Mycobiome and Cancer: What Is the Evidence?

3.1. The Role of Fungal Dysbiosis in CRC

Besides bacteria inhabiting the gastrointestinal (GI) tract, fungal phyla, such as Basidiomycota, Glomeromycota and Ascomycota, reside in high numbers in the digestive tract [47]. The most commonly found fungal genera inhabiting the physiologic GI are Candida, Saccharomyces, and Cladosporium [47]. Trojanowska et al. demonstrated that the intestinal tract is also inhabited by members of the oral mycobiome, as they have identified the same Candida albicans strain in the oral cavity and gut of subjects with inflammatory bowel disease (IBD) [52]. Unfortunately, there is a paucity of data regarding gut fungal commensals in cancer. Dysbiosis is well-known among patients suffering from IBD, who present higher odds of CRC occurrence [17]. It is noteworthy that decreased richness and diversity have also been reported in the bacterial community as well as the fungal microbiome [17,47]. For example, Cystofilobasidiaceae, Dioszegia genus and Candida glabrata were detected in abundance in the gut of subjects suffering from Crohn’s disease, when compared to healthy individuals [53]. Luan et al. have focused upon comparing the mycobiota composition in adenomas and their normal adjacent colon tissues. They have documented an increased number of Phoma and Candida genera as well as Candida tropicalis in adenomas [40]. These fungi may act as pathobionts, being implicated in tumor onset and progression.

Patients with CRC have been documented to present an increased ratio of Basidiomycota/Ascomycota [41,48]. Patients with colitis-associated CRC have also shown a similar ratio [47]. Coker et al. have detected 14 fungal biomarkers with a differential abundance in 184 CRC patients in comparison to 204 healthy participants [43]. Moreover, an abundance of Malassezia has been found among CRC patients by fecal shotgun metagenomic sequencing in conjunction with Moniliophthtora, Rhodotorula, Acremonium, Thielaviopsis and Pisolithus, whilst an increased number of Basidiomycota have been suggested to be related to more advanced stages of the disease [42,43,54].

Notably, a higher ratio of Basidiomycota/Ascomycota, an enhancement in C. albicans and C. tropicalis and a reduction in Saccharomyces cerevisiae were documented in individuals with IBD. It is noteworthy that C. albicans may produce a cytolytic toxic peptide called candidalysin, which is known to promote disruption of the epithelial barrier function, thus mediating dysbiosis. In addition, C. albicans and C. tropicalis may produce acetaldehyde to carcinogenic levels. Acetaldehyde is suggested to increase intracellular reactive oxygen species (ROS) and Ca⁺⁺ concentrations, thereby causing mitochondrial dysfunction, leading to cytoxicity as well as the disruption of epithelial tight junctions [47]. Figure 2 depicts various mechanisms by which fungal dysbiosis may participate in the etiopathogenesis of CRC. Overall, mycobiota dysbiosis is suggested to be a triggering factor of CRC among subjects with IBD through chronic inflammation and secretion of toxic metabolites, which may cause DNA damage.

3.2. The Interplay of Gut Microbiome and Mycobiome in Colon Physiology/Pathology and CRC Pathogenesis

Recent data have shown that the interplay between intestinal fungal and bacterial communities may affect the intestinal microbiome homeostatic balance maintaining overall intestinal health and protecting from gastrointestinal disorders. Prolonged antifungal treatment resulted in an exacerbation of colitis and alteration of the gut bacteriome in a mouse model treated with dextran sulfate sodium (DSS), which provokes colitis [55]. In a murine model, intake of the pathobiont fungus Mucor circinelloides resulted in a reduction of the beneficial Akkermansia and an augmentation of Bacteroides genus [56]. Candida albicans restored bacterial variability and influenced the bacterial colonization of the gut after broad-spectrum antibiotic treatment, such as the increase in Bacteroides species and the pathogen Enterococcus faecalis, and the reduction in Lactobacillus spp. [57,58].

Intestinal fungi and bacteria interact through a variety of ways, including the secretion of metabolites and toxins, the development of biofilms, and physical attachment, thus influencing host immune responses. Based on in vitro and in vivo research data on bacterial and fungal interplay, it was shown that synergistic associations generally enhance pathogenicity whilst antagonistic relations limit bacterial or fungal virulence [59].

