The efficacy and safety of a marine-derived Bacillus strain for use as an in-feed probiotic for newly weaned pigs Field of Invention

The present invention relates to the treatment of animals to improve their performance and health through the use of a Bacillus pumilus spore suspension. In particular the invention relates to the use of the probiotic Bacillus pumilus to decrease the bacteria within the animal which may adversely affect performance and health.

Background to the Invention

Weaning is a stressful event for young animals and in particular pigs which is characterized by gastrointestinal disturbances caused by physiological, immunological and microbiological changes within the gastrointestinal tract. During this period, animals are highly susceptible to enteric diseases, and those caused by Escherichia coli (e.g. post-weaning diarrhea and edema disease) are responsible for considerable economic losses in the pig industry. As a result, in-feed antibiotics have long been used for the elimination or reduction of pathogenic bacteria, in particular E. coli, during the post-weaning period. However, the routine use of in-feed antibiotics was banned in the EU in 2006 (Regulation EC/1831/2003), although their use is still permitted under veterinary prescription as the need arises. For instance, apramycin (Apralan G200, Elanco Animal Health, Eli Lilly & Co. Ltd) was prescribed for use on the pig farm where the current study was conducted to control persistent edema disease during the post-weaning period.

However, antibiotic -resistance is a major human health issue and effective alternatives to antibiotics are required. In-feed zinc oxide, at pharmacological concentrations (i.e. concentrations in excess of normal dietary requirements) is also commonly used for enteric disease prevention in weaned pigs but there are concerns about its accumulation in the environment.

Probiotics are defined as 'live microorganisms which when administered in adequate amounts confer a health benefit on the host'. They offer potential as an alternative to antibiotics for pigs, both as a means of controlling enteric pathogens and improving growth rate and feed conversion. Together with modulation of the immune system and competitive exclusion, antimicrobial production is one of the suggested mechanisms of action of probiotics. The latter can therefore be considered a probiotic trait and is often listed as one of the properties required of a strain for it to be considered probiotic. This is backed up by the fact that strains selected in vitro for their anti-E. coli activity, have proven successful in reducing E. coli shedding, preventing diarrhea and improving growth performance in pigs. However, as is usually the case with probiotics, benefits are strain-specific. The marine environment is largely untapped as a source of probiotics but should not be overlooked given that it is a potential source of novel microorganisms and that antimicrobial production is common amongst marine microflora. In fact, in vitro data has demonstrated that marine bacteria have potential as probiotics for animal production. One strain of Bacillus pumilus showed most promise as it satisfies a number of probiotic criteria and has activity against E. coli without being cytotoxic. However, as with any potential feed additive, observations made in vitro need to be substantiated with in vivo data and to date, this marine B. pumilus strain has not been tested in vivo.

The aim of the trial was to evaluate this pre- screened Bacillus strain for use as an in-feed probiotic for newly weaned pigs in comparison to a negative control treatment without antibiotic or pharmacological levels of zinc oxide (non-medicated treatment) and a positive control treatment containing apramycin and pharmacological levels of zinc oxide (medicated treatment). Key parameters including growth performance and health indicators were investigated in order to evaluate safety and efficacy in vivo.

Summary of the Invention

The B. pumilus treatment decreased ileal E. coli counts in a manner similar to medicated treatment but without the adverse effects on growth performance, Lactobacillus counts, cecal short chain fatty acid concentration and possible liver toxicity experienced with the medicated treatment.

Forty eight individual pigs (8.7 + 0.26 kg) weaned at 28 + 1 d of age were used in a 22-d study to evaluate the effect of oral administration of a Bacillus pumilus spore suspension on growth performance and health indicators. Treatments (n = 16) were: (1) non-medicated diet; (2) medicated diet with apramycin (200 mg/kg) and pharmacological levels of zinc oxide (2,500 mg zinc/kg) and (3) B. pumilus diet (non-medicated diet + 10 ¹⁰ spores/day B. pumilus). Final body weight and average daily gain tended to be lower (P = 0.07) and feed conversion ratio was worsened (P < 0.05) for the medicated treatment compared to the B. pumilus treatment. Ileal E. coli counts were lower for the B. pumilus and medicated treatments compared to the non- medicated treatment (P < 0.05), perhaps as a result of increased ileal propionic acid concentrations (P < 0.001). However, the medicated treatment reduced fecal (P < 0.001) and cecal (P < 0.05) Lactobacillus counts and tended to reduce the total cecal short chain fatty acid concentration (P = 0.10). Liver weights were lighter and concentrations of liver enzymes higher (P < 0.05) in pigs on the medicated treatment compared to those on the non-medicated or B. pumilus treatments. Pigs on the B. pumilus treatment had lower overall lymphocyte and higher granulocyte percentages (P < 0.001) and higher numbers of jejunal goblet cells (P < 0.01) than pigs on either of the other two treatments or the non-medicated treatment, respectively. However, histopathological examination of the small intestine, kidneys and liver revealed no abnormalities.

