Genetic selection of Eimeria parasites in the chicken for improvement of poultry health: implications for drug resistance and live vaccine development

ABSTRACT

Apicomplexan parasites of the genus Eimeria are widespread in poultry flocks and can cause the intestinal disease coccidiosis. Early studies, concerned with intraspecific variation in oocyst morphology, indicated that phenotypic changes may be induced by selection experiments conducted in vivo. Genetic selection driven by targeted selection for specific phenotypes has contributed to our understanding of the phenomenon of drug resistance and the development of live attenuated vaccines. Our present knowledge regarding genetics of Eimeria is largely based upon the utilization of such selected strains as genetic markers. Practical advantages of working with Eimeria spp. in the chicken are discussed. The selection of drug-resistant strains by serial propagation has provided useful information regarding the mechanisms of drug resistance and likely longevity of anticoccidial drugs when introduced in the field. Selection experiments to develop precocious strains of Eimeria and growth in chicken embryos have contributed to the development of safe and effective live attenuated vaccines for the control of coccidiosis. Establishment of protocols for genetic complementation by transient or stable transfection of Eimeria is now supporting direct manipulation of parasite genotypes, creating opportunities to expand the range and value of live parasite vaccines. Procedures for developing drug-resistant and precocious lines of Eimeria and/or genetic markers described here are likely to prove useful for researchers investigating the propensity for resistance development to novel compounds and the development of new attenuated vaccines. Such investigations can be helpful in providing a better understanding of biochemical and molecular aspects of the biology of these parasites.

KEYWORDS:

• Eimeria
• chicken
• drug resistance
• attenuation
• live vaccine
• genetics

Introduction

Parasites of the apicomplexan genus Eimeria infect a wide variety of vertebrate and invertebrate hosts and have been the subject of considerable research since the elucidation and description of the life cycle of these organisms (Chapman, 2003, 2014). Their economic significance is pathogenic effects in domestic livestock where they cause the enteric disease coccidiosis. Coccidiosis is especially important in intensive agricultural production systems where large numbers of animals are confined at high density; such is the case in much of modern poultry production. Since the 1950s, coccidiosis has largely been controlled by prophylactic chemotherapy, in which anticoccidial drugs are included in the feed, and to a lesser extent until recently by vaccination, in which live parasites are administered by various means to birds (Williams, 2002; Chapman, 2018). Soon after the introduction of drugs, parasites resistant to them were isolated from the field and this has led to attempts to duplicate the development of resistance in laboratory experiments, primarily in order to “anticipate” how rapidly resistance would develop to new anticoccidials following their introduction (McLoughlin, 1970a). In recent years, live coccidiosis vaccines comprising oocysts of various species of Eimeria have provided an alternative to medication for the control of coccidiosis. Early vaccines comprised unmodified parasites, intended to induce natural immunity but capable of causing the disease they were intended to prevent (Jeffers, 1974a). A significant advance was the demonstration by Jeffers (1974a) that it was possible to attenuate E. tenella by selection for early (precocious) development and that such parasites, in addition to a loss of pathogenicity, were capable of inducing a protective immune response. This has proved a valuable method for the safe vaccination of chickens (Shirley, 1989; Shirley & Bedrník, 1997). Genetic modification of Eimeria parasites by serial selection in the natural host, whereby it is possible to select novel phenotypes, may be a unique approach to the control of diseases caused by protozoan parasites of domestic livestock. In this article, we describe the laboratory experimentation that has made this possible.

Practical considerations

An advantage of genetic manipulation of Eimeria species in the fowl is that, unlike some other protozoa of veterinary importance, the parasites are of no threat to human beings and the entire life cycles can be completed in a short time period (6–7 days depending upon the species) in a host that can be reared relatively cheaply under laboratory conditions. Furthermore, the relatively synchronized and self-limiting nature of the Eimeria life cycle facilitates the collection of discrete populations of oocysts at precise times after the introduction of infection (Shirley, 1989). Nevertheless, the oro-faecal life cycle, in which the infective stage of the oocyst is transmitted via the faeces and hence the environment, poses difficulties in avoiding cross-contamination between species and accidental infection of birds intended as uninfected controls. Accidental contamination is most common with E. acervulina, which has a high reproductive index and is the most prevalent species described in many field surveys (e.g. Jeffers, 1974b; Clark et al., 2016). This is a particular problem with selection experiments concerned with developing drug-resistant strains and attenuated parasites, both of which require repeated serial exposure to infection in successive groups of birds. At the very least, experimental facilities require bird rooms, cages, etc. that can be sterilized, and ideally isolators for raising young birds. Several excellent monographs have been produced that describe optimal procedures for working with these parasites (e.g. Long et al., 1976; Shirley, 1995; Pastor-Fernández et al., 2019). An additional advantage for the genetic manipulation of Eimeria species in poultry is that methods have been developed to initiate infections with single sporocysts and sporozoites, although the procedures involved are technically difficult and have constrained progress in the genetic mapping of clonally derived traits (Shirley et al., 2004). The transmission stage, the oocyst, is diploid and the product of mating, but following a reduction division the sporocysts and each of their eight sporozoites are haploid (Canning & Morgan, 1975) therefore making it possible to work with clonal populations (Shirley & Millard, 1976; Chapman & Rose, 1986; Shirley & Harvey, 1996).

