Parasite load and genotype are associated with clinical outcome of piroplasm-infected equines in Israel

Abstract

Background: Equine piroplasmosis is a highly endemic protozoan disease of horses worldwide, caused by Theileria equi and Babesia caballi. While most horses in endemic areas are subclinically infected, the mechanisms leading to clinical outcome are vastly unknown. Moreover, since clinical signs of disease are not specific, and the prevalence in endemic areas is high, it is difficult to determine if equine piroplasmosis is the cause of disease. To identify possible mechanisms leading to the clinical outcome in an endemic area, we compared parasite loads and genotypes in clinically and subclinically infected horses.

Methods: Blood was collected from horses with clinical signs consistent with equine piroplasmosis, and from apparently healthy horses in Israel. Packed cell volume and total solids were measured. Quantitative and diagnostic polymerase chain reaction were used to identify, quantify and classify equine piroplasmosis infection. Phylogenetic analyses were used to determine the genotype of both parasites.

Results: For both parasites, clinical cases were associated with low mean packed cell volume and high mean parasite load (P < 0.001), enabling the determination of a cut-off value to distinguish between clinically and subclinically infected horses. Samples of Theileria equi from subclinical horses were classified into three different 18S rRNA genotypes, D (n = 23), A (n = 12) and C (n = 5), while samples from all clinical cases (n = 6) were classified as genotype A. The sequences of T. equi equi merozoite antigens 1 (ema-1, n = 9) and 2 (ema-2, n = 11) genes were fairly conserved and did not differ between clinical and subclinical cases. Babesia caballi rhoptry associated protein-1 (rap-1) was classified into sub-genotypes A1 (n = 14) and A2 (n = 5) with no association to clinical outcome. Classification of the 18S rRNA gene (sub-genotypes B1 and B2) agreed with the rap-1 classification.

Conclusions: The results of this study suggest that quantification of parasite loads of infected horses may be used to distinguish between infections resulting in disease and subclinical cases. Although number of clinical cases is limited, we identified T. equi 18S rRNA genotype A to be associated with clinical disease. This finding emphasizes the importance of in-depth genetic characterization of T. equi genotypes to identify possible markers for virulence.

Keywords: Babesia caballi; Clinical signs; Equine piroplasmosis; Parasitemia; Phylogeny; Theileria equi.

Fig. 1

Theileria equi (a) and B. caballi (b) parasite loads in clinical (red) and subclinical (blue) horses, as determined by qPCR. For each parasite a standard curve (black line) was created using serial dilutions of a clean PCR product of each gene (gray marks). Parasite gene copy number from field samples was calculated from the quantification cycle (Cq) and the standard curve. A diagnostic cut-off distinguishing between clinical and subclinical cases of each parasite was determined by ROC analysis and the cut-off value is marked by a vertical dashed line

Fig. 2

Phylogenetic analysis of T. equi sequences obtained from clinically infected horses (triangles) and subclinically infected horses (open circles) using three genes (sample names as detailed in Additional file 1: Table S1). a Analysis of 1079 nucleotide positions of T. equi 18S rRNA gene, from 6 clinical and 40 subclinical samples, along with additional published sequences (GenBank ID/parasite/host/location). The phylogenetic tree was constructed based on the Tamura-Nei model with gamma distribution (+G). b Analysis of 400 nucleotide positions of T. equi ema-1 gene sequences obtained from four clinical and five subclinical samples, along with additional published sequences (GenBank ID/parasite/host/location). The classification of each sample according to its 18S rRNA gene is states near the sample name (− 18SX). The phylogenetic tree was constructed based on the Kimura 2-parameter model with consideration on invariable sites (+ I). c Analysis of 782 nucleotide positions of T. equi ema-2 gene sequences obtained from six clinical and four subclinical samples, along with all 19 additional published sequences (GenBank ID/parasite/host/location). The classification of each sample according to its 18S rRNA gene is states near the sample name (− 18SX). The phylogenetic tree was constructed by based on the Kimura 2-parameter model with consideration on invariable sites (+ I). All phylogenetic trees were constructed using maximum likelihood method and 1000 bootstrap replicates. The percentage of trees in which the associated samples clustered together is shown next to the branches when it was above 70%. The analysis was constructed in MEGA7

Fig. 3

Phylogenetic analysis of B. caballi isolated from clinically infected horses (diamonds) and subclinically infected horses (open squares) (sample names as detailed in Additional file 1: Table S1). a Analysis of 1212 nucleotide positions of B. caballi 18S rRNA gene sequences from six clinical samples along with additional published sequences (GenBank ID/parasite/host/location). The phylogenetic tree was constructed based on the Tamura-Nei model with gamma distribution (+ G) and invariable sites (+ I). The percentage of trees in which the associated samples clustered together is shown next to the branches when it was above 70%. The analysis was constructed in MEGA7. b Analysis of 251 nucleotide positions of B. caballi rap-1 gene sequences from six clinical and 13 subclinical samples, along with additional sequences from GenBank. The phylogenetic tree was based on the Kimura 2-parameter model with consideration on invariable sites (+ I). Both phylogenetic trees were constructed using maximum likelihood method and 1000 bootstrap replicates. The percentage of trees in which the associated samples clustered together is shown next to the branches when it was above 70%. The analysis was constructed in MEGA7