Spread of an Experimental Salmonella Derby Infection in Antibiotic-Treated or Lawsonia intracellularis Vaccinated Piglets

Simple Summary Today, pigs repeatedly suffer from diarrhoea which requires treatment with antibiotics. Infections with Lawsonia intracellularis are one of the most common diseases, which, although usually subclinical, have a negative impact on performance. The alternative to antibiotic treatment is vaccinations against Lawsonia intracellularis, not least because antibiotic treatments are suspected of promoting the spread of certain zoonotic pathogens. A study was carried out with piglets from a farm that had a problem with Lawsonia intracellularis infections. In half of the animals, antibiotic treatments with tylosin were carried out in piglet rearing. In the other group, the piglets had been vaccinated against Lawsonia intracellularis as suckling piglets. Individual animals from both groups were subsequently artificially infected with Salmonella as piglets. A total of 72 animals were included in the study, 12 of which were primarily infected. The other animals had the possibility of becoming infected via direct animal contact or the faeces of infected animals. The detection of Salmonella in stool and intestinal lymph nodes was significantly higher in animals previously treated with antibiotics. Treatment with tylosin may significantly increase the spread of the Salmonella infection not observed after early Lawsonia intracellularis vaccination. Abstract Lawsonia intracellularis infections are a common reason for antibiotic treatment in pig production. Experimental studies in animals naturally infected with Lawsonia intracellularis comparing the course of an experimental Salmonella infection in piglets previously treated with tylosin or vaccinated against Lawsonia intracellularis are scarce. A total of 72 seven-week-old Salmonella-free pigs were taken from a herd with a Lawsonia intracellularis history in piglet rearing. The pigs were divided into two groups with three replicates each. Animals had either been previously treated with tylosin (10 mg/kg body weight) for seven days (AB+VAC−) or had been vaccinated as suckling pigs by drenching (Enterisol®Ileitis; AB−VAC+). Two animals per replicate were primarily infected with Salmonella Derby (1.04 × 108 colony-forming units per animal). The detection of Salmonella in faeces (p < 0.0001, odds ratio: 3.8364) and in the ileocaecal lymph nodes (p = 0.0295, odds ratio: 3.5043) was significantly more frequent in AB+VAC− animals. Overall, the odds ratio for detecting Salmonella in any substrate or organ was significantly higher in the AB+VAC− group animals (p = 0.0004, odds ratio: 5.9091). Treatment with tylosin can significantly increase the spread of a Salmonella infection, which is not observed after early Lawsonia intracellularis vaccination.

The public health aspects are of special importance as most human cases of salmonellosis (∼90%) are attributable to layers/eggs and pigs [23]. Salmonellosis is still a major cause of human bacterial gastroenteritis and the second most reported zoonosis in the European Union (EU) after campylobacteriosis [24]. In 2015, a total of 94,625 confirmed salmonellosis cases were reported by 28 EU member states, resulting in an EU notification rate of 21.2 cases per 100,000 population [24].
To date, there is no experimental study comparing the effects of the two possible treatments (antibiotic metaphylaxis or preventive vaccination) with regard to the spread of an experimental Salmonella infection in natural L. intracellularis infected pigs from herds with early infections. In our model with naturally infected piglets, the experimental Salmonella infection most closely reflects the conditions in practice under maximally standardised conditions. The aim of the investigations was to determine whether differences exist in the spread of an artificial Salmonella infection in groups with potentially subclinically L. intracellularis infected piglets, comparing animals that were either previously immunised as suckling piglets against the pathogen or therapeutically treated with an antibiotic against L. intracellularis just before the experimental infection with Salmonella Derby.

Materials and Methods
Animal experiments were performed in accordance with the German rules and regulations and approved by the Ethics Committee of Lower Saxony for the Care and Use of Laboratory Animals (LAVES: Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit; reference: 33-9-42502-04-12/0902).

Origin of Animals and their Preparation and Selection on the Farm
The crossbred piglets for the trial came from one German farm with 420 sows of Danish genetics (damline: Danish Landrace 50% × Yorkshire 50% (DK); sire line: Pietrain, artificial insemination). The production system on the farm was organised at two-weekly intervals with 21 days of suckling in general with 450 to 500 piglets weaned per group. The piglets in the herd were regularly vaccinated (porcine circovirus 2, Mycoplasma hyopneumoniae and Haemophilus parasuis). In the context of regular and long-term screening programmes, the herd showed no signs of Salmonella and Brachyspira infections. The selection of the farm from which the animals came was made in a targeted manner. Three essential characteristics had to be fulfilled:

•
In the past, the pathogen L. intracellularis had been repeatedly detected and was suspected of hindering the optimal performance on the farm.

