Impact of zinc oxide, benzoic acid and probiotics on the performance and cecal microbiota of piglets

Intestinal health remains a key factor in animal production because it is essential for digestion, absorption and bacterial fermentation. Feed additives have been used to attenuate the weaning stress such as Zinc Oxide (ZnO) and benzoic acid (C7H6O2). The objective of this study was to evaluate the impact of of benzoic acid and probiotics (BA + P) on performance, diarrhea and cecal microbiota of piglets in the nursery phase (23 to 65 days). One hundred and sixty weaned piglets with an initial weight of 6.335 ± 0.698 kg and 23 days of age were submitted to four treatments: supplementation with 2500 ppm of Zinc oxide (ZnO), supplementation with a commercial blend of benzoic acid and probiotics (Bacillus licheniformis, Bacillus subtilis and Enterococcus faecium NCIMB 10415; Vevogut P®) (BA + P), supplementation with Zinc oxide plus benzoic acid and probiotics (ZnO + BA + P), and controls receiving only the basal diet without any supplementation. At 65 days of age, 32 piglets (n = 8 per treatment) were slaughtered for the evaluation of the cecal microbiota. Supplementation with ZnO and BA + P were associated with better feed conversion (P < 0.05) in the early stage (23 to 49 days) and with an improvement in all performance parameters over the entire experimental period. The occurrence of diarrhea was lower (P < 0.05) in the BA + P group. The 4 most abundant phyla along with unclassified bacteria represented 93% of all sequences. Firmicutes dominated the cecal microbiota of all groups, followed by Bacteroidetes. Richness represented by the observed number of genera and by the Chao index were statistically lower in ZnO and ZnO + BA + P supplemented animals compared to controls. The beta diversity analysis that compares similarities between bacterial communities demonstrated formation of two distinct clusters containing samples with and without supplementation with ZnO, confirming a strong influence of ZnO on the intestinal microbiota. The use of Benzoic acid with probiotics yields similar performance results with lower impact on the gut microbiota compared to ZnO, and it should be considered as a potential alternative in swine production.

production. During the first weeks of life, the gastro intestinal tract passes through several adaptation processes that may have an impact on animal performance [1][2][3] Weaning is the most critical phase in swine production and post-weaning diarrhea remains the most common cause of morbidity and mortality in swine herds resulting in great economical losses [4][5][6].
Several strategies to attenuate the negative impact of the stress observed during weaning have been evaluated [7,8]. Feed additives, such as Zinc Oxide (ZnO) [9] and benzoic acid (C 7 H 6 O 2 ) [7] are globally used in swine production systems. The use of those products is associated with lower incidence of diarrhea and greater weight gain during weaning [10,11]. Zinc is essential for the intestinal mucosa integrity and protects against pathogenic organisms by modulation of the immune system [12], but the excessive use of ZnO during the last decade as an alternative for antibiotic growth promoters has been associated with the development of bacterial resistance to this product [13] and with environmental pollution [14]. Therefore, alternatives to this product such as essential oils and probiotics deserve further investigation.
The importance of the intestinal microbiota, especially of bacteria, has been revised after the development of next generation DNA sequencing (NGS) technologies, which are much more comprehensive than traditional culture-based approaches. Several studies have used NGS to reveal that the swine intestinal tract harbours a complex bacterial community that can vary according to age, diet and management [15]. The intestinal microbiota has been shown to be essential for the maintenance of intestinal health and to play a major role in the development of the local and systemic immune system. Therefore, microbiota manipulation has been attempted with the objective of improving health [16]. Amongst the different strategies to modulate the intestinal microbiota, the use of probiotics is a viable option to be used in swine production [17][18][19].
Dietary supplementation with ZnO affects the intestinal microbiota of pigs [13,20], but studies comparing the impact of different products in a controlled environment remain to be performed. This study aimed to investigate the impact of ZnO, benzoic acid and probiotics on growth performance, control of diarrhea and cecal microbiota composition of piglets at the nursery phase.

