Effect of dietary medium-chain fatty acids on nursery pig growth performance, fecal microbial composition, and mitigation properties against porcine epidemic diarrhea virus following storage

Abstract

An experiment was conducted to evaluate the effect of dietary medium-chain fatty acid (MCFA) addition on nursery pig growth performance, fecal microbial composition, and mitigation of porcine epidemic diarrhea virus (PEDV) following storage. A total of 360 pigs (DNA 400 × 200, Columbus, NE; initially 6.7 ± 0.07 kg) were randomized to pens (5 pigs per pen) on the day of weaning (approximately 20 d of age), allowed a 6-d acclimation, blocked by BW, and randomized to dietary treatment (9 pens per treatment). All MCFA (Sigma–Aldrich, St. Louis, MO) were guaranteed ≥98% purity, including hexanoic (C6:0), octanoic (C8:0), and decanoic (C10:0) acids. Treatment diets were formulated in 2 phases (7 to 11 and 11 to 23 kg BW) and formulated to meet or exceed NRC requirement estimates. Treatments (n = 8) were a dose response including 0%, 0.25%, 0.5%, 1.0%, and 1.5% added MCFA blend (1:1:1 ratio C6:0, C8:0, and C10:0), as well as treatments with individual additions of 0.5% C6:0, C8:0, or C10:0. Fecal samples were collected from pigs fed control and 1.5% MCFA blend diets on days 0 and 14 and analyzed using 16s rDNA sequencing. Following feed manufacture, feed was stored in bags at barn temperature and humidity for 40 d before laboratory inoculation with PEDV. Subsamples of retained feed were inoculated with PEDV to achieve a titer of 10⁴ TCID₅₀/g and separate sample bottles were analyzed on 0 and 3 d post-inoculation (dpi). Overall, ADG and ADFI were increased (linear, P ≤ 0.010) and feed efficiency (G:F) improved (linear, P = 0.004) with increasing MCFA blend. Pigs fed 0.5% C8:0 had greater (P = 0.038) ADG compared with pigs fed the control diet, and G:F was improved (P ≤ 0.024) when pigs were fed 0.5% C6:0, 0.5% C8:0, or 0.5% C10:0 compared with control. An inclusion level × day interaction was observed (quadratic, P = 0.023), where PEDV Ct values increased (quadratic, P = 0.001) on 0 dpi with increasing levels of MCFA blend inclusion and also increased on 3 dpi (linear, P < 0.001). Fecal microbial diversity and composition were similar between control and 1.5% MCFA blend. In summary, the use of MCFA in nursery pig diets improves growth performance, provides residual mitigation activity against PEDV, and does not significantly alter fecal microbial composition.

Introduction

Medium-chain fatty acids (MCFA) have been shown to mitigate the risk of porcine epidemic diarrhea virus (PEDV) transmission via feed and ingredients (Dee et al., 2016; Cochrane et al., 2016, 2017). In addition to feed pathogen mitigation, research has evaluated the impact of MCFA on growth performance (Zentek et al., 2011). Various mechanisms by which MCFA may influence growth performance have been described, including its use as a readily available energy substrate, modifier of gastrointestinal morphology, and antibacterial compound (Hanczakowska, 2017). However, uncertainty still exists largely due to variability in the MCFA content of various sources available to the industry, and what inclusion level is necessary to optimize performance. Thus, additional research is necessary to further characterize the effects of dietary MCFAs on growth performance, shifts in fecal microbiome composition, and the value of feed-borne pathogen risk reduction. One of the main advantages of adding MCFA to feed is residual mitigation characteristics following feed manufacture and storage (Gebhardt et al., 2019). The ability for MCFA to retain antiviral properties post-manufacturing provides the opportunity to mitigate pathogen contamination during transport or at the farm. However, to date, no research has been conducted quantifying residual mitigation properties of MCFA against PEDV beyond 1-d post-feed manufacturing. Therefore, the objective of this experiment is to 1) determine the impact of hexanoic (C6:0), octanoic (C8:0), and decanoic (C10:0) acid supplementation in nursery pig diets on growth performance and fecal microbial composition, and 2) PEDV mitigation activity in complete feed following several weeks of feed storage.

Materials and Methods

Animals and Diets

The Kansas State University Institutional Animal Care and Use Committee approved the protocol used in this experiment. The study was conducted at the Kansas State Segregated Early Weaning Facility in Manhattan, KS. Following arrival on the day of weaning (approximately 20 d of age), pigs were randomized to pens and allowed a 6-d acclimation period during which they were fed a commercial starter pellet containing no feed-grade antimicrobials. Following acclimation, 360 pigs (DNA 400 × 200; Columbus, NE, initially 6.7 ± 0.07 kg BW) were blocked by BW and randomized to dietary treatment. Treatment diets were formulated and manufactured in 2 dietary phases (phase 1 = 7 to 11 kg BW; phase 2 = 11 to 23 kg BW). Diets were formulated to meet or exceed NRC (2012) requirement estimates for nutrients and energy. Phase 1 diets included 1,910 mg/kg added zinc (1,800 mg/kg from zinc oxide and 110 mg/kg from zinc sulfate) and 17 mg/kg added copper from copper sulfate. Phase 2 diets included 110 mg/kg added zinc from zinc sulfate and 17 mg/kg added copper from copper sulfate. Medium-chain fatty acids (Sigma–Aldrich, St. Louis, MO) included C6:0, C8:0, and C10:0 and were guaranteed ≥98% purity. Treatments (n = 8) were constructed such that a dose response was created including 0%, 0.25%, 0.5%, 1.0%, and 1.5% added MCFA blend (1:1:1 ratio C6:0, C8:0, and C10:0), as well as treatments with 0.5% of either C6:0, C8:0, or C10:0. Each pen had tri-bar floors and contained a 4-hole, dry self-feeder and a cup waterer to provide ad libitum access to feed and water. Pens (1.22 × 1.22 m) each contained 5 pigs and allowed approximately 0.298 m²/pig. Pig weights and feed disappearance were measured on days 0, 7, 14, 21, 28, and 35 to determine ADG, ADFI, and feed efficiency (G:F). Fecal consistency was evaluated on days 8, 11, 14, 21, 28, and 35 by 3 trained, independent reviewers per day. Scoring was performed on a 5-point scale with 1 = hard, pellet-like feces; 2 = firm, formed stool; 3 = soft, moist stool that retains shape; 4 = unformed stool; and 5 = watery liquid stool. When consensus was not achieved within reviewers to provide single fecal score, the concordant score was considered the definite score to be included in the statistical model as described by Menegat et al. (2019).

