*Article* **Nonalcoholic Fatty Liver Disease—A Novel Risk Factor for Recurrent** *Clostridioides difficile* **Infection**

**Lara Šamadan <sup>1</sup> , Mia Jeliˇci´c <sup>2</sup> , Adriana Vince 1,2 and Neven Papi´c 1,2,\***


**Abstract:** Recurrent *Clostridioides difficile* infections (rCDI) have a substantial impact on healthcare systems, with limited and often expensive therapeutic options. Nonalcoholic fatty liver disease (NAFLD) affects about 25% of the adult population and is associated with metabolic syndrome, changes in gut microbiome and bile acids biosynthesis, all possibly related with rCDI. The aim of this study was to determine whether NAFLD is a risk factor associated with rCDI. A retrospective cohort study included patients ≥ 60 years hospitalized with CDI. The cohort was divided into two groups: those who were and were not readmitted with CDI within 3 months of discharge. Of the 329 patients included, 107 patients (32.5%) experienced rCDI. Patients with rCDI were older, had higher Charlson Age–Comorbidity Index (CACI) and were more frequently hospitalized within 3 months. Except for chronic kidney disease and NAFLD, which were more frequent in the rCDI group, there were no differences in other comorbidities, antibiotic classes used and duration of antimicrobial therapy. Multivariable Cox regression analysis showed that age >75 years, NAFLD, CACI >6, chronic kidney disease, statins and immobility were associated with rCDI. In conclusion, our study identified NAFLD as a possible new host-related risk factor associated with rCDI.

**Keywords:** *Clostridioides difficile* associated disease; CDI; nonalcoholic fatty liver disease; NAFLD; NASH; recurrent disease

## **1. Introduction**

*Clostridioides difficile* infection (CDI), with increasing prevalence and mortality rates, is the leading cause of healthcare-associated diarrhea [1]. On hospital admission, 7% of patients are already colonized with *C. difficile* and another 21% become infected during hospitalization [2]. Furthermore, 20 to 30% of patients will develop symptomatic CDI recurrence within 2 weeks of completion of accurate therapy, and each episode increases the risk of future episodes by about 20% [3,4]. Therefore, recurrent *C. difficile* infections (rCDI) cause substantial impact on healthcare systems with limited and often expensive therapeutic options [5]. This requires identification of patients at high risk of rCDI. So far, major risk factors suggested to be associated with rCDI were age, use of antibiotics for non-*C. difficile* infection, gastric acid suppression and infection with a hypervirulent strain (NAP1/BI/027) [5]. Age is the most frequently reported risk factor for rCDI; the probability of rCDI is 27% in the age group between 18 and 64 and 58.4% in patients older than 65 years [4,6]. However, the literature data on other host risk factors are inconsistent and often contradictory. Severe underlying comorbidities, chronic renal disease, CDI severity, prolonged hospital stay, and nursing home residency were inconsistently reported as additional risk factors [5–8]. Recently, it was suggested that metabolic syndrome and its components, primarily obesity and diabetes mellitus (DM), might be associated with rCDI [8,9]. Metabolic syndrome is linked with changes in gut microbiome that might serve as protective intestinal flora for *C. difficile* germination and disease development [10].

**Citation:** Šamadan, L.; Jeliˇci´c, M.; Vince, A.; Papi´c, N. Nonalcoholic Fatty Liver Disease—A Novel Risk Factor for Recurrent *Clostridioides difficile* Infection. *Antibiotics* **2021**, *10*, 780. https://doi.org/10.3390/ antibiotics10070780

Academic Editor: Guido Granata

Received: 14 May 2021 Accepted: 25 June 2021 Published: 27 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Nonalcoholic fatty liver disease (NAFLD) is the most common chronic liver disease, affecting about 25% of the adult Western population [11,12]. NAFLD represents a spectrum of chronic liver pathologies, from simple steatosis to nonalcoholic steatohepatitis, cirrhosis, and hepatocellular carcinoma [11,12]. There is a close relationship between NAFLD and metabolic syndrome, clustering visceral overweight, dyslipidemia, insulin resistance and arterial hypertension [11–13]. Some consider NAFLD as the hepatic manifestation of metabolic syndrome. Importantly, it is well described that changes in the gut microbiome promote the development of NAFLD through inflammatory processes, insulin resistance and bile acids metabolism, all possibly related with increased susceptibility to CDI [14,15]. We have recently shown that NAFLD is an independent risk factor for in-hospital CDI in elderly patients treated with systemic antimicrobial therapy [16]. Similarly, a case–control study showed that hospitalized patients with CDI more frequently have NAFLD and metabolic syndrome, as compared to age- and gender-matched controls hospitalized for non-CDI diarrhea [17]. However, the question remains if NAFLD is a risk factor for CDI recurrence. Due to the high and increasing burden of NAFLD in the Western population, identification of NAFLD as a novel and common risk factor could improve current CDI management strategies.

Therefore, the aim of this retrospective cohort study was to determine whether NAFLD is an independent risk factor associated with CDI recurrence in elderly patients who are at the highest risk for increased mortality, hospitalization, complications and healthcare costs.

#### **2. Results**

A total of 329 patients were included in the study (196, 59.6% females with the median age of 77, interquartile range (IQR) 71–83 years). Among the entire cohort, 107 (32.5%) patients developed recurrent CDI within 90 days of index hospitalization discharge or by the end of index CDI therapy (total of 171 episodes). Of those, 67 patients developed rCDI by 28 days (a cumulative probability of recurrence of 20.7%, 95% CI = 13.5–28.9%) and 40 patients by 90 days (33.3%, 26.5–40.4%). Three, four and five or more documented episodes were registered in the studied period in 38, 11 and 5 patients, respectively.

Patients with rCDI were older (78 years, IQR 74–84 vs. 77 years, 71–81), had higher Charlson Age–Comorbidity Index (CACI) (6, IQR 5–7 vs. 5, IQR 4–6) and were more frequently hospitalized within 3 months (81.3% vs. 68.9%). Except for chronic kidney disease (CKD) (26.2% vs. 6.8%) and NAFLD (34.6% vs. 18.5%), which were more frequent in the rCDI group, there were no differences in other comorbidities, as presented in Table 1. Notably, the majority of patients with CKD in our cohort did not undergo chronic dialysis; two patients in CDI and four in rCDI group were in a dialysis program.

High proportions of patients in both groups were nursing home residents (37.4% vs. 33.3%). Regarding chronic medications, patients without recurrence were more frequently receiving statins (23.4% vs. 13.1%), while there were no differences in prescription of perioral anti-diabetic, metformin or insulin. The majority of patients in both groups were receiving histamine-2 receptor antagonist and/or proton pump inhibitors (47.7% vs. 43.7%).

Use of specific of antibiotic classes per patient and duration of therapy prior the first episode of CDI were similar; fluoroquinolones were most frequently prescribed (29.9% vs. 27.0%), followed by 3rd generation cephalosporins (20.5% vs. 13.5%) and amoxicillin/clavulanate (19.6% vs. 22.1%). The reasons for antibiotic prescription prior the first episode of CDI were similar in both groups—most commonly for urinary tract infections (28.9% vs. 30.2%), respiratory infections (26.2% vs. 23.4%) and skin/soft tissue infections (11.2% vs. 5.9%). However, there were 9 patients with rCDI (8.4%) and 42 (18.9%) with CDI, for whom the reason for antibiotic prescription was not clear.

On index hospitalization, the majority of patients presented with severe CDI, median ATLAS score of 5 (IQR 4–7) in both groups, and elevated inflammatory markers (C-reactive protein and white blood cells count), as presented in Table 2. There were no differences in APRI or FIB4 scores between groups; APRI score >1.5 had 4.0% of patients in CDI and 2.8% in rCDI group. FIB4 score >3.25 had 10.8% in CDI and 10.3% of patients in rCDI group.


**Table 1.** Baseline patients' characteristics.

<sup>a</sup> Fisher exact or Mann–Whitney U test, as appropriate; <sup>b</sup> IQR, interquartile range; <sup>c</sup> data available for 299 patients.

