Erysipelas—A Review of an Emerging Disease in Layers Open Access

Erysipelas is a bacterial disease caused by Erysipelothrix rhusiopathiae (ER). The bacterium is a slender, slightly curved, Gram-positive rod with rounded ends, measuring 0.2–0.4 × 0.8–2.5 μm. It is facultatively anaerobic and tends to grow as long filaments. The colonies are small (0.3–1.5 mm), and a narrow zone of alpha hemolysis is visible on blood agar plates.

The first isolation of a bacterium of the genus that was later to be named Erysipelothrix was made by Robert Koch in 1876 when injecting putrefying blood into mice (1). In a letter to Dumas in 1882, Louis Pasteur (2) mentioned investigations on a bacterium causing “rouget” in pigs. This is of special interest from a poultry perspective because these bacteria were described as very similar to the ones causing fowl cholera, with the exception that they were smaller and differed in physiologic properties (no further details mentioned). Pasteur was able to induce disease in pigs, sheep, and rabbits when injecting them with cultures of the organism; chickens however remained unaffected (2). A few years later, Friedrich Löffler published a description of the disease in pigs and the causative organism (3). Around the same time, in 1884, the bacterium was also found to be a disease-causing agent in humans (4). According to a review by Beaudette and Hudson in 1936 (5), the first case of erysipelas in poultry was reported by Jarosch in a turkey in 1904, and the first report of a case in a chicken was reported by Hausser in 1909. This was followed by several additional reports in chickens and subsequently also in other poultry species (5).

At least 36 different names of the bacterium can be found in the literature (6). In 1876, Koch called it a bacillus causing septicemia in mice (1). In 1885, the first name of the species was introduced by Trevisan. However, based on a description of the bacterium as a spore-forming rod, it was named Bacillus insidiosus (7). The first species name to be linked to today’s name, Bacterium rhusiopathiae, was introduced by Migula (6), and in 1909, Rosenbach was the first to name the genus Erysipelothrix. At the same time, three separate species were suggested based on the host of origin: Erysipelothrix porci (pigs), Erysipelothrix murisepticus (mouse) and Erysipelothrix erysipeloides (human) (4). The genus-species combination Erysipelothrix rhusiopathiae was introduced in 1918 by Buchanan (8), and it was designated as the type species of the genus in 1920 (9). During the following years, several new names were proposed, and the bacterium was placed in different genera depending on the host of origin. Based on the recognition that all strains belonged to a single species, Langford and Hansen in 1954 proposed a common name, Erysipelothrix insidiosa, in recognition of the first name of the species by Trevisan (7). In 1966, a return to the previously used E. rhusiopathiae was suggested based on earlier common use (10). For more than 20 years, it was believed that the genus Erysipelothrix only consisted of this single species.

Erysipelothrix is the type genus of the family Erysipelotrichaceae in the phylum Bacillota (synonym Firmicutes) (11,12). The closest known relatives of Erysipelothrix include the genera Holdemania, Solobacterium, Bulleidia, and Anaerorhabdus (13). Many Erysipelotrichaceae species appear to be commensal intestinal bacteria in warm-blooded animals, and several representatives occur in feces from healthy poultry and pigs (14,15). The genus Erysipelothrix, however, contains both primary and opportunistic pathogens as well as commensal organisms. Erysipelothrix rhusiopathiae is the type species of the genus, which contains nine accepted species as of September 2024 (Fig. 1) (16). The other “classical” species of genus Erysipelothrix, Erysipelothrix tonsillarum, was described in 1987 and is closely related to ER (17). It is considered to be largely avirulent, and it does not cause erysipelas in poultry or pigs (17,18,19), but reports of disease in dogs (20) and rodents (21,22) can be found. Species described in recent years based on genomics include Erysipelothrix piscisicarius (previously named Erysipelothrix sp. strain 2) associated with disease in ornamental fish (23) and turkeys (24), the closely related Erysipelothrix anatis and Erysipelothrix aquatica recovered from multiple animal hosts (25), Erysipelothrix urinaevulpis recovered from fox urine (25), Erysipelothrix inopinata found in vegetable broth (12), and Erysipelothrix larvae isolated from beetle larvae (26). Most recently, Erysipelothrix amsterdamensis was formally described from outbreaks of seabird mortality (27), having earlier been published under the name Erysipelothrix enhydrae based on findings in sea otters (28).

Erysipelothrix rhusiopathiae and closely related species have traditionally been divided into numbered serogroups based on antigenic polysaccharide complexes, with serogroups 1a, 1b, and 2 most commonly associated with disease in pigs (30). These serogroups have also been applied to isolates from poultry, and there are currently 28 known serogroups (22,31,32,33). Serotyping is dependent on access to panels of species-specific reagents, and therefore it is only available to specialized laboratories, it has limited resolution, and not all isolates are typable. As for many other bacterial species, molecular and genomic methods have therefore gradually taken over.

Early studies on the molecular epidemiology of ER proposed the presence of subgroups within the species, using, e.g., restriction fragment length polymorphisms (34), pulsed-field gel electrophoresis (PFGE) (22,35), multilocus enzyme electrophoresis (36), and later multilocus sequence typing (37,38). In recent years, typing methods based on whole genome sequencing have taken over as the gold standard in microbiology, offering both robust surveillance and high-resolution outbreak investigation capabilities as well as the ability to identify individual genes and gene variants. The Fujisawa strain of porcine origin was the first ER to be whole genome sequenced in 2011, revealing a comparatively small genome (1.8 Mbp) consistent with reductive evolution and adaptation for intracellular parasitism in a nutrient-rich environment (39). A later comparison of genome sequences from a large set of isolates from diverse sources found a high degree of core genome recombination and a complex population structure with a distinctive clade 1 and two less well resolved clades 2 and 3, with some isolates deemed to be intermediate (I) between the two latter (40). In contrast with the other clades, clade 1 isolates were not recovered from poultry or pigs (40), and they were also absent in a later study investigating historical ER isolates from U.K. erysipelas cases in pigs (41). However, clade 1 was found among ER isolated from healthy pigs in Sweden and healthy wild boars in Sweden and Italy, while it was absent among pigs showing clinical signs of disease in the same countries (42), supporting the notion of lower virulence of clade 1 strains for terrestrial mammals.

