From birth and throughout life: Fungal microbiota in nutrition and metabolic health.

Abstract

The human gastrointestinal tract is home to a vibrant, diverse ecosystem of prokaryotic and eukaryotic microorganisms. The gut fungi (mycobiota) have recently rose to prominence due to their ability to modulate host immunity. Colonization of the gut occurs through a combination of vertical transmission from the maternal mycobiota, environmental and dietary exposure. Human and animal data demonstrate that nutrition strongly affects the mycobiota composition and that changes in the fungal communities can result in the aggravation of metabolic diseases. The mechanisms pertaining to the mycobiota’s influence on host health, pathology, and resident gastrointestinal communities through intra-kingdom, trans-kingdom and immune crosstalk are beginning to come into focus, setting the stage for a new chapter in microbiota-host interactions. Herein we present an examination of the inception, maturation, and dietary modulation of gastrointestinal and nutritional fungal communities with an inspection on their impact on metabolic diseases in humans.

Introduction

A diverse community of fungi coexists in the gut with other members of the microbiota (75, 99, 103, 128). In humans and mice, the gut mycobiota is dominated by members of the Ascomycota phylum that includes – among the most prevalent taxa – the genera Candida and Saccharomyces. A growing number of studies are uncovering how gut commensal fungi can influence host immunity in a number of immune-mediated diseases (75). However, recent findings implicate a role for gut fungi beyond their direct effect on the host immune responses. A complete understanding of the role of gut fungi in health and disease requires an effort to integrate different aspects of their interaction with the host immunity, the gut microbiota, and the host metabolism. Herein we review recent and historical perspectives on the reciprocal effects between fungi and human nutrition with a focus on the mycobiota’s ability to influence the course of gastrointestinal metabolic disease.

Early-life human mycobiota: Engraftment, breast milk and the transition to solid foods

While the establishment of a stable bacterial microbiota early in life has been linked to health later in life (4, 93), key questions about the pediatric sources of mycobiota engraftment and the importance of intestinal fungi in the context of disease development are just beginning to be investigated. A trend of increased fungal 18s rRNA in vaginal versus cesarean section birth has been observed, persisting until about one-year post delivery and mimicking what is seen for bacterial diversity(123). Principal coordinate analysis based on 18S rRNA sequencing revealed a cluster of exclusively vaginally delivered children with higher relative abundance of Candida spp., indicating that birthing method plays a role in setting mycobiota diversity on a genus level in infants. In very low birth weight neonatal infants, a population susceptible to gastrointestinal candidiasis, vaginal delivery is correlated with triple the risk of C. albicans colonization within one week of birth(15). Moreover, 76% of the mothers of these early-colonized infants were themselves colonized (compared to 56% of other infants), suggesting direct maternal transfer of fungi. Vertical transmission was confirmed in 65% of early colonization instances at birth by DNA fingerprinting, which dropped to only 41% of total colonizations 2 weeks after birth, indicating that mothers can be vectors for vertical mycobiota transfer with environmental transmission playing an increasingly significant role after the first few weeks of birth. Subsequent work has confirmed similar vertical transmission of obligate anaerobic bacteria(57), establishing a common, validated route for maternal microbial transfer.

Following birth, children are offered their first nutritional choice, namely: breast milk or formula, providing another opportunity for pediatric inoculation of microbes including fungi (Fig. 1). Several studies have investigated the fungi residing in breast milk, offering differing perspectives on the composition of the breast milk mycobiota based on both culture-independent sequencing approaches and fungal culturomics. Callado and colleagues reported a milk mycobiota consisting of Malassezia, Davidella, Sistotrema and Penicillium genera which varied only slightly across the four nations sampled (Finland, South Africa, China and Spain)(17). Additionally, 18S quantification of fungal loads were consistent geographically. This collection of genera is in contrast with an earlier investigation by the same laboratories which identified Candida, Malassezia and Saccharomyces as the highest abundance genera across breast milk samples(16). A more recent study of mothers of preterm infants corroborated a portion of these results showing a predominance of Candida spp. (C. albicans, C. parapsilosis, C. krusei and C. glabrata), and Saccharomyces cerevisiae by ITS2 amplification and sequencing(50). Breast milk was found to be devoid of cell-free fungal DNA, indicating that isolated sequences represent a fungal cell-derived community. Samples from this study were completely devoid of the genus Malassezia, common skin commensal fungi, suggesting a difference either in the donor populations or in the collection techniques across studies that might influence the extent of Malassezia shedding from the skin where this fungus is highly abundant. Furthermore, fungal isolates obtained from culturing breast milk were dominated by C. parapsilosis and Rhodotorula mucilaginosa. Interestingly, researchers have been unable to isolate Malassezia from breast milk samples despite high abundance by culture-independent identification. This suggests that live Malassezia is either not present in the breast milk or may be reflective of the culture and collection conditions implemented, although the latter is less likely as Malassezia culturing techniques are historically very well established (16, 17). Controlled studies comparing the milk, nipple and areola mycobiomes with a wider range of culture media and temperatures are needed to further delineate genuine inhabitants of lactation from serendipitous dermis carryover organisms and cell-free DNA. For bona fide inhabitants of breast milk, questions remain as to carbon and nitrogen sources available to fungi in this biological medium(16); their recognition by maternal immunity; and the time course for fungal expansion throughout pregnancy, after delivery, and after cessation of breast feeding.

Figure 1: Dietary factors affecting the gastrointestinal mycobiota throughout life.

