Longitudinal follow-up of gut and salivary microbiota in children with food protein-induced enterocolitis syndrome

Background: Gut microbiota and their metabolites play a key role in digestive dysimmunity and tolerance breakdown, leading to pathologies such as food protein-induced enterocolitis syndrome (FPIES). This study aimed to compare both gut and salivary microbiota over time between allergic and tolerant patients and matched controls. Methods Faecal and salivary samples were collected longitudinally from FPIES allergic patients on an elimination diet, and after their natural tolerance acquisition, and from matched controls (1:1 ratio). 16S rRNA gene sequencing and biostatistical analyses (ALDEx2, ANCOM-BC, DESeq2, LEfSe, MaAsLIN2) were used to compare microbiota composition between groups. The measurement of faecal short-chain fatty acids (SCFAs) (gas chromatography) evaluated gut microbiota functions. Results Thirty-eight allergic patients were included (median age: 1.3 years), and 22 became tolerant over time. Allergic patients exhibited lower gut and salivary alpha-diversities, and several different genera (12 in stools including less Ruminococcus, 2 in saliva), and higher relative abundance of acetate than controls. With tolerance acquisition, faecal alpha-diversities partially increased, whereas SCFAs normalized. Ultimately, there were more differences in faecal and salivary microbiota in allergic patients vs . after tolerance acquisition (i.e. more Blautia, Ruminococcus, Faecalibacterium, Prevotella -9, and less Escherichia - Shigella in tolerant patients) than between controls over time. Conclusion The faecal and salivary dysbiosis observed in allergic patients with FPIES partially resolved after tolerance acquisition. Understanding the underlying mechanisms of dysbiosis is crucial to prevent and help manage FPIES. A prolonged follow-up could determine if dysbiosis resolution would be complete, and if this early infancy dysbiosis has long-term consequences on health. Title: Longitudinal follow-up of gut and salivary microbiota in children with food protein-induced enterocolitis syndrome Short title: Gut and salivary microbiota in FPIES Authors, full names and institutional affiliations: LEMOINE Anaïs (MD-PhD): INRAe, Micalis Institute, UMR1319, Jouy-en-Josas, France; Sorbonne Université, APHP-Trousseau Hospital, Pediatric nutrition and gastroenterology department, Paris, France; Sorbonne University, Inserm UMRS1269, Nutriomics, Nutrition and obesities: systemic approaches, Paris, France; FHU-GLIMMER, Paris, France. ORCID: 0000-0001-8443-7207 ADEL-PATIENT Karine (PhD): Université Paris-Saclay, CEA, INRAE, DMTS, Gif-sur-Yvette, France. ORCID: 0000-0002-2242-0626 KAPEL Nathalie (PharmD-PhD): APHP-Pitié-Salpétrière Hospital, Functional coprology laboratory, Paris, France; INSERM, UMRS 1139, FPRM, Université Paris Cité, Paris, France; FHU-GLIMMER, Paris, France. ORCID: 0000-0002-2068-9015 MAYEUR Camille (PhD) : INRAE, Micalis Institute, UMR 1319, Paris-Saclay University, AgroParisTech, Jouy-en-Josas, France. ORCID: 0000-0003-1418-7657 ROSSIGNOL Marie-Noëlle, Université Paris-Saclay, INRAE, AgroParisTech, GABI, Jouy-en-Josas, France. ORCID: 0000-0003-1673-5258 BRUNEAU Aurélia, INRAE, Micalis Institute, UMR 1319, Paris-Saclay University, AgroParisTech, Jouy-en-Josas, France. ORCID: 0009-0004-8068-1041 TOUNIAN Patrick (MD-PhD): Sorbonne Université, APHP-Trousseau Hospital, Pediatric nutrition and gastroenterology department, Paris, France; FHU-GLIMMER, Paris, France. ORCID: 0000-0002-5976-2240 THOMAS Muriel (PhD): INRAE, Micalis Institute, UMR 1319, Paris-Saclay University, AgroParisTech, Jouy-en-Josas, France; FHU-GLIMMER, Paris, France. ORCID: 0000-0002-7608-3274 NICOLIS Ioannis (PhD): BioSTM, CNRS UAR 3612, Inserm US25, Université de Paris Cité, Paris, France. ORCID: 0000-0001-7207-7689 Acknowledgments: We acknowledge patients, controls and their parents for their participation. We also thank Dr G. Benoist, Dr P. Challier, Dr A. Cordesse, Dr E. Foy Foulques, Dr K. Garcette, Dr T. Guiddir, Dr J. Lemale, Dr G. Lezmi, Dr S. Pauliat-Desbordes for their help in the recruiting of the patients, and all the lab team, hospital technicians (Pitié-Salpétrière-APHP), A-J. Waligora, and Pr P. Seksik for their support. We are grateful to the INRAE MIGALE bioinformatics facility (MIGALE, INRAE, 2020. Migale bioinformatics Facility, doi: 10.15454/1.5572390655343293E12) for providing help and/or computing and/or storage resource. The sponsor was Assistance Publique – Hôpitaux de Paris (Délégation à la Recherche Clinique et à l’Innovation). The authors thank ITC Traductions for the English language review. Funding statement: Sodilac/Soredab, SFP (Société Française de Pédiatrie), SFNCM (Société Francophone Nutrition Clinique et Métabolisme), SNFGE (Société nationale française de gastro-entérologie), SFA (Société Française d’Allergologie). None of the funding sources was involved in the conduct of the research, results analysis and the preparation of the article. Disclosure Statement: The authors declare no conflict of interest in relation to this work. Abstract

Gut microbiota and their metabolites play a key role in digestive dysimmunity and tolerance breakdown, leading to pathologies such as food protein-induced enterocolitis syndrome (FPIES). This study aimed to compare both gut and salivary microbiota over time between allergic and tolerant patients and matched controls.

Faecal and salivary samples were collected longitudinally from FPIES allergic patients on an elimination diet, and after their natural tolerance acquisition, and from matched controls (1:1 ratio). 16S rRNA gene sequencing and biostatistical analyses (ALDEx2, ANCOM-BC, DESeq2, LEfSe, MaAsLIN2) were used to compare microbiota composition between groups. The measurement of faecal short-chain fatty acids (SCFAs) (gas chromatography) evaluated gut microbiota functions.

