Infection with Campylobacter jejuni is now considered to be the most common cause of acute bacterial gastroenteritis in humans worldwide. It occurs more frequently than infections caused by Salmonella species, Shigella species, or Escherichia coli O157:H7. Although C. jejuni is also recognized for its association with serious post-infection neurological complications, most patients with C. jejuni infections have a self-limited illness. Nevertheless, a substantial proportion of these infections are treated with antibiotics. These include severe and prolonged cases of enteritis, infections in immune-suppressed patients, septicaemia and other extra-intestinal infections. Under these circumstances, erythromycin is often recommended as the drug of first choice. However, erythromycin-resistant Campylobacter have emerged during therapy with macrolides. Moreover, the widespread use of macrolides, including erythromycin, in veterinary medicine has accelerated this resistance trend. Several countries including Canada, Japan and Finland have reported C. jejuni isolates with low and stable rates of macrolide resistance. In contrast, the increasing level of macrolide resistance in C. jejuni is becoming a major public health concern in other parts of the world such as the United States, Europe and Taiwan. Macrolide resistance in Campylobacter is mainly associated with point mutation(s) occurring in the peptidyl-encoding region in domain V of the 23S rRNA gene, the target of macrolides. Several rapid and practical techniques have recently been developed for the identification of macrolide-resistant isolates of C. jejuni. The aim of this mini-review is to give an overview of the worldwide distribution of macrolide resistance in C. jejuni and Campylobacter coli as well as its possible association with the massive use of these agents in food animals. Mechanisms implicated in macrolide resistance in C. jejuni and also techniques that have been developed for the efficient detection of macrolide-associated mutation(s) will be discussed in detail.
Campylobacter species are among the most frequently identified bacterial causes of human gastroenteritis in the United States and other industrialized countries.1 Annually, approximately 400 million cases of Campylobacter-associated gastroenteritis occur worldwide,2,3 with 2.5 million cases in the United States.4 The true public incidence, due to under-reporting, is estimated to be up to 10 times higher than the documented case numbers.5,6 As a result of campylobacteriosis, substantial economic losses are documented annually because of clinical treatment costs and lost working hours.7
In general, the occurrence of human Campylobacter gastroenteritis has been largely attributed to the consumption of contaminated food animal products, especially poultry, because of the high prevalence of Campylobacter in these animals.8,9 Other vehicles such as red meat, environmental water and unpasteurized milk are additional sources of infection.10–12 Person-to-person transmission is very rare.13
The most important Campylobacter species is Campylobacter jejuni, accounting for over 90% of infections,14 with Campylobacter coli accounting for most of the remaining infections.15,16 In addition to gastroenteritis, C. jejuni can cause post-infectious manifestations, including Guillain–Barré syndrome, an acute, immune-mediated disorder considered to be the most serious secondary complication.17 Furthermore, life-threatening systemic Campylobacter diseases are diagnosed more and more frequently.17
Most cases of enteritis do not require treatment as they are of short duration, clinically mild and self limiting. Antimicrobial treatment is, however, necessary for systemic Campylobacter infections, infections in immune-suppressed patients and severe or long-lasting infections.5,18 Erythromycin is considered the drug of choice for treating Campylobacter gastroenteritis, and ciprofloxacin and tetracycline are used as alternative drugs.19,20 The extensive development of resistance to tetracycline and ciprofloxacin in various countries has led to a decrease in their clinical use.21 In addition, the increased emergence of erythromycin resistance among isolates of C. jejuni and C. coli has prompted a search among newer macrolide derivatives for those useful against Campylobacter isolates. Clarithromycin is not generally prescribed for the treatment of Campylobacter infections, probably because the MIC90s of clarithromycin are 2-fold higher than those for erythromycin as observed by Hardy et al.22 Several studies have reported the in vitro activity of azithromycin against Campylobacter species;22–24 however, azithromycin did not exhibit increased potency in comparison with erythromycin.25 Although the ketolides exhibit an improved profile and more importantly show significant activity against some macrolide-resistant Gram-positive organisms,26–28 it was recently found that erythromycin-resistant Campylobacter isolates showed similar resistance patterns for the ketolide, telithromycin.29 Serious systemic infections may also be treated with an aminoglycoside such as gentamicin.30
Isolates of C. jejuni and C. coli with resistance to various antimicrobial agents have been reported in both developed and developing countries.31 Since the 1990s, a significant increase in the prevalence of resistance to macrolides among Campylobacter spp. has been reported, and this is recognized as an emerging public health problem.32 It has been suggested by some investigators that resistance to macrolides is mainly found in isolates of animal origin, especially C. coli from pigs and also C. jejuni from chickens.3,33,34 Furthermore, macrolide resistance may develop during the course of antibiotic treatment in humans.35 It also seems possible that the selection pressure arising from the use of macrolides in human medicine might also affect resistance in Campylobacter,36 probably by enhancing the selection of macrolide-resistant variants from susceptible isolates of Campylobacter.
