Broiler lines divergently selected for digestive efficiency also differ in their susceptibility to colibacillosis

Increasing feed efficiency of broiler chickens by selective breeding could lead to decreased feed cost and reduced environmental impact of poultry production. At INRA, two broiler chicken lines (D+/D-) were divergently selected for their digestive efficiency. Strong differences were shown between both lines for the anatomy and histology of the digestive tract, and for the intestinal microbiota composition. In the present study, we investigated whether this selection also had an effect on susceptibility to colibacillosis, which is one of the main causes of economic losses in poultry production. The broiler lines D+/D- were challenged with an avian pathogenic Escherichia coli strain. A first experiment was conducted to assess the 50% lethal dose by subcutaneous infection of hatchlings, whereas a second experiment reproduced colibacillosis by infecting air sacs of 23-day-old chicks. The 50% lethal dose was very low for both lines. However, the line with the higher digestive efficiency (D+) was the less susceptible to colibacillosis. This result is interesting for selection purposes and opens the way to integrative genetic studies of the interactions between digestion efficiency and resistance to colibacillosis.

Effects of microflora status, dietary bile salts and guar gum on lipid digestibility, intestinal bile salts, and histomorphology in broiler chickens

The aim of the present study was to reexamine the effects of intestinal viscosity, microflora, and bile salts, and their interactions in order to clarify the mechanisms that explain the effect of intestinal viscosity on lipid digestibility, especially those that could involve microflora. Effects of intestinal viscosity, microflora status, and dietary bile salts on lipid digestibility, intestinal bile salts, and intestinal histomorphology were tested in a 2 x 2 x 2 factorial experiment. The effect of microflora status was examined by comparing conventional chickens to chickens with limited microflora obtained by rearing birds in sterilized conditions. Viscosity and dietary bile salts were tested with guar gum (0 or 0.5% in diets) and sodium taurocholate (0 or 0.3% in diets), respectively. Guar gum was autoclaved and added to the gamma-irradiated diets by mixing inside the sterile isolators. The intestinal concentration of lactic acid and cecal concentration of short-chain fatty acids were both very low in birds with limited microflora compared to conventional birds (P = 0.0001). Chickens with limited microflora had higher gain:feed ratios (P = 0.002), higher fecal lipid digestibility (P = 0.0001), more intestinal conjugated bile salts (P = 0.0001), less intestinal unconjugated bile salts (P = 0.0001), and their gastrointestinal compartments were smaller (P < 0.05) than those of conventional chickens. Addition of bile salts in diets increased the lipid digestibility (P = 0.0001) with a more pronounced effect in conventional birds (P = 0.0001) and in birds fed on guar gum diets (P = 0.002). Feeding the guar gum diets increased the intestinal supernatant viscosity (P = 0.0001) for both microflora status. Guar gum addition increased (P < 0.05) the lactic acid concentration in the small intestine of conventional chickens. Guar gum decreased the fecal lipid digestibility (P = 0.0001) and the intestinal conjugated bile salts (P = 0.0001) for both microflora status. However, the magnitude of lipid digestibility decrease due to guar gum was lower (P = 0.05) in birds with limited microflora than in conventional birds, and the negative effect of guar gum on intestinal conjugated bile salts was more pronounced (P = 0.02) in birds with limited microflora. Bile salt addition reduced the negative effect of guar gum on lipid digestibility (P = 0.02) for both microflora status. The mean lipid digestibilities were negatively correlated (P = 0.0001) with the ratio [Ln(viscosity)/total bile salt] measured in intestinal digesta. Except for gizzard, size of digestive compartments and villus heights increased (P < 0.05) with feed efficiency and digestibility decreased. In conclusion, the results provided evidences that the negative effect of guar gum on lipid digestibility was mainly mediated by its direct effect on intestinal bile salt concentration and efficiency. The small interaction between guar gum and microflora status effects on lipid digestibility had to be accounted for by the low basal level of intestinal bile salts in conventional chickens.

