Naturally acquired bovine besnoitiosis: Disease frequency, risk and outcome in an endemically infected beef herd

The recent spread of bovine besnoitiosis warrants further epidemiological investigations to improve the knowledge on disease development. Thus, a 4-year longitudinal open cohort study was conducted in the first German cattle herd naturally infected with Besnoitia besnoiti. At seven herd-visits between 2008 and 2012, fourteen breeding bulls (>1.5 years) and 131 females (>1 year) were examined clinically and serologically. In females, clinical and serological prevalences, incidence and remission rates were determined. In addition, the association of age, antibody levels and number of visible parasitic cysts with clinical and serological outcome was investigated. The seroprevalence (89.4%-100%) and serological incidence rate (140.5 per 100 animal-years) were considerably higher than the clinical prevalence (23.5%-36.6%) and clinical incidence rate (16.7 per 100 animal-years). Of 33 new clinical and 12 new serological cases, only 6.7% (3/45) attracted attention with clinical signs of acute bovine besnoitiosis. The apparent serological remission rate (1.9 per 100 animal-years) was considerably lower than the clinical remission rate (37.3 per 100 animal-years). A median cyst score of <1 and mean immunofluorescent antibody test (IFAT) titre of ≤1,600 over the entire observation period was significantly associated with a negative clinical outcome at the end. Overall cyst score was not significantly associated with serological outcome and age had no significant influence on clinical and serological outcome. Within 4 years, there was a significant reduction in cyst scores and IFAT titres in the same animals, leading to eight clinically and serologically negative animals in the end. Two initially negative animals achieved clinical and apparent serological remission in about 2.5 years. In bulls, the time between herd entry and seroconversion was 7-30 months and the serological incidence rate was nearly identical to the rate in females (142.0 per 100 animal-years). This shows that a high B. besnoiti prevalence leads to infection of bulls within a short time period.


Summary
The recent spread of bovine besnoitiosis warrants further epidemiological investigations to improve the knowledge on disease development. Thus, a 4-year longitudinal open cohort study was conducted in the first German cattle herd naturally infected with Besnoitia besnoiti. At seven herd-visits between 2008 and 2012, fourteen breeding bulls (>1.5 years) and 131 females (>1 year) were examined clinically and serologically. In females, clinical and serological prevalences, incidence and remission rates were determined. In addition, the association of age, antibody levels and number of visible parasitic cysts with clinical and serological outcome was investigated.
The seroprevalence (89.4%-100%) and serological incidence rate (140.5 per 100 animal-years) were considerably higher than the clinical prevalence (23.5%-36.6%) and clinical incidence rate (16.7 per 100 animal-years). Of 33 new clinical and 12 new serological cases, only 6.7% (3/45) attracted attention with clinical signs of acute bovine besnoitiosis. The apparent serological remission rate (1.9 per 100 animalyears) was considerably lower than the clinical remission rate (37.3 per 100 animalyears). A median cyst score of <1 and mean immunofluorescent antibody test (IFAT) titre of ≤1,600 over the entire observation period was significantly associated with a negative clinical outcome at the end. Overall cyst score was not significantly associated with serological outcome and age had no significant influence on clinical and serological outcome. Within 4 years, there was a significant reduction in cyst scores and IFAT titres in the same animals, leading to eight clinically and serologically negative animals in the end. Two initially negative animals achieved clinical and apparent serological remission in about 2.5 years. In bulls, the time between herd entry and seroconversion was 7-30 months and the serological incidence rate was nearly identical to the rate in females (142.0 per 100 animal-years). This shows that a high B. besnoiti prevalence leads to infection of bulls within a short time period. . The lesions of this stage are induced by replicating tachyzoites within vascular endothelial cells (Basson, McCully, & Bigalke, 1970;. Severe chronic bovine besnoitiosis leads to thickening and wrinkling of the skin, alopecia and non-healing sole ulcers (Besnoit & Robin, 1912;. In bulls, infection can lead to orchitis, sterility and adverse effects on semen quality (Esteban-Gil et al., 2016;Kumi-Diaka, Wilson, Sanusi, Njoku, & Osori, 1981). In the chronic stage, typical parasitic cysts can be detected in various tissues. In some animals, large numbers of parasitic cysts are located in skin, scleral conjunctivae (SC), vestibulum vaginae (VV) and non-intestinal mucosa Majzoub et al., 2010;McCully, Basson, Van Niekerk, & Bigalke, 1966). In contrast to these peripheral tissues, visceral organs harbour markedly reduced parasite load, and the reasons for this remain unknown (Frey et al., 2013;Pols, 1960).
All breeds of cattle are considered intermediate hosts of the parasite, but despite several attempts to identify a potential definitive host, the presumed heteroxenous life cycle is still unknown (Basso, Schares, Gollnick, Rutten, & Deplazes, 2011;Diesing et al., 1988;EFSA, 2010).

