Tick-borne haemoparasite prevalence and Theileria parva strain diversity in African buffalo (Syncerus caffer) from northern Botswana and Gonarezhou National Park, Zimbabwe

Abstract

The African buffalo (Syncerus caffer) is host for many pathogens known to cause economically important diseases and is often considered an important reservoir for livestock diseases. Theileriosis, heartwater, babesiosis and anaplasmosis are considered the most important tick-borne diseases of livestock in sub-Saharan Africa, resulting in extensive economic losses to livestock farmers in endemic areas. In this study a variety of tick-borne haemoparasites (Theileria, Babesia, Anaplasma and Ehrlichia species) were identified either as mixed or single infections using the reverse line blot (RLB) hybridization assay from buffalo blood samples in the Chobe National Park (CNP) and Okavango Delta (OD), Botswana and in the Gonarezhou National Park (GNP), Zimbabwe. Also, a quantitative real-time PCR (qPCR) assay was used to identify Theileria parva more specifically in both these countries while the indirect fluorescent antibody test (IFAT) was used to identify Theileria parva more specifically in Botswana only. An attempt was made to characterize T. parva through the size differentiation of p67 genotypes and characterization of the variable regions of T. parva antigen genes, p104 and PIM, by using semi-nested PCR-RFLP profiles. This is the first report of tick-borne haemoparasites in northern Botswana and one of only a few from Zimbabwe. This study identified the following tick-borne haemoparasites: Theileria spp. present, T. parva (60%) and T. mutans (37%) were the most prevalent in the two wildlife areas from Botswana, while Theileria sp. (sable) (50%), T. parva (48%) and T. mutans (38%) were most prevalent in GNP, Zimbabwe. Other species of interest were Anaplasma marginale subsp. centrale (30%), A. marginale (20%), Babesia occultans (23%) and Ehrlichia ruminantium (6%) in Botswana and Anaplasma marginale subsp. centrale (25%) and Babesia occultans (15%) in Zimbabwe. Generally speaking, the buffalo population in the OD sample had lower levels of haemoparasite infection than the buffalo in the CNP and GNP, with the exception of Theileria sp. (buffalo) and to a lesser extent Anaplasma sp. Omatjenne and B. bovis (in the two later cases, where very few positives were detected). Interesting findings included: Anaplasma sp. Omatjenne identified in this study, another research group identified 16.5% to be positive in their samples, but the parasite was found in very low concentrations (3.1%) in our study. B. occultans causes a benign form of cattle babesiosis and was also reported in South Africa by by this research group for the first time. Our study identified 21.1% samples to be positive compared to the study in Hluhluwe-iMfolozi Park, South Africa (50.0%). This study serves as another report of the presence of these two parasites in buffalo. As in Uganda, the pathogenic B. bovis has previously been reported to be absent from buffalo in Botswana but were identified at a low concentration in OD. Similarly, E. ruminantium could be identified in a few CNP and OD buffalo tested. The significance of buffalo as possible reservoir host of some of these economically important haemoparasites (i.e. A. marginale, E. ruminantium) remains unknown. Theileria sp. (sable), which is fatal to sable (Hippotragus niger) and roan antelope (Hippotragus equinus), but non-pathogenic to buffalo was identified in some of the Botswana and Zimbabwe buffalo but positive RLB signals might be due to cross reactions of the Theileria sp. (sable) probe with T. velifera. Theileriosis is recognized as a major threat to the livestock industry as some members of the genus may cause severe disease and mortality, whereas others may only cause mild or subclinical infections. In this study the efficiency of IFAT, qPCR and RLB in identifying T. parva were compared to each other. qPCR was the most effective (81%) followed by IFAT (74%) and then RLB (60%) in Botswana. In Zimbabwe, qPCR (70%) identified more samples to be positive than RLB (48%). The level of agreement between the tests for detection of T. parva positive animals was higher between qPCR and IFAT (kappa=0.56), than between qPCR and RLB (kappa=0.26) or the latter and IFAT (kappa=0.15) in Botswana. The kappa agreement between qPCR and RLB in Zimbabwe was 0.27. The RLB, IFAT and qPCR tests all indicated a high prevalence of T. parva in the study areas. This indicates a high risk of spreading Corridor disease caused by T. parva from buffalo to cattle by the vector ticks at the wildlife-livestock interface. Several T. parva antigen genes have been identified as good candidates for differentiation between buffalo-derived and cattle-derived T. parva isolates. Some of these genes include: p67, p104 and the polymorphic immunodominant molecule (PIM). These genes were amplified in an attempt to differentiate between buffalo-derived and cattle-derived profiles. Amplification of p67 in this study led to the identification of three of the four known p67 alleles. Cluster analysis for p104 showed that samples from Botswana and Zimbabwe clustered together in clade B with themselves and with samples from Hluhluwe while all samples from the Kruger National Park clustered in clade A with samples from Ladysmith. The cluster analysis of PIM revealed that samples from this project (Botswana and Zimbabwe) cluster in four clades, distinct from all samples from South Africa, which, except for one sample from Hluhluwe, clustered in a single clade. In conclusion, this study highlights the diversity of haemoparasites present in African buffalo from northern Botswana and Zimbabwe and also the role of African buffalo as a sentinel species for livestock tick-borne pathogens. Important tick-borne haemoparasites identified in this study included: T. parva, A. marginale, B. bovis and E. ruminantium. This study reconfirmed that p67 profiles are too complex and could not be used to distinguish between cattle- and buffalo-derived T. parva isolates. Mixed infections of p104 and PIM profiles generated by PCR-RFLP analysis were too complex to successfully differentiate between known profiles. Further cloning and sequencing of single infections are needed.