In Vitro Activity and Single-Step Mutational Analysis of Rifamycin SV Tested against Enteropathogens Associated with Traveler's Diarrhea and Clostridium difficile

Abstract

Rifamycin SV is a broad-spectrum, poorly absorbed antimicrobial agent that, when coupled with MMX technology, is being targeted for the oral treatment of traveler's diarrhea (TD) and Clostridium difficile-associated disease (CDAD). Rifamycin SV was tested for activity against 911 TD-associated enteropathogens and 30 C. difficile isolates collected from several global surveillance studies. Rifamycin SV demonstrated similar antimicrobial activity levels against the Enterobacteriaceae, with MIC₅₀ values ranging from 32 to 128 μg/ml for all but one strain (an enterotoxigenic Escherichia coli at >512 μg/ml). For non-Enterobacteriaceae strains, MIC₅₀ values ranged from 2 to 8 μg/ml, with the exception of Campylobacter spp., for which all strains had MIC values of >512 μg/ml. Rifamycin SV also demonstrated excellent activity (MIC₅₀ of ≤0.03 μg/ml) against most C. difficile strains (including one hypervirulent NAP1 strain), and this activity was even superior to the potency observed for vancomycin, metronidazole, and rifaximin. In mutational passaging studies, rifamycin SV induced stable resistance and showed a mutation frequency in E. coli similar to that of rifampin. This study presents the potency of rifamycin SV for enteropathogens commonly recovered from patients with TD and CDAD. Additional in vitro and in vivo studies appear necessary to determine the utility of rifamycin SV as an oral agent for the prevention and treatment of TD and CDAD.

Rifamycin SV is a broad-spectrum semisynthetic antimicrobial agent of the rifamycin group with limited oral absorption that is active against Gram-positive bacteria and moderately active against Gram-negative organisms (18). Coupled with the proprietary drug delivery system MMX (Cosmo Technologies Ltd., Dublin, Ireland), designed to release the antibiotic in the colonic lumen, rifamycin SV has been formulated as a tablet for treatment of colonic bacterial infections, including traveler's diarrhea (TD) and Clostridium difficile-associated disease (CDAD). As with all members of the rifamycin group, rifamycin SV inhibits DNA transcription by interfering with bacterial RNA polymerases (2, 22).

TD is the most common gastrointestinal (GI) illness contracted by persons from developed countries when they visit resource-poor countries; annually, it is estimated that 100 million persons travel internationally, with an estimated 30 to 40% of them developing TD and approximately 3% progressing to postinfectious irritable bowel syndrome (15, 19). The most common bacterial etiologies of TD, accounting for nearly 60 to 80% of the cases, are pathogenic Escherichia coli, Salmonella spp., Shigella spp., and Campylobacter spp. although pathogen-specific prevalence rates do vary by geographic region (19).

Current travel medicine treatment guidelines recommend the use of antimicrobials for TD, in addition to fluid replacement and antimotility drugs, especially for severe disease (10). These guidelines recommend the use of a fluoroquinolone (FQ) or azithromycin for most travelers worldwide although regional pockets have significant levels of FQ-resistant (FQR) enteropathogens, especially in Southeast Asia and the Indian subcontinent. An alternative antimicrobial treatment option that is recommended for TD is rifaximin (12, 15), a compound closely related to rifamycin SV (18). Rifaximin is available in several countries and has been approved in the United States for the treatment of nondysenteric TD caused by noninvasive strains of E. coli in patients of ≥12 years of age (13). In clinical trials, rifaximin was shown to significantly shorten GI symptoms associated with TD compared to placebo and had similar outcomes compared to FQ treatment (13, 20). Since rifamycin SV is structurally closely related to rifaximin, it is possible that the biological activities of rifamycin SV will be similar to those of rifaximin. CDAD is the leading cause of nosocomial infectious diarrhea in developed countries, with a disease spec- trum ranging from asymptomatic carriage to life-threatening pseudomembranous colitis (7). In addition, CDAD results in a very high economic burden for health care systems (8). Although antimicrobial resistance to standard treatment options of oral vancomycin and metronidazole is still rare, decreased susceptibility to these two antimicrobials has been reported along with increasing and geographically variable rates of resistance to other antimicrobials, such as the fluoroquinolones and macrolides (11). In addition, treatment has been compromised by the emergence of hypervirulent clones (such as NAP1) that cause outbreaks with both recurrent and refractory CDAD within groups at increased risk, such as children and peripartum women (7, 11). Poor response rates to current standard therapies have resulted in follow-up investigations into many alternative approaches (such as vaccines and stool transplantations, among others). Among several new antimicrobial treatment approaches, rifamycins have been investigated, and initial in vitro and animal and human studies support their use (7).

