Evaluation of Sialyllactose Supplementation of a Prebiotic-Containing Formula on Growth, Intestinal Development, and Bacterial Colonization in the Neonatal Piglet.

BACKGROUND: Sialyllactose (SL) is a highly abundant oligosaccharide in human milk that has been shown to influence intestinal maturation and cognitive development and exert bifidogenic effects on the gut microbiota. The SL content of infant formula is significantly less than that of human milk, therefore there is interest in determining the effect of supplementing SL to infant formula at the levels in human milk on neonatal outcomes.
OBJECTIVE: The aim of this study was to investigate the effect of varying doses of dietary SL compared with a milk replacer formula on weight gain, gastrointestinal development, and microbiota composition in piglets.
METHODS: Thirty-eight intact male piglets were randomly assigned to 1 of 4 experimental diets from 2 to 32-33 d of age. Diets were formulated to contain SL at 0 mg/L (CON), 130 mg/L (LOW), 380 mg/L (MOD), or 760 mg/L (HIGH). At 32-33 d of age, blood was collected for serum chemistry and blood cellular analyses, and coagulation time. Immediately after humane killing, the small intestine was excised and intestinal segments fixed for quantification of mucin-producing goblet cells and morphologic analysis. In addition, mucosal disaccharide activity was assessed. Colonic luminal contents and feces were collected for measurement of pH, dry matter, volatile fatty acids, and the microbiota.
RESULTS: SL at ≤760 mg/L supported normal growth, intestinal development, and enzyme activity as well as serum chemistries and hematology (P > 0.05). In addition, SL supplementation did not affect overall microbiota structure and diversity in ascending colon contents and feces, but had minor effects on the relative abundances of specific microbes.
CONCLUSIONS: The findings in this study demonstrate that SL addition to a prebiotic-containing formula was well-tolerated by neonatal piglets, supported normal growth, and did not result in any adverse effects on serum chemistries or intestinal development.

Introduction

Despite changes in infant formula composition over the past 2 decades, substantial compositional differences with human milk persist, a notable example being the absence of structurally complex milk oligosaccharides (MOs) due to the low concentrations in the bovine milk starting material (1). MOs are synthesized in the mammary gland as elongation products of lactose with combinations of galactose, N-acetylglucosamine, fucose, and/or sialic acid (2). MOs are known to modulate the infant microbiota and to influence immune and neurologic development

(3–5) and their low concentrations in formula may be a factor that contributes to the differences in respiratory tract and gastrointestinal infections and cognitive outcomes between breast- and formula-fed infants (5, 6).

MOs are present in human milk at concentrations ranging from 5 to 15 g/L (7) and are mostly found as neutral or fucosylated, whereas sialic acid (SA)-containing MOs comprise <20% (8). SA in mature human milk is present as oligosaccharides (69–76%), whereas 21–28% is bound to proteins, and ∼3% is present in the free form and bound to gangliosides (5). In stark contrast, SA in cow milk–based infant formula is mostly bound to protein, with 27.8% present as oligosaccharides and <1% is found in its free form (5). The most abundant SA-containing MOs in human milk are 6′sialyllactose (6′SL), 3′sialyllactose (3′SL), disialyllactose-N-tetraose, and sialyllacto-N-tetraose. A recent worldwide survey of human MO (HMO) concentrations showed that 6′SL and 3′SL were present in all milk samples and concentrations ranged from 0.13 to 0.56 and 0.26 to 0.39 g/L, respectively (9). Others have reported broader variation, with higher concentration of 6′SL in colostrum than in mature milk, but no differences between early and later milk concentrations of 3′SL (10).

Studies in human infants and neonatal piglets have demonstrated that MOs, including SA-bound oligosaccharides, survive digestion and reach the large intestine intact (11, 12). Once in the intestine, MOs shape the composition of microbial communities by stimulating the growth of specific bacteria (13). Of interest is the study reported by Wiese et al. (14), in which 3′SL, 6′SL, 3′-fucosyllactose, and fructo-oligosaccharide fermentability was investigated in an in vitro colon model. 3′SL produced the strongest bifidogenic effect and higher butyrate production. In addition, in vivo studies have also shown the ability of SA-containing oligosaccharides to manipulate the microbiota. Studies in mice indicated that exposure to 3′SL in milk not only affected the microbiota, but also the resistance to dextran sodium sulfate-induced colitis (15). Similarly, neonatal pigs fed 6′SL presented a microbiome profile that differed from those fed a control (CON) diet (16).

Oligosaccharides are also known to modulate gut maturation as demonstrated in in vitro studies (17–19). Holscher et al. (19) used intestinal cell lines to build a model of the crypt-villus axis and were able to show that the presence of HMO [6′SL, 2′fucosyllactose (2′FL), lacto-N-neotetraose (LNnT)] reduced cell proliferation and barrier function, and increased gut maturation, but had no impact on alkaline phosphatase, an enzyme used as a marker of crypt-villus differentiation (20). Lactase and sucrase are among the digestive enzymes present at the villus tip and are often used as markers of cell differentiation. Berding et al. (21) demonstrated that the incorporation of galacto-oligosaccharides (GOSs) and polydextrose (PDX) in neonatal piglets increased jejunal lactase activity. Furthermore, GOSs have also been shown to influence goblet cell function by increasing the expression of mucin (MUC) genes (22). However, information on the in vivo effects of sialyllactose (SL) on digestive enzymes and goblet cells is lacking.

The most commonly added prebiotics to infant formula are fructo-oligosaccharide, GOS, and PDX. Compared to a standard infant formula, infants fed formula supplemented with GOS and PDX at 4 g/L (1:1 ratio) had stool consistency that more closely resembled that of breastfed infants; the supplemented formula also promoted an increase in Bifidobacteriuminfantis and Bifidobacteriumlongum in feces of healthy infants during a 60-d feeding period, while having no impact on weight gain (23, 24). More recently, neutral and fucosylated MOs have been added to infant formula and have shown health benefits compared with unsupplemented formula (6, 25, 26). To date, no commercial infant formula has been supplemented with SA-containing oligosaccharides, although the effects of SL have been tested in preclinical animal models. In neonatal pigs, adding 2 or 4 g/L of 3′SL or 6′SL to a formula devoid of other prebiotics did not affect weight gain, but increased brain ganglioside–bound SA and changed the gut microbiota profile (16). Pure 3′SL and 6′SL have not been tested in human infant clinical trials; however, a standard formula supplemented with 10 g/L bovine milk-oligosaccharides, a mixture containing 3′SL, 6′SL, and GOS, fed to infants for 4 mo led to microbiota changes (26).

These preliminary findings suggest a potential benefit of supplementing infant formula with SL; however, the bioactivity of SL as part of an infant formula containing other prebiotics was unknown. In our initial findings, prebiotic-enriched formula supplemented with SL in piglets influenced brain development with changes in levels of bound SA in the prefrontal cortex, the ratio of free SA to bound SA in the hippocampus, and axial and radial diffusivity in the corpus callosum (27). In addition to the effects of SL on brain development, our lab set out to investigate the impact of adding SL to a formula containing GOS and PDX on weight gain, serum chemistries, intestinal morphology, disaccharidase activity, and the composition of the microbiota in newborn piglets. The piglet is a commonly used preclinical model for formula ingredient testing owing to anatomic, developmental, and brain growth trajectory similarities to the human infant (28, 29). We hypothesized that dietary SL would be well-tolerated and would modify the microbiota compared with milk replacer without supplemental SL.

Methods

Animal study design

All animal care and experimental procedures were approved by the University of Illinois Institutional Animal Care and Use Committee and complied with the Guide for the Care and Use of Laboratory Animals (30). Beginning at postnatal day 2 (PND 2), 38 naturally farrowed intact male piglets were randomly assigned to 1 of 4 milk replacer diets until 32 or 33 d of age. The trial was completed in 4 replicates (5–12 piglets/replicate), with 8–11 piglets/diet, selected from 14 litters, to control for genetics and initial body weight (BW). A total of 12 pigs/dietary treatment were initially started on the study; however, several pigs were removed from the study for reasons not associated with the dietary treatments (27). All pigs were housed in custom piglet-rearing units and were maintained on a 12-h light/dark cycle. Researchers were blinded to treatment groups throughout the study and sample analyses.

