Typing of feces-derived Candida albicans strains using a novel seven-locus microsatellite panel reveals associations with yeast phenotype in individuals with inflammatory bowel disease

Abstract

Inflammatory diseases of the human gastrointestinal tract are affected by the microbes that reside in the mucosal surfaces. Patients with inflammatory bowel diseases (IBD) have altered bacterial and fungal intestinal compositions, including higher levels of fecal Candida yeasts. Ongoing research indicates that genetic and phenotypic diversity of Candida albicans may be linked with disease severity. Here, we set out to investigate feces-derived C. albicans strains from individuals with IBD and healthy volunteers through microsatellite-based genotyping and phenotypic assays. A seven-locus microsatellite panel was applied, of which six loci were newly developed. It appears that there is no specific lineage of C. albicans that is associated with IBD, but rather that the three study populations (Crohn's disease, ulcerative colitis, healthy volunteers) do have distinguishable distributions of genotypes. In addition, phenotypic characterization by means of enzyme release assays revealed trends between genotypes, virulence-related enzyme activity, and clinical biomarkers. We thus show that microsatellite typing can describe genetic diversity of feces-derived C. albicans strains, and that phenotypic diversity of these strains may indeed correlate with fungal genotype or disease. This study opens further possibilities to investigate fecal fungi in relation to severity of inflammation in IBD or in other (intestinal) diseases.

inflammatory bowel disease, yeast, mycobiota, genotype

Introduction

The intestinal fungal community resides in close contact with mucosal surfaces and is thereby able to influence the course and severity of inflammatory bowel diseases (IBD). In the two main subtypes Crohn's disease (CD) (Cheifetz 2013) and ulcerative colitis (UC) (Ungaro et al. 2017), different sections of the gastrointestinal tract can be affected by inflammation. The combined incidence is ∼0.3% albeit with a higher incidence in Western countries (Ng et al. 2017). Symptoms generally include diarrhea, bloody stools, abdominal pain, weight loss, and fatigue (Cheifetz 2013, Ungaro et al. 2017). The pathophysiology of IBD is not fully understood, but several major factors contribute to this disease, including genetics, environmental factors, and the intestinal microbial composition. The latter comprises roughly 40 trillion microbes (Sender et al. 2016), of which the vast majority is bacterial, and ∼0.1–1% is estimated to be of fungal origin (Qin et al. 2010). While ample data are available on the role of the bacterial community in IBD, data regarding the gut mycobiome are still limited.

The intestinal fungi composition in IBD patients is generally altered when compared to healthy volunteers (HV), and often characterized by elevated abundances of Candida species and lower diversity metrics (Liguori et al. 2016, Sokol et al. 2017). Several murine studies have indicated that colonization of mice with yeasts including Candida albicans (Iliev et al. 2012), Debaryomyces hansenii (Jain et al. 2021), or Malassezia restricta (Limon et al. 2019) worsens intestinal inflammation. Additionally, genetic mutations in several components of fungal recognition pathways (i.e. CLEC7A encoding Dectin-1, CARD9) were observed in patients with IBD (Zhernakova et al. 2008, Iliev et al. 2012). It is hypothesized that this genetic contribution is also reflected by elevated titers of anti-Saccharomyces cerevisiae antibodies (ASCAs) in IBD patients, and especially those with CD (Barnes et al. 1990, Annese et al. 2001, Walker et al. 2004, Schaffer et al. 2007). These ASCAs recognize yeast mannans and, despite their name, it is assumed that C. albicans is the more likely immunological trigger for the presence of ASCAs (Poulain et al. 2009). Recent mycobiome studies describe genetic and phenotypic variation of C. albicans yeasts that relate to disease severity and mucosal immune responses (Doron et al. 2021, Li et al. 2022). In a study by Li et al. (2022), genetic diversity of C. albicans strains was determined through whole genome sequencing and subsequently derived single nucleotide polymorphism (SNP) density analyses (Li et al. 2022). Historically, multiple other methods have been used to map genetic variation, including amplified fragment length polymorphisms fingerprinting (AFLP), random amplified polymorphic DNA (RAPD) (Mehta et al. 1999), multi-locus sequence typing analysis (Bougnoux et al. 2006), and polymorphic microsatellite loci assessment (Marakalala et al. 2013, Cavalieri et al. 2017, Di Paola et al. 2020, van Thiel et al. 2022).

