Survey of Infectious Etiologies of Bovine Abortion during Mid- to Late Gestation in Dairy Herds

Bovine abortion of unknown infectious etiology still remains a major economic problem. Thus, we investigated whether Brucella spp., Listeria monocytogenes, Salmonella spp., Campylobacter spp. and Coxiella burnetii are associated with abortion and/or stillbirth in Tunisian dairy cattle. Using a pan-Chlamydiales PCR, we also investigated the role of Chlamydiaceae, Waddlia chondrophila, Parachlamydia acanthamoebae and other members of the Chlamydiales order in this setting. Veterinary samples taken from mid to late-term abortions from twenty dairy herds were tested. From a total of 150 abortion cases collected, infectious agents were detected by PCR in 73 (48.66%) cases, 13 (8.66%) of which represented co-infections with two infectious agents. Detected pathogens include Brucella spp (31.3%), Chlamydiaceae (4.66%), Waddlia chondrophila (8%), Parachlamydia acanthamoebae (5.33%), Listeria monocytogenes (4.66%) and Salmonella spp. (3.33%). In contrast, Campylobacter spp. and Coxiella burnetii DNA were not detected among the investigated veterinary samples. This demonstrates that different bacterial agents may cause bovine abortion in Tunisia. This is the first report suggesting the role of Parachlamydia acanthamoebae in bovine abortion in Africa. Further studies with a larger number of samples are necessary to confirm whether this emerging pathogen is directly linked to abortion in cattle.


Introduction
Abortion among dairy cows is one of the major causes of economic losses in the cattle industry. Abortions may be idiopathic or the result of metabolic or hormonal abnormalities, nutritional deficiencies, trauma, toxicities, or infectious agents. The latter represents the leading etiology of reproductive disorders [1,2] A variety of infectious agents have been reported to cause bovine abortion throughout the world. The major bacterial agents that have been implicated in bovine abortion during mid-to lategestation are Brucella spp., Chlamydia spp., Salmonella spp., Campylobacter spp., Listeria monocytogenes and Coxiella burnetii [3][4][5]. Bovine brucellosis was distributed worldwide, and its importance is related not only to the economic losses in animal production but also to a significant risk for human health. In cattle, brucellosis is caused by Brucella abortus, a facultative intracellular Gram negative coccobacillus. The abortion generally occurs from 6 to 9 months of gestation. It is frequently followed by fetal membrane retention and endometritis, which may be the cause of infertility in subsequent pregnancies [6,7]. Even in the absence of abortion, excretion of the infectious organism occurs, for instance via placenta, foetal fluids, vaginal discharge, and milk. In addition to brucellosis, chlamydiosis and Q fever should be considered among the most common zoonotic diseases and are distributed all around the world. They are respectively caused by Chlamydia and C. burnetii, two strictly intracellular Gram negative bacteria. In cattle, C. burnetii infection is associated with abortions and stillbirths even if the infection is often asymptomatic. Inversely, chlamydial infections in cattle cause a variety of syndromes such as conjunctivitis, polyarthritis, encephalomyelitis, mastitis, infertility, abortion and other urogenital tract infections [8,9]. C. abortus, C. pecorum and W. chondrophila are additional etiologies of chlamydial abortions in cattle [10][11][12]. Moreover, there is increasing evidences supporting the role of another Chlamydia-related bacteria, P. acanthamoebae, in abortion in both cattle [13][14][15] and humans [16]. However, no information is available regarding the presence of this bacterium in Africa.
Listeriosis, salmonellosis and campylobacteriosis are also serious zoonotic diseases. Bovine listeriosis is caused by L. monocytogenes and the organism is excreted in feces, urine and milk. Abortions are usually sporadic but may affect 10-20% of a herd. Abortions occur most commonly during the last trimester of pregnancy. The aborted fetus is often autolyzed. Bovine salmonellosis is caused predominantly by S. enterica serotypes Typhimurium and Dublin [17,18]. Occasionally, Salmonella spp. cause abortion storms. The cows are usually sick and the fetuses and placentas are autolyzed and emphysematous. Salmonella can be isolated from the fetal tissues, vaginal fluids, feces and milk. Several Campylobacter species can be associated with abortion in cattle. Bovine genital campylobacteriosis is a venereal disease caused by Campylobacter fetus subsp. venerealis that can be found in the genital tract of cattle, in which it may cause genital tract infection and sporadic abortions [19].
Bacteriological isolation by culture on blood agar is usually used for the diagnosis of bovine brucellosis, but it is difficult, time consuming, hazardous, and sometimes inconclusive [20,21]. Routine diagnosis of Q fever is often made by the use of serological tests [22], which have the disadvantage of indicating post-exposure rather than ongoing infection. Diagnosis of chlamydial infections in animals still represents a considerable challenge [23]. Isolation in cell culture remains difficult, time consuming, and depends on the presence of sufficient numbers of viable bacteria. The reliability of standard diagnostic procedures for Campylobacter fetus, which are based on phenotypic methods, is difficult because of specific nutritional and atmospheric requirements [24]. Furthermore, all these traditional culture and serology-based methods can be relatively insensitive depending on the quality and timing of sampling. Rapid detection of abortigenic agents at the early stage of an outbreak using molecular methods contributes to minimizing the spread of infection, the outbreak burden and increasing treatment efficiency. Therefore, real-time PCR have been increasingly used as a diagnostic tool for etiologic diagnosis of abortion in cattle [3,15,25], as a complement or replacement of time consuming traditional diagnostic methods such as bacterial culture [26].
Thus, the present study aimed (i) to investigate the role of some abortigenic agents in cattle from different geographical regions of Sfax in Tunisia (ii) to define the role of P. acanthomoebae as a new abortigenic agent in African cattle and (iii) to detect a possible coinfection of members of the Chlamydiales order with other abortigenic agents.