| Clinical examination
Examinations and samplings were in accordance with national law and conducted by veterinarians during the specific herd health management routine in Herd-BbGer1. In 2008, the procedure was additionally supervised by the regional authorities. No ethical approval was required for this study.
Every clinical examination was performed by an experienced cattle veterinarian (NSG). Animals were restrained in a cattle chute, and the head was tethered. The examination premises for visual examination were illuminated with spotlights for direct and indirect lightening and headlamps. For each animal at each examination, the number of parasitic cysts in the mucous membrane of VV or prepuce and the mean number of cysts in the SC were estimated and assigned to the following scores: 1: 1-5 cysts; 2: 6-10 cysts; 3: 11-20 cysts; 4: 21-30 cysts; 5: >30 cysts. Animals with a cyst score of at least 1 in one location were defined as clinically positive.

| Blood sampling, serology and histology
Blood samples for serology were taken from the jugular or tail vein.
An in-house enzyme-linked immunosorbent assay (ELISA) using affinity-purified B. besnoiti tachyzoite antigens was performed to examine all sera. In the ELISA, a cut-off of SP = 1.756 was applied . Sera with a SP value above this cut-off were defined as positive. The serological status of every animal was defined by the ELISA result. To determine antibody levels, an IFAT was performed as previously described with a reciprocal titre of 200 as positive cutoff . In a few sera, serological results were confirmed by immunoblots using antigens of B. besnoiti tachyzoites and bradyzoites as previously described . Preputial samples of bull 215 were fixed, embedded, cut and HE stained as previously described .  For exposure category cyst score, animals were grouped into two groups according to their median overall cyst score, which was calculated for every individual animal from every clinical examination.

| Data analysis
One group comprised animals with a median overall cyst score <1 and the other animals with a median overall cyst score ≥1. For exposure category IFAT titre, animals were grouped into two groups according to their mean IFAT titre calculated for every individual animal from every serological examination. One group comprised animals with a mean reciprocal IFAT titre ≤1,600 and the other animals with a mean reciprocal IFAT titre >1,600. For exposure category age, animals were grouped into three groups (<2; 2-5; >5 years) according to their age at HC1.

| Clinical and serological findings and prevalence
In total, 771 examinations were performed, n = 739 in female and n = 32 in male animals (Table 1) tested seronegative at the respective HC. The four sera of these females were also tested with tachyzoite and bradyzoite antigen immunoblots, which revealed also negative results.
During the 4-year period, 46 animals were removed from the herd, leaving 85 of the initial 131 female cattle at HC7.

| Clinical and serological incidence and remission rates
Of 83 clinically negative female cattle at HC1, 12 were lost to follow-up and 33/71 (46.5%) acquired SC or VV cysts until HC7 ( appeared at the spring HCs, resulting in an autumn:spring ratio of 1:5. Of 13 serologically negative female cattle at HC1, one was lost to follow-up and the remaining twelve animals (100%) seroconverted. Animals were observed for 103 months, which equals 8.5 animal-years, resulting in an incidence rate of 140.5 per 100 animalyears. The autumn:spring ratio was 1:1.4.
Of 118 serologically positive female cattle at HC1, 42 were lost to follow-up and 9/76 (11.8%) lost seropositivity, all of them before HC7. Animals were observed for 4,345 months, which equals 362.1 animal-years, resulting in an apparent remission rate of 1.9 per 100 animal-years. Eight of these nine cows were also clinically negative.  (Table S1). To determine the association with clinical presentation, we calculated mean and median SC and VV cyst scores of all clinical examinations in all HCs of these animals.
Of the 26 clinically and serologically positive animals at HC1 (Table S1,  cyst scores between those two groups was also highly significant (Mann-Whitney test, p < .0001). To outline the clinical course from HC1 to HC7 of the two groups, the mean cyst scores are shown in Figure S1.
Of 51 clinically negative and serologically positive animals at HC1, 38 retained their status at HC7, four became clinically and seronegative in the end and nine developed SC or VV cyst at later HCs.
Of the group of 85, only seven animals were clinically and serologically negative at HC1 (Table S1, Table S3).
The association of the exposure category age with clinical or serological outcome and median cyst score and serological outcome was not significant (Fisher's exact test, p > .05).

