Longitudinal development of the gut microbiota in healthy and diarrheic piglets induced by age‐related dietary changes

Abstract Diarrhea is one of the most common enteric diseases in young piglets. Diverse factors such as an unstable gut microenvironment, immature intestinal immune system, early supplementary feeding, and weaning often induce dysfunction of gut microbiota, thus leading to a continuing high incidence of diarrhea in piglets. However, few studies have characterized the gut microbiota of diarrheic piglets following changes in diet and during the development of intestinal physiology. In this study, we used 16S rRNA gene sequencing to analyze the dynamic establishment of fecal microbiota in six healthy piglets in response to age‐related changes in the diet: sow‐reared, early supplementary creep‐feeding (sow‐reared + starter diet), and weaning (solid nursery diet). We compared the gut microbiota of these six healthy piglets with those of diarrheic piglets during each of the three dietary stages (n = 10 sow‐reared, n = 10 early supplementary creep‐feeding, and n = 5 weaning). We found that weaning (solid nursery feeding) was the primary factor leading to dynamic colonization by microbiota in healthy piglets, and diarrhea primarily affected the microbial communities of piglets before weaning. Healthy piglets showed a continuous decrease in Lactobacillus and Escherichia, as well as a gradual increase in Prevotella with the transition to solid food. An altered relationship between Prevotella and Escherichia may be the main cause of diarrhea in preweaned piglets, whereas reduced numbers of Bacteroides, Ruminococcus, Bulleidia, and Treponema that are responsible for the digestion and utilization of solid feeds may be related to the onset of postweaning piglet diarrhea. The Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) functional analysis indicated that a reduction in genes involved in carbohydrate metabolism induced by intestinal dysbacteriosis in diarrheic piglets was one of the major causes of diarrhea at the three dietary stages. These findings provide insights into developing an intervention strategy for better management of diarrhea in piglets.


