Differential Effects of Breed and Nursing on Early-Life Colonic Microbiota and Immune Status as Revealed in a Cross-Fostering Piglet Model

Early-life gut microbiota and immune status are pivotal for postnatal growth. By using an interspecific cross-fostering piglet model, we find that change in nursing mother transiently reshapes preweaning colon microbiota and immune status, while breed shows persistent effects both pre- and postweaning. Piglets nursed by Meishan sows had lower Streptococcus suis counts and higher anti-inflammatory cytokine expression. These results highlight the significance of nursing mother in regulating early-life gut health.

ABSTRACT

Nursing mother and breed can differently regulate early-life microbiota succession in pigs. However, it remains unclear whether they affect gastrointestinal microbiota and immune status, which are critical for early-life gut health. Here, an interspecific cross-fostering piglet model was employed by fostering neonatal Yorkshire and Meishan piglets to the same or another breed of sows. Jejunal and colonic microbiotas and mucosal immune parameters were analyzed at postnatal days 14 (preweaning) and 49 (postweaning). Nursing mother affected 10 genera in the colon and 3 minor genera in the jejunum. At day 14, Meishan sow-nursed piglets had lower Streptococcus suis and higher Cloacibacillus counts in the colonic digesta and larger amounts of interleukin 10 and Foxp3-positive cells in the colonic mucosa than did Yorkshire sow-nursed piglets. At day 49, nursing mother had no significant effects on cytokine expression. Breed effects were observed; Meishan piglets had lower relative abundances of Prevotella and lower gene expression of tumor necrosis factor alpha (TNF-α) than those of Yorkshire piglets at days 14 and 49. Collectively, nursing mother mainly affected preweaning colonic microbiota and immune status, while breed effects persisted after weaning. Piglets nursed by Meishan sows had different microbiota compositions and inflammatory cytokine profiles in the colon compared with those of piglets nursed by Yorkshire sows. These results highlight the different role of nursing mother and breed in affecting early gut microenvironment.

IMPORTANCE Early-life gut microbiota and immune status are pivotal for postnatal growth. By using an interspecific cross-fostering piglet model, we find that change in nursing mother transiently reshapes preweaning colon microbiota and immune status, while breed shows persistent effects both pre- and postweaning. Piglets nursed by Meishan sows had lower Streptococcus suis counts and higher anti-inflammatory cytokine expression. These results highlight the significance of nursing mother in regulating early-life gut health.

INTRODUCTION

Colonization of gut microbiota in early life is important for host health. The maternal environment during gestation and lactation affects offspring microbial colonization (1). The placenta was reported to have microbiota and could potentially affect early microbiota colonization (2). However, this idea is highly controversial and not universally accepted (3). Another study, using a different sample set, revealed that the microbiota in placental samples and contamination during DNA extraction could not be distinguished (4), which is completely contradictory to the results obtained by Aagaard et al. (2). Thus, the concept of microbiota in the placenta or in utero is still under debate (3). The discovery of Proteobacteria and Firmicutes species in the piglets’ umbilical cord blood also implies the vertical transfer of maternal microbiota to offspring (5). Maternal environment tends to be a key factor determining early-life gut microbiota colonization.

Recent evidence has indicated the importance of nursing mother in modulating offspring gut microbiota (6). Using cross-fostering experiments, previous studies found that different breeds of piglets showed certain identical microbial alterations in cecum content (7) and feces (6) when fostered by a nursing mother of a given breed. The nursing mother has also been found to shape the fecal microbiota composition in nonobese diabetes-resistant and nonobese diabetic mice, both before and after weaning (8). All of these studies highlight the great relevance of nursing mother to early-life gut microbiota.

Breed is another important factor affecting gut microbiota composition. The fecal microbial communities in Meishan and Yorkshire piglets showed that breed had a significant impact on the bacterial community structure on day 14 and day 49 (6). However, how breed affects gut microbiota inside the gut remains unclear.

