Salivary microbiome in children with EoE correlates with validated EoE activity indexes
In children with EoE, the relative abundance of Haemophilus (base mean = 1858.625, log2 fold change = 0.02, q value = 0.0002) at the genus level and a Pasteurellaceae_unclassified OTU (base mean = 2,977.989, log2 fold change = 2.058e-02, q value < 0.001), a Haemophilus OTU (base mean = 291.029, log2 fold change = 1.897e-02, q value = 0.001), and a Lactobacillus OTU (base mean = 3.127, log2 fold change = 1.239e-02, q value = 0.03) significantly increased with increasing density of eosinophilic infiltration (with log2 fold change given per unit of change of the infiltration density, eos/hpf). Similarly, the relative abundance of Haemophilus had a significantly positive correlation with esophageal mucosal abnormalities as assessed by the EREFS (base mean = 1942, log2 fold change = 1.4332, q value = 5.370e-10) and increasing histopathologic severity as assessed by the EoEHSS (base mean = 2014.595, log2 fold change = 5.8667, q value < 0.001) (Figure 3b,c). Neither microbial richness nor alpha diversity significantly correlated with the increasing density of eosinophilic infiltration, EREFS, or EoEHSS (see Figure 1, Supplementary Digital Content 2, http://links.lww.com/CTG/A38).
Clusters of salivary microbiome correlate with the histologic changes in EoE
Using a heatmap based on the top 30 most abundant genera, we delved deeper into the ongoing potential mechanistic processes that could unravel the relationship between alterations in the salivary microbiome and disease status in children with EoE. The EoEHSS correlated with 4 broad salivary microbial communities representing clades in the unsupervised hierarchical clustering of the taxonomic abundance profiles, identified as clusters A–D (Figure 4a); a Kruskal-Wallis test revealed significant differences in the EoEHSS by cluster designation (P = 0.0108) (see Figure 2, Supplementary Digital Content 2, http://links.lww.com/CTG/A39). Cluster A corresponded with the highest EoEHSS scores (range: 0.15–1) and was defined by higher abundances of a broad range of taxa, including Gemella, Neisseria, Rothia, Prevotella, and Veillonella. Cluster B was associated with the lowest EoEHSS scores (range: 0–0.09) and was also characterized by high abundances of Streptococcus, Gemella, Granulicatella, Neisseria, and Rothia, but unlike those from cluster A, Prevotella and Veillonella abundance was low. Cluster C was linked to intermediate EoEHSS scores (range: 0–0.41) and higher abundances of Oribacterium, Prevotella, Veillonella, and Atopobium. Cluster D was associated with higher EoEHSS scores (range: 0.08–0.75) and was defined by higher abundance of Haemophilus, Streptococcus, Corynebacterium, Moraxella, and Dolosigranulum. To further validate the presence of these microbial clusters observed at the genus level, we conducted additional analyses at the OTU level. At the OTU level, when samples were split by these clusters, the overall microbial communities were significantly distinct between the 4 clusters (PermANOVA P = 0.0002); however, samples from clusters A and C were relatively similar (Figure 4b). As would be predicted from the heatmap, samples from cluster D had the lowest microbial richness (Figure 4c). The salivary microbiome clusters did not correlate with any of the demographic and clinical metadata or the EREFS.
