Reducing Campylobacter colonization in broilers by active immunization of naive broiler breeders using a bacterin and subunit vaccine

Abstract

Campylobacter is the main cause of human gastroenteritis worldwide, with 50 to 80% of the cases related to consumption of poultry products. Maternal antibodies (MAB) from commercial breeder flocks may protect their progeny against infection during the first few weeks of life. We here studied the prevalence of Campylobacter antibody titers in broiler breeder flocks and to which extent immunization of broiler breeders increases maternal anti-Campylobacter titers in their progeny and protects the offspring against Campylobacter colonization. Two vaccines were used: a bacterin mix of 13 Campylobacter strains and a subunit vaccine comprising 6 immunodominant Campylobacter antigens.

All sampled on-farm breeder flocks were positive for anti-Campylobacter antibodies, yet in some breeder flocks only very low titers were detected. Vaccination of SPF broiler breeder flocks with both subunit and bacterin vaccines resulted in a prolonged presence of anti-Campylobacter antibodies in the serum and intestinal mucus of chicks. These bacterin- or subunit vaccine-induced MAB conferred protection against Campylobacter colonization in chicks until 7 and 21 d of age, respectively, but only at a low challenge dose (10^2.5 cfu). The concentration of MAB in the mucus is probably too low to sufficiently capture Campylobacter when higher challenge doses are used. In conclusion, vaccinating broiler breeders protects their offspring against Campylobacter colonization under low pathogen exposure conditions.

INTRODUCTION

Since 2005, campylobacteriosis has been the most often reported gastrointestinal illness in the European Union (EU), affecting over 127,840 people in 2021 (EFSA, 2022). However, these figures are vastly underestimated, because the true incidence was calculated to be 9.2 million cases annually, or 46.7 unreported cases for each reported case (Havelaar et al., 2013). The annual economic impact in the EU was estimated in 2014 at € 2.4 billion, due to work absences and medical costs (EFSA, 2014). In addition to the self-limiting gastroenteritis, Guillain-Barré syndrome, reactive arthritis, postinfectious irritable bowel syndrome, and inflammatory bowel disease are long-term consequences that could be encountered after Campylobacter infection (Keithlin et al., 2014). According to the European Food Safety Agency (EFSA), poultry meat may directly accounts for 20 to 30% of human cases and the poultry reservoir as a whole is responsible for 80% of the cases (EFSA, 2010). The majority of these cases are caused by the thermophilic species Campylobacter jejuni and Campylobacter coli. Being the main reservoir for human campylobacteriosis, 38.7% of the broiler carcasses in the EU were contaminated with Campylobacter sp. in 2020 (EFSA, 2021).

Campylobacter is considered a commensal in avian hosts (Keener et al., 2004), primarily colonizing the ceca to a high degree (Luangtongkum et al., 2006, Stern, 2008; Kittl et al., 2011). Despite having a high transmission rate with a prevalence increasing from 5 to 95% within 6 d in a flock (Van Gerwe et al., 2005), Campylobacter colonization is rarely seen during the first 2 wk of life, in what is called the “lag period” (Newell et al., 2011). It has been speculated that this “lag period” is due to maternal IgY antibodies (MAB), passed from the mother hens to the progeny, which possibly decreases the susceptibility or the shedding of the chicks for Campylobacter infections (Sahin et al., 2003; Cawthraw and Newell, 2010). Cawthraw and Newell (2010) state that most, if, not all, commercial breeder flocks are colonized with Campylobacter sp.

Studies investigating numbers of Campylobacter antibody titers that are present in these breeder flocks and that are passed on to their progeny are lacking. Therefore, the first aim of this study is to assess the prevalence of Campylobacter antibody titers in broiler breeder flocks in the field as well as in their offspring. In the field, broiler breeder flocks are vaccinated against several pathogens, for instance chicken infectious anemia virus, reovirus, Salmonella Enteritidis, Salmonella Typhimurium, and E. coli. Due to this active immunization of broiler breeders maternal antibodies against these pathogens are transferred to their progeny and protect chicks for up to 3 wk after hatching. However, boosting maternal immunity by vaccinating breeders with vaccines based on oil adjuvant has been shown to prolong this passive immunity up to 5 to 7 wk of age (Lucio and Hitchner, 1979; Baxendale and Lutticken, 1981; Heller et al., 1990). Therefore, the second aim of this study is to assess if hyperimmunization of broiler breeders can increase maternal anti-Campylobacter titers in their progeny and thus protect them against Campylobacter colonization throughout the growth period.

MATERIALS AND METHODS

Presence of Maternal Campylobacter IgY Titers in Broiler Breeder Flocks and Their Progeny in the Field

To investigate the prevalence of maternal Campylobacter IgY antibodies circulating in the field, eggs (n = 20/flock) from 10 different broiler breeder flocks, varying in age (range 26–43 wk) and geography (different Flemish regions) were acquired via a local hatchery (Belgabroed NV, Merksplas, Belgium). For quantification of the IgY titers, the egg yolks were pooled per flock and analyzed via ELISA as described below.

From the breeder flocks with the highest and lowest mean Campylobacter IgY titers respectively, cecal samples were collected and pooled, after which culture for Campylobacter was performed as discussed below. Next, 30 one-day-old chicks originating from the breeder flock with the highest mean Campylobacter IgY titers were purchased from the local hatchery and raised in Campylobacter-free conditions. The chicks received ad libitum commercial feed (Farm 1 and 2, Versele Laga, Belgium) and drinking water. Husbandry, euthanasia methods, experimental procedures, and biosafety precautions were according to the European ethical guidelines. At different time points after hatch (d 1, 3, 7, 9, 14, and 21) 5 chicks were euthanized by injection of 100 mg/kg of sodium pentobarbital (Kela, Hoogstraten, Belgium) in the wing vein. At 1 d and 28 d of age, 5 chicks derived from the breeder flock with the lowest IgY antibodies were also euthanized and analyzed. After euthanasia, birds were necropsied and bloodserum and cecal mucus samples collected, in order to determine the half-life time and distribution of the maternal antibodies as described below.

