In this study we compared the pathogenicity of oral and systemic Candida isolates. The pathogenesis of diverse candidal diseases depends upon both generalized virulence factors and those that function in specific environments dictated by immune function, tissue site and other host factors. The data presented here demonstrate that the pathogenicity of oral Candida isolates is similar to systemic Candida isolates, suggesting that the pathogenicity of Candida is not correlated with the infected site.
The pathogenesis of both oral and systemic candidiasis is closely dictated by properties of the yeast biofilms [28, 29]. Implanted devices, such as venous catheters or dental prosthesis, are a serious risk factor for Candida infections. They are substrates for the formation of biofilm, which in turn serve as reservoirs of cells to continually seed an infection . It has been estimated that at least 65% of all human infectious are related to microbial biofilms [30, 31].
A variety of methods have recently been used for the quantification of Candida biofilm on different substrata. These include counting of colony forming units (CFU), dry-weight assays, spectrophotometric analysis, and colorimetric assays, such as 2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide (XTT) reduction assay. However, each method carries its own advantages and limitations [7, 32, 33]. In our study, we used a dry-weight assay because this method allows the single quantification of a Candida biofilm on a clinically relevant substrate such as silicone and acrylic resin. Silicone is frequently used in the manufacture of medical devices and catheters and it is related to development of systemic candidiasis in hospitalized patients. Acrylic resin (methyl methacrylate) is a material widely used in preparation of dental prosthesis and it has significance for development of oral candidiasis.
Among all isolates tested in this study, the quantity of biofilm mass varied according to the Candida species. C. albicans and C. dubliniensis were the highest biofilm producers on silicone pads, followed by C. tropicalis, C. norvegensis, C. parapsilosis, C. glabrata, C. krusei, C. lusitaniae, and C. kefyr. Most studies have shown that the biofilm formation by clinical isolates of Candida was species dependent and generally the highest levels of biofilm formation were observed in C. albicans and the lowest in C. glabrata [5, 20]. Notably, unlike C. albicans and other Candida species, C. glabrata is unable to generate filamentous forms which may contribute to the impared ability of C. glabrata to form a biofilm . The observations for higher quantities of biofilm production by C. albicans and lower biofilm production from the non filamenting C. glabrata, given the same standards of in vitro test conditions, remained true for the clinical isolates from our study. Indeed, for both strains collected orally or systemically, there was very little in the way of quantity or quality of biofilm production for C. glabrata. C. albicans produced the greatest quantity of biofilm regardless of the adhesion platform material or whether it was isolated from oral or invasive infection sites.
Interestingly, the differences in biofilm formation among Candida species on acrylic resin were less significant than biofilm formed on silicone. This fact may be attributed to the methodology used which was previously developed for biofilm formation on silicone pads [23, 24]. The process of candidal adhesion to acylic resins is complex. Previous studies have shown that a number of factors including the nutrient source, the sugar used for growth (glucose or sucrose), and the formation of pellicules from saliva or serum may influence the adhesion and colonization of Candida [7, 29].
We also used an in vivo G. mellonella infection model to evaluate the pathogenicity of oral and systemic Candida isolates. There are some benefits to using G. mellonella larvae as a model host to study Candida compare to other invertebrate models. For example, the larvae can be maintained at a temperature range from 25°C to 37°C, thus facilitating a number of temperature conditions under which fungi exist in either natural environmental niches or mammalian hosts. High temperatures can be prohibitive for the growth of C. elegans or Drosophila infection models. Our study used 37°C to mimic mammalian infection systems. G. mellonella also has the benefit of facile inoculation methods either by injection or topical application, where injection inoculation provides a means to deliver a precise amount of fungal cells [12, 27, 34]. By contrast, other systems, such as C. elegans, require infection through ingesting the pathogen. Since we included both albicans and non-albicans strains in our study we thought it prudent to use a model that ensured equal pathogen delivery rather than a model that would have an aversion to consuming some of the infecting agents.
As with the biofilm assays, the virulence levels of Candida isolates in G. mellonella were dependent on the species studied. Surprisingly, within the same species, oral isolates were as virulent as isolates from candidemia, the most common severe Candida infection. Previously, Cotter et al.  reported that it is possible to distinguish between different levels of pathogenicity within the genus Candida using G. mellonella larvae. We observed that G. mellonella showed mortality rates of 100% after injection with 105 cells of C. albicans, C. dubliniensis, C. tropicalis, and C. parapsilosis, 87% with C. lusitaniae, 37% with C. novergensis, 25% with C. krusei, 20% with C. glabrata, and 12% with C. kefyr over a 96 hour period of incubation at 37°C. Cotter et al.  verified mortality rates of 90% for C. albicans, 70% for C. tropicalis, 45% for C. parapsilosis, 20% for C. krusei, and 0% for C. glabrata over a 72 hour period of incubation at 30°C after the injection with 106 cells of each Candida species. Probably, the virulence of the Candida strains in G. mellonella tested in this study were higher than the virulence of Candida strains observed by Cotter et al.  because of the difference of incubation temperature used. The temperature variations can affect gene expression and consequently the level of virulence of Candida strains .
Of note is that this is the first study to inoculate species of C. lusitaniae, C. norvegensis and C. dubliniensis in the G. mellonella model. Single isolates for C. lustaniae and C. norvegensis and two isolates of C. dubliniensis were included in our study. C. lusitaniae is considered an emerging non-albicans Candida species and isolates show resistance to amphotericin B. C. norvegensis appears to be a rare cause of human infection and the most of the isolates are resistant to fluconazole [36, 37]. There are limited data on the comparative virulence of C. lusitaniae and C. norvegensis in relation to C. albicans. In this study, C. lusitaniae and C. norvegensis were less virulent in G. mellonella than C. albicans.
Finally, in our study, C. dubliniensis isolates showed that the ability of biofilm formation and killing G. mellonella was similar to C. albicans. C. dubliniensis has been implicated in oropharyngeal candidiasis in HIV-infected patients, althought it has also been isolated from other anatomical sites, including lungs, vagina, blood, and feces [38, 39]. Despite the significant phenotypic and genotypic similarities shared between C. albicans and C. dubliniensis, the comparative virulence of the two species is clearly a very complex topic [40, 41]. Borecká-Melkusová  verified that the biofilm formation in C. albicans was significantly lower than in C. dubliniensis, and Koga-Ito et al.  observed that the survival rate and dissemination capacity of C. dubliniensis in mice were lower than C. albicans.