Animals and samples
Twenty dairy herds from different regions of Sfax (Tunisia) that experienced reproductive disorders (mainly abortions) from October 2010 to May 2012 were included in this study. Informations on individual animal such as age, sex and abortion history were recorded separately on sample data sheets. Herd sizes ranged from 20 to 1500 cows, knowing that all cows studied in this work are the descendents of pure Holsteins from the national production. All these dairy cows were vaccinated against foot and mouth disease (FMD) during national immunization campaigns. The majority of included cows were kept on limited pastures or tethered on a pasture. Seventy percent of the herds possessed small ruminants, and only twenty percent reported having bought new cattle. In dairy herds, nutrient requirements may not be the same depending on the animal's age and stage of production. Forages, which refer especially to hay or straw, are the most common type of feed used. Barley is an example of cereal grain that is extensively used in these herds. In all herds, samples from (i) cows with clinical signs (cases) and (ii) cows with normal pregnancies and normal parturitions (controls) were taken. A total of 214 animals were sampled: 150 cases and 64 controls. They were bled on the same day or up to 8 months after abortion. A total of 214 blood, 214 vaginal swabs and 214 milk samples were collected by local veterinarians and sent to the laboratory. Practically, blood samples of about 5 ml were aseptically collected using plain tubes from cows through jugular venipuncture. Serum samples were separated within 12 h of collection and transported to the laboratory using an ice box where they are stored at 220uC until tested.
Vaginal swab samples were collected from each cow, using sterile swabs after vulva disinfection with chlorhexidine solution. All swabs were put into tubes containing 1 ml of 2-sucrose phosphate medium (2-SP). They were kept at 4uC during transportation in ice-pack containers, and then stored at 280uC until use. All milk samples were stored at 280uC until tested.

DNA extraction and real-time PCR
The vaginal swab samples collected in 2-SP medium were thawed; 1 ml was transferred to a new microtube, and then centrifuged at 13,000 g for 20 min. The pellet was resuspended in 200 ml sterile water. The total volume was extracted by Quick-gDNA MiniPrep D3006 Kit (Zymo Research, Irvine, CA, USA) as recommended by the manufacturer. Extracted DNA was resuspended in 50 ml of elution buffer and stored at 220uC until subsequent analysis.
For milk samples, DNA was extracted by using 1 ml aliquot which was centrifuged at 6,0006 g for 10 min. The clear whey portion was suctioned out with a transfer pipette and discarded. The remaining milk solids and butterfat were resuspended in 200 ml sterile water and used for DNA extraction using Quick-gDNA MiniPrep D3006 Kit (Zymo Research, Irvine, CA, USA) as recommended by the manufacturer. Extracted DNA was resuspended in 50 ml of elution buffer and stored at 220uC until subsequent analysis. All samples were tested in duplicate.

