In this study, a total of 40 pregnant sows (Large White × Landrace) with the average parity of 4 at day 90 of gestation were obtained from a commercial pig farm (Sichuan, China). All sows were assigned to two groups: Control group (CONm, basal diet, n = 20), and COS group (COSm, basal diet supplemented with 100 mg/kg COS, n = 20). The basal diets were formulated to meet the nutrient requirements of gestating sows or lactation sows, respectively, as recommended by the National Research Council (NRC, 2012). The compositions and nutrients levels were shown in Tables 1, 2.

Nutrient levels in the experimental diet were measured according to procedures outlined by the Association of Official Analytical Chemists (2007) for the crude protein (method 976.05). The calcium (Ca) (GB/T 6436–2018) and phosphorus (P) (GB/T 6437–2018) levels were analyzed according to China National Standard, and for Lys using ninhydrin for postcolumn derivatization by an Amino Acid Analyzer (Hitachi High Technologies America Inc., Pleasanton, CA, USA), respectively. The metabolic energy was calculated according to the NRC (2012). The sows were provided with 3 kg of gestation diet daily (8,00 and 14,00) from day 91 of gestation until parturition. Following delivery, sows were fed with lactation diet and increased gradually until the sows could feed ad libitum (8,00, 14,00, and 21,00) from day 1 post-farrowing to day 24. Within the first 48 h after farrowing, the piglets were weighed and then the litter sizes were adjusted to 12 ± 1 piglets per litter by cross-fostering in the same treatment.

All piglets were individually weighed and weaned at 24 days post-birth. A 2 × 2 factorial treatment arrangement was employed with two fixed factors: Sow diet (CONm (basal diet) or COSm (basal diet + 100 mg/kg COS)) and piglet diet (CONp (basal diet) or COSp (basal diet + 100 mg/kg COS)). Within each sow treatment (CONm/COSm), piglets from multiple litters were assigned to piglet diets (CONp/COSp) based on body weight and gender. The piglets from multiple assigned litters were mixed to form replicate pens for each treatment combination, ensuring that littermates were distributed to different pens within their dietary treatment where possible. Thus, there were four piglet treatments: CONm-CONp, CONm-COSp, COSm-CONp and COSm-COSp. Each treatment consisting five replicates with 15 piglets per replicate, with sex strictly balanced such as half male and half female where possible. The piglet basal diet was formulated based on NRC 2012 recommendations (Table 3). During the 14-day period, the piglets had free access to feed and water. Feed intake of each replicate was recorded daily, and the body weight (BW) of each piglet was measured at the end of the experiment to calculate the average daily feed intake (ADFI), average daily gain (ADG), and the gain-to-feed ratio.

Chitosan oligosaccharide used in this study had a deacetylation degree of approximately 95% and an average molecular weight below 1,000 Da with maltodextrin is used as a carrier, it was provided by Zhongke Rongxin Biotechnology Co., Ltd. (Suzhou, China), with dosage based on prior research in sows (14).

After parturition, the total and average birth weight of piglets were recorded prior to nursing (1 sow/pen, n = 20). After injecting oxytocin (1 mg) intravenously, colostrum samples (2 h post-farrowing) and milk samples (day 7 of lactation) were collected manually from 8 sows with similar body weight, parity, and backfat thickness per diet, and stored at −20 °C until analysis.

During lactation, six male piglets from each group (per piglet from per sow) were bled on d 2 and d 24 of age, respectively. Blood samples of piglets were collected using vacuum blood collection tube with EDTA from the jugular vein. The samples were then centrifuged at 3000 × g for 10 min at 4 °C, and the plasma was subsequently collected and stored at −20 °C for later analysis.

