Irritable Bowel Syndrome Therapeutic Has Broad-Spectrum Antimicrobial Activity

ABSTRACT

Otilonium bromide is a poorly absorbed oral medication used to control irritable bowel syndrome. It is thought to act as a muscle relaxant in the intestine. Here, we show that otilonium bromide has broad-spectrum antibacterial and antifungal activity, including against multidrug-resistant strains. Our results suggest otilonium bromide acts on enteric pathogens and may offer a new scaffold for poorly absorbed intestinal antimicrobial therapy.

INTRODUCTION

Otilonium bromide (OB) is an effective orally administered therapy for treating irritable bowel syndrome (IBS), and it is considered to act as a smooth-muscle relaxant (1–6). Studies indicate that OB can act on several proteins, including calcium channels, muscarinic receptors, and tachykinin receptors, to affect muscle contraction (3, 7–10). As a poorly absorbed compound, OB is reported to be almost entirely excreted in feces and also largely unchanged during intestinal passage (3, 11). While the mechanism of action of OB is complex, it has clear benefits in alleviating patient symptoms. IBS is a multifaceted disease, and, more recently, changes in the composition of our gut microbiome have been associated with its onset and symptoms (12–14). Antibiotics, typically used to treat traveler’s diarrhea, such as rifaximin, can also control IBS symptoms, supporting a role for bacteria in IBS etiology (15–17).

Our previous studies of quaternary amine compounds led us to discover that OB had antibacterial activity against Acinetobacter baumannii, Staphylococcus aureus, and Clostridioides difficile (18). Further evaluation of OB’s substructure shows that it is related to the salicylanilide class of small molecules (Fig. 1A). Salicylanilide derivatives can have antibacterial activity against Gram-positive bacteria but generally lack activity against Gram-negative bacteria (19–21). The salicylanilide-like core of OB is appended with a quaternary amine and an eight-carbon alkyl chain. Quaternary amine groups are found in many classes of molecules, including neurotransmitters and biocides (22, 23). As biocides, lone quaternary amines require an alkyl chain length typically of ≥12 carbons for antimicrobial activity (24, 25), but this chain length is cytotoxic (26, 27). The eight-carbon chain length of OB is below the typical cytotoxic length and similar in length to the alkyl chains found on daptomycin and polymyxins. Both the quaternary amine and the alkyl chain likely promote bacterial membrane interactions. Thus, OB appears to be a combination of two antimicrobial backbones (salicylanilide and quaternary amine biocide), which could give it broad-spectrum antibacterial activity.

FIG 1.

(A) Structure of OB (top), salicylanilide (blue), and general quaternary amine biocide (red). Counter ions have been excluded for clarity. Related structures are color matched. (B and C) Time-kill curve of untreated or 0.5×, 1×, and 2× OB MIC-treated E. coli H10407 (B) and S. aureus USA100 635 (C). Cultures were serially diluted and spotted to determine viable numbers of CFU/ml at the indicated time points. Plots show the means and standard errors of the means from 6 biological replicates. The limit of detection was 10² CFU/ml.

We tested OB against a range of Gram-negative and Gram-positive bacteria, focusing on enteric pathogens, potentially pathogenic gut colonizers, and drug-resistant strains (Tables 1 and 2). For comparison we included rifaximin, another poorly absorbed broad-spectrum antibiotic of the rifampin family (15–17). We also tested carbenicillin against Gram-negative bacteria and vancomycin against Gram-positive bacteria. MICs were determined using Mueller-Hinton broth dilution methods in at least biological triplicate (28, 29). Changes in growth medium used for some strains are noted in the tables. OB stock solutions were dissolved in water at 5 mg/ml.

TABLE 1.

MIC of otilonium bromide, rifaximin, and carbenicillin toward Gram-negative bacteria

Strain	Note	MICa (μg/ml) of:
		OB	Rifaximin	Carbenicillin
Salmonella enterica serovar Typhimurium LT2		32	8	8
S. Typhimurium 14028S		32	8	8
E. coli B7A		32	8	8
E. coli H10407		32	8	8
E. coli O157 (CDC 0427)	MDR	32	4	>256
E. coli O157 (CDC 0429)	MDR	32	4	128
C. jejuni CG8421		1	32	ND
C. jejuni 81-176		4	>128	ND
A. baumannii AYE	MDR	16	16	>256
A. baumannii 5075	MDR	16	2	>256
S. flexneri SA100		32	4	4
S. sonnei (CDC 0422)	MDR	32	8	>256
S. flexneri (CDC 0423)	MDR	32	>256	>256
K. pneumoniae BAA-1705	MDR	64	64	>256
K. pneumoniae MKP103		64	>256	>256
K. pneumoniae ATCC 4208		64	16	4

^aND not determined.

TABLE 2.

MIC of otilonium bromide, rifaximin, and vancomycin toward Gram-positive bacteria

Strain	Note	MIC (μg/ml) of:
		OB	Rifaximin	Vancomycin
S. aureus USA100 635	MRSA	4	<1	1
S. aureus USA300 AH1263	MRSA	4	<1	<1
S. aureus MU50	MRSA	4	>256	2
E. faecium 1230933a	VRE	4	8	>256
E. faecium 1231502a	VRE	8	32	>256
E. faecalis HIP11704a	VRE	4	1	>256

^aMIC determined in BHI medium.