Gut microbes can interact with the host via providing fermentation products and regulating immune status (9). Among the microbial metabolites, butyrate participates in energy metabolism (10) and regulatory T cell differentiation in the colon (11). However, whether cross-fostering affects microbial activity and epithelial immune status in offspring remains unclear. In the present study, we used a cross-fostering piglet model to specifically focus on the colon, where dense host-microbe interactions happen in the gut. Two different pig breeds, Meishan and Yorkshire, were used. Meishan is a Chinese domestic obese breed and Yorkshire is a lean-type breed. Meishan sows had a higher prolificacy than Yorkshire sows. An earlier study found that Meishan sows had 12.4 fully formed piglets, while Yorkshire had 7.4 fully formed piglets (12). In addition, Meishan and Large White (Yorkshire) pigs differ in innate immune traits; for example, Meishan pigs have higher neutrophil and monocyte counts and lower lymphocyte counts (13). Therefore, it would be interesting to investigate whether nursing by Meishan sows may change the immune status of Yorkshire piglets.

Using the cross-fostering model, we found negative correlations between high concentration of milk lactose in Meishan sows and low relative abundance of potential pathogens, such as Haemophilus and Staphylococcus, in the feces of piglets (6). This novel finding raises the hypothesis that piglets might benefit from being nursed by Meishan sows, possibly through alteration in gut microbiota and epithelial immunity. To test the hypothesis, microbial composition in the jejunum and colon and cytokine expression were analyzed at day 14 (before weaning, exclusively suckling) and day 49 (after weaning, solid feed). We show that, relative to Yorkshire sow-nursed piglets, Meishan sow-nursed piglets had lower relative abundances of Streptococcus species and higher expression of interleukin 10 (IL-10) in the colon. These data further highlight the key role of nursing mother and breed in regulating early-life microbial communities inside the offspring’s gut.

RESULTS

In the present study, day 14 was chosen as a sampling time point in order to identify the nursing effect (discriminating from the solid feed effect). Solid feed was introduced on postnatal day 14, and the piglets received both solid food and sow’s milk until postnatal day 28. Solid food is one of the most important external factors affecting the gut microbiota of the offspring piglets (6) and infants (14). Postnatal day 49 was a key growing time point after weaning. Thus, in the present study, postnatal days 14 and 49, but not day 28, were chosen for sampling.

Day 0 was not included for tissue sampling. Our earlier studies found that the major difference of microbiota composition at birth is due to breed, as Yorkshire piglets had higher relative abundances of Escherichia species in feces at day 1 than that of Meishan piglets (6). In the present study, to dissect the breed effects after day 0, we included the breed as a major variable affecting the microbiota composition at days 14 and 49, which could objectively reflect the difference of microbial colonization in piglets. Since Meishan is a domestic breed, the cost to slaughter one piglet of a pure breed on the birth day is high. Therefore, postnatal days 14 and 49 were chosen as two time points to collect jejunal and colonic samples.

Effects of cross-fostering on microbial compositions in jejunal and colonic digesta.

To profile microbial compositions, we analyzed the gut microbiota in piglets of different ages (postnatal days 14 and 49) and gut locations (jejunal and colonic lumen). Five samples from jejunum and three samples from colon were not sequenced due to poor DNA quality for Illumina sequencing analysis. After initial filtering, at day 14, 5, 6, 4, and 6 samples from the jejunum and 6, 6, 5, and 6 samples from the colon were used for Yorkshire piglets fostered by their birth mother (Yy), Meishan piglets fostered by Yorkshire sows (Ym), Meishan piglets fostered by their birth mother (Mm), and Yorkshire piglets fostered by Meishan sows (My), respectively. At day 49, 6, 6, 6, and 4 samples from the jejunum and colon were used for Yy, Ym, Mm, and My groups, respectively. The PCR amplicon electropherogram of all samples used can be found in Fig. S1. In total, 3,320,275 qualified bacterial 16S rRNA gene reads were obtained from 88 samples (37,730 reads per sample) after Illumina sequencing and were used for subsequent analysis. The coverage was higher than 99% for all the groups. The Chao1 index and Shannon and Simpson diversity indices did not differ between groups in either the jejunal or colonic samples (Table S2). Principal-coordinate analysis (PCoA) results showed that the first coordinate explained 37.75% of the variance and the second coordinate explained 26.3% of the variance (Fig. 1B). Samples were clustered depending on gut location and ages (Fig. 1B, Fig. S2). The results representing the Chao1 and Shannon indices are also shown (Fig. 1C and D). The microbial composition at the phylum level is shown in Fig. 1E. Differences can be observed between days 14 and 49 in the jejunal digesta and between the jejunum and colon. At the genus level, a total of 515 genera were found, while 359 could be assigned to specific taxa, the relative abundance of which constituted 74.4% of the composition. At the species level, a total of 913 species were revealed, while only 257 could be assigned to known species, the relative abundance of which constituted 40.4% of the composition.

