Neither Hippurate-negative Brachyspira pilosicoli nor Brachyspira pilosicoli Type Strain Caused Diarrhoea in Early-weaned Pigs by Experimental Infection

A hippurate-negative biovariant of Brachyspira pilosicoli (B. pilosicolihipp-) is occasionally isolated in diarrhoeic pigs in Finland, often concomitantly with hippurate-positive B. pilosicoli or Lawsonia intracellularis. We studied pathogenicity of B. pilosicolihipp- with special attention paid to avoiding co-infection with other enteric pathogens. Pigs were weaned and moved to barrier facilities at the age of 11 days. At 46 days, 8 pigs were inoculated with B. pilosicolihipp- strain Br1622, 8 pigs were inoculated with B. pilosicoli type strain P43/6/78 and 7 pigs were sham-inoculated. No signs of spirochaetal diarrhoea were detected; only one pig, inoculated with P43/6/78, had soft faeces from day 9 to 10 post inoculation. The pigs were necropsied between days 7 and 23 after inoculation. Live pigs were culture-negative for Brachyspira spp., but B. pilosicolihipp- was reisolated from necropsy samples of two pigs. The lesions on large colons were minor and did not significantly differ between the three trial groups. In silver-stained sections, invasive spirochaetes were detected in colonic mucosae of several pigs in all groups. Fluorescent in situ hybridisation for genus Brachyspira, B. pilosicoli and strain Br1622 was negative. However, in situ detection for members of the genus Leptospira was positive for spirochaete-like bacteria in the colonic epithelium of several pigs in both infected groups as well as in the control group. L. intracellularis, Salmonella spp., Yersinia spp. and intestinal parasites were not detected. The failure of B. pilosicoli strains to cause diarrhoea is discussed with respect to infectivity of the challenge strains, absence of certain intestinal pathogens and feed and management factors.


Introduction
Porcine intestinal spirochaetosis (PIS) in weaned pigs is caused by an anaerobic, weakly ß-haemolytic spirochaete, Brachyspira pilosicoli (Taylor et al. 1980, Trott et al. 1996, and it occurs worldwide (Duhamel 1998, Møller et al. 1998, Barcellos et al. 2000, Heinonen et al. 2000, de Arriba et al. 2002, Choi et al. 2002. The pathogenicity of different B. pilosicoli strains has been shown to vary (Thomson et al. 1997). B. pilosicoli usually hydrolyses hippurate, which is a major criterion for the biochemical differentiation of B. pilosicoli from the other porcine Brachyspira species (Fellström & Gunnarsson 1995). However, some porcine B. pilosicoli strains in Finland are hippurate-negative and their pathogenicity is unclear since other potentially enteropathogenic bacteria, such as Lawsonia intracellularis and B. pilosicoli with a positive hippurate reaction, have simultaneously been diagnosed in the herds (Fossi et al. 2004). Co-infections with L. intracellularis and B. pilosicoli are highly prevalent among weaned pigs with diarrhoea in Sweden (Jacobson et al. 2003), and co-infections with L. intracellularis and various species of Brachyspira have also been reported in Denmark and the UK (Møller et al. 1998, Thomson et al. 1998). The concomitant ileitis caused by L. intracellularis or colitis due to the pathogenic species of Brachyspira aggravates the clinical outcome. L. intracellularis and pathogenic species of Brachyspira characteristically cause diarrhoea in pigs after weaning. The time-point of the intestinal colonisation by these organisms during the nursing period is unknown. In this work, we evaluated the pathogenicity of a single strain of hippurate-negative B. pilosicoli (B. pilosicoli hipp-), paying particular attention to the simultaneous presence or absence of other enteric pathogens. The pigs were weaned and moved to barrier facilities at the age of 11 days to enable the effect of early separation on transmission of opportunistic pathogens from sows to piglets to be assessed.

