Omadacycline Compared to Vancomycin When Combined with Germinants To Disrupt the Life Cycle of Clostridioides difficile

Clostridioides difficile infections (CDI) are commonly treated with antibiotics that do not impact the dormant spore form of the pathogen. CDI-directed antibiotics, such as vancomycin and metronidazole, can destroy the vegetative form of C. difficile and protective microbiota.

ABSTRACT

Clostridioides difficile infections (CDI) are commonly treated with antibiotics that do not impact the dormant spore form of the pathogen. CDI-directed antibiotics, such as vancomycin and metronidazole, can destroy the vegetative form of C. difficile and protective microbiota. After treatment, spores can germinate into vegetative cells, causing clinical disease relapse and further spore shedding. This in vitro study compares the combination of germinants with vancomycin or omadacycline to antibiotics alone in eradicating C. difficile spores and vegetative cells. Among the four strains in this study, omadacycline MICs (0.031 to 0.125 mg/liter) were lower than vancomycin MICs (1 to 4 mg/liter). Neither omadacycline nor vancomycin in medium alone reduced spore counts. In three of the four strains, including the epidemic ribotype 027, spore eradication with germinants was 94.8 to 97.4% with vancomycin and 99.4 to 99.8% with omadacycline (P < 0.005). In ribotype 012, either antibiotic combined with germinants resulted in 100% spore eradication at 24 h. The addition of germinants with either antibiotic did not result in significant toxin A or B production, which was below the limit of detection (<1.25 ng/ml) by 48 h. Limiting the number of spores present in patient gastrointestinal tracts at the end of therapy may be effective at preventing recurrent CDI and limiting spore shedding in the health care environment. These results with germinants warrant safety and efficacy evaluations in animal models.

INTRODUCTION

Clostridioides difficile is an anaerobic, sporulating, Gram-positive pathogen that has become the most commonly identified hospital-acquired infection (HAI) in the United States (1). C. difficile is one of the CDC’s five urgent-threat-level infections, requiring urgent and aggressive action (2). Hospital-onset infections alone cost an estimated $1 billion in 2017, attributed to 12,800 deaths and 223,900 infections (2). C. difficile infection (CDI) initiates through ingesting spores by either person-to-person spread via the fecal-oral route or direct exposure to the contaminated environment (1). Residing in a room previously occupied by a CDI patient is an independent risk factor for contracting the disease, regardless of other risk factors, such as antibiotic use (3). Once spores reach the intestine and colon, they are introduced to endogenously produced chemicals called germinants, which signal a safe environment for germination into vegetative cells. Vegetative cells multiply within the gastrointestinal (GI) tract, leading to the production of inflammatory toxins. The healthy host microbiome normally provides colonization resistance, preventing vegetative outgrowth, through resource competition and alterations to the metabolome that result in unfavorable growth conditions for C. difficile. Antibiotics are the most significant risk factor for C. difficile colonization progressing to CDI because they remove colonization resistance, altering the makeup of germinants and increasing resource availability (4).

Taurocholate, a bile acid found in human GI tracts, is the strongest known germination signal. Its germination potential increases when combined with cogerminants such as divalent cations and amino acids (5, 6). Taurocholate is hepatically cycled, and concentrations are highest in the small intestine, decreasing down to the colon, which is a resource- and germinant-limited environment. When vegetative cells reach the colonic environment, they form spores, which can be shed into the environment or remain in the GI tract to be activated later. This contributes to subsequent CDI recurrence (7). Even without recurrence following CDI-directed antibiotics therapy, asymptomatic spore shedding still occurs in up to 56% of patients (8). The risk of recurrent CDI (rCDI) depends on CDI treatment but is generally around 25% for the first recurrence, 45% for the second, and 65% for subsequent recurrences (9). During CDI treatment, antibiotics remove vegetative cells, resolving clinical symptoms, but do not impact spores and can create an environment favorable to rCDI by further damaging colonization resistance (10–12). The role of germinant/antibiotic combinations in preventing spore shedding and diminishing the spore reservoir has not been defined.