Dietary treatment

All diets were produced by Mead Johnson Nutrition with the use of a proprietary blend of nutrients formulated to meet the nutritional needs of growing piglets. The bovine milk-based CON and experimental diets are shown in SupplementalTable 1. The CON diet contained DHA (87 mg/100 g milk replacer powder; DSM) and arachidonic acid (174 mg/100 g milk replacer powder; DSM). In addition, diets contained GOS (1.0 g/100 g milk replacer powder; FrieslandCampina) and PDX (1.0 g/100 g milk replacer powder; Danisco). Experimental diets were formulated with the CON diet as the base and supplemented with bovine-derived modified whey enriched with SL (Lacprodan SAL-10; Arla Foods Ingredients Group P/S) to provide final SL concentrations of: low SL (65 mg SL/100 g milk replacer powder; LOW), moderate SL (190 mg SL/100 g milk replacer powder; MOD), and high SL (380 mg SL/100 g milk replacer powder; HIGH). The milk replacer powder was reconstituted fresh daily at 200 g of dry powder/800 g of water, and piglets were fed at 285 and 325 mL of reconstituted diet/kg of BW starting on 3 and 8 d of age, respectively. At this reconstitution rate, all diets contained arachidonic acid (348 mg/L), DHA (174 mg/L), and PDX/GOS (2 g/L each). The reconstituted experimental milk replacers were formulated to contain 130 mg SL/L (LOW), 380 mg SL/L (MOD), or 760 mg SL/L (HIGH). Analytic assessment conducted after study completion showed the levels of SL in the diets were: CON (55 mg SL/L), LOW (159 mg SL/L), MOD (429 mg SL/L), and HIGH (779 mg SL/L), owing to inherent SL.

Sample collection

At the end of the study, piglets were deprived of food for 8 h before being killed. Piglets were sedated with an intramuscular injection of Telazol [tiletamine hydrogen chloride (HCl) and zolazepam HCl, 3.5 mg/kg BW each; Zoetis] and blood samples were collected by cardiac puncture into heparin-laced and 3.2% sodium citrate–containing vacutainer tubes (BD Biosciences) for coagulation time and cell blood count with differential, respectively. In addition, blood samples were also collected in heparin-free tubes for the isolation of serum for chemistry analysis. Piglets were then killed by an intracardiac injection of sodium pentobarbital (72 mg/kg BW Fatal Plus; Vortech Pharmaceuticals). The small intestine was excised between the pyloric sphincter and ileocecal valve for measurement of total small intestinal weight and length. Thereafter, the small intestine was cut at 10% and 85% from the proximal end resulting in 3 segments: duodenum, jejunum, and ileum, respectively (31). Intestinal segments were flushed with PBS and 1- to 2-cm pieces were fixed in Bouin's solution for 24 h at room temperature for goblet cell analyses and intestinal morphology. In addition, a section of each segment was cut longitudinally, and mucosal scrapings were obtained and frozen in liquid nitrogen. Ascending and descending colon samples were obtained immediately following (distal to) the cecum or proximal to the rectum, respectively. Ascending content and feces were sampled for volatile fatty acids (VFA) analyses, 16S rRNA sequencing, and qPCR for quantification of bacteria species.

Hematologic analyses

Serum chemistries and coagulation time [partial prothrombin time (PT) and activated partial thromboplastin time (aPTT)] were determined with the use of an Olympus AU680 chemistry analyzer (Beckman Coulter) and an STA-Compact coagulation analyzer (Diagnostica STAGO), respectively, at the Clinical Pathology lab within the University of Illinois College of Veterinary Medicine. Serum was obtained after blood vials were centrifuged at 2200 × g for 20 min at 4°C in a benchtop centrifuge (CS-6R Centrifuge, Beckman Coulter Life Sciences). The present serum chemistry panel included minerals (calcium, phosphorus, magnesium), electrolytes (sodium, potassium, chloride), protein (total protein, albumin, and globulin), metabolites (glucose, total cholesterol, TGs, creatinine, urea, total bilirubin, and bicarbonate), and the enzymes alkaline phosphatase, aspartate transaminase, γ-glutamyltransferase, creatine phosphokinase, and glutamate dehydrogenase (GLDH). In addition, complete blood count was performed on CELL-DYN 3700 (Abbott), and trained technicians at the College of Veterinary Medicine performed differential analysis. The variables evaluated in our study were RBC count, hemoglobin concentration, hematocrit value, mean corpuscular volume, mean corpuscular hemoglobin, and mean corpuscular and hemoglobin concentration. Total white blood cell (WBC) count and differential WBC analyses (neutrophils and lymphocytes) were also performed. Platelet indexes were analyzed and included platelet count and mean platelet volume.

Intestinal analyses

Histomorphologic analyses

Bouin's-fixed sections of duodenum, jejunum, ileum, and ascending colon were processed into paraffin blocks, sectioned into 5-μm slices, and placed on glass slides. For the histomorphologic analyses, digital images (10× magnification) of hematoxylin-eosin–stained tissues were captured with the use of the NanoZoomer Slider Scanner (Hamamatsu Corporation; Institute for Genomic Biology Microscopy and Imaging Facility, University of Illinois). Small intestinal villus height (micrometers), villus area (millimeters squared), crypt depth (micrometers), crypt area (millimeters squared), and colon cuff depth (micrometers) and cuff area (millimeters squared) were measured from 10–15 well-oriented crypts-villi with the use of AxioVision 4.8 Digital Image Processing Software (Carl Zeiss MicroImaging, Inc.).

Goblet cell quantification

Slides with ileum and ascending colon sections were stained with High Iron Diamine/Alcian Blue to detect sulfated and carboxylated mucin-producing cells as described by Croix et al. (32). Digital images were obtained as described earlier. Goblet cell numbers were determined by enumerating the number of stained cells within an area delimited by a colon cuff, ileal villus, and ileal crypt. A total of 8 ascending colon cuffs, ileum villi, and crypts were used for the quantification. Data are presented as total number of cells or cell number normalized by crypt, cuff, and villus area (micrometers squared).

Disaccharidase activity

Intestinal lactase and sucrase activities were measured as previously described (33, 34). Briefly, mucosal homogenates were prepared and incubated with 0.6 M lactose (Thermo Fisher Scientific) or 0.3 M sucrose (Thermo Fisher Scientific) buffer (dissolved in 0.0625 M maleate buffer) for 60 min at 37°C. The reaction was stopped by the addition of 2.0% zinc sulfate (Thermo Fisher Scientific) and 1.8% barium hydroxide (Thermo Fisher Scientific), and the amount of glucose released was detected via a glucose oxidase reagent (Thermo Scientific). Enzyme activity was in μmoles glucose per minute per gram of protein. Protein was measured in intestinal homogenates via the Bradford method (Bio-Rad).

VFA concentrations

SCFA (acetate, butyrate, propionate) and branched-chain fatty acid (isobutyrate, valerate, and isovalerate) concentrations were determined in ascending colon (AC) contents and feces by GC as described previously (35), with data expressed as μmol/g of dry matter.

Microbiota analysis

DNA extraction

DNA was extracted from AC contents and feces by the QIAamp DNA Stool Mini Kit (Qiagen) in combination with the FastPrep-24 System (MP Biomedicals) as previously described (35). DNA concentration was determined with a NanoDrop 1000 spectrophotometer (NanoDrop Technologies).

Real-time qPCR

Bacterial DNA was analyzed for total bacteria, Enterobacteriaceae, Bifidobacterium spp., Lactobacillus spp., Prevotella spp., Roseburia spp./Eubacterium rectale, Clostridium perfringens, Clostridium difficile, Faecalibacterium prausnitzii, and Bacteroides fragilis with the use of primer/probe sequences indicated in Supplemental Table 2. Real-time qPCR was performed in the 7900HT Fast Real-Time PCR System (Thermo Fisher Scientific) and used TaqMan (for B. fragilis) or SYBR Green (for other bacterial groups/species) assays (21). After amplification, a dissociation step was included to analyze the melting profile of the amplified products. Standard curves (4–108 16S rRNA gene copies per reaction) were generated with the use of purified pCR 4 TOPO-TA plasmids (Thermo Fisher Scientific) containing the 16S rRNA genes of Eubacterium hallii 27751 (for total bacteria), Escherichia coli 11775 (Enterobacteriaceae), B. longum 15707 (Bifidobacterium spp.), Lactobacillus rhamnosus 53103 (Lactobacillus spp.), Prevotella ruminicola 19189 (Prevotella spp.), Eubacterium rectale 33656 (Roseburia spp./E. rectale), B. fragilis 25285 (B. fragilis), C. perfringens 13124 (C. perfringens), C. difficile 9689 (C. difficile), and F. prausnitzii 27766 (F. prausnitzii). Data analysis was processed with SDS version 2.3 software supplied by Thermo Fisher Scientific and data are presented as number of 16S rRNA gene copies per gram of AC contents or feces.

