A total of 42 primiparous German Landrace sows and their litters, bred and raised in the experimental pig unit of our institute, were used for the experiment with 6 independent replicates. Housing and breeding management were as recently described in detail 15 . The sows were fed an isoenergetic corn-barley and soybean meal diet (~13.7 MJ ME/kg) containing an adequate (AP, 12.1%; n = 13), a low (LP, 6.5%; n = 15) or a high (HP, 30%; n = 14) protein level corresponding to protein:carbohydrate ratios of 1:5, 1:10.4, and 1:1.3, respectively, throughout gestation 15 . Diets were fed between 2.3 and 2.9 kg/d from early to late pregnancy to achieve an average target energy intake of ~34 MJ ME/d during gestation following the recommendations for primiparous sows 27 . The sows were fed twice daily, and water was provided ad libitum. Farrowing was induced at gestation day (GD) 114 as described previously 15 .

After overnight food withdrawal, blood samples of sows were taken on GD −5, 24, 66 and 108 by jugular vein puncture. EDTA blood samples were centrifuged at 2000 × g for 15 min at 4°C to separate plasma, which was analysed for cortisol. Whole blood samples were allowed to clot for 4 h at room temperature and centrifuged at 1000 × g for 15 min at 4°C to obtain serum for analyses of total protein and immunoglobulins IgG, IgA and IgM. Plasma and serum samples were stored at −20°C until analysis.

Litter size, piglets born alive and dead, individual birth weights and sex of piglets were recorded at birth. Runt piglets weighing less than 800 g were excluded from further experiments. From each experimental litter, usually three to four piglets (the lightest one, 1–2 of medium weight, the heaviest one) were sampled between 24 to 36 h after birth (day 1 (D1): AP, n = 43; LP, n = 51; HP, n = 48). Three to four other piglets, in single cases only two piglets of each litter, were randomly assigned for sampling on D27 (AP, n = 42; LP, n = 46; HP, n = 41), and the remaining piglets were sampled on D80 (AP, n = 24; LP, n = 25; HP, n = 20) or D180 (AP, n = 23; LP, n = 24; HP, n = 21). Sex was almost equally distributed within diets. The timeline for blood sampling and distribution of animals in the different diet groups are summarised in Figure 1.

Piglets were cross-fostered within 48 h after birth to multiparous sows fed a standard diet during pregnancy (Provital RF R.324.0; Trede & Pein, Dammfleth, Germany). The litters were standardised to 11 piglets with piglets from foster sows. After birth, experimental sows and foster sows were fed a single standard lactation diet (Provital LAC R.325.0; Trede & Pein, Dammfleth, Germany). Male piglets were castrated at four days of age. The piglets were weaned at D28 and housed in groups of four littermates per pen (2.5 m x 1.8 m) in a post-weaning room until D32. Thereafter, they were transferred to single-housing cages for the rest of the experimental period. Piglets were offered a commercial pellet diet from an automatic feeder. Food and water were provided ad libitum.

Blood samples from the offspring were taken while animals were in a supine position by anterior vena cava puncture on D1, D27, D29 and D47 as well as by jugular vein puncture on D80 and D180 (the whole procedure lasted approximately 1 min). Blood samples without anti-coagulant were treated as described above to obtain serum for analyses of total protein and immunoglobulins IgG, IgM and IgA (D1, D27, D80 and D180). Additionally, heparinised blood samples were stored on ice until processing for the proliferation assay (D1 and D27).

One day before (D27) and one day after weaning (D29), EDTA-plasma samples were used for cortisol analyses, and heparinised blood samples were collected and stored on ice until processing for the proliferation assay and flow cytometric analysis (AP, n = 20; LP, n = 17; HP, n = 20). On D47, a subset of piglets (AP, n = 15; LP, n = 18; HP, n = 9) received an intraperitoneal (i.p.) injection of LPS from Escherichia coli (serotype O111:B4, Sigma Chemical Company, St. Louis, MO, USA) at a dose of 100 μg/kg of body weight dissolved in 2 mL sterile saline. Blood samples were taken immediately before LPS injection (0h, baseline for all measures), 1 h later for analysis of TNF-α, and 3 and 6 h later to measure cortisol and cytokines (TNF-α, IL-6 and IL-10). Blood samples were treated as described previously.