In order to investigate mechanisms of resistance and to select for attenuation, ideally experiments should be initiated with pure strains of Eimeria of known history that have never been exposed to anticoccidial drugs. In the UK such strains were originally isolated and maintained at the extant Houghton Poultry Research Station and the Central Veterinary Laboratory, Weybridge but unfortunately, many other strains described in the literature have been lost and are no longer available for research. Infections should be established with strains that comprise one species, ideally isolated at the very least from a single oocyst. Most experiments have involved E. tenella, an advantage being that oocysts can be collected directly from the caeca with less risk of contamination than from faeces. However, E. maxima has also been used in selection experiments because the large oocyst is readily distinguishable from other species with smaller oocysts, such as E. acervulina and E. mitis (Norton & Joyner, 1975).

Intraspecific variation in oocyst morphology

An early example of the malleability of the Eimeria genome was a study involving selection for maximal and minimal oocyst length in E. maxima (Jeffers, 1978). Variation in the dimensions of oocysts is one of the characteristics that are used to identify species that infect the fowl. Jeffers noted that there had been many investigations of the effects of environmental factors upon the dimensions of Eimeria oocysts and also that significant differences occurred between different strains within species that he attributed to genetic variation. He sought to minimize as many sources of variation as possible in experiments using “directional selection” with E. tenella in which small populations (25 oocysts) of extreme phenotype (maximal and minimal oocyst length) were selected and propagated. Five generations of selection for oocyst length indicated significant differences among progeny after each generation. A decrease in oocyst production was found in the line selected for minimal oocyst length indicating that attempts to alter this characteristic are opposed by natural selection. A change in oocyst morphology also occurred during a selection experiment with E. maxima in which drug resistance was developed to Lerbek®, a combination of methyl benzoquate and clopidol (Norton & Joyner, 1978). Poor sporulation and reduced oocyst output in successive passages was accompanied by the appearance of a small number of oocysts with only two sporocysts, each containing four sporozoites (bisporocystic) rather than the four sporocysts that normally occur in Eimeria. These oocysts were isolated and propagated and although initially the trait was unstable, after 10–14 passages the bisporocystic forms comprised 80% of the population although a few “normal” oocysts were still present.

Drug resistance

Anticoccidial drugs fall into two categories, those referred to as synthetic drugs (produced by chemical synthesis and sometimes referred to as “coccidiostats” or “chemicals”) and those produced by fermentation (the ionophorous antibiotics or ionophores). Drug resistance has been recorded for both classes of compounds and was first reported soon after the introduction of sulfonamides (Waletzky et al., 1954; Cuckler & Malanga, 1955). The quotation from Schnitzer & Grunberg (1957), that “drug resistance has accompanied the development of chemotherapy like a faithful shadow” has proved to be especially true for Eimeria species of poultry in which the practice of prophylactic medication, where drugs are included in the feed almost for the life of the bird, has resulted in the development of resistance to all the drugs that have been introduced. Indeed, in such an environment the parasites are continuously exposed to relentless selection pressure from agents intended to promote their demise. Nevertheless, it was observed that in the field resistance to some drugs such as glycarbylamide, the quinolones (amquinate, buquinolate, decoquinate, methyl benzoquate) and arprinocid appeared rapidly, soon after their introduction, whereas resistance to other drugs such as amprolium, nicarbazin, and monensin developed more slowly (Chapman, 1982). Experimental studies were therefore undertaken to try and simulate the acquisition of resistance in the laboratory with the objective of attempting to predict the useful field life of anticoccidial drugs. The most extensive series of such studies was conducted by McLoughlin and colleagues, at the USDA laboratory in Maryland, USA, who investigated many older drugs (Table 1). Although it is not possible to extrapolate directly from laboratory experiments to the field (Ryley, 1980), such studies can indicate the relative rate at which species can develop tolerance and it was considered reasonable to assume that if a strain develops resistance readily in such experiments, then it is likely resistance will develop rapidly in the field (McLoughlin, 1970a). Experimentation with arprinocid confirmed the utility of this approach. Thus, the number of serial propagations required for resistance to arprinocid was similar to that for the quinolone decoquinate to which resistance developed very rapidly in the field (Chapman, 1985; Williams, 2006). Soon after its introduction in the UK, resistance to arprinocid also appeared, a consequence of which was that this drug was withdrawn from use (Chapman, 1983).

