Background

Bovine babesiosis, the most economically important vector-borne disease of livestock globally, is mainly caused by protozoan parasites Babesia bovis and Babesia bigemina that are transmitted by tick species in the sub-genus Boophilus [1]. The disease causes major economic losses through animal mortality, poor growth rates, reduced milk yields in sick or recovered animals, and direct costs of tick control and disease treatment [2]. In Africa, B. bigemina is endemic and currently more widespread than B. bovis, reflecting the distribution of the indigenous tick vector Rhipicephalus decoloratus, which is more widely distributed in Africa, although Rhipicephalus evertsi evertsi is also a known vector [3]. Rhipicephalus microplus, which is a more competent vector of B. bovis, has become established in mainland Tanzania with evidence of displacement of the native R. decoloratus [4]. The tick had only been recorded in a small focus in coastal Kenya [5, 6]. However, in the past 15 years, R. microplus, which is a definitive vector for both B. bovis and B. bigemina, has been confirmed to occur in several countries in West Africa [7,8,9], Central Africa [10, 11] and recently East Africa [12,13,14,15,16]. The impact of this dispersal on the occurrence of B. bovis and the risk of pathogenic babesiosis in the affected African countries is yet to be assessed.

Clinical bovine babesiosis presents with significant haemolysis of the red blood cells, continuous fever, anaemia and often haemoglobinuria, which colours the urine reddish brown giving the disease the common name ‘red water’. Infections associated with B. bovis are often acute or subacute and have a shorter time course with more severe nervous symptoms rapidly leading to death or a protracted recovery rate in non-fatal cases [17]. In dairy cows, abortion and reduced or complete loss of milk (agalactia) are early signs of Babesia infection. Redwater causes high mortality and morbidity in susceptible livestock, especially in exotic and cross-breed cattle. Mortality rates of 30% for B. bigemina and 70–80% for the more pathogenic B. bovis infections have been observed [2]. Indigenous breeds of cattle can also be greatly affected by the less pathogenic B. bigemina under conditions of poor health or nutrition, a situation that is common in many vast areas of Africa, including Kenya [2], where other tick-borne diseases also occur. Babesiosis caused by B. bigemina is characterized by low parasitaemia of < 1%. In contrast, B. bovis infection has a high parasitaemia of > 10%, frequently with sequestration of infected red blood cells in cerebral capillaries resulting in symptoms which are often fatal. Cattle that recover from primary acute babesiosis, either naturally or after treatment, remain persistently infected and serve as a source of future tick infection and transmission [17, 18].

There is paucity of data on the status and occurrence of R. microplus outside the coastal counties in Kenya. Therefore, R. decoloratus, which is widely distributed in all agriculturally productive areas of eastern, central, Rift Valley and western Kenya [19], has been regarded as the major vector of bovine babesiosis in Kenya. Previously, McLeod and Kristjanson [20] predicted that 70% of Kenyan cattle were at risk from babesiosis and anaplasmosis with estimated annual economic losses amounting to $6.9 million per year. Co-infestation of animals with multiple tick species is typical, so co-infection with multiple tick-transmitted pathogens and significant disease burden are frequently encountered.

Infection with redwater in Kenya is mainly inferred from clinical signs; microscopic examination of blood smears for Babesia parasites is usually not performed because of the characteristic haematuria. Given the limited records of R. microplus in Kenya to date, redwater has always been attributed to the more ubiquitous B. bigemina vectored by R. decoloratus [21,22,23,24,25,26]. However, the recent confirmation of the occurrence of R. microplus in coastal Kenya by Kanduma et al. [13] indicates the existence of B. bovis. Babesia bovis DNA has been reported previously in cattle blood in Kenya [26, 27]. The use of molecular tests increases the sensitivity for detection and enables differentiation of B. bovis from other Babesia parasites. Kim et al. [28] and Zhang et al. [29] previously developed and validated highly sensitive quantitative qPCR TaqMan probes for detecting, quantifying and differentiating B. bovis from B. bigemina. In this study, we used these probes to investigate the occurrence of B. bovis in Kenya considering the recent reports of R. microplus in the local cattle populations.