Brucella spp. PCR
The real-time PCR assay was performed on a CFX96 TM realtime PCR cycler (Biorad, Hercules, CA, USA). Real-time amplifications were carried out in a total reaction volume of 20 ml containing 10 ml iTaq Supermix with ROX (Bio-Rad, Reinach, Switzerland), 0.3 mM of each primer, 0.2 mM of probe and 5 ml of purified DNA to a final volume of 20 ml using nuclease-free water. Primers and probe sequences are provided in Table 1. The optimal PCR efficacy was obtained using a cycling profile that included an initial denaturation step at 95uC for 3 min, then 45 cycles of 15 s at 95uC and 1 min at 60uC. In all experiments, each PCR run included a negative extraction control (sterile water) and a negative PCR control containing 5 ml Diethylpyrocarbonate (DEPC) treated H 2 O instead of DNA extract, to detect possible contamination of DNA. Samples were run in duplicate.

Chlamydiales-specific real-time PCR
A pan-Chlamydiales PCR that may amplify all members of the Chlamydiales order including Chlamydiaceae, Waddliaceae and Parachlamydiaceae was performed as described previously by Lienard et al. [27]. The CFX96 TM real-time PCR (Biorad) was used to perform the amplification and the detection of DNA with a cycling program of 3 min at 95uC, followed by 50 cycles of 15 s at 95uC, 15 s at 67uC, and 15 s at 72uC. Samples were tested at least in duplicates and were considered negative if no amplification was observed during all 50 cycles. Samples positive for members of the Chlamydiaceae family were further tested by corresponding commercial kits.
Commercial standard kits for the quantification of C. abortus and C. pecorum The PrimerDesign genesig Kits for C. abortus and C. pecorum Genomes are designed for the in vitro quantification of these two bacteria genomes. The two kits are targeting the ompA gene, allowing specific in vitro quantification of these species and detecting no other Chlamydia. A highly conserved sequence within the ompA gene has previously been shown to be a good target sequence in other clinical real time PCR based studies [28]. The primers and probe sequences in this kit have 100% homology with a broad range of clinically-relevant reference sequences.

Listeria monocytogenes PCR
Real-time PCR was performed on the CFX96 TM real-time PCR cycler (Biorad). Each reaction contained 10 ml SsoAdvan-cedTM SYBR1 Green Supermix (Biorad), 0.6 mM of each primer (Table 1), 2 ml template DNA and nuclease free H 2 O to a final volume of 20 ml. The cycling parameters consisted of 3 min incubation at 95uC followed by 45 cycles of 95uC for 15 s and 60uC for 1 min. A melting curve analysis was performed using the following cycling parameters: 60uC for 30 s, and 5uC temperature changes to the end temperature of 95uC. In all experiments, each PCR run included a negative extraction control (sterile water) and a negative PCR control; containing 2 ml DEPC treated H 2 O instead of DNA extract, to detect possible contamination of DNA. Samples were run in duplicate.

Salmonella spp. PCR
The real-time PCR was performed on the CFX96 TM real-time PCR cycler (Biorad). Amplification reactions were carried at a final volume of 25 ml containing 0.2 mM of each primer (Table 1), 12.5 ml of 26 SYBR Permix Ex Taq TM Tli RNaseH Plus (TaKaRa) and 1 ml of genomic DNA. PCR amplification was conducted by incubating the samples at 95uC for 30 s, followed by 40 cycles of 5 s at 95uC and 30 s at 60uC. A melting curve analysis was performed using the following cycling parameters: 60uC for 30 s, and 5uC temperature changes to the end temperature of 95uC. In all experiments, each PCR run included a negative extraction control (sterile water) and a negative PCR control; containing 1 ml DEPC treated H 2 O instead of DNA extract, to detect possible contamination of DNA. Samples were run in duplicate.

