Introduction

Candida albicans (CA), the most common human fungal pathogen, manifests as a number of distinct infectious disease phenotypes including a mucosal infection (oral candidiasis), localized organ infections (dermatitis or vaginitis), chronic or persistent infection (chronic mucocutaneous candidiasis), and acute invasive/disseminated infections, which will be the focus of this review.

As with many invasive infections, the first or antecedent step of CA invasive infection requires colonization of a host mucosal surface. In fact, 95% percent of all infectious agents enter through mucosal surfaces, most notably the linings of the respiratory, gastrointestinal (GI), and genitourinary tracts [1]. Interestingly, humans are considered a natural reservoir for CA, with the genitourinary tract and GI tract being the main repository of CA. CA GI colonization alone does not induce a pathophysiologic state in either humans or other mammals (e.g., mice)—despite reports in the lay press and internet attributing symptoms of fatigue and malaise to “Candida overgrowth.” CA is often classified as a commensal organism, in that it does not provide any known direct benefit to the mammalian host but itself likely benefits itself from possible nutrient access and a host niche or reservoir. CA is better defined as a pathobiont, a potentially pathological organism which, under normal circumstances, lives as a commensal or symbiont [2].

The major concern for pathobionts residing in the GI tract is that these microbes will translocate to extraintestinal organs, notably the liver and spleen, and ultimately in the bloodstream. In cancer and stem cell transplant patients, CA colonizes the GI tract with subsequent translocation into extraintestinal organs in the setting of chemotherapy-induced neutropenia and GI mucosal damage [3]. In these patients, the role of the gut as a source for disseminated candidiasis was first suggested by older autopsy studies [4] and recently substantiated by molecular methods [5•]. Interestingly, Candida parapsilosis bloodstream isolates do not correspond to rectal isolates [5•], confirming prior studies suggesting that C. parapsilosis infections do not originate from the GI tract [4] and also highlighting the important of Candida species-specific differences.

Three factors are critical for preventing both bacterial and fungal GI translocation in humans and mice: (1) a balanced gut microbiome, with low abundance of pathobiont bacteria and/or fungi; (2) robust intestinal barrier function, and (3) intact cellular immunity, particularly neutrophils [6,7,8,9]. It is becoming evident that all three factors can be modulated by (1) the host (innate and adaptive immune responses, mucosal and cellular immune responses); (2) the pathobiont itself, specifically genetic determinants that promote either GI colonization or dissemination; and finally, (3) the gut microbiome and its effect on both the host and the pathobiont.

Therefore, understanding the conditions both in CA and in the host that promote the change from commensal to pathogen could lead to significant insights into CA invasive infection pathophysiology. Hence, this review will focus on recent insights regarding (1) CA genetic determinants; (2) host factors, including the gut microbiome; and (3) environmental factors that modulate CA’s ability to colonize the mammalian GI tract.

CA Genetic Determinants of GI Colonization in the Mammalian Host

Morphogenesis

CA is a dimorphic fungus that can exist in the yeast (round/oval) and the filamentous/hyphal form. The ability to transition between the yeast and hyphal form (morphogenesis) has been intimately tied to CA virulence (e.g., CA mutants unable to filament are less virulent) [10]. Our group has shown that filament-locked CA is actually less virulent in a murine model of CA GI colonization and dissemination after immunosuppression [11••]. The decreased virulence in our preclinical model, however, was most likely secondary to a deficiency in GI colonization of the mutant (2–3 log fold lower than the wildtype CA strain), as both bacterial and fungal GI translocation is directly proportional to microbial gut burden [11••, 12••, 13]. These data suggest that the hyphal form of CA may be less suited for GI colonization.

Indeed, a recent study showed that CA almost uniformly adopts the yeast form in germ-free mice [14••]. Interestingly, CA mutants lacking the transcription regulator genes ZCF8, ZFU2, and TRY4 had reduced fitness in the GI tract of germ-free mice, which was attributable to a predilection for the filamentous form. Finally, ZCF8, ZFU2, and TRY4 promote CA adherence in the gut in a mucin-dependent fashion. As for a teleological explanation as to why the yeast form would be preferable in the mammalian GI tract, the authors postulate that the hyphal or filamentous form may be more immunogenic and thus less beneficial for the survival of CA in the gut—supported by the observation that a host immune effector (granulocyte colony stimulating factor) was significantly more abundant in intestinal tissue in mice colonized with filamentous CA compared to counterparts colonized with the yeast form [14••].