The impact of the prevalence of macrolide resistance among clinical isolates of C. jejuni and C. coli
Although there is increasing evidence of adverse events associated with antimicrobial drug resistance in Salmonella infections,37–39 there is limited information on the clinical consequences of resistance in the case of Campylobacter. A major methodological problem is that the number of well-defined, severe and relevant outcomes is found to be limited in most studies. However, a number of investigations from the United States, Thailand and Denmark have shown that infections with macrolide-resistant Campylobacter isolates could be associated with an increased risk of adverse events, development of invasive illness or death compared with infections with drug-susceptible isolates.40–42 The adverse events may include reduced efficacy of treatment and development of post-infectious manifestations such as Guillain–Barré syndrome or other severe reactive illnesses.40–42 Data supporting an increase in virulence of drug-resistant isolates of C. jejuni are beginning to emerge.41,43 The underlying mechanisms could be co-selection of virulence traits, up-regulation of virulence or improved fitness of the resistant isolates.41,43 In a recent study,42 it was suggested that macrolide resistance in clinical isolates of C. jejuni could be associated with some unknown virulence markers, and consequently the use of macrolides in any ecosystem may select for such isolates. As a result, macrolide resistance in Campylobacter isolates could be of public health importance even in the absence of evidence of therapeutic failure.
Incidence of macrolide resistance in C. jejuni and C. coli
Increased macrolide resistance among C. jejuni and C. coli has been reported in both developed and developing countries, but the situation seems to be deteriorating more rapidly in developing countries.31 The incidence of macrolide resistance among C. jejuni and C. coli is highly variable with respect to the country of isolation. Differences may well be related to the source of Campylobacter isolates and the frequency and type of antimicrobial agents used as feed additives or for treating animal and human infections in different geographical areas.32 While the rate of macrolide resistance in clinical isolates of Campylobacter is not yet alarming in most of the developed countries,32 a trend towards an increase has been documented both in animal and human isolates during macrolide treatment. In both human and animal isolates, nearly all studies report a higher frequency of macrolide resistance in C. coli than in C. jejuni.32,44
Resistance to macrolides in human isolates is now becoming a major public health concern. Table 1 shows the rate of erythromycin resistance among human isolates of C. jejuni and C. coli in countries where the increasing rates of resistance are becoming a public health concern. However, trends over time for macrolide resistance show stable and low rates in countries including Finland, Japan and Sweden.32,45
|Erythromycin-resistance rate (%) in|
|Location||C. jejuni||C. coli||C. jejuni and C. coli||Reference(s)|
|Erythromycin-resistance rate (%) in|
|Location||C. jejuni||C. coli||C. jejuni and C. coli||Reference(s)|
aThe majority of the strains (70%) isolated from children.
bAll strains isolated from children.
cIsolates acquired domestically.
dIsolates acquired abroad.
In the case of animal isolates of Campylobacter, several investigators have found high rates of macrolide resistance among C. coli isolates from swine18,46,47 which may be related to the extensive veterinary use of macrolides.32,47 Thus, pigs are likely to be a major source for infections with erythromycin-resistant C. coli.17,44 During a recent nationwide monitoring programme of antimicrobial resistance in Campylobacter isolated from food-producing animals in Japan,48 all C. jejuni isolates were susceptible to macrolides, whereas 48.4% of C. coli isolates were resistant. European antimicrobial susceptibility surveillance in animals coordinated by the European Animal Health Study Centre reported that the frequency of erythromycin resistance was variable, depending on the source of animal isolates of C. jejuni and C. coli.49 Resistance was highest in pig isolates from Denmark (36.7%) and the Netherlands (41.8%).49 Resistance was low in isolates from chickens (0–8%) and cattle (3.2–5.6%) in all countries sampled and in isolates from pigs from Sweden (9 %).49
The use of macrolides in veterinary medicine
Several antimicrobial agents including macrolides have a variety of therapeutic and preventive applications in food animals.50 Although macrolides have generally been used as second-line antibiotics against Gram-positive bacteria, they may, under specific conditions such as pneumonitis and mastitis, be of particular value because of their propensity to achieve high tissue concentrations.51,52 Several antimicrobial agents frequently used in cattle populations may select for erythromycin resistance in these herds, including spiramycin and erythromycin (used to treat bovine mastitis)32 or tylosin (used to prevent the formation of hepatic abscesses in feedlot cattle).53 Generally, tetracyclines, macrolides and pleuromutilins are frequently used in pigs for stabilization of the gut flora during the weaning phase.50 For the past 20 years, tylosin has also been the most commonly used agent for growth promotion in swine production worldwide, whereas spiramycin has been commonly used in poultry.32 Some antimicrobial animal-treatment practices may exert greater selective pressures for resistance than others. For example, the use of macrolide derivatives as growth promoters, which entails exposing bacteria to sub-lethal concentrations over long periods, would appear more conducive to selecting and maintaining resistant organisms.54 The use of antimicrobial agents, including macrolides, in food animals creates selective pressure for the emergence and dissemination of resistance among animal pathogens, commensal bacteria that are present in food-producing animals and human pathogens that have food animal reservoirs such as Salmonella and Campylobacter.55 Macrolide-resistant isolates of C. jejuni and C. coli from animals can infect or reach the human population not only by direct contact but also via food products of animal origin. The impact of the use of antimicrobials in food animals on human medicine could be minimized by reducing the potential for resistant enteric bacteria, selected during treatment, to move up the food chain.