Chemical Analysis

Complete diet samples were collected following feed manufacture using a feed probe from every fifth bag, subsampled, and submitted (Ward Laboratories, Inc., Kearney, NE) for DM (AOAC 934.01, 2006), CP (AOAC 990.03, 2006), crude fiber (AOAC 978.10, 2006), Ca (AOAC 965.14/985.01, 2006), P (AOAC 965.17/985.01, 2006), and ether extract (AOAC 920.39 A, 2006). In addition, free fatty acid [C6:0, C8:0, C10:0; AOCS (2017)] concentrations were determined at the University of Missouri Agricultural Experiment Station Chemical Laboratory (Columbia, MO).

Feed Sample Collection and PEDV Inoculation

Following manufacture, all treatment diets were stored in 22.7 kg paper bags at barn temperature and humidity at the Kansas State University Segregated Early Wean facility for 40 d (June to July 2017). Meteorological data were collected using Weather Underground (wunderground.com) for the closest sampling station (Manhattan, Kansas Regional Airport; 11.9 km) for the dates of diet storage (average ± SD; average daily average temperature = 25.8 ± 3.52 °C, average daily high temperature = 33.1 ± 3.67 °C, average daily low temperature = 18.3 ± 3.99 °C, average daily maximum humidity = 90 ± 8.8%, average daily minimum humidity = 41 ± 9.4%). Following storage, samples were collected from multiple bags per treatment and subsampled using a riffle-splitter. Six, 22.5-g samples of each treatment diet were placed in separate 250-mL high-density polyethylene bottles (ThermoFisher Scientific, Waltham, MA) to be inoculated with PEDV and analyzed on 2 sampling days post-inoculation (dpi; 0 and 3 dpi), with 3 replications of each sampling day and treatment combination. In addition, 22.5 g of feed without MCFA was added to 3 separate bottles as control samples, which were not inoculated with PEDV and were analyzed along with the 0 dpi inoculated bottles.

Inoculation was carried out at the Kansas State University Veterinary Diagnostic Laboratory. The PEDV inoculum was cell culture derived (USA/IN/2013/19338, passage 9) and had an initial concentration of 10⁵ TCID₅₀/mL. Inoculation occurred by pipetting 2.5 mL of inoculum into each bottle containing 22.5-g feed matrix, resulting in an inoculated feed matrix with a viral concentration of 10⁴ TCID₅₀/g of feed matrix. Following the addition of the viral inoculum to each bottle, the bottles were lightly shaken in a circular pattern for approximately 5 s to prevent material from sticking to the side of the bottles. After which, each bottle was vigorously hand shaken for approximately 10 s to mix the virus evenly throughout the feed. The 3 negative control bottles had 2.5 mL of PBS added to each bottle as a sham inoculation following similar procedures to the viral inoculation. Samples were stored at room temperature until analysis performed on appropriate dpi.

Real-Time PCR Analysis

Separate bottles were analyzed on 0 and 3 dpi. On the day of inoculation (0 dpi), processing of the inoculated 0 dpi bottles occurred within 3 h of inoculation. On each day of analysis, 100-mL PBS (pH 7.4 1X, Life Technologies, Grand Island, NY) was added to each bottle predetermined for analysis on that day. Bottles were shaken for approximately 10 s, at which point they could settle overnight at 4 °C. The following day, supernatant was pulled and aliquoted for further analysis. A total of 2 aliquots from each sample bottle were collected and stored at −20 °C until the conclusion of the trial, at which point qRT-PCR analysis was performed on one aliquot per sample bottle. The qRT-PCR was conducted at the Kansas State University Veterinary Diagnostic Laboratory as previously described (Gebhardt et al., 2019). Values are reported as cycle threshold (Ct), where an increase in Ct value indicates less genetic material present in the analyzed sample.

Fecal Microbial population analysis

Fecal samples were collected from one pig per pen within control and 1.5% MCFA blend (1:1:1 ratio C6:0, C8:0, and C10:0) treatments (9 pens per treatment) on days 0 and 14 of the study period. The same pig within each pen was sampled on both dates. Sterile cotton-tipped applicators (Puritan Medical Products, Guilford, ME) were inserted into the rectum to stimulate defecation. One applicator was used per pig per time point. Fecal samples were collected into clean, single-use zipper storage bags and were then transferred into 3-mL cryovials and stored at −80 °C until shipment to the University of Nebraska, Lincoln, for DNA extraction and bacterial community analysis.

