https://orcid.org/0000-0002-2513-8538
ABSTRACT
The study aim was to determine the best inoculation route for virulotyping Enterococcus cecorum in a chicken embryo lethality assay (ELA). Twenty-eight genetically different strains were used. Fourteen strains were isolated from cloaca swabs of broiler reproduction chickens (cloaca strains) and 14 strains from broilers with E. cecorum lesions (lesion strains). In all ELAs, 12-day incubated embryonated broiler eggs were inoculated with approximately 100 colony-forming units of E. cecorum/egg. Twenty embryos per inoculation route and strain were used in each of three experiments. In Experiment 1, four cloaca and four lesion strains were inoculated via various routes, i.e. albumen, amniotic cavity, allantoic cavity, chorioallantoic membrane, intravenous or air chamber. The albumen inoculation route showed low mortality with cloaca strains, high mortality with lesion strains and the largest difference in mortality between these groups of strains (≥60%). This route was therefore used in subsequent experiments. In Experiment 2, the same strains were used to test reproducibility, which proved to be generally good. All 28 strains were thereafter used in Experiment 3. In the three experiments, mortality caused by cloaca and lesion strains ranged from 0-25% and from 15-100%, respectively. Recovery rates, assessed in all experiments after albumen inoculation, were significantly lower from eggs inoculated with cloaca strains, compared to lesion strain-inoculated eggs (P < 0.05). However, the bacterial load of eggs with positive recovery was similar in both groups. In conclusion, the albumen inoculation route appeared to be the best to virulotype E. cecorum strains.
RESEARCH HIGHLIGHTS
The albumen route is the best to differentiate between E. cecorum strains.
Egg albumen likely affects cloaca E. cecorum strains more than lesion strains.
Based on SNPs, E. cecorum cloaca strains are clustered as well as lesions strains.
Introduction
Enterococcus cecorum is considered a normal inhabitant of the intestines of older chickens (Devriese et al., Citation1991). In more recent studies, it was found that E. cecorum can colonize the intestines of broilers as early as the first week of life. Early colonization was linked with E. cecorum outbreaks later in the production cycle (Borst et al., Citation2017; Jung, Petersen et al., Citation2017). Outbreaks of E. cecorum are characterized by lameness, increased mortality and rise in condemnation rates. Diseased chickens show typical lesions such as osteomyelitis of the proximal femur and the 6th thoracic vertebra (T6), arthritis and pericarditis (Herdt et al., Citation2009; Kense & Landman Citation2011; Borst et al., Citation2012; Robbins et al., Citation2012; Jung & Rautenschlein Citation2014; Borst et al., Citation2017; Jung, Petersen et al., Citation2017). Crucial steps in the pathogenesis of E. cecorum infections are the colonization of the intestines followed by bacteraemia, which subsequently can lead to the aforementioned E. cecorum lesions (Borst et al., Citation2017).
The T6 free thoracic vertebra is sometimes also denoted as T4, depending on whether vertebrae with floating ribs are considered as thoracic vertebrae or not. In the former case seven thoracic vertebrae exist in the chicken, in the latter case that number is reduced to five (King, Citation1975).
In vivo typing of bacterial strains is essential when differences in virulence between strains are investigated. Ideally, in vivo typing should be performed in an infection model using the animal species in which the disease occurs naturally. However, these models are time-consuming and expensive. Moreover, regarding E. cecorum, clinical disease and lesions have to date only been induced in a very low percentage of experimental chickens (Martin et al., Citation2011; Borst et al., Citation2017; Schreier et al., Citation2021), making these models unsuitable due to the very large number of experimental birds required to obtain meaningful differences.
Another effective screening method to assess the virulence of bacteria is the embryo lethality assay (ELA), as some studies have shown for Escherichia coli and Riemerella anapestifer (Nolan et al., Citation1992; Gibbs & Wooley Citation2003; Seo et al., Citation2013; Landman et al., Citation2021). In vivo typing of E. cecorum strains in an ELA has revealed significant differences in embryonic survival between strains after inoculation in to the allantoic cavity of 10–12-day-old chicken embryos (Borst et al., Citation2014; Jung, Metzner et al., Citation2017; Dolka et al., Citation2022). However, no clear cut-off values could be established to classify E. cecorum strains either as commensal/opportunistic or as virulent. In these studies, only the allantoic cavity route was used. Manders et al., (Citation2021) showed that embryonated chicken eggs can be successfully inoculated via various routes. Several of these routes (intravenous, allantoic cavity and chorioallantoic membrane (CAM)) have been previously applied to biotype different bacterial species (Bumgarner & Finkelstein Citation1973; Gibbs & Wooley Citation2003; Nix et al., Citation2006; Borst et al., Citation2014).
The aim of the present study, was to examine which inoculation route in a chicken ELA resulted in the best differentiation in virulence between E. cecorum strains from typical lesions in broilers (lesion strains) and E. cecorum strains from the cloaca of healthy broiler reproduction chickens (cloaca strains). Hereto, 14 lesions strains and 14 cloaca strains were selected. Three experiments were performed. (1) The embryonic mortality and mean death time (MDT) were assessed after inoculation of each of four lesion strains and four cloaca strains, each by six different inoculation routes (Experiment 1). (2) Research as described under 1, but inoculations were performed only by the albumen route based on the results of the first experiment (Experiment 2A). Furthermore, the bacterial load of the amniotic fluid was assessed at different days after albumen inoculation (Experiment 2B). (3) Each of the 28 E. cecorum strains was inoculated in to the albumen of embryonated broiler eggs, and mortality and MDT were determined (Experiment 3). Moreover, recovery rates of E. cecorum from eggs with dead and surviving embryos of experiments 1 and 3 were assessed.
