Institute for Animal Health, Compton,
United Kingdom
Public health problems associated with Salmonella infection in poultry remain of major significance. The economic constraints inherent in the poultry slaughter process indicate that the control of Salmonella infection on the poultry farm would be the most practical approach. It is possible to rear poultry that are totally free of Salmonella organisms, but this requires high cost housing, tight control on feed quality, hygiene and management. However, the economics of production still depends on the importation of poultry or meat from countries, where such levels of control may not be practiced. It is thus increasingly recognised that biological measures will form an integral part of control programmes. For a number of reasons vaccines are likely to become of increasing importance.
In poultry, two types of Salmonella infection need to be considered: (1) an infection of the reticuloendothelial system with little initial intestinal involvement, (2) an extensive intestinal infection by an organism, which is invasive to different degrees. S. gallinarum infection in immunologically mature birds is likely to be similar to S. typhimurium infection in mice.
Full protection can be obtained at the level of the reticuloendothelial system. The oral LD50 may be ca 104 cfu, although, in the field, the inoculum may vary considerably. Oral inoculation with large numbers of challenge organisms may induce bacterial penetration to the reticuloendothelial system, while a smaller number may be eliminated at the level of the gut (Barrow and Stocker, unpublished results). Infection of poultry with strains that are obviously invasive, such as S. typhimurium or S. enteritidis most likely belongs to the second type of Salmonella infection mentioned above.
In the study of uptake of S. enteritidis and S. thompson by the caecal mucosa, macrophages containing Salmonella organisms can be visualised in the process of crossing the epithelial layer and lamina propria across gaps in the basement membrane. This suggests that macrophage may carry bacteria to the circulation.
Vaccination Against Host-Specific Serotypes
Vaccination of poultry, man or other animals with live attenuated strains derived from the host-specific Salmonella serotypes induces a strong protective immunity against reinfection in the host. Over the years live attenuated strains derived from the host-specific serotypes that infect cattle, pigs, fowl, and man have been developed into effective vaccines that can be administered parenterally. Attenuated strains of S. gallinarum strain 9 have been assessed extensively as live vaccines for chickens since the 1950's. One of these, the 9R fowl typhoid vaccine (Smith, 1956), produced by passage in a medium of low nutritional quality, confers strong protection against the systemic disease in adult chickens, although the vaccine strain retains some virulence, may persist for many months and be transmitted through the egg (Gordon and Luke, 1959; Silva et al., 1981). The 9R vaccine is rough and therefore does not stimulate the production of antibodies directed against lipopolysaccharide. This is valuable since its use does not interfere with serological tests based on the presence of serotype-specific serum IgG used in the detection of Salmonella infection. A number of other smooth derivatives of these strains have been produced, which nevertheless conferred less protection than the 9R vaccine (Barrow, 1990; Hoiseth and Stocker, 1981; Stocker (1993); Griffin and Barrow, 1993). Such strains, which are more attenuated though less protective might be used in more susceptible breeds in conjunction with a second vaccination with the 9R vaccine. A nuoG mutation showing no NADH dehydrogenase I enzymatic activities, was shown to attenuate S. gallinarum and was highly protective against fowl typhoid in experimental infections (Zhang-Barber et al., 1998). It is generally regarded that killed vaccines are not protective against fowl typhoid, however, preliminary investigations have shown that protection against infection with S. gallinarum can be achieved using outer-membrane proteins (Bouzoubaa et al., 1987).
Vaccination Against Host Non-Specific Serotypes
The development of vaccines against non-host specific serotypes for use in poultry has been almost exclusively empirical. In assessing the following studies it is important to bear in mind the infection model used.
Various types of non-living vaccine have been used experimentally and in the field, and although they generate an immune response, protection is variable. Earlier work indicates that protection is generally no better than moderate. McCapes et al. (1967) demonstrated reduced mortality (from 85.6% to 39.7% and from 91.5% to 44.6%) in day old turkeys challenged via the yolk sac with S. typhimurium and S. schwarzengrund whose parents had been exposed to S. typhimurium infection early in life and vaccinated repeatedly subcutaneously with S. typhimurium bacterin. A less pronounced reduction in mortality (from 94.4% in controls to 75.1% in vaccinated birds) was obtained in experiments using crude endotoxin extract from S. typhimurium given intraperitoneally (Truscott and Friars, 1972). When young chickens were vaccinated directly, using boiled sonicates of several serotypes incorporated in the feed, then challenged a few weeks later, clearance from the faeces of the challenge serotypes was generally more rapid than for unvaccinated controls (Truscott, 1981) although the results were variable.
In contrast, Bisping et al., (1971) found that orally administered heated whole cell bacterins had little effect on excretion of Salmonella in the faeces. Similarly, vaccination of the mother birds with a bacterin did not significantly reduce excretion of Salmonella hadar from the challenged progeny (Thain et al., 1984). Parenteral vaccination with formal-killed bacteria, but also challenging parenterally either with S. virchow (Ghosh, 1989) or S. enteritidis (Timms et al., 1990), reduced mortality and intestinal colonisation. Ghosh (1989) also reported reduced mortality from 85% to 0% against log107.7 S. virchow organisms in young chicks, when inoculated intraperitoneally. Timms et al., (1990) showed reduced mortality from 100% to 50% against between log105 and log108 challenge organisms in similar experiments. Since the increase in S. enteritidis - associated human food poisoning in the late 1980s, a number of studies have been carried out vaccinating hens with this serotype killed by different methods. Gast et al.,(1993) demonstrated reductions in the rate of faecal excretion when birds were challenged two weeks after the second of two sub-cutaneous vaccinations, but not if they were challenged six weeks after. Fewer numbers of the challenge strain were isolated from the spleen, ovaries and oviducts compared to controls. Reduced excretion of Salmonella in faeces was also obtained when hens were vaccinated with acetone-killed S. enteritidis bacterins mixed with Freund's incomplete adjuvant (Barbour et al., 1993). This vaccination induced high levels of circulating specific IgG to unspecified protein antigens. Nakamura et al., also showed reduced excretion in faeces and of bacterial numbers in the tissues of birds vaccinated twice and challenged at laying age Nakamura et al., (1994). It is thought that a commercially available killed S. Enteritidis has played a significant role in the reduction of S. enteritidis in layers, combined with other improvements in biosecurity and hygiene (R. Davis, pers. comm).