Coxiella burnetii PCR
The real-time PCR assay was performed on a CFX96 TM realtime PCR cycler (Biorad). The optimal concentration of primers was assessed by testing different concentrations (0.05, 0.15, 0.2, 0.3, and 0.5 mM) and by defining the one that gave the highest recorded fluorescence and the lowest threshold cycle (Ct), defined as the point at which the fluorescence crosses the threshold. Each reaction was run in duplicate in a mastermix containing 10 ml of SsoAdvanced SYBR Green Supermix (Biorad), 0.2 mM of each primer (Table 1), and 1 ml of purified DNA to a final volume of 20 ml. The thermal cycling conditions were assessed by testing different annealing temperatures (between 54uC and 60uC) during different times (5, 10, 20, 30 and 60 s.). The optimal qPCR efficacy was obtained using cycling profile included an initial denaturation step at 95uC for 3 min, then 40 cycles of 10 s at 95uC and 10 s at 60uC. A melting curve analysis was performed using the following cycling parameters: 60uC for 30 s, and 5uC temperature changes to the end temperature of 95uC. In all experiments, each PCR run included a negative extraction control (sterile water) and a negative PCR control (containing 1 ml DEPC treated H 2 O instead of DNA extract, to detect possible contamination of DNA). Samples were run in duplicate.

Campylobacter spp. PCR
The real-time PCR assay was performed on a CFX96 TM realtime PCR cycler (Biorad). Each reaction was run in duplicate in a mastermix containing 12.5 ml of 26 SYBR Permix Ex Taq Tli RNaseH Plus (TaKaRa), 0.2 mM of each primer (Table 1), and 1 ml of purified DNA to a final volume of 25 ml using nuclease-free water. The thermal cycling conditions were assessed testing different annealing temperatures (between 54uC and 60uC) during different times (5, 10, 20, 30 and 60 s.). The optimal qPCR efficacy was obtained using cycling profile included an initial denaturation step at 95uC for 30 s, then 40 cycles of 5 s at 95uC and 30 s at 60uC. A melting curve analysis was performed using the following cycling parameters: 60uC for 30 s, and 5uC temperature changes to the end temperature of 95uC. The generated melt peak represented the specific amplified product. In all experiments, each PCR run included a negative extraction control (sterile water) and a negative PCR control (containing 1 ml DEPC treated H 2 O instead of DNA extract, to detect possible contamination of DNA). Rose Bengal plate test All serum samples collected were screened using the Rose Bengal plate test (RBPT), according to the procedures described by Alton et al. [29]. B. abortus antigen (obtained from the Institut Pourquier, Montpellier, France) was used to screen sera for the presence of antibodies to Brucella spp. In brief, 30 ml of serum was mixed with an equal volume of antigen suspension on a glass plate and agitated. After four minutes of rocking, any visible agglutination was considered as positive. The degree of agglutination was graded on an ordinal scale from 0 (no agglutination) to 3 (coarse clumping), with corresponding RBPT scores of 0, 1, 2 and 3. All doubtful reactions were recorded as negative or zero scores.

Detection of the amplification product and sequencing
Ten ml of each PCR product were electrophoresed in agarose gel, stained with ethidium bromide and observed under UV illumination. The fragments at correct size were excised and further purified for sequencing. Briefly, DNA was extracted using the QIAquick Gel extraction Kit (Qiagen, Courtaboeuf, France), was then subjected to cycle sequencing using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems) and processed by the ABI 3100 Genetic Analyzer. The obtained sequences were compared with the sequences available in GenBank by using the BLAST server from the NCBI website (http://www.ncbi.nlm.nih.gov/BLAST).

Statistical analysis
The association of abortion with positive results for different agents by real-time PCR and frequencies of positive results between vaginal and milk samples were assessed using the Fisher's exact test. The correlation between real-time PCR and the serological test (RBPT) for Brucella cases was assessed using Kruskal-Wallis test (nonparametric test). In all tests, a p value,0.05 was considered to be statistically significant. We decided to use a one-sided p value as all the evidence points to the directionality of this relationship only going one way. All tests were conducted using the Graph-Pad Prism 5 statistical package (Graph-Pad Software, San Diego, CA, USA).

Ethical considerations
Samples were collected by clinical veterinarians as part of the usual screening scheme on farms and Tunisian ethical guidelines and animal welfare regulations were strictly respected. All herd owners had given an informed consent prior to the study. All samples were collected as part of routine care.