Materials and methods
E. cecorum strains
Isolation, identification and selection
Cloaca strains. Cloaca swabs (REF 155C, Copan, Brescia, Italy) of 10 birds of each of 24 flocks of healthy broiler reproduction chickens were collected and submitted for bacterial examination. Swabs were plated on both, Columbia agar with sheep blood (CBA) (Oxoid, Wesel, Germany) and Kanamycin Aesculin Azide agar (KAA) plates (Oxoid), which were incubated at 37°C in a 5% CO2-enriched atmosphere. After 24 and 48 h of incubation plates were examined and colonies were morphologically compared to a reference E. cecorum strain, which had been identified by using a tRNA intergenic spacer PCR (Kense & Landman, Citation2011). In cases when colony morphology on both plates inoculated with the same swab was similar to the reference strain, one colony was selected. A small fraction of named colony was submitted for species determination by matrix-assisted laser desorption ionisation-time of flight mass spectrometry (MALDI-TOF MS) technique (Bruker MALDI Biotyper, Bruker Daltonics, Bremen, Germany). All strains used in this study were confirmed to be E. cecorum. The remainder of the same colony was inoculated on a CBA plate and incubated overnight as described before. The colonies were scraped off the plate and suspended in nutrient broth with a cryoprotectant (glycerol) (Tritium, Eindhoven, Netherlands) and cryopreserved at −80°C.
Lesion strains. Broilers, originating from flocks showing lameness and increased mortality, were submitted to Royal GD, Deventer, Netherlands and the Faculty of Veterinary Medicine, Utrecht University, Utrecht, Netherlands for necropsy and bacteriological examination. In case typical E. cecorum lesions were present, i.e. osteomyelitis of the proximal femur and/or in the 6th thoracic vertebra, pericarditis and/or arthritis, swab samples of the lesions were collected. Bacteriological examination of the above samples, identification of colonies from pure cultures only, and cryopreservation were performed as described under “cloaca strains”.
Selection of strains. Commensal E. cecorum bacteria vary in their ability to metabolize mannitol, while virulent E. cecorum strains are unable to do so (Borst et al., Citation2012; Dolka et al., Citation2017; Jung, Metzner et al., Citation2017). Therefore, in this study, only E. cecorum strains from cloaca swabs able to metabolize mannitol (mannitol-positive, avirulent/low virulent strains) and strains from lesions unable to metabolize mannitol (mannitol-negative, likely virulent strains) were selected: 14 cloaca strains and 14 lesion strains. Identification and characteristics of these 28 strains are presented in .
Table 1. Characteristics of Enterococcus cecorum strains used in the study.
Mannitol metabolism. The ability of strains to metabolize mannitol was assessed as follows: strains were cultured on CBA plates overnight. With a sterile inoculation loop (VWR, Radnor, PA, USA) one colony was inoculated in to a tube with Andrade peptone water and mannitol (Biotrading, Mijdrecht, Netherlands,) and incubated at 37°C in a 5% CO2-enriched atmosphere. After 48 h the colour of the suspension was evaluated: pink suspensions were considered mannitol-positive, and colourless to light straw coloured suspensions as mannitol-negative.
Confirmation of species and assessment of clonality
Whole genome sequencing
All strains were analysed by Next Generation Sequencing for confirmation of species and assessment of clonality on whole genome level. The DNA extraction was performed by BaseClear (Leiden, Netherlands) and was used for the preparation of libraries using the Nextera XT DNA library preparation kit (Illumina, San Diego, CA, USA). 300-cycle (2 × 150-bp paired-end) sequence reads were generated by BaseClear using the Illumina NovaSeq 6000 system. FASTQ read sequence files were generated using bcl2fastq version 2.20 (Illumina). Initial quality assessment was based on data passing the Illumina Chastity filtering. Subsequently, reads containing PhiX control signal were removed using an inhouse filtering protocol BaseClear. In addition, reads containing (partial) adapters were clipped (up to a minimum read length of 50 bp). The second quality assessment was based on the remaining reads using the FASTQC quality control tool version 0.11.5. De-multiplexed and adapter clipped reads were analyzed using the CLC Genomics Workbench version 21.0.5. (Qiagen Aarhus A/S, Aarhus, Denmark).
De novo assembly was performed using the following parameters: mapping mode = create simple contig sequences (fast), automatic bubble size = yes, minimum contig length = 500, automatic word size = yes, perform scaffolding = yes, auto-detect paired distances = yes, guidance only reads = no, min distance = 1, max distance = 1000.
Reference assembly was performed using the genome of E. cecorum strain NCTC 12421 (LS483306) as reference sequence. The script map reads to reference parameters: v.1.8. within the CLC Genomics workbench was used. The parameters for mapping the reads were: masking mode = no masking, match score = 1, mismatch cost = 2, cost of insertions and deletions = affine gap cost, insertion open cost = 6, insertion extend cost = 1, deletion open cost = 6, deletion extend cost = 1, length fraction = 0.5, similarity fraction = 0.8, global alignment = yes, auto-detect paired distances = yes, non-specific match handling = map randomly. Local realignment was performed using the following parameters: realign unaligned ends = yes, multi-pass realignment = 2. From the realigned reference assemblies consensus sequences were extracted using the following parameters: threshold = 15, action = remove regions with low coverage, post-remove action = join after removal, conflict resolution strategy = vote, use quality score = no, add consensus annotations (conflicts, indels, low coverage etc.) = no.
Confirmation of species previously identified with MALDI-TOF MS
Genetic species confirmation was performed by ribosomal multilocus sequence typing (rMLST). This method indexes variation of the 53 genes encoding the bacterial ribosome protein subunits (rps genes). The rps gene variation catalogued in this curated database permits rapid identification of the phylogenetic position of any bacterial sequence at the domain, phylum, class, order, family, genus, species and strain levels (Jolley et al., Citation2012, Citation2018). FASTA files containing de novo contigs from each strain included in this study were uploaded and submitted to the online available rMLST identify species tool at https://pubmlst.org/bigsdb?db=pubmlst_rmlst_seqdef_kiosk.