In the current study, we assessed microsatellite loci to investigate genetic and phenotypic variation of fecal C. albicans in the context of CD, UC, and HV as the method provides fast, reliable, and reproducible results (Alanio et al. 2017). We developed a novel seven-locus microsatellite panel for C. albicans for this purpose. Genetic variation between C. albicans isolates from one individual was limited and contrasted with more pronounced intraindividual differences. A subset of individuals with UC appeared to have a distinct cluster of C. albicans strains, even though the study population was very small. In addition, phenotypic evaluations revealed that C. albicans lipase activity correlated with detection of 1,3-β-glucans in serum. Taken together, the novel microsatellite panel is useful for assessment of multilocus genotypes and subsequent correlations with phenotypic or clinical parameters in the context of IBD.

Materials and methods

Patients and healthy volunteers

For this study, HV were defined as “no gastrointestinal abnormalities or disease”. Medication use was allowed, except for antifungal medications up to three months prior to inclusion. Clinical data were compiled in Castor EDC until analysis of the patient characteristics. HV only provided their sex and year of birth. Medical ethics approval was obtained at Amsterdam UMC, location Meibergdreef (number 2017_239). All participants gave written consent for the use of their fecal and blood samples, and patients additionally for the extraction of information from their electronic patient files.

Sampling and analysis of feces and serum

Patients collected freshly produced fecal samples at home maximally 24 h before processing. In the meantime, samples were refrigerated. Fecal calprotectin levels were determined on the frozen fecal samples at the routine clinical chemistry laboratory of Amsterdam UMC, location AMC. Blood was drawn at the Amsterdam UMC central phlebotomy laboratory. For determination of ASCAs and 1,3-β-glucans in serum, colorimetric kits were used according to the manufacturer's protocol, being the Orgentec ASCA IgA/IgG kit (Siemens Healthineers, Den Haag, The Netherlands) and FungiTell diagnostic kit (Nodia, Amstelveen, The Netherlands), respectively. Resulting absorbance of both assays were determined on a Synergy HT plate reader (BioTEK, Beun de Ronde, Abcoude, The Netherlands).

Culturing and identification of fecal fungi

Small samples of freshly provided stool specimens were diluted using sterile phosphate buffered saline (PBS) and thoroughly vortexed in a 1.5 ml Eppendorf tube. Approximately 100 µl of this suspension was spread onto at least three solid culture media, always including Sabouraud dextrose agar with 0.05 g/l chloramphenicol (SAB; Sigma Aldrich, St. Louis, MI, USA), yeast peptone dextrose agar (YPD; Sigma Aldrich) supplemented with 0.05 g/l chloramphenicol, and malt extract agar (MEA; Oxoid, Basingstoke, United Kingdom) supplemented with 1.68 mg/l penicillin G sodium salt and 1.332 g/l streptomycin sulfate (both Sigma Aldrich). In the early phase of our study, potato dextrose agar (PDA), modified Leeming–Notman agar (MLNA), or modified Dixon’s agar (mDA) supplemented with 0.4 g/l cycloheximide and 0.05 g/l chloramphenicol were included as well. PDA, MLNA, and mDA were later omitted as all observed fecal fungal growth was successfully captured by YPD, SAB, and MEA. Plates were incubated for up to 60 h at 37°C in an aerobic incubator. When microbial growth was observed, plates were individually sealed and stored at 4°C until further processing.

For identification of yeasts, ∼10 morphologically similar yeast colonies were sub-cultured per individual on glucose yeast peptone agar (GYPA) and subsequently identified in duplicate using MALDI–TOF MS (Bruker Daltronics, Bremen, Germany) using the manufacturer's protocol for extended direct transfer. Confirmed yeast strains were routinely frozen in 15% glycerol solutions. Filamentous fungi were omitted from further analysis.

Yeast DNA extraction

GYPA plates were inoculated with the yeasts of interest and allowed to grow for 16–48 h. Biomass was collected using a sterile culture loop, and samples were frozen until extraction of DNA, which was performed using the Wizard Genomic DNA Purification kit (Promega, Leiden, The Netherlands). In brief, yeast cells were mechanically disrupted using glass beads in Nucleic Lysis Solution. After a 30-min incubation period at 65°C, suspensions were subjected to RNase A treatment (Promega) for 15 min at 37°C. Protein precipitation solution (Promega) was added to remove proteins from the solutions, and DNA was lastly precipitated in isopropyl alcohol. After washing of the pellet in 70% ethanol, DNA was dissolved in DNA rehydration buffer (Promega). Presence and integrity of DNA were confirmed using 1.5% agarose Tris-acetate-EDTA (TAE) gel electrophoresis.