PCR amplification and sequencing of 16S rRNA genes

PCR amplification and sequencing of 16S rRNA genes were performed at the High-Throughput Sequencing and Genotyping Unit, University of Illinois. The V3-V4 regions of bacterial 16S rRNA genes were amplified through the use of primers V3f (5′-CCTACGGGAGGCAGCAG-3′) and V4r (5′-GGACTACHVGGGTWTCTAAT-3′) (36) with appropriate linkers and sample barcodes. PCR was performed on a Fluidigm Biomark HD platform (Fluidigm Corporation) as previously described (37) with the temperature profiles: 50°C for 2 min (1 cycle), 70°C for 20 min (1 cycle), 9°C for 10 min (1 cycle), followed by 10 cycles at 95°C for 15 s, 55°C for 30 s, and 72°C for 1 min, 2 cycles at 95°C for 15 s, 80°C for 30 s, 60°C for 30 s, and 72°C for 1 min, 8 cycles at 95°C for 15 s, 55°C for 30 s, and 72°C for 1 min, 2 cycles at 95°C for 15 s, 80°C for 30 s, 60°C for 30 s, and 72°C for 1 min, 8 cycles at 95°C for 15 s, 55°C for 30 s, and 72°C for 1 min, and 5 cycles at 95°C for 15 s, 80°C for 30 s, 60°C for 30 s, and 72°C for 1 min. Amplicons from 48 separate reactions were pooled for each sample and DNA concentration was measured on a Qubit 3.0 Fluorometer (Thermo Fisher Scientific). The size of amplicons was confirmed by running the samples on a Fragment Analyzer (Advanced Analytical Technologies). The amplicons were then mixed in equimolar concentrations and run on a 2% agarose E-gel (Thermo Fisher Scientific). The band of expected size (c.460 bp) was excised from the gel and the DNA was extracted with a Qiagen Gel Extraction kit (Qiagen). The extracted DNA was run on an Agilent 2100 Bioanalyzer (Agilent Technologies) to confirm an appropriate profile and determine the average size. Sequencing was performed on an Illumina MiSeq flow cell for 251 cycles from each end of the fragments with a MiSeq Reagent Nano Kit version 2 (2 × 250 nt paired end reads; Illumina) (38).

Sequence processing

Demultiplexed forward (read 1) and reverse reads (read 2) obtained from the sequencing facility were processed with the use of the QIIME software package (39). The paired-end reads were merged with the use of SeqPrep with a minimum overlap of 30 bp. Merged reads were filtered and split into libraries at Phred quality score ≥25. Sequences were assigned to operational taxonomic units (OTUs) via UCLUST with a 97% pairwise identity threshold. The representative sequences from each OTU were picked and the chimera sequences were identified via Chimera Slayer. After removal of chimeras, the remaining sequences were aligned to the greengenes reference alignment (gg_13_8_otus/rep_set_aligned/85_otus.pynast.fasta) with the use of PyNAST (40). The alignment was then filtered to remove highly variable regions and columns comprised of only gaps with the use of a lane mask. The phylogenetic tree was constructed from filtered alignment via FastTree (41). The representative sequence of each OTU was assigned to different taxonomic levels with the use of the Ribosomal Database Project naïve Bayesian rRNA Classifier at 80% confidence level (42) and an OTU table was created. The OTU table was further filtered to remove nonaligned and chimeric OTUs and singletons. α-Diversity (observed OTUs, Chao1, Shannon, and Simpson reciprocal indexes) and β-diversity analysis was performed from the filtered OTU table after rarefying to an equal number of reads.

Microbiome statistical analyses

Differences in bacterial communities among diet groups were evaluated with principal coordinate analysis and distance-based redundancy analysis (db_RDA) through the use of UniFrac distance matrices. Principal coordinate analysis and db_RDA were performed on both unweighted and weighted UniFrac distances with the use of QIIME and the function capscale of the vegan package of R (43, 44), respectively. Univariate statistical analysis (qPCR, α-diversity, relative abundances of bacterial taxa) was performed via the PROC MIXED procedure of SAS version 9.4 with diet as a main effect and replicate as a random effect (SAS Institute Inc.). Tukey’s post hoc tests were applied for multiple comparisons. When the relative abundance data were not normally distributed, arcsine-square root transformation was applied. Real-time qPCR data were log10 transformed to achieve normal distribution. False discovery rate (FDR) was computed with the use of PROC MULTTEST of SAS. The original P value was considered significant when the calculated FDR was ≤0.2.

Statistical analyses

All data, with the exception of the microbiota data, were subjected to an ANOVA via the MIXED procedure of SAS version 9.4. Most outcomes were analyzed by a simple ANOVA, except for daily piglet BWs, which were analyzed with repeated-measures ANOVA. All statistical models included replicate as a random effect and the level of significance was set at P < 0.05, with trends accepted at 0.05 < P < 0.10. Data are reported as means ± SEMs.

Results

BW and intestinal length and weight

Average daily and total weight gain, calculated as the difference in weight between PND 32 and PND 1, were similar across all dietary treatments (). The average formula volume offered to the piglets in periods of 5–8 d did not differ among the treatment diets (Supplemental Table 3), and, as expected, increased significantly as BW increased. Dietary SL supplementation did not significantly affect intestinal length or weight ().

FIGURE 1

Body weights (A) of piglets fed formula containing various concentrations of sialyllactose from postnatal days 2 to 31; total weight gain (kilograms) (B) calculated as the difference in weights on postnatal days 2 and 31. Values are means (A) or means ± SEMs (B). NS, not significant at P < 0.05. CON, control; HIGH, high-SL diet; LOW, low-SL diet; MOD, moderate-SL diet.

TABLE 1

BW, and small intestinal length and weight presented as absolute value and normalized by BW, of piglets fed formula containing various concentrations of sialyllactose

		Intestinal length		Intestinal weight
Diet and SL concentration	Body weight, kg	cm	cm/kg BW	g	g/kg BW
CON, 0 mg/L	5.93 ± 0.44	858 ± 40.1	148 ± 6.9	248 ± 19.2	42 ± 1.8
LOW, 130 mg/L	5.82 ± 0.68	826 ± 41.9	145 ± 9.8	238 ± 26.4	41 ± 1.9
MOD, 380 mg/L	6.14 ± 0.68	861 ± 34.3	149 ± 9.7	240 ± 22.9	40 ± 1.9
HIGH, 760 mg/L	5.81 ± 0.72	823 ± 35.1	155 ± 16.1	210 ± 19.5	37 ± 1.5

Values are means ± SEMs. No significant differences were detected at P < 0.05 (ANOVA). BW, body weight; CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

Complete blood count and cell differential results are presented in . All values were within the reference ranges (45–48) and measures of RBCs and platelets did not differ among the dietary treatments. Differences were observed in the distribution of WBC types. The total number of lymphocytes (× 103/µL) was lower in MOD than in LOW and HIGH (P < 0.05), but did not differ significantly from CON. Piglets in MOD had a higher percentage of neutrophils than did piglets in CON, LOW, and HIGH (P < 0.05). Conversely, the percentage of WBCs as lymphocytes was significantly lower in MOD than in the other 3 treatment groups.

TABLE 2

Complete blood count and differential analysis measured in the serum of piglets fed formula containing various concentrations of sialyllactose

			Concentration of sialyllactose in formula
Cell type	Unit	Reference ranges2	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
RBCs	×106/µL	4.1–8.2	7.2 ± 0.32	7.4 ± 0.24	7.0 ± 0.23	7.7 ± 0.16
Hemoglobin	g/dL	4.3–13.3	13.2 ± 0.45	13.7 ± 0.43	13.1 ± 0.44	14.3 ± 0.30
Hematocrit	%	16–41	41.5 ± 1.5	43.0 ± 1.4	41.1 ± 1.3	45.0 ± 0.86
MCV	Fl	34.2–61.3	58.3 ± 0.90	57.8 ± 0.83	59.0 ± 0.52	58.5 ± 0.56
MCH	Pg	9.4–19.8	18.5 ± 0.29	18.4 ± 0.22	18.8 ± 0.14	18.6 ± 0.14
MCHC	g/dL	26.5–33.6	31.8 ± 0.19	31.8 ± 0.16	31.8 ± 0.22	31.7 ± 0.26
Platelet	×103/µL	192–832	523 ± 57.7	606 ± 56.6	601 ± 36.3	572 ± 33.9
MPV	Fl	6.5–12.7	10.3 ± 0.54	9.2 ± 0.35	8.7 ± 0.28	8.9 ± 0.37
WBCs	×103/µL	5.6–18.5	8.6 ± 0.35	9.8 ± 0.66	8.6 ± 0.82	9.5 ± 0.43
Neutrophils	%	10.8–70.6	30.3 ± 5.0ab	21.2 ± 6.8b	41.9 ± 5.1a	21.8 ± 5.3b
Lymphocytes	%	26.2–82.9	61.5 ± 4.1b	62.2 ± 4.5b	47.8 ± 2.4a	64.5 ± 2.5b
Band	%	0–4	0	0.3 ± 0.24	0	0
Monocytes	%	1.4–8.3	4.3 ± 0.68	5.2 ± 0.68	5.0 ± 0.57	5.8 ± 1.1
Eosinophils	%	0–1.9	1.16 ± 0.27	0.79 ± 0.31	0.7 ± 0.26	0.7 ± 0.31
Basophils	%	0–0.9	0.2 ± 0.14	0.81 ± 0.23	0.4 ± 0.20	0.3 ± 0.17
Neutrophil count	×103/µL	0.8–9.7	2.5 ± 0.38	2.2 ± 0.69	3.5 ± 0.53	2.2 ± 0.53
Lymphocyte count	×103/µL	2.7–12.8	5.4 ± 0.46ab	6.1 ± 0.61b	4.1 ± 0.47a	6.1 ± 0.33b
Monocyte count	×103/µL	0.1–1.1	0.36 ± 0.06	0.5 ± 0.06	0.4 ± 0.08	0.6 ± 0.1
Eosinophil count	×103/µL	0–0.2	0.1 ± 0.02	0.09 ± 0.04	0.04 ± 0.02	0.07 ± 0.03
Basophil count	×103/µL	0–0.13	0.02 ± 0.01	0.08 ± 0.02	0.13 ± 0.1	0.02 ± 0.02

