Bovine campylobacteriosis in bulls: insights in the conventional and molecular diagnosis

Campylobacter fetus is a gram-negative motile bacterium, with two subspecies relevant to cattle health: C. fetus subsp. venerealis (Cfv) and C. fetus subsp. fetus (Cff). Both subspecies are associated with reproductive losses in cattle. In this study, we evaluated the identification of C. fetus for the diagnosis of bovine campylobacteriosis through bacteriological culture, direct immunofluorescence (DIF) and molecular tests in preputial smegma (PS) samples of three Angus bulls challenged with Cfv, Cfv biovar intermedius (Cfvi) or Cff, respectively, in an experiment imitating the natural infection. Two DNA extraction protocols were tested (in-house thermal extraction and commercial kit). Aspiration and scraping collection for PS were compared by conventional tests. Additionally, bacteremia was also evaluated in blood samples. Bulls were challenged by natural mating with heifers that had been experimentally infected with C. fetus subspecies; which led to infection. The Cfv- and Cfvi-bulls were positive for at least 9 months. Although Cff is not considered a venereal strain, in this study it was transmissible to bull from heifers experimentally infected, as evidenced by its colonization and persistence in the preputial cavity for 5 to 6 months. This finding suggests a potential risk of dissemination within herds. The results obtained by bacteriological culture or direct immunofluorescence (DIF) showed no significant differences, regardless the sampling device used (aspiration with Cassou pipette, metal and plastic scraper). C. fetus qPCR, on the other hand, yielded better results with an in-house DNA extraction method than with a commercial kit (75% vs 66.6%). Furthermore, qPCR diagnosis was more efficient than culture (66.6%) or DIF (56%). Bacteremia in whole blood samples was negative by qPCR and bacteriological culture in all samples. Altogether, this study demonstrated the transmission of Cff from heifers to bull and also showed that PCR-based methods are promising for the diagnosis of Bovine Genital Campylobacteriosis from clinical samples of PS.


Introduction
Campylobacter fetus is a gram-negative motile bacterium, with two subspecies relevant to cattle health: C. fetus subsp. fetus (Cff) and C. fetus subsp. venerealis (Cfv) [1]. C. fetus subsp. fetus inhabits the intestine of cattle and sheep, but may migrate to the genital tract and cause sporadic abortions [2]. Although to date there are no reports of venereal transmission of Cff, the bacterium can be present in bull's genitalia (as evidenced by positive preputial samples) and in the vagina and cervix of heifers [3]. Cfv, including the biotype intermedius (Cfvi), on the other hand, exclusively inhabits the genital tract of cattle and is the causal agent of Bovine Genital Campylobacteriosis (BGC), a global venereal disease causing pregnancy loss (up to 20%) [1,4]. This disease has high prevalence in countries with extensive  [4-7, 8, 9]. Differential virulence of both subspecies (Cff and Cfv) is still controversial [1,3,10]. In the male, the bacteria inhabit the preputial crypts and produce a chronic asymptomatic disease without affecting libido or fertility [1,11]. Although BGC transmission is primarily through coitus, the bacteria can survive cryopreservation in semen and, therefore, are also transmissible via artificial insemination [12]. The bull is the most important reservoir and disseminator of BGC and, in consequence, is the main target for diagnosis. In this regard, preputial smegma (PS) is the sample of choice to identify the organism and, in turn, to prevent and/or control the disease in a herd [13,14].
Researchers have assessed several methods for the collection of PS samples from bulls, with the purpose of identifying venereal pathogens, with different success rates [3,13,15,16]. The routine diagnosis of C. fetus from PS samples includes both direct immunofluorescence (DIF) and/or PCR, depending on the country. This technique, however, cannot differentiate between subspecies and is more laborious, since it requires two or more preputial samplings to reach an adequate test sensitivity [14,17]. The bacteriological culture (the gold standard), on the other hand, can differentiate subspecies, but requires special growth conditions that are not performed routinely [11,18]. Although recent PCR protocols for C. fetus have become more sensitive and specific [19][20][21][22][23], adequate subspecies differentiation is still a problem [23][24][25].
Herein, we evaluated and compared the identification of C. fetus within the diagnosis of bovine campylobacteriosis with DIF test, bacteriological culture and qPCR, from PS of three bulls naturally challenged with heifers experimentally infected with Cff, Cfv or Cfv biovar intermedius, respectively. Two different DNA extraction methods were used: In-house thermal extraction and a commercial kit. Concomitantly, three PS sampling methods were compared for DIF and bacteriological culture from all three bulls. In addition, C. fetus bacteremia was evaluated by bacteriological culture and qPCR in blood samples.