Relationship between medication exposure and the composition of salivary microbiome
As PPIs, inhaled/swallowed, and/or swallowed topical corticosteroids are routinely used in management of EoE, we examined the effect of these medications on the salivary microbiome composition. In total, 33 of the children were on PPIs and 12 were not on PPIs. The PPI use was nonsignificantly lower in non-EoE controls (63%) compared with those with inactive EoE (82%) or active EoE (80%). The PPI use was associated with a higher abundance of Streptococcus (base mean = 2,687.9599, log2 fold change = 3.1740, q value = 4.28e-05), Corynebacterium (base mean = 77.8230, log2 fold change = 3.0577, q value = 0.001), and Rothia (base mean = 38.4750, log2 fold change = 1.2574, q value = 0.01). Although the PPI use was not significantly associated with a difference in microbial richness or alpha or beta diversity, the richness and alpha diversity tended to be lower in children who were using PPIs (all P > 0.20). If analysis was restricted to children with EoE (active and inactive), PPI use (N = 21) compared with no PPI use (N = 5) was not significantly associated with any microbiome changes. Similarly, the use of inhaled and swallowed corticosteroids (N = 6) and/or swallowed corticosteroids (N = 2) in children with EoE (active and inactive) was not significantly associated with differential abundance of any taxa or with changes in the microbial richness or alpha or beta diversity (all relevant P- or q values > 0.30).
In this case-control study, we used 16S rRNA gene sequencing to characterize the composition of the salivary microbiota of children with EoE compared with non-EoE controls. We observed nonsignificant but notable differences in the overall salivary microbiome diversity and composition between children with EoE and non-EoE controls. In addition, in children with EoE, the richness of some distinguishing species positively correlated with validated disease activity indexes. Through an exploratory analysis involving 30 of the most common genera, we were able to discern salivary microbiome profiles which correlated with the intensity and severity of histologic changes observed in the esophageal biopsies—the current gold standard for the diagnosis and monitoring of EoE. These novel findings enhance our understanding of the role of the microbiome in EoE. In particular, it highlights the role of salivary microbiome in pathobiology of EoE.
The relative abundance of salivary Haemophilus was positively correlated with the validated EoE activity indexes after adjusting for potential confounders. Enrichment of Haemophilus in the hypopharyngeal region has been associated with an increased risk of other conditions characterized by eosinophil-mediated allergic inflammation such as recurrent wheezing and asthma in children (23). Similarly, an abundance of Haemophilus in the sinonasal cavity has been demonstrated in patients with chronic rhinosinusitis (24,25), and an abundance of Haemophilus in the sputum has been associated with bacterial exacerbations of chronic obstructive pulmonary disease (26). At the genus level, children with active EoE had a lower relative abundance of Leptotrichiaceae_unclassified, Actinomyces, Lactobacillus, and Streptococcus compared with non-EoE controls. Actinomyces and Streptococcus are among the most abundant genera in saliva collected from adults (27,28). It is unclear whether the decreased abundance of these genera in our cohort of children with EoE is related to their age and/or their disease. Finally, there was a nonsignificant trend toward highest microbial richness and alpha diversity in non-EoE controls compared with children with EoE, suggesting that decreased salivary microbial richness/alpha diversity could be indicative of EoE. Decreased richness of the gut microbiome has previously been associated with other disease states (29), and further research in larger cohort is warranted to investigate whether a decrease in microbial richness could be a predictor or mediator of EoE.
There was a modest overlap between the salivary microbiome composition in our cohort of children with EoE and previously published reports describing the esophageal microbiome in patients with EoE. Benitez et al. (10) evaluated the differences between the oral (collected by swab) and esophageal (in the esophageal biopsies) microbiomes in children with and without EoE and observed a shift in the relative abundance of Proteobacteria including Neisseria and Corynebacterium in children with active EoE, whereas Firmicutes (including Streptococcus and Atopobium) were enriched among non-EoE controls. They reported a modest correlation between oral and esophageal microbiomes for Bacteroides, Firmicutes, and Proteobacteria species. In another study, Harris et al. (11) observed significantly increased abundance of Proteobacteria (mostly Haemophilus) in esophageal mucosal samples collected from children and adults with EoE and a decrease in the extent of Firmicutes in patients with active EoE compared with inactive EoE and non-EoE controls. This is consistent with our current understanding that the esophageal microbiome can be broadly similar to the oral microbiome because both contain an abundance of anaerobes and a high ratio of Firmicutes and Bacteroidetes, and that the oral microbiome can shape the esophageal microbiome through migration of oral bacteria via swallowed or salivary secretions (30,31). Taken together, our findings suggest that although the oronasopharyngeal area and esophagus are anatomically 2 distinct locations, an allergen-mediated eosinophilic inflammation in the esophagus may be linked to oral and salivary dysbiosis. Our findings also raise important mechanistic questions regarding the role of salivary Haemophilus in development of EoE.
