Effect of heat-killed Escherichia coli, lipopolysaccharide, and muramyl dipeptide treatments on the immune response phenotype and allergy in neonatal pigs sensitized to the egg white protein ovomucoid

Abstract

Predisposition to food allergies may reflect a type 2 immune response (IR) bias in neonates due to the intrauterine environment required to maintain pregnancy. The hygiene hypothesis states that lack of early environmental stimulus leading to inappropriate development and bias in IR may also contribute. Here, the ability of heat-killed Escherichia coli, lipopolysaccharide (LPS), or muramyl dipeptide (MDP) to alter IR bias and subsequent allergic response in neonatal pigs was investigated. Three groups of three litters of pigs (12 pigs/litter) were given intramuscular injections of E. coli, LPS, MDP, or phosphate-buffered saline (PBS) (control) and subsequently sensitized to the egg white allergen ovomucoid using an established protocol. To evaluate change in IR bias, immunoglobulin isotype-associated antibody activity (AbA), concentrations of type 1 and 2 and proinflammatory cytokines released from mitogen-stimulated blood mononuclear cells, and the percentage of T-regulatory cells (T-regs) in blood were measured. Clinical signs of allergy were assessed after oral challenge with egg white. The greatest effect on IR bias was observed in MDP-treated pigs, which had a type 2-biased phenotype by isotype-specific AbA, cytokine production, and a low proportion of T-regs. LPS-treated pigs had decreased type 1- and type 2-associated AbA. E. coli-treated pigs displayed increased response to Ovm as AbA and had more balanced cytokine profiles, as well as the highest proportion of T-regs. Accordingly, pigs treated with MDP were more susceptible to allergy than PBS controls, while pigs treated with LPS were less susceptible. Treatment with E. coli did not significantly alter the frequency of clinical signs.

Fig 1

Pretreatment and sensitization protocol. Three groups of pigs, each group containing three litters of approximately 12 pigs/litter treated with intramuscular (i.m.) injections of heat-killed E. coli (10⁹ CFU/ml in 1 ml PBS; n = 11 pigs), lipopolysaccharide (LPS; 10 μg/ml in 1 ml PBS; n = 15 pigs), muramyl dipeptide (MDP; 10 μg/ml in 1 ml PBS; n = 16 pigs), or phosphate-buffered saline (PBS; n = 49 pigs, pooled across all litters) in a split-litter design. Six pigs in each litter of 12 received treatment with bacteria or bacterial components; the remaining pigs were treated with PBS as a negative control on each of the first 7 days of life. Injections were divided into two equal volumes and administered into lateral cervical areas. Pigs were then sensitized by intraperitoneal (i.p.) injection of ovomucoid (Ovm; 100 μg/ml) mixed with 10 μg/ml of cholera toxin (CT) on days 14, 21, and 35. Skin tests were conducted on days 14, 21, 35, and 45 to monitor sensitization by intradermal (i.d.) injection of Ovm (100 μg). On day 45, pigs were fasted overnight and challenged on day 46 with 100 ml of egg white and yoghurt mixed 3:2 (vol/vol), observed for at least 2 h for clinical signs of allergy, and assigned clinical scores (Table 1). Blood was taken via the retro-orbital sinus on days 14 and 45 to measure immunoglobulin isotype-associated antibody activity by enzyme-linked immunosorbent assay (ELISA) and on day 45 to isolate blood mononuclear cells (BMCs) for culture and quantification of cytokine concentrations and to measure the proportion of circulating T-regulatory cells (T-regs) by flow cytometry. Experimental days are equal to days of age.

Fig 2

Ovomucoid-specific antibody activity of pigs treated with heat-killed E. coli, LPS, MDP, or PBS (control). Serum immunoglobulin isotype-specific antibody activity (AbA) to ovomucoid (Ovm) was monitored by enzyme-linked immunosorbent assay. All pigs, regardless of treatment group, produced significantly more Ovm-specific IgG (H+L), IgE, IgG1, and IgG2 AbA at 45 days of age than at 14 (P < 0.0001). Escherichia coli-treated pigs (n = 11) produced more Ovm-specific IgG than PBS (n = 49 pigs, pooled across all litters)-treated controls and MDP (n = 16)-treated piglets. Escherichia coli-treated piglets had more IgE AbA than PBS-treated controls and LPS (n = 15)-treated pigs. A trend toward decreased activity of IgE in LPS-treated pigs compared to MDP was also noted. E. coli-treated pigs had more IgG1 activity than all other treatment groups. More Ovm-specific IgG2 activity was present in E. coli-treated pigs than in LPS- or MDP-treated pigs, while both LPS- and MDP-treated pigs had less IgG2-specific AbA than PBS controls. Individual serum means were expressed as percent of the positive-control values for each plate as follows: % of positive-control activity = (mean test serum OD/mean positive-control serum OD) × 100 (unpaired t test, Welch's correction applied when necessary, significance taken at P < 0.05, trends at P < 0.1).