The first objective of this study was to evaluate the antimicrobial activity of rifamycin SV against bacterial pathogens commonly associated with TD and against C. difficile using reference in vitro assays. The second objective was to evaluate the ability of selected E. coli isolates to develop/acquire resistance following subinhibitory exposures (passaging) to rifamycin SV over 7 days, and the final objective was to establish the rates of single-step mutations at drug concentrations of 4×, 8×, and 16× MIC.

MATERIALS AND METHODS

Organism collection.

A total of 941 enteropathogens were included in this study. Isolates with a known clinical source were collected from 33 countries (more than 100 medical centers) in five regions worldwide, including Latin America, North America, Europe, North Africa, and the Asia-Pacific region. The two primary sources for all enteropathogens were stool specimens (80.5%) and blood cultures (16.0%). All isolates were collected between 2000 and 2009, with 581 (63.8%) isolates collected between the years 2005 and 2009. Less frequently isolated species and those not within the scope of surveillance programs monitored by JMI Laboratories were collected from external sources dating from the year 2000.

The organism distribution was as follows: E. coli, 443 total isolates) including enterohemorrhagic E. coli ([EHEC]O157:H7; 105 isolates) enterotoxigenic E. coli ([ETEC] 201 isolates), enteropathogenic E. coli ([EPEC] 45 isolates), and enteroaggregative E. coli ([EAEC] 92 isolates); Salmonella spp. (102 isolates); Shigella spp. (105 isolates); Aeromonas hydrophila complex (101 isolates); Campylobacter jejuni and Campylobacter coli (102 isolates); Vibrio parahaemolyticus (42 isolates); Plesiomonas shigelloides (16 isolates); and C. difficile (30 isolates, including two NAP1 strains). For the passaging study, characterized E. coli strains were used, including the following: E. coli ATCC 25922 (control strain), E. coli 012-1222G (EHEC), E. coli 5347J (ETEC), E. coli 5753J (EAEC), and E. coli 915J (EPEC). All E. coli control strains used for the passaging experiments were susceptible to β-lactams, fluoroquinolones, aminoglycosides, tetracyclines, polymyxin B, and trimethoprim-sulfamethoxazole prior to and after passaging isolates over a 7-day interval (data not shown).

Antimicrobial susceptibility testing.

MIC values were determined using the reference CLSI broth microdilution methods (3-5). Microdilution panels were produced by JMI Laboratories and utilized cation-adjusted Mueller-Hinton broth (CAMHB). Quality control (QC) ranges and interpretive criteria for an antimicrobial agent with established QC ranges were as published in CLSI document M100-S20 (6). QC strains used included E. coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853. Antimicrobial susceptibility testing of C. jejuni and C. coli isolates was determined using the reference CLSI broth microdilution method with 48 h of incubation at 36°C (5). Microdilution panels utilized CAMHB supplemented with 2.5 to 5% lysed horse blood. QC was confirmed by using C. jejuni ATCC 33560, and an antimicrobial agent with an established QC range was used for these determinations. Antimicrobial susceptibility testing of C. difficile was determined using the reference CLSI agar dilution method (4). Agar dilution plates were produced by JMI Laboratories utilizing brucella agar supplemented with laked sheep blood. QC and interpretive criteria for comparator compounds were those published by the CLSI (4), and tested QC strains included C. difficile ATCC 700057, Bacteroides fragilis ATCC 25285 and Bacteroides thetaiotaomicron ATCC 29741. Rifamycin SV was provided by the sponsor in powder form and tested in a 12-log₂ serial dilution schedule (0.25 to 512 μg/ml). Rifampin, obtained from Sigma-Aldrich (St. Louis, MO), was tested in parallel as a control agent.