Vancomycin (VAN), the primary treatment option for CDI, is effective at eliminating vegetative cells but has been associated with decreased colonization resistance, shifting the bile acid pool to favor C. difficile growth by increasing taurocholate concentrations and decreasing growth-inhibiting bile acids (13). Pulsed vancomycin dosing, which allows intermittent antibiotic presence, may prevent the outgrowth of residual spores while allowing the partial return of colonization resistance, although this method is not well substantiated (1). Tetracyclines are one class of antibiotics that have not been associated with CDI and may even be protective (14). Omadacycline (OMC), a new aminomethylcycline antibiotic semisynthetically derived from tetracycline, is not approved for CDI treatment but has strong in vitro activity and a high fecal excretion rate.

This in vitro study evaluates the germination potential of fresh versus purified spores and whether exogenous concentrations of germinants combined with primary treatment with vancomycin or investigational omadacycline can reduce spore burden in media without causing significant increases in toxin production.

RESULTS

Susceptibility testing.

Table 1 describes strain-specific information and antibiotic susceptibility. Both VAN and OMC susceptibilities were within the quality control range for ATCC 700057 (VPI 11186). OMC MICs (0.031 to 0.125 mg/liter) were lower than VAN MICs (1 to 4 mg/liter) for the strains used in this study.

TABLE 1.

Details of strains used in this study^a

Strain	Ribotype	Toxin genes	Vancomycin MIC (mg/liter)	Omadacycline MIC (mg/liter)
ATCC 700057 (VPI 11186)	R038	None	2	0.125
4118 (ATCC 1870)	R027	tcdA, tcdB, cdtA, cdtB	2	0.031
ICCD 0715 (clinical isolate)	R027	tcdA, tcdB, cdtA, cdtB	4	0.031
VPI 10463 (ATCC 43255)	R087	tcdA, tcdB	1	0.063
630 (ATCC BAA-1382)	R012	tcdA, tcdB	4	0.125

^aThe presence or absence of combination toxin genes, cdtA and cdtB, was confirmed with PCR (27). For strains with an MIC range among the replicates, the highest concentration is reported.

Time-kill curves.

Time-kill curves followed the schematic shown in Fig. 1. For BHI alone (growth controls), total numbers of CFU (T-CFU; vegetative and spore growth) and numbers of spore CFU (S-CFU) remained relatively constant during the 96-h treatment period in each strain (Fig. 2). This indicated vegetative cells had appropriate growth conditions for cell viability over the duration of the experiment. The only control with a notable change was the ATCC 1870 S-CFU/ml, increasing from 1.25 × 10⁶ to 3.75 × 10⁷ over 96 h. In ATCC 1870 antibiotic-treated samples, with or without germinants, T-CFU equaled S-CFU from 24 to 96 h, indicating both omadacycline and vancomycin were effective at killing vegetative cells, including those germinating from spores, throughout the experiment. Vegetative cell killing to below the limit of detection by omadacycline and vancomycin within a 24-h period is consistent with previously published time-kill curves (15). BHI conditions with VAN or OMC did not affect S-CFU throughout the treatment period in any of the four strains tested (Fig. 2).

FIG 1.

Schematic of incubation and treatments. Toxin quantification was done on all treatment samples for specified strains at 24 h and at 48 h for controls and if there was a change in S-CFU.

FIG 2.

Time-kill curve results. T-CFU are only shown for controls. In VAN, OMC, VAN+G, and OMC+G samples, T-CFU equaled S-CFU from 24 to 96 h and are not shown for clarity. Control, BHI alone; VAN, vancomycin alone; OMC, omadacycline alone; VAN+G, vancomycin plus germinants; OMC+G, omadacycline plus germinants; T-CFU, total vegetative cell and spore counts; S-CFU, total spore count.

Spore eradication in germinant-treated samples was most effective between 24 and 48 h in most strains, with only moderate further eradication noted from 48 to 96 h. Overall, the degree of spore eradication in germinant-treated samples varied by strain type and antibiotic. The R027 strains ATCC 1870 and ICCD 0715 had the lowest spore eradication in response to BHI with germinants (BHIG) at 96 h. Spore eradication in these R027 strains was 94.8 to 97.4% in VAN plus BHIG (VAN+G) and 99.4 to 99.8% in OMC plus BHIG (OMC+G). Higher levels of eradication were seen in 630, with 97.3% in VAN+G and >99.9% in OMC+G. All final OMC+G S-CFU in these strains were significantly lower than the VAN+G S-CFU (P < 0.005). Finally, VPI 10463 showed the highest level of spore eradication, reaching the limit of detection (1 × 10² S-CFU/ml) at 24 h in both VAN+G and OMC+G.

