INTRODUCTION

Top Previous Next References

Candida species are ubiquitous, human fungal pathogens capable of initiating a variety of recurring superficial diseases especially in the oral and vaginal mucosae (129, 167). In the late 1950s there was a steadily increasing number of reports on superficial Candida infections associated with the administration of broad-spectrum antibiotics such as tetracycline (91, 178). In subsequent years, the extensive use of steroids, immunosuppressive agents in organ transplant recipients (158, 192) myeloablative radiation therapy (70, 74, 205), and antineoplastics in patients with hematologic malignancies (20, 62, 106) contributed to the increasing morbidity associated with Candida. More recently, mucosal Candida infections have received profuse attention due to the advent of the human immunodeficiency virus (HIV) infection. For instance, it is now known that up to 90% of HIV-infected individuals suffer from oropharyngeal candidiasis (161). This condition is a key feature in staging HIV disease and was once included as a marker in disease classification (64). Curiously, HIV-infected patients appear to be more susceptible than immunocompetent individuals to oropharyngeal but not vaginal or disseminated candidiasis (168, 177). The other general risk factors for oral candidiasis are age (mainly the very young and the very old), denture prostheses, smoking, diabetes mellitus, iron and vitamin deficiences (159), and salivary gland hypofunction (166).

Candida albicans is the species most often associated with oral lesions, but other, less pathogenic species such as C. glabrata, C. tropicalis, C. parapsilosis, and C. krusei are also occasionally but regularly isolated (105, 170). Recently, a novel species, C. dubliniensis, closely related to C. albicans, has been isolated, particularly from mucosal lesions in HIV-infected patients (39).

An important cofactor associated with the pathogenesis of oral candidiasis appears to be the virulence of the infecting organism (113, 153). The specific features of the fungus that contribute to the development of oral candidiasis include its ability to adhere to and colonize the oral mucosa (87), its ability to form cylindrical appendages termed germ tubes (33), and its cell surface hydrophobicity (68). In addition, phenotypic and genotypic switching (176, 186), extracellular aspartyl proteinase secretion (44, 208), and phospholipase production (98) appear to play a subsidiary role in the pathogenicity. Nonetheless, the hierarchy of the importance of these predisposing attributes is little known, although some animal studies described here have shed some light on this issue.

Isolation of Candida from the oral cavity does not imply disease, since its asymptomatic prevalence in healthy persons ranges from 3 to 48% (12) and is even higher in healthy children, 45 to 65% (129). In many epidemiological studies of oral candidiasis, the most commonly isolated yeast species is C. albicans (166). A median carriage rate of 38.1% has been observed for C. albicans alone in a number of surveys in community-dwelling outpatients (129), while a higher carriage rate (up to 78%) has been observed in hospitalized elderly patients (41, 204). Yeast carriage is even higher in those who are HIV seropositive and rises as the CD4⁺ T-cell count falls (67, 164, 196). Other pathogenic members of the genus Candida often isolated from the oral environment are (in descending order of virulence) C. glabrata, C. tropicalis, C. parapsilosis, C. pseudotropicalis, C. krusei and C. guilliermondi (129). C. dubliniensis is a recently discovered novel species, and its virulence potential is much like that of C. albicans due to their close genomic relatedness (190). Despite such diversity among the non-albicans species (143, 189), it is the general belief that they are of low virulence and that disease manifestation is determined mainly by the health of the host (128, 173, 183). Neither colonization with Candida species alone nor a significant increase in their salivary concentration (53) is necessarily a precursor of the development of oral candidiasis (121). Therefore, other local or systemic factors must be present for the organisms to initiate infection and cause disease.

Oral candidiasis may present in a variety of clinical forms, and the three main variants are the pseudomembranous type, commonly known as thrush, and the erythematous and hyperplastic variants (14) (Fig. 1). When two or more of these variants appear in unison, the term "multifocal candidiasis" is used (169). Other common lesions include Candida-associated denture stomatitis, angular cheilitis, and median rhomboid glossitis.

View larger version (36K): [in this window] [in a new window]   FIG. 1.   Simplified schematic diagram comparing the topographic distribution of common lesional sites in human oral candidiasis (A) and experimental candidal infection in an animal model (rat/mouse/hamster) (B). The most common sites of infection are shaded. The clinical variants of the disease and their preponderant sites are as follows: 1, erythematous candidiasis; 2, pseudomembranous candidiasis; 3, hyperplastic candidiasis; 4, Candida-associated denture stomatitis; 5, Candida-associated angular chelitis; 6, Candida-associated median rhomboid glossitis.

The recognition that Candida is an important pathogen, particularly in the immunocompromised host, has resulted in a vast body of in vitro investigations evaluating its virulent attributes in an attempt to elucidate the pathogenesis of the disease. The progress made in understanding some of these features, such as the mechanisms that result in adherence to host tissues (88), cell surface hydrophobicity (69), switching phenomena of the yeast (186, 187), secretion of aspartyl proteinases (208), and phospholipase production (98), is very impressive. Nonetheless, in vivo studies either in live humans or in animals are essential to elucidate and fully comprehend the mechanisms leading to candidal infection.

The host oral defenses against Candida essentially fall into two categories: nonspecific immune mechanisms (e.g., integrity of the mucosae, commensal bacteria, polymorphonuclear leukocytes, macrophages, and salivary factors) and specific immune mechanisms (e.g., serum antibodies, secretory antibodies, and cell-mediated immunity) (38).

The stratified squamous epithelium of the oral mucosa forms a continuous surface that protects the underlying tissues and functions as an impervious, mechanical barrier. The protection so provided is dependent on the degree of keratinization and the continuous desquamation or shedding of epithelial cells. Indeed, the latter mechanism is considered to play a pivotal role in maintaining a healthy oral mucosa and in limiting candidal colonization and infection. The interaction between Candida species and the commensal microbial flora is perhaps the next critical mechanism modulating oral candidal colonization (166). The commensal flora regulates yeast numbers by inhibiting the adherence of yeasts to oral surfaces by competing for sites of adherence as well as for the available nutrients. A number of studies have also shown, both in vivo in gnotobiotic mice and in vitro, that candidal colonization of epithelia could be suppressed by streptococci, which are the predominant resident commensals of oral mucosal surfaces (99, 123, 163).

The human oral cavity is a unique ecological niche because it is constantly bathed in saliva, a biological fluid with potent antifungal and antibacterial activity. In addition, the constant salivary flushing action mechanically inhibits the accumulation of microorganisms in various oral niches. A quantitative reduction in saliva or salivary flow, for instance in Sjögren's syndrome, leads to a xerostomic state with a concomitant increase in oral candidal carriage and infection, indicating the importance of salivary defenses against invading fungi (109, 162). Elements in saliva that inhibit the growth of Candida include nonspecific factors such as the histidine-rich proteins, the proline-rich proteins, the salivary peroxidase system, lactoferrin, and lysozyme (142, 171, 174, 193). The anticandidal nature of histidine-rich polypeptides in particular is noteworthy. Pollock et al. (142) found that the antifungal activity of purified salivary histidine-rich polypeptides is akin to that of imidazoles (104, 134, 142). Lysozyme and lactoferrin are two further nonimmunoglobulin salivary proteins that contribute to the regulation of oral Candida. A number of studies have documented the fungicidal effect of apolactoferrin against Candida (125, 171, 188, 197), while the relative sensitivity of different Candida species to lysozyme has been demonstrated by Tobgi et al. (193) using six different species.