Thirty-eight allergic patients were included (median age: 1.3 years), and 22 became tolerant over time. Allergic patients exhibited lower gut and salivary alpha-diversities, and several different genera (12 in stools including less Ruminococcus, 2 in saliva), and higher relative abundance of acetate than controls. With tolerance acquisition, faecal alpha-diversities partially increased, whereas SCFAs normalized. Ultimately, there were more differences in faecal and salivary microbiota in allergic patients vs . after tolerance acquisition (i.e. more Blautia, Ruminococcus, Faecalibacterium, Prevotella -9, and less Escherichia - Shigella in tolerant patients) than between controls over time.

The faecal and salivary dysbiosis observed in allergic patients with FPIES partially resolved after tolerance acquisition. Understanding the underlying mechanisms of dysbiosis is crucial to prevent and help manage FPIES. A prolonged follow-up could determine if dysbiosis resolution would be complete, and if this early infancy dysbiosis has long-term consequences on health. Key Words (up to 5) FPIES Microbiota Short-Chain Fatty acids Salivary Stools Word count : 3487/3500

The gut microbiota and their metabolites are widely explored to better understand the pathogenesis of disease. Oral pre-pro-syn-postbiotics (1,2) or even faecal microbial transfer (3) are being studied to prevent, alleviate or cure some pathologies. Indeed, the gut microbiota is a key element in immune system development during infancy, and any factor that could disturb the composition and function of the microbiota during the first years of life could lead to pathology later in life (4). Dysbiosis precedes the onset of food allergy in humans, suggesting microbiota plays a role in pathogenesis (5). Some correlations exist between salivary microbiota or metabolome and some extra-oral pathologies such as inflammatory diseases, diabetes, or cardiovascular pathologies (2,6,7). As with the gut microbiota, the oral microbiota matures during the first 2 years of life, influenced by pre- and post-natal factors (8,9). Food protein-induced enterocolitis syndrome (FPIES) is a non-IgE-mediated food allergy, primarily causing digestive symptoms (10), the pathophysiology of which has not yet been elucidated (11). Gut microbiota of infants with FPIES differs from that of healthy infants, as early as the first months of life, before the onset of FPIES symptoms (12–17). However, no study has investigated longitudinally the changes in microbiota in patients with FPIES, and saliva has not been studied in non-IgE-mediated gastrointestinal food allergy. We previously described increased faecal secretory IgA in allergic patients and no measurable chronic gut inflammation (18). In this current work, we hypothesised that gut and oral microbiota are transiently dysbiotic in allergic patients with FPIES, before an evolution towards those of healthy controls after tolerance. Thanks to the largest FPIES paediatric cohort to date, the objectives of this study were to describe the faecal and salivary microbiota longitudinally, from diagnosis to the acquisition of tolerance to the offending foods in FPIES patients, with comparisons to matched controls, and to identify microbial signatures or dysfunctions associated with this allergic disease.

Patients and controls This multicentre longitudinal case-control study was undertaken from December 2019 to May 2023 across 4 tertiary care centres of “Assistance Publique-Hôpitaux de Paris”. The detailed inclusion and exclusion criteria of the cohort and the matching procedure between patients and controls are described elsewhere (18). Allergic patients were on a strict elimination diet from the triggering food at inclusion. Patients were defined as tolerant once all triggering foods had been successfully reintroduced. Controls matched with allergic patients were designated as controls-A, and those matched with tolerant patients were designated as controls-T (Figure 1). The study protocol was approved by the French Committee for Protection of Persons (“Ouest I”) in May 2019 and March 2020 for patient and control inclusions (EudraCT/ID RCB: 2019-A00336-51, ClinicalTrials.gov: NCT04081415). The consent of the parents and of children, if possible, was obtained before samples were taken. Sample collection and preparation Stool and salivary samples were collected in allergic patients and controls-A, and in tolerant patients and controls-T. In case of acute infection, a one-month delay was observed before conducting sampling. Stools were promptly refrigerated, frozen at -20°C within 2 hours and then stored at -80°C until aliquoting and freezing again for further analysis. Saliva was collected with a pipette after a minimum of 2-hour fasting. The saliva was either stored at +4°C until centrifugation (5000x g, 15 minutes, 4°C) within 2 hours to separate the supernatant from the pellet and then frozen at -80°C or immediately frozen at -20°C and then centrifugated (18000x g, 15 minutes, 4°C) on the day of analysis. DNA extraction, quantification and amplification of 16S rRNA gene Faecal DNA was extracted from approximately 200 mg of faecal material. Salivary DNA was extracted from the pellet. QIAamp® PowerFecal® DNA kit (Qiagen) was used until September 2021, and then QIAamp® PowerFecal Pro® DNA kit (Qiagen) (similar faecal sequencing data with both kits, data not shown). A NanoDrop spectrophotometer (ThermoFisher Scientific, USA) was used to assess the quality of the DNA extracts. Faecal and salivary DNA were quantified using the Qubit® 3.0 fluorometric method (Life TechnologiesTM). PCR inhibition was checked (TaqManTM Exogenous IPC Reagent kit, ThermoFisher Scientific) to allow quantitative real-time PCR (qPCR) without inhibitor at the final dilution. qPCR was performed using 