Some examples of synergistic actions between bacteria and fungi that may enhance colitis are the following: (1) the requirement of Enterobacteriaceae that aggravate DSS-induced colitis mediated by Candida albicans [60]; (2) the enhancement of the strict anaerobe Clostridium difficile by Candida albicans because of the oxygen decrease in the proximity of the yeast [61]; (3) the survival at decreased pH of Helicobacter pylori in the vacuoles of Candida albicans [62].

Some examples of antagonistic actions between bacteria and fungi with beneficial or neutral actions in gut health include: (1) the antifungal activity of Serratia marcescens, Salmonella typhimurium and Acinetobacter baumannii on eliminating the hyphal and yeast forms of Candida albicans or limiting the formation of biofilm and infection [59]; (2) the restriction of hyphal growth of Candida albicans by Clostridium difficile via the production of p-cresol [63].

Biofilms represent aggregations of microorganisms that are embedded in an extracellular polymeric matrix sticking to biological or non-living surfaces. The development of biofilm is enhanced by the collaboration of C. albicans and bacterial microorganisms including, among others, E. coli, E. faecalis, Streptococcus spp., Staphylococcus aureus, and Staphylococcus epidermidis, while other bacteria including K. pneumoniae and P. aerugivalis limit the synthesis of biofilm [47]. The intestinal mucosal biofilm may be an inducing factor in colorectal carcinogenesis [64]. Biofilms from subjects with CRC or healthy individuals submitted to colonoscopy triggered tumorigenesis in mouse models of CRC through promotion of chronic inflammation, evasion from the host immune system and disruption of epithelial integrity [64]. A close physical contact between Candida tropicalis and E. coli facilitated by Serratia marcescens was observed by electron microscopy in subjects with Crohn’s disease, an IBD which predisposes patients to small bowel and colon cancer [65]. This fungal–bacterial interaction created a robust biofilm which triggered sustained intestinal inflammation. Based upon these findings, it can be inferred that similar biofilms create the perfect persistent inflammatory milieu for the promotion of colon carcinogenesis.

Another important aspect in colorectal carcinogenesis is that the majority of pathogenic bacteria associated with CRC resides also in the oral cavity. Interestingly, a plethora of studies have shown that the oral bacterial pathogens Fusobacterium nucleatum and Porphyromonas gingivalis could contribute to the initiation and progression of CRC and pancreatic cancer via chronic inflammation, inhibition of host immunity and the secretion of tumorigenic substances [66,67,68]. A microbiota and metabolomics profiling on feces from CRC patients and matched controls has shown that metabolites were linked to CRC via their association with Fusobacterium and Porphyromonas [69]. Interestingly, Fusobacterium nucleatum has been shown to form coaggregations with both the hyphal and yeast forms of Candida albicans through the involvement of genetic and structural cellular components [70]. This coaggregation may contribute to facilitate their synergistic colonization in the oral cavity and the gastrointestinal tract, as well as to enhance pathogenesis and chronic inflammation. However, more mechanistic and clinical studies are needed to decipher the implications of this coaggregation in colorectal carcinogenesis.

Finally, a growing body of evidence has suggested that there is a significant interplay between gut microbiota and the host at the intestinal stem cell niche level, where microbiota may affect directly or indirectly the proliferation, differentiation and reprogramming of the intestinal stem cells and their transformation to cancer stem cells, resulting in CRC initiation and progression through a plethora of mechanisms reviewed elsewhere [71]. The totality of studies has focused on the role of gut bacteria on the abnormal reprogramming of intestinal and cancer stem cells without examining the role of fungi in this interplay. More studies with the use of integrative system-based approaches (i.e., metagenomics) are required to decipher the interplay of gut bacterial and fungal communities with intestinal and cancer stem cells in the initiation and promotion of colorectal carcinogenesis.

5.1. Limitations and Challenges in Studies

Overall, fungal dysbiosis with decreased fungal richness and diversity is a common theme in cancer patients; however, a specific mycobiotic signature in HNSCC or CRC has not been revealed based on studies depicted in Table 1. Generally, studies on human mycobiome and cancer are scarce and heterogeneous, while they have included only a small number of individuals with cancer (from 14–184 patients), as they are expensive and need modern and sophisticated equipment. Therefore, due the small sample size, the statistical power to detect any differences in mycobiota between cases and controls is low. An important shortcoming of all studies is their retrospective or cross-sectional design, which is not appropriate to draw any conclusion regarding mycobiota as causal factors in cancer. Despite difficulties in performing such studies, more prospective, multicentric, larger and longitudinal studies over a long time period (≥5–10 years), are required to shed light on the role of the composition of fungal or other microbiota in cancer etiopathogenesis. The discrepancy of results between studies may also be attributed to the different specimens used (stool or saliva specimens versus tissue biopsies), different laboratory techniques summarized in 5.2, different populations and ethnic groups, different pathologic subtypes and stages of cancer examined, high inter-and intra-individual variability of the mycobiome, lack of evaluation criteria, etc. It is important to mention that more significant mechanistic evidence about the interplay of microbiota and host can be achieved by studying tissue biopsy specimens, due to the adherence of microbial communities to the epithelia, particularly the intestinal epithelium.

