Reduced spread of Campylobacter jejuni in broiler chickens by stimulating the bird's natural barriers

Authors


Correspondence

Birgitte Moen, Nofima – Norwegian Institute of Food, Fisheries and Aquaculture Research, 1430 Ås, Norway. E-mail: birgitte.moen@nofima.no

Abstract

Aim

We have tested the effect of feed structure and feeding regime to prevent the spread of the zoonotic pathogen Campylobacter jejuni in broiler chicken flocks.

Methods and Results

Birds were offered two types of feed, control diet and a diet supplemented with 15% oat/barley hulls for structure. In addition, the birds were either fed ad libitum or intermittent. One bird in each treatment group was infected with a three-strain-mix of Camp. jejuni, and the spread of Camp. jejuni within the group was investigated. Feed structure increased the gizzard weight, delayed the spread of Camp. jejuni within the group and reduced the relative amount of Camp. jejuni in the caecum compared with the control diet.

Conclusions

Our results show that stimulating the bird's natural barriers is a novel and promising intervention strategy to reduce the spread of Camp. jejuni in chicken flocks.

Significance and impact of the Study

Preventing Camp. jejuni in broiler chicken flocks is essential to ensure food safety because this bacterium is transferred to chicken carcasses during the slaughter process and readily survive in unprocessed poultry products. We have evaluated a novel approach for stimulation of the bird's natural barriers in the upper digestive tract with promising results.

Introduction

The zoonotic pathogen Campylobacter jejuni is the leading cause of food-borne bacterial diarrhoeal disease throughout the world (Blaser 1997). Campylobacter jejuni readily colonizes the gastrointestinal (GI) tract of poultry and represents a major problem for the poultry industry and is a potential public health hazard through its high ability to survive in unprocessed poultry products (Chynoweth et al. 1998; Lee et al. 1998; Chantarapanont et al. 2003; Solow et al. 2003). Despite major intervention efforts targeting the lower GI tract, there are currently no really successful strategies for reduction or elimination of this bacterium from the food chain (Hariharan et al. 2004).

The aim of this work was to investigate the effect of feed structure and feeding regime (ad libitum or intermittent) on the spread of Camp. jejuni in broiler chickens. To our knowledge, approaches modifying the upper digestive tract (crop and gizzard) have not yet been evaluated with respect to Campylobacter reduction. The chicken has several natural barriers in the upper digestive tract to kill pathogens. The crop contains lactic acid bacteria, and the feed is soaked (Saris et al. 2007). The purpose of the gizzard is to grind coarse feed particles to a certain size before entering the lower digestive tract (Clemens et al. 1975; Moore 1999; Svihus 2011). The gizzard also contains hydrochloric acid to aid the digestion of the feed and may also have a sterilizing effect where food-borne pathogens might get killed time dependently in the acid environment (Engberg et al. 2002). It is known that the development of the gizzard and gizzard activity is strongly influenced by feed structure and particle size (Svihus and Hetland 2001; Hetland et al. 2002). Feeding coarsely ground feed, whole grains or insoluble fibres have been found to increase gizzard weight significantly (Hetland et al. 2005; Svihus 2011). More specifically, inclusion of 10% oat hulls in broiler diets has been found to stimulate gizzard activity (Rogel et al. 1987; Svihus and Hetland 2001). Feeding of coarsely ground mash also stimulates gastric functions, including secretion of hydrochloric acid, and simultaneously increases the retention time of feed in the proventriculus and gizzard (Engberg et al. 2002; Svihus et al. 2010). It is also known that the physical properties of feed can influence pH, microbial populations and volatile fatty acids (VFA) in the digestive tract of broilers (Engberg et al. 2002). Feed particle size and feed form have been shown to influence the incidence of Salmonella in the gizzard, ileum and caecum of chickens (Bjerrum et al. 2005; Huang et al. 2006; Santos et al. 2008). Nothing has, however, been done to modify the chicken's natural barriers in the upper digestive tract with respect to reducing Camp. jejuni colonization.