Toxin production.

ATCC 1870 and ICCD 0715 strains were chosen to represent human epidemic strains, while VPI 10463 was chosen for high toxin production. Toxin A and B production in ATCC 1870 and toxin B production in ICCD 0715 and VPI 10463 were quantified in all 24-h time-kill curve supernatants, while only control, VAN+G, and OMC+G were measured at 48 h, after the largest drop in S-CFU in ATCC 1870 and ICCD 0715 (Table 1). Control, VAN+G, and OMC+G levels were measured at 48 h because a reduction in S-CFU was noted, indicating that spores germinated into vegetative cells, which are capable of toxin production. The S-CFU reduction in BHIG samples between 24 and 48 h did not result in significant toxin B production, and most samples were below the limit of detection of 1.25 ng/ml (Fig. 3). Toxin A had similar results (data not shown).

FIG 3.

Toxin quantification at 24 h and 48 h of antibiotic treatment in the kill curve. VAN, vancomycin alone; OMC, omadacycline alone; VAN+G, vancomycin plus germinants; OMC+G, omadacycline plus germinants. An asterisk indicates a P value of <0.05 versus vancomycin treatment in 24-h samples.

Using VAN as the comparison group, only controls (BHI alone) had statistically significant increases in toxins at 24 h (P ≤ 0.05). In addition, VAN was compared to VAN+G and OMC was compared to OMC+G at 24 h, and neither resulted in statistically significant differences in toxin production, indicating germination with these antibiotics did not increase toxin production.

Spore washing and germination.

The time-kill experiments used a combination of vegetative cells and newly produced spores, more akin to the C. difficile growth cycle environment, while prior studies often purify spores before germination. Therefore, a spore-washing experiment was performed to investigate the influence of in vitro spore preparation and vegetative cell presence on germination rates. The results of vegetative cell presence, heat shocking, water washing, and docusate exposure on percent germination are summarized in Table 2. Vegetative cell presence prevented germination in all strains compared to the other treatments, with less than 10% germination for 630, ATCC 1870, and ICCD 0715. VPI 10463 spores had the highest germination (60%). This was similar to the results of the time-kill curves at 24 h, with VPI 10463 showing the highest germination rate in both experiments and negligible germination for the other strains. BHIG (without docusate) exposure did not change the T-CFU compared to that of BHI (data not shown). Heat shocking increased germination in all strains compared to vegetative cell presence. Four distilled water washes further increased germination, reaching above 90% in all strains. Lastly, 0.5 mg/ml docusate exposure achieved the highest percent germination compared to all other treatments. These results indicate that extrapolating germination rates from studies using purified spores to living systems may not be reflective of in vivo systems, where germination would be affected by vegetative cell presence and unconditioned spores.

TABLE 2.

Effects of vegetative cell presence, heat shocking, washing, and docusate exposure on percent germination

Condition	Germination (%) for:
	630	VPI 10463	ATCC 1870	ICCD 0715
Vegetative cell presence	5.8	60.4	−5.9	−7.3
Heat shock	19.6	>99	40.4	63.7
Heat shock + four water washes	91.7	>99	95.7	97.0
Heat shock + docusate (0.5 mg/ml) exposure	92.8	>99	98.2	98.6

DISCUSSION

C. difficile is a significant infectious complication for patients (1). Treatment options for CDI are limited, and the most commonly used treatment, oral vancomycin, is associated with a ∼25% risk of relapse after treating the first CDI (16). The primary contributor of rCDI is remaining spores that are in the colon or taken up from the environment following antibiotic treatment that germinate prior to the reestablishment of the patient’s colonization-resistant microbiome. In this study, we demonstrate that the germination of spores, in the presence of antibiotics, may eradicate this reservoir without a noted increase in toxin production.

