Fungal and bacterial gut microbiota differ between Clostridioides difficile colonization and infection
0
, ... Abstract
Aim: The bacterial microbiota is well-recognized for its role in Clostridioides difficile colonization and infection, while fungi and yeasts remain understudied. The aim of this study was to analyze the predictive value of the mycobiota and its interactions with the bacterial microbiota in light of C. difficile colonization and infection.
Methods: The mycobiota was profiled by ITS2 sequencing of fecal DNA from C. difficile infection (CDI) patients
Results: CDI patients were characterized by a significantly higher abundance of Candida spp. (MD 0.270 ± 0.089, P = 0.002) and Candida albicans (MD 0.165 ± 0.082, P = 0.023) compared to controls. Additionally, they were deprived of Aspergillus spp. (MD -0.067 ± 0.026, P = 0.000) and Penicillium spp. (MD -0.118 ± 0.043, P = 0.000) compared to CDC patients. Network analysis revealed a positive association between several fungi and bacteria in CDI and CDC, although the analysis did not reveal a direct association between Clostridioides spp. and fungi. Furthermore, the microbiota machine learning model outperformed the models based on the mycobiota and the joint microbiota-mycobiota model. The microbiota classifier successfully distinguished CDI from CDC [Area Under the Receiver Operating Characteristic (AUROC) = 0.884] and CDI from controls (AUROC = 0.905). Blautia and Bifidobacterium were marker genera associated with CDC patients and controls.
Conclusion: The gut mycobiota differs between CDI, CDC, and controls and may affect Clostridioides spp. through indirect interactions. The mycobiota data alone could not successfully discriminate CDC from controls or CDI patients and did not have additional predictive value to the bacterial microbiota data. The identification of bacterial marker genera associated with CDC and controls warrants further investigation.
Keywords
INTRODUCTION
Clostridioides difficile is an anaerobic, Gram-positive, spore-forming bacterium and the main causative organism of nosocomial diarrhea. Upon secretion of toxin A and/or B, and in some strains binary toxin, colonic inflammation is induced, giving rise to clinical manifestations that range from mild diarrhea with abdominal cramping to life-threatening pseudomembranous colitis, toxic megacolon, and death[1].
The disruption of the microbiome due to antibiotic use is a major risk factor for C. difficile infection (CDI), as is an age of 65 years and above[1]. In CDI patients, the bacterial microbiome displayed low diversity and decreased abundance of Bacteroides, Prevotella, and Bifidobacterium species, while exhibiting an elevated abundance of Clostridioides and Lactobacillus species[2,3]. It remains unknown whether these microbial changes are associated with risk factors leading to CDI, such as antibiotic use, or due to the presence of
Although C. difficile can lead to (recurrent) infections, the organism can also be part of the human gut microbiota without causing symptoms. Such asymptomatic colonization is estimated to occur between 0% and 15% of the healthy adult population[13]. Carriage of C. difficile is potentially a risk factor for transmission to a susceptible population, and because of the bacterium’s opportunistic nature, it may progress to infection under circumstances in which the microbiome is disturbed[14-18], though carriage does not contribute to transmission in a setting with low prevalence of hypervirulent PCR ribotype 027[19].
Aside from bacteria, fungi are an integral part of the microbiome. Recent research has emphasized the role of fungi in human health[20]. In CDI specifically, the gut mycobiota has been shown to deviate from
Importantly, C. difficile was shown to withstand otherwise toxic aerobic conditions when cultured in the presence of C. albicans[27]. Further supporting the importance of fungal-bacterial interactions, it was shown that C. difficile-directed antibiotics, including metronidazole and vancomycin, lead to the outgrowth and emergence of potentially pathogenic fungi[25]. As such, an intriguing triad of antibiotics, bacteria, and fungi exists in C. difficile colonization and infection[28].
Despite the evidence suggesting fungal-bacterial interactions in CDI, the predictive value of the gut mycobiota alone is scarcely investigated and unexplored in combination with the bacterial microbiota. Studying the mycobiota composition, its predictive value for C. difficile asymptomatic colonization and (recurrent) infection, and interactions with the bacterial microbiota could contribute to the overall understanding of the changes associated with the presence of C. difficile in the intestinal tract. Such knowledge may help understand the value of fungi in diagnosing and treating CDI and may be of interest for future therapeutic microbiome intervention strategies. In the present study, we evaluated whether mycobiota data alone or in combination with bacterial composition data allows discrimination of C. difficile carriers from hospitalized, non-colonized controls or CDI patients.
METHODS
Subjects and sample collection
Samples were derived from a multicenter prospective, hospital-based, case-control study performed in the Netherlands, as described earlier[29]. The aim of that study was to assess the prevalence of C. difficile colonization at hospital admission and the risk of subsequent infection or onward transmission[19]. A subset of samples from this study from two of the participating hospitals [Leiden University Medical Centre (LUMC), a tertiary academic hospital; Amphia Hospital, a large general hospital] was included in our analysis.
Patients were part of one of three study groups: patients with CDI, asymptomatic C. difficile colonization (CDC), and patients without C. difficile colonization (controls). The majority of the patients from each group had received antibiotic treatment up to three months prior to sample collection[29] [Supplementary Table 1]. Feces was collected within 72 h of hospital admission and cultured for C. difficile. Suspicious colonies were tested to confirm the presence of C. difficile and free toxins as previously described[19,29]. Patients with a positive C. difficile culture but no diagnosis of CDI within the first 72 h of admission were considered C. difficile colonized. Patients with a negative stool culture at hospital admission were included as controls. Samples of the CDI group were derived from patients hospitalized at the LUMC and diagnosed with CDI during the study period. CDI diagnosis was based on strict clinical criteria in combination with laboratory CDI testing (including detecting free C. difficile toxins and culturing). C. difficile isolates from CDI and CDC were additionally ribotyped with capillary PCR[30]. The identified ribotypes among CDI and CDC patients are described in Supplementary Figure 1.
