Sample Collection

For the diagnosis of BGC, various methods that demonstrate either the
presence of Cfv (i.e., bacterial culture) or some of its components (i.e., DNA,
proteins), or host-developed immune response against Cfv (immunologic
assays) have been developed (8) (Table 1). The quality of the samples
submitted for diagnostic testing is extremely important because it directly
influences the accuracy of the results obtained (31).

Because bulls are lifelong asymptomatic carriers, and disseminators of the

disease, they are the animals of choice for the diagnosis of BGC in endemic
herds (21). Cfv establishes persistent colonization of the preputial crypts,
which may be linked to the modification of the molecular composition of
bacterial surface antigens that allow it to escape the local immune response
(32). Given the ecologic niche of Cfv, diagnostic samples for detecting or
isolating the agent should preferably be taken from the prepuce of bulls.

For this purpose, smegma may be obtained from the preputial and penile
mucosa by three different methods: (A) scraping: performed by scarifying the
foreskin and penile mucosa using a disposable plastic or sterilizable reusable
metal scraper (33) that is then rinsed in phosphate buffered saline (PBS); (B)
aspiration: performed through a disposable plastic sheath coupled to the
artificial insemination pipette to suction the smegma; and (C) washing:
performed by introducing about 20–30 mL of PBS into the foreskin,
massaging it with the closed ostium before collecting the material through a
siphon system (8, 34). Preputial scraping, compared to the other two methods,
is the technique of choice because more C. fetus-positive samples are
identified when samples are obtained by this method (33). McMillen et al. (35)
confirmed that the scraping method is more effective in recovering Cfv than
the other two collection methods. In addition, the authors emphasize that this
technique is easier and safer to perform.
Bulls should be kept in sexual rest for 15 days before sampling, and three
sample collections, with the same resting interval, should be performed to
increase the diagnostic sensitivity (34). Downsides of this practice include that
A-the bulls need to be out of service for ~45 days, and B-multiple visits to the
farms are required to complete sampling, which implies greater logistic efforts
by both veterinarians and farmers.

In contrast, cows and heifers infected with Cfv are temporally colonized by

this pathogen, thus females are not usually sampled and tested for herd
screening purposes. However, when gestational losses occur, the
cervicovaginal mucus and placenta from aborted dams, and fetal organs and
fluids represent adequate diagnostic samples (3, 21). The same plastic sheaths
coupled to the artificial insemination pipette used in bulls, can be successfully
used in dams to suction cervicovaginal mucus. This represents an easy and
practical sampling method. Eventually, a sterile speculum can be used to
collect a cervicovaginal washing (8, 34, 36).

Current BGC diagnostic methods include: A-tests that either demonstrate the
presence of Cfv (i.e., bacterial culture) or its components (i.e., direct
immunofluorescence or immunohistochemistry for the detection of surface
proteins, PCR-based methods for detection of DNA), or B-tests that aim at
detecting host-developed immune response against Cfv (i.e.,
immunoenzymatic assays, vaginal mucus agglutination test) (3, 8, 22, 31, 37–
41).
Bacterial Culture Approaches
Genital secretions (preputial smegma and cervicovaginal mucus), placenta
and fetal fluids (i.e., abomasal content) and/or tissues (i.e., liver and lung)
represent adequate samples for isolation and further identification of Cfv. In
cases where BGC is suspected, samples should be collected aseptically and
transported to the diagnostic laboratory (8, 34). To increase the chances to
isolate this fastidious microorganism, cultures should be carried out within 4
h of sample collection (42). However, the use of proper transport media
extends this period to about 24 h (42). Transport and enrichment media
(TEM) are available to optimize culture, such as Weybridge, Cary-Blair, Clark,
Thomann, Lander, and 0.85% sterile saline solution (8, 43–45). Usually,
Weybridge TEM is the medium of choice, as it is efficient in maintaining
viable Cfv and reducing contamination (44). Thomann TEM has proved
effective for both culture and PCR approaches (45).