Dietary modulation of the mycobiota begins at birth, where formula or breast milk serve as primary fungal inoculum. Transition to solid food and adherence to either a plant-based or ‘Western’ diet (high in simple carbohydrates, fat and protein) result in drastic alteration of both bacterial and fungal communities within the gut. Finally, the production of a range of fermented foods, or their contamination by fungi often leads to selection of a specific nutritional microbiota prior to ingestion, altering the consumer’s native intestinal microbiota.

The mycobiota transition from neonate to toddler occurs during a formative life period, when nutrition, metabolism, and immunity undergo rapid change and development. A progression from a high Malasseziales abundance to an increase in Saccharomycetales is observed between 1–5 months of age, a common timeframe for the introduction of soft, non-milk foods(40). In terms of bacterial populations, a decrease in both Enterobacteriaceae and Bifidobacteriaceae and increase in Lachnospiraceae is observed during the same time period, suggesting a concerted, trans-kingdom shift. A distinct gut microbiota state was identified in neonates with higher Candida spp. abundance and this was associated with an increase in allergen-specific IgE concentrations at 2 years of age. This suggests that a neonatal gut conducive to or containing Candida spp. expansion may influence atopy severity and asthma development later in life, an ability previously established for a defined group of intestinal bacteria(4). These findings are further corroborated by studies in mice showing that fungal dysbiosis and colonization with specific fungi can negatively influence the outcome of allergic airway inflammation both early and later in life(74, 104, 106, 118). Another cohort showed a steady increase of fungal alpha-diversity from 10 days to two years of age(102). The OTUs with the most relative abundance during this time period transitioned from Debaryomyces spp. (during formula or breastfeeding) to Saccharomyces spp. (with weaning and solid food). Detection of fungal DNA was correlated between mother-child pairs, with children of fungi-positive mothers being more likely to also be positive. This is possibly indicative of a maternal, bacterial community transfer more conducive to fungal colonization later in life. Further inspection of this critical transition period will require additional longitudinal studies, with a particular focus on species/strain acquisition and persistence.

Dietary modulation of intestinal fungi in adulthood

In contrast to the similarities between individuals and the longitudinally consistent nature of the bacterial microbiota, the adult intestinal mycobiota has been documented as being extremely unique and variable as made evident through both geographically diverse and longitudinal fecal material sequencing (34, 43, 45, 88, 109). Dietary and nutritional differences in cohorts could potentially explain some of the interindividual variability. In a pair of studies in the same geographical region, Hallen-Adams and colleagues evaluated the mycobiota of healthy individuals on conventional (Western) and plant-based diets (45, 109). While conventional diet seems to support the existence of a preferred ‘core mycobiota’ consisting of Debaryomycetaceae (Debaryomyces and Candida spp.), Dipodascaceae (Galactomyces spp., Geotrichum spp.), Davidiellaceae (Cladosporium spp.), and Malasseziaceae (Malassezia spp.) family members; the vegetarian mycobiota favors a greater abundance of spore-forming and dietary fungi (Fusarium spp., Penicillium spp. and Aspergillus spp., Cladosporium spp.). This alteration could be explained by a reduction in yeast-like fungal loads in the intestine, allowing for the observed domination of dietary and environmental fungi; a genuine expansion of filamentous fungi in the gastrointestinal tract; or increased intake of allochthonous mycobiota from a plant-based diet. Turnbaugh and coworkers further explored the changes in the human mycobiota induced by a plant- or animal-based diet (32). While Penicillium species seemed to flourish after 2–3 days of the animal-based diet (rich in cheeses, meat and eggs), Debaryomyces and Candida spp. were drastically reduced in relative abundance. In contrast, a plant-based diet (chiefly fruit, grains and vegetables) showed a corresponding increase of a diet-associated Candida sp. While this study makes interesting connections between the diet-derived fungi and the fecal mycobiota, interpreting abundance changes is confounded by the lack of dietary controls with minimal fungal components. Lewis and coworkers followed up on their seminal study examining bacterial-diet interactions in healthy individuals(132) to investigate the archaea and fungal members of the gastrointestinal microbiota(52). Specifically, they revealed a positive correlation between recent dietary intake of carbohydrates and Candida species while saturated fatty acids resulted in a negative correlation. When expanded to long-term dietary choices, these patterns resulted in trends in the same directions without statistical significance, possibly reflecting subtle changes in the diet not equated for in the volunteer diet questionnaire and highlighting challenges involved in conducting and enforcing a controlled nutrition study for extended periods of time.

Examining the impact of controlled, therapeutic diets on the intestinal mycobiota offers a unique vantage point to gauge the potential therapeutic benefit or detriment of these organisms to select patient populations. A recent examination of pediatric patients with inflammatory bowel disease uncovered a significant negative correlation between Crohn’s disease (CD) patient initiation of exclusive enteral nutrition (EEN), a first-line dietary treatment for the disease, and the relative abundance of Candida albicans, Clavispora lusitaniae and Cyberlindnera jandini –organisms that are typically increased in CD patients(72). Although the relative abundance of fungi in this cohort was not informative in predicting patient response to anti-TNF treatment, EEN patients maintained significantly decreased abundances of these fungi at the 8-week post-dietary modulation follow-up. Enteral nutrition is a broad term encompassing both elemental (amino acid-based) and non-elemental (isolated protein-based, such as casein) which past studies have determined have equivalent efficacy in the treatment of CD(87). Notably the diet is routinely low in both complex carbohydrate and fiber sources, correlating well with past nutritional studies with the Candida genus(52). Whether the observations of these preliminary studies are due to direct depletion of pro-fungal nutrients or indirect mediation through host immunity or bacterial microbiota alterations remains to be discerned. Looking ahead to future studies, the impact of therapeutic diets such as low FODMAP (for IBS patients(46)) - as well as regional or cultural diets may lead to further insight into the precise mechanism by which distinct elements within the diet alter the gastrointestinal mycobiota.