DNA was extracted using the manufacturer’s protocol for Mag-Bind Soil DNA 96 Kit (Omega Bio-tek, Inc., Norcross, GA) with the following modifications: precipitation of nucleic acids was done by using sodium acetate, isopropanol, and ethyl alcohol. Each sample tube received 0.1× volumes of 10 mM sodium acetate, which were vortexed and later incubated on ice for 5 min. Subsequently, 1 mL of ice-cold isopropanol was added and samples were incubated at −80 °C overnight to precipitate the DNA. The following day, samples were centrifuged at 4 °C for 15 min at 16,000 × g. The supernatants of the resulting samples were discarded, and the nucleic acid pellet was washed with 0.5 mL of ice-cold 70% ethyl alcohol. The samples were centrifuged for 2 min at 13,000 × g, the residual supernatant was discarded, and the nucleic acid pellet was air-dried for 3 min. The nucleic acid pellet was dissolved in a 0.45 mL of Tris [tris(hydroxymethyl)aminomethane; 10 mM, pH 8] and incubated for 1 h at 4 °C. For further purification of dissolved nucleic acids, the KingFisher (ThermoFisher Scientific) robot was used with reagents from the Mag-Bind Soil DNA 96 Kit. The resulting DNA was used for the tag-sequencing of the V4 region of 16S rDNA using the universal bacterial primers described previously (Kozich et al., 2013). A 20 µL PCR contained 1× TerraTM PCR Direct Polymerase Mix, 0.5 µL Terra polymerase, 20 mM of each primer, and 20 to 50 ng of DNA. The cycling conditions for PCR were the same as previously described (Paz et al., 2016). The PCR product size was confirmed by agarose gel electrophoresis. Normalization of the amplified PCR products was performed with Just-a-Plate 96 PCR Purification & Normalization kit (Charm Biotech, San Diego, CA) according to the manufacturer’s protocol. Following normalization, 10 µL from each sample was pooled and concentrated using Nucleospin Gel & PCR Cleanup kit (MACHEREY-NAGEL Gmbh & Co. KG, Duren, Germany) and was eluted using 20 µL of elution buffer. This pooled and purified sample was analyzed in an Agilent 2100 bioanalyzer (Agilent Scientific Instruments, Santa Clara, CA) using Agilent High Sensitivity DNA Kit (Agilent Technologies, Inc., Santa Clara, CA) to ensure the quality and quantity of the targeted V4 region of 16S rDNA. The concentration of the DNA library was determined using the DeNovix QFX Fluorometer (DeNovix Inc., Wilmington, DE) and using DeNovix dsDNA Fluorescence Quantification Assay (DeNovix Inc.). The resulting 16S rDNA libraries were sequenced using the Illumina MiSeq platform utilizing the 2 × 250 paired-end sequencing strategy using a MiSeq Reagent Kit V3 (Illumina Inc., San Diego, CA).

Data processing was performed on a custom pipeline utilizing several publicly available software tools. The paired-end reads were assembled into contiguous sets of overlapping clones (contigs) after quality filtering using MOTHUR v.1.38.1 as previously described (Schloss et al., 2009). Operational taxonomic units (OTUs) were generated from the quality-filtered sequences using the UPARSE pipeline (USEARCH v7.0.1090) at a threshold of 97% identity as previously described (Edgar, 2013). Chimeric sequences were removed using the ChimeraSlayer gold.fa as the reference database using UCHIME (Edgar et al., 2011). OTUs were aligned against the v128 (SILVA) database and mismatched sequences were discarded. A phylogenetic tree was generated using high quality aligned sequences within MOTHUR v.1.38.1 using the Clearcut algorithm as previously described (Sheneman et al., 2006). Taxonomies to the identified OTUs were assigned using QIIME v.1.9.1 pipeline (Caporaso et al., 2010) with the Greengenes reference database (gg_13_5_otus). OTUs representing Archaea and Cyanobacteria were removed as Cyanobacterial reads may be a result of contamination of plant chloroplast (Giovannoni et al., 1988), and the archaea sequences may be biased as the primers used are not designed to universally amplify all archaea. Alpha diversity matrices (Chao1 and Observed OTUs) were calculated using the QIIME v.1.9.1 pipeline. The rarefaction of the OTU table was performed using QIIME v.1.9.1 with the lowest number of reads (Edgar et al., 2011). For the experiment, 29,663 was used as the lowest depth.

Statistical Analysis

Growth data were analyzed as a randomized complete block design with pen as the experimental unit. Weight block was included in the statistical model as a random effect. Additionally, fecal score and PEDV inoculation data were evaluated as repeated measures over time. All comparisons incorporated Bonferroni multiple comparison adjustments within preplanned pairwise contrasts comparing individual 0.5% MCFA supplemented diets to 0.5% 1:1:1 MCFA treatment and control. Within outcomes, linear and quadratic effects of increasing MCFA blend were evaluated. Fecal consistency was evaluated assuming a multinomial distribution and considering the frequency distribution of pens within each fecal score category using a single determinate fecal score per pen per day of evaluation. An unstructured or first-order antedependence covariance structure was evaluated for fecal consistency scores due to the unequal nature of the evaluation days. The first-order antedependence was selected for use based on an improved Bayesian information criterion relative to the unstructured covariance matrix. Statistical analysis was performed using the GLIMMIX procedure of SAS version 9.4 (SAS Institute, Inc., Cary, NC). The FREQ procedure of SAS version 9.4 (SAS Institute, Inc.) was used to calculate frequency distribution of fecal scores within each fecal score category.

Fecal microbial population relative abundance was analyzed using a linear mixed model with individual pig as the experimental unit representing the single sample collected from each pen on a given sampling day. Relative abundance was considered the proportion of total reads for a specific sample classified into the designated microbial phyla or family and was analyzed assuming a binomial response distribution with the numerator being the number of reads within a sample for specific phyla or family and the denominator being the total number of reads for the sample. Day of analysis, treatment, and the associated interaction were included in the statistical model as fixed effects. A repeated measure statement was used to account for repeated sampling of the same set of pigs on both day 0 and day 14. Means separation was performed using the DIFF option to perform pairwise comparisons. The Kruskal–Wallis test was performed on the Chao1 and observed OTU’s to assess the bacterial richness difference among control and 1.5% MCFA blend treatments on day 14 using the NPAR1WAY procedure of SAS version 9.4 (SAS Institute, Inc.). All results were considered significant at P ≤ 0.05 and marginally significant between P > 0.05 and P ≤ 0.10 (Table 1).

Table 1.