According to the invention, there is provided, as set out in the appended claims, a Bacillus pumilus strain spore suspension for use as an in-feed probiotic in animals where the suspension satisfy a number of probiotic criteria and in particular for its ability to inhibit porcine pathogenic E. coli;

where that dietary supplementation with Bacillus pumilus strain spore suspension improved the growth performance ;

where the suspension resulted in reduced ileal E. coli counts and an improvement in feed conversion ratio compared to the medicated treatment due to an increase in average daily gain; where the intestinal survival of the administered strain spore suspension was proven; and where the decrease in E. coli was achieved without the reductions in potentially beneficial Lactobacillus that occurs with medicated feeds.

In one embodiment, the Bacillus pumilus strain spore suspension administered has antimicrobial activity against E. coli in vitro due to the higher concentrations of propionic acid which is used to reduce the growth of Enterobacteriaceae; where intestinal concentrations of short chain fatty acid are increased by the Bacillus pumilus strain spore suspension as a result of carbohydrate degradation by the administered strains;

where the reduction in ileal E. coli counts leads to the proliferation of endogenous bacteria which may increase propionic acid concentrations.

In one embodiment, the Bacillus pumilus strain spore suspension lowered E. coli counts without toxic effects;

where in contrast, liver weight was -10% decreased for animals fed the medicated treatment, which could be the result of chronic toxicity of apramycin leading to a reduction in size, when it is unable to regenerate following a period of ongoing insult.

In one embodiment, the Bacillus pumilus strain spore suspension resulted in no histopathological abnormalities in the liver;

where compared to overall serum concentrations of the liver enzymes, GGT, ALT and

ALP were 21, 30.2 and 67.7% higher, respectively, in pigs fed the medicated treatment, indicating possible liver damage;

where higher concentrations of serum creatinine found in pigs fed the medicated treatment could indicate kidney damage compared to the B. pumilus treatment.

In one embodiment, pigs fed the Bacillus pumilus strain spore suspension had lower percentages of lymphocytes and monocytes, compared to pigs on the non-medicated and medicated treatments, and to pigs on the non-medicated treatment, respectively;

where the histopathological examination revealed no signs of intestinal inflammation.

In one embodiment of the invention, there is provided a method for improving pig health, the method comprising administering to pig a composition comprising a suspension of spores from a strain of Bacillus pumilus, or a variant thereof, and a pharmaceutically acceptable excipient. In the specification that the term "improving pig health" should be understood to mean decreased ileal E. coli counts, improved growth performance of the pig receiving the composition, improved liver function and physiology, improved renal function, reduced inflammation, improved feed conversion ratio when compared to pigs receiving medical treatment such as antibiotics, and/or improved intestinal tract physiology and beneficial gut flora.

In one embodiment, the composition comprising the B. pumilus strain is formulated for administration in a probiotic form.

In one embodiment, the B. pumilus strain is a rifampicin-resistant strain isolated from seaweed.

In one embodiment, the suspension of spores from the strain of Bacillus pumilus, or variant thereof, has a concentration of 8 ¹⁰ to 12 ¹⁰ spores/ml, and ideally a concentration of 10 ¹⁰ spores/ml.

In one embodiment, the method comprises administering between 1 to 10 mis of the spore suspension. Preferably, the composition is administered orally. Preferably, the composition is administered daily. Ideally, the method comprises the step of administering the composition daily ad libitum.

Preferably, the method comprises the step of administering initially at least 5 x 10 ¹⁰ spores/ml to the pig in need thereof. Thereafter, the method comprises the step of administering 10 ¹⁰ spores/ml to the pig in need thereof.