•
The farmer had to be in the process of testing a commercial vaccine in part of the herd on his own initiative in order to further improve animal performance.

•
The trial phase of the partial vaccination of the herd had to be supervised by the herd supervising veterinary practice.
The farm from which the animals were collected fulfilled these criteria. Clinical symptoms of an L. intracellularis infection (confirmed by pathogen detection in faeces) were a common finding in piglets between the ages of seven to nine weeks. At the time of the investigations, the farmer was faced with the decision to establish a vaccination against L. intracellularis in the herd. The animals had not been vaccinated so far. Therefore, individual litters were vaccinated over several months in order for the farmer and veterinary surgeon to clinically observe the effects on the performance in piglet rearing and fattening.
During the trial period, every two weeks, approximately 50 piglets from a total of four to five complete litters were vaccinated with a commercial L. intracellularis live vaccine (Enterisol ® Ileitis, Boehringer Ingelheim Vetmedica GmbH, Ingelheim/Rhine, Germany) on the 21st day of life. The vaccination was administered via oral drenching. Subsequently, the piglets were marked individually.
After weaning, all 450 to 500 piglets in a group were housed in a piglet rearing compartment in boxes with a maximum of 30 animals on fully slatted floors. Non-vaccinated and vaccinated animals were mixed. From weaning to the presumed initial L. intracellularis infection time point, samples from the individual boxes were tested several times for their L. intracellularis status by real-time PCR using established methods [25]. A single-animal examination procedure started with positive proof (Figure 1). All piglets in a weaning group were clinically examined. Individual samples were collected from non-vaccinated conspicuous animals with regard to faecal quality (healthy pigs with moderate to soft faecal consistency, no diarrhoea) and analysed by means of real-time PCR depending on L. intracellularis status [25]. After obtaining the results, conspicuous animals were taken from the appropriate weaning group (subsequently called the "AB + VAC − group" due to the antibiotic treatment necessary in the further course of the study). At the same time, vaccinated animals of the same age and identical weight were chosen from different boxes (subsequently called the AB − VAC + group). In the course of the routine herd diagnostics, blood samples for L. intracellularis proof were simultaneously available from these animals (about four-five weeks after weaning). All the selected animals were culturally tested for Salmonella by rectal swabs to exclude the risk of being positive for Salmonella before starting the trials at the university. After obtaining the results, animals were transported to the Institute for Animal Nutrition, University of Veterinary Medicine Hannover, Foundation, Hannover, Germany. non-vaccinated conspicuous animals with regard to faecal quality (healthy pigs with moderate to soft faecal consistency, no diarrhoea) and analysed by means of real-time PCR depending on L. intracellularis status [25]. After obtaining the results, conspicuous animals were taken from the appropriate weaning group (subsequently called the "AB + VAC − group" due to the antibiotic treatment necessary in the further course of the study). At the same time, vaccinated animals of the same age and identical weight were chosen from different boxes (subsequently called the AB − VAC + group). In the course of the routine herd diagnostics, blood samples for L. intracellularis proof were simultaneously available from these animals (about four-five weeks after weaning). All the selected animals were culturally tested for Salmonella by rectal swabs to exclude the risk of being positive for Salmonella before starting the trials at the university. After obtaining the results, animals were transported to the Institute for Animal Nutrition, University of Veterinary Medicine Hannover, Foundation, Hannover, Germany. The clinical trial flow diagram on the selection process of animals on the farm for the experimental trial at the university with artificial Salmonella infection, allocation to groups (AB + VAC − : animals were treated with tylosin (10 mg/kg body weight) for seven days (AB + VAC − ) during the trial; AB − VAC + : animals were vaccinated as suckling pigs by drenching (Enterisol ® Ileitis)), as well as the number of animals.

Experimental Conditions, Pretreatment, and Infection Model
In total, 72 piglets (DK x Pietrain) were divided into two groups with three subgroups of 12 animals each. The first group was treated metaphylactically with tylosin (Tylan ® G 25%, Elanco Deutschland GmbH, Bad Homburg, Germany; dosage: 10 mg/kg body weight orally) for a minimum of five days (AB + VAC − ; Figure 2), whereas the second group (vaccinated as suckling pigs by drenching (Enterisol ® Ileitis)) was given no antibiotics (AB − VAC + ). The drug was mixed into a small amount of feed each morning during the treatment period. This mixture was offered to the animals until complete ingestion before the rest of the daily feed was offered.

