BΕ, and may have a possible role in the pathogenesis of inflammation and metaplasia. Therefore, a large-scale joint multicenter, multi-region, multi-race stud y about the alteration of the esop hageal microbiota in BΕ/ΕAC is needed to provide more evidence.

Damage of the esophageal epithelium could affect the normal barrier and ind uce the translocation of other bacteria, thus influencing the microenvironment and immune homeostasis[11]. Although studies of specific bacteria in the development of ΕAC in add ition to H. pylori[3]are rare, more attention has been p aid to the local microbiota changes. The microbial d iversity in the esophagus is d ecreased in ΕAC patients[12], regardless of the exact sampling locations. The decreased genera included some gram-negative and gram-positive taxa, such as Veillonella and Granulicatella. In contrast, Lactobacillus fermentum w as found to be enriched in ΕAC patients compared to controls and BΕ patients. Notably, lactic acid bacteria could be dominant and affect the microenvironment[12]. A low microenvironmental p H may facilitate the growth of Lactobacillus sp p. and Streptococcus spp. in the tumor niche[12]. Fermentation could produce more factors to inhibit the proliferation of other competitor microbes as well.Then, Lactobacillus may dominate the environment of the lower esophagus. Moreover,some specific species were d emonstrated to have higher abundance. At the phylum level, the p roportional abundance of Tenericutes w as higher. At the genus level, the p rop ortional abu nd ances of Fusobacterium, Megasphaera, Campylobacter,Capnocytophaga, and Dialister w ere greater[12]. How ever, Blackett KL et al.[37]d id not identify any specific taxa w ith significant differences, and Zaidi AH et al.[38]reported that Streptococcus pneumonia was present at a relatively higher abundance in control and BΕ groups compared to ΕAC in rat BΕ and ΕAC models. Interestingly, Peters BA et al.[39]found that the oral microbiota composition could reflect the prospective risk for ΕAC, and the genus Neisseria and the sp ecies Streptococcus pneumoniae w ere associated with ΕAC risk, which is consistent w ith the findings in the studies above. It w as reasonable to conclude that the esophageal microbiota is largely influenced by the oral microbiota and that the oral microbiota comp osition could provid e some evidence of ΕAC progression[17].

Furthermore, the microbiota may be associated w ith BΕ and ΕAC by interacting w ith their risk factors. One notable example is the case of obesity. As a chronic systemic d isease and a p roposed risk factor, obesity, particularly central obesity, is closely related to BΕ and ΕAC[20,40-42]. The linear pattern betw een increasing body mass ind ex (BMI) and increasing risks of BΕ and ΕAC has been verified in several stud ies[43-46], w hich partially accounts for the increasing prevalence of ΕAC. Central obesity is closely related to ΕAC, even after ad justment for BMI[47,48], w hereas the association betw een BMI and ΕAC risk d isapp eared after ad justment for central obesity. Moreover, the relationship betw een central obesity and BΕ has a similar pattern. Therefore, adiposity distribution may play an important role in BΕ and ΕAC p athogenesis. How ever, it is unclear w hether w eight loss could contribute to a red uced risk of BΕ and ΕAC. The possible mechanisms by w hich central obesity contributes to BΕ and ΕAC have been explored and discussed in several aspects. First,the increased abdominal ad ipose tissue might increase intra-abdominal pressure and gastric compression, disrupting the normal function of the gastroesophageal junction and promoting GΕR, w hich is also a w ell-recognized risk factor for BΕ and ΕAC[3].Second, excess ad ip ose tissue could secrete p ro-inflammatory cytokines and ad ipokines[20], and these active factors could provoke inflammatory and metabolic changes in the body[40], such as stimulation of cell proliferation, apoptosis inhibition and neoplastic transformation. Third, the gut microbiota is altered in obese patients and has been associated w ith the activation of inflammation, w hich may play an important role in the d evelop ment of BΕ and ΕAC[49]. Streptococcus and Prevotella species are the dominant bacteria in the upper gastrointestinal tract, and their ratio may be associated w ith central obesity and hiatal hernia length[27], w hich are tw o know n risks of BΕ and ΕAC. In ad d ition, the gut microbiota may be ad justed concomitantly along w ith d ietary changes that humans experience and that are the main cause of central obesity, but some 'lost' taxa may be d ifficult to regain over generations[50]. Therefore, a large gap in research must be bridged to elucidate the associations among central obesity, microbiota and ΕAC[51].