10 μL of diluted DNA (dilution of 1000 for faecal DNA, 20 for salivary DNA) to measure total bacterial load per sample (primers F-Bact 1369 (5’-CGGTGAATACGTTCCCGG-3’) and R-Prok 1492 (5’-TACGGCTACCTTGTTACGACTT-3’), fluorochrome FAM (6-carboxyfluorescein) TM1389F (5’-CTTGTACACACCGCCCGTC-3’), TaqMan® master mix (Applied Biosystems™)). The qPCR conditions on the StepOne thermocycler (ThermoFisher Scientific) were as follows: initial denaturation of genomic DNA by heating at 95°C for 10 minutes, followed by 40 cycles involving an initial 15-second denaturation of amplicons at 95°C and a second step of hybridisation and extension for 1 minute at 60°C (19). Amplification curves for qPCR and threshold cycles were generated using StepOneTM Software (Life Technologies Corporation), and the analyses were conducted using Microsoft Excel. All samples were studied in duplicate with a maximum tolerated difference of 0.5 x Ct (cycle threshold). The 16S rRNA gene was amplified targeting the V3-V4 hypervariable regions thanks to primers V3F (5’-CTTTCCCTACACGACGCTCTTCCGATCTACGGRAGGCWGCAG-3’) and V4R (5′-GGAGTTCAGACGTGTGCTCTTCCGATCTTACCAGGGTATCTAATCCT-3’), along with Taq polymerase (MolTaq 16S/18S for faecal and salivary DNA, with the mastermix 16S/18S Basic (Molzym) for the lowest DNA concentrated samples), following the Illumina 16S rRNA metagenomic sequencing library preparation protocol. Sequencing was performed using an Illumina MiSeq Sequencer (INRAe, @Bridge platform, Jouy-en-Josas) and generated 2 × 250 bp paired-end reads. FastQ files were generated after the run was completed (MiSeq Reporter software, Illumina, USA). Samples with fewer than 10,000 reads after quality control procedures were eliminated from the study. We used FROGS pipeline (v4.0.1; June 2023) (20) to merge read pairs (Flash v1.2.11), to cluster the reads with a distance of 1 between sequences (Swarm v3.0.0), to remove chimeric sequences ( Vsearch v2.17.0) and then to filter OTUs (removal of sequences with abundance less than 0.005% or present in databank of contaminants) (Blast v2.10) and affiliate them according to the reference database 16S SILVA Pintail100 138.1 and RDP assignment (EMBOSS v6.6.0). Multiple alignment of OTUs was generated with Mafft (v7.407) and we obtained an abundance table and a phylogenetic tree without data normalisation (Fasttree (v.2.1.0) and Phangorn R package (v.2.7.0) through FROGS). Short-chain fatty acids (SCFAs) SCFAs were quantified on a gas chromatograph (Agilent 7890B), following established protocol as described elsewhere (21). Faecal samples (0.4 g maximum) were solubilized in water (2 to 3 volumes of water per stool weight, mixed intensively using laboratory vortex, and stirred for 2 hours at 4°C). After centrifugation (12000 g, 15 minutes, 4°C), supernatants were deproteinized overnight with saturated phosphotungstic acid (10% of the supernatant volume) at 4°C. All samples with 20% of internal standard (20mM 2-ethyl butyrate) were tested in duplicate. The generated peaks were integrated using OpenLab Chemstation software. Values are expressed in µg/g of stools or in percentage of total SCFA and interquartile ranges (IQR). Statistical analysis All statistical analysis was performed using GraphPad Prism (22), Python and R statistical environments (23). Data were compared between matched subjects using paired t-test or Wilcoxon signed-rank test according to Shapiro-Wilk normality test, or Wilcoxon-Mann-Whitney test for non-paired comparisons. Whenever multiple tests were performed, p-values were adjusted for multiple comparisons (24). For all bacterial analyses except alpha-diversities, data were aggregated at the genus level and taxa were retained if detected in at least three subjects and if contributing to the top 99% cumulative abundance. Alpha-diversities (25,26) were assessed through three indices: Chao1 (richness estimate), Shannon and Inverse Simpson (diversity estimates, with Shannon taking into account rare and abundant species while Inverse Simpson being less sensitive to rare species). Beta-diversities explore inter-group differences from different viewpoints, and five different distances were used for beta-diversity analysis. Bray-Curtis is a metric sensitive to species abundance, Jaccard is a binary presence/absence metric, Unifrac (both with and without abundance weighting) considers the phylogenetic inter-species distance and Jensen-Shannon Divergence (JSD) is a symmetric probability distribution metric able to handle null abundances. Beta-diversities were mapped using unconstrained Multidimensional Scaling (MDS) or Constrained Analysis of Principal Coordinates (CAP). Homogeneity of multivariate group dispersions was verified (25) before applying permutation ANOVA analysis (PERMANOVA) with Adonis2 to test differences between groups (26). Permutations were constrained to consider pairing of subjects. Five different differential analysis algorithms were used to identify specific genera enriched or depleted between groups. Non-paired analyses were performed using DESeq2 (Differential expression analysis for sequence count data) (27) which fits negative binomial models to the data, LEfSe (Linear discriminant analysis Effect Size) (28) which identifies significant different taxa using Kruskal-Wallis and Wilcoxon tests followed by linear discriminant analysis to estimate the effect size, and finally ANCOM-BC (Analysis of Compositions of Microbiomes with Bias Correction) (29) which uses a log-linear regression model to correct bias arising from sampling depth differences. Paired analyses were performed using ALDEx2 (ANOVA-Like Differential Expression) (30) which fits Dirichlet-multinomial models and MaAsLin2 (Microbiome Multivariable Association with Linear Models) (31) which fits a linear model to log-transformed normalised data.