Experimental Conditions, Pretreatment, and Infection Model
In total, 72 piglets (DK x Pietrain) were divided into two groups with three subgroups of 12 animals each. The first group was treated metaphylactically with tylosin (Tylan ® G 25%, Elanco Deutschland GmbH, Bad Homburg, Germany; dosage: 10 mg/kg body weight orally) for a minimum of five days (AB + VAC − ; Figure 2), whereas the second group (vaccinated as suckling pigs by drenching (Enterisol ® Ileitis)) was given no antibiotics (AB − VAC + ). The drug was mixed into a small amount of feed each morning during the treatment period. This mixture was offered to the animals until complete ingestion before the rest of the daily feed was offered. and contact pigs (secondarily infected); AB + VAC − seeder − tylosin = two animals per group were administered tylosin (Tylan ® G 25%, Elanco Deutschland GmbH, Bad Homburg, active ingredient: tylosin phosphate, dosage: 10 mg/kg body weight orally) for five days; AB + VAC − contact pigs−tylosin = ten pigs per group were administered tylosin (Tylan ® G 25%, Elanco Deutschland GmbH, Bad Homburg, active ingredient: tylosin phosphate, dosage: 10 mg/kg body weight orally) for seven days; seeder pigs in each group were returned to the respective group at day 10; after that, the observation period started. Pigs in the AB − VAC + group had been vaccinated as suckling pigs by drenching (Enterisol ® Ileitis) without further antibiotic treatment.
The boxes (3.00 × 2.15 m) were equipped with a nipple drinker, two stainless steel troughs (each 0.80 m long at one of the narrow sides of the box), and an infrared warming lamp. The boxes had a concrete floor. Additional boxes (3.00 × 1.05 m) were placed adjacent to the subgroup boxes. After five days, two pigs per group (later referred to as seeder animals) were housed in these separate boxes for the purpose of artificial Salmonella infection. The infection took place two days later, two days after ending the tylosin treatment of these seeder animals. The remaining animals of the AB + VAC − group were administered tylosin for a further two days, i.e., a total of seven days. Therefore, it was guaranteed that at the time of artificial infection (seeder pigs), or rather, natural infection (contact pigs) the animals had been at least 48 without antibiotics.
The experimental infection was carried out in both groups in accordance with the model of Papenbrock et al. [26]. In each case, two seeder animals from one box of the subgroups from the AB + VAC − and AB − VAC + groups were infected. In total, three batches of infection broth were used (batch 1: 8.70 × 10 7 ; batch 2: 1.10 × 10 8 ; batch 3: 1.15 × 10 8 per animal, respectively). The strain A147/85 of Salmonella enterica ssp. enterica serovar Derby (antigenic formula O 1, 4, [5], 12, H f, g [1,2]) used for infecting the seeder pigs was provided by the Institute of Microbiology of the University of Veterinary Medicine, the Hannover Foundation, where it was stored in a lyophilised form. This strain had been previously used successfully for experimental infection of pigs [26].
In order to increase the gastric pH and thus prevent damage to the germs by the very acidic stomach environment of pigs [27], the seeder animals were each given 50 g of feed in the morning. After about 15 min, another 50 g of feed was mixed with 10 mL of freshly prepared infection broth and presented to the pigs for ingestion. The amount of feed was kept to a minimum to ensure a fast and complete intake. After complete ingestion of the feed containing the infection broth, the animals were given food ad libitum as usual. After successfully detecting the infection strain by means of rectal swabs, the experimentally infected animals were returned to their original subgroups (ten days after starting the trials, three days after experimental infection). In doing so, contact animals came into contact with the pathogen. From this point onwards, a four-week follow-up examination using rectal swabs (2,4,6,8,10,12,15,17,19,22,24, and 26 days after returning the seeder pigs to the group) was followed by a necropsy of all animals at days 38-40 after starting the trials. In each group, after confirming the successful infection of the seeder pigs, 144 swab samples (12 per animal; 864 swabs in the entire study in six subgroups) were taken over the entire four-week experimental period. and contact pigs (secondarily infected); AB + VAC − seeder − tylosin = two animals per group were administered tylosin (Tylan ® G 25%, Elanco Deutschland GmbH, Bad Homburg, active ingredient: tylosin phosphate, dosage: 10 mg/kg body weight orally) for five days; AB + VAC − contact pigs−tylosin = ten pigs per group were administered tylosin (Tylan ® G 25%, Elanco Deutschland GmbH, Bad Homburg, active ingredient: tylosin phosphate, dosage: 10 mg/kg body weight orally) for seven days; seeder pigs in each group were returned to the respective group at day 10; after that, the observation period started. Pigs in the AB − VAC + group had been vaccinated as suckling pigs by drenching (Enterisol ® Ileitis) without further antibiotic treatment.
The boxes (3.00 × 2.15 m) were equipped with a nipple drinker, two stainless steel troughs (each 0.80 m long at one of the narrow sides of the box), and an infrared warming lamp. The boxes had a concrete floor. Additional boxes (3.00 × 1.05 m) were placed adjacent to the subgroup boxes. After five days, two pigs per group (later referred to as seeder animals) were housed in these separate boxes for the purpose of artificial Salmonella infection. The infection took place two days later, two days after ending the tylosin treatment of these seeder animals. The remaining animals of the AB + VAC − group were administered tylosin for a further two days, i.e., a total of seven days. Therefore, it was guaranteed that at the time of artificial infection (seeder pigs), or rather, natural infection (contact pigs) the animals had been at least 48 without antibiotics.
The experimental infection was carried out in both groups in accordance with the model of Papenbrock et al. [26]. In each case, two seeder animals from one box of the subgroups from the AB + VAC − and AB − VAC + groups were infected. In total, three batches of infection broth were used (batch 1: 8.70 × 10 7 ; batch 2: 1.10 × 10 8 ; batch 3: 1.15 × 10 8 per animal, respectively). The strain A147/85 of Salmonella enterica ssp. enterica serovar Derby (antigenic formula O 1, 4, [5], 12, H f, g [1,2]) used for infecting the seeder pigs was provided by the Institute of Microbiology of the University of Veterinary Medicine, the Hannover Foundation, where it was stored in a lyophilised form. This strain had been previously used successfully for experimental infection of pigs [26].
In order to increase the gastric pH and thus prevent damage to the germs by the very acidic stomach environment of pigs [27], the seeder animals were each given 50 g of feed in the morning. After about 15 min, another 50 g of feed was mixed with 10 mL of freshly prepared infection broth and presented to the pigs for ingestion. The amount of feed was kept to a minimum to ensure a fast and complete intake. After complete ingestion of the feed containing the infection broth, the animals were given food ad libitum as usual. After successfully detecting the infection strain by means of rectal swabs, the experimentally infected animals were returned to their original subgroups (ten days after starting the trials, three days after experimental infection). In doing so, contact animals came into contact with the pathogen. From this point onwards, a four-week follow-up examination using rectal swabs (2,4,6,8,10,12,15,17,19,22,24, and 26 days after returning the seeder pigs to the group) was followed by a necropsy of all animals at days 38-40 after starting the trials. In each group, after confirming the successful infection of the seeder pigs, 144 swab samples (12 per animal; 864 swabs in the entire study in six subgroups) were taken over the entire four-week experimental period.