Some live vaccines are now available and licensed for use in several countries. Live attenuated vaccine strains developed for immunisation against the host-specific serotypes have been of less value in reducing excretion of those that are non-host specific. Rough mutants defective in lipopolysaccharide biosynthesis have been tested for use in poultry. S. typhimurium strain F98, which is virulent for chickens, was rendered avirulent by selecting for roughness as indicated by bacteriophage resistance. When challenged, chickens that had been vaccinated previously, either orally or intramuscularly with the mutant, showed reduced faecal excretion of the parent strain (Barrow et al., 1990a). A mutant of S. typhimurium defective in the UDP-galactose epimerase gene, galE, required for LPS synthesis (Germanier and Furer, 1975), has been shown to be an effective vaccine in mice and cattle (Wray et al., 1977). Such mutants are rough unless exogenous galactose is supplied. When one-day-old chicks were vaccinated orally with the galE mutant and were challenged, albeit only two weeks later, with an S. typhimurium isolated from turkeys, the number of Salmonellas excreted in the faeces, and the number of chicks reaching carrier status, was significantly reduced (Pritchard et al., 1978). Protection was more pronounced when chickens were inoculated intramuscularly rather than orally (Subhabphant et al., 1983). However it was found that galE mutants of S. choleraesuis and S. typhi can retain virulence for mice and humans respectively (Nnalue and Stocker, 1986; Hone et al., 1988), so the galE mutation is generally thought to have less value than previously.
An aroA mutant of S. typhimurium strain F98 was attenuated for chickens and gave protection to four-day-old chicks against challenge with 108 cfu of the parent strain, when vaccinated by the intramuscular or oral routes (Barrow et al., 1990b). Cooper obtained similar results (Cooper et al., 1992). In both cases the vaccine was administered orally within one day of hatching. The degree of protection was greatly improved when challenge was made by contact infection soon after oral vaccination of the newly hatched chicks (Cooper et al., 1993). These authors found no evidence of cross-protection between S. enteritidis and S. typhimurium in terms of excretion in the faeces. Others have found that a rough mutant of an S. enteritidis phage type 4 aroA strain protected laying hens against challenge with the wild type strain, but better protection against S. enteritidis was obtained by parenteral vaccination with the S. gallinarum 9R vaccine strain (Feberwee et al., 2001). However, it was recognised that the 9R vaccine strain persisted in the tissues (Cooper et al., 1993). Other auxotrophic mutations that have been found to attenuate virulence include those in the pur genes, which confer an adenine requirement. Such mutants were found to be poor vaccines when tested in mice, possibly because they are over attenuated, with the result that they have not been tested further. However, a purE mutant strain is being used extensively as a chicken vaccine in Germany (Meyer et al., 1992) and further undefined live vaccines are now used in a number of European countries (Barbezange et al., 2000). The cya crp genes encode the adenylate cyclase and the cAMP receptor protein respectively, and their products are required for growth on carbon sources other than glucose. A cya crp mutant of S. typhimurium was shown to be avirulent and immunogenic in chickens (Curtiss and Kelly, 1987) with claims of cross protection between serotypes, but the vaccine is almost produces gastro-enteritis in pigs (Barrow et al., 2001) and some of this work has now been discredited. These authors have also suggested that vaccines that are too virulent may be counter-productive because of their immunosuppression (Hassan and Curtiss, 1997). It should soon be possible to delete some of the genes responsible for fluid secretion in the intestine from the Salmonella chromosome (Galyov et al., 1997). Other attenuations have been used with claims for strong cross protection but the degree is relatively poor (Dueger et al., 2001)).
If live vaccines are administered to newly hatched chicks they multiply extensively because of the absence of the complex normal microbial flora found in adult birds. This in turn prevents colonization by other Salmonella strains inoculated within a few hours afterwards (Barrow et al., 1987). This is a purely microbiological effect, and may explain the degree of protection found by Alderton et al.,
(1991) where administration of an aroA S. typhimurium vaccine produced significant protection when the birds were challenged 10 days later. There is some evidence that this phenomenon does occur in very young birds administered with live, attenuated vaccines (P. Coloe, personal communication).
Several works suggests that a major determinant of the effects is the availability of carbon sources and electron acceptors in the local environment of the bacterial cells (Zhang-Barber et al., 1997). Protection may thus be obtained against homologous and heterologous serotypes and phage types. The practical consequences are that live vaccines can be administered to very young chickens by the oral route thereby obtaining protection within a matter of hours. The choice of strains from which to develop future live vaccines might depend on testing their inhibitory activity since a number of those currently available have a very poor inhibitory spectrum (Methner et al., 1997) and there are possibilities for combining the use of live vaccines with gut flora preparations (Methner et al., 2001)
References are available on request
From Proceedings of 11th European Poultry Conference, Bremen, Germany.