GUT Morphology

In terms of other CA-specific adaptive responses which would promote GI colonization, an unusual morphology, termed the gastrointestinally induced transition (GUT) phenotype [15••, 16], has been described as a specialized form of CA adapted to the GI tract. Noble and colleagues noted that the introduction of CA into the mammalian GI tract triggers a developmental switch, driven by the Wor1 transcription factor, to this commensal cell type. CA Wor1 deletion mutants showed a significant GI fitness disadvantage compared to wildtype CA. Overexpression of Wor1 resulted in a gain of function phenotype and a competitive advantage over the wildtype strain. Of note, Wor1 had previously been shown to be important for controlling CA white-opaque switch in mating [17, 18], and only rare cell types had been shown to be competent for WOR1 expression in vitro. Overexpression of Wor1 has also been shown to increase susceptibility to bile salts, which may explain a transient defect in the initial stages of GI colonization observed by different groups [16, 19] and also enhanced adhesion to the murine GI mucosa [19], perhaps explaining the competitive fitness advantage exhibited by the mutant in later stages of GI colonization. These data suggest that the GI tract environment may induce CA phenotype changes, such as the GUT phenotype, as an adaptive response to mammalian GI tract environmental signals and cues.

Candidalysin

GI epithelial damage is a critical factor required for CA GI translocation [11••, 12••]. Recently, the first fungal cytolytic peptide toxin (candidalysin) in CA was identified [20••]. Candidalysin is a short 31 amino acid long peptide generated from the hyphae-associated cell elongation 1 gene (ECE1). Interestingly, CA strains lacking candidalysin do not activate or damage epithelial cells and have a colonization defect in an animal model of oral candidiasis [20••]. These same mutants, however, do not exhibit a colonization defect in a mouse model of CA vulvovaginitis [21]. With regard to the lower GI tract, the importance of candidalysin in both colonization and dissemination is unclear. Candidalysin does, however, appear to be critical for translocation through intestinal epithelial cells in vitro and appears to have a direct effect on gut microbiota (personal communication, Bernard Hube). Of note, the three host factors that are critical for preventing both bacterial and fungal GI translocation in humans and mice include (1) balanced gut microbiome, with low abundance of potentially pathogenic bacteria and/or fungi; (2) intact intestinal barrier function, and (3) intact cellular immunity, particularly neutrophils [6,7,8,9]. Our group has shown that pathogenic bacteria (e.g., Pseudomonas aeruginosa) unable to induce gut intestinal damage (e.g., mutants unable to produce type III secretion exotoxins) are unable to translocate from the mouse GI tract despite being highly abundant in the gut and the host being severely neutropenic [22]. Thus, candidalysin may be a critical factor in determining CA’s ability to translocate from the mammalian GI tract.

Host Factors Promoting CA Colonization Resistance

Gut Microbiome

The concept of colonization resistance, the notion that the gut microbiota promotes colonization resistance to pathogens, was first noted over 50 years when mice treated with antibiotics developed Salmonella infection with an inoculum 100,000-fold less than required for untreated mice [23]. The relevance of colonization resistance to fungal infections has been well-established in the medical literature that treatment with anti-bacterial antibiotics can lead to the development of “yeast, infections,” specifically vulvovaginal candidiasis [24]. While the importance of the gut microbiota for providing protection against these infections was strongly suggested by these data, the underlying mechanisms of colonization resistance have only recently been elucidated.

One important mechanism by which commensal bacteria promote colonization resistance is via induction of host GI epithelial immune effectors. Our group leveraged the observation that mice are naturally resistant to CA GI colonization but can be colonized after administration of antibiotics. We used different clinically relevant antibiotics to induce variable CA GI colonization phenotypes [12••]. Interestingly, antibiotics most effective in depleting commensal anaerobes resulted in the highest CA GI colonization levels, comparable to levels seen in germ-free mice. Of note, the data supporting that mice are resistant to CA GI colonization, including our own, was generated using the popular laboratory CA strain SC5314. Various clinical CA isolates, particularly those recovered from the human gut, can colonize the murine GI tract without the use of antibiotics to various degrees—highlighting the importance of CA strain variability (A.Y Koh, unpublished observation).

Using gut microbiome profiling techniques, we were then able to identify specific commensal anaerobic bacteria which promoted reduction of CA GI colonization in germ-free mice [12••, 13]. Ultimately, we identified a mechanism by which specific commensal anaerobic bacteria induce the transcription factor HIF-1α, a key regulator of mammalian innate immunity [25], which then increases expression of the antimicrobial peptide LL-37/cathelin-related antimicrobial peptide (CRAMP) in intestinal epithelial cells. LL-37, which has activity against CA [26] and can also inhibit CA adhesion to epithelial surfaces [27]. By inducing HIF-1α, via the pharmacologic HIF-1α agonist mimosine, we were able to increase LL-37/CRAMP expression, reduce CA GI colonization, and decrease CA dissemination in mice, whereas these effects were nullified in mice lacking HIF-1α in their intestinal tissue [12••]. Not surprisingly, the protective effect of commensal anaerobic bacteria was abrogated in mice lacking HIF-1α or CRAMP.