Results
Detection of Brucella spp.
The results of this study showed that 47 out of 150 investigated samples (31.3%) were positive for Brucella using the RBPT  were considered as negative. Interestingly, the mean Ct-value of samples negative for Brucella was significantly lower than that of samples with a strong reactivity, which is lower than that of samples with a weak reactivity (p,0.0001) (Fig. 1). In conventional gel PCR, all positive samples by real-time PCR produced amplicon of 63 bp when resolved on an agarose gel (Figs. S1B, S2A). All of them were identified as B. abortus based on speciesspecific PCR [30]. To confirm that positive samples were due to the presence of B. abortus DNA, PCR products were sequenced and analyzed using the BLAST web interface (http://blast.ncbi. nlm.nih.gov/Blast.cgi). The amplicons of these positive samples were found to be 100% identical to the B. abortus gene in the database. Details of cases positive for B. abortus are given in Table 2.
Detection of Chlamydiaceae, W. chondrophila, P. acanthamoebae Chlamydial DNA was detected in 27 (18%) vaginal swab samples, all obtained from the 150 cows that suffered from an abortion (OR = 28.7, 95% CI [1.72-478.9], p,0.0001). Briefly, these cows exhibited symptoms of reproductive disorders like abortion and other symptoms like fever, appetite loss, weakness and diarrhea. No milk sample was positive (OR = 67, 95% CI [4.04-1111], p,0.0001). In conventional gel PCR, the 27 vaginal swab samples positive by real-time PCR produced as expected amplicons of 207 bp when resolved on agarose gel 1.5% (Fig. S3A, S3B). PCR products were sequenced and analyzed using the BLAST web interface (http://blast.ncbi.nlm.nih.gov/Blast.cgi). Among 27 sequences obtained, 12 (8%) belonged to the Waddliaceae family, 8 (5.3%) belonged to the Parachlamydiaceae family and 7 (4.6%) belonged to the Chlamydiaceae family (Fig. S3A, S3B). All sequences corresponding to a Waddliaceae species demonstrated $99% similarity with W. chondrophila. All of them had already been identified by the SYBR Green real-time PCR we used in our previous work [12] to detect the presence of W. chondrophila. Since both PCR are targeting two different DNA regions, these results may clearly not result from PCR contamination with amplicons. All the sequences belonging to Parachlamydiaceae family demonstrated $99% of similarity with P. acanthamoebae. Four among the 7 sequences belonging to the Chlamydiacaeae family demonstrated $99% sequence similarity with C. abortus, whereas the remaining 3 sequences exhibited $99% sequence similarity with C. pecorum. These results were confirmed by two commercial kits (PrimerDesign genesig) specific for the quantification of C. abortus and C. pecorum. All control samples remained negative by real-time PCR. Details of cases positive for P. acanthamoebae, C. abortus, C. pecorum and W. chondrophila are given in Table 4.

Detection of L. monocytogenes
L. monocytogenes DNA was detected in 7/150 (4.66%) vaginal swab samples all obtained from the 150 cows that suffered of an abortion (OR = 6.74, 95% CI [0.37-119.9], p value = 0.08) ( Table 3). Only 3/150 (2%) milk samples were found to be positive (Table 3). These cases originated from different herds in different regions of Sfax. All control samples remained negative by real-time PCR. The real-time PCR melt curve data identified one peak with a melting temperature of 87uC for all samples (Fig. S4A,  S4B). In conventional gel PCR, all samples positive by real-time PCR produced amplicon of 63 bp when resolved on an agarose gel 3% (Fig. S4A, S4B). To confirm that positive samples were due to the presence of L. monocytogenes DNA, PCR products were sequenced and analyzed using the BLAST web interface (http:// blast.ncbi.nlm.nih.gov/Blast.cgi). The amplicons of these positive samples were 100% identical to the L. monocytogenes hlyQ gene in the databases. All control samples remained negative by real-time PCR.
Detection of Salmonella spp.
Salmonella DNA was detected in 5 (3.33%) milk samples all obtained from the 150 cows that suffered of an abortion (OR = 4.87, 95% CI [0.26-89.5], p value = 0.1658) ( Table 3). No salmonella DNA was detected in vaginal swab samples (OR = 0.08, 95% CI [0-1.6], p value = 0.0302) ( Table 3). These cases were collected from the same herd located in the southern region of Sfax. The real-time PCR melt curve data identified one peak with a melting temperature of 76uC for all samples (Fig.  S5A). In conventional gel PCR, all positive samples by real-time PCR produced amplicon of 243 bp when resolved on an agarose gel 1.5% (Fig. S5B). To confirm that positive samples were due to the presence of Salmonella spp. DNA, PCR products were sequenced and analyzed using the BLAST web interface (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The amplicons of these positive samples were 100% identical to Salmonella enterica serovar Typhimurium in the databases.
Detection of C. burnetii and Campylobacter spp.
C. burnetii was absent in all bovine samples taken from the 150 cows that have aborted and from the 64 control cows tested by the

