Differential Effects of Early-Life and Postweaning Galacto-oligosaccharide Intervention on Colonic Bacterial Composition and Function in Weaning Piglets

ABSTRACT

Recently, we proved that the early-life galacto-oligosaccharides (GOS) intervention could improve the colonic function by altering the bacterial composition in suckling piglets. However, whether the early-life GOS (ELG) intervention could have a long influence on the colonic microbiota and whether the combined ELG and postweaning GOS (PWG) intervention would have an interacting effect on maintaining colonic health in weaning piglets remain to be explored. In this study, we illustrated the differential effects of the ELG and PWG interventions on colonic microbiota and colonic function of weaning piglets. Our results showed that the ELG and PWG interventions decreased the frequency of diarrhea in weaning piglets while the PWG intervention increased colonic indexes. After 16S rRNA gene MiSeq sequencing of the gut bacteria belonging to different colonic niches (mucosa and digesta), the increase in the α-diversity of the colonic mucosal bacteria during PWG intervention was revealed. In addition, we found that both the ELG and PWG interventions enriched the relative abundances of short-chain fatty acid (SCFA) producers in different colonic niches and increased the total SCFA concentration in colonic digesta. These changes selectively modulated the mRNA expression levels of pattern recognition receptors and barrier proteins in the colonic mucosa. Of note, the combined effect of ELG and PWG effectively enhanced colonic SCFA producer enrichment and upregulated the butyrate concentration. Meanwhile, the expression levels of MyD88-NF-κB signaling and the proinflammatory cytokines were markedly reduced under the combined effect of ELG and PWG.

IMPORTANCE Reducing the disorders of the gut ecosystem is an effective way to relieve weaning stresses of piglets and minimize economic losses in the modern swine industry. To this end, prebiotics have been often added to their diet during the weaning transition. In the present study, we demonstrated that the ELG and PWG interventions showed different effects on the bacterial composition of different colonic niches and on colonic function in the weaning piglets. Especially under the combined effect of ELG and PWG intervention, the expression levels of MyD88-NF-κB and the proinflammatory cytokines decreased with increasing concentrations of butyrate, which is an important microbial metabolite involved in the colon of weaning piglets. These findings further provided new insights into nutritional interventions that alleviate intestinal ecosystem dysbiosis and gut dysfunction in the piglets during the weaning transition.

INTRODUCTION

Currently, commercial and research interests in pigs focus on the interaction between the gastrointestinal function and gut microbiota composition, especially early in life (1). The early life of piglets is a critical time window for bacteria colonization in piglets’ gut. The microbiota, colonized on the intestinal mucosa during their early life, can stimulate local immune cell proliferation to educate the gut immune system, and may exert a long-lasting impact on a host health (2). Therefore, several recent studies have attempted to improve the gut function of piglets by modulating the gut bacterial composition with nutritional interventions in the early-life period (3, 4). Galacto-oligosaccharides (GOS), a type of functional oligosaccharide, have been used as a prebiotic in human formula milk or as a feed additive in husbandry production (5). Galacto-oligosaccharides cannot be digested within the host intestine or fermented by gut bacteria, which results in an altered bacterial composition (3). Our recent in vivo study found that early-life GOS (ELG) intervention is an effective method to alter the gut function and bacterial composition, especially that on the colonic mucosa, of suckling piglets (3). Although our data document the relationship between the microbiota colonization and intestinal function of piglets in the preweaning stage, whether the ELG intervention could have a long-lasting influence on the bacterial composition in the gut of piglets during the weaning stage requires further investigation.

In the modern swine industry, piglets undergo weaning around 3 to 4 weeks of age to improve reproductive efficiency of sows (1, 6). However, weaning is a sudden and stressful event for piglets, and causes digestive physiology disorders in the gut, diarrhea, and growth inhibition, which ultimately leads to economic losses (6). Current evidence suggests that the gut microbiota have a major role in maintaining piglet intestinal function, and that the imbalances of gut microbiota often occur in piglets that are under weaning stress (7). Thus, it is important to remodel the homeostasis of gut microbiota to ameliorate gut dysfunction in weaning piglets. Xing et al. (8) demonstrated that the diet supplemented with GOS altered the copies of Escherichia coli, Bifidobacteria, and Lactobacillus in feces, and exhibited positive effects on growth and immune function in weaned piglets. However, the effect of GOS supplementation on the relationship between the intestinal bacterial composition and gut function in weaning piglets is not fully understood. Upon the birth of piglets, the colon becomes the main site for bacterial colonization and the establishment of cross talk between the colonic bacteria and the host (7). Previous studies have documented that colonic bacterium can modulate the gut immune system and the host digestive physiology through their metabolites (9, 10). Additionally, GOS are a nondigestible carbohydrate that mainly enters the large intestine and are fermented by bacteria (3). Notably, Mu et al. (11) pointed out that the colonic bacteria, which distributed in different niches (mucosa and lumen) of piglets, showed marked differences in both composition and function. Thus, it is essential to describe the microbiota in different colonic niches and gut function response in postweaning GOS (PWG) intervention in weaning piglets.

