Exploring novel microbial metabolites and drugs for inhibiting Clostridioides difficile

ABSTRACT Clostridioides difficile is an enteric pathogen that can cause a range of illnesses from mild diarrhea to pseudomembranous colitis and even death. This pathogen often takes advantage of microbial dysbiosis provoked by antibiotic use. With the increasing incidence and severity of infections, coupled with high recurrence rates, there is an urgent need to identify innovative therapies that can preserve the healthy state of the gut microbiota. In this study, we screened a microbial metabolite library against C. difficile. From a collection of 527 metabolites, we identified 18 compounds with no previously identified antimicrobial activity and metabolites that exhibited potent activity against C. difficile growth. Of these 18 hits, five drugs and three metabolites displayed the most potent anti-C. difficile activity and were subsequently assessed against 20 clinical isolates of C. difficile. These potent agents included ecteinascidin 770 (minimum inhibitory concentration against 50% of isolates [MIC50] ≤0.06 µg/mL); 8-hydroxyquinoline derivatives, such as broxyquinoline and choloroquinaldol (MIC50 = 0.125 µg/mL); ionomycin calcium salt, carbadox, and robenidine hydrochloride (MIC50 = 1 µg/mL); and dronedarone and milbemycin oxime (MIC50 = 4 µg/mL). Unlike vancomycin and fidaxomicin, which are the standard-of-care anti-C. difficile antibiotics, most of these metabolites showed robust bactericidal activity within 2–8 h with minimal impact on the growth of representative members of the normal gut microbiota. These results suggest that the drugs and microbial metabolite scaffolds may offer alternative avenues to address unmet needs in C. difficile disease prevention and treatment. IMPORTANCE The most frequent infection associated with hospital settings is Clostridioides difficile, which can cause fatal diarrhea and severe colitis, toxic megacolon, sepsis, and leaky gut. Those who have taken antibiotics for other illnesses that affect the gut's healthy microbiota are more susceptible to C. difficile infection (CDI). Recently, some reports showed higher recurrence rates and resistance to anti-C. difficile, which may compromise the efficacy of CDI treatment. Our study is significant because it is anticipated to discover novel microbial metabolites and drugs with microbial origins that are safe for the intestinal flora, effective against C. difficile, and reduce the risk of recurrence associated with CDI.

Due to its role in C. difficile's colonic proliferation, intake of antibiotics is a significant risk factor for C. difficile infection (CDI) (7).Because it provides resistance to C. difficile colonization, the healthy gut microbiota serves as a protective barrier against CDI (8).When broad-spectrum antibiotics are used and the usual makeup of the gut microbiota is disturbed, patients are more susceptible to intestinal colonization by C. difficile (9,10).This leads to an increase of deadly toxins being produced by C. difficile during spore germination and vegetative cell outgrowth (11).The severe gastrointestinal inflamma tion that C. difficile toxins cause provides the ideal conditions for the bacteria to continue to survive and the infection to persist (12,13).
Moreover, the main factor responsible for CDI recurrence and spread, spores, has a strong correlation with the clinical severity of CDI (14).The germination and vegetative development of C. difficile spores can be inhibited by a robust and varied microbiome.Spore germination and toxin production will transpire if the metabolic and microbiolog ical environment of the gut has been disturbed (15,16).Individuals who are initially diagnosed with CDI have a 20% probability of experiencing a recurrence, and between 19% and 28% of them do not respond well to therapy after taking antibiotics (17,18).
The FDA-approved anti-CDI armamentarium includes three antibiotics, metronida zole, vancomycin, and fidaxomicin.Metronidazole should only be used for scenarios that are not severe when vancomycin and fidaxomicin are not readily accessible (19).Treatment failure rates are higher with vancomycin and metronidazole, frequent CDI recurrence, disruption of the normal intestinal microbiome, and promotion of vancomy cin-resistant enterococci (VRE) overgrowth (20)(21)(22)(23).Fidaxomicin is the only antibiotic approved against C. difficile in the last three decades (24).Despite its selectivity toward C. difficile over the gut microbiota, fidaxomicin is associated with multiple cases of treatment failure (25)(26)(27).In addition, fecal microbiota transplant (FMT) has emerged as a non-antibiotic therapy for CDI to restore the normal microbiome population.Currently, two live biotherapeutic products (LBPs) are approved by the FDA: Ferring's Rebyota and Seres' Vowst (28,29).However, FMT lacks standardization of the treatment protocol and poses a serious risk of transmitting infectious agents that could be fatal especially in immunocompromised and elderly patients (30)(31)(32)(33).Therefore, there is an unmet need to identify new effective therapeutics with high selectivity against C. difficile rather than the beneficial microbiota.
Despite recent technological advances, discovery of a novel drug can take up to 15 years and cost an average of $2-3 billion (34).We screened a library of microbial metabolites and drugs to identify novel anti-C.difficile metabolites of microbial origin and novel drugs that would restrict the recurrence rate and maintain healthy gut microbiota.Microorganisms are traditionally considered abundant reservoirs of rare bioactive compounds.Microbial metabolites have demonstrated efficacy as antibacte rial, anti-tumor, enzyme inhibition, and antiinflammatory agents (35,36).In this study, eight promising microbial metabolites and drugs against C. difficile were identified.The antibacterial activity of these metabolites against a panel of clinically relevant C. difficile strains was evaluated.Additionally, the activity of these metabolites against representa tive members of the human gut microbiota was investigated.

