Dynamics and within-host interaction of Theileria lestoquardi and T. ovis among naive sheep in Oman

Mixed species infections of Theileria spp. are common in nature. Experimental and epidemiological data suggest that mixed species infections elicit cross-immunity that can modulate pathogenicity and disease burden at the population level. The present study examined within-host interactions, over a period of 13 months during natural infections with two Theileria spp., pathogenic (T. lestoquardi) and non-pathogenic (T. ovis), amongst a cohort of naive sheep in Oman. In the first two months after exposure to infection, a high rate of mortality was seen among sheep infected with T. lestoquardi alone. However, subsequently mixed-infections of T. lestoquardi and T. ovis prevailed, and no further death occurred. The overall densities of both parasite species were significantly higher as single infection vs mixed infection and the higher relative density of pathogenic T. lestoquardi indicated a competitive advantage over T. ovis in mixed infection. The density of both species fluctuated significantly over time, with no difference in density between the very hot (May to August) and warm season (September to April). A high degree of genotype multiplicity was seen among T. lestoquardi infections, which increased with rising parasite density. Our results illustrate a potential competitive interaction between the two ovine Theileria spp., and a substantial reduction in the risk of mortality in mixed parasite infections, indicating that T. ovis confers heterologous protection against lethal T. lestoquardi infection.


Scientific Reports
| (2020) 10:19802 | https://doi.org/10.1038/s41598-020-76844-2 www.nature.com/scientificreports/ immunity to one pathogen reducing susceptibility to a second 9 . Consequently, the density and disease outcomes of virulent parasite strains may be less in mixed infections than in a single infections 10,11 . Studies in rodent malaria parasites have demonstrated that, in immune competent mice the avirulent Plasmodium strain suffered more competition than a more virulent strain, demonstrating a competitive advantage of virulent parasites in an immune-mediated interaction 11 . Similarly, an experimental study on T. annulata, showed higher density of the virulent clone in a mixed compared to a single infection 12 . This study reports the first detailed longitudinal investigation of the dynamics (within-host interactions) of coinfection of two Theileria species, the pathogenic parasite, T. lestoquardi, the causative agent of MOT and the non-pathogenic, T. ovis, in a naïve cohort of sheep in Oman. Our previous studies in Oman revealed a high prevalence of Theileria spp. among livestock 13 . MOT in sheep is associated with high morbidity of 30-40% and mortality can reach up to 100% among clinical cases of indigenous breeds during seasonal epidemics 14 . However, a large proportion of sheep carry asymptomatic mixed species infections of T. lestoquardi and T. ovis 7 . The data suggest the occurrence of interactions between the pathogenic and less pathogenic species of Theileria and supports the hypothesis that concurrent co-infection can lead to a reduction in parasitaemia and mortality associated with the pathogenic species 9 . This premise was tested further in the current study.

Results
Efficiency and sensitivity of qPCR for quantification of T. lestoquardi and T. ovis. T. lestoquardi and T. ovis-specific qPCR assays were developed to estimate parasite density and investigate within-host interaction between the two ovine Theileria spp. endemic in Oman. Initially, assay indices considered indicative of a well-optimized qPCR were assessed. This included the linearity of data (R 2 > 0.98), an efficiency (E) value within the range of 80-100% and consistency of Cq values across replicates. The qPCR amplification efficiency was 91% and 82% for T. lestoquardi and T. ovis 18s rRNA genes, respectively. The inter-assay variability between standard curves was 2% and 1% for the T. lestoquardi 18s rRNA and T. ovis 18s rRNA assays, respectively, while the correlation between log 10 18s rRNA copies and Cq values was significant for both species (T. lestoquardi adjusted R 2 > 0.99 for all PCRs with P < 0.001; T. ovis adjusted R 2 > 0.98 for all PCRs with P < 0.001). A high consistency of Cq values across replicates was seen, the standard deviation 15 ranged between 0.001-0.3 and 0.0003-0.3 for T. lestoquardi and T. ovis 18s rRNA qPCR, respectively. The limit of detection 3 was 9.26 log 10 and 5.3 log 10 T. lestoquardi and T. ovis 18s rRNA copies/μl blood, respectively (Fig. 1).

