No Beneficial Effects Evident for Enterococcus faecium NCIMB 10415 in Weaned Pigs Infected with Salmonella enterica Serovar Typhimurium DT104

Abstract

Salmonella enterica serovar Typhimurium DT 104 is the major pathogen for salmonellosis outbreaks in Europe. We tested if the probiotic bacterium Enterococcus faecium NCIMB 10415 can prevent or alleviate salmonellosis. Therefore, piglets of the German Landrace breed that were treated with E. faecium (n = 16) as a feed additive and untreated controls (n = 16) were challenged with S. Typhimurium 10 days after weaning. The presence of salmonellae in feces and selected organs, as well as the immune response, were investigated. Piglets treated with E. faecium gained less weight than control piglets (P = 0.05). The feeding of E. faecium had no effect on the fecal shedding of salmonellae and resulted in a higher abundance of the pathogen in tonsils of all challenged animals. The specific (anti-Salmonella IgG) and nonspecific (haptoglobin) humoral immune responses as well as the cellular immune response (T helper cells, cytotoxic T cells, regulatory T cells, γδ T cells, and B cells) in the lymph nodes, Peyer's patches of different segments of the intestine (jejunal and ileocecal), the ileal papilla, and in the blood were affected in the course of time after infection (P < 0.05) but not by the E. faecium treatment. These results led to the conclusion that E. faecium may not have beneficial effects on the performance of weaned piglets in the case of S. Typhimurium infection. Therefore, we suggest a critical discussion and reconsideration of E. faecium NCIMB 10415 administration as a probiotic for pigs.

INTRODUCTION

Salmonella enterica is one of the most common causes of food-borne disease, possibly affecting more than 90 million people globally each year (10). Salmonella enterica serovar Typhimurium is the predominant causative agent, as it is responsible for approximately 12% of salmonellosis outbreaks in the European Union (EU) (3). The main sources of salmonellae for humans are contaminated poultry products. However, pigs are natural carriers of salmonellae, even without clinical disease. In the past, antibiotic additives to the combined feed were widely used to prevent infection outbreaks in the herds. This led to the development of bacterial resistance against many antibiotics, and finally the EU banned their use in pig feed in 2006. Since then, much interest has been raised for the potential use of alternatives such as probiotics, i.e., live microorganisms which, when administered in adequate amounts, confer a health benefit to the host (5). Probiotics are believed to provide beneficial effects to the gastrointestinal tract (GIT) of farm animals, especially the lactic acid-producing bacteria (LAB), and thus result in increased performance of piglets. Among the different modes of actions of the LAB, the modulation of the mucosal and systemic immune systems are considered to play a major role in, e.g., prophylaxis against infections (15). The gut-associated lymphoid tissue (GALT) is the largest collection of lymphoid tissues in the body and is therefore an important first barrier against intestinal pathogens, e.g., Salmonella spp. or Escherichia coli (7). It consists of organized tissues such as jejunal and ileocecal mesenteral lymph nodes, Peyer's patches, and diffusely scattered intraepithelial lymphocytes (7), which are expected to respond to pathogen infection.

LAB, such as several strains of lactobacilli and some enterococci, are the focus of research because they are traditionally used in a range of industrial food fermentations with a long history of safe use (16). Evidence was provided (14) that Enterococcus faecium NCIMB 10415 reduces the Chlamydia sp. load of healthy piglets. Furthermore, experiments in piglets fed E. faecium have shown a reduction of diarrhea (22, 24). Those data suggested a beneficial effect of using E. faecium as a probiotic to reduce infectious diseases in pigs. In contrast, it was reported that E. faecium NCIMB 10415 tended to increase the shedding of S. Typhimurium in feces and the numbers of salmonellae in internal organs after gastric administration of S. Typhimurium DT104 to weaning piglets (crossbred German Landrace and Duroc) (21).

The aim of the present study was to test the probiotic effect of E. faecium on the immune response after infection with salmonellae. To investigate whether the effects published previously (21) were incidental or possibly race specific, the present study was performed under a similar approach using the same S. Typhimurium and E. faecium strains in pure German Landrace piglets. To assess the activation of the adaptive systemic and mucosal immune systems, we determined anti-Salmonella IgG levels as well as the levels of B and T lymphocyte subpopulations in the different tissues of the GALT. Additionally, the effect of the probiotic on performance and feed intake was investigated.

MATERIALS AND METHODS

Animals.