It is known that two independent systems, the systemic and the secretory immune systems, are both involved in defending the oral cavity against Candida. Lehner (100) was the first to suggest that salivary (secretory) immunoglobulin A (sIgA) may contribute to ameliorating the disease process. Individuals with lowered levels of sIgA are more often afflicted with mucosal candidiasis, and functional sIgA appears to prevent the attachment of C. albicans to the mucosal epithelium (200). Polymorphonuclear leukocytes and macrophages have the ability to phagocytose and kill Candida cells. However, the full expression of their activity is dependent on augmentation by cytokines synthesized or induced by T cells (13) and the length of time they survive in the hostile oral environment bathed in saliva.

Mucocutaneous and systemic candidiasis are both typically associated with defects in the cell-mediated immune response (129). A multiplicity of defects in cell-mediated immunity in subjects with chronic mucocutaneous candidiasis have been examined and defined (150). This is further exemplified in patients infected with the HIV, an agent which causes an impairment of the CD4⁺ T-helper lymphocytes, leading to frequent recurrences of oropharyngeal candidiasis (73). These and other host defenses against Candida have been reviewed recently by Greenfield (63).

A number of antifungal agents are available for the management of candidal infections (115). The major agents that are currently used for oropharyngeal candidiasis belong to either the polyenes (amphotericin B and nystatin), the imidazoles (clotrimazole, econazole, ketoconazole, and miconazole), or the triazoles (fluconazole and itraconazole) (52). Nystatin is ideal for topical treatment of oral infections since it is not absorbed from the gastrointestinal tract and hence the adverse effects are minimal. Amphotericin B is less widely used for this purpose due to its treatment-limiting adverse effects such as nephrotoxicity (86).

The introduction of the imidazole and azole groups of antifungals during the last two decades has revolutionized the management of fungal infections (86). The approved azole antifungal agents for the treatment of oral candidiasis are miconazole, clotrimazole, ketoconazole, fluconazole, and itraconazole (52). Miconazole is effective for almost all oral manifestations of candidiasis including chronic mucocutaneous candidiasis. Until the introduction of the triazoles (itraconazole and fluconazole), ketoconazole (an imidazole) was widely used as an alternative to amphotericin B (85), but it suffered from the drawbacks of hepatotoxicity and endocrine toxicity. However the more recently introduced triazole agents, itraconazole and fluconazole, are far superior since they are orally active and water soluble and have a significantly lower toxicity than do the imidazoles (85). Indeed, fluconazole is the drug of choice in the treatment of candidiasis in HIV infection.

The euphoria surrounding the efficacy of the azoles has now been tempered by the realization of moderate or high-level resistance to fluconazole in some species, such as C. glabrata, C. krusei, and C. albicans (148, 170). This phenomenon has been especially common in C. albicans isolated from patients in whom fluconazole has been extensively used, as in those infected with HIV (61, 138). In addition to these topical or systemic antifungals, antiseptic agents such as chlorhexidine gluconate have been used to supplement the drug regimens, especially in treating Candida-associated denture stomatitis. The animal models described herein have made major contributions to the evaluation of these drugs, especially during their developmental stages.

The most common form of oral candidiasis is Candida-associated denture stomatitis, seen in 50 to 69% of denture wearers at one time or another (29). Not surprisingly, therefore, the pathogenesis and management of this condition have been studied in a number of models by many investigators. Budtz-Jörgensen, who pioneerd such studies, employed Macaca irus monkeys with custom-fitted acrylic plates (26) for this purpose, while others have used the Wistar rat model to study Candida-associated denture stomatitis and its histopathology (132, 179, 181). All workers who successfully initiated the disease in animal models have claimed a striking similarity between the human and animal lesions and have stressed the utility of the respective animal model. After reproducing Candida-associated denture stomatitis, the next step was to demonstrate the cofactors involved and the efficacy of topical antiseptics and antifungals in the management of the condition. These therapeutic approaches have included the incorporation of chlorhexidine acetate (97) and azole antifungals (126) to denture base materials, as well as the delivery of systemic imidazoles by this route, using the Wistar rat model (8, 108, 201). However, translation of these into human therapeutic trials has met with little success (50).

Thrush, or pseudomembranous candidiasis, is the best-known form of mucosal candidiasis and has currently come to the forefront due to HIV infection (161). Efforts to produce oral thrush in rats and mice have been successful to varying degrees, and the etiology and therapy of this ailment have been elucidated. For instance, one recent study has shown that superficial candidal invasion and initiation of thrush is favored by topical application of corticosteroids, which dramatically shifts the host-parasite relationship in favor of the yeast (47). Although it is possible to obtain mice that are deficient in the quality and quantity of CD4⁺ T cells, thus mimicking HIV infection, surprisingly little work has been performed to explore this intriguing area (31, 32, 46).

Other predisposing factors for oral candidiasis that have come under scrutiny in a number of animal models include broad-spectrum antibiotic therapy (5, 82, 155-157), carbohydrate-rich diets (66, 154, 155), topical use of corticosteroids (47), corticosteroid inhalation (28), trauma (131), iron deficiency (147, 185), diabetes (51), xerostomia (1, 83, 84, 119, 133), decrease in CD4⁺ T-cell counts and phagocytic function (31, 32), defective T-cell function (16), and immunosuppressive therapy (27, 172). Undoubtedly, these studies have helped us to understand the etiology of oral candidiasis and the development of the management protocols that are currently prevalent.

From a histopathological and diagnostic point of view, most of the lesions described in animal models have faithfully reproduced human candidal lesions. For instance, Budtz-Jörgensen and Bertram (30) and Budtz-Jörgensen (26) experimentally induced palatal candidiasis in the monkey model, which closely mimicked the nonspecific inflammatory changes of the oral mucosa seen in Candida-associated denture stomatitis in humans. The palatal smears from the experimental infection yielded yeasts that were almost exclusively in the hyphal form, as in Candida-associated denture stomatitis (30, 34). Also, a number of studies by others with the Sprague-Dawley rat model of oral candidiasis have detected colonization patterns and lesions that were similar to human lesions both microscopically and histologically (4, 5). Earlier studies by Russell and Jones (156) and Jones et al. (82) using a rat model also demonstrated that Candida carriage and infectivity in this animal are similar to those in humans. The recently described murine acquired immune deficiency syndrome (MAIDS) mouse model (46) is an exciting new development resembling early stages of human HIV infection, which could be harnessed to elucidate the pathogenesis of oral candidiasis. Since 10 to 15% of candidal hyperplastic lesions progress to dysplasia and oral carcinoma (166) a few workers have attempted to investigate this relationship in animal models (117). Intriguingly, a putative correlation between specific biotypes of Candida and oral dysplasia has been demonstrated in one experiment (93) and yet no further studies have been conducted as a follow-up.

For all these reasons and more, animal models have served for more than five decades to illustrate the enigmatic and tempestuous relationship between this opportunist yeast pathogen and its human host. On perusal of the available literature, we were unable to find a comprehensive account of animal models in oral candidiasis. The following, therefore, is an attempt to review in detail the microbiological, histopathological, and therapeutic approaches and potential caveats pertaining to experimentally induced oral Candida infections in animal models described in the English language literature during the last half century.

In vitro tissue culture systems derived from nonhuman sources were used by a few investigators to study the pathological processes in candidal infection much earlier than the introduction of the in vivo experimental animal models. Partridge (136) was the earliest to confront this problem and used the chick chorioallantoic membrane to culture pathogenic fungi. Subsequently, Cawson (35) used the same assay to evaluate the hyperplastic response of the ectoderm to candidal invasion while Hurley and Stanley (77) experimented with cultured mucosal cells from the lingual dorsum of neonatal rats for the same purpose. They also assessed the yeast-induced cytopathic effect and the association between the growth phase of yeasts and the lethal effect on tissues.