The ER isolates belonging to clades 2/3/I have been suggested to be generalists in terms of host specificity, as they have been observed to intermingle within clades regardless of their avian or mammalian species of origin (37,40,42). This indicates a risk of cross-species transmission and spread of infection between wildlife and domestic animals. In contrast, and as previously mentioned, certain serotypes have been observed to more frequently cause severe outbreaks among pigs (43). Serotypes can be inferred via PCR or in silico detection of the genes responsible for producing the antigens (41,44,45,46), allowing comparison between clades and serotypes. This has revealed the serotypes to be paraphyletic; i.e., the members of each serotype do not reliably derive from a more recent common ancestor compared to isolates from other serotypes (41,42), explaining the apparent contradiction between genetic and serotype host specificity. Clonal outbreaks and long-term establishment of successful pathogenic clones in the pig population have been reported for serotype 1a in Japan and China (43), but, in contrast, a high diversity has been found among ER causing clinical cases in European pigs and poultry (37,41,42,47). Outbreaks of ER confirmed to be clonal by genome sequencing have occurred in wildlife such as muskox and other Arctic mammals (48) and harbor porpoises (49). Continued genomic study of isolates recovered from different countries and host species both with and without clinical signs will be necessary to better explain and predict the host preference and virulence of different ER strains.

Several genes and genomic regions have been implicated as being involved in ER virulence (39,50). The capacity to produce a protective polysaccharide capsule is considered to be a core virulence attribute that is dependent on products of the cps gene cluster (51). Mutant strains that lack the capsule were shown to lose their pathogenicity in a mouse model (50). Capsule production has also been linked to genomic regions responsible for production of antigen structures relevant for serotyping, possibly explaining the differences in clinical virulence between serotypes (44). Host antibodies to the protein encoded by surface protective antigen (Spa) genes of Erysipelothrix spp. have long been known to be protective against infection (52). The Spa protein occurs in three sequence variant groups (53), with genes encoding SpaA carried by ER, except ER clade 1 (40,41), as well as by E. amsterdamensis (27,28), SpaB carried by ER clade 1 (40,42), and SpaC carried by E. piscisicarius (23,24). The SpaA protein has been shown to act as an adhesion factor (54,55) and to protect against phagocytosis (56). Other putative virulence factors are less well characterized, but candidates include the hemolysin gene hlyIII (57), neuraminidase encoded by the nanH gene (58), and genes with possible roles in biofilm formation and survival in host phagocytes (39). A role in tissue dissemination would be expected from the hylC hyaluronidase gene based on the function of homologs in other pathogens (50); however, lethality was unaffected when mice were challenged with an hylC knockout strain (59).

In veterinary medicine and among farmers in general, erysipelas is mostly known as a disease in pigs. In pigs, the disease can occur in different clinical forms, including acute, subacute, and chronic. During the acute form, depression, high fever, characteristic diamond-shaped cutaneous lesions, and death of one or more animals can be seen. The chronic form may follow acute or subacute cases and is characterized by signs of arthritis and, in some cases, valvular endocarditis and death (60). Adult breeding pigs are commonly vaccinated against erysipelas, and vaccination of fattening pigs on previously affected farms is recommended (33). ER has also been isolated from organ samples from healthy pigs. In fact, the domestic pig has for a long time been considered to be the most important reservoir of the bacterium, with the epithelium of the tonsillar crypts being a likely site for a persistent carrier state (61). An estimate of 30%–50% of healthy pigs as carriers of ER in tonsils and other lymphoid organs was introduced early and is still often referred to in scientific publications (33,42). However, more recent studies have found only 0–3% of sampled pigs culture positive for ER in their tonsils (62,63).

Besides pigs, erysipelas may also be a considerable problem in sheep, in which the most common manifestation is polyarthritis in lambs. In some sheep-producing areas in the world, ewes are vaccinated in order to protect the lambs during their first weeks of life (33,64). In other domestic mammals, erysipelas is less common, but cases have been sporadically reported, e.g., in cattle, goats, horses, dogs, and cats (65,66,67,68,69). In parallel to pigs, ER has also been isolated from tonsils of healthy cattle (70).

Various wild mammals have been reported to have erysipelas, such as muskoxen, sea otters, numbat, and roe deer (28,71,72,73). In recent years, disease outbreaks, sometimes with mass mortality, have occurred in marine mammals such as dolphins and porpoises (49,74). Wild boars can be of special epidemiologic interest because of their wide geographic distribution, abundance, and potential interaction with farm animals. Studies suggest a high seroprevalence and ER isolation rate, and in parallel to domestic pigs, it has been suggested that wild boars may constitute a reservoir (75,76).

ER was for a long time considered nonpathogenic for fish, but isolation of the bacterium from the surface slime of fish was reported (77,78). It was, however, suggested early that this could represent contamination from the surrounding environment after the fish was caught (77). Interestingly, in recent years, several reports on mortality among fish associated with ER, as well as other Erysipelothrix spp., have been published (23,79,80,81).

When Rosenbach isolated ER from a human patient in 1884, he named the infection erysipeloid (4). This term is still in use today for the disease in humans, which most often manifests itself as a localized cutaneous infection. Cases with more general skin infection, fever, septicemia, arthritis, endocarditis, and sometimes death may also occur (82,83,84). It should be noted that there is a disease in humans called erysipelas, but it is caused by a streptococcal infection (85). Erysipeloid is an occupational hazard among people who handle animals and fish, and reports of suspected erysipeloid in poultry workers and slaughterhouse personnel have been published (86,87).

