Candida albicans is a polymorphic yeast species that often forms part of the commensal gastrointestinal mycobiota of healthy humans. It is also an important opportunistic pathogen. A tripartite interaction involving C. albicans, the resident microbiota and host immunity maintains C. albicans in its commensal form. The influence of each of these factors on C. albicans carriage is considered herein, with particular focus on the mycobiota and the approaches used to study it, models of gastrointestinal colonization by C. albicans, the C. albicans genes and phenotypes that are necessary for commensalism and the host factors that influence C. albicans carriage.
GENERAL INTRODUCTION
Candida albicans is a commensal yeast species that is found in the gastrointestinal (GI) tracts of humans and other animals. Vertical transmission of this microorganism from mother to infant during birth (Waggoner-Fountain et al.1996; Blaschke-Hellmessen 1998) means that humans often form a life-long association with this species. Nevertheless, this host–microbe relationship is not always benign, and infections with C. albicans among ‘at risk’ cohorts bear considerable morbidity and mortality (Brown et al.2012). Thus, as might be expected for a prevalent pathobiont, the capacity of C. albicans for both commensalism and pathogenesis and the transition between these lifestyle options are the focus of much ongoing research. In this review, we discuss the fungal component of the GI microbiota of humans, with a particular focus on C. albicans and its interaction with the rest of the microbiota and its host. We outline the approaches used to study commensalism and dissemination by this yeast in animal models, focusing in particular, on the recent research describing C. albicans commensalism (without an eventual transition to infection) in the GI tract. We describe the various C. albicans morphologies and the genes required for commensalism in mouse models. We also discuss the host response to C. albicans in the GI tract, describing the immune players involved, and we detail the inconsistencies of the emergent research in this area.
THE FUNGAL COMPONENT OF THE GI MICROBIOTA—THE MYCOBIOTA
The complete human gut microbiota includes microorganisms from each of the principal microbial lineages: bacteria, archaea, eukaryota and viruses (including bacteriophage) (Norman, Handley and Virgin 2014). The term ‘microbiota’ however, while formally intended to describe the totality of the microorganisms present in a particular environment (Medical Subject Headings www.ncbi.nlm.nih.gov/mesh), has often in practice been used to refer exclusively to the bacterial component of any given ecosystem. The widespread use of bacterial 16S rRNA gene targets to describe and evaluate ‘microbiota’ (without a ‘bacterial’ qualifier) composition is testament to the more selective usage of this word in the scientific literature. In contrast, the term ‘mycobiota’, which is derived from the Greek words myco (fungus) and bios (life), can be unambiguously used to refer exclusively to the fungal community of any given habitat. Moreover, the term ‘microbiome’ has an indiscriminate and all-inclusive sense, and shotgun metagenomics studies of the human gut microbiome (the DNA of all organisms present in a particular niche) have also yielded some insight into the fungal community present.
Usage of ITS, the universal fungal DNA barcode for high-throughput amplicon sequencing, lacks standardized implementation
The internal transcribed spacer (ITS) regions of the ribosomal DNA are the preferred molecular targets for compositional analysis of the mycobiota by high-throughput amplicon sequencing (Schoch et al.2012; Bates et al.2013). In fungal genomes, the genes encoding the 18S, 5.8S and 25S ribosomal RNA subunits are found contiguously in multicopy, and are separated from each other by stretches of DNA of variable length—the ITSs. Two such regions, ITS-1 and ITS-2, separate the three genes from each other. The ITS regions are intrinsically variable in length, ranging from ∼450 bp to ∼700 bp (Bellemain et al.2010). Thus, any given set of standard PCR oligonucleotides will yield amplicons of different sizes from different fungal species (Bellemain et al.2010; Bokulich and Mills 2013). This creates an inherent PCR bias, because short products are likely to be preferentially amplified.
A scientific consensus has not yet been established for the precise ITS region to target for high-throughput amplicon sequencing. Some authors have preferred amplification of the entire ITS region (ITS-1, 5.8S gene, ITS-2), while others have targeted either ITS-1 or ITS-2 (LaTuga et al.2011; David et al.2014). While the ITS-1 region is recommended for high-throughput sequencing (Bokulich and Mills 2013), the ITS-2 region is advocated for species in which spliceosomal inserts near the 3′ end of the 18S rRNA gene would bias amplification of ITS-1 (Bates et al.2013). Furthermore, the target amplicon size should be matched to an appropriate sequencing technology to maximize the phylogenetic signal that can be obtained through sequencing. For example, long amplicons, such as those encompassing ITS-1, 5.8S rRNA and ITS-2, can be successfully sequenced using 454-pyrosequencing technology, thereby maximizing signal for taxon assignment and phylogenetic analysis (Bokulich and Mills 2013). It has also been recommended to prepare multiple PCR reactions for each environmental DNA sample, some of which would target different ITS regions, and to perform the reactions at more than one annealing temperature, to overcome amplification biases (Diakarya vs non-Diakarya, Ascomycota vs Basidiomycota), which would otherwise confound a reliable community analysis (Bellemain et al.2010).
The utility of community profiling by modern high-throughput sequencing methods is also highly dependent on researchers’ ability to assign the sequenced reads to a taxon of defined nomenclature. Without the public availability of well-curated and up-to-date databases which include annotated fungal ITS (and for metagenome data, fungal genome sequences) for specific species, large proportions of generated sequence data would remain unclassified. While a number of sequence databases exist for fungal taxonomy assignment, [UNITE (Abarenkov et al.2010) (https://unite.ut.ee/), SILVA (Pruesse et al.2007) (http://www.arb-silva.de/), MaarjAM (Opik et al.2010) (http://maarjam.botany.ut.ee/)], the UNITE database is recommended for ITS-based taxon assignment while SILVA is appropriate for SSU-based targets (Lindahl et al.2013). Nevertheless, it has been estimated that less than 5% of fungal species that are thought to exist on earth are represented in databases (Kõljalg et al.2013). To address the problem of sequence reads for which no meaningful taxonomic information is available, the UNITE team has introduced the concept of the ‘species hypothesis’ (SH). A ‘species hypothesis’ is a group of two or more sequence reads which have been clustered to the approximate species level subsequent to an initial clustering to the subgenus/genus level, though singletons are also tolerated. If an expert-specified reference sequence has not been defined, a representative sequence for the new SH is designated and is assigned a unique SH number, which allows the sequence to be classified. The value of this approach is that it allows researchers to approximate the taxon assignment of sequence reads for unknown fungi and serves as a standardized reference for that particular sequence over time (Kõljalg et al.2013).
