Background

Equine piroplasmosis (EP), an infectious tick-borne disease of the Equidae, is caused by two intra-erythrocytic hemoprotozoans, Theileria equi (formerly Babesia equi) and Babesia caballi, in most tropical and subtropical areas, and some temperate zones of the world [1, 2]. The disease can cause significant economic losses in the equine industry globally. Theileria equi is considered a more virulent species than B. caballi, [3–5]. However, they share the same competent tick vectors, including Hyalomma excavatum [6] and mixed T. equi and B. caballi infections are common in endemic areas [7]. Both B. caballi and T. equi are responsible for hemolytic disease which is characterized by fever, anemia, red urine, jaundice, edema, drop in body mass and even death [8, 9]. Clinical signs of equine piroplasmosis are often non-specific, precluding accurate diagnosis [10]. Currently, specific diagnosis can be obtained by microscopic examination, indirect immunofluorescent antibody test (IFAT), enzyme-linked immunoassays (ELISA) and polymerase chain reaction (PCR). Sensitive and specific tests for EP are required for disease control and preventing introduction of parasites into countries that are regarded free of infection/disease.

Two competitive-ELISAs (cELISA) were developed for the detection of antibody to T. equi and B. caballi. The T. equi cELISA uses monoclonal antibody (mAb 36/133.97) to T. equi merozoite antigen (EMA)-1, while the B. caballi-specific cELISA is based on a monoclonal antibody to Bc48 (79/17.18.5), reactive with a member of the Babesia sp. rhoptry-associated protein (RAP)-1 family. Recent data suggests that the RAP-1 based cELISA lacks sensitivity for the detection of B. caballi - infected equids in South Africa and Israel, and the loss of sensitivity is hypothesized to be due to rap-1 sequence variations among strains [11, 12].

PCR-based detection of parasites has been reported to have higher sensitivity and specificity compared with serological assays, but these tests are not routinely performed as a diagnostic tool [13–18]. Detection of parasite DNA by PCR is also useful for the early detection of acute phase of infection, when antibodies are not yet detectable. Yet, serological and molecular methods are more reliable than microscopy for detecting persistent infections with low levels of parasitemia. For example, the ability of nested PCR (nPCR) to diagnose sub-clinical infections could help prevent exportation of infected animals as well as in determining the efficiency of pharmacological treatments [18].

Previous data collected in Egypt, suggest high prevalence of EP caused by T. equi, which is consistent with the current lack of control measures [19, 20]. Furthermore, no studies on infection of equids with B. caballi were previously reported, and surveys using sensitive and specific state-of-the-art diagnostic techniques are required to assess the overall impact of EP on the equine populations in Egypt. Importantly, the possible association between infection with the parasites causing EP, and the occurrence of hematological disorders in equine populations in Egypt also remains an important knowledge gap [20].

Using state-of-the-art diagnostic methods, this study provides an epidemiological snapshot of infection of B. caballi and T. equi infections in horses and donkeys in selected locations in Egypt, explores the association of hematological parameters with infection status and provides rationale for the selection of diagnostic tests depending on need. Detection of infection was correlated with occurrence of hematological changes suggesting possible long term health impact on infected equids.

Methods

Animals and ethical approval

A total of 88 horses from Cairo and 51 donkeys from the Giza zoological garden were investigated for infection with T. equi and/or B. caballi. The protocol used in these animal studies and all animal handling was approved by the Egyptian Medical Research Ethics Committee of NRC (No. 14–126).

Blood specimens, microscopic detection, and immunofluorescent assay

Whole blood was collected from animals by jugular vein puncture using Vacutainer tubes® with or without anti-coagulant. Blood with EDTA were used for preparation of blood films, for hematological evaluation and for DNA extractions using FTA® Elute cards (Whatman Cat. No. WB120410). Blood smears were stained with Giemsa stain for 15 min and examined by analyzing twenty five fields in each thin blood film at a 1000× magnification [21]. The indirect immunofluorescent slides of T. equi and B. caballi, as well as the control equine sera were kindly donated by VMRD Inc. (Pullman, WA, USA). The IFAT was performed using equine sera at a 1:50 dilution.