Values are means ± SEMs. Different letter superscripts indicate significant differences at P < 0.05 (ANOVA). CON, control; HIGH, high SL concentration; LOW, low SL concentration; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; MOD, moderate SL concentration; MPV, mean platelet volume; WBC, white blood cell.

From Ventrella et al. (45).

PT values did not differ among the treatment groups (). aPTT values of piglets in the HIGH group were 8–10% lower (P < 0.05) than in CON, LOW, and MOD. Serum minerals, electrolytes, proteins, metabolites, enzymes, biomarkers of liver and kidney function, and acid:base status were similar among the treatment groups ( and ), with the exception of glutamate dehydrogenase (GLDH, Table 5). Compared with piglets in the MOD group, GLDH values were 40% and 70% higher (P < 0.05) in CON and LOW piglets, respectively, whereas GLDH values in the HIGH group were not significantly different from the other dietary groups.

TABLE 3

Partial PT and aPTT measured in the serum of piglets fed formula containing various concentrations of sialyllactose

		Concentration of sialyllactose in formula
	Reference	CON	LOW	MOD	HIGH
Variable	ranges2	(0 mg/L)	(130 mg/L)	(380 mg/L)	(760 mg/L)
PT, s	9.3–13.3	13.7 ± 0.2	13.8 ± 0.2	13.9 ± 0.1	13.4 ± 0.4
aPTT, s	12.3–17.8	14.1 ± 0.2a	13.8 ± 0.4a	13.6 ± 0.2a	12.6 ± 0.5b

Values are means ± SEMs. Different letter superscripts indicate significant differences at P < 0.05 (ANOVA). aPTT, activated partial thromboplastin time; CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration; PT, prothrombin time.

From Rispat et al. (46).

TABLE 4

Serum minerals, electrolytes, metabolites, and proteins measured in the serum of piglets fed formula containing various concentrations of sialyllactose

			Concentration of sialyllactose in formula
Variable	Units	Reference ranges2	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
Minerals
Calcium	mg/dL	7.1–11.6	9.6 ± 0.2	9.8 ± 0.3	10.0 ± 0.2	9.9 ± 0.3
Phosphorous	mg/dL	5.3–9.6	10.7 ± 0.3	11.2 ± 0.9	10.7 ± 0.5	11.2 ± 0.7
Magnesium	mg/dL	2.7–3.7	2.3 ± 0.1	2.5 ± 0.2	2.5 ± 0.1	2.6 ± 0.2
Electrolytes
Sodium	mmol/L	135–150	136 ± 0.7	135 ± 0.8	136 ± 0.6	136 ± 0.6
Potassium	mmol/L	4.4–6.7	6.3 ± 0.6	6.3 ± 1.9	6.0 ± 0.5	5.6 ± 0.4
Sodium:potassium			23 ± 1.7	23 ± 5.8	24 ± 1.7	25 ± 1.6
Chloride	mmol/L	94–106	102 ± 1.0	102 ± 1.1	102 ± 0.9	107 ± 1.2
Metabolites
Glucose	mg/dL	85–150	137 ± 14.8	131 ± 15.6	141 ± 14.5	177 ± 31.5
Cholesterol total	mg/dL	78–116	77 ± 5.9	78 ± 5.6	80 ± 4.1	88 ± 4.4
TGs	mg/dL	29–80	51 ± 7.4	56 ± 6.1	54 ± 7.6	52 ± 6.7
Protein
Total protein	g/dL	3.7–4.8	3.8 ± 0.08	3.9 ± 0.07	3.9 ± 0.1	4.0 ± 0.1
Albumin	g/dL	1.9–3.9	2.8 ± 0.07	2.7 ± 0.05	2.8 ± 0.09	2.7 ± 0.08
Globulin	g/dL	1.9–3.9	1.0 ± 0.06	1.2 ± 0.06	1.2 ± 0.07	1.3 ± 0.09
Albumin:globulin			2.8 ± 0.2	2.3 ± 0.1	2.5 ± 0.2	2.2 ± 0.2

Values are means ± SEMs. No significant differences were detected at P < 0.05 (ANOVA). CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

From Ventrella et al. (45) and Merck Veterinary Manual (47).

TABLE 5

Enzymes, indicators of renal and liver function, and acid:base balance measured in the serum of piglets fed formula containing various concentrations of sialyllactose

			Concentration of sialyllactose in formula
Variable	Units	Ref ranges2	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
Enzymes
ALP	U/L	110–1292	396 ± 28.2	380 ± 55.8	351 ± 21.1	332 ± 23.9
AST	U/L	13–65	43 ± 6.7	40 ± 7.2	43 ± 3.6	33 ± 5.0
GGT	U/L	10–60	36 ± 5.8	30 ± 4.1	29 ± 1.6	34 ± 4.8
CPK	U/L	153–5427	1474 ± 407	802 ± 212	1194 ± 332	699 ± 126
GLDH	U/L		0.9 ± 0.1b	1.3 ± 0.1a	1.5 ± 0.2a	1.2 ± 0.08ab
Kidney function
Creatinine	mg/dL	0.51–1.39	0.8 ± 0.06	0.7 ± 0.03	0.7 ± 0.03	0.8 ± 0.06
Blood urea nitrogen	mg/dL	4–39	9.6 ± 1.3	9.7 ± 1.0	8.7 ± 0.8	8.6 ± 0.7
Liver function
Total bilirubin	mg/dL	0–10	0.5 ± 0.08	0.3 ± 0.04	0.5 ± 0.1	0.3 ± 0.03
Acid:base status
Bicarbonate	mmol/L		26 ± 1.1	27 ± 1.1	27 ± 1.1	26 ± 0.7
Anion gap3			13.8 ± 0.8	12.9 ± 1.0	12.9 ± 0.5	13.6 ± 0.8

Values are means ± SEMs. Different superscript letters indicate significant differences at P < 0.05 (ANOVA). ALP, alkaline phosphatase; AST, aspartate transaminase; CON, control; CPK, creatine phospholinase; GGT, γ-glutamyltransferase; GLDH, glutamate dehydrogenase; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

From Ventrella et al. (45); Merck Veterinary Manual (47); Cooper et al. (48).

(Sodium + potassium) – (chloride + CO2) = anion gap.

Assessment of intestinal development

Intestinal morphology and the number of goblet cells staining positive for sulfated-mucin did not differ among the dietary groups (Supplemental Tables4 and 5). The pattern of enzyme activity in the small intestine was similar to previously published findings (49), with lactase activity being highest in the jejunum and lowest in the ileum and sucrase activity being highest in the jejunum and lowest in the duodenum. However, disaccharidase activity did not differ among the dietary treatment groups ().

TABLE 6

Lactase and sucrase activity in the small intestine of piglets fed formula containing various concentrations of sialyllactose

	Concentration of sialyllactose in formula
Disaccharidase	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
Lactase, μmol · min−1 · g protein−1
Duodenum	107 ± 16	115 ± 29	125 ± 27	115 ± 18
Jejunum	444 ± 88	291 ± 55	505 ± 102	568 ± 140
Ileum	25 ± 5.8	27 ± 6.3	32 ± 9.4	21 ± 5.2
Sucrase, μmol · min−1 · g protein−1
Duodenum	4.55 ± 0.74	5.15 ± 3.4	8.44 ± 2.0	9.25 ± 2.4
Jejunum	117 ± 20	91 ± 10	113 ± 16	109 ± 8.3
Ileum	21 ± 3.7	15 ± 2.0	14 ± 1.9	19 ± 3.1

Values are means ± SEMs. No significant differences were detected at P < 0.05 (ANOVA). CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

VFAs and analysis of intestinal contents

The concentrations of VFAs in AC contents and fecal samples did not differ between the CON and SL-supplemented groups (). Similarly, pH and dry matter of the AC contents and feces were not significantly affected by dietary SL supplementation ().