Physiological characteristics of Cfv makes laboratory culture difficult as this

microaerophilic and fastidious microorganism requires special growth
conditions. Procedures for isolating Cfv involve the use of enriched culture
media with antibiotics to minimize the growth of contaminants, and
incubation in microaerobic conditions (5–10% O2, 5–10% CO2, preferably 5–
9% H2, and the rest of N2), at 37°C for a minimum of 48 h (8, 34).

Enriched culture media such as Skirrow agar, bright green agar, and blood
agar are recommended for Cfv isolation. When managing microbiologically
complex clinical samples, such as preputial smegma, Skirrow agar is the best
choice (44). This medium contains inhibitors that minimize the growth of
undesirable contaminating microorganisms, thus facilitating the development
of Cfv and the consequent observation of compatible bacterial colonies (44).
Passive filtration of the sample onto the media can increase the recovery
of Cfv and reduce fungal contamination (42). Additionally, Skirrow could also
be used for the isolation of Brucella spp., pathogens that share niche and are
present in South America (46).

The morphologic characterization of the bacterial colonies is not sufficient for

the identification of Campylobacter to the species and subspecies levels. For
this reason, it should be considered that Cff, which is also clinically important
but not related to BGC, can grow in the same selective media and show similar
colony characteristics as Cfv. These subspecies should be distinguished by
their phenotypic characteristics based on biochemical tests (8, 21) or
molecular approaches (see section on PCR-based methods below).
While Cff produces H2S and is able to grow in the presence of 1% glycine and
0.1% sodium selenite, Cfv does not (8, 12). However, there is concern about
the reliability of these biochemical characteristics for definite subspecies
identification, as some Cfv strains had acquired the glycine tolerance
characteristic. Chang and Ogg (47) described that this process occurred by
transduction and mutation events. In addition, a group of Cfv denominated
biovar intermedius also, as Cff, produces H2S, which may lead to reaching
erroneous conclusions regarding the identification of Cfv (10, 12).
Direct Immunofluorescence (DIF)
Direct immunofluorescence is widely used for the diagnosis of BGC on
samples of preputial smegma, cervicovaginal mucus, uterine tissue, placenta,
and abomasal fluid, lung and liver from aborted fetuses
(3, 22, 31, 36, 41, 48, 49), and is listed by the OIE as a prescribed diagnostic
method for international bull trade (8). This test proves to be very effective
even in contaminated field samples. The detection limit is 10 4 and
102 Cfv CFU/mL in non-centrifuged and centrifuged preputial washings,
respectively (31).

For DIF, samples should be stored in PBS with 1% formalin after collection.
Genital fluid samples should be centrifuged to remove debris and
contaminating particles. The fluid is placed on glass slides and subsequently
fluorescein isothiocyanate conjugated (FITC) antiserum is added in the
appropriate dilution. The slides are examined under ultraviolet light in a
fluorescence microscope. Samples showing fluorescent bacteria with the
typical C. fetus morphology are considered positive (8, 31).

A study performed in Argentina evaluating preputial smegma and

cervicovaginal mucus through DIF determined a sensitivity and specificity of
69.4 and 94.4%, respectively (36). A similar study in Brazil estimated a
sensitivity of 92.59% and a specificity of 88.89% (31). The adequate
interpretation of the slides may be challenging, especially when there are
contaminating particles and cellular debris with low concentration of Cfv in
the fields of microscopic observation (31, 41). The performance of this
technique is directly influenced by the quality of the sample and the
microscope used, and also by the observer's experience (50), which can lead to
a reduction in test sensitivity. Therefore, adequate training is key in diagnostic
settings. Contrary to this statement some authors consider that the effect of
experienced observers on test performance is minimal (31).

Routine DIF protocols generally employ polyclonal rabbit raised FITC

antibodies that do not discriminate between the two C. fetus subspecies (8).
There is a risk for false positive Cfv results when the sample
contains Cff (8, 31), therefore caution should be taken when interpreting DIF
results alone. Chicken raised antibodies (IgY) against C. fetus can also be
employed in DIF protocols. Although IgY based protocols had similar
sensitivity when compared with rabbit IgG they show lower unspecific
background fluorescence and are cheaper to produce (48).