The institutional review board of LUMC and the directing board of the Amphia Hospital had no objection to the performance of the study. A waiver for informed consent for feces collection of CDI patients was obtained. Feces of CDC and controls were collected under verbal consent; written informed consent was obtained for the collection of additional data.
Microbiota analysis
Microbiological and microbiota analyses have been performed and processed previously[29]. In this study, these data were supplemented with Internal Transcribed Spacer (ITS) 2 sequencing data. ITS2 sequencing was performed on available samples (Total: n = 105; CDI: n = 29; CDC: n = 38; controls: n = 38), and corresponding samples were selected from the available and processed 16S rRNA sequencing data.
Mycobiota analysis
DNA extraction, library preparation and sequencing
Based on sample availability, mycobiota profiling was performed on 105 of the 125 fecal samples, on which 16S rRNA gene sequencing was performed[29]. DNA extraction, quality control, library preparation, and ITS2 sequencing were performed according to the standard operating procedures of BaseClear B.V. (the Netherlands). DNA was extracted by a combination of chemical and mechanical lysis, in which two empty tubes were included as negative controls and two cellular microbial standards as positive controls (D6300 ZymoBIOMICS Microbial Community standard, Zymo Research, USA). The microbial standard contains ten microbial strains, of which two are fungi: Saccharomyces cerevisiae and Cryptococcus neoformans. After DNA extraction, concentrations were measured with Quant-iT dsDNA Assay kits [high sensitivity (Q33120) and broad range (Q33130), Invitrogen, USA], and DNA integrity was checked on an agarose gel. Subsequently, the ITS2 region was PCR-amplified with primers ITS3 (Forward:
Bioinformatic analyses: taxonomic assignment
Raw reads were processed according to the Q2-ITSxpress workflow[31]. Raw reads without barcodes and primers were imported in Qiime2 (version 2022-2)[32]. Subsequently, the conserved regions around the ITS gene were trimmed with ITSxpress, which has been shown to improve the accuracy of taxonomic classification[33]. The sequence variants were then identified in the unmerged, trimmed sequences with DADA2[34]. Next, the Qiime classifier was trained using the UNITE database (version 8.3, all eukaryotes)[35]. Fungal ITS classifiers were trained on the UNITE database on full reference sequences. Subsequently, sequence variants were classified with the trained classifier.
Bioinformatic analyses: pre-processing data
Data were imported in R (version 4.1.2)[36] with the Qiime2R package (version 0.99.6)[37]. In all samples, 959 taxa were identified, with 1,721,901 reads in total. The mean number of reads was 15,513 and ranged between a minimum of 255 and a maximum of 59,126. The sum of reads of three samples were outliers
Bioinformatic analyses: data quality control
Positive controls were included during DNA extraction (PCE) and sequencing (PCS), as well as negative controls for DNA extraction (NC) [Supplementary Figure 2]. The total number of reads from the filtered ASV table was visualized with ggplot2 (version 3.3.6)[38]. The mean total reads of positive (22.915 ± 2.686) and negative (748 ± 1.029) DNA extraction controls were not significantly different (P = 0.33, Wilcoxon Rank Sum Test, Supplementary Figure 2A). One of the negative controls, NC2, contained Malassezia globosa, which is often identified on human skin[39], as well as Cyberlindnera jadinii (teleomorph of Candida utilis) that has been isolated from distinct environments[40] [Supplementary Figure 2B].
Additionally, taxonomic compositions of positive controls were compared to the theoretical mock community. The two genera of the mock community - Cryptococcus and Saccharomyces - are present in a 1:1 ratio in the mock community, indicating a mean overrepresentation (1.48-fold) of Cryptococcus spp. and a mean underrepresentation (0.52-fold) of Saccharomyces spp. in the DNA and sequencing positive controls [Supplementary Figure 2C]. Correlation coefficients between the positive controls of DNA extraction (ρ = 1.00, P ≤ 0.001) and between positive controls of sequencing (ρ = 0.99, P ≤ 0.001) indicate high reproducibility between batches.
Bioinformatic analyses: downstream analyses
Pre-processed reads were used for downstream analyses in R (version 4.1.2)[36].
The number of reads and relative abundances from the ASV table were used to generate composition plots with microbiome (version 1.16.0)[41] and ggplot2 (version 3.3.6)[38].
Shannon diversity was computed at the ASV level with microbiome after reads had been rarified to the minimum sum of reads per sample with phyloseq (version 1.38.0)[42]. Overall significant differences were computed with a Kruskal-Wallis test from stats (version 4.1.2)[36]. P-values below 0.05 were considered statistically significant.
Differential abundant features were identified with ALDEx2 (version 1.26.0)[43] using centered log-ratio transformed values and default parameters, comparing CDI with controls and CDI with CDC at genus and species levels. The Wilcoxon Rank Sum Test of ALDEx2 was used for significance testing with Benjamini-Hochberg corrected P-values. P-values below 0.05 were considered statistically significant. Relative abundance of the differential abundant features was visualized for the three study groups and means were compared with the Wilcoxon Rank Sum Test of ggpubr (version 0.4.0)[44].
Machine learning was performed at the OTU and ASV level with the SIAMCAT package (version 2.1.3)[45] on 16S rRNA and ITS2 gene amplicon sequencing separately, as well as on the joint data. The CDI group was compared to the CDC group and to controls who received antibiotic treatment. Unsupervised feature filtering was performed with default parameters. Data was normalized with log.unit normalization and normalization parameters log.n0 = 1e-06, n.p = 2 and norm.margin = 1. A twice-repeated 5-fold cross-validation scheme was generated and the model was trained using the lasso method.