As opposed to these animal systems, cultured human explants and tissues have been used by a few investigators. Pemberton and Turner (137) used cultured human gingival epithelium to investigate C. albicans invasion, while cultured explants from the lingual dorsum were used in ultrastructural studies by Miles (120) and Howlett (75) to compare the invasive potential of different Candida species. The findings by these workers were very similar to those for clinical (oral) candidiasis, supporting the notion that in vitro cell culture systems were an appropriate model for the study of the disease.

At about the same time, Montes and Wilborn (122) demonstrated in clinical histopathologic studies that Candida penetrates the human oral epithelium in both acute and chronic phases of the infection and essentially behaves as an intracellular parasite. These findings gave impetus for more detailed studies on oral mucosal invasion of Candida. Subsequent electron microscopic studies by Cawson and Rajasingham (36), with biopsy tissues from patients, also demonstrated clearly the invasion of the hyperplastic oral epithelium by candidal hyphae. The results of these investigations were barely adequate to unravel the complexities of the disease process, and animal models (e.g., monkeys, rabbits, rats, and mice) have been continually used since then to study the genesis of oral Candida infections. We review below the advantages and disadvantages of these animal models and then the experimental details and outcomes of investigations related to each model.

PROS AND CONS OF CURRENT ANIMAL MODELS

Top Previous Next References

Monkey Model (Macaca irus)

Primates appear to be the ideal animal model for experimental Candida infections because of their relatively close kinship to humans. The composition of the oral microflora of monkeys, especially M. irus, is both qualitatively and quantitatively very similar to that of humans (23, 24), and C. albicans is a frequent oral saprophyte in monkeys (23, 152). In addition, monkeys are able to retain acrylic plates resembling denture prostheses in place, a prerequisite for experimental studies on Candida-associated denture stomatitis. However, primates are relatively expensive and difficult to maintain, especially for large-scale experiments. Some workers have also reported that artificial oral infestation of monkeys with Candida is difficult and unreliable (133). Hence, the monkey model has been largely replaced by smaller mammals such as rats and mice, which have gained popularity.

Two species of ratsSprague-Dawley (SD) and Wistarhave been widely used in experimental oral Candida infections. The two main advantages of the rat model are the low maintenance cost and the sufficient size of the oral cavity, which easily permits inoculation and sample collection. Furthermore, the tongue of this animal is fairly easily colonized by Candida, demonstrating conditions such as median rhomboid glossitis and atrophic candidiasis (2) (Fig. 1).

Candida infections in SD rats can be experimentally induced within a few weeks without traumatizing the mucosal epithelium, and a number of investigators claim this model to be satisfactory since it is known to yield consistently reproducible data (3-5, 54, 55, 82, 155, 156). However, the vast majority of these workers (with the exception of perhaps one group) had to provide antibiotic (e.g., tetracycline)-laced food to initiate the lesions. The clinical and histologic findings in experimental disease in SD rats are similar to those of humans. Clinically, small white patches of "thrush" can be visualized on the keratinized lingual mucosa and sometimes on the cheek mucosa.

According to some workers, rats are likely to harbor C. albicans in the oral cavity, albeit to a lesser extent than humans do (80). However, we were unable to find quantitative estimates of candidal colonization of the oral mucosa in wild-type rats. Therefore, one disadvantage of the rat model could be that the animal may harbor C. albicans as a low-level transient commensal (130, 206) and therefore the contribution of the innate immune response to the disease process would be difficult to fathom. However, it could be argued that such natural prevalance of oral Candida mimicks the human ecosystem since 30 to 50% of humans carry oral yeasts (166). Workers using this model for future studies should therefore bear in mind the critical importance of ruling out natural oral colonization by Candida prior to artificial inoculation.

As opposed to the rat, Candida is not a constituent resident oral microbe of the conventional laboratory mouse (96, 139). This appears to be a major advantage of the experimental mouse model of oral candidiasis. In addition, since the murine bacterial flora has been well characterized and recognized to consist of fewer than 20 species, (194), this model permits evaluation of the role of oral commensal bacteria in initiating or suppressing candidal infection. Moreover, the immunobiology of the healthy murine oral mucosa has also been fairly well characterized by a number of workers (48, 49, 94), making it ideal for unraveling adaptive immune responses of the mucosal tissues to candidal infection. Furthermore, mice are easily obtained in large numbers and their maintenance is cheap. Conventional infant mice can be readily colonized by topical inoculation of the oral mucosal surfaces with 10⁸ pelleted C. albicans blastospores per ml (95). Their small size could be considered an added advantage as it facilitates routine daily monitoring, especially when large numbers are used. Nevertheless, the size of the murine oral cavity can also be considered a distinct drawback due to the difficulty in monitoring mucosal changes by naked-eye examination. Hence, some workers have cultured the tissues or organs of the whole animal to ascertain infestation or infection (15).

Mouse mutants are also extremely useful for experimental studies. The sex-linked anemia (sla) mutant, for instance, is ideally suited for experimental candidiasis since it shows a consistent and a prolonged degree of anemia without artificial dietary restriction or bleeding (18, 65). Another mutant, an inbred strain with a metabolic disease resembling diabetes mellitus in humans, has been reported (76). Since uncontrolled diabetes mellitus is well known to predispose individuals to oral candidiasis, this model could be of potential value in understanding diabetes-related oral candidiasis.

Other mutants of this model appear useful for studying the effect of inherited disorders on the development of oral candidiasis. For instance, an autosomal recessive mutation responsible for severe combined immune deficiency (SCID syndrome) has been reported in mice (22). SCID is a rare congenital syndrome of humans that results in loss of both B- and T-cell immunity, and SCID mice are also severely deficient in these lymphocytes. Interestingly, the SCID-hu model has been used to study the infection of human lymphoid cells with HIV-1 a condition which is well known to initiate and aggravate mucosal candidiasis (114, 124). Perhaps the SCID mouse model may be used in future for studying oral candidiasis in HIV infection, together with the very recently described MAIDS model (see below).

Other immune disorders such as the athymic state, X-linked B-lymphocyte defects, and candidiasis related to these syndromes have been investigated using the mouse model (71, 175). A mutant mouse strain called the beige mouse, with a lysosomal defect resulting in deficient phagocytosis (60, 151) as well as deficient NK-cell activity (11, 57, 107, 149), has been used for additional studies. Indeed, beige mice handled thrushlike lesions less well than their littermates did. These mice, as expected from their lysosomal defect, which impairs phagocytosis, are also susceptible to systemic candidiasis. The foregoing mouse mutant variants have given us a fresh insight into the host defense mechanisms operational in superficial forms of oral candidiasis. Further details of oral Candida infections in these models are provided later in this review.

Although not as popular as the preceding animal models, the cheek pouch of the hamster has been used by some workers to investigate experimental oral candidiasis (116). McMillan and Cowell, (116) found that a single inoculation of the organism (10⁷ CFU per ml) was adequate to cause infection or infestation of the hamster cheek pouch mucosa. Artificial ligation of the cheek pouch with sutures after Candida inoculation was also noted to be a simple manouvere to retain the inoculum within the cheek pouch. These workers used the latter technique to study candidal infection in the hamster cheek pouch after induction of epithelial hyperplasia by turpentine (in liquid paraffin) application. The disadvantages of the hamster cheek pouch are its low oxygen tension and the lack of a natural salivary flow, which only poorly mimic the oral milieu. Hence, this model, for all intents and purposes is rarely used (118).

In the next section we sequentially review in detail the reports in the English literature on experimental oral Candida infections conducted in five different animal models, namely, monkey, rat (Wistar and SD), mouse, and hamster. Further experimental details of these studies are tabulated in chronological order in Table 1 for ease of reference.