| Reduction in cyst scores and IFAT titre over time resulted in few test negative animals in the end
To determine whether cyst numbers decreased over time, we compared cyst scores of 27 clinically positive animals at HC1 (Table 2) with the same animals' score at HC7. There was a highly significant reduction in SC cyst score from HC1 (median = 1, mean = 1.52, SD = 1.23) to HC7 (median = 0, mean = 0.56, SD = 1.2; Mann-Whitney test, p < .0001). There was also a significant reduction in VV cyst score from HC1 (median = 1, mean = 1.15, SD = 1.27) to HC7 (median = 0, mean = 0.7, SD = 1.43; Mann-Whitney test, p < .01).
Only eight female cattle were clinically and serologically negative at the end of the study (HC7). Only one of these was clinically negative at all HCs. Two of these were those animals which had developed first signs of bovine besnoitiosis during the study period and experienced apparent remission until HC7 (Section 3.4). All of these animals were additionally tested with immunoblots, which were also negative.

| Every breeding bull became infected with
B. besnoiti   prior to introduction into Herd-BbGer-1. All of them seroconverted during the study course. The time interval between herd entry and seroconversion was 7-30 months.
Ten seronegative bulls were observed for 84.5 months (7.0 animalyears), resulting in an incidence rate of 142.0 per 100 animal-years.
After introduction into the herd in April, the owner observed three   Clinical and serological incidence rates were quite high in this study. In a study on clinical and serological dynamics of bovine besnoitiosis over 4 years, a clinical incidence of up to 16.7 per 100 animal-years was reported in 26 animals studied (Guti errez-Exp osito, . In the present 4-year study on nearly three times as many animals (n = 71), the clinical incidence rate was the same. In contrast, the serological incidence in this study was considerably higher, that is, 140.5 per 100 animal-years compared with a maximum of 32 per 100 animal-years in the study by Guti errez-Exp osito, . This is most likely due to the very low number of seronegative animals at study start and the large proportion of infected animals. New clinical cases showed a tendency to occur between autumn and spring examinations, that is, were first detected at the spring HCs. This probably reflects the time interval from infection, which occurs preferably during summer, to appearance of first grossly visible tissue cysts. The incubation period of bovine besnoitiosis was reported as~13 days (Basson et al., 1970;Bigalke, 1968). The time period from first day of fever until first appearance of SC cysts ranged between 37 and 49 days (Bigalke, 1968). In a previous study, we could confirm this time period; we observed first SC cysts in two naturally infected animals 28 and 38 days after beginning of acute stage . This time frame of SC cyst development can explain the observed tendency, because animals infected in late summer need some time to display SC cysts and therefore appear first positive in the spring HCs.
The design of this study allowed the examination of the same individual cattle over a 4-year period, which enabled us to observe the disappearance of clinically visible parasitic cysts and to calculate a clinical and serological remission rate. Clinical recovery and negative seroconversion has been reported before (Guti errez-Exp osito, Jacquiet et al., 2010), but remission rates of bovine besnoitiosis have not been determined so far. Interestingly, the clinical remission rate was quite high (37 per 100 animalyears) and significantly higher than the clinical incidence rate (16 per 100 animal-years) in this study, which eventually led to a reduction in clinical prevalence in the last two HCs. In contrast, the apparent serological remission rate was very low (~2 per 100 animal-years), emphasizing the long time needed to overcome the infection and to become serologically negative again. It is striking that the clinical remission rate is more than two times higher than the clinical incidence rate. The majority of animals achieved clinical remission in the time period between autumn and spring, that is, were negative the first time at the spring HCs. It is tempting to speculate that the observed seasonality might be associated with increased insect activity and parasite contact in summer and autumn.
Using association measures and comparisons between different subgroups, we were able to show that overall cyst scores (reflecting cyst numbers) were significantly associated with disease status and disease outcome. With a RR of 3.64, animals with lower cyst scores are three times as likely to result in a negative clinical outcome as animals with higher cyst scores. Likewise, with a RR of 2.23, animals with lower IFAT titre are two times as likely to result in a negative clinical outcome as animals with a higher IFAT titre. A correlation between parasite load and disease signs has also been reported previously (Frey et al., 2013;Gollnick et al., 2015).