Analysis of T. lestoquardi and T. ovis infected sheep.
Fifty sheep (North Oman breed) were transferred to an endemic farm in April 2016. There was minimal difference in the mean age of the examined sheep (31-32 weeks) and all sheep were kept exclusively indoors, in a small open enclosure attached to the farmer's residence, with similar expected levels of exposure to ticks, Hyalomma anatolicum. At the early stages of the study, May and June 2016, 3 and 4 sheep died of suspected MOT, respectively. Theileria species identification and density were determined in blood samples collected from 43 sheep between May 2016 and May 2017 (13 times points per sheep). No parasites were seen when examined by Giemsa stained blood film and microscopy for the majority of the samples. However, occasionally animals showed parasites detected by microscopy during the follow up period. Interestingly, a very high parasite density (piroplasm stage) was seen in one dead animal examined few hours prior to death. Therefore, Theileria ssp were detected and infection levels assessed primarily by qPCR.
Infection was defined as single species, when the initial infection in May 2016 was diagnosed with either T. lestoquardi or T. ovis alone and continued in consecutive months till a mixed species infection (T. lestoquardi plus T. ovis) was seen. As the outcome of interaction between the two species can be influenced by the immune response 9 , the appearance of a single species following the detection of mixed species was not defined as single infection.
Theileria spp. could be detected in 485 (91%) of the 533 (95%) blood samples, collected during the study period; T. lestoquardi alone was detected in 125 (23%) and T. ovis alone was detected in 4 (1%) of the samples. Thus, a mixture of both species was detected in the majority of blood samples. The lower prevalence of single T. ovis differs from an earlier survey in Oman, (April and August 2014) when single T. ovis infection was more dominant than single T. lestoquardi 6 . The lower prevalence of single T. ovis among the present cohort is unlikely to be due to technical deficiency generating false negative results, as the limit of detection of the T. ovis specific qPCR assay was two-fold higher than that of T. lestoquardi.  Table 1). Disease symptoms (enlarged superficial lymph nodes, high fever and anorexia) consistent with theileriosis were observed in the first two months (May and June 2016). During this period, there were cases of mortality (7 out of 50 animals) and a high rate of single T. lestoquardi infection (50.8% of cohort). No other pathogens or clinical signs associated with ovine diseases known to occur in the study region were recorded. Parasite infection was detected by PCR and examination of the 7 dead showed T. lestoquardi schizonts and piroplasms in a few animals. Investigation of the 7 dead animals revealed enlarged super-  www.nature.com/scientificreports/ ficial lymph nodes, while necropsy of one animal also showed an enlarged liver and spleen. Potential increase in overt clinical signs assessed in surviving single infected animals may have occurred earlier in the infection period, prior to death, between the two time points, but were not taken. All but one of the animals that died, were infected by T. lestoquardi alone, and mortality was significantly associated with single T. lestoquardi infection (Fisher's exact test, P < 0.001).

Pattern of Theileria species infection and mortality.
The parasite population seen in each of the dead animals, prior to death, was comprised of a different, distinct combination of multiple genotypes, as identified by the pattern of alleles generated by five neutral microsatellite markers (Fig. 2, Supplementary Table 1). Thus a specific virulent population or over representation of certain genotypes could not be associated with death in these animals.

Dynamics and interaction between T. lestoquardi and T. ovis: parasite infection.
Following the acute phase of MOT, 43 (86%) sheep maintained chronic asymptomatic Theileria spp. infection throughout the 13 months of the study period, most often as a mixed infection detectable by qPCR. Figure 3 shows the persistence and temporal dynamics of T. lestoquardi and T. ovis detected among sheep with mixed species infection, using 18s rRNA gene copy number estimated by qPCR as a proxy measurement of parasite density.

Interaction between T. lestoquardi and T. ovis in mixed infection. The detection of mixed infec-
tions of T. lestoquardi and T. ovis increased from 36% in May 2016 to over 80% by Sep 2016 among the examined sheep and all animals with mixed infection, as well as the few persisting single T. lestoquardi infections, remained asymptomatic and apparently healthy (Fig. 2). We therefore tested whether the interaction between T. lestoquardi and T. ovis among these animals was associated with modulation of the load of either parasite.