Compounds and library
The microbial metabolite library, which contains a unique collection of 527 microbial metabolites and drugs (HY-L084), was purchased from MedChemExpress (Princeton, NJ, USA).

Library screening against C. difficile
A single screening of the microbial metabolite library was conducted against C. difficile ATCC BAA-1870 at a fixed concentration of 16 µM, using the broth microdilution method as reported previously (37)(38)(39)(40).Briefly, the bacterial strain was cultured onto BHI agar supplemented with yeast extract, vitamin K1, hemin, and L-cysteine (BHIS) and incubated anaerobically at 37°C for 48 h.Test agents were added to 0.5 McFarland bacterial solution, diluted in BHIS broth (~5 × 10 5 CFU/mL), and added to 96well plates.The plates were kept in an anaerobic environment at 37°C for 48 h.The OD 600 was measured using a SpectraMax i3 Multi-Mode Microplate Reader (Molecular Devices, Sunnyvale, CA, USA).With regard to the dimethyl sulfoxide (DMSO) negative control, the percent age of inhibition was calculated for each tested agent."Hits" were metabolites with an inhibition level of 90% and verified to be active and purchased commercially.The metabolites that showed no action had been eliminated from the study.GraphPad Prism version 8.0 was used to illustrate the growth inhibition percentage.

Activity of selected hits against C. difficile
The hits were subsequently tested against C. difficile ATCC BAA-1870 to ascertain the accurate minimum inhibitory concentrations (MICs) of the active hits from the library screening, using the broth microdilution method as previously described (40)(41)(42)(43)(44)(45)(46).Briefly, serial dilutions of the active hits and the control antibiotics (vancomycin and fidaxomicin) were anaerobically incubated with bacterial suspensions (~5 × 10 5 CFU/mL) at 37 °C for 48 h.The 96-well microtiter plates were then examined for growth, and the MIC was determined as the lowest concentration that completely inhibited the bacterial growth as determined by visual inspection of plates.
Based on the MIC data of the active hits (MIC range ≤4 µg/mL), eight microbial metabolites and drugs (ET770, two 8-hydroxyquinoline derivatives [broxyquinoline and choloroquinaldol], ionomycin calcium salt, carbadox, robenidine hydrochloride, dronedarone, and milbemycin oxime) were further investigated.These eight molecules were purchased and tested against a panel of 20 C. difficile clinical isolates, as described above.The MIC 50 and MIC 90 values (the concentration of test agent that inhibited 50% and 90% of the strains respectively) were determined.The experiment was done in triplicates.

Time-kill assay against C. difficile
A time-kill assay was performed against C. difficile ATCC BAA-1870 to evaluate the killing kinetics of the hits.The assay was performed as previously described (37,46,47).Briefly, an overnight culture of C. difficile in the logarithmic growth phase was diluted in pre-reduced BHIS ~ 10 6 CFU/mL (1:100) and exposed to 5 × MIC of the positive hits and the control antibiotics vancomycin or fidaxomicin.Aliquots (10 µL) were taken from each treatment at specific time points, diluted, and plated onto pre-reduced BHIS agar plates to determine the viable bacterial count for up to 24 h.The experiment was done in triplicates.