> There was no difference in the choice of CDI treatment between groups; 118 patients were treated with metronidazole, 182 with vancomycin and 29 received combined therapy. There were 39 patients considered metronidazole unresponsive who were switched to vancomycin. Importantly, the majority of our cohort received non-CDI antibiotics during hospitalization (62.6% in rCDI and 67.6% in CDI group). As shown in Table 2, the most commonly used antibiotics were piperacillin/tazobactam (23.4% vs. 18.5%), 3rd generation cephalosporins (23.4% vs. 21.6%) and carbapenems (10.3% vs. 14.9%). The reasons for non-CDI therapy were respiratory tract infection (15.5%), urinary tract infection (30.7%), sepsis (7.9%) and suspected bacteriaemia of gastrointestinal origin (6.1%). There were no differences between groups in the choice of antibiotic therapy, number of antibiotic classes prescribed or duration of antimicrobial therapy.


**Table 2.** Clinical, laboratory and treatment characteristics during the first episode of CDI.

<sup>a</sup> Fisher exact or Mann–Whitney U test, as appropriate.

In order to identify potential risk factors for CDI recurrence, we performed a multivariable Cox regression analysis. After adjustment for potential cofounders, Charlson Age–Comorbidity Index >6 (Hazard ratio (HR) 1.97, 95% CI 1.32–2.92), age > 75 years (HR 1.88, 95% CI 1.20–2.97), NAFLD (HR 1.81, 95% CI 1.19–2.74, Figure 1), chronic kidney disease (HR 1.86, 95% CI 1.19–2.88) and immobility (HR 1.73, 95% CI 1.16–2.56) were associated with rCDI, as presented in Table 3. Chronic therapy with statins was associated with a decreased risk of rCDI (HR 0.24, 95% CI 0.11–0.52). Interestingly, some of the expected risk factors such as diabetes mellitus or obesity, nasogastric tube feeding and nursing home residency, as well as previous hospital admissions, choice of previous antibiotic therapy or concomitantly used antibiotics were not associated with rCDI in our model.

In addition, when NAFLD was combined with age > 75 years, chronic kidney disease and immobility, the risk of rCDI was even higher, as shown in Figure 2. Statin use was associated with lower rCDI in both patients with and without NAFLD (Figure 2, Panel d).

disease (HR 1.86, 95%CI 1.19–2.88) and immobility (HR 1.73, 95%CI 1.16–2.56) were associated with rCDI, as presented in Table 3. Chronic therapy with statins was associated with a decreased risk of rCDI (HR 0.24, 95%CI 0.11–0.52). Interestingly, some of the expected risk factors such as diabetes mellitus or obesity, nasogastric tube feeding and nursing home residency, as well as previous hospital admissions, choice of previous antibiotic therapy or concomitantly used antibiotics were not associated with rCDI in our model.

In addition, when NAFLD was combined with age > 75 years, chronic kidney disease and immobility, the risk of rCDI was even higher, as shown in Figure 2. Statin use was associated with lower rCDI in both patients with and without NAFLD (Figure 2, Panel d).

**Figure 1.** Kaplan–Meier curves and Cox proportional hazard ratios (HR) for recurrence of *Clostridioides difficile* infection in patients with and without nonalcoholic fatty liver disease.

*ioides difficile* infection in patients with and without nonalcoholic fatty liver disease. **Table 3.** Multivariable Cox regression analysis of risk factors for the development recurrent *Clostridioides difficile* infection.

**Figure 1.** Kaplan–Meier curves and Cox proportional hazard ratios (HR) for recurrence of *Clostrid-*


Chronic Kidney Disease 1.86 1.19 to 2.88 0.006 Statins 0.24 0.11 to 0.52 <0.001 The strength of association was expressed as hazard ratio (HR) and its corresponding 95% confidence interval (CI). The area under the ROC curve in the fully adjusted model was AUC 0.72 (95% CI 0.66 to 0.77).

5 The strength of association was expressed as hazard ratio (HR) and its corresponding 95% confidence interval (CI). The area under the ROC curve in the fully adjusted model was AUC 0.72 (95% CI 0.66 to 0.77). Finally, we report the demographic and clinical characteristics of patients with NAFLD. As previously shown, they had significantly higher incidence of rCDI (47.4% vs. 27.9%). They were younger than patients without NAFLD (median of 76, IQR 71–81 vs. Finally, we report the demographic and clinical characteristics of patients with NAFLD. As previously shown, they had significantly higher incidence of rCDI (47.4% vs. 27.9%). They were younger than patients without NAFLD (median of 76, IQR 71–81 vs. 78, IQR 72–83 years), more commonly had BMI > 30 (38.5% vs. 9.6%), diabetes mellitus (41.0% vs. 22.3%) and hyperlipidemia (33.3% vs. 21.1%) (data not shown). Consequently, they were more frequently prescribed statins (26.9% vs. 17.9%), metformin (11.5% vs. 7.9%) and insulin (17.9% vs. 9.2%) (data not shown). There was no difference in other comorbidities, chronic medications or antibiotics used. The 1st episode disease severity was similar to patients without NAFLD (ATLAS score of 5, IQR 4–6 vs. 6, IQR 4–7, *p* = 0.103) (data not shown).

**Figure 2.** Kaplan–Meier curves for recurrence of *Clostridioides difficile* infection (CDI). Using the Kaplan–Meier method, a proportion of patients with CDI recurrence stratified by the concomitant presence of NAFLD and (**a**) age, (**b**) chronic kidney disease, (**c**) immobility and (**d**) statins during the follow-up period. **Figure 2.** Kaplan–Meier curves for recurrence of *Clostridioides difficile* infection (CDI). Using the Kaplan–Meier method, a proportion of patients with CDI recurrence stratified by the concomitant presence of NAFLD and (**a**) age, (**b**) chronic kidney disease, (**c**) immobility and (**d**) statins during the follow-up period.

#### **3. Discussion 3. Discussion**

0.103) (data not shown*)*.

Over the last decade, there has been a marked increase in the incidence and severity of CDI, with relapsing episodes now occurring at a higher frequency, especially in the elderly. Notably, recurrent CDI is associated with a significantly higher risk of death within six months after initial CDI treatment completion, compared with CDI patients who do not develop a recurrence [18,19]. Over the last decade, there has been a marked increase in the incidence and severity of CDI, with relapsing episodes now occurring at a higher frequency, especially in the elderly. Notably, recurrent CDI is associated with a significantly higher risk of death within six months after initial CDI treatment completion, compared with CDI patients who do not develop a recurrence [18,19].

78, IQR 72–83 years), more commonly had BMI > 30 (38.5% vs. 9.6%), diabetes mellitus (41.0% vs. 22.3%) and hyperlipidemia (33.3% vs. 21.1%) (data not shown*)*. Consequently, they were more frequently prescribed statins (26.9% vs. 17.9%), metformin (11.5% vs. 7.9%) and insulin (17.9% vs. 9.2%) (data not shown*)*. There was no difference in other comorbidities, chronic medications or antibiotics used. The 1st episode disease severity was similar to patients without NAFLD (ATLAS score of 5, IQR 4–6 vs. 6, IQR 4–7, *p* =

In this retrospective cohort study, we found a significant association between recurrent CDI and NAFLD in elderly patients. Moreover, this appears to be independent of a number of potential confounders, specifically other components of metabolic syndrome. In this retrospective cohort study, we found a significant association between recurrent CDI and NAFLD in elderly patients. Moreover, this appears to be independent of a number of potential confounders, specifically other components of metabolic syndrome.

While there are several well-established risk factors for primary infection with *C. difficile*, the studies examining risk factors for CDI recurrence had variable results depending on the population studied. To the best of our knowledge, none of them have analyzed the impact of NAFLD on CDI recurrence. While there are several well-established risk factors for primary infection with *C. difficile*, the studies examining risk factors for CDI recurrence had variable results depending on the population studied. To the best of our knowledge, none of them have analyzed the impact of NAFLD on CDI recurrence.