Materials and methods

Experimental infections, feeding regimes and feed structure

One hundred and fifty-six-day-old male birds (Ross 308) were raised in a group on commercial feed till 10 days of age. Thirteen birds were then be placed in 12 pens (with access to wood shavings) in one room, where half the birds were given a diet based on coarse feed, and half the birds were given a diet with normal, fine feed. In addition, each diet regime was divided into an ad libitum (ad lib) group and an intermittent feeding (IF) group. Thus, each of the four treatments was given to three pens. The pens were in a room with 18 h light and 6 h darkness (from 02:00 to 08:00). The IF regime the first week consisted of feed being available from 08:00 to 09:00, 12:00 to 13:00, 16:30 to 17:30 and from 21:00 till light went off at 02:00. The IF regime thereafter consisted of feed being available from 08:00 to 09:00, 12:00 to 13:00, 16:00 to 17:00, 20:00 to 21:00 and 24:00 till light went off at 02:00. The experiment was a two-factorial experiment where effect of feeding regime and diet structure was studied. The experimental feed was a diet containing coarse fibre material of cereal origin [15% hulls (7·5% oat, 7·5% barley)] or a diet without this fibre material, pelleted through a 3-mm pellet press. The feed (Table S1) was produced by the Centre for Feed Technology at Norwegian University of Life Sciences.

At 25 days of age, the total numbers of birds in each cage were reduced to 8. All birds were marked so that we could follow them individually, and cloacae swabs were collected to ensure that they were Camp. jejuni free. One bird from each pen was then inoculated orally with a specific Campylobacter mix. Age 25 was chosen partly because the birds needed time to develop the gizzard and that others have shown that the moment of colonization of the first bird in a flock is estimated to be from 21 days of age onward (van Gerwe et al. 2009). All in vivo experiments were approved by the Norwegian governmental committee for experimental animals (http://www.mattilsynet.no/fdu/).

Bacterial challenge procedure

The following Camp. jejuni strains were used in our experiments: C484 isolated from poultry leg (Rudi et al. 2005), G109 isolated from caecal dropping (Skånseng et al. 2007) and G125 isolated from dog faeces (Skånseng et al. 2007). The two latter strains are isolated from broiler farm environment (Johnsen et al. 2006), and all three strains have recently shown to be excellent colonizers of chickens (Skånseng et al. 2010). A mixture of the three strains was used to ensure that at least one of the strains would colonize the chickens and to simulate a natural infection. Direct sequencing of the gltA gene was later used to distinguish the Camp. jejuni strains. To prepare the inoculum, the strains were first grown microaerobically (CampyGen; Oxoid Ltd, Basingstoke, UK) at 42°C for 48 h on blood agar with Campylobacter Growth Supplement (Oxoid, SR0232). A single colony was inoculated into 5-ml Mueller Hinton (MH) broth (Oxoid Ltd) and incubated microaerobically at 42°C for 48 h. This culture was then diluted a 100-fold in buffered peptone water (BPW) and incubated at 37 ± 1°C for 24 h. The cultures were then mixed 1 : 1 : 1 and the inoculation mix contained approximately 107 CFU ml−1 of each strain. The birds were inoculated with 1 ml of the inoculation mix by crop instillation using a 2-ml syringe with an attached flexible tube.

Data collection

Feed and birds were weighed on 11th, 18th, 25th, 32th and 34th day of age, in the morning before having access to feed at 08:00. Growth performance was not impaired, as previously reported (Sacranie et al. 2012). Samples (swabs from the cloaca of each bird) were collected immediately before inoculation and examined for the presence of Camp. jejuni. All samples were found to be negative for Camp. jejuni. Swabs from all birds were thereafter taken one time per day for 5 days after inoculation (others have estimated that 95% of all birds in a flock will be positive 4–7 days after the first bird is infected (van Gerwe et al. 2009)). On day 9 postinoculation (p.i.) (birds were 34 days of age), four birds per pen were randomly collected for dissection. The weights of the gizzards (empty and full) were registered. In addition, the caeca were collected and immediately frozen. Luminal contents of the caeca were used for the quantitative real-time PCR examination of Camp. jejuni. Birds inoculated with Camp. jejuni appeared healthy and showed no signs of disease [although one of the infected birds (pen 9) died].