We used German Landrace piglets (n = 32) of sows that were fed either a diet supplemented with a probiotic (P group) or a diet without supplementation (C group). The probiotic-supplemented group contained Enterococcus faecium NCIMB 10415 as a probiotic additive, and the control group was fed the same basic feed without E. faecium. Sows of P group received E. faecium for 3 weeks before parturition, and the piglets were offered the respective combined feed with or without E. faecium from the age of 12 days on. Piglets of both feeding groups (n = 16 in both feeding groups) were weaned at 28 days of age and were pair-wise allocated to pens. The leftovers were collected daily, and feed intake was recorded on a dry matter basis. Water was provided via nipple drinkers ad libitum. The pens' floors and the feeding troughs were cleaned thoroughly twice daily. After 10 days of adaptation (at the age of 38 days), all piglets were challenged with Salmonella enterica serovar Typhimurium DT104 by intragastric application as described elsewhere (21). Six and 10 piglets from each group were sacrificed at 2 and 28 days postinfection (dpi), respectively.

The experiment was approved by the local authority (Landesamt für Gesundheit und Soziales, Berlin, Germany) under the number G0348/09.

Sampling.

After infection with S. Typhimurium, the following clinical and zootechnical parameters were recorded for all pigs: (i) general condition and fecal score (from 1 to 5, where 1 is liquid and 5 is firm); (ii) rectal temperature; (iii) shedding of salmonellae in feces, determined daily for 5 dpi and then once a week; (iv) prevalence of anti-Salmonella IgG in blood samples, which were obtained before infection and then at weekly intervals; (v) daily feed intake; and (vi) weekly body weight for the calculation of the feed conversion ratio (kg feed/kg weight gain).

During necropsy, liver; spleen; palatine tonsils; segments of midjejunum, ileum, cecum and ascendent colon (third centripetal gyrus); jejunal and colonic mesenterial lymph nodes; and muscle samples from the fore (musculus triceps brachii) and hind limbs (musculus gluteus superficialis) were collected for the cultivation of S. Typhimurium.

For the characterization of the immune cells, mesenteral jejunal and ileocecal lymph nodes, Peyer's patches, and ileal papilla were collected from six animals per group at 2 and 28 dpi.

Bacterial strains.

E. faecium strain NCIMB 10415 is a commercial probiotic feed additive (Cylactin LBC ME10; DSM Nutritional Products Ltd., Switzerland). It was provided in a microencapsulated form and mixed into the diets of sows and suckling and weaned piglets at concentrations of 1 × 10⁹ to 5 × 10⁹ CFU/kg feed.

For infection with S. Typhimurium, strain DT104 was chosen because it was obtained from a swine with sepsis (21) and was characterized by resistance against multiple antibiotics (chloramphenicol, ciprofloxacin, tetracycline, florfenicol, ampicillin, sulfamethoxazole, colistin, streptomycin, and nalidixic acid [NAL]). The resistance against nalidixic acid was used for later selective cultivation of the strain from the samples. S. Typhimurium was cultured in Luria-Bertani broth at 37°C for 20 to 22 h to reach an optical density of 0.69 to 0.72 (corresponding to 2 × 10⁹ to 3 × 10⁹ CFU/ml), which was subsequently confirmed by plating. Each piglet was infected with 7 ml of such culture broth (1.4 × 10¹⁰ to 2.1 × 10¹⁰ CFU in total).

Quantitation of salmonellae.

The quantitative detection of salmonellae in feces was performed using a spiral plater (Whitley, Meintrup DWS, Germany) with a detection limit of 4 × 10² CFU/g. For each sample, 50 μl from 5 dilutions was streaked onto 3 xylose-lysine-deoxycholate agar plates supplemented with 50 μg/ml of NAL (XLD-NAL) and incubated at 37°C for 20 to 24 h. For the quantitation of salmonellae in internal organs, tissue samples were immersed in 95% ethanol, flamed, minced aseptically, and homogenized with buffered peptone water (BPW; 1:10) in filter bags using a stomacher for 2 min at high speed. All samples were quantified for S. Typhimurium by the plating of 100 μl of filtrates using a spiral plater. The detection limit was 2 × 10² CFU/g. In addition, 1 ml of the 10-fold dilution was streaked directly onto 3 XLD-NAL agar plates to increase the detection limit to 10 CFU/g. Five hundred μl of blood collected in tubes containing K-EDTA was streaked onto 3 XLD-NAL agar plates to test if there was bacteremia after the challenge.