For a long time, the turkey was considered as the most susceptible and most commonly affected poultry species. This is reflected by the number of international reports on erysipelas during the early and mid-20th century, which describe outbreaks in this species. Internationally, this may still be true, despite the development of vaccines for turkeys and the increasing number of cases in laying hens. Recently, another species, E. piscisicarius, was also found to cause mortality in turkeys (24,88). Besides ER, this is so far the only other species within the genus Erysipelothrix that has been found to be pathogenic in poultry.

Outbreaks in turkey flocks are most often diagnosed in flocks over 10 wk of age (89,90). However, there is one case report of erysipelas in poults as young as 2–4 days old, following toe trimming (89). In many turkey cases, increased mortality is reported with no prior clinical signs. Birds may, however, show signs of depression and unsteady gait. Mortality in field cases may range from a slight increase to 40% (5,90,91). It has been suggested, but not proven, that genetic variation might be relevant for resistance to erysipelas in turkeys based on experiences when female turkeys of several genetic lines were intermingled (92). In males, cutaneous lesions on the head and neck indicate that infection may enter through the skin during fighting. In addition, male snoods may be swollen and purplish. Losses of turkey hens with perineal congestion and skin discoloration 4–5 days after artificial insemination may also occur. In turkeys and other poultry species, vegetative endocarditis associated with ER may cause loss of body condition and weakness with subsequent death.

Geese and ducks of all ages appear to be susceptible to ER infection (90,93,94,95). Mortality in young birds has been reported in these species, which is not often observed in turkeys or chickens. Clinical signs in geese, if reported, include incoordination, somnolence, and polydipsia. In ducks, in addition to mortality, clinical signs such as diarrhea and lethargy can be observed, and in breeder ducks, there may also be a drop in egg production (90). Chronic erysipelas with arthritis and endocarditis can develop in both species (96,97). Muscovy ducks are also susceptible (90).

In both breeder flocks and flocks of quail (Coturnix sp.) raised for meat, erysipelas may cause outbreaks with sudden high mortality. Also, lethargy is reported as a common clinical sign, if any, before death. Diarrhea, dehydration, facial edema, serous nasal exudation, and lacrimation have also been reported in acute disease (87,98), and arthritic conditions have been reported in chronic disease (98).

High mortality due to erysipelas also occurs in birds kept for restocking game, such as guinea fowl, partridges, and pheasants. Interestingly, most published case reports deal with outbreaks in young birds. Clinical signs include lethargy and sometimes drooping of wings, huddling, and crusty exudate on eyelids (99,100,101,102).

Erysipelas in ratites has primarily been reported in emus, where juveniles appear to be most often affected. Clinical signs associated with infected emus include lethargy, weakness, anorexia, dyspnea, bright green diarrhea, and sudden death with mortality rates up to 35% (103,104).

Erysipelothrix rhusiopathiae isolates from wild birds have been included in several studies, but case reports on outbreaks of erysipelas are relatively rare.

As mentioned previously, the first case of erysipelas in a chicken was reported in the beginning of the 20th century. Several reports in this species followed (31,86,105,106,107,108,109). For many years, erysipelas was considered to be a rare disease in chickens (106,109,110,111). Since the late 1990s, an increase in reports of erysipelas cases in layers can be noted, as well as research activities. The aim of this part of the review is to focus on experiences and research on erysipelas in chickens (especially in layers), with special emphasis on the advances in knowledge on the immune responses to ER infection in chickens.

During the 21st century, erysipelas has become an emerging problem in layer flocks in several parts of the world. The increased incidence was first observed during the initial stages of the change of housing systems for layers in Europe (112,113,114). Subsequently, following the European Union–wide ban of conventional battery cages in 2012 aiming to improve the welfare for layers in Europe, the number of erysipelas outbreaks increased further. Outbreaks of erysipelas have also occurred in layers from Australia and the South Pacific, and the occurrence in layers has also increased in parts of the United States (115). Interestingly, Silva et al. (2020) reported no cases in commercial layers in California during 2000–19 (90), but it remains to be seen whether this will continue after the in-state battery cage ban that went into effect in 2022.

In layers, the difference in erysipelas occurrence associated with housing system is a striking feature. Speculations on an association between housing system and erysipelas outbreaks was introduced in a Danish paper in 2002 (116) and was further supported by a Swedish study on causes of mortality in layers in different housing systems during 2001–04 (117). Subsequently it has been shown that flocks with outdoor access (free-range and organic) are at a higher risk than flocks in indoor litter-based systems, while flocks in cages are at the lowest risk (118). In fact, only two outbreaks in flocks in conventional battery cages (86,107), and none in furnished cages, can be found in the literature. The explanation for the difference between housing systems is not fully known. The very few reported cases in flocks in cages, combined with the fact that erysipelas for many years was considered a rare disease in layers, may indicate that the transmission routes between layers in cage systems are interrupted. The increased risk for flocks with outdoor access compared to flocks in indoor litter-based systems should be further considered. Epidemiologic aspects such as increased contact with potential sources and vectors in the wild fauna may constitute an important risk.

Outbreaks of erysipelas in layers most often occur in flocks that have passed peak production. Moreover, outbreaks in young chickens and pullets (layers before the onset of lay) are very rare. To the authors’ knowledge, only two reports on cases in young chickens can be found in the literature, one in broilers at 6 wk of age (106) and one in 8-wk-old pullets, in which the pathologic findings were consistent with the concurrent Eimeria necatrix infection and not with ER (111). These observations strongly indicate a relationship between age and mortality in chickens, i.e., that older birds are more susceptible.

A seasonal distribution of outbreaks, with more cases occurring during the winter months, has been observed for avian species in California (90), and seasonal differences in isolation rates when sampling broiler chickens at an abattoir have also been shown (119). This contrasts to an early report on cases in layers in Denmark during the 1970s indicating no seasonal distribution (31).

In affected layer flocks, erysipelas outbreaks usually start suddenly with an increase in mortality that in a few days can reach dramatically high levels (120). The total mortality during an outbreak may vary between 10% and 60% (105). In some cases, and for unknown reasons, the mortality is significantly lower (1%–2%) (31). In individual hens, the course of the clinical disease is of short duration (105), and the clinical signs observed resemble those in other poultry species. In some flocks, owners may observe a few drowsy hens (105) and, in some cases, also periorbital swelling (121) (Fig. 2). In some flocks, a drop in egg production is reported. In other flocks, egg-laying percentage for the remaining hens in the flock stays normal (105), presumably due to the short duration of the disease in individual birds.

