Sherry, Leighann (2014) Evaluating Candida albicans biofilm formation and novel antifungal treatment. PhD thesis, University of Glasgow.Full text available as:
Printed Thesis Information: http://encore.lib.gla.ac.uk/iii/encore/record/C__Rb3084848
Candida biofilms have become an increasingly important clinical problem. The widespread use of antibiotics, frequent use of indwelling medical devices, and a trend towards increased patient immuno-suppression has resulted in a creation of opportunity for clinically important yeasts to form biofilms. Whilst there is growing evidence of the importance of Candida biofilms in clinical medicine, not all clinical isolates are able to form biofilms. There is therefore a fundamental gap in understanding exactly what drives biofilm formation and its clinical implications. These structures have become increasingly recognised as a significant clinical problem. One of the major reasons behind this is the impact that these have upon treatment, as antifungal therapy often fails and surgical intervention is required. This places a large financial burden on health care providers. Therefore, the discovery of alternative antifungal agents to be used in the treatment of fungal biofilms is in great demand for the management of these infections. A panel of Candida albicans bloodstream isolates were assessed for their biofilm forming ability by using the crystal violet assay and measuring cellular surface hydrophobicity. Scanning electron microscopy was used to visualise differences in the clinical biofilms. The impact of amphotericin B (AMB) treatment was determined next by broth microdilution method to assess differences in susceptibility profiles of the clinical isolates. The virulence of these clinical isolates was evaluated in vivo using a Galleria mellonella model and transcriptional analysis used to assess the expression of various genes associated with C. albicans biofilm formation within clinical isolates. Extracellular DNA (eDNA) in clinical biofilms was quantified using a microplate fluorescence assay and chitinase activity measured using a biochemical assay. Moreover, the potential of a novel antimicrobial agent Carbohydrate-derived fulvic acid (CHD-FA) was assessed against a panel of fungal and bacterial species. The mechanism of action of CHD-FA was determined using membrane assays include ATP release, and propidium iodide fluorescence, with various inhibitors used to determine whether CHD-FA activity is affected by known resistance mechanisms. Finally, the immunomodulatory properties of CHD-FA were investigated using ELISA and PCR arrays. The results from this study have shown C. albicans biofilm formation is differential within clinical isolates, where those with high biofilm formation (HBF) predominately consisted of hyphal cells, were more virulent in vivo and had decreased susceptibility to AMB, when compared to those with low biofilm formation (LBF). Furthermore, transcriptional analysis identified a number of genes that positively correlated with C. albicans biofilm formation. The novel agent carbohydrate-derived fulvic acid (CHD-FA) was shown to not only be highly active against C. albicans biofilms, but also against a range or orally relevant bacteria through non-specific membrane activity. Furthermore, CHD-FA was shown to down-regulate a number of pro-inflammatory mediators in an oral epithelial cell line. In conclusion, this study has characterised C. albicans clinical isolates based on their biological characteristics, where clear difference in virulence and antifungal treatment have been shown. It may be possible to develop a panel of genetic markers that could be used as a diagnostic tool for detecting biofilm formation in clinical isolates. CHD-FA is a microbiocidal compound that may serve as a potential novel antiseptic agent for the treatment of oral candidiasis and other candidal biofilm infections, whereby the immunomodulatory properties of CHD-FA could be exploited for controlling inflammation in a number of diseases.
Actions (login required)
2. How Do We Investigate Biofilm Formation?
The key driver in understanding and evaluating biofilm formation from important Candida species lies in the quantitative methods utilized. When screening large collections of clinical isolates from different patient cohorts, several experimental strategies have been utilized, predominantly quantifying biomass using dry weight, stains such as crystal violet, and the metabolic dye XTT . Each technique has their own benefits and caveats, but caution must be taken when interpreting the data achieved from each assay, particularly when correlating it to clinical outcomes. Given the heterogeneity found between strains, alongside varying laboratory models and techniques, standardization becomes problematic. For example, two of the most commonly used media for biofilm formation are Roswell Park Memorial Institute (RPMI) media and Spider media. Studies have identified that RPMI is more supportive of biofilm formation, stimulating biofilms that are three times thicker than Spider media . Furthermore, these media are not physiologically relevant, with several studies employing more biologically relevant conditions for biofilm formation through use of artificial saliva, urine, and mammalian serum [4,5,6]. One of the most commonly used bioassays is the sodium salt XTT (2,3-bis(2-methoxy-4-nitro-5-sulfo-phenyl)-2H-tetrazolium-5-carboxanilide) [7,8]. This biofilm assay is highly reproducible and allows for a high throughput of multiple microtiter plates without compromising accuracy. Its usefulness comes with susceptibility testing, allowing for the direct comparison of antifungal treated samples compared to an untreated control . Given the metabolic variation observed between both different strains and species, caution must be taken when interpreting the assay, as a measurement for biofilm development may simply be a reflection of high cell numbers [10,11]. For example, scant biofilms of non-albicans yeasts may show a high XTT value, yet minimal biomass is present. Therefore, the output achieved from XTT is only cellular viability and it does not take into account other biofilm components such as the extracellular matrix (ECM), which are arguably the most important when it comes to biofilms .