Study design
One hundred and sixty weaned piglets (Camborough PIC x AG 337 sires) at 21 days of age were raised in an experimental unit (Akei Animal Research, Fartura, São Paulo, Brazil), according to a random design of four experimental groups: supplementation with 2500 ppm of Zinc Oxide (ZnO), supplementation with 0.4% of a commercial blend of benzoic acid and probiotics (Bacillus licheniformis, Bacillus subtilis and Enterococcus faecium NCIMB 10415; Vevogut P ® , DSM Animal Nutrition) (BA + P), supplementation with Zinc oxide plus benzoic acid and probiotic (ZnO + BA + P), and controls receiving only the basal diet without supplementation.
Piglets were housed in 2,55 m 2 pens (2 females or 2 males per pen) with controlled temperature. Therefore, each experimental group contained 40 animals (20 males and 20 females) housed in 10 different pens.
Additional file 1: Tables S1 and S2 specify the composition of diet used at the different growing phases: Pre starter I (21 to 28 days), pre-starter II (29 to 35 days), starter I (35 to 49 days) and starter II (49 to 63 days). Basal diet consisted of corn and soybean meal, with water and mashed ration offered ad libitum.

Performance
Piglets were weighed individually on day 0 (23 days of age), 5 (28 days of age), 12 (35 days of age), 26 (49 days of age) and 40 (63 days of age) of the trial to calculate average daily gain (ADG) for each period and considering all phases. Daily feed intake was determined by pen, and mortality was recorded to adjust the feed conversion ratio.
Fecal consistency was evaluated daily and classified as normal (score 0), soft (score 1), loose (score 2), or diarrhea (score 3) [21]. A diarrhea index expressed in percentage was calculated as 100 × number of piglets that had diarrhea/total number of piglets [22].
Thirty-two piglets (8 per treatment, being four barrows and four females, randomly chosen) were slaughtered at 65 days of life and cecal content was aseptically collected and immediately refrigerated for transportation and frozen at − 80 °C until DNA extraction. DNA was extracted using a commercial kit DNeasy PowerSoil (QIAGEN, Hilden, Germany) and the V4 region of the 16S rRNA gene amplified using the primers 515F: GTG CCA GCMGCC GCG GTAA and 806R: GGA CTA CHVGGG TWT CTAAT. Amplicons were then sequenced with a MiSeq Illumina sequencer (Illumina, Inc., San Diego, CA, USA) single end sequencing using a V3 kit for 300 bp reaction.
Bioinformatic analysis was carried out using the software mothur [23], following the protocol suggested by Kozich et al. (2013) [24]. Reads containing more than 300 bp or more than 8 homopolimers were excluded. Good quality reads were aligned to the SILVA reference alignment and classified according to the Ribosomal Database Project [25], 2016 release. Chimeras were removed with the vsearch algorithm. Reads classified as Chloroplast, mitochondria, Archaea or eukaryota were removed from the analysis. The phylotype approach was used by grouping all reads belonging to the same genus (94% similarity). Subsampling using the smallest number of reads obtained in a sample was used to standardize non-uniform samples in an attempt to avoid introducing bias into the analysis. The software mothur was also used to calculate richness by the total number of observed genera and by the Chao index. Diversity was estimated by the Simpson's index. Community membership, which takes into account each genus present in a community, was calculated by the Jaccard index, and community structure, which takes into account each genus and their relative abundances, was calculated by the Yue and Clayton index. The distance matrices created for both, membership and structure were subsequently used to generate dendrograms and Principal Coordinate Analysis (PCoA) graphs for visualisation of similarity between samples.

Statistical analysis
Body weight, ADG, DFI and FRC were submitted to analysis of variance (ANOVA) and means were compared by Tukey's test. The mean of the diarrhea index was compared between groups by the chi-squared test. Richness and diversity indices were compared between groups using an ANOVA test with Bonferroni correction for multiple comparisons using GraphPad Prism 7.0a (GraphPad Software Inc., California, USA), considering P values < 0.05 as significant. Relative abundances were compared by Tukey's test.
Beta diversity (Jaccard and Yue and Clayton indices) were compared using the analysis of molecular variance (AMOVA) test. Linear Discriminant Analysis Effective Size (LEfSe) was used to find bacterial taxa significantly associated to each treatment [26].

