Investigation of Intestinal Microbiota and Fecal Calprotectin in Non-Toxigenic and Toxigenic Clostridioides difficile Colonization and Infection

In this study, we aimed to evaluate the composition of the intestinal microbiota and level of fecal calprotectin in Clostridioides difficile-colonized patients. We included 102 C. difficile non-colonized (group I), 93 C. difficile colonized subjects (group II), and 89 diarrhea patients with C. difficile (group III). Chao1 index for alpha diversity and principal coordinate analysis was performed for beta diversity using QIIME. The mean relative abundance in each group was compared at the phylum and genus levels. Fecal calprotectin was measured using EliA calprotectin (Thermo Fisher Scientific). Group II showed significantly lower levels of Sutterella, Blautia, Ruminococcus, Faecalibacterium, Bilophila, and Ruminococcaceae and higher levels of Enterobacteriaceae compared to group I (p = 0.012, 0.003, 0.002, 0.001, 0.027, 0.022, and 0.036, respectively). Toxigenic C. difficile colonized subjects showed significantly lower levels of Prevotella, Phascolarctobacterium, Succinivibrio, Blautia, and higher levels of Bacteroides. The level of fecal calprotectin in group III was significantly higher than those in group I and group II (p < 0.001 for both). These data could be valuable in understanding C. difficile colonization process and the microbiota and inflammatory markers could be further studied to differentiate colonization from CDI.


Introduction
Clostridioides difficile, previously known as Clostridium difficile, is an anaerobic Gram-positive, spore-forming bacterium and a common cause of antibiotic-associated nosocomial diarrhea in the industrialized world [1,2]. After the use of antibiotics, microbiota-mediated colonization resistance is decreased partly by the reduced diversity of the intestinal microbiota [3][4][5][6][7]. Decreased diversity and alteration of the intestinal microbiota in C. difficile infection (CDI) has been shown in previous studies using various techniques from culture-based methods to high throughput sequencing [5,[7][8][9][10][11]. The current standard diagnosis of CDI is made from the presence of toxigenic C. difficile in patients with significant diarrhea, defined as unexplained and new-onset (≥3 unformed stools in 24 h) [12]. Unfortunately, current diagnostic tests for CDI detect toxigenic C. difficile or its toxins or toxin gene. Many assays cannot differentiate between colonization and infection. This causes a false classification of C. difficile-colonized patients with diarrhea as CDI [13,14]. Meanwhile, direct fecal biomarkers related to intestinal inflammation, such as calprotectin, were shown to significantly increase in CDI and were proposed as severity markers [15,16]. Although recent guidelines do not recommend inflammatory markers for diagnosis or prognosis determination in CDI due to insufficient data, the research is ongoing and could be evaluated for differentiation between colonized patients and infection [13,17]. To the best of our knowledge, data for the human microbiota composition and inflammatory status in C. difficile-colonized subjects is sparse and there are no data on the comparison of toxigenic and non-toxigenic C. difficile colonized subjects.
The present study aimed to evaluate the composition of the intestinal microbiota and inflammatory status in C. difficile colonized subjects and compare them with those in non-colonized control and patients with significant diarrhea. We also compared the differences between toxigenic and non-toxigenic C. difficile colonized subjects.