Materials and methods
The experimental infection was approved by the Ethics Committee for Animal Experiments of the National Veterinary and Food Research Institute and carried out under license number 2-2002. Piglets were purchased from a herd of 120 sows with a long history of good health status. Freedom from swine enzootic pneumonia, atrophic rhinitis and Salmonella was verified by regular laboratory investigations. Several studies for Brachyspira spp. performed within the last two years had occasionally yielded weakly ß-haemolytic spirochaetes other than B. pilosicoli. On the farm, nine sows were moved to a farrowing department two weeks before the expected farrowing date. Colostrum samples were collected on the day of farrowing, and faecal samples were taken two days after farrowing. Twelve Landrace piglets and 12 mixed-breed piglets were chosen from four litters which had been born within 24 hours of each other. Equal numbers of male and female piglets for the both races were chosen. No creep feed was served before weaning. At the age of 11 days, the piglets were transported to barrier facilities of the Finnish National Veterinary and Food Research Institute. The piglets were raised together in a pen of 6.20 m 2 until the age of 45 days. The pen had a solid floor which was scraped clean daily. The peat bedding was changed to wood shavings 4-7 days post-inoculation (p.i.). The piglets were fed a commercial sow milk substitute until the age of 24 days. Commercial creep feed was served from the age of 13 to 31 days. Commercial grower feed was based on wheat and barley until the age of 57 days, and thereafter, on barley. The crude protein content of the initial feed was 17.2%, and from the age of 57 days onwards, 17.4%. No growth-promoting feed additives were used, and the feed was granulated without heating. Most piglets had diarrhoea at the age of 15-18 days. Abundant growth of haemolytic Escherichia coli was observed in faecal cultures during this period. One piglet died. All piglets were medicated intramuscularly with trimeto-prime-sulphadiazine at the age of 16 and 17 days. At 45 days of age, the pigs were divided into three trial groups in which littermates, sex and race were evenly distributed. Two groups of eight pigs were placed in two identical isolated rooms, where pen space per pig was 0.60 m 2 . The third group of seven pigs was divided into two subgroups of four and three pigs and placed in two smaller rooms with pen spaces per pig of 0.68 m 2 and 0.51 m 2 . The health of the pigs was monitored daily. Body temperature and weight gain were recorded, and faecal samples were initially taken weekly, and after the challenge, three times a week. Blood samples were collected weekly.
Challenge B. pilosicoli hippstrain Br1622 had originally been isolated in the year 2000 from a pig herd in which weaners and young fatteners had diarrhoea. B. pilosicoli type strain P43/6/78 (ATCC 51139 T ) originated from a diarrhoeic pig in the UK (Taylor et al. 1980). The strains were stored at -70°C in beef broth supplemented with 12% horse serum and 15% glycerol. Br1622 and P43/6/78 were cultured on fastidious anaerobe agar and incubated anaerobically for three days at 39°C. The next passages were propagated anaerobically at 39°C in brain heart infusion broth supplemented with 10% foetal calf serum and 0.5% glucose. Broth cultures were propagated through three passages for a final volume of 300 ml. The final broth was incubated for 36±3 hours to achieve a density of 1.6-2.0 × 10 8 cells ml -1 . The second cascade of broth cultures was started one day after the first one. At the ages of 46 and 47 days, the eight pigs in one room were inoculated intragastrically with 3.2 × 10 9 cells of Br1622 on both days, and the eight pigs in the second room received 4.0 × 10 9 and 3.2 × 10 9 cells of P43/6/78 on two consec-utive days. The seven pigs in the two small rooms were sham-inoculated with a sterile broth on both days. Bacterial viability in the broth containers was controlled after inoculation by microscopy and cultivation.

Necropsy
One pig from both challenge groups and the control group was euthanised and necropsied as blind on days 7, 9, 11, 15, 16, 17, 21 and 23 p.i.. Blood samples were taken and the intestines were dissected immediately after the bolt stunning. The routine gross examination was performed paying special attention to intestinal mucosae. The caecum, the ascending, mid-spiral and descending colon, the rectum and the intestinal lymph nodes were sampled for histopathology. The tissue samples were fixed in 10% buffered formalin, processed routinely, sectioned at 5 µm and stained with haematoxylin and eosin, and Warthin-Starry silver stain. The sections were blind-examined. For transmission electron microscopy (TEM), 1.0 mm 2 samples from the mucosa of the large intestine were fixed in 2.5% glutaraldehyde in phosphate buffer. The samples were embedded in Epon TM , and thin sections were stained with uranyl acetate and lead citrate. Samples were then visualised using a Jeol Jem 100-s electron microscope. Mucosal scrapings for Brachyspira cultivation were taken from the caecum, ascending, midspiral and descending colon and rectum. Content from the large intestine and ileum was collected for further microbiological examination.

Fluorescent in situ hybridisation
The sections from the large colon were studied by fluorescent in situ hybridisation (FISH). The oligonucleotide probes used are presented in Table 2. The general bacterial probe EUB338, the genus-specific probe for Brachyspira and the species-specific probe for B. pilosicoli have been described elsewhere (Boye et al. 1998). The strain-specific probe for Br1622, the specific probe for Treponema (pallidum group) and the genus-specific probes for Leptospira and Borrelia were selected using ARB software (Strunk et al. 2000). The probes were 5'-labelled with fluorescein isothiocyanate (MWG-Biotech AG, Ebersberg, Germany). The probe for strain Br1622 had at least two mismatches to all other Brachyspira species in the ARB database. Processing of the sections and hybridisation were carried out as described previously (Jensen et al. 2000). An Axioplan2 epifluorescence microscope was used for simultaneous detection of red and green fluorescence.

