Contribution of the Gut Microbiota in P28GST-Mediated Anti-Inflammatory Effects: Experimental and Clinical Insights

An original immuno-regulatory strategy against inflammatory bowel diseases based on the use of 28 kDa glutathione S-transferase (P28GST), a unique schistosome protein, was recently proposed. Improvement of intestinal inflammation occurs through restoration of the immunological balance between pro-inflammatory T-helper 1 (Th1) responses and both T-helper 2 (Th2) and regulatory responses. However, detailed mechanisms explaining how P28GST prevents colitis and promotes gut homeostasis remain unknown. Considering the complex interplay between the adaptive and innate immune system and the intestinal microbiota, we raised the question of the possible role of the microbial ecosystem in the anti-inflammatory effects mediated by the helminth-derived P28GST protein. We first analyzed, by 16S rRNA sequencing, the bacterial profiles of mice fecal microbiota at several time points of the P28GST-immunomodulation period prior to trinitrobenzene sulfonic acid (TNBS)-colitis. The influence of gut microbiota in the P28GST-mediated anti-inflammatory effects was then assessed by fecal microbiota transplantation experiments from P28GST-immunized mice to either conventional or microbiota depleted naïve recipient mice. Finally, the experimental data were supplemented by the temporal fecal microbiota compositions of P28GST-treated Crohn’s disease patients from a pilot clinical study (NCT02281916). The P28GST administration slightly modulated the diversity and composition of mouse fecal microbiota while it significantly reduced experimental colitis in mice. Fecal microbiota transplantation experiments failed to restore the P28GST-induced anti-inflammatory effects. In Crohn’s disease patients, P28GST also induced slight changes in their overall fecal bacterial composition. Collectively, these results provide key elements in both the anti-inflammatory mechanisms and the safe therapeutic use of immunomodulation with such promising helminth-derived molecules.

l-1, Kabi). Taking the stability of antibiotics into account, a fresh mixture was reconstituted every 2-days and drinking bottles were protected from the light to avoid fluoroquinolone (ciprofloxacin) degradation. In addition, due to weak stability at room temperature, mice were orally fed daily with a freshly made solution of imipenem:cilastatin at 1:1 (250 mg l-1, Mylan). Control mice received standard drinking water by drinking bottle and gavage.

DNA Extraction
DNA extraction was carried out on 200 mg of mouse or human feces according to an optimized and standardized protocol developed by the company Genoscreen, Lille, France. Based on the use of the QIAamp stool DNA Kit (Qiagen, France), various preliminary steps of mechanical and chemical lysis have been added in order to favor the lysis of all bacteria and in particular Gram +. A control of the extracted DNAs was performed with a fluorometric assay method for the SybrGreen Nucleic Acid I molecule (LifeSciences Merck, Germany) according to a protocol also developed by the company GenoScreen.

16S targeted Metagenomic Analysis
Microbial diversity and taxonomic composition of samples were determined for each sample using the so-called metagenomic targeted methodology. Briefly, a fragment of the 16S rRNA gene framing the V3 and V4 hypervariable regions was amplified from 5 ng of gDNA of the samples according to an optimized and standardized amplicon library preparation protocol Metabiote® (Genoscreen, France). The final amplification products (each containing a nucleotide index to differentiate the samples as well as the adapters necessary for carrying out the sequencing) were purified on beads and then finally mixed in equal concentrations. The sequencing was performed with the 2*250 bp chemistry on the Illumina MiSeq platform (Illumina, San Diego) at Genoscreen.
The resulting raw sequences were subjected to a cleaning process comprising i) sorting the sequences according to the indexes and according to the 16S primers ii) the "trimming" of the sequences by truncating the bases from the 3 'end with a Phred threshold of 30 (iii) assembly the two paired sequences according to a minimum overlap of 30 bases and at least 97% identity over the overlapping area. The computer analysis was carried out on a fully automated (Metabiote® OnLine) pipeline built around the QIIME v 1.9.1 software [55]. Following the steps of pre-processing, the full-length 16S rRNA gene sequences go through a step in which chimera sequences are detected and eliminated (in-house method based on the use of Usearch 6.1). Then, a clustering step is performed in order to group similar sequences with a nucleic identity defined threshold (97% identity for an affiliation at the genus level on the V3-V4 regions of the 16S rRNA gene) with Uclust v1.2.22q [56] through an open-reference OTU picking process and complete-linkage method, finally creating groups of sequences or "Operationnal Taxonomic Units" (OTUs

Supplementary Tables
Supplemental Table S1: list of targeted genes and corresponding primers accession numbers.