Diet composition (as-fed basis)

Item	Weight range, kg
	7 to 11	11 to 23
Ingredient, %
Corn	54.92	62.55
Soybean meal (46.5% CP)	26.38	31.60
Whey powder	10.00	—
Enzymatically treated soybean meal2	2.50	—
Soybean oil	1.50	1.50
Calcium carbonate	0.95	1.00
Monocalcium phosphate (21% P)	1.30	1.15
Salt	0.60	0.60
l-Lysine HCl	0.50	0.51
Dl-Methionine	0.24	0.23
l-Threonine	0.21	0.21
l-Tryptophan	0.05	0.06
l-Valine	0.15	0.14
Trace mineral3	0.15	0.15
Vitamin premix4	0.25	0.25
Phytase5	0.07	0.07
Zinc oxide	0.25	—
Hexanoic acid, C6:06	±	±
Octanoic acid, C8:06	±	±
Decanoic acid, C10:06	±	±
Total	100	100
Calculated analysis7
Standardized ileal digestible (SID) AA, %
Lysine	1.35	1.35
Isoleucine:lysine	56	55
Leucine:lysine	111	113
Methionine:lysine	37.4	37.3
Methionine and cystine:lysine	58.2	58.1
Threonine:lysine	63.0	62.0
Tryptophan:lysine	20.1	20.3
Valine:lysine	70.3	70.1
Total lysine, %	1.48	1.49
ME, kcal/kg	3,349	3,347
NE, kcal/kg	2,502	2,485
SID Lys:ME, g/Mcal	4.03	4.03
SID Lys:NE, g/Mcal	5.40	5.43
CP, %	20.6	21.1
Ca, %	0.75	0.70
P, %	0.68	0.63
STTD P, %	0.57	0.50

¹Treatment diets were fed to 360 pigs [DNA 400 × 200 (Columbus, NE); initial BW = 6.7 ± 0.07 kg] for 35 d in a 2-phase feeding program with 5 pigs per pen and 9 pens per treatment.

²HP300 (Hamlet Protein, Findlay, OH).

³Premix provided per kg of premix: 110 g Fe from iron sulfate; 110 g Zn from zinc sulfate; 26.4 g Mn from manganese oxide; 11 g Cu from copper sulfate; 198 mg I from calcium iodate; and 198 mg Se from sodium selenite.

⁴Premix provided per kg of premix: 4,409,249 IU vitamin A; 551,156 IU vitamin D3; 17,637 IU vitamin E; 1,764 mg vitamin K; 15.4 mg vitamin B12; 19,842 mg niacin; 11,023 mg pantothenic acid; and 3,307 mg riboflavin.

⁵HiPhos 2700 (DSM Nutritional Products, Parsippany, NJ) provided an estimated release of 0.12% STTD P.

⁶Sigma–Aldrich (St. Louis, MO), guaranteed ≥98% purity added at the expense of soybean oil in appropriate treatment diets.

⁷NRC (2012).

Results

Chemical Analysis

Chemical analysis of diets (Table 2) reflected formulated values. The analyzed MCFA content closely matched formulated values, indicating manufacturing processes were effective in minimizing MCFA crossover between treatment diets. Ether extract decreased as MCFA was increased and soybean oil inclusion was decreased indicating that ether extract is not an efficient method to detect C6:0, C8:0, and C10:0 under these conditions.

Table 2.

Analyzed diet composition (as-fed basis)¹

Analyzed composition, %2	Added MCFA, %3
		C6:0:C8:0:C10:04				C6:0	C8:0	C10:0
	0	0.25	0.5	1.0	1.5	0.5	0.5	0.5
7 to 11 kg
DM	89.15	89.02	89.02	88.91	88.58	89.66	89.22	89.10
CP	20.90	20.70	21.20	20.90	20.45	20.85	20.50	19.80
ADF	3.90	3.35	3.20	2.45	2.30	4.35	2.80	2.10
Ether extract	4.40	3.95	3.80	2.75	2.15	3.15	2.80	2.90
Ca	0.91	0.91	1.00	0.92	0.95	0.88	0.98	0.93
P	0.76	0.75	0.76	0.67	0.71	0.71	0.72	0.70
Hexanoic acid	0.02	0.10	0.12	0.32	0.47	0.46	0.06	0.03
Octanoic acid	0.01	0.09	0.13	0.33	0.54	0.06	0.44	0.03
Decanoic acid	0.01	0.09	0.15	0.39	0.66	0.03	0.03	0.53
Total MCFA5	0.03	0.28	0.40	1.04	1.68	0.54	0.52	0.59
11 to 23 kg
DM	88.38	88.45	88.19	88.38	87.97	88.67	88.35	88.40
CP	20.90	21.00	21.15	21.10	20.65	20.90	21.50	20.95
ADF	2.80	2.70	2.50	2.35	3.10	3.15	3.05	2.95
Ether extract	3.90	2.90	2.70	2.35	1.90	3.05	3.35	2.90
Ca	0.94	0.87	0.85	0.83	0.84	0.85	0.81	0.85
P	0.69	0.62	0.63	0.65	0.62	0.64	0.64	0.64
Hexanoic acid	0.02	0.07	0.15	0.24	0.50	0.41	0.02	0.01
Octanoic acid	0.01	0.06	0.15	0.29	0.55	0.04	0.37	0.01
Decanoic acid	0.02	0.07	0.18	0.32	0.62	0.02	0.01	0.49
Total MCFA5	0.05	0.21	0.48	0.85	1.67	0.47	0.40	0.51

¹Treatment diets were fed to 360 pigs [DNA 400 × 200 (Columbus, NE); initial BW = 6.7 ± 0.07 kg] for 35 d in a 2-phase feeding program with 5 pigs per pen and 9 pens per treatment.

²Complete diet samples were collected following feed manufacture using a feed probe to create a composite sample, subsampled, and submitted to Ward Laboratories, Inc. (Kearney, NE) for proximate and the University of Missouri Agricultural Experiment Station Chemical Laboratory (Columbia, MO) for MCFA analysis performed in duplicate. Reported values are average of duplicate analysis.

³Sigma–Aldrich (St. Louis, MO). MCFA = medium-chain fatty acid.

⁴Consisted of a 1:1:1 blend of C6:0, C8:0, and C10:0.

⁵Sum of analyzed C6:0, C8:0, and C10:0 medium-chain fatty acids.