FIG 1.

(A) Schematic illustration of the four treatment groups in the present study. A half litter of neonatal Yorkshire and Meishan piglets was cross-fostered to the same or opposite breed of sow. The other half of each litter remained with the birth mother. Litters then contained both piglets born to the birth mother and piglets from the opposite Meishan or Yorkshire breed. (B) Unweighted principal-coordinate analysis of all the samples. Sample label J-Yy14 represents jejunum samples from Yorkshire piglets nursed by Yorkshire sows at day 14, etc. C-Yy14 represents colon samples from Yorkshire piglets nursed by Yorkshire sows at day 14, etc. (C) Chao1 diversity index. (D) Shannon diversity index. n.s., nonsignificant between groups. (E) The relative abundances of bacterial phyla in different groups. (F) The effects of breed and nursing mother on the microbial composition at the phylum level on days 14 and 49 in piglets. Each value is a mean ± standard error of the mean (SEM; n = 6 per group). An asterisk (*) represents a significant difference that is associated with the breed of the piglet or that of the nursing mother.

To compare the microbial composition in the colonic digesta of piglets with that in the feces of sows, we further conducted principal-component analysis. As shown in Fig. S3, the samples from piglet colon were clustered separately from those from sow feces, indicating that the microbial composition in piglet colon is different from that in sow feces. Therefore, it is hard to judge if the gut microbiota in the piglets could reflect the microbiota in sow feces.

In the jejunal digesta, nursing mother had no effects on microbiota at the phylum level and only affected the relative abundances of 3 genera. At the phylum level, nursing mother showed no effects. Breed effects were observed, as revealed by the higher relative abundance of Verrucomicrobia species at day 14 and higher relative abundance of Firmicutes species at day 49 in Meishan piglets than Yorkshire piglets (Fig. 1F). Verrucomicrobia is an important phylum containing the genus Akkermansia, which is a key species for mucin utilization in the gastrointestinal tract (15). At the genus level, nursing mother only affected 3 genera (Globicatella at day 14 and Peptostreptococcus and Weissella at day 49), all of which are nondominant genera. However, in the colonic digesta, more phyla and genera were affected by nursing mother at day 14, including the phylum Synergistetes and the dominant genera Streptococcus and Alistipes (Fig. 1F and 2). These results suggested that nursing mother showed marked impact on microbial composition in the colon, but not in the jejunum. Therefore, we then focused on the change of gut microbiota and immune status in the colon.

FIG 2.

Effects of breed and nursing mother on the microbial composition at genus level in the jejunal and colonic digesta of piglets. The mean relative abundance value of each taxon is shown. The red and blue colors represent high and low relative abundance, respectively.

In the colonic digesta, both nursing mother and breed affected microbial composition at days 14 and 49. Regarding the effects of nursing mother, at day 14, piglets nursed by Meishan sows had higher relative abundance of Synergistetes (Fig. 1F), especially Cloacibacillus (P < 0.05; Fig. 2), than those nursed by Yorkshire sows. The genus Streptococcus was predominant, especially in Meishan sow-nursed piglets. Compared to piglets nursed by Yorkshire sows, the relative abundance of Firmicutes species tended to decrease in piglets nursed by Meishan sows (Fig. 1F), while the relative abundance of Streptococcus species was significantly decreased (Fig. 2). Additionally, piglets nursed by Meishan sows had higher relative abundances of unclassified Prevotellaceae, Phascolarctobacterium, and Alistipes than those nursed by Yorkshire sows. At day 49, the effects of nursing mother weakened, as revealed by the higher relative abundance of unclassified Micrococcaceae and the lower relative abundance of unclassified Ruminococcaceae in piglets nursed by Meishan sows than in those nursed by Yorkshire sows (P < 0.05; Fig. 2). Effects of breed on microbial composition were also observed. At day 14, Meishan piglets had higher relative abundances of Actinobacteria and Collinsella and lower relative abundances of Prevotella and Desulfovibrio than those of Yorkshire piglets (P < 0.05; Fig. 1F and 2). At day 49, Meishan piglets had lower relative abundances of Phascolarctobacterium, Subdoligranulum, and Prevotella than those of Yorkshire piglets (Fig. 2). At the genus level, we present all the genera that are affected by the nursing mother or breed. Although the relative abundances are low, some genera may exert biological relevance in gut. For example, Ruminococcus is an important genus with carbohydrate-fermenting ability that produces short-chain fatty acids (SCFA) (16). Enterococcus is a common commensal in mammal gut.