Here, we explored the differential effects of the ELG and PWG interventions on different niches of colonic bacterial composition in the weaning piglets. To further reveal whether the changes in colonic bacterial composition would affect colonic function, we also investigated the concentrations of short-chain fatty acids (SCFAs) and the mRNA expression associated with immune homeostasis and barrier function in the weaning piglet colon.

RESULTS

Effect of the ELG and PWG interventions on growth performance, diarrhea frequency, and colonic indexes in weaning piglets.

After weaning at day 21, the growth rate and average daily gain of four groups of weaning piglets (day 21 to day 28) were reduced compared with that of those in suckling stage (day 0 to day 21), indicating the typical appearance of weaning stress appearance (Fig. S1A and S1B). The increased average daily food intake (ADFI), the ratio of colonic weight to colonic length and colonic weight to BW were observed when piglets received the PWG intervention (P < 0.05), while these parameters were not affected by the ELG intervention (P > 0.05; Fig. 1B to G). These results implied that the PWG intervention alleviated the growth inhibition caused by weaning transition. Both the ELG and PWG reduced the diarrheal frequency of weaning piglets (P < 0.05; Fig. 1D). There was no interaction effect between the ELG and PWG interventions observed in growth performance, diarrhea frequency, or colonic indexes of weaning piglets (P > 0.05; Fig. 1B to G).

FIG 1.

Differential effects of the ELG and PWG interventions on growth performance, diarrhea frequency, and colonic indexes of weaning piglets (n = 6). (A) Experimental scheme. (B) Average daily gain (ADG). (C) Average daily food intake (ADFI). (D) Diarrhea frequency. (E) Ratio of colonic weight/length. (F) Ratio of colonic weight/BW. (G) Ratio of colonic length/BW. When the P value for an interaction was <0.05, there was considered to be a significant interaction effect between the ELG and PWG interventions.

Diversity of colonic bacteria.

Several studies have demonstrated that lumen and mucosa have different gut bacterial composition (11, 12). Therefore, we examined the bacterial community in the colonic mucosa and digesta of weaning piglets using 16S rRNA gene MiSeq sequencing. All sample reads were processed together after 16S rRNA gene high-throughput sequencing. A total of 2,084,227 sequences were obtained in all samples (average of 43,421 sequences per sample) after the quality control. The overall number of detected operational taxonomic units (OTUs) was 907 with 97% sequence similarity between reads.

Beta-diversity can reflect the difference in gut bacterial composition. Therefore, based on principal coordinate analysis (PCoA) of β-diversity, we used the unweighted UniFrac distances to assess the separation distance observed between various groups. Plots of samples with similar bacterial composition tend to cluster together. Based on this principle, we found that the plots of colonic mucosa and digesta were significantly separated (Fig. S1C). Moreover, the plots representing the colonic mucosal bacteria in different groups were separated (Fig. 2A). The plots of colonic digesta bacteria in the C/C group and G/C group were only markedly separated from those in G/G group (Fig. 2B). These findings demonstrated that colonic mucosal bacterial composition has an increased tendency to change compared to that of colonic digesta in response to the different GOS interventions. Additionally, the stability of gut bacterial communities can be reflected by their calculated α-diversity, thus the α-diversity, including the Shannon diversity index, Simpson’s index, abundance coverage-based estimator (ACE), and Chao1, was evaluated in this study. Higher values of Chao1, ACE and the Shannon diversity index represent a higher richness of bacterial communities, while the higher value of Simpson’s index reflects a lower evenness of bacterial communities. Although the colonic mucosal bacterial α-diversity indexes were unchanged with ELG intervention (P > 0.05), the PWG intervention significantly increased the ACE and Chao1 of the colonic mucosal bacteria (P < 0.05; Fig. 2C). This indicates that PWG intervention improved the stability of colonic mucosal bacteria. As shown in Fig. 2D, neither ELG nor PWG intervention affected the Shannon diversity index, Simpson’s index, ACE, or Chao1 in colonic digesta (P > 0.05). The combined effect of ELG and PWG on changes of colonic bacterial α-diversity was not observed (P > 0.05).

FIG 2.

Differential effects of the ELG and PWG interventions on bacterial diversity of colonic mucosa and digesta (n = 6). (A) PCoA profile and AMOVA P values of the C/C, C/G, G/C, and G/G groups in colonic mucosa. (B) PCoA profile and AMOVA P values of the C/C, C/G, G/C, and G/G groups in colonic digesta. (C) Indexes of colonic mucosal bacterial community. (D) Indexes of colonic digesta bacterial community. When the P value for an interaction was <0.05, there was considered to be a significant interaction effect between the ELG and PWG interventions.

Impact of ELG and PWG intervention on the colonic bacterial composition.