MLST Analysis of Campylobacter Isolates

The species of the colonizing Campylobacter was determined via MALDI-TOF MS (Autoflex Speed LRF, Bruker, Billerica, MA). Subsequently, the genetic nature of the Campylobacter isolates from the on-farm breeders flocks was analyzed via multilocus sequence typing (MLST). Based on this, bacterial strains are divided into clonal complexes (CC) and sequence types (ST). DNA extraction of the isolates was performed with the DNeasy Blood and Tissue Kit (Qiagen, Venlo, the Netherlands). DNA yield was measured (Nanodrop 2.0) and diluted to 25 ng/µL. Primers, PCR conditions, sequencing and MLST scheme were as described by Jolley et al. (2018).

Bacterial Strains and Culture Conditions

For all experimental infections, C. jejuni strain KC40 from poultry origin was used as challenge strain, which has proven to be an excellent colonizer of the chicken ceca (Van Deun et al., 2008; Hermans et al., 2014; Vandeputte et al., 2019a). To compose the bacterin, 13 Campylobacter strains were used which were kindly provided by Sciensano (Brussels, Belgium), except for the C. jejuni KC40 reference strain which was previously isolated at the Flanders Research Institute for Agriculture, Fisheries and Food (ILVO, Melle, Belgium) (Table 1). These Campylobacter strains, all from chicken origin, were selected based on their genetic heterogeneity as analyzed via MLST, prevalence ratio in broilers and relationship with human campylobacteriosis cases (Vandeputte et al., 2019a).

Table 1.

Campylobacter strains included in the bacterin.

Species	Strain	Clonal complex	Sequence type	Poultry origin
C. jejuni	KC40	677	794	Broiler dunghill
	10kf-1.16	283	267	Carcass
	7P6.12	464	464	Feathers
	10c-6.1	574	305	Ceca
	10kf-4.12	433	51	Carcass
	10VTDD-5	UA	905	Unknown
	T124	658	1044	Ceca
	T84	354	1073	Ceca
	T70	21	50	Carcass
	3291	45	45	Carcass
	5970	UA	5970	Carcass
C. coli	2711	828	845	Carcass
	3250	UA	5163	Carcass

For bacterin and challenge inoculum preparation C. jejuni and C. coli bacteria were first grown on modified charcoal cefoperazone deoxycholate agar (mCCDA, CM0739; Oxoid) followed by microaerobic incubation (5% O₂, 5% CO₂, 5% H₂, 85% N₂) at 42°C for 22 h. Subsequently, bacteria were cultured in Nutrient Broth No. 2 (NB2, CM0067; Oxoid Ltd., Basingstoke, Hampshire, UK) supplemented with modified Preston Campylobacter-selective supplement (SR0204E; Oxoid) and Campylobacter-specific growth supplement (SR0232E; Oxoid), at 37°C for 17 h under microaerobic conditions. Subsequently, C. jejuni and C. coli bacteria were enumerated by plating 10-fold dilutions in Hank's balanced salt solution (HBSS; GIBCO-BRL, Invitrogen, Carlsbad, CA) on modified charcoal cefoperazone deoxycholate agar (mCCDA) supplemented with charcoal cefoperazone deoxycholate agar (CCDA)-selective supplement (SR0155E; Oxoid) and Campylobacter-specific growth supplement, followed by microaerobic incubation at 42°C for 22 h.

To monitor the Campylobacter status in cecal droppings of the breeders and broiler chickens, the same culture conditions were used as described above, with following adaptations: cecal samples were also diluted 1:9 (wt/vol) in enriched NB2 and incubated at 42°C for 17 h under microaerobic conditions before being plated on mCCDA agar supplemented with CCDA-selective supplement and Campylobacter-specific growth supplement.

Preparation of Recombinant C. Jejuni Antigens

Based on previous research (Hermans et al., 2014; Vandeputte et al., 2019a), 6 immunodominant antigens were selected to be incorporated in the subunit vaccine (Table 2). These proteins are expressed on the bacterial cell surface or are associated with the cell membrane (Tsugawa et al., 2007; Nieves et al., 2010; Lertsethtakarn et al., 2011; Zhang et al., 2012; Hermans et al., 2014).

Table 2.

Proteins incorporated in the subunit vaccine (Hermans et al., 2014).

Name	Function
AtpA	AtpA synthase
CheV	Chemotaxin protein V
Ef-Tu	Translational elongation factor
GroEL	Cochaperonin Groel
LivJ	Substrate binding protein
Tig	Trigger factor

For recombinant production of the immunodominant antigens, derived from the C. jejuni reference strain KC40, the Escherichia coli Expression System using Gateway Technology (Invitrogen) was used as described previously (Vandeputte et al., 2019a) with minor changes to the protocol. The addition of 10% glycerol (Sigma) to the elution buffer and a dialysis step in phosphate-buffered saline (PBS) and 10% glycerol (Sigma) overnight after elution were included in order to counter protein precipitation and conserve biological activity of the proteins.

Bacterin and Subunit Vaccine Preparation

Bacterin and subunit vaccine preparation was done as described previously (Vandeputte et al., 2019b). In short, for the bacterin, the 13 Campylobacter strains were cultured separately in NB2 to a concentration of 9 log₁₀ colony forming units (cfu)/mL and subsequently killed by incubation at 37°C for an overnight period with 5 mL of 36% formaldehyde (Sigma-Aldrich). After centrifugation for 30 min at 5,718 relative centrifugal force (rcf) at 20°C, the pellets were resuspended in 5 mL of 36% formaldehyde/L of PBS and incubated overnight at 37°C. The suspensions were then plated on mCCDA agar and incubated overnight at 37°C to ensure that all the bacteria were killed after which the suspensions were kept at 4°C for further use. A mix of the 13 Campylobacter suspensions was made, so that each bacterin dose consisted of 8.1 log₁₀ cfu inactivated Campylobacter (i.e., 7 log₁₀ cfu/Campylobacter strain).

For the subunit vaccine, HBSS was added to 75 µg of protein (12.5 µg of each recombinant antigen) until a volume of 125 µL/vaccine dose was reached. As placebo vaccination (the negative control group) 125 µL of pure HBSS was used. For the first immunization and the 3 boosters, respectively Freund's complete adjuvant (FCA, Sigma-Aldrich) and Freund's incomplete adjuvant (FIA, Sigma-Aldrich) were combined in a 250 µL dosage at a 1:1 ratio.