Clinical use of macrolides in human medicine
Erythromycin and other macrolides have enjoyed a renaissance in the 1970s, 1980s and 1990s following the discovery of ‘new’ pathogens such as Chlamydia, Legionella, Campylobacter and Mycoplasma spp. The efficacy of macrolides against Gram-negative bacteria is increased considerably when they are used in an environment of alkaline pH.56
The spectrum of activity of macrolides covers mainly Gram-positive cocci and bacilli.57 Macrolides show in vitro activity against some of Gram-negative organisms, including Campylobacter spp., Haemophilus spp., Bordetella pertussis, Pasteurella multocida, Neisseria gonorrhoeae, some intracellular and atypical organisms including Chlamydia spp., Ureaplasma spp. and Leptospira spp. in addition to some of the Mycobacteria spp.58 Telithromycin, which belongs to a new class of macrolides named ketolides, is the most active macrolide against Enterococcus faecalis.59 Telithromycin also displays good activity against intracellular pathogens such as Rickettsia spp. and Bartonella spp.59
Macrolides exhibit proven efficacy for a wide range of infections including upper and lower respiratory tract infections,60 skin infections,61 prophylaxis of endocarditis, acute rheumatic fever, ophthalmia neonatorum and pre-colonic surgery,62 campylobacteriosis,58 chlamydial and ureaplasmal infections,58 whooping cough, chronic bronchitis,63 community-acquired pneumonia,63 acute maxillary sinusitis63 and streptococcal pharyngitis.59 Azithromycin is approved for single-dose therapy of sexually transmitted diseases caused by Chlamydia trachomatis and Haemophilus ducreyi.60 The ketolides exhibit a pronounced concentration-dependent killing, a significant post-antibiotic effect and a reduction of intracellular Helicobacter pylori.59
Mechanism of action of macrolides and modes of resistance in other bacterial species
The mechanism of action of erythromycin (typical of the macrolide group) is the inhibition of protein synthesis via interference with the translocation step (Figure 1).64,65 During protein synthesis, the newly synthesized peptide chain passes through the 50S subunit tunnel and runs from the peptidyl transferase centre to emerge at the back of the ribosome.65 The narrowest portion of the tunnel is at a bend where the β-hairpin elongation from r-proteins L4 and L22 merge with the rRNA structures.65 This is the region where the macrolide antibiotics, including erythromycin, bind and make several common contacts around 23S rRNA nucleotide A2058.65 Macrolides block the entrance to the tunnel in the large ribosomal subunit66–68 inducing premature dissociation of the peptidyl-tRNAs from the ribosome.69,70 Such ‘drop-off’ events occur just after initiation of protein synthesis, when the nascent polypeptide chain is short.71 Different macrolides allow formation of peptides with different lengths depending on the space available between the macrolide and the peptidyl transferase centre.71 The effect is inhibition of the translocation of the developing peptide chain from the acceptor or A site to the donor or P site, which is required for the elongation of the peptide chain as the ribosome moves along the messenger RNA strand. The mode of action, although not identical, is probably similar for all macrolide antibiotics.72
Generally, resistance to macrolides has been reported to be associated with the following mechanisms:
Resistance attributed to target modification
Target modification mediated by rRNA methylases
The rRNA methylases can transform an adenine residue at position 2058 (E. coli-equivalent) of the 23S rRNA into either N6-monomethyladenine or N6,N6-dimethyladenine.73 As a consequence of methylation, binding of erythromycin to its target is impaired.65 Mono-methylation confers the so-called MLSB type I phenotype,74,75 whereas Erm dimethyltransferases confer the MLSB type II phenotype65,74 which is the more common resistance mechanism in bacterial pathogens. At least 22 different classes of erm genes have been identified in both Gram-positive and Gram-negative bacteria76 and the erm genes can be located on plasmids or transposons.73 Differences between various erm genes are seen in the regulation of the expression of the phenotype.76 A summary of the genera in which different erm genes have so far been detected has recently been published.76
Target modification mediated by specific point mutation in the 23S rDNA
In general, specific point mutation(s) within the peptidyl-transferase region in domain V of the 23S rRNA were initially observed in pathogens with one or two rrn copies, often with each copy carrying the mutation.77 More recently, macrolide-associated mutations have been identified in Streptococcus pneumoniae, which harbours four rrn copies,78 and they were also identified in bacteria harbouring up to six rDNA operons.79,80 One of several mutations such as A2058G, A2058C or U, A2059G and T2611C (E. coli numbering) is usually associated with macrolide resistance in a few genera of bacteria.73 More modest resistance limited to certain macrolide subgroups was found to be associated with mutations at or close to nucleotide 752.81,82
Target modification mediated by alteration of the 50S ribosomal proteins
Macrolide resistance may involve the alteration of the 50S ribosomal subunit proteins L4 and L22.83–85 This mechanism of macrolide resistance has been documented in laboratory-derived and clinical isolates of S. pneumoniae.78 Cryo-electron microscopy revealed that the r-protein mutations are invariably at the ends of the L4 and L22 hairpin structures situated close to the MLSB site.86 Mutations in the r-protein L22 were found to induce a local change in the L22 hairpin,87 which increases the tunnel width at the MLSB site to allow the passage of the nascent peptide without reducing the binding affinity for erythromycin.84,88
