In the last decade, there has been extensive research on the use of plant based ingredients in aquafeed in order to achieve the sustainability of aquaculture [45–47]. In this context, we recently shown that the use of a high-starch diet in a 100% plant-based diet did not resulted in adverse effects on the rainbow trout metabolism, but did not led to growth improvements [6]. Interestingly, the use of prebiotics such as inulin has already been investigated in teleosts and could have beneficial effects associated with a high-starch diet [31], but little is known about its effect on host metabolism and immunity, in contrast to mammals [48,49]. In this study, we wanted to understand the effects of inulin supplementation in rainbow trout fed a diet containing either high or low CHO/ protein ratio, with the basal diet comprising of 100% plant-based ingredients. Thus, we analyzed the intestinal microbiota, the host metabolism, growth parameters, and some immune markers.
The use of inulin did not significantly affect the gut microbiota in contrast to the use of a high CHO/ protein ratio in rainbow trout fed a 100% plant-based diet
The effect of inulin (or non-digestible polysaccharides) is mainly mediated by the intestinal microbiome via the production of bacterial metabolites such as SCFA or lactate [50,51]. Moreover, a high CHO/protein ratio can affect the rainbow trout gut microbiota and in particular the firmicutes/proteobacteria ratio as well as the lactic-acid bacteria [6]. In our study, the intestinal microbiota was dominated by Proteobacteria and Firmicutes regardless the diets, as previously described in salmonids [2,3,6,52–55]. Interestingly, in humans, the Firmicutes phyla is specialized in the degradation of non-digestible polysaccharides [8,56]. At genus level, Ralstonia (from Proteobacteria) dominates the gut microbiota. Ralstonia has already been described in the intestinal communities of rainbow trout [3,57], in tilapia (Oreochromis niloticus), in goldfish (Carassius auratus) and largemouth bass (Micropterus salmoides) [58–60]. However, in salmonids other genus, such as photobacterium or vibrio belonging to the proteobacteria are found in high abundance in the intestinal microbiota [53,54]. Interestingly we observed that the intestinal microbiota has been strongly modified by the change in the CHO/protein ratio in the fish fed with a 100% plant-based diet as previously shown by Defaix et al [6]. Indeed, the relative abundances of Proteobacteria and Firmicutes phyla were strongly affected by the change in the CHO/protein ratio, confirmed by the decrease of the Firmicutes / Proteobacteria ratio.
In the high-starch groups, the decrease in the Firmicutes / Proteobacteria ratio may seems surprising since the level of digestible carbohydrates is elevated in these groups and it is known that many bacteria belongings to the firmicutes are known to have the ability to encode for carbohydrate-active enzymes allowing the degradation of polysaccharides [61]. However, this decrease may be associated with the significant lower abundances of the Bacillus genus in the high-starch groups. Although in fish, some species of bacilli such as Bacillus cereus, Bacillus subtilis, Bacillus amyloliquefaciens are known to degrade polysaccharides [62], Bacilli found in this study such as Bacillus cytotoxicus in particular in low-starch groups (figure 4b) has not been demonstrated. Conversely, in the high-starch groups, many lactic acid bacteria belonging to the firmicutes groups were observed in higher proportion. Interestingly, these bacteria are known to metabolize dietary plant glucosides and externalizes their bioactive phytochemicals [63]. Additionally, these lactic-acid bacteria (LAB) were found in higher proportion in rainbow trout fed with plant-based diet in comparison to a FM/FO diet [64]. These LAB, in particular Lactobacillus, Lactococcus and streptococci, can present a symbiotic relationship with the host [65] and are able to produce lactate through homolactic acid fermentation [65,66]. Interestingly, two SCFAs i.e butyric acid, and valeric acid were found to be higher in high starch groups according to previous work [6]. The production of key cross-talk molecules such as SCFAs, in particular butyric acid, may have several beneficial effects on the fish gut health and are known to improve immunity [11,67], and an enhancement of the glucose homeostasis [68]. In contrast, we observed that the microbiota of trout fed with the low CHO/protein ratio presented a lower levels of LAB and higher proportion of Enterococcus and Bacillus cytotoxicus, which could be opportunistic pathogens [69,70] and could have detrimental effects on trout gut health. There were also some effects of inulin on the intestinal microbiota. Indeed, four genera were modified by the use of inulin in contrast to previous studies on fish where inulin strongly affected the gut microbiota including rainbow trout [71], and Nile tilapia [68]. Concerning Lactobacillus, a CHO/Protein: inulin interaction was observed. This result requires further investigations to elucidate the potential role of inuline. Additionally, we did not record a significant change in the SCFAs and lactic acid production with the use of inulin unlike a previous work in tilapia where the authors observed an increase of SCFA production in tilapia associated with a high-starch diet [68].