To facilitate assessment of mycobiota changes in a more controlled setting and possibly reveal individual the dietary components responsible, murine controlled-diet experiments offer an attractive alternative to human nutrition studies. Gale and colleagues evaluated changes in ITS2-fungal abundances following a high-fat diet in mice(49). After 8 weeks of diet, high-fat diet mice showed a significant decrease in the genus Saccharomyces (specifically S. cerevisiae) while Candida showed no change compared to conventional murine diet. Quantitative PCR of both diets revealed that species present in murine stool were overwhelming (80–90%) absent from the diet, effectively controlling for changes simply due to transient microorganisms. Importantly, S. cerevisiae represented <0.01% of diet sequences in the diets (compared to >80% in feces), supporting the notion that a high-fat diet induces a genuine outgrowth of the species within the murine gastrointestinal tract. Another study carried out by Ludwig and coworkers(122) examined the effect of both a Western diet (high sugar, cholesterol and fat) and caloric restriction compared to a conventional murine diet. While they observed only a marginal change in beta-diversity between regimens, an expansion in Wallemia spp. and the Basidiomycota genus Sporobolomyces was observed with western diet and caloric restriction, respectively. Interestingly, Candida spp. were highest in abundance with conventional diet compared to both alternative treatments. Importantly, the ability to manipulate increasingly intricate dietary components (e.g. sources of fiber, phenolic compounds, phytosterols and terpenes) in future studies may yield a new level of mechanistic insight into the observed effect of whole foods in the human diet.

Dietary fungi and alteration of the microbiota

Just as nutrition and diet may impact the mammalian mycobiota, so too can fungal dietary components modulate the bacterial and fungal communities in food preparation and the gastrointestinal tract. An early investigation of antagonist dietary fungi-fungal symbiont interactions revealed a negative sequencing correlation (Spearman coefficient of -0.31) between Debaryomyces hansenii, a fungus common in fermented products such as cheese(12), and Candida yeasts in feces(11). Isolation from 44 different cheeses and coculture with a variety of Candida species revealed several D. hansenii strains that produced peptide antifungal natural products or yeast killer toxins (YKTs). Yeast killer toxins, also known as mycocins, are antimycotic peptides produced by a variety of fungi in a similar manner as bacteriocins in bacteria. Inhibitory activity was found to be strongly dependent on temperature (most potent at 25 °C) and acidity (favoring a pH of 4.5). Although these properties cast doubt on the role of D. hansenii-derived YKTs in the small and large intestines, they do point to a plausible mechanism by which fermented products exclude Candida species, diminishing the inoculum of these opportunistic pathogens from dietary sources. Moreover, YKTs from Wickerhamomyces anomalus, a yeast previously identified in the normal murine mycobiota(34), have very recently been shown to be potent against Candida mesorugosa over a wide range of physiologically-relevant acidities (pH of 3–6) and temperatures (5–40 °C), increasing the possibility that YKTs play a role in directing fungal-mycobiota interactions within the mammalian intestinal tract(115). YKT susceptibility in C. albicans was aggravated by deletion of Hog1, a gene encoding a member of the MAPK phosphorylation pathway(82). Interestingly, Hog1-deficient mutants were previously shown to be more resistant to Congo red and calcofluor white, agents that bind to chitin in both yeasts and filamentous fungi(3), suggesting that alteration of the fungal cell wall increases susceptibility to pore-forming YKTs from D. hansenii. Another study determined that Pichia, a commensal genus found in a variety of fermented products such as miso soup and alcoholic beverages, displayed an inverse correlation in terms of relative abundance with Candida spp. in the oral cavity of HIV patients(86). Spent culture broths from a mycological repository Pichia strain (MRL1458, putatively assigned as P. farinosa) were found to be inhibitory towards the growth of Candida, Fusarium and Aspergillus spp. The causative agent was found to be a proteinase K-sensitive, heat- and alkali-insensitive diffusible molecule: presumably an antifungal peptide. This result matches well with the 30-year-old literature surrounding the discovery of salt-mediated killer toxins from P. farinosa and elucidation of its inhibitory activity against a variety of fungi(111, 112). Together, these studies highlight the importance of YKT peptides in preventing fungal contamination and spoilage in fermentation processes and hint at a potential mechanism for direct dietary modulation of the mycobiota by fungi.

The mycobiota inhabiting the intestinal tract not only have the potential ability to manipulate their own community, but also to influence the vastly more numerous neighboring bacteria, possibly indirectly influencing the mammalian host in the process. The increasing interest in the use of fecal microbiota transplantation (FMT) for the treatment of various microbial-related disease highlights the complexity of gut trans-kingdom interactions. In Clostridioides difficile infections, high burdens of Candida are associated with a negative outcome of FMT that is otherwise highly efficacious in this context(92, 139). In contrast, high levels of Candida have been linked to an increased responsiveness in patients with Ulcerative Colitis (UC) (71). In CDI, where C. difficile is the predominant bacterium, Candida and C. difficile appear to have a mutualistic relationship, thriving in a pro-inflammatory environment and becoming refractory to FMT(92). In UC patients, that retain a complex microbial community, high Candida burden were associated with an increased bacterial diversity that persisted following FMT. In these patients, a reduction of Candida burden following FMT was associated with an ameliorated disease severity, again suggesting a pro-inflammatory role for Candida although the underlying mechanisms remain unclear (71). These recent studies suggest that the influence of Candida on the colonization and persistence of different bacterial species is highly influenced by the preexisting disease context within the intestine.