Microsatellite typing

Short tandem repeat (STR) lengths were determined for seven loci of obtained C. albicans yeasts. One of these loci, CAI, was previously described (Sampaio et al. 2003, Chavez-Galarza et al. 2010). Primers for the remaining six loci were designed using Tandem Repeat Finder (Benson 1999) based on the genome of C. albicans SC5314 (=CBS 8758). After selecting suitable repeats, primers on flanking regions were designed using Primer3 v0.4.0 (Table 1) (Koressaar et al. 2018). Forward primers were labeled with 6-carboxyfluorescein on the 5’-terminus. A BioTaq polymerase kit was used for all PCR reactions, which consisted of 1 × NH₄ reaction buffer, 1.5 mM MgCl₂, 0.04 µM dNTPs, 0.5 U BioTaq (all Bioline, Meridian Bioscience, Cincinnati, OH, USA) and 1 pmol of each primer (Integrated DNA Technologies, San Diego, CA, USA). Amplification was performed according to the following program: 5 min at 94°C, followed by 35 cycles of 30 s at 94°C, 30 s at 60°C, and 60 s at 72°C. Final elongations were performed at 72°C for 5 min. The resulting amplicons were purified using Sephadex G-50 and diluted 200-fold in water. Hereof, 1 µl was mixed with 8.9 µl water and 0.1 µl Orange 600 DNA size standard (Nimagen, Nijmegen, The Netherlands) followed by incubation at 100°C for 1 min.

Table 1.

Primer sequences microsatellite loci.

Locus	Chr.	Motif	Forward primer sequences	Reverse primer sequences
A	1	(TTA)n	AACATTGCTCCTTCAAATAATTCAG	AACTTCATCCAACCGTGCAT
B	2	(AAC)n	TTCACCCGGTTCAAATTCA	TTGTTGATAATGCAGTTGTTGCT
D	2	(AAT)n	ACATCCGGGTGTTGAAGAAG	TGGTGGATATGAATAAGCATTGA
F	4	(TAT)n	TGGTCGAAGTGATAATGAAGAAGA	CATTCCTTATGCATCACCAAGAT
J	6	(TGA)n	GACTGGAAACGAAATTCATGG	TTTCCCTTCAATCTCATAATCGT
K	7	(AAC)n	GCCGTAAGATTATGGGAAGGT	GCTGAACTAACAACAGTGACATTTCT
CAI	4	(CAA)n	ATGCCATTGAGTGGAATTGG	AGTGGCTTGTGTTGGGTTTT

Forward primer sequences are labeled with 6-carboxyfluorescein on the 5’-terminus.

Fragment lengths were determined by capillary electrophoresis on an ABI3730xL Genetic Analyzer platform (Applied Biosystems, Palo Alto, CA, USA). BioNumerics v7.6 (Applied Maths, Sint-Martens-Latem, Belgium) was used for peak processing and generation of minimum spanning trees. Inability to read STR lengths for more than two loci led to exclusion of the sample from the analysis.

Bioinformatic analysis of STR lengths

The microsatellite dataset was converted into a Genind object and analyzed using the packages adegenet (Jombart 2008) v2.1.7 and poppr (Kamvar et al. 2015) v.2.9.3 in R v.4.2.1. A matrix based on Bruvo's genetic distances (Bruvo et al. 2004) was constructed and used to generate a minimum spanning tree based using the function plot_poppr_msn in poppr. A discriminant analysis of principal components was run using the adegenet package 2.1.7 (updated 2022) in R. DAPC analyses were conducted with de novo grouping without considering the three patients groups. We used the find.clusters() function to determine the number of groups (K) de novo. The optimal number of PCs to use in the DAPC was determined using the optim.a.score() and xvalDapc() commands and 1000 replicates.

