Higher levels of Bifidobacteria and tumor necrosis factor in children with drug-resistant epilepsy are associated with anti-seizure response to the ketogenic diet

Abstract

Background: Recently, studies have suggested a role for the gut microbiota in epilepsy. Gut microbial changes during ketogenic diet (KD) treatment of drug-resistant epilepsy have been described. Inflammation is associated with certain types of epilepsy and specific inflammation markers decrease during KD. The gut microbiota plays an important role in the regulation of the immune system and inflammation.

Methods: 28 children with drug-resistant epilepsy treated with the ketogenic diet were followed in this observational study. Fecal and serum samples were collected at baseline and three months after dietary intervention.

Findings: We identified both gut microbial and inflammatory changes during treatment. KD had a general anti-inflammatory effect. Novel bioinformatics and machine learning approaches identified signatures of specific Bifidobacteria and TNF (tumor necrosis factor) associated with responders before starting KD. During KD, taxonomic and inflammatory profiles between responders and non-responders were more similar than at baseline.

Interpretation: Our results suggest that children with drug-resistant epilepsy are more likely to benefit from KD treatment when specific Bifidobacteria and TNF are elevated. We here present a novel signature of interaction of the gut microbiota and the immune system associated with anti-epileptic response to KD treatment. This signature could be used as a prognostic biomarker to identify potential responders to KD before starting treatment. Our findings may also contribute to the development of new anti-seizure therapies by targeting specific components of the gut microbiota.

Funding: This study was supported by the Swedish Brain Foundation, Margarethahemmet Society, Stiftelsen Sunnerdahls Handikappfond, Linnea & Josef Carlssons Foundation, and The McCormick Genomic & Proteomic Center.

Keywords: Bifidobacterium; Bioinformatics; Epilepsy; Gut microbiota; Inflammation; Ketogenic diet; Machine learning; Omics; TNF.

Figure 1

Bacterial taxa contributing to the final models ranked by feature importance. A) Comparing all patients before vs during KD and B) responders vs. non-responders before KD. A) showing feature importance of taxa positively or negatively associated with KD and B) showing feature importance of taxa positively or negatively associated with anti-epileptic response to KD.

Figure 2

Inflammation profiles in children with refractory epilepsy and changes during the ketogenic diet. Volcano plots showing differences in inflammation markers (A) due to KD treatment or (B) comparing responders to non-responders before starting KD, and (C) after three months of treatment. Blue dots depict significant changes (p < 0.05), green dots represent markers with a log2 fold change of >0.3, while red dots fulfil both criteria and black dots none. D-F) show boxplots of TNF, IL-12B and MCP-4, respectively, for responders before KD (R_Ep1) and during KD (R_Ep2) as well as for non-responders before KD (NR_Ep1) and during KD (NR_Ep2). A centre line represents the median; box limits are upper and lower quartiles; whiskers show 1.5 × interquartile range; and points depict outliers. Asterisks indicate significance of Welch's t-test as follows: *** p < 0.001, ** p < 0.01, *p < 0.05.

Figure 3

Multivariate PLS-DA of associations between inflammation markers and gut microbes. In A) PLS-DA for all patients with metagenomics and inflammation profiles was performed using dietary treatment (before vs. during) as the discriminating factor. The contributing metagenomic variables are depicted in yellow and the contributing inflammation variables are grey. Strong positive correlations between variables of each dataset are shown as purple lines and strong negative correlations are pink. For visualization purposes only the top correlations were chosen in each figure. For A) a cut-off of 0.6 was chosen, while 0.7 was used in Figure B-E. In B) the same analysis was performed as in A) but for responders only and in C) for non-responders only. Circos plots of PLS-DA models comparing responders to non-responders before KD and during are shown in D) and E), respectively. In F) boxplots of levels of CCL3, CCL4 and PD-L1 are shown for responders before KD (R_Ep1) and during KD (R_Ep2) as well as for non-responders before KD (NR_Ep1) and during KD (NR_Ep2) and p-values from t-tests for each comparison were added above. Boxplots show a centre line representing the median; box limits are upper and lower quartiles; whiskers show 1.5 × interquartile range; and points depict outliers.

Figure 4

Responder-specific Bifidobacterium profiles. Hierarchical clustering of Pearson's correlations of Bifidobacteria in patients before starting KD is shown in (A). Spearman's correlation of members of the B. longum cluster with age are shown in B). Colour gradients in B) indicate the nature and strength of each correlation (Spearman's r), while Asterisks indicate significance as follows: *** p < 0.001, ** p < 0.01, * p < 0.05. Boxplots for each member of the B. longum cluster in children over eight years of age for responders before KD (R_Ep1) and during KD (R_Ep2) as well as for non-responders before KD (NR_Ep1) and during KD (NR_Ep2) with a centre line representing the median; box limits, are upper and lower quartiles; whiskers show 1.5 × interquartile range; and points depict outliers. Numbers above boxplots indicate p-values of Welch's t-test.