Targeting the microbiome in pediatric migraine: gastrointestinal manifestations and the therapeutic role of Bifidobacterium longum

Abstract

Migraine is a disabling neurological disorder that often begins in childhood or adolescence and is frequently accompanied by gastrointestinal (GI) symptoms. However, the microbiota signatures and gut-brain interactions underlying pediatric migraine, particularly in the presence of GI disorder, remain poorly defined. This study aimed to explore the clinical and microbial features of pediatric migraine, as well as the therapeutic potential of probiotics.We prospectively enrolled 126 pediatric migraine patients (ages 6-19) with or without GI disorder and 50 age-matched healthy controls. Fecal microbiota was profiled using 16S rRNA sequencing. Patients with migraine were stratified based on Rome IV-defined GI disorders and evaluated for headache characteristics, PedMIDAS scores (disability assessment), plasma calcitonin gene related peptide (CGRP, thought as a key biomarker of migraine), cytokines, and fecal calprotectin. Probiotic effects were tested in both young (3-4 weeks) and adult capsaicin-induced migraine rat models, and an exploratory pilot study involving 23 pediatric migraine patients.Compared to controls, migraine patients exhibited distinct gut microbiota with reduced Bifidobacterium longum. and elevated Bacteroides. GI disorders were present in 46.8% of migraine patients and were associated with significantly higher rates of abdominal pain (50% vs. 13%, p <0.001), greater migraine-related disability (PedMIDAS: 60 ± 13.2 vs. 29 ± 7.0, p = 0.042), elevated fecal calprotectin, and enrichment of Streptococcus gallolyticus. In contrast, Faecalibacterium prausnitzii, positively correlated with B. longum, was linked to milder symptoms and shorter disease duration in migraine patients without GI disorder. In animal models, B. longum attenuated trigeminal activation in both young and adult rats. An exploratory pilot study showed B. longum supplementation led to reductions in headache days, intensity, and frequency. These findings reveal distinct gut microbial signatures in pediatric migraine, and identify B. longum as a promising microbiota-targeted therapeutic strategy. Our work highlights the therapeutic potential of modifying the gut-brain axis in childhood migraine.

Keywords: Faecalibacterium prausnitzii; Gut microbiota; Streptococcus gallolyticus; gastrointestinal disorder; gut–brain axis; pediatric migraine.

Graphical abstract

Figure 1

Workflow of the present study. N, number of patients recruited; n, number of stool or blood samples.

Figure 2.

Richness and diversities of gut microbiota in migraine patients and healthy controls. (A) Shannon and chao1 indexes accessed microbial richness at the amplicon sequence variant (ASV) level. (B) Principal coordinates analysis (PCoA) using bray-cutis distance matrix distinguishes microbial diversities between groups. (C, D) Linear discriminant analysis (LDA) effect size (LEfSe) identified discriminating taxa at genus (C) and species (D) levels (p <0.05 and LDA > 4.5 considered significant). B. longum, Bifidobacterium longum; F. saccharivorans, Fusicatenibacter saccharivorans; R. timonensis, Romboutsia timonensis; E. hallii, Eubacterium hallii; P. merdae, Parabacteroides merdae; B. stercoris, Bacteroides stercoris; B. uniformis, Bacteroides uniformis; B. plebeius, Bacteroides plebeius; B. dorei, Bacteroides dorei. Statistical significance using FDR correction: * q <0.05; ** q <0.01; *** q <0.001. (E, F) Relative abundances (log₁₀ ± SEM) of taxa enriched in healthy controls (E) or migraine patients (F). *p <0.05, **p <0.005, ***p <0.0005 (Mann-Whitney U-test).

Figure 3.