Antibiotics have variable effects on the microbiome and, therefore, different rates of microbiologic relapse of CDI. Vancomycin and metronidazole have been associated with the destruction of colonization resistance, the removal of which is a risk factor for contracting CDI (10). Another recommended treatment option, fidaxomicin, is associated with lower rates of relapse, ∼15%, purportedly because of its narrow spectrum of activity, which preserves the microbiome (17). Tetracycline antibiotics may also preserve aspects of colonization resistance, as they have low association with, and perhaps a protective effect on, contracting primary CDI (14). For example, in a triple-stage chemostat model of CDI, omadacycline was unable to create the conditions needed for C. difficile outgrowth and toxin production (18). Omadacycline is a potent new aminomethylcycline antibiotic with in vitro activity against C. difficile and favorable CDI pharmacokinetic parameters, reaching high concentrations in the feces when given orally (19). In addition, the well-known drug interaction of decreased absorption with calcium and tetracyclines may be used to decrease systemic absorption without a notable effect on omadacycline activity (unpublished data). Coincidentally, calcium also increases the germination rate of taurocholate on C. difficile spores (6).

Spores are the main cause of rCDI, whether taken up from the environment or remaining in the colon until after CDI-directed antibiotic therapy. Unfortunately, antibiotics do not affect spores. As confirmed in this in vitro study, germinants are required for the removal of spores from culture by first converting them into vegetative cells. Antibiotic-only treatments removed vegetative cells within 24 h but had no effect on spore concentrations. Interestingly, germinants had limited effect on spores in the first 24 h for most strains. Based on the spore-washing component, we hypothesized this was due to vegetative cells consuming germinants or possibly changing them into antigerminants, as is seen with Bacillus anthracis (20). A precipitous drop in spore concentrations occurred from 24 to 48 h with daily BHIG replacement and antibiotic exposure, and spore reductions were consistently greater with OMC+G than VAN+G in ATCC 1870, ICCD 0715, and 630 strains.

One of the main concerns in using germinants to eradicate spores in vivo is an increase in toxin production. Germinant-treated samples did have numerically higher toxin amounts at 24 h, although none were statistically significant compared to VAN without germinants. Due to this numerical increase, it may be prudent to wait until clinical symptoms subside before administering germinants when exploring this approach in animal models of CDI. However, taurocholate has been shown to decrease toxin activity ex vivo, which may buffer this numerical increase (21). This is of interest for our further work. Critically, during the largest notable drop in spore concentration, between 24 and 48 h, toxin production was not appreciable in germinant/antibiotic combinations. Limited toxin production in newly germinated spores with antibiotics is encouraging for further evaluating the efficacy of germinant/antibiotic combinations.

Although this study shows encouraging results in using germinant/antibiotic combinations as a spore removal mechanism, there are some limitations that make extrapolation to living systems difficult. The first limitation is testing C. difficile without other microbes of the GI tract. The interactions between C. difficile, humans, and the microbiota are complex and difficult to recapitulate in in vitro models. Additionally, in vitro cultures remained stationary compared to the GI tract, where contents move through the system. We also only quantified ethanol-resistant spores and relied on colony formation for enumeration. Using another method, such as phase-contrast microscopy, would allow for greater resolution. Lastly, we monitored toxins in only three strains, which may not accurately reflect toxin production of other strains.

Based on the results of this experiment, germinant/antibiotic combinations can effectively diminish the C. difficile spore burden in media with minimal impacts on toxin production. Further investigation into the composition and delivery of germinant/antibiotic combinations in animal models is warranted.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

All growth media were reduced for a minimum of 24 h prior to use, and all manipulations took place within a type C vinyl anaerobic chamber (Coy Laboratory Products, Inc., Grass Lake, MI) using a gas mixture of 10% hydrogen and 10% carbon dioxide balanced with nitrogen. The chamber’s hydrogen concentration was kept between 1 and 2%. C. difficile strains were plated onto C. difficile brucella agar (CDBA) from spore stocks and incubated for 24 h in the anaerobic chamber at 37°C (22). A single colony from each strain was used to inoculate 35 ml of brain heart infusion broth (BHI; Becton, Dickinson). C. difficile strains ATCC 1870 and ICCD 0715 were incubated in BHI for 48 to 96 h to achieve a spore concentration of ∼1 × 10⁶ spore CFU (S-CFU)/ml and ∼1 × 10⁷ total CFU (T-CFU)/ml, which captured both spore and vegetative cell growth. VPI 10463 and 630 were incubated for 96 h and conditioned anaerobically at 23°C for 1.5 h on days two and three to achieve ∼1 × 10⁵ S-CFU/ml and ∼1 × 10⁷ T-CFU/ml. The temperature excursion for VPI 10463 and 630 was done to increase sporulation rates for a targeted starting concentration (10⁵ to 10⁶ S-CFU/ml) and was not necessary for ATCC 1870 and ICCD 0715.