The obtained colony numbers were transformed into log CFU/g to obtain nearly normal distributions of trait values. Mean values and standard deviations are provided in the text.

Isolation of immune cells.

Immune cells were characterized for six animals in the E. faecium-supplemented group and the control group at 2 and 28 dpi (Σ n [total number of samples] = 24). The dissected lymph nodes were transferred into 15-ml tubes filled with 5 ml phosphate-buffered saline containing 0.2% bovine serum albumin (PBS–0.2% BSA). The Peyer's patches and the ileal papilla were put into Hank's buffered salt solution (HBSS). The individual tissue samples were pressed through a 70-μm cell strainer (Becton, Dickinson GmbH, Heidelberg, Germany). The remaining cell suspensions containing lymphocytes and peripheral blood mononuclear cells (PBMC) were further purified by centrifugation for 30 min at 700 × g at 20°C in a Ficoll gradient (Histopaque 1077; Sigma-Aldrich Corporation). After final lysis of erythrocytes in erythrolysis buffer, pH 7.2 to 7.4 (Morphisto GmbH), for 5 min on ice, the immune cells were washed with 10 ml of PBS–0.2% BSA and centrifuged for 15 min at 390 × g at 4°C.

Flow cytometry.

Flow cytometry was performed with purified immune cells. The cells were stained with primary anti-CD4 conjugates in a one-step incubation. For each reaction, 1 × 10⁶ cells were exposed to saturating concentrations of the antibodies in a volume of 30 μl of PBS for 30 min on ice in the dark. Cells were washed with 1 ml of PBS, spun at 2,500 × g for 5 min, and resuspended in 300 μl PBS. Immune cell antigens CD2, CD25, CD8β, TcR1, and IgM (Table 1) were detected using unlabeled primary antibodies followed by washing and incubation with a fluorescence-labeled secondary antibody (Table 1). Table 1 displays the source of each antibody. Different lymphocyte subpopulations were designated by their phenotypes according to references 18 (for γδ T cells and B cells) and 8 (for αβ T cells). Per sample, 40,000 lymphocytes were assayed by flow cytometry using an BD FACSCalibur flow cytometer within the lymphocyte gate corresponding to their forward (FSC) and sideward (SSC) light scatter signals (Fig. 1A). From these cells, only living ones, which were negative for the propidium iodide (PI) staining (0.5 μg/ml), were further analyzed (Fig. 1B). T helper cells (Th), cytotoxic T cells (Tc), and regulatory T cells (Treg) were determined by the detection of the surface markers CD4, CD8β (Fig. 1C), and CD25 (Fig. 1D), respectively. For phenotypic analysis of γδ T cell receptor (TCR) cells, a marker for the δ chain and the surface marker CD2 were chosen in the same scheme as that shown for CD4 and CD8. Immature and mature naive B cells were identified by staining with the surface marker IgM.

Table 1.

Primary and secondary antibodies used for flow cytometry staining^a

Antibody	Clone	Isotype	Labeling	Company
Primary
CD2	MSA-4	IgG2a	None	VMRD
CD4	74-12-4	IgG2b	FITC	Southern Biotech
CD8β	PG164A	IgG2a	None	VMRD
TcR1-N4 (δ)	PGBL22A	IgG1	None	VMRD
CD25	K231.3B2	IgG1	None	Biozol
IgM	PIG45A	IgG2b	None	VMRD
Secondary
Goat anti-mouse IgG1	Pooled	IgG1	APC	Southern Biotech
Goat anti-mouse IgG2b	Pooled	IgG2b	FITC	Southern Biotech
Goat anti-mouse IgG2a	Pooled	IgG2a	PE	Southern Biotech

^aThe references for each antibody are provided by the manufacturers in the datasheets.

Fig 1.

Gating and selection strategies of the flow cytometry data for the analyses of lymphocyte subpopulations. The x and y axes show the intensity of the light scatter or fluorescent (FL) signal. Framed cells populations in panel A were further analyzed in panel B, those in panel B were further analyzed in panel C, and those in panel C were further analyzed in panel D. (A) Lymphocytes gated according to their forward and sideward light scatter signals. Circulating lymphocytes represent 40,000 cells. (B) Only living PI-negative cells were taken for analyses. (C) Relative cell counts of living lymphocyte subpopulations (CD8pos, cytotoxic T cells; CD4pos, CD8pos, pig-specific double-positive cells; CD4pos, T helper cells) detected with antibodies CD8β and CD4 were calculated. (D) The relative number of CD25^high cells among the CD4⁺ cells was calculated.