6 Surprisingly, the investigation of the role of NAFLD in bacterial infections has only recently been initiated. Although patients with NAFLD might have a higher risk for infections due to the concomitant presence of obesity or diabetes mellitus, few studies that included NAFLD in the analysis consistently showed its outcome impact independently Surprisingly, the investigation of the role of NAFLD in bacterial infections has only recently been initiated. Although patients with NAFLD might have a higher risk for infections due to the concomitant presence of obesity or diabetes mellitus, few studies that included NAFLD in the analysis consistently showed its outcome impact independently of the metabolic syndrome components [20]. So far, this was suggested for communityacquired pneumonia, bacteriaemia of gastrointestinal origin, sepsis and urinary tract infections [21–25]. Recently, we have shown that NAFLD is a risk factor for in-hospital CDI development in elderly patients treated with systemic antimicrobial therapy [16].

The possible explanation of increased risk of CDI in NAFLD patients includes changes in intestinal microbiota linked to the development and progression of NAFLD [13]. While *Bacteroides* and *Bifidobacterium* play an important role in the mechanism, preventing colonization by *C. difficile*, patients with NAFLD were shown to have a relative decrease in

the proportion of *Bacteroides* to *Firmicutes* [13,26,27]. Notably, NAFLD is now considered a multisystem disease due to the persistent low-level inflammation with impaired immune response that might predispose patients to a variety of infections [13,28].

In our cohort, DM and obesity were not associated with rCDI. The current medical literature is contradictory on the association between obesity and *C. difficile* infection. Two case–control studies that examined the association of CDI with BMI showed different results. While Bishara et al., based on data collected from 148 adult patients with CDI, showed an association of BMI with CDI (OR = 1.196 per 1-unit increase in BMI scale) [9], Punni et al. in their study on 189 patients did not [29]. There was no association of obesity with the risk of *C. difficile* infection among patients with ulcerative colitis and according to another study, obesity was even associated with decreased risk of CDI in hospitalized patients with pouchitis [30,31].

Meanwhile, several studies have shown that DM increases the risk of CDI recurrence, but none of them included NAFLD as a variable. A large Spanish cohort study showed a significantly higher incidence of CDI in DM patients, with an increasing trend between 2011 and 2015 [32]. In addition, patients with DM have significantly higher probability for hospital readmission due to CDI (adjusted OR of 3.79 to 5.46) and the development of severe CDI [8,33]. DM was also recognized as an independent risk factor in patients with toxigenic *C. difficile* colonization to develop *C. difficile*-associated diarrhea [7].

Interestingly, it was shown that metformin increases the *Bacteroidetes*/*Firmicutes* ratio; therefore, it may yield a protective effect against CDI in patients with DM [34]. A retrospective, case–control study compared CDI diabetic patients to diabetic patients without CDI and found metformin treatment to be associated with significant reduction in CDI (OR 0.58) [35]. This could be partially explained by the study that showed that metformin reduces vegetative cell growth of *C. difficile* in vitro, as well as ex vivo in human microbiome culture system [36]. Metformin modified human gut microbiome by decreasing *C. difficile* growth while increasing the growth of non-pathogenic Clostridium strains [36].

Other risk factors for rCDI in our study were age, significant comorbidities measured with Charlson Age–Comorbidity Index (CACI), chronic kidney disease and immobility.

Several studies examined the association of CKD with CDI. While the association of severe CKD requiring dialysis with CDI severity and mortality is clear, there are inconsistent data on if patients not undergoing dialysis are at higher risk for CDI [37–40]. We provide additional data that patients with CKD not requiring chronic dialysis are at increased risk for rCDI.

Meanwhile, antibiotics and proton pump inhibitors, which are well-known risk factors for initial CDI and rCDI, were not associated with rCDI in our study. This might be due to the characteristics of our cohort and widespread use of these medications, which might have predisposed the development of CDI in the first place.

Next, we found that statin use was associated with lower rCDI in patients both with and without NAFLD. Due to the global epidemic of cardiovascular diseases, obesity and metabolic syndrome, statins are considered as one of the most commonly used medications worldwide [41]. Other than their cholesterol-lowering effect, they also have an anti-inflammatory and immunomodulatory properties [42]. According to a recent metaanalysis of available data, the risk of developing CDI was approximately 25% lower in statin users compared with non-users [43]. However, this meta-analysis included eight observational studies with significant heterogeneity. There are no randomized control trials published so far. Although the exact mechanisms of risk decrease in CDI in statin users remains unknown, some studies demonstrated that statins have an influence on gut microbiota and could change its composition [44–46]. Vieira-Silva et al. recently identified statins as a key covariate of gut microbiome diversification [45]. In their study examining obesity-associated microbiota alterations, they showed that obesity is linked with intestinal microbiota configuration characterized by a high proportion of *Bacteroides*, a low proportion of *Faecalibacterium* and low microbial cell densities (the so called Bact2 enterotype) [45]. However, patients treated with statins had significantly lower Bact2 prevalence [45]. A

systematic review of both human and animal data showed that statins modulate the gut microbiome, but the effect of change is unclear, probably due to the differences in populations studied [46]. It seems reasonable to speculate that modulation of the gut microbiome by statins might aid in the restoration of colonization resistance and lower recurrence of CDI. Alternatively, as statins have anti-inflammatory properties, their use may decrease the inflammatory response to *C. difficile* infection, which may lead to the decreased severity of CDI [47].

The major limitation of this study comes from its retrospective, monocentric design, and despite adjustment for a variety of demographics, comorbidities and medications, residual confounding may exist. The diagnosis of NAFLD was based on abdominal US, which is operator-dependent, and patients were not systematically screened for other causes of liver steatosis, but from data available in medical charts. Importantly, US has limited sensitivity and does not reliably detect steatosis when <20% or in individuals with high body mass index (BMI) (>40 kg/m<sup>2</sup> ) and is inferior to MRI or CT scan for detection and grading of steatosis [48,49]. Next, while any potential effects of NAFLD should be interpreted in connection with metabolic syndrome, due to the retrospective design, we could not include waist circumference, type of dyslipidemia, levels of triglycerides or HDL cholesterol. However, we included diabetes mellitus and obesity (defined by BMI > 30 kg/m<sup>2</sup> ), which might have the highest confounding impact, and have both been previously shown to be associated with rCDI. Since only a minority of patients had a significant risk of advanced fibrosis, as measured by APRI and FIB4 score, we were not able to determine the effect of advanced NAFLD on rCDI. Another limitation was the lack of data on *C. difficile* strain, which might be important since 027/BI/NAP1 strain has been associated with increased risk of CDI recurrence [50]. The study was designed to investigate the effect of NAFLD in elderly who are at the highest risk for rCDI and included only patients who were rehospitalized during a 3-month period. Therefore, patients diagnosed and treated for rCDI entirely as outpatients would not have been identified as having a recurrence. However, this would underestimate, rather than overestimate, the impact of NAFLD on patient outcomes.

Nevertheless, we report the first data examining the association of NAFLD with rCDI.

#### **4. Materials and Methods**

#### *4.1. Study Design and Patients*

This was a retrospective cohort study conducted at the University Hospital for Infectious Diseases Zagreb (UHID), Croatia, which is a national referral center for infectious diseases. We reviewed the hospital records of all adult patients hospitalized at UHID with a diagnosis of *Clostridium difficile* infection over a 5-year period (2016–2019). We included patients > 60 years diagnosed with the first episode of CDI. Patients who had a previous episode of CDI within three months before index hospitalization were excluded, as well as patients with known alcohol abuse and/or those diagnosed with chronic viral hepatitis or with a history of other known liver diseases. Next, only patients with performed ultrasonography examination to assess liver steatosis were included in study. During the period studied, a total of 841 patients were hospitalized with diagnosis of CDI. Of those, 730 were >60 years (total of 999 episodes of CDI). A total of 401 patients were excluded: 31 had CDI within 3 months, significant alcohol intake in 38, chronic viral hepatitis in 21, cirrhosis in 11 and hepatotoxic medications in 14 patients. There were 267 patients who did not have abdominal imaging, and 19 patients died during initial hospitalization. In the end, 329 patients were included in the study. The cohort was divided into two groups, those who were and were not readmitted with CDI within 3 months of discharge, as described in a flowchart (Figure 3).