Assessment of clonality
The realigned reference contigs were also used for variant calling. Therefore the script Basic Variant Detection was used with the following parameters: ploidy = 2, ignore positions with coverage above = 5.000, restrict calling to target regions = not set, ignore broken pairs = yes, ignore non-specific matches = reads, minimum coverage = 15, minimum count = 10, minimum frequency = 35.0%, base quality filter = yes, neighbourhood radius = 5, minimum central quality = 20, minimum neighbourhood quality = 15, read direction filter = no, relative read direction filter = no, read position filter = no, remove pyro-error variants = no, create track = yes, create annotated table = yes. From the variant calling analysis the number of pairwise SNP differences was determined and subsequently a SNP tree was constructed using the following parameters: include multi-nucleotide variances = no, minimum coverage required in each sample = 20, create report = yes, minimum coverage percentage of average required = 15, prune distance = 0, minimum z-score required = 0.0, ignore positions with deletions = no, tree view settings = SNP tree default, tree construction algorithm = neighbour joining. No criteria are available to determine the genetic relatedness for E. cecorum bacteria based on pairwise SNP differences and neither is the mutation rate known.
Pinholt et al. (Citation2015) reported a cut-off value of 17 SNPs difference for clonality of E. faecium strains. Therefore, in this study, we used the same cut-off for E. cecorum strains.
Embryo lethality assay
Rationale and design of experiments
In Experiment 1, 12-day incubated embryonated eggs were inoculated with four lesion strains, four cloaca strains or a placebo via six routes resulting in nine (number of strains tested plus one placebo) × six (number of inoculation routes) = 54 groups of eggs. Inoculation sites were: albumen, air chamber, amniotic cavity, allantoic cavity, CAM and a vein. These sites were chosen as their inoculation previously showed a success percentage of 100 in 12-day incubated embryonated eggs, except for the amniotic cavity, which inoculation had a success rate of 95 percent. The yolk and embryo inoculation routes were not used as their success rates were considered to be too low (Manders et al., Citation2021).
In Experiment 2A, the reproducibility of the results obtained in Experiment 1 was examined. Hereto, the same strains as in Experiment 1 were used. However, only the albumen inoculation route was applied as this route had shown in Experiment 1 to best differentiate between cloaca and lesion strains.
The albumen and bacteria herein reach the amniotic cavity and subsequently the embryo by the sero-amniotic connection (Deeming, Citation1991; Manders et al., Citation2021). However, the albumen might have detrimental effects on E. cecorum strains, which may vary between strains. Therefore, in Experiment 2B the bacterial load of the amniotic fluid was determined after albumen inoculation.
Eggs used in experiments 2A and 2B belonged to the same batch. Also, the same inocula were used for experiments 2A and 2B. Each of these experiments consisted of nine groups of eggs.
Experiments 1 and 2A, showed that, generally, reproducibility of results obtained by the albumen route was good. Therefore, in Experiment 3 all 28 selected E. cecorum strains were examined for virulence by the albumen inoculation route. Including the placebo group, 29 groups of eggs were used in this experiment.
In experiments 1, 2A and 3, 20 12-day incubated eggs with live embryos were inoculated per group; in Experiment 2B, 21 eggs per group. Eggs were candled daily, up to and including day 6 after inoculation (18th day of incubation), mortality was recorded and the MDT (days) calculated.
In experiments 1 and 3, bacteriological examination was performed for eggs with embryos that died after inoculation and for eggs with surviving embryos at day 18 of incubation.
Hatching eggs
Hatchings eggs were obtained from one flock of Ross 308 broiler breeders. Eggs were collected when the birds were 43, 47 and 48 weeks of age for experiments 1, 2 and 3, respectively. They were incubated at 37.5°C, 53% relative humidity and turned every hour at the Faculty of Veterinary Medicine of Utrecht University.
Inocula
E. cecorum strains were recovered from cryopreservation, inoculated on to CBA plates and incubated as described under “E. cecorum strains”. Colonies were scrapped off the plate with a dry swab (Copan, Brescia, Italy) and suspended in physiological saline (Tritium) to an optical density of 0.5 McFarland (McFarland Densitometer, type DEN-1, Grant instruments Ltd., Shepreth, England) which corresponds with a concentration of approximately 108 colony-forming units (CFU)/ml. Subsequently, these suspensions were serial diluted in peptone physiological saline (PPS) (BioTrading, Mijdrecht, Netherlands) aiming at a concentration of 500 CFU/ml of which 0.2 ml was inoculated per egg, resulting in a dose of 100 CFU/egg. The latter dose is in agreement with doses used in previously published E. cecorum ELA studies (Borst et al., Citation2014; Jung, Metzner et al., Citation2017). Inocula were stored and transported on melting ice and E. cecorum concentrations were assessed by bacterial counting immediately after inoculation (Pakpour & Horgan, Citation2021).
Inoculations
Inoculations were performed according to previously described procedures (Manders et al., Citation2021). Briefly, the inoculations, except the intravenous and CAM inoculations, were performed via a small hole in the eggshell (diameter ± 1 mm) made with an electric engraver (Hugo Brennenstuhl GmbH & Co Kommanditgesellschaft, Electric Engraver Signograph 25 Set, article number 1500740, Tübingen, Germany). The air chamber and amniotic cavity were inoculated from the centre of the blunt end of the egg and the albumen from the centre of the apex. Injections were done following the central axis of the egg. The allantoic cavity was inoculated via a hole made in the eggshell at approximately 2 mm below the air chamber, whereby the needle was injected parallel to the long axis of the egg. Needles of 25G and 6, 16, 16 and 38 mm long were used for the inoculations in the air chamber, allantoic cavity, albumen and amniotic cavity, respectively. Needles were completely inserted into the egg. During the procedures, the eggs were always positioned with the blunt end upwards.
The intravenous inoculations were performed after a triangle with sides of approximately 4 mm long on top of a large blood vessel of the CAM was carved in the eggshell with the electric engraver and subsequently removed with a blunt needle, leaving the shell membranes intact. While candling the egg, the blood vessel was injected with a 30G × 13 mm needle. Disappearance of blood in the vessel during inoculation, and the return of blood flow thereafter, were indicators of a successful inoculation.