Passaging and selection of resistant strains.

From the MIC panel, the entire contents of the last well with growth was removed from both the rifamycin SV and rifampin panels and placed into tubes of broth medium. Tubes were placed in an ambient air incubator to allow growth to reach a 0.5 McFarland standard (1.5 to 3 h). Within 15 min of preparing a 0.5 McFarland, the appropriate amount of bacterial suspension was transferred to testing medium and vortexed. MIC panels were then inoculated using the appropriate volume and concentration, and this process was repeated through seven passage days. The strains were tested against numerous selected antimicrobial agents prior to resistance selection and upon completion of passaging to evaluate the emergence of cross-resistance. Reversion to the original rifamycin SV and rifampin MIC values was assessed by three passages performed on drug-free Mueller-Hinton (MH) medium, with final retesting by the broth microdilution method panel containing these agents. Passaging of each strain was performed in duplicate.

Single-step mutation rates.

Fresh colonies from an agar plate were emulsified in sterile broth or saline until at least a 4 McFarland standard (1.9 × 10⁸ to 4.6 × 10⁹ CFU/ml) was achieved. Final volumes of 0.1 and 1 ml of the inoculum suspension were plated onto agar plates containing 4×, 8×, and 16× MIC for both rifamycin SV and rifampin. Serial dilutions of the inoculum suspensions were plated onto antimicrobial-free MH agar plates to quantify the colony count (CFU/ml) at each exposure concentration. Each strain count was performed in triplicate, and the mutant counts were averaged for calculation of the mutation frequency. The mutant frequency was determined to be the ratio of the number of mutants to the total number of bacteria in the population.

Molecular methods.

Pulsed-field gel electrophoresis (PFGE) analysis of parent strain and mutant strains was assessed upon completion of the experiment. Digestion with the appropriate restriction enzyme (SpeI) was used to produce DNA fragments of a number and size that yielded useful patterns for strain typing (21).

RESULTS

Antimicrobial susceptibility study.

Similar rifamycin SV antimicrobial activity levels were observed for all Enterobacteriaceae, with MIC₅₀ values ranging from 32 to 128 μg/ml (Tables 1 and 2). Rifamycin SV inhibited all tested E. coli strains (MIC₅₀ and MIC₉₀ [MIC_50/90], 32/128 μg/ml), with the exception of one ETEC strain having a MIC value of >512 μg/ml (Table 1). Similar antimicrobial activities were noted among the pathogenic E. coli subgroups, with rifamycin SV being slightly more potent against EHEC and EAEC (MIC₅₀, 32 μg/ml for both) than against ETEC and EPEC (MIC₅₀, 64 μg/ml for both). Similar MIC values were observed for Shigella spp. (MIC_50/90, 32/64 μg/ml), with 100.0% of the Shigella spp. strains having MIC values at ≤128 μg/ml (Tables 1 and 2). Rifamycin SV was slightly less potent against Salmonella spp. (MIC_50/90, 128/256 μg/ml), with the vast majority of isolates (93.1%) exhibiting MIC values at 128 μg/ml (55.9%) and 256 μg/ml (38.2%).

TABLE 1.

Antimicrobial activity of rifamycin SV against a collection of 911 TD-associated enteropathogens

Organism (no. of isolates tested)	Rifamycin SV MIC (μg/ml)
	50%	90%	Range
E. coli (443)a	32	128	2->512
EHEC (105)	32	64	8-256
ETEC (201)	64	128	2->512
EPEC (45)	64	128	16-128
EAEC (92)	32	128	4-256
Salmonella spp. (102)b	128	256	16-256
Shigella spp. (105)c	32	64	8-128
Campylobacter spp. (102)d	>512	>512	>512
A. hydrophila (101)	4	16	2-512
P. shigelloides (16)	8	8	4-8
V. parahaemolyticus (42)	2	2	2-4

^aEHEC, enterohemorrhagic E. coli; ETEC, enterotoxigenic E. coli; EPEC, enteropathogenic E. coli; EAEC, enteroaggregative E. coli.