Campylobacter jejuni quantification

Cultivation procedures

Qualitative detection (positive/negative), after cultivation, of Camp. jejuni was chosen for the cloacal swabs to ensure the sensitivity because of variable amounts of cloacal content retrieved by swabs. Briefly, each cloacal swab sample was immersed in individual test tubes containing 5-ml Campy medium [500 ml of Campylobacter growth broth consisted of 475 ml Nutrient broth no. 2 (Oxoid, CM0067) supplemented with 25 ml Laked Horse Blood (Oxoid, SR0048C), one ampule of Campylobacter Growth Supplement (Oxoid, SR0232) and one ampule of CCDA Selective Supplement (Oxoid, SR0155)] and incubated microaerobically at 42°C for 48 h to cultivate Campylobacter. DNA was isolated from this cultivation medium (see DNA isolation below), and this DNA was later used as template in the qualitative detection of Camp. jejuni by real-time PCR (positive/negative). A sample was registered as positive for Camp. jejuni if the real-time PCR gave a positive detection during the 40 cycles.

DNA isolation

One hundred and fifty microlitres of the cultivation medium was mixed with 450 μl 4 mol l−1 guanidinium thiocyanate (GTC). DNA isolation and purification were further performed using an automated procedure with silica particles (Bioclone Inc., San Diego, CA, USA) as described earlier by Skånseng et al. (Skånseng et al. 2006). Briefly, the bacterial cells were lysed by mechanical lysis (FastPrep; Qbiogene Inc., Carlsbad, CA, USA), and DNA was purified using magnetic beads. At the day of dissection (day 9 p.i.), the caeca were collected and swabs with caecal lumen contents were mixed separately with 1 ml of Solution 1 (25 mmol l−1 Tris–HCl pH 8·0, 10 mmol l−1 EDTA pH 8·0). One hundred and fifty microlitres of this solution was mixed with 450 μl 4 mol l−1 GTC, and DNA isolation was further performed as described earlier.

Quantitative real-time PCR

Quantification of Camp. jejuni, relative to the total flora, was performed on DNA isolated directly from the caecal lumen contents. Quantification of Camp. jejuni was performed relative to the total flora as previously described by Skånseng et al. (Skånseng et al. 2006). Briefly, standard curves were made for both the Camp. jejuni and universal 16S rDNA primer-probe sets. Campylobacter jejuni NCTC 11168 DNA was serially diluted 10-fold in purified caecum DNA and subjected to real-time PCR. The standard curves were used for estimating the amount of Camp. jejuni (CJ) relative to the total flora (U16S), that is, [(Ct (CJ) × 1/a − b/a) − (Ct (U16S) × 1/a − b/a)] where a and b are the regression coefficients from the standard curves for Camp. jejuni and universal 16S rDNA primer-probe sets. Regression curves for all primer and probe sets had a square regression coefficient of 0·99. Universal 16S rDNA primers and a probe (Nadkarni et al. 2002) were used for the quantification of the total bacterial flora. A Camp. jejuni-specific real-time PCR was performed using the primer and probe set described by Nogva et al. (2000). The real-time PCR was performed as previously described by Skånseng et al. (2007). Briefly, universal 16S rDNA real-time PCR contained 0·2 μmol l−1 of each primer, 0·1 μmol l−1 probe, 1 U DyNAzyme II Hot Start DNA Polymerase (Finnzymes Oy, Espoo, Finland) and 0·5 μl DNA in a 25-μl PCR. Campylobacter jejuni-specific real-time PCR contained 0·3 μmol l−1 of each primer, 0·02 μmol l−1 probe, 1 U DyNAzyme II Hot Start DNA Polymerase and 2 μl DNA in a 25-μl reaction. The amplification profile was 40 cycles of 95°C for 30 s and 60°C for 1 min, with an initial heating step of 94°C for 10 min. The reactions were performed in an ABI Prism 7900 HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA), and the data were analysed using the SDS 2.2 Software (Applied Biosystems).