Limitations in our understanding of the human intestinal mycobiota
The paucity of information in the scientific literature regarding the eukaryotic component of the intestinal microbiota of healthy humans is a problem that has been acknowledged for several years. An early culture-independent study of the intestinal mycobiota of healthy adults targeted the eukaryotic 18S rDNA gene and the fungal ITS region (Scanlan and Marchesi 2008). Rather than using a high-throughput sequencing approach (a method that has only quite recently become widely available), traditional methods for community profiling by molecular methods, namely denaturing gel gradient electrophoresis and clone library preparation followed by Sanger sequencing, were used. As such, the sampling depth was probably quite limited. The overall diversity of eukaryotes was low, and fungal sequences were detected in ∼82% (14/17) of the individuals tested. Based on a Fungal Internal Transcribed Spacer Analysis, fungal communities were deemed relatively stable over time, although allochthonous fungal species (those that are transiently present and which were not detected at every sampling occasion) were also detected. Such species include food-associated fungi, such as Penicillium roqueforti, which is found in certain cheeses. In contrast, species of the genera Galactomyces, Paecilomyces and Gloeotinia, some of which may also be food-borne fungi, were present at all time points and were suggested as common and stable inhabitants of the human GI tract (Scanlan and Marchesi 2008). Furthermore, there was an incongruence between the results of the culture-dependent and culture-independent analyses of the fungal communities in these individuals, emphasizing the limitations and biases of each of these experimental approaches for comprehensive community profiling and assessment.
More recently, a meta-analysis of the ‘mycobiota’ literature, which focused on 75 research articles published in English between 1998 and 2014, provided a list of the fungal genera and species that occur in healthy individuals and those that were detected in individuals with a particular disease (Gouba and Drancourt 2015). The intestinal mycobiota of healthy individuals was reported to include 247 fungal species, which belong to the phyla Ascomycota (63%), Basidiomycota (32%), Zygomycota (3%) and non-classified fungi (2%) (Gouba and Drancourt 2015). These authors also acknowledge the discrepancies in the number of fungal species that can be identified by culture-dependent versus culture-independent techniques. They also indicate that the fungal community profile varies depending on the body site sampled, with the fungal communities of the oral and intestinal mycobiotas being different and with little overlap in species membership.
The application of modern, high-throughput sequencing and data analysis methods and the continuing improvement of reference databases thus provide a means to address the dearth of information on the GI mycobiota of healthy adult humans. The outcomes of several recent studies which availed of these technologies for mycobiota profiling will now be discussed.
Insights into the human intestinal mycobiota from high-throughput amplicon sequencing
A pyrosequencing-based study of the eukaryotic component of the fecal microbiota of eight healthy adult humans revealed that fungi accounted for most of the operational taxonomic units (OTUs) identified, whether assessed by 18S rRNA (∼62%) or fungal-specific ITS (∼90.5%) gene amplicons (Dollive et al.2012). The ITS sequences were assigned to 215 OTUs, while the 18S rRNA gene-derived reads yielded 93 OTUs. Nevertheless, most of the reads were assigned to few OTUs, implying that the taxa detected were present in unequal amounts. Ascomycota and Basidiomycota accounted for 81 and 17.2% of the reads from 18S rRNA gene-derived sequences, respectively while these phyla were represented at 57.4 and 25.7% respectively in the ITS data. The low number of food and plant reads among the sequenced amplicons resulted from a deliberate effort to minimize signal from mammalian and dietary plant sources through prudent primer design. Interestingly, 37% more reads were returned from the 18S rRNA gene amplicons than from the ITS-derived products, demonstrating how the choice of molecular target region and primer combinations can impact the outcome of mycobiota analyses.
The influence of diet and microbiota composition on the diversity of the fecal fungal and archaeal communities was studied using stool collected from 96 healthy adult humans (Hoffmann et al.2013). Each individual tested harbored a fungal community that was dominated at the phylum level by either the Ascomycota or the Basidiomycota. Substantial interindividual variation in the composition of the fungal community at the genus level was observed, with only 12 of the 66 fungal genera identified being common to 9 or more of the individuals tested. Saccharomyces and Candida were the most prevalent of these genera, being detected in 89 and 57% of the subjects respectively. While the prevalence of Saccharomyces reads could derive from food-associated consumption of this species (via bread or beer), the relative abundance of the genus Candida was found to be significantly influenced by an individual's short-term diet. The relative abundance of the genus Candida was positively correlated with carbohydrate consumption, but negatively correlated with the ingestion of saturated fatty acids (Hoffmann et al.2013).
The structure of the microbiota was examined by investigating the relationships between the fungi, bacteria and archaea present. The genus Candida tended to cooccur (Dice index > 0.5) with the bacterial genera Faecalibacterium, Roseburia and Parabacteroides, the fungal genus Saccharomyces and with uncharacterized species of the bacterial families Lachnospiraceae and Ruminococcaceae. The genus Candida was found to significantly and positively covary with the bacterial genera Xylanibacter, Catenibacterium and a member of the family Prevotellaceae. It also cooccurred (Dice index 0.61), but negatively covaried with the genus Bacteroides (Bacteria). Candida was found to be positively and negatively associated with Methanobrevibacter (Archaea) and Nitrososphaera (Archaea), respectively.
Furthermore, it was the proportion, rather than the types of microbes present which influenced microbiota composition. For example, the Prevotella/Bacteroides ratio correlated with the amount, but not the type, of fungi present. Similarly, significant correlations between the phyla Ascomycota and Basidiomycota and various bacterial taxa were returned only when taxon abundance and not simply taxon diversity was considered (Hoffmann et al.2013).
A study of 10 American adults who volunteered to follow either a plant- or animal- based dietary regime over 5-day period also revealed that the size and composition of the enteric fungal community is directly impacted by host diet (David et al.2014). Cured meats and several cheese products were found to harbor a fungal community consisting of Candida species and members of the fungal genera, Debaryomyces, Penicillium and Scopulariopsis. Furthermore, diets rich in animal-derived products supported a significant increase in the concentration of viable fungi in feces relative to baseline levels (before dietary intervention) (David et al.2014). Like in the study described above (Hoffmann et al.2013), Candida species became more abundant on the carbohydrate-rich plant diet. These analyses (Hoffmann et al.2013; David et al.2014), however, are limited to a discussion of the effect of diet on the genus Candida, rather than on C. albicans or any other Candida species specifically.
Thus, high-throughput amplicon sequencing provides considerable insight into the fungal communities of the human gut. The composition of the GI mycobiota is influenced both by ingested fungal species and the interaction of commensal fungi with the rest of the gut microbiota. Nevertheless, despite these initial studies, the autochthonous species of the healthy human gut mycobiota have not been well described and cannot at this time be distinguished from the allochthonous (most likely food-borne) fungal species transiting this niche.
Moreover, the technical choices made when isolating DNA and when selecting PCR oligonucleotides and a sequencing platform could have a considerable impact on the results. Standardization of protocols to establish best practice procedures is needed if mycobiota profiling by high-throughput sequencing methods is to become routine (Bates et al.2013; Lindahl et al.2013).