Table 1. Selection of resistance to anticoccidial drugs in laboratory experiments with E. tenella in the chicken.

Drug	Passages^a	Concentration	Reference
Amprolium	17	125	McLoughlin & Gardiner (1968)
Arprinocid	5	60	Chapman (1985)
Buquinolate	<6	82	McLoughlin (1970b)
Clopidol	17	125	McLoughlin & Chute (1973a)
Decoquinate	<6	30	McLoughlin & Chute (1971)
Diclazuril	10^b	1	Chapman (1989)
Glycarbylamide	9	30	McLoughlin & Gardiner (1961)
Halofuginone	9	3	Chapman (1986a)
Methyl benzoquate	5	20	McLoughlin & Chute (1973b)
Monensin	16^b	100	Chapman (1984a)
Nicarbazin	17	125	McLoughlin & Gardiner (1967)
Nitrofurazone	9	480	Joyner (1957)
Robenidine	16	66	Chapman (1976a)
Sulfaquinoxaline	5–10	250	Cuckler & Malanga (1955)
Zoalene	7	125	McLoughlin & Gardiner (1962)

^a Number of passages required to develop resistance to the concentration (ppm) indicated.

^b Partial resistance. More records provided by Chapman (1982).

Selection for drug resistance

The principal method employed to develop resistance involves repeated propagation of Eimeria (usually E. tenella) in birds initially fed suboptimal concentrations of drugs (McLoughlin, 1970a; Chapman, 1978a). As soon as any oocysts can be recovered from birds administered a given drug level they are propagated at higher concentrations, a procedure that is repeated until the approved use level or even higher concentrations are reached. Appropriate controls include parasites simultaneously propagated in the absence of medication; a comparison with those passaged in medicated birds can indicate the degree of resistance achieved. In order to compare the rate of development of resistance between different drugs, it is important to standardize a schedule of inoculations as far as possible, but in practice this can be difficult (Norton & Joyner, 1975). Variable yields of oocysts and poor sporulation can necessitate the use of different doses, particularly in early passages. Chapman, however, was able to standardize experimental inocula and other factors thus enabling comparison between drugs (Chapman, 1976a, 1978b). Simultaneous propagation of parasites at different drug concentrations was important to ensure that the selected line was not lost due to either the absence of oocyst production as the drug level was increased or poor oocyst sporulation. A different approach was necessary in the case of the ionophore monensin, as in this case oocyst production was not entirely prevented even at the recommended level of 100 ppm (Chapman, 1984a). Nevertheless, only partial resistance could be developed after 16 passages at this concentration. Interestingly, it proved possible to induce resistance to monensin following repeated passage of E. meleagrimitis in the turkey after only four generations of selection (Jeffers & Bentley, 1980). Experiments have also been carried out to compare the rate of development of resistance in different species. Thus, after 60 generations of selection E. acervulina showed high resistance to nicarbazin and a nicarbazin/monensin combination but, under the same conditions, E. tenella developed only partial resistance to these drugs (Bafundo & Jeffers, 1990).

Most workers have found it necessary to begin selection using concentrations of drug lower than the recommended level in order to recover sufficient oocysts for subsequent passage. However, according to Weppelman et al. (1977) it is possible to induce resistance in a single step without resorting to lower levels of drug by using large numbers of birds given very high doses of Eimeria. They were able to develop resistance to the optimal concentrations of glycarbylamide and amquinate in E. tenella but were unsuccessful with amprolium, nicarbazin, robenidine and monensin. Attempts to develop resistance to amprolium, clopidol, halofuginone and robenidine in a single step by this method were unsuccessful (Chapman, 1978a, 1986a). The genetic basis of drug resistance in coccidia is believed to involve mutation and the subsequent selection of resistant phenotypes. In the case of amquinate and glycarbylamide, the frequency of resistant mutants was calculated as 5.8 × 10⁻⁸ and 2.4 × 10⁻⁷ per wild-type oocyst, respectively (Weppelman et al., 1977). Failure to select resistance to amprolium, nicarbazin and robenidine in a single step was assumed to be due to the very low frequency of mutants resistant to these drugs (lower than 7.5 × 10⁻⁹). In the case of halofuginone it was estimated that, based upon the calculations of Weppelman et al. (1977), if resistance had been selected the frequency of resistant mutants would have been 1.9 × 10⁻⁹ (Chapman, 1986a). In view of this very low mutation rate, it was considered unlikely that resistance could result from a single mutation. While resistance to some drugs, such as the quinolones, may result from a single mutation, it was considered that for others, where resistance develops in a stepwise fashion, resistance may involve several mutations at multiple loci (Chapman, 1982). Thus, the most likely mechanism in the case of drugs such as amprolium etc. is that a mutant resistant to a low level of drug is first selected and that after several passages in the presence of drug this mutant predominates in the population. A second mutation may then ensue. The combined effects of several mutations could confer resistance to elevated levels of drug. Unfortunately, our limited understanding of the genetics of drug resistance, our lack of knowledge regarding the biochemical mode of action of most drugs and the basis for their selectivity, precludes the resolution of these alternative hypotheses (Chapman, 1997).