Resistance attributed to enzymatic inactivation of macrolides
Macrolide-inactivating enzymes have been detected in clinical and macrolide-producing strains and include enzymes that modify macrolides via phosporylation89–91 or glycosylation92,93 of the 2′-OH in the desosamine moiety. High-level resistance to erythromycin in members of the family Enterobacteriaceae has been shown to be due to macrolide-inactivating enzymes of two types: erythromycin esterases (EreA and EreB) and macrolide 2′-phosphotransferases.91,94 Phosphotransferases such as those encoded by the genes mph(A) and mph(B) have been identified in E. coli.95 Other macrolide phosphotransferases encoded by mph(C) genes have also been detected in clinical isolates of Staphylococcus aureus, although only a few resistant isolates have been reported to date.96 An esterase produced by a clinical strain of Staphylococcus haemolyticus is also believed to confer resistance to macrolides such as erythromycin, clarithromycin and spiramycin.95 Most of the genes for inactivating enzymes are associated with plasmids.95
Resistance attributed to enhanced efflux of macrolides
In Gram-negative bacteria, chromosomally encoded efflux pumps contribute to intrinsic resistance to a variety of antimicrobial agents including macrolides.77 These pumps often belong to the resistance nodulation cell division (RND) family.77 In Gram-positive organisms, two classes of efflux pumps are implicated in acquired macrolide resistance: members of the ATP-binding-cassette (ABC) transporter superfamily and the major facilitator superfamily (MFS).76 The major facilitator superfamily is exemplified by mef genes [mef(A) and mef(E)] which have been found in a variety of Gram-positive genera,97–99 suggesting a much wider distribution of this group of genes. Many of these genes are associated with conjugative elements located in the chromosome and readily transferred conjugally across species and genus barriers.97,98 The second type of efflux systems in Gram-positive organisms, the ABC transporter superfamily, is encoded by the msr(A), msr(SA), msr(SA)′ and msr(B) genes.96,100,101 This group of efflux genes differs from the mef genes because they confer resistance to both macrolides and streptogramin B antibiotics96,100,101 and they are located on plasmids.95
In pathogenic microorganisms, the impact of the above-mentioned mechanisms is unequal in terms of incidence and of clinical implications. Modification of the ribosomal target confers broad-spectrum resistance to macrolides and lincosamides, whereas efflux and inactivation affect only some of these molecules. Although not all mechanisms of macrolide resistance have been found in all bacterial species, it is clear that reservoirs of these resistance markers are present and that a horizontal spread of the different types of resistance markers can occur in nature.102 There is full cross-resistance among all available macrolides, lincosamides and group B streptogramins, giving rise to the so-called phenotype of MLSB resistance.103 This cross-resistance to the structurally related groups of antibiotics is explained by their known overlapping binding sites on 50S ribosomal subunits.104
Mechanisms of macrolide resistance in C. jejuni and C. coli
Resistance due to ribosomal mutation
To date, high-level macrolide resistance in C. jejuni and C. coli has mainly been attributed to mutations in domain V of the 23S rRNA target gene at positions 2074 and 2075 (corresponding to positions 2058 and 2059 in E. coli).105–108Figure 2 illustrates the locations of the erythromycin-associated mutations in the peptidyl transferase loop in domain V of the C. jejuni/coli 23S rRNA. The mutations at these positions alter the main anchoring point for macrolides109 and tend to confer resistance, presumably, by disturbing the binding of the drug to its inhibitory site on the bacterial ribosome.110–112 In the majority of macrolide-resistant C. jejuni isolates, the resistance-associated mutations were identified in the three copies of the target gene; however, at least two mutated copies are necessary to confer macrolide resistance.113 The predominant mutation detected among C. coli and C. jejuni was the transition mutation A2075G, which is usually associated with high-level resistance to erythromycin.105–108 The apparently high frequency of the A2075G mutation among Campylobacter isolates is possibly attributed to the biological features generated by this mutation, so that A2075G mutation in C. jejuni/coli has a survival advantage over the other mutations.113 An A2074G transition has been found so far in only one erythromycin-resistant C. jejuni isolate.113 This mutation was found to have a negative effect on the growth rate of its host which may explain its rare occurrence among C. jejuni isolates.113 This mutation was also observed to be relatively unstable in the absence of erythromycin-selection pressure.113 Another macrolide-associated mutation (A2074C transversion) occurs at relatively low frequency in resistant C. jejuni isolates only.108,113 An A2074C transversion may lead to a minor alteration in the ribosome structure, with minimal effect on the growth rate of the resistant isolates, which could explain the low frequency of this kind of mutation.113 In only two cases have macrolide-resistant isolates of C. jejuni carried both the A2075G and A2074C mutations.108,114 Although we previously reported that macrolide resistance in C. jejuni isolate UA710 was not associated with any observed mutation in the 23S rRNA gene,113 this now appears to be incorrect, as the observed macrolide-resistance phenotype was due to contamination of this isolate with a macrolide-resistant Gram-positive organism. Additional testing of the isolate UA710 has shown it to be susceptible to both erythromycin and clarithromycin.