Determination of enzymatic activity

Virulence-related enzyme activity of C. albicans strains was determined using multiple solid substrate media as previously described (van Thiel et al. 2022). Overnight C. albicans cultures (30°C, 200 rpm) in Yeast Extract Peptone Dextrose Broth (YPD; 1% yeast extract, 2% peptone, 2% dextrose) were used. Cultures were washed twice in Dulbecco's phosphate buffered saline (dPBS) before counting using a Nexcelom Cellometer K2 cell counter (Nexcelom Bioscience, Manchester, United Kingdom) and adjusting the inoculum to 1 × 10⁸ cells/ml in dPBS. A drop of 3 µl of this suspension was spotted in triplicate onto one of the specified culture media. For lipase activity, tributyrin agar was used (Buzzini and Martini 2002). Egg yolk-containing Sabouraud Dextrose agar was used to determine phospholipase activity (Price et al. 1982). Tween-80 containing agar served as medium for release of esterase (Slifkin 2000). For proteinase activity, bovine serum albumin (BSA) was added to the agar medium (Crandall and Edwards 1987). Plates were placed at 37°C for varying incubation times: phospholipase activity was determined after 3 days, proteinase and esterase after 5 days, and lipase activity after 6 days of incubation. Activity of the released enzymes was calculated as 1 − (diameter colony/diameter precipitation zone). Experiments were performed in three independent replicates and the C. albicans strain SC5314 served as reference.

Statistical analysis

Baseline characteristics of included individuals were summarized in R v4.1.1 using package tableone v0.13.0. Characteristics were, in line with STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for cross-sectional studies, not tested for significance (von Elm et al. 2007). In vitro data regarding C. albicans enzyme release is displayed as mean and standard deviation. Statistical differences between groups were tested using Mann–Whitney U test or Kruskal–Wallis test and Dunn's multiple comparisons test. Differences in enzyme activity versus the activity of the reference strain were determined using a one-sample t-test. Correlations were based on Spearman r. P values below .05 were considered statistically significant. Visualization of PCA plots and analysis of in vitro data were performed in GraphPad Prism v9.5.1.

Results

Culturing of fecal samples predominantly results in Candida spp.

A total of 46 patients (n = 33 CD, n = 12 UC, n = 1 IBD-unclassified) were recruited at the Amsterdam UMC and 17 HV were enrolled (Table 2, Supplementary Table 1) between January 2018 and July 2021. The majority of IBD patients (65.2%) was in clinical remission at the time of inclusion following physician assessment of the electronic patient files. Fecal cultures resulted in growth of any yeast in 72.6% of the cases (CD 36.4%, UC 75.0%, HV 70.6%; Table 3). Among these, species of the Candida genus were most frequently observed throughout all groups (CD 66.7%, UC 58.3%, HV 58.8%), and the number of C. albicans positive cultures was higher for UC patients compared to CD and HV (CD 42.2%, UC 75.0%, HV 53.0%).

Table 2.

Patient characteristics. All data are expressed as mean (standard deviation) or number of observations (%).

	CD	UC	HV
n	33	12	17
Female (%)	21 (63.6)	5 (41.7)	13 (76.5)
Fecal calprotectin (µg/g feces)	281.03 (525.14)	210.83 (575.09)	23.00 (23.53)
BMI (kg/m2)	25.19 (4.48)	25.41 (2.82)
Duration of disease (years)	14.90 (9.17)	16.25 (10.66)
Medication use (%)
No IBD medication	1 (3.0)	0 (0.0)
Aminosalicylates	3 (9.1)	7 (58.3)
Thiopurines	12 (36.4)	7 (58.3)
Steroids	4 (12.1)	0 (0.0)
Biologicals* or JAK inhibitors	18 (54.5)	6 (50.0)
Other IBD medications	3 (9.1)	2 (16.7)
Proton pump inhibitors	5 (15.2)	4 (33.3)
Intestinal resection (%)	1 (3.0)	0 (0.0)
Clinical remission based on EPF (%)
Yes	19 (57.6)	11 (91.7)
No	10 (30.3)	1 (8.3)
Unclear	3 (9.1)	0 (0.0)
Harvey–Bradshaw index	2.67 (3.38)
Simple clinical colitis activity index		1.50 (1.60)

BMI, body mass index; EPF, electronic patient file. *Biologicals include anti-TNF, interleukin inhibitors, vedolizumab, and JAK inhibitors. Clinical remission based on electronic patient file is physician guided. Number of observations: fecal calprotectin, n = 60; Disease duration, n = 44; BMI, n = 34; HBI, n = 18; SCCAI n = 8. For Montreal disease severity classification, see Supplementary Table 1.

Table 3.

Cultured fungal species from fecal samples.