Stability of resistance

The stability of resistance has been investigated by propagating experimentally selected resistant strains in unmedicated birds and birds given other drugs. Resistance in E. tenella to glycarbylamide, decoquinate and halofuginone was stable when propagated in the absence of medication (Gardiner & McLoughlin, 1963; Ball, 1968; Chapman, 1986b), as was a strain of E. acervulina resistant to amprolium, buquinolate, decoquinate and sulfaquinoxaline (Jeffers & Challey, 1973). However, resistance to arprinocid at the concentration recommended for commercial use (60 ppm) was lost following passage in unmedicated birds, although stable in a line resistant to 150 ppm of the drug (Chapman, 1986b). In the case of monensin, resistance was lost after passage in unmedicated chickens but was stable in a cloned line derived from a single sporocyst (Chapman, 1984a; unpublished observations). Strains of E. tenella resistant to amprolium, nicarbazin and zoalene became sensitive to these drugs if propagated in birds given other drugs (McLoughlin, 1971) and loss of resistance to decoquinate occurred following passage of a resistant line of E. acervulina in birds given clopidol although the converse did not occur (Jeffers & Challey, 1973). By contrast, resistance to amprolium, buquinolate, decoquinate and sulfaquinoxaline in E. acervulina persisted following passage in the presence of monensin (Jeffers & Challey, 1973). The poultry industry has adopted programmes involving alternation of drugs in successive flocks, the value of which would be enhanced if resistance were unstable, but this appears not to be the case as multiple resistance (strains resistant to more than one drug) is widespread.

Resistance may be lost, or at least diluted, by outgrowth of resistant parasites by drug-sensitive forms, a phenomenon that has been demonstrated experimentally (Jeffers, 1976a; Long et al., 1985), or by genetic recombination between drug-sensitive (vaccinal) strains and wild-type resistant parasites (Williams, 1998). It is possible that outgrowth by drug-sensitive parasites may occur in the field if particular drugs are no longer employed; this requires investigation. Replacement of drug-resistant strains by drug-sensitive parasites has also been reported following the use of vaccines based upon wild-type or attenuated parasites that comprise drug-sensitive strains of Eimeria (reviewed by Chapman & Jeffers, 2014). This has led to proposals involving the alternate use of chemotherapy and vaccination in broiler flocks (known as rotation programmes) that may contribute to the long-term control of coccidiosis.

Genetic segregation and recombination

Genetic mixing by segregation or recombination following cross-fertilization can permit multiple resistance to arise and has been demonstrated in E. tenella for various experimentally selected drug-resistant lines such as amprolium and decoquinate (Jeffers, 1974c), arprinocid and halofuginone (Chapman, 1984b), and amprolium, decoquinate and robenidine (Chapman, 1984b). The few cross-fertilization experiments involving characteristics other than drug resistance include a study involving the drug decoquinate and precocious development, between methyl benzoquate and the enzyme phosphoglucomutase, decoquinate and glucose phosphate isomerase, and between arprinocid resistance, precocious development, and glucose phosphate isomerase (Jeffers, 1976b; Rollinson et al., 1979; Sutton et al., 1986; Nakamura et al., 1988). In E. maxima, segregation/recombination has been shown for various combinations of robenidine and sulfaquinoxaline with clopidol or methyl benzoquate (Joyner & Norton, 1975, 1977). Interestingly, a hybrid could not be obtained between clopidol and methyl benzoquate (Joyner & Norton, 1977, 1978). These two drugs have a similar mode of action inhibiting different points on the electron transport pathway of coccidia and it is possible that determinants for resistance are present on the same chromosome and that genes for resistance are closely linked (Chapman, 1997). If this is the case, then it would be difficult to obtain a recombinant resistant to both drugs, but evidence to support this hypothesis has not been obtained so far. Other examples include cross-fertilization between lines defined by robenidine resistance and distinct antigenic type (Blake, Billington et al., 2011). Genetic recombination experiments have revealed a capacity for cross-fertilization, independent segregation, and the production of hybrid progeny (Chapman et al., 2013; Blake et al., 2015). The haploid nature of most of the Eimeria life cycle, involving repeated rounds of asexual multiplication and production of haploid sporozoites and merozoites, ensures that genes for any resistant mutant will be expressed in the absence of allelic heterozygosity, and have a selective advantage in medicated chickens compared with drug-sensitive counterparts. The existence of a sexual process provides opportunities for segregation/recombination that may permit innovative ways for coccidia to evade control by anticoccidials. Sexual reproduction may result in the generation of novel gene combinations and haploid endogenous development provides strong selection pressure to ensure that any adaptive changes are quickly fixed in the population. The development of resistant strains in the laboratory has provided selectable markers for investigation of the genetics of Eimeria and the creation of linkage maps (Shirley & Harvey, 2000; Blake, Oakes et al., 2011). Understanding the structure of the eimerian genome may contribute to the prospects for new anticoccidial drugs and possible future subunit vaccines as the frequency at which a genomic location hosting a target for anticoccidial selection can undergo recombination will impact upon the stability of that locus in field populations of Eimeria (Blake, Billington et al., 2011). Classical genetic analysis has been applied to the study of anticoccidial drug sensitivity and is the best-documented example of intraspecific variation in coccidia (Jeffers, 1978).