Growth Performance

During dietary phase 1, pigs fed increasing MCFA blend had increased (Table 3; linear, P ≤ 0.003) ADG and ADFI, as well as improved G:F (quadratic, P = 0.021). Pigs fed 0.5% C8:0 had greater (P = 0.007) ADG than pigs fed the control diet. Pigs fed 0.5% C6:0, 0.5% C8:0, or 0.5% C10:0 had improved G:F (P ≤ 0.044) compared with control-fed pigs. During phase 2, ADG increased (linear, P = 0.007) and ADFI marginally increased (linear, P = 0.052) with increasing MCFA blend. Overall, ADG, ADFI, and G:F increased (linear, P ≤ 0.010) with increasing MCFA blend. Pigs fed 0.5% C8:0 had greater (P = 0.038) ADG compared with pigs fed the control diets, and G:F was improved (P ≤ 0.024) when pigs were fed 0.5% C6:0, 0.5% C8:0, or 0.5% C10:0 compared with control. No evidence of a difference in BW, ADG, ADFI, or G:F was observed for dietary phase 1, dietary phase 2, or overall between individual MCFA at 0.5% inclusion compared with 0.5% inclusion of MCFA blend (P ≥ 0.467).

Table 3.

Effect of dietary addition of medium-chain fatty acids (MCFA) on nursery pig growth performance¹

Item	Added MCFA, %2
		C6:0:C8:0:C10:03				C6:0	C8:0	C10:0		Probability, P
	0	0.25	0.5	1.0	1.5	0.5	0.5	0.5	SEM	Linear4	Quadratic4	C6:0 vs. control	C8:0 vs. control	C10:0 vs. control
BW, kg
Day 0	6.7	6.7	6.7	6.7	6.7	6.7	6.7	6.7	0.07	0.605	0.641	1.000	1.000	1.000
Day 14	10.3	10.6	10.8	11.0	11.2	10.8	11.1	10.5	0.22	<0.001	0.359	0.178	0.009	1.000
Day 35	21.5	22.0	22.7	22.8	23.2	22.5	22.8	22.2	0.34	<0.001	0.177	0.194	0.033	0.798
Days 0 to 14
ADG, g	255	275	294	305	323	297	314	269	12.9	<0.001	0.376	0.105	0.007	1.000
ADFI, g	331	333	343	365	376	353	373	325	13.9	0.003	0.994	1.000	0.138	1.000
G:F	0.77	0.83	0.86	0.83	0.86	0.84	0.84	0.82	0.014	<0.001	0.021	0.004	0.003	0.044
Days 14 to 35
ADG, g	536	544	566	565	572	553	553	555	10.1	0.007	0.245	1.000	1.000	1.000
ADFI, g	829	814	842	868	855	837	834	834	16.8	0.052	0.513	1.000	1.000	1.000
G:F	0.65	0.67	0.67	0.65	0.67	0.66	0.66	0.67	0.008	0.367	0.503	1.000	0.811	0.497
Days 0 to 35
ADG, g	423	436	457	461	472	451	457	441	9.1	<0.001	0.187	0.159	0.038	0.927
ADFI, g	630	621	642	667	664	644	649	630	14.1	0.010	0.625	1.000	1.000	1.000
G:F	0.67	0.70	0.71	0.69	0.71	0.70	0.70	0.70	0.007	0.004	0.096	0.019	0.005	0.024
Feed PEDV Ct5,6
0 dpi	27.1	29.5	30.9	30.6	32.4	29.7	30.0	28.7	0.27	<0.001	0.001	<0.001	<0.001	0.003
3 dpi	28.3	28.3	29.9	30.9	32.1	30.5	29.5	29.9		<0.001	0.745	<0.001	0.061	0.004

¹A total of 360 pigs [DNA 400 × 200 (Columbus, NE); initial BW = 6.7 ± 0.07 kg] were fed for 35 d in a 2-phase feeding program with 5 pigs per pen and 9 pens per treatment. Bonferroni multiple comparison adjustments were applied to contrasts comparing 0.5% inclusion of individual MCFA compared with control and 0.5% inclusion of individual MCFA compared with 0.5% MCFA blend for all response variables.

²Sigma–Aldrich (St. Louis, MO).

³Consisted of a 1:1:1 blend of C6:0, C8:0, and C10:0.

⁴Linear and quadratic contrast statements include control and treatments fed 1:1:1 blend of MCFA.

⁵Each number is the mean of 3 samples of the 11 to 23 kg diet that were inoculated with porcine epidemic diarrhea virus (PEDV) 40 d following feed manufacturing with a calculated final titer of 10⁴ TCID₅₀/g feed. On 2 sampling days post-inoculation [dpi; day of inoculation (0 dpi) and 3 dpi], separate samples were analyzed for the presence of PEDV genetic material using qRT-PCR. Three bottles containing swine feed were not inoculated with PEDV as a negative control and analyzed on day 0, and qRT-PCR analysis did not detect genetic material using a threshold cutoff value of 45 cycles. A higher Ct value indicates a greater amount of PEDV genetic material present in the original sample. SEM representative for entire qRT-PCR analysis.

⁶Treatment × day, P < 0.001; Quadratic inclusion of MCFA blend × day, P = 0.023; 0 dpi C6:0 vs. 0.5% 1:1:1 and 0.5% C10:0 vs. 0.5% 1:1:1, P ≤ 0.047. SEM represents both 0 and 3 dpi.

No evidence for a dietary treatment × day interaction was observed (P = 0.663; Fig. 1) for fecal consistency. At all points of evaluation, the majority of determinant fecal scores were characterized as firm formed stool, soft, moist stool that retains shape, or unformed stool. As expected, a significant day effect was observed (P < 0.0001). Over the course of the study, fecal consistency transitioned to a firmer pattern, with a decrease in frequency distribution of pens with soft, moist feces that retains shape and an increase of pens with firm, formed stool. There was no evidence that treatment had an effect on fecal consistency (P = 0.211).

Figure 1.