This study has limitations. One of the major limitations is a relatively small sample size, which probably resulted in nonsignificant statistical trends. However, these results lay foundation for a larger and sufficiently powered study in the future. Although none of our participants were on dietary elimination therapy alone for EoE, some of them were avoiding foods not specifically as a part of their EoE therapy but for reasons such as texture issues, disliking certain foods, partially avoiding foods such as avoiding milk but consuming yogurt and cheese, empirically eliminating gluten for their gastrointestinal symptoms, and skin prick test–based dietary avoidances in children with known EoE. As a result, we were unable to account for the effect of variability in diet on the composition of salivary microbiome. However, because the participants were nil per os for at least 6 hours before providing saliva samples (per protocol for their EGD), we were able to minimize any immediate effect of diet on the salivary microbiome. Next, the saliva samples were collected as spit, a commonly used collection method, and preprocessed before microbial analyses. Although there is lack of consensus on the optimal method(s) to collect and preprocess saliva for microbial analysis, it is unclear how these steps may have affected our saliva sequencing analysis. However, it is reassuring that the salivary microbiome profiles are minimally affected by commonly used collection methods or DNA extraction protocol (32). This approach needs to be validated in larger and distinct populations before adoption into clinical practice. Our cohort predominantly consisted of males and whites. Although this is consistent with our current understanding of the age and sex distribution of patients with EoE, the results may have to be cautiously applied to patients with EoE who are in other age groups, female, and/or belong to other ethnicities. Finally, this study remains descriptive; future in vivo experiments would be necessary to establish a causal relationship between alterations in the salivary microbiome and EoE.
Despite these limitations, this study is among the first few studies to characterize the salivary microbiome in children and the first to our knowledge to characterize the salivary microbiome in children with EoE and compare it with non-EoE controls. Our participants were consecutively enrolled, allowing us to eliminate a selection bias. The saliva samples were collected within a narrow period, minimizing any influence of circadian rhythm on the composition of salivary microbiome. We collected and analyzed comprehensive clinical metadata (e.g., history of atopic comorbidities, including atopic non-EoE controls, and local corticosteroid use of all types), allowing us to assess how clinical factors influenced the relationship between the salivary microbiome and EoE. All the EGDs were performed by a single investigator, and all the biopsies were examined by a single blinded pathologist; this ensured a high level of uniformity of evaluating the EREFS and EoEHSS, respectively. Finally, we were able to identify salivary microbiome profiles associated with categories of the EoEHSS, indicating that variations in the salivary microbial communities may in general be stratified, and functional analysis might allow us to understand the contribution of microbial communities to EoE pathobiology.
In conclusion, saliva is a biofluid potentially rich in diagnostic indicators for both oral and systemic disorders. The composition of the salivary microbiome in children with EoE seems to be altered compared with that of non-EoE controls, and a relative abundance of Haemophilus may have a role in the pathobiology of EoE. Furthermore, alterations in the salivary microbiome could serve as a practical and noninvasive approach to identify and monitor EoE status (active or inactive). The exact mechanisms underlying the complex interactions between the salivary microbiome, innate immune system, an allergen specific inflammatory response, and esophageal inflammation warrants further research.
Sequencing data has been published at the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under BioProject PRJNA532939; runs SRR8903593 - SRR8903637.
CONFLICTS OF INTEREST
Guarantor of the article: Girish Hiremath, MD, MPH.
Specific author contributions: Study conception and design: G.H., H.C., S.A., and S.R.D. Collecting data and analyzing biopsies and saliva samples: G.H., M.H.S., H.C., S.V.R., and S.R.D. Generation, analysis, and interpretation of salivary microbiome data: G.H., M.H.S., H.H.B., A.T., S.V.R., and S.R.D. Critical revisions of the manuscript: G.H., M.H.S., H.H.B., H.C., S.A., A.T., S.V.R., and S.R.D.