Immunization of Broiler Breeders and Determination of Vaccine-Derived Maternal Antibodies

From a local hatchery (Avian Epi BV, Roermond, the Netherlands), 45 one-day-old commercial Ross 308 broiler breeder hens and 10 Ross 308 broiler breeder roosters were purchased. Throughout the rearing process, the juvenile birds were housed in the same stable unit according to the guidelines of the ethical committee. This means that they were housed on a bedding of wood shavings with a minimum enclosure space depending on the weight of the animals (ranging from 0.025 m² per bird during the first week to 0.21 m² when fully grown out). In the first weeks, the chicks were kept under a heating lamp and from 28 d onward, animals were housed at room temperature. The birds were raised in Campylobacter-free conditions and under high biosecurity levels (BSL2 facilities). To ensure the chickens were kept free of Campylobacter, cecal droppings were regularly checked for presence of Campylobacter sp. via quantitative polymerase chain reaction (qPCR) and culture. Also, serum and yolk were collected to monitor for presence of IgY antibodies against Campylobacter by means of ELISA. The birds were given restricted commercial food (Versele Laga, Belgium) and light schedule was adapted based on the ROSS Broiler Breeder Management Guide. Ad libitum access to drinking water was provided. At the start of production (around 20 wk of age) chickens were randomly assigned to 3 vaccine groups and each separated in 3 different units of the same stable: bacterin (n = 15), subunit (n = 15) and control (n = 15).

Broiler breeders were vaccinated by intramuscular injection in the pectoral muscle with the vaccines composed as described above. The first immunization (prime) was given at 28 wk of age, after which 3 booster immunizations were given in a 2-wk time interval. From the third immunization (second booster) onward, yolk and blood derived from the wing vein was collected to determine the IgY-titers as described below. The yolk and serum samples, derived 2 wk after the final (third) booster, were analyzed individually per hen and the average was calculated. After the final booster, 4 mo, 6 mo, and 8 mo later yolk and serum samples were pooled and each quantified per group (control, bacterin, and subunit). Fertilized eggs from these placebo and vaccinated broiler breeders were routinely incubated and after hatch the chicks were used in the in vivo trials as described below.

After the final booster, 36 chicks from the placebo and each vaccination group (bacterin, subunit, or control) were hatched. At different time points (d 1, 4, 10, 14, 21, and 28 of age) 6 chicks per group were euthanized by injection of an overdose (100 mg/kg) sodium pentobarbital (Kela, Hoogstraten, Belgium) and serum and cecal mucus samples were collected to determine the Campylobacter IgY titers as described below, in order to determine the half-life time and distribution of the maternal antibodies.

Husbandry, euthanasia methods, experimental procedures, and biosafety precautions were approved by the Ethical Committee (EC2020_036) of the Faculty of Veterinary Medicine, Ghent University, Belgium.

Protective Efficacy of Bacterin- and Subunit-Derived Maternal Antibodies on Cecal Colonization in Broilers Using a Seeder Model

By use of a seeder model, the protective efficacy of MAB against horizontal transmission of Campylobacter was studied. In this in vivo model, only a minority of the chickens (seeders) per group are inoculated with Campylobacter. These seeders are housed together with noninoculated sentinel chickens, thus mimicking the natural spread of Campylobacter in a poultry flock (Vandeputte et al., 2019a). From each broiler breeder group (bacterin, subunit, and control) chicks were hatched and housed in separate isolation units per vaccination group. The birds were again housed according to the guidelines by the ethical committee. Birds were raised on a bedding of wood shavings and under a heating lamp. The minimum enclosure space was 0.025 m² per bird. From the day of hatch until the end of the experiment, chicks received ad libitum commercial starting feed (Farm 1 and 2, Versele Laga, Belgium) and drinking water. At 7 d of age, 2 challenge groups per vaccination were made of 6 randomly selected birds in separate boxes. This means that they were raised on wood shavings and under a heating lamp. The total enclosure space in the challenge groups was 1 m². Two chickens (seeders) from each challenge group were randomly selected and orally inoculated with approximately 4.5 log₁₀ cfu of C. jejuni strain KC40. The birds that were not inoculated are referred to as contact animals or sentinels. Using this model, the Campylobacter infection will spread from the seeders to the other animals of the same group reproducing the natural way of infection in the stable and prevention of infection and transmission can be investigated. All animals were euthanized 3 d after inoculation by injection of an overdose (100 mg/kg) sodium pentobarbital (Kela, Hoogstraten, Belgium) in the wing vein and the cecal content was collected for C. jejuni enumeration (as described below). The experiment was performed in duplicate. Husbandry, euthanasia methods, experimental procedures and biosafety precautions were approved by the Ethical Committee (EC2020_036) of the Faculty of Veterinary Medicine, Ghent University, Belgium.

Protective Efficacy of Bacterin- and Subunit-Derived Maternal Antibodies on Cecal Colonization in Broilers Using a Threshold Model

To assess the protection of MAB against the onset of Campylobacter infection in a broiler flock, a threshold trial was used (Hermans et al., 2012). This model determines the Campylobacter colonization threshold by challenging individually housed chicks on different ages (7, 14, and 21 d of age) using different inoculation doses (10^2.5, 10^3.5, and 10^4.5 cfu). Fertilized eggs, derived from the immunized and placebo in-house broiler breeders, were hatched and housed per group in separate isolation units. Birds were housed according to the guidelines by the ethical committee. This means that they were raised on a bedding of wood shavings and under a heating lamp. The minimum enclosure space depended on the weight of the birds (spanning from 0.025 m² per bird at wk 1–0.09 m² at wk 3). From the day of hatch until the end of the experiment, chicks received ad libitum commercial starting feed (Farm 1 and 2, Versele Laga, Belgium) and drinking water. From different time points onward (7, 14, and 21 d of age) chicks were housed individually in a box on a bedding of wood shavings with an enclosure space of 0.16 m² per chick and audiovisual contact with the other chicks. After this they were orally inoculated with different doses of C. jejuni KC 40 (2.5 log₁₀ cfu, 3.5 log₁₀ cfu, or 4.5 log₁₀ cfu) according to Table 3. Numbers of animals included per time point (sample size, Table 3), were based on the available number of hatched chickens. Before challenge, birds were proved to be free of Campylobacter by examination of mixed fecal samples, as described above. Blood from the wing vein was collected from all chicks 48 h after inoculation for quantification of the Campylobacter IgY titers in the serum using ELISA (as described below). Next, all chicks were euthanized by injection of 100 mg/kg of sodium pentobarbital (Kela, Hoogstraten, Belgium) in the wing vein and the cecal content was collected for C. jejuni enumeration (as described below). Husbandry, euthanasia methods, experimental procedures and biosafety precautions were approved by the Ethical Committee (EC2021_086) of the Faculty of Veterinary Medicine, Ghent University, Belgium.