In addition, identification to the fungi species level is difficult, whilst it may be inconclusive and not illuminating. For example, C. albicans may be implicated in the etiopathogenesis of HNSCC, but it is also a component of the oral microflora among healthy individuals. Therefore, identification of different strains of C. albicans may be mandatory in order to shed light upon slightly differential, but potentially pathogenic, strains within the same species [24,30,31,32].

Finally, based on the paucity of available data, formal meta-analyses examining the association of a mycobiome with cancer occurrence, while significant, is very challenging to undertake.

5.2. Limitations and Challenges in Sample Collection and Laboratory Methodologies

Regarding fungal identification, as fungi are ubiquitously spread, there are concerns regarding the sampling and processing methods in order to avoid contamination [76]. Fungal DNA extraction is performed by mechanical cell lysis; however, there are some difficulties including their recalcitrant chitinous cell walls and the interference of secondary metabolites with DNA extraction. Therefore, the DNA extraction methods should be suitable for fungi. For example, species particularly rich in polysaccharides, such as encapsulated yeasts, may require further special DNA extraction techniques [77,78].

There is a plethora of laboratory methods to detect live fungi or a genome of fungal communities including, among others culture, fluorescent microscopy, amplicon sequencing and whole-genome shotgun (WGS) metagenomics, or high-throughput sequencing (HTS) metabarcoding studies [1,59,79]. Overall, sequencing methodologies reveal fungal DNA in a specimen independent of fungal viability.

In analogy to 16S sequencing for bacteria, amplicon sequencing employs distinct primers for fungi to amplify the internal transcribed spacer (ITS) region or the 18S region of the nuclear ribosomal RNA (rRNA) gene locus, permitting the discrimination of fungal species [59]. Sequencing the ITS region of the nuclear rRNA operon is the gold standard in Sanger sequencing-based species detection and HTS-based methods [80,81]. ITS constitutes the main fungal taxonomic marker gene or molecular identifier or “DNA barcode” [1]. Because the whole ITS region is too long for sequencing, spanning 500–700 bps, the majority of HTS-sequencing studies has concentrated on either the ITS1 or the ITS2 subregion within 250–400 bps [1,82]. The choice of what sub-region to use is still a discussed point. The use of ITS2 subregion may offer a more ubiquitous primer location and reduced length variability, resulting in fewer taxonomic errors in comparison to ITS1 [79]. Nevertheless, conflicting results have been emerged in the comparisons between the sub-regions ITS1 or ITS2 for the analyses of fungal profiles [83]. The advantages of using the full ITS region instead of ITS1 or ITS2 with third-generation HTS platforms such as Oxford Nanopore or Pacific Biosciences include the higher taxonomic discrimination and the decreased amplification of non-living microorganisms; however, a shortcoming is the low performance with poor-quality samples due to the DNA deterioration, where full ITS sequencing may be impossible [1]. Overall, the disadvantages of the amplicon sequencing comprise the paucity of a complete annotated reference database, taking into account the complex taxonomic composition of the fungal community as well as the amplification errors inherent to PCR [59].

Despite the fact that amplicon sequencing is the most widely used methodology to spot genes, shotgun HTS allows the detection of various genes together, based on the sequencing of all isolated genetic material in a given specimen without the employment of targeted PCR quantification of the ITS or other gene regions [79]. Notably, in comparison to amplicon sequencing, omics methods may confer taxonomic composition more precisely, by circumventing biases related to PCR and primer selection [84,85]. Besides, omics presents the ability to improve standardization as well as comparison of different phyla for future studies [86]. Moreover, metagenomics offers the opportunity to identify fungal taxa in conjunction with prokaryotes, and this dual ability may explain their usefulness in exploring different microbiomes [87,88].

