Abstract
Introduction
Antibiotic resistance is an increasing problem worldwide. The particular characteristics of developing countries, such as the lack of adequate sanitation, lack of regulation of antimicrobial use, lack of microbiology facilities, frequent inadequate and unnecessary antibiotic treatments, limited antibacterial options, inadequate storage conditions, unfinished treatments, or fraudulent compounds among others, may result in high levels of resistance to the most commonly available antimicrobial agents such as cotrimoxazole, ampicillin, or tetracycline. 4 In addition, in some areas such as India, resistance to quinolones has emerged in the last years. 7 TD reflects the situation in these countries.7–9 Thus, pathogens showing high antibiotic resistance profiles, including β-lactamases, quinolones, and others, have been described as a cause of TD.
This situation shows the need to develop studies in search of alternative treatments. Rifaximin (Rfx) is a semisynthetic derivative of rifamycin with in vitro activity against Gram-positive and Gram-negative bacteria. It is not absorbed in the intestinal tract because of its possession of an additional pyridoimidazole ring, thereby avoiding access to the systemic circulation.5,12 The evidence published suggest that Rfx should be considered as an option for the treatment of uncomplicated TD and, in some countries, is used to treat other intestinal diseases. Rfx possesses a superior efficacy to placebo or loperamide and comparable efficacy to a systemically absorbed antibacterial, such as ciprofloxacin, in reducing the duration of illness, rapidly restoring wellness in most of patients, and thereby minimizing the interruption caused by TD. 6 However, Rfx is unlikely to be effective against invasive bacterial strains. 6 Rfx acts by binding to the beta subunit of RNA polymerase (encoded in the rpoB gene), suppressing the replication activity of genetic material. 5 Different mutations in the rpoB gene have been related to the acquisition of resistance to rifamycin derivatives such as rifampicin. These mutations are clustered in three highly conserved regions in the mid portion of the rpoB gene. Cluster I includes codons 505–537, cluster II codons 563–575, and cluster III has codons 684–690.13,16 Efflux pumps or the presence of transferable mechanisms of resistance such as the arr genes have been also involved in resistance to rifamycin derivaties. 1 Thus, a possible co-selection of resistance by the use of unrelated agents cannot be ruled out. Different authors have studied the use of Rfx in the treatment of diarrheal infections, showing different properties such as in vitro or in vivo activity, pharmacokinetic parameters, or the ability to select in vitro resistant mutants.9,10 Nonetheless, although mechanisms of resistance to the rifamycin derivates have been established, to our knowledge no study has been designed to analyze the mechanisms of resistance of Rfx in enterobacteriaceae.
The main aim of this study was to analyze the role of the Phe-Arg-β-naphthylamide (PAβN)-inhibitible efflux pumps, point mutations in the rpoB gene, and alterations in outer membrane porin proteins (Omp) composition in the development of Rfx resistance in in vitro obtained mutants of diarrheogenic E. coli and their effect on the bacterial fitness.
Methods
Bacterial strains
Twenty-eight previously obtained Rfx-resistant mutants 10 and their respective four original parental E. coli strains isolated from patients with TD were used in the study.
Determination of the role of efflux pump
To determine this, the efflux pump inhibitor PAβN was used. Antibacterial susceptibility levels of Rfx and Rfx plus PAβN (20 mg/L) were determined by agar dilution methodology as previously described.10,11 Differences in the minimal inhibitory concentration (MIC) lesser than fourfold between MICs in the presence and absence of PAβN were considered as a methodological error and not an inhibitor effect.
Mutations in the rpoB
Mutations in the rpoB gene were determined by polymerase chain reaction (PCR) amplification of a fragment of 1,057 bp (containing clusters I, II, and III) in all the strains obtained using the following primers and conditions: 5′-AAG CTC ATC GAT ATC CGT AAC G-3′ and 5′-GCA CGT CGC CAC GTT CAA CC primers; 30 cycles at 94°C for 30 sec, 60°C for 30 sec, and 72°C for 30 sec; and final elongation at 72°C for 10 min. The PCR product was recovered using a commercial kit (Wizard SV gel and PCR Clean-up system; Promega, Madison, WI) and sequenced (Macrogen, Seoul, Korea). The sequences were compared with a sequence consensus of GenBank No. CP000800.1. Alignments were made using the Bioedit program.