In order to detect if the higher production of SCFA in the group fed with high CHO/protein ratio can act on the host metabolism, we studied the expression of the FFAR encoding genes on the mid-intestine. Indeed, these G protein-coupled receptors participate in both immune and metabolic regulation after activation by SCFA [72]. In humans, the activation of FFAR lead to signal molecules production (Gαq/11 or Gαi/o) enhancing the secretion of insulin by β-pancreatic cells [73]. These classes of receptors have already been characterized in the rainbow trout intestine, and were found to be regulated when the fish were fed with a plant-based diet [2,38]. Interestingly, we observed that all the FFAR receptors (ffar1, ffar2a1, ffar2a1a, ffar2b1a, ffar2b1b, ffar2b2a, and ffar2b2b1) in mid-intestine were significantly down-regulated by the high CHO/protein ratio. Inulin had no impact on the expression of these genes. In a previous studies, the ffar2b1a (previously ffar31) expression was significantly increased in trout fed a 100% plant-based diet with inulin [2]. Moreover, it is known that the chronic stimulation (here during 12 weeks of feeding) of G protein-coupled receptors lead to the recruitment of β-arrestins preventing further stimulation of the downstream signaling pathways [74,75]. In mid intestine, the decrease in the expression of these key receptors therefore probably reflects the desensitization (even remains to be demonstrated in fish) of these receptors after being stimulated by SCFA, demonstrating the high reactivity of FFARs in the presence of endogenous ligands in rainbow trout in mid intestine.
Effect of interactions between starch and inulin in the host metabolism of rainbow trout fed 100% plant-based diets
Increasing the CHO proportion by decreasing the proportion of plant protein in the high-starch diet did not affect the final weight of fish but resulted in an increase of the protein efficiency ratio as shown in our previous study [35]. CHO could prevent the protein catabolism for energy needs, as shown previously in fish fed with marine ingredients [76], and now in trout with plant based raw material [6]. But, unexpectedly, feeding rainbow trout with a diet supplemented with 2% inulin caused a decrease in the specific growth rate, final body weight, and feed efficiency at the end of the 12-weeks. Conversely, previous studies on rainbow trout did not observe a decrease of growth when fish were fed with 2% inulin in a 100% plant-based diet [2,3].
Rainbow trout are known to be “low users” of CHO when fed with a FM/FO diet. High levels of CHO generally lead to a persistent post-prandial hyperglycemia [77–79], explained in part, by a deregulation of gluconeogenesis pathways [80]. In this study, the high-starch diet, in combination with a 100% plant-based diet, did not induce post-prandial hyperglycemia, suggesting an efficient glucose homeostasis, which we previously observed in trout fed a 100% plant-based diet [6]. Interestingly, the use of these 2-factors experiments led to many significant interactions at molecular levels such as the glycolysis, gluconeogenesis, and lipogenesis pathways. Indeed, we observed an up-regulation of multiple genes in fish fed with the low-starch groups and inulin suggesting that inulin could induce a stimulation of the host’s metabolism with a low carbohydrate diet, already observed in a 100% plant-based diet [2]. In contrast, no significant differences were observed in the high-starch groups with the inulin intake. This could be explained by the strong effect of the high-starch diets on the microbiota composition and on the expression of many genes, limiting the potential effect of inulin. A significant interaction was measured for the pk mRNA expression along with a significant interaction effect on the pyruvate kinase enzymatic activity but not in the same way; indeed, surprisingly, in fish fed low starch, the increase of pk mRNAs is associated with a decrease of pyruvate kinase activity. In trout this gene is already known to be atypically controlled with high-starch diets, with a lower expression of the pk gene and a paradoxical increase of pyruvate kinase activity [81–83]. On the other hand, surprisingly also but not for the same reasons, the use of inulin in the high-starch diet resulted in a significant decrease of the glucokinase enzymatic activity, which can be caused by the significant decrease of the gckb gene expression.
Interestingly, while the poor glucose homeostasis in rainbow trout fed with fish meal and fish oil (FM/FO) is in part explained by a non-downregulation of the gluconeogenesis pathway with high levels of CHO [81,84,85], we did not observed higher expression of g6pcb2 genes related to gluconeogenesis in the high-starch groups. Additionally, a decrease of glucose-6-phosphatase enzymatic activity was observed with the high-starch diet. These observations suggest the existence of an efficient glucose homeostasis which may explained in part why no post-prandial hyperglycemia was detected in plasma.