Effects of study day on fecal consistency of nursery pigs assessed by fecal score. Graph bars show the frequency distribution of fecal scores (n = 72 pens) within each fecal score category according to day regardless of dietary treatment. Score was evaluated by 3 trained individuals on each day, and the concordant score was considered the definite score for each pen on each day of evaluation.

Mitigation Characteristics

A MCFA blend inclusion level × day interaction was observed (quadratic, P = 0.023; Table 3). This was the result of PEDV Ct values increasing in a quadric manner (P = 0.001) as inclusion rates increased on 0 dpi while increasing in a linear manner on 3 dpi (linear, P < 0.001). On 0 dpi, the addition of 0.5% C6:0, 0.5% C8:0, or 0.5% C10:0 resulted in significantly greater Ct values compared with control (P ≤ 0.003). Inclusion of 0.5% C6:0 or 0.5% C10:0 resulted in lower Ct values (P ≤ 0.047) on 0 dpi compared with 0.5% MCFA blend. Insufficient evidence was available on 0 dpi to demonstrate a difference in Ct value between 0.5% C8:0 and 0.5% MCFA blend treatments (P = 0.394). On 3 dpi, the addition of 0.5% C6:0 or 0.5% C10:0 resulted in greater Ct values compared with control (P ≤ 0.004), with marginally significant evidence that the 0.5% C8:0 treatment Ct value was greater than control (P = 0.061). Sufficient evidence was lacking on 3 dpi to demonstrate a difference in Ct value between 0.5% MCFA blend treatment and treatments consisting of individual MCFA at 0.5% inclusion (P = 1.000).

Fecal Microbial Population Analysis

The majority of bacterial sequences were classified in 2 phyla on days 0 and 14, and consisted of Firmicutes and Bacteroidetes (Fig. 2). A total of 6 phyla were found at ≥1% relative abundance for at least one of the treatment × day analysis combinations. Phyla and families with low (<1%) relative abundance were excluded from the data analysis. Overall, fecal microbial populations were similar between pigs fed 1.5% MCFA compared with control pigs. A marginally significant treatment × day interaction was observed in the Proteobacteria phylum (P = 0.057), where relative abundance was consistent over time in pigs fed the control diet (P = 0.668; 1.5 ± 0.52, 2.1 ± 0.60%; day 0, day 14, respectively), whereas a decrease over time was observed in pigs fed the 1.5% MCFA blend diet (P = 0.023; 1.2 ± 0.46, 0.1 ± 0.16%; day 0, day 14, respectively). There was no evidence of a difference in the effect of MCFA addition over time for the relative abundance of the remaining phyla (treatment × day, P ≥ 0.231).

Figure 2.

Relative abundance of microbial phyla and families by day presented as proportion of all reads for a specific sample classified into the designated microbial phyla or family ± SEM. A total of 18 pigs [1 pig per pen from each of 9 control diet-fed pens and 1 pig from each of 9–1.5% medium-chain fatty acid (MCFA) blend-fed pens] were selected randomly for sampling of fecal material for characterization of fecal microbial populations. Blend of MCFA consisted of 1.5% blend of 1:1:1 ratio of C6, C8, and C10; Sigma–Aldrich (St. Louis, MO). Sampling of the same set of pigs occurred on days 0 and 14. Samples were analyzed using 16s rDNA sequencing. Microbial phyla and families lacking at least one treatment × day combination having a relative abundance of ≤1% are not shown. A dietary treatment × day interaction was observed for Proteobacteria (P = 0.057), where relative abundance was consistent over time in pigs fed the control diet (P = 0.668), whereas a decrease over time was observed in pigs fed the 1.5% MCFA blend diet (P = 0.023). All other treatment × day, P ≥ 0.138. *Main effect of day, P < 0.05; ^xmain effect of day, 0.05 < P ≤ 0.10.

The main effect of day indicated a significant increase over time in the Tenericutes phylum (P = 0.016). A marginally significant increase over time was observed in the Firmicutes (P = 0.096) and Actinobacteria (P = 0.088) phyla. A significant decrease over time was observed for the Proteobacteria phylum (P = 0.026) and Spirochaetes phylum (P = 0.022). A Firmicutes:Bacteroidetes ratio was calculated, and there was no evidence of a treatment × day interaction (P = 0.338) or day effect (P = 0.211).

A total of 22 families were found at ≥1% relative abundance for at least one of the treatment × day of analysis combinations. The families with the greatest relative abundance were Prevotellaceae, Ruminococcaceae, and S24-7, which each had ≥10% relative abundance for at least one of the treatment × day analysis combinations. There was no evidence of a difference in the effect of MCFA addition over time for the relative abundance for any family (treatment × day, P ≥ 0.138). A reduction over time was observed for Ruminococcaceae (P = 0.032), Lachnospiraceae (P = 0.006), Christensenellaceae (P = 0.013), Spirochaetaceae (P = 0.033), Bacteroidaceae (P = 0.014), and Succinivibrionaceae (P = 0.039). An increase in relative abundance over time was observed for the unclassified Clostridiales (P = 0.014), Clostridiaceae (P < 0.0001), unclassified RF39 (P = 0.016), and Clostridiales; other (P = 0.001) families. At day 14, no evidence of a difference in alpha diversity was observed for Chao1 or observed OTU’s between control and 1.5% MCFA blend treatments (P > 0.10) was observed.

Discussion

Feedstuffs delivered to swine farms are a potential source of disease transmission (Pasick et al., 2014; Bowman et al., 2015, Aubry et al., 2017). Mitigation of transmission risk through feed has clear and definable economic benefits. The use of MCFA compounds to mitigate the risk of PEDV transmission via feed or feed ingredients has significant scientific promise (Cochrane et al., 2016, 2017; Dee et al., 2016). In addition to feed pathogen mitigation, MCFAs have been evaluated as compounds to improve growth performance as reviewed by Zentek et al. (2011) and Hanczakowska (2017). However, uncertainty still exists due to variability in the MCFA content of various sources available to the industry, and what inclusion level is necessary to optimize performance. Potential mechanisms by which MCFA may influence growth performance are diverse, and include use as a readily available energy source, modifier of gastrointestinal morphology, and modifier of gastrointestinal microbial populations through antibacterial properties (Zentek et al., 2011; Hanczakowska, 2017).