Ultimately, the optimal method for esophageal microbiota sampling is fundamental and needs to be explored. Many studies have used invasive endoscopy to obtain focal tissues that are similar to those obtained by biopsy or endoscopic brushing to examine the microbiota in a larger area. Gall A et al.[27]show ed that mucosal brush samples could enhance the detection of bacterial diversity in the esophagus and stomach and the microbiota compositions were similar after replicate sampling. Fillon SA et al.[52]p rop osed another minimally invasive samp ling method, namely the overnight esop hageal string test, and suggested that the comp ositions of the esop hageal microbiota were similar betw een the traditional biopsy and the string test. Due to the lack of standard sampling method s, Εlliott DRF et al.[12]stud ied the values of the Cytosp onge p rototyp e as a non-end oscop ic samp ling d evice to seek minimally invasive method s for samp ling the esop hageal microbiota. For comp arison,endoscopic biopsies, brushes and throat swabs were also included. Nevertheless, most of the microbial species overlapped among different sampling methods. However, the sampling area w as larger w ith the Cytosponge, and the microbial DNA yield and total microbial abund ance w ere higher. Moreover, the Cytosp onge also provid ed clues for histological data, thus making it a valuable sampling method.

The important find ings discussed above are summarized in Table 1. Nevertheless,some of these data w ere drawn from a specific population, and the sample sizes were not large. Consequently, these findings still need to be verified before any application to the general p op ulation, and more effort should be mad e to reveal the exact alteration of the microbiota in different diseases.

OTHER POSSIBLE FACTORS RELATED TO THE ESOPHAGEAL MICROBIOTA

Helicobacter pylori

It has been reported that more than 15% of human malignant cancers are related to infection or infection-associated inflammation[53], and many relevant stud ies are mainly focused on single microorganisms. There are tw o main theories to explain bacterial diseases[33]. Koch proposed the classic pathogen theory, which requires the presence of specific pathogens[54]. The other theory of microecological d iseases is a new concept in w hich the w hole microbiota contributes to pathogenicity[26]. As the only bacteria considered a class I human carcinogen, H. pylori is closely related to the progression of gastritis, gastric ulcer, gastric atrophy and gastric adenocarcinoma[55,56].The bacterium has infected more than 50% of the w orld's population[57]and continues to sp read. Although H. pylori is p rimarily localized to the gastric mucosa, its colonization could affect the gastric and esophageal microbiota[11,58], and the microbial composition in the esophagus and stomach overlaps to a certain extent[27].

The increasing incidence of BΕ and ΕAC may be inversely associated w ith H. pylori infection[59,60], and the inverse correlation between H. pylori and ΕAC risk has been well d ocumented[10,27]. Meta-analyses[61-63]based on ep id emiological and observational stud ies show ed that ΕAC coincid es w ith H. pylori erad ication[49]. That is, H. pylori might affect carcinogenesis in the low er esop hagus[27]. How ever, the inner mechanisms have only been partially revealed[10,27]. H. pylori harbors some factors that lead to chronic inflammation and cancer, such as cytotoxin-associated gene A (CagA),vacuolating cytotoxin (VAC) and ad hesins. The bacterium could p romote inflammatory responses by activating nuclear factor kappa B (commonly know n as NF-κB)[56]and may induce the production of certain cytokines such as IL-1β, IL-2, IL-8 and tumor necrosis factor-α (TNF-α), w hich trigger inflammatory resp onses in the gastric epithelium. H. pylori may also d irectly damage host DNA, d ysregulate DNA transcription factors such as caud al type homeobox 2 (Cdx2), and ind uce epithelial injury and acid secretory functions[55,56,58]. Another p ossible mechanism is that H.pylori-induced gastric atrophy causes a reduction in gastric acid, w hich is the main source of GΕR substances[20]. Some stud ies suggest that H. pylori erad ication could increase the serum level of ghrelin, w hich may lead to obesity and affect gastric emptying[61], subsequently initiating the risks of BΕ and ΕAC. Moreover, H. pylori may stimulate apoptosis of ΕAC cells via the Fas-caspase cascade, w hich may account for another protective mechanism[61].

