A synthetic bacterial community made of human milk isolated strains is an effective strategy to prevent food allergy in a mouse model

To date, there is no truly effective strategy to prevent food allergies (FA). FA can develop within the first months of life and result from the combined dysfunction of the immune system, the commensal microbiota and epithelial barriers. Early disruption of these systems compromises the induction of immune tolerance.1 In line with the developmental origins of health and disease (DOHaD) concept, breastfeeding plays a key role in their maturation, contributing to infant health and influencing the induction or prevention of allergies.2 Breastmilk contains numerous bioactive molecules, including proteins, lipids and prebiotic oligosaccharides.3 It is also a major source of bacteria that contribute to the establishment of the early intestinal microbiota, thereby supporting immune and epithelial development.4 We previously demonstrated that two synthetic bacterial communities (SynComs AI and HI), combining 11 human milk bacterial strains with similar overall taxonomic composition but distinct functional properties, influence intestinal homeostasis differently, particularly through contrasted effects on immune responses and epithelial barrier functions.5 Building on these findings, we aimed to evaluate whether these bacterial communities could protect against or promote FA when administered during lactation and dietary diversification. We hypothesized that these bacterial communities, by establishing specific microbial, epithelial and immune imprints, would modulate FA outcomes. We tested this hypothesis using a mouse model of wheat-induced FA (Figure 1A). Mouse pups were supplemented for 15 days, from 2 to 4 weeks of age, with SynCom AI (AI), SynCom HI (HI), or bacteria-free medium as control. One week after supplementation, FA was induced by three intraperitoneal injections followed by one oral challenge with wheat allergen (gliadins extract) in all groups, with non-allergic controls (CTL) included as negative controls. Interestingly, supplementation with AI completely abrogated FA symptoms (Figure 1B). In allergic pups, AI avoided the typical temperature drop observed in HI and control groups. Levels of specific IgE and IgG1 (Th2-associated biomarkers) were reduced in AI-supplemented allergic pups compared with CTL allergic pups, and IgG1 levels were also lower than in HI allergic pups. Conversely, the Th1-associated IgG2a was significantly higher in AI-supplemented allergic pups than in CTL or HI allergic pups. No differences in mMCPT1 secretion were observed among the allergic groups (Figure S1). AI supplementation also markedly decreased cellular biomarkers of FA and promoted tolerogenic immune profiles (Figure 1C). Th2 cells in mesenteric lymph nodes and ILC2 cells in Peyer's patches were reduced in the AI allergic group compared with CTL and HI allergic groups. Conversely, regulatory T and B cells increased in mesenteric lymph nodes of AI-supplemented allergic pups. On the contrary, HI was able to promote FA by increasing the drop of temperature (p value <.01) and the frequency of Th2 (p value <.0001). To further characterize the inflammatory or tolerogenic potential of AI and HI, we evaluated IFN-γ and IL-10 production in T and B lymphocytes from PBMC cultures (Figure S2). HI induced a more inflammatory response, whereas AI promoted tolerogenic T and B cell profiles known to protect against FA. AI also protected the intestinal epithelial barrier from FA-induced dysfunction (Figure S3). In allergic pups, paracellular permeability decreased in the AI group compared with HI and control groups, indicating preserved epithelial integrity. Finally, both SynComs AI and HI supplementation induced several changes in microbiota composition. Few of the SynCom bacteria were detected in the mouse microbiota after the supplementation period and/or after sensitization as shown by the presence of Operational taxonomic Unit (OTU) corresponding to these bacteria (>99% of sequence identity), suggesting persistence of some of these synCom bacteria after supplementation was stopped (Table S1). This is in agreement with previous observations using a piglet model where the majority of the SynCom bacteria were present in the intestine of piglets fed SynCom-supplemented formulas, at a low and decreasing abundance from ileum to colon and feces where they were poorly detected.6 Both SynComs (AI, HI) modified gut microbiota composition before FA induction, beyond the presence of SynCom bacteria, as illustrated by changes in α- and β-diversities (Figure 2A,B). After FA induction, sensitization reshaped the microbiota, but clear β-diversity differences were still observed between AI, HI, and CTL groups (Figure 2C). These findings indicate that the protective effect of AI was correlated with a lasting microbial imprint established before allergen exposure. Overall, we demonstrated for the first time that supplementation with a synthetic breastmilk-derived bacterial community can prevent FA in offspring by increasing immune tolerance, reinforcing intestinal epithelial barrier, and establishing a specific gut microbial imprint. We also showed that certain bacterial communities could also be promoting allergy by their proinflammatory properties. Interestingly, while the two bacterial communities (AI and HI) were similar in terms of taxonomic profile, their impact in terms of allergy differed greatly. This was likely related to the strain properties dependence, widely reported and confirmed through our previous characterization of 29 strains isolated from breast milk and belonging to 12 genera and 25 species.5 The strain properties dependence is notably linked to differences in surface components (proteins, peptidoglycan, exopolysaccharides).7, 8 These preliminary results pave the way for clinical studies testing the benefits of infant formulas enriched with specific breastmilk-derived bacterial communities to prevent the development of FA. In addition to these promising results and despite some limitations of this study as the bacterial communities used here do not strictly mimic the composition and dose of the human milk microbiota, this study suggests that the composition of the breast milk microbiota, which varies from one mother to another, could influence the development of allergies in children. Study conceptualization: B.M.A, S.E, M.B; methodology: B.M.A, S.E, M.B, M.L.R, M.D.C, C.L.B., LR; formal analysis and investigation: B.M.A, S.E, M.B, I.L.H.L; resources: B.M.A, M.B: writing (original draft preparation): B.M.A, S.E, M.B; writing (review and editing): B.M.A, M.B; supervision, project administration and funding acquisition: MB. We thank the Genotoul GeT-PlaGe plateform from Toulouse (France) for 16S rRNA sequencing. We are grateful to the INRAE MIGALE bioinformatics facility (MIGALE, INRAE, 2020. Migale bioinformatics Facility, doi:10.15454/1.5572390655343293E12) for providing computing and storage resources. Open access publication funding provided by COUPERIN CY26. This work was funded by the Région Bretagne and Région Pays de la Loire (France) and by the Bretagne Biotechnologie Alimentaire (BBA) Milk Valley association through the PROLIFIC project. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Table S1. Data S1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.