Mode of action of anticoccidial drugs

In a few cases, such as the quinolones and ionophores, the development of resistant strains in selection experiments has enabled a better understanding of their mode of action (Fry & Williams, 1984; Augustine et al., 1987; Chapman, 1997). Quinolones and clopidol inhibit respiration by blocking different points in the electron transport pathway in the parasite mitochondrion (Wang, 1975). Thus, electron transport in mitochondria from a drug-sensitive line of E. tenella was susceptible to inhibition but electron transport in a resistant line was not inhibited (Fry & Williams, 1984). The ionophores are able to transport cations across the cell membrane of sporozoites and cause an accumulation of sodium ions; water then enters the cell by osmosis causing the parasite to vacuolate, swell, and eventually rupture (Smith & Strout, 1979; Smith et al., 1981; Smith & Galloway, 1983). The uptake of monensin by sporozoites of a resistant line of E. tenella was significantly less than a sensitive line, the amount of monensin required to inhibit resistant parasites was 20–40 times higher than for those sensitive to the drug (Augustine et al., 1987). It was suggested that the mode of action of monensin, that of transmembrane sodium transport, was a general “biochemically-nonspecific” activity, and that this might explain why resistance was slow to occur (Smith & Galloway, 1983). Lines resistant to monensin were selected following 35 passages in medicated birds and compared with the drug-sensitive parental line (Wang et al., 2006). Membrane fluidity was significantly less in the resistant than in the sensitive line suggesting that altered membrane fluidity is involved in resistance to monensin. Transcriptome analysis has been carried out with maduramicin (an ionophore) and diclazuril-resistant lines of E. tenella and, by comparison with a drug-sensitive strain, differentially expressed genes were identified (Xie et al., 2020). ATPase EtASNA1 was upregulated in lines of E. tenella selected for resistance to these drugs compared with the drug-sensitive parental line (Yu et al., 2021). In a series of studies, a similar upregulation of malate dehydrogenase, serine/threonine protein phosphatase, citrate synthase, and glyceraldehyde-3-phosphate dehydrogenase was found in the diclazuril and maduramicin resistant lines (Chen et al., 2018; Yu et al., 2020; Wang et al., 2021; Huang et al., 2022). The significance of these findings is unknown.

An objective of several recent studies is the development of methods for the rapid and accurate diagnosis of resistance (Yu et al., 2020, 2021). Such methods might enable a more rational use of drugs and prolong their effective use in the field (Sangster et al., 2002). At present the only way to establish whether resistance has developed is by cumbersome in vivo drug sensitivity tests that are expensive to carry out, time consuming, and can only inform the poultry professional of the status of the drugs being used once resistance is already widespread. By that time, a decision to change to a different drug (or to vaccinate birds) will already have been taken and the determination of resistance will only be relevant to future flocks.

Attenuation

Selection in chicken embryos

The life cycle of E. tenella can be completed in the chorioallantoic membrane of embryonating chicken eggs by inoculating sporozoites, released from sporulated oocysts in vitro, into the allantoic cavity (Long, 1965). Initially, infection caused substantial haemorrhage and death but after 16 passages embryos were able to support the production of large numbers of oocysts (Long, 1972a). Thus an inoculum of 10⁴ sporozoites resulted in the production of more than 3 × 10⁶ new oocysts/embryo (Chapman 1976b). Growth in embryos has also been achieved for E. brunetti, E. necatrix and E. mitis (then considered to be E. mivati) but attempts to obtain endogenous development of E. acervulina, E. maxima and E. praecox in embryos were unsuccessful (Long, 1966). Serial passage of E. tenella, E. necatrix and E. mitis resulted in embryo adaptation that was accompanied by a loss of pathogenicity (Long, 1972a, b; Shirley, 1980; Long et al., 1982; Shirley et al., 1982). In the case of E. tenella, histological investigation of the endogenous life cycle stages indicated that, whereas the parent strain produced large subepithelial second-generation schizonts, the embryo-adapted strain produced small second-generation schizonts that developed only in epithelial cells of the chorioallantoic membrane (Long, 1973). Further studies, in which the embryo-adapted parasites were inoculated into young chicks that were subsequently challenged with the pathogenic parental strain, indicated that protection against challenge was achieved (Long, 1972a, b). An embryo-adapted strain of E. tenella has been incorporated into one commercial vaccine (Livacox®) (Shirley & Bedrník, 1997). However, Jeffers (1985) concluded that more studies were needed to confirm that embryo-adapted strains are truly effective immunogens when used under commercial conditions, and failure to obtain embryo adaptation for economically significant species, such as E. acervulina and E. maxima, was considered a major disadvantage for the practical use of this approach to develop coccidiosis vaccines.