Taken together, erysipelas outbreaks in layer flocks usually have an acute onset with a fulminant progression. Hence, differential diagnoses for erysipelas in poultry include other acute systemic bacterial and viral infections such as fowl cholera, colibacillosis, Newcastle disease, and highly pathogenic avian influenza. It has been shown that flocks housed in litter-based systems, especially with outdoor access, are particularly at risk. These characteristics make erysipelas an important factor to consider when layer management is altered to improve animal welfare.

Layers that die during the acute stage of erysipelas are often in good body condition at necropsy. Typical gross findings include generalized congestion of skeletal muscle and viscera, petechiae in coronary and abdominal fat, ecchymotic hemorrhages in subepicardial locations and organs such as the myocardium and spleen, prominent hepatosplenomegaly, and small to moderate amounts of clear coelomic fluid sometimes admixed with fibrin (86,115). In laying birds, there is often free yolk material in the coelomic cavity originating from acutely regressing egg follicles, which appear flaccid and congested (Fig. 3). Occasionally, hemorrhagic preovulatory follicles may be observed. Caseating exudate in the coelomic cavity is rarely observed, unless there is bacterial coinfection with, for example, Escherichia coli or Pasteurella multocida spp. multocida. The liver is friable, has rounded edges, and has dark or tan discoloration with scattered round to irregular pale foci of variable sizes (Fig. 4). The spleen is markedly enlarged, dark, and often mottled (Fig. 5). Swollen dark kidneys are commonly observed, and sanguinomucinous contents may be present in the small intestine. In layers, vent pecking and cannibalism are common sequelae in acute erysipelas, presumably because moribund hens are easily targeted by flock members. Postmortem pecking at the vent region may also occur. In turkeys, crusty skin lesions have been described as a common finding, especially on the head, snood, and wattles, whereas skin lesions in chickens in association with erysipelas are more rarely reported. However, periorbital swelling/swollen head has been described from layers and broiler breeder chickens (121,122). The overall occurrence of swollen head in affected flocks is not clear, presumably as the lesions may be mild and overlooked at necropsy, and the causal association with erysipelas has not been conclusively proven. In chickens that survive the acute stage, valvular endocarditis may develop, usually involving the left atrioventricular valve. At necropsy of such cases, ovarian regression has usually progressed to the point that laying has completely ceased. Carcasses may appear pale, and there may be accumulation of coelomic fluid and loss of body condition. Synovitis and arthritis are rarely diagnosed in chickens, but they may occur in ducks, geese, and turkeys (89,96,97).

Microscopic lesions in acute erysipelas of poultry are dominated by extensive fibrinoid necroses in liver and spleen, edema and hemorrhage, vascular lesions, and lymphocytic depletion in the spleen, usually accompanied by absence of or very mild inflammation with mononuclear cells and a few heterophilic granulocytes. Sometimes there are signs of widespread intravascular thrombus formation in liver sinusoids and capillaries in glomeruli and lungs, which suggests that disseminated intravascular coagulation may be an important step in the pathogenesis (123). Bacterial aggregates may be observed in close association to or within endothelial cells, in intravascular thrombi, and engulfed by mononuclear phagocytes such as splenic reticuloendothelial cells and Kupffer cells in centrilobular and periportal regions of the liver. In valvular endocarditis, necrotic debris admixed with numerous bacteria adhere to the endothelial lining of the valve, and there may be inflammation extending to the adjacent myocardium with edema and separation of myofibers and a mixed inflammatory cell population.

In summary, the gross findings at necropsy, including hepatosplenomegaly, are often indicative of an acute bacterial infection, but in severe outbreaks with high mortality, the possibility of a viral infection must not be ignored, and further diagnostic efforts are important.

Other bacterial and viral infections constitute important differential diagnoses (see above). From diagnostic experience of outbreaks in Sweden, organ samples from spleen and liver obtained with aseptic techniques from fresh layer carcasses are normally sufficient for isolation of ER following direct culture on nonselective (blood agar) plates. This is in contrast to results reported for samples from pigs, for which indications have been found that broth enrichment is more sensitive for isolation of ER than direct culture (124). When culturing from decomposed carcasses, bone marrow from long non-pneumatized bones may be a preferable sample material because ER may survive longer there than in organ samples. Bone marrow from muskoxen in an advanced state of decomposition was used successfully during an outbreak (125).

Because ER grows as very small colonies on nonselective blood agar plates, colonies may be difficult to detect from contaminated samples. Therefore, in cases when there is a risk of contamination, initial incubation in a selective broth for 24–48 hr followed by culture on nonselective blood agar plates may be necessary. In the case of heavily contaminated samples, environmental samples, or other samples containing a presumably diverse flora, selective broth could preferably be streaked on selective or inhibitory agar plates. ER has the ability to grow in the presence of sodium azide and crystal violet at concentrations that inhibit most other organisms. This combination was first described by Packer in 1943 (126), and several selective media have been developed based on this principle. Another selective medium based on a combination of antibiotics (kanamycin, neomycin, and vancomycin) was described by Wood in 1965 and named the “Erysipelothrix selective broth” (ESB) (127). Studies on samples from pig and layer farms have shown that modified versions of the original ESB, even with the exclusion of vancomycin, work well for investigating complex environmental samples (128,129,130).

For identification of ER colonies, colonial morphology, Gram staining, and biochemical tests have been used for a long time. In recent years, matrix-assisted laser desorption/ionization–time of flight mass spectrometry (MALDI-TOF MS) has replaced biochemical tests for species identification in many diagnostic laboratories and hence shortened the time to diagnosis; however, the ability to differentiate ER from other Erysipelothrix spp. using commercially available databases is unclear (84).