Porins
The porins were extracted using a methodology previously described. 14 Different expression levels of porins were determined with sodium dodecyl sulfate–urea–polyacrylamide gel electrophoresis 12% with urea 8 M. The electrophoresis conditions were 150 V and 40 mA.
The E. coli strains KAECF5 (OmpC−, OmpF−), KAEC5 (OmpC−, OmpF+), MH621 (OmpC+, OmpF−), and MKM505 (OmpC+, OmpF+) were used as control.
Growth rate measurements
Exponential doubling times were calculated by measuring the increase in optical density at 600 nm, OD600, using a Genequant spectrophotometer. Growth in luria broth (LB) agar medium at 37°C in aerobic atmosphere was followed for 15 hr. The experiments were performed twice.
Results
Mutations in the rpoB gene
No parental strain presented differences with respect to the wild-type GenBank consensus sequence, showing a wild-type sequence.
The analysis of the rpoB gene sequences of the mutant strains showed that 25% (7/28) did not present any mutation in the region analyzed, but a single mutation was found in the remaining 21 strains (75%) (Table 1). Amino acid codon substitutions at position 516 of the β-subunit of RNA polymerases were most frequently obtained (15/28). These 15 strains presented amino acid D516Y, D516N, and D516G changes. Two different amino acid changes (2/28) were found at position 526 (H526N and H526L), and in two isolates (2/28), amino acid change T525R was found. Finally, amino acid changes S512F (1/28) and S574Y (1/28) were found in two different isolates.
I, Phe-Arg-β-naphthylamide, efflux pump inhibitor.
Mutations found in rpoB gene.
Rfx, rifaximin.
Effect of the efflux pumps
PAβN presented an inhibitory effect decreasing the MIC of Rfx from 4- to ≥16-fold in 19 strains (67.9%). In 6 of these 19 strains, no mutations were detected in the rpoB gene, presenting a full reversion to the parental MIC levels in the presence of PAβN (Table 1).
Expression porins
The results did not show any difference between the expression levels of the porins and Omp A, C, and F in parental strains and Rfx mutants (data not shown).
Growth rate measurements
No significant differences were found in the growth measurements between strains (data not shown).
Overall
On analyzing the three mechanisms together, the overexpression of the PAβN-inhibitible efflux pumps in six mutants was able to explain the total resistance acquired. Mutations in the rpoB gene may explain high levels of resistance in eight mutants (28.6%), and the sum effect of at least two mechanisms could explain the resistance in the remaining mutants.
Discussion
The results of the present study suggest an overexpression of the PAβN-inhibitible efflux pumps, which together with target mutations play a relevant role in the development of Rfx resistance, but alterations in the OMPs were absent. However, other possible mechanisms of Rfx resistance such as the presence of efflux pumps non inhibitible by PAβN cannot be ruled out.
