Approaches to vaccination against Theileria parva and Theileria annulata

Summary Despite having different cell tropism, the pathogenesis and immunobiology of the diseases caused by Theileria parva and Theileria annulata are remarkably similar. Live vaccines have been available for both parasites for over 40 years, but although they provide strong protection, practical disadvantages have limited their widespread application. Efforts to develop alternative vaccines using defined parasite antigens have focused on the sporozoite and intracellular schizont stages of the parasites. Experimental vaccination studies using viral vectors expressing T. parva schizont antigens and T. parva and T. annulata sporozoite antigens incorporated in adjuvant have, in each case, demonstrated protection against parasite challenge in a proportion of vaccinated animals. Current work is investigating alternative antigen delivery systems in an attempt to improve the levels of protection. The genome architecture and protein‐coding capacity of T. parva and T. annulata are remarkably similar. The major sporozoite surface antigen in both species and most of the schizont antigens are encoded by orthologous genes. The former have been shown to induce species cross‐reactive neutralizing antibodies, and comparison of the schizont antigen orthologues has demonstrated that some of them display high levels of sequence conservation. Hence, advances in development of subunit vaccines against one parasite species are likely to be readily applicable to the other.


NENE aNd MORRISON
The pathogenic Theileria species cause acute lymphoproliferative diseases, with high levels of morbidity and mortality in susceptible populations of animals. 6,13,14 Like malaria parasites, Theileria undergo sequential development in nucleated cells and erythrocytes, but pathogenicity is largely attributable to parasite development during the nucleated cell stage. Theileria invade leucocytes, but unlike most other apicomplexan parasites, they reside free within the cytosol of the host cells. 15 Development to the schizont stage induces activation and proliferation of the infected host leucocytes, 16 and, by associating with the mitotic spindle during cell division, the parasites are able to divide at the same time as the host cells, ensuring that infection is retained in the daughter cells. [17][18][19] This process facilitates rapid parasite multiplication prior to differentiation to the erythrocyte-infective merozoite stage. In susceptible animals, large numbers of infected cells are found in the local lymph node draining the site of infection, from which they disseminate throughout the lymphoid system and to nonlymphoid tissues. 20 Infection usually results in death within 3-4 weeks. The mode of replication of the schizont stage of T. parva, T. annulata and T. lestoquardi enables the parasitized cells of these species to be cultured in vitro as continuously growing cell lines. 21 These Theileria are frequently referred to as "transforming species". Other species such as Theileria mutans and Theileria veliferi, which rarely cause disease, multiply predominantly during the intra-erythrocytic stage of development and undergo limited replication in nucleated cells (reviewed in Ref. 22).
The intra-erythrocyte piroplasm stage of T. parva undergoes little or no multiplication, whereas there is some replication of T. annulata piroplasms, 23 which is associated with higher levels of infection of erythrocytes. Infections with T. annulata may result in moderate anaemia and occasionally jaundice, although pathology produced by the schizont stage is usually the primary cause of mortality by both species.
Because of the acute and fatal nature of the Theileria infections in susceptible stock, control of the diseases is particularly challenging.
In the past, prevention of tick infestation by application of acaricides has been used successfully to prevent disease. However, the need for almost continuous use of these chemicals has proved to be expensive and difficult to sustain and runs the risk of selecting acaricide-resistant tick populations. A single therapeutic compound (buparvaquone, marketed as Buparvex) is available, 24 but its use is limited by cost and the need to treat animals during the early stages of disease to be effective.
Moreover, there are recent reports of the emergence of drug-resistant strains of T. annulata. 25,26 Due to the shortcomings of these control measures, it has long been recognized that vaccination is the most sustainable option for control of these diseases. Live vaccines were produced for T. parva and T. annulata over 40 years ago, 5,27,28 but have a number of practical disadvantages that have limited their use in many areas. Efforts to develop vaccines based on use of defined antigens have so far met with limited success. These recent studies have been the subject of several recent reviews. [29][30][31] Herein, we will discuss the current status of vaccination against T. parva and T. annulata and consider the potential value of comparative studies of these parasites for future development of improved vaccines.