Only four animals displayed SC and VV cysts but tested seroneg-  (Schares et al., 2016) are necessary to study this aspect in those animals with an apparent complete remission.
In this study, 88.9% (8/9) of the animals which showed apparent serological remission were also clinically negative at HC7. Obviously, a negative result in the clinical examination is likely a prerequisite for apparent negative seroconversion. This suggests that no or only very low parasitic loads are necessary for an animal to become serologically negative. Findings in previous studies are in accord with this: animals with high IFAT titres had a higher B. besnoiti DNA load (as determined by lower real-time PCR thresholds) in skin and vaginal scrapings, illustrating the association between parasitic load and antibody response (Schares, Maksimov, et al., 2011).
The most probable reason for the loss of SC and VV cysts is their destruction by the immune system. As the host reacts to the cysts, they are circumferentially surrounded by several layers of eosinophilic and granulomatous inflammation to destroy them (Langenmayer, . However, further studies are needed to investigate whether the immune response eventually leads to the effective destruction of every parasitic cyst and complete remission. As there are always cysts without surrounding inflammatory cells, there are indications for an effective immune evasion by B. besnoiti. Immune evasion is poorly studied, and immunosuppressive proteins and factors have yet to be determined.
The thick secondary outer cyst wall could act as one of the means of immune evasion providing effective sequestration from activated effector cells. Further research regarding that topic is certainly needed.
The effect of this rotation on transmission dynamics between bulls and females is likely minor due to the initial high herd-seroprevalence. In every subherd, there were few negative and many positive animals, which could facilitate the transmission between females or females and bulls. Every breeding bull used in Herd-BbGer1 got infected and many of them shortly after introduction into the herd.
During the summer period, six of nine bulls got infected. A recent study in an Italian herd also reported high incidences among bulls with twice as much seropositive animals within 1 year (Gazzonis et al., 2017). In contrast, in our study, serological incidence rates between bulls and females were nearly the same. The high disease prevalence in female cattle constituted a large reservoir for rapid infection via insect transmission and likely also favoured an infection via direct contact. Whether bulls are at an increased risk of getting bovine besnoitiosis is still not clear. In a cohabitation experiment, the bull remained negative despite frequent mating activity . Similar results were obtained in an Italian herd were mating with a seropositive cow did not enhance odds of infection in seronegative bulls (Gazzonis et al., 2017). Nevertheless, mating of seropositive with seronegative animals has to be regarded as a possible route of B. besnoiti transmission.
In conclusion, the endemic Herd-BbGer1 was characterized by a very high seroprevalence and significantly lower clinical prevalence.
The clinical and serological incidence rates were high. The ongoing herd health scheme, in which females with severe clinical signs and bulls with clinical disease were rapidly removed from Herd-BbGer1, was not sufficient. To control the disease, an integral control programme could be a feasible option. Obviously, the overwhelming presence of subclinically, mildly and moderately infected cattle was enough to cause these high incidence rates. On the opposite, the clinical remission rate was higher than the clinical incidence rate leading to a reduction in clinical prevalence in the end. In Herd-BbGer1, there was a significant reduction in total cyst numbers and IFAT titres over the 4 years leading to clinical and serological remission in eight animals. Low cyst numbers and low IFAT titres were significantly associated with negative clinical outcome. Thus, in this situation in an endemic herd, estimating individual animal cyst numbers can aid to assess degree of infection and provide a prognosis regarding the possible fate of the respective animal.

ACKNOWLEDG EMENTS
The authors express their gratitude for the excellent support pro-