Animals: naturally infected bull model
Three Aberdeen Angus bulls, of 4 to 5 years of age, were used in the study. Each bull was kept in a pen for 39 days (during the mating period) with one of the three groups of eight heifers experimentally infected with either Cfv, Cfvi or Cff, respectively, as previously reported [26] (Fig. 1). The animals of each group of heifers (Cff, Cfv or Cfvi) were maintained in separated pens. The strains used for the challenge of the heifers were Cff-C1N3, Cfv-97/608 and Cfvi-99/541 (Bacteriology Lab of INTA EEA Balcarce) [2,28]. A negative control group (NCG) of two bulls with six non-challenged heifers was allocated in a separate paddock. Schematic representation of the overall experimental procedure. Three groups of heifers (n=8 per group) were experimentally challenged with the respective C. fetus subsp. venerealis, C. fetus subsp. venerealis biovar intermedius or C. fetus subsp. fetus strains, and then each group was kept in the same pen with one bull each for 39 days. Postmating preputial smegma sampling is shown. Dark heifers: C. fetus non-infected animals; orange heifers: animals experimentally challenged with the respective C. fetus strain [27].

Sample collection
Nine PS samples were collected from each bull, every 30 days approximately (range of 13 to 43 days between samplings), by three different devices using two sampling methods (aspiration and scraping): a) aspiration by artificial insemination pipette (Cassou pipette) and disposable plastic sheats with 2 ml of PBS; b) scraping with metal scraper; and c) scraping with flexible plastic disposable scraper. Before collecting the samplings, the veterinarians induced urination by preputial massage, with subsequent trimming of the hairs and cleaning of the preputial orifice with a cloth moistened with 70° ethanol. The area of the fornix was reached with all the devices.
The scraping samplings consisted of 30 forward and backward movements while scraping the penis and internal prepuce, whereas the aspiration method was performed with the pipette device and no scraping ( Figure 2). Each sample was individually collected in Cary-Blair transport medium for bacteriological culture, in a tube with 1% phormolated physiological solution (FPS) for direct immunofluorescence (DIF) test and in a 1.5 ml microtube tube for qPCR analysis, and subsequently delivered to the lab within 2 h. Particularly, qPCR was performed with samples taken only with the aspiration method (which had been previously stored at -20°C until processed), because of its lower contamination.
Whole blood samples were collected from all three bulls on day 39, 132, 228 and 283, and processed by bacteriological culture and qPCR to evaluate possible bacteremia.

Bacteriological culture
The collected PS and whole blood samples were inoculated on modified 7% blood-Skirrow selective media agar (AS) plates (Oxoid, Hampshire, UK) enriched with 1.25 IU/ml polymyxin B sulfate, 5 mg/ml trimethoprim, 10 mg/ml vancomycin and 50 mg/ml cycloheximide (Sigma-Aldrich, St. Louis, MO, USA). The plates were incubated at 37°C under microaerophilic atmosphere (5% O 2 , 10% CO 2 and 85% N 2 ) and observed every 48 h for 10 days [13]. The bacterial genus was identified on the basis of colony characteristics (smooth, round, 1 to 3 mm in diameter, convex, white-gray, and translucent), and bacterial morphological characteristics, with darting motility in corkscrew, observed under a phase contrast microscope [13]. These colonies were subcultured in Kligler's iron agar tubes to confirm C. fetus species identity based on catalase and oxidase enzyme activity (positive) and sulfhydryl acid production (negative).

Direct immunofluorescence for C. fetus
The DIF test was performed with a commercial fluorescent conjugate anti-Campylobacter fetus according to the manufacturer's instructions (CONJUGADO-CAMPY®, Laboratorio Biológico Tandil SRL, Argentina). Briefly, 10 μl of each sample was applied in microscopy slides of 12 wells. The slides were dried in a 37°C stove and fixed in absolute ethanol, at room temperature, for 15 min, and then dried in the stove again. Finally, the slides were washed with distilled water (5 s) and dried again at 37°C. The fluorescent conjugate anti-Cfv was added at a dilution of 1/200 and incubated in a wet dark chamber at 37°C for 1 h. The slides were washed three times in PBS (pH 7.2) for 10 min each wash, and finally washed once with tap water. Subsequently, the slides were mounted with buffered glycerol (pH 9.2), and observed under an epifluorescent microscope (Olympus CX31, Zhejiang, China) by experienced personnel. Preparations with at least one bacterium presenting typical morphology and fluorescence characteristic of C. fetus were considered positive [13].