^bSalmonella entericaserovar Choleraesuis (1 strain), S. entericaserovar Derby (1 strain), S. entericaserovar Dublin (1 strain), S. enterica (1 strain), S. entericaserovar Enteritidis (10 strains), S. entericaserovar Hadar (1 strain), S. entericaserovarHeidelberg (three strains), S. entericaserovar Java (1 strain), S. entericaserovar Panama (2 strains), S. entericaserovar Paratyphi (12 strains), S. entericaserovar Reading (1 strain), S. entericaserovar Stanley (1 strain), S. entericaserovar Typhi (19 strains), S. entericaserovar Typhimurium (five strains), S. entericaserovar Virchow (1 strain), group B Salmonella (12 strains), group C Salmonella (2 strains), group D Salmonella (6 strains), and unspeciated Salmonella (22 strains).

^cShigella boydii (3 strains), Shigella dysenteriae (five strains), Shigella flexneri (33 strains), Shigella sonnei (51 strains), and unspeciated Shigella (13 strains).

^dC. coli (16 strains) and C. jejuni (86 strains).

TABLE 2.

In vitro activity of rifamycin SV in comparison to selected antimicrobial agents tested against C. difficile (30 strains)

Antimicrobial agent	MIC50 (μg/ml)	MIC90 (μg/ml)	MIC range (μg/ml)
Rifamycin SV	0.03	256	≤0.015-512
Rifaximin	0.12	>512	≤0.06->512
Metronidazole	0.25	0.5	0.12-0.5
Vancomycin	0.5	1	0.5-8

Among the non-Enterobacteriaceae species tested, rifamycin SV lacked measurable activity against C. coli and C. jejuni, with all strains having a MIC value of >512 μg/ml (Table 1). Against A. hydrophila, rifamycin SV MIC values ranged from 2 to 512 μg/ml (MIC_50/90, 4/16 μg/ml), but all P. shigelloides strains were inhibited by a rifamycin SV MIC value of ≤8 μg/ml. The most susceptible organism species among the non-Enterobacteriaceae was Vibrio parahaemolyticus (MIC₅₀ and MIC₉₀, 2 μg/ml), which demonstrated a very narrow MIC distribution, with 95.2% of strains inhibited at 2 μg/ml and the remaining two isolates at 4 μg/ml (Table 1). Rifamycin SV was very active (MIC₅₀, ≤0.03 μg/ml) against 26/30 (86.6%) C. difficile strains (including one hypervirulent NAP1 strain), but high MIC values (256 to 512 μg/ml) were observed against the remaining four C. difficile strains (also including one hypervirulent NAP1 strain). Rifamycin SV (MIC₅₀, 0.03 μg/ml) was 4-fold more active than rifaximin (MIC₅₀, 0.12 μg/ml), 8-fold more active than metronidazole (MIC₅₀, 0.25 μg/ml), and 16-fold more active than vancomycin (MIC₅₀, 0.5 μg/ml) against C. difficile (Table 2).

Passaging study.

Table 3 lists the passaging results for all tested strains. The MIC for E. coli ATCC 25922, the designated wild-type isolate, increased only 2-fold with rifamycin SV (32 μg/ml to 64 μg/ml) and rifampin (8 μg/ml to 16 μg/ml) after seven passage days, with slight MIC variations observed in the daily replicate assays. The E. coli 012-1222G (EHEC O157:H7) MIC of rifamycin SV increased 4- to 64-fold and that of rifampin increased 2- to 4-fold at day 7. For this strain, significant variation was observed between the daily replicated assays, with up to a 32-fold difference between replicates. The MICS of both rifamycin SV and rifampin for E. coli 5347J (ETEC) and E. coli 5753J (EAEC) increased 8- to 16-fold. The E. coli 915J (EPEC) MIC remained consistent at 32 μg/ml throughout the experiment for both replicates; however, rifampin MIC values increased 4- to 64-fold for this isolate, with significant variability between the two replicates in the last day of passaging. After passage on antimicrobial-free medium, most strains maintained the MIC value obtained after passaging for 7 days, and none of the isolates reverted back to the original (baseline) MIC value. For each strain, PFGE patterns of the parent strain (day 0) and matched day 7 strains with three additional days passaging on antimicrobial-free medium were considered to be identical (data not shown).

