Prebiotic effects of alfalfa (Medicago sativa) fiber on cecal bacterial composition, short-chain fatty acids, and diarrhea incidence in weaning piglets

Dietary alfalfa fiber (AF) is conceived to modulate gut microbial richness and diversity to improve the health and growth of weaning piglets. The objective of this study was to evaluate the prebiotic effects of AF on diarrhea incidence, the production of short-chain fatty acids (SCFAs), and microbiota composition in weaning piglets. This study utilized 100 crossbred piglets (Duroc × Landrace × Yorkshire) with a body weight of 8.42 ± 1.88 kg randomly assigned to the following treatments: 0.00% AF meal (A), 6.00% of AF meal (B), 12.00% AF meal (C), and 18.00% AF meal (D). The cecum samples were used to determine microbial community composition and diversity through high-throughput 16S rDNA sequencing. The results of this study show that the lowest average daily gain (ADG) was observed in treatment D, and the highest ADG was recorded in treatment C. However there was no significant difference between the treatment groups and the control. The average daily feed intake (ADFI) was significantly higher in treatment C compared to the other treatments. The feed conversion ratio was high in the control group compared to the AF treated groups. The highest diarrhea incidence was observed in treatment A and the lowest diarrhea incidence was observed in treatment C and D. The highest acetate and propionate levels were observed in treatment B, but there was no significant difference between the treatment groups and the control. The supplementation of AF significantly increased the butyrate level in treatment D compared with treatments A and B but was not significantly different from treatment C. The Observed_species richness and Simpson diversity values of the cecum bacterial composition in the AF fed piglets were higher than the control. In addition, the Chao 1 richness and Shannon diversity increased with an increase in AF supplementation, reaching a plateau at treatment B and C, then decreasing at treatment D. The Bacteroidetes, Firmicutes, Tenericutes, Proteobacteria, Cyanobacteria, Spirochaetae, Actinobacteria, Fibrobacteres, Saccharibacteria, Synergistetes, Chlamydiae, Elusimicrobia, Deferribacteres, Fusobacteria, and others were relatively abundant in all treatments. The Bacteroidetes and Firmicutes were the dominant phyla, accounting for 98% of all reads. AF treatment decreased the Bacteroidetes phylum and increased the Firmicutes phylum compared with treatment A. Therefore, the dietary inclusion of AF may decrease diarrhea incidence, increase cecal bacterial composition and richness, and consequently improve the growth performance of weaning piglets.


