Pathogenesis of Egg Infections by Salmonella and The Implementation of Preventive Measures

Most Salmonella strains belong to non-host specific or broad-host range serotypes, and thus can colonize the gut of many animal species, including humans. In contrast to host-specific serotypes that cause septicemia and severe disease (typhoidal serotypes), the broad-host range serotypes are asymptomatically colonising the host in most cases, but can cause diarrhoea when high numbers of bacteria are orally take up at once, as is the case in human food poisoning (non-typhoidal serotypes). Salmonellais one of the most important zoonotic bacterial pathogens, mainly caused by consumption of contaminated food products, and is of global importance. The global burden of gastroenteritis due to Salmonella has been studied by various authors. Majowicz et al. (2006) estimated the global number of Salmonella cases to be around 1600 million cases per year. An estimation of the global and regional disease burden by the World Health Organization shows non-typhoidal Salmonella to cause highest numbers of disability-adjusted life years (DALYs) of all foodborne pathogens (Kirk et al., 2015). DALYs are defined as the sum of the years of life lost due to premature mortality in the population and the years lost due to disability for people living with the health condition or its consequences. While many serotypes can be transmitted to humans due to contaminated meat products (mainly chicken and pork, but also cattle, fish and other sources), eggs are the main food vehicle for human infections. As an example, in the European Union in 2017, more than 50% of the human cases were egg-derived (37% egg-derived, 17% derived from bakery products). The other 50% was derived from various sources such as mixed food products (13%), meat products (8%), poultry meat (2.2%), porcine meat (4.5%), and to a lesser extent cheese, dairy, vegetables and fish (EU data, 2017). Interestingly, about 50% of all cases are caused by strains of the Salmonella serotype Enteritidis, and 25% by serotype Typhimurium strains (of which 8% are monophasic variant strains). The other 25% is caused by a variety of serotypes, including Salmonella Infantis (2.5%) (EU data, 2017). It is striking that, for many of the serotypes that are important for human food poisoning, there is an association with specific food sources. Salmonella Enteritidis is of particular importance because it can spread to the reproductive tract and contaminate eggs. A worldwide egg-associated salmonellosis pandemic started in the ‘70s and is currently fading away in many countries, thanks to huge efforts of policy makers and the poultry industry. This pandemic has been specifically caused by the serotype Enteritidis. Due to its preferential association with hen eggs, combined with the way humans tend to store (room temperature), handle and eat (uncooked) eggs, Salmonella Enteritidis had and still has a major impact on human health. Strains from the serotype Typhimurium are associated with many sources, including porcine, cattle, turkey and chicken meat, but also eggs. Strains from the monophasic variant of serotype Typhimurium are specifically associated with food poisoning cases and outbreaks after porcine meat consumption. In addition to the human infections caused by the 2 predominant serotypes, Enteritidis and Typhimurium, also many other serotypes can cause human gastroenteritis. These are mainly derived from meat sources, and the nature of the serotypes depends on the geographical location, and changes during time. Serotypes such as Hadar, Infantis, Paratyphi B, Heidelberg, Minnesota and many others can be derived from poultry meat. Strains from Salmonella Newport associate with turkey meat (EU data, 2017). A specific trend is the spread of clonal lineages of certain serotypes that are often multidrug resistant, in poultry. A metaanalysis of Ferrari et al. (2019) on the global epidemiology of Salmonella shows that there are some serotypes that are colonizing poultry worldwide (Enteritidis, Typhimurium, Infantis, Hadar, Kentucky), while there are serotypes that are specifically associated with certain regions (eg. Heidelberg in the Americas, Mbandaka in Europe). Oceania is an exception as Enteritidis, Hadar and Kentucky are not an issue (recently Enteritidis entered however), while serotypes such as Sofia and Kiambu are specific for this region (Ferrari et al., 2019).