Competitive-inhibition ELISA

The cELISA T. equi and B. caballi test kits used in this study were kindly provided by VMRD Inc. (Pullman, WA, USA). Assays were performed following the manufacturer’s instructions. Optical density (O.D.) values were determined at 650 nm using ELX 800 universal microplate reader Bio-TEK instruments, INC, USA. The results were expressed as a value of the percent inhibition (%I) according to the following formula:

$$ \left(\%I\right):\ \%I = 100\mathit{\hbox{-}}\left[\left( sample\ O.D.\times 100\right)\ /\left( mean\ negative\ control\ O.D.\right)\right]. $$

The manufacturer of the cELISA kit established the %I value of 40 as the cut-off [1, 22]. Thus samples with %I above 40 are considered as positive, and below 40, considered as negative.

Nested PCR

Total genomic DNA (gDNA) was extracted from whole blood using FTA® Elute cards following the manufacturer’s instructions. The PCRs were performed using 5–10 μl eluted DNA in a final volume of 25 μl containing 12.5 μl JumpStart RED Taq Ready Mix PCR reaction mix (Sigma-Aldrich), and 10 pmol of each primer. The oligonucleotide primer sequences are shown in Table 1. The thermocycling conditions used for either T. equi and B. caballi amplification were: 95 °C for 3 min, followed by 25 cycles, consisting of denaturation at 95 °C for 15 s in external reaction and 5 s in nested reaction, annealing at 60 °C for 15 s in external reaction and 5 s in nested reaction and extension at 72 °C for 15 s in external reaction and 5 s in nested reaction. A final extension cycle at 72 °C for 5 min was performed and reactions were cooled to 15 °C.

Table 1 Oligonucleotide primer pairs used in PCR amplifications for the detection of Theileria equi and Babesia caballi in equines
Full size table

Theileria equi and B. caballi gDNA derived from cultured Florida T. equi and Puerto Rico B. caballi strains, respectively, used as positive controls were obtained from the OIE Equine piroplasmosis reference laboratory, Animal Disease Research Unit, Agricultural Research Service, USDA, WSU, Pullman, WA. A negative control with no DNA template was included for PCR amplifications. The expected specific amplification products of T. equi and B. caballi were detected at 229 and 224 bp, respectively. Amplicons were purified for sequencing using the QIA quick Spin PCR Purification kit (Qiagen, Courtaboeuf, France). Sequencing of the PCR products was conducted by the GATC Company using an ABI 3730xl DNA sequencer. Each sequencing reaction was repeated three times in forward and reverse directions before being accepted for this study.

Analysis of rap-1 genes amplified from B. caballi - infected horses

Total gDNA was extracted of blood from infected horses as described above and PCR-amplified using primer set FS Rap-1 Forward: (5′-ATG AGG TGT TCT GCG AGT T-3′) and FS rap-1 Reverse (5′- GAT GGA CGT CAA AGG TG-3′) to amplify the full size rap-1 gene. The thermocycling conditions used were: 95 °C for 3 min, followed by 35 cycles, consisting of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s and extension at 72 °C for 2 min. A final extension cycle at 72 °C for 7 min was performed and reactions were cooled to 15 °C. Three PCR amplicons were cloned into pCR®2.1-TOPO® (Life Technologies) and fully sequenced.