Bushman and colleagues reported the existence of a significant positive correlation between the presence of Candida spp., detected in 57% of the fecal samples from healthy volunteers, and the genus Bacteroides(52). This finding has been subsequently corroborated by multiple independent studies(45, 133). In an attempt to exploit the bacterial community-modifying properties of fungi, Pu and coworkers trained a strain of S. cerevisiae for probiotic usage in freshly weaned swine(133). Livestock are typically weaned at an early age and supplemented with a regimen of antibiotics to promote rapid growth and stave opportunistic pathogen infection due to an underdeveloped intestinal microbiota. Unfortunately, this has led to the rapid evolution of drug resistant bacteria through direct selection in agricultural animals and indirect effects from wastewater run-off from grazing fields and animal housing facilities. Piglets were treated with a diet incorporating S. cerevisiae alone, S. cerevisiae and fermented egg white, or a control diet; each without the addition of antibiotics during weaning. After weaning on the yeast-containing diets, the fecal bacterial communities of both groups of S. cerevisiae-fed piglets were altered in beta diversity with an increase in Bacteroides spp. as well as a decrease in several Firmicutes including the genera Roseburia, Faecalibacterium and Anaerovibrio. Strikingly, these findings are in stark contrast to the results obtained from feeding chitin-glucan to mice on a high fat diet, wherein Delzenne and coworkers observed a dramatic reversal of high fat diet-induced Firmicutes Roseburia spp. and Eubacterium spp. depletion accompanied by an aggravated reduction in Bacteroides spp.(89) This apparent discrepancy suggests that live fungi may have altered, or even opposing, effects on the gastrointestinal microbiota when compared to their fungal cell wall components. In piglets, S. cerevisiae supplementation increased daily weight gain, resulting in a more efficient feed intake while decreasing diarrhea rate and death rate substantially. Additionally, fungi with increased pathogenic potential may also affect the mammalian mycobiota. Murine gastrointestinal exposure to Mucor circinelloides, a food contaminant and culprit in yogurt poisonings, was found to induce significant changes in the bacterial alpha and beta diversity, resulting in the expansion of Bacteroides and Butyricimonas spp. and a reduction in Akkermansia spp. and the family Verrucomicrobiaceae(85). Clearly, intestinal exposure to resident, probiotic and pathogenic fungi broadly affects the host intestinal bacterial communities and their responsiveness to diseases and therapy. As a consequence, the impact of and mechanisms underlying alteration of the gut ecological landscape within the host has emerged as a burgeoning field of future research.

Diet-derived mycotoxins and their effect on the host

Far from being solely beneficial organisms, fungi (in particular filamentous, spore-forming fungi) have a long and notorious history of food contamination, causing harm to both humans and livestock via their toxic secondary metabolites. Though hundreds of mycotoxins have been reported in the past century, we will focus on three of the most common classes of secondary metabolites: bicyclic polyketide toxins, aflatoxins and trichothecenes (Fig. 2). These toxic small molecules are prototypically synthesized by Aspergillus, Penicillium and Fusarium spp. through elaborate, specialized enzymatic machinery and display toxicity through a variety of mechanisms. For more information on the plethora of additional mycotoxins derived from Alternaria spp. and other molds, we invite readers to explore the extensive preexisting body of literature on the subject(18, 53).

Figure 2: Diverse filamentous fungi-derived mycotoxins adversely impact health.

Ascomycota in food sources cause a range of deleterious health consequences in livestock and humans via mycotoxin production. The representative structures of three major classes of fungal toxins: aflatoxins, bicyclic polyketides and trichothecenes are highlighted in the figure and text, canonically derived from Aspergillus, Penicillium and Fusarium spp.

A series of simple, bicyclic polyketide products were isolated from Penicillium and Aspergillus molds beginning in the 1930s. Citrinin, the first such toxin, was discovered by Hetherington and Raistick from a characteristically yellow mold, P. citrinum, in 1931(51). The carbon backbone of crystalline, golden-colored citrinin is constructed through an iterative, non-reducing polyketide synthase (citS(pksCT)). The mature pentaketide intermediate undergoes a series of enzyme-catalyzed oxidation and reduction events followed by final cyclization and dehydration to the quinomethide toxin(48). Producers of citrinin typically infect nuts, cereals, fruit, rice and soybeans; posing a risk to both livestock and human health(91). Upon ingestion, citrinin accumulates in the mammalian kidney where it acts as a potent nephrotoxin, leading to oxidative damage through disruption of the mitochondrial respiratory chain(26, 119), although the precise mechanism of action is still under investigation.