Details of the Invention

Forty-eight crossbred (Large White x Landrace) pigs (24 male and 24 female; 8.7 + 0.26 kg) were weaned at 28 + 1 d of age and blocked by sex, weight and litter origin, before being randomly assigned as individual pigs to one of three dietary treatments (n = 16 pigs/treatment) as follows: (1) non-medicated diet (negative control); (2) medicated diet containing 200 mg apramycin/kg (Apralan G200, Elanco Animal Health, Eli Lilly & Co., Basingstoke, Hampshire, UK) and zinc oxide (2,500 mg zinc /kg provided from Zincotec, Provimi Ltd., Lichfield, Staffordshire, UK) (positive control); and (3) non-medicated diet + ~10 ¹⁰ spores B. pumilus WIT 588 daily (prepared and administered as outlined below). Treatments were administered continuously for 22 d post-weaning and pigs were provided with ad libitum access to feed and water. The diets were manufactured in the Moorepark feed mill and were formulated to meet or exceed the National Research Council requirements for weaned pigs. All phase 1 diets were formulated to 16.2 MJ/kg digestible energy and 16.2 g/kg lysine using the same ingredients except that apramycin and pharmacological levels of zinc oxide were added to the medicated diet. Similarly, all phase 2 diets were formulated to 15.0 MJ/kg digestible energy and 15.0 g/kg lysine using the same ingredients except that apramycin and pharmacological levels of zinc oxide were added to the medicated diet. All diets were fed in 3 mm pellet form. Phase 1 diets were fed for 1 week after which phase 2 diets were fed for the remainder of the experiment. Bacillus pumilus WIT 588 is a rifampicin resistant variant of a strain previously isolated from seaweed. It was generated to facilitate enumeration in the porcine GIT and characterized in vitro as a probiotic for animal production. The strain was grown aerobically for -24 h in Brain Heart Infusion (BHI) broth (Oxoid Ltd, Basingstoke, Hampshire, UK) at 37 °C with agitation at 200 rpm. It was then induced to sporulate by spread-plating 1 mL of this culture onto sporulation agar and incubating for 7 d at 37 °C. The plates were then flooded with 10 mL sterile ice-cold water and the cells were suspended using a glass spreader. This suspension was heated at 80 °C for 15 min to kill any vegetative cells. The spore concentration was determined by diluting the suspension 10-fold in maximum recovery diluent (MRD; Merck, Darmstadt, Germany), and spread-plating on BHI agar incubated at 37 °C for 24 h. The concentration was adjusted to ~10 ¹⁰ spores/mL, and aliquots of this spore suspension were stored at -20 °C until use.

Aliquots of spore suspension were thawed at 4 °C each night before use, as required. On the day of weaning (d 0), pigs on the B. pumilus treatment received an oral dose of 5 mL of the spore suspension by syringe immediately after weaning (~5 x 10 ¹⁰ spores B. pumilus WIT 588 per pig). Pigs on the two other treatments received an oral dose of 5 mL sterile distilled water by syringe so that the handling of pigs was identical across treatments. Thereafter (from d 1 to 21), all pigs on the B. pumilus treatment received a top dressing of ~10 ¹⁰ spores on their morning feed.

Pigs were housed individually in fully slatted pens (1.2 m x 0.9 m) with plastic slats (Faroex, Manitoba, Canada) in a total of four rooms with 12 pigs per room. Each treatment group was represented in each room to avoid possible variation due to environment. The pigs had unlimited access to water from one nipple-in-bowl drinker (BALP, Charleville-Mezieres, Cedex, France) per pen. The temperature was controlled by a hot air heating system and an exhaust fan drawing air from under slat level, both controlled by a Stienen PCS 8400 controller (Stienen BV, Nederweert, The Netherlands). The temperature was maintained at 28-30°C in the first week and reduced by 2 °C per week to 24 °C. Pigs were observed closely at least three times daily. Any pig showing signs of ill health was treated as appropriate. All veterinary treatments were recorded including identity of pig, symptom, medication used and dosage. Individual body weight, recorded on d 0 and 22 of the study and feed disappearance, recorded on d 0, 8, 15 and 22 of the study, were used for calculation of average daily feed intake (ADFI), average daily gain (ADG), and feed conversion ratio (FCR; ADFI/ ADG). Fecal consistency scores (0 = normal, 1 = soft, 2 = mild diarrhea, 3 = severe, watery diarrhea) were recorded daily during the experiment.