Characteristics of study population The studied cohort has been described elsewhere (18). Briefly, 38 allergic patients were matched with 38 healthy controls (Figure 1). There was no statistical difference between the cases and the controls at inclusion, regarding age at study inclusion (1.3 years, IQR: 0.9-2.0 vs. 1.3 years, IQR: 0.9-2.2, p=0.08), the delivery mode (vaginal delivery: 27/38, 71.1% in both groups), and breastfeeding ratio (32/38, 84.2% in patients vs. 34/38, 89.5% in controls, p=0.74). The two main food allergens were cow’s milk (n=16, 42.1%) and eggs (n=7, 18.4%), followed by fish, meat and cereals (n=5, 13.2% each) (18). Comparison of allergic patients and controls at study inclusion The relative abundances of phyla and genera in stools are shown in Figures 2A-B. All alpha-diversity indices were significantly lower in allergic patients than in controls (Figures 2C, S1A-B) although faecal bacterial load was similar (6.2x10 10 and 7.2x10 10 bacteria/g of stools respectively, p=0.50) (Figure S1C). Statistical analysis did not evidence any differences when comparing alpha-diversities in allergic patients according to the main offending food (milk, egg, fish, other), or whether FPIES was single or multiple (n=6), or even according to the onset of sensitisation during follow-up (n=5 atypical FPIES), (all p-values>0.05). Faecal beta-diversities significantly differed between allergic patients and controls-A (Figures S1D-E). Considering that at least two concordant results must be obtained between abundance analysis to be relevant, 11 genera were significantly decreased (including Ruminococcus and Lactococcus ) (Figure 2D), and one genus was increased in allergic patients compared to controls-A (Figure 2E). In faecal samples, allergic patients had higher relative abundance of acetate (69.3% [62.5-77.2] vs . 64.6% [56.9-74.0], p=0.03) (Figure 2F), but SCFAs were at equivalent concentrations in allergic patients and controls-A (in µg/g, total SCFA: 47.4 [36.2-65.1] vs . 56.4 [42.0-69.0]; acetate: 32.5 [24.3-46.5] vs . 34.1 [25.3-48.0]; butyrate: 3.9 [2.4-9.1] vs . 6.3 [2.8-10.8]; propionate: 7.7 [5.4-11.4] vs . 8.9 [6.0-12.9], all p-values>0.05). Although less abundant than main SCFAs, other SCFAs were at lower relative abundances (Figure 2F) and concentrations in allergic patients compared to their matched controls (in µg/g: valerate: 0.1 [0.0-0.4] vs . 0.4 [0.0-1.1], p=0.03, but no significant difference in relative abundances; isobutyrate: 0.5 [0.2-0.75] vs . 0.7 [0.3-1.1], p<0.01; isovalerate: 0.4 [0.2-0.8] vs . 0.8 [0.3-1.6], p<0.01). Saliva was mainly composed of bacteria from the Bacillota phylum (genus Streptococcus ), and Pseudomonadota phylum (genus Neisseria ) (Figures 3A, 3B). Salivary alpha-diversities (Shannon, invSimpson) were significantly lower in allergic patients compared to controls-A (Figures 3C, S2A, S2B), with a similar bacterial load (8.6x10 8 and 8.4x10 8 bacteria/g of saliva respectively, p=0.54) (Figure S2C). Salivary beta-diversities were similar between allergic patients and controls-A (Figures S2D-E). According to at least two concordant differential abundance analyses, only two genera were significantly different in allergic patients compared to controls-A (Figure 3D). Longitudinal comparison of patients As previously described (18), 22 patients became tolerant to the offending food by the end of follow-up, with a median delay of 1.1 years [0.9-1.3] since inclusion (Figure 1). We then compared data obtained over time for these 22 patients, and that obtained for 22 matched controls over the same timeline. The increase in Bacillota phylum in stools over time was observed in patients (Figure 4A) (LEfSe), and was not observed in controls (not shown), whereas the faecal and salivary bacterial loads were stable in patients and controls (Figures S3C, S4C). Faecal alpha-diversities increased significantly with age in both patients and controls (Figures 4C, S3A-B), whereas it increased only in saliva in patients (Figures 5C, S4A-B). Faecal and salivary beta-diversities were significantly different in patients over time (Figure S3D-E, S4D-E), but not in controls (Figure S3F-G, S4F-G). Twenty-two faecal bacterial genera (including 17 increased genera from Bacillota phylum) (Figures 4D-E) and 17 salivary genera (Figures 5D-E) were different between allergic and tolerant patients, whereas only 2 faecal ( Enterobacteriaceae and Clostridioides ) and 1 salivary genera ( Corynebacterium) differed between controls during follow-up (data not shown). In patients developing tolerance, we observed a significant decrease in relative proportions of acetate (allergic vs. tolerant: 69.3% [62.0-77.2] vs. 58.8% [54.8-66.3], p<0.001) and increase in butyrate concentrations (3.4 µg/g [2.1-9.6] vs. 7.6 [4.0-11.0], p=0.007; 7.7% [4.3-15.8] vs. 14.1% [12.0-17.7], p<0.001) (Figure 4F), which was not due to age as not evidenced in controls (data not shown). Comparison of tolerant patients and matched controls at the end of follow-up We matched tolerant patients with the 20 initial controls at the end of follow-up and 2 new controls for paired analysis. At the end of follow-up, faecal alpha-diversities in tolerant patients almost reached the levels of controls-T (Figures 4C, S3A-B), whereas salivary alpha-diversities normalised (Figures 5C, S4A-B). Faecal and salivary bacterial loads of tolerant patients exceeded the values of controls-T (Figures S3C, S4C). Beta-diversities were similar in stools (Figures S3F-G), but discordant in saliva (Figures S4F-G). At the end of the study, only 2 genera in stools differed between tolerant patients and controls-T (Figures S3H-I), and none in saliva. Regarding SCFAs, both absolute and relative amounts were similar in tolerant patients and controls-T (Figure S3J). Comparison of patients based on duration of FPIES We tried to identify predictive factors of FPIES persistence in clinical and microbiota data. We only considered patients with sufficient hindsight to determine whether or not the patient recovered within 2 years of first symptoms (n=24). The age at inclusion was younger in patients who became tolerant faster (n=12 in each sub-group, 1.0 versus 2.8 years, p<0.001). However, neither faecal and salivary alpha-diversities (Shannon, Chao1, invSimpson), nor main SCFA (absolute and relative abundances) at inclusion discriminate patients based on duration of FPIES. These markers were also not different from matched controls (data not shown), except acetate (higher relative abundance in patients whose recovery period exceeded 2 years compared to matched-controls, 63.7% vs. 58.3%, p=0.003).