ORAL CANDIDIASIS IN ANIMAL MODELS

Top Previous Next References

Monkey Model

Denture stomatitis is the most common condition which affects the palatal mucosa of denture wearers, about 69% of whom are infected (29). The main etiologic agent responsible for the condition is Candida, which proliferates at the interface between the denture (almost always the upper prosthesis) and the mucosa (43). Budtz-Jörgensen (26), in an elegant series of studies using the monkey model, demonstrated that the palatal inflammatory changes observed in these animals resembled those of human denture wearers.

Early experiments were conducted by inoculating C. albicans under custom-made acrylic plates fitted to the palatal surface of monkeys (26). Candidal inoculation of control groups of monkeys without an acrylic plate, or those fitted with an uninoculated plate, did not reveal clinical or histologic changes in the palatal epithelium. However, when C. albicans was inoculated under the acrylic appliance, acanthosis and hyperplasia of the epithelium, together with a cellular infiltrate of the lamina propria, were noted. A diffuse erythema confined to the mucosa in contact with the acrylic plate was also observed, reminiscent of Candida-associated denture stomatitis (26, 30). Although all animals demonstrated Candida carriage, the yeast load on the palatal mucosa was not quantified by these workers. The authors also observed that (i) it was essential to cover the mucosa to produce the experimental infection and (ii) topical treatment with tetracycline enhanced candidal proliferation and the severity of the infection. The primary inflammatory lesion showed spontaneous healing 2 to 3 weeks after infection, clinically and histologically. Interestingly, repeat inoculation leading to reinfection of the healed mucosa resulted in a more intense erythema in comparison with the primary lesion. These primary and secondary inflammatory responses appeared to indicate that delayed hypersensitivity may be involved in Candida infections of the oral mucosa, although too few animals were investigated to give statistically valid information.

In a subsequent study, the same author showed that the natural healing process after candidal infection can be temporarily suppressed by systemic immunosuppressive therapy with azathioprine (27). Experimental Candida infection of the palatal mucosa was induced in 14 adult M. irus monkeys by inoculating C. albicans (serotype A) under acrylic plates. Seven animals were given azathioprine, and the remainder acted as controls. The control animals demonstrated an atrophic and erythematous epithelium which resolved within 2 to 3 weeks. Cellular hypersensitivity to C. albicans was measured by an in vitro leukocyte migration test (27). In the normal animals, the migration inhibition was significant from 1 week to 5 months after infection. Cellular hypersensitivity developed concomitantly with clearing of the infection, while antibody was not yet detectable as assessed by an agglutination reaction. On the other hand, in azathioprine-treated animals, the infection persisted and cellular hypersensitivity did not develop until 1 to 3 weeks after the drug treatment was discontinued. The antibody titer also rose consistently after 4 weeks, reaching a maximum at 8 months. In immunosuppressed monkeys, a depressed migration inhibition reaction, together with a delayed cellular immune response (up to 2 to 3 weeks), was seen; the humoral immune response was early and of shorter duration. These monkeys also developed thrush-like lesions on the palatal mucosa and a mild inflammatory response in the lamina propria; Candida hyphae were visible in the stratum corneum, as in human lesions. This study demonstrated that cellular hypersensitivity to Candida plays a critical role in host resistance to experimentally induced candidiasis.

Oral thrush is relatively common in those using steroid inhalers for asthma and other allergic conditions. To investigate this condition, Budtz-Jörgensen (28) used monkeys injected with the corticosteroid triamcinolone acetonide. Of 13 monkeys in this study, 6 were injected with the steroid triamcinolone acetonide intramuscularly 2 weeks before and 2 weeks after inoculation. In the control group, acute atrophic candidiasis without hyphal invasion was noted and healed within a period of 2 to 3 weeks, while in the steroid-treated group, thrush was seen in all animals together with hyphal invasion of the ortho- and parakeratinized epithelium. A depressed inflammatory response also persisted in 50% of the animals in the steroid group for 5 to 6 weeks until the plates were removed. The inflammatory response was marked under the area covered by the acrylic plate. These studies confirmed that the local environmental conditions such as restriction of salivary flow due to the acrylic prosthesis, as well as the systemic immunity, are important in the initiation and aggravation of oral Candida infections.

Another group of workers used the same model but a different monkey species, Cercopithecus aethiops, to induce thrush using maxillary acrylic plates and inoculations of C. albicans (133). They investigated the effect of a reduced salivary flow induced by systemic oxyphencyclimine chloride on pseudomembranous lesions under a maxillary plate. They reported that monkeys with reduced salivary flow developed larger lesions while reaffirming that the acrylic plate is a prerequisite for initiation of oral candidiasis.

To conclude, due to reasons such as the purchase and maintenance cost cited above, the monkey model has fallen into disfavor. Nonetheless, the pioneering work of Budtz-Jörgensen (26-28) and Olsen and Haanaes (133) using this model was instrumental in defining the basic pathological processes involved, especially in Candida-associated denture stomatitis. It is noteworthy that the monkey model served as the "gold standard" for subsequent animal modelsnamely the rat, the mouse, and the hamster.

At almost the same time as the experimental studies with the monkey model were being conducted in Scandinavia, Jones and Adams (78) were investigating an alternative, the rat model, in the United Kingdom. These workers found that the Wistar rat was a simple and less expensive alternative to the monkey model for experimental oral fungal infections. In their first investigation, which was done with 26 Wistar rats and lasted 10 days, Jones and Adams (78) demonstrated asymptomatic colonization of the mouths of all the animals and histologic evidence of candidiasis in some 50% of the rats orally inoculated with C. albicans. To demonstrate asymptomatic colonization, they sampled the oral cavities of the rats with sterile paper points (1 cm in length), which were then incubated in Sabouraud's broth for 24 h at 37°C to check for candidal growth. Histologically, infection occurred on the dorsal lingual surface, the buccal mucosae, and the free and attached gingivae (Fig. 1). Both the histology and the clinical appearance of the lesions closely resembled acute oral candidiasis in humans.

The authors subsequently extended these experiments to demonstrate the effect of xerostomia on oral candidiasis by desalivating the rats with hyoscine hydrobromide (1). A total of 42 Wistar rats were subjected to similar experimental conditions as before but for an extended period of 6 weeks. To determine oral candidal infection, one rat from each of the six groups was sacrificed weekly up to 6 weeks. The decapitated heads were fixed in 10% formol saline and decalcified, and sections were prepared for hematoxylin and eosin and periodic acid-Schiff staining, which showed evidence of candidiasis in 5 of 30 rats inoculated with C. albicans. The investigators observed epithelial abnormalities such as parakeratosis and thickening of the stratum corneum in lesional tissues, indicating that the model faithfully mimics the chronic hyperplastic variant of candidal infection.

Subsequently, Olsen and Bondevik (132) used the Wistar rat as an alternative model to study Candida-associated denture stomatitis. They used 38 Wistar rats in two experiments, each with an observation period of 2 weeks. The rats in the control and test groups were fitted with uninoculated or Candida-inoculated acrylic plates, respectively. After 1 week, a generalized simple palatal inflammation similar to that of humans was seen in the test group, and its histopathology resembled that of palatal inflammatory lesions in humans.

A similar but more extensive study, conducted by Shakir et al. (179) using 77 male albino Wistar rats, lasted for 6 weeks, in comparison to the 2-week observation period of Olsen and Bondevik (132). The results were similar since they observed that both an acrylic appliance and C. albicans inoculation were prerequisites for inducing palatal inflammation. The epithelial changes intensified with the duration of the experimental period, and after 6 weeks focal areas of the palatal mucosa were atrophic and markedly hyperplastic with hyphal penetration, resembling the later stages of Candida-associated denture stomatitis (Newton's type III) seen in humans. This study also helped dispel the theory that trauma alone from ill-fitting dentures can induce palatal inflammation since the presence of C. albicans was essential to the induction of inflammatory changes. Using a similar experimental design, Shakir et al. (179) further observed that C. albicans serotype A is more pathogenic than serotype B in inducing palatal candidiasis. Also, a single strain each of C. tropicalis and C. glabrata failed to induce pathologic changes (180), implying a heirarchy of virulence in Candida species. Although this simple experiment is indicative of the relative pathogenicity of Candida species, more comprehensive animal studies to illustrate this phenomenon are needed.