We next investigated how different means of GOS intervention regulated the colonic bacterial composition. After we categorized the sequencing data, 20 phyla were detected at the phylum level, where Firmicutes, Bacteroidetes, and Proteobacteria were 3 of the dominant phyla (the total of their relative abundance over 90%) in colonic mucosa and digesta (Fig. 3A). Thus, we selected these 3 phyla to statistically investigate the changes of their relative abundance in colonic mucosa and digesta. Upon ELG or PWG intervention, the relative abundances of Firmicutes, Bacteroidetes, and Proteobacteria did not change in the colonic mucosa or digesta (P > 0.05; Fig. 3B). Moreover, the combined effect of the ELG and PWG interventions did not influence the relative abundances of Firmicutes, Bacteroidetes, and Proteobacteria (P > 0.05).

FIG 3.

Differential effects of the ELG and PWG interventions on the relative abundance of bacterial phylum in colonic mucosa and digesta (n = 6). (A) The relative abundance of dominant bacterial phylum in colonic mucosa and digesta. (B) The relative abundance of Firmicutes, Bacteroides, and Proteobacteria in colonic mucosa. (C) The relative abundance of Firmicutes, Bacteroides, and Proteobacteria in colonic digesta. When the P value for an interaction was <0.05, there was considered to be a significant interaction effect between the ELG and PWG interventions.

Afterwards, we selected the dominant genus with more than 1% relative abundance in at least one group to statistically investigate the changes of their relative abundance in colonic mucosa and digesta (Fig. S2A and S2B). Using a heat map cluster can directly illustrate the changes in the relative abundance of different genera in different groups and clearly present clustering relationships. Thus, through an Euclidean hierarchical cluster analysis, we found that the changes of dominant genera relative abundances in C/C group were dramatically different from that in the G/G group in colonic mucosa and digesta. In contrast, the changes of dominant genera relative abundances in C/G group were similar to that of the G/C group (Fig. 4 and 5). In the colonic mucosa, the ELG intervention significantly increased the relative abundance of Prevotellaceae NK3B31 group, and the PWG intervention significantly increased the relative abundances of Prevotella 9 and Prevotellaceae Unclassified (P < 0.05; Fig. 4A and Fig. S2A). In addition, the PWG intervention significantly decreased the relative abundance of Halomonas (P < 0.05; Fig. 4A and Fig. S2A). In colonic digesta, the ELG intervention significantly elevated the relative abundances of Ruminococcaceae UCG-014 and Faecalibacterium (P < 0.05; Fig. 5A and Fig. S2B), while the PWG intervention significantly increased the relative abundances of Clostridium sensu stricto 1 and Terrisporobacter (P < 0.05; Fig. 5A and Fig. S2B). These data demonstrated that both the ELG and PWG interventions changed the dominant bacterial relative abundance in different colonic niches at the genus level.

FIG 4.

Differential effects of the ELG and PWG interventions on the relative abundance of dominant bacterial genus in colonic mucosa of weaning piglets (n = 6). (A) Heat map cluster analysis of the abundance of the colonic dominant genera in colonic mucosa. (B-F) The genus abundances with the interaction effect of the ELG and PWG interventions in the colonic mucosa. The (○) represents P value of < 0.05 for the ELG intervention, the (Δ) represents a P value of < 0.05 for the PWG intervention, and the (#) represents a P value of < 0.05 for the interaction effect. Among the four groups, identical letters represent no significant difference (P > 0.05); different letters represent significant difference (P < 0.05).

FIG 5.

Differential effects of the ELG and PWG interventions on the relative abundance of dominant bacterial genus in colonic digesta of weaning piglets (n = 6). (A) Heat map cluster analysis of the abundance of the colonic dominant genera in colonic digesta. (B and C) The genus abundances with the interaction effect of the ELG and PWG interventions in the colonic digesta. The (○) represents P value of < 0.05 for the ELG intervention, the (Δ) represents P value of < 0.05 for the PWG intervention, and the (#) represents P value of < 0.05 for the interaction effect. Among the four groups, identical letters represent no significant difference (P > 0.05); different letters represent significant difference (P < 0.05).

Of note, we also found that the combined effect of the ELG and PWG interventions changed the relative abundances of dominant genus in different colonic niches (P < 0.05; Fig. 4 and 5). In colonic mucosa, the combined effect of the ELG and PWG interventions significantly changed the relative abundances of seven dominant genus (P < 0.05; Fig. 4B to F and Fig. S3). Although the G/G group had a lower relative abundance of Lachnospiraceae Unclassified than the G/C group, the relative abundances of Prevotella 2 and Prevotella 1 in G/G group were higher than those in other groups. In addition, the relative abundance of Bacteroidales RF16 group norank was increased in C/G group while the C/C group had the highest relative abundance of Actinobacillus among all four groups. Moreover, as shown in Fig. 5B and C, the combined effect of ELG and PWG interventions significantly changed the relative abundances of two dominant genus in the colonic digesta (P < 0.05). The G/G group had higher relative abundances of Dorea and Phascolarctobacterium than other groups.

Function prediction of colonic bacteria.