Allergic patients with FPIES on an elimination diet exhibited gut and saliva bacterial dysbiosis associated with higher relative levels of faecal acetate than matched controls and lower relative butyrate levels than tolerant patients. Upon acquisition of tolerance to the offending food, gut dysbiosis partially recovered with intermediate alpha-diversity observed at the end of follow-up, fewer microbial differences and normalisation of SCFAs compared to non-allergic controls. In saliva, recovery of microbiota could be considered complete in tolerant patients. Gut microbiota was characterised by a lower richness and evenness in allergic patients with FPIES, along with significant differences in beta-diversity. Caparrós et al. observed contrasting results for alpha-diversities in a comparative Spanish study (14). Castro et al. (16) and Xiong et al . (17) found no difference in faecal alpha-diversities in their case-control studies. Su et al., like us, found lower diversity (Shannon) in FPIES cases (n=8) in a prospective case-control cohort, before the onset of FPIES (13). The differences in cohort age and size could explain these discrepant conclusions. Although lower alpha-diversities are typically expected in pathologies (32), results are conflicting in food allergy (33–38) and require clarification. Regarding gut bacterial composition, most differences were observed between allergic patients and matched controls at inclusion, as well as between still-allergic and tolerant patients at the end of follow-up. These differences were not solely explained by the gut microbiota maturation with age, as observed in controls. The taxonomic disparities between groups varied depending on the statistical methodology used (ALDEX2, LEfSe, ANCOM-BC, MaAsLin2 or DeSeq2). To ensure more relevant conclusions, we focused on genera identified by at least two differential compositional analyses. Although bacterial identification in our study was limited to the genus level, we found similar results to previous works on microbiota in FPIES, such as Escherichia-Shigella (12,16), Pseudomonadota (16) (increased in still-allergic patients compared to tolerant patients), Ruminococcus (14), Faecalibacterium (13,14) Clostridia (13) and Bacillota (formerly Firmicutes ) (13) (decreased in allergic patients compared to controls or tolerant patients). Conversely, we observed opposite results for Lactococcus (14), Lachnospiraceae (14), Lachnospiraceae NKA136 group (17) and Blautia (14) (decreased in allergic patients compared to controls or tolerant patients). The Ruminococcus gnavus group was associated with tolerance, and Su et al. found this bacterium to be decreased in patients who became allergic . In both IgE and non-IgE-mediated food allergies, there exists a wide range of results and contradictory conclusions regarding the abundances of bacteria such as Blautia, Roseburia or Ruminococcus (33,36–44). For most of these bacteria, their clear role in health has not been elucidated and could be strain-dependent (45–48). Factors like age at inclusion, FPIES management (17), geographic origin of patients, dietary habits and biostatistical methodologies could explain these differences. It is worth noting that the gut microbiota in FPIES has only been described in less than 120 children to date. Beyond describing a specific microbial signature in pathology, it is certainly more important to understand how global dysbiosis arises, the potential consequences of altered functions of a dysbiotic microbiota on the host (e.g.: synthesis of SCFAs or metabolites), and how and why dysbiosis spontaneously disappears. Longitudinal studies, as we proposed, could help answer these questions. Bacteria responsible for producing SCFAs were found to be less abundant, as well as the expression of the Bifidobacterium shunt responsible for SCFA production (13). Castro et al. also observed increased excretion of acetate in cow’s milk-induced FPIES patients compared to matched controls (16). In our cohort, we observed increased acetate and decreased butyrate (relative contents) in allergic patients, compared to matched controls or tolerant patients. Decreased butyrate has been associated with food allergy (37,41,42), as have acetate or propionate (37,42). SFCAs, particularly butyrate, play a crucial role in modulating the host immune response and activating Treg lymphocytes, cells that promote immune tolerance (49,50). We also found decreased branched SCFAs in allergic patients, whereas both Castro et al . and Diaz et al . found opposite results in children with FPIES to cow’s milk (16) and non-IgE-mediated food allergies (40). However, SCFAs represent only a fraction of the multiple metabolites produced by the microbiota. Metabolomic analysis of stools would provide additional information about the altered functions of the microbiota in FPIES. In a longitudinal study, oral microbial composition differs during early infancy between patients developing allergies or asthma and healthy subjects (51). Additionally, in food allergy, multi-omic analyses of saliva reveal distinct features in allergic patients (52), and the combined use of salivary parameters at 6 months of age could differentiate patients with food reactions in the first 2 years of life from controls (53). For the first time, we showed that salivary microbiota was less impacted in allergic patients with FPIES than gut microbiota. This confirms that FPIES is mainly a gut-brain disease and that saliva cannot be used as a tool for predicting prognosis. In practice, we now know that FPIES is associated with gut dysbiosis. A caregiver survey retrospectively revealed, in allergic patients, higher antibiotic usage during infancy and during the prenatal period in mothers before the first symptoms of FPIES (54). As a primary prevention, clinicians should limit unnecessary exposure to factors favouring dysbiosis in children, such as antibiotics in the absence of confirmed bacterial infection, or proton pump inhibitors for uncomplicated gastroesophageal reflux. Pre-pro-syn-postbiotics should be explored as a therapeutic perspective for a faster resolution of FPIES. Moreover, since tolerant patients achieve partial recovery of microbiota similar to that of healthy individuals, it is important to continue monitoring these children and to determine whether there is an increased risk of digestive or immune disease in the long term. The strengths of this case-control study relate to its longitudinal design, which extended until the acquisition of tolerance, the selection of controls with multiple demographic and clinical parameters for matching, the comprehensive analysis of bacterial composition to identify relevant differences, and the assessment of SCFAs. However, limitations included age variability among cases, diversity of offending foods, and use of 16S rRNA gene sequencing instead of shotgun metagenomic sequencing.

Allergic patients with FPIES exhibited both faecal and salivary dysbiosis with notable effects on the contents of SCFAs. The dysbiosis decreased spontaneously over time with natural acquisition of tolerance, but alpha-diversity remained lower at the end of the follow-up period. Identifying differences between tolerant and still-allergic patients at the species level would help future research investigating whether interventions targeting the gut microbiota with dietary or pre-pro-syn-postbiotics may expedite the acquisition of tolerance in FPIES.