In vitro activity against representative members of the human gut microbiota
The eight active metabolites and drugs were assessed for their antibacterial activity against reference strains of the human normal gut flora, following established proce dures (43,46,48).A bacterial solution equivalent to a 0.5 McFarland standard was prepared and diluted in MRS broth (for Lactobacillus strains) and BHIS broth (for Bacteroides and Bifidobacterium species) or TSB (for strains of Enterococcus) to achieve a bacterial concentration of ~5 × 10 5 CFU/mL.Serial dilutions of test agents were incubated with bacteria before determining the MICs.Anaerobic incubation was used for Bacteroides and Bifidobacterium species, whereas 5% CO 2 was used for Lactobacillus.Enterococci were incubated aerobically.The experiment was done in triplicates.

Cytotoxicity assessment against vero cells
To evaluate the potential toxic effect of the metabolites, cytotoxicity was assessed against Vero cells as has been described previously (37,38).Briefly, Vero cells were exposed to metabolites and incubated for 24 h.DMSO was included as a negative control.MTS/PMS reagent was subsequently added to the cells and incubated for 2 h.The absorbance values at OD 490 were measured using a Biotek Synergy H1 Hybrid microplate reader.The data were presented as percent viable cells (mean ± standard deviation) relative to DMSO-treated cells.

RESULTS AND DISCUSSION
First-line antimicrobials for C. difficile failed to guarantee clinical cure; patients frequently experience recurrent CDI (49)(50)(51).Therefore, researchers have been exploring alternative treatments for these challenging infections.Today, a large variety of bioactive metabo lites are utilized in medicine, and many antibiotics with microbial origins are available on the market (52).In light of this strategy, and the detrimental effects that antibiotics have on the gut microbiome, we focused on using compounds generated by microorganisms to target C. difficile in an effort to find novel scaffolds with anti-C.difficile action.

Screening a library of microbial metabolites against C. difficile
The microbial metabolite library (containing 527 compounds) was screened against C. difficile ATCC BAA-1870 at a concentration of 16 µM using the broth microdilution assay.According to the CLSI, broth microdilution is not a recommended method for antimicrobial susceptibility testing (AST) of C. difficile, but agar dilution is.However, broth microdilution is more accurate and practical especially in case of screening of drug libraries.The advantages of broth microdilution include reproducibility, the small amount of sample required, and the low cost allowing large numbers of replicates (53).In addition, AST via broth microdilution for C. difficile is well established and reported in multiple publications (46,(54)(55)(56).The initial screening identified 63 compounds that inhibited the growth of C. difficile (>90% of bacterial growth was inhibited) at 16 µM (Fig. 1; Table S3).After excluding antibacterial agents, we selected 18 metabolites (Table 1).
Six metabolites were excluded because they were previously reported, including tinidazole, tioconazole, miconazole, and choloroxine (48); lithocholic acid (57); and octenidine dihydrochloride (58).The other 12 unique compounds, which were not reported against C. difficile previously, were selected for further investigation (Table 2).Afterwards, the 12 metabolites were screened against C. difficile, and their MICs were determined.Of the 12 hits, eight metabolites and drugs (ET-770, broxyquino line, chlorquinaldol, ionomycin, carbadox, robenidine hydrochloride, dronedarone, and milbemycin oxime) displayed the most potent activity against C. difficile ATCC BAA-1870 with MIC values ranging from ≤0.06 μM to 4 µM.Due to their unique nature and lack of prior research, these promising hits were chosen for further investigation.