Table 3.

Threshold trial design.

Week	Challenge dose(s)	Amount of birds per treatment group (n)
Wk 1 Trial 1	102.5 cfu1	Bacterin, subunit, control (n = 9)
	103.5 cfu	Bacterin, subunit, control (n = 9)
	104.5 cfu	Bacterin, subunit, control (n = 9)
Wk 1 Trial 2	102.5 cfu	Bacterin, subunit, control (n = 27)
Wk 2	102.5 cfu	Bacterin, subunit, control (n = 16)
	103.5 cfu	Bacterin, subunit, control (n = 16)
Wk 3	102.5 cfu	Subunit, control (n = 2 × 18)
	103.5 cfu	Bacterin, subunit, control (n = 17)

¹cfu: colony forming units.

Determination of IgY Titers in Egg Yolk, Serum, and Mucus

IgY titers in egg yolks, serum, and cecal mucus were determined as described previously (Hermans et al., 2014; Vandeputte et al., 2019a) with minor changes to the protocol. Egg yolks were diluted 1/5 (vol/ vol) in PBS, mixed and incubated overnight at 4°C. Afterward the supernatants, containing the water-soluble fraction of the egg yolk, were collected for IgY quantification using ELISA. Chicken blood was centrifuged (1,425 rcf, 15 min) 30 min after collection, to allow the blood to clot, after which the serum was collected for IgY quantification using ELISA. Cecal mucus, collected at necropsy, was diluted 1/5 (w/v) in PBS with 1% proteinase inhibitor. After ultra-centrifugation (14,674 rcf, 15 min) the supernatant was collected for IgY quantification using ELISA.

To determine IgY titers against Campylobacter, 96 well flat bottom plates (Nunc MaxiSorp, Nalge Nunc Int., Rochester, NY) were coated (24 h, 4°C) with the 13 different Campylobacter strains (for the field screening of broiler breeders, bacterin breeder group, bacterin chicks and flock C chicks), the Campylobacter KC40 (bacterin chicks’ serum collected during the threshold trial), the Campylobacter isolate derived from breeder Flock A (for the flock A chicks) or a mixture of 6 subunit antigens (for the subunit breeder group and subunit chicks). Plates were coated with a mixture of 10⁶ cfu/well or a mixture of 6 µg of the subunit proteins, diluted in 50 µL coating buffer. After washing (3× washing buffer: 0.1% Tween-20 in PBS), the wells were blocked (90 min, 37°C) with 100 µL blocking buffer (5% skimmed milk powder in washing buffer). Next, 100 µL of a 1:2 dilution series of the sample was incubated in duplicate during 60 min at room temperature. Plates were washed as described above and incubated with 100 µL 1/10,000 horseradish peroxidase (HRP)-labeled antichicken IgY (Sigma-Aldrich) in washing buffer during 90 min at room temperature. Subsequently, the plates were incubated with 50 µL 3,3′,5,5′-tetramethyl benzidine (TMB) substrate (Sigma-Aldrich) for 10 min at room temperature in the dark. Next, 50 µL 0.5 M H₂SO₄ was added to each well and the absorbance at 450 nm (OD450) was measured using an automated spectrophotometer (Pharmacia LKB Ultrospec III, Gemini BV, Apeldoorn, the Netherlands).

The cut-off was determined by using the R package “changepoint” (Killick and Eckley, 2014; Lardeux et al., 2016). In this study the cut-off value was defined as the changepoint in the dilution series of the negative control. The dilution of the negative control at which this changepoint occurs is defined as the limit of detection. The IgY titer of the sample was reported as the dilution with the lowest absorbance value greater than the cut-off value. IgY titers, which were lower than the limit of detection were considered negative. For the field mucus samples, no changepoint could be established for the dilution series from the negative control and based on later results, an arbitrary cut-off of 0.1 optical density (OD) was used.

Cecal Campylobacter Enumeration by qPCR

For DNA-extraction, the QIAamp Fast DNA Stool Mini Kit (Qiagen, Venlo, the Netherlands) was used according to the manufacturer's instructions, but instead of 200 µL ATE buffer, the DNA was eluted with 100 µL. Quality control was verified via Nanodrop spectrophotometry and the DNA was stored at −20°C until further use.

The Campylobacter spp. status and level of cecal colonization was assessed by qPCR analyses as described by Lund et al. (2004) and Botteldoorn et al. (2008). Per qPCR reaction, 12.5 µL IQTM Supermix (Bio-Rad, Temse, Belgium), 0.25 µL of each primer (forward primer Camp2F: 5′ CACGTGCTACAATGGCATAT 3′, reverse primer Camp2R: 5′ GGCTTCATGCTCTCGAGTT 3′), 0.25 µL probe (Camp2P: 5′ 6FAMCAGAGAACAATCCGAACTGGGACA-BHQ1 3′), 6.75 µL high-performance liquid chromatography (HPLC) water and 5 µL sample DNA was mixed until a total volume of 25 µL. DNA samples were analyzed in duplicate. Primers and probe were purchased from Integrated DNA Technologies IDT (Leuven, Belgium). the qPCR program comprised 4 steps: 2 min at 50°C, 10 min at 95°C and 50 cycles of 15 s at 95°C and 60 s at 60°C (CFX96 Real-Time PCR Detection System, Bio-Rad). The number of Campylobacter in the cecal content was expressed as genomic equivalents (ge)/g cecal content.

Statistical Analysis

All figures were made in GraphPad Prism 9.

For the field trial GraphPad Prism 9 was used. To compare yolk titers between Flock A and C, a 2-tailed nonparametric Mann-Whitney test was used. The correlation between age and breeder flock antibody titer was determined using a Spearman rank correlation test.