Introduction
The gut microbiota contains over 10 14 microbial cells that collectively comprise at least 150 times more genes than their host. 1,2 The composition of the gut microbiota in humans and other monogastric animals varies based on age, diet, genetic characteristics, and sex. 1 Gut microbiota have some important functions in host health, nutrient metabolism and absorption, development of the host immune system, differentiation of the gut epithelium, and maintenance of the gut mucosal barrier functions. 3,4 Previous evidence indicated that gut microorganisms can directly inuence physiological conditions in animals such as improving the roles of the intestinal barrier, stimulating defence mechanisms against pathogens, improving the host immune system, increasing defensive mechanisms against inammatory bowel diseases (IBD), 5 regulating autoimmunity, producing biological metabolites, preventing obesity and regulating diabetes, and destroying cancer cells. 1,6 The interactions between the host gut-microbiota are dynamic and highly vulnerable to several environmental conditions, particularly diet. 1 Alfalfa is a high-yielding perennial legume that is cultivated worldwide with rich nutritional characteristics and bioactive compounds. 7,8 Alfalfa ber is mainly composed of insoluble dietary ber, such as cellulose, lignin, and xylans, representing more than 90% of the total dietary ber composition of alfalfa. 9,10 Interestingly, alfalfa meal contains 17.5% CP, 24.1% CF, and 1200 kcal kg À1 metabolizable energy. 11 Recent studies reported that pigs grown on alfalfa pastures produce goodquality natural pork, which is currently gaining more attention in Europe and North America. 12,13 Interestingly, for the production of good quality pork, alfalfa is used as fresh fodder in some traditional style pig production systems in China 13 and other parts of the world. AF is a high-quality protein feed; therefore, increasing its supplementation may improve the nutritional requirement of pigs and increase the growth of piglets. Few studies reported that the supplementation of natural compounds such as avonoids extracted from alfalfa had no estrogenic impacts, which may increase the production performance, enhance antioxidants, induce anti-stress function, and eradicate free radicals from animals. 12,14 Dietary ber inuence the role of the gut microbiota, as indicated by the low digestibility and availability of nutrients in feeds with high ber levels. 1,15 The type of dietary ber, soluble or insoluble, inuences its effects on gut microbial composition and function. 1,16,17 The intake of high dietary ber has been reported to inuence energy metabolism, nutrient utilization, and subsequently results in a reduction in the performance of monogastric animals, especially pigs. [18][19][20] Recent ndings indicated that dietary ber can prevent intestinal diseases such as constipation, diarrhea, bowel diseases, and improve intestinal health in both humans and animal subjects. 10,21,22 Different dietary ber has benecial roles in gut microbial composition and function. Such benecial roles include SCFAs synthesis, decreased intestinal pH, and fecal bulkiness. 1,21,23 It was reported that dietary ber may selectively modulate intestinal microbiota via the stimulation of constructive microbial species and the suppression of enteric bacterial species. 10,24 Dietary ber functions as a nutrient for intestinal microbiota and consequently motivates the growth of valuable intestinal bacteria, like Bidobacterium, which are involved in the reduction in cardiovascular disorders. 25 Weaning piglets are associated with a reduction in feed consumption and low nutrient metabolism due to the immature digestive systems, stress as a result of the separation from the sows and litter mates, and the sudden change from the easily digestible sow's milk to less digestible solid feed. 12,16,17,26 The treatment of high-ber diets to piglets results in an overall increase in the total empty-weight of the gut 27 and an increase in gut secretions. 28 Dietary ber increased the growth performance, nutrient digestibility, and immunity in piglets, 29,30 resulting in a decrease in diarrhea incidence. 31,32 There was a report that dietary ber supplementation may improve growth performance and decrease the intestinal permeability in piglets, 31,33 indicated that dietary ber extracted from Astragalus membranaceus increased growth performance, nutrient digestibility, VFA production, and immunity in piglets. However, the effects of AF on gut microbiota composition, health, and energy metabolism in piglets are limited.
Therefore, the objective of this study was to evaluate the prebiotic effects of AF on diarrhea incidence, SCFA production, and microbiota composition in weaning piglets. This objective was investigated by feeding different concentrations of AF to 28 day-old piglets for 48 days and the cecal samples were extracted for microbial analysis. The 16S rDNA pyrosequencing technique was used to examine the prebiotic effects of AF on the cecal microbial composition, and the growth parameters of the piglets were examined.

Animal care and ethics
All animal procedures were carried out in accordance with the Guidelines for the Care and Use of Laboratory Animals of Jilin Agricultural University, and experiments were approved by the Animal Ethics Committee of Jilin Agricultural University.

Experimental animals, diet, and sampling
A total of 100 crossbred piglets (Duroc Â Landrace Â Yorkshire) weaned at 28 days of age with a mean body weight of 8.42 AE 1.88 kg were randomly allocated to 4 treatment groups with 5 replicates per treatment and 5 pigs per replicate using the randomized complete block design. The piglets were housed in pens (1.8 Â 1.2 m) with a concrete slatted oor and ad libitum access to feed and water. The room temperature was maintained at 26 C at the time of weaning and progressively reduced to 22 C within the rst week of post weaning. The humidity was kept constant at 65-75%. The piglets in each group received one of the 4 experimental diets (Table 1) with cracked corn and soybean meal as the main source of energy and protein, respectively. The diets were formulated as follows: 34 the control diet of 0.00% AF (A), 6.00% AF diet (B), 12.00% AF diet (C), and 18.00% AF diet (D). All nutritional components in Table 1 were supplement based on the national nutrient requirements of piglets with the exceptions of AF supplementation. Therefore, the obtained results were analysed based on the added AF levels. The experiment lasted for 48 days and 7 days of prefeeding tests. There was no antibiotic supplementation before and during the experimental trial. Piglets were weighed individually at the beginning and the end of the experiment and daily feed intake per pen was recorded. At the end of the feeding trial, 10 pigs were randomly selected from each group. Aer fasting overnight, the pigs were slaughtered, and the gastrointestinal tract was immediately removed. The cecal content was collected into 5 mL cryotubes and snap-frozen immediately in liquid nitrogen and stored at À80 C until further microbial and metagenomics analysis. In addition, 50 mL centrifuge tube of cecum digest samples were collected and frozen in À20 C for SCFAs analysis.