The CAM inoculation procedure started by making a small hole in the eggshell and outer shell membrane at the centre of the blunt end of the egg. Subsequently, on a site halfway along the length of the egg, without visible major blood vessels on the CAM, a small hole was made in the eggshell. The shell membrane was perforated with a blunt needle while leaving the underlying CAM intact. A squeezed rubber pipette bulb was placed at the hole at the blunt end of the egg. By releasing the pressure on the rubber pipette bulb negative pressure was created resulting in an artificial air chamber on top of the CAM, which was then inoculated with a 25G × 6 mm needle.
Prior to inoculations, the eggshell around the intended inoculation site was disinfected with an iodine solution (Povidone iodine, Mylan Healthcare BV, Amstelveen, Netherlands). After inoculation, the inoculation site was sealed with sterile adhesive tape.
Recovery of E. cecorum from amniotic fluid after inoculation in the albumen
In Experiment 2B amniotic fluid of three embryonated eggs was collected daily per group of eggs from the day of inoculation up to and including day 6 after inoculation. Amniotic fluid was sampled only from eggs with live embryos as samples from eggs with dead embryos were often contaminated with other embryonic fluids. The amniotic fluid was collected after the embryonated eggs were gently tapped with forceps halfway along the longitudinal axis of the egg after which the egg was broken into two halves. The embryo and the extraembryonic compartments were carefully poured into a sterile Petri dish, leaving the amnion intact. At least 0.2 ml of amniotic fluid was collected per egg with a syringe fitted with a 21 G × 16 mm needle. Bacterial concentrations of the fluids were assessed by bacterial counts. Hereto, 0.1 ml was inoculated on each of two CBA plates, which were incubated as described under “E. cecorum strains”. After 24 h the number of colonies was counted on both plates and the mean number was calculated. Identification of colonies was performed as described under “E. cecorum strains”. A maximum of 250 colonies were countable on a plate, resulting in a detection range between 1 and 250 CFU/0.1 ml amnion fluid.
Recovery of E. cecorum from the content of eggs with dead and surviving embryos
Eggs with dead and surviving embryos on day 18 of incubation in experiments 1 and 3 were sampled for bacteriological analysis. Except for a few cases, all dead embryos to a maximum of five per group of eggs were examined. If less than five dead embryos were examined, live embryos were included to reach the maximum of five embryos. The number of eggs with dead and live embryos bacteriologically examined per group is given in . Eggs with dead embryos were sampled on the day that death was ascertained. Swabs were taken via an opening made in the blunt end of the eggs and processed as described under “E. cecorum strains”.
Statistics
Mortalities and recovery rates of E. cecorum were compared statistically using Fisher’s exact test. Bacterial loads of amniotic fluid were statistically analysed using the non-directionally Mann–Whitney U test (Statistix®, Citation2010). Statistically significant is hereafter referred to as significant. Differences were considered significant if P < 0.05.
Ethics
Embryonated eggs are not classified as experimental birds under Dutch law (Dutch Animal Procedure Act (Wet op de dierproeven)).
Results
E. cecorum strains
Isolation, identification and selection
Cloaca strains. Seventy-eight E. cecorum strains were obtained from cloaca swabs taken from birds belonging to 24 broiler reproduction flocks. In 23 flocks, 1–8 out of 10 swabs contained E. cecorum. Fifty percent (39/78) of E. cecorum strains metabolized mannitol. In 17/24 (71%) flocks at least one mannitol-positive E. cecorum strain was isolated. One mannitol-positive strain from each of 14 flocks, which were randomly chosen, was selected for the present study ().
Lesion strains. Fourteen mannitol-negative E. cecorum strains used in the present study were isolated as pure culture from separate broiler flocks. Isolations were performed from typical E. cecorum lesions. Characteristics of these strains are given in .
All strains were identified as E. cecorum by MALDI-TOF MS.
Confirmation of species and assessment of clonality
All bacterial strains were confirmed as E. cecorum with the rMLST species identification tool. No hits for other species were found.
In total 137,738 single nucleotide polymorphisms (SNPs) were informative and used for the construction of the SNP tree (). This SNP tree showed that there was no clonal relationship between the draft genome sequences of all 28 strains. For the majority of strains the number of pairwise SNP differences was very high (mean SNP difference of 26,259). Lesion strains differed at least 85 SNPs from each other (strains 21 and 24), while cloaca strains differed at least 510 SNPs (strains 6 and 9). All strains were therefore regarded as not clonally related as the number of SNPs difference between strains was more than 17.
Figure 1. Single nucleotide polymorphism (SNP) phylogenetic tree based on whole genome sequencing of 28 Enterococcus cecorum strains. The difference in number of SNPs between strains was at least 85 (between strains 21 and 24), and as a cut-off of 17 SNPs was used (see “Materials and methods” [under “Assessment of clonality”] and “Discussion”), all strains were considered non-clonal. Cloacal strains, excluding strain 4, form cluster I and lesion strains cluster II. The average SNP difference between strain 4 and cluster I is 90,607 and between strain 4 and cluster II 90,033. Samples are numbered according to .
![Figure 1. Single nucleotide polymorphism (SNP) phylogenetic tree based on whole genome sequencing of 28 Enterococcus cecorum strains. The difference in number of SNPs between strains was at least 85 (between strains 21 and 24), and as a cut-off of 17 SNPs was used (see “Materials and methods” [under “Assessment of clonality”] and “Discussion”), all strains were considered non-clonal. Cloacal strains, excluding strain 4, form cluster I and lesion strains cluster II. The average SNP difference between strain 4 and cluster I is 90,607 and between strain 4 and cluster II 90,033. Samples are numbered according to Table 1.](/cms/asset/8507b207-afc2-4231-b55e-5f9fc0522c4b/cavp_a_2130174_f0001_ob.jpg)
In the SNP tree two different clusters were observed. Cluster I consisted of all cloaca strains except cloaca strain 4. Within cluster I the lowest number of pairwise SNP differences was 510 between cloaca strains 6 and 9. The highest number of SNP differences was 16,525 between cloaca strains 9 and 14. The average number of pairwise SNP differences for cluster I was 13,920. Cluster II consisted of all lesion strains. Within cluster II the lowest difference in number of SNPs was found between lesion strains 21 and 24 (85 SNPs). The highest difference was found between lesion strains 16 and 19 (26,456 SNPs). The average number of SNP differences in cluster II was 10,209. Between cluster I and cluster II an average of 21,572 SNP differences was found, indicating high genetic diversity. Cloaca strain 4 appeared to be genetically distinct from cluster I (average difference 90,607 SNPs) as well as from cluster II (average difference 90,033 SNPs).