Blood samples were taken from each of 12 pigs/treatment by venipuncture from the anterior vena cava on d 0, 8 and 15. On d 22 of the experiment, 10 pigs/treatment were euthanized by captive bolt stunning followed by exsanguination and blood samples were taken at this time. All samples were collected in plastic blood collection tubes (Vacuette®, Labstock, Dublin, Ireland) and immediately inverted 10 times. Blood samples for serum biochemistry were collected in serum collection tubes, and allowed to clot at room temperature for 2-3 h prior to centrifugation (2000 x g for 10 min). Serum was collected and stored at -20 °C for subsequent biochemical analysis. Whole blood samples were collected in EDTA tubes and stored at room temperature for hematology analysis within 6 h of sampling.

For hematological analysis, the EDTA blood samples were analyzed on a Beckman Coulter Ac T Diff (Beckman Coulter Inc., Brea, CA, USA), as outlined previously. Serum samples were analysed using an ABX Pentra 400 Clinical Chemistry Analyser (Horiba ABX, Northampton, UK), as outlined previously.

The kidneys, spleen and liver were removed, trimmed of any superficial fat or blood, blotted dry and weighed (n = 10/treatment). Samples of tissue were excised from two anatomical regions of the small intestine: the jejunum (55 cm distal to the pyloric junction) and the ileum (15 cm proximal to the ileo-cecal junction). Samples of liver (centre of quadrate lobe) and kidney (cortex and medulla) were also taken and all samples were immediately placed in No-Tox fixative (Scientific Device Laboratory, Des Plaines, IL, USA) on a shaker for a minimum of 48 h. Intestinal and organ samples were then treated, sliced and mounted and stained with haematoxylin and eosin (Sigma Aldrich, Ireland) for light microscopic examination. Determination of gross morphological parameters of the structure of the jejunum and ileum (villus height and crypt depth) was conducted. For each pig 10 villi and 10 crypts were measured on five fields of view, where villi were attached to the lumen. Measurements were taken from images obtained using a light microscope (Olympus, Southend-on-Sea, UK) fitted with an Optikam PR05 camera (Optika SRL, Ponteranica, Italy) using Optika- Vision Pro software. The goblet cell number was determined in jejunal and ileal sections by periodic acid-Schiff staining. Positively stained periodic acid-Schiff cells were enumerated on 10 villi/sample. The means of all parameters were utilized for statistical analysis. All intestinal and organ tissue samples were also examined for histological evidence of abnormality by an experienced histopathologist.

Fecal samples were obtained by digital rectal stimulation from 12 pigs/treatment on d 0, 8, 15 and 20 and collected in sterile containers. On d 22, digesta samples from the cecum (terminal tip) and ileum (15 cm proximal to the ileo-cecal junction) were collected aseptically into sterile plastic containers. Both digesta and fecal samples were stored at 4 °C until analysis (within 12 h). Samples were homogenized and 500 μΐ of each homogenate was heated to 80 °C for 15 min. Both heated and unheated homogenates were diluted. Appropriate dilutions were plated, as follows; (1) unheated and heated samples were spread-plated on BHI agar containing 200 μg rifampicin/mL, 3.5% NaCl and 50 U/mL nystatin (Sigma) and incubated aerobically for 2 d at 37 °C to enumerate the vegetative cells + spores and spores alone, respectively of the administered B. pumilus strain; (2) unheated samples were pour-plated on ChromoCult® tryptone bile X- glucuronide (CTBX) agar (Merck) incubated at 44 °C for 24 h to enumerate E. coli; and (3) unheated samples were pour-plated on Lactobacillus selective (LBS) agar (Becton Dickinson, Franklin Lakes, NJ, USA) following anaerobic incubation at 37 °C for 5 d to enumerate Lactobacillus. Representative putative E. coli isolates (four colonies per pig from fecal samples at each time point and ileal and cecal digesta) were streaked onto nutrient agar and then onto CTBX agar to obtain pure cultures. They were then streaked onto eosin methylene blue (EMB) agar (Acumedia Manufacturers, Neogen Europe, Ltd., Auchincruive, Scotland, UK) to confirm identity. To determine if these representative E. coli isolates were hemolytic, they were then streaked onto Columbia agar (Sigma) containing 5% sheep blood (TCS Biosciences Ltd., Buckingham, UK).