In general, no correlation was observed between the level of macrolide resistance in Campylobacter isolates (erythromycin MIC values range from 128 to >1024 mg/L) and the number of mutated copies of the 23S rRNA gene or the site of the mutation.105,113,115 The erythromycin-resistance phenotype in C. jejuni and C. coli, due to a modified target gene is usually stable, regardless of the level of resistance or the number of the mutated copies of the 23S rRNA gene.113 Members of the ketolide family of macrolides such as telithromycin may be less affected by the A2075G mutation in the 23S rRNA gene than the other related molecules such as erythromycin, azithromycin and tylosin.116 In spite of the presence of the target point mutation, MICs of telithromycin are usually lower (32–128 mg/L) than the corresponding MICs of erythromycin (1024–2048 mg/L).116 This enhanced activity of telithromycin is probably due to a higher binding affinity to the 50S ribosomal subunit,82,117 largely because of the alkyl-aryl substituent extending from the macrolactone ring.82,117 Therefore, telithromycin represents a new generation of antimicrobials that have generally been developed with a view to overcoming the problem of macrolide resistance.
Resistance due to efflux of macrolides
Efflux was first postulated as a mechanism of multiple-antibiotic resistance in Campylobacter spp. some 10 years ago.118 Based on bioinformatic data, 10 putative efflux pumps in addition to the CmeABC system have been identified in C. jejuni.119 The CmeABC system has been well characterized as the major multi-drug efflux pump system in Campylobacter.120–123 It extrudes a wide variety of compounds such as dyes, detergents and antimicrobial agents of various families including quinolones, β-lactams and macrolides.120–122,124 The contribution of the CmeABC system to intrinsic macrolide resistance in Campylobacter was first demonstrated by the mutagenesis of the pump which rendered the susceptible strains 81–176 and NCTC11168 hyper-susceptible to macrolides.120,121 In the case of strains 81–176 and NCTC11168, inactivation of the CmeABC system caused a decrease in the MIC of erythromycin from 0.078 to 0.02 mg/L120 and from 0.5 to 0.25 mg/L,121 respectively.
The CmeABC efflux pump shares significant homology with typical multi-drug efflux pumps belonging to the RND superfamily in Gram-negative bacteria.119 It consists of three components (Figure 3): a periplasmic fusion protein, CmeA; an inner membrane drug transporter, CmeB; and an outer membrane protein, CmeC.120 These three proteins function together to form a membrane channel that extrudes toxic substances, including several antimicrobial agents, directly out of the Campylobacter cells.120 The cmeR gene, which is located upstream of the cmeA gene, encodes a protein which shares sequence and structural similarities with members of the TetR family of transcriptional repressors.125 The CmeR protein represses the transcription of the cmeABC operon by binding directly to the promoter region (specifically to the inverted repeat18) of the efflux operon.125 The CmeR protein acts as a modulator in Campylobacter to maintain a basal level of the CmeABC efflux pump to meet the physiological needs and facilitate the adaptation of Campylobacter to environmental changes, including antibiotic exposure.125–127
The presence of various antimicrobial compounds such as ciprofloxacin, norfloxacin, tetracycline, erythromycin, chloramphenicol, rifampicin and cefotaxime does not cause any induction of the expression of the CmeABC efflux pump.128 Bile salts, which occur naturally in the animal intestinal tract, play a major role in the induction of the expression of the cmeABC operon.128 The action of bile salts is based on inhibiting the binding of the CmeR repressor protein to the promoter of the efflux operon suggesting that bile salts are inducing ligands of CmeR.128 Based on the structural similarity between the CmeR protein and homologous regulator proteins, such as TetR and QacR,129,130 it is likely that the interaction of bile salts with the C-terminal region of the repressor protein induces conformational changes in the repressor resulting in a considerable reduction in its binding affinity.128 Some bile salts such as taurocholate activate the expression of the cmeABC operon via both CmeR and CmeR-independent pathways.128 It is unclear, however, which regulator is involved in the CmeR-independent activation pathway.