Patulin, a natural product produced by Penicillium and Aspergillus spp., was discovered during a resurgence in fungal antibiotic discovery stoked by both Sir Alexander Fleming’s elucidation of penicillin in 1928 and the advent of World War II(38). As many as four independent laboratories may be credited with the first documentation of an antibiotic substance derived from an unknown Penicillium sp., P. claviforme, or A. clavatus from 1938–1942(6, 27, 117, 129, 130). Patulin’s bicyclic unsaturated lactone structure was deciphered by Woodward and Singh(131), and results from a fungal, non-reducing PKS that produces 6-methylsalicylic acid (6-MSA)(77). Critically, the study of this exact process in P. griseofulvum by A. J. Birch formed the basis of polyketide natural product chemistry(14). This product undergoes a decarboxylation reaction followed by a series of biosynthetic oxidation and cyclization steps to yield the mycotoxin patulin. Due to the intrinsic electrophilic nature of its dienoate moiety, patulin is reactive to Michael donors. Sulfhydryl groups of glutathione and cysteine efficiently form patulin adducts and are able to abolish the molecule’s toxic properties(78). The exact etiology of acute patulin mycotoxicosis is still debated, but recent studies suggest a role for mucosal tight junction disruption and bacterial translocation(78, 137). Apples, other fruit and fruit juices are the primary reservoirs for patulin, resulting in the strict regulation of mycotoxin levels in these consumer products in most Western countries(96).

Ochratoxin A was originally isolated from a bulk maize meal culture of A. ochraceus (K-804 strain) which was found to be acutely toxic to young mice, rats and ducklings(81). Bioassay-guided fractionation yielded a pure chlorophenol which induced sudden death in ducklings with an LD₅₀ of 500 mg per kg. Histology of the liver revealed fatty infiltration of parenchymal cells throughout the gross structure of the organ. The 3,4-dihydroisocoumarin structure of ochratoxin A is furnished by a type I polyketide synthase(125). This pentaketide scaffold subsequently undergoes benzylic oxidation followed by adenylation, thiolation and condensation with phenylalanine catalyzed by a non-ribosomal peptide synthetase to afford ochratoxin B. Finally, a dedicated halogenase installs the aryl chloride moiety, delivering ochratoxin A. Ochratoxin A acts primarily as a nephrotoxic agent of multiple mechanisms of unclear ranking in importance and has been linked to historic cases of human endemic nephropathy in the Balkan states in the 1950s(94, 138). The possible food stocks harboring contamination are extremely varied, encompassing both plant-based (e.g. cereals, wine, vegetables, coffee, peanuts) and animal-based (e.g. dairy, meat) products which serve as suitable substrates for the ochratoxin-producing fungi(91).

In the spring and summer of 1960, England was subjected to a massive agricultural outbreak affecting domesticated turkeys, known only as Turkey “X” Disease, that swept the country devastating the poultry market. Animals presented with a characteristic liver disease resulting in compression of sinusoids and loss of granular appearance with regions of eosinophilic cells with homogenous cytoplasm, indicative of damage and possible initiation of carcinogenesis(126). The source of the disease was discovered to be a newly initiated groundnut supply from Brazil and the feedstocks were quickly altered sparing half a season of fowl(67). Over the course of the next five years, researchers elucidated the causative agents: the aflatoxins, a series of extremely potent mycotoxins arising from the secondary metabolism of Aspergillus flavus(5). The identity of four highly unsaturated scaffolds were revealed by Wogan and coworkers for two blue (aflatoxin B1 and B2) and two green fluorescing (aflatoxin G1 and G2) compounds. Their tetracyclic skeleton arises through the sequential action of two fatty acid synthases (aflA and aflB) and a type I, non-reducing fungal polyketide synthase (aflC (pksA)) resulting in the formation of norsolinic acid anthrone(28). An anthrone oxidase (hypC) catalyzes the aerobic conversion to the anthraquinone, norsolinic acid which undergoes a series of post-polyketide synthase tailoring steps yielding the mature aflatoxins(36). Interestingly, A. nidulans and A. versicolor produce only the anthraquinone sterigmatocystin, a toxic intermediate in the biosynthesis of the aflatoxins, through an otherwise homologous pathway(20). Aflatoxin-producing Aspergillus spp. are notorious for their ability to invade stores of peanuts, grain (e.g. corn, wheat, linseed) and dairy products(101), posing risks to humans and livestock alike. Aflatoxins produced by A. flavus, A. nomius and A. parasiticus are bioactivated by cytochrome P450 epoxidation in the liver forming genotoxins (e.g. aflatoxin B1 exo 8,9-epoxide). These highly electrophilic epoxides intercalate into DNA, by virtue of their unsaturated, planar scaffold, and rapidly alkylate N⁷-guanine residues(56). This, in turn, may result in errors in DNA-adduct replication or repair, leading to carcinogenesis in the liver.

The large family of trichothecene natural products arise from a collection of filamentous fungi typically from the genera Fusarium and Trichoderma. The trichothecene secondary metabolites were originally isolated from Trichothecium roseum, a typical phytopathogen found in fruits such as melons, apples and grapes(39). In 1968, Prentice and Dickson at the University of Wisconsin investigated the occasional instances of Fusarium spp. associated with a wheat disease known as ‘scab’(95). Upon ingestion of Fusarium-infected feed, cattle present with emesis and, upon gratuitous consumption, death. Similar cases of mycotoxicosis have been known for nearly one hundred years. In an incident from early Soviet Russia, contaminated cereal incorporation into rye bread resulted in human poisoning, with affected individuals displaying weakness, vertigo, headache and nausea(35). Prentice and Dickson were successful in isolating emetics from corn infected by several Fusarium spp. and the structures of T-2 toxin and HT-2 toxin (named for the T-2 strain identification code and the deacetylated hydroxy-derivative of the former, respectively) were disclosed by Strong and colleagues(9, 10). Around the same time of these discoveries, two closely related cyclohexenone-containing natural products from F. nivale and F. roseum were isolated and designated as nivalenol and deoxynivalenol, respectively(83, 114, 135). Collectively these fungal secondary metabolites arise from terpene biosynthesis involving the cationic cyclization of farnesyl diphosphate to form the sesquiterpenoid trichodiene. This committed intermediate undergoes an elaborate series of oxidations, a cyclization and a varying degree of acylation reactions to form an extremely diverse series of toxins from a common biosynthetic origin(33). The toxic properties of these compounds are dependent on the nucleophilic epoxide moiety adorning their intricate scaffold, as chemoenzymatic removal leads to a drastic loss in cytotoxicity(110). The established central mechanism for toxicity of many members of this family involves inhibition of ribosomal protein synthesis(31). Continued efforts examining their ability to activate innate immunity and affect intestinal homeostasis may prove as interesting avenue of future research(42, 58).