Serology.

Blood samples of all animals were collected from the cranial vena cava at 2 and 28 dpi. Blood was incubated at 37°C for 2 h and centrifuged for 15 min at 1,400 × g, and serum was collected and stored at −20°C until analysis.

The presence of anti-Salmonella antibodies was tested with the Salmotype Pigscreen enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions (Labordiagnostik Leipzig, Leipzig, Germany). The IgG levels were calculated using a reference standard method and are presented as percentages of the optical density (OD%).

Haptoglobin concentrations in the sera were analyzed by use of the a pig haptoglobin ELISA test kit (Life Diagnostics, West Chester, PA) by following the manufacturer's instructions. The haptoglobin concentrations in sera were determined by applying standard curves and calculated in mg/ml.

Statistical analysis.

Statistical analysis was performed using R 2.11.1 and SPSS 12.0.2. (SPSS, Inc., Chicago, IL). For the phenotypes of immune cells, we analyzed 6 samples per group (combined effect of time and probiotic [dpi × diet]). The experiment had a statistical power level of 0.8 for an anticipated correlation (r²) between diet and phenotype of ≥0.35 and a P value of 0.1. Alternatively, it had a power of 0.7 to detect effects with a P value of 0.05. All data values higher than twice the interquartile range (IQR), below the first quartile, and above the third quartile were identified as outliers. Outliers had no significant effects on test statistics due to having less than 1% outliers in our data. The Shapiro-Wilk test was used to check raw data for normal distribution. Since raw data were not normally distributed, numbers of salmonellae were log transformed.

To test the effects of supplementation with E. faecium on the phenotypes, we applied an analysis of variance (ANOVA) with E. faecium supplementation in the diet, days postinfection, and sex as fixed effects. The following linear model was used to assess the variation (V) for every trait for each tissue: V_trait = V_{diet group} + V_{days postinfection} + V_sex + residual error.

The amounts of S. Typhimurium in feces and organs were compared using the nonparameteric Kruskal-Wallis test, because salmonellae could not be continuously detected in all samples. Differences were considered significant at P < 0.05. Pearson's correlation coefficients were calculated between IgG, haptoglobin, and log counts of S. Typhimurium in feces. Box-whisker plots were chosen for the graphical presentation of the results. The boxes indicate the medians (horizontal lines) and the lower and upper quartiles (lower and upper sides of the boxes). Outliers are indicated as circles and asterisks in the plots.

RESULTS

Zootechnical parameters.

The mean body weight was equal in both groups at weaning (10.4 kg in C and 10.0 kg in P). During the study, the piglets from C group gained more body weight than those from P group (31.0 and 28.9 kg, respectively; P = 0.05). The feed conversion ratio was not affected by the probiotic. Based on feed intake, daily E. faecium intake was 3.7 × 10⁸ CFU in the week before infection with S. Typhimurium and 4.7 × 10⁸, 8.0 × 10⁸, 1 × 10⁹, and 1.3 × 10⁹ CFU during the first, second, third, and fourth week postinfection, respectively. Two animals died during the study, one in each group, due to reasons unrelated to the salmonella infection.

Clinical parameters.

The infection with S. Typhimurium caused mild clinical signs. One piglet of the E. faecium-treated group developed systemic disease immediately after infection. The serum concentration of haptoglobin of this piglet increased from 1.79 mg/ml at the time of infection to 3.90 mg/ml at 7 dpi and remained elevated (2.82 mg/ml) until 21 dpi. In this case, constipation occurred and the fecal score was 5 (hard firm stool) until 4 dpi, indicating high water demand.

Increased rectal body temperatures (≥40°C) were measured in three and two piglets at 1 dpi and in seven and eight piglets at 2 dpi in C and P groups, respectively. On average, piglets in C and P groups had elevated rectal temperatures for 1.8 ± 1.0 and 0.9 ± 1.4 days within the first 7 dpi (P = 0.13), and thereafter all piglets had a physiological temperature.