Fig 3

Ratio of isotype-specific antibody activity to ovomucoid in pigs treated with heat-killed E. coli, LPS, MDP, or PBS (control). Serum isotype-specific antibody activity (AbA) to ovomucoid was measured by enzyme-linked immunosorbent assay, and the ratio of AbA was calculated for each pig. The ratio of IgG1 to IgG2 was increased in MDP (n = 16)-treated piglets compared to that in E. coli (n = 11)- and PBS (n = 49, pooled across all litters)-treated piglets. This ratio was also increased in LPS (n = 15)-treated piglets compared to in E. coli-treated piglets. The IgE/IgG2 ratio was greater in MDP-treated piglets than in piglets in all other treatment groups. There was a trend toward a greater IgE/IgG2 ratio in PBS-treated pigs than in those treated with E. coli (unpaired t test, Welch's correction applied when necessary, significance taken at P < 0.05, trends at P < 0.1).

Fig 4

Cytokine expression by PHA-P-stimulated BMCs from piglets treated with heat-killed E. coli, LPS, MDP, or PBS (control). Blood was collected from piglets at 45 days of age, postsensitization to ovomucoid. Blood mononuclear cells (BMCs) from each pig were isolated on a density gradient of Histopaque (Sigma-Aldrich) and cultured at 2.5 × 10⁶ cells/ml for 96 h, stimulated with 10 μg/ml of PHA-P (Sigma-Aldrich). Culture supernatant was collected and stored at −80°C. Cytokine concentration was measured using capture enzyme-linked immunosorbent assay (Invitrogen, IL-10 and IL-4; Kingfisher, IL-17; R&D Systems, IL-12). The concentration of IL-17 was lower in the supernatant of BMCs from E. coli-treated (n = 11) pigs than from all others. No difference in the expression of IL-4 was noted from the supernatant of cells from pigs in each treatment group. Cells from piglets treated with MDP (n = 16) and PBS (n = 49, pooled across all litters) had an increased expression of IL-10 compared to that in E. coli-treated pigs. Cells from MDP-treated pigs trended toward increased IL-10 production compared to cells from LPS-treated (n = 15) pigs. The concentration of IL-12 was lower in supernatant from cells of E. coli-treated pigs than in any other treatment group (unpaired t test, Welch's correction applied when necessary, significance taken at P < 0.05, trends at P < 0.1). Please note the difference in scale on the y axis.

Fig 5

Ratio of cytokine expression by PHA-P-stimulated BMCs from pigs treated with heat-killed E. coli, LPS, MDP, or PBS (control). Blood was collected from pigs at 45 days of age, postsensitization to ovomucoid. BMCs were isolated using a density gradient of Histopaque (Sigma-Aldrich) and cultured at 2.5 × 10⁶ cells/ml for 96 h, stimulated with 10 μg/ml of PHA-P (Sigma-Aldrich). Culture supernatant was collected and stored at −80°C. The concentrations of cytokines were measured using capture enzyme-linked immunosorbent assay (Invitrogen, IL-10; Kingfisher, IL-17; R&D Systems, IL-12). Ratios of cytokines were calculated to elucidate possible bias as a result of treatment. There was a trend toward an increased ratio of IL-17 to IL-10 in cells from LPS-treated pigs (n = 15) compared to those from pigs treated with E. coli (n = 11). The ratio of IL-17 to IL-12 was greater in supernatant from cells of PBS-treated pigs (n = 49, pooled across all litters) than in cells from LPS-treated pigs. In cells from LPS-treated pigs, the ratio of IL-10 to IL-12 was less than that in cells from all other treatment groups (unpaired t test, Welch's correction applied when necessary, significance taken at P < 0.05, trends at P < 0.1). Please note difference in scale.

Fig 6

Percentage of T-regulatory cells from blood of pigs treated with heat-killed E. coli, LPS, MDP, and PBS (control). Blood was collected from pigs at 45 days of age, postsensitization, prior to oral challenge with egg white and prepared for identification of CD25⁺ FoxP3⁺ T-regs by flow cytometry. Percentage of T-regs was measured from 10 μl of whole blood from each pig. Pigs treated with E. coli (n = 11) had higher percentages of T-regs than all others. Pigs treated with MDP (n = 16) had a lower percentage of T-regs than pigs treated with PBS) (n = 49, pooled across all litters) (unpaired t test, Welch's correction applied when necessary, significance taken at P < 0.05).