Hematological investigations

A set of 48 blood samples were selected for hematological analysis according to the following criteria: (i) a control group of 12 samples from horses that are negative for infection to B. caballi and/or T. equi; (ii) a positive group of 12 samples from horses that were positive to blood film examination; (iii) and a positive control group with nine samples derived from horses which demonstrated positive cELISA reaction; (iv) nine samples derived from horses with nPCR positive results; and (v) a group of six samples derived from horses that were positive in all tests. All samples were evaluated using an automatic counter (MEDONIC CA620, Sweden). The parameters evaluated in the hemogram included erythrogram: red blood cell count (RBCs), hematocrit (HCT), hemoglobin (Hb) concentration and red cell indices; mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), red blood cell distribution width absolute (RDWA), red cell distribution width percentage (RDW%), Leukogram: white blood cell count (WBCs), lymphocytes, granulocytes and mid cells (MID) and platelet count (PLT), mean platelet volume (MPV), platelet distribution width (PDW), plateletcrit (PCT), large platelet concentration ratio (LPCR).

Statistical analysis

Chi-square (χ2) test was applied at probability of P < 0.05 to compare infection rates between equids, and to estimate the relative efficacy of the diagnostic methods for the detection of infected animals. Hematological data were subjected to statistical analysis including calculation of mean and standard error. Differences between groups were tested for significance using a one-way analysis of variance (ANOVA) followed by Duncan’s multiple range tests. Differences were considered significant at P < 0.05 level [23] using the SPSS version 14 software.

Results

Microscopic, serological, and molecular analysis of equine piroplasmosis

Direct microscopic examination of 139 equids samples detected 19 (13.6 %) positive samples [10 (11.4 %) positive horses and 9 (17.8 %) positive donkeys] (Table 2).

Table 2 Results of infected T. equi and/or B. caballi horses, donkeys and total equines detected by cELISA, nPCR, IFAT and blood film
Full size table

Testing of sera from the 139 equines by IFAT detected 21 (23.9 %) and 15 (17.0 %) out of 88 horses positive for T. equi and B. caballi respectively, and 16 (31.4 %) out of 51 donkeys tested positive for T. equi and 16 (31.4 %) for B. caballi (Tables 2 and 3). Therefore, 26.6 % of the combined equine serum samples tested contained antibodies reactive with T. equi, whereas 22.3 % of the equine serum samples contained antibodies reactive with B. caballi parasites as tested by IFAT (Table 2). Statistical analysis of these data using a χ2 test results shows no significant differences in the rate of infections with piroplasms between donkeys and horses using IFAT (Table 3).

Table 3 Results of cELISA, nPCR, IFAT and blood film for detection of T. equi and/or B. caballi in horses and donkeys
Full size table

Collectively, of the 88 horses and 51 donkeys tested by cELISA, 13 (14.8 %) and 12 (23.5 %) were positive for T. equi, respectively. Statistical analysis of these data using a χ2 test result showed, again, no significant differences in the rate of T. equi infections between donkeys and horses (Table 3).

Unexpectedly, and despite the identification of antibodies against B. caballi in 31 equine serum samples using IFAT (Tables 2 and 3), the B. caballi cELISA test failed to provide concordance for any positive samples when performed on all equid samples, despite the correct performance of the positive and negative controls provided by the cELISA manufacturer.

Nested-PCRs results to detect T. equi and B. caballi performed in all 139 equids sampled (88 horses and 51 donkeys) are shown in Tables 2 and 3. The prevalence of T. equi and B. caballi in equids were 38.8 % and 18 %, respectively. However, χ2 testing (Table 3) demonstrated no significant differences among B. caballi and T. equi infections in horses and donkeys using the nPCR data. Co-infection with both T. equi and B. caballi was found by nPCR in fifteen equids, 2 horses (2.3 %) and 13 donkeys (25 %).

Sequence analysis of two 229 bp nPCR amplicons, using T. equi internal specific primers (Table 1), demonstrated 100 % identity to the B. equi E12 ema-1 reference sequence (GenBank Accession number AF261824). In addition, sequencing of two 224 bp nPCR amplicon products of B. caballi obtained from Egyptian horses demonstrated identical sequences between them (GenBank accession number KJ094947). Comparison of the 224 bp amplicon and the predicted resulting amino acid sequences of the Bc48 Egyptian genotype sequences (87-1 rap-1) with the B. caballi clone X6 rap-1 reference sequence (GenBank Accession number AF092736), demonstrated three synonymous substitutions in this rap-1 gene region of Egyptian genotypes (Fig. 1a, b).