TABLE 7

SCFA and branched-chain fatty acid concentrations (μmole per gram of dry matter) in ascending colon contents and feces of piglets fed formula containing various concentrations of sialyllactose

	Concentration of sialyllactose in formula
	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
Ascending colon
Acetate	200 ± 53	201 ± 53	170 ± 39	151 ± 37
Propionate	64 ± 19	58 ± 17	49 ± 11	42 ± 12
Butyrate	20 ± 4.7	20 ± 5.9	16 ± 4.1	17 ± 5.1
Isobutyrate	3.6 ± 1.0	2.9 ± 0.5	2.3 ± 0.4	2.2 ± 0.3
Isovalerate	4.7 ± 1.5	3.7 ± 0.8	2.7 ± 0.4	3.0 ± 0.5
Valerate	3.8 ± 0.9	3.9 ± 0.9	2.9 ± 0.6	3.0 ± 0.8
Feces
Acetate	55 ± 5.6	67 ± 10	53 ± 8.2	69 ± 17
Propionate	17 ± 2.3	24 ± 4.2	18 ± 3.7	22 ± 6.1
Butyrate	6.8 ± 1.0	8.4 ± 1.6	5.3 ± 0.7	8.2 ± 2.5
Isobutyrate	1.6 ± 0.3	2.2 ± 0.5	1.6 ± 0.3	2.0 ± 0.5
Isovalerate	2.2 ± 0.4	2.8 ± 0.6	2.3 ± 0.4	2.5 ± 0.6
Valerate	1.6 ± 0.3	1.8 ± 0.4	1.4 ± 0.3	1.7 ± 0.5

Values are means ± SEMs. No significant differences were detected at P < 0.05 (ANOVA). CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

FIGURE 2

pH and dry matter (percentage dry matter) of AC contents and feces of piglets fed formula containing various concentrations of sialyllactose. Values are means ± SEMs. pH was measured immediately after removal of contents from the large intestine. Dry matter was calculated as the weight difference before and after dehydration at 80°C for 16 h. NS, not significant at P < 0.05. AC, ascending colon; CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

Real-time qPCR

The densities of total bacteria, Enterobacteriaceae, Bifidobacterium spp., Lactobacillus spp., Prevotella spp., Roseburia spp., E.rectale, C.difficile, F.prausnitzii, and B.fragilis, as determined by qPCR, were similar among the treatment groups in both AC contents and feces (). In the AC contents, C. perfringens density in MOD was significantly higher than in CON and HIGH, with no differences among CON, LOW, and HIGH. However, differences among the groups disappeared after correction for false discovery (FDR = 0.286). In feces, C. perfringens density was greatest in MOD, intermediate in CON and HIGH, and lowest in LOW, and differences were not affected by FDR correction.

TABLE 8

Bacterial densities (Log10 16S rRNA gene copies per gram) in the ascending colon contents and feces of piglets fed formula containing various concentrations of sialyllactose

	Concentration of sialyllactose in formula
Bacteria	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (780 mg/L)
Ascending colon
Total bacteria	12.3 ± 0.15	12.4 ± 0.09	12.2 ± 0.06	12.3 ± 0.14
Enterobacteriaceae	10.4 ± 0.22	10.8 ± 0.14	10.5 ± 0.11	10.4 ± 0.19
Bifidobacterium spp.	6.70 ± 0.42	6.66 ± 0.43	6.87 ± 0.45	6.02 ± 0.09
Lactobacillus spp.	10.6 ± 0.17	10.7 ± 0.22	10.6 ± 0.15	10.6 ± 0.25
Prevotella spp.	11.6 ± 0.25	11.7 ± 0.12	11.7 ± 0.10	11.7 ± 0.19
Roseburia spp./E. rectale	6.88 ± 0.32	7.08 ± 0.44	6.42 ± 0.45	6.65 ± 0.18
B. fragilis	9.02 ± 0.55	9.46 ± 0.23	9.63 ± 0.25	9.58 ± 0.33
C. perfringens	6.55 ± 0.43ab	6.14 ± 0.46a	7.33 ± 0.54b	6.18 ± 0.39ab
C. difficile	5.23 ± 0.22	5.14 ± 0.26	4.79 ± 0.18	5.23 ± 0.28
F. prausnitzii	9.30 ± 0.16	9.15 ± 0.29	9.00 ± 0.10	9.11 ± 0.15
Feces
Total bacteria	12.0 ± 0.16	12.1 ± 0.14	11.9 ± 0.19	11.9 ± 0.26
Enterobacteriaceae	10.2 ± 0.24	10.4 ± 0.17	10.0 ± 0.25	9.85 ± 0.43
Bifidobacterium spp.	6.48 ± 0.37	6.45 ± 0.39	6.58 ± 0.51	5.93 ± 0.23
Lactobacillus spp.	10.3 ± 0.24	10.5 ± 0.17	10.1 ± 0.33	9.96 ± 0.29
Prevotella spp.	11.2 ± 0.23	11.4 ± 0.19	10.9 ± 0.28	11.1 ± 0.34
Roseburia spp./E. rectale	7.18 ± 0.30	7.16 ± 0.52	6.72 ± 0.27	6.61 ± 0.23
B. fragilis	8.67 ± 0.56	9.14 ± 0.30	8.97 ± 0.47	8.96 ± 0.43
C. perfringens	6.43 ± 0.33b	6.49 ± 0.45ab	7.54 ± 0.58a	5.83 ± 0.21b
C. difficile	5.26 ± 0.18	4.90 ± 0.19	4.78 ± 0.15	4.97 ± 0.22
F. prausnitzii	9.21 ± 0.15	9.05 ± 0.20	8.96 ± 0.19	8.88 ± 0.20

Values are means ± SEMs. Labeled means in a row without a common superscript letter differ at P ≤ 0.05 (ANOVA).

16S rRNA gene sequencing

Effect of diet on the microbiota structure and relative abundances

After the assembling of paired-end reads and performing the quality control depletions as described, 7,221,345 sequences with a mean of 91,409 sequences/sample were used for further analysis. db_RDA of both unweighted and weighted UniFrac distance matrices revealed that the overall bacterial communities were similar among treatment groups in both AC contents and feces (P > 0.05; ). Similar results were also obtained by permutational multivariate ANOVA (data not shown) with the use of the vegan function Adonis of R. Statistical analysis of the relative abundances of bacterial phyla and genera showed that several bacterial taxa were modulated by the presence of SL in the diet. Proportions of Proteobacteria were highest in the AC contents of piglets fed LOW diet, intermediate in CON and HIGH, and lowest in MOD (; P < 0.05). At the genus level, relative abundances of Alistipes in AC contents were highest in piglets fed HIGH diet, lowest in CON, and intermediate in LOW and MOD, whereas Bilophila were highest in LOW and lowest in CON, with diets MOD and HIGH being intermediate (). In feces, Butyricicoccus was highest in diet HIGH, intermediate in diets LOW and MOD, and lowest in diet CON (). After correction for FDR, no differences in relative abundances of bacterial phyla or genera among the diets persisted (FDR > 0.20).

FIGURE 3

Principal coordinate analysis based on unweighted and weighted UniFrac distances generated from ascending colon contents (A and B) and feces (C and D) of piglets fed formula containing various concentrations of sialyllactose. P values were obtained by distance-based redundancy analysis. The percentages (in parenthesis) on each axis express the contributions of variables in accounting for the variability in a given PC. CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration; PC, principal component.