Another commonly used assay for biofilm formation is crystal violet staining. This method provides the total quantification of the biofilm biomass (cells and ECM) and also allows for rapid, high-throughput processing of multiple samples. However, variability of the washing step can result in both over- and under-estimation of biomass, with the assay also unable to differentiate subtle differences between samples . An interesting example of this was described in a recent study, where these techniques were used to stratify the ability of Candida bloodstream isolates to form biofilms . There was no evident standard for their stratification to denote strains as biofilm or non-biofilm formers, with a crystal violet values of OD570 > 0.09 simply denoted as a biofilm former. By doing so, it was concluded that non-Candida albicans species (NCAS) form greater biofilms than C. albicans, and that biofilm formation does not correlate to clinical outcomes. This is contrary to a wealth of previous literature, whereby the ability of Candida isolates to form a biofilm does associate with mortality [14,15,16,17].
Discrepancies between these findings illustrates the necessity for standardised testing to elucidate biofilm-related risk factors. Our group has taken a “belt and braces” approach, using a combinational approach of crystal violet, XTT, and SYTO®9 fluorescence quantitative biofilm assays (Thermo Fisher Scientific, Paisley, UK). Here, significant correlations were observed for C. albicans biofilm formation, which was subsequently used to stratify biofilm-forming ability . Irrespective of the particular quantitative approach, wide-spread biofilm heterogeneity is observed within different clinical panels of isolates [13,14,18,19]. Collectively, these data suggest that different Candida strains function differently, and that consideration should be given to the individual isolates as we try and understand their clinical importance with respect to antifungal resistance and pathogenic potential.
3. Is Heterogeneity Clinically Important?
Since the earliest descriptions of Candida biofilms, great strides have been made to unequivocally demonstrate their clinical significance, despite perceived contention in the field. Throughout the human host, Candida biofilms colonize a wide variety of anatomical locations, as shown in Table 1. The oral and vaginal epithelium provide a mucosal niche for biofilm formation, whilst indwelling medical devices such as prosthetic heart valves and central venous catheters provide an inert, abiotic substrate for subsequent biofilm adherence and proliferation [20,21]. Irrespective of isolation site, biofilm heterogeneity has been reported, including the oral cavity, bloodstream, and urinary tract [14,22,23,24,25,26].
Within a clinical setting, intravascular catheters provide an optimal environment for Candida spp., allowing for the development and maturation of biofilms to which cells can disperse and subsequently cause candidaemia. Dispersed biofilm cells have been shown to be more pathogenic than their planktonic counterparts, exhibiting greater cytotoxicity and virulence in vivo . Therefore, the role of the biofilm phenotype has potentially profound implications within the clinical environment. An initial study from Tumbarello and colleagues (2007)  aimed to identify the top risk factors associated with mortality rates in candidaemia patients. Using multivariate analysis, they were able to distinguish inadequate antifungal therapy (odds ratio (OR) 2.36, p = 0.03), APACHE III (OR 1.03, p < 0.001), and overall biofilm-forming Candida species (OR 2.33, p < 0.007) as significant variables associated with mortality . When scrutinized at the Candida species level, only C. albicans (OR 3.97, p < 0.001) and C. parapsilosis (OR 4.16, p = 0.03) were shown to significantly correlate to biofilm-based mortality. A follow up study subsequently identified that central venous and urinary catheters, use of total parenteral nutrition, and diabetes mellitus as independent entities of bloodstream infections caused by biofilm forming isolates . Furthermore, they demonstrated the potential economic burden of these isolates resulting from increased lengths of hospital stays and use of antifungals and ultimately resulted in an increased possibility of mortality . A more recent, prospective analysis subsequently identified line removal (p = 0.032) as a significant risk factor associated with mortality rates from a candidaemia patient cohort, with the removal of an indwelling line correlating with a more positive patient outcome . Interestingly, when this was then subsequently assessed at Candida species level, survival analysis demonstrated significantly higher survival rates for patients with C. albicans associated line removal compared to no removal, with no differences observed in NCAS . Furthermore, Tascini and colleagues used a random forest model of analysis to cluster candidaemia associated mortality and to identify its accompanying risk factors . It was shown that azole use and high APACHE II, as well as biofilm formation, significantly correlated with the highest mortality group . Published guidelines have suggested that catheter-related bloodstream infections should result in the direct removal of such devices, if possible [53,54,55]. Furthermore, a meta-analysis of seven clinical trials revealed that the removal of central venous catheters significantly correlated with reduced mortality rates (OR 0.50, p < 0.001) . Conversely, a study assessed the efficacy of catheter removal within 24 h to 48 h of antifungal therapy and demonstrated no clinical improvement. This study, however, looked at echinocandins and liposomal amphotericin B, two highly active Candida biofilm agents . What these studies do provide is an insight into differential responses to biofilm-active therapies, and suggest clinical isolates respond differently depending on their capacities to form biofilms. Despite the majority of studies focusing on the potential for Candida biofilms to develop on hard, abiotic surfaces, there are a variety of mucosal niches within the host to which Candida can colonise as a biofilm an induce tissue damage.