TABLE 3.

MIC results of serial passaging experiments with rifamycin SV and rifampin tested against five E. coli strains

E. coli strain	Passage day or period	MIC(μg/ml)
		Rifamycin SV		Rifampin
		Assay 1	Assay 2	Assay 1	Assay 2
ATCC 25922	0	32	32	8	8
	1	32	32	8	16
	2	32	32	8	16
	3	64	32	16	16
	4	64	32	16	16
	5	64	64	16	16
	6	128	64	16	16
	7	64	64	16	16
	Ra	64	64	16	16
012-1222G	0	32	32	8	8
	1	32	64	16	8
	2	64	64	16	8
	3	128	64	16	16
	4	1,024	64	16	16
	5	1,024	128	16	16
	6	2,048	128	16	8
	7	2,048	128	32	16
	R	2,048	64	16	16
5753J	0	8	8	4	8
	1	32	32	16	8
	2	32	32	32	16
	3	64	64	16	16
	4	64	64	16	64
	5	64	64	32	64
	6	128	128	32	128
	7	128	128	32	128
	R	64	64	16	64
5347J	0	8	8	4	4
	1	16	16	8	8
	2	16	32	16	16
	3	64	64	32	32
	4	64	64	32	32
	5	64	64	32	32
	6	128	128	32	32
	7	64	64	32	32
	R	64	64	32	32
915J	0	32	32	8	8
	1	32	32	8	8
	2	32	32	8	8
	3	32	32	8	8
	4	32	32	8	8
	5	32	32	16	16
	6	32	32	32	32
	7	32	32	1,024	16
	R	32	32	1,024	16

^aPostpassaging reversion subcultures on drug-free medium.

Mutation frequency determination study.

Rifamycin SV and rifampin mutants were observed when the strains were exposed to 4×, 8×, and 16× the MIC value for all five E. coli strains tested; however, the mutation frequency was independent of the antimicrobial concentration. Among the E coli strains tested, the mutation frequencies for rifamycin SV ranged from 1.4 × 10⁻⁶ to <5.0 × 10⁻¹⁰, and for rifampin, they ranged from 3.3 × 10⁻⁷ to <5.0 × 10⁻¹⁰ (Table 4). PFGE patterns of rifamycin SV for the single-step mutants were identical to those of the initial (parent) isolate for all five strains at each of the tested concentrations (data not shown).

TABLE 4.

Mutation frequency with a single exposure to rifamycin SV and rifampin in five E. coli strains

Drug and strain	Mutation frequency at the indicated drug concna
	4× MIC	8× MIC	16× MIC
Rifamycin SV
E. coli ATCC 25922	5.9 × 10−8	5.9 × 10−8	8.8 × 10−8
E. coli 012-1222G	1.4 × 10−7	1.4 × 10−7	<5.0 × 10−10
E. coli 5347J	7.8 × 10−7	8.7 × 10−8	1.7 × 10−7
E. coli 915J	1.4 × 10−6	6.3 × 10−7	1.6 × 10−7
E. coli 5753J	8.7 × 10−7	2.0 × 10−7	2.0 × 10−7
Rifampin
E. coli ATCC 25922	1.3 × 10−8	8.5 × 10−9	4.8 × 10−9
E. coli 012-1222G	3.5 × 10−8	3.9 × 10−8	3.3 × 10−8
E. coli 5347J	4.7 × 10−9	2.9 × 10−9	1.8 × 10−9
E. coli 915J	4.0 × 10−9	<5.0 × 10−10	<5.0 × 10−10
E. coli 5753J	1.4 × 10−8	5.0 × 10−9	6.7 × 10−8

^aDrug concentration is based on the MIC for each individual strain. The mutation frequency was determined to be the ratio of the number of mutants to the total number of bacteria in the population.