After incubation in 35 ml of BHI in a 50-ml centrifuge tube, or at time zero, 10 centrifuge tubes were filled with 1 ml of medium for duplicate treatments. T-CFU were determined daily for controls by dilution in prereduced 1× phosphate-buffered saline (PBS) and plating on CDBA to ensure proper growth conditions. Total counts also were done for all treatment groups in ATCC 1870 to evaluate the continuous removal of vegetative cells with antibiotics. Dilutions for S-CFU were done daily for all treatments and controls in a 70:30 ethanol-1× PBS mixture at a maximum ratio of 1:10 and allowed to settle for ≥15 min to remove vegetative cells before plating on CDBA. Dilutions plated on CDBA were anaerobically incubated for 24 h before enumerating T-CFU and S-CFU for ATCC 1870, VPI 10463, and 630. Counts were enumerated after 48 h for the clinical strain, ICCD 0715, which required increased incubation time to capture complete growth. VPI 10463 and 630 were evaluated because of their frequent use in animal models and for VPI’s ability to produce large amounts of toxins (23). ATCC 1870 and ICCD 0715 represent R027 human epidemic strains. Baseline data about 1870 indicated a high sporulation rate; only ribotype information was available for clinical isolate 0715.

Susceptibility testing.

The MICs of all isolates to vancomycin and omadacycline were determined in triplicate by broth microdilution in reduced BHI containing 2-fold dilutions of antibiotics, similar to previous studies (15). Microwell plates containing omadacycline and BHI were used on the same day as their creation, and all media were reduced for 24 h prior. Media used to inoculate plates were created by suspending 24-h colonies in BHI to a McFarland standard of 0.5, which was added to each well, sufficient to ∼1 × 10⁶ CFU/ml. A total volume of 200 μl was used, and plates were read after 24 h of incubation in the anaerobic hood. The lowest concentration without growth was determined to be the MIC. In accordance with CLSI recommendations, C. difficile ATCC 700057 was used as a control for each plate (24).

Time-kill curves.

Time-kill curves follow the schematic shown in Fig. 1. Experimental conditions included BHI alone (control), BHI with vancomycin or omadacycline (VAN and OMC), or BHI with germinants (BHIG) and antibiotics (VAN+G and OMC+G). BHIG consisted of sodium taurocholate 0.54% (Biosynth), glycine 0.23% (Fisher Scientific), calcium chloride dihydrate 0.0118% (Fisher Scientific), taurine 0.40% (RPI), and calcium docusate 0.01% (USP). BHIG medium was filter sterilized within the anaerobic chamber and frozen at −20°C. Omadacycline at 25 mg/liter and vancomycin at 200 mg/liter were the antibiotic concentrations used in each treatment. These concentrations were chosen to reflect a midpoint between fecal concentrations reached during clinical dosing and our MIC values for C. difficile strain ATCC 700057. Fecal excretion of omadacycline is 81%, equating to approximately 430 mg/liter, and the MIC for ATCC 700057 was 0.125 mg/liter (19). Clinical dosing of oral vancomycin for CDI can achieve 2,000 mg/liter in the feces, while the MIC for 700057 was 2 mg/liter (25).