Results representing microbial composition at the species level further showed that piglets nursed by Meishan sows had a 9.69-fold higher relative abundance of an unclassified Cloacibacillus species and lower relative abundance of Streptococcus suis at day 14 than those nursed by Yorkshire sows (P < 0.05; Fig. 3). Interestingly, piglets nursed by Meishan sows had a lower relative abundance of Streptococcus gallolyticus subsp. macedonicus than those nursed by Yorkshire sows at days 14 and 49. These results indicated the significant main effects of cross-fostering in altering early-life colonic microbial communities.

FIG 3.

Effects of cross-fostering on the representative phylotypes in the colonic digesta of piglets. Each value is a mean ± SEM (n = 6 per group). An asterisk (*) represents a significant difference that is associated with the breed of the piglet or that of the nursing mother.

Effects of cross-fostering on microbial fermentation product.

Short-chain fatty acids are carbohydrate fermentation products of gut microbes and serve as indicators of microbial activity. Measurement of short-chain fatty acids in the colonic digesta found that both nursing mother and breed did not affect the concentrations of acetate, propionate, and butyrate in the colon at day 14 (P > 0.05; Fig. 4). At day 49, no effects were observed for nursing mother, whereas the main effect of breed was observed. Meishan piglets had lower butyrate concentrations than those in Yorkshire piglets (P < 0.05). These data indicated that cross-fostering had minor effects on microbial fermentation in the colon.

FIG 4.

Concentrations of short-chain fatty acids in colonic digesta at days 14 and 49. Each value is a mean ± SEM (n = 6 per group). An asterisk (*) represents a significant difference (P < 0.05) that is associated with the breed of the piglet or that of the nursing mother.

Effects of cross-fostering on cytokine expression and cell numbers.

Gut microbiota and metabolites can regulate epithelial gene expression and cell populations. Therefore, we analyzed the expression of genes involved in mucosal defense and cytokine production and the numbers of regulatory T cells (Fig. 5A). At day 14, nursing mother affected the mRNA expression of interleukin 10 (IL-10) in the colonic mucosa (P < 0.05). Piglets nursed by Meishan sows had higher mRNA expression of IL-10 than those nursed by Yorkshire sows (P < 0.05). Breed affected the mRNA expression of tumor necrosis factor alpha (TNF-α) and antimicrobial peptide PMAP-23 (P < 0.05). Meishan piglets had lower mRNA expression of TNF-α at days 14 and 49. In addition, Meishan piglets also had lower mRNA expression of PMAP-23 at day 49 (P < 0.05).

FIG 5.

(A) Expression of genes involved in antimicrobial defense and cytokine production in the colonic mucosa. (B and C) Protein concentrations of IL-10 (B) and TNF-α (C) in the colonic mucosa. Each value is a mean ± SEM (n = 6 per group). An asterisk (*) represents a significant difference (P < 0.05) that is associated with the breed of the piglet or that of the nursing mother.

Analysis of cytokine protein expression by enzyme-linked immunosorbent assay (ELISA) further validated the effect of nursing mother on IL-10 (Fig. 5B). At day 14, piglets nursed by Meishan sows had higher concentrations of IL-10 in the colonic mucosa than those of piglets nursed by Yorkshire sows (P < 0.05), while nursing mother had no impact at day 49. Breed, but not nursing mother, had main effects on TNF-α expression (Fig. 5C). Meishan piglets had lower concentrations of TNF-α than those of Yorkshire piglets at day 14 (P < 0.05) and no difference was observed between groups at day 49.