The half-life of maternal antibodies in chick serum was determined using the “PKNCA” package in R 4.2.0. The cut-off value of negative controls in ELISAs was determined using the “changepoint” package in R 4.2.0 with the “SIC” and “Exponential” parameters for penalty and test.stat, respectively.

For analysis of the challenge in vivo trials R.4.2.0 was used. C. jejuni counts were transformed to log₁₀ counts.

The threshold prevalence data were analyzed using a generalized (i.e., binomial) linear mixed-effects regression model with Campylobacter presence or absence as the dependent variable and age, inoculation dose (thresholdtrial), vaccine (seeder- and thresholdtrial) and seeder and sentinel (seedertrial) as independent variables. A random intercept term was included for each trial level (thresholdtrials) or repetition (seedertrial) to allow the intercept to vary across trials. The resulting treatment effects were compared via the “emmeans” function. The threshold colonization data were analyzed using a zero-inflation model with the Campylobacter titer as the dependent variable and vaccine as the independent variable. For the age of 7 d, a random factor Trial was included as a random slope term to allow the effect of independent variables to vary across different trial levels. The analysis was repeated for each age and each dose individually at age 14 and 21 d. The seeder colonization data were also analyzed using a zero-inflation model with the Campylobacter titer as the dependent variable and vaccine as the independent variable. A random intercept term was included for each pen to allow the intercept to vary across different pens.

The decrease of maternal antibodies, both in serum and mucus, and the ratio of antibodies transferred from the serum to the mucus was analyzed in R 4.2.0. The data set was being simplified by grouping it by vaccine and age, and then the mean of the antibody titer values or the ratio was taken. A linear regression model was then fitted to the log-transformed data with age, vaccine, and their interaction term as predictor variables. Model residuals were approximately normally distributed (Shapiro-Wilk W ∼ 0.90). Eventually, the slopes of the regression line for each vaccine were obtained and compared using the “emmeans” package, using the Tukey method to adjust P values for multiple pairwise comparisons.

RESULTS

Anti-Campylobacter Antibody Titers Are Highly Prevalent in Broiler Breeder Flocks

All broiler breeder flocks sampled in the field showed anti-Campylobacter IgY titers. Mean anti-Campylobacter IgY titers per flock observed in the yolk ranged from 1:200 to 1:3,200 (Figure 1). No correlation could be found between anti-Campylobacter titer and age (Spearman, r = 0.2301, P value = 0.517). Based on these results, the flock with the highest (Flock A) and the flock with the lowest titer (Flock C) were selected for further analysis of individual yolk titers and checked for presence of Campylobacter in the flock. Campylobacter could be isolated from a pooled cecal sample from Flock A, while Flock C was negative for Campylobacter at the moment of sampling. Via MALDI-TOF MS (Autoflex Speed LRF, Bruker, Billerica, MA) the strain colonizing Flock A was identified as C. coli. Subsequently, this strain was typed via multilocus sequence typing (MLST) as ST 830 belonging to the ST-828 complex, a clonal complex also used in the bacterin vaccine and - coating.

Figure 1.

Average anti-Campylobacter yolk IgY titers originating from 10 broiler breeder field flocks ranked by age.

All egg yolks derived from breeder Flock A had anti-Campylobacter IgY titers of 1:800 or higher (Figure 2) with an average titer of 1:4,520 ± 507. Fifteen out of 20 egg yolks derived from breeder Flock C had an anti-Campylobacter IgY titer below the detection limit (<1:200). Five yolks had IgY titers of 1:200. Titers observed in the 2 breeders flocks proved to be significant different from each other (P < 0.0001).

Figure 2.

Distribution of the anti-Campylobacter IgY antibody titers in individual yolks (n = 20) derived from the broiler breeder flock with the highest (Flock A) and the lowest (Flock C) mean yolk IgY titer. The statistical significance between groups is summarized as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.

Vaccination Leads to High Anti-Campylobacter Titers in Broiler Breeders

Vaccination of broiler breeders with the bacterin or subunit vaccine leads to high anti-Campylobacter IgY titers in both yolk and serum. Compared to the bacterin vaccine, the subunit vaccine results in higher IgY titers against their respective antigen (pooled serum titers of 1:1,600,000 vs. 1:12,000 after the second booster, respectively) (Table 4). In both groups, a third booster did not result in higher antibody titers compared to the titers observed after the second booster. Moreover, after the third booster a decline of 90.6% in both yolk and serum titers could be seen in the subunit vaccinated broiler breeder group. Here after, the level of the serum and yolk titers were maintained until the end of the experiment (8 mo after the final booster). The bacterin-induced titers remained stable throughout the whole experiment. In both vaccine groups, yolk titers were highly comparable to the serum titers with a maximum difference of approximately one 2-fold dilution. Limited reaction of the negative controls could be observed on the bacterin coated ELISA plates, however not consistent in time. Therefore, the signal is probably related to background or cross reactions (Table 4).

Table 4.

Mean bacterin- and subunit vaccine-induced egg yolk and serum IgY titers in broiler breeders as determined by ELISA at different time points.

Group	Bacterin		Control		Subunit		Control
Matrix	Yolk	Serum	Yolk	Serum	Yolk	Serum	Yolk	Serum
Antigen	Bacterin coating				Subunit coating
Booster 2	1:12,000	1:12,000			1:1,600,000	1:1,600,000
Booster 3	1:9,200± 2,375	1:9,733± 2,513	Negative	1:1,600	1:1,706,667± 440,659	1:693,333± 179,017	Negative	Negative
4 mo	1:8,000	1:8,000			1:160,000	1:80,000
6 mo	1:8,000	1:8,000			1:160,000	1:80,000
8 mo	1:16,000	1:16,000	1:800	1:1600	1:160,000	1:80,000	Negative	Negative

Half-Life Curve of Maternal Campylobacter Antibodies in Broilers Originating From Vaccinated Breeders and Breeders Seropositive After Field Infection

After the final booster, chicks from each vaccinated broiler breeder group (bacterin, subunit, or control) were hatched and chicks originating from the breeders flocks with the highest (Flock A) and lowest IgY titers (Flock C) encountered in the field were purchased from the hatchery to determine the half-life time of the anti-Campylobacter antibodies in the serum and cecal mucus.