Raw sequence reads were uploaded to the National Centre for Biotechnology Information Sequence Read Archive and BioProject PRJNA853961. References to BioSample numbers for individual samples can be found in .
Embryo lethality assay
Results of ELAs performed in Experiments 1, 2A and 3 as well as inoculation doses used in these experiments are presented in . Mean dose in CFU/egg (range) was 77 (36–125) in Experiment 1, 131 (24–215) in experiments 2A and 2B, and 89 (21–168) in Experiment 3. Placebo groups were inoculated with 0.2 ml PPS/egg.
Table 2. Broiler embryo mortality and mean death time (MDT) up to and including day 6 after inoculation of Enterococcus cecorum strains on day 12 of incubation. Three experiments were performed.
Placebo inoculations did not result in embryo mortality in any of the experiments, except for a few cases in CAM- and air chamber-inoculated groups in Experiment 1 due to mechanical damage.
Experiment 1. Differentiation between cloaca strains (strains 1–4) and lesion strains (strains 15–18) could not be made by means of air chamber inoculation as embryo mortality did not occur with any strain. However, clear differentiation could be made using all other inoculation routes, i.e. the albumen, the allantoic cavity, the amniotic cavity, the CAM and the intravenous route. Mortality induced by the lesion strains was higher than that obtained by the cloaca strains, for each of the aforementioned inoculation routes. Overlap in mortality was not observed. The gap between mortality in the groups inoculated with lesion strains and those inoculated with cloaca strains, i.e. the difference in mortality between the highest mortality score in the groups inoculated with the cloaca strains and the lowest mortality score in the groups inoculated with the lesion strains was largest after inoculation in the albumen (60%), followed by CAM application (47%), amniotic cavity and intravenous injection (each 10%), and allantoic cavity inoculation (5%). Each of the cloaca strains inoculated in the amniotic cavity induced higher mortality compared to the mortality induced by the same strains inoculated in the albumen at the same doses. Differences ranged from 10-55%. The aforementioned finding was not or hardly observed when lesion strains were used (differences of 0–11%).
MDTs ranged from 1-4.6 days depending on E. cecorum strain and inoculation route. The shortest MDTs (1–2.2 days) were observed after inoculation of lesion strains by the intravenous and amniotic cavity routes, and after CAM application of both lesion and cloaca strains. The longest MDTs (3.1–4.1 days) were noticed after albumen inoculation with both sets of strains.
Experiment 2A. This experiment in which albumen inoculations were performed showed good reproducibility of results of Experiment 1: no significant differences in mortality compared to Experiment 1 were found except for strains 16 and 17 (lesion strains). Regarding these strains, relatively large differences existed in E. cecorum dose between both experiments (strain 16: 125 CFU/egg in Experiment 1 versus 24 CFU/egg in Experiment 2A; strain 17: 36 CFU/egg in Experiment 1 versus 108 CFU/egg in Experiment 2A). Nevertheless, also in this experiment a clear differentiation between cloaca strains (strains 1–4) and lesion strains (strains 15–18) could be made with the albumen inoculation route. Mortality in groups inoculated with cloaca strains ranged from 0 to 25%, in groups inoculated with lesion strains from 30 to 100%.
MDTs were similar to those observed in Experiment 1, except for strains 15 and 16. The latter strains gave MDTs of 5.1 and 5.5 days, respectively, while in Experiment 1 MDTs of 3.3 and 3.1 days, were found, respectively.
Experiment 3. In this experiment, all 28 E. cecorum strains (cloaca strains 1–14; lesion strains 15–28) were inoculated in the albumen of fertile hatching eggs. Twelve cloaca strains did not induce embryo mortality, while two cloaca strains, strains 2 and 11, gave 10 and 15% mortality, respectively. Twelve lesion strains produced mortalities between 80 and 100%. Strains 16 and 23 induced 45% and 15% mortality, respectively. Overlap in mortality caused by cloaca and lesion strains was found once: strains 11 (cloaca strain) and 23 (lesion strain) both gave 15% mortality. Mortalities induced by strains 4 and 16 were significantly different from the results obtained with these strains in experiments 2A and 1, respectively. Relative large differences in dose of these strains between these experiments were present.
MDTs after inoculation of lesion strains ranged from 4.0 to 5.7 days, which were similar to MDTs found in Experiment 2A.
Recovery of E. cecorum from amniotic fluid after inoculation in the albumen
Experiment 2B. Results are presented in . Recovery rates increased with time after inoculation in groups of eggs inoculated with cloaca strains (strains 1–4) as well as in groups inoculated with lesion strains (strains 15–18), resulting in maximum positive scores at 3 days post-inoculation (dpi) (7/11 in cloaca strain groups and 10/10 in lesion strain groups). From 3 dpi onwards, E. cecorum was recovered from all samples of the lesion strain groups, while this was not the case for samples of the cloaca strain groups (5/12, 6/11 and 2/6 at 4, 5 and 6 dpi, respectively). In the period from the day of inoculation up to and including day 6 pi the total number of positive amniotic fluid samples differed significantly between cloaca strain groups (26/75 = 35%) and lesion strain groups (33/52 = 63%).
Table 3. Experiment 2B. Recovery of Enterococcus cecorum from amniotic fluid of live embryos after inoculation of the bacterium into the albumen at day 12 of incubation.
Mean bacterial load of all positive samples and the number of positive samples with >250 CFU/0.1 ml in cloaca strain groups and lesion strain groups, were >205 CFU/0.1 ml and 19/26 (73%) and >222 CFU/0/1 ml and 29/33 (88%), respectively. Differences were not significant.
