1. Introduction
In intensive swine production systems, optimizing gut function is critical for ensuring both nutrition and animal health. This is particularly true during weaning, as it represents a critical stage of adaptation to major and abrupt changes that young pigs experience [
1]. Indeed, from the switch in diet to dealing with multiple stressors such as separation from the dam, settling into an unfamiliar environment, and adjustment to a new social hierarchy, weaning can have short- and long-term consequences on gut function, which ultimately can impact productivity and profitability [
2,
3].
The change in diet combined with reduced feed intake at weaning impairs the expression of digestive enzymes as well as the absorptive capacity of the epithelial lining [
2,
4,
5,
6,
7,
8,
9]. In addition, the gut of weaned pigs is affected by increased permeability, alterations in enteric neuron numbers and cholinergic activity, as well as by upregulation of proinflammatory cytokines and hyperplasia of intestinal mast cells [
10]. Together, these factors contribute to increased disease susceptibility in young pigs. For instance, impaired host digestion and absorption provide a source of nutrients for opportunistic pathogens, whose colonization and proliferation are facilitated by a compromised immune response. These conditions can lead to further damage from impaired gut barrier function, triggering the onset of diarrhea or other digestive dysfunctions [
11,
12,
13], which ultimately result in economic losses to producers [
5,
6,
14].
Traditionally, diarrhea and digestive dysfunctions when weaning have been managed through the prophylactic use of antibiotics. Under these conditions, antibiotics also exhibit growth-promoting effects. However, higher incidences of antibiotic resistance in food animals, as well as increased consumer concerns, have led to the implementation of policies that restrict or ban the use of antibiotics as growth promoters [
15]. While the mechanisms responsible for the growth-promoting effects of antibiotics remain poorly understood [
16], they are likely to involve modulation of the gut microbiota, since these agents act as antimicrobials. This is consistent with a growing body of literature reporting that the composition of gut microbial communities can be associated with the health status and performance of individual animals [
13,
17,
18]. Thus, compounds or agents that can change gut symbiont profiles have the potential to be developed as tools to improve critical livestock production parameters.
To date, a number of candidate alternatives to antibiotics have been developed. These include formulations based on plant compounds and extracts, such as prebiotics, essential oils, or other phytochemicals, as well as microbial-based products, often referred to as direct-fed microbials (DFMs), which include probiotics and postbiotics. Supplementation of nursery diets with probiotics, which consist of live microorganisms [
19], is thought to reduce pathogen proliferation by establishing beneficial bacteria and lowering gut pH, as well as to enhance gut health and barrier function by providing substrates such as short-chain fatty acids (SCFAs) to host epithelial cells [
20]. Postbiotics, which consist of microbial products derived from probiotics grown in culture [
21], contain active compounds with antimicrobial, immunomodulatory, and/or anti-inflammatory properties. Postbiotics are considered preferable to probiotics under certain conditions, as they contain the same beneficial components, in the form of lactic acid, bacteriocins, short-chain fatty acids, exopolysaccharides, and cell wall fragments, without concerns such as having to ensure the survival of live organisms or the potential production of undesirable compounds when exposed to the gut environment [
21].
In this context, the study presented in this report aimed to evaluate the effects of feeding a Lactobacillus-based probiotic or a Bifidobacterium-based postbiotic to weaned pigs until the beginning of the grow-finisher phase. Overall, the DFMs tested were beneficial in terms of performance and intestinal morphology, and they were found to impact the fecal bacterial composition of pigs.