Real Time-Polymerase Chain Reaction (qPCR) analysis
Two DNA extraction protocols for C. fetus were carried out: a slightly modified In-house method loaned by BIO-TANDIL® (Laboratorio Biológico Tandil®) (LBT) and a commercial kit (Inbio Highway ®). Briefly, in the In-house-LBT, 500 μl of physiological solution was added to each PS samples and then the samples were homogenized and centrifuged at 12,000 xg for 5 min. The supernatant was discarded and the pellet resuspended with 100 μl Tris-EDTA (Inbio Highway®). Then, the samples were placed in a thermoblock at 100°C for 15 min under agitation, and then were centrifuged at 12,000 xg for 5 min. The supernatant was recovered and placed in a new 1.5 ml tube. The extraction with the commercial kit was carried out following the manufacturer's instructions (Inbio HighWay® DNA Puriprep-S kit, protocol for preputial smegma samples).
The whole blood samples were analyzed with the same commercial kit for the DNA extraction following the manufacturer's instructions for blood samples (Inbio HighWay® DNA Puriprep-S kit).
The qPCR was performed as previously described [29] with a 10 μL reaction volume. The set of primers to amplify a sequence of 78 bp from the 16SRNA gene were 16SFw (5′GCA CCT GTC TCA ACT TTC 3′) and 16SRv (5′CCT TAC CTG GGC TTGAT3′), with a fluorescent TaqMan-MGB probe (16SPb: 5′-VIC-ATC TCT AAG AGA TTA GTT G-MGB/NFQ-3′). This probe targets a polymorphic region of 19 bp that discriminates the C. fetus strains from other Campylobacter species or other bacteria. A SsoAdvanced Universal Probes Supermix (Biorad®) (5 μL) was used: 1 μL (900 nM) of each of the forward and reverse primers, 1 μL of DNA template, 1.25 μL (250 mM) of the probe and 0.75 μL RNase-free ddH2O RNase-free ddH 2 O. Negative (DNA from C. jejuni, C. sputorum subsp. bubulus, C. coli and C. hyointestinalis strains) and positive controls (DNA from Cff, Cfv y Cfvi strains) as well as a standard curve duplicate were added in the assay. The standard curve was constructed with a serial 10-fold dilution from an aliquot of purified genomic DNA from C. fetus subsp. fetus strain 04-554 (range 50 pg/ μL to 5 fg/μL).
The reaction conditions were as follows: polymerase activation and predenaturation at 95°C for 3 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s, using StepOnePlus TM Real-Time (Thermo Fisher, Walhtma, MA, USA).

Statistical analyses
The full data set of samples was evaluated by Cohen's Kappa coefficient to measure agreement between the results of the different C. fetus detection tests (qPCR, DIF and bacteriological culture) and between sampling devices (aspiration and metal and plastic scraper). The comparisons between diagnostic tests and between sampling devices for DIF test and bacteriological culture were performed by a generalized mix linear model with binomial distribution, logit link function, by using Rcmdr as statistic software.
The analyses were performed considering bulls as independent (categorical or continuous) variables and the diagnosis tests or sampling devices as dependent variables. The significant difference was set at P<0.05.