| THEILERIA PARVA AND THEILERIA ANNULATA INFECT DIFFERENT CELL TYPES BUT CAUSE SIMILAR IMMUNOPATHOLOGY
The ability to infect bovine cells in vitro with tick-derived sporozoites has enabled the cell tropism of the different Theileria species to be determined. Early studies demonstrated that T. parva sporozoites can bind to and infect B and all subsets of T lymphocytes in vitro with similar efficiency, whereas T. annulata was found to infect monocytes and B lymphocytes but not T lymphocytes. 32,33 Subsequent analyses of the cells infected by T. parva in vivo showed that the vast majority of the infected cells in animals undergoing primary infection were T cells, of both CD4 and CD8 T-cell lineage. 34 Moreover, experiments in which purified cell populations were infected in vitro with sporozoites and then administered to the autologous animals after 24 hours (when no viable sporozoites remained) demonstrated that infected T cells produced lethal infections, whereas infected B cells resulted in transient mild infection from which they recovered. 35 In the animals that received T. parva-infected B cells, infection was first detected in the regional lymph node around the same time as in animals receiving infected T cells, but they were able to clear the infection around 11-12 days after infection and were subsequently immune to parasite challenge. We have recently shown that animals inoculated with infected autologous T cells that have been cultured in vitro for 6 weeks or more also develop similar mild self-limiting infections (Morrison WI and Connelley T, unpublished data). Based on these findings, we have concluded that transformation of the infected cells alone is not sufficient for virulence and that additional as yet undefined properties of recently infected T cells enable them to produce disease. Recent studies have identified a Zambian isolate of T. parva that infects CD8 T cells but not CD4 or γ/δ T cells. 36 Infection with this isolate results in disease, although the authors suggested that it is of lower virulence than other isolates. The relative contribution of infected monocytes and B cells to infections with T. annulata in vivo has not been determined.
Despite the different cell tropisms of T. parva and T. annulata, the pathology they produce is remarkably similar. In both infections, the parasitized cells migrate from the site of infection and become disseminated throughout the lymphoid system. In both cases, there is also evidence of an early powerful nonspecific T-cell response in animals experimentally infected with a lethal dose of sporozoites, but this response appears to be ineffective in controlling the infection. In the case of T. parva, the regional lymph node was found to contain 15%-20% lymphoblasts at a time when there was <1% parasitized cells. 37 A majority of these cells were CD8 T cells, including a phenotypically unusual subset of CD2 − CD8 + cells; they did not show any cytotoxic activity against parasitized cells and also failed to respond to antigenic or mitogenic stimulation in vitro. Separate studies of animals infected with T. annulata revealed a broadly similar picture. [38][39][40] These observations have not been investigated further, nor has there been any comparison of the responses to infections induced by needle and tick challenge. Nevertheless, currently available data indicate that the progressive nature of infections with the two parasite species is not due to a failure to stimulate an immune response but rather that the infection stimulates a dysregulated immune response that fails to differentiate to generate appropriate effector functions. This presumably reflects a strategy by the parasites to delay parasite clearance.

What is remarkable is that infection of different cell types by the two
Theileria species results in a very similar outcome. Further research using contemporary methodologies is required to understand the molecular basis of these immune responses, to better inform what is required for appropriate differentiation of a protective vs a nonprotective immune response.

| Vaccination with parasitized cell lines
The advent in the 1960s of cell culture systems to propagate Theileria-

| Vaccination by infection with sporozoites
The unfeasibility of using cultured parasites for vaccination against T. parva prompted alternative methods of immunization with live parasites. With the development of methods to cryopreserve large stocks of sporozoites, in the form of homogenized infected ticks, 45 titrations of this material in cattle were undertaken with the aim of identifying a dose that would reproducibly result in mild transient infections and immunity. However, this was not achievable because the lowest doses that produced infection in all animals still resulted in severe reactions in some of the animals. 46 An alternative approach involving infection and simultaneous treatment with oxytetracycline was developed, which successfully achieved mild transient infections in all animals. 28,47 Use of a long-acting formulation of oxytetracycline that provided 5-6 days of activity was required to control the infection. This so-called infection and treatment immunization procedure resulted in long-lasting immunity in all animals against high challenge doses of the T. parva isolate used for immunization, but only a proportion of immunized animals withstood challenge with other parasite isolates. 48,49 However, based on a series of experiments involving immunization and challenge of cattle with different combinations of parasite isolates, a mixture of three isolates was identified (known as the Muguga cocktail), which when used to immunize cattle gave broad protection against experimental challenge with different parasite isolates and against field challenge with T. parva. Despite evidence of efficacy, 50 until recently use of the Muguga cocktail vaccine in the field has been limited.

| Practical constraints
Vaccination using the Muguga cocktail requires production of three large batches of T. parva sporozoites by feeding ticks on cattle infected with each parasite isolate, and each batch needs to be carefully titrated in cattle to determine a dose that will reproducibly infect and immunize all animals but will not break through the tetracycline treatment. This complex protocol coupled with the requirement for a liquid nitrogen cold chain to distribute the vaccine presents challenges for quality control and marketing. However, recent initiatives have led to increased field uptake. This has included the establishment of a centre for vaccine production and systems to facilitate distribution of the vaccine.