Growth performance, diarrhea incidence, and SCFAs
The initial body weight of the piglets was subtracted from the nal body weight and divided by the days of the experiment to obtain the ADG. The ADFI was calculated by estimating the total feed intake per day. The FCR was calculated by dividing the feed intake by the body weight gain. All piglets were checked for the consistency in fecal samples as an indicator of their health status. The fecal samples were recorded as normal, pasty, or watery. 35 The diarrhea incidence (%) was calculated using the method. 36 The acetate, butyrate, propionate, and the total SCFAs were analyzed according to the method described. 37

Total DNA extraction
Total DNA was extracted and quantied from 0.5 g of each cecal samples using a Qubit 2.0 Fluorometer (Thermo Fisher Scien-tic Inc. Massachusetts, USA) following the manufacturer's instructions. The amount of total DNA was estimated by the NanoPhotometer spectrophotometer (Thermo Fisher Scientic Inc. Massachusetts, USA), and the molecular size and DNA quality were measured using 1% agarose gel electrophoresis. Cecal sample preparation was performed as described. 31 All DNA samples were stored at À20 C until further analysis.

16S rDNA gene amplication and sequencing
Sequencing was performed at the Annotech Genentech (Beijing Co., Ltd). Briey, 10 ng of DNA template was used to amplify the target region, according to the different sequencing regions, the corresponding primers: 341F (5 0 -CCTA-CACGACGCTCTTCCGATCTN-3 0 ) and 805R (5 0 -GACTG-GAGTTCCTTGGCACCCGAGAATTCCA-3 0 ) using TaKaRa's Ex Taq enzyme to ensure amplication efficiency and accuracy. The preliminary quantication was performed using Qubit 2.0. Then, the insert size of the library was detected using Agilent 2100. The Bio-Rad CFX 96 real-time PCR instrument was used to perform the qPCR to accurately quantify the effective concentration of the library to ensure library quality. The PCR was performed as described. 31 Then, the mixture of the PCR product was puried with a GeneJET Gel Extraction Kit (Thermo Fisher Scientic). The sequencing libraries were generated using a NEB Next Ultra DNA Library Prep Kit for Illumina (NEB, United States) following the manufacturer's instructions. The library quality was assessed on the Qubit@2.0 Fluorometer (Life Technologies, CA, United States) and the Agilent Bioanalyzer 2100 system. Finally, the library was sequenced on an Illumina HiSeq platform, and paired-end reads were generated.