What was most striking about these findings was that a 1–2 log-fold reduction in CA GI colonization was sufficient to significantly decrease CA dissemination or mortality. Similar findings have been reported with regard to bacterial dissemination from the gut [28]. These data suggest that complete eradication or absence of pathobiont GI colonization is not needed to achieve a significant decrease in dissemination. As to whether these findings are relevant to humans, an expansion of GI Enterococcus spp. or Enterobacteriaceae (along with a concomitant depletion of commensal anaerobic microbiota) in adult stem cell transplant patients is associated with a significantly increased risk of developing bloodstream infection with the same bacterial species [29, 30]. While there are data showing that CA bloodstream isolates recovered from patients are genetically similar to CA GI isolates from the same patient [5•], there are no data confirming that CA GI burden is directly proportional to the risk of invasive CA infection in patients.

Lactobacillus spp. probiotic therapies have been used with some success in both animal models and human patients to reduce CA GI colonization. For example, L. acidophilus can reduce the size of Candida-induced gastric ulcers and decrease Candida GI colonization in animals [31]. L. rhamnosus oral therapy induces significant reductions in Candida GI colonization in both premature babies [32] and elderly adults [33]. While these results are intriguing, one caveat that must be considered is that some Lactobacillus probiotic therapies have no effect on host gut microbiota composition or levels [34, 35]. Further studies using probiotics, particularly preparations including commensal anaerobic gut microbiota, will need to be conducted in order to determine whether precision probiotic therapy can modulate CA GI colonization in humans.

As to whether gut microbiota induce a direct effect on CA growth or colonization, we have shown that specific commensal anaerobic gut microbiota do not directly inhibit CA growth in vitro and vice versa [12••]. Furthermore, co-colonization with other pathobiont bacteria, such as P. aeruginosa or Escherichia coli, has no effect on CA GI colonization in germ-free mice [13]. Gut microbiota, however, also produce metabolites (e.g., short-chain fatty acids, SCFAs) that could potentially have a direct effect on CA. SCFAs inhibit CA yeast to hyphal transition [36] and inhibits its growth in vitro [37]. To further confound issues, however, SCFAs can also induce GI epithelial cells to produce immune effectors, such as antimicrobial peptides (e.g., LL-37) [38]. Ultimately, the maintenance of CA colonization resistance in the mammalian host is most likely dependent on both gut microbiota and gut microbiota-derived metabolite effects.

Host Immune System

While lymphocyte deficiency (e.g., patients with HIV and AIDS) results in oral and esophageal candidiasis [39, 40], lymphocytes do not appear to be important for modulating CA GI colonization, as noted by studies using athymic mice [41] and recombinase-activating gene-deficient mice [11••]. In contrast to the critical importance of neutrophils for controlling disseminated fungal disease [11••, 42,43,44], their role in modulating CA GI colonization appears to be negligible [11••]. Similarly, neither macrophages [11••] nor NK cells [45] appear to affect CA GI colonization.

The mammalian innate immune system utilizes pattern recognition receptors (PRRs) to recognize fungi. Dectin-1, a C-type lectin receptor PRR, has been shown to be critical for the control of fungal infections, including CA, in both mice and humans [46, 47]. Interestingly, dectin-1 is essential for the control of GI invasion or translocation during systemic infection in mice, manifested as impairment in fungal clearance and dysregulated cytokine production [48•]. Surprisingly, dectin-1, however, is not required for the control of mucosal colonization of the GI tract, in terms of either fungal burdens or cytokine response [48•]. In light of the importance of commensal gut microbiota in maintaining CA colonization resistance, there is both older [49] and more recent data [50, 51] to suggest that Toll-like receptors (TLR), specifically TLR2 and TLR4, are important for promoting CA colonization resistance. These data are consistent with studies showing that TLR4 and TLR5 are essential to for maintaining colonization resistance to pathobiont bacteria [28, 52]. In total, these data suggest that the mechanisms of maintaining colonization resistant to bacterial and fungal pathobionts may share common pathways: commensal gut microbiota signal through PRRs (e.g., TLRs) to induce gut epithelial immune effectors, most notably antimicrobial peptides (i.e., LL-37/CRAMP [12••, 26], alpha-defensins [53, 54], beta-defensins [55, 56]) that have activity against a variety of pathogens, including CA.