Diarrhea was recorded at score 1 (watery) and 2 (mushy). Before infection, diarrhea (score 1) was observed in one animal from P group and in two control animals (score 2). On the day of infection, no diarrhea was noticed. At 1 dpi, one control piglet had feces of score 2. At 2 dpi, watery or mushy diarrhea was observed in three and one C piglet and in two and three piglets from P group, respectively. At 3 dpi, an additional two control and one P piglet developed diarrhea. The diarrhea lasted for 1 to 2 days without statistical differences between the treatment groups (P = 0.442).

Detection of Salmonella Typhimurium in feces and organs.

None of the animals shed salmonellae before challenge. At 1 dpi, all animals shed S. Typhimurium, except one and two piglets from the C and P groups, respectively, which started to shed salmonellae at 2 dpi. The shedding of salmonellae lasted from 1 to 28 days, with levels of 4 to 8 log CFU per g feces. At the end of the trial (at 28 dpi), S. Typhimurium was detected in feces from 1 pig from group C (2.8 log CFU/g) and 3 pigs from group P (0.9 to 3.0 log CFU/g) without significant differences between groups (Fig. 2A).

Fig 2.

Colonization of piglets with salmonellae and the reaction of their immune systems to infection. Shedding of Salmonella Typhimurium in feces (A) and in organs after either 2 (B) or 28 days (C), as well as serum anti-salmonella IgG titers (D) and haptoglobin concentrations (E) after intragastric S. Typhimurium infection in groups of piglets that were either untreated controls (C) or were treated with E. faecium (P). CFU/g feces were log transformed to obtain normal distributions. Box-whisker plots indicate the medians (horizontal lines) and the lower and upper quartiles (lower and upper borders of the boxes); minimum and maximum values, as well as the outliers (mild, open circles; extreme, asterisks), are presented.

S. Typhimurium was detected in the wall of GIT (2.4 to 6.4 log CFU/g) and in mesenteral lymph nodes (2.5 to 3.1 log CFU/g) of all pigs at 2 dpi. S. Typhimurium was also present at 2 dpi in tonsils from 1 pig from C (2.8 log CFU/g) and 2 pigs from P (2.1 to 3.7 log CFU/g) and at 28 dpi in 1 pig from C (5.7 log CFU/g) and 4 pigs from P (1.2 to 5.9 log CFU/g). S. Typhimurium could be also detected sporadically in liver and spleen at 2 dpi. At 28 dpi, S. Typhimurium was detected in the P group in the wall from ileum of 2 piglets and in the wall from jejunum of 5 piglets. At this time, 3 piglets from C and 3 piglets from P had S. Typhimurium in the wall of cecum, with 2 piglets from C and 1 piglet from P had it in colon. No effect of the probiotic on the CFU numbers could be observed (Fig. 2B and C). Furthermore, no salmonellae could be isolated from the blood, showing that no bacteremia had occurred.

Humoral immunity and anti-Salmonella IgG.

Before infection, for all samples measured, the anti-Salmonella IgG titers were below the 20% cutoff value (Fig. 2D). Although low throughout, the levels of specific antibodies were significantly higher in E. faecium-treated piglets than in untreated piglets (P < 0.01). Specific IgG titers continuously increased in both groups starting from 7 until 28 dpi, when the animals were euthanized. Possibly because of the short duration of the study, no plateau was reached at 28 dpi, and titers in the two groups did not differ at this time point (P > 0.5). However, the high standard deviations of the mean IgG titers in both groups implicate high individual variability in the immunological response to the S. Typhimurium challenge which did not correlate with the numbers of salmonellae shed in feces.

Acute-phase protein haptoglobin as part of innate immunity.

At weaning, haptoglobin concentrations in serum were 2.18 ± 1.1 and 1.98 ± 0.44 mg/ml in C and P groups, respectively. They remained at the elevated level in C and increased in P (P = 0.048) 1 week later (Fig. 2E). After the adaptation period, at the day of challenge the haptoglobin levels decreased (P < 0.001) in both groups without significant differences (P = 0.867). After infection with S. Typhimurium, the haptoglobin levels increased in the pigs from both groups (P < 0.01); however, their levels at 2 dpi did not differ between groups (P = 0.264). In pigs from C group, the haptoglobin level decreased at 7 dpi to the level measured at day 0 and decreased further until 14 dpi (P = 0.006). In pigs from P group, the decrease of serum haptoglobin was slow and reached a significantly lower level only on 28 dpi (P < 0.001). During the whole period of 28 days after infection, a significant effect of the probiotic on the haptoglobin level was observed (P = 0.014), and the haptoglobin concentrations were negatively correlated with the OD% values obtained for anti-Salmonella IgG (r² = −0.325; P < 0.01).