Selection for precociousness

Genetic alteration of the development rate was first investigated in selection experiments with E. tenella in order to elucidate the mechanisms controlling the length of the prepatent period (Jeffers, 1974a, 1975, 1985). The prepatent period is the time interval between the inoculation of oocysts and the appearance of a new generation in the faeces and was studied by selecting the earliest oocysts produced in infected birds (Jeffers, 1975). Oocysts were first harvested at 125 hr post-inoculation and propagated separately from oocysts collected at 168 hr (relaxed selection controls), representing a 43-hr differential. In subsequent generations, attempts were made to increase this differential in increments of 5 hr. Selection was continued for 46 generations. After nine and 27 generations, the duration of the endogenous cycle had been substantially shortened, the prepatent period having been reduced by 12–18 hr, a phenomenon Jeffers referred to as “precociousness”. This was accompanied by a loss of pathogenicity that proved to be a stable trait through 25 generations of relaxed selection. Histological examination indicated that the second-generation schizonts of the selected and unselected lines were present at 84 and 96 hr respectively; those of the selected line were either defective or smaller containing fewer merozoites. Continued selection resulted in a complete loss of second-generation schizogony and a prepatent period reduced by more than 40 hr (McDougald & Jeffers, 1976a, b). Furthermore, experiments in cultured cells indicated that the sexual phase of the life cycle (gametogony) occurred after the first generation of schizogony. Studies on immunogenicity indicated that the precocious strain was able to protect birds against challenge with virulent strains of E. tenella (Johnson et al., 1979) and it was concluded that it might have potential use in live coccidiosis vaccines.

Biochemical and molecular investigations

There have been few studies investigating biochemical and molecular differences between parent and precocious lines of Eimeria. A study of cytoplasmic proteins in a precocious line of E. tenella demonstrated an s100-like molecule with a different molecular weight to that found in the parent strain (del Cacho et al., 1998). They also reported that the expression of Heat Shock Protein 70 in sporozoites of E. tenella showed a gradual decrease as attenuation progressed, and suggested that the protein might be involved in pathogenicity and that low levels of the protein in precocious lines might lead to impaired adaptation resulting in lower multiplication in the host (del Cacho et al., 2005). Dong et al. (2011) investigated changes in gene expression in oocysts of a precocious line and a parent line of E. maxima and found that 21 genes were downregulated, and 11 genes upregulated in the precocious line. Six genes encoded proteins homologous with previously reported proteins, whereas the function of 26 was unknown. Transcriptional profiles of virulent and precocious lines of E. tenella indicated that some genes involved in carbohydrate metabolism, genes related to proteins secreted from the apical complex, cell attachment proteins, mitochondrial proteins, and transporters were strongly upregulated in the virulent line (Matsubayashi et al., 2016). By contrast, the expression of genes associated with cell survival, development and proliferation was strongly upregulated in the precocious line and it was concluded that they might be involved in pathogenicity or attenuation.

Vaccine development

Methodology employed for the selection of strains utilized in the first attenuated vaccine (Paracox®) followed the procedures described by Jeffers (1975). Detailed documentation of these procedures and subsequent examination of efficacy under field conditions has resulted in the most extensive series of publications concerned with the development of any coccidiosis vaccine (e.g. McDonald & Shirley, 1985; Shirley, 1989, 1993; Williams et al., 1999). Depending upon the number of serial passages, selection for precocity resulted in a decrease in the prepatent period of 19, 26, 20, 14, 21, 36 and 35 hr for E. acervulina, E. mivati, E. praecox, E. maxima, E. necatrix, E. brunetti and E. tenella respectively (Table 2). Attenuation of six species of Eimeria isolated from the USA from 1969–1983, has also been described (Long & Johnson, 1988). Depending upon the species, selection for early development resulted in a reduction in the number of generations of schizogony, a reduction in the size of schizonts and the number of merozoites they produce, and a decrease in the time required for schizont maturation. A consequence of the abbreviated endogenous life cycle is that precocious parasites have a lower reproductive potential that may impact the economic cost of vaccine production. Thus, the fecundity of precocious lines is lower and in Paracox®, depending upon the species, these produce only 2–18% of the oocysts produced by the parent strains (Chapman et al., 2002). According to Shirley (1989), this problem may be overcome by the manipulation of parasite doses to obtain the large numbers of oocysts required. However, reduced replication results in greater production costs compared with their wild-type counterparts and limit the vaccinal doses that can be achieved (Blake et al., 2017). The Paracox® vaccine incorporated attenuated strains of all seven species that infect the chicken and, like non-attenuated vaccines such as Coccivac®, is primarily employed to vaccinate long-living and breeding birds. Improvements in vaccine delivery systems, such as spray cabinets, has enabled vaccination of broilers to become a feasible prospect, and versions of these and other vaccines containing fewer species deemed necessary are available (Chapman, 2018).