2. Materials and Methods
2.1. Animals, Housing, and Diets
The trial described in this report was conducted at the South Dakota State University (SDSU) Off-Site Wean-to-Finish Barn, with all procedures approved by the SDSU Institutional Animal Care and Use Committee (IACUC) before the start of the study. A total of 1040 pigs (initial body weight (BW) 6.1 ± 0.1 kg) were selected based on overall health by visual assessment from an initial population of 1244 weaned pigs (PIC 800 × PIC) that were delivered to the facility from a single local producer. Selected pigs were allocated to 40 pens, blocked by weight and barn location, in order to populate individual pens with 26 pigs (13 barrows and 13 gilts per pen). Each pen had an area of 3.1 m × 6.9 m, with approximately 0.82 m
2 for each pig. The experimental design was a complete block, with 10 blocks per treatment and every treatment represented in every section of the barn. Diets were formulated to provide lysine and energy that met or exceeded the NRC (2012) recommendations for each diet ([
22];
Supplementary Material Document S1). Pens were assigned to one of four dietary treatments: Control diet (Control), Control diet + 0.1% of a
Lactobacillus-based probiotic (LPr-0.1%), Control diet + 0.2% of a
Lactobacillus-based probiotic (LPr-02%), or Control diet + 0.2% of a
Bifidobacterium-based postbiotic (BPo-0.2%). They were each included in their respective treatment diets starting at the very beginning of the nursery phase (D0, Phase 1) until D47, which corresponded to the beginning of the grower phase (Phase 3). LPr and BPo were both manufactured according to proprietary protocols. The LPr product consisted of two proprietary
Lactobacilli strains at 2 × 10
9 CFU/g.
2.2. Growth Performance Measurements and Calculations
Pigs were weighed by pen at barn entry (D0), as well as on D10 (end of Phase 1), D21 (Phase 2), D47 (beginning of Phase 3, last day of DFM supplementation), D70, D105, and D135. Each pen used in the trial was equipped with one dry feeder and two cup waterers for ad libitum access to feed and water, respectively. A Feedlogic M-Series system (Feedlogic ComDel Innovation, Wilmar, MN, USA) was used for dispensing feed. On each weigh day, the amount of feed remaining was determined using an in-house calibration curve by measuring the distance from the top of the feeder to the top of the leveled feed; the calibration curve was designed to also correct for feed density. Feed disappearance was then calculated by subtracting the remaining feed from the total feed delivered since the previous weigh day, which had been recorded by the Feedlogic software (v.2.4.3).
Average daily gain (ADG), average daily feed intake (ADFI), and gain-to-feed ratio (G/F) were calculated using the following:
To avoid compromising the calculations of performance results, ‘pig days’ took into account the removal date(s) and weight(s) of any pig(s) between weigh days.
2.3. Histological Analysis
Samples from a total of 40 pigs (n = 10/treatment) harvested on D10 were analyzed to determine the potential effect of DFMs on intestinal morphology; one representative from each pen was randomly selected based on weight and health status. Selected pigs were euthanized using a non-penetrating captive bolt gun (CASH®, Frontmatec Accles & Shelvoke Ltd., Sutton Coldfield, UK), which was applied to deliver a concussive force sufficient to cause effective stunning; this procedure was performed according to guidelines for the humane euthanasia of swine as prescribed by the American Association of Swine Veterinarians (AASV) and the National Pork Board (NPB). Euthanasia was confirmed by verifying that respiration had ceased and that no corneal reflex was present before proceeding to sample collection. The intestinal tract was dissected and removed from the body cavity and then samples 5 cm in length were collected from the jejunum (midpoint of the small intestine) and ileum (10 cm proximal from the ileocecal junction).
Intestinal samples were fixed, sectioned, and then stained with hematoxylin and eosin at the SDSU Animal Disease Research and Diagnostic Laboratory. Villus height (VH), measured from the top of the villus to the villus–crypt junction, and crypt depth (CD), measured from the villus–crypt junction to the base, were determined at 4× magnification using a Micromaster® microscope (Fisher Scientific, Waltham, MA, USA) equipped with a 0.55× wifi camera eyepiece (MC500-W 3rd Gen., Meiji Techno Co., Ltd., Saitama, Japan) and Micro-Capture software (v.6.9.11) (Meiji Techno Co., Ltd., Saitama, Japan). The average of ten measurements from each sample was used as a representative of each pen.