Phenotyping of immune cell populations.

No significant differences could be observed for CD8⁺ cytotoxic T cells (Fig. 3A) and for the CD4⁺ T helper cells (Fig. 3D) between the groups. There was a tendency toward higher frequencies of CD8⁺ and CD4⁺ cells in the lymph nodes of the ileum and jejunum (data not shown), Peyer's patches, and papilla and a tendency toward lower levels of circulating CD8⁺ cells in PBMC in C group at 2 dpi. The pig-specific CD4⁺ CD8⁺ double-positive cell population (2) was also detectable (Fig. 3B) but without significant differences between groups. The CD2⁺ δ TCR⁺ T cells were on a comparable level among the PBMCs in the examined tissues between control and E. faecium-treated piglets. The only remarkable difference (P = 0.014) found was an increase of the CD2⁻ δ TCR⁺ T cells in the PBMCs of blood from 2 to 28 dpi (Fig. 4A). CD2⁻ δ TCR⁺ T cells were detectable in all examined tissues, but they were present predominantly in blood at a level of up to 24% of the whole lymphocyte population (Fig. 4B). However, there were no differences between the control and the E. faecium treatment groups.

Fig 3.

Cell counts of αβ T cells relative to the living lymphocyte population. (A) CD8β pos (cytotoxic T cells), (B) CD8β and CD4 pos, and (D) CD4 pos (T helper cells) in peripheral blood mononuclear cells (PBMCs), ileal lymph nodes (IL LN), ileal Peyer's patches (IL PP), and papilla (PAP IL) of the E. faecium-supplemented group (P) and the non-E. faecium-supplemented control group (C). (C) Fluorescence-activated cell sorting (FACS) scatter plots of one female piglet derived from P group at 2 dpi as an example for the appearance of T helper cells, cytotoxic T cells, and CD4⁺ CD8⁺ T cells in the PBMCs, IL LN, IL PP, and PAP IL. All piglets were infected with S. Typhimurium.

Fig 4.

Cell counts of γδ T cells relative to the living lymphocyte population. (A) CD2- and γδ TCR-positive cells. (B) γδ TCR-positive (CD2-negative) cells in the PBMCs, IL LN, IL PP, and PAP IL of piglets from the E. faecium-supplemented group (P) and the control group (C) 2 and 28 days after challenge with S. Typhimurium.

The level of monomeric cell surface-bound IgM, correlating with the presence of B cells, was numerically higher in the P group than in the C group in the lymph node of the ileum and jejunum lymph nodes at 2 dpi (Fig. 5A and B) (P < 0.05), but they were at the same levels in both groups at 28 dpi. The PBMC levels of B cells were higher in the P group than in the C group at 2 dpi, and this effect was even stronger at 28 dpi (Fig. 5C). In the GALT, the B cell population of continuous Peyer's patches and the papilla decreased significantly at 28 dpi compared to that at 2 dpi (P < 0.01), without significant differences between the groups (Fig. 5D).

Fig 5.

Cell counts of B cells relative to the living lymphocyte population. (A) IgM-positive cells (B cells) of one piglet representative for the E. faecium-supplemented group (P) at 2 dpi in the ileal lymph node (IL LN) and one piglet representative of the control (C) at 2 dpi in IL LN. (B to D) IgM-positive cells (B cells) of 12 piglets from P group and of 12 piglets from C group at 2 (n = 6) and 28 dpi (n = 6) in mesenteric lymph nodes (LN) of jejunum (Je LN) and ileum (IL LN) (B), in PBMCs (C), and in the ileal Peyer's patch (IL PP) and papilla (PAP IL) (D). All piglets were infected with S. Typhimurium.

In an ANOVA covering the whole test period, there were no significant effects in the tested immune cell populations between the probiotic-treated group and the control group, and there were no effects between sexes.

DISCUSSION

E. faecium NCIMB 10415 has been widely approved as a probiotic feed supplement for use in animals in the European Union. To obtain approval for commercial use, a feed additive must possess at least one of the characteristics stated in Article 5 (sentence 2) of Regulation 1831/2003 (4). When applying for approval of a bacterial strain as a probiotic feed additive, it is sufficient to show that a given strain will “favorably affect animal production, performance or welfare, particularly by affecting the gastrointestinal flora or digestibility of feeding stuffs” (4). Studies to test the beneficial effect of E. faecium mostly focused on zootechnical parameters, such as feed intake, daily weight gain, and feed conversion ratio under commercial farm conditions, without focusing on the targeted infection (11, 24).