Effect of B. longum on capsaicin-induced neuronal activation in the trigemino-cervical complex (TCC) and trigeminal ganglia (TG). (A) Experimental design: male Wistar rats (adult 7-8 wk, young 3-4 wk) were pretreated for 2 wk with B. longum (3 × 10⁹ CFU adults; 1.5 × 10⁹ CFU young) or BbLa (250 mg containing B. bifidum (10⁹ CFU) and Lactobacillus acidophilus (10⁹ CFU) adults; 125 mg young) via oral gauge. Rats (n = 4-6/group) then received intracisternal capsaicin (10 nmol). (B) Schematic of migraine pathology showing increased c-Fos and CGRP in the trigeminal system. (C) c-Fos immunohistochemistry (TCC) and CGRP immunofluorescence (TG). Sham rats received vehicle instead of capsaicin. Three TCC and TG sections was analyzed per rat. Scale bar: 100 µm. (D) Quantification of c-Fos-immunoreactive TCC neurons and CGRP density in TG sections. CGRP density (%) = (activated cell area/full cell area) × 100% in V1 + V2 region. Data are means ± SE. *p <0.05, **p <0.005 (Mann-Whitney U-test).

Figure 4.

B. longum alters gut microbiota composition and diversity in a migraine rat model. (A, B) PCoA plots based on bray-curtis distance matrix showing beta diversity of fecal microbiota in young (A) and adult (B) rats. (C, D) Genus-level taxa with LDA thresholds of 4.5 (young, C) and 4 (adult, D). (E) Heatmap of Spearman rank correlations coefficient r between Bifidobacterium and other microbiota genera detected in the fecal samples of young and adult rats. (F) Linear regression depicts the negative correlation between Bifidobacterium abundance and total c-Fos-ir cell number in TCC sections of young and adult rats that fed with B. longum. *p <0.01, **p <0.001, ***p <0.0001 (Mann-Whitney U-test).

Figure 5.

Bacterial diversity and clinical correlations between migraine patients with and without GI disorders. (A) Shannon and Chao1 indexes at the ASV level. (B) LEfSe identified discriminating taxa (p <0.05 and LDA > 4): S. gallolyticus, Streptococcus gallolyticus; A. inops, Alistipes inops; R. torques, Ruminococcus torques; F. saccharivorans, Fusicatenibacter saccharivorans; F. prausnitzii, Faecalibacterium prausnitzii. (C) Spearman correlations between bacteria and clinical factors (pain score, PedMIDAS and disease duration) in patients with (M + G) or without (M + noG) GI disorder. (D) Fecal calprotectin levels among M + G (n = 35), M + noG (n = 49), and healthy controls (n = 48). (E) Heatmap of Spearman correlations between F. prausnitzii and S. gallolyticus/B. longum during non-attack (NA) and attack (A) periods. *p <0.05, **p <0.005, ***p <0.0005 (Mann-Whitney U-test).

Figure 6.

B. longum negatively associates with migraine activity of patients without GI disorders. (A) Heatmaps of Spearman rank correlation r values between gut microbiota and clinical variables indicated. Fecal samples collected during attack phase of migraine patients with GI disorders (M + G) and those without GI disorders (M + noG) were analyzed. *p <0.05, **p <0.005 by Mann-Whitney U-test. (B) Linear regression plots depicting correlations between B. longum and pain score (left) or disease duration (right) of M + noG patients during attack phase. (C) Violin plots show the range and distribution of pain score (upper) and disease duration (lower) of patients with M + G or M + noG. (D) A comparison of the relative abundance (%) of B. longum detected during non-attack (NA) and attack (A) stages of M + G and M + noG patients.

Figure 7.

Therapeutic effects of B. longum in pediatric migraine patients compared to the group using B. bifidum/L. acidophilus. Over the 12-week study period, patients treated with B. longum (BL, n = 17) demonstrated significant improvements in clinical outcomes by least-squares (LS) mean comparison, including significant reduction of pain scores after treatment (p = 0.025) (A-B), and weekly headache days (p = 0.014) (C-D). No significant changes were observed in the group taking B. bifidum/L. acidophilus (BbLa, n = 6) for any parameters (all p > 0.05) (A-D). *p <0.05 by the Wald test. The beneficial effects of BL on weekly trends in the median number of headache days (E) and attack frequencies (F) are shown.