Several PCR-based methods have been developed for detection of ER that can be used for chicken samples (e.g., 131,132). For detection of ER nucleic acid in chicken blood samples, PCR sensitivity can be increased by gradient centrifugation to simultaneously enrich leukocytes that may harbor intracellular bacteria in the sample and exclude red blood cells to reduce the amount of chicken DNA that has an inhibitory effect on the PCR (133). The primers and probes specific for ER developed by Pal et al. (132) for real-time PCR have also been successfully used for digital PCR quantification of ER genomes in chicken blood samples (133) and for relating the number of ER genomes to the number of chicken genomes in organ samples (134).

Already in 1975, Bisgaard and Olsen (105) concluded that “considering the economic aspects of erysipelas infections in large flocks producing eggs for consumption and the severely reduced possibilities of treating such flocks, serological determinations of the isolated strains and further examination of the environment, together with experimental investigations will be of importance to elucidate the epidemiological conditions.” Despite further work during the years that have followed, the epidemiology of erysipelas in layers is still not fully understood. Several aspects related to the sources of the infection and the transmission routes between and within flocks are still unknown.

ER has been considered to be ubiquitous (33,135). In addition, early belief was that ER was a soil saprophyte (136,137) that could survive indefinitely and even multiply in soil (138). Thus, soil in itself was considered to be a possible source of the infection. However, studies on the survival time of ER in soil under experimental conditions have shown a maximum survival time of 72 days (136,138). The survival of ER in different matrices, including soil, has previously been reviewed by Mitscherlich and Marth (139). Moreover, as mentioned, the reduced genome size of ER suggests this bacterium is adapted to replication in a host (39). Thus, the importance of animal hosts for the long-term maintenance of ER is apparent, and several animal species, e.g., pigs, wild boars, and cattle, have been shown to be or assumed to be reservoirs of ER and thus potential sources of the infection (33,70,75). Indeed, as mentioned, molecular epidemiology has indicated a risk for cross-species transmission, including from wild and domestic mammals to layers and other poultry. Whether chickens themselves may act as carriers of ER is still not clarified. Limited numbers of reports suggest that ER may be isolated from healthy chickens, i.e., from Japanese broilers at slaughter and African backyard chickens (119,140), which may indicate that chickens might be carriers of ER. However, the possibility of environmental contamination, including transfer from other species, needs to be further studied to establish if chickens indeed are true carriers. An observation of interest in this context is that several serologic studies have shown antibodies that recognize ER in healthy chickens without any signs of ER related disease (118,141,142,143,144). The seroprevalence of antibodies to ER varied from 5% to 100% between studies, which can be attributed to several factors, including the methodology used and the cohorts studied. This serologic evidence hence suggests that ER or antigenically similar bacteria are frequently present in conventional chickens. Evidence for other Erysipelothrix spp. in environmental samples from healthy chickens was recently found (15). Accordingly, it seems likely that exposure to ER in the environment could be a reason for the high prevalence of laying hens with antibodies recognizing ER. It remains to be shown whether other Erysipelothrix spp. in the chicken environment can also induce antibodies that cross-react with ER.

PFGE subtyping of ER isolates from layer flocks with ongoing outbreaks suggested that these isolates were of clonal nature and therefore probably originated from a single extraneous source (120,129). On some farms with repeated outbreaks in subsequent flocks, molecular epidemiologic studies have shown that these outbreaks may be caused by identical farm-specific strains (22). Repeated outbreaks in different flocks on the same farm may also be caused by different strains, suggesting new introductions (22,129).

In summary, for long-term survival of ER, animal reservoirs are needed, and the wide host range of the bacterium offers ample opportunities for both wild and domestic supplies of such. In addition, the survival capacity of ER in, e.g., soil suggests this as a short-term source of the infection, if there is a breach in biosecurity barriers or direct access to newly contaminated soil, which may be the case in free-range production. The acute and often fulminant nature of erysipelas outbreaks in layer flocks may suggest introduction of a novel pathogen into a naïve population. However, this stands in contrast to the serologic evidence showing that many healthy hens seem to have been exposed to ER during the production period. Hence, further studies on the potential multifactorial pathogenesis of this disease are needed.

Knowledge on the transmission between layers in a flock is still limited, but it has been suggested that ER may be spread through broken skin and mucous membranes and also that feather pecking, vent pecking, and cannibalism may enhance transmission. Pecking at carcasses may also be a risk for bacterial transmission. Because ER has been isolated from jejunal contents and manure from infected layer flocks (129), it is suspected that, once introduced into a flock, the spread of ER occurs via the fecal-oral route. The difference between housing systems in outbreak risk, including the historically low occurrence of outbreaks in Europe when layers were predominantly housed in conventional cages, may support this theory as layers in litter-based systems are continuously exposed to feces in contrast to caged layers. Moreover, noncaged layers have more frequent interactions with other hens compared with hens in cages. For breeders, it can be noted that ER is not vertically transmitted and does not adversely affect hatching egg quality (145).

It has been shown that insects, e.g., flies and mosquitoes, may transmit ER to pigs and pigeons (146,147,148). The internationally widespread poultry red mite (PRM; Dermanyssus gallinae) may carry and transmit several pathogens (149), and ER has been isolated both from the integument and the interior of this parasitic mite (150). Therefore, PRM has been suggested as a potential reservoir and/or vector of ER (150). Both of these aspects have been studied but not proven (120,151). The possible role of other arthropods in transmission of ER also requires further consideration.

Despite the long history of, often relatively successful, vaccination against erysipelas in, e.g., pigs and turkeys (reviewed in Opriessnig et al. [33]), the general knowledge on host immune responses upon ER infection is still incomplete across species. Herein, the current knowledge on chicken immune responses to ER infection is reviewed, with some comparisons to studies in other avian and mammalian species.