Fungal-bacterial network analyses were performed with SParse InversE Covariance Estimation for Ecological Association Inference, SpiecEasi (version 1.1.2)[46] combining 16S rRNA and ITS2 amplicon sequencing data with multi.spiec.easi. Prior to the network analyses, taxa were aggregated to genus level and pruned at 100 reads using microbiome and phyloseq. Network analyses for each study group were performed using meinshausen-Buhlmann's neighborhood selection with lambda.min.ratio = 0.01, nlambda = 20 and rep.num = 99. Networks were visualized with ggraph (version 2.0.5)[47].
RESULTS
C. difficile-infected patients are enriched with Candida spp.
To assess differences in the mycobiota of CDI, CDC, and controls, taxonomic profiles of the mycobiota using ITS2 sequencing were generated [Figure 1]. This revealed the presence of eukaryotic kingdoms, including Fungi, Viridaeplantae, and Metazoa. [Figure 1A]. For further analyses, fungal and unassigned kingdom reads were retained and comprised 468 out of 959 taxa and 68.8% of total reads. The relative abundance of the fungal kingdom was 0.75, 0.60, and 0.57 in CDI, CDC, and controls, respectively (± 0.32, 0.39, and 0.39 S.D.).
Figure 1. Taxonomic profiles of CDI patients, asymptomatic CDC patients, and controls (Control). (A) Eukaryotic taxonomic profiles at kingdom level are shown for the three study groups; (B) Fungal taxonomic profiles on genus level are shown for the ten most abundant genera. Taxa not belonging to the ten most abundant genera were summarized into “Other”. CDC: C. difficile colonization; CDI: C. difficile infection.
The most abundant genus in CDI patients was Candida spp. [Figure 1B], with a significant mean difference (MD) compared to controls (Figure 2A, MD 0.270 ± 0.089, P = 0.002, Wilcoxon Rank Sum Test). More specifically, the mean relative abundance of C. albicans was higher in CDI patients compared to controls (Figure 2B, MD 0.165 ± 0.082, P = 0.023). This was accompanied by a decrease in Saccharomyces spp., with a significantly higher Candida to Saccharomyces ratio in CDI patients compared to controls
Figure 2. Relative abundance of Candida spp. (A) and Candida albicans (B) in CDI patients, asymptomatic CDC patients, and controls (Control). Individual data points are displayed as circles. **P ≤ 0.01; *P ≤ 0.05 (Wilcoxon Rank Sum Test). CDC: C. difficile colonization; CDI: C. difficile infection.
Differential abundance testing confirmed the significant increases in CDI patients of Candida spp. (Punadj = 0.003) and C. albicans (Punadj = 0.027) compared to the controls, as well as the significant deprivation of Aspergillus (Padj = 0.044) and Penicillium species (Punadj = 0.002) compared to CDC patients. No differential abundant features were identified between CDC patients and controls.
Besides differences in the mycobiota composition, the median Shannon diversity increased from CDI patients to CDC patients to controls, though differences between groups were not significant
Together, these results indicate that CDI patients were most deviant from CDC patients and controls regarding the taxonomic composition and the diversity of the mycobiota, even though the majority of the patients from each group had received antibiotic treatment up to three months prior to sample collection[29].
The microbiota can discriminate C. difficile infection from colonization
Machine learning was used to investigate if models of the microbiota, mycobiota, and joint data could discriminate between CDI, CDC, and controls. With regard to the controls, antibiotic-treated controls were selected to exclude bias introduced by antibiotic treatment. The performance of the models was evaluated based on the Area Under the Receiver Operating Characteristic (AUROC), indicating the probability of classifying a randomly selected patient to the correct study group.
First, the bacterial microbiota, mycobiota, and joint models were used to assess their performance in discriminating between CDI and CDC patients, as well as between CDI patients and controls
The two microbiota models with an AUROC above 0.800 were selected to further assess the bacterial OTUs that were collectively characteristic for CDI and CDC patients [Figure 3] and for CDI patients and controls [Supplementary Figure 7]. In both cases, Clostridioides was identified as a distinct taxon in CDI patients. The nucleotide sequence of this Clostridioides OTU (OTU 1648581238) demonstrated a 100% sequence identity with Clostridioides difficile. Compared to CDC, CDI patients were additionally characterized by the genera Lactobacillus, Haemophilus, Bacteroides, and Enterococcus [Figure 3B]. Interestingly, CDC and controls shared OTUs from the Bifidobacterium and Blautia genera. The nucleotide sequences belonging to the identified Bifidobacterium OTUs resulted in 100% sequence identity with species Bifidobacterium breve and Bifidobacterium longum with subspecies longum, infantis, and siullum, while those belonging to the identified Blautia OTUs resulted in 100% sequence identity with Blautia wexlerae, Blautia luti, and Blautia massiliensis. In the CDC patients, Subdoligranulum was additionally part of the microbial signature [Figure 3B], while Collinsella was associated with the signature in the controls [Supplementary Figure 7B]. Together, these results demonstrate that CDC patients could not be successfully distinguished from controls based on machine learning with bacterial data. In comparison to CDI, Bifidobacterium and Blautia were bacterial marker genera for asymptomatic C. difficile carriage and non-carriage.
Figure 3. Signature of gut bacterial OTUs associated with CDI patients and asymptomatic CDC patients. (A) Cross-validation accuracy of the microbiota classifier is shown in the ROC curve with the AUC; (B) Median relative feature weights of the selected gut bacterial OTUs show the contribution of each marker OTU to the classification. Robustness of the selected features indicates the fraction of models containing the specific feature. The normalized values of the selected features across CDI and CDC are shown in the heatmap. AUC: Area Under the Curve; CDC: C. difficile colonization; CDI: C. difficile infection; ROC: Receiver Operating Characteristic.