Relationship between the Tsh autotransporter and pathogenicity of avian Escherichia coli and localization and analysis of the Tsh genetic region

The temperature-sensitive hemagglutinin Tsh is a member of the autotransporter group of proteins and was first identified in avian-pathogenic Escherichia coli (APEC) strain chi7122. The prevalence of tsh was investigated in 300 E. coli isolates of avian origin and characterized for virulence in a 1-day-old chick lethality test. Results indicate that among the tsh-positive APEC isolates, 90.6% belonged to the highest virulence class. Experimental inoculation of chickens with chi7122 and an isogenic tsh mutant demonstrated that Tsh may contribute to the development of lesions within the air sacs of birds but is not required for subsequent generalized infection manifesting as perihepatitis, pericarditis, and septicemia. Conjugation and hybridization experiments revealed that the tsh gene is located on a ColV-type plasmid in many of the APEC strains studied, including strain chi7122, near the colicin V genes in most of these strains. DNA sequences flanking the tsh gene of strain chi7122 include complete and partial insertion sequences and phage-related DNA sequences, some of which were also found on virulence plasmids and pathogenicity islands present in various E. coli pathotypes and other pathogenic members of the Enterobacteriaceae. These results demonstrate that the tsh gene is frequently located on the ColV virulence plasmid in APEC and suggest a possible role of Tsh in the pathogenicity of E. coli for chickens in the early stages of infection.

Increased tracheal colonization in chickens without impairing pathogenic properties of avian pathogenic Escherichia coli MT78 with a fimH deletion

Several studies suggest that the expression of F1 fimbriae could be involved in the virulence of Escherichia coli for chickens. F1 fimbriae display multivalent properties such as adhesion to epithelia or interaction with the immune system that imply specific interactions between the adhesin FimH and different cell receptors. We constructed a delta fimH mutant of the avian pathogenic E. coli MT78 and evaluated its in vivo colonization and pathogenicity, as compared to that of the parent strain. The generated mutant PA68 was unable to adhere in vitro to chicken epithelial pharyngeal or tracheal cells; mutant bacteria were mostly afimbriated although a minority of them displayed altered piliation phenotypes. Two inoculation routes were used to compare the ability of MT78 and PA68 to colonize the respiratory tract and to induce colibacillosis in chickens. In the first model, 2-wk-old axenic chickens were inoculated intratracheally with one or both E. coli strains, after primary infection with infectious bronchitis virus. In the second model, 3-wk-old specific-pathogen-free chickens were inoculated via the caudal thoracic air sac. After intratracheal inoculation, the delta fimH mutant was found to be a better colonizer than MT78 in the trachea of inoculated chickens. Furthermore, when both strains were inoculated simultaneously, the delta fimH mutant constituted 98% of the bacterial population in the trachea at day 7 postinoculation. Irrespective to the inoculation route, MT78 and PA68 showed similar abilities to induce macroscopic lesions in chickens, to provoke bacteremia, and to colonize the internal organs. However, 4 days after intra-air sac inoculation, bacterial counts of the mutant were lower in the spleen and liver than those of MT78. Our results show that FimH is not required for colonization of the trachea of axenic chickens by E. coli and that it is not a major determinant of bacterial pathogenicity. On the contrary, the lack of expression of FimH seems to favor the in vivo colonization of the trachea of chickens by E. coli.