Microbiology, parasitology and haematology
The selective culture and the biochemical differentiation for Brachyspira spp. were performed as previously described (Jenkinson & Wingar 1981, Fellström & Gunnarsson 1995, Olson 1996. The primary cultures from intestinal scrapings were studied also by B. pilosicolispecific polymerase chain reaction (PCR) assays targeting 16S rDNA and 23S rDNA (Fellström et al. 1997, Leser et al. 1997. The PCR studies were repeated with frozen gut samples when invasive spirochaetes were observed by histopathology. Pulsed-field gel electrophoresis (PFGE) with MluI restriction enzyme was used to compare DNA macrorestriction profiles between isolated spirochaetes and the challenge strains, as described earlier (Fossi et al. 2003). The banding patterns of DNA were compared visually. Faecal and intestinal samples were cultured for facultative anaerobic bacteria and anaerobic bacteria, selectively for Campylobacter spp. and by enrichment methods for Salmonella spp. and Yersinia spp., using standard laboratory methods. Cultivation for Campylobacter spp. was done only from colon and caecum scrapings at necropsy. Faecal samples of the sows and 26-day-old piglets were tested for L. intracellularis by nested PCR. Sample preparation was performed according to McOrist et al. (1994), and primer pairs used for DNA amplification were as described by Jones et al. (1993). Antibodies for L. intracellularis were analysed by using an indirect fluorescent antibody test (IFAT) (McOrist et al. 1987, Lawson et al. 1988, Knittel et al. 1998) from the sera of pigs at the age of 42 days and at necropsy. L. intracellularis antibodies were analysed also from the colostrum of the pigs' dams, as well as from the colostrum of the other five sows farrowing in the same department during the same time. Faecal samples from the challenge groups were pooled and investigated for rotavirus by Ani ™ Rotatest (Ani Biotech Ltd., Vantaa, Finland) according to the manufacturer's instructions, and for parasite eggs by the flotation method. Blood samples were studied for total and differential leucocyte count, haematocrit value, haemoglobin concentration and platelet count.

Results
After the challenge, the pigs remained clinically healthy. Only one pig (pig no. 13,     (DWG) was 0.48, 0.39 and 0.49 kg in the Br1622-, P43/6/78-and sham-inoculated groups, respectively. During the second week, the corresponding mean DWG was 0.58, 0.44 and 0.55 kg. The difference in mean DWG between the groups was statistically not significant. The body temperatures of the pigs remained within the normal range throughout the trial period.

Pathology
At necropsy, a moderate hyperaemia was noted on colonic and/or caecal mucosae of six, seven and four pigs infected with Br1622, P43/6/78 and sterile broth, respectively. The colonic mucosae were also slightly oedematous in three pigs infected with Br1622 as well as in three pigs infected with P43/6/78 (Table 1). In light microscopy, minor inflammatory lesions were seen in sections of the large colon of all pigs (Table 1), but the severity of the microscopic lesions did not differ notably between the groups. In silver-stained sections from the large intes-tine, spiral-shaped bacteria were seen in the epithelium of five, six and two pigs in the Br1622, P43/6/78 and control groups, respectively. Spiral-shaped bacteria were more common in sections from the colon than from the caecum. The spiral-shaped bacteria crossed the epithelial cell layer into the lamina propria of the caecum and/or colon in samples from four, three and two pigs in these groups, respectively (Figs 1  and 2). Invasion of spiral-shaped bacteria below the lamina muscularis mucosae of the colon was seen in pig no.13, which was inoculated with P43/6/78. In TEM, trans-sections of endoflagellated bacteria were seen between and below the colonic epithelial cells of one pig inoculated with Br1622 (no. 6) and one control pig (no. 23). The bacteria were 0.28-0.32 µm in diameter and had 10-14 endoflagella.