In the present study, we examined the effect of E. faecium administration with respect to the response to S. Typhimurium infection in weaned piglets of the German Landrace breed. In our bacterial challenge trial, E. faecium-fed piglets gained even less weight than the untreated control animals. This shows that there are no positive effects of this probiotic on the pig's bodily development in the face of an infection with S. Typhimurium. Similarly, the same probiotic (Cylactin) reduced daily weight gain in uninfected piglets fed the probiotic in the same way without further effects on the feed conversion ratio (12, 22). In contrast, an improvement of daily weight gain (17 g/day) of unweaned piglets kept under farm conditions and fed the same probiotic strain suspended in liquid or gel form also was reported (24). Therefore, it remains unclear if different forms of application (liquid, microencapsulated, once or for several days, etc.) or other factors, such as farm conditions, genetic background of the animals, or gestation number, could influence the effects of E. faecium on body weight gain.

We investigated the effect of E. faecium as a probiotic on the shedding of pathogenic bacteria in a model challenge experiment with Salmonella enterica serovar Typhimurium DT104 infection. The virulence of the S. Typhimurium strain used for infection could be confirmed in one infected piglet. All other piglets developed mild clinical signs only. Nevertheless, the colonization of the gastrointestinal tract with S. Typhimurium occurred in all pigs, since all of them shed salmonellae for at least 2 days after infection. It is well known that in the course of pathogenesis after salmonellae infection, the wall of the gastrointestinal tract is actively invaded by the pathogen by different mechanisms (9) before it multiplies and spreads to other organs through macrophages, thereby reaching other lymphatic organs, such as the mesenteral lymph nodes of the small and large intestine, the spleen, and tonsils (6). In the present study, we could not observe any protective effects of E. faecium on clinical symptoms of salmonellosis or on S. Typhimurium shedding or distribution into internal organs. This is in line with the results obtained from a similar experimental setting (21). In contrast, Casey et al. (1) reported a reduced number of shed S. Typhimurium without effects on the duration of the shedding in Large White × Landrace weaned piglets fed a cocktail of five different probiotic strains of LAB. It is intriguing that reports on the effects of probiotics on pathogen infection often provide controversial results. We suspect that the reasons are manifold. The choice of pathogen, the method of infection (oral once or for several days, intragastric, so-called troyan model, etc.), the infection pressure, and other controlled or uncontrollable experimental conditions, as well as genetic predisposition, could play a role in yielding variable effects of probiotics. We emphasize for the present study that the pens and feeding troughs were cleaned thoroughly twice daily, and that this rigorous hygienic regimen was obviously sufficient to reduce the number of pathogens in the environment and therefore to eliminate or at least reduce the source of reinfection. This could have an impact on the lack of differences between the two experimental groups. On the other hand, the control pigs coped very well and recovered from the infection without any additional help, which underlines the importance of proper management.

The present study confirms that feeding E. faecium to piglets resulted in higher numbers of S. Typhimurium in tonsils, as reported previously (21). The greater abundance of salmonellae in the tonsils lasting for at least 28 days after infection could become a source for reinfection of the pigs and thus could be responsible for subclinical and chronic infections of a pig herd. Furthermore, infected tonsils also could be a source of contamination of meat at slaughter. From this perspective, our data provide evidence against the use of E. faecium NCIMB 10415 as a probiotic feed supplement in pigs to prevent the infection of meat and thereby humans via the slaughter house.

Probiotics are considered to modulate the host's immune system, resulting in an improvement of protection against infections (20). In a previous study (21), higher levels of serum IgM and IgA against S. Typhimurium could be observed in the group fed E. faecium without any effects on serum anti-S. Typhimurium IgG. In the present study, E. faecium had no increasing effect on serum IgG after infection. However, we found that the increase of monomeric cell surface-bound IgM was more prominent in the E. faecium-treated group (P < 0.05). This indicates a higher infection level followed by a swift recognition of the agent by the immune system of the piglets fed the probiotic, but with a less effective clearing of salmonellae.