Samples of cecal and ileal digesta were taken from individual pigs to measure short chain fatty acid concentrations and pH. The pH was measured using a Mettler Toledo pH meter, short chain fatty acid concentrations were determined using gas chromatography. A 5 g sample was centrifuged at 1,810 x g for 10 min., with 1 mL of the resultant supernatant mixed with ImL of internal standard. A 1 μL· aliquot of centrifuged filtered sample was then injected into an Agilent 5890 gas chromatograph with a 15 m x 0.53 mm i.d. Econo-Cap EC- 1000 100% polyethylene glycol-acid modified column (Alltech Associates Applied Science Ltd, Carnforth, Lancashire, UK). Nitrogen was used as the carrier gas at a flow rate of 5.6 mL/min. Oven, detector and injector temperatures were set at 82, 280, and 240 °C, respectively.

Data for growth performance, digesta microbiology, histology, organ weights, short chain fatty acid concentrations and pH of digesta were analyzed as a complete randomized block design using the mixed models procedure of SAS (SAS Institute, Inc., Cary, NC, USA) . Initial (d 0) body weight and the final (d 22) body weight were used as covariates in the model for analysis of growth performance (BW, ADG, ADFI and FCR) and organ weights, respectively. Fecal microbiology, whole blood hematology and serum biochemistry data were analyzed as repeated measures using the MIXED procedure of SAS with sampling day as the repeated variable. D 0 values were used as a covariate in the model for analysis of hematology, serum gamma glutamyltransferase (GGT) and serum total protein (TP). The appropriate covariance structure, as indicated by the model fit statistics, was fitted to the data. The denominator degrees of freedom were computed using the Satterthwaite approximation. Fixed effects were treatment and sex. Block was included as a random effect. Simple main effects were obtained using the 'slice' option in SAS. Least squares means were computed and P values were adjusted for multiple comparisons using the Tukey-Kramer adjustment. Significance was reported for P < 0.05 and tendencies towards significance were reported for 0.05 < P < 0.10. For all response criteria, the individual pig was considered the experimental unit.

On d 12 one pig from the B. pumilus group displaying symptoms of pneumonia was treated with injectable enrofloxacin (Baytril, 5% v/v, Bayer Ireland, Dublin, Ireland; lmL/day) but died on d 13. Also on d 12 one pig from the medicated group displaying symptoms of pneumonia was treated with injectable enrofloxacin (lmL/day for 3 d) and penicillin (Norocillin, 300 mg/mL, Norbrook, Monaghan, Ireland; lmL/day for 3 d). On d 13 another pig from the B. pumilus group with symptoms of pneumonia was treated with injectable enrofloxacin and penicillin as before. The latter two pigs made a complete recovery and their data were included in the analysis of growth performance, but excluded from the analysis of all remaining parameters. The initial and final body weight of pigs did not differ between treatments (P > 0.05), although pigs on the B. pumilus treatment tended to be heavier at the end of the experiment than pigs fed the medicated treatment (P = 0.07). Although no overall treatment effect was observed for ADFI, there was a tendency for pigs on the B. pumilus treatment to have a greater ADFI between d 15 and 22 than pigs fed the medicated treatment (P = 0.07). Similarly, a tendency was observed for pigs on the B. pumilus treatment to have a higher ADG than pigs fed the medicated treatment (P = 0.07). Consequently, pigs on the B. pumilus treatment also had improved FCR when compared to pigs on the medicated treatment (P < 0.05), whereas the FCR of pigs on the non-medicated treatment was similar to that of pigs on the two other treatments.

The mean piglet diarrhea scores for the entire experimental period were 0.34, 0.02 and 0.23 (SE = 0.058; P < 0.01) for the non-medicated, medicated and B. pumilus treatments, respectively. No treatment x time interaction or treatment effect was observed for total white blood cell (WBC) counts (P > 0.05). There was no treatment x time interaction for the lymphocyte percentage (P > 0.05). Lymphocyte percentage was, however, lower for the B. pumilus treatment than for the other two treatments for the overall experimental period (P < 0.001) and at d 15 (P < 0.05). At d 8 the lymphocyte percentage was lower for the B. pumilus treatment than the medicated treatment (P < 0.001), with the non-medicated treatment being similar to that of both other treatments (P > 0.05). There was a treatment x time interaction for monocytes (%) (P = 0.01). Monocyte percentage was higher for the non-medicated treatment than for the other two treatments for the overall period (P < 0.001) and at d 22 (P < 0.001). At d 8 the percentage of monocytes was higher for the non-medicated treatment than the B. pumilus treatment (P < 0.01), with that of the medicated treatment being similar to both other treatments (P > 0.05). There was no treatment x time interaction for granulocyte percentage (P > 0.05). The percentage of granulocytes was higher for the B. pumilus treatment than for all other treatments for the overall period (P < 0.001), at d 8 (P < 0.001) and at d 15 (P < 0.01). At d 22, granulocyte percentage was higher for the B. pumilus treatment than for the non-medicated treatment (P < 0.01), with that of the medicated treatment being similar to that of both other treatments (P > 0.05).