In vitro activity of the active hits against a panel of C. difficile strains
Next, we evaluated the anti-C.difficile activity of eight active hits against 20 C. diffi cile clinical isolates (Table 3).ET-770 showed strong anti-C.difficile activity (MIC 50 = 0.06 µg/mL) and inhibited the growth of the tested isolates at doses ranging from 0.008 to 0.5 μg/mL.ET-770 is naturally produced from Ecteinascidia thurstoni, which has potent anticancer qualities and can make human lung cancer cells more sensitive to anoikis (59).ET-770 possesses antibacterial activity against Bacillus subtilis, methicil lin-resistant S. aureus (MRSA), and Escherichia coli, with MIC values of 1.01, 2.02, and 32.43 µM, respectively (60).Furthermore, with (half-maximal inhibitory concentration [IC 50 ] of 0.13 µM), ET-770 previously demonstrated extremely strong activity against Mycobacterium tuberculosis (60).
The derivatives of 8-hydroxyquinoline, broxyquinoline, and chlorquinaldol, exhibited strong anti-C.difficile activity with an MIC 50 value of 0.125 µg/mL, which was compara ble to the MIC of fidaxomicin (Table 3).Chlorquinaldol is a common chelating agent that possesses robust antibiofilm action against Pseudomonas aeruginosa and S. aureus biofilms, with MIC values ranging from 0.016 to 0.5 μg/mL against staphylococci (61,62).Chlorquinaldol is administered topically as a cream for skin infections (63).Broxy quinoline is an antiprotozoal agent that is poorly absorbed from the gastrointestinal tract (64,65); broxyquinoline exhibited potent anti-C.difficile activity against the tested isolates (MIC 50 = 0.125 µg/mL).Therefore, broxyquinoline's potency and low absorption make it an intriguing molecule to further pursue for development as a novel anti-C.difficile agent.Against Staphylococcus epidermidis, S. aureus, and Acinetobacter bauman nii, broxyquinoline exhibited modest antibacterial activity with MIC values of 12.5 µM (4 µg/mL) (66).
Ionomycin is a calcium ionophore agent that was isolated from the bacterium Streptomyces conglobatus.Ionomycin exhibits high potency and selectivity against Gram-positive bacteria (67).Ionomycin can activate Ca 2+ /calmodulin-dependent kinase and phosphatase to induce gene expression (68).Ionomycin inhibited the tested C. difficile isolates at concentrations ranging from 0.5 to 4 μg/mL (MIC 50 = 1 µg/mL), which   was similar to the MIC values for vancomycin, the drug of choice for the treatment of CDI (Table 3).
Carbadox is frequently used as a feed additive in swine to stimulate growth and manage swine dysentery.Additionally, carbadox inhibits the growth of spirochete bacteria isolated from dogs (69,70).However, the drug has been associated with the development of tumors and birth defects in laboratory animals.Despite these concerns, carbadox displayed potent activity against the strains of C. difficile that were tested (MIC 50 = 1 µg/mL) (Table 3).
Another promising discovery in our screening is robenidine hydrochloride, an anticoccidial drug with MIC values of 4.7 and 8.1 µM, respectively, that has been shown to be effective against vancomycin-resistant enterococci (VRE) and MRSA (71).Addition ally, robenidine hydrochloride exhibits bactericidal activity against A. baumannii (MIC = 8 µg/mL) and Acinetobacter calcoaceticus (MIC = 2 µg/mL) (72).Robenidine is generally regarded as a safe drug with no observed genotoxicity or mutagenicity (73).Here, we report that robenidine hydrochloride exhibited strong antibacterial activity against C. difficile (MIC 50 = 1 µg/mL), which was comparable to the control antibiotic vancomycin (Table 3).Dronedarone is an antiarrhythmic drug that is used to treat cardiac arrhythmias (74).Dronedarone is a benzofuran molecule, the same as amiodarone; however, it lacks the iodine-containing groups that amiodarone is known to cause thyroid issues with.
Because it has a methyl sulfonyl group in its structure, dronedarone is likewise more lipophilic and has a shorter half-life than amiodarone.Consequently, the probability of drug buildup is reduced in bodily tissues and organ toxicities, such as thyroid and lung concerns (75).Furthermore, dronedarone was previously found to possess antibacterial and antiparasitic activity against S. aureus and Leishmania mexicana (76,77).Dronedar one also exhibits modest anti-C.difficile activity (MIC 50 = 4 µg/mL) in these previous studies.

Time-kill assay against C. difficile
To determine the rate at which the positive hits from the library of microbial metabolites eliminated a high inoculum of C. difficile, we performed a time-kill assay.As shown in Fig. 2, in contrast to the control antibiotics, vancomycin and fidaxomicin, which eliminated C. difficile within 24 h, positive hits exhibited a bactericidal effect on C. difficile ATCC BAA-1870 at 5 × MIC within 2 h, with the exception of broxyquinoline and ET-770, which cleared C. difficile after 6 h.Ionomycin, by contrast, began to show a bactericidal effect after 8 h and completely cleared C. difficile after 12 h.