The combination of a HTS approach with barcoding has been termed “metabarcoding”, and consists of several laboratory processes needing bioinformatics and computational statistical analyses. Briefly, a typical fungal metabarcoding workflow presents the following major items: specimen preparation, extraction of DNA, DNA amplification, HTS-sequencing, processing of sequencing, quality control (demultiplexing, elimination of chimeric molecules and flanking genes, etc.) and sequence data analysis (clustering, operational taxonomic unit/OTU positioning and taxonomic annotation) [1]. Therefore, the selection of an adequate modus operandi for all processes in the metabarcoding workflow is of paramount importance. Inadequate methodology in metagenomic studies may result in erroneous biological assumptions. HTS significant biases may happen from additive systematic and random errors, which include, among others, biases in genetic material extraction, markers, primers, PCR, library preparation, sequencing, bioinformatics analysis, index-switching, low clustering, unequal sequencing depth, etc. [1,83]. Moreover, DNA metabarcoding for microbiota analyses presents significant shortcomings, including the variability of copy numbers of the targeted barcodes in microorganism genomes, the poor taxonomic discrimination at the species level for some microbiota, and errors in the taxonomic assignment of sequences based on the selected variable region [83].

WGS metagenomic studies in a plethora of human specimens have shown the reduced quantity of fungi (≤0.1% of total microbiota) in comparison to bacterial DNA. Fungal DNA has a large quantity of non-coding regions compared to bacterial DNA, which further complicates fungal metagenomics [89]. Furthermore, genomic fungal databases have been lacking until today, especially when compared to genomic bacterial databases [89,90]. Due to the low sequence quantity rates, it is hard to implement OTU studies based on metagenomic data.

Fungal metatranscriptomics is a more beneficial, emerging and analytical approach because of the lack of introns in the expressed genes and the better annotation. It is noteworthy to mention that RNA-based studies may help in the quantification of fungal taxa as well as their functionality in the least biased way [1]. Metaproteomics and metabolomics help to understand the properties and the interplay between intestinal microbiota (e.g., bacteria and fungi), by providing information of the role of bacterial and fungal proteins and metabolites (i.e., small and low molecular weight molecules, being the downstream products of gene and protein processes), respectively. In the framework of cancer microbiota, a variety of metabolites may originate from bacterial, fungal and host metabolisms, or cometabolic networks between the host and microbial communities. In order to circumvent the compositional nature of metabarcoding data in the metagenomics strategy, a multi system-based approach (integrative multi-omics) could broaden our knowledge of the function of the interplay of microbial communities and cancer stem cells in carcinogenesis [71]. Llyod-Price et al. used an innovative multi-omics network analysis of ten omics data (metagenomic, metabolomic, metatrascriptomic data, etc.) to highlight the pathogenetic mechanisms in IBD [91]. A similar approach could be useful in deciphering the pathogenetic mechanisms in cancer. Furthermore, using integrative analytical approaches of metagenomic and small noncoding RNA-sequencing data from stool samples, a human and microbial small RNA signature was detected, and could be used for diagnostic purposes in CRC [92].

Finally, prospective, larger and long-term research studies are required to examine if alterations in the mycobiome are causal or a consequence of cancer.

6.1. Preventive and Therapeutic Implications

Fungal dysbiosis may be reversed through nutrition, administration of probiotics and prebiotics, and fecal microbiota transplantation (FMT).

Dietary factors are pivotal with regards to the composition of human mycobiome and bacteriome, and the risk of specific types of cancer in the future. As vegetarians have been found to possess different compositions of their mycobiome, in comparison to subjects who consume a Western-based nutrition, it is highly likely that the nutritional status may represent a significant way of preventing these specific types of cancer [93].

The human mycobiome, just like the human microbiome, is greatly influenced by dietary factors [94,95,96]. Hoffman et al. gathered 98 samples from healthy participants to assess the association between the type of nutrition and the composition of intestinal microbiota [49]. Interestingly, the gut mycobiome was linked only to the recent type of nutrition. Candida was positively correlated with the consumption of carbohydrates, particularly polysaccharides and araboxylan, and was negatively associated with total saturated fatty acids, whilst Aspergillus was negatively related to the intake of short-chain fatty acids [49]. Intake of animal food was correlated with Saccharomyces [10].