**Figure 3.** Study design flow chart.

#### *4.2. Data Collection, Outcomes, and Definitions*

We collected multiple variables that could be associated with patient outcomes. These variables included demographic data (age, sex and residence in nursing homes) and medical history (comorbidities (measured by Charlson Age–Comorbidity Index, CACI [51]), chronic medications, hospital admission within 3 months, administration of antibiotics within 90 days). CDI was defined as the presence of 3 or more unformed stools in 24 or fewer consecutive hours, confirmed with positive two-stage *C. difficile* stool tests (screening GDH test confirmed with toxin A/B PCR), according to current guidelines [52–54]. CDI severity was determined by ATLAS score calculation [55]. CDI treatment regimen, its duration and the use of other non-CDI antimicrobials were collected. Selected blood laboratory data at the admission were analyzed: C-reactive protein level, white blood cell count, platelet count, hemoglobin, blood urea nitrogen, serum creatinine, aspartate aminotransferase, alanine aminotransferase, gamma-glutamyl transferase, alkaline phosphatase, bilirubin and serum albumin concentration. In addition, as a surrogate marker of liver injury, APRI and FIB-4 score were calculated for all patients [56]. Using the CKD-EPI equation, we calculated estimated glomerular filtration rate (eGFR) [57]. The diagnosis of NAFLD was made based on the results of abdominal ultrasound and by the absence of a secondary cause of NAFLD, according to current guidelines [11,12]. The liver steatosis was assessed by ultrasound in all patients by an experienced radiologist, and defined as finding liver parenchyma with increased echogenicity and sound attenuation [49].

The primary study outcome was rCDI, defined as CDI occurring within 14–90 days of the initial CDI diagnosis date. Same as CDI, rCDI was defined based on compatible symptoms accompanied by a positive laboratory test.

1

## *4.3. Statistical Analysis*

The clinical characteristics, laboratory and demographic data were evaluated and descriptively presented. We used Fisher's exact test and the Mann–Whitney U test to compare the groups. All tests were two-tailed; a *p* value < 0.05 was considered statistically significant. Time to CDI recurrence was evaluated using the Kaplan–Meier method, and the comparison of CDI recurrence risk between patients with and without NAFLD was made using the log-rank test. Risk factors for the development of the rCDI were investigated using a univariate, and subsequently, a multivariable Cox regression model by estimating the hazard ratio (HR) and its 95% CI for the time from cure date of the primary CDI to the first episode of rCDI. Variables with *p* < 0.2 on univariate analysis or with clinical/biological plausibility were included in initial multivariable models. Multivariable Cox proportional hazards models were developed using backward elimination with *p* < 0.1 to retain variables in the model. Statistical analyses were performed using the GraphPad Prism Software version 9.1.1. (San Diego, CA, USA) and MedCalc version 20.008 (MedCalc Software, Ostend, Belgium).

## **5. Conclusions**

In conclusion, we have shown that NAFLD is a novel host-related risk factor for recurrent CDI in elderly patients. This might be relevant for several reasons. Firstly, this highlights the need to include NAFLD as a variable in future studies, examining both CDI and rCDI. Secondly, patients with NAFLD might benefit from screening strategies, preemptive treatment or prophylactic measures, such as antibiotic prophylaxis that has recently been investigated [58,59]. The finding that statins reduce the risk of rCDI in both patients with and without NAFLD might be a novel treatment option that warrants further examination. Finally, there is growing evidence that immunological changes in patients with NAFLD might have a profound impact in the course of bacterial infections, the place where we have not been looking so far.

**Author Contributions:** Conceptualization, N.P. and A.V.; methodology, N.P.; validation, A.V.; resources, N.P.; data curation, L.Š., M.J.; writing—original draft preparation, L.Š., M.J.; writing—review and editing, N.P., A.V.; supervision, A.V.; funding acquisition, N.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Croatian Science Foundation, for the project entitled "The role of immune semaphorins in NAFLD and sepsis" (principal investigator Neven Papi´c, project number UIP-2019–04–7194).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of the University Hospital for Infectious Diseases Zagreb (protocol code 01–1247–2–2019, date of approval 30 August 2019).

**Informed Consent Statement:** Patient consent was waived due to retrospective design of the study.

**Data Availability Statement:** The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


**Suvash Chandra Ojha 1,2 , Matthew Phanchana <sup>3</sup> , Phurt Harnvoravongchai <sup>4</sup> , Surang Chankhamhaengdecha <sup>4</sup> , Sombat Singhakaew <sup>4</sup> , Puey Ounjai <sup>4</sup> and Tavan Janvilisri 5,\***


**Abstract:** In recent decades, the incidence of *Clostridioides difficile* infection (CDI) has remained high in both community and health-care settings. With the increasing rate of treatment failures and its ability to form spores, an alternative treatment for CDI has become a global priority. We used the microdilution assay to determine minimal inhibitory concentrations (MICs) of vancomycin and teicoplanin against 30 distinct *C. difficile* strains isolated from various host origins. We also examined the effect of drugs on spore germination and outgrowth by following the development of OD600. Finally, we confirmed the spore germination and cell stages by microscopy. We showed that teicoplanin exhibited lower MICs compared to vancomycin in all tested isolates. MICs of teicoplanin ranged from 0.03–0.25 µg/mL, while vancomycin ranged from 0.5–4 µg/mL. Exposure of *C. difficile* spores to broth supplemented with various concentrations of antimicrobial agents did not affect the initiation of germination, but the outgrowth to vegetative cells was inhibited by all test compounds. This finding was concordant with aberrant vegetative cells after antibiotic treatment observed by light microscopy. This work highlights the efficiency of teicoplanin for treatment of *C. difficile* through prevention of vegetative cell outgrowth.

**Keywords:** teicoplanin; *Clostridioides difficile*; spore; antibiotics

#### **1. Introduction**

*Clostridioides difficile*, previously known as *Clostridium difficile*, is a gram-positive anaerobic spore-forming bacterium. It accounts for about 20–25% of antibiotic associated diarrhea [1,2] and almost all cases of pseudomembranous colitis [3]. *C. difficile* infection (CDI) normally occurs after antibiotic administration, especially ampicillin and amoxicillin, cephalosporins, clindamycin, fluoroquinolones, and meropenem [3,4]. Studies have suggested that gut microbiota dysbiosis after antibiotic treatment allows colonization and growth of *C. difficile* [5]. CDI can cause clinical manifestations ranging from asymptomatic to severe diarrhea, pseudomembranous colitis, bowel perforation, and multi-organ dysfunction [6]. Ultimately, CDI can be fatal, mostly in older patients [7]. The total CDI incidence has decreased in the US according to the Center for Disease Control and Prevention (CDC) [8]. Although the number of cases is not on the rise, *C. difficile* is classified as a pathogen posing an urgent threat due to antibiotic resistance [9].

Treatment for CDI is now limited to a few antibiotics including fidaxomicin and vancomycin according to the new guideline by the Infectious Diseases Society of America

**Citation:** Ojha, S.C.; Phanchana, M.; Harnvoravongchai, P.; Chankhamhaengdecha, S.; Singhakaew, S.; Ounjai, P.; Janvilisri, T. Teicoplanin Suppresses Vegetative *Clostridioides difficile* and Spore Outgrowth. *Antibiotics* **2021**, *10*, 984. https://doi.org/10.3390/ antibiotics10080984

Academic Editor: Guido Granata

Received: 13 July 2021 Accepted: 10 August 2021 Published: 15 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

(IDSA) and the Society for Healthcare Epidemiology of America (SHEA) [10]. Metronidazole, which was suggested as a first line, is now recommended only when fidaxomicin or vancomycin is not available or is limited, owing to its inferiority to vancomycin and fidaxomicin, higher recurrence rate, and neurotoxicity in prolonged or repeated use [10,11]. Additionally, treatment failures have been reported for most regimens, mostly due to the recurrence of *C. difficile*, hence new antibiotics for CDI are of utmost importance [12].