We next performed PICRUSt (phylogenetic investigation of communities by reconstruction of unobserved states) to further predict the differences in bacterial function profiles with different means of GOS intervention in the colon. We obtained the gene categories using the normalized OTUs and then predicted the gene categories at level two of KEGG orthology groups in the KEGG database. Here, 39 gene families were identified in the colonic digesta and mucosa at level two (Fig. S4A and S4B). In line with the results of β-diversity mentioned above (Fig. S1C), we found that the pathway abundance values of different colonic niches were clearly distinguished using a principal-component analysis (Fig. S5). At level two, approximately 50% of the genes of colonic bacteria were mainly enriched in membrane transport, carbohydrate metabolism, amino acid metabolism, replication and repair, translation, and energy metabolism. Therefore, we selected these dominant pathways for further statistical analysis (Fig. S6). In colonic mucosa, both the ELG and PWG interventions significantly enriched the relative abundance of translation in piglets (P < 0.05; Fig. S6E). Accordingly, the PWG intervention significantly decreased the relative abundance of membrane transport (P < 0.05) while it significantly increased the relative abundance of energy metabolism (P < 0.05; Fig. S6A and S6F). In colonic digesta, the relative abundance of dominant pathways was not changed according to the ELG intervention (P > 0.05; Fig. S6H to S6M). The relative abundance of amino acid metabolism was significantly increased in PWG intervention piglets (P < 0.05), and the relative abundance of carbohydrate metabolism was significantly decreased (P < 0.05; Fig. S6I and S6J). However, there was no combined effect of the ELG and PWG interventions on the relative abundance of the dominant pathways in colonic mucosa and digesta (P > 0.05; Fig. S6).

SCFA production in the colon.

Galacto-oligosaccharide intervention has been reported to alter SCFA production of colonic bacteria. Thus, we measured the SCFA concentration in colonic lumen. As shown in Fig. 6 and Fig. S7, ELG and PWG interventions significantly increased the concentrations of isobutyrate and acetate, respectively. In addition, we found that the ELG and PWG interventions significantly increased the concentrations of total SCFAs (P < 0.05; Fig. 6E). Of note, the concentrations of butyrate and valerate significantly changed in the colonic lumen with the combined effect of the ELG and PWG interventions (P < 0.05), and the concentrations of butyrate and valerate in G/G group were higher than those in other groups (P < 0.05; Fig. 6C and D).

FIG 6.

Differential effects of the ELG and PWG interventions on SCFA concentration in the colon of weaning piglets (n = 6). (A) Acetate concentration. (B) Propionate concentration. (C) Butyrate concentration. (D) Valerate concentration. (E) Total SCFAs. Among the four groups, identical letters represent no significant difference (P > 0.05); different letters represent significant difference (P < 0.05).

Colonic mucosal immune homeostasis.

To investigate whether the changed bacterial composition would regulate the colonic mucosal immune homeostasis, we first measured the gene expression of pattern recognition receptors (PRRs). As shown in Fig. S8A, the ELG and PWG interventions significantly decreased the colonic mucosal toll-like receptor 2 (TLR2) expression. In addition, the ELG intervention decreased the relative mRNA expression of TLR5 and nucleotide-binding oligomerization domain-containing protein 1 (NOD-1) (P < 0.05) in the colonic mucosa while the PWG intervention significantly decreased the relative mRNA expression of TLR2 and TLR4 (P < 0.05; Fig. S8). No combined effect of the ELG and PWG interventions was observed in the relative gene expression of PRRs in colonic mucosa (P > 0.05; Fig. S8). To further assess the alteration of PRRs downstream signaling, we investigated the gene expression of NF-κB in the colonic mucosa. With the combined effect of the ELG and PWG interventions, the gene expression of NF-κB and MyD88 significantly changed in colonic mucosa (P < 0.05; Fig. 7A and B), and the expression of NF-κB and MyD88 in the G/G group was markedly lower than those of other groups (P < 0.05).

FIG 7.

Differential effects of the ELG and PWG interventions on the mRNA expression of MyD88-NF-κB signaling (A and B) and the protein expression of cytokines (C to F) in the colonic mucosa of weaning piglets (n = 6). Among the four groups, identical letters represent no significant difference (P > 0.05); different letters represent significant difference (P < 0.05).

In addition, we also measured the colonic mucosal cytokines concentration. As shown in Fig. 7C, the ELG and PWG interventions significantly decreased the colonic mucosal interleukin 8 (IL-8) concentration (P < 0.05). Of note, consistent with the changes in NF-κB signaling (Fig. 7A and B), under the combined effect of the ELG and PWG interventions, we also found that the concentrations of IL-10, IL-1β and tumor necrosis factor α (TNF-α) significantly changed in colonic mucosa (P < 0.05; Fig. 7D to F). As shown in Fig. 7D to F, the level of IL-10 in the G/C and G/G groups significantly increased compared to that of the C/C group (P < 0.05) but was significantly decreased compared to that of the C/G group (P < 0.05). The level of IL-1β in the C/C and C/G groups was significantly increased compared to that of G/G group (P < 0.05) but significantly decreased compared to that of the G/C group (P < 0.05). Moreover, the level of TNF-α in the G/G group was significantly decreased compared to that of other groups (P < 0.05).

Colonic mucosal barrier function.

