Resistance mechanisms in Campylobacter jejuni

Abstract

Campylobacter jejuni is a major cause of food-borne gastroenteritis worldwide. While mortality is low, morbidity imparted by post-infectious sequelae such as Guillain-Barré syndrome, Reiter syndrome/reactive arthritis and irritable bowel syndrome is significant. In addition, the economic cost is high due to lost productivity. Food animals, particularly poultry, are the main reservoirs of C. jejuni. The over-use of antibiotics in the human population and in animal husbandry has led to an increase in antibiotic-resistant infections, particularly with fluoroquinolones. This is problematic because C. jejuni gastroenteritis is clinically indistinguishable from that caused by other bacterial pathogens, and such illnesses are usually treated empirically with fluoroquinolones. Since C. jejuni is naturally transformable, acquisition of additional genes imparting antibiotic resistance is likely. Therefore, an understanding of the antibiotic resistance mechanisms in C. jejuni is needed to provide proper therapy both to the veterinary and human populations.

Introduction

Campylobacter jejuni is a small, Gram-negative, curved rod and is the most common cause of bacteria-mediated diarrheal disease globally.¹ For the first time in 2005, campylobacteriosis exceeded salmonellosis as the most commonly reported zoonosis in the European Union, and the number of cases continues to increase.^2,3 Campylobacteriosis is also the most common notifiable disease in New Zealand and Australia.^4,5 There is little human-to-human transmission, probably due to its microaerophilic nature. Instead, it is primarily a zoonosis because it is a commensal of food animals, particularly poultry, which serves as the main reservoir for human infection.⁶ Meat becomes contaminated during the slaughtering process, and C. jejuni survives in the crevices of animal carcasses where oxygen tension is low.⁷ Although implementation of Hazard Analysis and Critical Control Points (HACCP) in the food industry in the mid 1990s markedly reduced the rate of Campylobacter infections,⁸C. jejuni remains second only to Salmonella as the cause of food-borne disease in the United States.⁹ However, other modes of transmission, such as drinking contaminated water, are also important means of disease spread.¹⁰

The indiscriminate use of antibiotics in the human population as well as the use of antibiotics in animal husbandry, for treatment, growth promotion and off-label uses, has led to an increase in antibiotic-resistant Campylobacter infections, particularly with regard to fluoroquinolones (FQ).^9,11-16 There is evidence to support the hypothesis that resistance patterns in poultry may predict human resistance patterns; this has been most clearly shown with FQ.^9,11-20 Although not all cases of Campylobacter infection require treatment,²¹ many cases of acute diarrhea are empirically treated with FQ, which likely further contributes to the emergence of FQ resistance.

The use of veterinary antibiotics varies greatly throughout the world. Of greatest concern are situations in which antibiotics can be used for growth-promotion purposes (as opposed to therapeutic) because the low levels of antibiotics used in this setting and over long periods of time set the stage for the emergence of resistant bacteria. In some areas including Indonesia, Thailand, India and parts of Africa, veterinary antibiotics can be obtained without prescription or other controls.^20,22 In contrast, the general use of antibiotics for growth promotion is banned in the European Union and Japan,²³ and FQ cannot be used in food producing animals in Australia.

Although Campylobacter has an extensive formidable restriction modification system that would tend to decrease the uptake of foreign genetic material, it is also naturally transformable, and the acquisition of resistance genes from other organisms has been described.^24-35 For all these reasons, the study of the resistance mechanisms present in C. jejuni is important to both human and veterinary health.

The genetic elements that underlie these mechanisms may be chromosomal or plasmid-borne, and represent a combination of endogenous and acquired genes. In general, mechanisms of antibiotic resistance include (Table 1):

Table 1. Antibiotic resistance mechanisms of Campylobacter.