	CD	UC	HV
Samples cultured, n	33	12	17
Any yeast, n (%)	24 (36.4)	9 (75.0)	12 (70.6)
Babjeviella inositovora		1 (8.3)
Candida aaseri	1 (3.0)
Candida africana	1 (3.0)	4 (33.3)
Candida albicans	14 (42.4)	6 (50.0)	9 (53.0)
Candida dubliniensis	2 (6.1)	1 (8.3)
Candida maltosa/albicans	1 (3.0)		1 (5.9)
Candida orthopsilosis		1 (8.3)
Candida parapsilosis	6 (18.2)	4 (33.3)	2 (11.8)
Candida pararugosa	1 (3.0)
Candida tropicalis	2 (6.1)		1 (5.9)
Clavispora lusitaniae (=Candida lusitaniae)	1 (3.0)	1 (8.3)	1 (5.9)
Cyberlindnera jadinii (=Candida utilis)		1 (8.3)
Debaryomyces hansenii (=Candida famata)	1 (3.0)
Exophiala dermatitidis		1 (8.3)
Filobasidium magnum (=Cryptococcus magnus)	1 (3.0)
Geotrichum silvicola		1 (8.3)
Hanseniaspora uvarum	1 (3.0)
Kluyveromyces marxianus (=Candida kefyr)	1 (3.0)
Metschnikowia pulcherrima	1 (3.0)
Myerozyma guillermondii (=Candida guilliermondii)			1 (5.9)
Nakaseomyces glabratus (=Candida glabrata)		1 (8.3)
Ogataea polymorpha	1 (3.0)
Pichia fermentans (=Candida lambica)			1 (5.9)
Pichia kudriavzevii (=Candida krusei)	1 (3.0)	1 (8.3)	1 (5.9)
Saccharomyces cerevisiae	2 (6.1)	1 (8.3)
Torulaspora delbrueckii	1 (3.0)
Wickerhamomyces anomalus	1 (3.0)

Table indicates number of samples positive for each fungal species detected and percentage of total cultures in brackets. CD, Crohn's disease; UC, ulcerative colitis; HV, healthy volunteers.

Genetic evaluation of C. albicans strains shows inter- and intra-individual variation

Given earlier reports on genetic variation of C. albicans yeasts in patients with IBD and HV (Marakalala et al. 2013, Cavalieri et al. 2017, Di Paola et al. 2020, Li et al. 2022, van Thiel et al. 2022, Anderson et al. 2023), we performed genetic analysis of the obtained C. albicans strains. To this end, a novel panel of seven microsatellite loci was developed. For six loci, primers were newly developed and not in particular relation to any gene, and the seventh marker was the previously described locus CAI (Sampaio et al. 2003, Chavez-Galarza et al. 2010). Variation in microsatellite length is seen in all loci, and this variation also occurs across the different IBD diagnoses (Supplementary Fig. 1). A minimum spanning network (MSN) was generated based on the resulting STR analyses of 257 strains derived from 29 individuals (Fig. 1A). In total, 77 multilocus genotypes of C. albicans were observed. Genetic variation mainly occurred between individuals rather than within individuals (Supplementary Fig. 2A and B). Yet, visually striking, no UC strains are found in the lower right branch of the MSN (Fig. 1A). In addition, it appears that strains derived from UC patients show a lower number of genotypes per individual (Fig. 1B), albeit not statistically significant in this cohort. No statistical conclusions can thus be drawn with respect to the association between C. albicans genotype and disease using the microsatellite-based MSN analyses.

Figure 1.

Analysis of genetic diversity of feces-derived C. albicans strains through MSNs and principle component analysis. N = 29 individuals (n = 14 CD, n = 5 UC, n = 9 HV) were included in this analysis, resulting in n = 257 C. albicans strains. (A) MSN was determined based on seven microsatellite loci and is not clone corrected. Thickness and style of the line indicates similarity between different clusters based on Bruvo's distance. (B) Number of genotypes observed per patient. (C) PCA analysis of all included C. albicans strains mainly shows separation of diseases along PC2 rather than PC1. (D and E) Quantification of principle component scores along PC1 and PC2. UC has a distinct position along PC1 and PC2 compared with CD and HV. (F–H) PCA and corresponding boxplots based on corrected data in which the ‘median’ clone per patient is shown. Kruskal–Wallis test, * P < .05, **P < .01, ^****P < .001.