Sequence analysis
For the sequences obtained by the sequencing, data ltering was completed by removing the low-quality base, Ns and joint contamination sequences to obtain a reliable target sequence for subsequent analysis. The ltered sequences were referred to as the Clean Reads. First, the corresponding Read1 and Read2 of the sequence fragments were obtained from the 5 0 and 3 0 ends, respectively using the sequence stitching method PEAR; 38 the spliced sequence was analyzed using the QIIME soware version 1.80, [39][40][41] including the extraction of operational taxonomic units (OTUs), overlap analysis, cluster analysis, Lefse analysis, phylogenetic tree construction, alpha diversity analysis, beta diversity analysis, etc. 42 The operational taxonomic units were clustered with a 97% similarity threshold. Species identication was based on the comparison of the corresponding recognition sequences of each species in the database with the OTUs. Operational taxonomic units with fewer than 2 sequences in all samples were eliminated to rule out the impact of less reliable OTUs on subsequent analysis. Change in OTUs between samples was counted and represented in heat maps. The number of unique OTUs for each group was represented on a Venn diagram. The distribution of species in each sample at each classication level was represented in histograms. According to the species annotation and abundance information of all the samples at the class, family, and genus levels, the top 25 subjects were selected based on the families, genera, and their abundance information in each sample to draw sample heat maps, and classify them according to the classication standards. Sample clustering was perform using the information available at each sample level for the easy display of results and the discovery of information to identify more species or samples that were concentrated in the sample study for the obtained species distribution data of each sample aer normalization. The Bray-Curtis distance value was calculated, and each sample was clustered by a hierarchical clustering algorithm. The cluster analyzed the similarities between samples.
According to the species content of the samples, the distribution results at the ve taxonomic levels comprised of the phylum, class, order, family, and genus were calculated. The information available at the phylum, class, order, family, and genus levels were used to draw species distribution histograms. For the identication of the dominant species in each sample, the dominance index for each species in each sample was calculated. For species similarities and diversity, the alpha diversity analysis using the Shannon diversity index, Simpson diversity, Chao 1 richness, and observed species richness were used. The principal component analysis (PCA) was performed at the genus level, 43 and the results were visualized using the Statistical Analysis of Metagenomics Proles (STAMP) version 2.13 program. 44 The linear discriminant analysis coupled with effect size (LEfSe) was performed to identify the bacterial taxa differentially represented between groups at the genus or higher taxonomy levels. 45 LefSe analysis combines linear discriminant analysis with the KW rank sum test and Wilcox test to determine the effect of different microbial species in different treatment groups. Finally, the Metastats soware and T-test soware were used to analyzed the species abundance data between groups.

Data analysis
Data were analyzed using a one-way analysis of variance (ANOVA) procedure of SPSS soware version 20.0 (SPSS Inc., Chicago, IL, USA). A probability value of p < 0.05 was considered statistically signicant and differences between means were noticeable. The Duncan Multiple Range Test (DMRT) was employed to ascertain the differences between the treatments.

Growth performance and piglet diarrhea
The growth performance of piglets increased with the increase in AF supplementation, but at higher AF concentrations (D), there was a decrease in the growth performance of piglets. There was no signicant difference (p > 0.05) in the ADG among all the treated groups. The ADFI of treatment C was signicantly different (p < 0.05) from treatment A and treatment D but not signicantly different (p > 0.05) from treatment B. Treatment D recorded the highest F : G with no signicant difference (p > 0.05) between the treatments. The diarrhea incidence in treatment A was signicantly higher (p < 0.05) than the other treatment groups. Meanwhile, the lowest diarrhea incidence was recorded in treatment C and treatment D ( Table 2).

Cecal short-chain fatty acids production in pigs
As shown in Table 3, the supplementation of AF increased the acetate and propionate levels. The highest acetate and propionate levels were observed in treatment B, but there was no signicant difference (p > 0.05) between the treatment groups. The butyrate concentration increased with the increase in AF concentration. There was a signicant increase (p < 0.05) in butyrate production in treatments C and D in comparison with treatments A and B. The highest total SCFAs was observed in treatment B, but there was no signicant difference (p > 0.05) between the groups.

Sequence description analysis
The high throughput pyrosequencing of the cecal samples produced 65 590 overlapping pair-end reads from the 16 samples; 61 618 clean reads were obtained representing 93.94% of reads to be subjected to further analysis. The clean reads were combined to tag bases. About 30 809 total reads were obtained with a mean of 460.32 tags per sample, and all reads were assembled. 27 867 normalized tags of the OTUs were obtained with total aligned tags of 25 094. The aligned rate per sample was 90.05%, and a total number of OTUs of 210 was obtained by clustering the 30 809 tags at 97% similarity ( Table 4). The overlap of OTUs between groups was represented in a Venn diagram (Fig. 1). The Venn diagram estimates the overlapping OTUs composition and species overlap between the treatment groups. Fig. 1 showed that 334 OTUs were shared among all sample groups. Higher species specicity was observed in treatment B with 10 OTUs, while treatment A, C, and D have OTUs specicity of 4 each. The shared OTUs between the groups were 37 and between treatments B, C, and D; 7 OTUs were shared between treatments A, B, and D; 10 OTUs were shared between treatments A, B, and C. 18 OTUs were shared between A, C, and D. The shared OTUs between two sample treatments were 18 between treatment C and D, 1 OTU between treatment A and C, and 1 OTU between treatment A and D, 3 OTUs between B and C, and 8 OTUs between B and D (Fig. 1).