Methods

Study site

A cross-sectional baseline survey was conducted in May 2019 in 12 sites in Kwale County, Kenya (Fig. 1) to determine the occurrence of B. bovis. The county is situated along the Kenyan coastline neighbouring the Indian Ocean on the East and Southeast and Tanzania on the Southwest. The county has a tropical type of climate with an average temperature of 23 °C with a high of 25 °C in March and a low of 21 °C in July. Annual precipitation is < 800 mm with the coastal parts of the county receiving > 1000 mm of precipitation per year, while most of the central to west areas receive around 500–750 mm. Rainfall is bi-modal with a short rain season from October to December and a long rain season from April to July. Detailed geo-climatic characteristics of the county have previously been described [13].

Fig. 1
figure 1

Map of Kwale County showing the 12 localities (shown in numbers) where the 506 cattle were sampled. The sampling localities were: (1) Miaji farm (n = 40), (2) Matuga (n = 13), (3) Ukunda (n = 28), (4) Mwanjaba (n = 10), (5) Kipabane (n = 43), (6) Kichaka simba (n = 107), (7) Majiboni (n = 32), (8) Tangini (n = 59), (9) Shimoni Kidimu (n = 34), (10) Shimoni (n = 32), (11) Kiwambale (n = 33) and (12) Vanga (n = 75). Topography data were acquired from ASTER Global Digital Elevation Model Version 3, while the shapefiles for administrative boundaries were obtained from the World Resource Institute website (https://www.wri.org). ArcGIS Desktop version 10.8 was used to prepare and analyse the layers and produce the final map

Full size image

Cattle blood sampling

In total, 506 adult cattle were randomly sampled across 12 sites located in three sub-counties, namely, Matuga, Msambweni and Lunga Lunga (Fig. 1). Since this was a baseline survey to screen for presence of Babesia parasites, opportunistic sampling was done with no stratification at farm/household level and data on risk factors were also not collected. Five millilitres of blood was collected from the jugular vein into EDTA vacutainers (BD, USA). The samples were stored in a cool box with ice packs and later transported to the regional laboratory in Kwale where they were refrigerated at 4 °C for 3 days. The samples were later transported to the International Livestock Research Institute (ILRI) Laboratories in Nairobi where subsequent investigations were conducted.

Whole-blood genomic DNA isolation

Whole cattle blood in EDTA tubes was thawed and thoroughly mixed by gentle rocking. Whole-blood genomic DNA was isolated from 300 µl of blood using the Promega Wizard® genomic DNA purification kit (Promega Corporation, Madison, WI, USA) following the manufacturer’s protocol. DNA yield and purity were determined by spectrophotometry using Nanodrop2000 spectrophotometer (Thermo-Scientific, USA). The DNA was stored at − 20 °C until use.

Detection of Babesia DNA

A TaqMan multiplex real-time PCR targeting two Babesia bovis and one B. bigemina genes was used to detect Babesia DNA. One B. bovis primer and probe set targeted the nuclear 18S rDNA [28] while the other targeted the mitochondrial cytochrome b gene [29]. The B. bigemina primers and probe were derived from the cytochrome b gene. Details of the primer and probe sequences used in the multiplex assay are listed in Additional file 1: Table S1. The real-time PCR was conducted in 200-µl 96-well plates in a QuantStudio™ 5 detection system (Thermo Fisher Scientific, Waltham, MA, USA). Samples were prepared in 20-µl reactions, which included 10 µl of 2 × Universal Probe qPCR Master Mix (New England Biolabs, Ipswich, MA, USA), 0.8 µl of 10 pmol of each oligonucleotide primer, 0.2 µl of 10 pmol of the fluorescence-labelled probe, 3 µl of 10 ng genomic DNA template and 5 µl of RNase/DNase free water. Temperature cycling parameters were 95 °C for 3 min followed by 45 cycles of 95 °C for 10 s and 54 °C for 25 s. Each PCR run included a B. bovis control sample extracted from an in vitro cell culture and B. bigemina-positive control sample obtained from an infected calf and a no-template PCR grade water sample as negative control. The cycle quantification (Cq) scores corresponding to the PCR cycle number at which the amplification curve of each sample intersected the threshold line were recorded for each sample.

