Fecal Supernatant from Adult with Autism Spectrum Disorder Alters Digestive Functions, Intestinal Epithelial Barrier, and Enteric Nervous System

Abstract

Autism Spectrum Disorders (ASDs) are neurodevelopmental disorders defined by impaired social interactions and communication with repetitive behaviors, activities, or interests. Gastrointestinal (GI) disturbances and gut microbiota dysbiosis are frequently associated with ASD in childhood. However, it is not known whether microbiota dysbiosis in ASD patients also occurs in adulthood. Further, the consequences of altered gut microbiota on digestive functions and the enteric nervous system (ENS) remain unexplored. Therefore, we studied, in mice, the ability offecal supernatant (FS) from adult ASD patients to induce GI dysfunctions and ENS remodeling. First, the analyses of the fecal microbiota composition in adult ASD patients indicated a reduced α-diversity and increased abundance of three bacterial 16S rRNA gene amplicon sequence variants compared to healthy controls (HC). The transfer of FS from ASD patients (FS-ASD) to mice decreased colonic barrier permeability by 29% and 58% compared to FS-HC for paracellular and transcellular permeability, respectively. These effects are associated with the reduced expression of the tight junction proteins JAM-A, ZO-2, cingulin, and proinflammatory cytokines TNFα and IL1β. In addition, the expression of glial and neuronal molecules was reduced by FS-ASD as compared to FS-HC in particular for those involved in neuronal connectivity (βIII-tubulin and synapsin decreased by 31% and 67%, respectively). Our data suggest that changes in microbiota composition in ASD may contribute to GI alterations, and in part, via ENS remodeling.

Keywords: autism; bacterial metabolite; enteric nervous system; intestinal permeability; microbiota.

Figure 1

Fecal microbiota composition of HC and ASD patients. (A) Summary of demographic characteristics of HC and ASD patients. (B) Alpha diversity in HC and ASD individuals was estimated by the richness (Observed ASVs), evenness (Pielou index), and Shannon index, n.s.: non-significant, ** p < 0.01, (C) Principal coordinate analysis (PCoA) of bacterial beta-diversity in HC and ASD individuals generated by subsampling and Bray–Curtis distance. (D) Bacterial taxonomic profile in HC and ASD patients at phylum level. (E) Firmicutes: Bacteroidetes ratio in HC and ASD individuals. Values are represented as mean ± SEM (HC: n = 13–15, ASD: n = 34–36). Statistical analyses were performed with the Mann–Whitney test, * p < 0.05. (F) Bacterial taxonomic profile in HC and ASD patients at genus level (top 11 most abundant genus). (G) Genera significantly increased in ASD compared to HC individuals according to DESeq2 analysis. Data are presented as Log2 of the fold-of-change (FC) between ASD and HC.

Figure 2

Transfer of FS–ASD to mice decreased pellet weight and reduced colonic permeability. (A) Intestinal transit rate measured as the distance of migration of carmine red over total intestinal length. (B) In vivo colonic propulsive motor function assessed by the measure of the number of pellets expelled for 2 h. (C) Total weight and (D) fecal water content of fecal pellets. All values represent means ± SEM (HC: n = 9–10; ASD: n = 11–12). Statistical analyses were performed with the Mann–Whitney U-test or Student t-test, * p < 0.05. (E) in vivo and (F) ex vivo paracellular (sulfonic acid) and transcellular permeability (horseradish peroxidase, HRP) in proximal colon segments. For (F), colonic permeability was determined in Ussing chambers by measuring the mucosal to serosal flux of the markers. All values represent means ± SEM (HC: n = 8–10; ASD: n = 11–12). Statistical analyses were performed with the Mann–Whitney U-test or Student t-test, * p < 0.05. (G) Morphological parameters of the proximal colon characterized by a histological score integrating quantification of the muscle thickening, mucosa integrity, and cellular infiltration (scale bar 250 µm).

Figure 3

Transfer of FS–ASD modulates the expression of tight junction and inflammation-related molecules in the colon. (A) Protein expression of tight junction-forming molecules measured by Western blot in the proximal colon of mice treated with FS–HC or FS–ASD. For each protein, quantification of the signal intensity was normalized to theβ-actin signal of the same sample and expressed as a percentage of controls. (B) Gene expression of molecules with proinflammatory (IL-1β and TNFα) and anti-oxidative (HO-1 and GCLC) activity in the proximal colon of mice treated with FS–HC or FS–ASD. Western blot data are expressed as relative values to β-actin normalized to control, and q-PCR data are expressed as relative values to S6. All values represent means ± SEM (HC: n = 9–10; ASD: n = 11–12). Statistical analyses were performed with the Mann–Whitney U-test or Student t-test, * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 4

Transfer of FS–ASD to mice modulates the expression of glial and neuronal molecules. Protein expression of the glial molecules S100β (A) and GFAP (B) in the proximal colon of mice treated with FS–HC or FS–ASD. Protein expression of the neuronal molecules βIII-tubulin (C) and Synapsin 1 (D) in the proximal colon of mice treated with FS–HC or FS–ASD. Western blot data are expressed as relative values to β-actin normalized to control. All values represent means ± SEM (HC: n = 10; ASD: n = 12). Statistical analyses were performed with the Mann–Whitney U-test or Student t-test, * p < 0.05, *** p < 0.001.

Figure 5

FS–ASD applied to ENS primary cultures induces a remodeling of glial and neuronal molecule expression. Protein expression of the glial molecules S100β and GFAP (A) in ENS cultures treated with FS–HC or FS–ASD. Protein expression of the neuronal molecules βIII-tubulin (B) and Synapsin 1 (C) in ENS cultures treated with FS–HC or FS–ASD. Western blot data are expressed as relative values to β-actin normalized to control. All values represent means ± SEM (HC: n = 10; ASD: n = 16). Statistical analyses were performed with the Mann–Whitney U-test or Student t-test, * p < 0.05, ** p < 0.01.