Gut microbiota richness and diversity
There was no signicant difference (p > 0.05) in species richness, as represented by the Chao 1 and Observed_species values in the treatments (Table 5 and Fig. 2). The species richness increased with the increase in the supplementation of AF, reaching a plateau and then decreasing as the AF content increased (treatment D). The highest Chao 1 and Observ-ed_species values were recorded in treatment B and the lowest in the control group (A). There was no signicant difference (p > 0.05) in species diversity in all treatment groups. The highest Shannon diversity index value was found in treatment B, and the lowest Shannon diversity index was observed in treatment D. The Simpson diversity value increased with the increasing AF content. The highest Simpson diversity index was observed in treatment D and the lowest in treatment A. The PCA and nonmetric multi-dimensional scale method (NMDS) reects the differences between samples in the two-dimensional coordinate map (Fig. 3B). The more similar the composition of the microora between the samples, the closer the distance representing their coordinate points in the PCA plot. The contribution rates of PC1 and PC2 in Fig. 3A were 45.73% and 19.71%, respectively, with a cumulative contribution rate of 65.44%. On the PC2 axis, treatments A and B were distributed towards the positive direction, while treatments C and D were mainly distributed towards the negative direction. The supplementation of AF increased the microbial community distribution in each sample, but the microbial composition between the treatments was similar. The closer the PCA plot in the distribution principal components in Fig. 3A, the more similar the community composition of the samples.

Cecal microbiota composition
The effects of AF on piglets cecal microbial composition were observed at different taxonomic levels. With regard to the phyla, a total of 15 phyla were identied in each sample. There was no signicant difference (p > 0.05) in the phylum microbial composition between the treatment groups and the control.
Bacteroidetes and Firmicutes were the main dominant phyla, accounting for about 98% of all reads (Fig. 4A). The Bacteroidetes phylum was signicantly higher (p < 0.05) in all treatment groups compared with the other bacterial phyla. The Bacteroidetes was the most dominant phylum; the relative abundance of the Bacteroidetes decreased slightly upon the addition of AF and then increased at a steady rate as the AF concentration increased. The Firmicutes concentration increased with the increasing AF concentration; at the highest AF level, there was a decreasing trend observed (Fig. 4A). At the class level, a total of 23 bacterial classes were observed across all treatment groups (Fig. 4B). Our results indicated that Bacteroidia, Bacilli, Clostridia, and Negativicutes were the dominant bacteria classes. The relative abundance of Bacteroidia was signicantly different (p < 0.05) in all treated groups compared to the other bacteria classes. The Bacilli signicantly increased (p < 0.05) in treatment C compared to the other treated groups. The Negativicutes class was not signicantly affected by the addition of AF in the diet. According to the increasing order of the dominant bacteria classes, Bacteroidia > Bacilli > Clostridia > Negativicutes, accounting for 95% of the total reads (Fig. 4B).
A total of 26 bacteria orders were identied in each treatment (Fig. 4C). The cecal microbial composition was dominated by the following orders: Bacteroidales (65%), Lactobacillales (15%), Clostridiales (10%), and Selenomonadales (5%) (Fig. 4C). The Bacteroidales was signicantly different (p < 0.05) in all the treatment. The Bacteroidales decreased as the concentration of AF increased. The Clostridiales increased with an increase in AF levels. The Selenomonadales order increased with the increase in AF supplementation. The Lactobacillales levels maintained a steady state level with the increase in AF; at the highest AF level (D), there was a reduction in the relative abundance of Lactobacillales.
At the family level, 26 families were characterized within each treatment group (Fig. 4D). Bacteroidaceae, Lactobacillaceae, Streptococcaceae, and Veillonellaceae were the dominant bacteria  Table 3 represents the data sequencing analysis involving data ltering statistics, sequencing splicing, and OTUs statistics.
families. The supplementation of AF had no signicant effect on the Bacteroidaceae and Streptococcaceae families in each treatment group. There was a slight increase in the relative abundance of Lactobacillaceae and Veillonellaceae as the AF content increased (Fig. 4D). At the genera level, 26 genera were characterized (Fig. 4E). Prevotella 2, Prevotella 9, uncultured, and Streptococcus were the dominant species in all treatments (Fig. 4E). The relative abundance of Prevotella 9 was signicantly dominant (p < 0.05) in all treatment groups compared with the other bacterial genera. The supplementation of AF in the diet of piglets slightly increased the relative abundance of Prevotella 2 and Prevotella 9 at higher concentrations of AF (D). AF treatment has no effect on the Uncultured and the Streptococcus bacterial populations. Fig. 5A showed that the bacterial members with signicantly high abundance mainly belonged to the Bacteroidia, Clostridia, and Negativicutes classes. The bacterial taxa with signicantly high population abundance were dominated by the Bacteroidales, Clostridiales, and Selenomonadales orders in the treatment groups. The bacterial taxa with signicantly greater population were dominated by the Prevotellaceae family (Fig. 5A). From Fig. 5B, the R-value was À0.083 and the p-value was 0.662, implying that the R-value was closer to 0 and the p-value was greater than 0.05. This means that there was no signicant difference (p > 0.05) between and within the treatment groups.