Innate immune mechanisms are the first responses activated upon encounter with an infectious agent and are imperative for swift control of pathogen replication and spread of infection in the host. Moreover, these responses are crucial for the initiation and regulation of ensuing specific immune reactions. Nonetheless, information about the innate immune recognition of ER in general is very limited. Using experimentally ER-infected, specific-pathogen-free (SPF) chickens, a rapid and transient innate response in the circulation of naïve chickens was observed (134,152). Collectively, these studies showed prompt heterophilia, lymphopenia, monocyte activation, and increased levels of the acute phase protein mannose binding lectin (MBL) in the circulation early after ER infection.

A transient increased expression of the mannose receptor 1 like-B (MRC1L-B) on the surface of circulating monocytes was consistently observed on day 1 after infection (134,152). In one of these studies, it was also shown that the monocytes had a MRC1L-B^high–major histocompatibility complex class II (MHCII^low) phenotype on day 1 after ER infection (134). It has been shown that spleen macrophages of the MRC1L-B^high–MHCII^low phenotype have high phagocytic capacity and that this population increased upon lipopolysaccharide challenge of chickens (153). Moreover, monocytes with decreased MHCII expression and concurrent increased phagocytic activity were observed in the circulation of chickens early after Avibacterium paragallinarum infection (154), and monocytes with increased MRC1L-B expression were likewise observed in the circulation of chickens early after Escherichia coli infection (155). Hence, it seems that a shift of monocyte phenotype in the circulation to MRC1L-B^high–MHCII^low, either by alteration of receptor cell surface expression or by redistribution of monocyte populations, is a general feature in response to bacterial infection and may even confer rapid phagocytic clearance of the infection.

Prompt and transient heterophil and MBL responses were also a general feature upon ER infection of naïve chickens (134,152). The magnitude and persistence of these responses were often connected to the presence and levels of bacteremia in the individual chicken. Heterophils are important effector cells in many bacterial infections in birds (156,157) and may thus have a role in killing ER, e.g., by phagocytosis. Phagocytosis and killing of ER in vitro have been shown in porcine and murine neutrophils (50), i.e., the mammalian counterpart to heterophils. In addition, significant proportions of the bacteria recovered from chicken blood during early ER infection, i.e., concurrent with heterophilia, were intracellular in leukocytes (133). Heterophils may also have further roles in the innate response to ER infection, e.g., as sentinel cells and as regulators of immune responses.

The serum MBL responses observed upon experimental ER infection of chickens were very prominent, at least twofold higher than other infection-induced increases previously monitored in chickens (158,159,160,161). The acute phase protein MBL is a C-type collectin and soluble pattern-recognition receptor with several roles in the innate response to pathogens in mammals and chickens (162,163). It has high affinity to carbohydrate residues on the surface of microorganisms that mediates, e.g., direct opsonic properties. In the naïve host, innate opsonins such as MBL and complement proteins can be important for effective phagocytosis of infectious agents (162,164). For example, early studies in mice indicated a survival-critical role for complement components, putatively as opsonins, in the innate response to ER infection (165). In view of the MBL responses in ER-infected chickens, and considering that the polysaccharide capsule of ER has a high content of mannose (51), a known MBL ligand (162,163), one could hypothesize that MBL has a role in ER elimination as an innate opsonin. Conversely, it seems that ER infection in chickens may still persist despite high serum levels of MBL (134). Hence, further work is needed to determine the possible effector role of MBL in chicken ER infections.

In addition to experiments with SPF chickens, immune reactions in layers in conventional flocks with ongoing natural erysipelas outbreaks have also been monitored. Preliminary results comprising four flocks have been compiled (166). These results revealed some distinct features of circulating leukocytes depending on the level of ER bacteremia. Hens with high bacteremia predominantly had low numbers of circulating monocytes, lymphocytes of all subtypes, and thrombocytes, and a high expression of CD45 on heterophils. Hens with medium and low bacteremia had high numbers of circulating heterophils and monocytes. Regarding the phenotype of circulating monocytes, high bacteremic hens had monocytes of MRC1L-B^highMHCII^low phenotype, while hens negative for ER predominantly had monocytes of MRC1L-B^lowMHCII^high phenotype. Interestingly, a mixed group of medium, low, and negative hens had monocytes with a MRC1L-B^lowMHCII^low phenotype. Decreased MRC1L-B expression on monocytes subsequent to the initial increased expression upon ER infection of SPF chickens was also observed (152), and in naïve chickens, this decrease, similar to that in the layers during natural ER infection, occurred concurrent with heterophilia. In addition, reanalysis of data (for the benefit of this review) from experimentally ER infected chickens (134) showed that the MRC1L-B^highMHCII^low monocyte phenotype was found predominantly in chickens on day 1 after infection and in chickens with bacteremia measuring >10⁴ cfu of ER/ml blood, while the majority of infected, non-bacteremic chickens on day 2 after infection had monocytes of an MRC1L-B^lowMHCII^low phenotype (Fig. 6). One may thus speculate that the MRC1L-B^lowMHCII^low monocyte phenotype is associated with remission/clearance of ER infection. Recent results from single-cell transcriptome mapping of chicken leukocytes (167) have indeed shown that with respect to the levels of MRC1L-B and MHCII mRNA expression, several monocyte subtypes occur in blood, and different functions for these subtypes were indicated. For the future, more basic knowledge on chicken monocyte phenotypes will thus aid further understanding of chicken immune responses to ER infection.

In summary, among innate responses to ER infection in chickens, a dominance of the MRC1L-B^highMHCII^low monocyte phenotype in the circulation seems to be strongly associated with the early phase of infection. Heterophilia and sometimes monocytophilia seem to be associated with moderate and subclinical ER infections. In such infections, phagocytosis by heterophils and MRC1L-B^highMHCII^low monocytes appears to be a likely effector mechanism for elimination of the infection, putatively aided by innate opsonins such as MBL. In chickens with signs of severe disease and high ER bacteremia, severe leukopenia, including thrombocytopenia, is a striking feature. This thrombocytopenia is also consistent with the suggestion that disseminated intravascular coagulation could be a part of the pathogenesis in chicken erysipelas (123).