Antibiotic class	Resistance mechanisms
Aminoglycoside	Modification of the antibiotic by aminoglycoside-modifying enzymes (AphA, AadE, Sat) Contribution of efflux is not clear
Beta-Lactam	Enzymatic inactivation of the antibiotic by β-lactamase (penicillinase, OXA-61) Decreased membrane permeability of most anionic and MW > 360 kDa antibiotics due to MOMP Efflux through CmeABC and possibly others
Fluoroquinolone	Modification of the DNA gyrase target (Thr-86-Ile; also Asp-90-Asn, Ala-70-Thr) Efflux through CmeABC
Macrolide	Mutations in 23S rRNA Contribution of mutations in ribosomal proteins L4/L22 is likely minor Efflux through CmeABC and possibly others Decreased membrane permeability due to MOMP
Tetracycline	Modification of the target ribosomal A site by TetO binding Efflux through CmeABC and possibly others Contribution of decreased membrane permeability due to MOMP is not clear

(1) Modification of the antibiotic’s target and/or its expression (i.e., DNA gyrase mutations)

(2) Inability of the antibiotic to reach its target (i.e., expression of the major outer membrane protein or MOMP)

(3) Efflux of the antibiotic (i.e., multidrug efflux pumps such as CmeABC)

(4) Modification or inactivation of the antibiotic (i.e., β-lactamase production)

In Campylobacter, a recurring theme is synergy between antibiotic efflux and a second mechanism. The best-described multi-drug efflux pump in Campylobacter is CmeABC, consisting of three components: an outer membrane protein (encoded by cmeC), an inner membrane drug transporter (encoded by cmeB), and a periplasmic protein (encoded by cmeA) that bridges CmeB and CmeC.^36,37 This efflux pump also contributes to resistance to bile acids.³⁸ Other putative efflux pumps including CmeDEF and CmeG, may also contribute to antibiotic resistance.^39,40 Sequencing reveals that C. jejuni has a total of 14 possible efflux pumps, but most have not been characterized functionally.⁴¹ In addition to intrinsic resistance mediated by efflux,^{36,37,39,40,42-44} antibiotic exclusion [via the major outer membrane porin (MOMP),⁴⁵ lipooligosaccharide and possibly capsule]⁴⁶ also contribute to intrinsic resistance. Campylobacter exhibits intrinsic resistance to novobiocin, bacitracin and vancomycin, polymyxin/colistin, presumably due to the absence of appropriate targets and/or low affinity binding to targets.^47-50 In the case of intrinsic resistance to trimethoprim,^26,47,51-53 variant forms of dihydrofolate reductases (encoded by dfr1 most often but also by dfr9) that are not inhibited by trimethoprim are found in > 90% of C. jejuni that have been examined.²⁶

Approximately 90% of Campylobacter infections in humans are caused by C. jejuni (C. coli accounts for ~9%),⁵⁴ and the majority of the literature on human infection focuses on C. jejuni. Therefore, this review will focus on the antibiotic resistance mechanisms found in C. jejuni for commonly-used antibiotics.

Fluoroquinolone Resistance

Fluoroquinolones manifest concentration-dependent, bactericidal activity against a wide variety of both Gram-negative and Gram-positive organisms, are available in both oral and intravenous forms, are conveniently dosed once or twice daily usually, and are well-tolerated; all these attributes make this a heavily-used class of antibiotic in humans. Nalidixic acid is the parent, non-fluorinated compound of this antibiotic class. The fluoroquinolones include the most commonly used antibiotics (i.e., ciprofloxacin) to treat acute bacterial diarrhea, although macrolides are the drug of choice if campylobacteriosis is strongly suspected.²¹ However, campylobacteriosis is clinically indistinguishable from other causes of bacterial diarrheal illness, and so without epidemiology suggestive of Campylobacter infection, many cases are treated empirically with FQ. As such, FQ resistance is of great clinical concern.

Worldwide, FQ resistance was unusual in the late 1980s to early 1990s.^12,13,55 However, the combination of indiscriminate use of FQ in humans and increased FQ use in the poultry industry in particular, has contributed to an increase in the prevalence of FQ-resistance in both animals and humans.^11,12,14 The surveillance of FQ susceptibility in Campylobacter in animals is important not only for purposes of food production, but also because the emergence of resistant strains in animals portends an increase in resistant human infections.^11-14,56 Recognition of this connection has supported the limitation or outright banning of FQ for veterinary purposes in many countries including the United States,⁵⁷ Denmark, and Australia among other countries. Accordingly, FQ-resistance in > 150 Campylobacter isolates from broilers in Australia was reported to be between 0–2.4%,^58,59 which likely contributes to the similarly low-level of FQ-resistance (2%) in human isolates.⁵⁹ Additionally, data from the Danish Integrated Antimicrobial Resistance Monitoring and Research Programme (DANMAP) show that C. jejuni isolates from domestic broilers was 11%, compared with 57% from imported broilers,⁵⁵ and that while one-third of domestically-acquired Campylobacter infections were FQ-resistant in 2011, 84% of infections acquired abroad were FQ-resistant.⁵⁵ Similarly, 6.5% of human Campylobacter infections acquired in Norway were FQ-resistant, compared with 67% FQ-resistance in infections acquired abroad.⁶⁰