Cecal microbial network
From Fig. 5C, 11 bacterial taxa were shown to be specic to treatment D and, hence, inuence the community structure. A total of 10 bacterial taxa were specic to treatment C; these bacterial taxa were specic and inuenced the community structural composition in treatment C. 2 bacterial taxa were specic and inuenced the community structural composition in treatment B. In treatment A, 5 bacterial taxa were specic and inuenced the bacterial composition. Fig. 5D showed that the bacterial members with signicantly high abundance mainly belonged to the f_Bacteroidales RF16 group and f_Helicobacteraceae families in treatment D. In addition, the bacterial taxa with signicantly greater population were dominated by the f_family, f_family III, and O_campylobacterales order in treatment A. In treatment B, the relative abundance of the bacterial taxa with signicantly greater population were the O_campylobacterales and C_Epsilonproteobacteria classes. There were three families with signicantly dominant bacteria population, f_Clostridiaceae, f_Clostridiaceae 1, and f_Peptostreptococceae, in treatment group C.

Discussion
Early weaning is a normal husbandry practice in most global swine production. 46 Weaning is a harsh condition that is mostly associated with growth disorders and health effects in piglets. Piglets oen exhibit a reduction in feed consumption and low nutrient metabolism due to their immature digestive systems, stress as a result of the separation from the dam and littermates, and the sudden change from the easily digestible sow's milk to a less digestible solid feed. 26 These piglets have limited susceptibility to diseases and infections from enteric bacteria in addition to frequent diarrhea, stunted growth, low nutrient utilization efficiency, and even mortality. 31,47 Due to the reduction in small intestinal composition, limited immunological response, and other protections derived from the maternal milk, piglets are more susceptible to diseases and infections by enteric bacteria like E. coli and Salmonella. 46,47 There have been reports that the adaptation of piglets to a weaning diet takes more than 14 days. 48 The composition and functions of the gut microbiota can be inuenced by dietary nutrient Fig. 1 Venn diagram showing the overlap between groups. The Venn diagram represents the OTUs similarities and dissimilarities shared between samples. The letters A, B, C, and D represent the treatment groups, and the numbers represent the OTUs.  Table 4 represents the effects of AF on microbial richness and diversity. The microbial richness was represented with the Chao 1 index and the Observed_species index, and the microbial diversity was represented with the Shannon and Simpson diversity indices. The letters A, B, C and D represent the treatment groups. There was no signicant difference (p > 0.05) in the microbial richness and diversity.
supplementation. 49 Therefore, the objective of this study was to investigate the prebiotic effects of AF supplementation to the diet of early weaning pigs on cecum microbial composition and function, piglet diarrhea, and SCFAs production. In global pig farming, several dietary-ber-based feeds have been used to increase the nutrient digestibility and growth performance of pigs 50 and improve their gut microbial composition. However, the effects of dietary ber supplementation on growth performance in pigs are inconsistent. 36 Many studies have indicated that high dietary ber supplementation increased growth, 51-53 while others observed no impact on growth performance. 29,30,54 The changes in the gut milieus and nutrient digestion can improve the gut microbial composition, production performance, and health of pigs. Previous studies in our group 28,46 reported that Astragalus membranaceus ber supplemented in a corn-soybean-based diet signicantly increased the growth performance of piglets. Similarly, the ndings of this study showed that the supplementation of AF increased the growth performance of piglets, but at higher AF concentrations, there was a decrease in growth performance. This result implies that increasing the AF content of the diet may improve the gut milieus, increase growth, and consequently improve the gut microbial composition of piglets, but at higher AF levels, there may be a detrimental effect on the growth performance of piglets. 