Chickens usually are infected by oral uptake of bacteria from the environment. Salmonella bacteria are able to survive gastric acidity and can pass the stomach to reach the intestinal tract of the animal. The caeca are the predominant colonisation sites. The bacteria can adhere to and invade caecal epithelial cells, by rearranging the actin cytoskeleton of the epithelial cells in such a way that bacteria are engulfed by ruffles on the host cell membrane, resulting in uptake by the epithelial cell. This process of invasion is mediated by a type three secretion system, encoded by genes of the Salmonella pathogenicity island I, and is essential for caecal colonisation (Bohez et al., 2006). Immune cells are attracted to the gut wall and the macrophages may take up bacteria penetrating through the caecal mucosa. This is the start of the systemic phase of the infection as Salmonella bacteria can survive within and replicate in these macrophages. These cells spread the bacteria to the internal organs, such as liver, spleen, ovary and oviduct (Bohez et al., 2008), where the bacteria can be found in large numbers. All this is serotype and strain dependent, and some strains are more invasive than others while others are less or not efficient in persistent caecal or organ colonisation. Shedding can occur intermittently. Contamination of poultry meat can thus be caused by contamination in the slaughterhouse when faecal material or gut content (or internal organ material) is contaminating the carcasses during the slaughter process (for example evisceration, defeathering). Eggs can be contaminated either externally (on the shell) or internally (Gantois et al., 2009). Shell contamination is caused by contamination during or after lay, because faecal material or Salmonella bacteria present in the environment contaminate the outer shell. Internal egg contamination can be caused by Salmonella bacteria that are transported through the eggshell after shell contamination. In addition, internal egg contamination can be caused by Salmonella bacteria that are incorporated in the forming egg during passage in the oviduct. Salmonella can colonise the oviduct after systemic spread and thus contaminate the egg components, depending on the site of colonisation (magnum, egg white; isthmus, shell membrane). While all Salmonella serotypes, to a greater or lesser extent depending on the serotype and strain, are able to colonize the gut and internal organs, Salmonella Enteritidis is far more capable of persisting in the oviduct as compared to other serotypes (Gantois et al., 2008). In addition, Salmonella Enteritidis strains have been shown to be superior in egg white survival (De Vylder et al., 2013). The lipopolysaccharide (LPS) structure and multi drug resistance efflux pumps have been shown to be involved (Raspoet et al., 2014, 2019). The egg white is very antibacterial (high pH, a variety of antimicrobial proteins and peptides) and is capable of killing most bacteria, including most Salmonella serotypes, but Enteritidis strains are rather resistant. These characteristics of Salmonella Enteritidis explain its success in contaminating and infecting humans. In addition, Salmonella Enteritidis is not growing in egg white, but only staying alive and thus no sensory or visual changes occur in the contaminated eggs. Consumers are thus not alerted.

Monitoring programs are of utmost importance in a global strategy of controlling Salmonella, because they assess the prevalence of infected flocks (or even the within-flock prevalence, depending on the method used) and detect changes in prevalence, as well as the serotype distribution, clonal spread, etc. They can also be used to evaluate the efficacy of control methods and programs. Periodic testing using bacteriological detection methods is the most widely used method, but serological methods can also be of value. Bacteriological testing methods are often based on excretion of Salmonella, and thus have inherent problems with sensitivity because infected chickens shed Salmonella intermittently. This can partly be overcome by using mixed faecal samples so that faecal material of many animals is analysed. The within-flock prevalence can however also be low and, if only a low number of animals shed Salmonella, this method will most likely often not detect these infections. Mostly, if a number of samples is analysed and 1 sample is positive, the flock is considered to be Salmonella positive. This positivity is thus not giving information about the actual number of infected animals and the colonization level in the animals. The analytical methods used to detect Salmonella are based on enrichment of the samples for Salmonella and plating of the enriched material on different selective media, often followed by serotype identification. The frequency of the sampling depends on the animal type (breeders, layers, broilers) and the production stage (e.g. pullets vs layers). Often the frequency is higher for breeders as compared to layers, because these animals can contaminate the whole production chain by vertical transmission. As an example, under EU legislation (2160/2003), sampling and detection of all Salmonella serotypes with public health significance should be done according to the following schemes: (a) breeding flocks: day-old, 4 weeks, 2 weeks before transport to the laying unit and every 2 weeks during lay; (b) laying hens: day-old, 2 weeks before transport to the laying unit and every 15 weeks during lay; (c) broilers: before transport to the slaughterhouse. In addition to bacteriological detection, also antibody responses in serum can be used to monitor the Salmonella status of a flock. Antibody detection tests are available in ELISAs and typically detect either O antigens (LPS) or H antigens (flagelllin). While bacteriological detection methods have a higher chance of detecting positive animals in the early period post-infection due to higher excretion, serological tests can detect positive animals a long time post-infection and do not detect antibodies in the early post-infection period due to the dynamics of antibody production after infection. Not all animals, however, generate an efficient antibody response, and also here the number of samples to be taken is not easy to calculate, and this depends on the actual minimal to be detected within-flock prevalence and the accuracy that is defined beforehand. Both methods thus have advantages and disadvantages.

Although control tools are available to reduce Salmonella colonisation, there needs to be a general strategy on the control methods to be used and defining the situations in which specific measures need to be implemented, but also on the actual consequences of finding Salmonella positive samples. For example, for breeding flocks this can mean that the hens need to be eradicated when certain Salmonella serotypes are detected. For layers and broilers, the finding of certain serotypes could imply that the eggs or meat have to be treated in a way that kills the bacteria before the food is marketed. For example, in the EU (Regulation 2160/2003) in case of an infection with Salmonella Enteritidis or Typhimurium in breeding flocks, non-incubated hatching eggs should be destroyed or used for human consumption following treatment in a manner that guarantees the elimination of Salmonella Enteritidis and Typhimurium. All birds from these flocks must be slaughtered or destroyed, even the day-old chicks. Eggs derived from these birds that are still present in a hatchery, also have to be destroyed or treated as described above. Another specific requirement is that eggs must not be used for human consumption as fresh table eggs unless they originate from a commercial layer flock subject to a national control programme. Moreover, eggs originating from flocks with unknown health status, suspected of being infected or from infected flocks may only be used for human consumption if treated in a manner that guarantees the elimination of all Salmonella serotypes with public health significance. The use of control methods on the farms can be made obligatory, depending on the Salmonella status of the flocks, or even the Salmonella status of the flocks in a region.