Colonization ability and pathogenic properties of a fim- mutant of an avian strain of Escherichia coli

Several studies suggest that the expression of type 1 fimbriae is involved in the virulence of Escherichia coli in chickens, by promoting adhesion of bacteria to the respiratory tract, which is most probably the first step to occur in the infection, and by interacting with the immune response. In order to determine to what extent type 1 fimbriae were involved in the pathogenic process, the fim cluster of an avian pathogenic strain of E. coli, MT78 (O2:K1:H+), was modified in vitro and reintroduced in the parent strain via allele exchange using suicide vector pCVD442. The mutant strain thus generated (DM34) had its entire fim cluster removed. Its pathogenic properties were compared to those of the parent strain in an experimental reproduction of avain colibacillosis in 15-day-old chickens, after primary infection with infectious bronchitis virus followed by intratracheal inoculation of the challenge strain. In specific-pathogen-free (SPF) animals, mutant DM34 was less pathogenic than the parent strain and colonized the lungs of infected animals to a lower level. In germ-free chickens, although DM34 was less pathogenic than MT78 according to the differences in weight gains, it colonized the trachea, lungs and internal organs to the same extent as MT78. Our results suggest that, whereas type 1 fimbriae are not strictly required in colonization of the respiratory tract of germ-free chickens, they might be important in establishing a bacterial population in the lungs of SPF animals. The difference regularly observed in weight gains between mutant- and wild-type-inoculated chickens reflects a decreased pathogenicity of the fim- mutant. However, the isolation of E. coli in target organs and the observation of colibacillosis symptoms and lesions in mutant-inoculated chickens suggest that type 1 fimbriae by themselves play a limited role in pathogenicity.

Localization of the in vivo expression of P and F1 fimbriae in chickens experimentally inoculated with pathogenic Escherichia coli

Escherichia coli causing septicemia in poultry often possess F1 (type 1) and/or P fimbriae which may be involved in bacterial colonization and infection. To investigate the expression of these fimbriae in vivo, two pathogenic E. coli strains with different fimbrial profiles, TK3 (fim+/pap+) and MT78 (fim+/pap-), were administered to 2-week-old chickens by either the intratracheal or caudal thoracic air sac inoculation route. Antibodies specific for native F1 fimbriae were detected by ELISA and immunodot in the serum of chickens inoculated with either strain MT78 or strain TK3, irrespective of the route of inoculation. Antibodies specific for P fimbriae of serotype F11 were detected by ELISA and immunoblotting in the serum of chickens inoculated by either route with strain TK3. F1, but not P fimbriae, were expressed by bacteria colonizing the trachea of chickens inoculated by the air sac route with strain MT78 or TK3, as demonstrated by examination of frozen tissue sections using immunofluorescence. F1 fimbriae were also expressed by bacteria colonizing the air sacs and lungs, but not by bacteria in the blood or other internal organs, of chickens inoculated with either strain. P fimbriae were expressed by bacteria colonizing the air sacs, lungs, kidney, blood, and pericardial fluid, but not by bacteria colonizing the trachea, of chickens inoculated with strain TK3. Fimbriae-like structures were observed by electron microscopy on bacteria adhering to the epithelial cells of the air sacs of chickens inoculated with strain TK3. These results demonstrate that both strains MT78 and TK3 undergo in vivo phase variation with respect to their fimbrial profiles and site of bacterial colonization in different organs of infected chickens and suggest that F1 fimbriae are important for initial bacterial colonization of the upper respiratory tract whereas P fimbriae are important for later stages of the infection.

Bacterial colonization and in vivo expression of F1 (type 1) fimbrial antigens in chickens experimentally infected with pathogenic Escherichia coli

Escherichia coli strains that cause septicemia of poultry often possess F1 (type 1) fimbriae (encoded by pil [fim] homologous gene clusters) and/or P fimbriae (encoded by pap homologous gene clusters). These fimbriae are thought to be involved in infection and colonization. To study the dynamics of infection due to E. coli with different virulence determinant profiles and to examine the expression of these fimbriae in vivo, three pathogenic E. coli isolates--O1 (pil+/pap+), O2 (pil+/pap), and O78 (pil+/pap+)--were administered intratracheally to 1.5-week-old chickens. Chickens were euthanatized from 3 to 144 hr after infection. The three isolates caused lesions in 30 to 55% of birds. Colonization rates of the trachea, lungs, internal organs, and pericardial fluid were similar for all three isolates, whereas significant differences among isolates were observed in colonization of the air sacs and blood. Bacteria appeared rapidly in the blood, liver, and spleen, whereas presence in the pericardial fluid generally occurred only after 24 hr postinoculation. The dynamics of colonization of the air sacs varied among isolates. Immunofluorescence of frozen tissue sections demonstrated F1 fimbriae (pil expressed) but not P fimbriae on all three isolates colonizing the trachea and on the O1 and O78 isolates colonizing the air sacs. Results suggest that F1 fimbriae are involved in the early stages of development of colisepticemia by promoting association of pathogenic E. coli with the trachea and air sacs of chickens.