In analogy to the favorable role of bacterial probiotics and prebiotics, fungal probiotics have shown beneficial anti-neoplastic effects, and could be helpful in cancer prevention and therapeutics [67]. The probiotic cocktail of Saccharomyces, Lactobacillus rhamnosus, Lactobacillus acidophilus and Bifidobacterium breve has been shown to exhibit anti-biofilm and anti-tumor actions in the colon [97]. Galinari et al. have shown the antioxidant and pro-apoptotic properties of the yeast Kluyveromyces marxianus, that presents a phylogenetic association with Saccharomyces cerevisiae. S. cerevisiae is generally employed in the classic nutrition industry [98]. In another study, the cytoprotective actions of β-glucan, originating from S. cerevisiae, played an important role in the prevention of genotoxicity [99]. Notably, the treatment of animal models with selenium-enriched S. cerevisiae has been shown to exhibit a remarkable efficiency in comparison to the separate administration of selenium or S. cerevesiae. [100].

In line with the properties of S. cerevisiae, the commonly employed probiotic Saccharomyces boulardii has been reported to arrest the expansion of bacteria by synthesizing elevated concentrations of acetic acid [101]. Chen et al. have demonstrated that S. boulardii regulated inflammatory responses and suppressed gut cancer expansion in a mice model, by inhibiting the EGFR-Mek-Erk signaling network, while it exhibited pro-apoptotic actions in tumor cells by suppressing Akt, a central actor of the cell cycle [102,103]. Besides, S. boulardii has been shown to sufficiently decrease levels of inflammation and rebuild gut microbiota, resulting in an amelioration of cancer-associated colitis [102,103]. Apparently, beneficial fungi exist and seem to be capable of creating a favorable environment for anti-cancer effects. To date, Schizophyllum commune, Saccharomyces cerevesiae and Saccharomyces boullardii have been mainly reported to exert beneficial and anti-oxidant properties. Moreover, treatment targeting immune components, such as MBL or the complement C3 cascade, may prove to be effective in the therapeutic armamentarium against cancer [102,103].

Finally, FMT, which is the process of transferring a fecal microbiome from a healthy donor to another subject, restoring gut microbial homeostasis, may be a promising tool in CRC prevention and treatment. The composition of stool mycobiota of the donor could represent an important parameter in the efficacy of FMT as a preventive and therapeutic approach in CRC, as it was shown in other diseases such as Clostridium difficile-associated diarrhea and ulcerative colitis [104,105]. Indeed, the efficacy of FMT in these disease entities was associated with an increased load of Saccharomyces and Aspergillus, and a lower load of Candida in donor feces [105]. FMT with donor stool enriched with beneficial fungi may be also used as a potential adjunct treatment in augmenting the efficacy of immunotherapy with anti-PD-1 agents in epithelial cancer, including CRC [67,106].

More large-scale and long-term randomized placebo-controlled trials are needed to examine the safety, efficacy and sustainability of results in all these preventive and therapeutic interventions. Based on detected fungal signatures, novel targeted treatment modalities in precision medicine may emerge with the alteration or restoration of a healthy fungal community in patients with cancer.

6.2. Fungal Dysbiosis and Biomarkers as a Diagnostic and Prognostic Tool in Cancer

Alterations in the intestinal mycobiome could be used as an adjunct screening, diagnostic and prognostic tool in CRC, distinguishing early from advanced stages. Fungal biomarkers and dysbiosis could differentiate cases with CRC from healthy subjects, as was shown in a study where an altered load of 14 fecal fungal biomarkers, including an augmentation in Malasseziomycetes and a reduction in Saccharomycetes and Pneumocystidomycetes, presented a good diagnostic discriminative ability in CRC diagnosis [43]. The ratio of Basidiomycota to Ascomycota was also increased [43]. A proteomic study examining the contribution of fecal microbial secretome in colorectal carcinogenesis has revealed the presence of unique proteins from Schizosaccharomyces pombe in subjects with CRC, where 4 fungal proteins characterized the advanced stage [107]. Interestingly, a number of fungi detected in the oral cavity of individuals suffering from CRC, such as Rhodotorula, were indicative of CRC progression based on their enzymatic activity [108].

However, it is too early to draw conclusions regarding the real clinical significance and relevance of fungal dysbiosis and fungal biomarkers in cancer. It is of paramount significance to investigate if fungal dysbiosis is a causal factor in cancer promotion and progression, or a mere consequence of cancer development. Based on detected fungal signatures, novel targeted treatment modalities in personalized medicine may emerge with the modification or restoration of a healthy fungal community in patients with cancer.