Duodenal and faecal microbiota of celiac children: molecular, phenotype and metabolome characterization

Abstract

Background: Epidemiology of celiac disease (CD) is increasing. CD mainly presents in early childhood with small intestinal villous atrophy and signs of malabsorption. Compared to healthy individuals, CD patients seemed to be characterized by higher numbers of Gram-negative bacteria and lower numbers Gram-positive bacteria.

Results: This study aimed at investigating the microbiota and metabolome of 19 celiac disease children under gluten-free diet (treated celiac disease, T-CD) and 15 non-celiac children (HC). PCR-denaturing gradient gel electrophoresis (DGGE) analyses by universal and group-specific primers were carried out in duodenal biopsies and faecal samples. Based on the number of PCR-DGGE bands, the diversity of Eubacteria was the higher in duodenal biopsies of T-CD than HC children. Bifidobacteria were only found in faecal samples. With a few exceptions, PCR-DGGE profiles of faecal samples for Lactobacillus and Bifidobacteria differed between T-CD and HC. As shown by culture-dependent methods, the levels of Lactobacillus, Enterococcus and Bifidobacteria were confirmed to be significantly higher (P = 0.028; P = 0.019; and P = 0.023, respectively) in fecal samples of HC than in T-CD children. On the contrary, cell counts (CFU/ml) of presumptive Bacteroides, Staphylococcus, Salmonella, Shighella and Klebsiella were significantly higher (P = 0.014) in T-CD compared to HC children. Enterococcus faecium and Lactobacillus plantarum were the species most diffusely identified. This latter species was also found in all duodenal biopsies of T-CD and HC children. Other bacterial species were identified only in T-CD or HC faecal samples. As shown by Randomly Amplified Polymorphic DNA-PCR analysis, the percentage of strains identified as lactobacilli significantly (P = 0.011) differed between T-CD (ca. 26.5%) and HC (ca. 34.6%) groups. The metabolome of T-CD and HC children was studied using faecal and urine samples which were analyzed by gas-chromatography mass spectrometry-solid-phase microextraction and 1H-Nuclear Magnetic Resonance. As shown by Canonical Discriminant Analysis of Principal Coordinates, the levels of volatile organic compounds and free amino acids in faecal and/or urine samples were markedly affected by CD.

Conclusion: As shown by the parallel microbiology and metabolome approach, the gluten-free diet lasting at least two years did not completely restore the microbiota and, consequently, the metabolome of CD children. Some molecules (e.g., ethyl-acetate and octyl-acetate, some short chain fatty acids and free amino acids, and glutamine) seems to be metabolic signatures of CD.

Figure 1

Clustering of denaturing gradient gel electrophoresis (DGGE) profiles of biopsies from thirty-four children (1-34). Universal V6-V8 (A) and Lac1/Lac2 Lactobacillus group (B) primers were used. Clustering was carried out using the unweighted pair-group method with the arithmetic average (UPGMA) based on the Pearson correlation coefficient. T-CD, treated celiac disease children; and HC, non-celiac children; band a, L. plantarum; band b, human DNA. See materials and methods for correspondence of numbered duodenal biopsies.

Figure 2

Clustering of denaturing gradient gel electrophoresis (DGGE) profiles of faecal samples from thirty-four children (1-34). Universal V6-V8 (A), Lac1/Lac2 Lactobacillus group (B), g- Bifid F/g-BifidRGC Bifidobacterium group (C) primers were used. Clustering was carried out using the unweighted pair-group method with the arithmetic average (UPGMA) based on the Pearson correlation coefficient. T-CD, treated celiac disease children; and HC, non-celiac children. See materials and methods for correspondence of numbered faecal samples.

Figure 3

Cultivable cells (log cfu/g) of the main microbial groups in faecal samples of treated celiac disease (T-CD) children and non-celiac children children (HC). The data are the means of three independent experiments (n = 3). The top and bottom of the box represent the 75th and 25th percentile of the data, respectively. The top and bottom of the error bars represent the 5th and 95th percentile of the data, respectively.

Figure 4

Dendrogram of combined RAPD patterns for Enterococcus using primer P7, P4 and M13. Isolates were from faecal samples of treated celiac disease (T-CD). Cluster analysis was based on the simple matching coefficient and unweighted pair grouped method, arithmetic average. Enterococcus and Lactobacillus isolates (I) are coded based on partial 16S rRNA, recA and pheS gene sequence comparisons and correspond to those of Table 2.

Figure 5

Dendrogram of combined RAPD patterns for Enterococcus using primer P7, P4 and M13. Isolates were from faecal samples of non-celiac children (HC). Cluster analysis was based on the simple matching coefficient and unweighted pair grouped method, arithmetic average. Enterococcus and Lactobacillus isolates (I) are coded based on partial 16S rRNA, recA and pheS gene sequence comparisons and correspond to those of Table 2.

Figure 6

Dendrogram of combined RAPD patterns for Lactobacillus using primer P7, P4 and M13. Isolates were from faecal samples of treated celiac disease (T-CD) (A) and non-celiac children (HC) (B). Cluster analysis was based on the simple matching coefficient and unweighted pair grouped method, arithmetic average. Enterococcus and Lactobacillus isolates (I) are coded based on partial 16S rRNA, recA and pheS gene sequence comparisons and correspond to those of Table 2.

Figure 7

Canonical Discriminant Analysis of Principal Coordinates (CAP) loading coefficient plot of the volatile organic metabolites from faecal (A) and urine (B) samples of treated celiac disease (T-CD) and non-celiac children (HC). Data are the means of three independent experiments (n = 3).