Foxp3-postive regulatory T cells are responsible for secreting IL-10. Therefore, we measured the numbers of Foxp3-postive cells in colon mucosa. At day 14, piglets nursed by Meishan sows had higher numbers of Foxp3-postive cells in the colonic mucosa than those nursed by Yorkshire sows (P < 0.05), while no impact was detectable at day 49 (Fig. 6).

FIG 6.

Immunohistochemical analysis of Foxp3-positive cells in the colonic mucosa. The integrated optical density (IOD)/area is shown. An asterisk (*) represents a significant difference (P < 0.05) that is associated with the breed of the piglet or that of the nursing mother.

Concentration of mucosal IgA.

In addition to defensive peptides, cytokines, and regulatory T cells, immunoglobulin A (IgA) is an important component linking host-microbiota interactions in the gut (17). IgA was measured as an indicator of mucosal immunity. Nursing mother did not affect the concentration of IgA in the colonic mucosa (Fig. 7). At day 14, the effect of breed was observed in that Meishan piglets had higher concentrations of IgA than those of Yorkshire piglets. No difference was detectable at day 49.

FIG 7.

Concentrations of immunoglobulin A in the colonic mucosa. Each value is a mean ± SEM (n = 6 per group). An asterisk (*) represents a significant difference (P < 0.05) that is associated with the breed of the piglet or that of the nursing mother.

DISCUSSION

The role of maternally provided environment in offspring development has recently received more research attention. In this study, we demonstrated in a piglet model that nursing mother has major impacts on early colonic microbiota and the mucosal immune system but minor effects on gut microbiota in the jejunum. A distinct microbial community with a lower relative abundance of Streptococcus species was characteristic for Meishan sow-nursed piglets. Some effects of nursing mother on the microbial composition even persisted after weaning. Of particular interest is the result that higher expression of anti-inflammatory cytokine IL-10 and Foxp3-positive cells was found in Meishan sow-nursed piglets. All of these results further highlight the importance of nursing mother in regulating offspring colonic microbiota at early life.

Using cross-fostering piglet models, it has been demonstrated that nursing mother had diverse effects on offspring, including growth performance (18), fecal microbial succession (6), and cecal microbial composition (7). In the present study, our results provide new evidence that colonic homeostasis could also be modulated by nursing mother.

In the colon, the nursing mother was found to affect the dominant Streptococcus species at day 14, including Streptococcus suis and Streptococcus gallolyticus, with the Meishan sow lowering the relative abundances of these species. Depending on the serotype, S. suis can be highly pathogenic, weakly pathogenic (hypovirulent), or nonpathogenic (19). In piglets, S. suis generally colonizes the upper respiratory tract, such as the tonsils and nasal cavities. Other tropism locations include genital and alimentary tracts of piglets (19). S. suis has been detected in the small intestine of pigs by real-time PCR (20) and Illumina sequencing (21) and in the large intestine of the pig by phylogenetic microarray-based pig intestinal tract chip (PITChip) analysis (22). In the present study, to avoid sample contamination from blood or upper respiratory fluids, the intestine was carefully dissected and separated for sampling digesta. Our results in the present study confirmed the presence of S. suis in the gut of the pig.

Among the bacteria affected by the nursing mother, we noted an increase in a species similar to Cloacibacillus porcorum. Although the relative abundance is below 1% of the microbial community, it increased by almost 10-fold in Meishan sow-nursed piglets. This bacterium was first identified as a mucin-degrading bacterium from porcine cecum (23). It can use mannose, fucose, and sialic acid for growth (23). Both mucin and sow milk can provide these substrates. Our previous study found that milk of Meishan sows had a higher relative abundance of lactose than that of Yorkshire sows (6). Whether differences in milk lactose composition may be involved in the increase of Cloacibacillus porcorum is an interesting question to be further studied. The physiological effect of Cloacibacillus porcorum is as yet unclear. Nevertheless, many mucin-degrading bacteria, such as Akkermansia muciniphila (24) and the newly identified species Peptostreptococcus russellii (25), have beneficial functions in host metabolism and immune status. These results suggest the potential importance of the mucin-degrading bacterium Cloacibacillus porcorum in regulating the gut environment.

