Commercial TTO (1 kg) was mixed with phosphatidyl choline (1 kg) isolated from soybean, and an antioxidant (0.2 kg vitamin E) in n-hexane solution (2 L; Figure 1; Hebei Chenguang Biotech Group Co., Ltd.). The solution was heated to 50°C and stirred for 3 h followed by vacuum drying to form 1 layer on TTO particles. Carrier material was prepared by mixing lactose (4 kg) and gum arabic (3.75 kg) in water (25 L, 50°C). This carrier solution was emulsified at 9,600 g. Dried 1-layer TTO particles were mixed with the carrier solution and further emulsified at 9,600 for 30 min and homogenized at a pressure of 20 to 60 MPa for two to three times. To this second step, two-layer encapsulated TTO was produced. Finally, another carrier, mixture of maltodextrin and cornstarch (1:2 wt/wt), was spray-dried on two-layer TTO and sieved (60-mesh screen) forming encapsulated three-layer TTO particles. The diameter of the one-layer particle composed of TTO and phosphatides was from 2 to 20 μm, and the diameter of the three-layer encapsulated TTO was from 75 to 250 μm.

Commercial TTO was obtained from Hebei Chenguang Biotech Group Co., Ltd. (Handan, China), and extraction was performed by the steam distillation technique, utilizing leaves of tea trees. The analysis was conducted using gas chromatography–mass spectrometry (QP2010, Shimadzu Company, Japan) according to Adams (19). The ionization voltage, mass range, and the condition of GC condition were set according to previously described methods (20). Peaks were identified that corresponded to the data from mass spectra library (National Institute of Standard and Technology). Finally, quantitative determination was conducted based on the peak area integration.

Piglets (n = 144, Duroc × Landrace × Large white, weaned at 28 days after birth) with an average initial weight of 8.5 ± 0.24 kg were used in this experiment. This study designed six replicates per treatment with six piglets (three male and three female piglets) per pen, and the experiment was separated into phase 1 (days 1–14) and phase 2 (days 15–28). Dietary treatments included basal diet (NC), basal diet supplemented with 75 mg/kg aureomycin (PC), basal diet supplemented with 0.4% unencapsulated TTO (Un-encp TTO; unencapsulated TTO was sprayed on porous starch and dried at 20°C for 30–50 min), and basal diet supplemented with 0.4% encapsulated TTO (Encp TTO; the encapsulation strategy was described in “preparation of encapsulated TTO”). The actual content of TTO in each product was 11% (Hebei Chenguang Biotech Group Co., Ltd.). We added 0.4% encapsulated or unencapsulated TTO to the basic diet, so the actual value of TTO among treatments was 0.04%. All diets were formulated to meet or exceed the requirements suggested by National Research Council (NRC, 2012) estimates of the nutrient requirements of weaned piglets (Table 1). Chromic oxide (0.3%) was added to each treatment diet as an indigestible marker to calculate the apparent total tract digestibility (ATTD) of nutrients. Pigs were housed in pens that measured 1.2 ×2.1 m and were equipped with stainless-steel feeder and one nipple waterer with slatted plastic floors. The environmental temperature was controlled and maintained at 24–26°C. Pigs had ad libitum access to feed and water.

Each piglet was weighed on days 0, 14, and 28, and feed consumption was also recorded on a pen basis to determine average daily gain (ADG), average daily feed intake (ADFI) and the feed to gain ratio (F:G). Fecal consistency was assessed visually three times per day by observers blinded to treatments according to previously described methods (21). The scoring system was applied to determine the rate of diarrhea as following: 0 = firm and normally shaped, 1 = soft (retained some shape), 2 = brown liquid, 3 = frequent passage of watery feces (22). The occurrence of diarrhea was defined as a continuous 2-day fecal score at level 2 or 3. Diarrhea rate was calculated using the following formula. Diarrhea rate (%) = (number of piglets with diarrhea × diarrhea days) / (number of piglets × total observational days) ×100%.

On the morning of days 14 and 28, one pig closest to the average body weight in each pen was selected to collect blood samples (8 mL) via the anterior vena cava puncture into vacuum tubes following a 12-h fasting. Serum samples were separated by centrifugation at 3,000 g at 4°C for 15 min and stored at −20°C until further analysis.