Clinical Samples
This study was approved by the Institutional Review Board of the Konkuk University Medical Center, Seoul, Korea (a tertiary referral hospital). In addition to this, all methods were performed in accordance with the relevant guidelines and regulations. From March 2018 to June 2019, we screened for C. difficile colonization in 812 residual fecal samples from individuals who visited our center for a general health examination. Korean adults (>30 years old) were included and subjects with obesity (body mass index > 25), other ethnicity or history of recent hospitalization were excluded. C. difficile colonization was firstly assessed by glutamate dehydrogenase (GDH) testing using VIDAS C. difficile GDH on the VIDAS instrument (bioMérieux, Marcy-l'Etoile, France) according to the manufacturer's instructions. Among the screened samples, 93 (11.4%) were GDH-positive. We defined "C. difficile colonization" as positive GDH in the absence of CDI symptoms [14]. We included randomly selected 102 GDH-negative samples (group I, C. difficile non-colonized) and 93 GDH-positive samples (group II, C. difficile colonized) for further study. Toxin status was determined by tcdB gene real-time PCR (Xpert C. difficile system, Cepheid, Sunnyvale, CA, USA). The Xpert C. difficile is an automated, real-time multiplex PCR assay, which detects the genes for tcdB, binary toxin (cdt), and a point mutation at position 117 associated with hypervirulent 027/NAP1/BI strains. A maximum valid Ct is 37 for tcdB and cdt according to the manufacturer's instructions. Among the 93 C. difficile-colonized subjects (group II), 13 (14%) were tcdB gene positive. In addition, we collected 89 GDH-positive samples from hospitalized patients with significant diarrhea (≥3 unformed stools in 24 h) [3,18]. They received antibiotics and most patients were from the hematologic and gastrointestinal department. We excluded the patients with laxative use in 48 h before sample collection to rule out other reasons for diarrhea.
They included 20 tcdB gene-negative and 69 tcdB gene-positive samples (group III, diarrhea with C. difficile). A total of 284 samples were analyzed and the characteristics of each group are summarized in Table 1. Upon reception, stool samples were stored at −70 • C until use. This study required neither study-specific nor any other interventions and the data were analyzed anonymously. Therefore, written informed consent from the enrolled patients was waived by the ethics committee.

Microbiome Analysis
DNA extraction from a stool sample was performed using a QIAamp DNA stool mini kit (Qiagen, Valencia, CA, USA) and bacterial 16S rRNA genes were amplified using an Ion 16S™ Metagenomics Kit (ThermoFisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol. The kit included 2 primer tubes with 3 primer sets that amplify the hypervariable regions of 16S rRNA (V2, V4, V8 and V3, V6-7, V9, respectively). After PCR amplicons were purified using Agencourt AMPure ® XP beads (Beckman Coulter, Indianapolis, IN, USA), sequencing libraries were prepared using an Ion Plus Fragment Library Kit and Ion Xpress™ Barcode Adapters (ThermoFisher Scientific, Waltham, MA, USA). Prepared libraries were quantified using a High Sensitivity DNA kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Template preparation and sequencing were performed using the Ion Chef™ System and Ion S5™ XL system with Ion 530™ Chip Kit (ThermoFisher Scientific, Waltham, MA, USA). After filtering out low quality and polyclonal reads, and trimming any adaptor sequences at the 3 end, the sequencing data were exported as FASTQ files and were processed using the Quantitative Insights into Microbial Ecology (QIIME) pipeline 1.9.1 [19]. After quality filtering, 26,634,236 sequences were obtained, with a mean of 89,165 sequences per sample (min: 16,585, max: 335,275). Operational taxonomic units (OTUs) were clustered based on 97% sequence similarity with at least 10 identical sequences and assigned against the curated Greengenes v.13.8 reference database at the QIIME web site (http://qiime.org/home_static/dataFiles.html).
Alpha and beta diversity measures were calculated by QIIME [20]. Alpha diversity assessment was based on observed OTUs, including unidentified OTUs, and Chao1 index. Microbial diversity was visualized using principal coordinate analysis (PCoA) and microbial diversity between samples was assessed by qualitative (unweighted UniFrac) and quantitative (weighted UniFrac) distances. The mean relative abundance (percentage among all reads) in each group was compared at the phylum and genus levels.

Calprotectin Measurement
After thawing, fecal samples were extracted using EliA Stool Extraction Kit (Thermo Fisher Scientific, Waltham, MA, USA). The extraction tubes were pre-filled with 750 µL of EliA Calprotectin Extraction Buffer. Fecal calprotectin was measured using EliA calprotectin on the Phadia 250 system (Thermo Fisher Scientific, Waltham, MA, USA). The wells of the EliA Calprotectin were coated with monoclonal antibodies against calprotectin, and the calprotectin levels were quantified by fluoroenzyme immunoassay. Phadia 250 measures calprotectin concentrations in ng/mL, and by using a conversion factor given by the lot-specific code of the EliA Calprotectin Well, the results were automatically converted to mg/kg.