Recovery of E. cecorum from the content of eggs with dead and surviving embryos
Results are presented in . E. cecorum was not isolated from placebo-inoculated eggs of experiments 1 and 3, and from eggs of Experiment 1 inoculated in the air chamber with E. cecorum strains, which will therefore not be mentioned further in the results below.
Table 4. Recovery of Enterococcus cecorum from eggs with dead and live embryos from experiments 1 and 3 after inoculation of this bacterium on incubation day 12.
Experiment 1. E. cecorum was recovered from all samples of eggs inoculated with lesion strains by either inoculation route, except one sample (a sample of one egg with a live embryo, inoculated with strain 16 in the allantoic cavity). Samples were predominantly taken from eggs with a dead embryo. Sampled eggs inoculated with cloaca strains (strains 1-4) consisted of approximately 50% eggs with a dead embryo. Intravenous and amniotic cavity inoculation with cloaca strains resulted in E. cecorum recovery from all samples. Recovery scores of 45 to 80% were achieved for other routes.
All samples from eggs inoculated in the amniotic cavity with cloaca strains or lesion strains were E. cecorum-positive, while recovery from eggs inoculated by the albumen route differed significantly between cloaca and lesion strain inoculated groups: 9/20 in cloaca strain-inoculated groups and 20/20 in lesion strain-inoculated groups.
Experiment 3. E. cecorum recovery was significantly more successful from eggs inoculated with lesion strains (strains 15–28) compared to those inoculated with cloaca strains (strains 1–14): 68/70 versus 11/70. Almost all sampled eggs inoculated with lesion strains contained a dead embryo, while nearly all samples from eggs inoculated with cloaca strains came from eggs with a live embryo at day 18 of incubation.
Discussion
The aim of the study was to assess which embryo inoculation route is the best to determine the virulence of E. cecorum strains. To address this, two sets of E. cecorum strains are required; a group of avirulent strains and a group of virulent strains. Mannitol-positive cloaca strains are avirulent/low virulent, therefore only mannitol-positive strains were chosen for this set. In the set of virulent strains, mannitol-negative lesion strains were included. In the category mannitol-negative strains, avirulent counterparts may be found. However, this was the best option at hand. Indeed, typing in broilers would have been even better; however, this was not possible for obvious reasons already given in the introduction. It should be noted that it was not the aim of the study to determine the virulence of E. cecorum isolates, but to determine which embryo inoculation route best discriminates virulent from avirulent strains. The virulence should therefore be known beforehand.
Mortality induced in Experiment 1 by cloaca strains was much lower than that obtained with lesion strains for each of the inoculation routes, except the air chamber route. Using the latter route no mortality was observed with either cloaca or lesion strains (, Experiment 1), likely due to the inability of the bacteria to cross the air chamber membrane.
Albumen inoculation can be considered as indirect amniotic cavity inoculation with a delayed effect, as from embryonic day 12 onwards the albumen is transferred to the amniotic cavity via the sero-amniotic connection. This process is completed at day 18 of incubation (Deeming, Citation1991; Manders et al., Citation2021). In the aforementioned period, inoculated bacteria are exposed to the hostile environment of the albumen consisting of antimicrobial proteins and physico-chemical properties (Willems et al., Citation2014; Guyot et al., Citation2016). Although the antimicrobial activity of the albumen declines during the first 12 days of incubation (Guyot et al., Citation2016), it may still affect E. cecorum bacteria and may vary between strains.
The previously described delayed effect is reflected in longer MDTs after albumen inoculation with lesion strains (approximately 3–5 days) in comparison with amniotic inoculation (approximately 2 days) (, experiments 1, 2A and 3). In this context, the increasing immune competence of the embryo from day 12 of incubation onwards may also play a role (Hincke et al., Citation2019). A meaningful comparison in the same way regarding cloaca strains is not possible due to (very) low mortality rates in albumen-inoculated eggs.
The gap in embryo mortality induced by cloaca and lesion strains (the difference in mortality between the highest mortality score in the groups inoculated with the cloaca strains and the lowest mortality score in the groups inoculated with the lesion strains) was 10% after amniotic cavity inoculation. This gap was extended to 60% after albumen inoculation, using the same strains and doses (, Experiment 1).
Recovery of E. cecorum was (nearly) 100% successful after inoculation in the amniotic cavity with both, cloaca and lesion strains (, Experiment 1) and after inoculation of lesion strains in the albumen (, Experiment 2B from day 3 p.i.; , experiments 1 and 3). Recovery rates after albumen inoculation with cloaca strains were significantly below those obtained by lesion strains (, Experiment 2B; , experiments 1 and 3).
In view of the foregoing, it is concluded that the discrimination between strains based on embryo mortality after albumen inoculation can likely be explained by the added outcome of two effects: (1) mortality differences due to differences in virulence between strains and (2) a significant decrease of the proportion of infected embryos when cloaca strains are inoculated, as they are likely more sensitive to the harmful effects of the albumen. In a number of cases, this may result in the inactivation of bacteria to the extent that embryo infection does not occur. In the present study, a bacterial dose per egg in order of magnitude of 100 CFU was used. In case a substantially higher dose is applied, the aforementioned second effect may be reduced or even disappear completely. In which case, the discriminatory power of the albumen route is levelled to that of the amniotic cavity route.
In general, reproducibility of the results obtained by albumen inoculation was good: mortality induced by the same strains did not differ significantly between experiments, except for three strains (, experiments 1, 2A and 3, strains 4, 16 and 17). These exceptions are likely due to relative large differences in bacterial dose between experiments, as a positive correlation exists between dose and mortality (Borst et al., Citation2014; Blanco et al., Citation2017).
In the three experiments, mortality induced by cloaca and lesion strains following albumen inoculation ranged from 0-25% and from 15-100%, respectively. Lesion strains mostly caused 80–100% mortality (, experiments 1, 2A and 3). The overlap in mortality between cloaca and lesion strains is due to mortality of 15% by one lesion strain (, Experiment 3, strain 23). Although this strain was isolated from a typical E. cecorum lesion it might still be an avirulent/low virulent strain.
