A role for flies (Diptera) in the transmission of Campylobacter to broilers?

SUMMARY Campylobacter is the leading cause of bacterial diarrhoeal disease worldwide, with raw and undercooked poultry meat and products the primary source of infection. Colonization of broiler chicken flocks with Campylobacter has proved difficult to prevent, even with high levels of biosecurity. Dipteran flies are proven carriers of Campylobacter and their ingress into broiler houses may contribute to its transmission to broiler chickens. However, this has not been investigated in the UK. Campylobacter was cultured from 2195 flies collected from four UK broiler farms. Of flies cultured individually, 0·22% [2/902, 95% confidence interval (CI) 0–0·53] were positive by culture for Campylobacter spp. Additionally, 1293 flies were grouped by family and cultured in 127 batches: 4/127 (3·15%, 95% CI 0·11-6·19) from three broiler farms were positive for Campylobacter. Multilocus sequence typing of isolates demonstrated that the flies were carrying broiler-associated sequence types, responsible for human enteric illness. Malaise traps were used to survey the dipteran species diversity on study farms and also revealed up to 612 flies present around broiler-house ventilation inlets over a 2-h period. Therefore, despite the low prevalence of Campylobacter cultured from flies, the risk of transmission by this route may be high, particularly during summer when fly populations are greatest.


INTRODUCTION
Campylobacter is the leading cause of human bacterial diarrhoeal disease worldwide. In the European Union (EU) alone, 214 779 cases were reported in 2013 [1]. However, due to under-reporting, this figure underestimates the true impact of the disease. For example, in the UK only one in ten cases are reported to national surveillance schemes [2]. Aside from possible complications of acute infection, campylobacteriosis can lead to a number of chronic sequelae [3]. These significant public health implications are coupled to a serious socioeconomic burden, with campylobacteriosis costing the EU 0·35 million disability-adjusted lifeyears and €2·4 billion per annum [4].
The majority (>90%) of cases are caused by C. jejuni, with the second most frequently isolated species, C. coli, accounting for a further 5-10% of cases [5]. Both species colonize the intestinal mucosa of food-producing animals, including poultry [3]. Raw chicken meat can have very high (>10 7 cells per carcass) contamination levels [6] and may account for up to 70% of cases of campylobacteriosis [7]. An EU-wide survey conducted on Campylobacter in broiler chickens at slaughter, found 75·3% of broiler batches colonized with Campylobacter in the UK, Europe's largest producer of broiler meat [8]. The prevention of broiler flock colonization has become a European food safety priority.
However, Campylobacter colonization of broiler flocks is difficult to prevent even if good biosecurity is maintained, particularly during the summer months. Campylobacteriosis in temperate regions, such as the UK, has distinct seasonality, with the peak incidence of human cases occurring in summer [9]. This correlates with the peak in broiler Campylobacter colonization, which occurs in late spring and summer [10]. This peak in flock colonization remains unexplained and hypotheses include seasonal changes in climatic factors, which concurrently result in peak fly (Diptera) populations during summer months [10][11][12].
It has been established previously that flies are carriers of Campylobacter spp. and are proven vectors of other enterobacterial pathogens [13][14][15]. Flies act as mechanical vectors, transmitting microorganisms as fomites, and as biological vectors, multiplying pathogens within the gut [13,16]. Campylobacter spp. are common in the farm environment and livestock faeces represent a major environmental source [17,18]. Studies conducted within the poultry farm environment have demonstrated the carriage of Campylobacter spp. by flies [13,19,20]. In addition, flies have been shown to transmit C. jejuni to chickens under laboratory conditions [21], and molecular typing in a Danish field study strongly implicated flies in the transmission of a strain of C. jejuni from sheep to broilers [22]. Further, Danish studies strongly suggest that the ingress of flies into broiler houses contributes to the transmission and colonization of broiler chickens with Campylobacter spp. [23,24].
While data are available from several countries to support the hypothesis that flies are implicated in the epidemiology of Campylobacter colonization of broiler flocks, field studies have not previously taken place in the UK and there are no data on the carriage of Campylobacter spp. by flies on and around UK broiler farms. Therefore, this study aimed to investigate the role of flies in the transmission of Campylobacter spp. to broiler chickens in the UK.