In another investigation with the Wistar rat, the same group observed that after inducing palatal candidiasis (with C. albicans CA 3091 serotype A) by using an acrylic appliance, removal of the appliance resulted in complete resolution of the lesion, although Candida still persisted as a commensal for up to 2 weeks (181). Nonetheless, the organisms transformed into the pathogenic form when the appliance was refitted without further inoculation. Microbiological sampling was conducted by swabbing the palatal mucosa immediately after killing and observing the resultant growth on Sabouraud's agar. To confirm whether the recovered yeasts were C. albicans 3091 serotype A, the isolated colonies were serotyped. This experiment, which parallels the clinical experience in denture wearers, confirms the critical role of the denture in initiating Candida-associated denture stomatitis and the importance of good denture hygiene in the management of the disease.

The association between filament formation in yeasts and oral candidiasis is still unclear. The superior virulence of both forms of Candida, i.e., Candida blastospores and hyphal forms (129), in human tissue has been reported. Germ tube formation, which precedes hyphal growth in C. albicans, is generally associated with increased adherence to epithelial cells (89) and resistance to phagocytosis by virtue of their large physical dimensions. Some studies also suggest that the hyphal structures are better than the individual yeast cells of Candida at gaining a foothold during the primary invasion process of the host (111, 184). These views have generally led to the belief that germ tube and hyphal formation in C. albicans accentuates disease induction in humans (127). To investigate this phenomenon, Martin et al. (111) compared the pathogenic potential of two germ tube-negative strains and a single germ tube-positive strain of C. albicans. When inoculated into three groups of rats fitted with an appliance covering the palatal mucosa, the germ tube-negative strains (MS997 and XTM2) did not produce palatal histologic changes; no changes were observed in rats not fitted with an appliance. In contrast, the germ tube-positive strain (C. albicans 3091 serotype A) elicited a chronic inflammatory response together with hyphal invasion and epithelial hyperplasia. Nonetheless, in the absence of an appliance, no pathologic changes were noted. These results reinforced the contention that hyphal formation or filamentation is an important pathogenic attribute of Candida species (42, 184, 209).

Candida species have a predilection for specific anatomical sites of the oral cavity. They commonly reside on both the nonkeratinized and keratinized oral mucosae of humans, particularly the lingual dorsum and the buccolingual surfaces, while gingivae are not normally favored. The oral colonization profile of Candida was determined by Fisker et al., following short-term oral inoculation of C. albicans in Wistar rats on a tetracycline-laced diet (54). This experiment revealed preferential C. albicans colonization of four main areas of the oral mucosa. Almost 98.8% of infective foci evidenced by pseudohyphal penetration of mucosal epithelium were found on the buccal mucosa, the buccal and lingual sulci, and the crest of the molar gingivae, and in the interpapillary areas of the dorsum of the tongue (Fig. 1). The remaining foci (1.2%) were in the mucosa of the hard palate and the attached gingivae. Associated ultrastructural studies clearly revealed that the lingual surfaces which show preferential yeast colonization, particularly the interpapillary areas, were characterized by an uneven irregular epithelium with a loosely structured stratum corneum (54). The investigators surmised that the loss of cell cohesion and the abundant intercellular clefts between keratinized cells having a microplicated surface facilitated the colonization and initiation of hyphal penetration (140). The contention that the profile of oral candidal colonization and hyphal penetration is related to the degree of keratinization and/ or surface morphology of the mucosa was well supported by these findings.

The Wistar rat palatal candidiasis model has also been used to evaluate the therapeutic efficacy of topical oral antiseptics and antifungals used to treat this condition. Lamb and Martin (97) incorporated chlorhexidine acetate into an autopolymerizing resin appliance at a sufficient concentration to prevent palatal candidiasis in the Wistar rat and proposed that the effect was due to the slow release of the antiseptic. Similarly, Norris et al. (126) examined the therapeutic efficacy of the azole antifungal miconazole incorporated into autopolymerizing acrylic resin and observed that palatal candidiasis could be prevented by fitting Wistar rats with appliances supplemented with 10% (wt/wt) miconazole in acrylic polymer powder. The appliances were well tolerated, since the rats remained healthy during the experimental period, and indeed the rats wearing drug-laced appliances gained weight more rapidly than did their drug-free counterparts. In contrast, in the investigation with chlorhexidine acetate, the test animals lost weight, probably due to the adverse effect of chlorhexidine (97).

The efficacy of imidazole and triazole antifungals (ketoconazole and fluconazole) in denture stomatitis has also been studied in Wistar rats using palatal acrylic appliances inoculated with C. albicans (108). The authors observed that a ketoconazole dose of 7.0 mg/kg of body weight¹ and a fluconazole dose of 0.75 to 1.0 mg/kg of body weight¹ for 14 days was necessary to prevent the recrudescence of palatal candidiasis. Although human trials of drug-laced palatal appliances have not been conducted to our knowledge, some workers have used this principle and incorporated antifungal agents into denture-lining materials with some degree of clinical success in Candida-associated denture stomatitis patients (50).

The oral mucosa serves as a rugged, impenetrable barrier against a multitude of physiological and pathological insults. This primary host defense mechanism is highly effective due to the prolific and incessant epithelial cell turnover, and it has been postulated that this activity may accelerate under slowly progressing chronic disease conditions. However, histologic investigations of Candida-associated denture stomatitis patients have revealed that the mitotic activity of the palatal epithelium is similar to that of the healthy palatal mucosa (198, 203). Nonetheless, experiments by Shakir et al. (182) with Wistar rats indicate that Candida infection results in a significant increase in the mean numbers of mitotic figures per unit length of basement membrane in the palatal epithelium of the inoculated animals fitted with appliances. Since this increased epithelial proliferation and desquamation could be considered a protective measure that wards off systemic fungal invasion, Van Mens et al. (198) have suggested that hyperplastic lesions in Candida-associated denture stomatitis are defense mechanisms of the host. Indeed, the exuberant granulomas of chronic mucocutaneous candidiasis and similar syndromes could be considered an extreme evasive reaction of the body to fungal invasion (129).

Since the presence of an oral prosthesis traumatizes the palatal epithelium (19, 29), some workers have conducted Wistar rat studies to investigate the effect of candidal infection on the barrier properties and permeability of the palatal epithelium (110). They observed that in the healthy rat, the palatal epithelial barrier was impermeable to the passage of lanthanum, whereas in the presence of candidal infection, the permeability barrier was selectively operational, with a predominant leakage of low-molecular-weight proteins and selective permeability of macromolecules. Furthermore, an electron-dense material was noted throughout the subepithelial tissue. Removal of the prosthesis resulted in healing of the epithelium and a reversal of the barrier properties to its original state, implying that permeability changes are intimately associated with palatal inflammation in Candida-associated denture stomatitis. The pathological effects, if any, of the loss of permeability of the diseased human palatal epithelium are unknown.