The gut mycobiota and metabolic diseases

Whether through mycotoxin, bacterial modulation or nutrient metabolism several studies suggest that select fungal strains have the ability to influence host metabolism. This influence might either be a direct effect on the metabolism of selected nutrients or an indirect modulation of the host microbiota. Microbial communities in the intestine are unavoidably connected and these microbial networks are crucial to various host metabolic pathways (52).(45, 133)(89). Despite this continuous cross-talk, recent studies suggest that gut fungi might be directly involved in disease progression independently from their interaction with gut bacteria. High levels of Candida and Saccharomyces species have been found in children with beta-cell autoimmunity that progressed to clinical type 1 diabetes (T1D) early in life(54). Children with beta-cell autoimmunity also presented with bacterial dysbiosis that was nonetheless not associated with early T1D development. In a HFD-induced obese mouse model, administration of a probiotic beverage containing high levels of yeasts (including Candida spp.) altered the gut bacterial composition and reduced both liver damage and the serum levels of cholesterol and IL-6(62). In leptin-resistant (db/db) mice, which develop fatty livers associated with severe obesity and type 2 diabetes, intensive gavage with S. boulardii altered the gut bacterial microbiota and resulted in decreased body weight, fat mass, hepatic steatosis, and inflammatory state. S. cerevisiae has also been shown to reduce levels of liver damage and alter the bacterial microbiota composition in mouse and rat models of acute liver injury(73, 136). Fungal-derived sugars might contribute to these metabolic changes(37, 89). Administration of fungal-derived mannan to mice with chemically induced acute lipemia reduced the levels of atherogenic LDL, cholesterol, and triglycerides. Mannan also causes labilization of lysosomal membranes and a decrease in the volume of the lipid droplets in hyperlipidemic mice(41). β-glucan of fungal origin has also been reported to reduce SGLT‐1 expression on the intestinal epithelium and to improve glucose control and fatty liver in genetic and high-fat diet obese rodent models(19, 25, 59, 65). Besides its immunological effect, β-glucan might also act as a dietary fiber by slowing gastric emptying; hindering triglyceride and cholesterol absorption and consequently reducing cholesterol levels(105). The ability of fungi to modulate host uptake of key nutrition components important to the development of obesity opens the door for future precise experiments investigating their effect on obesity-related diseases.

Akin to nutrients absorbed throughout the gastrointestinal tract, gut-derived microbial products – including metabolites and toxins – are constantly transported to the liver via the portal circulatory system where they may undergo metabolism prior to entering the systemic circulation (Fig. 3). This continuous exposure to immune-reactive substances has the potential to induce potent immune responses. In the healthy liver, such responses are effectively counteracted by unique regulatory mechanisms such as regulatory soluble mediators and local tolerogenic antigen-presenting cells (APCs). This immune tolerance is crucial to avoid unnecessary responses that can lead to the development of liver pathologies through detrimental inflammatory responses. In most cases, the etiology of these diseases appears to involve a failure of immune tolerance possibly incited by environmental triggers in a genetically predisposed host. These environmental triggers could include xenobiotics, diet and components of the microbiota. Importantly, recent evidence suggests that intestinal fungi could affect the gut-liver axis and might play a critical role in the development of liver pathologies.

Figure 3: Fungi modulate host metabolism directly by secreting enzymes and toxins, and indirectly through their cell wall polysaccharides.

The toxin candidalysin secreted by C. albicans, induces liver damage by promoting the release of IL-1β by liver phagocytes. β-Glucans on the surface of the fungal cell wall decrease intestinal absorption of cholesterol and reduce epithelial expression of the glucose transporter SGLT‐1. β-Glucan might slow gastric emptying thus reducing triglycerides and cholesterol absorption. Mannan also exerts a hypolipidemic effect on the mouse liver. Fungi secrete enzymes that directly digest the luminal nutrients and polyamines that modulate the expression of host metabolic enzymes and nutrient transporters in the intestinal epithelium.