| INTRODUC TI ON
The gastrointestinal microbiota of mammals play numerous roles in maintaining host health, such as regulating intestinal nutrient metabolism, synthesizing vitamins, promoting the development and maturation of the gut-associated immune system, and protecting against pathogenic bacteria (Buffie & Pamer, 2013;Kabat, Srinivasan, & Maloy, 2014;Richards, Gong, & Lange, 2005). Diarrhea and gut microbial dysbiosis show reciprocal causation (Zoetendal, Akkermans, Vliet, Visser, & Vos, 2000). In the diarrheic condition, some normal indigenous microflora are discharged and invasive pathogens are abnormally increased, which preferentially lead to a disproportionate amount of beneficial microbes in the gastrointestinal tract; conversely, harmful substances produced by abnormal flora cause abnormalities in the intestinal function and immune response, which in turn cause the development of diarrhea and enteritis (Ward et al., 2016).
Piglet diarrhea is the most common enteric disease in swine, which leads to low growth rates and high mortality rates in neonatal and young piglets. Morris, Davies, and Lawton (2002) reported that almost half (49%) of piglet deaths are caused by diarrhea, which has a detrimental economic impact on the swine industry. To improve the utilization efficiency of sows on modern, intensive pig farms, piglets are usually given supplementary feeding at 1-2 weeks of age and weaned (with a solid nursery diet) at 3 or 4 weeks of age. However, the microenvironment of the gastrointestinal tract is particularly vulnerable to disruptions during the dynamic development of the intestinal mucosal structure and function in young piglets (Isaacson & Kim, 2012). Furthermore, the immune system of piglets undergoes significant postnatal development and has been suggested as a factor influencing the structure of intestinal community (Bailey et al., 2001). These diverse factors often induce the dysfunction of intestinal microflora, thus resulting in the high incidence of diarrhea in piglets.
Several investigations on intestinal microbiota have been focused on dynamic colonization at different ages or following dietary changes in healthy piglets, particularly at weaning (Frese, Parker, Calvert, & Mills, 2015;Slifierz, Friendship, & Weese, 2015). These studies provide evidence that the switch from sows' milk to solid diets critically influences microbial colonization in the immature gastrointestinal tract. Furthermore, several lines of evidence suggest that both specific bacterial species and microbial communities exert either pathogenic effects that facilitate diarrheal disease or probiotic effects that enhance intestinal health (Azcarate-Peril et al., 2011;Jonach, Boye, Stockmarr, & Jensen, 2014). For instance, the occurrence of neonatal piglet diarrhea is closely related to increases in Prevotella, Sutterella, Campylobacter, and Fusobacteriaceae, as well as a decrease in several members of the phylum Firmicutes (Hermann-Bank et al., 2015;Yang et al., 2017). The development of mucohemorrhagic diarrhea in Brachyspira hyodysenteriae-infected piglets was associated with the perturbation of intestinal microbiota, with inoculated diarrheic pigs having lower bacterial numbers and a lower Bacteroidetes:Firmicutes ratio than uninoculated pigs (Costa, Chaban, Harding, & Hill, 2014). Dietary supplements of Enterococcus faecalis LAB31 in weaned piglets may increase the numbers of Lactobacillus and consequently reduce the incidence of diarrhea (Hu et al., 2015).
Given these crucial implications, animal scientists have started to perform microbial intervention strategies to improve animal health status, as well as prevent or treat enteric diseases. Zhang et al. (2017) revealed that oral administration of a Bacillus mix reprograms the gut microbiota and enhances goblet cell function to ameliorate pig diarrhea caused by enterotoxigenic Escherichia coli infection. A recent study indicated that fecal microbiota transplantation (FMT) could reduce susceptibility to epithelial injury and modulate tryptophan metabolism in the microbial community of a piglet model . Hu, Geng, et al. (2018) and Hu, Ma, et al. (2018) established a standardized model of exogenous FMT in pigs and subsequently found that FMT and oral administration of Lactobacillus gasseri LA39 and Lactobacillus frumenti strains could modulate the structure of gut microbiota and prevent diarrhea induced by early weaning stress in recipient piglets. In children, microbiota reconstitution has been used to prevent and treat diarrhea caused by Clostridium difficile infection (Buffie et al., 2015). Thus, it is conceivable that identification of the gut microbiota in diet-matched diarrheic piglets could also provide translatable knowledge regarding digestive physiological mechanisms, as well as microbial prevention and therapy for diarrheic disease. However, very few studies have characterized the intestinal microbiota of diarrheic piglets during intestinal development and the associated dietary changes.
In this study, we first characterized the dynamic establishment of individual microbiota in healthy piglets following age-related dietary changes: sow-reared (sows' milk), early supplementary creep-feeding (sows' milk + starter diet), and weaning (solid nursery diet). Second, we examined the differences in the intestinal microbial balance between diarrheic and healthy piglets during these dietary changes.
We demonstrated an obvious continuous decrease in Lactobacillus and Escherichia, as well as a gradual increase in Prevotella in healthy piglets with the transition to solid food. An altered relationship between Prevotella and Escherichia may be the main cause of diarrhea in preweaned piglets, while reduced numbers of Bacteroides, Ruminococcus, Bulleidia, and Treponema may be related to the onset of postweaning piglet diarrhea. These findings provide insights for the development of an intervention strategy for better management of diarrhea in piglets.