MEDICATIONS

Acid suppression therapies have been highly effective in acid-induced diseases, such as gastritis, esophagitis, BΕ and ΕAC. Proton pump inhibitors (PPIs) are considered benign and are commonly used in clinical p ractice[64]. The sup p ression of acid secretion could affect gastric acid ity, volume and GΕR and even the bacterial composition in the stomach and esophagus, w ith potential consequences for human health and d iseases[36,64]. The ad ministration of PPIs could change the microbial composition in the esophagus and stomach in BΕ patients[34,36], which may contribute to the pathogenesis of BΕ, though this has not been w ell established. Studies have suggested that PPIs may directly target certain bacteria that contain P-type ATPase enzymes as p art of their p roton pump s, such as Streptococcus pneumoniae and H.pylori[65]. Moreover, the microenvironment could also be affected by the increased p H in the stomach and esophagus after PPI therapy. PPIs could reduce the number of gram-negative bacteria and d ecrease the risk for neop lasia in the esop hagus[64]. To detail the changes, Amir I et al.[36]collected esophageal samples before and after 8 wkof PPI treatment in BΕ patients and found that PPI usage changed the esophageal microbiota. At the family level, Comamonadaceae w as d ecreased, whereas other families, such as Clostridiaceae and Lachnospiraceae, were increased. However, longterm PPI therapy induced hypergastrinemia, which may upregulate cyclooxygenase-2(COX-2) exp ression, cell p roliferation and esophageal carcinogenesis[66]. PPI administration plays an important role in H. pylori eradication therapy. A study[67]conducted by Fischbach LA et al. showed that PPIs augment anti-H. pylori activity,and H. pylori appears to exert a protective role in esophageal neoplasia. The possible reason might be that PPIs have some direct protective effects in BΕ, which extend far beyond other effects[64]. At present, there is no direct evidence of a relationship between PPIs and esophageal carcinogenesis, and long-term preclinical and clinical studies with larger samples are needed to reveal the precise associations among PPIs,H. pylori and BΕ/ΕAC.

Table 1 Esophageal microbiota studies on Barrett's esophagus and esophageal adenocarcinoma

These studies displayed are sorted by the publication year. BΕ: Barrett's esophagus; ΕAC: Εsophageal adenocarcinoma; GI: Gastrointestinal.

The introd uction and world w id e application of antibiotics might also have contributed to the increasing incidence of BΕ and ΕAC[20]. The discovery and wide usage of antibiotics have cured many infectious diseases, but several unexpected effects have appeared, some of which may influence the progression of BΕ and ΕAC.Antibiotics may definitively change the gastrointestinal microbiota[3], and alteration of the microbial abundance and/or diversity might contribute to disease pathology[68].Cho I et al.[69]found that the sub-therapeutic administration of antibiotics increased the abundance of Firmicutes and decreased the abundance of Bacteroidetes, which are the two main phyla in the colonic microbiota. Moreover, the overweight population harbored a higher ratio of Firmicutes to Bacteroidetes than the controls[26], and weight loss increased the abundance of Bacteroidetes. These data[26,69]indicated that long-term exposure to antibiotics changes the colonic microbiota, which may induce obesity and GΕR. In the same way, the use of antibiotics can change the microbiota in the esophagus. Tian ZY et al.[70]have suggested that H. pylori infection and antibiotic treatment changes the microbiota composition in the esophagus in a mouse model. In addition, an altered esophageal microbiota might play a more direct role than H.pylori or obesity in inflammation and in BΕ and ΕAC carcinogenesis[20]. On the other hand, antibiotics may help restore the normal esophageal microbiota to type I from type II by increasing the relative abundance of Streptococcus. Interestingly, antibiotics are another important part of H. pylori erad ication therap y. The complicated associations among these factors need further investigation and validation[20].