Fungi and bacteria interact in C. difficile infected and colonized patients
Fungal-bacterial network analysis using SParse InversE Covariance Estimation for Ecological Association Inference (SpiecEasi) was performed for the CDI and CDC patients [Figure 4] and for the controls
Figure 4. Fungal-bacterial networks in CDI patients (A) and asymptomatic CDC patients (B). Interactions between fungi and bacteria are highlighted in green, while other interactions are displayed in grey. Kingdoms are represented by color; the mean relative abundance of the genus is represented by point size. Bacterial taxa in bold are part of the microbial signature as identified by machine learning. Fungal taxa in bold are part of the three most abundant fungi in the respective study group. (A) In CDI patients, interactions with the Clostridioides genus are additionally displayed. CDC: C. difficile colonization; CDI: C. difficile infection.
In CDI patients, twenty-nine fungal-bacterial associations were observed [Figure 4A]. Abundant fungi positively associated with bacteria included interactions between Pichia and Ruminococcus 2; Nakaseomyces and Ruminococcaceae UCG-002; and between Issatchenkia and Ruminococcus gauvreauii group. Interestingly, no associations were observed for the Candida genus. From the bacterial perspective, fungal-bacterial associations were observed for the bacterial marker genera associated with CDI [Figure 3]. Haemophilus was positively associated with an unidentified species of the Teratosphaeriaceae genus. Yet, the Clostridioides genus did not associate with fungi, but did show a negative association with Fusicatenibacter. Additionally, no fungal-bacterial associations were observed for the CDI bacterial marker genera Bacteroides, Bifidobacterium, and Lactobacillus, although Bacteroides did negatively associate with Bifidobacterium and Lactobacillus [Supplementary Figure 8A].
Nine fungal-bacterial associations were observed in CDC patients [Figure 4B]. Associations between the abundant fungi and bacteria included interactions between Debaryomyces and Holdemanella; Cladosporium and Erysipelotrichaceae UCG-003; and Wickerhamiella and Acidaminococcus. Similar to CDI, Candida spp. did not show any associations. From the bacterial viewpoint, no interactions were observed for the CDC bacterial marker genera [Supplementary Figure 8B].
In the controls who received antibiotics, Collinsella and Bifidobacterium were few among the bacterial marker genera [Supplementary Figure 7]. Collinsella abundance was positively associated with Anaerostipes, while Bifidobacterium negatively associated with Aspergillus [Supplementary Figure 8C].
DISCUSSION
Here, we show distinct fungal profiles in CDI compared to CDC and controls, most notably enrichment of Candida spp. and deprivation of Aspergillus spp. and Penicillium spp. in CDI. We also describe that mycobiota data alone could not successfully discriminate CDC from controls or CDI patients. Lastly, fungal-bacterial networks showed no association for Candida and no fungal-bacterial associations for Clostridioides in CDI and CDC. A negative bacterial-fungal association was observed between Bifidobacterium and Aspergillus in controls.
There is contradictory evidence for the role of Candida spp. in CDI in literature. Some studies are suggestive of a protective effect of Candida in CDI[48-50]. A lower prevalence of Candida colonization was observed in CDI patients[48,49], and susceptibility to CDI was reduced upon C. albicans colonization prior to a lethal challenge with C. difficile[50]. In contrast to the protective effect of Candida spp., other studies report an increase in Candida spp. in CDI[22,23,26]. Specifically, C. albicans could be involved in the pathogenesis of CDI by allowing C. difficile to thrive[27], reducing the healing of inflammatory lesions[51], driving Th17-mediated immune responses, and disrupting the gut microbiome[20]. In line with the pathogenic role of Candida spp. in CDI, a high prevalence of C. albicans was reported upon infection with C. difficile hypervirulent PCR ribotype 027[52,53]. Within our cohort, differential ribotype profiles were observed in CDI and CDC patients. Interestingly, the C. difficile binary toxin positive ribotype 078 was more frequently found in CDI patients compared to CDC patients (24.10% vs. 2.63%, P = 0.017, Fisher’s Exact Test), but the number of patients was not sufficiently high to analyze this group separately. The differential ribotype profiles may affect the mycobiota compositions and interactions with the host, although this triad of interactions has not been assessed hitherto. Our findings corroborate the pathogenic hypothesis, showing a higher mean relative abundance of Candida spp. and C. albicans in CDI patients compared to controls, which was accompanied by an increase in the ratio between Candida spp. and Saccharomyces spp. Antibiotic treatment is commonly associated with the overgrowth of Candida species[54,55]. Within our study, 96.6% of CDI and 60.5% of CDC patients received antibiotic treatment in the last three months prior to fecal sample collection. When we compared antibiotic-treated CDI patients (n = 28) to antibiotic-treated controls (n = 17), the mean relative abundance of Candida spp. was still significantly higher in CDI patients (P = 0.021, data not shown). This suggests Candida spp. could be involved in the pathogenesis of CDI rather than being an effect of antibiotic treatment per se. Nevertheless, the contradicting findings on the role of Candida spp. reflect the highly variable nature between patients and might, therefore, limit the use of Candida spp. as a marker for CDI.
Other fungi also show differential abundance between CDI, CDC, and controls. CDI patients were deprived of Aspergillus and Penicillium. Aspergillus has previously been related to healthy status[22,23,26] and therefore highlights its potentially beneficial role against CDI. Hypothetically, Aspergillus could interact with the immune system by affecting cytokine levels, as this genus was previously negatively correlated to IL-4 and TNF-α serum immune levels in a non-CDI group[22]. Moreover, the significant reduction of the Candida to Saccharomyces in controls might point towards a protective effect, as the fungal probiotic S. boulardii has been associated with a reduced risk of C. difficile-associated diarrhea[11,56]. Moreover, its supplementation during antibiotics prevented the development of CDI and reduced rCDI. Yet, evidence is insufficient for the clinical application of S. boulardii as a probiotic[57].