Statistical Analysis
The difference between the continuous variables was analyzed using Student's t-test or the Mann-Whitney U test, and that between categorical variables was analyzed using the chi-squared test, Fisher's exact test, or the McNemar test. The Kruskal-Wallis test and one-way analysis of variance (ANOVA) were used to assess the differences between groups. Permutational multivariate analysis of variance (PERMANOVA) between groups was performed using QIIME. Statistical analysis was performed using MedCalc Statistical Software (v. 15.8, MedCalc Software, Mariakerke, Belgium) and IBM SPSS Statistics 22.0 (IBM Corporation, Armonk, NY, USA). p values less than 0.05 were considered statistically significant.

Comparison of Alpha Diversity of the Intestinal Microbiota Across Different Groups
We evaluated the differences in alpha diversity between group I (C. difficile-non colonized), group II (C. difficile-colonized), and group III (diarrhea with C. difficile). The distribution of the Chao1 indexes in each group is presented in Figure 1. The median Chao1 index of group III [median, interquartile range (IQR), 102.3, 85.7-127.1] was significantly lower than that of group I and group II (120.8, 105.7-136.9 and 119.5, 109.2-136.3, p < 0.001 for both). The median Chao1 index between group I and group II was not significantly different (p = 0.797). The median Chao1 index between subjects with non-toxigenic and toxigenic C. difficile was not significantly different in group II and III (p = 0.287 and 0.988, respectively).
indexes in each group is presented in Figure 1. The median Chao1 index of group III [median, interquartile range (IQR), 102.3, 85.7-127.1] was significantly lower than that of group I and group II (120.8, 105.7-136.9 and 119.5, 109.2-136.3, p < 0.001 for both). The median Chao1 index between group I and group II was not significantly different (p = 0.797). The median Chao1 index between subjects with non-toxigenic and toxigenic C. difficile was not significantly different in group II and III (p = 0.287 and 0.988, respectively).

Comparison of Beta Diversity of the Intestinal Microbiota Across Different Groups
PCoA using unweighted and weighted UniFrac matrix was performed to evaluate the beta diversity between each group. In unweighted UniFrac analysis, group I vs. III and group II vs. III clustered separately (both p = 0.001 by PERMANOVA), while group I and II could not be separated (p = 0.057 by PERMANOVA). In weighted UniFrac analysis, all three groups clustered separately (p = 0.017, 0.001, and 0.001 for group I vs. II, group I vs. III, and group II vs. III, respectively) ( Figure 2). In group II, subjects with non-toxigenic and toxigenic C. difficile clustered separately both in unweighted and weighted UniFrac matrix (both p = 0.001). In group III, subjects with non-toxigenic and toxigenic C. difficile clustered separately only in unweighted UniFrac matrix (p = 0.001 in unweighted and 0.721 in weighted UniFrac matrix).

Comparison of Beta Diversity of the Intestinal Microbiota Across Different Groups
PCoA using unweighted and weighted UniFrac matrix was performed to evaluate the beta diversity between each group. In unweighted UniFrac analysis, group I vs. III and group II vs. III clustered separately (both p = 0.001 by PERMANOVA), while group I and II could not be separated (p = 0.057 by PERMANOVA). In weighted UniFrac analysis, all three groups clustered separately (p = 0.017, 0.001, and 0.001 for group I vs. II, group I vs. III, and group II vs. III, respectively) ( Figure 2). In group II, subjects with non-toxigenic and toxigenic C. difficile clustered separately both in unweighted and weighted UniFrac matrix (both p = 0.001). In group III, subjects with non-toxigenic and toxigenic C. difficile clustered separately only in unweighted UniFrac matrix (p = 0.001 in unweighted and 0.721 in weighted UniFrac matrix). indexes in each group is presented in Figure 1.