Insights into the human intestinal mycobiota from shotgun metagenomics studies
Metagenomics studies of the human gut microbiota have rarely focused on the functions contributed by the non-bacterial microbes present. For the eukaryotic community specifically, there may be several reasons for this. First, the non-bacterial community is overall quite small, and fungi comprise only a fraction of this. For example, the Metahit consortium (Arumugam et al.2011) identified eukaryotic DNA fragments in their metagenome data at a low level (<1.3% of DNA from any sample; 0.5% on average, based on homology to eukaryotic sequences in the STRING database, version 8). Together, fungi and metazoa accounted for the largest average fraction of the eukaryotic sequences (∼0.3%), with fungi specifically contributing an average of 0.11% of these (Arumugam et al.2011).
Furthermore, eukaryotic genomes are often large, with a low coding density (Wood et al.2002). Consequently, a considerably greater depth of sequencing would be required to access the genetic content of the eukaryotic community at a level comparable to that routinely obtained for the prokaryotic community, particularly in niches harboring a diverse fungal community. This is exemplified in a non-redundant microbial gene catalogue representing the fecal microbiota of 124 Europeans, in which only ∼0.1% of the genes were of eukaryotic origin while 99.1% of the genes were derived from bacteria (Qin et al.2010). The additional sequencing that would be required to achieve a depth sufficient to access fungal genes is likely to be prohibitive, rendering an analysis of the specific contributions of the intestinal eukaryotic community beyond the scope of most metagenome studies performed to date.
Nevertheless, with the falling cost of sequencing combined with the greater availability of curated fungal genomes in publically accessible reference databases, and the development of the necessary tools and expertise which will enable labs to handle and analyse their sequencing data ‘in house’, we can expect that metagenomics studies will incorporate more detailed descriptions of the contributions of the fungal and archaeal communities in the future. One such initiating metagenome project provides insight into the intestinal mycobiota on an international scale.
In a comparative study of the intestinal microbiomes of infants and adults from Malawi, the United States and Amerindians from Venezuela (Yatsunenko et al.2012), in which a total of 5.9 Gb of sequencing data was generated, only 7 ± 8% of the reads representing 110 metagenomes, could be mapped to non-bacterial sequences. The majority of these non-bacterial reads corresponded to either archaea or fungi. For each human population examined, the fungal proportion of the microbiota was always significantly greater in adults than in children. Furthermore, the microbiomes of individuals from the USA contained significantly fewer fungal sequences than those of either the Malawian or Venezuelan adults. These differences in mycobiota size and structure could possibly reflect differences in the diets of adults and children within and across the different societies. Sequences of the phyla Ascomycota and Microsporidia were the most abundant fungal sequences in each of the three populations tested (Yatsunenko et al.2012).
Thus, the burgeoning field of ‘gut microbiomics’ has not yet well integrated the resident and transient fungi nor the non-prokaryotic microbes in general, though this is likely to change with time. These hitherto neglected microbial populations cannot continue to go unnoticed. The microbiota–mycobiota dynamics in the gut are starting to be appreciated (Cuskin et al.2015) and given their contribution to biomass in the gut (Underhill and Iliev 2014), enteric fungi are likely to bear significant immunomodulatory potential.
MODELS OF GI COLONIZATION BY C. ALBICANS
The study of the C. albicans—host relationship in the laboratory often relies on the development of appropriate in vivo models that recapitulate the characteristics of either infection or commensalism in humans. Although several models have been developed for these purposes (Van Cutsem and Thienpont 1971; Hube et al.1997; Braun et al.2000; Schinabeck et al.2004; Hoeflinger et al.2014), models involving mice have been most commonly used to study C. albicans commensalism in the stomach and lower GI tract (Schofield et al.2005; Koh 2013; Pérez, Kumamoto and Johnson 2013; Prieto et al.2014).
In general, the mouse is a desirable model organism because its small size means that large numbers of animals can be housed and handled easily and inexpensively. Furthermore, mouse models often recapitulate major characteristics of human diseases and the availability of methods to manipulate the mouse genome means that researchers can generate or purchase mice of specific mutant genotypes of interest if necessary. The establishment of rodent models of intestinal colonization by C. albicans is, however, complicated by the fact that this yeast does not naturally colonize these animals (Huppert, Cazin and Smith 1955). The gut microbiota of laboratory rodents is quite unlike that of humans (Loan et al.2015), and the colonization resistance that it imposes is sufficient to prevent the establishment of a large enteric C. albicans population. Thus, experimental interventions are required to facilitate stable intestinal colonization in these animals (Romani 2001; Clancy, Cheng and Nguyen 2009; Koh 2013).
Antibiotic treatments have often been the intervention of choice to establish long-term and stable colonization of the murine gut by C. albicans. Advantages of this system include the potential for the use of immunologically competent wild-type (WT) mice, the widespread and cheap availability of the antimicrobial drugs, the high levels of colonization that can be achieved with the system, the non-invasive dosing method (if antibiotics are delivered via drinking bottles) and its simple set-up, with no specific training or apparatus required. Disadvantages of the system include variation in the amount of antibiotic solution consumed by the animals resulting in variable doses, the possibility of wastage of antibiotic solutions, widespread loss of microbiota structure following broad spectrum antibiotic treatment, antibiotic-associated diarrhoea as a side effect of treatment and an unwanted impact on the immune system or other physiological consequences of antibiotic exposure. Furthermore, dosages found to be effective in animals may not be easily applied to humans.
While, genetic interventions centred on immunosuppression (reviewed in Koh 2013) have been used for C. albicans colonization of the GI tract, methods that do not require antibiotic treatments or immunosuppression have also been reported. For example, the use of purified diets (Yamaguchi et al.2005), regular C. albicans dosage via food (Samonis et al.1990, 2011), immunodeficient germ-free animals (Schofield et al.2005) and the inoculation of mouse pups with C. albicans soon after birth may allow the establishment of a sizable C. albicans population in these animals (reviewed in Romani 2001). In each case, however, it is necessary to characterize the model to verify the absence of dissemination and disease.
THE ROLE OF THE INTESTINAL MICROBIOTA IN THE REGULATION OF ENTERIC C. ALBICANS POPULATIONS
A tripartite interaction involving the host's immune system, the intestinal microbiota and C. albicans regulates the population size of this yeast in the gut. Evidence from animal models (Zelante et al.2013) and human trials (Romeo et al.2011; Kumar et al.2013; Roy et al.2014) has demonstrated that certain bacterial populations can specifically restrain the outgrowth of C. albicans in this niche. The identification of bacterial genera, species or strains with anti-Candida properties could potentially be developed as probiotics against this yeast. Although certain products of bacterial metabolism may offer generic anti-Candida properties, such as short-chain fatty acid production which has been demonstrated to inhibit fungal growth and morphogenesis by C. albicans in vitro (Noverr and Huffnagle 2004), bacteria may also possess strain-specific anti-Candida properties. Several publications testify to the in vivo anti-C. albicans properties of various bacteria.