Although fungi have been studied as a marker for CDI in combination with host immune factors[23], this study provides novel information by combining mycobiota and microbiota data in a well-defined cohort of C. difficile infection patients, asymptomatic C. difficile colonization patients and hospitalized non-colonized controls. As opposed to CDI patients, CDC patients and controls were more alike in their bacterial and fungal microbiome. In fact, CDC patients could not be successfully distinguished from controls based on machine learning with bacterial data alone or when combined with fungal data. Bifidobacterium spp. and Blautia spp. were bacterial marker genera associated with both CDC and controls. In line with these findings, other studies have reported depletions of Bifidobacterium and Blautia in CDI patients compared to non-CDI patients[58] and were associated with less severe CDI-associated outcomes[59]. Among the Bifidobacterium OTUs, Bifidobacterium longum subsp. longum has been identified as a predictor for negative C. difficile status[60]. Both Bifidobacterium and Blautia include species that are known to be involved in bile acid metabolism and short-chain fatty acid production[61,62]. In most of the bifidobacterial marker OTUs, bacterial bile salt hydrolase (BSH) genes have been identified. Such mechanisms could contribute to colonization resistance against C. difficile[9,63-65].
The combination of machine learning and network analyses can infer potentially important fungal-bacterial interactions in CDI and CDC patients. Interestingly, the abundance of Clostridioides was not associated with fungi in CDI and CDC in these analyses, while it did identify the previously reported negative association with the Gram-positive Fusicatenibacter genus in CDI[29,66]. Moreover, the CDI-associated genus Bacteroides showed a negative association with Bifidobacterium. As Bifidobacterium was a bacterial marker genus associated with CDC patients and controls, increased relative abundance of Bacteroides could have a negative impact on beneficial bacterial-fungal interactions such as those between Bifidobacterium and Aspergillus as observed in antibiotic-treated controls. Although explorative, the fungal-bacterial network analyses described herein provide insight into differential interactions that occur during C. difficile colonization and infection.
FMT has proven highly effective for the treatment of rCDI, in which the restoration of the bacterial microbiota undoubtedly contributes to the successful eradication of CDI symptoms[9]. Emerging data suggest similar roles for the mycobiota. Increased fungal engraftment rates were observed in rCDI patients who responded to FMT compared to patients who did not respond to FMT[26]. More specifically, the genera Aspergillus, Penicillium, and Saccharomyces were significantly higher in responders post-FMT, while a dominance of Candida and C. albicans was observed in non-responders post-FMT[26]. Additionally, the efficacy of FMT in clearing C. difficile infection could be restored when C. albicans-colonized mice were treated with the antifungal agent fluconazole before FMT[26]. Within this study, CDI patients were depleted of Aspergillus, Penicillium, and Saccharomyces species, while the relative abundance of Candida spp. and
It was noted in our previous work that patient groups may not be comparable with regard to solid organ transplants, previous hospitalization, immunosuppressant use, and chemotherapy, as the majority of diagnosed CDI patients were recruited at the university-affiliated hospital (LUMC), whereas CDC patients and controls were more evenly recruited at a general hospital (Amphia hospital) and the LUMC[29]. Several of those clinical variables affect microbiota composition[67,68], yet its effect on the mycobiota remains to be elucidated. In a subgroup analysis on LUMC patients, results remained similar (data not shown). While the relative abundance of Candida spp. and C. albicans was not significantly higher in CDI compared to all controls, the mean ratio between Candida spp. and Saccharomyces spp. remained significantly higher in CDI compared to antibiotic-treated controls (P = 0.018). Moreover, CDI patients were, on average, depleted of Aspergillus spp. (P = 0.003) and Penicillium spp. (P = 0.015) compared to all controls and to CDC patients, respectively. The relative abundance of specific fungal genera, such as Candida spp., might thus be affected by the aforementioned clinical variables. In addition to that, the history of antifungal use remained unregistered, but this may have impacted the mycobiota composition significantly. Yet, fungal infections are uncommon in patients at risk for CDI in The Netherlands, except for severely immunocompromised patients and ICU patients receiving antifungal agents as part of selective gut decontamination. Moreover, 16S rRNA gene sequencing and ITS2 sequencing were performed separately. For ITS sequencing specifically, sequencing the full ITS region may provide greater taxonomic resolution compared to sequencing the ITS1 or ITS2 region alone[69]. However, future studies should rely on shotgun metagenomic sequencing to characterize additional kingdoms as well as genes to perform functional characterization of the microbiota and mycobiota. The characterization of additional kingdoms could be especially relevant to investigate the role of the relatively understudied Archaea and bacteriophages, which gained increasing attention with regard to FMT efficacy in CDI patients[70]. Lastly, the identification of bacterial markers for
In conclusion, we have shown that the gut mycobiota differs between C. difficile infection, asymptomatic carriage, and non-carriage. Most notably, C. difficile infected patients were enriched in Candida spp. and deprived of Aspergillus spp. and Penicillium spp. The mycobiota data alone could not successfully discriminate C. difficile carriers from hospitalized, non-colonized controls or CDI patients and did not have additional predictive value to the bacterial microbiota data. While not valuable as a diagnostic tool, indirect fungal-bacterial interactions may affect Clostridioides spp. and subsequent treatment effectiveness. The identification of bacterial marker genera associated with carriage and non-carriage warrants further investigation for diagnosing and treating CDI.
DECLARATIONS
Acknowledgments
We are indebted to all patients who participated in this study. We would like to thank Melanie Srodzinski, Inge van Duijn, Michelle de Raaf, and René Vermaire for their help in obtaining fecal samples. Additionally, we would like to thank the authors of the microbiota-based paper of this study and the persons acknowledged therein[29]. Preliminary results of the current study were presented (poster) at the annual meeting of the Royal Dutch Society of Microbiology (KNVM) (Arnhem, the Netherlands, 4-5 April 2023).