Direct sequencing of the three Campylobacter jejuni strains

Amplification of the Camp. jejuni gltA genes in the caecum samples was performed using glt1F and glt1R (Berget et al. 2007) as previously described by Skånseng et al. (Skånseng et al. 2007). The PCR amplification contained 1× Hot Start Buffer (Finnzymes), 200 μmol l−1 dNTP mix, 1U DyNAzyme II Hot Start DNA Polymerase (Finnzymes), 0·2 μmol l−1 of each primer and 1 μl DNA in a 25-μl reaction. The amplification profile was an initial step of 94°C for 10 min, then 35 cycles of 94°C for 30 s, 50°C for 2 min, and 72°C for 30 s and a final extension at 72°C for 7 min. Purification of the PCR products were performed using 0·4 μl of ExoSap-IT (USB Corp., Cleveland, OH, USA) to 5 μl of PCR product, with a thermal profile of 37°C for 30 min and 80°C for 15 min. The sequencing reaction contained 0·75× BigDye® v1.1/3.1 Sequencing Buffer (Applied Biosystems), 1 μl BigDye® Terminator v1.1 Cycle Sequencing Kit, 0·25 μmol l−1 of primer glt1F and 1·0 μl of purified PCR product in a 10-μl reaction. The sequencing reactions were carried out in 25 cycles of 96°C for 15 s, 50°C for 10 s and 60°C for 4 min. BigDye XTerminator Purification Kit (Applied Biosystems) was used according to the manufacturer's recommendations to clean up the sequencing reactions. Sequencing was performed on an ABI Prism 3130xl Genetic Analyzer (Applied Biosystems).

Statistical analysis

Chi-square tests (Systat 12; Systat Software, Inc, San Jose, CA, USA) were performed for each day, to see whether there were differences with respect to Camp. jejuni-positive and Camp. jejuni-negative birds between the four feeding regimes: coarse and fine feed, and intermittent and ad libitum feeding. Four separate linear regressions were performed on the data, pens with 100% positive birds being kept out after the first registration (Systat 12).

An analysis of variance (anova) was performed on the results from the relative Camp. jejuni values (caecum samples) and the weight of the empty gizzard [MiniTab® 16.1.0.0 software (Minitab Inc., State College, PA, USA)]. A correlation test was performed between the relative Camp. jejuni values and the empty gizzard weight (MiniTab® 16.1.0.0). Pen 9 (IF, coarse feed) was removed from the analysis because of the death of the inoculated bird. The relative proportions of Camp. jejuni strains were determined by multivariate decomposition of mixed gltA gene sequence electropherograms according to the direct PLSR method, as previously described (Trosvik et al. 2007).

Results

Effect of feeding regimes and feed structure on Campylobacter jejuni colonization and gizzard weight

Birds given hulls for feed structure had delayed spread of Camp. jejuni in the group compared with birds given a control diet (Table 1). There was a significant difference between Camp. jejuni-positive and Camp. jejuni-negative birds (based on real-time PCR of cloacal swabs after enrichment) over time between the four feeding regimes (P < 0·001 for day 2, 3 and 4 p.i., respectively, and P = 0·002 for day 5 p.i.), and between fine and coarse feed (P < 0·001 for day 2, 3 and 4 p.i., respectively, and 0·003, for day 5 p.i.). Pens where coarse feed were used had fewer Camp. jejuni-positive individuals over time than pens where fine feed were used. There was no significant difference between the two feeding regimes (ad libitum and IF), although the P-value for day 3 p.i. was very close to 0·05. (P = 0·459, 0·051, 0·3507 and 0·887, for day 2, 3, 4 and 5 p.i., respectively). Day 1 p.i. was not included in the calculations because all birds were negative. The results from the linear regressions (Fig. 1) demonstrate that fine feed leads to a faster increase in Camp. jejun-positive birds, with an estimated 100% infected birds within <4 days. There was a tendency of birds fed coarse feed intermittently having a slower increase in Camp. jejuni-positive birds compared with birds fed coarse feed ad libitum. There was a delayed detection of Camp. jejuni of the inoculated bird in pens 2 and 4 (coarse feed). In pen 2 and 4, the inoculated bird was positive on day 3 and day 2 p.i., respectively. For all other pens, the inoculated bird was positive on day 1 p.i. Four birds per pen were randomly collected for dissection and Camp. jejuni quantification (relative to the total flora) of caecal lumen contents by real-time PCR. Increased feed structure gave significantly lower levels (P < 0·001) of Camp. jejuni compared with the total flora of the caecum at the end of the experiment (day 9 p.i.) (Table 2 and Fig. 2). Feeding regime did not have a significant effect on the spread of Camp. jejuni in the flocks or the final Camp. jejuni levels in the caeca. Quantification of the colonizing Camp. jejuni strains at day 9 p.i. is shown in Fig. S1, and all birds were colonized by at least two of the challenge strains at the end of this experiment. There was no correlation between colonizing Camp. jejuni strains in caecum at day 9 p.i. and feeding regimes or structure. The dominating strain was G125 except in one of the pens fed intermittent fine feed and four of eight pens fed ad libitum coarse feed where strain C484 dominated.