Fig. 1
figure 1

Alignments of the DNA a and predicted amino acid sequences b among Egyptian isolate termed 87-1 rap-1 (GenBank accession no. KR811096) and its equivalent region in the BC48 rap-1 reference gene sequence (GenBank accession no. AF092736)

Full size image

Comparative diagnostic performance of the cELISA, IFAT and nPCR tests used for surveillance of equine piroplasmosis

The diagnostic performances of the three methods used in the study were then statistically compared using a χ2 test (Table 4). The results confirmed that the nPCR technique detected significantly higher T. equi infection rates compared to cELISA in horses (χ 2 = 10.65, df = 1, P < 0.001). However, no differences among the nPCR and IFAT tests were found for the detection of equine piroplasmosis, and no significant differences were found among the cELISA and IFAT tests for the detection of T. equi infections.

Table 4 Representation of statistical analysis comparing number of infected animals for T. equi and B. caballi by cELISA, IFAT and nPCR
Full size table

Sequence analysis of rap-1 genes amplified by PCR in Egyptian B. caballi - infected equines

Sequence analysis of the nPCR amplicons and IFAT demonstrated the presence of B. caballi - infected equines in Egypt. While none of the horse or donkey samples tested in this study were demonstrated to be positive in the B. caballi RAP-1 based cELISA, comparisons among rap-1 Egyptian sequences with the reference rap-1 sequence shows high nucleotide conservation in a short portion of the full B. caballi rap-1 gene. The failure of the B. caballi RAP-1 based cELISA assay (shown in Additional file 1: Table S1 and Additional file 2: Table S2) prompted further investigation of the full size rap-1 gene sequences present in the Egyptian B. caballi isolates. nPCR using a set of primers designed to amplify the full size B. caballi rap-1gene (1943 bp) was performed on gDNA extracted from blood of three Egyptian B. caballi infected horses. Three distinct types of PCR amplicons termed “2-5 rap-1”, “1-1 rap-1”, and “87-3 rap-1” were initially generated using this approach. The in silico translation of the 87-3 rap-1 gene (GenBank Accession number KR811096), demonstrates a truncated open reading frame (ORF) (Additional file 2: Table S2 and Additional file 3: Figure S1). In contrast, alignments of predicted amino acid sequences derived from sequences of the 2-5 rap-1 and 1-1 rap-1 amplicons (GenBank accession number KR811097 and KR811095 respectively) with the reference B. caballi RAP-1 sequence (GenBank accession number AF092736) demonstrated uninterrupted open reading frames in both genes encoding for a predicted 55 kDa rap-1 protein, containing a predicted signal peptide. (Fig. 2). A matrix representing the percentage of amino acid identities among these putative RAP-1 proteins is shown in Additional file 1: Table S1.

Fig. 2
figure 2

Alignments of the predicted amino acid sequences among the PCR amplicon derived from the reference gene sequence (GenBank accession no. AF092736), and two predicted RAP-1 amino acid sequences, termed 2-5 RAP-1 (GenBank accession number KR811097) and 1-1 RAP-1 (GenBank accession number KR811095), derived from Egyptian isolates of B. caballi

Full size image

Hematological findings

Results from the hematological analysis are shown in Table 5. Statistical analyses of the hematological results using one-way analysis of variance followed by Duncan’s multiple range tests and the χ2 revealed that the number of erythrocytes were markedly decreased  (ANOVA: F (4.488), P = 0.004), while MCV values were significantly increased (ANOVA: F (4.578), P = 0.004) in all EP positive equids, as detected by different techniques [blood film, cELISA, and nPCR testing or by all of them (blood film + ELISA + nPCR)] compared to the non-infected control group (12 horses that were negative for EP in blood film examination, cELISA, and nPCR). The Hb concentrations were significantly decreased (ANOVA: F (2.813), P = 0.037) in infected groups detected by blood film, cELISA and nPCR techniques. However, there was a non-significant difference in HCT, MCHC and RDWA values in all EP positive equids compared to non-infected animals. This result indicated that EP potentially caused macrocytic hypochromic anemia in persistently infected horses (Table 5).