| Animals and sample collection
Piglets were the progeny of six third-parity, healthy Large White sows that were subjected to artificial insemination within 2 days and maintained under identical husbandry practices and epidemic prevention systems on a commercial farm in Gansu province, China, during July 2015. The farm had no previous history of bacterial or viral infections. At birth, the piglets were marked by ear notching for individual identification. They all received identical standards of housing and management, and were exclusively sow-reared during the first 2 weeks, followed by supplementation with piglet early starter diets at 2 weeks of age. After weaning at 4 weeks of age, piglets were raised on solid nursery feed. The starter diet provided 14.6 MJ/kg digestible energy (DE), 21.0% crude protein, 4.0% crude fiber, 7.0% crude ash, 0.5%-1.2% calcium, 0.55% total phosphorus, 0.3%-1.2% salt, and 1.3% lysine from the 15th to 28th days. For the next 2 weeks after weaning, the diet contained 14.6 MJ/kg DE, 20% crude protein, 4.0% crude fiber, 7.0% crude ash, 0.5%-1.2% calcium, 0.55% total phosphorus, 0.3%-1.2% salt, and 1.45% lysine.
The piglets were given ad libitum access to water throughout the experimental period. The general health of each piglet was closely monitored with special attention to fecal consistency and disease history. According to these observations, piglets were defined as diarrheic or healthy and their feces were collected as described previously (Yang et al., 2017). Briefly, diarrheic piglets were those showing signs of diarrhea for at least two consecutive days with thickened and watery feces; healthy piglets never experienced diarrhea or other diseases. Diarrheic samples were collected directly using sterile tools (week 1) or sterile cotton swabs (week 3 and week 5) and immediately frozen in liquid nitrogen.
The experimental design of this study is illustrated in Figure 1.
To analyze the microbial structure and diversity of healthy piglets at the sow-reared (sows' milk), early supplementary creep-feeding (sows' milk + starter diet) and weaning (solid nursery diet) stages, fecal samples from six healthy piglets with similar birth weights and from different litters were collected using sterile cotton swabs at 1 week (H_week1), 3 weeks (H_week3), and 5 weeks (H_week5) of age, respectively. A case-control study was conducted to investigate the fecal microbial profiles of diet-matched diarrheic and healthy piglets. Diarrheic samples collected at 7-9 days were classified as week 1 (D_week1, n = 10), those at 24-28 days were classified as week 3 (D_week3, n = 10), and those at 35-40 days were classified into week 5 (D_week5, n = 5). The six H_week1 samples and 10 D_week1 samples were also used in our previous study (Yang et al., 2017).   filtering of raw reads was performed using the Quantitative Insights into Microbial Ecology (QIIME) pipeline (Caporaso et al., 2010). The acquired sequences were chimera filtered using the UCHIME algorithm (Haas et al., 2011) by aligning them to the reference database (Gold database, http://drive5.com/uchim e/uchime_downl oad.html).

| Bioinformatics and statistical analyses
High-quality sequences were assigned to distinct operational taxonomic units (OTUs) using the UPARSE pipeline (v7.0.1001) with a 97% similarity threshold (Edgar, 2013). Representative sequences of each OTU were then taxonomically classified at different levels (phylum, class, order, family, genus, and species) by comparing them to sequences in the GreenGene database (Desantis et al., 2006) using the RDP 3 classifier algorithm (v2.2). The OTU abundance was normalized using a standard with sequence number corresponding to the sample with the least number of sequences.
Alpha diversity measurements (i.e., observed species, Shannon index, Chao1, and ACE) and beta diversity were calculated using QIIME (Caporaso et al., 2010). Intergroup differences in alpha and beta diversity of bacterial communities among treatments (diarrheic vs. healthy) and weeks (1, 3, and 5) were analyzed by a nonparametric test: the Mann-Whitney U test was chosen for analysis of two groups, while the Wilcoxon rank-sum test was chosen for analysis among more than two groups. Dissimilarity matrices of intragroup and intergroup beta distances were visualized using principal coordinate analysis (PCoA) and the unweighted pair-group method with arithmetic means (UPGMA) analysis. A one-way analysis of similarity (ANOSIM) test was performed to determine significant differences in bacterial communities among groups using the R software (v3.2.2) (Clarke & Gorley, 2006).
Linear discriminant analysis (LDA) coupled with effect size (LEfSe) was performed to identify the core bacterial taxa differentially represented among healthy piglet groups at the species or higher taxonomic levels. A size-effect threshold of 4.0 for the logarithmic LDA score was used for discriminative functional biomarkers.
Differentially abundant bacterial taxa between diarrheic and healthy piglets at each age (week) stage were detected using the Kruskal-Wallis test, and only taxa with a mean relative abundance >0.1% in at least one group were considered. Significance was considered at p < .05.

| Quantitative PCR (qPCR)
To assess the results of 16S rRNA gene sequencing, E. coli, Lactobacillus spp., and Prevotella spp., which have been reported to perform crucial functions in maintaining the intestinal health or disease of newborn animals (Bordin et al., 2013;Larsen et al., 2010), were selected for absolute quantification qPCR using previously validated bacterial group-specific 16S primers (Table A1) Data obtained from the amplification were transformed to give the number of bacterial log copies/µl feces according to the copy number of recombinant DNA plasmids. The Mann-Whitney U test was used for comparisons between two groups.