The D516Y change has been previously described in rifampicin-resistant Mycobacterium tuberculosis but not in E. coli, and D516N, D516G, and S512F changes have been previously described in E. coli rifampicin-resistant mutants. 15 Although H526 and S574 have been described as point mutations involved in the development of rifampicin resistance, 15 to our knowledge, the specific H526N and S574Y changes have not been reported in the literature. Finally, the previously undescribed T525R point mutation was detected but does not seem to play a relevant role in the development of Rfx resistance, because the MIC of the two strains carrying this change (mutants 7 and 8) reverts toward the parental MIC in the presence of PAβN (Table 1). Interestingly, R525 is present as a wild-type amino acid in Thermus aquaticus, not being involved in the development of resistance to rifamycin. 15 Similarly, strain 13 carrying the amino acid substitution S512F reverts toward parental MIC in the presence of PAβN, suggesting that, despite reported being involved in the development of rifampicin resistance, its role in the development of Rfx resistance is low or null. The selection of this kind of mutations in the rpoB gene may result in the development of stable Rfx-resistant microbial populations. This fact, together with the relatively high facility to select Rfx-resistant strains in vitro, 10 as well as the lack of effect of these alterations on bacterial fitness, suggests the need to perform surveillance of the use of Rfx to detect and limit the selection of stable Rfx-resistant diarrheagenic pathogens. However, it has been shown that the number of Rfx-resistant coliforms and Enterococci isolated was low at baseline (pretreatment) and did not significantly increase at day 3 or 5 posttreatment, indicating no significant development of resistance during this period. 6 This apparent nonconcordance might be explained by the high levels of Rfx that might be achieved in the intestinal lumen after oral treatment, with 4,000–8,000 μg/g of stool being 160–250 times higher than the MIC90 for the various enteropathogens. 12 Although the bioavailability of Rfx has not been established, such high concentrations in the intestine may prevent or strongly limit the development of resistance to Rfx. However, no current studies have been performed on the selection of Rfx-resistant microorganisms after excretion in feces (e.g., sewerage systems). Although amino acid substitutions in the rpoB gene have been related to high levels of resistance to rifampicin, 15 it is of concern that the high levels of Rfx resistance exhibited by these mutants, in some cases, were not exclusively associated with the presence of target mutations, but rather needed concomitant overexpression of the PAβN-inhibitible efflux pumps. Similar results analyzing the mechanisms of resistance to other antimicrobial agents have been observed, in which the apparently high levels of resistance conferred by target alterations are, in fact, due to the association of target alterations with efflux pump systems. 11 On comparing mutations in the rpoB gene and the role of the PAβN-inhibitible efflux pumps with the previously obtained stability of the mutants, it was observed that three of the seven Rfx-resistant mutants lacking mutations in the rpoB gene presented a modest decrease in the MIC levels. This may suggest a certain degree of reversion toward the parental MIC after 20 consecutive cultures in the absence of Rfx, 10 whereas this phenomenon was absent in all isolates in which an amino acid substitution was detected. However, in in vivo situations, the reversion of PAβN-inhibitible efflux pump overexpression may be limited by the effect of bile salts.
Rifampicin plays an important role in clinical practice, as a part of standard therapy of tuberculosis, prophylaxis of meningitis, and against staphylococcal infections. 16 Previous results have shown the presence of cross-resistance between these two agents, Rfx and rifampicin, and accordingly, the present results show similarity between resistance mechanisms in these two agents. Thus, extensive use of Rfx might have a possible negative effect on the activity of rifampicin. Future studies are of special concern in that they refer to the possible use of Rfx as an alternative to antibacterial agents currently in use in developing countries, because of the relevance of tuberculosis in these areas. However, referring to its possible use in treating diarrhea in children, no data are currently available to support the efficacy and safety in persons under 18 years of age. 6
In summary, different mechanisms were found to be involved in the development of Rfx resistance in E. coli strains. The present results show the easy selection of rpoB mutations as well as the concomitant overexpression of PAβN-inhibitible efflux pumps. Further studies with clinical Rfx-resistant isolates are needed to determine the transfer of these results to clinical settings.
Footnotes
Acknowledgments
The strains used as control were kindly provided by L. Martínez, Hospital Universitario Marqués de Valdecilla. This study was supported by the grant CP05/0130 from the FIS (Fondo de Investigaciones Sanitarias, Spain) to J.R. and by Laboratorios Bama-Geve.
Disclosure Statement
The authors do not have any commercial associations that might create a conflict of interest in connection with this manuscript. This study was partially funded by Laboratories Bama-Geve. The sponsor of the study had no role in study design, data collection, data interpretation, or the writing of the report. J.R. had full access to all the data in the study and was responsible for the decision to submit for publication. We declare that we have no conflict of interest.