In vitro activity against representative members of the human gut microbiota
Dysbiosis, induced by the administration of broad-spectrum antibiotics, can disrupt the normal intestinal microbial composition, which can lead to opportunistic infections, such as C. difficile colonizing the gut (81,82).Therefore, it was crucial to ascertain if the active microbial metabolites detected by our screen may have a negative impact on commensal organisms that are often found in the normal gut microbiota.Targeted metabolites were tested for their antibacterial properties against species of Bacteroides, Bifidobacterium, Lactobacillus, and Enterococcus, which are typical of the human gut microbiota.The selected microbial metabolites and drugs exhibited poor activity against the representative members of normal microbiota strains tested.As presented in Table 4, ionomycin, dronedarone, broxyquinoline, and milbemycin oxime exhibited limited activity against Bacteroides with MIC values ranging from 8 to >256 µg/mL.By con trast, the remaining metabolites exhibited potent activity against species of Bacteroides (MIC values, ≤2 µg/mL).Regarding activity against Bifidobacterium species, ionomycin, broxyquinoline, and chlorquinaldol did not show significant inhibition (MIC values ranged from 64 to >256 µg/mL); dronedarone and milbemycin oxime were slightly active (MIC values ranged from 2 to 8 μg/mL), whereas the remaining metabolites significantly inhibited the growth of species of Bifidobacterium (MIC values, ≤2 µg/mL) similar to the control antibiotics vancomycin and fidaxomicin.Additionally, ionomycin, drone darone, broxyquinoline, carbadox, milbemycin oxime, and chlorquinaldol exhibited limited activity against enterococcal species (MIC values ranged from 4 to >256µg/mL).ET-770 and robenidine hydrochloride exhibited more potent activity against entero cocci compared with other compounds (MIC values, ≤2 µg/mL).With the exception of ET-770, all other microbial metabolites exhibited limited activity against species of Lactobacillus.In contrast, vancomycin and fidaxomicin significantly inhibited the growth of the majority of gut microbiota strains tested, which is in agreement with previous reports (16,49,50).

Cytotoxicity assessment against vero cells
Most of these hits are drugs that have been approved for human or veterinary use and their pharmacokinetics and toxicity profiles are well studied.For instance, dronedarone is approved for treatment of atrial fibrillation in humans with dosages up to 400 mg twice daily (83,84).Broxyquinoline is also highly tolerable and its LD 50 in mice is 7.4 g/kg when administered orally (85).Robenidine is generally regarded as a safe drug with no observed genotoxicity or mutagenicity (73).Although, milbemycin oxime and cholroqui naldol are approved drugs with low to no toxicity (61,86).Also, ionomycin showed no toxicity on embryo development in mice (87).Additionally, we performed the in vitro cytotoxicity assessment for ET-770 and carbadox against monkey kidney epithelial (Vero) cells.The IC 50 values of both drugs exceeded 128 µg/mL, indicating that Vero cells were able to tolerate them.This value indicates the high safety index for these metabolites, since it represents more than 1024 and 64 times the MIC 90 of ET770 and carbadox, respectively, against strains of C. difficile (Fig. S1).
Additionally, most of these compounds have a relatively simple structure, such as carbadox, which can be further modified for optimization or to enhance specific anticlostridial activities, such as avoiding absorption from the gut, and reduce toxicity.The information we presented will provide an opportunity for medicinal chemists to further investigate these molecules and synthesize analogs with more potent activity and better pharmacokinetic characteristics, including reduced toxicity.This approach is standard in drug discovery To summarize, screening a library of microbial metabolites led to the discovery of novel lead molecules, such as ionomycin (calcium salt), broxyquinoline, chlorquinaldol, and carbadox that possess potent in vitro anti-C.difficile activity and limited activity against some members of the gut normal microbiota.These metabolites warrant further investigation for the development of new anti-C.difficile therapeutics.

FIG 1 (
FIG 1 (A) Screening the library of microbial metabolites against C. difficile identifies novel lead scaffolds.Results from screening 527 microbial metabolites at a concentration of 16 µM against C. difficile ATCC BAA-1870.Agents designated as hits (red in color) showed >90% reduction of bacterial growth.(B) A flow chart showing how many compounds are evaluated at each step.

FIG 2
FIG 2 Time-kill assay of compounds from the library of microbial metabolites against C. difficile.The positive hits, vancomycin (VAN), and fidaxomicin's (FDX) time-kill assays were performed at 5 × MIC against C. difficile ATCC BAA-1870.On pre-reduced BHIS medium, the start of C. difficile development was seen as CFUs during a 24-h period.

TABLE 1
Active hits identified from initial screening against C. difficile ATCC BAA-1870 (Continued) (Continued on next page)

TABLE 1
Active hits identified from initial screening against C. difficile ATCC BAA-1870 (Continued)

TABLE 3
The MIC values selected for microbial metabolite hits and control anti-C.difficile agents against clinical isolates of C. difficile a The minimum inhibitory concentration (MIC50) against 50% of isolates is represented by the bold values. b