Protective immunity to any infection is dependent on development of antigen-specific immune effector mechanisms that are effective in the elimination of the infectious agent in question, preferably at their site of entry into the host. For bacterial infections, effective specific responses against primarily extracellular bacteria often involve the production of antigen-specific antibodies that aid the killing of bacteria, e.g., by mediating bacterial lysis or by opsonization for enhanced killing by phagocytic cells. For primarily intracellular bacteria on the other hand, antibody production may be less effective, and T-helper 1 (Th1)–type responses focused on killing of intracellular pathogens will be among the effector mechanisms that confer protection (168). After infection of a host, ER bacteria are often located intracellularly (50,66,133,169,170), and it has also been shown that ER may survive and replicate in murine macrophages (50,170). The intracellular aspects of ER infection must consequently not be overlooked when considering effector mechanisms in protective immunity to this infection.

In chicken experimental studies aiming to induce ER immunity either by vaccination (152) or by previous ER infection (E. Wattrang, pers. comm.), heterophil, monocyte, and MBL responses were rapidly induced similarly to those in naïve chickens, but they were subsequently swiftly downregulated already on day 2 after infection. As a response to previous infection, this could potentially reflect a more efficient innate response by so-called trained innate immunity (171). However, trained innate immunity seems to be less likely as a response to the inactivated vaccine used (152). Hence, the observed outcome most likely involved specific immune mechanisms that quickly aided in the clearance of ER, which shortened the innate response. Nevertheless, the effector mechanisms that mediated this protection have not been conclusively identified.

Regarding antibody-mediated immunity to ER, early work in mice showed that antibodies to ER enhanced both in vitro phagocytosis and intracellular killing of the bacterium (50). ER-specific immunoglobulin Y (IgY)–mediated enhancement of bacterial phagocytosis by chicken cells in in vitro tests have also been observed (E. Wattrang, pers. comm.). Upon experimental ER infection, antigen-specific IgY responses in serum have readily been observed approximately 1 wk after infection of conventional Leghorn-type hybrids (143,151), although infection of an inbred research line, L10 of UM-B19 × white Cornish origin, induced no or very brief ER-specific immunoglobulin M (IgM) and IgY responses (134). An observation from the experimentally infected chickens was that high ER-specific titers were frequently observed in individuals that had undergone subclinical infections (143).

Moreover, a recurring observation from serologic field studies was an increased occurrence of positive individuals as well as increasing ER-specific titers in older animals (141,143,144). Considering that outbreaks of erysipelas frequently occur in older layer flocks, it seems that seropositive animals can succumb to the disease. High titers of IgY recognizing ER were indeed observed in layers sampled during acute outbreaks of clinical erysipelas (143). Thus, it seems like these ER-recognizing antibodies confer no or very limited protection against systemic ER infection.

Conversely as discussed below, in vaccine-induced immunity against ER infection, antibodies are generally considered crucial (33,50,135). Nonetheless, while it induces no or very low IgY titers to ER prior to challenge, an inactivated ER vaccine conferred protection against experimental ER infection of young SPF-reared laying hen hybrids (152).

Apart from issues regarding antibody production, very little research on other antigen-specific immune mechanisms in ER immunity has been reported. Considering that ER infections may comprise intracellular phases, one can speculate that Th1-type T-cell responses and interferon (IFN)-γ-mediated enhancement of phagocytic killing would be of importance (168). This is, for instance, the case for clearance of Salmonella by the chicken immune system, a bacterium that infects macrophages (157). Studies of T cells in ER infection are very few, but in pigs, T-cell activation measured as ER-antigen-induced proliferation of blood mononuclear cells has been observed after vaccination and infection (172,173). In addition, an association between vaccine-induced in vitro proliferative responses and better protection after ER challenge infection of pigs was reported (173). ER-specific T-cell proliferation upon in vitro recall stimulation of spleen cells from chickens 18 days after experimental ER infection has been observed, and the proliferating cells were identified as CD4+ (i.e., T-helper cells), TCRγ/δ-CD8αβ+ (i.e., cytotoxic T-cells [CTL]), and TCRγ/δ+CD8αβ+ T-cells (134). The observed activation of CTL suggests that a Th1-type response was induced in these chickens. Likewise, after ER infection of Atlantic bottlenose dolphins (Tursiops truncatus), antigen-specific IFN-γ mRNA expression was observed upon in vitro ER antigen stimulation of blood mononuclear cells, which is also an indication of a Th1-type response (174). Interestingly, ER-antigen-specific activation of TCRγ/δ+CD8αβ+ T-cells upon in vitro recall stimulation of chicken spleen cells was also observed (134). This cell population is a subtype of so-called γ/δT-cells that constitutes an “unconventional” T-cell type, which in mammals has a role in specific immunity against bacterial pathogens (175,176). The role of γ/δT-cells in the chicken immune system is still under investigation, but, so far, they have been implied to have a role in the innate response to Salmonella infection (177,178,179).

Taken together, specific antibodies to ER may aid in phagocytosis and killing of the bacterium, but their role in protective immunity remains unclear, and seropositivity per se seems insufficient as a correlate for protection against systemic disease in laying hens. There is evidence that Th1-type responses and IFN-γ-mediated functions might be important in ER infection, and further research into ER immunity in chickens and development of effective vaccines against erysipelas should focus on elucidating these factors.

Outbreaks of erysipelas do not normally cease spontaneously, and, therefore, with regard to both animal welfare and economic aspects, prompt action should be taken in affected flocks. Due to the acute nature of the disease and the fact that diseased and dead birds may contain large numbers of ER bacteria, it is strongly recommended that dead carcasses should be collected and removed as frequently as possible from the barn, at least several times a day, to minimize the spread within the flock. Due to the zoonotic potential of the bacterium, protective gloves should be worn whenever live birds and carcasses are handled. Biosecurity measures are essential to prevent spread to nearby flocks.

