Exploring the effects of dietary inulin in rainbow trout fed a high-starch, 100% plant-based diet

Abstract

Background: High dietary carbohydrates can spare protein in rainbow trout (Oncorhynchus mykiss) but may affect growth and health. Inulin, a prebiotic, could have nutritional and metabolic effects, along with anti-inflammatory properties in teleosts, improving growth and welfare. We tested this hypothesis in rainbow trout by feeding them a 100% plant-based diet, which is a viable alternative to fishmeal and fish oil in aquaculture feeds. In a two-factor design, we examined the impact of inulin (2%) as well as the variation in the carbohydrates (CHO)/plant protein ratio on rainbow trout. We assessed the influence of these factors on zootechnical parameters, plasma metabolites, gut microbiota, production of short-chain fatty acids and lactic acid, as well as the expression of free-fatty acid receptor genes in the mid-intestine, intermediary liver metabolism, and immune markers in a 12-week feeding trial.

Results: The use of 2% inulin did not significantly change the fish intestinal microbiota, but interestingly, the high CHO/protein ratio group showed a change in intestinal microbiota and in particular the beta diversity, with 21 bacterial genera affected, including Ralstonia, Bacillus, and 11 lactic-acid producing bacteria. There were higher levels of butyric, and valeric acid in groups fed with high CHO/protein diet but not with inulin. The high CHO/protein group showed a decrease in the expression of pro-inflammatory cytokines (il1b, il8, and tnfa) in liver and a lower expression of the genes coding for tight-junction proteins in mid-intestine (tjp1a and tjp3). However, the 2% inulin did not modify the expression of plasma immune markers. Finally, inulin induced a negative effect on rainbow trout growth performance irrespective of the dietary carbohydrates.

Conclusions: With a 100% plant-based diet, inclusion of high levels of carbohydrates could be a promising way for fish nutrition in aquaculture through a protein sparing effect whereas the supplementation of 2% inulin does not appear to improve the use of CHO when combined with a 100% plant-based diet.

Keywords: Aquaculture; Fish nutrition; Gut microbiota; Immune markers; Intermediary metabolism; Inulin; Prebiotic; Rainbow trout; Short-chain fatty acids.

Fig. 1

Experimental design of the feeding trial. Four experimental diets were produced as extruded pellets. These diets were made of 100% plant raw material and contained either 19% of digestible starch (high-starch diet) or 2% digestible starch (low-starch diet) and with 2% of inulin or 0 of inulin were produced as extruded pellets. The fish were fed the experimental diets by hand, twice a day, during 12 weeks to 324 females rainbow trout (~ 31 g) distributed in 12 tanks (3 tanks per group). The trout were weighed every 3 weeks to record the zootechnical parameters (n = 3 tank per experimental diet)

Fig. 2

Bacterial alpha diversity (A) is represented in terms of observed OTUs, Chao1, Shannon, Simpson, InvSimpson, in the mid-intestinal section of rainbow trout after 12 weeks of feeding. Feeding groups are symbolized as LS-0: Low-starch with 0 inulin; LS-In: low-starch with 2% inulin; HS-0: High-starch with 0 inulin; and HS-In: High-starch with 2% inulin. Beta diversity (B) is presented by a PCoA representation (Bray–Curtis distance, Weighted-Unifrac analysis) in mid intestine section, according to the experimental diets. Beta diversity was compared using a pairwise PERMANOVA test and data were considered statistically different for P < 0.05. n = 12 fish per group

Fig. 3

Microbial composition in the mid-intestinal section at phylum (a) and genus (b) levels after 12 weeks of feeding. In panel b, only the top 15 abundant genera are presented. Feeding groups are symbolized as LS-0: Low-starch with 0 inulin; LS-In: low-starch with 2% inulin; HS-0: High-starch with 0 inulin; and HS-In: High-starch with 2% inulin. On panel c, the relative abundance of the Proteobacteria and Firmicutes phyla and the relative abundances of the Ralstonia, Bacillus, and Lactobacillus genus are indicated. Statistical differences were analyzed with a two-way ANOVA test and were considered statistically significant for P < 0.05. Significant differences are represented by asterisk. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. n = 12 fish per group

Fig. 4

PLS-DA analysis for fish fed the high-starch and low-starch diets (independently of inulin) based on OTUs abundance (a). Each red points or blue triangles represent a sample. Samples can be discriminated according to experimental group on component 1. Contribution level of the top 15 OTUs are presented (b). Red bars correspond to the high-starch group and blue bars to the low-starch group. n = 12 fish per group

Fig. 5

Levels of acetic acid, butyric acid, propionic acid, valeric acid, caproic acid as well as lactic acid were measured (µg) relative to the mass of the mid-intestinal digestive contents (g) with SIFT-MS mass spectrometry. Feeding groups are symbolized as LS-0: Low-starch with 0 inulin; LS-In: low-starch with 2% inulin; HS-0: High-starch with 0 inulin; and HS-In: High-starch with 2% inulin. Statistical differences were analyzed with a two-way ANOVA test and were considered statistically significant for P < 0.05. Significant differences are represented by an asterisk. ^*P < 0.05, ^**P < 0.01, ^***P < 0,001. n = 6 fish per group

Fig. 6

Hepatic enzymatic activities of glucokinase (a), pyruvate kinase (b), glucose-6-phosphatase (c), and fatty-acid synthase (d) after 12 weeks of feeding. Enzymatic activities were measured in liver samples relative to the average of milligram of proteins. Feeding groups are symbolized as LS-0: Low-starch with 0 inulin; LS-In: low-starch with 2% inulin; HS-0: High-starch with 0 inulin; and HS-In: High-starch with 2% inulin. Statistical differences were analyzed with a two-way ANOVA test and were considered statistically significant for P < 0.05. Significant differences are represented by an asterisk. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. n = 12 fish per group

Fig. 7

Mid-intestinal mRNA expression levels of tight-junction protein associated genes (tjp1a, tjp3, marveld1, and marveld3) (a), mid-intestine mRNA expression levels of C-X-C motif chemokine receptor 4 paralogues (cxcr4 and cxcr4.1.1) (b), il1b, il8, and tnfa (c) after 12 weeks of feeding. Feeding groups are symbolized as LS-0: Low-starc h with 0 inulin; LS-In: low-starch with 2% inulin; HS-0: High-starch with 0 inulin; and HS-In: High-starch with 2% inulin. Statistical differences were analyzed with a two-way ANOVA test and were considered statistically significant for P < 0.05. Significant differences are represented by an asterisk. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. n = 12 fish per group

Fig. 8

Plasma immune markers after 12 weeks of feeding. Feeding groups are symbolized as LS-0: Low-starch with 0 inulin; LS-In: low-starch with 2% inulin; HS-0: High-starch with 0 inulin; and HS-In: High-starch with 2% inulin. Data are presented as the mean ± SD. Statistical differences were analyzed with a two-way ANOVA test and were considered statistically significant for P < 0.05. Significant differences are represented by an asterisk. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001. n = 12 fish per group

Fig. 9

Correlations between OTUs abundances and immune markers (a), and metabolic parameters (b). Heatmap correlations, were calculated using regularized canonical correlation analysis (rCCA). n = 12 fish per group