MECHANISMS

Some stud ies have alread y proposed several p otential mechanisms by w hich the microbiota participates in human carcinogenesis[20]by complicated interactions with the human host immune system and signaling pathways[56](Figure 1). First, alteration of the microbiota may result in inflammation, and persistent chronic inflammation may promote carcinogenesis[71]. The disequilibrium betw een human immunity and microbiota may change the compositions of essential bacterial molecules at certain organs or sites and then form microorganism-associated molecular p atterns(M AMPs)[71], such as toll-lik e recep tors (TLRs) and nucleotid e-bind ingoligomerization-domain (NOD)-like receptors[72,73]. Then, the subsequent activation of related pathw ays may lead to the p rod uction and release of some target genes involved in inflammation[73], such as cytokines, chemokines and other inflammatory factors. For example, specific bacterial invasion could promote the production of IL-17 and IL-23 and then induce an inflammatory response[74].

Bacterial prod ucts, or even the microbes themselves, could be sensed by some recep tors on the ep ithelial membranes. The esophageal type II microbiota could p rod uce larger amounts of gram-negative bacterial comp onents, inv olving lipopolysaccharides (LPS)[19,75]. LPS could d elay gastric emptying via COX1/2[76]and contribute to the d evelop ment of GΕR by increasing the intra-gastric p ressure.Notably, LPS could also affect the function of the lower esophageal sphincter, w hich may promote GΕR and carcinogenesis[75].

TLRs recognize know n molecules from microbes[77,78]and have a well-recognized role in carcinogenesis[38]. As p art of the pathogen-associated molecular p atterns(PAMPs) or danger-associated molecular patterns (DAMPs), TLRs serve as receptors of various ligands, such as bacterial cell wall components and DNA and viral doublestranded RNA[77,78]. Therefore, TLRs med iate the interaction of the immune system w ith the microbiota[79]. Due to the connective roles betw een innate and ad ap tive immune responses, TLRs represent an important linking factor betw een inflammation and cancers[38,77,78,80]. As the natural ligand of LPS, TLR4 is expressed in the human esophageal epithelium, and its expression increases in BΕ and ΕAC[75]. TLR4 activation triggers the NF-κB p athw ay, w hich is related to inflammation-associated carcinogenesis[81]and med iates the initial metap lastic BΕ changes[82]. COX-2 is also upregulated as one of the dow nstream genes, which is related to gastric emptying[75]and occurs along the progression of BΕ and ΕAC[83]. Therefore, the activation of the LPS-TLR4-NF-κB p athw ay may contribute to inflammation and malignant transformation[20,75]. Similarly, activation of the Wnt signaling pathw ay could induce defects in cellular tight junction proteins and decrease the production of some mucins,which may have positive roles in the protection from carcinogenesis[84].

Second, bacteria may generate genotoxins that could cause genomic damage. For instance, the cytolethal distending toxin, w hich may be prod uced by gram-negative bacteria, can ind uce DNA d amage and genomic instability[23,85]. Some bacterial products have tumor-promoting effects[86], such as CagA and VacA, and some bacteria may also activate procarcinogens to p rovoke inflammation and cancer d evelopment[87].

CONCLUSION

The number of studies on how microbial communities contribute to the pathogenesis of BΕ and ΕAC is increasing, and greater attention has been paid to the etiology and molecular mechanisms[49]. Although some observations are promising, the sample sizes of the related studies are not large enough, and the existing data are too limited to draw any convincing conclusion. The most appropriate methods and controls need to be p roven as w ell. The mechanisms by w hich the microbiota affects the pathogenesis of BΕ and ΕAC are still not clear, and further studies are required.Whether the direct interactions betw een microbes and epithelia or the released prod ucts from microbes regulate local inflammation and immunity are und er investigation[88]. Furthermore, it is important to identify when changes occur in the microbial composition during disease progression.

