In vitro Modeling of Chicken Cecal Microbiota Ecology and Metabolism Using the PolyFermS Platform

Abstract

Continuous in vitro fermentation models provide a useful tool for a fast, reproducible, and direct assessment of treatment-related changes in microbiota metabolism and composition independent of the host. In this study, we used the PolyFermS model to mimic the conditions of the chicken cecum and evaluated three nutritive media for in vitro modeling of the chicken cecal microbiota ecology and metabolism. We observed that our model inoculated with immobilized cecal microbiota and fed with a modified Viande Levure medium (mVL-3) reached a high bacterial cell density of up to approximately 10.5 log cells per mL and stable microbiota composition, akin to the host, during 82 days of continuous operation. Relevant bacterial functional groups containing primary fibrolytic (Bacteroides, Bifidobacteriaceae, Ruminococcaceae), glycolytic (Enterococcus), mucolytic (Bacteroides), proteolytic (Bacteroides), and secondary acetate-utilizing butyrate-producing and propionate-producing (Lachnospiraceae) taxa were preserved in vitro. Besides, conserved metabolic and functional Kyoto Encyclopedia of Genes and Genomes pathways were observed between in vitro microbiota and cecal inoculum microbiota as predicted by functional metagenomics analysis. Furthermore, we demonstrated that the continuous inoculation provided by the inoculum reactor generated reproducible metabolic profiles in second-stage reactors comparable to the chicken cecum, allowing for the simultaneous investigation and direct comparison of different treatments with a control. In conclusion, we showed that PolyFermS is a suitable model for mimicking chicken cecal microbiota fermentation allowing ethical and ex vivo screening of environmental factors, such as dietary additives, on chicken cecal fermentation. We report here for the first time a fermentation medium (mVL-3) that closely mimics the substrate conditions in the chicken cecum and supports the growth and metabolic activity of the cecal bacterial akin to the host. Our PolyFermS chicken cecum model is a useful tool to study microbiota functionality and structure ex vivo.

Keywords: PolyFermS; broiler; cecum; in vitro model; microbiota.

FIGURE 1

Experimental design of in vitro continuous PolyFermS fermentations inoculated with immobilized chicken cecal microbiota. Single cecal microbiota was immobilized in a gellan–xanthan gel for each fermentation. (A,B) The immobilized cecal microbiota of chicken 1 was used to inoculate two bioreactors (30% v/v); F1-A and F1-B and continuously fed with mVL-1 and mVL-2 fermentation medium, respectively. (C) F2 consisted of an inoculum reactor (IR) containing cecal beads of chicken 2, connected to four second-stage reactors (SSR) and continuously fed with 5% fermentation effluent of IR and 95% fresh fermentation mVL-3 medium. (D) F3 consisted of an inoculum reactor (IR) containing cecal beads of chicken 3, connected to seven second-stage reactors (SSR) and continuously fed with 5% fermentation effluent of IR and 95% fresh fermentation mVL-3 medium. For activity, microbial composition and metabolites in reactor effluent samples were monitored daily using MiSeq and HPLC-RI, respectively, to show to the model stability and compared with the cecal inoculum.

FIGURE 2

Analysis of cecal content of Cobb-500 broiler chicken (mean ± standard deviation; n = 10). (A) pH; (B) relative metabolite concentrations (%); (C) microbial composition obtained by 16S rRNA gene amplicon sequencing and expressed as relative abundance at phylum; and (D) genus level. When the genus assignment was not possible, the highest-level taxonomy assignment was shown. Values < 1% are summarized in the group “others”.

FIGURE 3

Metabolite and microbial compositions in the cecal inocula and reactor effluents of different fermentations after initial stabilization. (A) Relative metabolite ratios and standard deviation (%). (B–D) Quantification of bacterial populations by qPCR, expressed as mean ± SD log gene copies/g or mL.

FIGURE 4

Microbial composition and diversity in cecal inocula and reactor effluents of the different fermentations. (A,B) Microbial composition measured by 16S rRNA gene amplicon sequencing, represented by relative abundance at phylum (A) and genus levels (B). When assignment at the genus level was not possible, the highest-level taxonomy assignment was shown. Values < 1% are summarized in the group “others”. (C) α-Diversity measured by Shannon index and observed ASVs.

FIGURE 5

Daily fermentation metabolite concentrations in the effluents of the inoculum reactor (IR) for F2 (A) and F3 (B) measured by HPLC-RI. End metabolites (acetate, butyrate, propionate, and formate), intermediate metabolite (succinate), BCFAs (isovalerate and isobutyrate), and valerate. Two consecutive batch fermentations were used for bead colonization. Period 1: The period within which the second-stage reactors were connected to IR and stabilize.

FIGURE 6

Bacterial activity and composition of inoculum reactor (IR) and second-stage reactors (SSR) after initial stabilization estimated from days 6–10 in period 1. Relative metabolite with standard deviation of F2 (A) and F3 (C). Quantification of key bacterial populations by qPCR, expressed as mean ± SD log gene copies/g cecal content or mL effluent of F2 (B) and F3 (D).

FIGURE 7

Microbial composition and diversity in the inoculum reactor (IR) and four second-stage reactors (SSR) of F2 after initial stabilization and estimated from days 6–10 in period 1. (A,B) Microbial composition by 16S rRNA amplicon sequencing represented by relative abundance at phylum (A) and genus levels (B). When the genus assignment was not possible, the highest-level taxonomy assignment was shown. Values < 1% are summarized in the group “others”. (C,D) Principal coordinate analysis (PCoA) of reactor microbiota based on weighted (C) and unweighted (D) UniFrac analysis matrix. α-Diversity measured by (E) Shannon index and (F) observed ASVs.

FIGURE 8

Microbial composition and diversity in the inoculum reactor (IR) and seven second-stage reactors (SSR) of F3 after initial stabilization in period 1. (A,B) Microbial composition by 16S rRNA amplicon sequencing represented by relative abundance at phylum (A) and genus levels (B). When the genus assignment was not possible, the highest-level taxonomy assignment was shown. Values < 1% are summarized in the group “others”. (C,D) Principal coordinate analysis (PCoA) of reactor microbiota based on weighted (C) and unweighted (D) UniFrac analysis matrix. α-Diversity measured by (E) Shannon index and (F) observed ASVs.

FIGURE 9

Predictive functional profiling of microbial communities of chicken cecal samples and modeled microbiota of F1-A, F1-B, F2, and F3 using PICRUSt2 analysis. Heatmap depicts the log-transformed gene abundance of microbiota-associated predicted KEGG pathways. Numbers in scale represent log range of gene abundance for this dataset. Darker shades of light blue represent higher relative abundance as indicated in the legend; white color represents absence.