Exploring the Fecal Metabolome in Infants With Cow's Milk Allergy: The Distinct Impacts of Cow's Milk Protein Tolerance Acquisition and of Synbiotic Supplementation

Abstract

Scope: Cow's milk allergy (CMA) is one of the most prevalent food allergies in early childhood, often treated via elimination diets including standard amino acid-based formula or amino acid-based formula supplemented with synbiotics (AAF or AAF-S). This work aimed to assess the effect of cow's milk (CM) tolerance acquisition and synbiotic (inulin, oligofructose, Bifidobacterium breve M-16 V) supplementation on the fecal metabolome in infants with IgE-mediated CMA.

Methods and results: The CMA-allergic infants received AAF or AAF-S for a year during which fecal samples were collected. The samples were subjected to metabolomics analyses covering gut microbial metabolites including SCFAs, tryptophan metabolites, and bile acids (BAs). Longitudinal data analysis suggested amino acids, BAs, and branched SCFAs alterations in infants who outgrew CMA during the intervention. Synbiotic supplementation significantly modified the fecal metabolome after 6 months of intervention, including altered purine, BA, and unsaturated fatty acid levels, and increased metabolites of infant-type Bifidobacterium species: indolelactic acid and 4-hydroxyphenyllactic acid.

Conclusion: This study offers no clear conclusion on the impact of CM-tolerance acquisition on the fecal metabolome. However, our results show that 6 months of synbiotic supplementation successfully altered fecal metabolome and suggest induced bifidobacteria activity, which subsequently declined after 12 months of intervention.

Keywords: bifidobacteria; early life; food allergy; fructooligosaccharides; metabolomics.

FIGURE 1

RM‐ASCA+ combined effect matrix showing the common metabolome development throughout the study for the CM‐allergic (blue solid line, n = 15) and CM‐tolerant (orange dashed line, n = 24) groups as scores (A) and loadings (B). Only the metabolites with 12 highest and 12 lowest loadings are shown in the plot. Error bars representing 95% CI were estimated based nonparametric bootstrapping. CI, confidence interval; CM, cow's milk; RM‐ASCA+, repeated measures analysis of variance simultaneous component analysis+.

FIGURE 2

RM‐ASCA+ interaction effect matrix showing the metabolome differences between the CM‐allergic (blue solid line, n = 15) and CM‐tolerant group (orange dashed line, n = 24) over time as scores (A) and loadings (B). Only the metabolites with 12 highest and 12 lowest loadings are shown in the plot. Error bars representing 95% CI were estimated based nonparametric bootstrapping. CI, confidence interval; CM, cow's milk; RM‐ASCA+, repeated measures analysis of variance simultaneous component analysis+.

FIGURE 3

RM‐ASCA+ interaction effect matrix showing the metabolome differences between the AAF (purple solid line, n = 16) and AAF‐S (green dashed line, n = 23) group over time as scores (A) and loadings (B). Only the metabolites with 12 highest and 12 lowest loadings are shown in the plot. Error bars representing 95% CI were estimated based nonparametric bootstrapping. AAF, amino acid‐based formula; AAF‐S, amino acid‐based formula with synbiotic; CI, confidence interval; RM‐ASCA+, repeated measures analysis of variance simultaneous component analysis+.

FIGURE 4

Volcano plot showing the resulting p value of the interaction coefficient for TP1 (left) and TP2 (right) in intervention LMM, dashed (p = 0.05), solid line (Q = 0.1) for TP1 (A) and TP2 (B). Red symbols indicate metabolites with Q < 0.1 after Benjamini–Hochberg procedure. LMM, linear mixed model.

FIGURE 5

Spearman's rank correlations between the changes in Bifidobacterium and ILA, 4‐OH‐PLA#, glutamine in AAF (purple solid line, n = 16) and AAF‐S (green dashed line, n = 23) groups from baseline to TP1 (TP1–TP0) and TP2 (TP2–TP0). The rank of the changes in metabolite response and relative abundance of Bifidobacterium within each group were used for plotting. The figure shows p values; the Q values after Benjamini–Hochberg procedure are provided in Table S9. 4‐OH‐PLA#, 4‐hydoxyphenyllactic acid; AAF, amino acid‐based formula; AAF‐S, amino acid‐based formula with synbiotic; ILA, indolelactic acid.