There was a treatment x time interaction for the erythrocyte index, mean corpuscular hemoglobin (MCH) (P < 0.001). The MCH was also lower for the medicated treatment compared to the B. pumilus treatment for the overall experimental period (P < 0.05) and at d 15 (P < 0.05), with that for the non-medicated treatment being similar to that of both other treatments for the overall period (P > 0.05). At d 22, MCH was higher for the B. pumilus and non-medicated treatments than the medicated treatment (P < 0.001), with no difference observed between the former two treatments (P > 0.05). There was a tendency for a treatment x time interaction for the erythrocyte index, mean corpuscular volume (MCV) (P = 0.07), but no treatment effect was observed for the overall period (P > 0.05). At d 15 there was a tendency for MCV to be lower in pigs on the medicated treatment than for pigs on the two other treatments (P = 0.08). There was a treatment x time interaction for the erythrocyte index, mean corpuscular hemoglobin concentration (MCHC) (P < 0.001) and although there was no overall treatment effect (P > 0.05), MCHC tended to be higher at d 8 for the B. pumilus treatment than the non-medicated treatment (P = 0.08). There was no treatment x time interaction or treatment effect for the erythrocyte index, red blood cell distribution width (RDW) (P > 0.05). Nevertheless, at d 22 RDW was 24.9, 27.7 and 23.0% (SE = 1.25%; P < 0.05) for the non-medicated, medicated and B. pumilus treatments, respectively (data not shown). For the remainder of the hematological parameters measured [red blood cell (RBC) counts, hemoglobin, hematocrit, mean platelet volume (MPV) and platelet counts] no treatment x time interaction or treatment effects were observed (P > 0.05; data not shown). There was no treatment x time or overall treatment effect on creatinine concentration (P > 0.05). However, at d 15 the creatinine concentration was 77.5, 83.1 and 76.9 μΜ/L (SE = 1.93 μΜ/L; P < 0.05) for the non-medicated, medicated and B. pumilus treatments, respectively (data not shown). There was a treatment x time interaction for urea concentration (P < 0.001) but no overall treatment effect (P > 0.05). At d 8, urea concentration was 3.1, 1.9 and 3.0 mM (SE = 0.29 mM/L; P < 0.01) for the non-medicated, medicated and B. pumilus treatments, respectively (data not shown). At d 22 urea concentration was 3.9, 4.4 and 3.6 mM (SE = 0.22 mM/L; P < 0.05) for the non-medicated, medicated and B. pumilus treatments, respectively (data not shown).

There was a tendency for a treatment x time interaction for serum alanine aminotransferase (ALT) concentration (P = 0.10) and there was an overall treatment effect (P < 0.01). Serum ALT concentration was increased in the medicated compared with the non-medicated and B. pumilus treatments for the overall period (P < 0.01), as well as at d 15 (P < 0.01) and d 22 (P < 0.01). There was a treatment x time interaction (P < 0.001) for serum alkaline phosphatase (ALP) concentration. Serum ALP was higher for the medicated than the non-medicated and B. pumilus treatments for the overall period (P < 0.001) and at d 8 (P < 0.001), d 15 (P < 0.001) and d 22 (P < 0.001). There was a treatment x time interaction for serum GGT concentration (P < 0.05). Serum GGT concentration was higher for the medicated treatment than for all other treatments for the overall period (P < 0.01) and at d 15 (P < 0.01) and d 22 (P < 0.001). There was no treatment x time interaction or treatment effect for serum aspartate aminotransferase (AST). There was a treatment x time interaction for serum TP (P < 0.05) but no overall treatment effect (P > 0.05). Treatment did not affect renal or splenic weight; however, pigs on the medicated treatment had lighter livers than pigs on either the non-medicated or B. pumilus treatments (P < 0.05). Histopathological examination of the kidneys did not reveal any abnormalities for pigs on any of the treatments. Subtle inflammation was observed in the liver of one pig on the medicated treatment and in one pig on the B. pumilus treatment, but in both cases the inflammation was classified as 'very mild' in character and was most likely subclinical.