Table 2. Some characteristics of attenuated strains of Eimeria species from the chicken.

Species	Prepatent period (hr)		Passage number	Reference
	Attenuated	Parent
E. acervulina	70	89	14	McDonald et al. (1982)
E. mivati^a	67	93	14	McDonald & Ballingall (1983)
E. praecox	<64	84	15	Shirley et al. (1984)
E. maxima	<107	121	12	McDonald et al. (1986a)
E. necatrix	<117	138	17	Shirley & Bellatti (1984)
E. brunetti	<84	120	25	Shirley et al. (1986)
E. tenella	109	144	33	McDonald et al. (1986b)

^a Subsequently considered to be E. mitis.

Techniques for the selection of precocious strains of Eimeria have also been utilized for the turkey and rabbit with the ultimate goal of developing live attenuated vaccines for these hosts (Matsler & Chapman, 2007; Rathinam et al., 2016; Fang et al., 2019).

Application of ‘omics technologies

Understanding the genetic basis of selectable traits such as drug resistance, precocious development or strain-specific immunogenicity remains challenging, precluding development of molecular assays to inform choice of optimal anticoccidial chemoprophylaxis or establish efficacy of vaccine application. The highly repetitive nature of eimerian genomes has added further complexity, with even reference sequence assemblies fragmented into many thousands of scaffolds and more than 65% of predicted protein-coding genes unannotated within the E. tenella genome (Reid et al., 2014). Improvements in next-generation sequencing technologies to reduce cost and improve sequence depth (e.g. Illumina) or increase read length (e.g. PacBio, Oxford Nanopore) now offer opportunities for considerable innovation. Application of these tools has improved the reference E. tenella genome sequence assembly from 4664 scaffolds in 2014 to just 33 in 2021 (Aunin et al. 2021), providing a foundation for genome-based research. When combined with genetic mapping studies, it has been possible to identify loci that contribute to arprinocid resistance and precocious development on chromosomes 1 and 2 of the E. tenella genome, respectively (Shirley & Harvey, 2000). Published genome resources for other Eimeria species that infect chickens are less developed than those available for E. tenella, represented by assemblies comprising between 3415 and 15,978 (Reid et al., 2014).

Sequencing bacterial artificial chromosomes identified by genetic mapping that contained more than 100 Kb of the E. maxima Weybridge strain genome has been used to identify a panel of loci whose inheritance associate with susceptibility or escape from strain-specific immune killing (Blake, Billington et al., 2011). Targeted sequencing within each locus from the antigenically distinct Houghton strain was used to identify new vaccine candidates and predict epitopes, informing work towards subunit vaccines. In the future, genome sequencing Eimeria lines with different responses to selection can be used to identify signatures of selection, markers for use in molecular diagnostics and even causative polymorphisms as described for other apicomplexan parasites (e.g. Borges et al., 2011; Miotto et al., 2015). Comparative transcriptome sequencing from closely related Eimeria populations defined by varied susceptibility to anticoccidial drugs, such as diclazuril, maduramycin and monensin, has also been used to identify differentially expressed genes with functional and diagnostic relevance (as outlined above).

Transfection and genetic complementation of Eimeria

Stable genetic complementation of Eimeria by transfection was first described in 2008, creating opportunities for direct manipulation of Eimeria genomes and accelerated selection for phenotypes such as drug resistance (Clark et al., 2008; Liu et al., 2008). Optimal genetic complementation is currently achieved using restriction enzyme-mediated integration and Nucleofector technology (Pastor-Fernández et al., 2019), although targeted gene editing has been described recently using CRISPR-Cas9 (Clustered regularly interspaced short palindromic repeats – CRISPR associated protein 9; Hu et al., 2020). Several researchers have used transfection to test the capacity of Eimeria to function as a vaccine vector, although all studies remain experimental.