Feed and Feed Analysis
The animals were given a complete diet (meal) for pigs of a common composition during the experiment (For Farmers-Bela GmbH, Langförden, Germany; Table 1). The diet contained wheat, barley, soybean meal, soybean oil, fried bread, wheat bran, calcium carbonate, monocalcium phosphate, sodium chloride, and minerals in descending order. No organic acids were added to the diet. The pigs were fed ad libitum once a day at 08:00. The diets were analysed by standard procedures in accordance with the official methods of the VDLUFA [28]. The analyses were always performed in duplicate. The dry matter content was determined by drying to a constant weight at 103 • C. The crude ash was analysed by means of incineration in the muffle furnace at 600 • C for six hours. The total nitrogen content was determined by means of the analyser Vario Max (Elementar, Hanau, Germany), which operates according to the principle of a catalytic tube combustion (DUMAS combustion method). The molecular nitrogen formed by the reduction from nitric oxide was detected by a thermal conductivity detector and the nitrogen content was calculated by the device software. The crude protein content of the sample was calculated by multiplying it with a constant factor of 6.25. The crude fat content was determined after acid hydrolysis in the Soxhlet apparatus. The content of crude fibre was determined after washing in diluted acids and alkalis using established methods. The starch content was determined polarimetrically (Polatronic E, Schmidt und Haensch GmbH and Co., Berlin, Germany). The sugar content was analysed in accordance with Luff-Schoorl by titration with sodium thiosulphate. The mineral content was determined in accordance with the official methods [28] by atomic absorption spectrometry (UnicamSolaar 116, Thermo, Dreieich, Germany).