FIGURE 4

Heat maps representing relative abundances of 15 bacterial phyla in ascending colon contents (A) and feces (B) of piglets fed formula containing 0 mg/L (CON), 130 mg/L (LOW), 380 mg/L (MOD), or 760 mg/L (HIGH) SL. Relative abundances are presented on a progressive color scale with dark and light color depicting the lowest and highest abundances, respectively. *Significant differences at P ≤ 0.05 (ANOVA). CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

TABLE 9

Relative abundances of bacterial genera detected in ascending colon contents of piglets fed formula containing various concentrations of sialyllactose

			Concentration of sialyllactose in formula
Genus	Family	Phylum	CON (0 mg/L) (n = 11)	LOW (130 mg/L) (n = 9)	MOD (380 mg/L) (n = 10)	HIGH (760 mg/L) (n = 8)	P value
Bifidobacterium	Bifidobacteriaceae	Actinobacteria	0.06 ± 0.05	0.08 ± 0.08	0.06 ± 0.05	0 ± 0	0.7566
Eggerthella	Coriobacteriaceae	Actinobacteria	0.05 ± 0.01	0.06 ± 0.01	0.06 ± 0.01	0.09 ± 0.02	0.1379
Butyricimonas	[Odoribacteraceae]	Bacteroidetes	0.85 ± 0.17	0.81 ± 0.18	0.84 ± 0.19	0.52 ± 0.15	0.5176
Odoribacter	[Odoribacteraceae]	Bacteroidetes	0.09 ± 0.03	0.32 ± 0.25	0.10 ± 0.04	0.10 ± 0.03	0.5151
[Prevotella]	[Paraprevotellaceae]	Bacteroidetes	7.66 ± 2.36	5.08 ± 1.99	4.40 ± 1.03	5.55 ± 1.35	0.4089
Bacteroides	Bacteroidaceae	Bacteroidetes	20.0 ± 5.25	18.2 ± 3.54	24.1 ± 4.01	22.4 ± 3.66	0.2738
Parabacteroides	Porphyromonadaceae	Bacteroidetes	12.6 ± 3.00	13.2 ± 3.90	10.9 ± 2.52	9.36 ± 2.34	0.7977
Prevotella	Prevotellaceae	Bacteroidetes	0.72 ± 0.45	2.50 ± 1.89	0.82 ± 0.35	0.16 ± 0.08	0.2382
Alistipes	Rikenellaceae	Bacteroidetes	0.13 ± 0.05a	0.18 ± 0.09ab	0.29 ± 0.08ab	0.55 ± 0.16b	0.0341
Christensenella	Christensenellaceae	Firmicutes	0.11 ± 0.03	0.06 ± 0.02	0.10 ± 0.03	0.08 ± 0.02	0.3831
Clostridium	Clostridiaceae	Firmicutes	0.48 ± 0.15	0.20 ± 0.08	0.59 ± 0.32	0.23 ± 0.13	0.4713
SMB53	Clostridiaceae	Firmicutes	0.04 ± 0.01	0.07 ± 0.02	0.04 ± 0.01	0.05 ± 0.02	0.5219
RFN20	Erysipelotrichaceae	Firmicutes	1.02 ± 1.02	0.52 ± 0.52	1.43 ± 1.41	0.03 ± 0.02	0.7891
[Eubacterium]	Erysipelotrichaceae	Firmicutes	0.16 ± 0.12	0.20 ± 0.08	0.10 ± 0.04	0.28 ± 0.07	0.2251
p-75-a5	Erysipelotrichaceae	Firmicutes	0.26 ± 0.16	0.07 ± 0.06	0.18 ± 0.11	0.01 ± 0.01	0.1701
Blautia	Lachnospiraceae	Firmicutes	5.96 ± 1.61	3.05 ± 0.86	5.15 ± 1.77	8.16 ± 2.66	0.2423
[Ruminococcus]	Lachnospiraceae	Firmicutes	1.14 ± 0.23	1.62 ± 0.42	1.50 ± 0.55	1.52 ± 0.57	0.8828
Dorea	Lachnospiraceae	Firmicutes	0.87 ± 0.41	0.55 ± 0.29	1.03 ± 0.41	0.47 ± 0.15	0.5612
Coprococcus	Lachnospiraceae	Firmicutes	0.26 ± 0.05	0.20 ± 0.06	0.27 ± 0.07	0.32 ± 0.09	0.6449
Clostridium	Lachnospiraceae	Firmicutes	0.07 ± 0.02	0.17 ± 0.05	0.10 ± 0.03	0.14 ± 0.07	0.2421
Lactobacillus	Lactobacillaceae	Firmicutes	6.23 ± 4.27	4.61 ± 2.01	2.46 ± 0.85	2.88 ± 1.33	0.5687
Leuconostoc	Leuconostocaceae	Firmicutes	0.35 ± 0.13	1.24 ± 0.79	1.39 ± 0.96	1.17 ± 0.45	0.5284
Oscillospira	Ruminococcaceae	Firmicutes	1.16 ± 0.22	1.57 ± 0.39	1.20 ± 0.23	3.32 ± 1.37	0.1313
Ruminococcus	Ruminococcaceae	Firmicutes	1.34 ± 0.36	0.77 ± 0.23	0.98 ± 0.19	1.19 ± 0.30	0.3378
Anaerotruncus	Ruminococcaceae	Firmicutes	0.22 ± 0.09	0.51 ± 0.17	0.24 ± 0.08	0.61 ± 0.38	0.2135
Faecalibacterium	Ruminococcaceae	Firmicutes	0.68 ± 0.30	0.29 ± 0.17	0.09 ± 0.03	0.13 ± 0.07	0.0741
Butyricicoccus	Ruminococcaceae	Firmicutes	0.03 ± 0.01	0.04 ± 0.02	0.09 ± 0.04	0.06 ± 0.01	0.2356
Lactococcus	Streptococcaceae	Firmicutes	0.44 ± 0.15	0.34 ± 0.13	0.33 ± 0.16	0.27 ± 0.18	0.7112
Streptococcus	Streptococcaceae	Firmicutes	0.04 ± 0.01	0.02 ± 0.01	0.14 ± 0.08	0.02 ± 0	0.0996
Turicibacter	Turicibacteraceae	Firmicutes	0.13 ± 0.04	0.07 ± 0.04	0.10 ± 0.03	0.20 ± 0.12	0.5348
Phascolarctobacterium	Veillonellaceae	Firmicutes	0.57 ± 0.45	1.29 ± 1.14	0.32 ± 0.19	0.53 ± 0.31	0.6845
Anaerovibrio	Veillonellaceae	Firmicutes	0 ± 0	0.15 ± 0.13	0.66 ± 0.64	0.01 ± 0.01	0.4252
Acidaminococcus	Veillonellaceae	Firmicutes	0.08 ± 0.05	0.06 ± 0.05	0.33 ± 0.23	0 ± 0	0.5096
Fusobacterium	Fusobacteriaceae	Fusobacteria	0.37 ± 0.24	0.48 ± 0.29	0.16 ± 0.09	0.13 ± 0.13	0.5627
Sutterella	Alcaligenaceae	Proteobacteria	0.08 ± 0.03	0.18 ± 0.10	0.05 ± 0.02	0.12 ± 0.05	0.3538
Desulfovibrio	Desulfovibrionaceae	Proteobacteria	1.13 ± 0.29	0.89 ± 0.33	1.04 ± 0.30	1.56 ± 0.54	0.4517
Bilophila	Desulfovibrionaceae	Proteobacteria	0.12 ± 0.02b	0.43 ± 0.17a	0.22 ± 0.05ab	0.21 ± 0.06ab	0.0287
Escherichia	Enterobacteriaceae	Proteobacteria	4.25 ± 1.28	12.6 ± 4.45	3.29 ± 0.50	5.40 ± 2.21	0.0538
Citrobacter	Enterobacteriaceae	Proteobacteria	0.10 ± 0.03	0.06 ± 0.02	0.09 ± 0.03	0.03 ± 0.02	0.0613
Pasteurella	Pasteurellaceae	Proteobacteria	0.49 ± 0.43	0.02 ± 0.01	0.02 ± 0.01	0.01 ± 0.01	0.3409
Synergistes	Synergistaceae	Synergistetes	1.21 ± 0.82	0.43 ± 0.13	1.17 ± 0.80	1.20 ± 0.56	0.8587
Akkermansia	Verrucomicrobiaceae	Verrucomicrobia	1.70 ± 1.20	1.67 ± 1.19	2.92 ± 2.06	0.01 ± 0.01	0.4214

Values are means ± SEMs. Labeled means in a row without a common superscript letter differ at P ≤ 0.05 (ANOVA). Only the genera with mean relative abundances ≥0.05% were analyzed. Taxa that have brackets around the names are proposed taxonomies that greengenes recommend based on the whole genome phylogeny, but not officially recognized by Bergey's manual of systematic bacteriology. CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

TABLE 10

Relative abundances of bacterial genera detected in feces of piglets fed formula containing various concentrations of sialyllactose