26 The ndings of this current study were consistent with our previous investigation 31,46 and that of other reports. 51,55 The reason for the observed decreased in growth performance of piglets at high AF levels may be that piglets cannot metabolize high-ber diets due to their immature digestive tracts and the low microbial composition. 26 In contrast to our ndings, 56,57 no effects were reported on the ADFI, ADG, or FCR aer supplementing dietary corn bran to the diet of weaned piglets. Our results further elucidated that the supplementation of AF to the diet of piglets signicantly decreased diarrhea incidence in comparison with the control. This observation was consistent with the results. 3,16,31,32,58 This observation may be due to the rich bioactive compound composition of AF, which may exert anti-inammatory and anti-microbial effects on the gut of weaning piglets thereby reducing the population of enteric bacteria in the gut. Therefore, decreasing the diarrhea incidence in pigs. We further evaluated the effects of AF on SCFAs production in pigs. Short-chain fatty acids, also known as volatile fatty acids (VFA), mainly include acetic acid, propionic acid, and butyric acid, which are involved in organ metabolic processes in the body and exert their respective functions. The synthesized acetic acid enters the blood and participates in the metabolic processes of the liver, heart, spleen, and the brain. 59 Propionic acid and butyric acid not only provide energy for liver This journal is © The Royal Society of Chemistry 2019 metabolism but also inhibit cholesterol synthesis; these SCFAs regulate the expression of genes and maintain stability within the intestinal tract. From the results of this experiment, the contents of acetate, propionate, and butyrate in the cecum of the AF-fed groups were higher than that of the control group. This study was consistent with the ndings. 15,31 Also, Gidenne et al. found that an increase in ber levels can increase the concentration of VFAs in the cecum and concluded that a high-ber diet can be fermented in the cecum to produce SCFAs to provide energy for the body. 60 However, acetate and propionate content decreased at higher dietary AF levels, indicating that excess ber may decrease the microbial activities in the cecum. The butyrate levels signicantly increased with the increase in dietary AF levels. In contrast to our ndings, Zhang et al.
reported lower ileal VFA concentrations in pigs fed corn-based diets. 61 In addition, Freire et al. reported low VFA production in pigs aer changing the diet from sugar-beet pulp to soybean hulls. 62 Conversely, insoluble dietary ber can decrease the transit time, leading to a reduction in nutrient absorption in the small intestine. 52,56 Short-chain fatty acids such as propionic and butyric acids were reported to activate the GPR43 on enteroendocrine L-cells to increase gut hormonal synthesis, such as PYY and GLP1, which play vital roles in appetite regulation through the gut-brain axis. 56 High ber intake increased the synthesis of PYY and GLP in addition to inhibiting energy consumption and increase satiety. 56,63 Therefore, the increase in the synthesis of appetite-stimulating hormones through the consumption of dietary ber coupled with the nutrient-rich AF diet may be a vital insight into the observed increase in the growth performance and synthesis of energy metabolites. Moreover, the mechanism of action may be due to the positive impact of AF on the cecal microbial composition of piglets.
The gut contains large volumes of microorganisms that play vital roles in both human and animal well-being. High throughput 16S-rDNA sequencing was used in this study to evaluate the abundance and diversity of the cecum microbial composition in piglets. The supplementation of AF regulates the cecum microbial richness and diversity by enhancing the production of SCFAs through the fermentation of bers. The knowledge of the diversity of the gut microbial composition during the production period in animals is very important. 64 The microbial richness and diversity analysis indicates that there was no signicant difference in species richness and diversity between the gut microbiota in all treatment groups. However, AF supplementation increased species richness in treatment B and decreased as the ber content of the diet increased. The Shannon diversity rst increased and later decreases as the AF content of the diet reached higher levels, while the Simpson diversity increased with the increase in AF levels. Our ndings were consistent with the results of Che et al., who observed an increase in species richness and diversity aer incorporating Astragalus membranaceus ber in the diet of piglets. 