Regarding lipid metabolism at the interface with the glucose metabolism, significant interaction CHO/Protein: inulin was observed for lipogenesis. In fact, dietary inulin supplementation induced an up-regulation of serbf1, aclyb, and aclyc genes in fish fed the low-starch diets. In fish fed high-starch diets, these genes were down-regulated. Additionally, as previously shown in trout fed plant-based diets [6], the use of a high-starch diet did not result in a higher activity of the fatty acid synthase, with no change in the fasna and fasnb mRNA expression, while usually the use of a high-starch diet with FM/FO induce an increase of the lipogenesis pathway [79,85]. Moreover, a decrease of plasma triglycerides was observed with the use of high-starch diets. Finally, no differences in the final body lipid content (Figure S3) were observed, showing that with a 100% plant-based diet the dietary starch in excess does not appear to have been stored as fat (only an expected increase of glycogen was found). In plasma, there was also a significant decrease in cholesterol levels, and it was consistent with the down-regulation of genes related to cholesterol biosynthesis in liver. The lower cholesterol level could be linked to the higher proportion of intestinal LAB, and in particular the presence of Lactobacillus species in the in high-starch diet. Indeed the presence of this bacteria and the production of SCFA have already been linked to a decrease of the cholesterol biosynthesis in human [86].
Effects of the CHO/protein ratio and inulin on the immunological status
In carnivorous fish species, the use of plant-based diets, containing for instance pea [87] or soybean protein [88], may induce enteritis of the digestive tract due to the presence of anti-nutritional factors (i.e. non-starch polysaccharides, lectins, tannins) [89]. In Salmonids, enteritis can lead to an increase in intestinal epithelial permeability, which can induce an inflammation and leukocyte infiltrations of lamina propria [90]. Moreover, the use of a high CHO diet increased the intestinal permeability and induce inflammation in chinese perch [91], and largemouth bass [60]. Thus, we studied tight junction proteins (tjp) gene expression in mid-intestine. These genes tighten the junctions between epithelial cells to prevent the passage of pathogens through the epithelial barrier, inducing the host’s immune response [92]. In our study, we observed a reduction in tjp1a expression in the high-starch diet and down regulation of tjp3 expression in fish fed with the diets supplemented with inulin. Additionally, two genes coding for tight-junction associated transmembrane proteins, marveld1 and marveld3, were significantly affected by the interaction between starch and inulin, with a reduction of these genes’ expression for HS-0 in comparison with LS-0. The lower expression of these genes in the HS-0 group could be a direct effect of the reduction of antinutritional factors known to cause epithelial damage. Moreover, the use of a high-starch diet did not induce an up-regulation of the tjp genes.
We also studied the effects of the change in the CHO/protein ratio as well as the inulin factors on immunity actors in intestinal mucosa and liver. Indeed, the presence of anti-nutritional factors in the diet can induce innate immune response of epithelial cells and in the liver [93–95]. In this study, a significant decrease of plasma nitric oxide (NO) in fish fed with the high-starch diet was also observed. Interestingly, NO is involved in immune defense in rainbow trout [96], and is particularly linked to the activation of macrophages in site of inflammation in fish [97]. Further studies must be made to assess the inflammatory status of trout, but these results may suggest that the diets formulated with a high CHO/plant protein ratio can have partially reduce the inflammation. Moreover, induction of NO can led the activation of gene encoding for pro-inflammatory receptors, such as cxcr4 and cxcr4.1.1 [98,99]. In mid intestine, significant interaction between factors was detected for both cxcr4 and cxcr4.1.1 gene expression with a lower expression in the HS-0 diet than the LS-0 diet. This result may also be related with the lower proportion of plant protein containing antinutritional factors in the HS-0 group, as antinutritional factors are known to cause enteritis in salmonids [100].
While the use of anti-nutritional factors can led to mucosal inflammation [94], a high-starch diet in rainbow trout may also result in higher production of liver pro-inflammatory cytokines [95], the liver in teleost being known to be involved in immune responses in teleosts [101,102]. Interestingly, here, high starch diets induced a significant decrease of the il1b, il8, and tnfa genes, suggesting that reducing of the proportion of plant-proteins may result in a decrease of cytokines responses in the liver. We also observed, surprisingly, that the dietary inclusion of inulin induced an increase in the il8 expression in the liver. Activation of il8 may result from tissue damage or infection, which may highlight a potential unexpected adverse effect of dietary inulin.
Finally, as an integrative biomarker of immune status in fish, we analyzed the lysozyme known to play a key role in innate immunity by eliminating pathogens [103]: lysozyme is produced and secreted by granulocytes and monocytes during pathogens and parasites infection [104–106]. In our study, we did not observe any difference between the high and low starch groups, but a reduction of this enzymatic activity was observed in the inulin groups. This result with inulin was unexpected since most studies using prebiotics or probiotics in fish have led to an enhancement of the lysozyme activity [107].
In conclusion, in our experiment, the high CHO/protein ratio in a full plant-based diet appears to be beneficial for the health of fish regarding the gut integrity (through tight junction expression genes) and the reduction in the production of pro-inflammatory cytokines in liver. Conversely, inulin reduced lysozyme activity and had no beneficial effect on several immune markers.