Table 5 Hematological profile in non-infected and infected horses with T. equi and/or B. caballi
Full size table

Statistical analysis performed on the leukogram data using the same statistical tools referred above showed that there was a marked (ANOVA: F (2.664), P = 0.045) decrease in granulocyte counts (neutrophil + eosinophil + basophil) in EP positive horses. No significant difference occurred in WBCs, lymphocytes and MID (monocytes) counts in all EP positive equids in comparison with the non-infected control ones (Table 5).

Regarding the platelets; the PLT and PCT values revealed significant decrease (ANOVA: F (9.743), P < 0.0001) and (ANOVA: F (8.692), P < 0.0001), respectively, in all EP-positive equids compared to non-infected animals. Moreover, the MPV, PDW and LPCR values markedly decreased (ANOVA: F (4.441), P = 0.004; F (3.908), P = 0.009; and F (4.835), P = 0.003) in piroplasma-infected animals detected by blood film and cELISA techniques compared to the other groups (Table 5).

Discussion

Designing robust definitive studies on the incidence of EP in Egypt and globally, requires methods of proven efficacy. A goal of this study was to compare currently available EP diagnostic methods and to define their suitability for more definitive surveys in the country. Thus, the presence of T. equi and B. caballi infections in horses and donkeys in Egypt was evaluated using blood film, IFAT, nPCR, and cELISA methods [13, 24]. In addition, and to our knowledge, this study is the first report of the presence of B. caballi infections of equids in Egypt.

Microscopic examination of blood smears in the current study suggested lower EP incidence compared to what was reported before by Salib et al. [20] who found an incidence of 34 % for T. equi in horses using a similar microscopic examination method, and Farah et al. who reported incidences of 38.8 % for different horse populations in Egypt [19]. Comparison of microscopic examination data with the other techniques used in this study suggests that this method, although convenient due to its simplicity, likely, due to limited sensitivity, grossly underestimates the incidence of infection with parasites causing equine piroplasmosis.

In the present study, cELISA detected lesser number of T. equi infected animals than nPCR; however since only a small subset of nPCR products were sequenced, the accuracy of this comparison is unknown. However, based on the sequencing of the nPCR products and the consistent size of the amplicons produced in the reactions, it is reasonable to assume that nPCR would detect acute infections prior to serological assays. Consistently, the data in Table 4 shows statistically significant differences among the nPCR and cELISA for detection of T. equi infections in horses. These data are not unexpected since serological detection of infection with T. equi by cELISA requires four to six weeks post-infection [25]. The limited correlation found among these methods may be due to the fact that these two tests detect different entities (DNA vs antibodies) and thus differ in principle. Hence, while PCR can be considered reliable for diagnosis of active infection, serological tests are usually considered as the method of choice for detecting persistently infected animals [22, 24, 26].