31 In addition, Ran et al. reported an increase in the species richness and observed OTUs aer supplementing thymol and carvacrol in the diet of hybrid tilapia. 65 In contrast with our study was the ndings of Liu et al., who reported that there were no differences in the species diversity and richness in the colonic microbial composition of pigs. 56 In the cecal gut microbiota of piglets, Bacteroidetes (72%) and Firmicutes (26%) were the dominant bacterial phyla. Consistent with our ndings was the study by Lamendella et al., who reported that Bacteroidetes and Firmicutes were the dominant bacteria in the pig cecum. 66 Similarly, other studies 31,33 reported that the Bacteroidetes and Firmicutes were the dominant bacterial phyla in the cecum of piglets. In addition, recent researchers reported that Firmicutes and Bacteroidetes were the predominant phyla in the gut microbiota of pigs. 33,56,[67][68][69] We observed that the relative abundance of Bacteroidetes was higher than the relative abundance of Firmicutes. Similarly, Klein-Jöbstl et al. reported that Bacteroidetes was the highest relative abundant phylum in the gut of calves. 64 In contrast, Mateos et al. and Kim et al. reported that Firmicutes was the highest relatively abundant phylum in cattle. 70,71 The supplementation of AF in the diet increased upon increasing the relative abundance of Firmicutes, while the Bacteroidetes population decreased upon the addition of AF. The changes in these two microbial strains may affect the metabolism and functions of the gut microbiota. Similarly, Li et al. reported that the addition of essential oils increased the Firmicutes phylum population while decreasing the relative abundance of the Bacteroidetes phylum. 3 In contrast to this observation, Liu. et al. reported a decrease in the Firmicutes phylum population and an increase in the Bacteroidetes phylum aer feeding a corn-based diet to piglets. 56 The reason for the high abundances of Bacteroidetes was reported in previous studies 72, 73 that the Bacteroidetes were the main degraders of complex polysaccharides in the gut of pigs due to their ability to encode genes for glycoside hydrolases such as b-xylosidases, endo-1,4-b-xylanases, a-N-arabinofuranosidases, and polysaccharide lyases compared with the Firmicutes phylum or other bacterial phyla in the gastrointestinal tract. Meanwhile, recent studies have indicated that the Firmicutes to Bacteroidetes ratio is of signicance in human gut microbial composition and functions in regards to energy metabolism and health. 74 For example, Turnbaugh et al. reported that a higher Firmicutes to Bacteroidetes ratio improved the digestibility of nutrients in pigs. 75 Related studies indicated that the Bacteroidetes in the cecum can encode glycosidic hydrolysis and polysaccharide cleavage genes that promote polysaccharide degradation, 76 while Firmicutes play a role in digesting carbohydrates, 72 which can extract more energy through the fermentation of ber to synthesize more SCFAs. The relative abundance of the Firmicutes in groups B and group C was slightly higher than that in the control group. Consistently, this indicating that dietary ber can promote the growth of the Firmicutes phylum.

Conclusions
In summary, piglets were fed with different concentrations of dietary AF for 48 days. The prebiotic effect of dietary AF was investigated on the diarrhea incidence, SCFAs production, and microbiota composition in weaning piglets. Dietary inclusion of AF in the diet of weaning piglets increased their growth performance and decreased diarrhea occurrences. In addition, the added AF content promoted the composition and diversity of the gut microbiota of piglets. However, the functions of the different microbiota inhabiting in the cecum of piglets were not estimated. Therefore, further research is recommended to elucidate the functional roles of all the microbial species involved in growth promotion, SCFAs synthesis, nutrient digestion, immunity, and energy metabolism.

Conflicts of interest
The authors declare no conicts of interest.