Weaning-associated feed deprivation stress causes microbiota disruptions in a novel mucin-containing in vitro model of the piglet colon (MPigut-IVM)

Abstract

Background: Risk factors for the etiology of post-weaning diarrhea, a major problem in swine industry associated with enormous economic losses, remain to be fully elucidated. In concordance with the ethical concerns raised by animal experiments, we developed a new in vitro model of the weaning piglet colon (MPigut-IVM) including a mucin bead compartment to reproduce the mucus surface from the gut to which gut microbes can adhere.

Results: Our results indicated that the MPigut-IVM is able to establish a representative piglet archaeal and bacterial colon microbiota in terms of taxonomic composition and function. The MPigut-IVM was consequently used to investigate the potential effects of feed deprivation, a common consequence of weaning in piglets, on the microbiota. The lack of nutrients in the MPigut-IVM led to an increased abundance of Prevotellaceae and Escherichia-Shigella and a decrease in Bacteroidiaceae and confirms previous in vivo findings. On top of a strong increase in redox potential, the feed deprivation stress induced modifications of microbial metabolite production such as a decrease in acetate and an increase in proportional valerate, isovalerate and isobutyrate production.

Conclusions: The MPigut-IVM is able to simulate luminal and mucosal piglet microbiota and represent an innovative tool for comparative studies to investigate the impact of weaning stressors on piglet microbiota. Besides, weaning-associated feed deprivation in piglets provokes disruptions of MPigut-IVM microbiota composition and functionality and could be implicated in the onset of post-weaning dysbiosis in piglets.

Keywords: Colon; Dysbiosis; In vitro gut model; Microbiota; Mucin; Piglet; Weaning.

Fig. 1

Description of the new mucin implemented in vitro of the piglet colon (MPigut-IVM). a Schematic view of the MPigut-IVM. b Schematic representation of the mucin bead compartment of the MPigut-IVM. c Denomination of the MPigut-IVM samples throughout this publication. d Experimental design of the fermentation experiments performed using the MPigut-IVM

Fig. 2

Set-up and validation of the MPigut-IVM: microbiota activity. a Relative abundance of gas produced by fermentation activity of the microbiota inhabiting the MPigut-IVM during control assays. b Short chain fatty acids (SCFA) relative abundance produced by fermentative activity of the microbiota inhabiting the MPigut-IVM during control assays in the bioreactor medium

Fig. 3

Set-up and validation of the MPigut-IVM: microbiota composition. a Evolution of the structure and colonization of a mucin bead before (I) and after (II) 48 h of incubation in the MPigut-IVM during the fermentation #1 and observation of the specific adherent microbiota at high magnitude (III & IV) by scanning electron microscopy. b Relative abundance of the 15 principal bacterial families in the bioreactor medium of the MPigut-IVM during control assays measured by 16S Illumina sequencing. c Relative abundance of the 15 principal bacterial families on the mucin beads of the MPigut-IVM during control assays measured by 16S Illumina sequencing

Fig. 4

Set-up and validation of the MPigut-IVM: microbiota diversity. a Alpha diversity indices based on bacterial OTUs in the bioreactor medium of the MPigut-IVM during the fermentations #1, 2 and 3. b Alpha diversity indices based on bacterial OTUs on the mucin beads of the MPigut-IVM during the fermentations #1, 2 and 3. c Principal component analysis (PCoA) plot with Bray-Curtis dissimilarity on the bacterial communities in the bioreactor medium and on the mucin beads of the MPigut-IVM from the end of stabilization phase (day 7) to the end of fermentation (day 15) of the fermentations #1, 2 and 3

Fig. 5

Set-up and validation of the MPigut-IVM: comparison with in vivo data. I. Comparison of the top 15 bacterial families between in vivo proximal colon luminal samples and in vitro fermentation samples from the bioreactor medium at day 7 corresponding to the end of the stabilization phase. The black colour corresponds to an abundance of 0. On the X axis: “F” = “Fermentation” and “D” = “Day”. J. Comparison of the top 15 families between in vivo proximal colon mucosal samples and in vitro mucin beads at day 7 corresponding to the end of the stabilization phase. The black colour corresponds to an abundance of 0. On the X axis: “F” = “Fermentation” and “D” = “Day”

Fig. 6

Effects of a feed deprivation stress of 12, 24, and 48 h on the MPigut-IVM microbiota composition. On this figure, the “12h feed deprivation stress”, “24h feed deprivation stress” and “48h feed deprivation stress” correspond to the fermentations #4, 5 and 6, respectively. a Evolution of the redox potential. An example of redox measurement for a control assay is provided one the figure. These data are from the fermentation 2. b Effect of a feed deprivation period of 12, 24 or 48 h on the relative abundance of the main bacterial families in the bioreactor medium of the MPigut-IVM as measured by 16S Illumina sequencing. c Effect of a feed deprivation period of 12, 24 or 48 h on the relative abundance of the main bacterial families on the mucin beads of the MPigut-IVM as measured by 16S Illumina sequencing

Fig. 7

Quantification of bacterial and methanogen archaeal in the bioreactor medium (a) and on the mucin beads (b) of the MPigut-IVM during the fermentations #6, 7, 8 and 9 which were subjected to a feed deprivation stress of 48 h (n = 4 for each time point)

Fig. 8

Relative abundance of the 15 main bacterial families in the bioreactor medium (a) and on the mucin beads (b) of the MPigut-IVM during the fermentations #6, 7, 8 and 9 which were subjected to a feed deprivation stress of 48 h, as measured by 16S Illumina sequencing

Fig. 9

Differentially abundant genera between the end of the stabilization (day 7) and the recovery phase (day 9.5, 10, 11 and 15) in the bioreactor medium, mucin beads and bead medium of the fermentations #6, 7, 8 and 9. Only significative Log₂FoldChanges are represented on the figure. P-value codes are indicated on the figure

Fig. 10

Evolution of the mean total concentration (a) and of the relation abundance (b) of SCFAs during the fermentation #6, 7, 8 and 9 which were subjected to a feed deprivation stress of 48 h (n = 4 for each time point)

Fig. 11

Analysis of the in vitro metabolome by nuclear magnetic resonance (NMR). Metabolomics analysis were performed by using NMR in the bioreactor medium (a and b) and bead medium (c and d) compartments. Individual plots of partial least square-discriminant analysis using metabolites as variables and time points as predictors (a and c). Heatmap representing the relative concentrations of all identified metabolites (rows) in individual samples (columns). The color represents the Z-scores (row-scaled relative concentrations) from low (blue) to high (red) values. Metabolites (rows) were clustered by the average method. The mean relative abundances were analyzed by a mixed model and Anova.*: P < 0.05, **: P < 0.01, ***: P < 0.001 (adjusted P-values by the false discovery rate method). The bead medium of the fermentation number 6 was not analyzed by NMR which explains the three replicates for the figure c

Fig. 12

Spearman's correlation between the relative abundance of the main bacterial families and the metabolites in the bioreactor medium (a) and on the mucin beads (b) of the MPigut-IVM during the fermentation 6, 7, 8 and 9 at day 7, 9, 9.5, 10, 11 and 15. Cells are colored based upon the Spearman correlation coefficient. Asterisks indicate significant P values (< 0.05)