Salivary Metabolomic Signatures in Pediatric Eosinophilic Esophagitis

Abstract

Little is known about metabolomic imbalances in eosinophilic esophagitis (EoE)—an allergen-mediated immunoinflammatory condition of the esophagus [1]. To address this knowledge gap, we conducted global untargeted salivary metabolomics in children with EoE. We analyzed saliva as oral mucosa is the initial interface between the triggering allergens and the host mucosal immune system. Additionally, saliva is rich in metabolites and is uniquely suited for noninvasive sampling, especially in children [2].

In an institutional review board-approved study (#151341), unstimulated saliva samples were collected from 28 children (6–18 years) with known EoE and symptoms of esophageal dysfunction immediately before their esophagogastroduodenoscopy (EGD). Participants were nil per os for > 6 h prior to their EGD. Demographic, clinical, endoscopic, and histologic features were also analyzed (Table 1).

Per the current guidelines [3], 19 (68%) had EoE (active EoE [aEoE] = 9 [47%], and inactive EoE [iEoE] = 10 [53%]), and 9 (32%) were non-EoE controls (controls). Liquid chromatography-high-resolution mass spectrometry was used to analyze saliva samples. Downstream analyses of confidence levels 1, 2, and 3a identified compounds matching searched libraries or databases were performed (see Supporting Information for details) [4].

In all, an abundance of 434 compounds was significantly different ([p ≤ 0.05] with > 2-fold change [FC]) between EoE and controls, and PLS-DA identified well-defined clustering of salivary metabolites classifying the two groups. Ten compounds from endogenous origins belonging to nucleosides, nucleotides, organic acid derivatives, and organoheterocyclic compounds robustly differentiated EoE from controls. Specifically, 4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanol (FC = 6.85; p = 0.003), and N-Heptanoylyglycine (FC = 3.62, p = 0.02), were significantly increased in EoE compared to controls. While deoxyadenosine (FC = −4.0, p = 0.01), deoxycytidine (FC = −3.42, p = 0.01), and adenosine (FC = −3.55, p = 0.02) were significantly lower in EoE than controls (Figure 1). Adenosine is known to have anti-inflammatory effects and is actively converted to inosine due to its long-term instability [5]. We hypothesize that a decrease in adenosine and a corresponding increase in inosine (inosine-41N5) in EoE (compared to control; FC = 1.24; p = 0.02) could be related to the underlying allergic inflammation. Next, 57 compounds differed between aEoE and iEoE. Of these, urea was significantly high (p = 0.03), and serylarginine was significantly low (p = 0.007) in aEoE compared to iEoE, suggesting that EoE activity can impact salivary metabolites (Figure 2b).

Finally, we found 102 compounds with negative FC and 207 with positive FC differentiated aEoE from controls. Of these, eight molecules had a maximum fold-change of ≥ 3. All nucleotide molecules (including deoxyadenosine, deoxycytidine, adenosine, deoxyguanosine, and guanosine) showed significant negative FCs. In contrast, the organic acids and derivatives (such as urea, N-Heptanolyglycine, and L-Arginine) were significantly increased in aEoE compared to controls (Figure 2a). Arginine has an essential role in the inflammatory response, and esophageal microbiota of aEoE patients have a higher abundance of arginine, which treatment can disrupted [6]. A concurrent increase in salivary urea and arginine needs to be further investigated. The adenosine (p-value = 0.019, FC = 3.82) and pentadecanoylcarnitine (p-value = 0.031, FC = 3.38) were decreased in iEoE compared to controls. Pentadecanoylcarnitine is a fatty acid that plays a role in cellular metabolism, but its role in esophageal dysfunction remains to be understood.

This is the first study to conduct global untargeted salivary metabolomics in children with EoE. We identified eight salivary metabolites of significant fold change altered in children with EoE compared to controls and two metabolites altered between aEoE and iEoE, irrespective of age and medication exposure. Studies with large sample sizes involving all ages and longitudinal follow-up are warranted to validate our findings and test their clinical utility.

G.H. is a consultant to Bristol Myers Squibb, Regeneron, and Sanofi.

The authors declare no conflicts of interest.