Standardization and construction of real-time PCR calibration curves

A standard curve was run singly for each of the primer and probe sets using tenfold serially diluted B. bovis control DNA sample (48 ng/µl) isolated from an in vitro culture and a B. bigemina positive control DNA sample (17 ng/µl) isolated from blood of an infected calf at the Tick Fever Centre (TFC), Queensland Department of Agriculture, Australia, ranging from 10−1 to 10−8. The diluted DNA was then amplified using a multiplexed assay of three primer and probe sets (B. bovis 18S, B. bovis cytochrome b and B. bigemina cytochrome b; Additional file 1: Table S1). Resulting data were analysed using the QuantStudio design and analysis software version 2.6.0. To generate the calibration curves, the cycle quantification (Cq) scores for the diluted samples were determined based on a preset baseline threshold of 0.074 ΔRn for B. bovis 18S and 0.02 ΔRn for B. bovis cytochrome b and B. bigemina cytochrome b (ΔRn is defined as the Rn value of an experimental reaction minus the Rn value of the baseline signal). The Cq scores were then plotted against corresponding DNA dilutions. The efficiency (E) of the primer sets expressed as a percentage of each of the individual real-time PCR assays was calculated from the slope of the respective standard curve. The coefficient of correlation (R2) of each standard curve was also determined using the QuantStudio analysis software.

Babesia bovis spherical body protein-4 (SBP-4) nested PCR amplification and sequencing

A sub-set of 59 real-time PCR B. bovis-positive samples was subjected to standard PCR amplification and sequencing using specific primers targeting a 503 fragment of the highly conserved B. bovis spherical body protein-4 (SBP-4) in a nested PCR (nPCR) as previously described [30]. Primer sequences used in the nested PCR are listed in Additional file 1: Table S1. Initial PCR amplifications were done in a 25 μl-reaction mixture having 12.5 μl 2 × MyTaq™ Red Mix (Meridian Bioscience, USA) PCR buffer mix, 1 μl (10 pmol) of each primer, 3 μl DNA template and 7.5 μl nuclease free PCR water. The thermocycling conditions for the PCR amplifications were as follows: initial denaturation of 5 min at 95 °C followed by 35 cycles (1 min of denaturation at 94 °C, 1 min of annealing at 55 °C and 1 min of extension at 72 °C) and final extension at 72 °C in 10 min in a AllInOneCycler™ PCR system (Bioneer). A nested PCR was done using 2 μl of DNA template obtained from the primary PCR under the same amplification conditions. DNA control samples used in the real-time PCR were included as positive controls while nuclease free PCR water was used as a negative control. The nPCR products were purified and directly sequenced with both forward and reverse nested PCR primers at Macrogen Europe B.V (The Netherlands).

Data analysis

Real-time PCR data were exported to Ms Excel where the mean and range of Cq scores was calculated. A box plot of Cq values detected was generated using STATA 15. The prevalence of each Babesia was estimated as the proportion of total samples for each gene and prevalence rates between sub-counties compared using chi-square. The P values for statistical significance were set at 0.05. To estimate the confidence intervals around the prevalence estimate, the function prop.test of the stats package implemented in R (R Core Team) was used, setting the confidence level at 95% to test the true proportion of the sample genes. The 95% confidence interval was estimated using the formula:

$$\overline{x } \pm z\alpha /2 \frac{\sigma }{\sqrt{n}}$$

where \(\sigma\) is standard deviation, n is sample size, and \(\overline{x }\) is proportion of estimate.

Spherical Body Protein-4 (SBP-4) amplicons were obtained with 29 of 59 field isolates amplified. Sequence chromatograms were obtained from 24 of the 29 field samples and three control samples sequenced. The chromatograms were visually inspected and resulting sequences edited manually using CLC Main Workbench 20 software (CLC bio, Qiagen GmbH, Germany). Five of the sequenced field samples returned low quality sequences and were omitted from further analysis. Obtained nucleotide sequences were trimmed to remove low-quality reads at the 5′ and 3′ ends and consensus sequences generated from both the assembled forward and reverse fragments. Molecular identity of the isolates was confirmed via Basic Local Alignment Search Tool for nucleotides (BLASTn) [31] searches of the SBP-4 sequences against the GenBank’s non-redundant nucleotide sequence database. The consensus nucleotide sequence from each isolate was then processed for further multiple sequence alignments using CLC Main Workbench 20 software (CLC bio, Qiagen GmbH, Germany). Percent identity analyses were performed using Clustal Omega multiple sequence analyses tool [32].