Serological conservation and location of the adhesin of avian Escherichia coli type 1 fimbriae

By inoculation of mice with purified type 1-like fimbriae isolated from an avian Escherichia coli strain, a monoclonal antibody (mAb G5) was obtained. mAb G5 reacted in an enzyme-linked immunosorbent assay (ELISA) with type 1-like and type 1A fimbriae differing in the molecular masses of their major fimbrial subunit and isolated from several avian E. coli strains. The specificity of mAb G5 for type 1 fimbriae was assessed in a whole bacteria ELISA with 16 reference E. coli strains expressing different types of fimbriae. Immunoblotting experiments showed that mAb G5 recognized the 29 kDa minor component of reference type 1A fimbriae which has been identified as the adhesin. mAb G5 also recognized the 29 kDa component of type 1-like and type 1A fimbriae expressed by avian E. coli strains, suggesting that the adhesin is antigenically conserved among these fimbriae. Immunoelectron microscopic studies gave evidence that the adhesin could be located mainly at the tip or both at the tip and along the fimbriae, depending on the strain.

[Effect of intravenous human immunoglobulins on bacterial clearance and mortality in experimental Escherichia coli K 1 septicemia in chicken]

The antibacterial effect of intravenous immune globulins (IVIG) was tested in 4 to 6-week-old chickens experimentally infected with a K1 E. coli strain from a neonate with meningitis. Following inoculation of 10(7) E. coli K1 per animal, the chickens were randomized to receive a placebo, 0.25 g/kg, 0,50 g/kg or 1 g/kg of intravenous immune globulins. Injections were performed 4 hours after the bacterial inoculation. Numbers of animals in each group were 30, 19, 30 and 10 for 6-week-old chickens and 9, 10, 9 and 0 for 4-week-old chickens. Bacterial diffusion was evaluated using quantitative cultures of blood from breast capillaries. In the 6-week-old chickens given 0.25 g/kg or 0.50 g/kg of intravenous immune globulins, bacteremia 48 hours after inoculation and mortality were significantly lower than in the placebo group. Administration of intravenous immune globuline in a dose of 1.0 g/kg, however, was followed by increased blood bacteria counts 48 and 72 hours after inoculation. In the 4-week-old chickens, no reduction in bacterial counts or improvement in survival was seen after administration of intravenous immune globulins. These data should provide a basis for a more rational use of intravenous immune globulins in neonates and children.

Surface antigens from Escherichia coli O2 and O78 strains of avian origin

Fimbriae from O2 and O78 virulent strains of avian Escherichia coli were compared with type 1A fimbriae with regard to the apparent molecular weights of their subunits and their antigenic relationships. Under static broth culture conditions, most O78 strains expressed fimbriae closely related to those of type 1A. Under the same culture conditions, another type of fimbriae, sharing some common properties with type 1A fimbriae, was observed only on O2 strains; however, these fimbriae differed from type 1A fimbriae in the apparent molecular weights of their subunits and in the expression of specific epitopes. They were called type 1-like fimbriae. Homologies in lipopolysaccharide and outer membrane protein profiles were also demonstrated among the strains expressing type 1-like fimbriae, which suggests the existence of a clonal relationship among O2:K1 avian E. coli strains. The O78 strains studied did not appear to be clonally related.