Comparison of Beta Diversity of the Intestinal Microbiota Across Different Groups
PCoA using unweighted and weighted UniFrac matrix was performed to evaluate the beta diversity between each group. In unweighted UniFrac analysis, group I vs. III and group II vs. III clustered separately (both p = 0.001 by PERMANOVA), while group I and II could not be separated (p = 0.057 by PERMANOVA). In weighted UniFrac analysis, all three groups clustered separately (p = 0.017, 0.001, and 0.001 for group I vs. II, group I vs. III, and group II vs. III, respectively) ( Figure 2). In group II, subjects with non-toxigenic and toxigenic C. difficile clustered separately both in unweighted and weighted UniFrac matrix (both p = 0.001). In group III, subjects with non-toxigenic and toxigenic C. difficile clustered separately only in unweighted UniFrac matrix (p = 0.001 in unweighted and 0.721 in weighted UniFrac matrix).

Comparison of Mean Relative Abundance in Each Group at the Phylum Level
The comparison of mean relative abundance in groups I, II, and III at the phylum level is shown in Table 2. Group III showed significantly lower levels of Bacteroidetes (p < 0.001 and 0.001 compared with groups I and II, respectively) and significantly higher levels of Proteobacteria compared to group I (p < 0.001). Comparing subjects with non-toxigenic and toxigenic strains in group II, those with toxigenic C. difficile colonized subjects showed significantly lower levels of Firmicutes (p < 0.001). In group III, there was no significant difference in composition between patients infected with non-toxigenic and toxigenic C. difficile strains (Table 3).

Comparison of Mean Relative Abundance in Each Group at the Genus Level
The comparison of mean relative abundance in groups I, II, and III at the genus level is shown in Table 4. Group III showed a significantly lower mean relative abundance of many genera (Prevotella, Lachnospira, Blautia, Ruminococcus, Phascolarctobacterium, Faecalibacterium, Bilophila, unassigned genus of Ruminococcaceae, and Lachnospiraceae) compared with those in groups I and II. Group III also showed a significantly higher mean relative abundance of Parabacteroides, Serratia, Veillonella, Enterococcus, and an unassigned genus of Enterobacteriaceae. Group II showed a significantly lower abundance of Sutterella, Blautia, Ruminococcus, Faecalibacterium, Bilophila, and an unassigned genus of Ruminococcaceae and higher abundance of an unassigned genus of Enterobacteriaceae compared with those in group I (p = 0.012, 0.003, 0.002, 0.001, 0.027, 0.022, and 0.036, respectively). Moreover, toxigenic C. difficile colonized subjects showed a significantly lower abundance of Prevotella, Phascolarctobacterium, Succinivibrio, Blautia, and an unassigned genus of Ruminococcaceae, and higher abundance of Bacteroides and an unassigned genus of Enterobacteriaceae compared with non-toxigenic C. difficile colonized subjects (p = 0.014, <0.001, 0.019, 0.006, <0.001, 0.035, and 0.044, respectively). In group III, there was no significant difference between patients with non-toxigenic and toxigenic C. difficile (Table 5). Only genera with mean relative abundance >1% were described in Table 4. * unspecified to a specific genus within this family.

Comparison of Calprotectin Levels between Groups
We compared fecal calprotectin levels in groups I, II, and III and its distribution is presented in Figure 3. The fecal calprotectin level in group III (median and IQR, 199.0 and 34.3-658.3 mg/kg) was significantly higher than those in group I and group II (15.0, 15.0-39.5 and 23.0, 15.0-57.0 mg/kg, respectively, p < 0.001 for both). There was no significant difference between group I and group II (p = 0.256). The fecal calprotectin levels were not significantly different between individuals with non-toxigenic and toxigenic C. difficile in both group II and group III (p = 0.496 and 0.273, respectively).

Comparison of Calprotectin Levels Between Groups
We compared fecal calprotectin levels in groups I, II, and III and its distribution is presented in Figure 3. The fecal calprotectin level in group III (median and IQR, 199.0 and 34.3-658.3 mg/kg) was significantly higher than those in group I and group II (15.0, 15.0-39.5 and 23.0, 15.0-57.0 mg/kg, respectively, p < 0.001 for both). There was no significant difference between group I and group II (p = 0.256). The fecal calprotectin levels were not significantly different between individuals with nontoxigenic and toxigenic C. difficile in both group II and group III (p = 0.496 and 0.273, respectively).