Results

Evaluation of Babesia real-time PCR assays: statistical parameters, standard curves and efficiency

Initially, control gDNA was used to evaluate the utility of a previously published set of real-time PCR primers targeting B. bovis and B. bigemina [28, 29]. A summary of the statistical parameters, standard curves and efficiencies of the primer sets used in the multiplex real-time PCR is shown in Table 1 and Fig. 2a, b. The lowest Cq value detected by the B. bovis 18S primer pair at the lowest serial dilution of 10−1 was 23.3 while the cytochrome b detected a Cq of 17.7. The B. bigemina cytochrome b primers detected a Cq of 22.3 at the lowest dilution of 10−1. The B. bovis 18S primers returned a maximum Cq value of 44.7 at 10−7 dilution, which was the detection limit of the control DNA, while the cytochrome b returned a high of 35.7 at a dilution of 10−6, which was the detection limit of the control DNA used. The B. bigemina cytochrome b primers returned a maximum of 32.4 at a dilution of 10−4, which was also the detection limit of the control DNA used (Table 1).

Table 1 Summary details of the statistics and efficiency parameters of the three real-time PCR primer sets
Full size table
Fig. 2
figure 2

Real-time PCR amplification and calibration curves of the three primer and probe sets used in the detection of Babesia parasites in the current study. a Amplification plots. b Standard curves. The Babesia bovis 18S primer set had an efficiency of 98.96%, a slope of –3.34 and a correlation coefficient (R2) of 0.981, while the cytochrome b primers had an efficiency of 90.4%, a slope − 3.57 and a R2 of 0.999. The B. bigemina set had an efficiency of 99.51%, a slope of − 3.33 and a R2 of 1.0. The plots and standard curves were generated using QuantStudio design and analysis software version 2.6.0. The detection limit of the gene in the control DNA was 10−7 dilution for B. bovis 18S primers, 10−6 for cytochrome b and 10−4 for B. bigemina cytochrome b

Full size image

A standard curve was run singly for each primer and probe set. The B. bovis 18S primer set had an efficiency of 98.96%. Its standard curve had a slope of -3.34 and a correlation coefficient (R2) of 0.981. The Y-Intercept was 20.2. The B. bovis cytochrome b primers had an efficiency of 90.4%, a slope of − 3.57, a R2 of 0.999 and a Y-intercept of 14.0. The B. bigemina set had an efficiency of 99.51%, slope of − 3.33, R2 of 1.0 and Y-intercept of 18.97. The amplification plots and standard curves of each primer set as determined from the Cq values of the serially diluted standard solutions are shown in Fig. 2a, b, respectively. The detection limit in the control DNA was 10−7 dilution for 18S gene, 10−6 dilution for B. bovis cytochrome b and 10−4 dilution for B. bigemina cytochrome b. The B. bovis 18S and cytochrome b, which were labelled with different fluorescence dyes, were not detected in the B. bigemina control, nor was B. bigemina cytochrome b detected in the B. bovis control included in each PCR run. Thus, the primer sets employed were species-specific for B. bovis and B. bigemina as no cross-reaction was observed with the respective controls. In all the PCR reactions carried out, a no-template water sample was included as a negative control to rule out false positives due to cross-contamination.