Comparative infectivity for axenic and specific-pathogen-free chickens of O2 Escherichia coli strains with or without virulence factors

Adhesion to epithelial respiratory cells, iron acquisition, and production of K1 polysaccharide capsules have been proposed as potential virulence factors of avian Escherichia coli. These factors were studied by inoculating groups of axenic or specific-pathogen-free (SPF) chickens intratracheally with O2 E. coli strains after previous challenge with a wild strain of infectious bronchitis virus (IBV). In all experiments, the association between IBV and an E. coli strain endowed with the three virulence factors previously mentioned resulted in the most severe pathological effects, as measured by mortality, weight gains, lesions, and reisolation of E. coli from internal organs. An E. coli strain devoid of virulence factors was able only to induce mild pathological effects restricted to the respiratory tract when combined with IBV. Both E. coli strains were more invasive in axenic chickens than in SPF chickens. These results confirm the probable involvement of the three factors studied in the pathogenic properties of avian E. coli. This model can be used to assess the role of virulence factors, by comparing pairs of positive and negative isogenic strains.

[Venous and capillary diffusion of Enterococcus faecalis during experimental bacteremia in the chicken]

Capillary and venous diffusion of bacteria were studied in chickens after intravenous inoculation with a strain of Enterococcus faecalis isolated from a case of septicaemia in a human neonate. One group of 13 chickens was inoculated with 10(8) bacteria/animal and another group of 20 chickens with 10(7) bacteria/animal. The level of bacteraemia in the first group of chickens was similar in capillary and venous blood. With 10(8) E. faecalis, the clearance of bacteria was very weak. The level of bacteraemia in the second group was significantly higher (after 180 min) in capillary than in venous blood, which suggests slower bacterial clearance in capillary than in venous blood. However, clearance was more efficient with the inoculum of 10(7) E. faecalis than with the first type of inoculum. These experimental data confirm the greater sensitivity of capillary rather than venous sampling, which we have already observed in chickens, under similar experimental conditions, with Escherichia coli K12. Persistence of bacteria in the capillaries could account for the higher level of sensitivity of capillary microblood culture than venous blood culture, which is in agreement with the capillary trapping hypothesis of Knudson and Alden.

[Quantitative study of capillary and venous bacterial diffusion in experimental bacteremia in chickens]

Capillary and venous diffusion of bacteria were studied in chickens after intravenous inoculation with Escherichia coli K12. A first group of 35 chickens was inoculated with 10(7) and a second group of 49 chickens with 10(8) E. coli. Bacterial enumeration showed two types of bacterial blood distributions. Early after inoculation, the level of bacteraemia was higher in venous than in capillary blood; later, the level of bacteraemia was higher in capillary than in venous blood. The difference was significant for the two kinds of inoculum. These results showed a slower clearance of E. coli in capillary than in venous blood. These experimental data confirm the greater sensibility of capillary than of venous sampling, already observed in neonates.

Experimental study of some factors limiting 'competitive exclusion' of salmonella in chickens

Some factors affecting the efficiency of competitive exclusion of Salmonella typhimurium var copenhagen in chicks were studied experimentally under gnotobiotic conditions. Axenic chickens were given dilute suspensions of adult faeces and exposed to the salmonellae. Prevention of colonisation of the gut by salmonellae ('competitive exclusion') was variable and depended possibly in part on the source of the adult faeces used to protect the chicks. Exclusion was also dose-dependent, a large inoculum of the salmonellae (10(7) viable organisms per animal) leading to colonisation in treated chicks. An inapparent carrier state was sometimes produced by lower doses of the salmonellae (10(2) viable organisms per animal), but bursts of excretion still appeared after inoculation of the salmonellae. In inapparent healthy carriers, the inoculation of a dose of Eimeria tenella oocysts that was known to produce subclinical caecal coccidiosis led to the shedding of large numbers of salmonellae for over two weeks, and the use of 'competitive exclusion' in poultry as a preventive measure for salmonella infections might thus be limited by the frequent occurrence of subclinical coccidiosis in the field.