Authors’ contributions
Conception and design of the study: Smits WK, Kuijper EJ, Zwittink RD
Obtained funding: Kuijper EJ, Zwittink RD
Performed data acquisition and supervised the clinical study: Crobach MJT, Terveer EM, Kuijper EJ
Performed bioinformatic analyses: Henderickx JGE
Involved in result interpretation and manuscript preparation: Henderickx JGE, Smits WK, Kuijper EJ
Critically revised the manuscript: Crobach MJT, Terveer EM, Smits WK, Kuijper EJ, Zwittink RD
Reviewed and agreed to the published version of the manuscript: All authors
Availability of data and materials
The raw ITS (mycobiota) and 16S rRNA (microbiota) gene amplicon sequencing data have been deposited at the European Nucleotide Archive (ENA) at EMBL-EBI under accession number PRJEB61263 and PRJEB30586, respectively.
Financial support and sponsorship
This work was supported by the Netherlands Organization for Health Research and Development, ZonMw Grant 522008007.
Conflicts of interest
All authors declared that there are no conflicts of interest.
Ethical approval and consent to participate
As described previously[29], the institutional review board of Leiden University Medical Center (LUMC) and the directing board of the Amphia Hospital had no objection to the performance of the study. A waiver for informed consent for feces collection of CDI patients was obtained. Feces of CDC and hospitalized controls were collected under verbal consent; written informed consent was obtained for the collection of additional data.
Consent for publication
Not applicable.
Copyright
© The Author(s) 2023.
Supplementary Materials
REFERENCES
1. Smits WK, Lyras D, Lacy DB, Wilcox MH, Kuijper EJ. Clostridium difficile infection. Nat Rev Dis Primers 2016;2:16020.
2. Rea MC, O’Sullivan O, Shanahan F, et al. Clostridium difficile carriage in elderly subjects and associated changes in the intestinal microbiota. J Clin Microbiol 2012;50:867-75.
3. Hopkins MJ, Macfarlane GT. Changes in predominant bacterial populations in human faeces with age and with Clostridium difficile infection. J Med Microbiol 2002;51:448-54.
4. Sehgal K, Khanna S. Gut microbiome and Clostridioides difficile infection: a closer look at the microscopic interface. Therap Adv Gastroenterol 2021;14:1756284821994736.
5. Terveer EM, Vendrik KEW, Ooijevaar RE, et al. Faecal microbiota transplantation for Clostridioides difficile infection: four years’ experience of the Netherlands Donor Feces Bank. United European Gastroenterol J 2020;8:1236-47.
6. Hvas CL, Dahl Jørgensen SM, Jørgensen SP, et al. Fecal microbiota transplantation is superior to fidaxomicin for treatment of recurrent Clostridium difficile infection. Gastroenterology 2019;156:1324-32.e3.
7. Quraishi MN, Shabir S, Manzoor SE, et al. The journey towards safely restarting faecal microbiota transplantation services in the UK during the COVID-19 era. Lancet Microbe 2021;2:e133-4.
8. Wei S, Bahl MI, Baunwall SMD, et al. Gut microbiota differs between treatment outcomes early after fecal microbiota transplantation against recurrent Clostridioides difficile infection. Gut Microbes 2022;14:2084306.
9. Baktash A, Terveer EM, Zwittink RD, et al. Mechanistic insights in the success of fecal microbiota transplants for the treatment of Clostridium difficile infections. Front Microbiol 2018;9:1242.
10. Khoruts A, Sadowsky MJ. Understanding the mechanisms of faecal microbiota transplantation. Nat Rev Gastroenterol Hepatol 2016;13:508-16.
11. Szajewska H, Kołodziej M. Systematic review with meta-analysis: saccharomyces boulardii in the prevention of antibiotic-associated diarrhoea. Aliment Pharmacol Ther 2015;42:793-801.
12. Zhang Y, Saint Fleur A, Feng H. The development of live biotherapeutics against Clostridioides difficile infection towards reconstituting gut microbiota. Gut Microbes 2022;14:2052698.
13. Crobach MJT, Vernon JJ, Loo VG, et al. Understanding Clostridium difficile colonization. Clin Microbiol Rev 2018;31:e00021-17.
14. Eyre DW, Griffiths D, Vaughan A, et al. Asymptomatic Clostridium difficile Colonisation and Onward Transmission. PLoS One 2013;8:e78445.
15. Kong LY, Eyre DW, Corbeil J, et al. Clostridium difficile: investigating transmission patterns between infected and colonized patients using whole genome sequencing. Clin Infect Dis 2019;68:204-9.
16. Zacharioudakis IM, Zervou FN, Pliakos EE, Ziakas PD, Mylonakis E. Colonization with toxinogenic C. difficile upon hospital admission, and risk of infection: a systematic review and meta-analysis. Am J Gastroenterol 2015;110:381-90.
17. Blixt T, Gradel KO, Homann C, et al. Asymptomatic carriers contribute to nosocomial clostridium difficile infection: a cohort study of 4508 patients. Gastroenterology 2017;152:1031-41.e2.
18. Rodríguez C, Romero E, Garrido-Sanchez L, et al. Microbiota insights in Clostridium difficile infection and inflammatory bowel disease. Gut Microbes 2020;12:1725220.
19. Crobach MJT, Hornung BVH, Verduin C, et al. Screening for Clostridioides difficile colonization at admission to the hospital: a multi-centre study. Clin Microbiol Infect 2023;29:891-6.
20. Zhang F, Aschenbrenner D, Yoo JY, Zuo T. The gut mycobiome in health, disease, and clinical applications in association with the gut bacterial microbiome assembly. Lancet Microbe 2022;3:e969-83.
21. Stewart Sr DB, Wright JR, Fowler M, et al. Integrated meta-omics reveals a fungus-associated bacteriome and distinct functional pathways in clostridioides difficile infection. mSphere 2019;4:e00454-19.