The average maternal IgY antibody titer in the serum at d 1, against a whole cell coating of the C. coli strain isolated from the breeder flock for Flock A chicks, bacterin coating for bacterin chicks and Flock C chicks and subunit coating for subunit chicks, amounted to 1:10,880 ± 3,997 for Flock A chicks, 1:4,800 ±1,753 for bacterin chicks and 1:62,933 ± 10,826 for subunit chicks. All serum samples were under the detection limit at d 21 for the Flock A chicks and at d 28 for the bacterin chicks. For subunit chicks 50% of the serum samples were still above the detection limit at d 28 (Figure 3A). Some background reaction in the control animals was seen on both d 1 and 4, ranging to 1:400 on bacterin coating and 1:1,600 on subunit coating. All serum samples of the Flock C chicks proved negative. The half-life time of maternal anti-Campylobacter antibodies in the serum for bacterin, subunit and Flock A chicks numbered 3.5, 5.2, and 4.0 d, respectively. The decline of maternal antibodies in the serum was slower in subunit chicks in comparison with Flock A chicks (P = 0.001) and bacterin chicks (P = 0.021). Significant difference in decrease in maternal antibodies between Flock A and bacterin chicks could not be noticed.

Figure 3.

Half-life curve of maternal anti-Campylobacter antibody titers in serum (A) and cecal mucus (B) of broilers derived from immunized breeders (bacterin, subunit) or breeders with the highest titers encountered in the field (Flock A). Titers are displayed as relative values to the average titers of d 1 from each group, respectively.

The average maternal IgY antibody titer in the mucus at d 1, against a whole cell coating of the C. coli strain isolated from the breeder flock for Flock A chicks, bacterin coating for bacterin chicks and subunit coating for subunit chicks, amounted to 1:352 ± 78 for Flock A chicks, 1:347 ± 64 for bacterin chicks and 1:7,040 ± 25,120 for subunit chicks. At d 18 all mucus samples were under the detection limit for bacterin chicks or at d 9 for the Flock A chicks. For subunit chicks, 50% of the mucus samples were still above detection limit at d 28 (Figure 3B). Some background reaction in the control animals was seen on both d 1 and 4, ranging to 1:40 on bacterin coating and 1:20 on subunit coating. In Flock C some reaction was seen on d 1, ranging to 1:40. The half-life time of maternal anti-Campylobacter antibodies in the mucus is 3.4 d for bacterin and subunit chicks. For Flock A chicks, the half-life time could not be calculated, as there were not enough time points with values greater than 0. Significant difference could not be noticed in the decrease of maternal antibodies in the cecal mucus between Flock A, bacterin and subunit chicks.

Overall, a decline in ratio cecal mucus-to-serum of maternal anti-Campylobacter antibodies could be seen in the bacterin, subunit and Flock A chicks. A transfer around 10% of the antibodies from serum to the cecal mucus was observed at d 1 in the 3 groups (subunit: 11.0%, bacterin: 12.1%, and Flock A: 6.3%). In the following days, antibody transfer declined between 2 and 6% (subunit: 5.9% (d 4); bacterin: 3.0% (d 4) and Flock A 2.6% (d 3)). The ratio was 1% or less after 18 d for subunit chicks, 10 d for bacterin chicks, and 9 d for Flock A chicks, respectively. Differences in decline between groups were not statistically different.

Protective Efficacy of Bacterin- and Subunit-Derived Maternal Antibodies on Cecal Colonization in Broilers Using a Seeder Model

Experimental infection of seeder chickens at 1 wk of age with a dose of 10^4.5 cfu C. jejuni did not lead to a statistical difference in the percentage of Campylobacter positive chickens (both seeders and sentinels) between the bacterin, subunit and control group. With the exception of one seeder animal in one of the bacterin groups (replicate 1), all seeders were positive for Campylobacter 3 d after experimental infection. In addition, most sentinels from the subunit and bacterin group (all sentinels in the replicate experiment) proved positive for Campylobacter (Figure 4A).

Figure 4.

Results of the seeder trial. The broilers, cohoused per group, are progeny from broiler breeders immunized with a bacterin or subunit vaccine. The positive controls are progeny from broiler breeders immunized with the placebo vaccine. In figure A, the percentage Campylobacter positive broilers after experimental infection of the seeders with 10^4.5 colony forming units (cfu) C. jejuni at 1 wk of age. In figure B, the individual and mean numbers of cecal Campylobacter genomic equivalents (log₁₀ GE/g). The experiment was performed in duplicate. The statistical significance between groups is summarized as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.

Also differences with regard to Campylobacter numbers encountered in the ceca could not be observed between the 3 experimental groups, both in seeders and sentinels (Figure 4B).

Protective Efficacy of Bacterin- and Subunit-Derived Maternal Antibodies on Cecal Colonization in Broilers Using a Threshold Model

Next, chicks originating from the immunized broiler breeders were experimentally infected with Campylobacter using a threshold model. After inoculation of the chickens with a dose of 10^2.5 cfu at 1 wk of age, a lower percentage of Campylobacter positive animals was noticed in the subunit and bacterin group compared to the control group (P < 0.0001 and P = 0.001, respectively). No difference in number of positive animals could be seen between bacterin and subunit chicks. After inoculation with a dose of 10^2.5 cfu at 14 d of age, a lower percentage of Campylobacter positive animals compared to the control group could be noticed in chicks from the subunit group (P = 0.015), no significant difference in number of positive animals could be seen between bacterin and control chicks. Inoculation at 21 d of age, results in a lower number of positive chickens in the subunit group compared to the control group, however the difference was not statistically significant (Figure 5A).

Figure 5.

Percentage Campylobacter positive broilers after experimental inoculation with 10^2.5 cfu (A) and 10^3.5 cfu (B). C. jejuni at the age of 7, 14, and 21 d using a threshold model. The individually housed broilers are progeny from broiler breeders immunized with a bacterin or subunit vaccine. The positive controls are progeny from broiler breeders immunized with the placebo vaccine. Since there was no statistically significant difference observed between the bacterin and control group at wk 2, the bacterin group was no longer included in the challenge trial at wk 3. The statistical significance between groups is summarized as **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.