			Concentration of sialyllactose in formula
			CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
Genus	Family	Phylum	(n  =  11)	(n  =  9)	(n  =  10)	(n  =  8)	P values
Bifidobacterium	Bifidobacteriaceae	Actinobacteria	0.08 ± 0.07	0.02 ± 0.02	0.24 ± 0.16	0 ± 0	0.2710
Eggerthella	Coriobacteriaceae	Actinobacteria	0.37 ± 0.14	0.16 ± 0.06	0.18 ± 0.04	0.24 ± 0.04	0.3496
Butyricimonas	[Odoribacteraceae]	Bacteroidetes	0.69 ± 0.13	0.85 ± 0.24	0.76 ± 0.26	0.90 ± 0.31	0.9630
Odoribacter	[Odoribacteraceae]	Bacteroidetes	0.39 ± 0.23	0.23 ± 0.13	0.21 ± 0.14	0.14 ± 0.05	0.8768
[Prevotella]	[Paraprevotellaceae]	Bacteroidetes	4.10 ± 1.41b	2.01 ± 1.05a,b	0.62 ± 0.20a	2.13 ± 0.72a,b	0.0455
Bacteroides	Bacteroidaceae	Bacteroidetes	16.1 ± 3.64	21.7 ± 4.10	20.6 ± 4.68	24.1 ± 5.08	0.0541
Parabacteroides	Porphyromonadaceae	Bacteroidetes	10.5 ± 3.03	12.5 ± 3.83	9.11 ± 1.66	9.15 ± 2.82	0.7478
Prevotella	Prevotellaceae	Bacteroidetes	0.52 ± 0.22	0.75 ± 0.33	0.46 ± 0.28	0.33 ± 0.18	0.6304
Alistipes	Rikenellaceae	Bacteroidetes	0.64 ± 0.25	0.43 ± 0.16	0.76 ± 0.20	0.94 ± 0.29	0.2557
Christensenella	Christensenellaceae	Firmicutes	0.35 ± 0.09	0.21 ± 0.05	0.35 ± 0.07	0.39 ± 0.25	0.4295
Clostridium	Clostridiaceae	Firmicutes	0.87 ± 0.25	1.09 ± 0.57	2.05 ± 1.17	1.03 ± 0.63	0.4564
SMB53	Clostridiaceae	Firmicutes	0.07 ± 0.03	0.06 ± 0.02	0.05 ± 0.02	0.05 ± 0.01	0.9838
[Eubacterium]	Erysipelotrichaceae	Firmicutes	0.29 ± 0.16a,b	0.34 ± 0.11a,b	0.13 ± 0.05a	0.70 ± 0.26b	0.0236
p-75-a5	Erysipelotrichaceae	Firmicutes	2.45 ± 2.14	0.11 ± 0.07	2.08 ± 1.99	0.01 ± 0.01	0.2833
[Ruminococcus]	Lachnospiraceae	Firmicutes	1.46 ± 0.33	2.16 ± 0.54	2.42 ± 0.83	1.39 ± 0.53	0.3824
Blautia	Lachnospiraceae	Firmicutes	4.90 ± 0.94	4.22 ± 1.44	4.32 ± 1.46	7.40 ± 3.08	0.5165
Clostridium	Lachnospiraceae	Firmicutes	0.28 ± 0.14	0.31 ± 0.12	0.23 ± 0.05	0.35 ± 0.09	0.7577
Coprococcus	Lachnospiraceae	Firmicutes	0.26 ± 0.08	0.28 ± 0.08	0.23 ± 0.07	0.21 ± 0.05	0.9057
Dorea	Lachnospiraceae	Firmicutes	0.80 ± 0.34	0.67 ± 0.39	0.85 ± 0.22	0.44 ± 0.13	0.7124
Lactobacillus	Lactobacillaceae	Firmicutes	1.87 ± 0.61	2.32 ± 0.81	2.12 ± 1.05	1.12 ± 0.42	0.6332
Leuconostoc	Leuconostocaceae	Firmicutes	1.33 ± 0.70	1.15 ± 0.42	0.74 ± 0.40	0.70 ± 0.26	0.7906
Anaerotruncus	Ruminococcaceae	Firmicutes	0.32 ± 0.09	0.68 ± 0.16	0.61 ± 0.18	0.65 ± 0.15	0.0554
Butyricicoccus	Ruminococcaceae	Firmicutes	0.03 ± 0.01a	0.05 ± 0.02a,b	0.04 ± 0.01a,b	0.15 ± 0.07b	0.0239
Faecalibacterium	Ruminococcaceae	Firmicutes	0.64 ± 0.28	0.16 ± 0.05	0.21 ± 0.15	0.23 ± 0.10	0.2503
Oscillospira	Ruminococcaceae	Firmicutes	2.99 ± 0.62	2.38 ± 0.33	3.42 ± 1.04	4.00 ± 1.45	0.6158
Ruminococcus	Ruminococcaceae	Firmicutes	2.39 ± 0.66	1.95 ± 0.37	2.15 ± 0.43	1.72 ± 0.39	0.4842
Lactococcus	Streptococcaceae	Firmicutes	1.13 ± 0.38	0.96 ± 0.63	0.61 ± 0.34	0.18 ± 0.08	0.0619
Streptococcus	Streptococcaceae	Firmicutes	0.07 ± 0.03	0.05 ± 0.03	0.04 ± 0.01	0.02 ± 0.01	0.5977
Turicibacter	Turicibacteraceae	Firmicutes	0.38 ± 0.11	0.26 ± 0.11	0.76 ± 0.50	0.60 ± 0.37	0.7676
Acidaminococcus	Veillonellaceae	Firmicutes	0.15 ± 0.12	0.18 ± 0.18	0.37 ± 0.28	0 ± 0	0.8525
Megasphaera	Veillonellaceae	Firmicutes	0.03 ± 0.02	0.04 ± 0.02	0.05 ± 0.03	0.11 ± 0.09	0.5973
Phascolarctobacterium	Veillonellaceae	Firmicutes	0.65 ± 0.42	0.63 ± 0.35	0.27 ± 0.13	0.52 ± 0.27	0.7991
Bilophila	Desulfovibrionaceae	Proteobacteria	0.31 ± 0.10	0.65 ± 0.17	0.48 ± 0.11	0.58 ± 0.23	0.2002
Desulfovibrio	Desulfovibrionaceae	Proteobacteria	0.91 ± 0.19	1.22 ± 0.38	1.27 ± 0.28	3.07 ± 1.76	0.2497
Escherichia	Enterobacteriaceae	Proteobacteria	4.33 ± 1.14	5.10 ± 1.00	3.22 ± 1.03	3.04 ± 0.98	0.3927
Synergistes	Synergistaceae	Synergistetes	2.35 ± 1.58	0.86 ± 0.23	2.44 ± 1.04	2.71 ± 0.95	0.5715
Akkermansia	Verrucomicrobiaceae	Verrucomicrobia	1.72 ± 1.25	3.13 ± 2.46	2.97 ± 2.03	0.03 ± 0.02	0.1049

Data are means ± SEMs. Labelled means in a row without a common superscript letter differ, P ≤ 0.05 (ANOVA). Only the genera with mean relative abundances ≥ 0.05% were analyzed. Taxa that have brackets around the names are proposed taxonomies that greengenes recommend based on the whole genome phylogeny, but not officially recognized by Bergey's manual of systematic bacteriology. CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration.

Effect of diet on the α-diversity

α-Diversity was calculated with the use of QIIME after removing OTUs containing only 1 sequence and rarefying to an equal number of reads (34,500) for all samples. No significant differences were detected on observed OTUs, Chao1, Shannon, and Simpson reciprocal indexes among dietary treatment groups ().

TABLE 11

Diversity measures obtained from the microbiota in ascending colon contents and feces of piglets fed formula containing various concentrations of sialyllactose

	Concentration of sialyllactose in formula
	CON (0 mg/L)	LOW (130 mg/L)	MOD (380 mg/L)	HIGH (760 mg/L)
	(n = 11)	(n = 9)	(n = 10)	(n = 8)	P value
Ascending colon
Observed OTUs	3386 ± 257	3205 ± 230	3541 ± 170	3521 ± 264	0.6331
Shannon	7.1 ± 0.3	6.5 ± 0.4	7.0 ± 0.3	7.1 ± 0.4	0.3625
Simpson reciprocal	26.6 ± 4.7	16.9 ± 3.4	27.7 ± 6.3	26.4 ± 4.9	0.2931
Chao1	6368 ± 529	6313 ± 354	7061 ± 387	6667 ± 545	0.5228
Feces
Observed OTUs	3717 ± 277	3788 ± 126	3881 ± 313	3704 ± 134	0.8358
Shannon	7.4 ± 0.3	7.3 ± 0.2	7.4 ± 0.3	7.5 ± 0.2	0.9648
Simpson reciprocal	36.3 ± 7.2	27.1 ± 4.3	34.1 ± 5.6	29.5 ± 5.6	0.3494
Chao1	6934 ± 706	7408 ± 381	7572 ± 781	6677 ± 456	0.6267

Values are means ± SEMs. No significant differences were detected at P < 0.05 (ANOVA). CON, control; HIGH, high SL concentration; LOW, low SL concentration; MOD, moderate SL concentration; OTU, operational taxonomic unit.