As stated above, diabetes mellitus is a common disease that predisposes to oral candidiasis (160). Dourov and Coremans-Pelseneer (51) conducted experimental studies with streptozotocin-treated diabetic rats to investigate the oral candidal carriage and histopathology induced over a 40-week experimental period. The oral flora was quantified before and after inoculation. Tongue swabs were taken and cultured on Sabouraud's agar for candidal growth. Results were scored according to the yield of CFU. Diabetic rats given a single lingual inoculation of C. albicans remained positive for the yeast throughout the 40 weeks, in contrast to three other control groups, namely, nondiabetic rats inoculated with C. albicans, normal rats, and diabetic rats without C. albicans inoculation. Moreover, these controls were devoid of the histologic changes seen in the test group that were consistent with long-term mycotic lesions of the lingual mucosa, such as loss of filiform papillae, parakeratosis, irregular thickening, and a diffuse lymphocytic infiltration of the deeper layers of the epithelium. Although this model appears useful for investigating diabetes-induced oral candidiasis, no other researcher to our knowledge has exploited in full the etiopathology of this condition using the Wistar rat.

Pathologic changes in salivary glands due to diseases such as Sjögren's syndrome, cytotoxic therapy, and irradiation may lead to reduced salivary flow and xerostomia (102, 103, 109, 162, 191). Oral candidiasis is a common manifestation of xerostomia, which also promotes chronic candidal colonization (101). The relationship between xerostomia and oral candidiasis was investigated in the monkey model (133), as described earlier in this review. Jorge et al. (83) used Wistar rats to further study this phenomenon. They rendered 20 Wistar rats xerostomic by surgical removal of the major salivary glands (parotid, sublingual, and submandibular) and orally inoculated them with C. albicans three times a week for 32 weeks. When the rats were sacrificed and examined, candidiasis and hyphal infiltration of the lingual mucosa were found in 70% of the sialoadenectomized animals compared with 20% of the controls, confirming the critical importance of saliva and salivary flow in preventing oral candidiasis.

The vast majority of workers to date have resorted to the unnatural use of antimicrobials to eradicate the antagonistic population pressure of the commensal oral flora and thus initiate oral candidal colonization (Table 1). Since the broad-spectrum antibiotics used, such as tetracycline, adversely affect the immune response, a model that obviates the use of antimicrobials is a desirable alternative to mimic the clinical status. Jorge et al. (84) claim that the sialoadenectomized Wistar rat fits this requirement, since they noted 100% oral Candida carriage in xerostomic rats (after consecutive once-weekly oral inoculation for 5 weeks) compared with 50% carriage in controls. Candida was totally eradicated from the latter group within 18 weeks, whereas 66.6% of sialoadenectomized rats continued to harbor the yeasts. This model therefore appears suitable for the investigation of oral candidiasis since it maintains the normal oral flora with its competitive, colonization pressure akin to the clinical conditions in humans. However, since no researcher thus far has substantiated the claims of these authors, further studies with the sialoadenectomized Wistar rat models are urgently warranted.

As can be seen above, the Wistar rat model has been used by a number of workers to induce experimental oral candidiasis. However another species, the SD rat, has been more extensively used by others and appears to be the most popular model by far for the study of mucosal candidiasis (2-5). Jones and Russell, who pioneered Candida studies with the SD model, demonstrated that animals that succumb to infection show histologic changes similar to chronic candidiasis of the posterior dorsum of the human tongue (80). Furthermore, they have also shown ultrastructurally that C. albicans changes into the mycelial phase and penetrates the cornified layer of the rat lingual epithelium as in humans (81). The following studies with the SD model have contributed significantly to our understanding of the host-fungus interactions in oral candidiasis.

In early investigations, Bowen and Cornick (25) demonstrated that a carbohydrate-rich diet (CRD) positively encourages the oral carriage of C. albicans in SD rats. A number of workers have confirmed this finding using both in vitro and in vivo studies and have elucidated the role of dietary carbohydrates in the pathogenesis of oral candidiasis (66, 91, 154, 165). Therefore, Russell and Jones (154) tested the effect of a CRD containing 42% powdered icing sugar and 30% starch against the "standard 3/8" mouse and rat diet on the oral carriage of C. albicans. In experiments with both the mycelial- and yeast-phase C. albicans, the organism was recovered more frequently from the CRD-fed animals than from the controls on a normal diet. It was also observed that oral infection, noted as mycelial penetration of the superficial epithelial layers, was more frequent in CRD-fed animals.

Armed with this information and the effect of broad-spectrum antibiotics on the genesis of candidal infection, Russell and Jones (155) further studied the effect of both tetracycline-laced drinking water and a CRD on disease progression. The persistence of Candida in the mouth of each rat was examined by swabbing the tongue and mucosal surfaces throughout the experimental period. The oral yeast carriage was monitored semiquantitatively by counting the number of oral swabs positive for C. albicans (and not by quantifying the yeast growth). Tetracycline administration resulted in oral persistence of C. albicans in all rats over a period of 24 days. The prolonged carriage induced by a tetracycline-laced diet was far superior to that achieved by feeding a CRD alone. Furthermore, increased frequency and severity of lingual infection were seen in tetracycline-fed rats compared with the controls. For instance, in contrast to controls, the tetracycline-fed rats sacrificed on day 13 demonstrated lingual, gingival, and buccal mucosal infections. The posterior of the oral cavity was affected more than the anterior (Fig. 2), and candidal infection was seen with pseudomembranes, mimicking human oral candidiasis. Histologically, mycelia penetrated the orthokeratotic epithelium, which also demonstrated inter- and intracellular edema, and there was a marked inflammatory cell infiltration of the corium. A concomitant increase in the severity of infection was also seen with prolongation of the experiment.

View larger version (63K): [in this window] [in a new window]   FIG. 2.   Macroscopic appearance of typical lesions observed on the dorsal surface of SD rat tongue infected with C. albicans after tetracycline and cyclophosphamide administration. Note the areas of hyperplasia or leukoplakia (arrows) and the conical papillae appearing as a crescent in between.

The onset of thrush in neonates is usually seen 4 days after birth (129), and the predisposing conditions are thought to be immature immune defenses, antibiotic therapy, maternal cross-infection, and cross-infection from nursery staff (166). Jones and Russell (79) explored the importance of these host factors, including infancy, leading to the transition of C. albicans from saprophytism to parasitism. When infant (12-day-old) SD rats were inoculated with the yeast form of C. albicans (without tetracycline and a CRD), they were unable to demonstrate either candidal carriage or infection. However, inoculations of the mycelial form produced histologically demonstrable infection after 15 days. This study suggests that infancy per se may not be a predisposing factor in the initiation of candidal infection and reaffirmed the generally held belief that the mycelial form of C. albicans is more pathogenic than the yeast form. Since the authors did not include a control adult group of rats in this study, the question of high oral carriage of Candida due to infancy per se remains unresolved.

After these preliminary experiments, Russell and Jones (156) studied the effect of prolonged candidal inoculation and tetracycline treatment on murine oral infection. This experiment is perhaps the most extensive animal study to date, conducted over a period of 12 months with 60 rats to identify oral histologic changes that were wholly due to candidal infection. The oral carriage of C. albicans after fortnightly inoculation and recorded at 1-, 3-, 6-, 9-, and 12-month intervals was 58.6, 48.3, 38.3, 40.0, and 45.0% respectively. These results were rather disappointing since the animals were inoculated regularly. Nonetheless, the authors recorded in detail the lingual histologic changes and, after 21 weeks, observed a loss of papillae together with flat-surfaced hyper- or parakeratotic stratified squamous epithelium. Changes in the deeper layers included a mononuclear cell infiltrate of the corium, degenerative changes of superficial muscle cells with a giant cell reaction, and sarcolemmal proliferation and perivascular inflammatory infiltrate in deep muscle layers. These observations tended to suggest that the candidal infestation, though restricted to the superficial cornified epithelium, may also produce pronounced histologic changes in the deeper corium and the underlying muscle. This was the first observation in an animal model that Candida may elicit pathologic effects in subjacent tissues in addition to the immediate viscinity of hyphal infiltration. Other studies have now confirmed that candidal extracellular enzymes, such as secreted aspartyl proteinases and phospholipases, may account for such effects (21, 144).