Discussion
The asymptomatic colonization rate of toxigenic C. difficile among healthy individuals ranges from 1.8 to 15% [14,21,22]. However, most studies were based on point prevalence detection and methodology and the target population was variable. In our study, C. difficile colonization among the screened population were 11.4% and the toxigenic portion among colonized C. difficile was relatively low (14%), which was lower than our previous studies performed in the same institution using the same method [23]. This difference might be due to different populations and different criteria in this study, which only included adult, non-obese individuals among visitors for a general health examination. The toxigenic proportion was much higher in hospitalized patients with diarrhea [23].
As expected, the alpha diversity index (Chao1), as shown in Figure 1, was significantly lower in group III (diarrhea with C. difficile) compared with those in group I (C. difficile non-colonized) and group II (C. difficile colonized). Previous studies also showed decreased alpha diversity in CDI or antibiotic exposed group compared with the control [8,24]. As decreased species diversity could occur in many hospitalized patients without CDI, this feature might not be specific for CDI and could be a common feature in hospitalized patients with antibiotic exposure. The development of overt CDI needs additional host factors, such as immunity, age, or hospital stay [3,4,25]. In this study, the alpha diversity index between group I and group II did not significantly differ. Thus, the C. difficile colonization itself in healthy individuals does not seem to affect species diversity. The present study also showed that the alpha diversity in group II and group III was not affected by toxigenic status.
In PCoA analysis, group III clustered separately from group I and II by unweighted and weighted UniFrac analysis, which was consistent with previous studies [8,24]. This indicates that hospitalized patients with diarrhea have different beta diversity to healthy individuals. We could find a significant difference between group I and II by weighed UniFrac analysis. Thus, although the

