The contribution of animal models of aspergillosis to understanding pathogenesis, therapy and virulence

Abstract

Animal models of aspergillosis have been used extensively to study various aspects of pathogenesis, innate and acquired host-response, disease transmission and therapy. Several different animal models of aspergillosis have been developed. Because aspergillosis is an important pulmonary disease in birds, avian models have been used successfully to study preventative vaccines. Studies done to emulate human disease have relied on models using common laboratory animal species. Guinea pig models have primarily been used in therapy studies of invasive pulmonary aspergillosis (IPA). Rabbits have been used to study IPA and systemic disease, as well as fungal keratitis. Rodent, particularly mouse, models of aspergillosis predominate as the choice for most investigators. The availability of genetically defined strains of mice, immunological reagents, cost and ease of handling are factors. Both normal and immunosuppressed animals are used routinely. These models have been used to determine efficacy of experimental therapeutics, comparative virulence of different isolates of Aspergillus, genes involved in virulence, and susceptibility to infection with Aspergillus. Mice with genetic immunological deficiency and cytokine gene-specific knockout mice facilitate studies of the roles cells, and cytokines and chemokines, play in host-resistance to Aspergillus. Overall, these models have been critical to the advancement of therapy, and our current understanding of pathogenesis and host-resistance.

animal models, aspergillosis, Aspergillus

Introduction

Animal models of infectious diseases have been used extensively over the past century to study various aspects of pathogenesis, innate and acquired host-response, disease transmission and therapy. Among the organisms studied, Aspergillus has come to the forefront in recent years as a result of its increasing importance as a cause of serious opportunistic infections in various immunocompromised patient populations, which have poor prognosis in spite of treatment [1–4].

Several factors should be considered before any particular model of infection is employed. First, the choice of model and animal should strive to mimic the clinical disease as well as possible. That is to say, is it a model of acute or clinical disease seen in humans or animals? The model needs to be highly reproducible, which requires standardization of all methods needed to perform the model, and should be economical. The parameters of infection, such as mortality and fungal burden in the target organs, should be controllable. Thus, the model needs to be carefully defined. In studies of antifungal drug efficacy, it is wise in most instances to let the infection become established prior to initiating therapy, in order to more closely mimic the clinical situation.

Why are animal models of infection performed? This question has been answered in part above, but these models allow investigations into the evolution and progression of disease (i.e. pathogenesis), aspects of innate and acquired immunity, how disease transmission might occur through fomites, contact or aerosols, and lastly, studies on therapeutics and diagnostics that might improve patient care and outcome.

Several different animal models of aspergillosis have been developed by numerous investigators. In veterinary medicine, aspergillosis is an important pulmonary disease in birds, and several avian models have been developed. Studies done to emulate human (or mammalian) disease have relied on models using common laboratory animal species, such as guinea pigs, rabbits and rodent models. Murine models of aspergillosis predominate as the choice for most investigators. The availability of genetically defined strains of mice, immunological reagents, cost and ease of handling are factors in this choice. Both normal and immunosuppressed animals are used for the various species mentioned. Each of the species mentioned have been used in studies of pathogenesis, immune response, virulence and therapy.

Avian models of aspergillosis

Among the different animals or humans that are affected by aspergillosis, birds are one of the few species that can have naturally acquired infection in the absence of immunosuppression. The most frequent agents of avian aspergillosis are Aspergillus fumigatus, A. flavus, A. terreus, A. nidulans, A. niger and A. glaucus. In general, models have used a variety of birds, with ages ranging from hatchlings to adult animals. Both intravenous inoculation and pulmonary routes (intratracheal or aerosol) have been used to infect the animals. For the most part, a rapidly fulminant lethal disease is established, with deaths occurring as early as two days postinfection.

Studies on pathogenesis have shown that not all birds are equally susceptible to aspergillosis. Rank order for decreasing susceptibility has shown Japanese quail more susceptible than turkeys, followed by guinea fowl and then chickens [5–7]. In these studies, birds were less than six days old and were exposed via aerosol or intratracheal inoculation [5–7]. Although it is possibly not surprising that different species of birds have differing susceptibilities to infection, studies have also shown that strains of chicken vary in their susceptibility to aspergillosis [5]. Temporal studies on the pathology of infection in quail showed focal hemorrhagic lesions in the lungs, air sacs and trachea, which further developed into white nodules [8]; the authors noted that dissemination did not occur [8]. In contrast, three-week-old turkey poults exposed to A. fumigatus or A. flavus conidia showed a hematogenous dissemination of the conidia within 15 minutes of exposure [9]. Others have reported similar findings on the pathology of infection [10–12]. Pathogenesis is also related to the virulence of the organism, with isolates from different sources showing a range of virulence for causing lethal infection of turkeys inoculated intra-air-sac [13].