Fluorescent in situ hybridisation
FISH with probes specific for Br1622, B. pilosicoli and the genus Brachyspira did not give a signal in any section of the large intestines.
However, hybridisation with the probes for genus Leptospira and for domain bacteria revealed multiple spiral-shaped bacteria in and below the epithelial lining of the colon and/or caecum in five, three and three pigs in the groups inoculated with Br1622, P43/6/78 and sterile broth, respectively (Fig. 3). These spiralshaped bacteria seemed uniform and were 6-11 µm in length.
Microbiology, parasitology and haematology Brachyspira spp. were not isolated from the faecal samples of live pigs, but B. pilosicoli hippwas recovered from colon scrapings of two pigs challenged with Br1622 (pig nos. 4 and 5, Table  1) and necropsied on day 15 or 16 p.i.. According to the PFGE macrorestriction profiles, the B. pilosicoli hippisolates were the same as those of the challenge strain Br1622. By B. pilosicolispecific PCR, all primary cultures for Brachyspira, excluding these two Brachyspira isolations, were negative. Haemolytic E. coli was isolated from the faeces of most pigs until one week before challenge. Neither Salmonella spp. nor Yersinia spp. were isolated. Campylobacter coli or Campylobacter spp. were isolated from colon scrapings of four, five and four pigs inoculated with Br1622, P43/6/78 and sterile broth, respectively. Intestinal parasites were not detected. The latex test for rotavirus was positive in the group infected with Br1622 from the age of 46 days onwards, and in all other groups from the age of 48 days onwards. Beyond the age of 55 days, the test results became ambiguous. Pigs' sera were negative for L. intracellularis antibodies. Four of the nine sows in the same farrowing group had colostral antibodies to L. intracellularis. However, these sows were not the dams of the trial pigs. L. intracellularis was not detected by PCR in faeces of the sows or the pigs.
The leucocyte count of one pig in both inocu-lated groups and two pigs in the control group exceeded the upper limit of reference (20 000 cells µl -1 ) once in the first week p.i.. The values of the other haematological parameters remained within the normal range throughout the trial period.

Discussion
Neither B. pilosicoli type strain nor hippuratenegative field strain Br1622 caused diarrhoea in pigs. Reisolation of Br1622 succeeded for only two pigs at necropsy, whereas type strain P43/6/78 was not detected in any of the samples. The inoculated pigs possibly may have shed spirochaetes below the detection limits. The sensitivity of the cultivation for B. pilosicoli is approximated as 5.4 × 10 5 CFU g -1 faeces (Stege et al. 2000) or 1.5 × 10 2 CFU g -1 faeces (Fellström et al. 1997). The sensitivity of our cultivation method lies between these figures (unpublished data). In our experience, PCR from a primary culture rarely yields a positive result in cases where the culture has been negative for Brachyspira. One interpretation is that true colonisation did not occur in most of the pigs. An oral inoculation of 10 9 cells ml -1 of B. pilosicoli given once (Thomson et al. 1997(Thomson et al. , 2001, twice (Neef et al. 1994) (Trott et al. 1996). An attenuation of this strain before its deposition can not be excluded. Strain Br1622 has undergone from four to six passages before its deposition, and a further four passages for the final volume of inocula. We assume that strain Br1622 is no more attenuated than the type strain at the stage of inoculation in this experiment. Some loss of pathogenicity of Br1622 through the passages can not, however, be excluded. No infection of pigs by L. intracellularis was detected. Their early separation from the sows and the apparently low prevalence of infection in the herd of origin could explain the freedom from L. intracellularis. Co-infections, especially with L. intracellularis and B. pilosicoli, have been suggested to aggravate clinical diarrhoea among growers in field conditions (Thomson et al. 1998, Jacobson et al. 2003. The absence of L. intracellularis could also promote the health of the pigs during the trial. Several environmental and management-related risk factors which enhance the outcome of diarrhoea among weaners were missing in this trial. The feed was not heat-treated and its crude protein content was relatively low. Heat treatment of feed during pelleting increases the prevalence of weakly ß-haemolytic intestinal spirochaetes (Stege et al. 2001) and non-specific colitis in herds (Smith & Nelson 1987). Moreover, high protein content in feed has been shown to cause loose stools in pigs (Reese 1999). The feed used in this trial might have supported the clinical health of the pigs to some extent. Peat bedding is popular due to its high capacity for fluid absorption and low pH. In this trial, the peat was changed to wood shavings at 4-7 days p.i.. However, we can speculate that sour peat may have controlled the number of circulating spirochaetes at the early stage of infection. The other environmental conditions in the barrier facilities were undoubtedly more favourable than those on conventional farms, possibly promoting the health of the trial pigs. The ventila-tion in the barrier facilities was filtrated, and diurnal variation in room temperature and draught was minimal. The pen area per pig increased gradually after seven days p.i. due to the necropsy schedule, which might have decreased the number of circulating bacteria. Further experiments for pathogenicity of B. pilosicoli hippstrains should be done for several B. pilosicoli hippstrains of various genotypes in a conventional-like environment using an unattenuated porcine B. pilosicoli strain as a positive control. The bacteria observed in TEM sections of two pigs were deemed not to belong to the species B. pilosicoli because of their inconsistent endoflagella count (Sellwood & Bland 1997). In FISH, the positive signal merely by the genusspecific probe Leptospira was somewhat unexpected. Further investigations are needed to clarify the taxonomic position of these intestinal bacteria.