The innate immune response, as measured by the serum haptoglobin concentrations, supports this hypothesis. The initial serum haptoglobin levels, measured in blood collected immediately after transport of the pigs to the facility, were elevated compared to published reference values (13), likely as a result of the effect of stress at weaning. The haptoglobin concentrations decreased during the time of adaptation to the new conditions. An increase of serum haptoglobin caused by the salmonella challenge was observed in both groups independently of E. faecium treatment. The maximal concentrations were of the same magnitude as those in viral infection studies of pigs (17). The level remained elevated for a longer time in probiotic-fed pigs than in controls, indicating either longer-lasting synthesis and release of the acute-phase protein or its slower degradation when E. faecium was in the diet. Thus, these data suggest that the nonspecific reaction of the immune system was activated for a longer time under the probiotic treatment. As we failed to find differences in the relative cell count of B cells with cell surface-bound monomeric IgM between the E. faecium-supplemented and the control groups in GALT, we hypothesize that there was no stimulation of the immune response against S. Typhimurium by E. faecium. This is in agreement with previous results obtained from Landrace × Duroc piglets (21). The same strain of E. faecium that we fed to sows and piglets could reduce the rate of naturally occurring infection with Chlamydia sp. in pigs (14). Previous studies examined the effects of E. faecium on infection with salmonellae or chlamydiae (14, 21), and those authors observed a reduction of the cytotoxic CD8⁺ subpopulation of intraepithelial and circulating lymphocytes and therefore discussed, as a possible mechanism, that E. faecium activates the production and secretion of IgM and IgA, thus the reduction of CD8⁺ lymphocytes could be a result of increased binding to the intestinal pathogen bound by the antibodies. In our study, the frequency of cytotoxic CD8⁺ T cells in tissues responsible for the removal of cells infected by S. Typhimurium was lower in E. faecium-treated pigs than in control animals. This is in line with the finding that the frequency of salmonellae was higher in tissues of the probiotic feed group. Otherwise, the proportion of circulating effector CD8⁺ cytotoxic T lymphocytes was enhanced in the peripheral blood mononuclear cells in the E. faecium-treated group 2 dpi. Therefore, one can assume an alternative mechanism triggered by E. faecium. E. faecium may inhibit the efficient homing of the effector cytotoxic T cells. This would explain the lower clearance of salmonellae in the E. faecium-treated group. This finding goes with the assumption that E. faecium enhances IgM and IgA production and secretion, and therefore the stimulation of more naive T cells to differentiate into circulating effector T cells. Although the assumptions of the possible mechanisms of the effects of an E. faecium supplementation on S. Typhimurium infection made here are in line with previous studies, one has to consider that we did not measure antigen-specific T or B cell responses. Therefore, the mechanism has to be elucidated in further research. However, the same pattern of low shedding of salmonellae in both groups suggests that this mechanism does not result in an improved elimination of the pathogen from the gastrointestinal tract of the piglets. According to previously published evidence (19, 23), we detected a high percentage of CD2⁻ γδ T cells in the peripheral blood, with only a few such cells accumulating in the lymphoid tissues of the intestine. CD2⁺ γδ T cells were present in all examined tissues and in blood with a frequency of 1 to 2%. Neither the CD2⁻ γδ T cells nor the CD2⁺ γδ T cells showed a difference between the E. faecium-treated and control group. Although the CD2⁺ γδ T cells, which additionally express CD8α, are believed to be cytotoxic (18) and hence are able to clear salmonella-infected cells, γδ T cells did not seem to be a target of activation for E. faecium-derived signaling in our salmonella challenge experiment. The number of B cells in ileal Peyer's patches and ileal papilla significantly (P < 0.001) decreased from 2 to 28 dpi. Since IgM is a marker for immature and mature naive B cells, it is likely that at 28 dpi the B cells are activated and proliferate into Ig-producing plasma cells. This is accordance with the observed uptake of IgG in the blood of the piglets.

In the present study, we were able to show that Enterococcus faecium NCIMB 10415 had no beneficial effect on the performance of weaned piglets after infection with Salmonella enterica serovar Typhimurium DT104, and that it fails to reduce the shedding of salmonellae and to stimulate a positive cellular immune response. Furthermore, treatment with E. faecium could result in the higher abundance of the pathogen in tonsils of infected animals. Therefore, we suggest that the use of E. faecium NCIMB 10415 as a probiotic for pigs undergoes a critical discussion and reconsideration.