Developing new drugs is a costly and time-consuming process. Therefore, drug repurposing or repositioning has come under the limelight in pharmaceutical research in recent years as it can cut down the development process to minimal [13,14]. Teicoplanin, a mixture of glycopeptide antibiotics, belongs to the same class as vancomycin, distinguished by glucosamine as the basic sugar and the presence of aliphatic acid residues. It binds to the terminal D-Ala-D-Ala sequence of peptides forming the bacterial cell wall and, by sterically hindering the transglycosylation reaction, inhibits the formation of peptidoglycan by an intracellular accumulation of UDP-*n*-acetylmuramyl-pentapeptide [15]. Teicoplanin exhibits great activity against multiple gram-positive bacteria, which fail for other regimens [16], including *C. difficile* [17]. Even though the activity of teicoplanin against vegetative cell *C. difficile* is well documented [17–19], no experimental evidence has been presented so far for spore and outgrowth of *C. difficile*. We investigated the effect of teicoplanin on spore germination and outgrowth in *C. difficile* isolates from different host origins and compared it to its vancomycin counterpart. In addition, we also examined the cell stage alterations marked by staining the affinity of germinating cells following antibiotic treatment. The data presented here provide experimental evidence for the inhibitory effect of teicoplanin on germinating *C. difficile* cells.

#### **2. Results**

#### *2.1. MICs of Teicoplanin on C. difficile*

Minimum inhibitory concentrations (MICs) of teicoplanin and vancomycin against 30 *C. difficile* isolates obtained from various sources were determined by the broth microdilution method as described earlier [20]. All strains tested were sensitive to both antibiotics with MIC ranges for teicoplanin and vancomycin of 0.03–0.25 µg/mL and 0.5–4.0 µg/mL, respectively. Teicoplanin showed lower MICs among all strains tested (Table 1). We also compared the effect of teicoplanin at sub-MICs to vancomycin on selected strains. At sub-MICs, both antibiotics were at least 1-log less potent than at their respective MICs. However, teicoplanin at sub-MIC concentrations reduced the number of colonies more than that of vancomycin (Figure 1). The MICs of antimicrobial agents for each respective vegetative strain were used as a platform to evaluate the effect of these drugs on *C. difficile* spore germination and outgrowth. We also evaluated minimum bactericidal concentrations (MBC) of both teicoplanin, and vancomycin and the results revealed that the MBC of both antibiotics were at the concentration of 2× MIC.


**Table 1.** Minimal inhibitory concentrations for teicoplanin and vancomycin for 30 *C. difficile* strains.

**Figure 1.** Antibacterial activity at sub-MICs of antibiotics using spot assay. A representative strain of *C. difficile* R20291 was attained to 0.1 OD600 in BHIS (brain heart infusion supplemented with 0.1% sodium taurocholate) medium supplemented with 0.25×, 0.5×, 1× MICs of teicoplanin or vancomycin. After 24-h incubation, cultures were 10-fold serially diluted and spotted onto BHIS agar plates, and plates were photographed following 24 h anaerobic incubation at 37 °C. **Table 1.** Minimal inhibitory concentrations for teicoplanin and vancomycin for 30 *C. difficile* strains. **Figure 1.** Antibacterial activity at sub-MICs of antibiotics using spot assay. A representative strain of *C. difficile* R20291 was attained to 0.1 OD<sup>600</sup> in BHIS (brain heart infusion supplemented with 0.1% sodium taurocholate) medium supplemented with 0.25×, 0.5×, 1× MICs of teicoplanin or vancomycin. After 24-h incubation, cultures were 10-fold serially diluted and spotted onto BHIS agar plates, and plates were photographed following 24 h anaerobic incubation at 37 ◦C.

F101 Food 0.06 2 F102 Food 0.06 2 F103 Food 0.125 2 F104 Food 0.125 2 A121 Animal 0.25 4 A122 Animal 0.125 2 A123 Animal 0.25 4 A124 Animal 0.25 4 A125 Animal 0.25 4 A126 Animal 0.125 2 R20291 Human 0.03 0.5 H201 Human 0.06 1 H203 Human 0.125 2 H204 Human 0.06 1 H205 Human 0.25 4 H206 Human 0.25 4 H207 Human 0.25 4 H208 Human 0.25 4 H209 Human 0.125 2 H210 Human 0.125 2 H211 Human 0.25 4 H212 Human 0.125 2 H213 Human 0.25 2 H214 Human 0.25 4 H215 Human 0.25 4 H216 Human 0.125 2 H217 Human 0.125 2 H218 Human 0.25 4 32g57 Human 0.125 2

**Teicoplanin Vancomycin**

#### *2.2. Teicoplanin Does Not Inhibit C. difficile Spore Germination 2.2. Teicoplanin Does Not Inhibit C. difficile Spore Germination*

To determine the role of teicoplanin and vancomycin in its ability to block *C. difficile* spore germination, purified spores of 6 *C. difficile* strains were incubated for 1 h in the presence of BHIS supplemented with 0.5×, 1×, or 32× MIC of teicoplanin, and germination kinetics were monitored by observing the changes in OD<sup>600</sup> of *C. difficile* cultures over a 1 h period. A drug-free control and 1× MIC of vancomycin were included as a comparator. As expected, *C. difficile* isolates germinated poorly in BHI broth without supplementation of 0.1% taurocholate. Nevertheless, teicoplanin and vancomycin did not inhibit spore germination at all concentrations tested (*p* > 0.05) (Figure 2), resulting in a significant drop of OD600. The initiation of spore germination under treatment conditions was comparable to the drug-free control over time based on ANOVA at each time point. To determine the role of teicoplanin and vancomycin in its ability to block *C. difficile* spore germination, purified spores of 6 *C. difficile* strains were incubated for 1 h in the presence of BHIS supplemented with 0.5×, 1×, or 32× MIC of teicoplanin, and germination kinetics were monitored by observing the changes in OD600 of *C. difficile* cultures over a 1 h period. A drug-free control and 1× MIC of vancomycin were included as a comparator. As expected, *C. difficile* isolates germinated poorly in BHI broth without supplementation of 0.1% taurocholate. Nevertheless, teicoplanin and vancomycin did not inhibit spore germination at all concentrations tested (*p* > 0.05) (Figure 2), resulting in a significant drop of OD600. The initiation of spore germination under treatment conditions was comparable to the drug-free control over time based on ANOVA at each time point.

**Figure 2.** Teicoplanin and vancomycin do not inhibit spore germination. *C. difficile* spores were exposed to different concentration of antibiotics in relation to their respective MICs, and germination was followed by measuring loss of OD600 for 1 h at 1 min time interval. The growth control contained no antibiotic, and as a comparator, vancomycin was included. Data points represent the mean of the relative OD600 at the indicated time points normalized to *t* = 0 (control). All the experiments were performed in triplicates and error bars represent the standard errors. The results include the strains: (**A**) F101; (**B**) F102; (**C**) A125; (**D**) A126; (**E**) R20291; (**F**) H203. Circle (●), diamond (◆), triangle (▲), inverted tringle (▼), cross (×), and plus (+) denote the exposure to BHIS, 0.5× MIC teicoplanin, 1× MIC teicoplanin, 1× MIC vancomycin, 32× MIC teicoplanin, and BHI, respectively. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHI = Brain heart infusion broth without 0.1% sodium taurocholate, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate. **Figure 2.** Teicoplanin and vancomycin do not inhibit spore germination. *C. difficile* spores were exposed to different concentration of antibiotics in relation to their respective MICs, and germination was followed by measuring loss of OD<sup>600</sup> for 1 h at 1 min time interval. The growth control contained no antibiotic, and as a comparator, vancomycin was included. Data points represent the mean of the relative OD<sup>600</sup> at the indicated time points normalized to *t* = 0 (control). All the experiments were performed in triplicates and error bars represent the standard errors. The results include the strains: (**A**) F101; (**B**) F102; (**C**) A125; (**D**) A126; (**E**) R20291; (**F**) H203. Circle (•), diamond (), triangle (N), inverted tringle (H), cross (×), and plus (+) denote the exposure to BHIS, 0.5× MIC teicoplanin, 1× MIC teicoplanin, 1× MIC vancomycin, 32× MIC teicoplanin, and BHI, respectively. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHI = Brain heart infusion broth without 0.1% sodium taurocholate, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate.