A number of studies demonstrated the involvement of CmeABC efflux pump in both intrinsic and acquired resistance to erythromycin, in both C. jejuni and C. coli, mostly by the use of the efflux pump inhibitor (EPI), phenylalanine-arginine β-naphthylamide (PAβN).29,115,131 PAβN was discovered as such by empirical screening of a small molecule library in Pseudomonas aeruginosa.132 PAβN meets the following criteria which indicate that it is acting as an EPI. First, PAβN enhances the activity of levofloxacin and other antibiotics that are transported by efflux in P. aeruginosa strains containing functioning pumps.133 Second, PAβN does not significantly potentiate the activities of antibiotics in strains that lack efflux pumps.133 In addition, PAβN increases the level of accumulation and decreases the level of extrusion of efflux pump substrates in accumulation and efflux assays.133 PAβN is now used as a broad-spectrum EPI in a variety of Gram-negative bacteria.133
The effect of PAβN on macrolide resistance in C. jejuni and C. coli was reported to be variable depending on the level of resistance as well as the concentration of the inhibitor incorporated.29,116 Erythromycin-resistant Campylobacter isolates could be divided into two groups: low level (LLR; erythromycin MICs range from 8 to 16 mg/L with no mutation in the target gene) and high level (HLR; erythromycin MICs over 128 mg/L in which a point mutation was always detected in the 23S rRNA gene).115 A number of studies have showed that PAβN restored erythromycin susceptibility of the LLR isolates, which did not carry any mutation in the 23S rRNA gene, regardless of the concentration of PAβN used.29,107,116 In the case of HLR isolates, the use of a low concentration of PAβN (20 mg/L) caused no change in erythromycin-resistance level,29 whereas a higher concentration of the inhibitor (40 mg/L) could lead to 2- to 4-fold decrease in erythromycin-resistance level in some HLR isolates,115 in spite of the presence of the target gene mutation. It is worth mentioning that several investigators reported that the use of 40 mg/L of PAβN inhibited the growth of many HLR Campylobacter strains.29,113 This suggests that the effect of PAβN on the growth of Campylobacter could be isolate dependent, necessitating careful optimization of the inhibitor concentration when used in similar investigations.
Insertional inactivation of the cmeB gene (or addition of PAβN) was found to have a variable effect on the MICs of different macrolides.116 The MICs of tylosin and telithromycin usually decrease from 32 to 1 mg/L and from 32–16 to 0.5–0.25 mg/L, respectively (32-fold and 32- to 64-fold) upon CmeABC inactivation in LLR isolates compared with 4- to 8-fold decrease in the MICs of azithromycin (from 1–2 to 0.25 mg/L).116 This indicates that tylosin and telithromycin appear to be better substrates for this pump than azithromycin.116 This difference was also observed for HLR isolates since MICs of azithromycin remained unchanged (2048 mg/L) in cmeB-knockout mutants, whereas MICs of tylosin and telithromycin greatly decreased by more than 128-fold (from 2048 to 16–8 mg/L and from 128–32 to 0.5–0.25 mg/L, respectively).116
Lin et al.125 obtained a spontaneous multidrug-resistant mutant of strain 81–176 by stepwise selection on ciprofloxacin-containing plates. This mutant showed a single nucleotide deletion between the two half-sites of the inverted repeats in the promoter region of the cmeABC operon (Figure 3) resulting in a reduction in the level of CmeR binding to the promoter sequence and a consequent over-expression of the CmeABC efflux system.125 As a result, the mutant exhibited an increase in the level of resistance to several antimicrobial agents including tetracycline (from 0.098 to 0.196 mg/L), ampicillin (from 0.625 to 2.5 mg/L), cefotaxime (from 0.390 to 1.56 mg/L), erythromycin (from 0.039 to 0.156 mg/L) and fusidic acid (from 39 to 78 mg/L).125 In another study,123 sequence analysis of the regulator cmeR gene of the laboratory-selected mutant that over-expressed cmeB gene revealed two mutations, one leading to the substitution of glutamine 9 with proline and the other substituting glycine 86 with alanine. These substitutions were at conserved residues predicted on the basis of the E. coli AcrR structure and within the substrate binding site.123 The contribution of CmeABC over-expression to acquired antibiotic resistance in Campylobacter isolates, however, awaits further investigation.