However, the veterinary use of FQ varies widely throughout the world depending on the production setting (i.e., commercial vs. family-owned) as well as on a country-wide basis.^11,52,53,61 For example, as in many countries, antibiotics are used both prophylactically and therapeutically in industrialized and commercial free-range poultry farms in South Africa, but not in small-scale family farming. The rates of FQ resistance were highest in commercial free-range broilers, at > 95%, but were lower in industrial broiler (16%) and lowest in poultry from family farms (8%).⁶² Thailand also reports very high rates of FQ resistance in C. jejuni from broilers, upwards of 80%.^13,52,63 In Japan, the rate of nalidixic acid resistance in C. jejuni from broiler flocks was 55%, and for a veterinary FQ (enrofloxacin) the rate was 30%.⁶⁴ The experience in Europe has been quite variable, ranging from very low FQ-resistance (1.2%) in broilers in Norway,⁶⁰ intermediate in Belgium⁵⁶ and Poland,⁶⁵ (44% and 56%, respectively), to alarmingly high in Spain, where it was reported in 2000 that 99% of Campylobacter isolates from broilers were FQ-resistant.⁶⁶

In the United States in 2004, the Federal Drug Administration reversed its prior approval for the therapeutic use of the veterinary FQ enrofloxacin because of concerns that the rising level of FQ-resistant Campylobacter in poultry was being reflected in increasing FQ resistance in human isolates.^12,57 However, it is not yet clear if the elevated rates of FQ resistance will decline after FQ restriction. Studies in several countries have shown that FQ-resistant Campylobacter persists in poultry populations after the withdrawal of FQ,^67-70 suggesting that there is little/no fitness cost to FQ-resistance in Campylobacter.^70,71 The latest available data from the United States shows that resistance to ciprofloxacin in human isolates of C. jejuni peaked in 2007 at nearly 26%, and has since declined somewhat to ~22% in 2010;⁵⁴ whether this is a true decline attributable even in part to the enrofloxacin ban is debatable.

In Campylobacter, there are two well-described mechanisms that underlie resistance to FQ: (1) inactivation of the target of FQ; (2) efflux of FQ. These two mechanisms work together synergistically.^72-74 In general, the two intracellular enzymatic targets of FQ are DNA gyrase (encoded by gyrA and gyrB) and the structurally related topoisomerase IV (encoded by parC and parE).⁷⁵ Fluoroquinolones form a stable complex with these enzymes and traps them onto DNA, leading to decreased DNA replication, transcription, and ultimately to cell death.^76,77 However, several studies have demonstrated that C. jejuni and C. coli lack the parC and parE genes^41,78-80; therefore, they cannot be a source of FQ resistance. Instead, FQ-resistance in C. jejuni and C. coli occurs via specific point mutations in the quinolone resistance-determining region (QRDR) of the gyrA gene, with the Thr-86-Ile mutation being both the most common. This single mutation in gyrA leads to high-level resistance to nalidixic acid and FQ (ciprofloxacin minimum inhibitory concentration (MIC) > 16 g/ml), unlike FQ resistance in E. coli or Salmonella, which requires the accumulation of several point mutations in the QRDR before high-level resistance is achieved.⁸¹ Interestingly, the less common Thr-86-Ala mutation confers resistance only to nalidixic acid but not FQ.⁸² Perhaps since only a single point mutation is required for high level resistance, FQ-resistant mutants appear rapidly both in animals and humans^{16,67,72,73,83-88} The less common Asp-90-Asn and Ala-70-Thr mutations in gyrA confer intermediate-level resistance to FQ (ciprofloxacin MIC 6–16 g/ml).^89,90 While mutations in gyrB have been reported, they do not confer FQ resistance.^78,80,91