Detection of B. bovis and B. bigemina DNA

The real-time PCR primer sets employed in this study allowed for species-specific detection of B. bovis and B. bigemina DNA present in the screened blood samples. Since the efficiency parameters of the primers were within the recommended limits of a good quantitative PCR (qPCR) efficiency of between 90 and 110%, a slope that falls between − 3.6 and − 3.3 and a coefficient of correlation (R2) of > 0.98 (Table 1; Fig. 2b), all samples that returned Cq values of < 45 for the B. bovis 18S primers with a detectable amplification plot above the baseline threshold magnitude of 0.072 ΔRn were regarded as positives since the corresponding Cq values obtained with the cytochrome b primers for the same samples were lower (Additional file 2: Table S2). For the B. bovis and B. bigemina cytochrome b primer sets, all samples that returned Cq values of < 42 with a detectable amplification plot showing an exponential growth of the PCR product above the signal threshold of 0.02 ΔRn were also regarded as positives. A total of 131/506 samples returned Cq values that met these criteria and were thus classified as positive for either B. bovis or B. bigemina, or both. Detectable amplifications with the B. bovis 18S primers were observed in 39 samples, while a total of 87 samples resulted in detectable amplicons using the B. bovis cytochrome b primers. A list of the 131 field samples and their corresponding Cq values is shown in Additional file 2: Table S2. The range distribution of the Cq values detected by the three assays is shown in Fig. 3. The highest Cq value obtained with the field samples using the B. bovis 18S was 44.7 and the lowest was 12.9. The B. bovis cytochrome b returned a high Cq value of 41.2 and a low of 29.9 with a mean Cq value of 35.4 (SD ± 2.997). The B. bigemina cytochrome b primers returned a maximum of 41.4 and a minimum of 21. No detectable amplification was observed in 375 of the screened DNA samples.

Fig. 3
figure 3

Box plot showing Cq score distribution range of the three real-time PCR target genes. The minimum Cq value was 12.9 for Babesia bovis 18S, 29.9 for B. bovis cytochrome b and 20.9 for B. bigemina cytochrome b. The median values were 37.5, 35.1 and 33.6 while maximum values were 44.7, 41.2 and 41.4 for B. bovis 18S, B. bovis cytochrome b and B. bigemina cytochrome b, respectively. The plot was generated using STATA 15

Full size image

Nucleotide BLAST analysis of 24 SBP-4 sequences obtained showed that they were homologous with B. bovis SBP-4. The percent sequence identities among isolates analysed in this study and top matching reference sequences from GenBank repository ranged between 97 and 100% (Table 2). They matched previously reported Kenyan isolates with a similarity of 96–100% [27]. Nine isolates shared 99–100% similarity with isolates from Benin reported by Moumouni et al. [33], while 13 isolates shared a similarity of 99.0–100% with South African isolates. Two isolates shared a 97.0% similarity with sequences from Indonesia reported by Guswanto et al. [34]. The 27 SBP-4 nucleotide sequences obtained were deposited in the GenBank database under the accession numbers ON012652-ON012678 (Table 2).

Table 2 Summary of BLASTn analysis hits showing the sequence accession numbers, top matching GenBank sequence, percent identity, animal host and country of origin for each of the 24 field and 3 control samples whose spherical body protein 4 gene was sequenced in the present study
Full size table

Prevalence of B. bovis and B. bigemina infection

A total of 506 cattle blood samples were screened for presence of both B. bovis and B. bigemina. The distribution of the number of positive cases across the sampled sub-counties is shown in Table 3. The overall bovine babesiosis prevalence rate was 25.9% (131/506). Of the 131 Babesia-positive animals, 87 were found to be infected with B. bovis, indicating a prevalence of 17.2% (95% CI; 14.1–20.8%) (87/506), while 70 were positive for B. bigemina, giving a prevalence of 13.8% (70/506) (95% CI; 11.0–17.2%), and 27 (5.3%) (95% CI; 3.6–7.8%) had mixed infections of both B. bovis and B. bigemina. The 18S B. bovis primers detected 39 positive cases (7.7%) (95% CI; 5.6–10.4%), while the B. bovis cytochrome b primers detected 81 (16%) (95% CI: 12.9–19.6%) positive cases (Table 3). A total of 61 animals (12.1%) were found to have single B. bovis infection while 44 animals (8.7%) were positive for B. bigemina only.

Table 3 Distribution of Babesia-positive blood across the study sites
Full size table

Babesia bovis infections were detected in all the three counties sampled. The highest B. bovis prevalence was observed in Lunga Lunga sub-county (52.9%; 46/87), while the lowest was observed in Msambweni (8%; 7/87). The highest prevalence of B. bigemina was also observed in Lunga Lunga (42.9%; 30/70), while the lowest was observed in Msambweni (18.6%; 13/70). Mixed infections were also highest in Lunga Lunga (59.3%; 16/27). The sampling site (sub-county) was statistically associated with prevalence of Babesia infection (χ2 = 72.4, df = 8, P < 0.0001).