Considering all experiments, the E. cecorum dose per egg ranged from 21-215 CFU. Further limitation of this range will likely improve reproducibility/reliability of results.
In contrast to findings of other scientists (Borst et al., Citation2014; Jung, Metzner et al., Citation2017; Dolka et al., Citation2022), in our study no embryo mortality was induced by E. cecorum cloaca strains administered by the allantoic cavity route. However, the same strains inoculated in the same doses by other routes, did result in mortality (, Experiment 1), showing that these strains were detrimental for chicken embryos. Moreover, it is clear that the bacteria were not inactivated as they were recovered from 16/20 eggs at day 18 of incubation (, Experiment 1). The cause of the aforementioned conflicting results remains obscure.
Mortality rates after allantoic cavity inoculation of lesion strains (, Experiment 1) were in agreement with those obtained by other workers (Borst et al., Citation2014; Jung, Metzner et al., Citation2017; Dolka et al., Citation2022).
E. cecorum was isolated from nearly all broiler reproduction flocks (23/24), indicating widespread distribution of this bacterium in this poultry category in the Netherlands. This observation is in agreement with the common view that E. cecorum is part of the normal intestinal flora of chickens (Devriese et al., Citation1991).
In the present study, the genetic relatedness of 28 E. cecorum strains was analysed by the identification of SNPs. General cut-off values for the difference in the number of SNPs between strains to establish clonal relatedness are not available, thus species-specific cut-offs have to be defined, which has not been done for E. cecorum yet. Therefore in our study the cut-off was based on that of Enterococcus faecium. Multiple cut-off values, ranging from 6-17 differences in SNPs based on core genome MLST or WGS, have been proposed to identify putative transmission links and therefore genetic relatedness of E. faecium strains (Pinholt et al., Citation2015; Higgs et al., Citation2022). In our study, we used a difference of 17 SNPs between strains as cut-off value for clonality (Pinholt et al., Citation2015). The minimum difference in SNPs between E. cecorum strains was 85 (between lesion strains 21 and 24), which is above the cut-off value and therefore we considered all strains as non-clonal. However, it is important to consider the cut-offs of genetic relatedness as guidelines rather than absolute rules, and they should subsequently be applied with certain flexibility (Schürch et al., Citation2018).
The whole genome sequence data are subject to ongoing research on the molecular differences between lesion and cloacal strains.
In conclusion, we showed that the albumen inoculation route is the best to differentiate between mannitol-positive E. cecorum cloaca strains and mannitol-negative lesion strains in an ELA. To our knowledge, this is the first report on the use of albumen inoculation in an ELA.
Acknowledgements
We thank M. Spaninks for her technical assistance, and Dr R. Dijkman for critical reading of the manuscript.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
- Blanco, A.E., Barz, M., Icken, W., Cavero, D., Sharifi, A.R., Voss, M., Preisinger, R. & Buxade, C. (2017). Chicken embryo lethality assay for determining the lethal dose and virulence of Enterococcus faecalis. Avian Pathology, 46, 548–555.
- Borst, L.B., Suyemoto, M.M., Keelara, S., Dunningan, S.E., Guy, J.S. & Barnes, H.J. (2014). A chicken embryo lethality assay for pathogenic Enterococcus cecorum. Avian Diseases, 58, 244–248.
- Borst, L.B., Suyemoto, M.M., Robbins, K.M., Lyman, R.L., Martin, M.P. & Barnes, H.J. (2012). Molecular epidemiology of Enterococcus cecorum isolates recovered from enterococcal spondylitis outbreaks in the southeastern United States. Avian Pathology, 41, 479–485.
- Borst, L.B., Suyemoto, M.M., Sarsour, A.H., Harris, M.C., Martin, M.P., Strickland, J.D., Oviedo, E.O. & Barnes, H.J. (2017). Pathogenesis of enterococcal spondylitis caused by Enterococcus cecorum in broiler chickens. Veterinary Pathology, 54, 61–73.
- Bumgarner, L.R. & Finkelstein, R.A. (1973). Pathogenesis and immunology of experimental gonococcal infection: virulence of colony types of Neisseria gonorrhoeae for chicken embryos. Infection and Immunity, 8, 919–924.
- Deeming, D.C. (1991). Reasons for the dichotomy in egg turning in birds and reptiles. In D.C. Deeming & M.W.J. Ferguson (Eds.), Egg Incubation: Its Effects on Embryonic Development in Birds and Reptiles (pp. 307–324). Cambridge: Cambridge University Press.
- Devriese, L.A., Hommez, J., Wijfels, R. & Haesebrouck, F. (1991). Composition of the enterococcal and streptococcal intestinal flora of poultry. Journal of Applied Bacteriology, 71, 46–50.
- Dolka, B., Chrobak-Chmiel, D., Czopowicz, M. & Szeleszczuk, P. (2017). Characterization of pathogenic Enterococcus cecorum from different poultry groups: broiler chickens, layers, turkeys, and waterfowl. PLoS One, 12, e0185199.
- Dolka, B., Czopowicz, M., Dolka, I. & Szeleszczuk, P. (2022). Chicken embryo lethality assay for determining the lethal dose, tissue distribution and pathogenicity of clinical Enterococcus cecorum isolates from poultry. Scientific Reports, 12, 10675.
- Gibbs, P.S. & Wooley, R.E. (2003). Comparison of the intravenous chicken challenge method with the embryo lethality assay for studies in avian colibacillosis. Avian Diseases, 47, 672–680.
- Guyot, N., Réhault-Godbert, S., Slugocki, C., Harichaux, G., Labas, V., Helloin, E. & Nys, Y. (2016). Characterization of egg white antibacterial properties during the first half of incubation: a comparative study between embryonated and unfertilized eggs. Poultry Science, 95, 2956–2970.