Results
According to the bacteriological culture and DIF tests (Figure 3), the three bulls naturally exposed resulted infected with the respective strain corresponding to the group of infected heifers (Cfv, Cfvi y Cff) [27]. The infection was persistent until day 283 for bulls infected with Cfv and Cfvi, while the bull infected with Cff was sustained for more than 5 months (day 165).
The post mating PS samples consisted of 81 samples (three bulls per nine sampling periods and per three devices, for the PS sampling) for the bacteriological and DIF tests. Table 1 details the number of positive samples to C. fetus by culture with the different devices for PS from each bull (27 PS samples per bull). Most samples were positive regardless the sampling device (67% for aspiration and plastic scraper; 63% for metal scraper). No significant differences were observed between the devices, according to the bacteriological culture (P=0.940). However, more samples  Table 2 shows the results of C. fetus detection by DIF according to the device used (27 PS samples per bull). The number of positive samples was similar to those of the culture analysis, regardless of the sampling device (62.9%, with aspiration; 59.2%, with metal scraper; and 66.6% with plastic scraper). No significant differences were observed between devices for the DIF test (P=0.830). Regarding the subspecies analysis, almost all the Cfvi samples (88.8%) were detected as positives, while around half of the Cfv (55.5%) and Cff (44.4%) samples were positive. Indeed, like in the culture analysis, DIF also showed significant differences between the detection of positive samples from the bull infected with Cfvi and those from bulls infected with the other two subspecies (Cfv and Cff) (P=0.006). Including all the devices evaluated, 62.3% (51/81) of the samples were positive. The analysis of the venereal subspecies (Cfv and Cfvi) yielded a higher positivity (72.2%; 39/54).
No significant differences were detected between the performance of the bacteriological culture and DIF (P=0.772). However, the concordance between these two techniques (accounting all three sampling devices) for the 81 PS samples was moderate (kappa index: 0,463; IC:0,259-0,667).
The assessment of the qPCR detection was performed only with samples collected with the aspiration method, in eight sampling times per bull (24 PS samples). The two different DNA extraction methods used showed no significant differences ( Table 3). The determined standard curve showing the logarithm of the copy numbers yielded a strong linear correlation (slope: -3.791, Eff%: 83.5%, and R 2 =0.995) (Figure 4). For qPCR, 75% (18/24) and 66.6% (16/24) of the samples resulted positive by IH-LBT and the commercial Kit DNA extraction method, respectively. Regarding only the venereal subspecies (Cfv and Cfvi), 87.5% (14/16) and 81% (13/16) of the samples were positive by IH-LBT and the commercial Kit DNA extraction method, respectively.
Concordance between DNA extraction methods used to analyze the PS samples through qPCR was quite adequate (kappa index: 0.77; IC:0,562-0,982). On the other hand, concordance between bacteriological culture and qPCR (IH-LBT DNA extraction method) for the analysis of PS samples acquired through aspiration sampling was moderate (kappa index: 0,599; IC:0,282-0,916). However, the concordance between DIF and qPCR (IH-LBT DNA extraction method) with PS samples collected through aspiration sampling was adequate (kappa index: 0,769; IC:0,523-1,105). No significant differences were obtained between culture, DIF test and qPCR (P=0.726). All the diagnostic tests (DIF, bacteriological culture and qPCR) identified the samples from bulls of the NGG as negatives throughout the assay.
All the whole blood samples (n=12, four period samplings per three bulls) were negative by bacteriological culture and qPCR.