| Spearman's correlation analysis
To capture extreme exclusion relationships among key microbial taxa, Spearman's correlation coefficient (r) was analyzed among marker bacterial genera for diarrheic and healthy piglets at week 3 and week 5 stages using the SPSS 18.0 software (SPSS Inc., 2009).
Taxa with r > .6 or <−.6, and p < .05 were considered to have a strong correlation.

| Description of sequencing data
We

| Bacterial composition and community diversity associated with age-related dietary changes in diarrheic and healthy piglets
The relative abundances of the top 10 phyla and top 10 genera present in diarrheic and healthy piglets at different dietary stages are displayed in Figure 2. Bacteroidetes, Firmicutes, and Proteobacteria were the dominant phyla in both diarrheic and healthy piglets; proportions of these reads accounted for more than 95% of reads in healthy piglets at the three dietary stages (H_week1, H_week3, and H_week5), and the weaning diarrheic piglets at week 5 (D_week5). The alpha diversity of the fecal microbiota varied considerably following dietary changes in healthy pigs. The richness index (observed species, Chao1 and ACE) of healthy piglets subjected to supplementary creep-feeding (H_week3) was significantly lower than that of either healthy piglets that were sow-reared (H_week1) or those subjected to full solid feeding (H_week5) (p < .01, We found no significant difference in beta diversity of the OTU community structure in healthy piglets among the different sampling stages (Wilcoxon rank-sum test, p > .05). Beta diversity was lower in healthy piglets than in diarrheic piglets at the different sampling stages (Wilcoxon rank-sum test, p < .01; Figure 3a), indicating that diarrhea causes variance in fecal microbial structure and diversity in individual piglets. The principal coordinate analysis (PCoA)-based trajectory plot also revealed distinct structures among healthy piglets of different ages as well as diarrheic and healthy piglets of the same age, and the ANOSIM for differences was significant (Rvalue > .5, p < .01; Figure 3b). The structure of the gut microbiota became increasingly similar with age and the transition to solid food; before weaning, diarrheic and healthy piglets had distinct microbiota, while there was little difference between the two groups after weaning. Similarity cluster analysis using the UPGMA showed good agreement with PCoA analysis (Figure 3c), indicating the fecal microflora in healthy piglets is ever changing with progressive change of dietary stage, and diarrhea primarily affects the microbial communities in piglets before weaning.

| Bacterial biomarkers in healthy piglets at different age-related dietary stages
We performed LEfSe analysis to reveal the significant ranking of abundant bacterial taxa among H_week1, H_week3, and H_week5 samples. The cladogram in Figure 4a highlights 15

| Microbial taxa increased or decreased in diarrheic piglets at different age-related dietary stages
We analyzed differences in the major microbial taxa (mean relative abundance > .1% in either of the two groups) between diarrheic and healthy piglets at 3 and 5 weeks of age using the Kruskal-Wallis test ( Figure 5).
For piglets at 3 weeks of age, three phyla, 11 families, and 20 genera showed significant differences between healthy and diarrheic piglets ( Figure 5a). The relative abundances of the phyla Firmicutes (D_week3 vs. [Ruminococcus], Bulleidia, and Treponema showed significant differences between healthy and diarrheic piglets (Figure 5b). The relative abundances of all these bacterial taxa were decreased in diarrheic piglets compared with healthy piglets.
To confirm the results of 16S rRNA gene sequencing, three specific microbial communities (E. coli, Lactobacillus, and Prevotella) were subjected to qPCR ( Figure 6). Significant differences in the absolute abundance of Lactobacillus, E. coli, and Prevotella were compared between preweaned and postweaned healthy piglets (Figure 6a). After weaning, the abundance of Lactobacillus and E. coli was markedly decreased, while the abundance of Prevotella was increased in healthy piglets (p < .05). The absolute abundance of Prevotella in preweaned diarrheic piglets was greater than that in preweaned healthy piglets (p < .05), while the abundance of Lactobacillus and E. coli was significantly lower in preweaned diarrheic piglets than in preweaned healthy piglets (p < .05) (Figure 6b). These data were consistent with the results of 16S rRNA gene sequencing, indicating our differential bacterial taxonomy and abundance estimations are highly reliable.