Recently, it was observed that in a cmeB− background, PAβN was able to increase telithromycin susceptibility suggesting that another efflux system was still active.29 This additional efflux system is unlikely to be the CmeDEF system because Pumbwe et al.134 found no correlation between the expression of CmeDEF and macrolide resistance in Campylobacter isolates. This suggests that a third efflux system could play a role in macrolide resistance in Campylobacter isolates.29
The role of other mechanisms of resistance
Although previous work in E. coli and other organisms has implicated the ribosomal proteins L4 and L22 in macrolide resistance,74,77 there is no evidence of the contribution of these ribosomal proteins to the high-level macrolide resistance in C. jejuni and C. coli.113,131 Similarly, no rRNA modifying enzymes or macrolide-inactivating enzymes have been described to date in Campylobacter spp.34,113,115
In general, the interplay between 23S rRNA mutations and efflux in the development of macrolide resistance in Campylobacter remains unclear, requiring further evaluation. Many points must be satisfied with the aim of improving our understanding of the mechanism of macrolide resistance in Campylobacter. An important question which needs to be addressed is the predominance of the A2059G mutations among resistant Campylobacter isolates. Also, more data regarding the stability of macrolide resistance, especially among clinical isolates, is potentially important from the clinical point of view. Limited data revealed that telithromycin is less affected by an A2059G mutation in the 23S rRNA gene compared with other related macrolides;116 however, more studies are needed to confirm this point. Interestingly, resistance to erythromycin has been reported in some Campylobacter isolates in the absence of either target gene modification or an over-expressed efflux pump,135 suggesting that novel mechanisms of resistance to macrolides in Campylobacter need to be explored. The role of bile-induced expression of the cmeABC operon in the emergence of antibiotic-resistant isolates of Campylobacter in vivo needs to be investigated. Deciphering antibiotic transport through the CmeABC efflux system and also detection of the drug binding sites might offer new clues to block the production of drug transporter. In addition, further information about the affinity between the substrates of the CmeABC efflux pump and efflux inhibitors could be of major interest for the rational design of antibiotics that are not recognized by the efflux pump and also for the design of an efficient EPI. While reports have demonstrated the benefits of the EPI approach in combating macrolide resistance in some isolates of Campylobacter, more data are needed before inhibitors could be used clinically.
Techniques used for detection of the target site alteration associated with macrolide resistance in Campylobacter isolates
Antibiotic resistance in Campylobacter isolates from food animals is now recognized as an emerging public health concern.32 Campylobacteriosis is primarily a zoonosis;4,136 therefore, the study and monitoring of macrolide resistance in Campylobacter spp. has become necessary. The availability of an efficient and reliable method for screening the macrolide-associated mutations is also of importance in epidemiological studies of antibiotic resistance.
In routine clinical practice, the detection of macrolide resistance in Campylobacter isolates is mainly based on phenotypic methods performed after culture and isolation: agar dilution method or agar diffusion using the Etest. These methods are generally time-consuming and costly. In addition, it is difficult to compare susceptibility testing results from various studies due to a large number of testing variables as well as to the different interpretive criteria used. There are three erythromycin-resistance breakpoints used by researchers working in the field: ≥2, >4 and ≥8 mg/L as recommended by BSAC,137 CA-SFM138 and CLSI (formerly known as NCCLS),139 respectively. This leads to some variability regarding the definition of erythromycin LLR and HLR isolates of C. jejuni and C. coli reported in the literature. As a result, the perception of the relative contribution of the 23S rRNA mutations and efflux pumps to macrolide resistance in C. jejuni and C. coli is likely to differ among various research groups. Since the phenotypic methods do not identify the type of macrolide-associated mutation present, which may be interesting from an epidemiological point of view, detection of point mutations conferring resistance to macrolides by molecular methods constitutes a more up-to-date approach. Another advantage of using the genetic-based assays includes the possibility of direct detection from a sample eliminating the need for culture.140,141 Molecular methods can also facilitate analysis of organisms that may be sub-lethally damaged and difficult to grow.140,141 These methods also offer the possibility of screening large numbers of isolates for a specific mutation within a single assay.140,141 The disadvantages of using molecular detection methods include the failure to detect resistance if a new, unexpected or rare resistance mechanism is present.140,141 Another disadvantage of molecular detection methods is that they cannot detect the resistance mediated by efflux mechanisms. Therefore, a combination of phenotypic and genotypic methods for resistance characterization would be optimal.