22. Cao Y, Wang L, Ke S, et al. Analysis of intestinal mycobiota of patients with clostridioides difficile infection among a prospective inpatient cohort. Microbiol Spectr 2022;10:e0136222.
23. Cao Y, Wang L, Ke S, et al. Fecal mycobiota combined with host immune factors distinguish Clostridioides difficile infection from asymptomatic carriage. Gastroenterology 2021;160:2328-39.e6.
24. Sangster W, Hegarty JP, Schieffer KM, et al. Bacterial and fungal microbiota changes distinguish C. difficile infection from other forms of diarrhea: results of a prospective inpatient study. Front Microbiol 2016;7:789.
25. Lamendella R, Wright JR, Hackman J, et al. Antibiotic treatments for Clostridium difficile infection are associated with distinct bacterial and fungal community structures. mSphere 2018;3:e00572-17.
26. Zuo T, Wong SH, Cheung CP, et al. Gut fungal dysbiosis correlates with reduced efficacy of fecal microbiota transplantation in Clostridium difficile infection. Nat Commun 2018;9:3663.
27. van Leeuwen PT, van der Peet JM, Bikker FJ, et al. Interspecies interactions between Clostridium difficile and Candida albicans. mSphere 2016;1:e00187-16.
28. Romo JA, Kumamoto CA. Characterization of the effects of Candida gastrointestinal colonization on Clostridioides difficile infection in a murine model. In: Calderone R, editor. Candida species: methods in molecular biology. New York: Humana; 2022. pp. 271-85.
29. Crobach MJT, Ducarmon QR, Terveer EM, et al. The bacterial gut microbiota of adult patients infected, colonized or noncolonized by Clostridioides difficile. Microorganisms 2020;8:677.
30. Fawley WN, Knetsch CW, MacCannell DR, et al. Development and validation of an internationally-standardized, high-resolution capillary gel-based electrophoresis PCR-ribotyping protocol for clostridium difficile. PLoS One 2015;10:e0118150.
31. Rivers AR. Q2-ITSxpress: a tutorial on a QIIME 2 plugin to trim ITS sequences. Available from: https://forum.qiime2.org/t/q2-itsxpress-a-tutorial-on-a-qiime-2-plugin-to-trim-its-sequences/5780. [Last accessed on 5 Dec 2023].
32. Bolyen E, Rideout JR, Dillon MR, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat biotechnol 2019;37:852-7.
33. Rivers AR, Weber KC, Gardner TG, Liu S, Armstrong SD. ITSxpress: software to rapidly trim internally transcribed spacer sequences with quality scores for marker gene analysis. F1000Res 2018;7:1418.
34. Callahan BJ, McMurdie PJ, Rosen MJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: high-resolution sample inference from Illumina amplicon data. Nat Methods 2016;13:581-3.
36. R Core Team. R: a language and environment for statistical computing. In: R Foundation for Statistical Computing; 2022. Available from: https://www.bibsonomy.org/bibtex/7469ffee3b07f9167cf47e7555041ee7. [Last accessed on 5 Dec 2023].
37. Bisanz J. qiime2R: Importing QIIME2 artifacts and associated data into R sessions. Available from: https://rdrr.io/github/jbisanz/qiime2R/. [Last accessed on 5 Dec 2023].
39. Ianiri G, Heitman J, Scheynius A. The skin commensal yeast Malassezia globosa thwarts bacterial biofilms to benefit the host. J Invest Dermatol 2018;138:1026-9.
40. Sousa-Silva M, Vieira D, Soares P, Casal M, Soares-Silva I. Expanding the knowledge on the skillful yeast Cyberlindnera jadinii. J Fungi 2021;7:36.
41. Lahti L, Shetty S. microbiome R package. Available from: https://doi.org/10.18129/B9.bioc.microbiome. [Last accessed on 5 Dec 2023].
42. McMurdie PJ, Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 2013;8:e61217.
43. Fernandes AD, Macklaim JM, Linn TG, Reid G, Gloor GB. ANOVA-Like differential expression (ALDEx) analysis for mixed population RNA-Seq. PLoS One 2013;8:e67019.
44. Kassambara A. ggpubr: ‘ggplot2’ based publication ready plots. Available from: https://rpkgs.datanovia.com/ggpubr/. [Last accessed on 5 Dec 2023].
45. Wirbel J, Zych K, Essex M, et al. Microbiome meta-analysis and cross-disease comparison enabled by the SIAMCAT machine learning toolbox. Genome Biol 2021;22:93.
46. Kurtz ZD, Müller CL, Miraldi ER, Littman DR, Blaser MJ, Bonneau RA. Sparse and compositionally robust inference of microbial ecological networks. PLoS Comput Biol 2015;11:e1004226.
47. Pedersen TL. ggraph: an implementation of grammar of graphics for graphs and networks. Available from: https://rdrr.io/cran/ggraph/. [Last accessed on 5 Dec 2023].
48. Manian FA, Bryant A. Does Candida species overgrowth protect against Clostridium difficile infection? Clin Infect Dis 2013;56:464-5.
49. Nerandzic MM, Mullane K, Miller MA, Babakhani F, Donskey CJ. Reduced acquisition and overgrowth of vancomycin-resistant enterococci and Candida species in patients treated with fidaxomicin versus vancomycin for Clostridium difficile infection. Clin Infect Dis 2012;55 Suppl 2:S121-6.
50. Markey L, Shaban L, Green ER, et al. Pre-colonization with the commensal fungus Candida albicans reduces murine susceptibility to Clostridium difficile infection. Gut Microbes 2018;9:497-509.
51. Kumamoto CA. Inflammation and gastrointestinal Candida colonization. Curr Opin Microbiol 2011;14:386-91.
52. Blanco N, Walk S, Malani AN, et al. Clostridium difficile shows no trade-off between toxin and spore production within the human host. J Med Microbiol 2018;67:631-40.