| Correlations among bacterial genera in healthy and diarrheic piglets
We performed Spearman's correlation analysis of different bacterial genera of diarrheic and healthy piglets at week 3 and week 5. TA B L E 2 Comparison of fecal bacterial α diversity between diarrheic and healthy piglet groups at three age-related dietary stages For piglets given early supplementary creep-feed (week 3), a strong negative correlation was observed between Prevotella and Escherichia in healthy samples (r < −.6, p < .05), while this correlation was relatively weaker in diarrheic piglets. Diarrhea also caused the correlation between Lactobacillus and [Ruminococcus] seen in healthy piglets to change from negative to positive (r > .6, p < .05) ( Table 3)

| Functional differences of fecal microbiota between diarrheic and healthy piglets
We compared the functional gene composition of fecal microbiota between diarrheic and healthy piglets using the PICRUSt software. As shown in Figure 7a, notable differences in KEGG pathways between 3-week-old diarrheic and healthy piglets included the enrichment of functional genes involved in metabolic pathways, and genetic and environmental information processing (p < .05). The metabolic functional genes showing a significantly increased presence in diarrheic piglets were involved in energy metabolism, metabolism of cofactors and vitamins, glycan biosynthesis and metabolism, metabolism of terpenoids and polyketides, and the biosynthesis of other secondary metabolites.
In addition, genes associated with nucleotide metabolism were enriched in diarrheic piglets. However, crucial carbohydrate metabolism was reduced in diarrheic piglets. For the genetic and environmental information processing category, functional genes for signal transduction and membrane transport were decreased in diarrheic piglets compared with healthy piglets. At the postweaning stage (week 5), only the functional pathways of carbohydrate metabolism and the excretory system exhibited different gene abundances between diarrheic and healthy piglets (p < .05; Figure 7b). These findings suggest that one of the major causes of diarrhea is the reduced uptake and fermentation of available carbohydrates caused by enteric dysbacteriosis in diarrheic piglets.

| Microbiota variations in healthy piglets at different age-related dietary stages
Several studies have shown that the gut microbiome of piglets is rapidly colonized with increasing age and dietary changes (Pajarillo, Chae, Balolong, Bum Kim, & Kang, 2014;Slifierz et al., 2015). In the present study, we analyzed microbial composition and diversity in F I G U R E 4 Core biomarkers in healthy piglet groups by LEfSe analysis. (a) The cladogram shows the microbial species exhibiting significant differences in the three groups. Red, green, and blue nodes in the phylogenetic tree represent microbial species that play an important role in the week 1, week 3, and week 5 groups, respectively. (b) Species with significant differences in abundance that have an LDA score > 4.0. The length of the histogram represents the LDA score six healthy piglets at 1, 3, and 5 weeks of age. Alpha diversity was first decreased at 3 weeks of age and then increased at 5 weeks of age, in agreement with the results of Frese et al. (2015), and this is likely linked to the reduction in maternal antibodies absorbed in piglets by 3 weeks old. After weaning, however, with immune system development and the establishment of intestinal microflora of F I G U R E 5 Bacterial taxa showing differences between diarrheic and healthy piglets at 3 weeks (a) and 5 weeks (b) of age. Bacterial taxa with mean relative abundances >0.1% in at least one group are included piglets, microbial diversity was increased to accommodate digestion and absorption solid food.
The developing fecal microbiota of piglets in our study was similar to that observed in previous studies, which found Firmicutes, Proteobacteria, Bacteroidetes, Fusobacteria, and Actinobacteria as the predominant and stable bacterial phyla in piglets (Slifierz et al., 2015). Bacterial diversity and the core microflora of piglets varied with age, with the relative abundance of Firmicutes increasing, and the relative abundance of Proteobacteria, Fusobacteria, and Actinobacteria decreasing as piglet age increased (Slifierz et al., 2015). At the genus level, we found that Lactobacillus, Escherichia, and Clostridium were more abundant in preweaned healthy piglets than in postweaned piglets. Escherichia and Lactobacilli are commonly found in newborn animal intestines (Collado, Cernada, Baüerl, Vento, & Pérez-Martínez, 2012). Escherichia coli is a representative species of the genus Escherichia; most E. coli strains are harmless and exert a barrier effect against colonization by intestinal pathogenic bacteria (Hudault, Guignot, & Servin, 2001), producing vitamin B and K2 for their hosts (Meganathan, 2001). Lactobacilli can protect against enteric pathogens by competing for nutrients and mucosal binding sites, and producing antibacterial substances (i.e., lactic acid, bacteriocin, etc.) (Li et al., 2012;Mann et al., 2014). Liu et al. (2017 reported that Lactobacillus casei can improve porcine intestinal immunologic function and consequently reduce the number of cases of piglet diarrhea and death. After weaning, the proportion of Lactobacillus in the gut microbiome of piglets declines Su, Yao, Perez-Gutierrez, Smidt, & Zhu, 2008). In the present study, Lactobacillus and Escherichia were the core microflora of piglets at 1 week of age, indicating that they play crucial roles in establishing and maintaining the intestinal flora of piglets after birth.
Several strains representing the genus Clostridium are butyrate producers, which convert oligosaccharides (i.e., soybean fibers, fructo-oligosaccharide, and maltose) to organic acids and alcohols, and thus maintain the balance of intestinal microflora (Louis & Flint, 2009).