Medium-chain fatty acids are unbranched, saturated monocarboxylic acids containing between 6 and 12 carbon atoms (Zentek et al., 2011). In nature, these compounds are often in triglyceride form [medium-chain triglyceride (MCT)] with greatest concentrations in coconut and palm kernel oil (Zentek et al., 2011). To date, MCFA studies conducted by our research group regarding mitigation of feed-borne pathogens have utilized the free fatty acid forms. Depending on the chemical structure (MCFA vs. MCT), digestion, absorption, and utilization differ. Due to the water-soluble nature of MCFA, they are rapidly absorbed within the small intestine and can be used by enterocytes as an energy substrate, or travel to the liver to undergo β-oxidation in hepatocytes (Zentek et al., 2011). To accurately assess the value of MCFA utilization for enhancement of growth performance in combination with previous research evaluating efficacy as a feed-borne pathogen mitigant, the current experiment utilized the free fatty acid form.

The effects of MCFA inclusion in swine diets on growth performance have recently been reviewed by Hanczakowska (2017) with multiple studies demonstrating positive effects on gain and feed efficiency. In the current experiment, the addition of MCFA resulted in an improvement in growth performance compared with control, particularly C6:0 and C8:0 fatty acids for ADG and C6:0, C8:0, or C10:0 for G:F. Hanczakowska et al. (2011b) observed an increase in ADG when 0.2% C8:0, 0.2% C10:0, and a combination of 0.1% C8:0 and 0.1% C10:0 were included in swine diets beginning at 7 d of age compared with control, and feed efficiency was also improved when pigs were fed 0.2% C8:0 compared with control. The levels of MCFA used in the current experiment were significantly higher than Hanczakowska et al. (2011b); however, similar improvements in growth performance were observed. Hanczakowska et al. (2013) observed an increase in ADG when pigs were fed C8:0, C10:0, or combination of C8:0 and C10:0 in diets containing propionic and formic acids compared with control-fed pigs; however, ADG did not significantly differ comparing pigs fed MCFA, propionic, and formic acids to propionic and formic acids alone. Mohana and Kim (2014) did not observe any improvement in growth performance when pigs were supplemented 0.2% MCFA product, which contained 58% C8:0 and C10:0, compared with control-fed pigs. Largely, published literature has shown that MCFA supplementation improves growth performance. Variability in presence and magnitude of effect is probably driven by differences in MCFA origin and dose, further reinforcing the need for greater understanding of optimum dose.

Several studies have shown evidence that inclusion of MCFA in pig diets of varying ages can improve gastrointestinal morphology through increased villus length and associated increased villus height:crypt depth ratio (Dierick et al., 2013; Hanczakowska et al., 2011a,b; Chwen et al., 2013). Contrary to these findings, other studies did not find evidence of a difference in villus height or crypt depth when MCFA were fed to young pigs (Hanczakowska et al., 2016; Ferrara et al., 2017). Increased surface area for nutrient digestion and absorption is generally believed to be beneficial, and Han et al. (2011) observed an increase in nutrient digestibility including energy, crude protein, calcium, phosphorous, and AA when pigs were fed a micro-encapsulated eucalyptus-MCFA product. The present study did not evaluate morphologic changes in the gastrointestinal tract or nutrient digestibility; however, feed efficiency was improved with MCFA supplementation. Further evaluation of gastrointestinal morphology using the free fatty acid form used in the current experiment at the dose range utilized would be necessary to draw any associations between the observed improvements in growth and feed efficiency and changes to morphology attributed to MCFA.

Pure MCFA commonly have a foul odor and have been shown to reduce feed intake as described by Hanczakowska (2017). Cera et al. (1989) observed a reduction in intake when C8:0 and C10:0 fatty acid mixture was fed to weaned pigs; however, it was not known whether the reduction was due to odor characteristics or nutritional adjustment to increased energy availability. On the contrary, multiple studies have shown no evidence of reduced feed intake when MCFA are fed (Hanczakowska et al., 2011a, 2016; Mohana and Kim, 2014), which is consistent with our data. The observed differences among previous studies may be attributed to differences in MCFA inclusion level as well as origin of the MCFA. The MCFA used in the current experiment created a strong odor both in pure form and when mixed into feed. The strong odor of the MCFA used in the current experiment did not reduce feed intake, but rather led to a linear increase with increasing MCFA blend inclusion.

It has been observed that weaned pigs have firmer stool when fed pharmacological copper and zinc doses shortly after weaning, which also correlates with improved growth (Hill et al., 2000; Shelton et al., 2011). It is also accepted that clinical enteric disease can create diarrhea and can significantly reduce growth performance. In the current experiment, no known significant disease pressure was present, and firmer stools were observed over time. Contributing factors for the observed impact on stool firmness could include changes in gastrointestinal microbial populations unable to be identified in fecal analysis or modulation of gastrointestinal water and ion balance over time. A greater understanding of the underlying physiology is necessary prior to inferring causation for changes in fecal consistency.