Discussion
This natural challenge study revealed the long-lasting persistence (over a period of 13 months) of T. lestoquardi and T. ovis parasites as asymptomatic mixed species infections in naïve sheep in Oman, and provided evidence of competitive interactions between these two parasites. The more pathogenic T. lestoquardi was associated with higher density and mortality when present as a single infection relative to mixed infection with the nonpathogenic T. ovis. Thus, co-infection with the two parasites species correlated with reduced mortality of MOT in sheep, supporting the hypothesis of heterologous protection proposed for other combinations of Theileria spp. mixed infection 9 .
T. lestoquardi was associated with higher density when present as a single infection than where it was mixed with the non-pathogenic T. ovis (Figs. 3 and 4). A similar pattern was seen with T. ovis in mixed infection compared to the few occasions when it existed as single infection (4 time points), demonstrating within host interaction between the two parasites. This is evident by the fact that density of both species in mixed infection fluctuated around a threshold over the study period, while the animals were apparently healthy, with a tendency to return to an equilibrium value (Fig. 3), indicating a cross-species mechanism of regulation. In mixed parasite infection, within-host interactions resulting in lower virulence can occur through several mechanisms, including heterologous innate immunity 9 . Within-host selection may also favour traits that allow the parasite to avoid immune responses and survive at low densities 16 . A low degree of serological cross-immune reactivity has been detected between T. ovis and T. lestoquardi together with the relatively long duration of a detectable antibody response 17 , but it is unclear whether this provides cross protection against disease. There is evidence that immunity to T. lestoquardi in sheep involves T cell mediated responses 18 against the macroschizont infected leukocyte and similar immune responses have been reported for T. annulata and T. parva 19 . Genetic and antigenic similarities are evident for T. lestoquardi and T. annulata, and cross-immunity between the two species has been demonstrated 17,20 . However the disease produced by these parasites are distinct from T. ovis, where amplification of the macroschizont infected leukocyte does not occur to any known extent, and replication is more apparent within erythrocytes similar to T. orientalis 21 . Thus, it is debatable whether the reduced parasite density in mixed infections found in this study relate to an acquired immune response: non-specific mechanisms linked to innate immunity and/or growth inhibition should be considered for further investigation.
However, T. lestoquardi was associated with higher density when present as a mixed infection with the nonpathogenic T. ovis. This suggests a prevailing competitive advantage of T. lestoquardi over the non-pathogenic, T. ovis. This is evident by the relatively lower density of T. lestoquardi in mixed infection compared to single infection. The relative density pattern in mixed species infection could be, explained by a competitive advantage of the pathogenic parasite (T. lestoquardi) in an immune-selection (potentially via innate immunity) environment, where it is more likely to overcome nonspecific immune responses, and build up higher densities 22 . However, the suppressive environment of mixed infection on T. ovis, has not been tested, as few animals in cohost displayed single T. ovis. Nonetheless, on the few occasions when single T. ovis infection (n = 4) appeared during the follow up period, as in these few samples T. ovis density was always relatively higher compared to mixed infection (Samples, 3981, 3986, 3988 and 3995; Supplementary Table 1). The above findings may be consistent with experimental studies using genetically-distinct strains of the rodent malaria parasite, Plasmodium chabaudi, in mice, where the more virulent parasite competitively suppresses a less virulent one in mixed infection 11 . Similarly, observational data from multiple Plasmodium parasites also suggest the existence of competitive suppression 23 . The dynamics of multispecies Plasmodium infection in asymptomatic carriers, under intense transmission provide evidence for a density-dependent regulation that transcends species as well as genotypes 23 . Whether this form of interspecies competition operates in ovine Theileria infection, or one species is simply more resistant to detrimental conditions than the other requires further study.
The delayed appearance of T. ovis among the examined cohort (Fig. 3) was not expected from previous reports on species prevalence, but may be attributed to differences in the kinetics of proliferation of the two parasites. T. ovis parasites are not detected in the blood until the large macroschizont stages mature and merozoites invade erythrocytes, whereas for T. lestoquardi, asexual multiplication occurs within leukocytes within the first week of infection with parasites entering red blood cells in week two 24 . Nevertheless, despite the possibility that T. lestoquardi may competitively suppress T. ovis in mixed infections, T. ovis is widespread in Oman 6 , indicative of a high transmission capacity. A high level of T. ovis transmission is also reflected in the increased rate of mixed infection over time seen our study. The two parasites are transmitted by Hyalomma anatolicum, which was the main tick species identified on the examined animals and is widely distributed in Oman 6 .
A high level of genetic diversity and genotype multiplicity of T. lestoquardi was detected in samples from the infected cohort of sheep. Almost all infected animals displayed a novel (or combination of) genotype (defined by unique combinations of alleles detected by the examined microsatellites), including strains extracted from dead animals. This is consistent with the high genetic diversity and multiplicity of infection reported for T. lestoquardi populations in Oman 7 and Sudan 25 . Moreover, the observed increase in T. lestoquardi multiplicity of infection over time can be attributed to super-infection and the consecutive acquisition of novel genotypes combined with delayed clearance of initial infecting parasites. The study site is known for high abundance and infestation of Hyalomma anatolicum 6 , and this is predicted to promote cross-mating, genetic recombination and generation of novel parasite genotypes.