Additionally, Clostridium can modulate systemic immune responses by
inducing the expansion of regulatory T regs and the production of inflammatory cytokines, and is associated with colitis resistance (Atarashi et al., 2011). Similar to the finding of Gorham, Williams, Gidley, and Mikkelsen (2016), we found that Clostridium was a core bacterial genus in piglets after supplementary creep-feeding (3 weeks), suggest that an increase of Clostridium is conducive for the digestion of cellulose and hemicellulose substrates in piglets. Alternatively, certain dietary constituents provide main energy source or affect gastrointestinal acidity for the growth of gut microbiota, and it is likely a more pH neutral digesta is more conducive for Clostridium to thrive in the intestine.
Although Slifierz et al. (2015) found that Megasphaera and Lactobacillus were the predominant genera in postweaned healthy piglets, there is evidence that Prevotella spp. dominate swine fecal microbial communities and gradually increase in numbers with age (Lamendella, Domingo, Ghosh, Martinson, & Oerther, 2011;Pajarillo, Chae, Balolong, Bum Kim, et al., 2014). Prevotella is a key microbial member of the gastrointestinal tracts of adult animals; it is crucial for the degradation of starch and plant polysaccharides but also has a strong capacity for mucoprotein catabolism (Ivarsson, Roos, Liu, & Lindberg, 2014;Rho et al., 2005). Frese et al. (2015) revealed that the relative abundance of Prevotellaceae increased nearly 50-fold from an average of 0.3% in suckling piglets to 14.8% in weaned piglets. We found a low relative abundance of Prevotella in preweaned healthy piglets and, typically, a much higher abundance in postweaned healthy piglets. This difference is believed to be linked to a diet containing less-digestible solid feed after weaning, suggesting F I G U R E 6 Fecal bacterial concentrations of Lactobacillus, Escherichia coli, and Prevotella quantified by qPCR. (a) Differences in relative abundance between preweaning and postweaning healthy piglets. (b) Differences in relative abundance between healthy and diarrheic piglets before weaning. Significance is considered at p < .05 that the increase in Prevotella is likely conducive to digestion and the utilization of feed nutrients in weaned piglets.