Since the administration of tetracycline encourages the oral carriage of C. albicans in SD rats, it was postulated that a germ-free gnotobiote would be ideal for the study of oral candidiasis. To test this hypothesis, Jones et al. (82) compared the oral carriage of C. albicans in germ-free and conventional (specific-pathogen-free) SD rats with and without tetracycline treatment. The mouth and the rectum of the rats were swabbed and the swabs were cultured for Candida in Sabouraud's agar at the beginning of the experiment, before inoculation, and at regular intervals afterwards, and the number of positive swabs was recorded. The authors observed that the oral cavity of all germ-free animals, whether treated with tetracycline or not, remained colonized with C. albicans until the end of the experimental period. In contrast, only 50% of the tetracycline-free, as opposed to 85% of the tetracycline-treated, conventional rats harbored C. albicans, reconfirming that the antibiotic does favor oral yeast carriage (P < 0.05). Infection was clearly evident in both the germ-free and conventional rats as mycelial penetration of the cornified epithelium, particularly the dorsal lingual surface.

Further experiments by Jones et al. (82) reconfirmed that (i) germ-free animals can remain colonized for up to 19 weeks with or without receiving tetracycline and (ii) colonization in conventional rats receiving tetracycline is longer lasting than in those without the antibiotic (P < 0.01). Interestingly, they found no evidence of oral infection, as opposed to superficial infestation, in any of the conventional rats whereas they saw histologic evidence of infection in gnotobiotes. Contradictory findings on infectivity have been reported by others using conventional rats (157). The latter group tested the effect of different schedules of tetracycline administration in two groups of SD rats (60 in each group) that were either maintained on tetracycline throughout the experimental period of 22 weeks or given the drug only during the first fortnight. All animals were inoculated with C. albicans orally on three alternate days in the second week. The results showed that initial administration of tetracycline fosters long-term oral candidal colonization with no significant difference in the incidence of infection.

Further studies by Hassan et al. (66) have shown that oral carriage of C. albicans of SD rats rapidly diminished when animals were fed a normal diet free of tetracycline and given only an initial inoculum of the yeast at the beginning of the experiment, compared with the following combinations (i) CRD, (ii) normal diet and tetracycline, and (iii) CRD and tetracycline. Significant differences in the recovery of C. albicans between the last three groups of rats were maintained irrespective of whether the inoculum was continuous or given only once at the beginning of the experiment.

To conclude, the foregoing studies indicate that a combination of tetracycline treatment and a CRD favor the oral carriage of C. albicans in SD rats regardless of whether an adequate inoculum of the challenge strain is administered once or on several occasions. It should, however, be noted that at least one group has found that tetracycline exposure is not prerequisite for oral Candida colonization provided that the SD rats are infected with a mucosally virulent strain of C. albicans. Allen et al. (5) studied the development of oral candidiasis in test and control groups of 20 SD rats each, receiving tetracyclinelaced and drug-free water, respectively. The animals were all inoculated with a mucosally pathogenic strain of C. albicans that was noted to produce infection in 80% of the animals in an earlier study (141). There were no significant differences in C. albicans carriage rate in the two groups, and after 20 weeks grossly visible lesions were seen in 50 to 55% of both the test and control groups. Nonetheless, the lesions in the tetracycline-treated group were significantly larger than those in the controls (P < 0.05). These results suggested that the establishment of yeast infection is not necessarily dictated by antibiotic exposure. However, the degree and severity of infection are likely to be related to the synergistic effect of tetracycline and the virulence of the infecting strain.

Van Wyk et al. (199) also investigated the possibility of using germ-free SD rats as a model for oral candidiasis. They observed that the daily inoculation (10⁶ CFU) of C. albicans in drinking water for 14 days was adequate to infect the oral epithelium of the gnotobiotes sufficiently to produce epithelial changes such as acanthosis, loss of papillae, and a chronic inflammatory infiltrate of the lamina propria. In a second experiment, they investigated the chronological events leading to oral candidiasis in germ-free animals supplied with C. albicans-laced drinking water, over a period of 36 days. Invading yeasts were seen in the superficial lingual, palatal, and cheek epithelium within 72 h but were scanty after day 8. The resulting lesions were pronounced from 72 h to 6 days and resolved after day 15. This implied that host immunity to the invading pathogen was the major force in eliminating the infection in the SD model. They also proposed that the genetic differences among the rats may result in variant oral lesions and that genetically homogeneous inbred animals should be used to reproduce similar lesions.

An animal model that resembles human oral candidiasis is of value not only for studying the pathogenesis of the disease and the virulence of the organism but also for evaluating new antifungals which are introduced from time to time. The SD rat model has therefore been evaluated for antifungal drug testing by a few workers. Walrath and Blozis (J. Dent. Res. Spec. Issue, abstr. 975, 1986) produced clearly visible hyperkeratotic tongue lesions in female SD rats using tetracycline-laced drinking water and oral inoculation of C. albicans. (The authors observed continuous oral carriage for up to 8 months and tongue lesions for 7 months.) The rats were then treated with food pellets laced with the antifungal clotrimazole, and the lingual lesions resolved within 1 week. Hence, the authors suggested that the SD rat model was a satisfactory in vivo tool to study the effect of antifungals in the management of chronic oral candidiasis.

A study was also designed by Allen et al. (8) to investigate the effect of ketoconazole on lingual candidal infection in SD rats. Two control and two test groups of animals were orally inoculated with a C. albicans isolate known to produce mucosal lesions. Several animals in the test and control groups developed lingual candidal lesions (9 of 20 and 8 of 20) respectively. All lesions in the test group of animals treated with ketoconazole resolved, while 2 of 9 animals in the untreated control group also showed spontaneous resolution of lesions. These findings confirmed that the observed leukoplakic lesions were indeed caused by the Candida inoculum and that the SD model was suitable for testing antifungal therapy. However, it should be borne in mind that natural resolution of the lesions is common with this model and that appropriate controls need to be used to obviate spurious results. Despite the availability of this satisfactory model, very few workers appeared to have ventured into studies of the efficacy of newer antifungals using the SD rat model.

There is a comprehensive body of data on the oral histopathology of Candida infection in SD rats, and these are discussed below. A number of workers have reported that most candidal lesions are concentrated on the posterior midline dorsum of the tongue (6). In studies by Allen et al. (6), hyphae were present in the parakeratotic layer, together with chronic inflammation of the underlying connective tissues. The authors speculated that the topography of the conical papilla region might favor the retention of yeasts in the interpapillary crevices and thus provide them with an increased opportunity to invade the epithelium. Hence, it would seem that the surface architecture of the mucosa plays a role in selective candidal colonization, a view that has been echoed by Fisker et al. (54) and Philipsen et al. (140). Further investigations were performed by Fisker et al. (55), who subjected SD rats to prolonged oral candidiasis to localize the infection foci and evaluate the mucosal response. Candidal infection confirmed by histologic examination was observed in 15 of 60 animals; the majority of the infective foci were localized in the buccal sulcular folds, the gingival margin, the cheek, and the interpapillary area of the tongue. These areas accounted for 92.2% of the infective foci, and the remainder were in densely keratinized attached gingivae and palatal epithelium. A noteworthy observation was the apparent similarity of the tongue lesions and the histologic features of human median rhomboid glossitis (40, 207). The mucosal response in median rhomboid glossitis comprises an inflammatory reaction without degenerative changes in the subepithelial tissues.