After inoculation with a dose of 10^3.5 cfu, statistical differences in percentage of Campylobacter positive animals could not be observed between any of the experimental groups at any investigated time point (i.e., 7, 14, or 21 d of age) (Figure 5B). This was also the case after inoculation with a dose of 10^4.5 cfu (at 1 wk of age, data not shown). Besides the difference (P < 0.05) with regard to the number of Campylobacter positive chickens using an inoculation dose of 10^2.5 cfu, lower mean numbers of Campylobacter as genomic equivalents (GE) could be detected in the ceca from the bacterin group in comparison with the ceca from the control group at wk 1 (P = 0.023) (Figure 6A). At 21 d of age, the number of Campylobacter GE encountered in the subunit group were also lower than those in the control group (P = 0.021) (Figure 6C). Although the mean number of cecal GE detected in the bacterin group at wk 2 and 3 and in the subunit group at wk 1 and 2 were lower than those in the control groups, the difference proved not statistically significant (Figure 7).

Figure 6.

Individual and mean numbers of cecal Campylobacter genomic equivalents (log₁₀ GE/g) in broilers inoculated with 10^2.5 colony forming units (cfu) C. jejuni at the age of 7 d (A), 14 d (B), or 21 d (C) of age using a threshold model. The individually housed broilers are progeny from broiler breeders immunized with a bacterin or subunit vaccine. The positive controls are progeny from broiler breeders immunized with the placebo vaccine. The statistical significance between groups is summarized as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and ****P ≤ 0.0001.

Figure 7.

Individual and mean numbers of cecal Campylobacter genomic equivalents (log₁₀ GE/g) in broilers inoculated with 10^3.5 colony forming units (cfu) C. jejuni at the age of 7 d (A), 14 d (B), or 21 d (C) of age using a threshold model. The individually housed broilers are progeny from broiler breeders immunized with a bacterin or subunit vaccine. The positive controls are progeny from broiler breeders immunized with the placebo vaccine. The statistical significance between groups is summarized as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, and **** P ≤ 0.0001.

The mean number of Campylobacter GE encountered in chickens of the subunit and bacterin groups were not lower than the ones detected in the chickens of the control groups at any investigated time point using an inoculation dose of 10^3.5 cfu. This was also the case after inoculation with a dose of 10^4.5 cfu (at 1 wk of age, data not shown).

DISCUSSION

All sampled (n = 10) on-farm broiler breeder flocks proved positive for anti-Campylobacter IgY antibodies in the yolk, with titers ranging from 1:200 to 1:3,200. This confirms the widespread contact of breeder flocks with Campylobacter. Yet, although not age correlated, significant differences in titers could be observed between flocks (Flocks A vs. Flock C titers, P < 0.0001). This might be explained by the time of a Campylobacter infection in a certain flock. Campylobacter could be isolated from Flock A (mean yolk titers of 1:3,200 ± 2,267), demonstrating for an ongoing Campylobacter infection. The isolated Campylobacter strain was identified as C. coli ST 830 belonging to the ST-828 complex, a clonal complex also used in the bacterin vaccine and coating. Campylobacter, however, could not be isolated from Flock C (mean yolk titers of 1:200 ± 21). In this flock, lower titers observed might be related to a Campylobacter infection in the past. Messens et al. (2009) showed that multiple genotypes of thermotolerant Campylobacter may be present in a commercial broiler flock during rearing, either due to subsequent introductions of Campylobacter or by frequent mutations of the dominant clones. This can also explain the variations in antibody titers between breeder flocks and the presence or absence of Campylobacter in correlation to the antibody titer. Yet, it cannot be ruled out that the low titers (1:200) observed in certain breeder flocks could be due to ELISA background reactions against the bacterin coating, because low positive signals were also sometimes encountered in our specific pathogen-free (SPF) breeders (Table 4), being regularly checked for Campylobacter. Based on the rather low mean antibody titers encountered in certain breeder flocks (1:200 in Flock B and C), protection conferred by maternal antibodies in the field against Campylobacter infections in these flocks is likely rather limited.

Active immunization of broiler breeders resulted in higher IgY antibody titers than the maximum titers encountered in the field against the bacterin coating (maximum pooled yolk titer of 1:3,200). After the third booster, the bacterin-induced mean yolk titers of 1:9,200 ± 2,375 and the subunit vaccine even mean yolk titers of 1:1,706,667 ± 440,659. However, one should keep in mind that these titers were obtained using ELISA coated with their respective antigens. As mentioned before, the yolks of the on-farm broiler breeders were analyzed using a coating of mixture of Campylobacter strains, which might result in an underestimation (diluting effect) of the titers observed. After the third booster, higher IgY antibody titers could not be achieved in comparison with the second booster. This might indicate that a final booster in the immunization scheme might not be necessary. The titers induced with the bacterin remained stable until at least 8 mo after the third booster (end of the experiment), however a drop of 90% was seen in subunit breeders between the final booster and 4 mo later, after which the antibody titers remained stable until the end of the experiment. Both for the subunit vaccine and the bacterin serum IgY titers are comparable to the yolk IgY titers. That IgY antibody levels in the egg yolk are directly proportional to the antibody levels in the hens’ serum was already observed previously (Loeken and Roth, 1983; Brown et al., 1989; Trampel et al., 2006). The current vaccines have been used in a previous study to immunize ISA Brown layer hens with the same immunization scheme (Vandeputte et al., 2019A). After the third bacterin booster, yolk titers induced in broiler breeders were lower than the yolk titers induced in the layers (1:9,200 ± 2,375 vs. 1:65,536). In contrast, the subunit vaccine resulted in much higher titers in the breeders compared to the layers (1:1,706,667 ± 440,659 vs. 1:65,536). In case of the subunit vaccine, changes to the protocol were made in the current study to optimize the preservation of the recombinant proteins. Differences in IgY antibody titers induced might also be related to the difference in genetics. A difference in antibody response in broiler breeders vs. layers was also seen by Cathraw and Newell (2010), where White Leghorns had a consistent higher antibody response than Ross broiler breeders, after being challenged with the same Campylobacter stain.