Performance
Animals supplemented with Benzoic acid plus probiotic and ZnO + Benzoic acid plus probiotic had significantly better FCR than the control group during the Pre-Starter I phase. Those groups, along with ZnO supplemented animals also performed better compared to controls during the Starter I phase I (Table 1). Final weight at Starter II was significantly higher in Benzoic acid plus probiotic likely because they started this phase heavier due to better performance at Starter I, as there was no statistical difference in ADG (despite a numerically greater average in that group). Overall, ZnO and Benzoic acid plus probiotic treatments significantly improved daily feed intake (DFI) and average daily gain (ADG) compared to controls.
The diarrhea ocurrence (number of piglets that had diarrhea) and the diarrhea index (100 × number of piglets that had diarrhea/total number of piglets) are presented in Table 2. Results of statistical analysis showed that controls presented higher frequency of severe diarrhea (score 3) and of intermediary and severe diarrhea together (score 2 + 3), suggesting that supplementation with ZnO and/or Benzoic acid plus probiotic was associated with lower incidence of diarrhea. No deaths were recorded during the study period.

Microbiota analysis
A total of 2,432,186 good quality sequences were retained for the final analysis, with an average of 76,006 reads per sample (SD 7048). A subsample using the smallest number of reads obtained in a sample (1310 reads) was used to standardize non-uniform samples in an attempt to avoid introducing bias into the analysis. This approach yielded 99.21% (SD, 0.15%) coverage indicating that most of the genera present in the samples were adequately detected. Figure 1A represents the overall relative abundance of the main phyla (> 1% abundance) found in the swine cecum. The 4 most abundant phyla along with unclassified bacteria represented 93% of all sequences. Firmicutes dominated the cecal microbiota of all groups, followed by Bacteroidetes. Figure 1B represents the relative abundances of the main genera found in the cecum of pigs supplemented with ZnO and/or Benzoic acid plus probiotic. The most abundant taxa were Roseburia, Prevotella, Gemmiger, and Streptococcus spp. Although some visual differences between treatments can be observed from Fig. 1B, there were no statistical differences found at the phylum or genus levels. The Linear Discriminant Analysis Effective Size (LefSe) analysis also failed to identify bacterial taxa associated to each of the specific treatments.
Results of the statistical analysis comparing richness and diversity indices between experimental groups are shown in Table 3 and Fig. 2. Richness represented by the observed number of genera and by the Chao index were statistically lower in ZnO and ZnO + Benzoic acid plus probiotic supplemented animals compared to controls, suggesting a strong influence of ZnO on the number of different bacteria comprising the cecal microbiota. Results also show that supplementation had no influence on bacterial diversity of studied animals addressed by the Simpson's index.
The beta diversity analysis that compares similarities between bacterial communities is represented by the Principal Coordinate Analysis (PCoA) (Fig. 3). This analysis demonstrated formation of two distinct clusters containing samples with and without supplementation with ZnO. Those differences were even more evident in communities' membership ( Fig. 3B) confirming that ZnO changes the overall microbiota composition, affecting mainly the less abundant organisms. This finding was further evidenced by dendrograms that demonstrated the similarity of the microbiota membership of animals receiving ZnO (Fig. 4). The strong impact of ZnO on the cecal microbiota of pigs was confirmed by the statistical analysis (AMOVA and Parsimony) comparing beta diversity between the experimental groups (Table 4).