Generally, MCFA are considered to have inhibitory effects on growth and composition of bacterial populations (Zentek et al., 2011). Although control of pathogenic bacterial populations is beneficial, commensal bacterial populations play a significant role in nutrient metabolism and immune system development (Mach et al., 2015). The mechanism by which MCFA exert their antibacterial properties is not fully understood; however, it is likely due to amphiphilic structure that allows for disruption of cellular membranes, leakage of intracellular materials, and subsequent cellular death (Zhou et al., 2019). Zentek et al. (2011) and Hanczakowska (2017) summarized both in vivo and in vitro experiments evaluating the effects of MCFA on microbial populations, and concluded the antimicrobial properties of C8:0 and C10:0 are most consistent against Gram-positive bacteria, with limited activity against Gram-negative species. Recent in vitro experiments have shown that using a 1:1:1 ratio of C6:0:C8:0:C10:0 MCFA has the lowest minimum inhibitory concentration (MIC) against generic Escherichia coli, and C6:0 has the lowest MIC for enterotoxigenic E. coli (Swanson et al., 2018). Additionally, Sylvester et al. (2018) observed that C6:0 and C8:0 had lower MIC for Salmonella Typhimurium compared with C10:0 or 1:1:1 ratio of C6:0:C6:0:C10:0 MCFA. Shilling et al. (2013) observed that C12 was more inhibitory toward growth of Clostridium difficile than C6:0 or C8:0 MCFA. Escherichia coli and Salmonella are both Gram-negative bacterial species, whereas C. difficile is a Gram-positive bacterium. Thus, the efficacy of shorter length MCFA compounds is generally more effective at inhibiting the growth of Gram-negative species, whereas Gram-positive species are more effectively controlled with longer length MCFA. Previous literature would suggest that although MCFA have been shown to have antibacterial properties, the effectiveness differs among the combination of MCFA and bacterial species. In an effort to further characterize the impact of MCFA supplementation on the fecal microbial populations, the current experiment evaluated the fecal microbial populations on days 0 and 14 using 16s rDNA sequencing.

Consistent with previous literature evaluating microbial populations at multiple locations within the gastrointestinal tract (Li et al., 2018; Pollack et al., 2018; Yang et al., 2017), the largest proportion of microbes in the present study were characterized as part of the Firmicutes and Bacteroidetes phyla. At the family level, Li et al. (2018) found that Prevotellaceae to be most abundant family in the colon of piglets around the time of weaning, which is consistent with the present study. The microbial composition within different locations in the gastrointestinal tract substantially differs (Zhao et al., 2015), and fecal characterization probably only represents a subset of total gastrointestinal microbial populations (Quan et al., 2018). The present study only evaluated fecal microbial populations; thus, differences in microbial populations in the small or large intestine would have gone undetected using current methodologies.

Addition of MCFA resulted in minor alterations in fecal bacterial populations. A marginally significant interaction characterized by a reduction in Proteobacteria over time in pigs fed MCFA compared with no change in controls was the only change associated with MCFA supplementation. Proteobacteria is a diverse phyla of Gram-negative bacteria including the genera Escherichia and Salmonella (Gupta, 2000). In humans, high levels of Proteobacteria are associated with unstable microbial communities, commonly known as dysbiosis (Shin et al., 2015). In the present study, the marginally significant reduction in Proteobacteria over time caused by MCFA may have been a contributing factor to improved growth performance. However, no evidence for differences on a family level was associated with MCFA supplementation over time. It is possible that significant changes in community structure could take longer than 14 d and thus would not be identified within the current experiment. Additionally, further characterization of the microbiome, using metagenomic sequencing or the evaluation of nonbacterial microbial composition, may reveal a more significant impact of MCFA on the gut microbial populations. Nonetheless, the present study suggests that the positive impact associated with feeding MCFA on growth may not be directly related to substantial shifts in fecal microbial populations.

With recent advances in high-throughput quantification of microbial populations, research focusing on the impact of gastrointestinal microbial populations on feed efficiency is present (McCormack et al., 2017, 2019; Maltecca et al., 2019). The results of these studies identify multiple taxonomic classifications that are observed to be associated with feed efficiency and growth performance outcomes. In one such investigation, pigs with greater feed efficiency than their counterparts were observed to have greater abundance of the Lachnospiraceae and Prevotellaceae families along with the Escherichia–Shigella and Streptococcus genera (Quan et al., 2019). In addition, advances in knowledge of the role-specific organisms play in metabolism of starch, fructans, and lactose has improved, which may lead to improved theories regarding efficiency of nutrient utilization and growth performance (Wang et al., 2019). Although the overall microbial community remained largely unaffected by the addition of 1.5% MCFA in the diet, time had a significant influence on relative microbial abundance. It is understood that as young pigs age and transition from a milk-based diet to dry feed, their digestive enzyme profile and gastrointestinal microbial profile adjust (Hu et al., 2016; Chen et al., 2017; Guevarra et al., 2019). Multiple changes in relative abundance were observed over time, indicating that the experimental methods were sensitive to changes in microbial populations and provide further understanding of the changes in microbial populations post-weaning.

The use of MCFA to mitigate PEDV in swine diets has been evaluated by Cochrane et al. (2016) and reported to reduce quantifiable genetic material and infectivity characteristics. To date, no research has been conducted quantifying the residual mitigation properties of MCFA beyond 1-d post-feed manufacturing. In contrast, the use of 37% formaldehyde has been approved to be added to livestock feed or feed ingredients to maintain Salmonella-negative status for up to 21 d (FDA, 2003). We believe that this is the first study to demonstrate long-term mitigation potential for MCFA against PEDV. In the current experiment, feed treated with MCFA retained mitigation activity as quantified by an increase in Ct value when inoculated with PEDV following a storage period of 40 d. Thus, the present study provides a baseline understanding of the timeline associated with the residual duration of activity that treatment of feed with MCFA provides. An important limitation of the present study was lack of evaluation of infectivity characteristics post-inoculation using methods such as virus isolation or bioassay. The use of swine bioassay models to evaluate the infectivity of feed samples inoculated with PEDV has been shown to be a robust experimental model (Schumacher et al., 2016). The current experiment was preliminary in nature and set out to determine residual mitigation activity as quantified using qRT-PCR as a baseline for further investigation. Additional investigation is required to fully understand the level of risk reduction for post-processing contamination that can be expected when using MCFA in swine feed.

In summary, treating swine nursery feed with MCFA improved growth performance, resulted in sustained mitigation properties when inoculated with PEDV following a period of feed storage, and does not significantly alter fecal microbial populations. The use of such compounds may provide benefits to swine producers from both a reduction in disease transmission risk and improved growth performance. Further investigation into the role of added dietary MCFA including sources that would be practical and cost-effective on growth performance, gastrointestinal bacterial flora, and residual mitigation properties against feed-borne pathogens is warranted.