However, DIF is rapid and advantageous compared to more time-consuming

assays for C. fetus testing. Using samples of vaginal mucus, Marcellino et al.
(41) compared the diagnostic efficiency of the DIF technique with the
bacteriological culture, obtaining a moderate agreement (kappa coefficient =
0.52) between these techniques. However, usually vaginal mucus samples do
not harbor as many microbial contaminants as preputial samples. Considering
that Cff and Cfv are both subspecies of clinical importance and may both cause
reproductive losses, DIF is an easily applied technique suitable for the
screening of C. fetus in endemic BGC herds and abortion episodes. However,
for the definitive diagnosis of BGC, more specific and sensitive tests are
needed to provide evidence of Cfv infection.

Currently, DIF is widely used in most diagnostic laboratories in South America

as the only diagnostic test for BGC, possibly because cultivation is a difficult
technique to use for routine testing, and PCR protocols are not yet
standardized and validated under South American conditions. The lack of
experience, equipment and funding required to conduct more sensitive and/or
specific tests is still a limiting factor in most veterinary diagnostic laboratories
in the region.
ELISA and Other Immunologic Assays
An enzyme-linked immunosorbent assay (ELISA) was developed to detect
secretory IgA antibodies specific for Cfv antigen in the vaginal mucus and used
in the diagnosis of Cfv induced abortion (51). In infected cows, especially those
that aborted recently, there is a strong immune response of local antibodies in
the vaginal and uterine mucosa (51). This test has been used to screen for BGC
in cattle herds with infertility and abortions; the specificity was found to be
98.5%, although sensitivity could not be estimated (43). It has been postulated
that vaccination against campylobacteriosis does not interfere with the IgA
ELISA test, as only IgG, but not IgA, is secreted in the vaginal mucus of
vaccinated cows (43). As the immune response in the preputial mucosa of
bulls carrying the bacterium is fleeting (32), tests aiming at detecting
antibodies in preputial smegma should be avoided.

ELISA tests have also been developed for the detection of C. fetus antigens in
enriched bacterial cultures. In a study, field samples including preputial
washings, placenta from aborted cows, and abomasal fluid from aborted
fetuses, were incubated for 4 days in Clark's TEM and then tested with a
monoclonal antibody-based antigen capture ELISA for the detection of C.
fetus as a screening test. The sensitivity and specificity were of 100 and 99.5%,
respectively, compared to conventional culture (52).

An indirect ELISA for the detection of serum antibodies against C. fetus was

developed with rSapA-N, a recombinant protein codified by the C. fetus-
specific virulence gene sapA. The specificity and sensitivity of this method
were 94.3 and 88.6%, respectively (53). This test does not discriminate
between antibodies against the two C. fetus subspecies.

ELISAs can be used as a first screening of herds with C. fetus infection,

allowing for rapid large-scale sample processing (43, 52–54). However, there
are no reports of the application of these techniques for BGC diagnosis in
South American countries, and commercial kits are not readily available in
this region.

The vaginal mucus agglutination test (VMAT), was used in the past decades to
detect antibodies in vaginal mucus washes. The sensitivity was ~50% (15, 55).
This assay resulted in many false negative results not only because of the
intrinsic low sensitivity, but also because antibodies are detected in vaginal
mucus after 2 and before 7 months of infection (56). Therefore, the use of
VMAT was discouraged. This test has not been widely used with diagnostic
purposes in South America.
PCR-Based Methods
Culture-independent methods such as polymerase chain reaction (PCR) (10),
real time PCR (35, 39, 57) and multiplex PCR (37–39, 58) are also used for the
diagnosis of BGC. These approaches have improved reported sensitivity and
specificity compared to other techniques such as microbiological culture (57).
These techniques can detect specific Campylobacter DNA sequences by means
of the design of specific primers (37, 39, 59).