| Differences in fecal microbiota between diarrheic and healthy piglets
Abundant evidence has revealed an intimate biological interaction between host animals and their gut microbiota. Pop et al. (2014) showed that fecal microbial diversity gradually increases with age and that children with diarrhea have lower microbial diversity than healthy children. In this study, analysis of microbial diversity revealed a higher diversity index in preweaned diarrheic piglets (D_ week1 and D_week3) than in preweaned healthy piglets (H_week1 and H_week3). However, the diversity index in postweaned diarrheic piglets (D_week5) was lower than that in postweaned healthy piglets (H_week5). These results suggest that early supplementary TA B L E 3 Spearman's correlation analysis among maker bacterial genera in diarrheic and healthy piglets at 3 weeks of age creep-feeding has little effect on microbial diversity and composition before weaning, while the imbalanced microbial diversity observed in diarrheic piglets is correlated with particular changes in their solid feed after weaning.
Earlier findings underscore the importance of microbial composition variation during the development and progression of intestinal diseases (Kamada, Seo, Chen, & Núñez, 2013). In this study, differences were also observed in bacterial species and their relative abundance between diarrheic and healthy piglets at three age-related dietary stages, especially in those receiving early supplementary creep-feed (H_week3 vs. D_week3). Furthermore, most of these major differences in bacteria coincided with those identified in our previous study in healthy and diarrheic piglets receiving full sows' milk (H_week1 vs. D_week1) (Yang et al., 2017). Diarrheic piglets showed decreased relative abundance of Firmicutes, the major genera being Ruminococcus/[Ruminococcus], Anaerovibrio, and Clostridium, which is consistent with previous research on neonatal healthy and diarrheic piglets (Hermann-Bank et al., 2015). Several members of the phylum Firmicutes are believed to produce short-chain fatty acids (SCFAs), related to energy acquisition and immune regulation (Atarashi et al., 2011). For instance, several members of Clostridiales are metabolized to produce butyrate (Devillard, Mcintosh, Duncan, & Wallace, 2007;Louis & Flint, 2009). Reduced abundance of these main Firmicutes types in the fecal microbiome of diarrheic piglets may concomitantly affect their digestive physiology and immune functions.
The phylum Proteobacteria including members of the genera Sutterella and Campylobacter is reportedly associated with intestinal inflammatory disease (Minamoto et al., 2014). In dogs with acute diarrhea, the relative abundances of Sutterella and Clostridium perfringens are decreased while those of Blautia, Faecalibacterium, and Turicibacter are increased when compared with levels in healthy dogs (Suchodolski et al., 2012). These findings concur with our results that Sutterella and Campylobacter showed increased relative abundance in preweaned diarrheic piglets, suggesting these two genera may be considered potential pathogens in the etiology of piglet diarrhea.
Research has shown that the co-occurrence of E. coli and Enterococcus increases the risk of neonatal piglet diarrhea (Hermann-Bank et al., 2015;Jonach et al., 2014 (Vlasblom et al., 2015). For instance, the probiotic E. coli Nissle 1917 has been demonstrated to limit the expansion of competitors such F I G U R E 7 PICRUST analysis of KEGG metabolic pathways at the second level. Graphs show the abundance ratios of different functions between diarrheic and healthy piglets at 3 weeks old (a) and 5 weeks old (b). p < .05 represents a significant difference as adherent-invasive E. coli and the related pathogen Salmonella enterica in the inflamed intestine and prevent infectious diarrhea (Sassone-Corsi et al., 2016). Based on these findings, it is understandable that the E. coli identified in the preweaned piglets of the present study is a nonpathogenic microorganism that confers benefits to the host.
Although Prevotella has important roles in nutrient digestion and absorption (Ivarsson et al., 2014;Sandberg, Kovatcheva-Datchary, Björck, Bäckhed, & Nilsson, 2018), its pro-inflammatory effect of Prevotella has been gaining increasing attention as it is correlated with enhanced susceptibility to arthritis (Hofer, 2013). Increases in the abundance of Prevotella copri correlate with reduction in the abundance of several beneficial microbes, such as Bacteroides, Clostridium cluster XIV, Blautia, and Lachnospiraceae, resulting in reduced resistance to colitis (Atarashi et al., 2011;Scher et al., 2013).