Figure 1.

Regression analyses, percentage of Campylobacter jejuni positive birds over time (days p.i.). Four separate linear regressions were performed on the data, pens with 100% positive birds being kept out after the first registration. That is if a pen was registered as 100% positive on 1 day, the succeeding days was not included in the analysis. Although the pen effect is not included in the model, its contribution to the noise is included since the regressions are based on results from the individual pens. Grey symbols = Ad lib feeding; black symbols = intermittent feeding (IF); circle symbol and dotted line = coarse feed and triangle symbol and solid line = fine feed. The regression coefficients were: Ad lib, coarse: y = 20·03 × days-50·53; Ad lib, fine: y = 71·50 × days-143·00; IF, coarse: y = 12·24 × days-20·61; IF, fine: y = 61·67 × days-123·33.

Figure 2.

Colonization of Campylobacter jejuni in the chicken caecum at day 9 p.i. The amount of Camp. jejuni ralative to the total flora (log10) measured by real-time PCR (the relative values were calculated using the regression lines from the standard curves for Camp. jejuni and the total flora (using a universal 16S rDNA primer-probe set) as described in the material and method section). The mean value for each treatment is marked with (−). Campylobacter jejuni was not detected in the samples marked in grey (a Ct-value of 40 was used to calculate relative values in these samples).

Table 1. Campylobacter jejuni colonization over time. The number of positive (pos) compared with total (tot) birds (cloacal swabs) in the different treatment groups (IF fine, IF coarse, ad lib fine, ad lib coarse) at day 2–5 p.i. Day 1 p.i. was not included because all birds were negative. The data presented were obtained by real-time PCR on cloacal swabs after enrichment
  Day 2 p.i.Day 3 p.i.Day 4 p.i.Day 5 p.i.
TreatmentPenPos/totPos/totPos/totPos/tot
  1. IF, intermittent feeding.

  2. a

    The total number of chickens at day 5 p.i. is reduced by one bird because of the death of one bird in one of the pens (IF, coarse).

  3. b

    Pen 9 (ad lib, fine) was removed from the analysis because of the death of the inoculated bird.

IF fine14/77/77/77/7
34/77/77/77/7
55/77/77/77/7
Sum13/2121/2121/2121/21
IF coarse20/70/70/72/6
40/64/63/64/6
62/76/77/77/7
Sum2/2010/2011/2014/19a
Ad lib fine76/77/77/77/7
114/77/77/77/7
Sum10/14b14/1414/1414/14
Ad lib coarse80/70/71/74/7
100/72/73/76/7
120/73/76/76/7
Sum0/215/2110/2117/21
Total 25/7650/7656/7666/75a
Table 2. Campylobacter jejuni colonization relative to the total flora in caecum and empty gizzard weight. Colonization of Camp. jejuni relative to the total flora in caecuma (log10), measured by real-time PCR and empty gizzard weight (g) at day 9 p.i
 IF fineAd lib fineIF coarseAd lib coarseP-value
DietFeedingD × F
  1. IF, intermittent feeding.

  2. a

    Four birds per pen were randomly collected for dissection and Camp. jejuni quantification by real-time PCR. Pen 9 (ad lib, fine) was removed from the analysis because of the death of the inoculated bird, resulting in only eight birds in this group. Some samples had high Ct values for the universal 16S rDNA PCR (probably due to some error in the sample preparation) and were not included in the calculation: IF fine n = 1, IF coarse n = 2 and ad lib coarse n = 2. Duplicate real-time PCR were performed for each sample and the average used for further calculation. The relative values were calculated using the regression coefficients from the regression lines from the standard curves for Camp. jejuni and the total flora (using a universal 16S rDNA primer-probe set) as described in the material and method section.