Discussion
As expected, animals supplemented with feed additives had better performance compared to controls, which might be related with intestinal health (lower rates and severity of diarrhea) and to the improved digestibility of nutrients [27,28].
Zinc oxide aids to preserve intestinal mucosa integrity, modulating the immune system and protecting it against pathogenic bacteria [12]. Higher concentrations of ZnO have been associated with greater weight gain [17]. In the present study, the use of ZnO had a strong impact on the microbiota composition of the studied piglets as revealed by community analysis and reduced richness. Many studies have reported changes in intestinal bacteria of pigs   caused by ZnO [29,30]. The mechanisms by which ZnO improves growth performance are not completely understood, but it might increase the intestinal absorpiteve capacity by increasing the villus height [31]. Non-dissociated benzoic acid passes through the cellular membrane of bacteria releasing protons that acidify the medium [32], changing the bacterial metabolism and suppressing essential enzymes, such as decarboxylases and catalases [33,34] Furthermore, benzoic acid improves the structure of the intestinal mucosa, reducing the crypt depth and increasing the villus/crypt ratio  culminating in improved digestibility and performance [35][36][37]. Probiotic organisms (i.e. Bacillus cereus, E. faecium, Saccharomyces cerevisae boulardii) can improve intestinal absorption and increase the transport of L-glutamine and ion secretion in piglets, preserving the mucosa integrity by reducing enterocyte mortality and improving protein digestibility [21,37,38]. Bacillus and Enterococcus are among the most used probiotics because of their potential to depleat pathogenens from the digestive tract and benefit commensal bacteria, controlling for diarrhea and improving performance [21,39,40]. The use of a specific strain of Ligilactobacillus salivarius was associated with increased clostridia and lactobacilli species as well as greater production of volatile fatty acids [19]. The use of symbiotics (a combination of probiotics with prebiotics) have also been shown to benefit fibrolytic bacteria, improving performancing and reducing inflammation in pigls [18].
Firmicutes and Bacteroidetes were the two major phyla found in the cecum of the studied piglets, which was in agreement with other studies [41][42][43][44]. Although not statistically significant, the abundance of Proteobacteria was higher in the groups not receiving ZnO (Fig. 1B). The inhibition of some Proteocateria species, such as Escherichia and Salmonella may have contributed to the lower rates of diarrhea and better performance observed in the present study [45][46][47][48][49]. Further studies using larger sample sizes are necessary to confirm this hypothesis. Specific taxa could not be associated to each of the treatment groups, likely because of interindividual variabilities in their relative abudances. The high abundance of Prevotella reported here can be explained by the increase in those organisms after weaning in piglets mainly related  with the fibrolytic activity of the genus [47,50,51]. The genus Roseburia was also highly abundant in the intestinal tract of all studied piglets. These butyrate producing bacteria have been shown to be part of the normal microbiota of piglets [52] and belongs to the family Lachnospiraceae (Firmicutes), which is normally associated good intestinal health [53]. The genus Gemmiger is another Firmicutes of the Ruminococcaceae family that was highly abundant in this study. This bacterium seems to be favoured by the insertion of soybean meal in the diet of piglets [54], which might explain our results since this protein source was an important component of the diet used in the present study. The abundance of Megasphaera was numerically lower in the cecum of piglets receiving ZnO only, but the role of these commensal bacteria in the swine intestines deserves better attention [55].
Weaning is the most stressing event in swine production due to abrupt diet changes, displacement to new environments and housing with new individuals. Not surprisingly, this is the phase with the highest incidence of diarrhea and with the highest use of antibiotics [56]. The association of benzoic acid with probiotics and prebiotics has been suggested to have a synergic effect to favor weaned piglets' performance [37,57,58], which are consistent with the present study. A recent study reported that supplementation of ZnO (1200 ppm) with a probiotic containing Bacillus coagulans, B. licheniformis, B. subtilis and C. butyricum had similar effect than higher doses of ZnO (3000 ppm) on many performance paramethers [1,17].
Although the methods employed for this study are rather descriptive and not intended to prove a cause consequence relationship, it can be hypothesized that the microbiota changes induced by ZnO may play a role in weight gain by acting at different levels, such as modulating the local immunity, benefiting species that are more efficient in extracting energy from food or by competing with pathogenic organisms. Noteworthy, the beneficial effects of benzoic acid plus probiotic observed in the present study (i.e. on growth performance and diarrhea control) were similar to the ones obtained with the use of ZnO without inducing detectable changes in microbiota composition, avoiding the negative aspects of ZnO usage, such as development of resistance and environmental contamination.

Conclusions
The use of ZnO is associated with decreased richness and with changes in microbiota composition of weaning piglets. The use of Benzoic acid with probiotics yields similar performance results with lower impact on the gut microbiota compared to ZnO, and it should be considered as a potential alternative in swine production.