Additionally, six male piglets with average BW per treatment were used for intestinal sampling on d 2, d 24 and d 38, respectively. The piglets were euthanized via jugular puncture under anesthesia with an intravenous injection of a 4% sodium pentobarbital solution. Mucosal scrapings of jejunum and ileum were rapidly snap-frozen and stored at −80 °C for further analysis. Furthermore, on d 38, the entire small intestines were immediately removed from the abdomen of each piglet, removed digesta in the intestine with physiological saline, weighed and measured for the length of jejunum and ileum. Jejunal tissues samples were fixed in a 4% paraformaldehyde solution for morphological analysis. Cecum contents were quickly collected and stored at −80 °C for future analysis of gut microflora and short-chain fatty acids (SCFA).

Reproductive performance of each sow was assessed by monitoring the number of piglets born (total, alive and stillborn) and the number of piglets with a birth weight less than 1,000 g. Individual piglet weights at birth and weaning were recorded. Daily feed intake during lactation was recorded for each sow. The weaning-to-estrus interval from weaning to estrus was also recorded.

The fat, protein, lactose, and solids-not-fat content of the colostrum and milk samples were measured with an automated milk analyzer (Milk-Yway-CP2; Zhejiang, China), with results reported as percentage. The concentrations of transforming growth factor -β1 (TGF-β1) in colostrum and milk were determined using a specific assay kit (Beijing 4A Biotech Co., Ltd., Beijing, China) according to the manufacturer’s instructions.

The commercial available assay kits were used for ELISA measurements of plasma fructose and immunoglobulin A (IgA) concentrations (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer’s instructions. Plasma concentrations of triglyceride (TG) and immunoglobulin G (IgG) were determined using corresponding commercial kits (Sichuan Maker Biotechnology Inc., Chengdu, China) with an automatic biochemical analyzer (Hitachi 7,020, Hitachi High Technologies Corporation, Tokyo, Japan).

The densities of jejunum and ileum were calculated using the formula: intestinal density (g/m) = intestinal weight/intestinal length. The jejunal samples were fixed in paraformaldehyde solution, rinsed with physiological saline, and then embedded in paraffin wax. Sections were dewaxed, hydrated, and then stained with hematoxylin and eosin for morphology determination. Ten well-orientated and intact villi and crypts were measured with an image processing and analysis system (Image-Pro Plus 6.0, Media Cybernetics, USA), and the mean villus height and crypt depth, as well as the villus height-to-crypt depth ratio were calculated. The goblet cell counts per villus were also assessed.

Ileal mucosal were homogenized in ice-cold physiological saline (1:9, w/v) and then centrifuged at 3000 × g for 15 min at 4 °C. The supernatants were used to measure the interleukin-10 (IL-10), interferon-γ (IFN-γ), and secretory immunoglobulin A (sIgA) levels according to the manufacturer’s instructions by using the respective ELISA kits (Beijing 4A Biotech Co., Ltd., Beijing, China).

Gene transcriptional abundances of intestine were detected using reverse transcription polymerase chain reaction (RT-PCR). Briefly, total RNA was extracted with Trizol reagent and reverse-transcribed into cDNA using the Prime-Script® RT Master Mix kit (Takara Bio Inc., Dalian, China) following the manufacturer’s instructions. Real-time PCR assays were performed with SYBR® green system (Takara Bio Inc., Dalian, China) on complementary DNA samples using a QuanStudio 6 Flex Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The most stable housekeeping genes (HPRT1, PPIA, and TBP) were chosen for normalization, and the relative mRNA abundance of the analyzed genes was calculated using the 2^−ΔΔCT method (15). Primer sequences for individual genes were listed in Supplementary Table S1.