#### *2.3. Teicoplanin Inhibits C. difficile Spore Outgrowth* Since both teicoplanin and vancomycin did not have either a positive or negative *2.3. Teicoplanin Inhibits C. difficile Spore Outgrowth*

influence on the initiation of spore germination, we next evaluated the outgrowth to vegetative cells by monitoring the change in OD600, to examine whether the later stage of germination would be affected. As expected, under the influence of 0.5× MIC of teicoplanin, growth differences were observed in contrast to the untreated control, although not significantly different (Figure 3). Sub-inhibitory concentrations of teicoplanin appeared to delay the onset of spore outgrowth and increased the time until the stationary phase was Since both teicoplanin and vancomycin did not have either a positive or negative influence on the initiation of spore germination, we next evaluated the outgrowth to vegetative cells by monitoring the change in OD600, to examine whether the later stage of germination would be affected. As expected, under the influence of 0.5× MIC of teicoplanin, growth differences were observed in contrast to the untreated control, although not significantly different (Figure 3). Sub-inhibitory concentrations of teicoplanin appeared

to delay the onset of spore outgrowth and increased the time until the stationary phase was reached. However, they did not affect the outcome or the later stage of spore outgrowth. The outgrowth of spores in all tested strains exposed to a minimum of MIC of drugs was significantly inhibited compared to the drug-free control (*p* < 0.0001). reached. However, they did not affect the outcome or the later stage of spore outgrowth. The outgrowth of spores in all tested strains exposed to a minimum of MIC of drugs was significantly inhibited compared to the drug-free control (*p* < 0.0001).

*Antibiotics* **2021**, *10*, 984 5 of 12

**Figure 3.** Teicoplanin and vancomycin at MIC inhibit spore outgrowth, while sub-MIC teicoplanin substantially delays outgrowth. Data points represent the mean of the relative OD600 at the indicated time points normalized to *t* = 0 (control). All the experiments were performed in triplicates and error bar represents the standard errors. The results include the strains: (**A**) F101; (**B**) F102; (**C**) A125; (**D**) A126; (**E**) R20291; (**F**) H203. Circle (●), diamond (◆), triangle (▲), inverted tringle (▼), and cross (×) denote the exposure to BHIS, 0.5× MIC teicoplanin, 1× MIC teicoplanin, 1× MIC vancomycin, and 32× MIC teicoplanin, respectively. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate. The dashed and dotted lines represent the data trend between the Y-axis break. **Figure 3.** Teicoplanin and vancomycin at MIC inhibit spore outgrowth, while sub-MIC teicoplanin substantially delays outgrowth. Data points represent the mean of the relative OD<sup>600</sup> at the indicated time points normalized to *t* = 0 (control). All the experiments were performed in triplicates and error bar represents the standard errors. The results include the strains: (**A**) F101; (**B**) F102; (**C**) A125; (**D**) A126; (**E**) R20291; (**F**) H203. Circle (•), diamond (), triangle (N), inverted tringle (H), and cross (×) denote the exposure to BHIS, 0.5× MIC teicoplanin, 1× MIC teicoplanin, 1× MIC vancomycin, and 32× MIC teicoplanin, respectively. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate. The dashed and dotted lines represent the data trend between the Y-axis break.

To further investigate the alteration under the influence of antibiotics on spore germination and outgrowth, we included strains that were the least and the most susceptible to teicoplanin, R20291 and A125, respectively. We next performed Wirtz–Conklin staining of untreated spores or spores treated with 1× MIC of teicoplanin or vancomycin following 3 h incubation in the growth medium. The number of spores estimated by visual inspection using light microscopic analysis revealed that >98% of spores germinated in both antibiotic treatment groups, which was not significantly different from the untreated control (Figures 4 and 6). The control spores appeared as greenish-blue spheres, and the germinated cells appeared as pink spheres without shape alteration and no bacilli were detected. However, the inhibition of outgrowth by these antibiotics was more evident as determined by microscopic analysis (Figure 5), confirming that antibiotic exposure did inhibit outgrowth to vegetative cells. After 24 h incubation in a growth medium supplemented with 1× MIC of antibiotics, spores treated with both antibiotics appeared to lose their ability to change to bacillus vegetative cells. Spores were stained pink/purple or faintly stained with spherical or blunted rounded ends and showed structural degeneration and clear extrusion from the spore germinated structure (Figure 5). This suggested that the outgrowth was inhibited at the initial stage of germination. Both antibiotics at 1× MIC inhibited up to 80% of spore outgrowth to vegetative cells (Figures 5 and 6). Statisti-To further investigate the alteration under the influence of antibiotics on spore germination and outgrowth, we included strains that were the least and the most susceptible to teicoplanin, R20291 and A125, respectively. We next performed Wirtz–Conklin staining of untreated spores or spores treated with 1× MIC of teicoplanin or vancomycin following 3 h incubation in the growth medium. The number of spores estimated by visual inspection using light microscopic analysis revealed that >98% of spores germinated in both antibiotic treatment groups, which was not significantly different from the untreated control (Figure 4 and Figure 6). The control spores appeared as greenish-blue spheres, and the germinated cells appeared as pink spheres without shape alteration and no bacilli were detected. However, the inhibition of outgrowth by these antibiotics was more evident as determined by microscopic analysis (Figure 5), confirming that antibiotic exposure did inhibit outgrowth to vegetative cells. After 24 h incubation in a growth medium supplemented with 1× MIC of antibiotics, spores treated with both antibiotics appeared to lose their ability to change to bacillus vegetative cells. Spores were stained pink/purple or faintly stained with spherical or blunted rounded ends and showed structural degeneration and clear extrusion from the spore germinated structure (Figure 5). This suggested that the outgrowth was inhibited at the initial stage of germination. Both antibiotics at 1× MIC inhibited up to 80% of spore outgrowth to vegetative cells (Figures 5 and 6). Statistical analysis suggested that teicoplanin and vancomycin at their respective MICs inhibited

cal analysis suggested that teicoplanin and vancomycin at their respective MICs inhibited

spore outgrowth significantly (*p* < 0.001) when compared to untreated control. However, the effect of teicoplanin and vancomycin on spore outgrowth was not significantly different (Figure 6). spore outgrowth significantly (*p* < 0.001) when compared to untreated control. However, the effect of teicoplanin and vancomycin on spore outgrowth was not significantly different (Figure 6). spore outgrowth significantly (*p* < 0.001) when compared to untreated control. However, the effect of teicoplanin and vancomycin on spore outgrowth was not significantly different (Figure 6).

**Figure 4.** Teicoplanin and vancomycin at their respective MICs do not inhibit initiation of spore germination as revealed by differential staining. *C. difficile* spores were incubated with 1× MIC of either teicoplanin or vancomycin supplemented with BHIS medium for 3 h. Germinated cells were stained by Wirtz–Conklin staining (5% malachite green/0.5% safranin). Spores were greenish-blue spheres and germinated cells appeared to be pink/purple spheres. All the micrographs were taken at a magnification of 1000×. All the experiments were repeated in triplicates to ensure the reproducibility of the results. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate. **Figure 4.** Teicoplanin and vancomycin at their respective MICs do not inhibit initiation of spore germination as revealed by differential staining. *C. difficile* spores were incubated with 1× MIC of either teicoplanin or vancomycin supplemented with BHIS medium for 3 h. Germinated cells were stained by Wirtz–Conklin staining (5% malachite green/0.5% safranin). Spores were greenish-blue spheres and germinated cells appeared to be pink/purple spheres. All the micrographs were taken at a magnification of 1000×. All the experiments were repeated in triplicates to ensure the reproducibility of the results. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate. **Figure 4.** Teicoplanin and vancomycin at their respective MICs do not inhibit initiation of spore germination as revealed by differential staining. *C. difficile* spores were incubated with 1× MIC of either teicoplanin or vancomycin supplemented with BHIS medium for 3 h. Germinated cells were stained by Wirtz–Conklin staining (5% malachite green/0.5% safranin). Spores were greenish-blue spheres and germinated cells appeared to be pink/purple spheres. All the micrographs were taken at a magnification of 1000×. All the experiments were repeated in triplicates to ensure the reproducibility of the results. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate.