Co-infections
From a total of 150 abortion cases collected, infectious agents were detected in 73 (48.66%) cases, 13 (8.66%) of which represented co-infections with two infectious agents. These observations may suggest that co-infections commonly exist in cattle herds. Table 5 shows the number of positive samples for each pair of pathogens.

Discussion
In many developing countries, studies conducted to estimate the true prevalence of abortigenic agents were only occasional and limited to certain farms or regions. Consequently, the control programs are not very effective, even for established agents of abortion in cattle. Moreover, the etiology remains unknown in many cases of bovine abortion. In this study, we thus focused on new or not well-defined emerging agents in Tunisian cattle. Vaginal swab samples were used in many previous studies for the detection and isolation of many abortive bacteria [28,[31][32][33][34][35], because this type of sampling (i) is simple, fast and it is not necessary to have access to the placenta and (ii) limits the external contamination, which is often the case in the collection of the cotyledons if the placenta hit the ground [36]. The choice of milk specimens is based on the fact that milk is the most frequent shedding route in chronically infected cows [37][38][39] and the transmission of infection to humans is mainly due to the consumption of raw milk and dairy products [39][40][41][42][43].
This work studies the etiological agents of some infectious causes of bovine abortion during mid-to late-gestation in Tunisian dairy herds. The high percentage of positive cases for B. abortus detected in the present study both through RBPT (31.3%) and PCR (30.6%) was unexpected. Quantitatively, the current results are relatively similar to other studies based on the detection of Brucella in African veterinary samples by serology tests [44][45][46] and by the IS711 real-time PCR [47]. There is a strong correlation between the DNA copies (Ct values) obtained using the real-time PCR on  (Table 3), a statistically significant finding. Reports on the ability to detect Brucella spp. in milk samples from infected animals by PCR can be somewhat conflicting when compared to standard serological [48] and bacteriological methods. Thus, the results of our study confirm that milk is not the best sample to detect Brucella DNA. No correlation was found between B. abortus positivity and age, herd size and breeding system. Considering abortion as the only clinical sign, there was an association between the occurrence of abortion and positive results by the RBPT (31.3% vs 7.81%, p,0.0001) and real-time PCR (30.6% vs 0%, P,0.0001) (Table 3), a statistically significant and clinically important finding. The percentage of Brucella contamination showed that brucellosis is endemic in cows from different regions of Sfax, despite the extensive vaccination programs implemented as one of the main control measures. Chlamydiosis seem also to play an important role in bovine abortion in Tunisia, since twenty-seven samples out of 150 (18%) were found to be positive by real-time PCR. The combination of real-time PCR and two commercial diagnostic kits identified 7 positive cases for C. abortus (n = 4) and C. pecorum (n = 3). Using the specific real-time PCR, three samples were shown to be infected by both C. abortus and B. abortus. The association of these two known abortigenic agents was not described yet. In these cases of mixed infection, it remains unclear if both or only one is the etiological agent of bovine abortion. The 3 cases positive for C. pecorum were negatives for all other agents investigated in this study. C. pecorum is known as the causing agent of conjunctivitis, arthritis, pulmonary inflammation and fertility disorders in Tunisian ruminants and could also be detected in ovine placenta possibly leading to abortion [49]. On the other hand, C. pecorum occurs in asymptomatic gastrointestinal infections of ruminants. Thus, a transmission of infection from sheep to cattle cannot be excluded especially in areas where there is a contact between the two animal species. Chlamydia-related organisms seem to play an important role in bovine abortion since 20 samples were positive (13.3%). Similar results were given by Swiss studies based on the detection of Chlamydia-related organisms in veterinary samples using another PCR [50]. The involvement of W. chondrophila in abortion in Tunisia was previously confirmed in a previous work [12]. In the present study, 8 (5.33%) bovine vaginal samples were also positive for P. acanthamoebae in the subset of 150 samples. The results for all 8 positive samples were confirmed by the P. acanthamoebae PCR previously described by Casson et al. [51]. This is the first description of this Chlamydia-related organism in bovine