OB was broadly active against Gram-negative bacteria, with MICs between 1 and 64 μg/ml (Table 1). The majority of strains tested had a MIC of 16 to 32 μg/ml. Campylobacter jejuni was the most sensitive species, with MICs of 1 and 4 μg/ml. Interestingly, OB was significantly more active against C. jejuni than rifaximin. A recent clinical trial using rifaximin to prevent Campylobacter infection with strain CG8421 was unsuccessful (30), possibly due to the tolerance of this strain. Klebsiella pneumoniae strains were the least sensitive to OB, with MICs of 64 μg/ml. Klebsiella species can express a thick capsule layer that can inhibit the action of antimicrobials (31, 32), which may also lead to decreased activity of OB. OB showed similar activity between multidrug-resistant (MDR) and non-MDR strains, suggesting its action is not affected by traditional mechanisms of drug resistance.

OB was more potent against Gram-positive bacteria than Gram-negative bacteria (Table 2). Nearly all strains tested had a MIC of 4 μg/ml OB. This MIC was quite consistent compared to those of rifaximin and vancomycin, which ranged from <1 μg/ml to >256 μg/ml depending on the strain. Similar to our observations with Gram-negative bacteria, OB activity was largely unaffected by current drug resistance mechanisms present in the Gram-positive bacteria tested.

We performed time-kill studies using representative Gram-negative (Escherichia coli H10407) and Gram-positive (S. aureus USA100 635) strains (Fig. 1B and C). The same medium and inoculum used for determining MICs were used for these time-kill studies. Exponentially growing bacterial cultures were untreated or treated with 0.5×, 1×, or 2× OB MIC. At the indicated time points, samples were drawn, serially diluted, and plated to determine remaining viable CFU numbers (CFU/ml). We found 0.5× MIC had only a slight effect on E. coli viability in the first hour before growth resumed. In contrast, both 1× and 2× OB MIC E. coli cultures quickly lost viability below the level of detection of our assay (10² CFU/ml) within the first 0.5 to 1 h. Membrane-disrupting agents frequently cause rapid loss of Gram-negative viability (33, 34). In comparison, 0.5×, 1×, and 2× MIC treatment of S. aureus cultures resulted in a graded dose-dependent loss in viability over the time course. OB at 2× MIC caused loss of S. aureus viability below our detection limit by 4 h. At 1× MIC, viable S. aureus cells were still observed at 4 h. The chemical nature of OB suggests a membrane lytic mechanism of action, which is supported by our previous studies that showed OB promoted the influx of a cell membrane-impermeable dye (propidium iodide) into the bacterial cytoplasm (18). The difference in MIC and time-kill results between E. coli and S. aureus may be related to OB accumulation at the cytoplasmic membrane, required to induce cell death. A critical OB concentration may be required to first breach the outer membrane of Gram-negative bacteria, like E. coli, leading to an all-or-nothing loss in viability at the MIC, whereas no such barrier is present for Gram-positive bacteria.

Bacteria generally show low rates of spontaneous resistance to membrane-active antimicrobials. Consistent with a membrane-targeting mechanism of action, we observed low spontaneous OB resistance of both E. coli and S. aureus compared to rifaximin (Table 3). Bacteria were plated on solid growth medium plates with 5× MIC OB or rifaximin, and resistant colonies were identified. We have been unable to isolate any OB-resistant strains after repeated attempts.

TABLE 3.

Rate of emergence of mutants spontaneously resistant to rifaximin and OB

Strain	Concn	Rifaximina	OBa,b
S. aureus USA100 635c	5× MIC	2.8 × 10−7	>10−9
E. coli H10407d	5× MIC	2.3 × 10−8	>10−9

^aThe resistance frequency is expressed as the mean from three biological replicates.

^bThe detection limit for this assay was 10⁻⁹.

^cTSA plates were used.

^dRPMI 1640 plates were used.

While less studied than bacteria, fungi are also important intestinal colonizers that influence our health and disease (35, 36). Colonization of Candida species is associated with their subsequent infection (37, 38). We tested three fungal pathogens and found all were susceptible to OB between 0.5 and 4 μg/ml (Table 4). In each case, OB was much more active than the antifungal fluconazole.

TABLE 4.

MIC of otilonium bromide and fluconazole toward fungi

Straina	MIC (μg/ml)
	OB	Fluconazole
C. albicans SC5314	1	2
C. parapsilopsis ATCC 22019	0.5	2
C. krusei ATCC 6258	4	64

^aMICs performed in YPD medium.

Our study shows that the IBS therapeutic otilonium bromide is a broadly active antibacterial and antifungal. It retains activity against drug-resistant strains and has low spontaneous levels of resistance. Given the importance of our microbiome in IBS, our results suggest that part of the action of OB in these patients is due to the activity against pathogenic gut colonizers. Efforts to elucidate the effect on the commensal gut microbiota are currently in progress. Furthermore, OB may represent an unexplored chemical scaffold for the development of nonabsorbed intestinal antibacterials and antifungals.