Of the molecular methods, DNA sequencing remains the ‘gold standard’ for identifying macrolide-associated mutations, not only for Campylobacter but also for other microorganisms. DNA sequencing has recently become much cheaper and faster by virtue of automation.142 Instruments are now available for semi-automated running and analysing of sequencing gels.141 This technology is, however, not always available for routine microbiology laboratories, and therefore alternative protocols have been recently developed. During the last few years, several PCR-based techniques have been developed to detect macrolide-associated mutations, including PCR–restriction fragment length polymorphism (PCR–RFLP),108 PCR–line probe assay (PCR–LIPA),106 real-time PCR,143 pyrosequencing144 and rapid mismatch amplification mutation assay (MAMA).145
PCR–RFLP is based on performing a combination of PCR and RFLP analysis using BasI and BceAI restriction enzymes.108 Erythromycin-resistant Campylobacter isolates can be easily distinguished from the susceptible ones by the pattern characteristic of the specific substitution.108
PCR–LIPA is a genetic method that can recognize single nucleotide changes. It is based on PCR and reverse hybridization to a number of oligonucleotide probes immobilized as parallel lines on a nylon membrane strip.106 This technique was used for the detection of rifampicin-resistant strains of Mycobacterium tuberculosis and clarithromycin-resistant strains of H. pylori.146,147
The real-time PCR assay for detection of macrolide-associated mutations is based on the amplification of a fragment of the 23S rRNA gene surrounding bases 2074 and 2075.114 The amplification is performed using a PCR mixture that incorporates the DNA double-strand specific fluorophore SYBR Green 1.148 Melting curve analysis takes place after all thermocycling is complete, and the mutation is detected with a probe labelled at the 5′ end with an acceptor fluorodye.148 Hybridization of the probe to the target sequence leads to an increase in fluorescence from the acceptor fluorodye as a result of fluorescent resonance energy transfer between the donor and the acceptor fluorophores.148 When there are mismatched bases between the probe and the target, the melting peak occurs at a temperature lower than that of a perfectly matched hybrid. This assay was able to detect the correct genotype of erythromycin-resistant Campylobacter isolates in the majority of cases.114 It is more sensitive and more rapid than other PCR assays; the entire procedure requires less than 2 h.148
In pyrosequencing, a sequencing primer is annealed to a single-stranded PCR product and nucleotides are added to the reaction.149 Incorporation of the nucleotides by DNA polymerase leads to the release of pyrophosphate, which is further processed by sulphurylase and luciferase, producing light in proportion to the amount of pyrophosphate.149 The light is detected by the instrument and presented as a pyrogram, in which the peak heights are proportional to the number of nucleotides incorporated.149 Excess nucleotides are enzymatically degraded before the next nucleotide is added.149 Wild-type and mutated strains are readily distinguished, and the number of mutated alleles can be estimated by interpreting the pyrogram and analysing the numerical peak heights.144 The pyrosequencing technique has several advantages over other techniques discussed previously including lower cost, and the ease of both performing the sequencing without gels or capillaries and analysing the data.149 The entire procedure can be performed in a single working day.144 In addition, parallel processing of a large number of samples can easily be envisioned with the use of high-density microtitre plates.149
The MAMA makes use of the principle that short oligonucleotides with a perfect match at their 3′ ends, complementary to the mutation to be detected, will initiate the polymerization by Taq polymerase far more efficiently than primers with a single mismatch in this position.150 For the determination of macrolide-associated mutations in Campylobacter, a conserved forward primer and a reverse mutation detection primer are used to generate a PCR product that will be a positive indication of the presence of the A2075G or/and A2074C mutation.145 MAMA–PCR is a simple and rapid method for the detection of point mutations in the 23S rDNA, overcoming the need for DNA sequencing.145 This protocol could serve as a portable alternative to methods such as PCR–LIPA106 and PCR–RFLP108 because it is faster and cheaper.
Other molecular methods have recently been developed for the rapid identification of the genotype associated with macrolide resistance in other pathogens including H. pylori and Gram-positive organisms.151–153 These methods include high-performance liquid chromatography,151 hybridization in liquid phase,153 PCR/DNA enzyme immunoassay,152 single-stranded chain polymorphism,154 PCR amplification of specific alleles155 and denaturing gradient gel electrophoresis.156 Although these methods have not yet been applied for characterization of macrolide resistance in Campylobacter isolates, they may well be useful in this regard.
Campylobacter is currently the leading cause of bacterial diarrhoea in the developed world and therefore presents a significant challenge to public health. There is accumulating evidence of the clinical and public health consequences of macrolide resistance in C. jejuni and C. coli. Infection with macrolide-resistant Campylobacter strains is associated with increased risk of invasive illness or death, compared with infection with drug-susceptible strains. Macrolide resistance in Campylobacter species is mainly a consequence of the use of antimicrobials in food animal production, emphasizing the need for the limitation of the use of antimicrobial drugs in agriculture. Campylobacter species have a natural ability for acquiring heterologous DNA. Therefore, the potential for interspecies exchange of genetic material in reservoirs of mixed bacterial flora, such as the intestines, emphasizes the need for surveillance of resistance rates and elucidation of transmission processes. To maintain up-to-date treatment regimens and to understand the transmission dynamics of macrolide-resistant Campylobacter isolates, global antimicrobial susceptibility surveillance studies need to be implemented and maintained. The improvement of safety and the prudent use of antimicrobial agents are the two cornerstones in the management of the risks associated with antimicrobial drug resistance transferred from food animals to humans. In fact, the two options probably only work in concert. The requirement of antibiotics in veterinary therapy and bacterial infection prevention in food animals should be minimized by improving the methods of animal husbandry and disease eradication, optimal usage of existing vaccines and development of new vaccines. This will not only protect public health but also safeguard the future efficacy of antibiotics in both veterinary and human medicine.
None to declare.
Work in Edmonton on Campylobacter was supported by the Canadian Institute of Health Research and the Natural Science and Engineering Research Council. D. E. T. is a Medical Scientist with the Alberta Heritage Foundation for Medical Research.