Discussion

Bovine babesiosis is one of the four major tick-borne diseases that affect livestock in Kenya [21,22,23,24,25,26]. Until recently, bovine babesiosis in the country was presumed to be caused by B. bigemina, transmitted principally by R. decoloratus, which is endemic in many regions in Kenya and tropical Africa [19]. Recent recordings of R. microplus in Kwale [13] indicated an urgent need to investigate the prevalence of pathogens transmitted by this tick including B. bovis to update epidemiological maps of these diseases. Using specific molecular assays, this study has confirmed the presence of pathogenic B. bovis DNA in cattle blood from coastal Kenya, with the findings representing the first report of a substantial level of occurrence of the parasite in Kenyan cattle.

Screening of the 506 cattle samples with two B. bovis and one B. bigemina specific probes, we observed that B. bovis was present in 17.2% (87/506) of the animals, while 13.8% (70/506) of the samples were positive for B. bigemina. Co-infections with the two Babesia spp. were observed in 27 (5.1%) of the animals screened (Table 3). This is also the first major study to report occurrence of the pathogenic B. bovis in the East African region in relatively high numbers. The findings indicate that B. bovis infections do occur in significant numbers in this region but are probably disregarded in most studies screening for tickborne pathogens in cattle because of the belief that only the endemic B. bigemina is present. This calls for implementation of surveillance and control measures because the economic burden of ticks and tick-borne diseases in Africa [2] and globally is very high [1].

In this study, prevalence of Babesia infection was significantly associated with sub-county (χ2 = 72.4, df = 8, P < 0.0001). The highest number of B. bovis infections was observed in Lunga Lunga sub-county, which borders Tanzania. There is a livestock market in Tanga, a port city in Northeast Tanzania, and at the Kenya border post at Lunga Lunga where cattle from Tanzania are purchased by Kenyans from neighbouring counties. Moreover, there are two holding sites, one at Lunga Lunga and the second at Msambweni, both in Kwale, where animals are held pending movement to final destinations of neighbouring counties and beyond. The Lunga Lunga market is frequented by Kenyan traders from as far as Tana River County while some are from Somalia. The extensive movement and transborder trade of potentially tick-infested and infected cattle is regarded as the major driving factor responsible for introduction of Babesia-infected tick vectors into new areas [17]. A similar pattern of transboundary cattle trade contributing to the spread of R. microplus and possibly its associated pathogens has been reported in Central and West Africa where animal movement across national boundaries is common [11]. High prevalence of the pathogenic B. bovis found in this study suggests fatal bovine babesiosis may become a severe threat to cattle across eastern African countries where the tick vector has been confirmed to be present. Our current report alongside recent confirmation of the occurrence of R. microplus in Kenya provides a rationale to implement surveillance to monitor potential disease outbreaks related to B. bovis infections. Effective measures should also be instituted to control the spread of R. microplus to limit transmission of B. bovis babesiosis especially further inward where both were previously absent.