Collection and identification of flies
Flies were caught on four broiler farms within a 2·5 mile radius in North Wales, UK, from July to August 2011. The four broiler farms consisted of three or four broiler houses stocking about 26 000-30 000 broilers per house per flock cycle. All four farms operated an all-in/all-out system, with thinning at around day 35 and a 36-to 42-day crop cycle. The houses operated a tunnel ventilation system. All four broiler farms were part of food assurance schemes requiring adherence to a number of strict biosecurity protocols, including, but not limited to, restricted and monitored access, an anteroom at the entrance to the broiler house containing a physical barrier delineating a biosecure area, widespread use of disinfectant footbaths, farm-and broilerhouse-specific clothing and footwear and rigorous policies preventing the introduction and spread of disease. The four farms were in close proximity to livestock, including ruminants and horses. Live flies were collected for culture during at least one whole crop cycle on each farm, by repeat visits once or twice weekly. At each sampling event, flies were individually collected in sterile re-sealable plastic bags outside broiler houses (<10 m periphery) for 2 h. An open bag was placed into a plastic jar and the jar placed over a fly resting on a surface to capture it. The bag containing the fly was removed from the jar and transported to the laboratory for processing.
Captured flies were anaesthetized through placement at −18°C for 10 min and identified taxonomically to family level under a microscope. Dipteran families were categorized as (i) filth flies, (ii) livestock, dung or carrion associated flies and (iii) other flies. Filth flies commonly consume and breed in excreta and carrion and include those of the families Calliphoridae, Fanniidae and Muscidae [14]. Where possible, the species of Calliphoridae, Fanniidae or Muscidae were determined, as these have previously been most highly implicated in the carriage of Campylobacter spp. [12,14,22]. Category (ii) contains families comprising a large proportion of species dependent upon animal excrement or organic matter for part of their lifecycle, i.e. mating, oviposition or feeding. The families contained in category (iii) are generally not associated with livestock, dung or carrion.

Bacterial isolation and identification
Following identification, each fly was individually and manually macerated for 10 s, enriched in 10 ml modified Exeter selective enrichment broth [1100 ml nutrient broth, 11 ml lysed defibrinated horse blood, Campylobacter enrichment supplement SV59 (Mast Group Ltd, UK) and Campylobacter growth supplement SV61 (Mast Group Ltd)] and incubated under microaerobic conditions at 42°C. After 24 h and 48 h incubation, 5 µl of enrichment broth was streaked onto Campylobacter-selective blood-free agar [modified charcoal-cefoperazone-deoxycholate agar (mCCDA)] supplemented with cefoperazone (32 mg/l) and amphotericin B (10 mg/l). These plates were incubated microaerobically at 42°C for 48 h and up to four colonies per plate were selected based on morphology and streaked onto Columbia agar containing 5% (v/v) defibrinated horse blood. Plates were incubated for 48 h under both microaerobic conditions at 42°C and aerobic conditions at 37°C, to distinguish morphologically similar Campylobacter and Arcobacter species. Four Campylobacter isolates were stored at −80°C in Microbank™ vials (Pro-Lab, UK). All media were obtained from Lab M Ltd (UK) and all blood from Southern Group Labs (UK).
Genus-and species-specific polymerase chain reaction (PCR) assays were performed on the four selected suspected Campylobacter colonies. To extract DNA, ∼10 µl of bacterial cells were suspended in 300 µl chelex solution [20% w/v in 100 ml Trizma HCl (Sigma-Aldrich, UK)] and heated at 95°C for 10 min, centrifuged at 16200 g for 3 min and 50 µl of the clear supernatant added to 450 µl sterile distilled water. A confirmatory multiplex 16S rDNA PCR assay for Campylobacter and Arcobacter spp. [25,26] was followed by a multiplex C. jejuni and C. coli specific PCR assay [27]. Each PCR reaction consisted of 21 µl of 1·1 × (2·5 mM MgCl 2 ) ReddyMix PCR Master Mix (Thermo Scientific, UK), 1 µl of bovine serum albumin (Sigma-Aldrich), 0·25 µl of each primer and 2 µl template DNA. DNA amplification was performed for 30 cycles using an annealing temperature of 50°C. PCR amplicons were visualized after electrophoresis on a 2% agarose gel (Alpha Laboratories, UK).