Lawsonia intracellularis and Salmonella Diagnostics
The serological tests (at the start of the trial and at necropsy) concerning L. intracellularis were carried out using a sandwich blocking ELISA in accordance with Keller et al. [29]. This ELISA has a specificity of 98.7% and a sensitivity of 96.5% and works with specific monoclonal antibodies. Cut-off values for the blocking ELISA test are given as per cent inhibition (PI) with a cut-off value of PI 30. Direct confirmation of the presence of L. intracellularis in the faeces of pigs in the AB + VAC − group was made from the faeces collected on the farm via real-time PCR with established methods [25].
The serological tests concerning Salmonella were performed with an IDEXX Swine Salmonella Ab Test at the start of the trial (verification of being Salmonella unsuspicious) and at necropsy (an indicator of contact intensity with the pathogen). This test detects antibodies against the most common serotypes (B, C1, D) isolated in Europe, Asia, and America. The results are available as OD% values using IDEXX xChek ® software. The samples were considered positive if the optical density (OD) was ≥10%.
For bacteriological Salmonella testing, the qualitative examination of different samples was carried out after liquid pre-enrichment (feed samples, organ samples). Feed and organ samples were non-selectively pre-enriched in peptone water (Oxoid, Wesel, Germany) for 24 h at 37 • C. Subsequently, the liquid selective Rappaport-Vassiliadis (RV) enrichment medium (Oxoid, Wesel, Germany) and Tetrathionate Brilliant Green Galle Broth (TBG; Merck, Darmstadt, Germany) were supplemented with 0.1 mL (for 9.9 mL RV) and 1 mL (for 9 mL TBG) of the suspension. These were further incubated for 48 h at 42 • C. Both after 24 h and after 48 h, one loop from each of the media was streaked onto both of the two selective nutrient media (Brilliance TM Salmonella, Oxoid, Wesel, Germany; brilliant green phenol red lactose sucrose agar-BPLS, Oxoid, Wesel, Germany). The culture medium was incubated at 37 • C (Brilliance) or 42 • C (BPLS). Salmonella-suspected colonies were determined by morphology, the staining of colonies, and the medium. Salmonella-suspected colonies were isolated and the surface antigens were characterised by means of rapid slide agglutination (anti-Salmonella O 4, anti-Salmonella O 5, Sifin GmbH, Berlin, Germany). Further differentiations were made using a Kligler agar (lactose negative, glucose positive, H 2 S-formation; [30]). For more far-reaching differentiation, colony material from the Kligler agar was subjected to further slide rapid agglutination for ascertaining the presence of flagellum antigens Hg and Hf (anti-Salmonella Hf, anti-Salmonella Hg, Sifin GmbH, Berlin, Germany). Both flagellum antigens are present in S. Derby.

Statistical Analyses
Analyses were carried out with the statistical software SAS, version 9.3 (SAS Institute, Cary, NC, USA), using SAS ® Enterprise Guide 5.1.
The comparison for quantitative variables between groups (independent samples; AB + VAC − , AB − VAC + ) and within a group (paired t-test) was carried out by two-sample t-tests (BW; ADWG; FCR (normal distribution was assumed for n = 3)) for normal distributed data and by means of the Wilcoxon test in non-normally distributed data.
The analysis of the qualitative features (Salmonella detection in rectal swabs, caecal content, ileocaecal lymph nodes) was carried out using the Pearson's chi-square homogeneity test. The odds ratio was also calculated for group comparison.
Differences were taken to be statistically significant when p < 0.05.

Results
The investigations ran without any abnormalities dependent on animal health. Even after the experimental infection with S. Derby, there were no clinical signs of infection. There were no animal losses during the trial.

Performance
The body weights at the start of the investigations, the average daily feed intake, the final body weight, the average daily weight gain, and the feed conversion ratio did not differ between groups (Table 2).

Serological and Microbiological Salmonella Statuses
The non-vaccinated animals (AB + VAC − ) showed significantly higher L. intracellularis antibody titres at the time of selection of the animals on the farm compared to the vaccinated group and were characterised as serologically positive (Ø PI values at group level ≥30; Table 3). The vaccinated animals, however, were serologically negative. After transport to the university, at the start of the trials, the animals were regrouped in new boxes. At the end of the experiment, the mean L. intracellularis, PI-values were significantly higher in both groups, whereas the final values in the AB − VAC + group were significantly higher than in the AB + VAC − group. Table 3. The mean antibody titres in the blood (L. intracellularis and Salmonella Derby) of tylosin-treated (AB + VAC − ) and vaccinated animals (AB − VAC + ).