2.4. Isolation of Microbial Genomic DNA and Sequencing of 16S rRNA Gene Amplicons
On D10 and D47, fecal samples were collected by rectal palpation from one representative pig in each pen for the Control, LPr-02%, and BPo-02% groups. The same pigs were sampled at each time point. Samples (total
n = 60) were stored frozen at −20 °C until further processed. Microbial genomic DNA was extracted from individual samples using a bead-beating plus column approach [
23] using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). The V1–V3 regions of the bacterial 16S rRNA gene were selected for analysis and targeted by PCR using the universal forward 27F 5′AGAGTTTGATCMTGCTCAG and reverse 519R 5′GWATTACCGCGCGCGCTG primers [
24]. Purified microbial genomic DNA samples were processed by Molecular Research DNA (MRDNA, Shallowater, TX, USA) for V1–V3 amplification, and for amplicon sequencing with the Illumina MiSeq 2 × 300 platform to generate overlapping paired-end reads.
2.5. Bioinformatics Analyses
A combination of custom-written Perl scripts and publicly available software was employed for 16S rRNA data analysis. The overlapping paired-end reads generated from the same V1–V3 amplicons were first merged, then screened for quality by filtering for (1) the presence of both intact 27F and 519R primer sequences, (2) a minimal average Phred quality score of Q33, and (3) a length between 400 and 580 nt [
25]. Quality-filtered sequences were then aligned and clustered into Operational Taxonomic Units (OTUs) using a threshold of 4% sequence dissimilarity. This cutoff is more suitable for the V1–V3 hypervariable regions than the typical 3% threshold that is commonly used for 16S rRNA sequence data clustering [
26,
27]. The OTUs were then screened using three independent approaches to flag artifacts needing to be removed [
25]. Identification of chimeric sequences was performed first, with the commands ‘chimera.slayer’ [
28] and ‘chimera.uchime’ [
29] using the open-source software package MOTHUR (v.1.44.1) [
30]. A sequence-alignment-based approach with BLAST [
31] was then used on the remaining OTUs to assess the quality of their ends; OTUs were considered artifacts when more than five nucleotides were missing from the 5′ or 3′ ends of their respective alignments with their closest match of equal or longer sequence length from the NCBI ‘nt’ database. For OTUs with only one or two assigned reads, an additional screen for artifacts was performed, where only OTUs showing a perfect or near-perfect match to a sequence in the NCBI ‘nt’ database were kept for analysis; a maximum of 1% dissimilarity was tolerated. All OTUs and their assigned reads that were flagged during these screens were removed from further analyses. For all artifact-free OTUs, phylum and family-level affiliations were determined using the RDP Classifier [
32]. For the most abundant OTUs, the closest valid relatives were also identified using BLAST against the ‘refseq_rna’ database [
31].
The alpha diversity indices ‘Observed OTUs’, ‘Chao’, ‘Ace’, and ‘Shannon’ were generated using the ‘summary.single’ command in MOTHUR (v.1.44.1) [
30]. For beta diversity analysis, Bray–Curtis distances were calculated first, then used as input for principal coordinate analysis (PCoA), using the commands ‘summary.shared’ then ‘pcoa’, respectively, in MOTHUR (v.1.44.1) [
30]. A PCoA plot was created using the Tableau Visualization Software (Version 2020.4,
https://www.tableau.com/products/new-features accessed on 22 April 2024).
2.6. Statistical Analyses
Data for BW, ADG, ADFI, G/F, and histological measurements were analyzed using the PROC GLIMMIX procedure of SAS (Version 9.4, SAS Inst. Inc., Cary, NC, USA), with ‘dietary treatment’ as the main effect and ‘pen’ as the experimental unit. Statistical testing of microbiome data was performed in ‘R’ (Version 3.6.0). Analysis of variance (ANOVA) with a post hoc ‘HSD.test’ pairwise function was used to compare alpha diversity indices. The Kruskal–Wallis rank-sum test was performed for the analysis of taxonomic groups and most abundant OTUs (non-parametric data), concomitant with the pairwise Wilcoxon rank-sum test, which included the Benjamini–Hochberg correction for controlling the false discovery rate. For all statistical tests described above, a threshold of p ≤ 0.05 was considered significant.