The B. bovis real-time PCR and SBP-4 nested PCR tests used currently have added both reliability and sensitivity to the detection of B. bovis infections in Kenya. Classically, active B. bovis infection is diagnosed by Giemsa staining of blood smears with positive cases showing two pairs of small pear-shaped bi-lobed parasites [17]. In contrast, the presence of B. bigemina in a blood smear may not necessarily indicate clinical babesiosis, as symptoms can be due to resurgence of a chronic infection. In Kenya, diagnosis of redwater is confirmed by observation of the characteristic red urine, whereas microscopic examination of blood smears is seldom performed. In national diagnostic and research laboratories, detection of circulating B. bigemina antibodies is used for disease surveillance but this has its own shortcomings [21, 22, 35]. Molecular PCR techniques have been used to detect and differentiate Babesia parasites with high sensitivity and specificity [36,37,38]. Based on PCR amplification and sequencing of the B. bovis spherical body protein-4 (SBP-4) gene, Moumouni et al. [27] previously reported a B. bovis prevalence of 12.3% and 23.7% in Kajiado and Machakos Counties in Kenya, respectively. Using the reverse line blot (RLB) assay targeting the V4 hypervariable region of Babesia rRNA, Njiiri et al. [26] reported a B. bovis prevalence of 2.2% in Busia County, Western Kenya. In this study we used a well-validated TaqMan probe multiplex assay based on B. bovis 18S [28], B. bovis cytochrome b and B. bigemina cytochrome b [29] that can detect these species and differentiate B. bovis from B. bigemina. High specificities and efficiencies of > 96% were observed with these primer and probe sets (Fig. 2), confirming the efficiency and usefulness of these assays in detecting and discriminating B. bovis infections. In our study, a total of 61 (12.1%) and 44 (8.7%) animals were found to have single B. bovis and B. bigemina infections, respectively, while 27 (5.1%) animals were found to have B. bovis and B. bigemina co-infections. The presence of co-infections indicates the co-occurrence of the two Boophilus spp., Rhipicephalus decoloratus, which transmits B. bigemina and is endemic in Kenya, and R. microplus, the known vector for B. bovis that also transmits B. bigemina. Further studies on tick infections, species densities and their distribution are required to define the contribution of each species to the epidemiology of redwater in the country. The difference in sensitivities between the B. bovis 18S and cytochrome b primers may be due to differences in abundance of the two genes. The B. bovis genome has been shown to contain three rRNA operons [39]. Although information on the number of mitochondria in Babesia parasites is lacking, some apicomplexan parasites such as Toxoplasma gondii and Plasmodium falciparum have been reported to have only a single mitochondrion per parasite [40]. The mitochondrial DNA of P. falciparum comprises approximately 20 copies of a 6-kbp linear genome per cell [41, 42] encoding three protein coding genes including cytochrome b [39]. Therefore, the quantities detected by different genes could reflect the gene copies per individual organism as well as level of parasitaemia, which was not determined in this study. More B. bovis-positive samples were detected with the cytochrome b primers compared to the 18S indicating the higher sensitivity obtained by using this gene compared to the nuclear-encoded 18S rRNA gene. Therefore, based on the data obtained in this study and theoretical predictions, the B. bovis cytochrome b primers would be the best target for a routine field diagnostic assay.

Animals that recover from babesiosis become carriers of babesia parasites for life and can develop the disease again if they undergo physiological stresses such as nutritional restriction or co-infection with another infection. Babesiosis is therefore a costly chronic animal disease in endemic areas because of frequent resurgence, especially if the animals are experiencing stress [2]. Factors such as animal breed, type of agro-ecological zone (AEZ) and livestock production systems are important risk factors associated with babesiosis [22]. Most farmers in Kwale County, which has an estimated 190,988 zebu cattle and 5475 dairy crosses [43], practice an open grazing system, which has previously been shown to be significantly associated with high prevalence of B. bigemina infections in Murang’a County in Kenya [22]. Kwale County, which is a gateway to mainland Kenya for livestock purchased from the Tanzanian border market, could be acting as a focal source of B. bovis infections to the rest of the country.

Ticks and tick-borne diseases are the biggest threat to sustainable cattle production in Kenya [2], which has close to 18 million cattle [44]. Although this study reports high prevalence of B. bovis in three sub-counties in Kwale, the situation in the rest of the country is unknown. The emergence of B. bovis infections in Kenya threatens the already fragile livestock sector, plagued by other tick-borne diseases such as East Coast fever, and the introduction of susceptible taurine cattle breeds [45]. Therefore, the demonstration of a high prevalence of B. bovis revealed its establishment and local transmission. It is recommended that molecular diagnostics including the real-time PCRs used in this study be added to routine surveillance to detect and differentiate the two Babesia parasites with the aim of elucidating the current status of bovine babesiosis in the country. Such efforts will underpin the deployment of strategic disease control strategies for Kenya and its neighbours.

Conclusion

In conclusion, we found a relatively high number of B. bovis and B. bigemina infections in asymptomatic cattle from Kwale County in Kenya with potential to cause disease outbreaks in susceptible animal populations. The Babesia multiplex real-time PCR used in this study is specific and can detect and differentiate the two Babesia parasites. The B. bovis cytochrome b primers would be the best target for a routine field diagnostic assay. Active surveillance to monitor the spread of B. bovis babesiosis across the country is recommended.