For each strain, starting T-CFU and S-CFU concentrations were similar between treatments. At time zero, baseline T-CFU and S-CFU samples were taken from the 35 ml of BHI that had been incubating for 48 to 96 h. One-milliliter samples from the 35 ml of incubated BHI then was distributed into 10 microcentrifuge tubes for duplicate treatments. The remaining inoculated medium from the 35-ml container was then discarded. Treatment samples were removed from the anaerobic chamber for centrifugation at 3,000 × g for 10 min. After centrifugation, samples were returned to the anaerobic chamber and supernatant was removed and saved for toxin quantification. Centrifugation, supernatant removal, and resuspension were minimized to once daily, as spore washing increased our germination rates in pilot experiments and does not replicate in vivo conditions. Pellets were resuspended in fresh BHI or BHIG without antibiotics. Dilutions to capture T-CFU and S-CFU were taken after resuspension and before antibiotic addition. Centrifugation, supernatant removal, pellet resuspension in fresh BHI or BHIG, dilutions for T-CFU and S-CFU, and antibiotic addition were done sequentially every 24 h. This was done for 4 days unless S-CFU fell below the limit of detection (1 × 10² CFU/ml) in any sample, at which the experiment was stopped, since no further germination would be detected. All samples were mixed via pipetting and vortexing for 15 s. Spore eradication was calculated as a percentage with the equation [(starting S-CFU − ending S-CFU/starting S-CFU) × 100]. Statistical significance for differences in germination between treatments was determined on log-transformed S-CFU concentrations at 96 h using analysis of variance and t tests.

Toxin quantification.

To determine if germination increased toxin production during antibiotic treatment, C. difficile toxins were quantified in two strains from time-kill curve supernatants after 24 and 48 h of exposure (Fig. 1). C. difficile toxins A and B for strain 1870 and toxin B for ICCD 0715 and VPI 10463 were quantified separately by a C. difficile toxin A or B Quanti kit (tgcBiomics GmbH, Germany) using sandwich enzyme-linked immunoassay according to the manufacturer’s instructions. ATCC 1870 and ICCD 0715 were chosen to represent human epidemic strains with increased sporulation, while VPI 10463 was chosen for its ability to produce large amounts of toxins (23). We were particularly concerned with toxin B production, as it is reported to be 100 to 1,000 more potent than toxin A ex vivo (26). All treatments were quantified at 24 h. Toxins were quantified at 48 h in controls and when germination was detected by a decrease in S-CFU/ml from time zero, because only vegetative cells can produce toxins. Combination toxin was not quantified, but the presence or absence of combination toxin genes cdtA and cdtB was confirmed with PCR, as shown in Table 1 (27).

Results were summarized by treatment groups using means and standard deviations. The equality of variances was tested using Bartlett’s test. Linear regression was used to evaluate statistical differences between VAN and other treatment groups at the 24-h time point. All P values of ≤0.05 were considered statistically significant. We used STATA SE version 16 to analyze data.

Spore washing and germination.

A single 24-h colony of each strain from CDBA was used to inoculate 35 ml of BHI and incubated anaerobically for 96 h at 37°C. Inoculated medium was then exposed to BHIG (without docusate) or BHI after four separate treatments: vegetative cell presence, heat shock, water washing, and docusate exposure. The first treatment, vegetative cell presence, was done anaerobically. After incubation, 35-ml cultures were split into four centrifuge tubes, 1 ml each, pelleted, and resuspended in reduced BHIG or BHI and incubated at 37°C for 1 h before dilution in 70:30 ethanol-1× PBS mixture for S-CFU and 1× PBS for T-CFU. Dilutions for T-CFU were done to ensure vegetative cell presence. The next treatments, heat shock, water washing, and docusate exposure, all started with 5 ml of the 35-ml culture and were done aerobically except for the final step of plating for S-CFU. The 5-ml aliquots for these three treatments were all heat shocked at 60°C for 20 min to destroy vegetative cells. The samples used for testing heat shock and germination were then aliquoted into four centrifuge tubes, resuspended in BHIG or BHI, incubated for 1 h, resuspended in 1× PBS, and heat shocked again before S-CFU enumeration. After heat shocking the water washing treatments, 5-ml samples were resuspended in sterilized water four times before being aliquoted into four centrifuge tubes. Samples were then resuspended in BHIG or BHI, incubated for 1 h, and resuspended in 70:30 ethanol-1× PBS mixture for plating in the anaerobic chamber. For docusate exposure, after heat shock, the 5-ml samples were resuspended in 0.5 mg docusate/ml water and allowed to sit at room temperature for 1 h. Docusate-exposed samples were then aliquoted into four centrifuge tubes, resuspended in BHIG or BHI, incubated for 1 h, resuspended in 1× PBS, and heat shocked before S-CFU enumeration. All treatments were done in duplicate. Percent germination was determined by comparing samples exposed to BHIG versus BHI with the following equation: [(BHI S-CFU − BHIG S-CFU/BHI S-CFU) × 100].