Interestingly, toxins produced by Aspergillus spp. may or may not play a role in pathogenesis or in altering susceptibility to infection. In particular, aflatoxin added to the diet of turkey poults did not affect susceptibility to infection [14]. However, gliotoxin may contribute to pathogenicity via its immunosuppressive properties [15].

Another question that has been studied is the source of Aspergillus conidia that cause infection. In a case report following a severe outbreak of aspergillosis in a flock of turkey poults, mould counts of >10⁶ per gram were reported in soft wood shavings that were used for litter for the flock [16]. After treatment with copper sulfate, the counts were reduced as was mortality in the flock [16]. Other investigators showed that contamination of egg shells with A. fumigatus and subsequent handling of the eggs resulted in dispersal of the conidia into the hatchery air [17].

Because domestic fowl, as well as other birds, are susceptible to infection, an outbreak of aspergillosis becomes economically important. Thus, investigations into the immunology of aspergillosis have been performed — often in conjunction with vaccine or diagnostic studies. With respect to diagnostics, antigen and antibody detection, polymerase chain reaction (PCR), ELISA and other methodologies have been used [18–23]. Both the PCR and ELISA methodologies appear to have utility in diagnostics. Immunologically, birds respond in much the same way as mammals to infection, and a type I response appears most beneficial. Birds also respond with specific antibody production that is similar in its kinetics to mammals, including the maternal transfer of specific antibodies, but to the embryonated egg rather than to a fetus [22,24,25]. Several studies have examined the utility of vaccines against aspergillosis of birds. Among these, fractions derived from germlings of A. fumigatus have been shown to be effective in preventing subsequent experimental pulmonary disease in turkeys [26–29]. Along with prevention, treatment of avian aspergillosis is important. In a systemic model using pigeons, saperconazole was demonstrated to be highly effective [30]. In other studies, enilconazole was effective in preventing mortality of chickens [31].

Guinea pig models of aspergillosis

Guinea pigs are another animal used for studies of experimental aspergillosis, including models of invasive pulmonary aspergillosis (IPA), systemic disease and a model of endocarditis. These models have been used in therapy studies of IPA established in immunosuppressed animals; these animals are less commonly used than other species. Examples of these types of studies include the efficacy of itraconazole and saperconazole in models using immunocompromised animals [30,32]. Others have more recently used a model of IPA in guinea pigs to study voriconazole and caspofungin [33,34]. In each instance the drug under study showed some efficacy against experimental infection. The endocarditis model is established by catheterization of a femoral vein and inoculation of conidia to establish a left-side endocarditis [35]. This model showed an improved efficacy for voriconazole versus itraconazole in treatment [35].

There are few other studies using guinea pigs. In a report by Reichard et al. [36], an aspergillopepsin PEP⁻ (protease) deficient mutant was compared to the wild-type parental A. fumigatus for its capacity to invade tissue and cause mortality in a guinea pig model. The results indicated that the mutant and the parental were equivalent in those respects and, thus, other secreted or cell-wall associated proteases were likely to be involved [36].

Rabbit models of aspergillosis

Similar to the avian and rodent systems, several species of Aspergillus have been studied in the rabbit, including A. fumigatus, A. flavus and A. terreus [37–40]. Rabbits have been used to study IPA and systemic disease, as well as fungal keratitis. To establish pulmonary or systemic disease, animals are highly immunosuppressed and the choice of immunosuppressive regimen alters the type of disease manifested. However, there appears to be some question as to the need for immunosuppression in keratitis models.

Rabbit models of keratitis have been used primarily for studies of therapy, showing various antifungals (e.g. ketoconazole, oxiconazole, etc.) to have some therapeutic effect in improving either clinical symptomology or cure [39–43]. The models vary somewhat in the manner of establishment, but often the route of infection is injection of the conidia into the corneal stroma [40]. The importance of immunosuppression in the model is in question. However, the use of glucocorticoids, usually given subconjunctivally, alters the course of disease and facilitates the establishment of infection. In nonsuppressed animals, the fungal burden in the cornea declines to undetectable by day 7, whereas if corticoids are given the fungal burden remains stable through 15 days of infection and the inflammatory response in the cornea is worsened [40,44]. Interestingly, during studies to establish a rabbit model of keratitis due to Aspergillus, O'Day [40] noted substantial variation in virulence of different isolates of A. fumigatus. He also noted that the usual effective duration of the model is about 5–10 days.