1. Sestito S, D’Auria E, Baldassarre ME, Salvatore S, Tallarico V, Tarsitano F et al. The Role of Prebiotics and Probiotics in Prevention of Allergic Diseases in Infants. Front Pediatr 2020; 8 :583946.2. Khor B, Snow M, Herrman E, Ray N, Mansukhani K, Patel KA et al. Interconnections between the Oral and Gut Microbiomes: Reversal of Microbial Dysbiosis and the Balance between Systemic Health and Disease. Microorganisms 2021; 9 :496.3. Gulati AS, Nicholson MR, Khoruts A, Kahn SA. Fecal Microbiota Transplantation Across the Lifespan: Balancing Efficacy, Safety, and Innovation. Am J Gastroenterol 2023; 118 :435–439.4. Parkin K, Christophersen CT, Verhasselt V, Cooper MN, Martino D. Risk Factors for Gut Dysbiosis in Early Life. Microorganisms 2021; 9 :2066.5. Bunyavanich S, Berin MC. Food allergy and the microbiome: Current understandings and future directions. J Allergy Clin Immunol 2019; 144 :1468–1477.6. Basilicata M, Pieri M, Marrone G, Nicolai E, Lauro M Di, Paolino V et al. Saliva as Biomarker for Oral and Chronic Degenerative Non-Communicable Diseases. Metabolites 2023; 13 :889.7. Chattopadhyay I, Lu W, Manikam R, Malarvili MB, Ambati RR, Gundamaraju R. Can metagenomics unravel the impact of oral bacteriome in human diseases? Biotechnol Genet Eng Rev 2023; 39 :85–117.8. Dzidic M, Collado MC, Abrahamsson T, Artacho A, Stensson M, Jenmalm MC et al. Oral microbiome development during childhood: an ecological succession influenced by postnatal factors and associated with tooth decay. ISME J 2018; 12 :2292–2306.9. Kennedy B, Peura S, Hammar U, Vicenzi S, Hedman A, Almqvist C et al. Oral Microbiota Development in Early Childhood. Sci Rep 2019; 9 :19025.10. Nowak-Wegrzyn AH, Chehade M, Groetch ME, Spergel JM, Wood RA, Allen K. International consensus guidelines for the diagnosis and management of food protein–induced enterocolitis syndrome: Executive summary—Workgroup Report of the Adverse Reactions to Foods Committee, American Academy of Allergy, Asthma & Immunology. J Allergy Clin Immunol 2017; 139 :1111-26.e4.11. Lemoine A, Adel-Patient K. FPIES: Immune Mechanisms. In: Food Science . 2024: 50003.12. Boyer J, Scuderi V, Keene NH, Katial R, Albers F, Prazma C et al. P504 Comparaison of the gut microbiome between food protein-induced enterocolitis syndrom (FPIES) infants and allergy-free infants. Ann Allergy Asthma Immunol 2017; 119 :e3.13. Su K-W, Cetinbas M, Martin VM, Virkud Y V., Seay H, Ndahayo R et al. Early infancy dysbiosis in food protein‐induced enterocolitis syndrome A prospective cohort study. Allergy 2023; 78 :1595–1604.14. Caparrós E, Cenit MC, Muriel J, Benítez-Páez A, Moreno MV, González-Delgado P et al. Intestinal microbiota is modified in pediatric food protein–induced enterocolitis syndrome. J Allergy Clin Immunol Glob 2022; 1 :217–224.15. Lijing X, Mengyao Z, Jing L, Yang L, Xiaoli X. Analysis of the characteristics of intestinal microbiota after oral tolerance in infants with food protein–induced proctocolitis. Front Pediatr 2024; 12 :1338294.16. Castro AM, Sabater C, Navarro S, Rodriguez S, Molinos C, Jiménez S et al. The intestinal microbiome of infants with cow’s milk-induced FPIES is enriched in taxa and genes of enterobacteria. J Pediatr Gastroenterol Nutr 2024;:841–849.17. Xiong J, Ma YJ, Liao XS, Li LQ, Bao L. Gut microbiota in infants with food protein enterocolitis. Pediatr Res 2024;:1–11.18. Lemoine A, Kapel N, Nicolis I, Tounian P, Bruneau A, Kapandji N et al. Identification of gut biomarkers of FPIES in a longitudinal comparative pediatric study. Allergy 2025; 80 :1389–1399.19. Mayeur C, Gratadoux J-J, Bridonneau C, Chegdani F, Larroque B. Faecal D/L Lactate Ratio Is a Metabolic Signature of Microbiota Imbalance in Patients with Short Bowel Syndrome. PLoS One 2013; 8 :e54335.20. Escudié F, Auer L, Bernard M, Mariadassou M, Cauquil L, Vidal K et al. FROGS: Find, Rapidly, OTUs with Galaxy Solution. Bioinformatics 2018; 34 :1287–1294.21. Noordine ML, Seyoum Y, Bruneau A, Baye K, Lefebvre T, Cherbuy C et al. The microbiota and the host organism switch between cooperation and competition based on dietary iron levels. Gut Microbes 2024; 16 :1–19.22. GraphPad Prism 10.0.03 for Windows, GraphPad Software, Boston, Massachussetts USA. 23. R Core team (2024). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org. 24. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J R Stat Soc Ser B 1995; 57 :289–300.25. Oksanen J, Simpson GL, Blanchet GF, Kindt R, Legendre P, Minchin PR et al. Vegan: Community Ecology Package. Published Online First: 2022.https://github.com/vegandevs/vegan26. Mcmurdie PJ, Holmes S. phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. PLoS One 2013; 8 :e61217.27. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol 2014; 15 :1–21.28. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS et al. Metagenomic biomarker discovery and explanation. Genome Biol 2011; 12 :R60.29. Lin H, Peddada S Das. Analysis of compositions of microbiomes with bias correction. Nat Commun 2020; 11 :3514.30. Gloor GB, Macklaim JM, Fernandes AD. Displaying Variation in Large Datasets: Plotting a Visual Summary of Effect Sizes. J Comput Graph Stat 2016; 25 :971–979.31. Mallick H, Rahnavard A, McIver LJ, Ma S, Zhang Y, Nguyen LH et al. Multivariable association discovery in population-scale meta-omics studies. PLoS Comput Biol 2021; 17 :e1009442.32. Ghosh S, Pramanik S. Structural diversity, functional aspects and future therapeutic applications of human gut microbiome. Arch Microbiol 2021; 203 :5281–5308.33. Berni Canani R, Sangwan N, Stefka AT, Nocerino R, Paparo L, Aitoro R et al. Lactobacillus rhamnosus GG-supplemented formula expands butyrate-producing bacterial strains in food allergic infants. ISME J 2016; 10 :742–750.34. Chen CC, Chen