Fungal dysbiosis has been reported in cirrhosis patients with HCV, alcohol-related liver disorders and non-alcoholic fatty liver disease (NAFLD)(8, 68). In particular, cirrhosis patients had an increased burden of Candida and a lower Basidiomycota/Ascomycota ratio was observed in patients with advanced and infected cirrhosis(8). In cirrhosis patients, this increased Candida burden has been previously associated with increased mortality(63) and recent evidence suggests that intestinal fungi contribute to the development of alcohol‐associated liver disease(134). Similarly, an increased burden of Candida, including C. albicans, has been reported in the feces of alcohol-associated liver disease patients(29, 134). ASCA IgG, that can be generated in response to C. albicans, is increased in alcoholic hepatitis and is associated with increased mortality(134)(28). In mice, colonization with C. albicans led to a worsened outcome of ethanol-induced liver disease whereas treatment with the antifungal amphotericin resulted in lower levels of liver injury, inflammation, and hepatic steatosis (29, 134). A recent analysis showed that 30% of alcoholic hepatitis patients had fecal samples positive for C. albicans strains carrying the candidalysin expressing gene ECE1, whereas ECE1 was not detected in control patients. Importantly, these patients had an increased disease severity and were at higher risk of mortality(29). Candidalysin is a cytolytic enzyme that damages the host’s cell membrane by forming pore-like structures that lead to LDH release and calcium influx into the cytoplasm. These trigger a danger response signaling pathway that activates epithelial immunity(2, 84, 120, 121) as well as IL-1β secretion by macrophages (60, 100) (113). In mice, C. albicans exacerbation of alcoholic liver disease was dependent on ECE1 expression, leading to increased levels of IL-1β and inflammatory chemokines in the liver via the release of candidalysin(29). Several pattern recognition receptors (PRR) have been shown to recognize carbohydrates on the surface of the fungal cell wall and to be essential for the initiation of immuno-modulatory responses. In particular, C-type lectin receptors (CLRs) such has Dectin-1, Dectin-2, and macrophage-inducible C-type lectin (Mincle) signal via the tyrosine kinase (Syk)/Card9 pathway to induce NF-κB activation(107). In the gut, fungal-recognition and responses are mediated by a subset of myeloid cells, characterized by the expression of the chemokine receptor CX3CR1(70, 74, 75). While it remains to be established whether these cells and pathways play any role in influencing the host metabolic status in response to intestinal fungi, mice lacking Dectin-1 on hematopoietic cells are protected from ethanol induced liver damage(134). Systemic β-glucan injection has been shown to promote local IL-1β increase in the liver and, in vitro, β-glucan promotes the release of IL-1β by Kupffer cells(134). However, this in vitro release is Dectin-1 independent(134), and Dectin-1 appears to be dispensable for C. albicans-induced hepatic damage(29). This discrepancy could be due to insufficient levels of β-glucan released systemically by gut C. albicans, or by β-glucan signaling through other pro-inflammatory pathways such as TLR2/TLR4(29).

A further mechanism by which fungal colonization might contribute to liver damage is through the induction of adaptive T cell responses. Gut fungal colonization can induce strong Th17 responses both locally and systemically(69, 104). Th17 cells have been implicated in the development of immune-mediated liver diseases(66, 76, 90). Among such diseases, primary biliary cirrhosis, primary sclerosing cholangitis (PSC) and autoimmune hepatitis are the most common indications for liver transplant. Peripheral blood mononuclear cells isolated from PSC patients had an increased Th17 response to stimulation with C. albicans when compared to cells from healthy individuals(61). Biliary candida infections have been reported in PSC and are associated with a worsened disease prognosis(61, 64), providing one plausible locale for origination of a T cell interaction. Whether intestinal Candida can also contribute to the development of immune-mediated liver diseases remains unclear.

The C-type lectin receptor Mincle, an essential component of the innate immune response to systemic C. albicans(127), is highly expressed in hepatic innate inflammatory cells and endothelial cells of autoimmune hepatitis (AIH) patients(44). AIH patients also present with increased levels of Syk phosphorylation, a tyrosine kinase immediately downstream in the CLR signaling pathway. These findings could be replicated in the mouse model of concanavalin A-induced hepatitis, and both chemical and genetic blockade of the Mincle pathway protected against the development of the disease. However, SAP130, the only endogenous, non-pathogen–derived Mincle ligand was also found to be increased suggesting that the role of Mincle in the development of autoimmune-liver diseases might be fungal-independent(44).

The clinical presentation of NAFLD is very similar to the pathology spectrum observed in patients with alcoholic liver disease. The disease presentation ranges from simple steatosis and non-alcoholic steatohepatitis to the development of cirrhosis and hepatocellular carcinoma. NAFLD is characterized by the accumulation of fat in hepatocytes. Furthermore, NAFLD is associated with obesity and with a high risk of type 2 diabetes (T2D), hypertension, dyslipidemia, and hypertension. Supporting a role of the intestinal microbiota in the disease pathogenesis, germ-free mice are protected from diet-induced obesity and insulin resistance(97). Further, microbiota transplant has been shown to transfer NAFLD features from high-fat diet fed mice to GF animals and the gut microbiota appears to modulate lipid metabolism in the liver, independently of obesity(29, 60). In humans, gestational weight gain is associated with changes in the metabolic potential of the infant microbiota, including abnormalities in vitamin synthesis and carbohydrate utilization pathways of the infant microbiota(13, 30, 108). These data suggest that these metabolic imbalances can be affected by the gut microbiota. Although the role of fungi in the development of NAFLD has not been studied, sequencing of the intestinal fungal communities has demonstrated fungal dysbiosis in obese subjects with a less diverse mycobiota characterized by an increased abundance of the phylum Ascomycota and class Saccharomycetes(80). Interestingly, the mycobiota composition in these individuals also correlated with various metabolic indexes including levels of fasting triglycerides and HDL-cholesterol suggesting a potential link between gut fungi and lipid metabolism. As mentioned above, the differences in fungal composition might be driven by the different dietary intake of obese patients and further studies are needed to clarify the driving forces behind the observed changes.