Item
Time The serological Salmonella status between the groups did not differ at the beginning of the investigations. Both groups scored average within the range that is considered negative by the test kit manufacturer (OD% values < 10). At the end of the experiment, the mean serological values in the AB + VAC − group were significantly higher (AB + VAC − -OD%: 25.3 ± 26.3) than in the AB − VAC + group (OD%: 14.1 ± 16.3). At the end of the experiment, 19 out of 33 pigs in the AB + VAC − group and 15 out of 33 pigs in the AB − VAC + group showed an OD% value higher than or equal to ten. Three samples in each group could not be examined.
All primary infected animals showed a Salmonella excretion in the faeces before being returned to the group. Additionally, during the trial, these animals again showed Salmonella shedding. Salmonella Derby was significantly more frequently detected in faecal samples from the AB + VAC − group throughout the experiment ( Table 4). The same applied to the number of positive animals regarding the faeces and the Salmonella prevalence in samples of ileocaecal lymph nodes and caecal content. If one of the microbiologically tested samples was Salmonella positive, the animal was classified as Salmonella positive overall. Therefore, the number of pigs (both seeder and contact pigs) which were detected positive in a minimum of one sample after being returned to the subgroup was significantly higher in the AB + VAC − group at days 4,12,15,17,19,22,24,26, and at necropsy (Figure 3).  If one of the microbiologically tested samples was Salmonella positive, the animal was classified as Salmonella positive overall. Therefore, the number of pigs (both seeder and contact pigs) which were detected positive in a minimum of one sample after being returned to the subgroup was significantly higher in the AB + VAC − group at days 4,12,15,17,19,22,24,26, and at necropsy (Figure 3).

The Interaction between Serological and Microbiological Salmonella Status
The serological Salmonella status of animals prior to the start of the study was identical for animals which remained negative or became positive after the experimental Salmonella infection during the study (Table 5).

The Interaction between Serological and Microbiological Salmonella Status
The serological Salmonella status of animals prior to the start of the study was identical for animals which remained negative or became positive after the experimental Salmonella infection during the study (Table 5). The comparison of animals that were considered Salmonella negative with respect to the defined samples (faeces, ileocaecal lymph node, caecal content, sum of all locations) showed significantly lower serological Salmonella responses in the ELISA than animals in the AB − VAC + group at the end of the experiment for all sample types (Table 6). In addition, a significant difference in the serological response to Salmonella was seen in this group when comparing animals that were either negative (OD%: 11.2 ± 13.1) or positive (OD%: 42.7 ± 5.03) for the detection of Salmonella in the caecum. Table 6. The serological Salmonella status (OD%) at necropsy in relation to Salmonella detection during the trial in tylosin-treated (AB + VAC − ) and vaccinated animals (AB − VAC + ).

Sample
Microbiological S. AB + VAC − = treated with tylosin before artificial S. Derby infection, unvaccinated; AB − VAC + = vaccinated with Enterisol ® Ileitis in the suckling period (day 21); IDEXX Swine Salmonella Ab Test: OD% values were considered positive if the optical density (OD) was ≥10%; upper case letters ( A, B ) signify differences in columns (vertical) within a group for analysed parameters at p < 0.05; lower case letters ( a, b ) signify differences in rows (horizontal) between groups depending on parameter at p < 0.05.
The comparison of animals depending on ∆-values concerning serological Salmonella responses in the ELISA showed significantly lower results for animals with a Salmonella negative status in ileocaecal lymph nodes and in the caecal content in the AB − VAC + group (Table 7). In the AB − VAC + group, the ∆-values concerning serological Salmonella responses were also significantly lower in animals with a negative Salmonella status in the caecal content in comparison to pigs where Salmonella was found in the caecal content. Table 7. The delta-values in the serological Salmonella status (OD%) between the start of the experiment and necropsy in relation to Salmonella detection during the trial in tylosin-treated (AB + VAC − ) and vaccinated animals (AB − VAC + ).

Discussion
Lawsonia intracellularis is of great importance as a causative agent of digestive diseases in today's pig production [1-4,31]. Nowadays, vaccination or antibiotic treatment are the two alternatives used to deal with the pathogen. The antibiotic treatment of L. intracellularis infections is discussed against the background regarding whether the treatment can accelerate the spread of zoonotic-relevant pathogens like Salmonella. To deal with these issues, artificial Salmonella infection experiments were conducted with naturally L. intracellularis-infected piglets.