On the morning of day 28, six pigs per treatment group were randomly selected and were euthanized by electric shock and then bled. The selected pigs were from different pens, and their body weights were close to the average body weight of pigs in each pen. The small intestine was dissected from the mesentery and immediately placed on ice. Segments (2 cm in length) of the middle parts of the duodenum, jejunum, and ileum without any digesta were fixed in 10% neutral-buffered formalin for subsequent morphological examination. Liver tissues were collected and quickly washed with ice-cold saline solution and then homogenized in ice-cold phosphate buffer (phosphate-buffered saline: 1:9 wt/vol, pH 7.4) and centrifuged at 2,000 g for 20 min. The supernatant was collected and stored at −80°C for inflammatory cytokine analysis. Liver microsomes were prepared by differential centrifugation (11,000 and 2 ×100,000 g) in a Tris/KCl buffer (20 mM, pH 7.4) and stored at −80°C for cytochrome P450 (P450) analysis (23).

Fecal samples for determination of nutrient digestibility were collected from each pen during the last 3 days (days 25–27) and were immediately stored at −20°C. At the end of the experiment, the 3-day collection of fecal samples was pooled by pen and then dried in a forced-air drying oven at 65°C for 72 h. Feed and dried fecal samples were ground finely to pass through a 1-mm screen (40-mesh) before chemical analysis. For analysis of gut microbiota, digesta in the cecum and colon of slaughter piglets were sampled (each sample from one pig per pen) in sterile tubes on day 28 and immediately immersed in liquid nitrogen and then stored at −80°C (24).

Dry matter (DM; method 934.01), crude protein (CP; method 990.03), and ether extract (EE; method 920.39) of feed and fecal samples were determined according to the Association of Official Analytical Chemists (25). Neutral detergent fiber (NDF) concentration of these samples were determined according to the method of McCarthy et al. (26). The content of chromium in feed and fecal samples was determined by atomic absorption spectrophotometer (Hitachi Z-5000, Tokyo, Japan) (27).

In the serum, concentrations of glutathione peroxidase (GSH-Px) and superoxide dismutase (SOD) were determined according to the manufacturer's instructions using commercial kits (Jiancheng Biochemical Reagent Co., Nanjing, China). The content of serum immunoglobulins A, M, and G (IgA, IgM, and IgG) were measured using commercially available enzyme-linked immunosorbent assay (ELISA) kits following manufacturer's instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Serum biochemical parameters including aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP) were measured using corresponding commercially available kits (BioSino Bio-technology and Science Inc., Beijing, China).

In serum and liver, concentrations of interleukin 10 (IL-10), tumor necrosis factor α (TNF-α), and IL-1β were determined using the corresponding commercially available ELISA kits (Beijing Sino-UK Institute of Biological Technology, Beijing, China). Concentrations of the microsomal cytochromes P450 were determined using the spectrophotometric method of Omura and Sato (28).

Intestinal morphology was characterized according to the methods of our previous study (22). Briefly, fixed intestinal samples were dehydrated and embedded in paraffin. Embedded samples were cut into slices (6 μm thick) and stained with hematoxylin-eosin. Then six villous height (VH, determined as the distance between the crypt opening to the end of the villous), crypt depth (CD, measured from the crypt–villous junction to the base of the crypt), and VH/CD (VH/CD) ratio were measured under a light microscope (Nikon, Tokyo, Japan) at a 40 × magnification and were analyzed using Image-Pro Plus 6.0 software.

Microbial DNA was extracted from cecal and colonic digesta using the E.Z.N.A.® stool DNA kit (Omega Bio-tek, Norcross, GA, USA) according to the manufacturer's instructions. Markers and adaptors-linked universal primers 338F (ACTCCTACGGGAGGCAGCAG) and 806R (GGACTACHVGGGTWTCTAAT) targeting the V3–V4 region were used to amplify microbial 16S rRNA. The polymerase chain reaction (PCR) amplification of 16S rRNA gene was performed as follows: 95°C for 3 min, followed by 25 cycles at 95°C for 30 s, 55°C for 30 s, and 72°C for 45 s and a final extension at 72°C for 10 min (29). The PCR product was extracted with 2% agarose gel and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA). Purified amplicons were pooled in equimolar proportions and paired-end sequenced (2 ×300) on an Illumina MiSeq platform (Illumina, San Diego, CA, USA). The processing of sequencing data was conducted as previously described (30). The data presented in the study are deposited in the NCBI repository, accession number PRJNA677843. Operational taxonomic units (OTUs) with a 97% similarity cutoff (31) were clustered using UPARSE (version 7.1, http://drive5.com/uparse/), and chimeric sequences were identified and removed. Taxonomy of each OTU representative sequence was analyzed by RDP Classifier (http://rdp.cme.msu.edu/) against the 16S rRNA database (Silva SSU128) using confidence threshold of 0.7.