There was a tendency for jejunal villus height to be higher for the B. pumilus and the medicated treatments compared to the non-medicated treatment (P = 0.10). Treatment had no effect on crypt depth, villus width or villus height:crypt depth ratio in the jejunum. The number of goblet cells/villus in the jejunum was higher for the B. pumilus and the medicated treatments compared to the non-medicated treatment (P < 0.01); however, the number of goblet cells/μιη of villus was not affected by treatment (P > 0.05). Treatment had no effect on any of the histological parameters investigated in the ileum. Histopathological examination revealed crypt inflammation in the ileum of one pig on the B. pumilus treatment, but this was categorized as very subtle and unlikely to be of clinical significance.

There was a treatment x time interaction for E. coli counts (P < 0.01). E. coli counts were lower for the medicated than both the non-medicated and B. pumilus treatments for the overall period (P < 0.01) as well as at d 8 (P < 0.001) and d 15 (P < 0.01). However, none of the representative fecal E. coli isolates examined were hemolytic. There was no treatment x time interaction for fecal Lactobacillus counts (P > 0.05). Lactobacillus counts were lower for the medicated than both the non-medicated and B. pumilus treatments for the overall period (P < 0.001) and at d 8 (P < 0.01). At d 15 and d 20, Lactobacillus counts were lower for the medicated than the B. pumilus treatment (P < 0.01) while Lactobacillus counts for the non-medicated treatment were similar to those of both other treatments (P > 0.05). The administered B. pumilus strain (vegetative cells plus spores as well as spores alone) was detected in the feces of all pigs on the B. pumilus treatment at all time points except d 0. The administered strain was not recovered from any of the pigs on either the non-medicated or medicated treatments throughout the experiment.

E. coli counts in the ileum were lower for the medicated and B. pumilus treatments compared to the non-medicated treatment (P < 0.05), but no treatment effect was observed for cecal E. coli counts (P > 0.05). However, none of the representative E. coli isolates recovered from either the ileum or cecum were hemolytic. Lactobacillus counts in the ileum were not affected by treatment (P > 0.05), while cecal Lactobacillus counts were lower for the medicated than the non- medicated and B. pumilus treatments (P < 0.05). The administered B. pumilus strain (vegetative cells plus spores as well as spores alone) was detected in both the ileum and cecum of all pigs on the B. pumilus treatment but not from pigs on either the non-medicated or medicated treatments.

The pH of the ileal digesta was higher for pigs on the B. pumilus treatment compared to those on the non-medicated treatment (P < 0.05), with pigs on both treatments having ileal digesta with a similar pH to those on the medicated treatment (P > 0.05). The pH of the cecal digesta was higher for pigs on the medicated treatment compared to those on the non-medicated and B. pumilus treatments (P < 0.01) with the latter two treatments having cecal content with a similar pH (P > 0.05).

The total concentration of short chain fatty acid in the ileum was higher for the medicated treatment than for the non-medicated treatment (P < 0.01) with that for the B. pumilus treatment being similar to that of both other treatments (P > 0.05). This was also the case for acetic acid concentrations in the ileum (P < 0.05). The ileal concentration of propionic acid was similar for the medicated and B. pumilus treatments, but both treatments had higher concentrations than that found in the non-medicated treatment (P < 0.001). There was a tendency for butyric acid concentration in the ileum to be higher for both the medicated and B. pumilus treatments than for the non-medicated treatment (P = 0.08). Total short chain fatty acid concentrations in the cecum tended to be higher for the non-medicated and B. pumilus treatments than for the medicated treatment (P = 0.10). This pattern was significant for propionic acid concentrations in the cecum (P < 0.01). Higher cecal valeric acid concentrations were found for the non-medicated treatment compared to the medicated treatment (P < 0.05), while both treatments had similar concentrations to those of the B. pumilus treatment (P > 0.05). There was a tendency for cecal acetic acid concentrations to be higher in pigs on the B. pumilus treatment than the other two treatments (P = 0.08). Pigs on the medicated treatment tended to have a higher concentration of isovaleric acid in the cecum than those on the B. pumilus treatment (P = 0.10).