4. Discussion
In recent years, an increasing number of commercial feed additives have been developed with the aim of optimizing gut function, which has long been recognized as critical in ensuring animal health and performance [
1,
2,
37]. Feed additives encompass a wide range of products, from formulations based on plant compounds and extracts such as prebiotics, essential oils, or other phytochemicals, to microbial-based products, often referred to as direct-fed microbials, which include probiotics and postbiotics. In the latter category, a wide spectrum of bacterial strains have been developed for use as DFMs, with many affiliated with species in the genera
Lactobacillus,
Bacillus,
Enterococcus,
Bifidobacterium, or
Clostridium [
38]. In this study, the effects of a
Lactobacillus-based probiotic and a
Bifidobacterium-based postbiotic were tested in pigs through dietary supplementation, which started right after weaning and lasted for the duration of the nursery phase. Overall, the DFMs tested provided benefits in terms of performance and intestinal morphology, and they were found to impact the fecal bacterial composition of pigs.
To date, reported effects of DFMs on growth performance have been varied and inconsistent overall. Indeed, the inclusion of probiotics has been found to either decrease, increase, or have no significant effect on ADG, feed conversion ratio, or ADFI in weaned pigs. For instance, it was reported that a
Lactobacillus reuterii-based probiotic resulted in higher ADG with no significant effect on ADFI [
39], which is in contrast to the results described in this report; perhaps differences in experimental design could explain these contrasting outcomes, such as fewer pigs per pen and fewer pigs enrolled in the study by Yi et al. (2018), as well as comparison of the probiotic treatment group to a control group whose diet was supplemented with two antibiotics [
39]. Intriguingly, supplementation with a probiotic formulation consisting of two different strains, one affiliated with
Lactobacillus spp. and the other with
Bifidobacterium thermacidophilum, resulted in multiple performance metrics (BW, ADG, and ADFI) being significantly higher [
40]; notably, the strains used in that study had originally been isolated from fresh feces collected from Tibetan pigs, and would therefore correspond to normal residents in the swine gut [
40]. However, strains of bacterial species from other microbial habitats, such as soils, have also been reported to be beneficial when used as DFMs. For instance, pigs supplemented with a
Bacillus toyonensis-based product had higher BWs and a lower feed conversion ratio compared to the control group [
41]. A formulation combining strains of
Bacillus subtilis and
Bacillus amyloliquefaciens was also reported to be effective, as pigs fed the DFM had greater ADG and G/F than pigs fed the control diet [
42].
While varying effects on performance could be explained by the wide range of bacterial species that are used to generate DFM strains, the same level of variation in effects can be observed when comparing microbial products developed from the same bacterial species. For instance, strains of
Clostridium butyricum, a butyrate-producing member of the swine gut microbiome [
43], have been tested as DFMs in a number of independent studies, with mixed and varying impacts on performance ranging from a lack of observable effects [
44] to higher ADG [
45]. One possible variable could be the health status of the pigs on trial. Indeed, feeding a
C. butyricum-based DFM to lipopolysaccharide (LPS)-challenged pigs resulted in higher final BWs [
46], and a number of DFMs have been reported as beneficial in mitigating the impacts of rotavirus infections [
47]. At this point, it also remains unclear to what extent factors such as DFM inclusion levels, properties of the bacterial strains, and DFM formulation can affect performance. Additional research is clearly needed to elucidate the varying outcomes that DFM supplementation can have on pig performance.
Low feed intake is one of the main causes of reduced performance during early weaning [
48]. Indeed, as they adapt to a solid diet that is less digestible and palatable, a large proportion of pigs have severely reduced metabolizable energy intake during this period [
49], and it may take up to two weeks for them to recover to pre-weaning energy intake levels. In addition to a direct reduction in nutrient availability, low feed intake also causes the shortening of villi, hyperplasia of crypt cells, as well as damage to gut mucosal integrity [
5,
50,
51]. Since these morphological changes negatively impact the absorption capacity of the gut, they further impair the nutrition and performance of weaned pigs.