Fungi also secrete various digestive enzymes and small molecules that can influence host metabolism. Secretion of catabolic enzymes is a unique feature that fungi, being eukaryotes devoid of digestive organs, rely on to obtain nutrients from their environmental niches. S. cerevisiae boulardii administration to healthy volunteers resulted in an increased activity of lactase, α-glucosidase, and alkaline phosphatase activity measured in the intestinal brush border(55). These properties have been successfully used to treat enzymatic deficiency in patients with sucrase–isomaltase deficiency, an inherited disaccharidase deficiency that hinders sucrose digestion leading to diarrhea and abdominal cramps(47, 116). The unabsorbed sucrose is converted to hydrogen by gut bacteria leading to an increased concentration in the breath that can be used for diagnosis. S. boulardii can relieve this enzymatic deficiency, reducing gastrointestinal hydrogen production and clinical symptoms in these patients while consuming a sucrose-containing diet via the production of the enzyme sucrase(47, 116). A similar mechanism is also thought to underlie the protective effect of S. boulardii in patients with drug-induced disaccharidase inhibition and trehalose intolerance(24, 98). Loss of epithelial surface caused by mechanical injury can also profoundly affect nutrient absorption. In mice, intestine resection causes an increased circumference of the intestinal wall, mucosal hyperplasia and reduced enzymatic activity. Several studies have shown that S. boulardii can restore gut-associated enzymatic activity (including disaccharidases, sucrose, lactase and maltase activity while rescuing the levels of D-glucose uptake) by increasing expression of the sodium/glucose cotransporter 1 (SGLT-1) in the brush border membrane (21, 23). The effect is likely mediated by the release of polyamines from S. boulardii cells that are able to directly promote synthesis of intestinal glycoproteins, nutrient transporters and digestive enzymes(1, 21–23). Studies on S. boulardii highlight the unexplored role that intestinal fungi and their enzymatic machinery have as direct modulators of intestinal metabolic activity.

Concluding remarks

The gut mycobiota is comprised of a diverse community of fungi inhabiting the human gastrointestinal tract. This collection of intestinal fungi plays a key role in several aspects of human nutrition. Inoculation at birth and through breastfeeding introduces neonates to their foundational GI mycobiota(15, 16, 50, 123). This community shifts as infants begin transitioning to soft and solid foods, and later adapts to the dietary choices of adults(40, 45, 102, 109). Changes within the mycobiota correlate with the much larger bacterial community dynamics, resulting in increases in obligate anaerobic genera such as Bacteroides(52, 133). Additionally, fungi are central elements in fermentation processes. More than simply providing a unique source of nutrition, colonizing fungi alter the mycobiota of these food sources by producing selective antimycotic peptides, providing a unique selection mechanism for gastrointestinal inoculation. In a darker turn, infection of produce, grain and silage by filamentous fungi presents a challenge to livestock and humans alike. Mycotoxins biosynthesized by Fusarium, Aspergillus, Penicillium and other molds, varying in their intricate structures, induce genotoxic, nephrotoxic and hepatotoxic effects(56, 78, 138).

The gastrointestinal mycobiota is broadly recognized by innate and adaptive immunity, serving to prevent opportunistic pathogens from causing systemic infection. Mucosal fungal immunity has been further implicated in the genesis and exacerbation of a broad spectrum of metabolic diseases and liver pathologies (7, 29, 70, 75, 79, 124). Future studies investigating the exact mechanism of how fungi promote these disease states and their role in other metabolic disease etiologies remain under investigation.

Summary Points List:

• The human gastrointestinal fungal community, or mycobiota, is a diverse facet of the larger microbial community of eukaryotes and prokaryotes thriving within the host.

• With the advent of modern sequencing technologies, a sharper image of the early-life acquisition and variability throughout life of the mycobiota is beginning to emerge.

• Nutritional sources of fungi (e.g. fermented foods and plant-association) as well as general dietary choices appear to manipulate the intestinal communities.

• Fungal contaminations within food sources have a global impact on human and livestock health through the production of hepatotoxic, nephrotoxic and carcinogenic mycotoxins.

• The gastrointestinal mycobiota broadly affects immunity and physiology, resulting in aggravation of gastrointestinal cancers, obesity and other metabolic disorders. ·

• Intra-kingdom (fungal-fungal) and trans-kingdom (fungal-bacterial) interactions can shape the gut microbiota to affect host immunity and disease.

• Further research is needed to understand whether other members of the gut mycobiota can directly modulate host immunity, metabolism, and cancer development by secreting enzymes, toxins or metabolites.

Glossary

Intra-kingdom interaction

interaction that occur within the same kingdom of life. Examples relevant to this review include fungal–fungal and bacterial-bacterial interactions.

Trans-kingdom interaction

interaction that occur within the different kingdoms of life. Example relevant to this review include fungal-bacterial interactions.

Xenobiotics

substances that are foreign to the body

Steatohepatitis

a type of fatty liver disease characterized by concurrent inflammation and fat accumulation in the liver.

Dyslipidemia

an imbalance in the concentrations of triglycerides and HDL cholesterol in the plasma

FODMAP

Fermentable Oligo-, Di- and Mono-saccharides And Polyols, highly fermentable but poorly absorbed short-chain carbohydrates and polyols that might exacerbate intestinal symptoms by increasing small intestinal water retention, colonic gas production, and intestinal motility.

ITS

The nuclear ribosomal Internal Transcribed Spacer region of fungal DNA, widely used to characterize the diversity and composition of fungal communities.