Financial support: This study was funded by the American College of Gastroenterology Clinical Research Award. G.H. is supported by the American College of Gastroenterology Junior Faculty Career Development Award, Vanderbilt University Turner Hazinski award, Vanderbilt University Katherine Dodd Faculty Scholar program, and the Consortium of Eosinophilic Gastrointestinal Disease Researchers (U54 AI117804) training award. Consortium of Eosinophilic Gastrointestinal Disease Researchers (CEGIR) is part of the Rare Disease Clinical Research Network, an initiative of the Office of Rare Diseases Research, National Center for Advancing Translational Sciences, and is funded through collaboration between the National Institute of Allergy and Infectious Diseases, the National Institute of Diabetes and Digestive and Kidney Diseases, and the National Center for Advancing Translational Sciences. The CEGIR is also supported by patient advocacy groups including the American Partnership for Eosinophilic Disorders, Campaign Urging Research for Eosinophilic Diseases, and Eosinophilic Family Coalition. S.D. is supported by the Vanderbilt Institute for Clinical and Translational Research grant support (NIH/NCATS UL1 TR000445, U54 RR24975). S.R.D. is also supported by NIH-funded Tennessee Center for AIDS Research (P30 AI110527) and U19AI095227. Content is solely the responsibility of the authors and does not represent official views of the CDC and the NIH.
Potential competing interests: A.T. is employed by MedImmune, the biologics arm of AstraZeneca, and owns AstraZeneca stock.
WHAT IS KNOWN
- ✓ EoE is a chronic, food and/or aeroallergen-mediated, eosinophil-predominant inflammatory condition affecting the esophagus.
- ✓ Little is known about the role of the microbiome in EoE.
- ✓ To date, the composition of the salivary microbiome in children with EoE and its relationship with disease activity status have not been studied.
WHAT IS NEW HERE
- ✓ The composition of the salivary microbiome is altered in children with EoE compared with non-EoE controls.
- ✓ In children with EoE, abundance of salivary Haemophilus positively correlates with validated endoscopic and histologic disease activity indexes.
- ✓ This study provides novel insights into the role of the salivary microbiome in EoE pathobiology.
- ✓ Alterations in the salivary microbiome may hold potential to serve as a noninvasive marker to monitor EoE activity in children.
1. Davis BP, Rothenberg ME. Mechanisms of disease of eosinophilic esophagitis. Annu Rev Pathol 2016;11:365–93.
2. Wechsler JB, Bryce PJ. Allergic mechanisms in eosinophilic esophagitis. Gastroenterol Clin North Am 2014;43:281–96.
3. Liacouras CA, Furuta GT, Hirano I, et al. Eosinophilic esophagitis: Updated consensus recommendations for children and adults. J Allergy Clin Immunol 2011;128:3–20.e6; quiz 21–2.
4. Moawad FJ. Eosinophilic esophagitis: Incidence and prevalence. Gastrointest Endosc Clin N Am 2018;28:15–25.
5. Soon IS, Butzner JD, Kaplan GG, et al. Incidence and prevalence of eosinophilic esophagitis in children. J Pediatr Gastroenterol Nutr 2013;57:72–80.
6. Rochman M, Azouz NP, Rothenberg ME. Epithelial origin of eosinophilic esophagitis. J Allergy Clin Immunol 2018;142:10–23.
7. Spergel J, Aceves SS. Allergic components of eosinophilic esophagitis. J Allergy Clin Immunol 2018;142:1–8.
8. Jensen ET, Dellon ES. Environmental factors and eosinophilic esophagitis. J Allergy Clin Immunol 2018;142:32–40.
9. Kho ZY, Lal SK. The human gut microbiome: A potential controller of wellness and disease. Front Microbiol 2018;9:1835.