GLM procedure of SAS (SAS Institute Inc., Cary, NC, USA) was used for data analysis. Each pen of pigs was considered as an experimental unit. Diarrhea rates were analyzed by χ² test (SAS Institute Inc., Cary, NC, USA). Planned orthogonal contrasts were used to evaluate the overall effect of TTO ([NC] vs. [Un-encp TTO, Encp TTO]) or encapsulation effect ([Un-encp TTO] vs. [Encp TTO]). All data were checked for normal distribution and homogeneity of variance using the Shapiro–Wilk normality test and Bartlett test, respectively. The differences and composition of bacterial abundance were analyzed by Kruskal–Wallis rank-sum test (29). p < 0.05 was considered statistically significant, and 0.05 ≤ p < 0.10 was indicative of a differential trend.

The total yield of volatile compounds from encapsulated TTO was 2.74% ± 0.22%. Fourteen compounds from the encapsulated TTO were detected by gas chromatography–mass spectrometry (Table 2). Two volatile compounds, terpinen-4-ol (50.31%) and γ-terpinene (20.06%), accounted for the majority of volatiles measured.

Regardless of encapsulation, compared with the negative control treatment, TTO and positive control treatment increased (p < 0.05) ADG at phase 1 (days 0–14), phase 2 (days 15–28), and overall (days 0–28) and increased (p < 0.05) ADFI at phase 2 and overall (Table 3). TTO and PC treatment decreased F:G (p < 0.01) and diarrhea rate (p < 0.01) compared with the NC treatment. In addition, there were no differences in growth performance between TTO treatment and PC treatment. Concerning the encapsulation effect, compared with the Un-encp TTO treatment, the treatment of Encp TTO had higher ADG at phase 1, phase 2, and overall phases (p < 0.05) and higher ADFI in phase 2 and overall phases (p < 0.05). Encp TTO treatment had lower F:G at phase 1 (p < 0.05) and tended to have lower diarrhea rate compared with Un-encp TTO treatment (p = 0.083).

TTO treatment (Un-encp TTO and Encp TTO) increased ATTD of CP, EE, and DM compared with the negative control treatment (p < 0.05, Table 4). TTO treatment tended to increase ATTD of EE (p = 0.089) compared with the PC treatment. There were no differences in ATTD of DM, CP, and NDF between TTO treatment and PC treatment. Encapsulation of TTO (Encp TTO vs. Un-encp TTO) had no significant effect on ATTD of these nutrients.

Pigs fed the TTO and positive control diets had higher VH and VH/CD ratio in the jejunum than those fed NC diet (p < 0.05, Table 5). The VH, CD, and VH/CD ratio in the duodenum and ileum did not differ among NC, PC, and TTO treatments (Un-encp TTO and Encp TTO). Encp TTO tended to increase the VH/CD ratio (p = 0.089) in the jejunum compared with the Un-encp TTO treatment.

Serum SOD, GSH-Px, and IgG in piglets fed TTO and PC diets were higher than the NC treatment (p < 0.05, Table 6). Serum IgG was greater in pigs fed the TTO diet than those fed the PC diet. Compared with the Un-encp TTO and PC treatment, piglets fed with Encp TTO showed higher serum SOD and GSH-Px activities (p < 0.05).

Pigs fed TTO exhibited lower serum concentrations of ALT and AST compared with pigs fed the NC and PC treatments (p < 0.05, Table 7). In addition, supplementation with TTO increased serum level of IL-10 (p < 0.05) and tended to decrease the level of TNF-α (p > 0.05) compared with NC and PC treatments. Compared with pigs fed Un-encp TTO and PC, Encp TTO decreased concentrations of AST and increased concentration higher level of IL-10 (p < 0.05).

Pigs fed TTO treatments had lower levels of IL-1β (p = 0.004) in the liver than those in the negative control treatment (Figures 2B,C). Compared with the Un-encp TTO and positive control treatments, lower level of IL-1β and higher level of IL-10 were observed in pigs fed Encp TTO (p < 0.05). There were no significant differences in concentration of TNF-α and P450 in liver among experimental treatments (Figures 2A,D).

We observed a shift in the microbial composition of cecal and colonic digesta (Figure 3). The dominant bacteria phyla across treatments were mainly Firmicutes, Bacteroidetes, and Proteobacteria. At the genus level, abundance of Subdoligranulum in cecal and colonic microbial communities in the TTO (Un-encp TTO and Encp TTO) and PC treatments was higher than that in the NC treatment (Figures 3C,D), whereas Escherichia–Shigella were lower (p < 0.05, Figures 4B,D). Compared with the Un-encp TTO and PC treatments, pigs fed the Encp TTO diets had higher relative abundance of Clostridium_sensu_stricto_1 in the colonic digesta (Figure 3D) and lower relative abundance of Escherichia–Shigella (Figure 4B) in the cecal and colonic digesta (p < 0.05). Moreover, the relative abundance of Streptococcus in cecum and colon was not altered among all treatments (Figures 4A,C).