To further illustrate the relationship between the changed bacterial composition and barrier function in the colon, we measured the expression of genes associated with barrier function in colonic mucosa. As shown in Fig. 8, the ELG intervention significantly increased the gene expression of occludin, claudin-2, and mucin-4 (P < 0.05) in the colonic mucosa. In addition, the PWG intervention significantly increased the gene expression of zonula occludens 1 (ZO-1) and mucin-1 (P < 0.05). However, there was no combined effect of the ELG and PWG interventions observed in barrier function gene expression in the colonic mucosa (P > 0.05).

FIG 8.

Differential effects of the ELG and PWG interventions on the mRNA expression of barrier protein (A to C) and mucins (D to F) in the colonic mucosa of weaning piglets (n = 6).

DISCUSSION

Weaning is one of the critical events in the life cycle of a pig and could lead to intestinal ecosystem and immune system disorders that induce piglet growth inhibition, particularly during the first week after weaning (1, 7). Hence, we selected piglets that were under weaning stress as research objects to understand the different effects of ELG and PWG interventions on the colonic microbial composition and colonic function. As expected, the growth rate of weaning piglets was sharply retarded compared to that of piglets at the suckling stage, which indicated the weaning stress that piglets experience. Although it has been reported that the early-life functional oligosaccharides intervention could improve the growth of suckling piglet (4, 13), our observations indicated that the piglets who received the PWG intervention exhibited a higher growth performance than those who received ELG intervention during the weaning transition. In general, previous studies mainly showed how to decrease diarrhea frequency by alleviating the upper intestinal infection in the piglets during the weaning transition (14, 15), but microbial dysbiosis of the hindgut could also induce diarrhea in weaning piglets (16). Therefore, several recent studies focused on the changes of colonic bacteria in weaning piglets (17, 18). In this study, we found that the ELG and PWG interventions could reduce diarrhea frequency in weaning piglets. These findings prompted us to further investigate whether the ELG and PWG interventions would alter the colonic bacterial composition and their metabolism. Moreover, we also sought to understand whether the changes in bacterial composition and metabolism might mediate the alterations of colonic function during the weaning transition.

To comprehensively understand the differential effects of the ELG and PWG interventions on colonic microbial composition in weaning piglets, we investigated the microbial composition in different colonic niches (mucosa and lumen). In agreement with previous findings, the patterns in the PCoA showed that the colonic lumen and mucosa were dramatically distinguished (11, 12), and the results of the PCoA suggested that it is more amenable to alter the colonic mucosal composition than that of colonic lumen by using GOS intervention. Meanwhile, comparing to the results of the ELG intervention, the weaning piglets that received the PWG intervention had higher α-diversity in the colonic mucosa. In the weaning piglets, the decreased microbial diversity is a hallmark of gut ecosystem dysbiosis, suggesting that gut microbial communities with higher diversity would be beneficial to maintain intestinal ecosystem homeostasis (7, 19). Thus, the increased colonic mucosal microbiota diversity in the piglets during weaning transition that was caused by the PWG intervention could reduce the severity of colon dysfunction induced by weaning stress.

Besides losing the diversity of microbiota, piglets with weaning stress underwent a disrupted state of the gut microbial composition, which can lead to severe gut dysfunction (7). In different colonic niches, we found the ELG intervention increased the abundance of Prevotellaceae NK3B31 group, Ruminococcaceae UCG-014, Ruminococcaceae NK4A214 group, and Faecalibacterium while the PWG intervention increased the abundance of Prevotella 9, Prevotellaceae Unclassified, and Bacteroidales S24-7 group norank. The bacteria that belonged to the Prevotella genus are attracting more interests as a commensal in the hindgut because of their ability to degrade diet fiber and oligosaccharides that could modulate the host metabolism (20). Moreover, the members of Ruminococcaceae, Bacteroidales and Faecalibacterium are defined as important commensal bacteria in the piglet intestine where they play a key role in maintaining gut immune homeostasis (21). Thus, the ELG and PWG interventions could maintain a higher abundance of beneficial commensal bacteria during the piglet weaning stage. Interestingly, the abundance of Clostridium sensu stricto 1 was sharply increased during PWG intervention in different colonic niches. Previous studies pointed out that intestinal Clostridium sensu stricto 1 abundance is positively associated with diet protein level because it is a protein-utilizing bacterium (22). Because the PWG intervention promoted the average food intake of weaning piglets, we speculated that many undigested protein-associated substrates that accumulated in colonic lumen promoted the enrichment of Clostridium sensu stricto 1. Notably, under the combined effect of the ELG and PWG interventions, we found several commensal bacteria, such as the Lachnospiraceae Unclassified, Prevotella 1, Prevotella 2, Bacteroidales RF16 group norank, Dorea, and Phascolarctobacterium, that were enriched in the piglet colon. These bacteria proved to be beneficial to maintain gut health (23, 24). Moreover, the abundance of Actinobacillus was enriched in colonic mucosa of the C/C group. A previous study showed that some species that belong to Actinobacillus can trigger infection and inflammation (25). Thus, these findings demonstrated that the ELG and PWG interventions may relieve colonic ecosystem dysbiosis by increasing the abundance of beneficial commensal bacteria, and that these alterations were enhanced by the combined effect of the ELG and PWG interventions.