The disruption of competitive relationships between Prevotella
and Escherichia was found to be associated with neonatal piglet diarrhea (Yang et al., 2017). These observations highlight attenuated or altered bacterial inter-relationships in individuals with diarrheal disease; thus, synergistic or competitive relationships of intestinal bacteria are closely related to microbial diversity and health status in the intestinal microenvironment. Indeed, our evaluation of correlations among marker genera associated with piglet diarrhea during early supplementary creep-feeding showed consistent results with the study of neonatal piglets receiving sows' milk (Yang et al., 2017).
Therefore, we consider that Prevotella confers a high risk for preweaning piglet diarrhea, resulting in decreases in E. coli and beneficial Firmicutes, which maintain both a balance in the gastrointestinal microecological environment, and the normal function of the mucosal barrier and immunity.
The current consensus is that Fusobacterium is a common pathogen of the gastrointestinal mucosa, the high abundance of which is associated with increased expression of inflammatory factors (Nosho et al., 2016). The relative abundance of Fusobacteriaceae was higher in animals with neonatal porcine diarrhea or horse colitis than in healthy subjects (Costa et al., 2012;Hermann-Bank et al., 2015). It is noteworthy that Fusobacterium is a small group with low abundance in the intestinal tract of humans and animals, and this is possibly the reason it has not been specifically reported in these studies possibly. Uniformly, we found that the proportion of members of the Fusobacteriaceae family in the fecal microbiome of diarrheic piglets was higher than that in both neonatal healthy piglets (Yang et al., 2017)  of the fecal microbiome of 6-month-old healthy pigs (Lamendella et al., 2011). Taken together, a large discrepancy was noted in the fecal microflora associated with diarrhea between preweaned (sow-reared and early supplementary creep-fed) and postweaned piglets. As the diet of piglets is completely transformed to solid foodstuffs after weaning, a reduction in the numbers of beneficial bacteria reduces the ability of intestinal epithelial cells to digest and absorb proteins and carbohydrates, which in turn leads to permeability disorders and diarrhea.
The PICRUSt analysis of the fecal microbial function between diarrheic and healthy piglets gave similar results to the aforementioned analyses. Diarrhea caused by an intestinal microbiota disorder significantly changed the specific metabolic pathways present in preweaned piglets, which is consistent with the results of another study on diarrheic captive musk deer .
Compared with healthy piglets, the most obvious change in diarrheic piglets was the reduced abundance of functional genes involved in carbohydrate metabolism at all three age-related dietary stages. Many of the bacterial taxa that showed reduced numbers in piglets with diarrhea are associated with the production of SCFAs from carbohydrates, which has also been observed in humans with Crohn's disease (Erickson et al., 2012), suggesting that carbohydrate dysmetabolism may be an important feature of piglet diarrheal disease. It may be as a consequence of increased levels of other important functions like energy metabolism, the metabolism of vitamins, cofactors, or terpenoids and polyketides, or glycan biosynthesis in preweaned diarrheic piglets.

| CON CLUS IONS
Our study identified additional characteristics of the gut microbiota in diarrheic piglets during age-related dietary changes and the development of intestinal physiology. Postweaning feed is a primary factor influencing the dynamic colonization by microflora in healthy piglets. Obvious changes include a continuous decrease of Lactobacillus and Escherichia, as well as a gradual increase of Prevotella during the transition to a solid diet. Piglets receiving sows' milk or early supplementary creep-feed had identical major microbial communities that were associated with diarrhea. An altered relationship between Prevotella and Escherichia may be the main cause of diarrhea. Reduced numbers of Bacteroides, Ruminococcus, Bulleidia, and Treponema may be related to the onset of postweaning piglet diarrhea. These findings provide insights for the development of intervention strategies for better management of diarrhea in piglets. However, further studies should be considered that are focused on the reciprocal interactions between gut microbiota and host genetics and/or metabolic substances or the intestinal immune system in response to microbial alterations.

ACK N OWLED G M ENTS
This work was supported by the special funds for the discipline construction project of Gansu Agricultural University (Grant No. GAU-XKJS-2018-042).

CO N FLI C T O F I NTE R E S T S
None declared. Zhao.

E TH I C S S TATEM ENT
All animal experimental protocols were conducted according to the

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated and/or analyzed during the current study are available in the NCBI sequence read archive (SRA) database at https ://