In the present study, nursing mother and breed had no significant effects on microbial diversity and richness, as shown by the results representing Chao1 and Shannon indices, whereas they significantly changed the relative abundances of some taxa. Namely, the composition of the gut microbiota, rather than its diversity, was greatly affected. It is possible that the same solid feed may lead to similar alpha diversity, as all the piglets across groups received the solid feed after day 14. Food composition is a major driver of microbial structure.

Together with the alteration in microbial composition, nursing mother further affected colonic immune status. Interestingly, fostering by Meishan sows increased IL-10 expression and Foxp3-positive cells in piglets relative to fostering by Yorkshire sows. IL-10 is a well-known anti-inflammatory factor to suppress intestinal inflammation (26) and pathogen invasion (27). Upregulation of IL-10 has been related to a decreased incidence of rotavirus-induced diarrhea in piglets fed with human milk oligosaccharides (28). In a study comparing breastfeeding and formula-fed neonatal piglets, breastfeeding increased IL-10 expression in the ileum mucosa and decreased Streptococcaceae abundance and diarrhea incidence (29). Foxp3-positive cells constitute a major population of regulatory T cells, which exert anti-inflammatory function by secreting Il-10 (30). Collectively, an increase of IL-10 and Foxp3-positive cells can be possibly linked with an improved anti-inflammatory function in piglet epithelium. These findings may support our hypothesis that piglets might benefit from being nursed by Meishan sows.

In the present study, the breed had major effects on the microbial composition in the colon on day 14 and 49. Meishan piglets had a higher relative abundance of Enterococcus species in the colon digesta than that of Yorkshire piglets, which was also found in an earlier study of fecal microbiota (6). These results further reinforce our earlier finding that breed was an important factor in shaping gut microbiota in the pig (6). Another noteworthy result is the identification of specific taxa and cytokine expression affected by breed in our data. A typical taxon is Prevotella, which was less abundant in Meishan piglets than in Yorkshire piglets at both days 14 and 49. This may reveal their selective adaptation to different host genotypes. (31). In the present study, we found that Meishan piglets had lower expression of proinflammatory TNF-α and higher IgA levels in colonic mucosa at day 14. IgA is important for the mucosal immune system to maintain the intestinal barrier function (32). The higher IgA levels may aid the epithelial defense against potential pathogens. It might be speculated that an increase in IgA levels may be a factor accounting for higher disease resistance in Meishan pigs than in Yorkshire pigs.

The breed-related differences may also involve reproductive tract environmental differences. Vaginally delivered infants have a bacterial composition similar to their own mother's vaginal microbiota, e.g., with Lactobacillus spp. dominant (33). This can be a factor accounting for the early-life microbiota difference. Although the microbiota in the reproductive tract of the sow has not yet been profiled, the fecal microbiota differs in Meishan and Yorkshire sows (6). Meishan sows have higher relative abundances of Firmicutes but lower abundances of Bacteroidetes than those of Yorkshire sows (6). This may potentially affect postnatal development in piglets.

In conclusion, we have demonstrated that breed affects colonic microbiota and immune status at postnatal days 14 and 49. Nursing mother has a weakened imprint on microbial composition after weaning, and no imprint was found for the immune parameters studied in the present work. Piglets nursed by Meishan sows had a higher expression of IL-10. These results highlight the key role of nursing mother in shaping early-life microbial communities inside the offspring gut. Findings from the current study further provide reference for early nutritional interventions to regulate gut microbiota and immune status.

MATERIALS AND METHODS

Experimental design.