Plasma cortisol concentrations were analysed in duplicates using a commercially available ¹²⁵I-RIA kit (DSL, Inc., Sinsheim, Germany) according to the manufacturer’s instructions. The test sensitivity was 8.1 nmol/L, and intra- and inter-assay coefficients of variation were 8.2% and 9.8%, respectively. Total protein content in serum was determined by the biuret-method (Bioquant® Protein 110307; Merck, Darmstadt, Germany). Concentrations of immunoglobulins IgG, IgA and IgM were determined in duplicate with porcine specific enzyme-linked immunosorbent assays (ELISA) according to the manufacturer’s instructions (Bethyl, Laboratories Inc., Montgomery, TX USA). The intra-assay and inter-assay coefficients of variation for the ELISAs were <5% and <10%, respectively. TNF-α and IL-10 concentrations were analysed in plasma samples using commercially available pig ELISA kits (Biosource Invitrogen, Carlsbad, California, USA) according to the manufacturer’s instructions. The sensitivity of the TNF-α assay was 3 pg/mL, and intra- and inter-assay coefficients of variation were 6.2% and 8.2%, respectively. The detection limit of the IL-10 assay was 3 pg/mL. Intra- and inter-assay coefficients of variation were 6.3% and 9.4%, respectively. Plasma concentrations of IL-6 were determined with pig IL-6 ELISA kits (R&D Systems Inc., Minneapolis, MN, USA) according to the manufacturer’s instructions. The sensitivity of this assay was 10 pg/mL, and the intra- and inter-assay coefficients of variation were 3.5% and 8.1%, respectively.

The mitogens Concanavalin A (5 μg/mL; ConA) and lipopolysaccharide (10 μg/mL; LPS) were used in lymphocyte proliferation/viability assays as previously described 28 . Briefly, peripheral blood mononuclear cells (PBMC) were isolated from heparinised blood by density gradient centrifugation, and the cell concentration was adjusted to 5 × 10⁶ cells/mL in complete RPMI-1640 medium (Sigma-Aldrich, Deisenhofen, Germany). PBMCs were incubated in 96-well microplates for 72 h in a 5% CO₂ humidified incubator at 37°C. Cell proliferation/viability was evaluated using the 3-[4,5-dimethyldiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) assay (Roche Diagnostics, Mannheim, Germany). The optical density (O.D.) was measured by a microplate reader (Dynatech, Denkendorf, Germany) using a test wavelength of 550 nm and a reference wavelength of 690 nm. The results were expressed as a proliferation index (PI), which was calculated as follows: PI = O.D. in the presence of mitogen/O.D. in the absence of mitogen.

Immunofluorescence double staining for flow cytometry was performed by incubating 200 μL of the PBMC suspension (1 × 10⁶ cells) with 10 μL of fluorescein isothiocyanate (FITC)-conjugated mouse-anti-porcine CD4α (IgG2b isotype) and 10 μl R-phycoerythrin (RPE)-conjugated mouse-anti-porcine CD8α (IgG2a isotype) monoclonal antibodies (AbD Serotec, Oxford, UK) in the dark at room temperature for 30 min. Excess antibodies were removed by washing plates twice with PBS/1% fetal bovine serum (FBS) and centrifuging at 180 × g for 5 min at 20°C. The cell pellet was resuspended in 1 mL of PBS/1% FBS for immediate analysis using an EPICS XL flow cytometer (Beckman Coulter, Krefeld Germany). The cells were gated on previously established bitmaps based on the forward and side scatter characteristics of each population. The proportion of cells was calculated after acquisition of at least 6000 events as described by Löhrke et al. 29 . To estimate the specific binding of mouse monoclonal antibodies to cell surface of lymphocytes mouse-IgG2b negative control-FITC and mouse-IgG2a negative control-RPE (Serotec, Oxford, UK) were used.