| Parasite strain-restricted immunity
As referred to above, vaccination against T. parva by infection and treatment was found to require incorporation of three parasite isolates in the vaccine to provide immunity against field challenge. This followed on from field studies in which animals immunized with a single parasite isolate were not protected, providing the first convincing evidence of antigenic heterogeneity in T. parva. 51 More extensive testing of the Muguga cocktail vaccine has demonstrated that it does not provide complete protection against field challenge in all circumstances, in particular against challenge with buffalo-derived parasites. 52,53 Indeed in one study, vaccinated animals introduced into an area grazed only by buffalo showed no protection. 52 These findings are consistent with sequencing data on parasite genes encoding two polymorphic antigens (Tp1 and Tp2) and genomewide SNP density, which revealed much greater genotypic diversity in parasites isolated from buffalo compared with those of cattle origin. 54 The results of two recent studies of the Muguga cocktail vaccine, showed a remarkably high level of sequence similarity, but differed significantly from the third isolate (Kiambu). The Serengeti parasite was originally isolated from a buffalo and adapted to tick transmission between cattle following several tick passages. However, the results of these two studies both suggested that the Serengeti isolate had at some point become contaminated with parasites from the Muguga isolate. Amplicon sequencing and satellite DNA typing also demonstrated that the vaccine components contained minor genotypic components present at <5% within the vaccine parasites. 56 If these minor components contribute to the broad protective capacity of the vaccine, then the possibility that these components might not be present in all vaccine batches or indeed all vaccine doses is of concern regarding standardization of vaccine content. The authors proposed formulating an alternative vaccine comprising a mixture of antigenically divergent parasite clones, to standardize the content of the vaccine and potentially enhance its ability to generate broadly cross-reactive immunity.
This material would have to be derived through tick passage, and the influence on parasite homogeneity through sexual recombination would need to be investigated.
It is of note that antigenic heterogeneity among isolates of T. annulata has also been documented in a number of early studies, which showed incomplete cross-protection between some isolates, with a proportion of the animals succumbing to disease (reviewed in Ref. 4,5). Analyses of these cell lines with isoenzymes and polymorphic genotypic markers indicated that prolonged passage resulted in reduced genetic diversity in the parasites within the cell lines, and these authors suggested that this loss in diversity may account for the reduced capacity to provide protection against challenge with heterologous parasite isolates. Nevertheless, the reasons why strain-restricted immunity, as observed with T. parva, is not a major issue when vaccinating cattle against T. annulata are unclear.

| Acceptability of live vaccines
A key feature of infections with Theileria parasites is their ability to establish persistent infections in the face of immune responses that control the infection. In the case of T. parva and T. annulata, persistent infections (referred to as the carrier state) are usually not detectable microscopically but can be revealed by polymerase chain reaction (PCR) assays and can be transmitted by ticks. 58 The live para-

| Immune protection
There is a large body of evidence indicating that immunity generated by infection with T. parva or T. annulata is mediated by cellular immune responses directed against schizont-infected leucocytes. This information has been reviewed elsewhere 29,30 and will therefore only be summarized briefly here.
Because animals can be immunized by administration of schizont-

| The antigens recognized on schizontinfected cells
A series of studies using T. parva-specific CD8 T-cell lines to screen for recognition of cells transfected with parasite cDNAs has identified 10 antigens ( Table 1) that are recognized by CD8 T cells from immune cattle. 56,71 The genes encoding these antigens are unrelated and are distrib-  respectively, demonstrated that in each case, approximately 70% of the parasite-specific CD8 T cells in MHC-homozygous immune animals were specific for the respective antigen and that these animals did not recognize the other defined antigens. 65  shown that this antigen is highly polymorphic in both parasite species 56,74 and that the corresponding CD8 T-cell lines exhibit parasite strain specificity. Although the repertoire of antigens recognized by CD8 T cells specific for these parasites is far from complete, these initial findings suggest that there is likely to be substantial commonality in the gene products that they recognize. Hence, information from one species can help to focus efforts on antigen screening in the other species.