All Meishan and Yorkshire pigs were raised under the same conditions on a commercial farm in Jiangsu Province, China. This study was approved by the Animal Care and Use Committee of Nanjing Agricultural University, in compliance with the Regulations for the Administration of Affairs Concerning Experimental Animals, 1998. This cross-fostering operation was described in our previous study (6). A randomized complete block design (2 treatments [breed] × 2 blocks [nursing], with 10 replicates per group) was adopted. Briefly, sows from the two breeds with similar expected delivery dates were chosen in the study. After the vaginal delivery and before suckling the colostrum, the piglets in a litter from one sow of one breed were split into two groups (half were moved to the fostering sow of another breed and half remaining with the sow), according to their equal average body weights and balanced sexes. Furthermore, an earlier study found no effects of sex on the fecal microbiome at day 20 before weaning in piglets (34). Therefore, sex was not a confounding variable for the piglets in the present study. Only candidate pairs of sows that differed in delivery time by less than 2 h were included in this study. The cross-fostering operation created the following four groups: Meishan piglets fostered by their birth mother (Mm), Yorkshire piglets fostered by Meishan sows (My), Meishan piglets fostered by Yorkshire sows (Ym), and Yorkshire piglets fostered by their birth mother (Yy). Each group contained piglets from ten litters. A total of 10 Meishan sows and 10 Yorkshire sows with a gestational length of 113 to 114 days and a litter size of 10 or 12 piglets each were used for the delivery-matched pairs during the cross-fostering operation. Litters then contained both piglets born to the birth mother and piglets fostered by a fostering mother. Fostered piglets were differentiated by the skin color (black for Meishan piglets and white for Yorkshire piglets). From day 14 after birth, all suckling piglets were offered creep feed ad libitum and had free access to water. All piglets were weaned at day 28 of age. At days 14 and 49 of age, one of each breed of piglet from each of 6 litters was sampled at each time point. To fulfill the minimum requirement of statistical analysis and cost, a total of 6 litters from each treatment were used (n = 6 per group).

Prior to the sampling, all the sampling equipment (scalpels, blades, glass slides, and tubes) was sterilized by autoclave. The sampling operators wore facial masks and sterile gloves during sampling. To prevent cross-contamination during sampling, the sampling operators were grouped and assigned to sample different sites within one pig, and one specific operator would only collect a given specific site from all the pigs. The middle jejunum and proximal colon were segmented and separated using stainless wires. Digesta samples from the middle jejunum and proximal colon were collected and placed immediately into sterilized tubes, snap-frozen, and stored at −80°C for microbial and short-chain fatty acid analyses, respectively. The mucosa was gently scraped off using sterilized glass slides and immediately put into sterilized tubes at −80°C before further analysis. Feces from each sow before delivery were sampled. During sampling, the jejunal and colonic digesta were collected and handled by different operators. For the subsequent measurements, the digesta were thawed only once and then mixed homogeneously. Then 0.3 g of each sample was used for DNA extraction and 0.3 g for metabolite analyses. The leftover samples were refrozen in −80°C. These procedures ensured avoidance of cross-contamination between samples and measurements.

Microbial genomic DNA extraction, 16S rRNA sequencing, and data analysis.

Microbial genomic DNAs were extracted using the bead-beating and phenol-chloroform method (35). DNA quality and quantity were determined using a NanoDrop 2100 spectrophotometer (Thermo Fisher Scientific, WI). The V3 to V4 region of the 16S rRNA gene was amplified with primers (forward, 5′-barcode-TAC GGR AGG CAG CAG-3′, and reverse, 5′-AGG GTA TCT AAT CCT-3′) (36). The barcode is an eight-base sequence specific to each sample. The procedures of 16S rRNA gene amplification and amplicon purification were conducted as previously described (21). Blank controls, in which no DNA was added to the reaction, were included to exclude negative amplifications. DNA extraction controls were not used in the study and could be included in further studies. Purified amplicons were pooled in equimolar and paired-end sequenced (2 × 250 bp) on an Illumina MiSeq platform according to the standard protocols (37).

Raw fastq files were demultiplexed and quality-filtered using QIIME (version 1.17) as previously described (21). After trimming the primer, barcode, and chimeras, the unique sequences were identified and aligned against a high-quality 16S rRNA sequence from the Greengenes reference alignment (38). Operational taxonomic units (OTUs) were clustered with a 97% similarity cutoff using UPARSE (version 7.1; http://drive5.com/uparse/), and chimeric sequences were identified and removed using UCHIME. Diversity indices, including the Chao1, abundance-based coverage estimator (ACE), and Shannon diversity indices, were calculated. Unweighted UniFrac principal-coordinate analysis (PCoA) (39) based on OTUs was performed to examine dissimilarities in community composition. Principal-component analysis was conducted using MetaboAnalyst (40) to compare the microbial composition in the colonic digesta of piglets with that in the feces of sows. The data of microbial compositions in the feces of sows (NCBI SRA accession number SRP066893) were reported in our previous paper (6).