- Herdt, P., Defoort, P., Steelant, J., Swam, H., Tanghe, L., Goethem, S.V. & Vanrobaeys, M. (2009). Enterococcus cecorum osteomyelitis and arthritis in broiler chickens. Vlaams Diergeneeskundig Tijdschrift, 78, 44–48.
- Higgs, C., Sherry, N.L., Seemann, T., Horan, K., Walpola, H., Kinsella, P., Bond, K., Williamson, D.A., Marshall, C., Kwong, J.C., Grayson, M.L., Stinear, T.P., Gorrie, C.L. & Howden, B.P. (2022). Optimising genomic approaches for identifying vancomycin-resistant Enterococcus faecium transmission in healthcare settings. Nature Communications, 13, 509.
- Hincke, M.T., Da Silva, M., Guyot, N., Gautron, J., McKee, M.D., Guabiraba-Brito, R. & Réhault-Godbert, S. (2019). Dynamics of structural barriers and innate immune components during incubation of the avian egg: critical interplay between autonomous embryonic development and maternal anticipation. Journal of Innate Immunity, 11, 111–124.
- Jolley, K.A., Bliss, C.M., Bennett, J.S., Bratcher, H.B., Brehony, C., Colles, F.M., Wimalarathna, H., Harrison, O.B., Sheppard, S.K., Cody, A.J. & Maiden, M.C.J. (2012). Ribosomal multilocus sequence typing: universal characterization of bacteria from domain to strain. Microbiology, 158, 1005–1015.
- Jolley, K.A., Bray, J.E. & Maiden, M.C.J. (2018). Open-access bacterial population genomics: BIGSdb software, the PubMLST.org website and their applications. Wellcome Open Research, 3, 124.
- Jung, A. & Rautenschlein, S. (2014). Comprehensive report of an Enterococcus cecorum infection in a broiler flock in Northern Germany. BMC Veterinary Research, 10, 311.
- Jung, A., Metzner, M. & Ryll, M. (2017). Comparison of pathogenic and non-pathogenic Enterococcus cecorum strains from different animal species. BMC Microbiology, 17, 33.
- Jung, A., Petersen, H., Teske, L. & Rautenschlein, S. (2017). Colonization patterns of Enterococcus cecorum in two different broiler production cycles detected with a newly developed quantitative real-time PCR. BMC Microbiology, 17, 106.
- Kense, M.J. & Landman, W.J.M. (2011). Enterococcus cecorum infections in broiler breeders and their offspring: molecular epidemiology. Avian Pathology, 40, 603–612.
- King, A.S. (1975). Aves: respiratory system. In R. Getty (Ed.), Sisson and Grossman’s Anatomy of the Domestic Animals 5th ed., vol. 2, (p. 1909). London: W.B. Saunders.
- Landman, W.J.M., Buter, G.J., Dijkman, R. & Van Eck, J.H.H. (2021). In vivo typing of Escherichia coli obtained from laying chickens with the E. coli peritonitis syndrome. Avian Pathology, 50, 436–446.
- Manders, T.T.M., Matthijs, M.G.R., Veraa, S., Van Eck, J.H.H. & Landman, W.J.M. (2021). Success rates of inoculation of the various compartments of embryonated chicken eggs at different incubation days. Avian Pathology, 50, 61–77.
- Martin, L.T., Martin, M.P. & Barnes, H.J. (2011). Experimental reproduction of enterococcal spondylitis in male broiler breeder chickens. Avian Diseases, 55, 273–278.
- Nix, E.B., Cheung, K.K., Wang, D., Zhang, N., Burke, R.D. & Nano, F.E. (2006). Virulence of Francisella spp. in chicken embryos. Infection and Immunity, 74, 4809–4816.
- Nolan, L.K., Wooley, R.E., Brown, J., Spears, K.R., Dickerson, H.W. & Dekich, M. (1992). Comparison of a complement resistance test, a chicken embryo lethality test, and the chicken lethality test for determining virulence of avian Escherichia coli. Avian Diseases, 36, 395–397.
- Pakpour, N. & Horgan, S. (2021). General Microbiology lab Manual. Hayward: California State University East Bay.
- Pinholt, M., Larner-Svensson, H., Littauer, P., Moser, C.E., Pedersen, M., Lemming, L.E., Ejlertsen, T., Søndergaard, T.S., Holzknecht, B.J., Justesen, U.S., Dzajic, E., Olsen, S.S., Nielsen, J.B., Worning, P., Hammerum, A.M., Westh, H. & Jakobsen, L. (2015). Multiple hospital outbreaks of vanA Enterococcus faecium in Denmark, 2012-13, investigated by WGS, MLST and PFGE. Journal Antimicrobial Chemotherapy, 70, 2474–2482.
- Robbins, K.M., Suyemoto, M.M., Lyman, R.L., Martin, M.P., Barnes, H.J. & Borst, L.B. (2012). An outbreak and source investigation of enterococcal spondylitis in broilers caused by Enterococcus cecorum. Avian Diseases, 56, 768–773.
- Schreier, J., Rautenschlein, S. & Jung, A. (2021). Different virulence levels of Enterococcus cecorum strains in experimentally infected meat-type chickens. PLoS One, 16, e0259904.
- Schürch, A.C., Arredondo-Alonso, S., Willems, R.J.L. & Goering, R.V. (2018). Whole genome sequencing options for bacterial strain typing and epidemiologic analysis based on single nucleotide polymorphism versus gene-by-gene-based approaches. Clinical Microbiology and Infection, 24, 350–354.
- Seo, H.S., Cha, S.Y., Kang, M. & Jang, H.K. (2013). Chicken embryo lethality assay for determining the virulence of Riemerella anatipestifer isolates. Avian Pathology, 42, 387–392.
- Statistix®. (2010). User's manual Statistix 8.2 for Windows [Computer software]. Retrieved from www.statistix.com.
- Willems, E., Decuypere, E., Buyse, J. & Everaert, N. (2014). Importance of albumen during embryonic development in avian species, with emphasis on domestic chicken. World's Poultry Science Journal, 70, 503–518.