In contrast to their varying effects on performance, DFMs have been more consistently associated with improved intestinal morphology, expressed in the form of longer villi, reduced crypt depth, as well as higher VH/CD ratios. These beneficial effects on intestinal tissue have been reported for a wide range of DFMs, including yeast-based products [
52],
C. butyricum [
53,
54,
55,
56],
Enterococcus faecalis [
57],
Lactiplantibacillus plantarum [
58,
59],
L. reuteri [
39],
L. salivarius [
60],
B. subtilis [
61], and
B. amyloliquefaciens [
62]. Product formulations containing a combination of DFM strains such as
B. subtilis and
B. amyloliquefaciens [
42],
C. butyricum and
E. faecalis [
63], as well as others consisting of more complex mixtures [
64,
65], have also proven effective. These reported benefits to intestinal morphology align well with the histological analysis of pigs that were fed either a
Lactobacillus-based probiotic or a
Bifidobacterium-based postbiotic in this study.
Conceptually, DFMs would be expected to safely pass through the acid environment of the stomach and then act by changing or modulating the composition and/or activity of the gut microbiota. Considering the chemical and biological intricacies of the gut microbial environment as well as the broad spectrum of strains and species that have been developed as DFMs, the range of potential mechanisms that could be involved is likely very complex. Accordingly, to date, available reports on the impact of DFMs on gut microbial composition indicate that a range of different responses can be expected. Among possible outcomes,
Lactobacillus-related or affiliated symbionts so far represent one of the main types of gut bacterial groups that can be modulated by DFMs [
46,
55,
56,
64,
66], an effect that was also observed in this study. This finding was not surprising, considering the importance of these lactic acid-producing bacteria for gut health. Indeed, several effects that are beneficial to the host have been reported for
Lactobacilli, such as preventing adhesion of pathogens to the gut mucosa, suppressing the growth of intestinal pathogens, and/or promoting the colonization of beneficial bacteria [
67,
68,
69,
70,
71,
72,
73].
Lactobacilli have also been reported to enhance immune functions [
74,
75].
In addition to
Lactobacilli, bacterial groups from other lineages have also been reported to be impacted by DFMs, including
Bifidobacterium [
54] and Erysipelotrichaceae [
53], as well as more complex combinations [
40,
44,
76,
77]. The identification, in this study, of three OTUs that were not affiliated with
Lactobacilli and whose abundances differed in treated pigs is thus consistent with these reported findings. Based on the V1–V3 region of their 16S rRNA genes, Ssd-00950 and Ssd-01187 were predicted to be strains of
C. massiliense [
78] and
A. porci [
79], respectively, which were both originally isolated from gut environments. At the time of submission of this report,
A. porci has been reported as a Gram-positive obligate anaerobe that can ferment glucose into lactate, acetate, and formate [
79]; while information on cellular or metabolic characteristics of
C. massiliense was not available. As both OTUs were found in significantly higher abundance in pigs fed a DFM, it is possible that compounds released from the feed additives provided a growth or survival advantage. However, future research will be necessary to determine the mechanism of action involved. Ssd-00039, the third OTU showing a difference in abundance in DFM-supplemented pigs, was predicted to be a strain of
S. alactolyticus, a predominant commensal in the swine hindgut [
80,
81,
82] that can produce lactate from fermenting cellobiose, maltose, galactose, fructose, glucose, or mannose [
80,
81]. The high representation of Ssd-00039 in D47 samples could then be attributed to the likely abundance of these compounds in feed ingredients, which is supported by a previous report that this OTU was present in higher abundance in nursery pigs fed a mannose oligosaccharides-based feed-additive [
25]. However, as currently available information is not sufficient to explain how DFM supplementation can result in a lower representation of Ssd-00039, additional investigations will be necessary to gain further insights.