Multiple genotypes were common, among a single and mixed T. lestoquardi infection, while total parasitemia exhibited fluctuation around a limit (Fig. 4), suggesting a similar mechanism for regulation of the dynamics of species-and/or genotypes. The similar level of T. lestoquardi diversity detected in animals with single and mixed species infection suggests that both forms of infection are equally susceptible to establishment of distinct multi genotype populations. The frequent cross-mating, and random re-assortment of alleles on different gens, generate novel genotypes in infected ticks which can readily infect sheep. Following the acute phase of infection, Scientific Reports | (2020) 10:19802 | https://doi.org/10.1038/s41598-020-76844-2 www.nature.com/scientificreports/ however, protective immunity acquired by the animals does not shield from reinfection, a scenario similar to that reported for Plasmodium 26 . However, over time chronic multi-clonal (mixed genotypes) super-infection is considered to promote a status of 'premunition' with elevated protection against a re-occurrence of disease 26 , leading to higher genotype complexity in these individuals. Such findings are in agreement with the significant increase in level of T. lestoquardi genotypic diversity in animals, with both mixed and single infection, without mortality, over time among the examined cohort in this study.
In conclusion, the limited data from this study on the outcome of ovine theileriosis in Oman indicates that single T. lestoquardi infection is associated with high mortality, and that mixed species infections are associated with T. lestoquardi density regulation, indicative of within-host interaction, and lower mortality. Ethical statement. The project adhered to the guidelines and ethics code of animal welfare, approved by the Animal Care and Use Committee of Sultan Qaboos University, Oman. The blood samples were collected under supervision of a veterinarian, a non-invasive method was used to restrain animals, and blood samples collected by experienced technical staff of the Ministry of Agriculture and Fisheries, Oman. The blood samples collected solely for the purpose of this study and have not been used or will be used for another purpose. The volume of blood collected from each animal, at each time point, is compliance with the guidance of the Animal Care and Use Committee of Sultan Qaboos University, Oman. Informed consent was taken from the farm owner before drawing blood from animals.
This research was also performed in accordance with the relevant guidelines and regulations of ethical and animal welfare of the Ministry of Agriculture and Fisheries, Oman.
Detection and quantification of T. lestoquardi and T. ovis. The density of T. lestoquardi and T. ovis parasites in whole blood was quantified using a real-time polymerase chain reaction (qPCR) assay. Primers and probes targeting 18S rRNA gene of T. lestoquardi and T. ovis were designed (Table 1) using available GenBank sequences for T. lestoquardi and T. ovis, as described 6 . The specificity of T. lestoquardi and T. ovis 18S rRNA gene primers was tested against pure T. lestoquardi, T. ovis and T. annulata DNA samples. The T. lestoquardi 18S rRNA primer amplified T. lestoquardi DNA samples while no amplification was detected with T. ovis and T. annulata DNA. Similarly, T. ovis 18S rRNA gene primers amplified T. ovis DNA but not T. lestoquardi or T. annulata DNA.
For assay optimization, genomic DNA was extracted from cultured T. lestoquardi and a natural infection with T. ovis. The 18S rRNA genes were amplified using primers described in Table 1, and purified using Wizard SV Gel and a PCR clean-up system (Promega). Purified PCR product concentration was measured by Spectrophotometry (Nanodrop), and the copy numbers of 18S rRNA genes for both species were calculated using the following equation. DNA concentration (ng) × 6.022 × 10 23 / length of DNA product (bp) × 1 × 10 9 × 650 27 . The estimated 18S rRNA gene copies were 9.26 × 10 10 /μl DNA and 5.3 × 10 10 /μl DNA for T. lestoquardi and T. ovis, respectively. Table 1. T. lestoquardi 18s rRNA and T. ovis 18s rRNA specific qPCR assays. Sequences for forward primers (FW), reverse primers (RV) and TagMan probe (Pr) are shown with their location, based on genbank sequence.
Genotyping of T. lestoquardi. Five polymorphic T. lestoquardi specific microsatellites, TL_MS05, TL_ MS07, TL_MS13, TL_MS280 and TL_MS281 were used to genotype samples, as described previously 7 . Multiplicity of infection was then determined as any infection with more than one allele at the examined loci, and the minimum number of clone/genotypes (MNC) defined by the maximum number of alleles at any locus per infection. Response variables and model fitting strategy. A variety of distributions (Binomial, Negative binomial, Poisson and Gaussian, gaussian zero-inflation) were tested to describe the distribution of T. lestoquardi and T. ovis density between animal hosts. A Gaussian and Gaussian zero inflation distribution gave the most parsimonious fit for T. lestoquardi count data (Log Likihood: − 662.70, x 2 :178.220, P < 2e-16) and for T. ovis data (Log Likihood: − 591.87, x2: 418.4, P < 2.2e−16), respectively. Zero inflation parameters considered the fact that many samples were negative for T. ovis at different time points, especially during the first months of the experiment. The T. lestoquardi and T. ovis density was allowed to vary between hosts by including them as a random effect in the models. This facilitates detection of changes in density over time, as there is substantial inter-host variability. Models were built using the forward stepwise selection method and the Log likelihood ratio test (LRT) was used to determine the most parsimonious fit. When more than one model significantly fitted the data, the one with the lowest Akaike Information Criterion (AIC) was selected 28 .