The titer of the maternal antibodies (MAB) found in the serum of field chicks, hatched from on-farm breeders, was higher on d 1 compared to the titer of MAB found in bacterin chicks (1:10,880 vs. 1:4,800), although the difference was close to only one 2-fold dilution used in the ELISA assay. This might be explained because of the difference in coating (field strain vs. bacterin coating) or the more immunogenic nature of the colonizing Campylobacter strain of the breeder flock in comparison with the overall immunogenicity of the bacterin strains. The calculated half-life times of the MAB present in the sera of the bacterin and subunit chicks were 3.5 and 5.2 d, respectively. The half-life time of the MAB present in the field chicks was set on 4.0 d. More than half of the bacterin chicks was seronegative after 28 d, for field chicks this was the case at 14 d. For subunit chicks still more than half of the chicks were still seropositive after 28 d. In the study of Cathraw and Newell (2010) commercial chicks had MAB against the acid extractable and outer membrane protein Campylobacter antigens until about 8 d of age.

Previously, Gharaibeh and Mahmoud (2013) investigated the course of MAB titers in chicks derived from breeders vaccinated against a range of poultry pathogens (avian influenza virus, avian encephalitis virus, infectious bronchitis virus, mycoplasma synoviae, …) using commercially available vaccines. Different half-life times for maternal antibodies in the chick's serum against different pathogens were described, ranging from 3.8 to 7 d depending on the pathogen. For most tested pathogens the maternal antibodies in the serum were undetectable after 25 d. In our study the maternal antibodies derived from the Campylobacter subunit vaccine could still be detected in the serum at 28 d of age in more than half of the chicks (endpoint of the trial). During cecal colonization, Campylobacter can be mainly found in the mucus layer (Beery et al., 1988) which it uses as the site of multiplication. Campylobacter only invades the enterocytes temporarily and briefly (Knudsen et al., 2006), probably to hide from mucosal clearance (Van Deun et al., 2008). Therefore, an increased presence of IgY antibodies in the cecal mucus might be essential to protect chickens against Campylobacter infection. This is supported by the results from Vandeputte et al. (2019A). They recently demonstrated that supplementation of feed with hyperimmune egg yolks, loaded with maternal anti-Campylobacter IgY, could prevent the onset of Campylobacter colonization in broilers. Anti-Campylobacter IgY could be detected in the cecal mucus of the bacterin chicks until 18 d of age. In the subunit chicks, albeit at very low levels, maternal antibodies could still be observed in the cecal mucus at 28 d of age (endpoint of the trial). In the field chicks, mucosal antibodies could be observed until 7 d of age. Yet, the ratio mucus-to-serum of maternal antibodies decreased rapidly in the first days of the chick's life, probably due to the fast expansion of the cecal surface size. This decrease in ratio was similar for all 3 groups.

The protective potential of the bacterin- and subunit-induced maternal antibodies against Campylobacter infection was first evaluated using a seeder model. Progeny of the vaccinated broiler breeders was challenged in this model at 1-wk-old with 10^4.5 cfu, which did not result in protection against Campylobacter infection. No difference was observed between the control group and the vaccinated groups, both in terms of the percentage of Campylobacter-positive animals and the degree of colonization. This suggests that induced maternal antibodies are not capable of preventing Campylobacter colonization after challenge with 10^4.5 cfu and that once the first chicks are colonized, horizontal transmission to the other chicks in the flock cannot be prevented. Therefore, it was decided to further evaluate the protective efficacy of the bacterin- and subunit-induced maternal antibodies using a threshold model. A clear protective effect of the vaccine-induced MAB could be noticed after challenge with the lowest challenge dose of 10^2.5 cfu/chicken. In the subunit immunized chicks, the Campylobacter prevalence (% of Campylobacter positive birds) was lower (P < 0.05) in comparison with the control chicks at 7 and 14 d of age. Also, infection loads (number of Campylobacter present in the ceca) were lower (P < 0.05) at 21 d of age. In the bacterin immunized chicks, a protective effect of the MAB was only observed at 1 wk of age after challenge with 10^2.5 cfu Campylobacter. Compared to the control chicks, a reduction (P < 0.05) in Campylobacter prevalence and infection load was obtained. Significant differences between the bacterin, subunit, and control chicks could not be observed at any time point when higher challenge doses were used (10^3.5 and 10^4.5 cfu). The results of these threshold trials indicate that the vaccine-induced maternal antibodies do protect broiler flocks against Campylobacter entry at low dose, but not against intraflock dispersal from infected chicks. When infection doses of 10^3.5 cfu or higher are used, the concentration of MAB in the mucus are probably too low to sufficiently capture the bacteria, which enables Campylobacter to establish itself in the ceca. As evidenced by the results of the seeder trial, this results in a stable population within 24 h, followed by the excretion of very high numbers of bacteria, making the pathogen able to quickly spread throughout the stable. These results are comparable to the results of Cathraw and Newell (2010), where chicks, coming from challenged and unchallenged SPF breeder flocks, were inoculated with a homologous strain 308 (at doses of 6 × 10² to 1.6 × 10³ cfu) at d 1, 8, 15, and 22 and cecal colonization levels were determined 5 d later. In this study initial susceptibility in newly hatched birds was observed, followed by a period of resistance, which was observable from at least 8 to 15 d of age, and after which, susceptibility was reestablished. Similar results were seen with heterologous strains, but to a lesser extent.

To conclude, all sampled breeder flocks were positive for anti-Campylobacter antibodies, yet significant differences in titers were observed between flocks. In some breeder flocks only very low titers were detected. As a result, the protective effect of MAB in the field for the progeny derived from these breeder flocks is likely to be of little relevance. Vaccination of SPF broiler breeder flocks with both subunit and bacterin vaccines results in a prolonged presence of anti-Campylobacter antibodies in the serum and mucus of chicks. Protective efficacy of these vaccines-induced MAB against Campylobacter infection was observed in both bacterin and subunit immunized chicks until 7 and 21 d of age, respectively. The extended efficacy in the subunit chicks is probably related to the higher induced MAB titers in both serum and mucus after vaccination with the subunit vaccine. The vaccine-induced MAB, however, only significantly reduces the prevalence of Campylobacter infections in chicks at low challenge doses (10^2.5 cfu). The concentration of MAB in the mucus are probably too low to sufficiently capture Campylobacter when higher challenge doses are used, which enables Campylobacter to itself in the ceca.