Data collection 2
Collection and identification of flies A second sampling phase was conducted on the same four broiler farms sampled in data collection 1 from June to August 2012. As described, flies were collected on farm and transported to the laboratory. Captured flies were anaesthetized with CO 2 gas and identified taxonomically to family level. While anaesthetized, flies were grouped into batches of ∼10 flies of similar dipteran families using sterile forceps.

Bacterial isolation and identification
Following grouping of flies into batches, microbiological culture and PCR of Campylobacter spp. was carried out as in data collection 1.

Broiler chicken flock sampling and testing for Campylobacter
One broiler chicken flock on each of the four farms was tested daily for Campylobacter from June to August 2012. To sample the broiler flock, one pair of disposable fabric boot socks (overshoes) was worn over rubber boots as the sampler walked 100 m inside the broiler house. Boot socks were premoistened with sterile physiological saline to allow maximum uptake of Campylobacter from the litter. Each pair of boot socks was then placed into an individual sterile re-sealable plastic bag for transport. In the laboratory, 200 ml sterile physiological saline was added to each pair of boot socks and agitated by hand to release attached matter. Boot socks and saline were left to settle for 10 min, then 1 ml clear upper fluid was transferred to a sterile 1·7 ml Eppendorf tube and centrifuged at 16200 g for 7 min. The supernatant was discarded and DNA extracted from the pellet using the Promega Genomic DNA Wizard Extraction kit (Promega, UK). Presence of Campylobacter was confirmed by performing PCR as in data collection 1. To confirm flock positivity following a positive PCR result, 30 swabs of excreted chicken faeces were collected from the broiler flock. Faecal swabs were streaked directly onto mCCDA and incubated microaerobically at 42°C for 48 h. If <28 swabs were Campylobacterpositive by culture, a further 30 faecal swabs were collected every other day until at least 28 positive swabs were obtained. Additionally, the flock continued to be boot-sock sampled daily until Campylobacter had been isolated from at least 28 faecal swabs. Finally, ten Campylobacter isolates were randomly selected from these final 28-30 cultures to represent ten individual faecal samples. Isolates were stored at −80°C in Microbank™ vials (Pro-Lab).

Multilocus sequence typing (MLST) analysis
To assess genetic diversity, MLST profiles were obtained for C. jejuni and C. coli isolates. From the four isolates stored from each Campylobacter-positive batch of flies and ten isolates from each Campylobacter-positive flock, two isolates were randomly selected for MLST from each positive batch of flies or broiler flock. These were subjected to the Campylobacter-specific MLST scheme as described by Miller et al. [28], with alternative PCR and sequencing primers used for the tkt locus where necessary [29]. Sequence data were analysed using ChromasPro (Technelysium Pty Ltd, Australia) and allele numbers and sequence types (STs) assigned using the Campylobacter MLST website (developed by Keith Jolley and sited at the University of Oxford; http://pubmlst.org/campylobacter/) [30].

Malaise trap survey of flies (Diptera)
Malaise traps were used for community analysis to survey the species and numbers of Diptera gaining access to the broiler houses through the ventilation inlets and to analyse the diversity of dipteran species present on the study farms. A Malaise trap is a tent-like structure, where flies and other insects fly into the tent wall and are funnelled up to the highest point and into a collection vessel. Two Malaise traps were placed perpendicular to broiler-house walls, at a right angle to the insect flight line, at each sampling event for 2 h. Seventy per cent ethanol was used in the collection vessel as a killing agent and preservative prior to identification.