Principal component analysis of microsatellite data reveals separation of UC-derived C. albicans strains

A principal component analysis (PCA) was additionally performed to statistically differentiate the three study groups from one another. In order to perform this PCA, the STR length of each of the alleles of the seven loci was binarized. For each of the 14 measurements (7 loci × 2 alleles), a sample can thus only be positive for one length option (Supplementary File 1). This analysis excludes any assumptions on genetic relatedness between the different length options. Based on this analysis, a statistically significant separate cluster of UC-derived C. albicans strains can be observed, both along PC1 and PC2 (Fig. 1C–E), and all study populations have a significantly different distinct distribution along PC2 (Fig. 1E). However, as the inclusion of multiple strains per individual may skew statistical analysis, the same analysis was repeated but with only one strain per individual, in which the most abundant strain per individual was selected as representative, or in case of a tie, the median coordinates of all strains of an individual was used instead (Fig. 1F–H). Despite the loss of power, the distinct position of C. albicans strains from patients with UC on PC2 remained significant in comparison to HV (Fig. 1G–H). In contrast, discriminant analyses of principle components (DAPC) analysis is specifically able to determine relations between (possible) different clusters by optimizing variance between groups while minimizing within group variance (Jombart et al. 2010). The clone-corrected data were subjected to DAPC analysis, which resulted in four clusters (Supplementary Fig. 3A). The three disease groups were scattered across the four clusters, although UC-derived strains were not observed in clusters 1 and 3 (Supplementary Fig. 3B). Thus, it appears that C. albicans strains from UC patients have a distinct genotype compared to HV and patients with CD, represented by scoring higher on average on PC2.

Clinical parameters and enzyme activity of feces-derived C. albicans strains correlate

To investigate relations between genotype and phenotype of the feces-derived C. albicans strains, we proceeded with phenotypic investigation as described by means of enzyme release for a single strain per individual (van Thiel et al. 2022). Activities for four enzyme groups were determined, being proteinases, phospholipases, lipases, and esterases (Fig. 2A–E and Supplementary Fig. 2). The enzyme activity levels of some individual strains deviated significantly from the reference strain SC5314 (Supplementary Fig. 3A–D), but when all strains are compiled into their respective study populations, no statistically significant differences were established (Fig. 2A–E). We additionally examined clinical markers in serum samples of all included individuals (Supplementary Fig. 4). Fecal calprotectin, a commonly used clinical marker for intestinal inflammation, was only elevated in the CD group compared with HV (Fig. 2F). ASCA titers (IgA, IgG, and total) were also significantly altered between the total groups of CD and HV, although this statistically significant difference disappeared when the analysis was only focused on individuals of whom fecal culture was positive for C. albicans due to the reduced group sizes (Fig. 2F–J). In general, only marginal differences were observed in regards to biomarkers and the three study groups.

Figure 2.

Enzymatic activity of feces-derived C. albicans strains reveals no differentially active enzymes. (A–E) Enzyme activity was based in triplicate for one strain per individual. Each data point represents mean activity of a strain as determined in triplicate. (A) Proteinase activity. (B) Phospholipase activity. (C) Lipase activity. (D) Esterase activity. (E) Average activity across all enzymes. Horizontal dashed line indicates activity of reference strain SC5314. (F–J) Clinical fecal and serum titers. (F) Fecal calprotectin. (G) ASCA titers, IgA immunoglobulins. (H) ASCA titers, IgG immunoglobulins. (I) Total ASCA titers. (J) Serum 1,3-β-glucan levels. (K) Correlation plot of all determined enzyme activities and clinical parameters based on Spearman r. Letters and symbols denote significance levels. Spearman r, ^aP < .05, ^bP < .005. (L) Correlation between PC2 and proteinase activity. Spearman r = 0.417, P = .043, n = 24. (M) Correlation between esterase activity and PC2. Spearman r = .408, P = .048, n = 24. Data of all included patients, except for enzyme activity (n = 8 HV, n = 12 CD, n = 4 UC). Kruskal–Wallis or Mann–Whitney U test, *P < .05, **P < .01.

As previous research has shown that phenotypic characteristics of C. albicans strains may correlate with disease severity, Spearman correlation coefficients were determined to investigate relations between each of the acquired parameters (Fig. 2K). Of interest is that significant correlations were observed between fecal calprotectin levels and (total) ASCA levels (r = 0.464, P = .04, n = 20), lipase activity of feces-derived C. albicans strains and serum 1,3-β-glucan levels (r = −0.448, P = .036, n = 21). Lastly, significant trends were observed between fecal calprotectin and proteinase (r = 0.441, P = .046, n = 21). Thus, although the enzymatic activities are not significantly higher or lower in any of the three study groups, several correlations exist between clinical parameters and enzymatic activities.