KJ, Kong MS, Chang HJ, Huang JL. Alterations in the gut microbiotas of children with food sensitization in early life. Pediatr Allergy Immunol 2016; 27 :254–262.35. Savage JH, Lee-Sarwar KA, Sordillo J, Bunyavanich S, Zhou Y, O’Connor G et al. A prospective microbiome-wide association study of food sensitization and food allergy in early childhood. Allergy 2018; 73 :145–152.36. Fazlollahi M, Chun Y, Grishin A, Wood RA, Burks AW, Dawson P et al. Early-life gut microbiome and egg allergy. Allergy 2018; 73 :1515–1524.37. Goldberg MR, Mor H, Magid Neriya D, Magzal F, Muller E, Appel MY et al. Microbial signature in IgE-mediated food allergies. Genome Med 2020; 12 :92.38. Martin VM, Virkud Y V., Dahan E, Seay HL, Itzkovits D, Vlamakis H et al. Longitudinal disease-associated gut microbiome differences in infants with food protein-induced allergic proctocolitis. Microbiome 2022; 10 :154.39. Ling Z, Li Z, Liu X, Cheng Y, Luo Y, Tong X et al. Altered fecal microbiota composition associated with food allergy in infants. Appl Environ Microbiol 2014; 80 :2546–2554.40. Díaz M, Guadamuro L, Espinosa-Martos I, Mancabelli L, Jiménez S, Molinos-Norniella C et al. Microbiota and derived parameters in fecal samples of infants with Non-IgE cow’s milk protein allergy under a restricted diet. Nutrients 2018; 10 :1481.41. Berni Canani R, De Filippis F, Nocerino R, Paparo L, Scala C Di, Cosenza L et al. Gut microbiota composition and butyrate production in children affected by non-IgE-mediated cow’s milk allergy. Sci Rep 2018; 8 :12500.42. De Filippis F, Paparo L, Nocerino R, Della Gatta G, Carucci L, Russo R et al. Specific gut microbiome signatures and the associated pro-inflamatory functions are linked to pediatric allergy and acquisition of immune tolerance. Nat Commun 2021; 12 :5958.43. Zubeldia-Varela E, Barker-Tejeda TC, Mera-Berriatua L, Bazire R, Cabrera-Freitag P, Ubeda C et al. Further Insights into the Gut Microbiota of Cow’s Milk Allergic Infants: Analysis of Microbial Functionality and Its Correlation with Three Fecal Biomarkers. Int J Mol Sci 2023; 24 :9247.44. Castro AM, Gutiérrez‐Díaz I, Saiz ML, Navarro S, Suárez M, Carbajal I et al. Gut microbiota and inflammatory mediators differentiate IgE mediated and non‐IgE mediated cases of cow’s milk protein at diagnosis. J Pediatr Gastroenterol Nutr 2024; 78 :836–845.45. Crost EH, Coletto E, Bell A, Juge N. Ruminococcus gnavus: friend or foe for human health. FEMS Microbiol Rev 2023; 47 :1–23.46. Liu X, Mao B, Gu J, Wu J, Cui S, Wang G et al. Blautia—a new functional genus with potential probiotic properties? Gut Microbes 2021; 13 :1–21.47. Nie K, Ma K, Luo W, Shen Z, Yang Z, Xiao M et al. Roseburia intestinalis: A Beneficial Gut Organism From the Discoveries in Genus and Species. Front Cell Infect Microbiol 2021; 11 :1–15.48. Mclellan P, Auger S, Goudiaby MT, Brot L, Benech N, Grill JP et al. Faecalibacterium Diversity in the Gut Microbiome of Crohn’s Disease Patients. United Eur Gastroenterol J 2025;:1–11.49. Tan JK, Macia L, Mackay CR. Dietary fiber and SCFAs in the regulation of mucosal immunity. J Allergy Clin Immunol 2023; 151 :361–370.50. Zhu L Bin, Zhang YC, Huang HH, Lin J. Prospects for clinical applications of butyrate-producing bacteria. World J Clin Pediatr 2021; 10 :84–92.51. Dzidic M, Abrahamsson TR, Artacho A, Collado MC, Mira A, Jenmalm MC. Oral microbiota maturation during the first 7 years of life in relation to allergy development. Allergy 2018; 73 :2000–2011.52. Wang Y, Zhang Q, Wang C, Chen J, Cheng Z, Luo L et al. Variations in oral microbiota and salivary proteomics reveal distinct patterns in polysensitized individuals. Allergy 2022; 77 :1899–1902.53. Beheshti R, Stone S, Chandran D, Hicks SD. Multi-Omic Profiles in Infants at Risk for Food Reactions. Genes (Basel) 2022; 13 :2024.54. Boyer J, Sgambelluri L, Yuan Q. Association of Antibiotic Usage With Food Protein-Induced Enterocolitis Syndrome Development From a Caregiver’s Survey. JPGN Reports 2021; 2 :e132. Figure Legends Figure 1: Flow chart Figure 2: Gut microbiota in allergic patients compared to matched controls 1. Taxonomic relative abundances at phyla level 2. Taxonomic relative abundances at genus level (Top 10) 3. Alpha-diversity (Shannon) 38 pairs, Wilcoxon test, p-value<0.01 Less abundant genera in allergic patients than in controls-A (Venn diagram) More abundant genera in allergic patients than in controls-A (Venn diagram) Genera are specified if they are common in at least 2 differential compositional analyses among LEfSe, DeSeq2, ANCOM-BC, ALDEx2, MaAsLIN2. Relative abundances of short-chain fatty acids 38 pairs, Acetate: Wilcoxon text, p=0.03; Butyrate: paired t-test, p=0.07; Propionate: Wilcoxon text, p=0.81; Valerate: Wilcoxon text, p=0.21; Isobutyrate, Wilcoxon text, p=0.04; Isovalerate, Wilcoxon text, p=0.04 Figure 3: Salivary microbiota in allergic patients compared to matched controls 1. Taxonomic relative abundances at phyla level 2. Taxonomic relative abundances at genus level (Top 10) 3. Alpha-diversity (Shannon) 36 pairs, Wilcoxon test, p<0.001 Less abundant genera in allergic patients than in controls-A (Venn diagram) Genera are specified if they are common in at least 2 differential compositional analyses among LEfSe, DeSeq2, ANCOM-BC, ALDEx2, MaAsLIN2. No shared increased genera in allergic patients compared to controls-A. Figure 4: Gut microbiota in allergic patients compared to tolerant patients, and controls 1. Taxonomic relative abundances at phyla level in patients 2. Taxonomic relative abundances at genus level in patients 3. Alpha-diversities (Shannon) Allergic versus tolerant patients: 22 pairs, paired t-test, p<0.01; Controls over time: 22 pairs, paired t-test, p=0.001; Tolerant patients versus controls-T: 22 