Environmental Factors

A major disparity in studies focused on CA GI colonization rests with the observation that mice are resistant to CA colonization, whereas 40–80% of humans are colonized with CA [57]. The human CA colonization data, however, is based on culture-based data from humans living in Western societies [58]. More recent studies of humans living in remote and traditional societies, however, exhibit widespread Candida GI carriage (e.g., C. krusei), but CA GI carriage rate of less than 10% [59••, 60]. Thus, CA might not be a “normal” commensal of the human gut, but a more recently acquired commensal resulting from medical advances (particularly antibiotics) and adoption of Western diets.

Antibiotics and Chemotherapy

The impact of antibiotics on CA GI colonization resistance cannot be overstated. As noted before, almost all murine models of CA GI colonization utilize antibiotics to establish sustained CA GI colonization, but the CA GI colonization levels achieved can vary widely depending on the antibiotics used [11••, 61,62,63,64,65,66]. Antibiotics that are most effective in depleting anaerobic bacteria, which are the majority of commensal gut microbiota, are the most effective in promoting high CA GI colonization levels. In fact, penicillin (and not clindamycin or metronidazole) has been shown to be most effective in depleting endogenous murine anaerobic gut microbiota and promoting overgrowth and translocation of Enterobacteriaceae [67]. Of note, mice treated with penicillin achieve CA GI colonization levels comparable to those seen in germ-free mice [12••]. These data further suggest that commensal anaerobic gut microbiota are essential for GI colonization resistance to both bacterial and fungal pathobionts and that the host mechanisms for maintaining pathobiont GI colonization resistance may utilize redundant functional pathways and/or strategies.

A recent study screened more than 1000 marketed drugs against 40 representative gut microbiota (all bacteria) and found that a large number of non-antibiotic medications inhibited the growth of gut microbiota, including commensal anaerobes [68]. One of the major medication classes that had a negative effect on gut microbiota were cancer chemotherapy agents. Interestingly, there are prior reports suggesting that cancer chemotherapy can lead to reduced overall numbers of gut microbiota and also lead to changes in gut microbiota taxonomic composition in both preclinical models [69, 70] and human patients [71, 72]. The two classes of medications most frequently given to cancer and stem cell transplant patients are chemotherapeutic agents and antibiotics. This begs the question as to whether these medications contribute to the fact that these patients are at such high risk of developing CA invasive infections originating from the gut [3]. Unfortunately, there are no studies in either animals or humans that have examined the effect of cancer chemotherapeutic agents on the gut mycobiome; so, further studies are merited.

Diet

One obvious environmental factor that differs between mice and humans is diet. Thus, if a mouse were to adopt a human “Western society” diet, could the mouse be colonized with CA without the use of antibiotics? In fact, when mice are a fed “purified” diets consisting of significant amounts of cornstarch, sucrose, and soybean oil [73, 74••], they can be colonized with CA in the absence of antibiotics, albeit at levels 1–2 log fold lower than seen in antibiotic-treated or germ-free mice. There is some data to suggest that the use of this specific diet leads to gut microbiota taxonomic changes that results in decreased amounts of gut microbiota-derived organic acid, which as noted previously may have a direct inhibitory effect on CA [74••]. Yet, it is still unclear whether the effects of the diet on CA GI colonization resistance are more a result of gut microbiota taxonomic changes (and thus changes in host immune effectors) or whether this may be due to direct effects on CA metabolism secondary to the increased availability of refined carbohydrates and fat. Further studies are needed to dissect the mechanisms by which dietary changes can modulate CA GI colonization resistance in the mammalian host.

Conclusion

In sum, the regulation of CA GI colonization and dissemination is a dynamic and complex process that involves CA genetic determinants, host factors (including the gut microbiome), and environmental elements, such as exposure to medication and dietary choices (Table 1). Commensal microbiota outnumber mammalian host cells by a factor of 10:1, with the microbial genetic repertoire 100-fold more abundant than the host. The gut microbiome contributes to the metabolic, nutritional, and immunological status of the mammalian host. As such, the field of host-pathogen interactions must now include host-pathogen-commensal interactions, which inherently increases the complexity of exploring these pathophysiologic processes. But given the advances in molecular biology approaches and multi-omic analyses, we can now pursue more mechanistic insight into how and why the commensal CA transforms into the pathogenic CA and thus provide the platform for innovative diagnostic and therapeutic approaches to preventing harmful CA infections in the future.

Table 1 Summary of CA genetic determinants, host factors, and environmental factors that modulate CA GI colonization and dissemination in the mammalian host
Full size table