Systemic and IPA models of aspergillosis in rabbits have been used extensively for experimental treatment studies [1,29,37,38,45–64]. The course of the model is influenced greatly by the choice of immunosuppressive agents. Granulocytopenic animals with experimental IPA had progressive lethal infections with many hyphal elements in the tissues, angioinvasion and extra-pulmonary infection, whereas cyclosporine A and steroid suppression results in reduced lethality, and few hyphal elements or angioinvasion [65]. Both types of model have been very useful for studies of therapeutics (see references above) and for diagnostic studies where samples can be taken temporally from individual animals, and allow for the correlation of the diagnostic results with disease progression or resolution. The rabbit model has also been used to examine alternative parameters of disease, such as tomography, that would correlate well with progression of disease and severity [66].

RIA has been developed for use with bronchoalveolar lavage (BAL) fluids for the detection of Aspergillus antigens and correlation with disease progression or outcome [67]. PCR methodologies have been used on samples from infected animals and correlated with fungal burden and disease progression [48,68]. In one study, PCR methods were used diagnostically in a model of systemic disease [68], with those authors indicating the superior sensitivity of the PCR versus conventional plating techniques of blood cultures. Another investigation used a quantitative fluorescent PCR technique to determine fungal burden from BAL or lung tissues in a model of IPA [48]. Those authors also showed the superior sensitivity of the technique and indicated it to be useful in following the therapeutic response to amphotericin B therapy [48]. More recently, the diagnostic procedure of following galactomannan antigenemia by an EIA assay (GM-EIA) has been used [69]. The results of those studies show that the concentrations of circulating antigen in the blood of rabbits with IPA correlate with fungal tissue burdens and that clinical response to therapy also correlates with GM-EIA results [69].

Murine models of aspergillosis

Rodent models of aspergillosis have predominated as the model of choice over the years. Larger numbers of animals can utilized, handling is easier, genetically-defined animals are available and these model are usually more economical. Both mice and rats have been used; for the purposes of this review we will focus on the mouse models. Information on the rat models can be found in other reviews [1,29,38].

There are three primary models of aspergillosis performed in mice [1,29,38]. Systemic infection can be established in normal animals by intravenous inoculation; the model can also be performed in immunosuppressed animals. Most often, immunosuppression is induced by glucocorticoids or cyclophosphamide administration prior to infection. IPA in mice after intranasal or intratracheal inoculation can only be induced in immunosuppressed animals and more closely mimics the human bronchopneumonia. Several inhalation models have been reported as alternatives to intranasal models. CNS infection with Aspergillus is the most common extra-pulmonary site of infection, and one that results in >80% mortality. We recently developed a model of CNS disease studied in pancytopenic animals after direct intracerebral inoculation of conidia, and found this to be useful for drug efficacy studies and for studies on progression of disease [70,71]. A murine model of allergic bronchopulmonary aspergillosis (ABPA) is a special instance where disease is due to host-response after sensitization with Aspergillus antigens rather than invasive infection. This unique model has been developed and used for pathological and immunological studies [72–76].

Mouse models have been used for a variety of studies on aspergillosis, including examination of the comparative virulence of different isolates of Aspergillus, which genes are involved in virulence, comparative susceptibility to infection with Aspergillus, and preclinical antifungal drug efficacy. Mice with genetic immunological deficiency, and the availability of cytokine gene-specific knockout mice, facilitates studies of the roles that cells, cytokines and chemokines play in host-resistance to Aspergillus.

Because of the various strains of mice available, investigators use both outbred and inbred animals for experiments. The susceptibility of these different strains appears to be about the same, regardless of genetic background of the mice or of the model of aspergillosis being used; even T-cell deficient nu/nu mice are not more susceptible to aspergillosis than are their normal littermates [1,77,78]. However, mice that are C5⁻ appear to be the exception and are more susceptible to both pulmonary and systemic infection [79]. In spite of this increased susceptibility, the C5⁻ mice must be immunosuppressed to establish progressive pulmonary infection [79]. Establishment of disease may depend somewhat on the strain of A. fumigatus used, as there appears to be a correlation between virulence and the presence of an undefined 0.95 kb fragment of genomic DNA [80,81]. However, there are few studies addressing variations in virulence among strains of A. fumigatus or comparisons with other species of Aspergillus (e.g. A. flavus, A. terreus, A. nidulans). In our own experience, we have found A. fumigatus and A. terreus to be about equal in virulence in the systemic model of aspergillosis [82]. This is a question that should be addressed in future studies.