Statistical analyses were performed using the SAS System for Windows, release 9.2 30 . Data were evaluated by analysis of variance (ANOVA) using the Mixed procedure. The model for blood parameters of sows comprised the fixed effects diet (AP, LP, HP), replicate (1 to 6), repeated factor time (GD − 5, GD24, GD66, GD108) and all two-way interactions. Gestation length and litter data were analysed with the model comprising the fixed effects diet, replicate and the interaction diet × replicate. The mixed model for birth weight and early weight gain of piglets included the fixed effects diet, sex (male, female) and replicate, the corresponding 2-way interactions and the random effect of sow. Frequencies of stillborn and dead piglets per litter were analysed by fitting a logistic model using the GLIMMIX procedure. The model included the fixed effects diet, replicate and the interaction diet × replicate. The model for basal blood parameters in offspring comprised the fixed effects diet, age (D1, D27, D80, D180), sex, replicate, all two-way interactions between the fixed effects and the random sow effect. The model for weaning data and for measurement with LPS challenge in piglets included the fixed effects diet, sex, replicate, sampling time as repeated factor (for weaning: D27, D29 and for LPS challenge: 0 h, 1 h, 3 h, 6 h), all two-way interactions between these fixed effects and the random sow effect. In addition, least-squares means (LS-means) and their standard errors (SE) were computed for each fixed effect in the models, and all pairwise differences of LS-means were tested by the Tukey-Kramer procedure. Correlations between plasma cortisol levels in sows on GD108 and immune parameters in their neonatal offspring were estimated by Spearman’s rank correlation coefficients within each diet group using the Corr procedure. Effects and differences were considered significant if P < 0.05.

No effects of diet were found on the gestation length, the total number born or the frequency of stillborn piglets (P > 0.3; data not shown). Further, the litter size was not influenced by diet (AP: 11.5 ± 0.7; LP: 12.7 ± 0.7; HP: 11.5 ± 0.8; P = 0.42). However, the maternal diet significantly affected the weight of piglets at birth. Both the LP and the HP diet caused a reduction in offspring birth weight compared to control AP piglets (AP: 1.35 ± 0.04 kg; LP: 1.20 ± 0.03 kg; HP: 1.23 ± 0.03 kg; P < 0.05), which is consistent with the results obtained for all piglets born⁽¹⁵⁾. LP piglets gained less weight compared with AP and HP piglets during the first day of life (AP: 78 ± 15 g; LP: 29 ± 14 g; HP: 85 ± 15 g; P = 0.05). Piglet mortality during the suckling period was higher in the LP than in the AP and HP groups (AP: 3.0 ± 1.7%; LP: 11.2 ± 1.7%; HP: 4.0 ± 1.8%; P < 0.01).

Plasma cortisol concentrations of primiparous sows were affected by diet and time (Table 1). HP sows had lower cortisol levels than AP (P < 0.01) and LP sows (P < 0.05). The Tukey-Kramer means separation procedure indicated that the cortisol concentration in the plasma from LP sows was significantly higher compared with HP sows on GD108 (P < 0.05). In AP and LP sows, cortisol levels increased from early to late pregnancy (P < 0.05). In this subset of sows, serum protein concentrations tended to be affected by diet, but were significantly influenced by time and diet × time interaction (Table 1). On GD108, protein concentrations were significantly lower in LP than in AP (P < 0.01) and HP sows (P < 0.001). In addition, the protein levels decreased within the LP group at the end of pregnancy (P < 0.01), whereas the protein levels increased within the HP group from GD24 to GD66 (P < 0.05). Serum immunoglobulin levels were affected by time during pregnancy (P < 0.001), but not by pregnancy diet (Table 1). Independent of the diet, IgG and IgM levels increased from GD24 to GD66 (P < 0.01). In AP and LP sows, serum IgA decreased from GD − 5 to GD24 (P < 0.01). Furthermore, ANOVA revealed a significant diet × time interaction for IgA levels.

GD, gestation day.

Sows were fed diets with adequate (12.1%, n = 13), low (6.5%, n = 15) and high (30%, n = 14) protein levels throughout gestation.

Data are expressed as least-squares means (LS-means) ± standard errors (SE).

P values for effects of diet, time and diet × time interaction.

^A,BWithin parameter and time, LS-means with unlike capital superscript letters were significantly different between diets (P < 0.05; Tukey-Kramer test).

^a,bWithin a row, LS-means with unlike lower case superscript letters were significantly different between times (P < 0.05; Tukey-Kramer test).