Effects on Performance
In the present study, no significant differences concerning performance data could be seen. The animals were the same age and weight at the start of the experiment. The one-week treatment with tylosin did not have a positive or a negative effect on ADWG and FCR. For tylosin, the growth-promoting effects are known, especially from earlier times when tylosin was still used as a growth promotor [32,33]. In a study on pigs at 38 to 58 days of age reduction, the feed intake was 3.79% after experimental infection (1.26 × 10 10 L. intracellularis organisms in the inoculum) in comparison with animals treated with 50 ppm tylvalos for 14 further days after experimental infection [34]. In terms of performance, in our study, there was no significant difference between the experimental groups (AB + VAC − : therapeutically 10 mg/kg BW, AB − VAC + : prophylactic vaccination). Therefore, in the present study, it can be assumed that both alternatives work equally well in terms of performance.

Lawsonia intracellularis and Salmonella Infection
The pigs used in these trials came from a farm with a history of early L. intracellularis infections. Animals in the AB + VAC − group had already shown a serological response immediately before the start of the experiments with artificial Salmonella infections, which indicated an intensive field contact. The PCR samples of faecal material in the AB + VAC − group support this finding. By contrast, animals in the AB − VAC + group were serologically negative at the start. With regrouping after transport, a significant increase in antibody levels was seen. At the end of the experiments, the animals vaccinated as suckling piglets against L. intracellularis (AB − VAC + ) showed the highest antibody titers. Intensive contact with the pathogen fosters the antibody response in pigs. Positive correlations (Pearson's correlation coefficient) were observed between antibody concentrations in pigs at 21 days post-infection (orally dosed at 6 weeks of age) and the number of L. intracellularis shed 14 days (r = 0.52), 17 days (r = 0.58), and 21 days (r = 0.50) after infection [11]. Additionally, serum titers were higher in pigs challenged with a pathogenic isolate than in those exposed to the vaccine, which ranged from 30 to 3820 and from 30 to 480, respectively [35]. In most cases, pigs become seropositive two-three weeks after primary inoculation with L. intracellularis [36]. In vaccinated pigs, sometimes there can be a delayed onset of seroconversion [35]. In the named study, the authors explained the delayed seroconversion and the lower IgG titers observed in the pigs in the vaccine group compared to the animals in the pathogenic isolate group as being caused by a combination of factors, such as a lower concentration of L. intracellularis organisms in the vaccine compared to the pathogenic inoculum (5.3 × 10 5 and 1.76 × 10 8 organisms/mL, respectively), the route of administration (drinking water for the vaccine group and intragastrically for the pathogenic isolate group), and the immunogenicity of the isolate [35]. The higher antibody titers in the vaccinated group in this study may be explained by the fact that systemic antibody responses were boosted following challenge [37]. This development of the serological measurements clearly shows that the model worked well. In both groups, we had a serological reaction within the experimental period, indicating a more intensive contact with L. intracellularis.
Regarding the serological Salmonella status, all animals were serologically negative at the beginning of the experiments. During the trial, serological reactions to Salmonella increased significantly in both groups. Nevertheless, individual animals showed titers above the cut-off value already at the start of the experiment. In piglet rearing, maternal antibodies to Salmonella can be detected up until the seventh week of life [38]. Individual positive ELISA responses in both experimental groups at the start of the experiment are therefore likely to be due to the presence of maternal antibodies.
At the end of the trial, the prevalence of serologically positive samples was significantly higher in the AB + VAC − group. A major advantage of serodiagnostics compared with bacteriological tests is those specific IgG antibodies generally persist, whereas Salmonella are excreted intermittently and pigs may harbour infections without excreting Salmonella over a prolonged period [39]. However, it should be mentioned that these tests are not suitable as an individual pig test because not all pigs seroconvert [39].
The odds ratio for positive Salmonella detection in individual samples or animals in general (independent of location, i.e., in faeces, ileocaecal lymph nodes, caecal content), was greatly increased in the AB + VAC − group. This observation is in line with the results from European epidemiological studies, which see a clear risk of increased Salmonella seroprevalence when using antibiotic substances, both in the lower dosage as growth promotor or in therapeutic dosages [20,21] In the German study, the application of antibiotics increased the odds ratio for being serologically Salmonella conspicuous by a factor of 5.21 (p < 0.001) compared to untreated pigs [21]. A direct effect of tylosin on the commensal gastrointestinal flora has already been described for concentrations in the range of a body weight of 2 mg/kg [15]. In swine, the adverse effect of antibiotic treatment on gut flora persisted for only about a week, but animals are five to six times more susceptible to gastrointestinal tract infections during this time [40]. According to a US study, using the antimicrobial growth promoter tylosin did not pose a public health risk. There was no effect on the carriage of S. enterica [19]. From the point of view of the present investigations, it can be assumed that the antibiotic treatment fostered the spread of the Salmonella infection. However, it cannot be ruled out that the vaccination itself has a protective effect. Vaccination against L. intracellularis at three weeks of age significantly reduced S. Typhimurium shedding (p < 0.05) in co-infected animals in comparison to the co-infected group without vaccination and the group challenged with S. Typhimurium alone [41]. Twenty-one days post vaccination, the animals were challenged with a pure culture of L. intracellularis (2 × 10 9 organisms per pig; strain PHE/MN1-00). One week post-L. intracellularis challenge, the pigs were challenged orally with S. Typhimurium (strain 798; 1 × 10 8 organisms per pig) [41]. At seven days post-infection, the co-challenged non-vaccinated group shed 2.94 log 10 S. Typhimurium organisms per gramme faeces, while the vaccinated co-challenged group shed 0.82 log 10 S. Typhimurium organisms per gramme faeces (p = 0.003) [41]. These results indicate that the vaccination against L. intracellularis impacts the microbiome and reduces shedding of S. Typhimurium in co-infected animals [41]. These previous findings are in agreement with the results of the present study. In our study, the animals were also vaccinated against L. intracellularis in the suckling pig phase, selected about three weeks later and transported from the farm to the university. On arrival, the animals were regrouped and thus the L. intracellularis infection was re-established. It is known that regrouping and/or mixing animals from different herds in the finisher unit might increase the rate of transmission of L. intracellularis [42]. In both groups in our trial, the serological response to L. intracellularis increased significantly during the trial. In week two at the institute, the experimental infection with S. Derby took place. This experimental model is therefore comparable in principle to the design described above [41]. Therefore, this study supports the recommendation that the L. intracellularis vaccination may be used as a novel tool to aid the control of Salmonella on swine farms, as well as using it as an alternative measure to reduce the need for the antibiotic treatment of pigs, thereby improving food safety [41].