DISCUSSION

Rifamycin SV is a semisynthetic antimicrobial agent closely related to another rifamycin-derived antimicrobial agent, rifaximin, which is licensed for the treatment of TD in the United States and is widely available in several countries. Rifaximin, a poorly absorbed oral antimicrobial agent, achieves luminal concentration up to 8,000 μg/gram of feces after a 3-day treatment regimen (12), which is manyfold higher than measured MIC values obtained from in vitro testing for several enteropathogens. Likewise, rifamycin SV has been shown to have very limited absorption in humans after oral administration (1). When rifamycin SV was administered in tablets to be delivered directly to the colon, with little or no loss of antimicrobial activity while transiting the gastrointestinal tract, the total recovery in the feces was >80% of the administered dose, which represents in the colonic environment a concentration far exceeding the MIC values for the isolates tested (data on file; Cosmo Technologies Ltd., Dublin, Ireland).

In our study, rifamycin SV showed consistent potency among the groupings of enteropathogens, both Enterobacteriaceae and non-Enterobacteriaceae, with the exception of Campylobacter spp. Compared to rifaximin in vitro MIC studies (9, 16), rifamycin SV was slightly less potent against Enterobacteriaceae (rifaximin MIC₉₀ of 8 to 64 μg/ml versus a rifamycin SV MIC₉₀ of 64 to 256 μg/ml), had similar potency for non-Enterobacteriaceae (rifaximin MIC₉₀ of 4 μg/ml versus a rifamycin SV MIC₉₀ of 2 to 16 μg/ml), and was much less potent for Campylobacter spp. (rifaximin MIC₉₀ of 32 μg/ml versus a rifamycin SV MIC₉₀ of >512 μg/ml). Nevertheless, these differences of microbiological activity between the two antibiotics should not have a great clinical impact considering the high concentrations of rifamycin SV reached in the lumen of the colon due to the MMX formulation, which is specifically designed for colonic release.

Our data also show that rifamycin SV was very active against most C. difficile strains tested, including one hypervirulent NAP1 clone, resulting in 4-, 8-, and 16-fold more potency than rifaximin, metronidazole, and vancomycin, respectively. This new and interesting activity could be regarded as a promising option for a possible role of rifamycin SV in the treatment of C. difficile-related enteritis.

For an antimicrobial agent to remain effective, the selection of resistance among targeted species should be minimized during antimicrobial exposure, and then any increase in resistance should revert back to baseline once the exposure is removed. Previous studies have noted that rifamycins, including rifaximin and rifampin, are prone to inducing resistant mutants in E. coli at approximately the rate of 10⁻⁸, and these mutants are then stable, maintaining increased MIC values even after successive passage on antibiotic-free medium. The results of our passaging experiments (4- to 64-fold stable increases in MIC values) and the single-step mutation experiments (approximately 10⁻⁷ for rifamycin SV and approximately 10⁻⁹ for rifampin) were consistent with these previously published studies referring to rifaximin and rifampin resistance (14, 17, 18, 22). In Ruiz et al. (17), the authors suggest that sustained rifaximin resistance was likely due to the presence of chromosomal mutation in the rpoB gene or a stable deregulated efflux pump. Specific molecular studies will be required of our rifamycin SV-resistant mutants to determine the underlying mechanism(s) of resistance, as no obvious genetic changes between parent and progeny strains were observed in the PFGE results.

In conclusion, the data presented here quantitate the baseline potency of rifamycin SV when it is tested against enteropathogens causative TD and CDAD. The results of our study should be interpreted in the context of the high concentrations of rifamycin SV achieved in the intestinal lumen and the promising results of initial clinical trials. Antimicrobial agents with bioavailability restricted to the gastrointestinal tract (high intraluminal and fecal concentrations) that largely exceed the pathogen MIC values are needed against a wide range of organisms for the treatment of bacterial diarrheal disease. Pending further clinical trials examining clinical outcomes and microbial eradication, rifamycin SV may prove to be a valuable oral agent for the prevention and treatment of acute gastroenteritis for several enteropathogens associated with TD and CDAD.