As might be expected, when inocula are chosen for use in the initiation of the infection there is a dose-dependent correlation with the severity of infection, with higher numbers of conidia in the inoculum resulting in more severe disease. This has proven true regardless of the model being studied and strain of Aspergillus being used [70,79,83–85].

It is critical to know what outcome to expect with different inocula, which allows the investigator to control the course of disease in the model. This is particularly important when performing models for preclinical drug efficacy, where too severe an infection may result in early deaths before therapy has begun or after a minimal number of doses have been administered. In contrast, an inoculum that is too low may result in no mortality or even clearance of the infection without therapy. Numerous investigations have been carried out using a murine model of aspergillosis for the study of antifungal drug efficacy [38,45,71,82,86–114]. Among these studies are numerous variations in model and immunosuppression. In general, these types of models have been useful and reflective of future clinical application. A caveat to the interpretation of these data is that the same drug may behave differently against the same organism when used in different murine models of aspergillosis. We have observed that micafungin and nikkomycin Z act synergistically in a systemic model [96], and show no synergistic activity in a pulmonary model of infection [115].

One of the largest uses for murine models of aspergillosis is for the study of virulence factors (genes) in Aspergillus [1,116]. Genes that have been related to virulence included elastase and various proteases [117–126], catalases [127–130], phospholipase [131], toxins [16,119,132–135], adhesins [136–138], restrictocins [119,139–143], and conidial pigments [144–146]. Also, histidine kinase [86,147], PABA synthetase [119,148–150], temperature sensitivity and even ploidy [151–153] have been shown to play a role in virulence. Likewise, some genes have been demonstrated to be less important or not important to virulence, such as alkaline protease [119,154–156] and chitin synthase (CHSE) [1,116,157]. As the tools of molecular biology become more sophisticated and more widely available, many additional studies will be done using a variety of models of aspergillosis.

Over the past decade the use of genetically deficient knockout mice (KO) has increased as the animals have become more readily available. In most cases these mice are deficient in a single cytokine (or factor in the immune response). Use of these types of animals has shown that IL-4 and IL-10 KO mice are more resistant to infection, whereas IL-6, IL-12 and interferon-γ deficient mice show decreased resistance [83,158–161]. The results with IL-10 are of particular interest, since in a model of ABPA IL-10 has been shown to be beneficial [76].

Mention must be made of the mouse mutants that mimic human chronic granulomatous disease, an hereditary condition in which patients are naturally susceptible to various infections including pulmonary aspergillosis [162]. These mutant mice are deficient in NADPH oxidase, which results in deficiency of the oxidative burst in phagocytic cells that contributes toward clearance and killing of Aspergillus. Mice with mutations in either gp91 or gp47 of NADPH oxidase can be infected via a pulmonary route without the use of immunosuppression [127,163–165]. Thus far, the use of these animals for models of aspergillosis has been limited, but may become greater if the mice become more readily available.

Conclusions

Overall, the use of animal models has been critical to the advancement of therapy and our current understanding of pathogenesis and host-resistance. Investigators have determined that various species of animals differ in their susceptibility to infection, as well as that the capacity of the different species and strains within a species of Aspergillus differ in their capacity to cause disease. The use of animal models has clearly demonstrated that a variety of Aspergillus genes contribute to virulence, but that no single gene is responsible for virulence in its entirety. Thus, additional studies on the quantitative genetic aspects of virulence remain to be done. The evolution and specificity of the host response has been demonstrated in animal models of infection and further dissected to show the beneficial nature of a Th1 response and the deleterious consequences of a Th2 response. Extremely important are the studies done on therapy of aspergillosis, which have led to improved patient treatment and improving survival rates. The study of new antifungal compounds in vivo could not be accomplished without the use of these models. Lastly, animal models of aspergillosis allow scientists to address questions of biology and medicine that cannot be answered completely by in vitro studies and provide us with a better understanding of the complexities of the biology of host parasite interactions.