Maternal diet did not affect the concentrations of serum total protein (P = 0.75) or immunoglobulins in the offspring (P > 0.2) over all ages examined. There were lower protein levels on D1, D27 and D80 compared with D180 for all pigs (P < 0.001, Figure 2A). We also found significant differences in the concentrations of immunoglobulins between the diet groups on the first day of life. At D1, the IgG and IgM levels in HP piglets were lower than in AP piglets (P < 0.05, Figure 2B and C, respectively), and the IgA levels were lower in both LP and HP piglets compared with AP piglets (P < 0.05, Figure 2D). Concentrations of IgG, IgM and IgA were significantly affected by age (P < 0.001), with similar patterns in all dietary groups. The lowest immunoglobulin levels were measured in piglets on D27 compared with D1, D80 and D180 (P < 0.001).

Further, there was no significant effect of maternal diet on the number of PBMCs or the mitogen-induced PBMC proliferation in response to either ConA or LPS in piglets at D1 and D27 of age (P > 0.2; data not shown). The factor sex and the interaction diet × age had no significant effects on any of the traits of the offspring that were investigated (P > 0.12).

To investigate the relationships between plasma cortisol levels in sows on GD108 and immune parameters in their neonatal offspring, Spearman’s rank correlation coefficients were estimated within each diet group (AP: n = 41; LP: n = 49; HP: n = 46). In the LP and HP groups, negative correlations were found between the cortisol levels in sows at the end of pregnancy and total protein concentrations in the serum of piglets on the first day of life (rs = −0.306, P < 0.05, and rs = −0.378, P < 0.01, respectively), whereas no significant correlation was observed in the AP group (rs = −0.102, P = 0.52). There were also negative correlations between the cortisol levels of sows and IgA levels of piglets in both the LP and HP groups (rs = −0.374, P < 0.01, and rs = −0.443, P < 0.01, respectively), but not in the AP group (rs = +0.107, P = 0.52). In addition, the cortisol levels of sows and the IgM concentrations in piglets were negatively correlated in the LP group (rs = −0.481, P < 0.001), but were not in the AP and HP groups (rs = −0.201, P = 0.21, and rs = −0.198, P = 0.19, respectively). No further significant relationships were found between the cortisol levels in late-pregnant sows of the three diet groups and immune parameters in neonatal piglets (P > 0.2).

The plasma cortisol level was not affected by maternal diet and the interaction diet × sampling time (P > 0.14), but sampling time had a significant effect (P < 0.01). The Tukey-Kramer means separation test showed that only the LP piglets displayed a significantly higher cortisol level one day after weaning than before (P < 0.05, Figure 3).

The lymphocyte proliferation in response to both mitogens ConA and LPS was not influenced by maternal diet and diet × sampling time interaction (P > 0.21). There was an effect of sampling time for proliferation indices in mitogen-stimulated cultures (P < 0.01). The Tukey-Kramer procedure revealed in AP piglets that the ConA-stimulated lymphocyte proliferation was significantly higher on D29 than D27 (P < 0.001, Figure 4A).

No significant effects of maternal diet and interaction diet × sampling time were found for the percentages of CD4⁺ and CD8⁺ cells or their ratio (P > 0.12), but these blood parameters were affected by sampling time (P < 0.01). For the percentage of CD4⁺ cells and the CD4⁺/CD8⁺ ratio, the Tukey-Kramer test indicated an increase in HP piglets (P < 0.05, Figure 5A and P < 0.01, Figure 5C, respectively). None of these variables was significantly affected by sex (P > 0.28).

Intraperitoneal injection of 100 μg LPS/kg body weight induced significant increases in the plasma concentration of TNF-α 1 h later (Figure 6B) and in concentrations of cortisol (Figure 6A), IL-6 (Figure 6C) and IL-10 (Figure 6D) 3 h after LPS in piglets on day 47 (P < 0.001). However, only IL-6 concentration was influenced by both maternal diet and diet × sampling time interaction (P < 0.05). Three hours after the challenge, LP and HP piglets responded to LPS with higher IL-6 concentrations than AP piglets (P = 0.09 and P < 0.01, respectively). Furthermore, LP piglets showed higher IL-10 levels compared with AP piglets (P < 0.05). No significant effects of sex were observed (P > 0.21).