abortion in African cattle. All positive samples were collected from cows belonging to a same herd including 1500 dairy cows. The percentage of abortion in this herd was 12% during 2011 knowing that all cows in this herd were vaccinated against Brucella, Bovine Viral Diarrhoea (BVD) and Infectious Bovine Rhinotrachetis (IBR) viruses. These 8 samples were however negatives for C. abortus, C. pecorum, B. melitensis, C. burnetii, L. monocytogenes, Salmonella spp., Campylobacter spp., BVD and IBR Viruses. The co-infection of P. acanthamoebae and B. abortus (three cases) has not been described yet. Again it remains unclear if both or only one agent caused bovine abortion. For the five remaining cases, no other agents have been detected and no other conditions explaining abortion were documented. Thus, our data suggest that P. acanthamoebae is a possible factor in some cases of abortion. This is consistent with a Swiss work that identified P. acanthamoebae in placenta of aborted bovines by PCR, immunohistochemistry and electron microscopy [13]. Based on the present work, Chlamydia and other Chlamydiarelated bacteria do not appear to be excreted in the milk of cows. This is in concordance with the reports of other authors stating that shedding of the C. abortus in milk is possible but uncommon [52]. It is likely that the excretion of Chlamydia organisms was more strongly manifested with vaginal discharge than with milk (18% vs 0%, p,0.0001), a statistically significant and clinically important finding.  L. monocytogenes is another agent that may be involved in bovine abortion. This is the first time that bovine abortion samples from Tunisia were investigated for L. monocytogenes. Besides the 2 cases of mixed infection with L. moncytogenes and W. chondrophila already described above [12], 5 new cases positive for L. monocytogenes were found, in which no other agents have been detected. There was no association between the occurrence of abortion and positive results for this bacterium by real-time PCR (4.6% vs 0%, p value = 0.08). In addition, our results suggest that Listeria may be readily detectable in vaginal samples and also in milk (2.6% vs 2%, p value = 0.5). Infection with L. monocytogenes is transmissible to humans through milk and milk products [53], in which case it can cause systemic and often fatal diseases. There are other agents such as Salmonella that may be involved in cattle abortion and can be transmitted to humans through the consumption of milk. In this study, S. enterica Typhimurium was detected only in 5 milk samples taken from 5 cows that suffered an abortion. In these 5 cases, B. abortus DNA was also detected by a specific real-time PCR in vaginal and milk samples. When mixed infection occurs, it remains unclear if both or only one agent caused bovine abortion. If we consider abortion as the only clinical sign, there is no association between the occurrence of abortion and positive cases for Salmonella by real-time PCR (3.3% vs 0%, p value = 0.1658). In many previous studies, it was revealed that abortion was a quite unusual clinical finding of S. enterica Typhimurium infection in cattle because it hasn't a higher potential for systemic dissemination in body [54]. S. enterica Typhimurium is often associated with enteritis that usually affects young calves, resulting in marked acute diarrhea [55]. All these data suggest that B. abortus may be the most probable agent of abortion in these 5 cows belonging to the same herd located in the south of Sfax. Unlike to other agents cited above, Campylobacter spp. and C. burnetii were not detected in the limited number of cases studied. This is in contrast with previous studies in ruminants and human beings that supported a role of C. burnetii in different mammals. In fact, the seroprevalence of Q fever in sheep flocks in different Tunisian regions was 40% and C. burnetii was the abortigenic agent in 17% of sheep and goats [56]. C. burnetii does not seem to have a crucial role in bovine abortion in contrast to the small ruminants. Previous studies already showed negligible percentages for the involvement of Campylobacter spp. in bovine abortion [34,57]. More samples should be collected to survey the infections caused by these two bacteria in dairy herds.
This study highlighted the real prevalence of brucellosis and the presence of diseases that are not subject to routine control diagnosis, such as chlamydiosis, salmonellosis and listeriosis. It also demonstrated the importance of Chlamydia and Chlamydia-related organisms as major causes of abortion. Strong evidence suggests that P. acanthamoebae should be considered as a new abortigenic agent in Africa. As well, this study underlined the notion of coinfection, which emphasizes the great interest of a differential diagnosis of different abortion causes.