C. albicans enzymatic activity correlates with PCA-based genetic separation

When linking phenotypic characteristics with PCA coordinate localization various other relevant differences emerge. A strong correlation was observed between PC2 and proteinase activity (Fig. 2K–L; r = 0.417, P = .043, n = 24) and esterase activity (Fig. 2M; r = 0.408, P = .048, n = 24) highlighting that the lack of significance in regard to UC and the above disease parameters is likely a power issue. For the DAPC analysis, no clear trends were found between genetic cluster and enzyme activity or clinical parameters, with the exception of a possible link between DAPC cluster and ASCA IgA titers (Supplementary Fig. 3C–L). In conclusion, genetic and phenotypic data modalities of feces-derived C. albicans strains were shown to be linked through PCA analysis.

Discussion

Genetic and functional differences among the intestinal fungal populations may contribute to intestinal inflammation. Earlier research already described a potential role for C. albicans in the IBD UC and CD (Barnes et al. 1990, Schaffer et al. 2007, Zhernakova et al. 2008, Iliev et al. 2012), and the most recent studies indicate that differences at the strain level were related to disease severity and clinical parameters like serum ASCA titers (Doron et al. 2021, Li et al. 2022). In the current study, we focused on feces-derived C. albicans strains from patients with either UC or CD, and included a cohort of control individuals. Using a newly developed microsatellite panel of seven loci, we investigated the genetic diversity of C. albicans strains within and between individuals using MSN and PCA analyses. PCA analyses revealed that the strains derived from UC patients tend to be more alike. Attempting to correlate genetic and functional diversity, one strain per individual was subjected to phenotypic examination by means of enzymatic activity determination. Correlations between enzymatic activity and clinical markers were observed, as well as associations between genetic distribution by PCA and enzymatic markers. Taken together, we show that our novel microsatellite panel is able to discriminate feces-derived C. albicans strains and that functional capacities may contribute to intestinal disease.

The intestinal tract harbors many different fungal species, but which ones of these are viable in the challenging and competitive intestinal environment is not fully elucidated yet. Recent efforts have led to a cultivated gut fungi compendium, but culturing of all feces-derived fungi remains challenging due to the wide variety in nutritional and atmospheric requirements (Hallen-Adams and Suhr 2017, Yan et al. 2024). In our current study, we cultured fecal samples of 62 individuals under aerobic conditions on various culture media, resulting in 27 different yeast species. Since C. albicans was the most frequently observed yeast in the fecal cultures, we proceeded our investigations with this species. Earlier reports of C. albicans in IBD mainly reported high abundances or overgrowth of this yeast in fecal samples of patients at the species level, and similar observations have been made through Internal Transcribed Spacer 1 (ITS1)-based metabarcoding on fecal or mucosal samples. Whether the patients included in this study have increased abundances of C. albicans was not determined, especially since recent studies indicated that genetic and functional sub-strain differences contribute to inflammation.

Given the large number of feces-derived C. albicans strains, we were able to perform genetic analysis on the obtained specimens. In the current study, we decided to employ analysis of seven microsatellite loci as this technique provides a relatively rapid and reproducible manner of assessing genetic variation (Alanio et al. 2017). Although polymerase enzymes are needed to determine said profiles, the results are not batch-dependent as would be the case for some other typing approaches. In addition, the technique is accessible and affordable, and does not require highly specialized knowledge to analyze the data. As such, we were able to map genetic diversity between and among individuals, either with an IBD or for healthy subjects. Earlier, similar analyses were performed on feces-derived C. albicans strains using AFLP, RAPD analysis, and whole-genome sequencing (Mehta et al. 1999, Bougnoux et al. 2006, Sciavilla et al. 2021, Li et al. 2022, van Thiel et al. 2022). Across all studies, a similar pattern arises describing mainly inter-individual genetic variation. Therefore, the microsatellite panel developed for this study is a useful tool for genotyping of feces-derived C. albicans strains. However, this technique also has a few limitations. The first limitation entails the development of microsatellite loci: the here suggested loci are specific to C. albicans and not related to any gene function in particular. Second, at this moment there is only a limited database available of the here presented C. albicans strains, whereas whole-genome data of various C. albicans isolates is currently also available. In future studies, it may thus be possible to select loci based on highly variable regions in the genome of C. albicans. Nevertheless, the advantages of microsatellite fingerprinting outweigh the limitations of this technique, and the here-described novel seven-locus panel is able to capture genetic variability of feces-derived C. albicans strains. As result hereof, genetic differences were observed between UC and CD or HV-derived C. albicans strains.