Discussion and Conclusion
The bull is the main target for diagnosis and management of BGC. However, there are still difficulties to establish effective preventive and control strategies [14].
As previously reported [26], an experimental model imitating natural infection can be exploited to study different diagnostic approaches. Here, we studied all three C. fetus subspecies relevant to cattle health.
According to the diagnostic tests used, the bulls infected with the venereal subspecies (Cfv y Cfvi) were positive for at least 9 months. This result reinforces its known carrier status between breeding seasons [4,14]. Regarding the bull exposed to Cff contact, the infection was persistent for at least 5 months and allowed passive or active mechanical transmission by traditional (3 months) or continuous breeding systems. Importantly, this is also a risk for cattle reproductive health.
Although researchers have previously isolated Cff from preputial samples of experimentally and naturally infected bulls [3,30], it is still unclear if this bacterium is a venereal pathogen in cattle. However, Cff is associated with infertility and induces abortion [5,6,31]. These pieces of evidence suggest that bulls could have a significant role in the dissemination and maintenance of Cff within herds. The results from our study demonstrated the transmission of Cff from heifers to bulls and the bacterial infection was characterized by colonization and persistence in the preputial cavity. Although this study includes the assessment of only one strain of Cff, our findings show that this subspecies should be considered a potential risk of transmission during mating seasons. In this line, differential diagnosis of C. fetus subspecies should not be mandatory.
The lower persistence of Cff could be due to its low adaptation in the prepuce, probably because of the lack of surface proteins, genomic islands and mobile elements that predominate in Cfv; which would enhance long term survival in the urogenital tract [32][33][34]. Particularly, the variant Cfvi had a higher rate of recuperation from PS than other variants. This could be associated with different genetic profiles in relation to genes affecting colonization and attachment capacity [33].
Bacteriological culture showed higher efficiency detecting the venereal strains (Cfv and Cfvi), with higher bacteriological recuperation of Cfvi. The detection through culture was more efficient in our study in relation to other reports (63-67% vs. 33-38%) [3,21].
Regarding the DIF test, there were significant differences between Cfvi in relation to Cff and Cfv. This may be explained by a higher concentration of the bacterium in the prepuce and may be associated with a more virulent strain. In effect, the higher detection rates of Cfvi through all the diagnostic tests reinforce the idea of a higher bacterium concentration.
Previous studies have reported detection variations, from 69.4% to 92.6%, with DIF test for Cfv [4,20] and these percentage values are coincident with that of the present research (72%). In our study, the DIF detection rate decreased when Cff was included in the analysis (63%). The lack of previous reports regarding the detection of this subspecies through this technique makes comparison impracticable. Despite the moderate detection rate of DIF (56% to 67%), this technique is still widely used because of its low cost, simplicity and fast processing [4,5,14].
No significant differences were detected between culture and DIF. This is in contrast with studies showing significantly better results by bacteriological culture, which were associated with a low bacterial concentration or inadequate sampling [3]. The moderate concordance between bacteriological culture and DIF opens the debate around the use of both tests in herds with endemic BGC in order to reduce false negative results in bulls.
In accordance with this study, some researchers have suggested that the detection efficiency of PS collection by DIF or bacteriological culture seems to be independent of the devices used [35,36], while others have described better results with scraping devices [ 22,41]. According to McMillen [37], the scraping method would increase the possibility of isolating C. fetus at low concentrations within the preputial crypts, since friction against the mucosa generates detachment of epithelial cells along with the bacteria. As mentioned before, in the present study the results were independent of the device used. Thus, according to our results and those of some other researchers, the use of the device should be based practicality or feasibility for veterinarian depending on the length of the preputial cavity. More studies should be conducted to resolve these discrepancies and to find the most efficient device for this purpose.
Our study showed better results for the detection of C. fetus by qPCR (66.6%-75%) than through conventional diagnostic tests (culture and DIF; 66.6% and 56%, respectively). This is in accordance with previous studies [20-22, 24, 37]. Variations in the molecular technique associated with detection failure may be due to different local strain variants, cross-reactions, and contaminants inhibiting the reaction (such as blood, urine, or pus) that may limit the effectiveness of PCR, mainly in clinical samples [24,[37][38][39][40]. The inclusion of a TaqMan probe-based qPCR in the study was because its high sensitivity [29]. Culture fastidious growth and lower capacity to process large volumes of samples, as well as the lower sensitivity of DIF, makes qPCR an adequate and promising diagnostic tool for C. fetus detection. The In-house DNA extraction method resulted in higher C. fetus detection. Indeed, its lower cost and fast processing make this method a practical option.
The detection of subspecies of these bacteria continues to cause problems. This is due to the use of sequences or genomic island present both in Cff and Cfv, or even in other species, for PCR assays designed to discriminate between C. fetus subspecies. The use of these common sequences, needless to say, leads to false positives [24,25,39,41,42].
The differences on detection between samplings for either diagnostic test could be associated with variations of local immunity and bacterial concentration in the bulls, as well as strain differences (virulence factors) [7,11]. Because of these various factors and other variations, such as sampling and diagnostic test effectiveness, the sampling needs at least two scraping samples [15,[20][21][22]. Similarly, in our study concordance between tests varied from moderate to adequate, thus leading to false negative bulls. It should be notice that this diagnostic test may display lower sensitivity with animals infected in natural conditions, since naturally infected animals are exposed to lower quantities of bacteria than animals from experimental models.
Campylobacter spp., mainly Cff, is associated with bacteremia and extraintestinal infections in humans [43] and ruminants [2]. Here we were unable to detect any C. fetus in blood. This suggests different tropism associated with dissimilar C. fetus pathogenesis in bulls and humans, as suggested for Cff, which causes abortion in cows, ewes, alpacas, among others [2,4,27,44].
The detection of the infection of bulls upon C. fetus challenge was similar with any of the diagnostic test evaluated. The easy implementation and low cost of DIF makes it an adequate test, particularly in Argentina, where the conjugate used in the test is still largely produced. In the case of culture, the availability of reference autochthonous strains in future research is highly important to confirm the etiology, as well as to develop and validate new diagnostic techniques. Finally, PCR-based methods are promising for the diagnosis of BGC from clinical samples of PS, thus ensuring higher detection rate and fast results. As previously reported, PCR is currently part of the routine diagnosis of campylobacteriosis in many countries. Where BGC is endemic, as in Argentina, more investment in infrastructure and laboratory training is necessary to establish control and eradication programs.
Future research should assess the efficacy of qPCR in field cases of bulls naturally infected during mating seasons with more diverse clinical samples collected by veterinarians.