**Figure 5.** Antibiotic treatments limit the outgrowth of germinated *C. difficile* spores and prevent outgrowth to vegetative cells. *C. difficile* spores were incubated with BHIS medium supplemented with either 1× MIC of teicoplanin or vancomycin for 24 h. Germination inhibition and outgrowth were marked by Wirtz–Conklin staining. Vegetative cells were long filamentous and stained pink/purple, germinated cells were pink/purple spheres, outgrowths were pink/purple/faintly stained, blunted ends or extrusion from germinated cell. All the micrographs were taken at a magnification of 1000×, 5 fields counted, and repeated in triplicates to ensure the reproducibility of the results. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate (as negative control). **Figure 5.** Antibiotic treatments limit the outgrowth of germinated *C. difficile* spores and prevent outgrowth to vegetative cells. *C. difficile* spores were incubated with BHIS medium supplemented with either 1× MIC of teicoplanin or vancomycin for 24 h. Germination inhibition and outgrowth were marked by Wirtz–Conklin staining. Vegetative cells were long filamentous and stained pink/purple, germinated cells were pink/purple spheres, outgrowths were pink/purple/faintly stained, blunted ends or extrusion from germinated cell. All the micrographs were taken at a magnification of 1000×, 5 fields counted, and repeated in triplicates to ensure the reproducibility of the results. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate (as negative control). **Figure 5.** Antibiotic treatments limit the outgrowth of germinated *C. difficile* spores and prevent outgrowth to vegetative cells. *C. difficile* spores were incubated with BHIS medium supplemented with either 1× MIC of teicoplanin or vancomycin for 24 h. Germination inhibition and outgrowth were marked by Wirtz–Conklin staining. Vegetative cells were long filamentous and stained pink/purple, germinated cells were pink/purple spheres, outgrowths were pink/purple/faintly stained, blunted ends or extrusion from germinated cell. All the micrographs were taken at a magnification of 1000×, 5 fields counted, and repeated in triplicates to ensure the reproducibility of the results. Abbreviations: Tec = teicoplanin, Van = vancomycin, BHIS = Brain heart infusion broth supplemented with 0.1% sodium taurocholate (as negative control).

**Figure 6.** Teicoplanin and vancomycin significantly reduce outgrowth in *C. difficile* (**A**) A125 and (**B**) R20291 as determined by differential staining assay. *C. difficile* spores were incubated with BHIS medium alone or medium supplemented with 1× MIC of either teicoplanin or vancomycin for 24 h, then thoroughly washed and stained with a Wirtz–Conklin stain. All the micrographs were taken at a magnification of 1000×. Relative percentage of germinated spore and vegetative cells were calculated based on the count of five different fields by CellSens Dimension software. Bar graphs represent geometric means of three independent experiments and error bars represent the standard errors. Abbreviations: Tec = teicoplanin, Van = vancomycin and \*\*\* denotes *p*-value < 0.0001. **Figure 6.** Teicoplanin and vancomycin significantly reduce outgrowth in *C. difficile* (**A**) A125 and (**B**) R20291 as determined by differential staining assay. *C. difficile* spores were incubated with BHIS medium alone or medium supplemented with 1× MIC of either teicoplanin or vancomycin for 24 h, then thoroughly washed and stained with a Wirtz–Conklin stain. All the micrographs were taken at a magnification of 1000×. Relative percentage of germinated spore and vegetative cells were calculated based on the count of five different fields by CellSens Dimension software. Bar graphs represent geometric means of three independent experiments and error bars represent the standard errors. Abbreviations: Tec = teicoplanin, Van = vancomycin and \*\*\* denotes *p*-value < 0.0001.

#### **3. Discussion 3. Discussion**

CDI continues to be a major nosocomial pathogen and a particular source of morbidity and mortality among elderly and immune suppressed individuals [21]. Treatment failures have been more evident recently, which has raised a serious concern for clinicians across multiple specialties. Hence, there is a medical need to explore potential therapeutic drugs with improved properties. Previous studies have encouraged the use of teicoplanin over commonly used antibiotics in CDI treatment due to its longer half-life, lower relapse rate, relatively uncommon nephrotoxicity or cytotoxicity, and lack of requirement for routine serum monitoring [18,19,22,23] A study by Wenisch et al. (1996) also claimed to have 100% cure rate with the use of teicoplanin in patients endoscopically confirmed with pseudomembrane colitis [19]; however, direct comparisons with vancomycin are difficult be-CDI continues to be a major nosocomial pathogen and a particular source of morbidity and mortality among elderly and immune suppressed individuals [21]. Treatment failures have been more evident recently, which has raised a serious concern for clinicians across multiple specialties. Hence, there is a medical need to explore potential therapeutic drugs with improved properties. Previous studies have encouraged the use of teicoplanin over commonly used antibiotics in CDI treatment due to its longer half-life, lower relapse rate, relatively uncommon nephrotoxicity or cytotoxicity, and lack of requirement for routine serum monitoring [18,19,22,23] A study by Wenisch et al. (1996) also claimed to have 100% cure rate with the use of teicoplanin in patients endoscopically confirmed with pseudomembrane colitis [19]; however, direct comparisons with vancomycin are difficult because of inherent differences between studies.

cause of inherent differences between studies. As teicoplanin is fast acting at low concentrations and has poor absorption in the gut [24], it is incontestably an effective antimicrobial agent for the control of pathogens in the gut, including *C. difficile,* without permitting spore formation [25,26]. The activity of teicoplanin was at least 8–16 fold more potent than that of vancomycin, which was consistent with the studies done by Kunishima [27]. A set of 6 distinct target isolates obtained from humans, animals and food were used to account for the variation in the *C. difficile* spore germination and outgrowth under the influence of these antimicrobial agents. Examining the 0.5×, 1× or 32× of teicoplanin or 1× MIC of vancomycin on *C. difficile* spore germination revealed that none of the antimicrobial agents affected the initiation of germination compared to the drug-free control. At MIC or above, we observed that the As teicoplanin is fast acting at low concentrations and has poor absorption in the gut [24], it is incontestably an effective antimicrobial agent for the control of pathogens in the gut, including *C. difficile,* without permitting spore formation [25,26]. The activity of teicoplanin was at least 8–16 fold more potent than that of vancomycin, which was consistent with the studies done by Kunishima [27]. A set of 6 distinct target isolates obtained from humans, animals and food were used to account for the variation in the *C. difficile* spore germination and outgrowth under the influence of these antimicrobial agents. Examining the 0.5×, 1× or 32× of teicoplanin or 1× MIC of vancomycin on *C. difficile* spore germination revealed that none of the antimicrobial agents affected the initiation of germination compared to the drug-free control. At MIC or above, we observed that the outgrowth was inhibited by these antimicrobial agents. This is predictable as

spores lose their dormancy upon germination, resume metabolism at the core region and subsequently an outgrowth begins by synthesizing new cell wall peptidoglycan [28,29]. The germinated spores are vulnerable to these glycopeptide antibiotics, which inhibit spore outgrowth. Although teicoplanin is functionally similar to vancomycin, its potent activity at relatively low concentration on vegetative cells and spore outgrowth may contribute to the lower recurrence rate in clinical trials [19,30–33]. Furthermore, transition from spores to vegetative cells is important and involves various metabolic changes. There are reports showing that some antibiotics can inhibit vegetative cells but not outgrowth [34]. Certain antibiotics inhibit both spore outgrowth and vegetative cells but at different concentrations, implying that there are underlying differences between these stages [35].

At sub-inhibitory concentration of teicoplanin, late growth had begun for most strains, which took an extended duration to reach their stationary phase; this could be due to the stress generated by the antibiotic at an early stage of spore germination.