Total DNA was extracted from caecum digesta samples using the EZNA® Stool DNA kit (Omega Bio-Tek, Doraville, CA, USA) following the manufacturer’s instructions. Quantitative RT-PCR for Total bacteria, Lactobacillus spp. and Escherichia coli was performed with fluorescent oligonucleotide probes using the commercial available SuperReal Premix Kit (Tiangen Biotech, Ltd., Beijing, China). Amplification was performed using a CFX 96 System (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Primer sequence data for Total bacteria forward primer (5′-3′) was ACTCCTACGGGAGGCAGCAG, reverse primer (3′-5′) was ATTACCGCGGCTGCTGG, Lactobacillus spp. forward primer (5′-3′) was GAGGCAGCAGTAGGGAATCTTC, reverse primer (3′-5′) was CAACAGTTACTCTGACACCCGTTCTTC, and for Escherichia coli forward primer (5′-3′) was CATGCCGCGTGTATGAAGAA, reverse primer (3′-5′) was CGGGTAACGTCAATGAGCAAA. The results are expressed as the percentage of Lactobacillus and Escherichia coli within the total bacterial population.

Extraction of SCFA, including acetate, propionate and butyrate, from cecum digesta by organic solvent was performed as described in previous study (16). Cecum digesta (approximately 1 g) was mixed with deionized water and allowed to stand for 30 min at 4 °C before being centrifuged (1,000 × g, 4 °C) for 10 min. The supernatant was then mixed with 0.24 mL of 25% metaphosphate and centrifuged again (1,000 × g, 4 °C) for another 10 min to obtain the final supernatant, which was subsequently prepared for gas chromatography analysis (Varian, Inc., Artaud, CA, USA).

The data on reproductive performance and milk composition in sows, plasma parameters, jejunal relative mRNA expression, and ileal immunity parameters in suckling piglets (day 2 and day 24 of age) were analyzed using a t-test within SPSS 22.0 (IBM SPSS Company, Chicago, IL, USA). The growth performance, intestinal morphology, jejunal relative mRNA expression, ileal immunological parameters, and cecal microbes and SCFA productions in weaning piglets (d 38 of age) were analyzed as a completely randomized design with a 2 × 2 factorial treatment arrangement by ANOVA using the GLM procedure of SPSS, and Tukey multiple comparisons were performed when interactions were significant. The statistical model included the main effects of sow dietary COS treatment (CONm or COSm), piglet dietary COS treatment (CONp or COSp), and their interactions. The growth performance of weaning piglets (d 38 of age) was analyzed with the pen as the statistical unit, while all other parameters were evaluated using individual sows/piglets as the experimental unit. All results obtained in this experiment were expressed as means and the SEM. Differences were considered as statistically significant when p < 0.05. The 0.05 ≤ p ≤ 0.1 was considered a trend.

As shown in Table 4, maternal COS supplementation decreased the weaning-to-estrus interval in sows (p < 0.05) and tended to increase the number of piglets with a birth weight greater than 1.0 kg (p = 0.081) and the piglet ADG before weaning (p = 0.100). However, there was no effect on the numbers of total born, born alive, and stillborn piglets and mean piglet weight at birth due to maternal COS supplementation, and no treatment effect was observed for sows’ ADFI during lactation.

According to Table 5, there were no differences in the concentrations of fat, protein, lactose, solids-not-fat, and TGF-β1 in colostrum between CONm sows and COSm sows. Maternal COS supplementation increased (p < 0.05) the milk contents of protein, lactose, solids-not-fat. The TGF-β1 tended to be higher (p = 0.053) in COSm group compared to CONm group.

As shown in Figure 1, compared with nursing piglets from CONm sows, nursing piglets from COSm sows exhibited higher concentrations of plasma fructose at 2 days of age (p < 0.05, Figure 1A), as well as tended to elevate concentrations of fructose (p = 0.072) and TG (p < 0.05, Figure 1B) in plasma at 24 days of age. However, the contents of plasma IgG (Figure 1C) and IgA (Figure 1D) in nursing piglets were not affected by dietary COS supplementation in sows.