Many antigens have been described as candidates for recombinant anticoccidial vaccines, but realistic strategies for mass administration to large cohorts of chickens are lacking. One suggestion has been to use live transfected Eimeria lines to deliver additional coccidial antigens protective against other Eimeria species, creating a multivalent vaccine. Current live vaccines include between three and eight distinct Eimeria lines, making them difficult to produce and manage. Opportunities to express one or more antigens protective against one Eimeria species in a different species could create a single, more cost-effective multivalent vaccine line. Proof of concept has been demonstrated using transgenic E. tenella expressing the E. maxima proteins Apical Membrane Antigen 1 or Immune Mapped Protein 1 to vaccinate chickens, with greater protection reported after co-immunization of both parasite lines (Tang et al., 2019; Pastor-Fernández et al., 2020). Taking the concept further, transgenic E. tenella lines have been tested for expression and delivery of vaccine antigens protective against other apicomplexans such as Toxoplasma gondii (Tang et al., 2016), and other parasites such as Dermanyssus gallinae (Price et al., 2019). Vaccination using E. tenella expressing the anti-Campylobacter jejuni vaccine candidate C. jejuni antigen A reduced caecal C. jejuni colonization by up to 91% following oral challenge, while other parasite lines have been created expressing M2e from avian influenza virus H9N2, VP2 from infectious bursal disease virus, or glycoprotein I from infectious laryngotracheitis virus (Marugan-Hernandez et al., 2016; Zhang et al., 2021). Future development using less pathogenic, or precocious Eimeria host lines may be required before such transgenic vaccines might be considered for licensing.

Genetically complemented transgenic Eimeria can also be used as tools to identify novel immunoprotective antigens or to test drug efficacy in vitro. Transient transfection of E. maxima has been used in an in vivo screen of immunoprotective capacity to prioritize anticoccidial vaccine candidates identified by genetic mapping (Blake, Billington et al., 2011). Quantification of E. tenella expressing a Yellow Fluorescent Protein reporter has been used as a measure of parasite replication during in vitro cultivation that is amenable for use as a screen of anticoccidial drug efficacy (Clark et al., 2008; Marugan-Hernandez et al., 2020). While the assay is restricted by the limited ability of E. tenella to develop past the first round of schizogony in vitro, the approach can be deployed as a preliminary screen of new drugs and phytobiotics before moving to in vivo studies.

Discussion

Pioneering experiments conducted by Jeffers, Joyner, Long, McDonald, McLoughlin, Shirley, and others have provided insight into the genetics of Eimeria and our understanding of phenomena including intraspecific variation in oocyst morphology, drug resistance, embryo adaptation and precocious development. The procedures employed are remarkably similar, involving the repeated serial propagation of parasites over many generations in birds. In the case of drug resistance, the selection pressure of drug concentration is applied, whereas for attenuation the selection pressure comprises intense selection for the first oocysts produced following infection. Both procedures have been successfully utilized to develop strains resistant to commercially significant drugs and attenuation of all species that infect the fowl. From a practical point of view, the eventual fate of most anticoccidial drugs has been the acquisition of resistance that has been extensively documented for isolates obtained from the field and in several cases has resulted in the withdrawal of drugs from regular commercial use. The ability to predict in advance the likelihood of the rapid development of resistance could reduce the enormous expenditure involved in obtaining approval and registration of new products (Ryley, 1980). The principal drugs employed today are ionophores, and their combination with nicarbazin but a threat to their continued use is the desire by some poultry companies, in order to satisfy perceived consumer demands, to raise chickens without antibiotics. This has led to compromised production including reduced livability, increased feed costs, poor feed conversion, increased incidence of various disease conditions and negative effects upon poultry welfare (Cervantes, 2015; Karavolias et al., 2018; Parker et al., 2021). These factors have proved a major disincentive for the discovery of new compounds for the control of coccidiosis and an incentive for the development of novel vaccines. Eventually, it seems possible that our understanding of the immunology and molecular biology of Eimeria will advance to the point where molecular-based vaccines become available. Until that time, attenuated vaccines will play a key role in the sustainable control of coccidiosis in poultry flocks.

We have a superficial knowledge of the biochemical pathways involved in the mode of action of a few anticoccidial drugs but the basis for their selectivity is poorly understood. The first step in such understanding requires the intentional development of pure lines of Eimeria resistant to these drugs that can provide the basis for further research. Similarly, attenuation of pure lines of coccidia in the laboratory may provide strains that can be utilized to increase our knowledge of fundamental processes underlying control of different phases of the parasite life cycle. The development of drug-resistant and attenuated strains as described here may provide the building blocks necessary for such research. Future control of Eimeria that infect chickens would be improved by the identification of genetic markers for drug resistance or, preferably, causative mutations that contribute to resistance and can be used in molecular diagnostics to replace anticoccidial susceptibility testing. Similarly, markers that can be used to differentiate infected from vaccinated animals would be invaluable to manage and monitor vaccination using live parasite formulations. While it is possible that naturally occurring vaccine line-specific markers could be detected, it might be necessary to introduce unique tags by genetic complementation if such approaches become acceptable to the general public and industry. Other opportunities include understanding the genetic basis of naturally occurring variation in virulence, such as that described for E. praecox (Williams et al., 2009). Progress in these and related areas is anticipated as understanding of Eimeria genetics and resources for Eimeria research improve and extend to species that infect other host species.

Disclosure statement

No potential conflict of interest was reported by the author(s).