One such study showed how the metabolism of dietary tryptophan by selected members of the stomach microbiota of mice influenced colonization resistance towards C. albicans (Zelante et al.2013). Specifically, Lactobacillus species that harbored a gene for an aromatic amino acid aminotransferase (araT), which enabled the bacterium to produce the indole metabolite, indole-3-aldehyde (IAld) from dietary tryptophan, were able to stimulate IL-22 production in the murine stomach and in specific immune cells via the host aryl-hydrocarbon receptor (AhR). Although the AraT enzyme is phylogenetically conserved, it is not present in all bacterial lineages, and is absent from Clostridia species of clusters IV and XIVa that are dominant in the murine gut. Thus, not every species is capable of antagonizing C. albicans or of potentiating an antifungal effect in this manner.
Experiments in which antibiotics or dietary tryptophan were used to modulate the indigenous Lactobacillus community in the mouse stomach demonstrated how the C. albicans population of intragastrically inoculated mice was correlated with the size and composition of the gastric Lactobacillus population (Zelante et al.2013). Colonization resistance towards Candida was increased in WT mice during tryptophan feeding (promotes expansion of the Lactobacillus community) and was decreased following ampicillin treatment (decreases the Lactobacillus community). In WT mice consuming a diet rich in tryptophan, this anti-Candida resistance was also associated with the increased local expression of several parameters, including transcription of a bacterial gene encoding a presumptive araT, IL-22 production and the proportion of IL-22-producing innate lymphoid cells in the stomach (IL-22 had an anti-Candida activity). This shows how microbiota-derived signals can elicit a host response that underlies colonization resistance to fungi.
Finally, therapeutic administration of IAld to WT mice with either mucosal candidiasis or dextran sodium sulphate induced colitis, restored anti-Candida resistance, promoted mucosal protection from injury and ameliorated the symptoms of colitis. These effects were not found in mice lacking the AhR receptor, and which were therefore insensitive to the immunostimulatory effects of IAld.
Thus, specific microbiota interventions that would harness the immunomodulatory potential of its constituent microbes to boost the host's intrinsic antifungal resistance could be developed.
Several clinical trials have shown that probiotic strains of bacteria or fungi can be used prophylactically to prevent fungal outgrowth or infection following antibiotic treatment. One such prospective, double-blind, randomized controlled trial based on the Pediatric intensive care unit of a hospital in India (Kumar et al.2013) enrolled 150 critically ill children who were being treated with broad-spectrum antibiotics, and gave them either a probiotic cocktail or a placebo as part of their treatment regime for 14 days. The probiotic cocktail contained a mixture of bacteria and yeast species—two Lactobacillus (L. acidophilus and L. rhamnosus), two Bifidobacterium (B. longum and B. bifidum) and two Saccharomyces (S. boulardii and S. thermophiles) species—, in addition to prebiotics (fructooligosaccharides) on a lactose base, while the placebo group received only lactose. The probiotic treatment led to significantly fewer individuals testing positive for the presence of Candida species in rectal swabs taken 14 days after enrolment in the study. The incidence of candiduria was also significantly decreased in the probiotic-treated group, although there was no significant effect on the prevalence of candidemia in these children (Kumar et al.2013).
In a separate prospective study involving 249 preterm newborns in Italy (Romeo et al.2011), Lactobacillus probiotics (L. reuteri ATCC55730 and L. rhamnosus ATCC53103) were also found to be effective at reducing enteric Candida colonization and improving several other clinical parameters. Another prospective, randomized, double-blind, placebo-controlled study which focused on 112 preterm low birth weight neonates also found that supplementation with a probiotic cocktail of Lactobacillus and Bifidobacterium species led to a significant reduction in enteric Candida levels, and a reduced incidence of invasive fungal sepsis (Roy et al.2014). Evidently, dietary supplementation with probiotic bacteria can be a useful clinical strategy to overcome fungal infections in humans, but little is known of the utility of fungal probiotics alone for the same purpose. The fact that yeasts are likely to be resistant to most antibiotics is a favourable attribute of yeast-based probiotics. Thus, unlike bacterial probiotics which may be sensitive to a given antibiotic, yeast probiotics could potentially confer beneficial effects on the host, even during antibiotic treatment. Although, the literature describing the anti-fungal potential of fungal probiotics is sparse, several animal trials have attested to the probiotic potential of S. boulardii (Berg et al.1993; Jawhara and Poulain 2007). This yeast is the active biological component of commercially available probiotic products [UltraLevure, (France), Perenterol (Germany), Reflor (Turkey), Florastor (USA), Sacchaflor, (Denmark)]. To our knowledge, the only published clinical trial to date to examine the anti-fungal potential of S. boulardii in humans was a prospective, randomized comparative study of very low birth weight infants (Demirel et al.2013). This trial showed that consumption of S. boulardii was as effective as the anti-fungal agent nystatin, for the prevention of fungal colonization and invasive fungal infections in these premature infants. Additional clinical trials are needed if S. boulardii is to be adopted as a prophylactic treatment for fungal infections in humans, particularly given that sepsis caused by S. boulardii is potentially a serious problem for immunocompromised individuals or those consuming antibiotics (Thygesen, Glerup and Tarp 2012).
WHAT IS THE EFFECT OF C. ALBICANS CARRIAGE ON THE GI MICROBIOTA UNDER DYSBIOTIC CONDITIONS?
This issue was addressed by several studies from the same research group, in which the impact of the introduction of C. albicans into a GI microbial community that was previously disturbed by antibiotic treatment was investigated (Mason et al.2012a; Erb Downward et al.2013). The antibiotic of interest was cefoperazone, a broad-spectrum antibiotic which functions by inhibiting cell wall synthesis (DrugBank DB01329), and which is poorly absorbed from the GI tract (Erb Downward et al.2013). It was administered ad libitum via the drinking bottles for 7 days prior to the introduction of C. albicans. Data from these studies have demonstrated, as expected, that broad-spectrum antibiotic treatment substantially alters the structure of the GI bacterial community. In some instances, recovery to the pre-treatment state can be lengthy (Mason et al.2012b; Erb Downward et al.2013) and during the intervening period, the microbiota may be considered as being in a state of flux. Consequently, the colonization resistance that is usually imparted by the undisturbed microbiota towards opportunistic pathogens and transient colonizers is lowered during this recovery period. As a result, outgrowth of particular microbial species, such as C. albicans, may ensue (Fig. 1).
Impact of antibiotic treatment on microbiota dynamics and on C. albicans colonization and establishment in the gut. (A) The ‘normal’ microbiota contains a diverse community of bacteria (rods and small cocci in diagram), fungi (large white and coloured circles) and other microorganisms. Together, these impart colonization resistance against infections by opportunistic pathogens. (B) Antibiotic treatment eliminates subpopulations of the bacterial community. This microbiota perturbation disturbs the colonization resistance and allows C. albicans to establish or outgrow in this niche. In doing so, C. albicans may outcompete the other yeast species of this ecosystem. (C) Withdrawal of the antibiotic treatment allows for microbiota recovery and a reduction in the C. albicans load. Candida albicans carriage may alter the recolonization dynamics, and could therefore delay or prevent restoration of the pre-antibiotic treatment state. Furthermore, C. albicans may become established at low levels in this niche, which means that relapsing infections could occur if antibiotic treatment is resumed.