| Protective immune responses to sporozoites
Although infections with T. parva 77  Moreover, monoclonal antibodies with neutralizing activity (nmAbs) have been produced for both parasites by immunizing mice with sporozoites. [79][80][81] The majority of such antibodies recognize a sporozoite surface protein, called p67 in T. parva 82 and SPAG1 in T. annulata. 81 Allelic variants of the SPAG1 protein are detected in parasites derived from both cattle and Asian buffalo. 83 While allelic variants of p67 have been detected, they appear to be found primarily in parasites derived from the African buffalo rather than from cattle-derived parasites. 84,85 The p67 86 and SPAG1 81 proteins are encoded by single copy genes and consist of ~700 and ~900 amino acid residues, respectively. They been recently reviewed. 31 In brief, recombinant protein gave superior results to the vectored systems used and immunity to severe ECF ranged from 20% to ~70%. A range of different clinical responses to challenge was observed, from no-reaction to mild, moderate and severe disease. It is likely that in the first of these categories of animals, no infection occurred, as these cattle were negative by PCR. Immunity induced by SPAG1 and different forms of SPAG1 has been less well studied, but as with p67 has been found to induce significant levels of immunity (~50%) to challenge. 90 Interestingly, the SPAG1 protein has been shown to synergize with the protective efficacy of the Tams-1 protein 91,92 and a live attenuated T. annulata cell line, 93 indicating that there is merit in assessing the role of multiple parasite antigens in increasing the efficacy of candidate vaccine antigens.
Expression of recombinant p67 in a native and stable form remains a technical challenge. Mapping of linear B-cell epitopes on p67 has revealed that the bovine immune response to recombinant full-length p67 is primarily directed to N-and C-terminal domains, which also harbour linear epitopes recognized by nmAbs. 94 This has led to the demonstration that an easy-to-produce 80 amino acid peptide at the C-terminal end of p67, called p67C, induces the same level of immunity to fulllength p67. 95 This immunogen is now being used to further optimize protective immune responses to ECF. A 108 amino acid C-terminal peptide of SPAG1 fused to hepatitis B core antigen has also been shown to induce immunity equivalent to that of full-length SPAG1. 89 The polymorphic immunodominant molecule (PIM) in T. parva is also a target of murine nmAbs, 96 Herein, we will explore this further to consider whether there is scope for inducing protective responses that are active against both parasite species.

| Conservation in genome architecture and gene content
The genome architecture and protein-coding capacity of T. parva and T. annulata are remarkably similar. Theileria parva 104  Comparative data on the physical map of chromosomes suggest that the total size of the T. parva genome can vary between different parasite isolates. 109 The impact of genome size variations on parasite gene content and biology, however, remains to be fully documented. Such data are beginning to accumulate for T. parva as more strains are sequenced 55,110,111 and their impact on genotypic diversity has been briefly reviewed elsewhere. 31 Both genomes exhibit a highly compact structure. In brief, chromosomal DNA does not contain highly repetitive DNA, and telomeres are short, and noncoding subtelomeric sequences in T. parva are simpler in sequence than those found in T. annulata. 105

| Conservation of candidate Theileria parva and
Theileria annulata vaccine antigens

| Sporozoite antigens
Although the p67 and SPAG1 proteins exhibit only 47% sequence identity, there is sufficient conservation of epitopes between them so that anti-p67 serum recognizes SPAG1 in immunoblots and neutralizes in vitro the infectivity of T. annulata sporozoites, and vice versa. 113 This functionality even extends to some nmAbs. mAb 23F raised to p67 inhibits T. annulata sporozoite infectivity, and mAb 1A7 raised to SPAG1 inhibits T. parva sporozoite infectivity. By Pepscan analysis on p67, 1A7 has been shown to bind to the core peptide sequence PSLVITD.
The sequence PSLVI is present in SPAG1. 113 mAb 23F binds a conformational epitope in p67, which remains to be mapped. An orthologue of p67/SPAG1 is also present in T. lestoquardi, called SLAG1. 114 Encoding 723 amino acid residues, the protein shares 42% and 58% sequence identity with p67 and SPAG1, respectively. The higher level of sequence identity with SPAG1 is perhaps not unexpected as T. lestoquardi is a closer relative to T. annulata than T. parva (reviewed in Ref. 115) Abs to a fragment of SLAG1 bind to p67 and SPAG1 and SLAG1 also contains the sequence PSLVI, which predicts that 1A7 should bind and inhibit T. lestoquardi sporozoites and that these three molecules are antigenically related. 115 The ability of p67 and SPAG1 to induce cross-species immunity has been confirmed, 116 namely that p67 immunized cattle can exhibit immunity to T. annulata sporozoite challenge and vice versa. Both antigens induced approximately 50% immunity to the homologous and heterologous sporozoite challenge. The role of SLAG1 as a candidate vaccine antigen remains to be tested, but given that p67 also protects a proportion of cattle against ECF under field conditions, that is, through parasite tick challenge, 117 this group of related proteins are prime candidate vaccine antigens.

| Schizont antigens
As discussed above, CD8 T-cell responses specific for T. annulata have been shown to recognize orthologues of two of the T. parva CD8 T-cell target antigens identified to date. Moreover, the amino acid sequence of the epitope identified in one of these antigens (Ta5) 74   In the absence of a proven antigen delivery system for induction of protective CD8 T-cell responses with defined antigens, it is currently not possible to determine the protective potential of these conserved antigens. In the meantime, further studies to examine potential crossprotection between the two species, linked to analyses of the specificity of the T-cell responses, may help to shed light on this question.