A lot of experimental vaccines have been produced for chickens, and also a variety of commercial vaccines are available on the market. These comprise both live and inactivated vaccines. The currently available live vaccines are produced by chemical mutagenesis or are selected on culture media as slow growing natural mutants (metabolic drift mutants). In general, it is believed that live vaccines induce better protection because they stimulate both cell-mediated responses and antibody responses, while inactivated vaccines mainly induce antibody production, but both methods are in use, singly or in combination in vaccination regimens. Triple dose vaccination schemes are common for layers and breeders, and also combinations of live and inactivated vaccines are given. Live vaccines are mostly administered in the drinking water (or using a coarse spray) and inactivated vaccines need to be administered parenterally. Autologous vaccines are used in some countries, made by killing a strain isolated from the flock where the vaccine is administered. Cross-protection is shown to be occur but it is believed that intra-serotype and intra-serogroup protection is more pronounced. For example, Eeckhaut et al. (2019) showed that live Enteritidis vaccines significantly reduce Salmonella Infantis colonization in layers.

Vaccines have been used extensively in laying hens and should a) reduce or prevent the intestinal colonisation resulting in reduced faecal shedding and thus egg shell contamination and b) prevent systemic infection resulting in a decreased colonisation of the reproductive tissues, in this way reducing internal egg contamination. Inactivated vaccines are often used in parent flocks. Parenteral administration of inactivated Salmonella vaccines to breeder birds will induce a strong production of antibodies. These antibodies will be transferred to the progeny. The maternally transferred antibodies persist for a few weeks but, although there seems to be some protective effect against disease in the early post-hatch period, there is little effect on intestinal colonisation by challenge strains (Methner and Steinbach, 1997; Methner et al., 1994). There is a report on the efficacy of inactivated vaccines in prevention of egg contamination in layers (Woodward et al., 2002). Gantois et al. (2006) showed that oral vaccination with live vaccines at day 1, week 4 and week 16 decreased internal organ colonisation, including reproductive tract colonisation, and egg contamination. Although it is very difficult to prove reduction of egg contamination following vaccination under field conditions owing to the low and variable percentage of contaminated eggs laid, a European baseline study showed that vaccinated layer flocks were less frequently contaminated by Salmonella as compared to non-vaccinated flocks (4% vs 12%). In theory, an ideal live vaccine strain should possess following characteristics (Van Immerseel et al., 2005):

Multiple scientific groups have reported a phenomenon, in which oral administration of Salmonella wild type and attenuated strains can confer resistance to infection by a virulent Salmonella challenge strain within 24 h of administration. This ‘competitive exclusion’-like phenomenon is called colonization-inhibition. These data suggest that it might be possible to administer live Salmonella vaccine strains to newly hatched chicks such that they would colonize the gut extensively and very rapidly, inducing a profound resistance to colonization by other Salmonella strains of epidemiological significance, which may be present in the poultry house or may also have arisen from the hatchery (Van Immerseel et al., 2005). Colonisation of the gut by the colonisation-inhibition strains would prevent gut colonisation by virulent strains, while invasion in the gut tissue would evoke an inflammatory response that would prevent invasion to the internal organs by virulent strains. This means that live vaccines can thus also be used in broilers to control gut colonisation and shedding. An issue is to administer the strains as early post hatch as possible to the birds; this is not ideal using drinking water applications but can be done using coarse sprays (De Cort et al., 2014).

It is difficult to speculate about the nature of future vaccines but good methods are available to rationally design live vaccines that have defined mutations so that both detection methods and safety aspects are highly controlled. These are, however, genetically modified organisms and their use is still under debate although some are already marketed. Many research groups have designed genetically modified live vaccines with a very good safety and efficacy profile, and with markers that are differentiating the strains and the serological response from wild type strains and serum responses, respectively. In relation to emerging phenotypes and the variety of Salmonella phenotypes in broilers, developing vaccines against other serotypes can become a need, but the registration process is long, hampering development of these vaccines.

To be complete, one needs to mention that, in addition to vaccines, other methods are available and a multiple hurdle approach is needed. Biosecurity is crucial, and it is evident that rodent and insect control and general hygienic and biosecurity measures are a prerequisite for keeping Salmonella out of the farms. In addition, many drinking water and feed additives are being used, including organic acids such as butyrate, prebiotics, probiotics and phytochemicals, amongst others. This is not within the scope of this paper but reviews can be consulted: Van Immerseel et al. (2002), Micchiche et al., (2018), Clavijo and Flores (2018) and many more. It is a utopia to eradicate Salmonella from chickens and the environment, but one should try to aim for appropriate levels of protection, and thus low flock prevalence and within-flock prevalence, and low individual colonization levels.

ACKNOWLEDGEMENTS: We thank all scientists, technicians and personnel from the Department of Pathology, Bacteriology and Avian Diseases, all funding agencies and industrial partners, and all research groups that have contributed to the data used to produce the paper.