The RAP-1 based cELISA demonstrated sensitivity for detecting antibodies against Venezuelan isolates of B. caballi [13] and for the detection of persistent-chronic infections [26]. Remarkably, the B. caballi cELISA test failed to detect any positives using equids sampled from Egypt. The cELISA data is not concordant with detection of B. caballi infection by IFAT and nPCR, since several horses and donkeys tested positive for B. caballi infection by these methods. Interestingly, a similar finding was previously reported in a study performed in South Africa, where no positive samples were identified using the same cELISA, and was attributed to extensive polymorphisms in the RAP-1 region containing the B-cell epitope defined by the mAb 79/17.18.5 used in the test [11]. In addition, a similar scenario was recently reported while analyzing the serological status to B. caballi in horses in Israel [12]. Sequence analysis on nPCR partial rap-1 amplicons also demonstrated the presence of the Bc48/rap-1 gene in Egyptian B. caballi strains. The portion of the Bc48 gene amplified and sequenced in this study using nPCR primers has a predicted amino acid sequence that is almost identical to the Bc48/rap-1 gene of the Florida strain, used in the development of the mAb included in the B. caballi cELISA commercial kit used in this study [22]. Overall, the amplification of distinct full-sized rap-1 sequences in each of the samples analyzed in this study suggests the presence of a diverse repertoire of rap-1 genes in the Egyptian B. caballi isolates. Yet, while the South African and Israel B. caballi strains demonstrated dramatic rap-1 polymorphisms including indel mutations [11, 12] when compared to the reference strain, the Egyptian strains showed a relatively limited degree of polymorphisms. However, whether the rap-1 genes identified in this study are expressed during equid infections by the parasites, and if the observed sequence changes affect the immunogenicity of RAP-1 and/or the ability of the mAb 79/17.18.5 to bind the RAP-1 molecules from the Egyptian B. caballi isolates, remains unknown, and is beyond the scope of this study. Regardless of the reasons justifying the inability of the current B. caballi cELISA to detect infected equines in Egypt, these data suggest the need to improve serological diagnosis of B. caballi infection using alternative antigens and assays.

In this study, the most frequent blood alterations caused by EP include a reduction of the number of RBCs, PLT, HCT levels, Hb, MPV, PDW and LPCR, with increased values of MCV. While it cannot be disregarded that the blood alterations found in the samples analyzed may be due to other factors, collectively, these associations suggest that the piroplasma-infected horses suffered from macrocytic hypochromic anemia. Similar finding was recorded by Zobba et al. [27] in equine experimentally infected with piroplasms. Three mechanisms of hemolysis have been described: mechanical by trophozoite intra-erythrocyte binary fission [28], immune-mediated by auto-antibodies directed against components of the membranes of infected and uninfected erythrocytes, and toxicity by hemolytic factors produced by the parasite [29, 30]. There was regenerative anemia in infected horses associated with marked increase in RDWA due to anisocytosis and thrombocytopenia found in horses infected by T. equi and/or B. caballi has been rarely described in experimental and natural infections [31–33]. Although thrombocytopenia may be rarely reported in T. equi and/or B. caballi - infected horses, this is a very characteristic feature of canine babesiosis [31, 34]. The mechanism of thrombocytopenia in EP may be caused by local and systemic disseminated intravascular coagulation, immune-mediated destruction, and sequestration of platelets in the spleen [35]. Significant differences in total leukocyte counts and its differential between all groups of horses were found. Interestingly, taking together the data suggest numerous disturbances of the hemogram in chronically affected equines in the absence of acute signs of infections.

In summary, several currently used diagnostic tests were used in order to assess the presence of EP in a diverse equine population in Egypt. The overall incidence of infection with T. equi and/or B. caballi found in the population under study was relatively high regardless of the diagnostic method used (56.8 % using nPCR and 48.9 % using IFAT), and is in general agreement with previous assessments. Performance comparison of the IFAT and PCR tests used in this study shows no significant differences in their ability to detect infection of the causative agents of EP. Lack of concordance between the B. caballi cELISA, IFAT and nPCR suggests that a new diagnostic test based on additional antigen(s) will be needed. EP may be a contributing factor to hematological abnormalities of horses in Egypt.

Conclusions

Data presented here support the use of IFAT, nPCR and cELISA for the detection of T. equi to assure identification of acute and chronic infection. Importantly, these assays represent key tools to design and implement more effective control measures. Absence of reactivity using serum samples from Egyptian equines in the B. caballi cELISA test indicates that improvements in the serological detection of B. caballi are needed. It is likely that EP is a cause for hematological abnormalities in horses in Egypt.