Interactions between Serological and Microbiological Findings
At the end of the experiment, animals with a negative status depending on different samples (faeces, ileocaecal lymph nodes, caecal content, one of the named samples positive for S. Derby) from the AB − VAC + group had significantly lower Salmonella antibody values in the blood in each case than animals from the AB + VAC − group. The delta values between the end of the experiment and the start of the experiment were also significantly lower for animals with negative findings in the lymph nodes and the caecal content in the AB − VAC + group compared to the AB + VAC − group.
High antibody levels also reflect a certain degree the history of cultural detection of Salmonella. According to the results of a previous experimental Salmonella infection study (4.4 × 10 9 CFU Salmonella Typhimurium DT104), the majority of pigs produced no anti-Salmonella immunoglobulin G within the third week in spite of high Salmonella excretion rates [39]. Afterwards, all pigs showed a high seroprevalence, but in contrast, a low prevalence of Salmonella Typhimurium DT104 was shown in the faeces [39]. Another study on 1658 finishers from 167 herds suspected that there was a correlation between the herd seroprevalence (cut-off 20 OD%) and the probability of Salmonella isolation from 25 g of caecal content [43]. Overall, pigs with Salmonella-positive caecal content showed the highest mean optical densities in meat juice in another study, namely, 37.1 OD% (n = 124; [44]). Completely microbiologically negative animals (n = 178) instead had an LsMean OD% of 10.9 [44]. From these results, one can conclude that the differences in bacteriological findings in the present study are not overestimated because microbiologically negative animals from the AB + VAC − group also had significantly higher Salmonella antibody values.

Conclusions
In conclusion, the present results clearly show that the antibiotic therapy of an infection with L. intracellularis leads to a higher susceptibility to the S. Derby infection in comparison to the alternative prophylactic vaccination. In the AB + VAC − group pretreated with antibiotics, Salmonella spreading significantly increased after experimental infection compared to the AB − VAC + group. In vaccinated pigs, therapy is not required. Therefore, with regard to the prevalence of Salmonella infections in pigs, excellent protection against Salmonella is expected from the vaccine. supported by the Deutsche Forschungsgemeinschaft (German Research Foundation), Bonn, Germany and the University of Veterinary Medicine Hannover, Foundation within the funding programme Open Access Publishing.