Other histopathologic features of experimental oral candidiasis in animal models that have been documented thus far include increased epithelial mitotic activity, epithelial proliferation leading to hyperplasia (Fig. 3), and rapid desquamation of the oral mucosa. The factors causing enhanced epithelial proliferation are not clear, although they could involve either host immune mediators or enzymes or metabolites released by the organism. To determine the impact of the latter attributes of Candida on epithelial cell turnover, Reed et al. (145) injected yeast-free culture supernatants into the buccal epithelium of young adult SD rats and assessed the mitotic activity by using a metaphase arrest technique at 11 and 31 h. They observed a significant rise in mitotic activity 31 h after injection of a 5-h culture supernatant of C. albicans, indicating that extracellular products of the yeast may induce proliferation of the buccal epithelium. Although they postulated that a range of candidal products such as hydrolytic enzymes and cell wall polysaccharides may adversely affect the epithelial and connective tissue cell turnover, they have performed no further experiments to substantiate these assertions.

View larger version (117K): [in this window] [in a new window]   FIG. 3.   Photomicrograph of a biopsy specimen illustrating the hyperplastic epithelial response to candidal invasion of the rat oral mucosa. Note the hyphal elements in the superficial layers. Periodic acid-Schiff stain. Magnification, ×40.

It has been stated that the different incidences of infection seen in different animal models, and even within the same model, could be due to strain variations and the related virulent attributes of C. albicans. Hence, studies have been conducted to elicit differences in pathogenic traits among C. albicans isolates (3). In one study, four groups of SD rats (with 10 animals per group) were orally inoculated weekly for 25 weeks with four disparate strains of C. albicans. Of these, oral candidal carriage of various degrees was seen in three groups while the fourth group was completely devoid of infection. Further, it was observed that some strains exhibited consistently high colony counts while others invariably produced low colony counts. In addition, histologic evidence of infection was observed only in two of the three groups exhibiting candidal carriage and only 4 of 10 (40%) and 2 of 10 (20%) rats in each such group exhibited characteristic candidal lesions of the lingual mucosa. Allen and Beck (4) extended their experiments with 16 strains of C. albicans and an experimental period of 16 weeks without tetracycline supplements. They demonstrated significant variations in the oral recovery of Candida, ranging from 0 to 65%, and intraspecies differences in pathogenicity in terms of lingual infection, yielding results consistent with their earlier study (3).

Allen et al. (7) also observed that a single oral inoculation with a mucosally virulent strain of C. albicans without the help of any antibiotics or immunosuppressive agents was adequate to induce dorsal tongue lesions in SD rats. In a study of 210 animals sequentially killed over a 20-week period to follow up the clinical evolution of the lesions, they observed the most extensive epithelial changes, such as papillary atrophy and the destruction of dorsal lingual papillae, during the weeks 2 and 3 of infection (Fig. 4). Between 4 and 20 weeks, the percentage of animals with clinically evident lesions ranged from 10 to 30% although after week 18 all tongue lesions had been resolved. These observations and the histopathology concurred well with those seen in human mucosal lesions in candidal infections while reaffirming that Candida is an opportunistic pathogen easily overcome by innate body defenses. It should be noted that this is one of the few experiments described in the literature where the authors were able to induce infection without antibiotic or carbohydrate supplements in the food.

View larger version (126K): [in this window] [in a new window]   FIG. 4.   Scanning electron micrographs of a depapillated lingual lesion observed in experimental candidiasis in a rat, resembling erythematous candidiasis of humans (magnification, ×38) (A) and the lesion showing a hyphal element of C. albicans penetrating the epithelium, together with a multitude of commensal bacteria (magnification, ×2,400) (B). Photographs provided courtesy of Carl M. Allen, College of Dentistry, Ohio State University.

The same workers noted the distinct strain-related patterns of C. albicans infections on the dorsal lingual mucosa of immunocompetent rats (3, 4). They showed that while some isolates produced lesions particularly on the posterior-dorsal lingual mucosa accompanied by flattening of the normal papillary architecture of the epithelium, another group failed to produce any mucosal lesions. Allen et al. (9) further evaluated a lesion-inducing isolate and a nonpathogenic isolate by using both normal and cyclosporin-immunosuppressed rats. The lesion-inducing isolate showed a significantly increased rate of infection in normal as well as cyclosporin-treated rats compared with the nonpathogenic strains.

An association between oral candidiasis and iron deficiency has been documented by a number of investigators (34, 56, 159). The SD rat model has an added advantage for use in studies of this relationship, since there are several dietary methods for producing iron deficiency in these rats (10, 112). In one such study, Rennie et al. (147) demonstrated that iron deficiency may not necessarily predispose SD rats to oral candidiasis since some malnourished animals did not acquire the infection. Further, they observed a reduced capacity of anemic rats to recover from candidal infection. Similar results have been reported by Sofaer et al. (185) using anemic mice (see below).

The protective role of saliva in preventing oral candidiasis has been studied previously in the Wistar rat model (83, 84). To further investigate this in SD models, hyposalivatory rats have been used (119). The latter authors conducted a series of studies and observed that all desalivated rats were susceptible to C. albicans infection and that the oral carriage in the infected animals was 30-fold greater than that in the normal control animals (i.e., 3.8 × 10⁵ and 1.1 × 10⁴ CFU, respectively) (P < 0.05). The importance of an intact salivary response in preventing C. albicans infection was clearly shown when transmission of infection from one desalivated animal to its counterpart occurred in 1.2 days while transfer from a normal donor to a recipient took 4.3 days. Interestingly, in the hyposalivatory model of oral candidiasis, pretreatment with tetracycline was unnecessary to initiate infection. The authors therefore concluded that the hyposalivatory-rat model is useful to assess the infectivity, pathogenesis, and virulence of different Candida strains in both qualitative and quantitative terms.

A relationship between Candida infection and mucosal trauma has been addressed by other workers (30, 131, 195). When the role of thermal trauma to the oral mucosa was investigated, it was noted that thermal ulceration facilitated candidal invasion of the dorsal lingual mucosa (131). These results also suggested that mild, long-term trauma due to chronic irritation from unstable dentures may contribute to the initiation or aggravation of Candida-associated denture stomatitis. The authors further postulated that the inflammatory exudate consequential to trauma might enhance the adhesion of the yeasts and thus facilitate infection.

One major host factor preventing fungal infections in general and oral candidiasis in particular is the cell-mediated limb of the immune system. The number of patients with immunological problems and hence susceptible to candidiasis is increasing in the community and includes those undergoing organ transplantation, those undergoing cancer therapy, and HIV-infected individuals (38). Recent experimental studies to investigate the impact of immunosuppression on mucosal candidiasis have been carried out by Samaranayake et al. (172), using both C. albicans and C. krusei. They noted that oral colonization by C. albicans was 12-fold greater than that of C. krusei prior to immunosuppression during an initial experimental period of 21 weeks with weekly inoculation of organisms. However, none of the animals succumbed to candidal infection. This was confirmed by histopathologic studies of a few selected animals within each group. However, when the animals were immunosuppressed with cyclophosphamide, the leukocyte counts of all the animals were significantly depressed and both Candida species produced histopathologic changes on the lingual mucosa characteristic of mucosal candidiasis (Fig. 5). C. albicans produced 100% infection in animals (three of three), while only 25 to 30% infection was observed with two different C. krusei isolates. Both species produced fungal hyphae that penetrated the lingual epithelium and stopped short of the prickle cell layer. However, the C. albicans hyphae penetrating the lingual mucosa were longer than C. krusei hyphae (17 and 8 µm, respectively) and tended to be relatively more profuse (Fig. 6). The results of this study substantiated that (i) immunosuppresion facilitates candidal infection and (ii) C. krusei i