Discussion

MOs are complex glycans that are resistant to digestion and exert a number of important physiologic functions including shaping the composition of the gut microbiota, influencing the proliferation and maturation of intestinal cells, regulating intestinal and systemic immunity, and modulating brain structural and cognitive development (4, 19–21, 50). MOs are either absent or found in very low concentration in infant formula (10). Currently, some commercially available infant formulas contain the prebiotics GOS and PDX, which have bifidogenic effects in the infant's intestine and influence early bacterial colonization, resulting in a microbiota that closely resembles that of breastfed infants (23, 24). With technologic advances in the production of MOs, the supplementation of infant formula with these compounds has become viable. Herein, we investigated the effects of a bovine-derived modified whey enriched with SL on the growth and gut development of neonatal pigs. The results of the study demonstrated that the addition of bovine-derived SL at concentrations similar to those present in human milk to a GOS- and PDX-enriched formula was well tolerated, supported normal growth, and did not adversely affect piglet development during the first 4 wk of life. SA-containing compounds have been described as essential for the neonatal period owing to its involvement in brain development and cognition, inhibition of pathogen binding, and immunity (10, 50). In addition, several studies have demonstrated the beneficial effects of SA-containing oligosaccharides on intestinal health. Oligosaccharides (including the acidic oligosaccharides 3′SL and 6′SL) isolated from human milk induced differentiation of transformed and nontransformed human intestinal cells that directly affect the absorptive and digestive capacity of the intestine (17). In our in vivo pig study, diets were formulated to contain SL concentrations above those endogenously present in bovine milk–based human infant formula and which were similar to SL concentrations reported for mature bovine and human milk (10). SL supplementation did not significantly affect intestinal growth, morphology, goblet cell numbers, disaccharidase activity, stool consistency, and cecal and colonic content pH.

Similarly, only a few minor hematologic and histologic differences were detected with the addition of SL to formula. Coagulation (clotting) time is routinely evaluated to monitor blood clotting functions and detect coagulation disturbances. One-stage PT and aPTT are indicators of the activation of the intrinsic and extrinsic pathways, respectively, and they differ in the initial factor triggered in the coagulation pathway. Values obtained for both PT and aPTT in our study were within the normal range for young pigs. Owing to the relative insensitivity of such tests and the remarkably small change in coagulation time in the HIGH group (<2 s), the aPTT difference between the groups may not be of clinical significance. In fact, a reduction of 30% in the activity of a single factor in the coagulation cascade is needed before a noticeably prolonged time is observed in the clotting process (51). Similarly, there was a slight, but significant, increase in the activity of GLDH in the MOD group. GLDH is a mitochondrial enzyme, and is indicative of hepatocellular injury, especially in large animals (52). The use of plasma enzyme activity as a measure of hepatic injury should be interpreted cautiously. Karlsson et al. (53) did not find a correlation between the serum levels of commonly assayed enzyme activities and pathologic liver changes in newborn pigs. In addition, a reference value for GLDH activity is missing in neonatal piglets, but it has been reported to be <10 U/L in small healthy animals (53) and to range between 0.1 and 3.2 U/L in >20-kg pigs (54). Nonetheless, GLDH activity observed in CON- and SL-fed piglets was within the ranges reported for older pigs. Furthermore, the absence of a dose-response effect on GLDH and lack of abnormal values of other enzymes suggest that the change in GLDH activity in MOD animals has little physiologic significance.

Modest changes in neutrophil and lymphocyte counts were observed in the MOD relative to the other groups. However, we did not observe a dose-response change in the number of neutrophils and lymphocytes and the values for both measures were within reference ranges (45), suggesting no immune impairment. Given the potential importance of the gut in educating cells of the immune system, more research in this area could be informative.

Although previous studies have reported microbial composition changes with the addition of higher doses of SL to the diet, we observed that the addition of SL-enriched bovine-derived modified whey to a milk replacer containing PDX + GOS had only minor effects on the gut microbiota relative to CON formula. The only statistically significant change was higher fecal density of C. perfringens from animals fed SL at the MOD (380 mg/L) SL concentration relative to CON and HIGH. Among the factors that may explain differences between the findings reported herein and previous reports are the presence of PDX and GOS in all dietary treatments and the lower SL dose used. The relevance of this finding is unclear, but it is worth noting that there was no dose response in C. perfringens abundance among the groups. Furthermore, the presence of C. perfringens in the pig intestine is not unusual. Bacteria from the Clostridium genus (Firmicutes phylum) are commonly found during the first 2 wk of life in the nursing piglet intestine, although its abundance decreases after weaning (55). C. perfringens is found in high abundance in the microbiota of sow-reared piglets; however, its presence was not correlated with enteritis (56, 57).

The addition of prebiotics such as PDX and GOS to formula has produced changes in microbial communities in human infants, rats, and piglets. Mika et al. (58) demonstrated an increase in Lactobacillus spp. abundance in rats fed diets supplemented with GOS and PDX at 7 g/kg diet for 4 wk. In piglets, the addition of PDX and GOS to formula (4 g/L at a 1:1 ratio) increased the number of Parabacteroides and Lactobacillus species in colonic contents relative to control formula (59). Infants fed formula with a prebiotic blend of PDX and GOS (4 g/L total at a 1:1 ratio) for 60 d had increased total Bifidobacterium counts and a gut microbial profile that more closely resembled that of breastfed infants (23). Thus, PDX and GOS at the concentration present in the diets may have resulted in similar gut bacterial communities among the groups.

Other prebiotics, including SA-containing oligosaccharides, have been tested in formula for their ability to modulate the gut microbiota. Jacobi et al. (16) observed significant changes in the colonic microbial ecology in neonatal piglets fed formula supplemented solely (no other prebiotics) with 6′SL, but not 3′SL, at 4 g/L. The most significant change observed was an increase in bacterial species that are capable of metabolizing 6′SL, namely Collinsella aerofaciens (phylum Actinobacteria), genera Ruminococcus and Faecalibacterium (phylum Firmicutes), and genus Prevotella (phylum Bacteroidetes), in 6′SL-fed piglets relative to controls (16). Tarr et al. (60) demonstrated that dietary supplementation with 5% 3′SL or 6′SL mitigated stress-induced changes in the gut microbial population. In human infants, the addition of 10 g/L bovine MO enriched in 3′SL and 6′SL resulted in a moderate change in the microbiota, which included lower Clostridia counts relative to formula without bovine MO (26). As described earlier, the concentration of SL used herein was significantly lower than in other studies, which may explain differences in the extent to which SL affected the gut microbiota. No effect of SL on VFA concentrations was observed. Previous studies have shown that although 6′SL is fermented by the human infant microbiome, it produced lower concentrations of SCFAs compared with GOS, LNnT, or 2′FL (61). Therefore, small effects of SL fermentation were likely not detectable owing to the presence of PDX and GOS in all diets.

Currently, most infant formulas include prebiotics, but the addition of MOs has been limited by the lack of availability and high cost of these compounds. However, a few clinical trials have been conducted to test the tolerance and safety of MO supplementation in formula. Marriage et al. (62) reported that the addition of GOS + 0.2 or 1.0 g/L of 2′FL (total amount of oligosaccharide was 2.4 g/L) for 4 mo was well tolerated, and that infant growth and 2′FL uptake were similar to those observed in exclusively breastfed infants (62). Similarly, a clinical study testing the safety of the addition of 2′FL (1.0 g/L) and LNnT (0.5 g/L) to a cow milk–based formula showed that HMO supplementation supported age-appropriate growth and was safe (6). Our study confirms that adding dietary SL at ≤760 mg/L to formula containing prebiotics is well tolerated by neonatal piglets. We demonstrated that it supported weight gain and did not have any adverse effect on the intestine or other organ weights and structure. The addition of SL to a prebiotic-containing diet did not affect overall microbiota structure and diversity in AC contents and feces, but had minor effects on the relative abundances of specific microbes. We hypothesize that the lack of distinct changes in the microbiota and VFAs can be attributed to the presence of prebiotics, specifically PDX and GOS, in the base formula. However, our group recently reported that the combination of prebiotics and SL affected the development of the prefrontal cortex and corpus callosum in these piglets (27), suggesting that SL-containing oligosaccharides have the potential to influence brain development without detrimentally affecting the development of other organ systems in the piglet.

Although the piglet is considered to be one of the best preclinical animal models for the human infant (28, 29), a limitation of the study is that the native microbiome of the piglet differs from that of the human infant, thus it is not possible to directly extrapolate microbiome findings to infants. Therefore, future studies with humanized gnotobiotic piglets could better inform SL-induced changes in microbiome structure and function. Another limitation is the use of a complex SL-enriched whey fraction. As pure forms of 3′SL and 6′SL become available, future research should seek to investigate their unique roles in neonatal development. Lastly, the presence of PDX and GOS in the formulas likely masked potential SL-specific outcomes on microbiome composition or VFA concentrations; however, the goal of this experiment was to investigate the impact of SL in the presence of prebiotics present in commercially available infant formula.

Click here for additional data file.