Here, PICRUSt was used to further predict the effect of the ELG and PWG interventions on colonic bacterial function. After analyzing the dominant pathways, we found that the pathways, such as the replication and repair pathways, translation, and energy metabolism, that associated with bacterial survival and proliferation were increased with the PWG intervention. These findings are consistent with the changes in colonic mucosal bacterial α-diversity. As expected, the PWG intervention increased the abundance of amino acid metabolism in colonic digesta while it decreased the abundance of carbohydrate metabolism. Indeed, the digestive physiology of the small intestine in weaning piglets was also disrupted, which, in turn, caused the difference of fermentable components in the hindgut (26). Mountzouris et al. (27) showed that growing pigs that consumed the diet with 1% trans-GOS exhibited a higher digestibility of carbohydrate components than those feeding on the control diet. Here, we postulated that the alterations in colonic bacterial function were determined by the digestive capacity of the foregut, but this speculation requires future exploration.

The SCFAs, mainly acetate, propionate, and butyrate, are produced through bacterial fermentation of undigested substrates enter the colon (28). Our results indicated that the ELG and PWG interventions could increase the abundances of SCFA producers in the piglet colon, such as the members of Prevotellaceae, Ruminococcaceae, Bacteroidales, Faecalibacterium and Clostridium, and enhance SCFA concentrations in the colonic lumen. Interestingly, under the interaction effect of ELG and PWG intervention, the butyrate concentration was significantly increased in the G/G group, while there were no classical butyrate-producing bacteria enriched in the G/G group. In contrast, the enrichment of acetate producers, such as Prevotella 1 and Prevotella 2, was observed in the G/G group. In addition to directly utilizing the substrates, the gut butyrate producer can also produce butyrate through the metabolic cross-feeding pathway (29). For example, members of the Faecalibacterium family are capable of transforming acetate to butyrate by the butyryl-CoA:acetate CoA-transferase pathway in vitro. A previous study confirmed that mutualistic cross-feeding interactions, which are more common in the hindgut than those in the foregut, were promoted by anaerobic conditions (30, 31). Therefore, the interaction effect of the ELG and PWG interventions may promote the bacterial metabolic cross-feeding in the colon in the G/G group piglets, and these changes contribute to the increment of butyrate concentration. In short, our results indicated that the ELG and PWG interventions increased the total SCFA concentration by enriching the colonic SCFA producer abundances. Notably, under the combined effect of the ELG and PWG interventions, the concentrations of butyrate were increased in the G/G group piglets.

Under normal physiological conditions, most colonic SCFAs are in the ionic form and require H⁺- or Na⁺-coupled transporters for absorption (32). Monocarboxylate transporter 1 (MCT-1) and sodium-coupled MCT-1 (SMCT-1) are two main SCFA transporters located on the apical membrane of colonic epithelial cells and directly participate in electrolyte absorption with increases of Na⁺ and Cl⁻ absorption and release of HCO₃⁻ in the colonic lumen (33). Because SCFAs could improve electrolyte exchange in the lumen, which is related to the water absorption and diarrhea development in piglets, we postulated that the ELG and PWG interventions would reduce the frequency of piglet diarrhea partly through increasing colonic SCFA concentrations. In addition, the SCFAs are a main energy source for colonic epithelial cells (3). Zhou et al. (34) illustrated that the weaning process has a significant effect on colonic growth and development, which might be associated with the change in SCFA concentrations in colon. In line with these findings, we found that the PWG intervention could improve colonic indexes by increasing the total SCFA concentrations. Thus, our results indicated that the ELG and PWG interventions effectively maintained colonic physiology function while the PWG intervention improved colonic growth. Each of these changes were mediated by the increased SCFA concentration in the colon.

Beyond the maintaining of colonic physiology function and growth, previous studies also highlighted that a high intestinal SCFA level in weaning piglets is essential for modulating the colonic immune responses and barrier function (35, 36). We first evaluated the colonic immune response in current study. Toll-like receptors (TLRs) and NOD-like receptors (NLRs) are two mainly PRRs in mucosal immune system (37). According to our results, the ELG and PWG interventions selectively reduced the mRNA expression of TLRs or NLRs in colonic mucosa, while the gene expression of MyD88 and NF-κB was sharply downregulated in the G/G group when under the combined effect of the ELG and PWG intervention. MyD88-NF-κB is the main signal located downstream of PRRs and is considered a crucial role in modulating gut mucosal immune responses (37). The previous study confirmed that the influences of butyrate on colonic inflammatory responses resulted in the inhibition of NF-κB activation (38, 39). Consistent with this review, the enhancement of butyrate concentrations contributed to the reduction in mRNA expression of MyD88-NF-κB signaling in the colonic mucosa in the G/G group. Meanwhile, we also observed that proinflammatory cytokines, such as IL-1β and TNF-α, were decreased in the colonic mucosa of the G/G group piglets under the interaction effect of the ELG and PWG interventions. These results were consistent with the changes of MyD88-NF-κB signaling in our present study. Mucins and barrier proteins are the key factors to maintain the colonic epithelial barrier function (3, 40). We found that the ELG and PWG interventions upregulated the mRNA expression of mucins in colonic mucosa. Paassen et al. (41) illustrated that butyrate could improve MUC2 expression in LS174T cells by modulating the activation of the MUC2 promoter. Thus, the enhanced SCFA concentration contributed to the increased expression of mucins in the colonic mucosa. Ours and others’ in vivo and in vitro studies have confirmed that the SCFAs produced by gut microbiota can enhance the gut barrier function by the activation of the AMP-activated protein kinase pathway (3, 42). In agreement with these findings, the ELG and PWG interventions increased the mRNA expression of barrier proteins by enhancement of SCFA concentrations. Here, we found the ELG and PWG interventions modulated the colonic mucosal mRNA expression of PRRs and barrier proteins by enhancing the colonic SCFA production in piglets during the weaning transition. Under the combined effect of the ELG and PWG interventions, the gene expression of MyD88-NF-κB and the proinflammatory cytokines were reduced in the G/G group piglets, and these alterations may be caused by the upregulation of colonic butyrate concentration.