Discussion
The asymptomatic colonization rate of toxigenic C. difficile among healthy individuals ranges from 1.8 to 15% [14,21,22]. However, most studies were based on point prevalence detection and methodology and the target population was variable. In our study, C. difficile colonization among the screened population were 11.4% and the toxigenic portion among colonized C. difficile was relatively low (14%), which was lower than our previous studies performed in the same institution using the same method [23]. This difference might be due to different populations and different criteria in this study, which only included adult, non-obese individuals among visitors for a general health examination. The toxigenic proportion was much higher in hospitalized patients with diarrhea [23].
As expected, the alpha diversity index (Chao1), as shown in Figure 1, was significantly lower in group III (diarrhea with C. difficile) compared with those in group I (C. difficile non-colonized) and group II (C. difficile colonized). Previous studies also showed decreased alpha diversity in CDI or antibiotic exposed group compared with the control [8,24]. As decreased species diversity could occur in many hospitalized patients without CDI, this feature might not be specific for CDI and could be a common feature in hospitalized patients with antibiotic exposure. The development of overt CDI needs additional host factors, such as immunity, age, or hospital stay [3,4,25]. In this study, the alpha diversity index between group I and group II did not significantly differ. Thus, the C. difficile colonization itself in healthy individuals does not seem to affect species diversity. The present study also showed that the alpha diversity in group II and group III was not affected by toxigenic status.
In PCoA analysis, group III clustered separately from group I and II by unweighted and weighted UniFrac analysis, which was consistent with previous studies [8,24]. This indicates that hospitalized patients with diarrhea have different beta diversity to healthy individuals. We could find a significant difference between group I and II by weighed UniFrac analysis. Thus, although the total species diversity did not change in C. difficile colonization (based on alpha diversity), the composition and abundance of microbiota were changed in C. difficile colonization. Regarding toxin status, subjects with non-toxigenic and toxigenic C. difficile clustered separately both in unweighted and weighted UniFrac matrix, indicating that the intestinal microbiota is further changed distinctly upon colonization of toxigenic C. difficile strains compared to non-toxigenic strains. In patients of group III, non-toxigenic and toxigenic C. difficile also clustered separately, but significant difference was observed only in the unweighted UniFrac matrix. Disruption of intestinal microbiota composition occurs in many hospitalized patients, which could weaken the difference between those colonized with non-toxigenic and toxigenic strains.
Our study showed different patterns of the composition of fecal microbiota in patients with diarrhea, as shown previously [8,26]. Specifically, Group III patients showed lower and higher abundance of Bacteroidetes and Proteobacteria, respectively, at the phylum level, and lower relative abundance of many genera, including butyrate-producing organisms and health-promoting bacteria [9,27]. Increased Parabacteroides or Enterococcus, which reflect a blooming phenomenon by reduced ecological niche competition [7,8,28,29], were also observed in group III which included various hospitalized patients. These features were noted similarly in non-toxigenic and toxigenic C. difficile, suggesting again that this disruption of microbiota was not specific to patients with toxigenic C. difficile.
There is very limited data on the composition of intestinal microbiota in C. difficile colonization and a few studies only described a very limited number of subjects [8,14,30,31]. In this study, we could find several distinct features in C. difficile colonization, including those due to the toxigenic status of the strains. Group II (C. difficile colonized) did not show significant differences compared with non-colonized subjects (group I) at the phylum level and the results at phylum levels were the sum of the changes of many genera. At the genus level, group II showed a significantly lower abundance of many genera and a higher abundance of Enterobacteriaceae compared to group I. This pattern of change was partly similar to that observed in CDI or hospitalized patients, indicating that the genus compositions of C. difficile colonized subjects lies between non-colonized and hospitalized patients. Moreover, toxigenic C. difficile-colonized subjects showed significant changes, such as a lower abundance of Prevotella, Phascolarctobacterium, Succinivibrio, Blautia, and an unassigned genus of Ruminococcaceae and higher abundance of Bacteroides and an unassigned genus of Enterobacteriaceae than non-toxigenic C. difficile-colonized subjects. These features were closer to the patterns observed in hospitalized patients with diarrhea. We could not find a significant difference in hospitalized patients with diarrhea based on the presence or absence of toxigenic C. difficile.
Fecal calprotectin is a 36.5 kDa molecule derived from the cytoplasm of neutrophils, mononuclear cells, and squamous epithelial cells and reflects the inflammatory status of the intestine [32]. The proposed clinical utility of this marker has been focused on inflammatory bowel disease (IBD) [33]. In CDI, fecal calprotectin has been shown to reflect severity [15,16,34]; however, this marker showed a non-specific increase in various inflammatory conditions of the intestines, thereby limiting its diagnostic utility [35,36]. In this study, we showed that C. difficile-colonized subjects, regardless of toxin status, had comparable fecal calprotectin levels to those of non-colonized subjects and significantly lower levels than in diarrhea patients with C. difficile. The C. difficile-colonized subjects seemed to show a non-inflammatory status, although they showed changes in the composition of intestinal microbiota (beta diversity). The diarrhea patients had significantly higher fecal calprotectin levels, reflecting in high inflammatory status, and the toxin status did not show significant differences in calprotectin levels, indicating the non-specific increase of this marker, as suggested by previous studies [34][35][36].
This study had several limitations. The cause-effect relationship between specific alterations of the microbiota and clinical status or colonization could not be elucidated in this study design. Many covariates such as age or food intake habit could also affect the intestinal microbiota composition [37]. The median age of group III was significantly higher than the other groups and old age itself could result in decreased diversity and change composition of gut microbiota in group III [7,37]. The old aged subjects were rare in general health examination and enrollment of elderly subjects was difficult.
Several studies recently showed data regarding differences in the gut microbiota between sexes [38,39]. In this study, we could not evaluate data based on sex differences and could be further studied. In addition, the results of the OTU assignment could be different between algorithms or programs used for OTU analysis [37,40].

Conclusions
There were significant alterations in the alpha and beta diversity in intestinal microbiota in hospitalized patients with diarrhea, but these alterations were not significantly affected by the toxin status of C. difficile. We found significant differences in the composition of intestinal microbiota in apparently healthy C. difficile colonized subjects in comparison to C. difficile non-colonized subjects. Despite the significant difference in C. difficile colonization such as lower levels of Sutterella, Blautia, Ruminococcus, Faecalibacterium, Bilophila, and Ruminococcaceae and higher levels of Enterobacteriaceae, inflammatory status and alpha diversity were not significantly different between C. difficile colonized and non-colonized subjects. Further investigation of the intestinal microbiota and immunological studies could provide some insight into the conditions that allow for colonization and protect against disease progression.