Key to successful colonisation and host damage to a mucosal niche is the secretion of various hydrolytic enzymes. These secreted proteins are a primary attribute within the virulence armamentarium of the organism allowing it to invade host tissue, and include proteinases, haemolysins, and phospholipase. Of these enzymes, the secreted aspartyl proteinases (Saps) are the most studied, comprising a family of ten genes (SAP1–10). The secretion of these enzymes has been attributed with disease, with high levels of expression observed from a variety of diseases including infections of the bloodstream, vagina, oral cavity, and diabetes mellitus [58,59,60]. Given the diversity of the Sap family, then differential expression of independent genes has been associated with varying anatomical location [59,61]. During biofilm formation, SAP5 is up-regulated, significantly correlating with biomass . Indeed, an integrated global substrate and proteomics approach identified SAP5 and SAP6 as the major biofilm-related proteases utilised by C. albicans. Manipulation of both of these genes resulted in decreased adhesion and impaired biofilm development both in vitro and in vivo, highlighting their role as potential biofilm biomarkers . Recent studies have identified a novel fungal toxin termed candidalysin, a hyphae-specific peptide critical for epithelial damage  and expression of the gene coding this toxin (ECE1) was shown to be highly up-regulated in C. albicans isolates capable of forming biofilms .
An area worthy of consideration for mucosal biofilm formation is vulvovaginal candidiasis (VVC). Although not life-threatening per se, this infection will affect up to 75% of women in their lives at least once and are one of the most common fungal infections globally . While the majority of these cases are sporadic and will clear after one episode, some women will emerge with persistent occurrences (>4 episodes a year), despite being completely asymptomatic between these episodes (recurrent VVC (RVVC)) . The reasoning for RVVC is multi-factorial, yet given that azoles are first line topical drug of choice and have widespread availability from over the counter, then inadequate therapy is extremely problematic. While biofilm formation is regarded as a pathogenic attribute of bacterial vaginosis, its role in RVVC remains equivocal, despite a growing body of evidence to suggest otherwise [23,67,68,69]. Candida biofilms have been shown to form on the vaginal mucosa in vivo, as well as on inert substrates such as intrauterine contraceptive devices [67,70]. A recent study from our group screened a cohort of 300 VVC isolates for their epidemiology, biofilm formation, and azole susceptibility . Interestingly, an epidemiological shift towards NCAS was observed, and that biofilm formation was heterogeneous between these isolates regardless of Candida species. For C. albicans, it was demonstrated that the planktonic MIC50 for fluconazole was 4 mg/L, yet when the susceptibility profile of these isolates was tested as biofilms, the MIC50 escalated to >32 mg/L. This highlights the role for the biofilm phenotype, and may go towards explaining the chronic phenotype in this patient cohort and irresponsiveness to treatment.
4. How Does Heterogeneity Impact Antifungal Treatment?
Antifungal resistance is a complex, multifactorial process to which can be induced in response to a compound or as an irreversible genetic alteration as a result of prolonged drug exposure. While resistant planktonic cells predominantly arise from inherited traits to maintain a resistant phenotype, biofilm resistance rises through mechanisms such as over-expression of target molecules, efflux pump activity, and through the protective barrier of the extracellular matrix (ECM) allowing limited diffusion. Undoubtedly, the most defining characteristic of biofilms is this intrinsic and adaptive recalcitrance to many antimicrobial therapies. Compared to their free-floating planktonic equivalents, up to 1000-fold higher concentrations of antifungal agents can be required to effectively kill Candida biofilms in vitro, with the same decreased sensitivities also observed in vivo [71,72].
Several clinical observations have associated the ability to form biofilms with mortality, but also with azole and inadequate antifungal use. Many studies have sub-categorised C. albicans isolates as low biofilm formers (LBF) and high biofilm formers (HBF) [14,73,74]. Phenotypically, biofilms formed by these isolates are distinct, with LBF existing predominantly as sparse populations of yeast cells and pseudohyphae, whereas HBF have a dense, tenacious hyphae based morphology (Figure 2