In summary, we demonstrated that the ELG and PWG interventions increased the abundance of SCFA producers and the SCFA content and that these changes selectively altered the gene expression of pattern recognition receptors and colonic function in weaning piglets. In addition, under the combined effect of the ELG and PWG interventions, the butyrate concentration was increased in the colon of weaning piglets while the proinflammatory cytokines were decreased through inhibition of the gene expression of MyD88-NF-κB. The findings of this study provide new insights into prebiotic intervention to alleviate intestinal ecosystem dysbiosis and gut dysfunction in the weaning piglets.

MATERIALS AND METHODS

All animal care procedures in our study were according to the Experimental Animal Care and Use Guidelines of China, and in vivo experimental protocols were approved by the Animal Care and Use Committee of Nanjing Agricultural University (Nanjing, Jiangsu province, China).

Animal experimental design.

One hundred and twenty healthy neonatal piglets (Duroc × Landrace × Yorkshire) with an average birth weight of 1.47 ± 0.04 kg were obtained from 12 sows (10 piglets in each litter) with the similar parity (3 or 4 parities). All piglets were raised under the same condition in a commercial farm in Jiangsu Province, China. Twelve sows with their 120 neonatal piglets were equally divided into a control group (C/C group), an ELG intervention group (G/C group), a PWG intervention group (C/G group), and an ELG + PWG intervention group (G/G group). Food grade GOS (purity is approximately 90%) was purchased from Quantum Hi-Tech Biological Co., Ltd. (GuangDong, China), and the composition was described by Tian et al. (19). From the age of 1 to 7 days, the piglets in the G/C and G/G groups were administered 10 mL GOS solution (1g/kg) by oral gavage every day, whereas the piglets in the C/C and C/G groups were administered the same dose of physiological saline solution by oral gavage. All piglets had free access to water and were not given creep feed during the suckling stage. All piglets were weaned on day 21. During the postweaning period, the piglets in the C/C and G/C groups received a basal diet and piglets of the C/G and G/G groups received a GOS-supplemented diet (basal diet + 2% GOS). The composition and nutrient level of weaning diets are shown in Table S1.

Growth performance and diarrhea frequency.

Bodyweight of piglets was recorded on days 7, 14, 21, and 28. Daily feed intake of the piglets was recorded in postweaning period. Diarrhea incidence of weaning piglets was visually assessed each morning at 09:00. Diarrhea frequency (%) is the number of piglets with diarrhea in each pen × diarrhea days/(total number of piglets × experimental days) × 100.

Sampling.

On day 28, six piglets from each group were randomly selected for sampling. After the piglet was slaughtered, the colon was ligated, separated, and removed. The length of the colon was measured after stripping off the mesentery by holding the intestine against a ruler. Digesta of the proximal colon, which was approximately 10 cm behind the junction of cecum and colon, were collected and stored at −80°C until analysis. The wet weight of the colon was determined after gently squeezing out the contents. In addition, the proximal colon tissue was excised and rinsed in ice-cold phosphate buffer saline (PBS). Mucosal samples were scraped from colon tissue using a sterile glass microscope slide. The samples were frozen in liquid nitrogen immediately and kept at −80°C until analysis.

DNA extraction, 16S rRNA gene amplification, and high-throughput sequencing.

The method of DNA extraction, 16S rRNA gene amplification, and high-throughput sequencing were the same as that employed in our previous study (12). In brief, according to the manufacturer’s instructions, total genomic DNA of colonic mucosa and colonic digesta was extracted by using the PowerSoil DNA isolation kit (MoBio Laboratories, Carlsbad, CA). The concentration of extracted genomic DNA was measured by a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA), and DNA samples were stored at −80°C until further 16S rRNA gene MiSeq sequencing. A universal primer (F 5′-ACTCCTRCGGGAGGCAGCAG-3′ and R 5′-GGACTACCVGGGTATCTAAT-3′) was used to amplify the V3-V4 region of the bacterial 16S rRNA gene. The PCR process included initial denaturation (95°C for 2 min), 25 cycles of denaturation (95°C for 30 s), annealing (55°C for 30 s), elongation (72°C for 30 s) and a final extension (72°C for 5 min). The PCR products were detected by 2% agarose gel electrophoresis and purified by using AxyPrep DNA Gel Extraction kit (Axygen Biosciences, Union City, CA). Purified products were pooled in equimolar and paired end sequenced (2 × 250) on an Illumina MiSeq platform according to the standard protocols.