Measurement of short-chain fatty acids.

The concentrations of acetate, propionate, and butyrate in colonic digesta (0.3 g) were measured using gas chromatography with an Agilent 7890A system. The measurement was conducted as described previously (41).

Mucosal RNA extraction and quantitative real-time PCR.

Total RNA was extracted from mucosal scrapings (100 mg) using the RNAiso Plus (TaKaRa Bio, China), in accordance with the manufacturer's instructions. RNA concentration and purity were determined using a NanoDrop 2100 spectrophotometer (Thermo Fisher Scientific, WI). RNA (1,000 ng) was used for reverse transcription using the PrimeScript real-time (RT) reagent kit with gDNA Eraser (Perfect Real-Time) (TaKaRa). The cDNA was used for quantitative real-time PCR analysis using SYBR Premix Ex Taq dye (TaKaRa, Japan) on an ABI StepOne platform (Applied Biosystems). The mRNA expression of genes involved in mucosal defense and cytokine production was studied using the primers in Table S1. Each cDNA was analyzed in duplicate, and the average threshold cycle was calculated. The results were normalized to the expression of the cyclophilin gene, and relative expression levels were calculated using the threshold cycle (2^−ΔΔCT) method.

Immunohistochemical analysis.

Paraffin-embedded sections of colonic tissue were dewaxed and hydrated through a graded ethanol series to distilled water. Endogenous peroxidase was blocked by incubation with 3% H₂O₂ in methanol for 15 min. Antigen retrieval was achieved by microwaving in citrate buffer solution for 4 min followed by a washing step with Tris-buffered saline. The sections were incubated with blocking reagent for 45 min at room temperature to block nonspecific binding sites. Sections were then incubated with the rabbit polyclonal anti-Foxp3 antibody (1:100, catalog no. YT5446; ImmunoWay, Plano, TX) overnight at 4°C. The bound primary antibody was detected by incubation with horseradish peroxidase-conjugated goat anti-rabbit IgG (D110063; BBI, Sangon Biotech, China) for 45 min at room temperature. After rinsing with Tris-buffered saline, the antigen-antibody complex was detected using 3,3′-diaminobenzidine. Sections were then counterstained with hematoxylin. Each slide was microscopically analyzed, and the percentage of the positively stained area was enumerated using Image-Pro Plus 6.0 software. The area of interest and the integrated optical density (IOD) were measured, and the IOD/area was calculated to obtain the mean IOD in each image. Finally, the average of the mean IOD in each image was used to determine the final mean IOD of each piglet.

Enzyme-linked immunosorbent assay.

The concentrations of interleukin 10 (ANG-E31026P; Angle Gene, Nanjing, China), TNF-α (ANG-E31006P; Angle Gene, Nanjing, China) and immunoglobulin A (IgA; ANG-E31164P; Angle Gene, Nanjing, China) in colonic mucosa were determined by using commercial kits according the manufacturers' instructions. The tissue protein concentration was determined using a Bradford assay kit (Solarbio, Beijing, China) with bovine serum albumin as the standard.

Statistical analysis.

All of the data were analyzed by using a full-factorial general linear model (GLM) analysis of variance (ANOVA) with type III sums of squares in SPSS (version 20.0; SPSS Inc., Chicago, Illinois). Breed and nursing mother were used as fixed factors. To avoid type I errors during microbiota analysis, the resulting P value table was adjusted with false-discovery-rate (FDR) correction by the Benjamin-Hochberg method (42). An FDR (q value) of <0.05 was considered to be statistically significant. A q value between 0.05 and 0.1 indicates a trend toward a significant main effect. For data of short-chain fatty acids (SCFA), gene expression, cell numbers, and IgA concentrations, a P value of <0.05 was considered statistically significant.

Data availability.

Raw reads described in this study were deposited into the NCBI Sequence Read Archive (SRA) database under accession number SRP145753.