Without a d oubt, exp loring BΕ p athogenesis is a good w ay to stud y the carcinogenesis of ΕAC[11]. However, the alterations of microbial diversity in BΕ and ΕAC are modest[36], and the particular species of bacteria that can discriminate BΕ and ΕAC from controls have not yet been identified. Moreover, some low-abundance genera might be difficult to detect. Amir I et al.[36]reported that they could not identify any biomarker taxa for distinguishing BΕ from controls. Yang L et al.[33]showed that some gram-negative bacteria are enriched in BΕ and might thus be related to BΕ.Moreover, some genera might contribute to the etiology and pathogenesis of BΕ and ΕAC, or result from BΕ and ΕAC[11]. We still need to search for distinct microbial species that could be biomarkers for BΕ and ΕAC in individuals at higher risk, and the subtypes of ΕAC should be taken into consideration.

Nevertheless, finding the true causal mechanisms of dysbiosis is complicated, and the identification of a single species or a collection of species responsible for a particular disorder is sophisticated. The introduction of new techniques, such as nextgeneration sequencing, w ill definitely assist in revealing the mechanisms of the gastrointestinal microbiome in BΕ and ΕAC development, which might provide some evidence of their relations and the pathogenesis of BΕ and ΕAC. The microbiota in BΕ and ΕAC patients remains to be explored, particularly with the adjustment of other risks, such as sex and central obesity. It is crucial to improve our understanding of the process by which the microbial composition may promote disorder progression[20].

Figure 1 Hypothetical mechanisms by which the esophageal microbiota participates in the pathogenesis of Barrett's esophagus (BE) and esophagealadenocarcinoma (EAC). The microbial dysbiosis in the esophagus is associated with abnormal esophagus. Normal esophagus harbors a larger proportion of grampositive bacteria, whereas the microbiota in abnormal esophagus is dominated by gram-negative bacteria. This shift from a gram-positive aerobic microbiota to agram-negative anaerobic microbiota may interact with inflammatory cells and promote the production and secretion of inflammatory factors, such as cytokines andchemokines. In addition, the increase in gram-negative bacteria and their components/products, including LPS, DNA and RNA, may stimulate TLRs (mainly TLR4).TLR4 expression in the esophageal epithelium of BE/EAC is upregulated. As the natural ligands of LPS, TLR4 may play an important role in pathogenesis, whoseactivation could trigger the NF-κB pathway. These interactions mentioned above may stimulate activation of certain intercellular signaling pathways, such as NF-κB.This activation may upregulate the expression of target genes. Moreover, genotoxins generated by some bacteria may cause genomic instability and DNA damage.The end effects might be the induction of inflammation, BE transformation and carcinogenesis. However, whether the microbiota plays a causative role in BE/EACprogression is still unclear. LPS: lipopolysaccharides; NF-κB: nuclear factor kappa B; TLRs: toll-like receptors; BE: Barrett's esophagus; EAC: Esophagealadenocarcinoma.

The prognosis of ΕAC is poor, and the most p ivotal factor is the tumor stage at diagnosis[3]. Given this situation in ΕAC patients, screening certain individ uals w ith higher risks could be useful for early diagnosis. The cost-effectiveness and feasibility of endoscopic usage urge us to seek a less invasive and effective way of screening and early detection for BΕ and ΕAC[89]. Εarly detection of ΕAC will improve the survival rate and quality of life. Id entification of BΕ p atients w ho have higher risks of developing ΕAC could provide diagnostic clues and avoid unnecessary procedures,and the medical resources could be re-distributed to those w ho truly need attention.Understand ing the p athogenesis and exploring biomarkers could also lead to early d etection and p revention and imp rove the survival of some ΕAC p atients. The therap eutic manip ulation of the microbiota, such as p rebiotics, p robiotics or microbiota transp lants, could be a p otent ap p roach in the management of inflammation and cancers[56].

The convincing associations betw een the microbiota and BΕ/ΕAC have indicated the imp ortance of these stud ies. This exciting field of gastrointestinal microbiota allows us to unravel the mystery of carcinogenesis from another perspective, and the integration of different biomarkers may lead us to rapid advances. The introduction of animal models could be the proverbial icing on the cake. Future perspective studies with sophisticated techniques are need ed to explore whether the microbiota changes before or after disease onset, to improve our understanding of the pathogenesis, and to find novel targets for prevention, diagnosis and therapy, w hich could offer more cost-effective and relatively safe choices[19,56,90].