Bioinformatics analysis.

As described by Mao et al. (12), raw FASTQ files were demultiplexed and quality-filtered using QIIME (version 1.70) with a standard criterion. OTUs were clustered with a 97% similarity cutoff using UPARSE (version 7.1 http://drive5.com/uparse/), and chimeric sequences were identified and removed using UCHIME.20. The most abundant sequences within each OTU were designated as representative sequences and were classified using the Ribosomal Database Project (RDP) classifier with a standard minimum support threshold of 80%. Diversity of colonic mucosal and digesta bacteria (Shannon diversity index, Simpson’s index, ACE, and Chao1) was assessed using mothur v.1.29.0.

Predicted molecular function based on 16S rRNA data using phylogenetic investigation of communities by reconstruction of unobserved states.

Phylogenetic investigation of communities by reconstruction of unobserved states (PICRUSt, http://picrust.github.io/picrust/) was used to predict the molecular functions of each sample based on 16S rRNA gene sequencing data. Based on the first picking closed OTUs against the Greengenes 16S rRNA gene database (13.5) using QIIME 2, OTUs with known bacterial genomes were precalculated in PICRUSt. To reflect the true abundances of the underlying bacteria, PICRUSt normalized the OTU table with 16S copy number predictions. The data were used for metagenome inference of KEGG orthologs by using PICRUSt. The difference in predicted molecular functions of the colonic mucosal and digesta bacterial communities was determined using a PCA using MetaboAnalyst web server (http://www.metaboanalyst.ca).

Measurements of microbial metabolites.

As our previous report described, a gas chromatography method (Shimadzu, GC-14A with a flame ionization detector [FID], Japan) was used to determine the concentrations of SCFAs in the colonic digesta (43). Briefly, a capillary gas chromatography (GC) column (30 m × 0.32 μm × 0.25 mm; Sigma-Aldrich, St. Louis, MO) was used. The temperatures of the injector, column, and detector were 180°C, 110°C, and 180°C, respectively.

RNA extraction and quantitative real-time PCR.

Total RNA was extracted from colonic mucosa with TRIzol (TaKaRa Bio, Otsu, Japan) according to the manufacturer’s instructions. Total RNA (1 μg) was reverse transcribed to complementary DNA (cDNA) using a PrimeScript RT reagent kit (TaKaRa Biotechnology Co., Ltd., Otsu, Japan) according to the recommended procedures. The expression of target genes was measured by quantitative real-time PCR with SYBR Premix Ex TagTM (Tli RnaseH Plus) qPCR kit (TaKaRa Biotechnology Co., Ltd., Otsu, Japan) according to the manufacturer’s guidelines, and the fluorescence was detected on a sequence detector system (Applied Biosystems 7300 SDS; Foster City, CA) according to the description of a previous study (13). Primers in this study were synthesized from Invitrogen Life Technologies (Shanghai, China), and their sequences are shown in Table S2. The 2^-ΔΔCt method was used to calculate the expression of target genes relative to housekeeping gene (β-actin) (44).

ELISA.

The concentrations of the cytokines, including interleukin 8 (IL-8), interleukin 10 (IL-10), interleukin 1β (IL-1β), and tumor necrosis factor-α (TNF-α), in the colonic mucosa were determined by using porcine enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems Co., Ltd., Minneapolis, MN) according to the manufacturer’s instructions. For standardization, the concentrations of protein were determined by the bicinchoninic acid (BCA) method using a protein assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).

Statistical analysis.

Data were analyzed by using a full-factorial generalized linear model (GLM) analysis of variance (ANOVA) with type III sums of squares in SPSS software (SPSS version 20.0, SPSS, Inc.) and expressed as the mean ± standard deviation (SD). The principal coordinate analysis (PCoA) was conducted based on the Bray-Curtis distance to compare groups of samples, and then the analysis of molecular variance (AMOVA) was conducted based on an unweighted distance to assess significant differences among samples by using mothur. The ELG intervention, PWG intervention, and their interaction were treated as the fixed factors, and the piglet was the experimental unit. All data were analyzed by a two-way ANOVA, and the differences were considered significant at P values of <0.05. When a significant effect of the interaction (P < 0.05) between the ELG and PWG interventiona was observed, the data were further analyzed by a one-way ANOVA with Tukey’s post hoc test. The P value of the bacterial analysis was adjusted with false-discovery-rate (FDR) correction by the Benjamini-Hochberg method. A P value of <0.05 was considered statistically significant.

Data availability.

The raw reads were deposited into the NCBI Sequence Read Archive database (accession number SAMN16481017).