The effect of C. albicans on the microbiota of antibiotic-treated and untreated mice was assessed by 16S rRNA gene targeted amplicon sequencing of DNA recovered from the cecal mucosa on days 7 and 21 after the cessation of antibiotic treatment (Erb Downward et al.2013). A substantial loss of diversity was apparent on day 7 in mice treated with cefoperazone alone, but this was largely restored by day 21. The introduction of C. albicans into a disturbed microbial community further reduced its diversity, though the reduction was not statistically significant (Erb Downward et al.2013). Microbiota diversity was largely but incompletely restored in animals that had received antibiotics and C. albicans by day 21.
Candida albicans carriage influenced recovery dynamics. For example, on day 7, in the family Ruminococcaceae the representation of the genus Oscillibacter was substantially reduced in antibiotic-treated animals in which C. albicans was also present. Among the Lachnospiraceae, Coprococcus and Dorea levels were reduced while an outgrowth of Roseburia and Robinsoniella species was observed. In fact, the animals inoculated with C. albicans and treated with antibiotics had a microbiota profile that was significantly different from those of untreated or antibiotic-only treated animals, implying that the presence of C. albicans influenced community reassembly. In contrast, introducing C. albicans into immunocompetent mice without an antibiotic-induced microbiota disturbance did not significantly alter bacterial diversity on either day 7 or 21.
Another study also investigated the associations between antibiotic treatment, GI microbiota and immune responses on C. albicans colonization of the murine GI tract (Shankar et al.2015). These authors found that antibiotic treatment was the factor that most consistently influenced C. albicans colonization. Upon inspection of the mycobiota of the antibiotic-treated mice, it was found that once it was introduced, C. albicans tended to displace the resident GI fungi, such as C. tropicalis. In stool samples, species of the genera Streptococcus and Parabacteroides were associated with high levels of C. albicans, while the genera Lactobacillus and Prevotella were associated with lower C. albicans colonization levels. Enterococcus and Veillonella were associated with higher levels of C. albicans in the terminal ileum. The proportional abundance of each of these genera tended to be influenced by the antibiotic treatment regime chosen. Together these studies exemplify the close relationship that exists between C. albicans, the microbiota and their host.
By analyzing both the bacterial and fungal components of the microbiota following defined perturbations and over time, we broaden and deepen our understanding of the microbial population dynamics and interactions occurring in the gut ecosystem. A thorough mining of such data could potentially identify specific bacterial populations or communities that usually keep opportunistic fungal pathogens, such as C. albicans, in check. While some studies of this nature have already been published (Shankar et al.2015; Erb Downward et al.2013), these examples often study C. albicans in the context of the murine gut microbiota. Equivalent data from human or humanized subjects are lacking. Thus, before the natural colonization resistance of the human gut microbiota towards C. albicans can be fully exploited, we need to better understand the communities of microbes or the specific microbial functionalities that are responsible for the anti-C. albicans effect. Methods, models and study cohorts that would allow the research community to address these needs are therefore highly sought and are likely to direct future work in this area.
PHENOTYPES AND GENES REQUIRED FOR C. ALBICANS COMMENSALISM
Even though C. albicans’ default lifestyle is commensalism in the GI tract, its ability to cause potent disseminated disease means that its reputation as a pathogen often overshadows its commensal tendencies. Many infections with C. albicans are, however, typically opportune. Candida albicans is a polymorphic fungus, capable of occurring in a unicellular yeast form, in one of two filamentous forms (pseudohyphae or hyphae), and also as a chlamydospore (reviewed in Whiteway and Bachewich 2007). Particular morphotypes and the phenomenon of phenotypic switching in response to environmental triggers have been associated with commensalism and pathogenesis (Gow et al.2011) (Fig. 2). Thus, a precise understanding of the attributes of the various C. albicans phenotypes and their role in either commensalism or pathogenesis is important.
The C. albicans and host factors affecting C. albicans colonization levels of the murine GI tract. The laboratory mouse, which is the focal point of this figure, has been extensively used as a mammalian host to decipher many aspects of the biology of C. albicans in the GI tract. Colonization results from the interplay between the host and the colonizing yeast. Features that are known to influence GI colonization are listed in panels within larger boxes representing either the colonizing C. albicans cell (top) or the host (bottom). Within each panel, arrows indicate if a given factor increases or decreases the C. albicans colonization levels. (A) A list of proteins which regulate C. albicans colonization levels. Of the proteins listed, all but Hog1 (a mitogen-activated protein kinase) are transcription factors. (B) When compared to the white, yeast form cells, the other C. albicans morphologies have enhanced or reduced colonization levels as indicated. (C) Microbiota composition and interventions that modify it, such as probiotic treatment or host diet, can also influence the C. albicans GI colonization levels as indicated. (D) A functional host immune system, genetic polymorphisms or genetic background may also impact on C. albicans levels in the GI tract as indicated. P = pathogenic infection, ? = unknown effect, - = no effect. O/E = overexpression, KO = knockout. WT = wild-type.
In its unicellular form, C. albicans may adopt one of three distinct, non-genetically determined phenotypes named for their colony appearance: white, opaque or gray (Tao et al.2014). Each of these three morphologically distinct phenotypes is stable and can be inherited over several generations. In vitro, cells of the different types exhibit distinguishable transcriptional profiles, secreted aspartyl protease activities and mating competencies (Tao et al.2014). They also have varying virulence in in vivo models of candidiasis and cutaneous infection. White cells were more virulent than either opaque or gray cells in a mouse model of systemic candidiasis, while gray and opaque cells caused more skin damage than white cells when assessed 24 hours after being applied to the skin of a newborn mouse (Tao et al.2014). Gray cells were found to grow at a faster rate than either the opaque or the white cells in an ex vivo murine tongue infection assay, suggesting that gray cells were better adapted for nutrient acquisition from host tissues (Tao et al.2014).
It has been suggested that passage through the GI tract induces a specific developmental program in C. albicans which primes the yeast for commensalism (Pande, Chen and Noble 2013). Consequently, its morphology, transcriptome and metabolism may become altered, though colonizing cells usually occur in the yeast form (White et al.2007; Pierce and Kumamoto 2012). WT white C. albicans cells colonize the GI tract of antibiotic-treated mice at high levels, while opaque cells are comparatively attenuated for colonization (Pande, Chen and Noble 2013). White cells lacking the transcription factor, Wor1, whose levels dictate whether C. albicans is in the white or opaque state, are rapidly cleared from the mouse gut, implying that activation of this gene is essential for commensalism. Accordingly, overexpression mutants of WOR1 demonstrated enhanced colonization of the murine GI tract relative to the white WT strain (Pande, Chen and Noble 2013). The predicted alteration of morphology during GI colonization or transit was stimulated by the recovery of WOR1 overexpression mutants with an unexpected ‘dark’ phenotype. These cells were distinguishable from opaque cells by having a phenotype that was stable at elevated temperatures (>25°C), bearing no surface pimples, containing prominent vacuoles, having a heterozygous mating-type locus, not responding to mating pheromones and by being less efficient at mating (Pande, Chen and Noble 2013). The ‘dark’ cells also had an altered transcriptome which appeared to optimize yeast behaviour and metabolism for GI colonization. However, the ‘dark’ phenotype was only observed with the WOR1 overexpression strain, prompting the suggestion that the host factors that are responsible for the white-to-dark transition are not well replicated in vitro. It was hypothesized that the stable overexpression of WOR1 (which for WT cells would ostensibly be triggered by GI transit) in the mutant strain allowed the altered ‘GUT’ (Gastrointestinally-indUced Transition) phenotype to be observed in the laboratory.