Penicillin has been the drug of choice for treatment of ER infections for a long time (180). Few isolates from layers have however been investigated for minimal inhibitory concentration values. Previous studies, including isolates from other animal species, have indicated that ER isolates are still susceptible to beta-lactam antibiotics (i.e., ampicillin and penicillin) (22,181,182). Recently, indications on penicillin resistance among European ER isolates from poultry have been published (183,184). Updated studies on a significant number of isolates from more countries and several hosts are needed to investigate the current situation of antimicrobial resistance among ER isolates, in particular, in consideration of the unexplained difference in resistance to the beta-lactam antibiotics penicillin G and ampicillin that was reported in recent studies (183).

Antibiotic treatment is effective in reducing the mortality in a layer flock, but disease often reoccurs upon suspension of treatment (86,112) because ER is still present in the barn environment. To stop an ongoing outbreak, antibiotic treatment should therefore preferably be combined with vaccination (see below). Also, in case there is no beta-lactam antibiotic available, or if the beta-lactam antibiotic available has no withdrawal time for eggs, or if the withdrawal time for eggs is too long, vaccination alone may be a suitable measure to stop an outbreak.

Vaccination against ER has a long history (33) and has predominantly been used as a prophylactic measure to prevent disease in high-risk species and situations. Similar to many other bacterial vaccines, vaccine-induced immunity against ER infection is considered to be dependent on antibodies (33,50,135). It has been shown that vaccination of layers with a commercial inactivated ER vaccine for pigs induced antibody production (185). Under field conditions, IgY titers to ER were monitored during the production period of conventional Swedish layer flocks vaccinated once at placement at 14–16 wk of age (144). In that cohort, the levels of IgY to ER were significantly higher in hens vaccinated against erysipelas compared to age-matched unvaccinated hens, and this difference was most prominent for the youngest age category tested, i.e., 35 wk of age. It thus seemed that vaccination induced an earlier ER-antibody response compared to that naturally evoked in this environment. However, all vaccinated flocks were housed in facilities where outbreaks of erysipelas had occurred on some occasion prior to the study. Therefore, it cannot be completely ruled out that the increased ER antibody levels in the vaccinated flocks were induced by a higher environmental presence of ER on these farms compared to the farms where vaccination was not performed. Experimentally, it has been shown that vaccination with an inactivated ER vaccine confers protection against ER challenge infection in young chickens despite not inducing high levels of ER-specific IgY prior to challenge (152).

There is currently no vaccine registered for use in chickens. An inactivated vaccine against erysipelas licensed for turkeys and pigs has therefore often been used in layers in Europe (Sweden [129] and Denmark [S. Kabell, pers. comm.]). In a few countries, e.g., the United States, a live attenuated vaccine developed for turkeys might be available (130). According to Crespo et al. (130), information on the effectiveness of this vaccine is limited. These authors reported on the use of an attenuated live ER vaccine for pigs during ER outbreaks in three flocks of layers and described a 90% reduction of mortality after administration of two doses via drinking water (130). Autogenous vaccines, i.e., vaccines based on farm-specific isolates and manufactured for use on the particular farms of origin, may also be used if there is no vaccine available or if there is a perceived lack of protection by other vaccines (33).

For many ER vaccines, the manufacturers recommend that vaccines should be administered twice to provide full protection. In addition, the vaccines registered for use in turkeys should be administered subcutaneously, which requires training and is more time-consuming than intramuscular injection. Hence, vaccination with inactivated vaccines requires large numbers of layers to be caught and handled, potentially twice, which will be a laborious procedure in flocks on litter, especially in aviaries.

As previously mentioned, in case there is no appropriate antibiotic treatment available, vaccination alone may be a suitable measure to stop an outbreak. Depending on the age of the flock at the time of the outbreak (remaining time in egg production), vaccination may be economically justifiable. After vaccination, mortality decreases gradually and returns to normal levels after approximately 2 wk (115).

Vaccination of subsequent flocks on affected farms is strongly recommended, since ER may persist both in the poultry house and the environment, despite thorough cleaning and disinfection. In addition, there is a risk of ER being reintroduced from an environmental source. Flocks in other houses on the farm should also preferably be vaccinated. Due to the labor-intensive vaccination procedure in layer flocks in litter-based systems, these flocks are usually vaccinated only once, i.e., at placement. This is usually effective, but for some farms, a second dose may be necessary (115).

If antibiotic treatment and/or vaccination are not viable options for the affected flock, e.g., old flocks with high mortality, euthanasia may be the sole remaining option to prevent animal suffering.

In conclusion, more work on antibiotics suitable for treatment of flocks during egg production, including resistance patterns, is needed together with work on vaccines that can be easily mass administered, e.g., live attenuated vaccines, during outbreaks and to flocks at risk.

Despite being known as a disease-causing agent in mammals for 150 years, and almost as long in poultry, many important aspects of erysipelas, not least in layers, are still unknown. Continued research into the pathogenesis, immunologic aspects, and, perhaps most importantly, effective prophylactic measures for flocks at risk is therefore urgently needed.

The authors’ ER work has been financially supported by the Swedish Research Council Formas (grant numbers 942-2015-00766 and 2019-01270), the Animal Health and Welfare ERA-Net (ANIHWA) under the European Union Seventh Framework Network (ID number 119, in Sweden grant number 221-2015-01895), the Swedish Board of Agriculture, the European Union’s Horizon 2020 research and innovation program under grant agreement no. 731014 (VetBioNet project; a transnational access grant), and the Albert Hjärre foundation. We wish to thank all our research collaborators, especially Tina Sørensen Dalgaard for invaluable support.

CTL=

cytotoxic T-cells;

ER =

Erysipelothrix rhusiopathiae;

ESB =

Erysipelothrix selective broth;

IFN =

interferon;

IgY =

immunoglobulin Y;

MBL =

mannose binding lectin;

MHCII =

major histocompatibility complex class II;

MRC1L-B =

mannose receptor 1 like-B;

PFGE =

pulsed-field gel electrophoresis;

PRM =

poultry red mite;

Spa =

surface protective antigen;

SPF =

specific pathogen free;

TCR =

T-cell receptor;

Th1 =

T-helper 1