To investigate further spore germination and outgrowth by light microscopy, we included 2 strains that were the least and the most susceptible to teicoplanin. Wirtz–Conklin staining displayed a clear distinction between different stages of spore germination following antibiotic treatment or untreated spores, which agreed with the studies done by Hamouda [36]. Untreated spores were dormant in structure and acquired greenish blue color following Wirtz–Conklin staining instead of pink/purple color spheres or rods [21,36,37]. Incubating spores with growth medium for 3 h displayed initiation of germination without complete transition to bacilli, where they appeared as pink spheres. This change in their staining affinity is associated with the initiation of germination in the presence of growth medium without complete outgrowth to the filamentous vegetative bacilli. The variability in spore germination appeared due to the asynchrony in spore population germination following exposure to spore germinants, which followed studies done by Moir [38]. Following complete germination, spores were transformed into filamentous vegetative cells that were stained purple rods. However, spores following treatment with antibiotics did not develop into bacilli for up to 24 h. The inhibitory action of those antibiotics affected their outgrowth to vegetative cells, which appeared as spheres or short rods with blunted ends, supporting the notion of being sporostatic agents.

#### **4. Materials & Methods**

#### *4.1. C. difficile Strains and Growth Conditions*

A total of 30 *C. difficile* isolates obtained from various sources, including food, animal, and human were used in this study (Table 1) [39]. *C. difficile* strain R20291 was kindly provided by Prof Nigel Minton, University of Nottingham. As described previously [40,41], all *C. difficile* strains were grown at 37 ◦C in an anaerobic workstation (85% N2, 10% H2, and 5% CO2; Don Whitley Scientific, UK) in the brain heart infusion supplemented with 0.1% (*w*/*v*) sodium taurocholate, 0.1% L-cysteine, and 5 mg/mL yeast extract broth or agar (BHIS).

#### *4.2. Spore Preparation*

A total of 6 strains of *C. difficile* obtained from various sources were included for spore purification. Briefly, a single colony was inoculated into BHIS broth and incubated overnight at 37 ◦C. A 100-µL aliquot of overnight culture was spread onto BHIS agar, supplemented with 250 µg/mL cycloserine and 8 µg/mL cefoxitin, and incubated anaerobically at 37 ◦C for 10 days to allow efficient sporulation. Sporulation-induced lawns were harvested in 1 mL sterile distilled water (dH2O) using cell scrapers. The suspension was then centrifuged at 5000× *g* for 15 min and washed 5 times with sterile dH2O. To inactivate viable vegetative cells, spore suspensions were then treated with 0.3 mg/mL proteinase K at 37 ◦C for 2 h with gentle shaking, followed by incubation at 65 ◦C for 1 h. Subsequently, spore suspensions were washed 5 additional times to remove any residuals from proteinase K. Purified spores were examined by phase-contrast microscopy to ensure that they were free of vegetative cells and debris, and subsequently stored at 4 ◦C.

#### *4.3. Antimicrobial Assay*

To determine minimal inhibitory concentrations (MICs), a single colony from overnight culture was resuspended in 5 mL BHIS broth and incubated anaerobically for 12 h. Next day, 100 µL of aliquots were transferred to a new BHIS broth and incubated for another 6 h to minimize spore carryover and dilute the pre-formed toxin effect. The diluted vegetative cell suspension (100 µL) was aliquoted to the wells of flat-bottomed 96-well plates containing equal volumes of BHIS medium supplemented with defined concentrations of antibiotics, with the initial inoculum concentration maintained at OD<sup>600</sup> of 0.6. An antibiotic-free control was included in each experiment. After 24 h incubation, the plates were measured for OD<sup>600</sup> as an indicator of bacterial growth using a microplate reader (Tecan, Switzerland). MIC is defined by the concentration that has no visible growth. These assays were repeated at least 3 times to ensure reproducibility of the results. To compare the effect of sub-inhibitory concentrations of teicoplanin and vancomycin, *C. difficile* was exposed to 0.25×, 0.5×, and 1× MIC values, and was serially diluted before stamping on to BHIS plate, then incubated anaerobically for 24 h.

Minimal bactericidal concentration (MBC) was performed as previously described [42]. Briefly, the assay plate containing various concentrations of antibiotics was inoculated with bacterial suspension as per MIC and incubated for 24 h, then the bacterial suspensions around MIC value were transferred to the BHIS plate by the stamping technique and incubated for 24 h. MBC is defined by the concentration where no bacterial colony was observed on the BHIS plate.

## *4.4. Spore Germination and Outgrowth*

Spore suspensions were heat activated at 65 ◦C for 30 min, vortexed to obtain a homogenous suspension and checked for clumping by microscopy. The time-kill kinetics of teicoplanin against 6 *C. difficile* strains were performed at the 0.5×, 1×, 2×, and 32× MIC of antibiotics supplemented in BHIS medium, with the final inoculum concentration maintained at OD<sup>600</sup> of 0.6. Spore germination was followed anaerobically at 37 ◦C by measuring the loss of OD<sup>600</sup> at 1 min time intervals for 1 h using a microplate reader. Reduction of OD<sup>600</sup> reflects spore germination as it changes with the refractility of the spore coat [43]. Following germination, the differences in spore outgrowth efficiency were recorded by measuring OD<sup>600</sup> at 20 min time intervals for 22 h using the same protocol. The ratio of the OD<sup>600</sup> at time t and the control (t = 0) was then plotted against time. A drug-free control, and as a comparator, the spore suspensions treated with 1× MIC of vancomycin were included for every strain tested.

#### *4.5. Wirtz–Conklin Staining*

The staining protocol for differentiation of spores from germinating and vegetative cells was adapted from Hamouda [36]. The spore suspension was anaerobically incubated at 37 ◦C for 3 h or for 24 h with BHIS medium alone or medium supplemented with MIC concentrations of antibiotics. Following incubation, the spore suspension was washed thoroughly 5 times with sterile distilled water and resuspended with 15 µL dH2O (OD600~0.1). Five microliters of spore suspension were transferred to a clean glass slide and smeared uniformly. Henceforth, slides were air dried, and heat fixed and stained with Wirtz–Conklin stain. The specimen was visualized under oil immersion objective of a light microscope (Olympus BX53, Tokyo, Japan). Five fields of each slide were imaged from 3 independent biological replicates, counted by using software CellSens Dimension software version 1.11 (Olympus Software, Imaging System, Hamburg, Germany). Based on their color differences, relative percentages of germinated spores and vegetative cells were calculated. The percentage of spores in each image was also calculated as [Number of spores/Total number of cells (spores + vegetative cells)] × 100.

## *4.6. Statistical Analysis*

All data presented were of at least 3 independent experiments. Statistical analyses were performed by the nonparametric one-way analysis of variance (ANOVA), using GraphPad Prism software (GraphPad Software Inc., La Jolla, CA, USA) to compare each condition with the corresponding controls. *p*-values less than 0.05 indicated statistically significant difference.

## **5. Conclusions**

In conclusion, our results indicated that teicoplanin could be a potential therapeutic drug for *C. difficile* due to its potent activity at low concentrations, as well as having predetermined broad-spectrum activity against gram-positive anaerobes. Teicoplanin did not interrupt spore germination, but instead inhibited the outgrowth to vegetative cells from the germinated spore. Our data bridge the experimental gap on the effect of teicoplanin on spores as its effects on spore germination and outgrowth in *C. difficile* have not yet been reported. As most *C. difficile*-associated diseases are multifactorial, further in-depth studies, including many hypervirulent strains, animal models and human trials must be warranted to elucidate the therapeutic role of teicoplanin.

**Author Contributions:** Conceptualization, T.J.; formal analysis, S.C.O. and M.P.; methodology, S.C.O., M.P., P.H., S.C., S.S., and P.O.; investigation, S.C.O.; validation, T.J.; writing-original draft preparation, S.C.O.; writing-review and editing, M.P. and T.J.; project administration, T.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research project is supported by Research Cluster: Multi-generation Researchers Grant from Mahidol University to S.C., T.J. and P.H. M.P. received the ICTM grant from the Faculty of Tropical Medicine, Mahidol University.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data are contained within the article.

**Conflicts of Interest:** The authors declare no conflict of interest.

## **References**