Thus, the phenotypes expressed by C. albicans may reflect different functional specializations and developmental programs that have been optimized for either commensalism or pathogenesis in different niches. While these phenotypes most likely reflect the coordinated expression of a particular panel of genes, several studies have identified specific genes with critical roles in either commensalism or pathogenesis.
One such gene is SFU1, which promotes C. albicans commensalism in the GI tract (Chen et al.2011). This gene is part of the unique tripartite iron utilization system of C. albicans and confers resistance to toxic levels of iron in the gut by restricting iron uptake in this iron-rich environment (Chen et al.2011). In contrast, Sef1, a transcriptional activator of both iron uptake and iron-utilization genes (amongst other targets), was necessary for virulence in bloodstream infections in which iron was poorly available due to its sequestration by the host. Nevertheless, relative to the WT strain, mutants lacking either SEF1 or SFU1 were impaired in their ability to persist in the murine gut over a 15-day period (Chen et al.2011). Interestingly, the GUT cells described by Pande, Chen and Noble (2013) were also found to downregulate expression of iron-acquisition genes, which is consistent with their apparent optimization for GI colonization.
Thus, signalling pathways that relay signals from the host or the environment are likely to influence gene expression in the colonizing yeast. The mitogen-activated protein kinase Hog1 was found to be essential for colonization of the murine GI tract by C. albicans (Prieto et al.2014). The HOG pathway is involved in adaptation to osmotic and oxidative stresses and has a role in cell wall biogenesis (Prieto et al.2014). Mutants lacking HOG1 were rapidly cleared from the gut when in competition with the parental strain. A Hog1 conditional mutant, in which excision of HOG1 was made possible by induction of a flippase system, revealed that HOG1 was needed for long-term colonization of the mouse gut. Ex vivo assays indicated that the hog1 mutant was sensitive to bile salts and displayed poorer adhesion relative to the WT strain (Prieto et al.2014). Two other MAP kinases, Cek1 and Mkc1, were also tested in a competition model with a CAF2 mutant (parental strain with a reporter gene, considered as WT), and while they initially colonized the GI tract as well as the WT, this stable colonization was not maintained for longer than 2 to 3 weeks (Prieto et al.2014).
The transcription factor Efg1, which is central to the regulation of the C. albicans yeast-to-hypha transition, has also been identified as a major regulator of GI colonization in mice (Pierce and Kumamoto 2012). The expression levels of this gene are influenced by the immune status of the host and the time course of colonization (Pierce and Kumamoto 2012). With reference to expression levels in vitro, transcription of EFG1 was low during the early phases of colonization of the murine ileum and Cecum but tended to increase with time in immunocompetent hosts (Pierce and Kumamoto 2012). In contrast, in immunocompromised mice, EFG1 expression remained low throughout the experimental period. Thus, C. albicans’ gene expression during colonization is influenced by the immune status of the host. A natural heterogeneity of EFG1 expression levels is thought to exist in the colonizing yeast population. This ‘phenotypic diversity’ may allow the C. albicans community to respond appropriately to host-imposed signals to ensure successful persistence and niche colonization (Pierce and Kumamoto 2012).
The transcriptome of C. albicans cells in the GI tract differs markedly from that of laboratory grown cultures (Pierce et al.2013). Moreover, while the transcriptional profile of yeast cells colonizing different regions of the GI tract overlap considerably, they can also be distinguished, reflecting adaptation by the yeast to colonization of the anatomically and physiologically distinct niches of either the ileum or the Cecum (Pierce et al.2013). In general, according to GO terms enrichment, colonization requires expression of genes apparently involved in pathogenesis (but which are more likely to represent cell surface restructuring for adhesion or immune evasion during colonization, given that most colonizing cells occur in the yeast rather than hyphal forms), carbohydrate metabolism and stress response. Comparative studies involving various EFG1 knockout mutants revealed that Efg1 directly influenced expression of the Als3 adhesin, hypha-regulated cell surface protein Hyr1, the secreted aspartyl proteases Sap4 to Sap6 and the Sod5 superoxide dismutase (Pierce et al.2013). Genes involved in pathogenesis and oxidative stress responses were generally expressed at lower levels in the efg1 mutant strain than in the WT during colonization, while genes involved in lipid catabolism were expressed at higher levels in the mutant. This suggests that Efg1 also influences the in vivo expression of these ‘host-response’ genes, and the influence of Efg1 was more pronounced for cells colonizing the Cecum than the ileum (Pierce et al.2013).
An earlier study from the same group showed that Efh1, a transcription factor paralogous to Efg1, was also highly expressed in the porcine and murine oro-GI tracts during colonization by C. albicans (White et al.2007). A mutant strain lacking EFH1 was capable of enhanced colonization relative to the WT strain. This echoes the results achieved with Efg1 knockout mutants above (Pierce and Kumamoto 2012). Thus, expression of EFH1 and EFG1 in vivo provides a reversible way to limit the size of the colonizing C. albicans population (White et al.2007; Pierce and Kumamoto 2012) perhaps in response to host-derived signals. Transcription factors are therefore interesting candidates for studies aiming to decipher the means by which C. albicans persists as a commensal in its host. Nevertheless, the influence of transcription factors on GI colonization likely represents the first step in a cascade of events in which the transcriptome of the colonizing yeast is reorientated so that the genes encoding the true ‘effector’ molecules are expressed. The challenge therefore lies in determining not only which genes are essential for GI tract colonization, but also in elucidating the mechanism(s) by which these genes exert their effect.
This was the approach taken in a comprehensive study of 77 C. albicans transcription regulator mutants (Pérez, Kumamoto and Johnson 2013). Using mouse models, eight such factors were identified as being required for GI colonization, systemic infection or both. Of these, three regulators (Rtg1, Rtg3, Hms1) were involved in both colonization and infection, while two (Lys144, a zinc-finger cluster transcription factor and Tye7, which is involved in carbohydrate metabolism) were needed for intestinal colonization only. Considerable overlap exists in the targets of these regulators, suggesting that colonization and systemic infection are controlled by the same network. The subset of effector genes under the influence of Rtg1 and Rtg3 was found to be particularly upregulated in the GI tract, implying a significant role for these two regulators during colonization. These and some of the aforementioned regulators were shown to bind upstream of genes for amino acid permeases and allantoate transporters, expression of which is thought to facilitate nitrogen acquisition in the gut environment (Pérez, Kumamoto and Johnson 2013). Homozygous deletion mutants of GAL10, DFI1 and HAP41, genes that are believed to be regulated by Hms1 and Rtg1/3, demonstrated impaired GI colonization. Expression of these genes may lead to cell surface remodelling, which therefore influences colonization.
