Gut Microbiota Dysbiosis Associated With Altered Production of Short Chain Fatty Acids in Children With Neurodevelopmental Disorders

Abstract

While gut microbiota dysbiosis has been linked with autism, its role in the etiology of other neurodevelopmental disorders (NDD) is largely underexplored. To our knowledge this is the first study to evaluate gut microbiota diversity and composition in 36 children from the Republic of Serbia diagnosed with NDD and 28 healthy children. The results revealed an increased incidence of potentially harmful bacteria, closely related to Clostridium species, in the NDD patient group compared to the Control group: Desulfotomaculum guttoideum (P < 0.01), Intestinibacter bartlettii (P < 0.05), and Romboutsia ilealis (P < 0.001). On the other hand, significantly lower diversity of common commensal bacteria in the NDD group of patients was noticed. Enterococcus faecalis (P < 0.05), Enterococcus gallinarum (P < 0.01), Streptococcus pasteurianus (P < 0.05), Lactobacillus rhamnosus (P < 0.01) and Bifidobacteria sp. were detected in lower numbers of patients or were even absent in some NDD patients. In addition, butyrate-producing bacteria Faecalibacterium prausnitzii (P < 0.01), Butyricicoccus pullicaecorum (P < 0.05), and Eubacterium rectale (P = 0.07) were less frequent in the NDD patient group. In line with that, the levels of fecal short chain fatty acids (SCFAs) were determined. Although significant differences in SCFA levels were not detected between NDD patients and the Control group, a positive correlation was noted between number of rDNA amplicons obtained with universal primers and level of propionic acid, as well as a trend for levels of total SCFAs and butyric acid in the Control group. This correlation is lost in the NDD patient group, indicating that NDD patients' microbiota differs from the microbiota of healthy children in the presence or number of strong SCFA-producing bacteria. According to a range-weighted richness index it was observed that microbial diversity was significantly lower in the NDD patient group. Our study reveals that the intestinal microbiota from NDD patients differs from the microbiota of healthy children. It is hypothesized that early life microbiome might have an impact on GI disturbances and accompanied behavioral problems frequently observed in patients with a broad spectrum of NDD.

Keywords: Bifidobacterium; Clostridium like species; Lactobacillus; autism; gut-brain axis; microbial diversity.

Figure 1

The identity of the rDNA clones obtained from DGGE rDNA amplicons related to the individual children. The numbers on the y-axis correlate to the numbers of the rDNA amplicons observed. The rDNA amplicons whose incidences were statistically different between NDD patient and Control groups (or if trend existed) were excised, cloned, and sequenced and identities of those bacterial species are displayed. The numbers on the x-axis correspond to the numbers of lanes. Each lane represents sample of an individual child. Total of 64 subjects (36 NDD patients and 28 healthy controls) was analyzed. Each symbol represents a corresponding band on DGGE gel. MSDD, mixed specific developmental disorder; PDD-NOS, pervasive developmental disorder-non-specified; ERLD, expressive and receptive language disorder; CHA, childhood autism; NDD, neurodevelopmental disorder.

Figure 2

Box-plot diagram based on Dice similarity coefficient. Homogeneities of the groups and similarities between the groups were evaluated by Dice similarity coefficient (%). Homogeneities of the groups were evaluated by comparing DGGE profiles obtained by universal, Lactobacillus (LB)-, Bifidobacteria (BB)-specific primers and all three sets of primers together (A–D) within Control (CTRL) and particular NDD groups. Similarities between particular NDD groups and similarities of NDD groups with the CTRL group were evaluated by comparing DGGE profiles of samples from particular NDD groups, obtained by universal, LB-, BB-specific primers and all three sets of primers together, between NDD groups (E–H) and with DGGE profiles of the CTRL group (I–L), respectively. Relevant statistical values of Dice analysis are presented in Supplementary Table 3. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 3

Microbial diversity comparison. Student's unpaired t-test was used to compare mean values of range-weighted richness index (Rr) derived from DGGE patterns obtained with universal (Uni) primers (A) and Lactobacillus (LB)-specific primers (B) between NDD patient and Control group. Due to absence of normal distribution of range-weighted richness indexes derived from DGGE patterns obtained with Bifidobacteria (BB)-specific primers, difference between groups was assessed by non-parametric Mann–Whitney U test (C). Comparison of mean values of Rr derived from DGGE patterns obtained with universal (D) and LB-specific primers (E) between particular NDD diagnoses and Control group was performed with one-way ANOVA followed by Tukey post-hoc test, while for comparison of Rr derived from DGGE patterns obtained with BB-specific primers Kruskal-Wallis test was used (F). MSDD- mixed specific developmental disorder, PDD-NOS- pervasive developmental disorder-non-specified, ERLD, expressive and receptive language disorder; CHA, childhood autism; NDD, neurodevelopmental disorder. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 4

Amounts of SCFAs in fecal samples of NDD patients and healthy children. The amounts of acetic acid (A), propionic acid (B), butyric acid (C), and total SCFAs (D) in NDD, Control group and particular NDD diagnoses groups were presented in mmol/kg of feces. Quantification of SCFAs in fecal samples was performed by HPLC analysis. Results are presented as Mean ± SEM. MSDD, mixed specific developmental disorder; PDD-NOS, pervasive developmental disorder-non-specified; ERLD, expressive and receptive language disorder; CHA, childhood autism; NDD, neurodevelopmental disorder.

Figure 5

Correlation between average number of rDNA amplicons and amounts of SCFAs in Control and NDD patient group. Average number of rDNA amplicons obtained with universal primers was correlated with amount of propionic acid (A), butyric acid (B), and total SCFAs (C) in Control group and in NDD group (D–F). Each dot corresponds to one subject. The line denotes linear regression of individual data point; r stands for Pearson's correlation coefficient. P < 0.05 was considered statistically significant; NDD, neurodevelopmental disorder.