Supplementation of COS in sow diet resulted in increased jejunal mRNA levels of CLDN3 (p < 0.05, Figure 2A), SDCBP2 (p < 0.05, Figure 2D), G6PC (p = 0.083, Figure 2E), LDHA (p < 0.05, Figure 2G), MAP2K1 (p < 0.05, Figure 2H) and PCK1 (p < 0.05, Figure 2J) but decreased the mRNA level of IGF1R (p < 0.05, Figure 2F) in 2 days of age nursing piglets. The mRNA levels of jejunal CLDN7 (p < 0.05, Figure 2B) and HNF4A (p < 0.05, Figure 2C) were increased, and mRNA level of SLC27A2 (p = 0.092, Figure 2L) tended to increase in nursing piglets from COS sows at 24 days of age.

As shown in Figure 3, elevated IL-10 and IFN-γ (p < 0.05) levels were observed in the ileum of suckling piglets from COSm sows compared to CONm sows at 2 days of age, and the ileal IL-10 (p < 0.05) and IFN-γ (p = 0.058) levels were increased in piglets at age of 24 days from COSm sows. However, the levels of sIgA in ileal mucosa of piglets were not affected by COS diet in sow. Moreover, the ileal MHCII mRNA expression (p = 0.076) tended to increase in suckling piglets from COSm sows compared to CONm sows at 2 days of age, and the ileal mRNA levels of CD163 (p < 0.05) and MHCII (p = 0.074) were increased in piglets at age of 24 days from COSm sows.

As shown in Figure 4, the supplementation of COS after weaning did not have a effect on piglet ADFI, ADG, and gain-to-feed ratio.

Based on Table 6, sows supplemented with COS improved the density of the jejunum and ileum in weaned piglets (p < 0.05). There was an interaction between sow dietary COS and piglet dietary COS on ileal density (p < 0.05) as well as jejunal crypt depth (p < 0.05), and there was a tendency of interaction between sow dietary COS and piglet dietary COS on jejunal villus height-to-crypt depth ratio (p = 0.057). The group supplemented with both maternal and piglet dietary COS had the highest ileal density compared to CONm-COSp and CONm-CONp groups (p < 0.05). The jejunal crypt depth of piglets from the groups of CONm-COSp and COSm-CONp were lower than that of CONm-CONp group (p < 0.05). No differences were observed in the goblet cells and villus height between different treatments in weaning piglets.

As shown in Table 7, COS supplementation in sows increased the mRNA level of PPARGC1A (p < 0.05), and tended to evaluate mRNA levels of LDHA (p = 0.053) and CD163 (p = 0.084), while decreased the mRNA levels of CCL2 (p = 0.094) with tendency in weaned piglets at d 38 of age. In weaned piglets, COS supplementation in nursery diets increased the mRNA levels of IGF1R (p = 0.074) and CD163 and HNF4A (p < 0.05), but decreased the mRNA level of MHCII (p = 0.002). An interaction between sow dietary COS and piglet dietary COS was observed for jejunal mTOR expression (p < 0.05). The mRNA levels of CLDN3, CLDN7, SDCBP2, G6PC, MAP2K1, PCK1, TGFB2 and SLC27A2 were not affected by dietary COS supplementation in either sows or piglets.

The results of ileal immunological parameters in weaning piglets were presented in Table 8. There were no interactions between sow dietary COS and piglet dietary COS on ileal IL-10, IFN-γ, and sIgA. Meanwhile, piglets from COS-treated mothers exhibited lower concentrations of ileal IL-10 and IFN-γ compared to non-treated CONm piglets (p < 0.05).

The Table 9 presented the Lactobacillus spp., Escherichia coli, and SCFA levels in cecal chyme. Dietary supplementation of COS in sows increased the Lactobacillus spp.-to-Escherichia coli ratio (p < 0.05), butyrate (p < 0.05), and propionate (p < 0.05), as well as increased total SCFA in tendency (p = 0.071), while decreasing the concentration of Escherichia coli (p < 0.05) in the cecum. Chitosan oligosaccharide supplementation in piglets resulted in an increase (p < 0.05) in acetate, propionate and total SCFA. No interaction between sows and piglets supplemented COS was found among these cecal-related indexes.