Data collection 1
Culture of Campylobacter spp.
Flies were sampled on each broiler farm 6-9 times, totalling 29 sampling events on all four farms. A mean of 31 flies were caught during each sampling event. A total of 902 flies, representing 28 families, were collected and cultured individually (Supplementary  Table S1). Two flies were found to be carrying Campylobacter spp.; a Calliphora vomitoria or 'bluebottle' was positive for C. lari and a fly of the Heleomyzidae, a dipteran family associated with livestock, dung or carrion, was found to be carrying C. coli. The prevalence of Campylobacter spp. from flies was calculated to be 0·22% [2/902, 95% confidence interval (CI) 0-0·53].

Data collection 2
Culture and MLST of Campylobacter spp.
Flies were sampled on each broiler farm 5-6 times, totalling 21 sampling events on all four farms. A mean of 62 flies were caught during each sampling event. A total of 1293 flies, representing 29 families, were collected and cultured in 127 batches with a mean of ten flies per batch (range 1-17) (Supplementary Table S2). Batch size varied due to grouping of similar families of Diptera to aid taxonomical identification of dipteran species carrying Campylobacter spp.

Malaise trap survey of flies (Diptera)
Two Malaise traps were set up during 20 sampling events, collecting a total of 1771 insects (class: Insecta), including 1644 flies, from 28 dipteran families ( Table 2, Supplementary Table S3). The range of Diptera caught in one sampling event was 0-612.