Unravelling the genetic programs behind C. albicans commensalism and pathogenicity (Fig. 2) could potentially allow for the development of drug-based interventions which would reduce the risk of C. albicans infections in colonized or ‘at risk’ individuals. While much progress has already been made towards this goal, more is needed, and research into this area is likely to continue into the future.
CANDIDA ALBICANS–HOST INTERACTIONS AND INTERPLAY DURING COLONIZATION AND OUTCOME DURING INFECTION
Several pattern recognition receptors (PRRs) are involved in sensing fungal microbe-associated molecular patterns (MAMPs), and include members of the Toll-like receptor family, the C-type lectin receptor family, the galectin family receptors and indirectly, the Nod-like receptors (Romani 2011). These receptors respond to fungal molecules such as cell wall carbohydrates, surface proteins and fungal nucleic acids (Romani 2011). The host can respond differently to the yeast and the filamentous forms of C. albicans and certain cell types and receptors may specifically recognize or mount a tailored immune response to either one or the other of these fungal morphotypes (reviewed in Romani, Bistoni and Puccetti 2002; Naglik et al.2011). This capacity of the host to distinguish yeast forms from filamentous C. albicans cells is critical to the host's ability to discriminate benign commensalism from pathogenic infections (Naglik et al.2011). Moreover, the host can sense and respond to the increased fungal burden that is a hallmark of candidiasis, implying that a threshold level of C. albicans exists, and levels exceeding this threshold distinguish pathogenesis from commensalism (Moyes et al.2014). What follows here is a discussion of C. albicans recognition by the host during commensalism in the GI tract specifically, with particular attention given to the PRRs and cytokines that are involved. Disagreements on these topics in published experimental research articles will also be described.
Of the PRRs responsible for detecting fungal ligands, the C-type lectin receptor, Dectin-1 (also known as CLEC7A) has received considerable attention. This receptor is considered by some as essential for responding to systemic but not GI infections with C. albicans (Vautier et al.2012). Dectin-1 is expressed on myeloid cells (neutrophils, macrophages, dendritic cells) that are found in the lamina propria of the GI tract and recognizes β-1,3 glucans, which are part of the yeast cell wall. Following systemic infection with C. albicans, mice lacking Dectin-1 exhibited higher mortality rates, higher GI tissue colonization and a greater C. albicans load in stool when compared to WT mice with a functional Dectin-1 receptor (Vautier et al.2012). The cytokine profile of Dectin-1-deficient animals was also altered, particularly in the stomach and colon. However, there was no major alteration of inflammatory parameters in the Dectin-1-deficient animals following systemic C. albicans infection, but evidence for tissue invasion and altered bile salt production was recorded among these animals. Under conditions of direct oral GI inoculation with these Dectin-1-knockout animals, no difference in the stool load nor the tissue colonization levels of WT nor knockout animals was recorded, even when mice were cohoused (and therefore subjected to identical antibiotic and C. albicans dosing regimens). This observation held true, even when two clinical C. albicans isolates, which had been recovered from mucosal infections, were used instead of SC5314, which is suspected to poorly colonize GI tissues. Candida albicans elimination from the GI microbiota was also rapid in both WT and Dectin-1-deficient mice in the absence of antibiotic treatment. Cytokine responses were not affected by the Dectin-1 deficiency in mouse models of direct GI inoculation, except for IL-4, which although expressed at low levels overall (<10 pg ml−1 protein) was slightly, though significantly, more abundant in the stomachs of the mice lacking this receptor on day 14 following the introduction of C. albicans (Vautier et al.2012).
The findings of Vautier et al. (2012) suggest that Dectin-1 does not play a role in host-immune responses to C. albicans carriage in the GI tract; however, this contradicts an earlier study (Galès et al.2010) in which Dectin-1 was found to play a role. In the earlier study, mice were bred with a macrophage-specific Dectin-1 deficiency. They were also treated or not with a PPAR-γ stimulatory ligand, rosiglitazone. PPAR-γ is a nuclear receptor, stimulation of which (by interleukin-13 or rosiglitazone) activates anti-fungal responses. Following oral inoculation, greater C. albicans loads were recorded in the stomachs and ceca of mice with the macrophage-specific Dectin-1 deficiency and which were treated with rosiglitazone (a PPAR-γ ligand) than in the control mice. Furthermore, in the absence of rosiglitazone treatment, C. albicans levels were higher in the stomachs and ceca of macrophage-specific Dectin-1-deficient mice than in the controls. Subsequent treatment with rosiglitazone resulted in C. albicans clearance only in the control mice, thus emphasizing the role of Dectin-1 in the antifungal host response towards C. albicans. The mannose receptor on macrophages also plays a role in the host's antifungal response, and its expression on macrophages was increased in response to rosiglitazone treatment, but without stimulating C. albicans elimination. This implies that the mannose receptor alone is insufficient to clear C. albicans infections of the GI tract (Galès et al.2010). Vautier et al. (2012) reconcile the differences between their findings and those of Galès et al. (2010) by referring to differences in experimental design (presence/absence of antibiotic treatment or cohousing of WT and mutant mice). They also suggest that microbiota differences between mouse colonies could have a contributory role (Vautier et al.2012).
Critically, the genetic background of the mouse line has been shown to affect the Dectin-1-knockout phenotype (Carvalho et al.2012), and this may affect experimental outcomes and reproducibility across laboratories and studies. Differences in innate and adaptive immune responses occur. C57BL/6 Dectin-1-deficient mice were susceptible to infection with C. albicans. These mice harbored a higher fungal load in the stomach, showed evidence of dissemination to the kidneys and histological analysis revealed considerable inflammation and tissue invasion, when compared to WT mice. In contrast, BALB/c Dectin-1-knockout mice harbored a lower fungal load in the stomach and showed no signs of dissemination or inflammatory cell recruitment (Carvalho et al.2012). Furthermore, the Dectin-1 deficiency resulted in different cytokine profiles upon C. albicans challenge in each of the different mouse lines. Proinflammatory cytokines, tumor-necrosis factor-α and interleukin-6 were higher and lower in C57BL/6- and BALB/c- Dectin-1-knockout mice, respectively, when compared to the WT controls. In addition, the interleukin-17 family cytokines were more highly expressed in the BALB/c Dectin-1-deficient mice than in the controls, and these cytokines were produced at lower levels in the C57BL/6 Dectin-1-knockout animals than in the WT mice. These differential responses were also extended to the Ahr/IL-22 signalling pathway in these different mouse lines, with these responses being either suppressed or activated in C57BL/6- and BALB/c- Dectin-1-deficient mice, respectively (Carvalho et al.2012).