The differentiation of Cfv from other species and subspecies can be carried out
through the molecular genotype classification by means of multiplex PCR
techniques. In these approaches, specific primers designed to
detect Campylobacter spp. or the different subspecies are used (37, 38).
Another sensitive variation is real time PCR that is theoretically capable of
detecting and quantifying DNA from a single Cfv cell in field samples
(35, 57, 59). PCR-based methods have advantages over microbiological
cultures, such as simplicity in interpretation of results. A multiplex PCR
protocol targeting sequences present in Cfv but not in Cff, was developed in
Uruguay (39). Interestingly, genomic DNA is directly extracted from the
abomasal fluid of aborted bovine fetuses and the downstream process involves
only one step (39). Applying this protocol, Cfv biovar intermedius could be
detected when cultures and biochemical tests were not consistent (39, 60, 61).
However, a recent study demonstrates that the virB11 gene, used as a target
gene for differentiation between subspecies in the above-mentioned PCR
protocol (39), is not exclusively present in Cfv strains and is not always absent
in Cff strains (62).

The complete genome of an Argentinian Cfv biovar intermedius strain was

sequenced (61). The genomic information of such atypical strains would
optimize the development of tools for more specific molecular diagnosis.
Recently, the study of two genomes, reported incongruities between
the Cff and Cfv lineages and the biochemical characteristics used for their
differentiation, questioning the clinical relevance of subtyping mammalian
strains (63, 64).

Van der Graaf-van Bloois et al. (60) tested five PCRs that are routinely used by
diagnostic laboratories, and surprisingly none of them were able to correctly
identify strains of C. fetus at the subspecies level. All the tests were compared
with the methods of amplified fragment length polymorphism (AFLP) (65)
and multilocus sequence typing (MLST) (10), which so far are the only
techniques that have proven to reliably differentiate the two subspecies.
Unfortunately, these assays are cumbersome, costly and impractical for
routine use.

The differentiation of C. fetus subspecies is essential for the implementation of

efficient Cfv control and eradication programs, and for investigating the public
health burden of C. fetus subspecies, however its genomic evolution in
mammals remains poorly understood. A recent study provides the
phylogenetic and evolutionary structure of C. fetus which could guide the
development of methods for differentiation and epidemiological surveillance
of bovine and human strains (66).

In South America, even though molecular tests have been developed and used
in recent years, their applicability to routine laboratory diagnosis is currently
subject to discussion. This is due to the discrepancy between results obtained
by different protocols, and because most protocols cannot be performed with
confidence directly on DNA extracted from field samples (i.e., preputial
smegma). These factors have led so far to the non-acceptability of the
technique by field veterinarians and diagnostic laboratories. Even though it is
likely that these technologies will be increasingly adapted by veterinary
laboratories in years to come, particularly considering that they are accepted
by the OIE standards (8). Phenotypic tests are still the only ones that are
reliable and available in South America for the identification of subspecies.
However, they are poorly reproducible and do not correspond to the genomic
characteristics of some strains (63, 64). Due to these aspects, further research
is needed to validate molecular techniques with locally isolated strains and to
assess their applicability in the region.
Histochemical and Immunohistochemical Methods
Histology and immunohistochemistry are useful and effective diagnostic
methods in natural cases of bovine abortions by C. fetus (22), particularly in
fetuses and placentas from which C. fetus cannot be successfully isolated
because of poor preservation of the samples (i.e., aborted fetuses that are
autolytic, frozen or deteriorated upon arrival to the laboratory) (23).

The main histologic lesions in aborted fetuses are suppurative pneumonia,

myocarditis, fibrinous serositis, interstitial nephritis, hepatitis, gastroenteritis,
meningitis, and necrotizing placentitis; however, these lesions are not
pathognomonic of C. fetus infection (22, 23, 40). Immunohistochemistry for
the detection of C. fetus antigens in formalin-fixed paraffin-embedded tissues,
in the absence of bacterial isolation and molecular evidence of infection, is an
alternative and practical method to establish the diagnosis in aborted fetuses
(22, 23, 36).

Pathological examination coupled with DIF and/or bacterial culture has been
widely used in Brazil, Argentina, and Uruguay for the diagnosis
of Campylobacter-induced abortion (3–6, 22, 23, 67). However, this does not
seem to be a broadly used diagnostic approach in other countries in the
region.