Genetic and phenotypic diversity of UC-derived fecal C. albicans was previously also described by Li et al. (2022). Genetic diversity was assessed through two methods, being whole genome sequencing and subsequent SNP density analyses. No immediate correlation between disease status and genotype was observed, although a separation of patient-derived strains was visible based on how aggressively the C. albicans strains behaved towards murine bone-marrow-derived macrophages. The damage-inducing potential of C. albicans was also translated to an in vivo experimental setting, where a high-damage strain would induce more severe inflammation (Li et al. 2022). In our study, we assessed phenotypic aspects of the obtained C. albicans strains by means of hydrolytic enzyme activity, divided in four categories, being proteinases, esterases, lipases, and phospholipases. By targeting enterocytes, these hydrolases may contribute to intestinal barrier dysfunction which is highly relevant to IBD pathophysiology (Lim et al. 2021, Talapko et al. 2021, Rao and Grover 2023). Thus, next to adhesion and hyphae formation, release of these enzymes entails an important aspect of IBD-related virulence mechanisms of C. albicans. While the enzyme release assays showed little variation between groups (i.e. UC vs. CD vs. HV), we did observe vast differences between individual strains (Supplementary Fig. 4). Based these results, we cannot yet conclude whether such differences possibly contribute to disease in preclinical or human inflammatory settings. In order to determine the exact relevant mechanisms of C. albicans-induced worsening of intestinal inflammation, the array of in vitro phenotypic assessments should be expanded and next evaluated in in vivo preclinical experiments. However, the in vivo situation of IBD is highly complex and influenced by multiple factors, including the (interplay between) bacterial and fungal communities, as well as genetic variations within the host that may influence immunological responses (Leonardi et al. 2018, Doron et al. 2021, Martini et al. 2023). Thus, considering this complexity, future strain-level identification of mechanisms contributing to disease (severity) will be a challenging albeit not impossible task.

In conclusion, we here present a novel seven-locus microsatellite typing panel that can describe inter- and intra-individual genetic variation among feces-derived C. albicans strains. PCA-based analysis, and visually the analysis of MSNs, shows separation and clustering of UC-derived fungal strains. Phenotypic analysis through assessment of enzyme release shows correlations with genotype and clinical markers, although not significantly with the disease subtype. Together, this research opens the possibility to further study yeasts in (inflammatory) bowel diseases through microsatellite locus assessment.

Acknowledgments

S.E.M.H. and F.H. received funding from the Dutch Ministry of Economic Affairs (Health∼Holland, PPP Allowance number LSHM20085). C.Y.P. received grants from Gilead and Perspectum, and consulting fees from Chemomab, Pliant, and NGM. T.B. wants to acknowledge support from the Distinguished Scientist Fellow Program of King Saud University, Riyadh, Saudi Arabia after retirement from the Westerdijk Fungal Biodiversity Institute. W.J.d.J. received consulting fees from Janssen Research, European Commision, and Alimentiv, and received honoraria from Alimentiv. W.J.d.J. is Chief Scientific Officer to and has stocks in AlBiomics BV. F.H. received grants from the European Society for Human and Animal Mycology, and has received materials from Pathonostics, OLM Diagnostics, EWC Diagnostics, Bruker MDx, and CHROMagar.

Author contributions

Isabelle A.M. van Thiel (Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing), Irini A.M. Kreulen (Formal analysis, Investigation, Writing – review & editing), Mèlanie V. Bénard (Investigation, Resources, Writing – review & editing), Marcus C. de Goffau (Formal analysis, Methodology, Visualization, Writing – review & editing), Bart Theelen (Methodology, Visualization, Writing – review & editing), Sigrid E.M. Heinsbroek (Conceptualization, Funding acquisition, Methodology, Writing – review & editing), Patrycja K. Zylka (Investigation, Visualization, Writing – review & editing), Cyriel Y. Ponsioen (Resources, Writing – review & editing), Teun Boekhout (Conceptualization, Methodology, Project administration, Supervision, Writing – review & editing), Wouter J. de Jonge (Funding acquisition, Writing – review & editing), Søren Rosendahl (Formal analysis, Methodology, Visualization, Writing – review & editing), René M. van den Wijngaard (Conceptualization, Funding acquisition, Methodology, Supervision, Writing – review & editing), and Ferry Hagen (Conceptualization, Data curation, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing)

Conflict of interest

None declared.

Funding

The here described research was funded by the Dutch Ministry of Economic Affairs (Health∼Holland, PPP Allowance number LSHM20085).

Data availability

Medical data will not be made available due to privacy regulations. Data generated in this study are available in the Supplementary Materials.

Author notes

Shared second authorship position