Dectin-1 deficiency has also been shown to regulate the severity of chemically induced colitis in mice (Iliev et al.2012). Moreover, a SNP variant of CLEC7A (a Dectin-1 homolog) was found to be associated with a severe ulcerative colitis phenotype in humans (Iliev et al.2012). This receptor variant was not associated with the establishment, but rather the severity of disease.
Human genetic polymorphisms may impact the outcome of C. albicans infections, as is well demonstrated in the example of chronic mucocutaneous candidiasis (CMC). This disease involves persistent or recurrent infections of the mucosae of the oral and genital tracts, the skin and nails with Candida species, often C. albicans, and typically affects individuals with T-cell deficiencies (reviewed in Cypowyj et al.2012). The interleukin-17 (IL-17) family is a group of six cytokines (IL-17A to IL-17F) (Gaffen 2009) which play a significant role in the host's anti-fungal response (Gladiator and LeibundGut-Landmann 2013). Several IL-17 receptors (IL-17RA to IL-17RF) are known in humans (Gaffen 2009). IL-17A and IL-17F, which are principally secreted by T cells, play a role in the resistance to mucocutaneous infections by pathogens (Gladiator and LeibundGut-Landmann 2013).
Impaired IL-17 immunity may increase an individual's susceptibility to CMC, and genetic studies have identified several IL-17 deficiencies as being responsible for CMC in humans. For example, a homozygous, premature stop codon which abolished IL-17RA receptor expression and the subsequent cellular responses to IL-17A and IL-17F homo- and heterodimers was responsible for CMC in a child of consanguineous parents (Puel et al.2011). Moreover, a dominant hypomorphic missense mutation in the IL-17F gene negatively impacted the formation of homo- and heterodimers involving this isoform (Puel et al.2011). In particular, receptor binding by complexes involving IL-17F was impaired. Variations in genes such as CARD9 and STAT1 or STAT3, which act downstream of the PRRs that detect fungal ligands, and which impact IL-17 T-cell development are also involved in susceptibility to CMC (Cypowyj et al.2012). Despite these findings from human studies, it has recently been suggested that IL-17-dependent immune responses are not involved in the regulation of GI tract colonization by C. albicans in mice (Vautier et al.2015). Mice lacking either the IL-17A or the IL-17RA genes were inoculated with a C. albicans strain (MBY38; tetO-UME6) in which filamentation can be readily induced by the withdrawal of doxycycline from the system. Typically, C. albicans filamentous forms colonize the GI tract at lower levels than the yeast form cells, thus induction of the switch from the yeast to the filamentous morphology in vivo was expected to result in reduced C. albicans colonization levels. Indeed, for both WT and mutant mice, the recovery of C. albicans from stool was reduced upon induction of filamentation. However, there was no difference in the kinetics of colonization in either the WT or knockout mice lacking either the IL-17A or the IL-17RA genes during or after doxycycline treatment, and the C. albicans load in stool from these different treatment groups was deemed similar, implying that neither IL-17A nor IL-17RA influenced GI colonization by C. albicans (Vautier et al.2015). Other authors have also acknowledged that members of the IL-17 cytokine family, including IL-17A and IL-17F, are perhaps not essential for anti-fungal immunity; however, these authors have shown that IL-22, an IL-10 family cytokine that is also secreted by Th-17 cells, is required for early-stage fungal resistance in mice lacking IL-17RA (De Luca et al.2010).
Clearly, a complex network of immune mediators is involved in the host response to C. albicans and the outcome of infection is likely to be influenced by the site and stage of infection, as well as by the immune status of the host (Romani 2011). Given that carriage of C. albicans is typically benign in immunocompetent hosts, host immunity clearly plays a major role in preventing infections with this yeast.
The widespread presence in healthy humans of antibodies against serodominant C. albicans antigens including those associated with virulence traits is indicative of an ongoing commensal-host interplay (Mochon et al.2010). It has been suggested that the strong humoral response of healthy individuals towards these serodominant antigens could be one way in which colonization and outgrowth by C. albicans is controlled in these individuals. Nonetheless, these authors acknowledge that the strong IgG response of this healthy cohort could also reflect their prior exposure to C. albicans during a superficial infection, such as vaginitis (Mochon et al.2010).
CONCLUSIONS AND OUTLOOK
The yeast C. albicans is a common member of the human gut mycobiota with a worldwide, though variable, distribution (Yatsunenko et al.2012; Angebault et al.2013). Even so, the interactions of C. albicans with the enteric yeast community and with the gut microbiota in general are presently quite poorly understood. While the bacterial component of the human gut microbiota has been studied under various healthy and disease states, the mycobiota and its potential role in host health and well-being has been largely overlooked. In addition, few studies have attempted to access or assess the fungal gene catalogue of the GI mycobiome.
Although the bacterial component of the microbiota is known for its ability to prime the immune system and maintain homeostasis in the intestine (Rakoff-Nahoum et al.2004), only very recently has any progress been made towards understanding the equivalent contribution made by the viral (Kernbauer, Ding and Cadwell 2014) or indeed the eukaryotic communities present. Given their many surface MAMPs which can be recognized by several different classes of host PRR, yeasts are very likely to contribute. A better understanding of the interactions between the colonizing enteric yeasts and their hosts could elevate the role of these yeasts from commensals to symbionts, engaged in the promotion and the establishment of host health.
An understanding of the gene networks involved in either commensalism or pathogenesis in C. albicans could provide candidate genes or signalling pathways that could be targeted to prevent the transition to pathogenesis. As such, investigations of the expression and regulation of these signalling networks during in vivo commensalism and infection are worthwhile, and many forward genetics screens have already yielded considerable insight into the genes regulating C. albicans commensalism in the GI tract. The next step will be to decipher the downstream targets of these regulators and to understand the mechanisms by which they function.
Millennia of coevolution underscore the relationship between the human host and its commensal microbiota, and C. albicans is thought to have been an early colonizer (though it could also represent a more recent arrival) (Odds 2010). Although its pathogenicity has stimulated much of the research on C. albicans to date, its commensal form is also worthy of attention, given its potential for immunomodulation, microbiota interaction and its contribution of metabolites and biomass to the gut environment.
The authors thank Natacha Sertour and Sadri Znaidi for advice, Martin Frank and Mariachiara DiMatteo for etymological assistance and Muriel Derrien for critical reading of the manuscript.
FUNDING
This work has been supported by grants from the Agence Nationale de la Recherche (KANJI, ANR-08-MIE-033–01; ERA-Net Infect-ERA, FUNCOMPATH, ANR-14-IFEC-0004), Danone Nutricia Research, and the French Government's Investissement d'Avenir program (Laboratoire d'Excellence Integrative Biology of Emerging Infectious Diseases, ANR-10-LABX-62-IBEID; Institut de Recherche Technologique BIOASTER, ANR-10-AIRT-03) to CdE and by grants from DIM Malinf – Région Ile-de-France to MEB. BAN was the recipient of a post-doctoral fellowship from Danone Nutricia Research.
Conflict of interest. None declared.
REFERENCES