Rel. Camp. jejuni−2·90 ± 0·48 (n = 11)−2·34 ± 0·66 (n = 8)−4·12 ± 1·15 (n = 10)−3·70 ± 1·32 (n = 10)<0·0010·1260·815
Empty gizzard weight (g)33·58 ± 3·52 (n = 12)27·20 ± 7·92 (n = 8)51·08 ± 6·21 (n = 12)44·48 ± 9·26 (n = 12)<0·0010·0040·959

The empty gizzard weight was significantly higher in birds given coarse feed in both feeding regimes, IF giving the highest weight. The empty gizzard weight was also significantly higher in birds that were fed intermittent compared with birds fed ad libitum (Table 2 and Fig. 3). There was a negative correlation between the relative amount of Camp. jejuni and the empty gizzard weight; Pearson correlation −0·614, P < 0·001. Performance was normal for all birds, but weight gain tended (not significant) to be lower for birds given coarse feed structure.

Figure 3.

Empty gizzard weight (g). The mean value for each treatment is marked with (−).

Discussion

In this work, we have demonstrated that chickens offered feed with increased structure (oat/barley hulls) were better protected against Camp. jejuni during the infection trial than birds offered control feed. Feed structure resulted in a delayed spread of Camp. jejuni between the birds and also in a lower abundance of Camp. jejuni in the caecum (relative to the total flora). The use of Camp. jejuni relative to the total flora has previously been shown to correlate with traditional cultivation (Skånseng et al. 2010), and screening of changes in the total flora by direct sequencing [an in-house developed method to analyse changes in dominating bacterial flora (Trosvik et al. 2007; Zimonja et al. 2008)] indicated that the dominating bacterial flora was not changed by the diets (data not shown). Increased particle size has previously been shown to reduce the numbers of Salmonella in poultry (Bjerrum et al. 2005; Huang et al. 2006; Santos et al. 2006, 2008). An interesting observation was the negative correlation between the empty gizzard weight and the relative amount of Camp. jejuni. It has previously been reported that increased feed structure stimulated the gizzard resulting in a larger and more active gizzard. Coarse feed has also been shown to reduce the gizzard pH (Engberg et al. 2004; Bjerrum et al. 2005; Svihus 2011). When Camp. jejuni enter the gizzard, the bacteria might get killed time dependently in the acid environment due to the fact that coarse hull particles are retained in the gizzard until they are ground to a certain critical size allowing them to pass through the pyloric sphincter (Clemens et al. 1975; Moore 1999; Hetland et al. 2002, 2003). The gizzard activity will determine the amount of Camp. jejuni killed in the upper digestive tract and correspondingly the chances to colonize the lower digestive tract. There was a significant effect of feeding on the empty gizzard weight, probably caused by birds consuming more litter when fed intermittent. There was also a tendency of birds fed intermittently having a lower relative abundance of Camp. jejuni in the caecum. A positive effect of intermittently feeding on Camp. jejuni colonization may be caused by feed being retained in the crop during periods without access to feed. The crop contains lactic acid bacteria that may have influenced the survival of Camp. jejuni (Saris et al. 2007).

We conclude that stimulation of the gizzard by increased structure in feed is a promising intervention strategy to reduce the zoonotic pathogen Camp. jejuni in poultry. Our hypothesis is that reduction in Camp. jejuni in the upper digestive tract reduces the ability of this bacterium to colonize the lower digestive tract, thereby reducing the load in poultry products.

Acknowledgements

This work was supported by Grant 178267/I10 from the Norwegian Research Council, The Norwegian Centre for Poultry Science and the Found for Research Levy on Agricultural Products and Research funds from the Norwegian Agricultural Authority. The authors want to thank Nortura SA for supporting the project, Merete Rusås Jensen, Janina Berg and Tove Maugesten for excellent laboratory assistance, Adam Sacranie and Frank Sundby for assistance with the animals, Per Lea and Pål Trosvik for statistical assistance.

Ancillary