53. Raponi G, Visconti V, Brunetti G, Ghezzi MC. Clostridium difficile Infection and Candida colonization of the gut: is there a correlation? Clin Infect Dis 2014;59:1648-9.
54. Gutierrez D, Weinstock A, Antharam VC, et al. Antibiotic-induced gut metabolome and microbiome alterations increase the susceptibility to Candida albicans colonization in the gastrointestinal tract. FEMS Microbiol Ecol 2020;96:fiz187.
55. Seelbinder B, Chen J, Brunke S, et al. Antibiotics create a shift from mutualism to competition in human gut communities with a longer-lasting impact on fungi than bacteria. Microbiome 2020;8:133.
56. Wombwell E. Saccharomyces boulardii prophylaxis for targeted antibiotics and infectious indications to reduce healthcare facility-onset Clostridioides difficile infection. Microbes Infect 2023;25:105041.
57. van Prehn J, Reigadas E, Vogelzang EH, et al. European Society of Clinical Microbiology and Infectious Diseases: 2021 update on the treatment guidance document for Clostridioides difficile infection in adults. Clin Microbiol Infect 2021;27 Suppl 2:S1-21.
58. Berkell M, Mysara M, Xavier BB, et al. Microbiota-based markers predictive of development of Clostridioides difficile infection. Nat Commun 2021;12:2241.
59. Lesniak NA, Schubert AM, Flynn KJ, et al. The gut bacterial community potentiates Clostridioides difficile infection severity. mBio 2022;13:e0118322.
60. Skraban J, Dzeroski S, Zenko B, Mongus D, Gangl S, Rupnik M. Gut microbiota patterns associated with colonization of different Clostridium difficile ribotypes. PLoS One 2013;8:e58005.
61. Jarocki P, Targoński Z. Genetic diversity of bile salt hydrolases among human intestinal bifidobacteria. Curr Microbiol 2013;67:286-92.
62. Song Z, Cai Y, Lao X, et al. Taxonomic profiling and populational patterns of bacterial bile salt hydrolase (BSH) genes based on worldwide human gut microbiome. Microbiome 2019;7:9.
63. Mullish BH, McDonald JAK, Pechlivanis A, et al. Microbial bile salt hydrolases mediate the efficacy of faecal microbiota transplant in the treatment of recurrent Clostridioides difficile infection. Gut 2019;68:1791-800.
64. Brown JRM, Flemer B, Joyce SA, et al. Changes in microbiota composition, bile and fatty acid metabolism, in successful faecal microbiota transplantation for Clostridioides difficile infection. BMC Gastroenterol 2018;18:131.
65. Aguirre AM, Sorg JA. Gut associated metabolites and their roles in Clostridioides difficile pathogenesis. Gut Microbes 2022;14:2094672.
66. Takada T, Kurakawa T, Tsuji H, Nomoto K. Fusicatenibacter saccharivorans gen. nov., sp. nov., isolated from human faeces. Int J Syst Evol Micr 2013;63:3691-6.
67. Gibson CM, Childs-Kean LM, Naziruddin Z, Howell CK. The alteration of the gut microbiome by immunosuppressive agents used in solid organ transplantation. Transpl Infect Dis 2021;23:e13397.
68. Huang J, Liu W, Kang W, et al. Effects of microbiota on anticancer drugs: current knowledge and potential applications. eBioMedicine 2022;83:104197.
69. Nilsson RH, Anslan S, Bahram M, et al. Mycobiome diversity: high-throughput sequencing and identification of fungi. Nat Rev Microbiol 2019;17:95-109.
70. Kazemian N, Ramezankhani M, Sehgal A, et al. The trans-kingdom battle between donor and recipient gut microbiome influences fecal microbiota transplantation outcome. Sci Rep 2020;10:18349.
71. Hernández Medina R, Kutuzova S, Nielsen KN, et al. Machine learning and deep learning applications in microbiome research. ISME Commun 2022;2:98.
Cite This Article
Export citation file: BibTeX | RIS
OAE Style
Henderickx JGE, Crobach MJT, Terveer EM, Smits WK, Kuijper EJ, Zwittink RD. Fungal and bacterial gut microbiota differ between Clostridioides difficile colonization and infection. Microbiome Res Rep 2024;3:8. http://dx.doi.org/10.20517/mrr.2023.52
AMA Style
Henderickx JGE, Crobach MJT, Terveer EM, Smits WK, Kuijper EJ, Zwittink RD. Fungal and bacterial gut microbiota differ between Clostridioides difficile colonization and infection. Microbiome Research Reports. 2024; 3(1): 8. http://dx.doi.org/10.20517/mrr.2023.52
Chicago/Turabian Style
Henderickx, Jannie G.E., Monique J.T. Crobach, Elisabeth M. Terveer, Wiep Klaas Smits, Ed J. Kuijper, Romy D. Zwittink. 2024. "Fungal and bacterial gut microbiota differ between Clostridioides difficile colonization and infection" Microbiome Research Reports. 3, no.1: 8. http://dx.doi.org/10.20517/mrr.2023.52
ACS Style
Henderickx, JGE.; Crobach MJT.; Terveer EM.; Smits WK.; Kuijper EJ.; Zwittink RD. Fungal and bacterial gut microbiota differ between Clostridioides difficile colonization and infection. Microbiome. Res. Rep. 2024, 3, 8. http://dx.doi.org/10.20517/mrr.2023.52
About This Article
Copyright
Data & Comments
Data
0
Cite This Article 7 clicks
Like This Article 4
likes
Comments
Comments must be written in English. Spam, offensive content, impersonation, and private information will not be permitted. If any comment is reported and identified as inappropriate content by OAE staff, the comment will be removed without notice. If you have any queries or need any help, please contact us at support@oaepublish.com.