pairs, paired t-test, p=0.045. Less abundant genera in allergic patients than in tolerant patients (Venn diagram) More abundant genera in allergic patients than in tolerant patients (Venn diagram) Genera are specified if they are common in at least 2 differential compositional analyses among LEfSe, DeSeq2, ANCOM-BC, ALDEx2, MaAsLIN2. Relative abundances of short-chain fatty acids in allergic and tolerant patients 22 pairs; Acetate, butyrate: paired t-tests, p<0.001; Propionate, paired t-test, p=0.49; Valerate, Wilcoxon test, p=0.005; Isobutyrate, Wilcoxon test, p=0.02; Isovalerate, Wilcoxon test, p=0.01. Figure 5: Salivary microbiota in allergic patients compared to tolerant patients, and controls 1. Taxonomic relative abundances at phyla level in patients 2. Taxonomic relative abundances at genus level in patients (Top 10) 3. Alpha-diversities (Shannon) Allergic versus tolerant patients: 22 pairs, Wilcoxon text, p=0.003; Controls over time: 21 pairs, Wilcoxon test, p=0.10; Tolerant patients versus controls-T: 21 pairs, paired t-test, p>0.99. Less abundant genera in allergic patients than in tolerant patients (Venn diagram) More abundant genera in allergic patients than in tolerant patients (Venn diagram) Genera are specified if they are common in at least 2 differential compositional analyses among LEfSe, DeSeq2, ANCOM-BC, ALDEx2, MaAsLIN2. Supplemental figures Figure S1 : Gut microbiota in allergic patients and matched controls Alpha-diversity (Chao1) 38 pairs, paired t-test, p-value<0.001 Alpha-diversity (invSimpson) 38 pairs, Wilcoxon test, p-value=0.03 Bacterial load 38 pairs, Wilcoxon test, p-value=0.50 Beta-diversity (JSD) Beta-diversity (Unifrac) Allergic patients in red, controls in blue. Lines between pairs and numbers of pairs. Figure S2 : Salivary microbiota in allergic patients and matched controls Alpha-diversity (Chao1) 36 pairs, Wilcoxon test, p-value=0.13 Alpha-diversity (invSimpson) 36 pairs, Wilcoxon test, p-value<0.0001 Bacterial load 36 pairs, Wilcoxon test, p-value=0.54 Beta-diversity (JSD) Beta-diversity (Unifrac) Allergic patients in red, controls in blue. Lines between pairs. Figure S3 : Fecal microbiota at baseline and during follow-up Alpha-diversities (Chao1) Allergic versus Tolerant patients: 22 pairs, paired t-test, p-value=0.01 Tolerant versus Controls-T: 22 pairs, paired t-test, p-value=0.14 Controls-A versus Controls-T: 22 pairs, paired t-test, p-value<0.001 Alpha-diversities (invSimpson) Allergic versus Tolerant patients: 22 pairs, paired t-test, p-value=0.01 Tolerant versus Controls-T: 22 pairs, Wilcoxon test, p-value=0.04 Controls-A versus Controls-T: 22 pairs, Wilcoxon test, p-value=0.007 Bacterial loads Allergic versus tolerant patients: 22 pairs, Wilcoxon test, p-value=0.11 Tolerant versus Controls-T: 22 pairs, paired t-test, p-value=0.03 Controls-A versus Controls-T: 22 pairs, paired t-test, p-value=0.29 Beta-diversity (JSD) in allergic and tolerant patients Beta-diversity (Unifrac) in allergic and tolerant patients Allergic patients in red, tolerant patients in green. Lines between pairs and numbers of pairs. Beta-diversity (JSD) in tolerant patients and matched controls Beta-diversity (Unifrac) in tolerant patients and matched controls Tolerant patients in green, Controls-T in violet. Lines between pairs and numbers of pairs. More abundant genera in tolerant patients than in matched controls (Venn diagram) Less abundant genera in tolerant patients than in matched controls (Venn diagram) Genera are specified if they are common in at least 2 differential compositional analyses among LEfSe, DeSeq2, ANCOM-BC, ALDEx2, MaAsLIN2. Relative abundances of short-chain fatty acids in tolerant patients and matched control 22 pairs; Acetate: paired t-test, p=0.26; Butyrate: paired t-test, p=0.79; Propionate, paired t-test, p=0.30; Valerate, Wilcoxon test, p=0.97; Isobutyrate, paired t-test, p=0.04; Isovalerate, paired t-test, p=0.46. Figure S4 : Salivary microbiota at baseline and during follow-up Alpha-diversities (Chao1) Allergic versus Tolerant patients: 22 pairs, Wilcoxon test, p-value=0.02 Tolerant versus Controls-T: 21 pairs, paired t-test, p-value=0.58 Controls-A versus Controls-T: 21 pairs, paired t-test, p-value=0.06 Alpha-diversity (invSimpson) Allergic versus Tolerant patients: 22 pairs, paired t-test, p-value=0.003 Tolerant versus Controls-T: 21 pairs, paired t-test, p-value=0.43 Controls-A versus Controls-T: 21 pairs, paired t-test, p-value=0.31 Bacterial loads Allergic versus Tolerant patients: 22 pairs, Wilcoxon test, p-value=0.07 Tolerant versus Controls-T: 21 pairs, Wilcoxon test, p-value=0.02 Controls-A versus Controls-T: 21 pairs, Wilcoxon test, p-value=0.26 Beta-diversity (JSD) in allergic and tolerant patients Beta-diversity (Unifrac) in allergic and tolerant patients Allergic patients in red, tolerant patients in green. Lines between pairs and numbers of pairs. Beta-diversity (JSD) in tolerant patients and controls-T Beta-diversity (Unifrac) in tolerant patients and controls-T Tolerant patients in green, controls in violet. Lines between pairs and numbers of pairs. Information & Authors Information Version history Peer review timeline Published Allergy Version of Record28 Apr 2026Published Copyright This work is licensed under a Non Exclusive No Reuse License. Collection

Authors Metrics & Citations Metrics Article Usage 355views 236downloads Citations Download citation Anaïs Lemoine, Karine Patient, Nathalie Kapel, et al. Longitudinal follow-up of gut and salivary microbiota in children with food protein-induced enterocolitis syndrome. Authorea. 01 August 2025. DOI: https://doi.org/10.22541/au.175405345.52856923/v1 DOI: https://doi.org/10.22541/au.175405345.52856923/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.