DISCUSSION
This is the first study to investigate the prevalence of Campylobacter spp. carried by flies on UK commercial broiler farms and demonstrates that flies are carriers of Campylobacter in this environment. The prevalence of Campylobacter spp. was calculated as 0·22% (2/902, 95% CI 0-0·53) from flies cultured individually in data collection 1 and 3·15% (4/127, 95% CI 0·11-6·19) from batches of flies cultured in data collection 2. Following data collection 1, it was decided to group flies into batches for logistical purposes. However, this prevented determining prevalence at the individual fly level. In data collection 1, a fly of the Heleomyzidae was found to be carrying C. coli and a Calliphora vomitoria was positive for C. lari. This indicates that positive batches may be the result of a single fly carrying Campylobacter, suggesting that the prevalence of Campylobacter from flies is lower than the estimated 3·15% batch prevalence. There was one minor difference in the microbiological methodology for the two data collections; in data collection 1, flies were anaesthetized by being placed at −18°C for 10 min; however, in data collection 2, flies were anaesthetized with CO 2 gas. This difference in anaesthetic technique is not anticipated to have affected the rate of recovery of Campylobacter spp. from flies. Ultimately, carriage of Campylobacter by flies is at low frequencies in the broiler-farm environment.
Few studies have caught flies in numbers comparable to this study. However, prevalence estimates compare favourably to European studies with similar sample sizes. Hald et al. [12] caught 2186 flies around five Danish broiler farms and found 31 (1·1%) to be positive for Campylobacter spp. However, this estimate may have been influenced by a single broiler farm, which also had a pig farm on the same site, as very few positive flies were found (<1·0%) on the four farms without additional livestock.
In data collection 2, isolates from Campylobacterpositive broiler flocks and batches of flies were typed using MLST. No direct correlation of STs was found between isolates from flies and broilers. However, the STs of three flock isolates could not be determined as amplicons were not produced with any MLST loci primers. One C. jejuni ST [ST257 (CC257)] and one C. coli ST [ST828 (CC828)] were identified in three flocks from the broiler farms. The STs identified in the broiler flocks belong to chickenassociated CCs [32]. Three C. jejuni STs (ST25, ST137 and ST1701) belonging to CC45 were identified in four batches of flies from three broiler farms. CC45 is associated with human enteric illness and food animal populations, particularly chicken flocks [32]. While flies were only found positive for Campylobacter before the broiler flock on one occasion (Table 1), this indicates that flies may transmit Campylobacter to humans either directly or indirectly via broiler chickens.
Transmission of Campylobacter spp. from flies to broilers may be successful if a source exists within a short distance of the broilers and there is a consistently high turnover rate of new flies acquiring the pathogen [33,34]. This is viable in the UK due to the general proximity of pastured livestock to broiler farms. The four broiler farms in this study were all in close proximity to livestock, including ruminants and horses. Given the association of CC45 with food animal populations, the origin of the C. jejuni isolated from flies in this study may have been the surrounding livestock. Previous studies have provided evidence that flies are carriers of Campylobacter over short distances; Hald et al. [22] discovered the same macrorestriction PFGE fingerprint in 27/28 C. jejuni isolates from broilers, flies and sheep on one farm and Stern et al. [35] found flaA type 15 in both flies and chicken intestinal samples from one broiler farm. A Norwegian study found a clustering of broiler flocks up to 4 km apart positive for Campylobacter spp. in summer, which indicates the presence of factors acting on a narrow scale, such as flies [11].
The Malaise traps captured a high diversity of flies, totalling 28 families of Diptera. The Malaise traps do not confirm which fly families and species are entering broiler houses and in what frequencies. However, up to 612 flies were caught adjacent to the broiler-house ventilation inlets in just 2 h, presenting multiple opportunities for flies to enter the house. This supports the findings of other studies [12,22,31], which also conclude that despite the low prevalence of Campylobacter spp. from flies, risk of transmission is potentially high as flies are present in large numbers around broiler-house ventilation inlets. It has been estimated that the average influx of insects into a broiler house per broiler rotation is 30 728 insects (range 2233-180 300) [12]. Twenty-one per cent of these were flies, of which 20·3% were filth flies and the remaining majority composed of families that are attracted to livestock, dung or carrion. A similar distribution was seen in this study; where in total, 86·9% of flies caught in Malaise traps on broiler farms were associated with livestock, dung or carrion; however, only 1·5% were filth flies. Increased ventilation airflow during warmer weather also increases the likelihood of insects being carried passively into houses [10]. Additionally, if the Campylobacter prevalence from flies is compared to other vector-borne diseases, similarly low prevalences are observed, which indicates that the prevalence seen in this study may be sufficient for flies to be a regular source of colonization for the broiler flock. For example, 0·114% (n = 70 937) of flies were carrying Shigella spp. in Southwestern USA where flies were heavily implicated in the epidemiology of local cases [36].
In summary, Campylobacter spp. were isolated from flies caught on broiler farms and this study provides the first estimate of the prevalence of Campylobacter spp. from flies on UK broiler farms. This study demonstrates that flies may play a role in the transmission of Campylobacter to broilers. Despite the low prevalences of Campylobacter spp. detected from flies on broiler farms (0·22% and 3·15%), risk of transmission is high as flies are present in large numbers around broiler-house ventilation inlets. MLST demonstrated that flies commonly carry broiler-associated STs, responsible for human enteric illness. Flies threaten the microbiological safety of the food chain, with an ability to breach existing biosecurity measures and colonize broiler flocks. Future studies should aim to quantify the risk of colonization that flies present to UK broiler farms and implement strategies to reduce this risk. In Denmark, fly screens have been shown to significantly (P = 0·0002) reduce the prevalence of Campylobacterpositive flocks at slaughter [23] and the use of a similar case-control study in the UK would be beneficial in assessing the risk that flies present and the effectiveness of available control measures.

ACKNOWLEDGEMENTS
We thank Elli Wright, Ruth Ryvar and Cathy Glover for their technical assistance and staff from the study farms for their cooperation. The research leading to these results has received funding from the European

DECLARATION OF INTEREST
None. Values given are n (%).