10. Benitez AJ, Hoffmann C, Muir AB, et al. Inflammation-associated microbiota in pediatric eosinophilic esophagitis. Microbiome 2015;3:23.
11. Harris JK, Fang R, Wagner BD, et al. Esophageal microbiome in eosinophilic esophagitis. PLoS One 2015;10:e0128346.
12. Dellon ES, Gonsalves N, Hirano I, et al. ACG clinical guideline: Evidenced based approach to the diagnosis and management of esophageal eosinophilia and eosinophilic esophagitis (EoE). Am J Gastroenterol 2013;108:679–92; quiz 693.
13. Hirano I, Moy N, Heckman MG, et al. Endoscopic assessment of the oesophageal features of eosinophilic oesophagitis: Validation of a novel classification and grading system. Gut 2013;62:489–95.
14. Collins MH, Martin LJ, Alexander ES, et al. Newly developed and validated eosinophilic esophagitis histology scoring system and evidence that it outperforms peak eosinophil count for disease diagnosis and monitoring. Dis Esophagus 2017;30:1–8.
15. Kozich JJ, Westcott SL, Baxter NT, et al. Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the MiSeq Illumina sequencing platform. Appl Environ Microbiol 2013;79:5112–20.
16. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer, New York, 2009.
17. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014;15:550.
18. Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol 1995;57:289–300.
19. Oksanen J, Blanchet FG, Kindt R, et al. Vegan: Community Ecology Package. R Package Version 2.0-10. 2014. https://kityna.ga/548893pdf.pdf#
Accessed on May 3, 2019.
20. Anderson MJ. A new method for non-parametric multivariate analysis of variance. Austral Ecol 2001;26:32–46.
21. Chao A. Estimating the population size for capture-recapture data with unequal catchability. Biometrics 1987;43:783–91.
22. Hill MO. Diversity and evenness: A unifying notation and its consequences. Ecology 1973;54:427–32.
23. Bisgaard H, Hermansen MN, Buchvald F, et al. Childhood asthma after bacterial colonization of the airway in neonates. N Engl J Med 2007;357:1487–95.
24. Lal D, Keim P, Delisle J, et al. Mapping and comparing bacterial microbiota in the sinonasal cavity of healthy, allergic rhinitis, and chronic rhinosinusitis subjects. Int Forum Allergy Rhinol 2017;7:561–9.
25. Chalermwatanachai T, Vilchez-Vargas R, Holtappels G, et al. Chronic rhinosinusitis with nasal polyps is characterized by dysbacteriosis of the nasal microbiota. Sci Rep 2018;8:7926.
26. Wang Z, Bafadhel M, Haldar K, et al. Lung microbiome dynamics in COPD exacerbations. Eur Respir J 2016;47:1082–92.
27. Shaw L, Ribeiro ALR, Levine AP, et al. The human salivary microbiome is shaped by shared environment rather than genetics: Evidence from a large family of closely related individuals. MBio 2017;8:e01237–17.
28. Chen C, Hemme C, Beleno J, et al. Oral microbiota of periodontal health and disease and their changes after nonsurgical periodontal therapy. ISME J 2018;12:1210–24.
29. Le Chatelier E, Nielsen T, Qin J, et al. Richness of human gut microbiome correlates with metabolic markers. Nature 2013;500:541–6.
30. Segata N, Haake SK, Mannon P, et al. Composition of the adult digestive tract bacterial microbiome based on seven mouth surfaces, tonsils, throat and stool samples. Genome Biol 2012;13:R42.
31. Yang L, Francois F, Pei Z. Molecular pathways: Pathogenesis and clinical implications of microbiome alteration in esophagitis and Barrett esophagus. Clin Cancer Res 2012;18:2138–44.
32. Lim Y, Totsika M, Morrison M, et al. The saliva microbiome profiles are minimally affected by collection method or DNA extraction protocols. Sci Rep 2017;7:8523.
Supplemental Digital Content
© 2019 by Lippincott Williams & Wilkins, Inc.