It is somewhat surprising that the Thr-86-Ile mutation in gyrA seems to increase the fitness of Campylobacter in a chicken model,⁷¹ although this observation supported by the previously mentioned studies demonstrating that these resistant strains persist even after FQ are withdrawn from poultry flocks for several years.^67-69 Conflicting reports exist about whether this mutation translates into more severe infections in humans.^92-97

Another mechanism of FQ-resistance that seems to work in concert with gyrA mutations is efflux via the chromosomally-encoded CmeABC multidrug efflux pump, which reduces the intracellular concentration of FQ and several other antibiotics.^36,37 This efflux pump acts synergistically with DNA gyrase mutations to effect high-level FQ-resistance;^72-74 for example, strains carrying DNA gyrase mutations that alone lead to intermediate-level FQ resistance manifest high-level resistance when CmeABC is also expressed.^73,74 Also, CmeABC assists in the emergence of gyrA mutants that otherwise could not survive selection by even low dose FQ.⁷⁴ Unlike efflux pump mechanisms in other Gram-negative bacteria which require overexpression to lead to clinically relevant resistance, the basal, constitutive expression of CmeABC is sufficient to mediate FQ resistance (although experimental overexpression does increase the level of FQ-resistance).⁷⁴

An additional putative efflux pump, CmeG, has also recently been described as conferring both resistance to structurally unrelated antibiotics as well as oxidants.⁴⁰ Insertional mutagenesis of cmeG led to a 4-fold reduction in ciprofloxacin resistance vs. the wild-type parent, and an 8- to 32-fold increase in resistance to ciprofloxacin and other FQ when cmeG was overexpressed.⁴⁰

Interestingly, FQ-resistance has emerged on poultry farms even in the absence of FQ administration.^64,98 Since the major mechanism of nalidixic acid and FQ-resistance in Campylobacter is via point mutations in gyrA, it is difficult to attribute this phenomenon to co-inheritance of multi-resistance mobile elements. It has been suggested that other antibiotics could select for FQ-resistance in Campylobacter,⁹⁸ but whether this occurs via expression of the CmeABC pump or other mechanism remains to be clarified.

Macrolide Resistance

Macrolide antibiotics and the closely-related ketolides are large molecules (MW > 700) that inhibit bacterial protein synthesis. The macrolide antibiotic erythromycin is the treatment of choice for campylobacteriosis.²¹ Other members of this class of antibiotics include clarithromycin, azithromycin, telithromycin (technically a ketolide), tylosin and tilmicosin; the latter two are approved for veterinary use only (erythromycin also has a veterinary indication).⁹⁹ Macrolides inhibit protein synthesis by binding reversibly to the P site on the 50S subunit of bacterial ribosomes. The main mechanisms of resistance to macrolides in Campylobacter are (1) target modification, (2) efflux and (3) altered membrane permeability. The first two mechanisms act synergistically to confer high-level macrolide resistance.^100,101 A fourth mechanism of macrolide resistance, enzymatic modification of macrolides, has not been described in Campylobacter.¹⁰²

As in other bacteria, point mutations in the peptidyl encoding region in domain V of the 23S rRNA gene at positions 2074 and 2075 (corresponding to positions 2058 and 2059 in E. coli numbering) confer high-level macrolide resistance,^101,103-109 with the 2075 substitution being more common.^104,110 C. jejuni and C. coli carry three copies of 23s rRNA gene,^41,111 all of which are usually mutated in macrolide-resistant strains. However, some strains with lower MICs to macrolides have been found to have only two mutated gene copies, suggesting a gene dosage effect.^110,112,113 Strains harboring single mutations in 23S rRNA have not been reported. Mutations (usually insertions) in the ribosomal proteins L4 and L22 leading to macrolide resistance but are not the major means of tetracycline resistance.^100,106

The barrier to the generation of macrolide resistance in Campylobacter appears to be much higher than that of FQ-resistance. In two studies, several weeks of tylosin administration to poultry at a growth-promotion dose was necessary to select for macrolide resistance.^101,114 In another departure from FQ-resistant Campylobacter, macrolide resistance imparts a fitness cost, at least when analyzed in competition experiments.^114-118 These two factors combined with a low spontaneous mutation rate leading to macrolide resistance (~10⁻¹⁰ per cell per generation)¹⁰¹ and clinical efficacy, make macrolides the drug of choice to treat campylobacteriosis. Through 2008, there was concern in the US about increasing macrolide resistance in both C. jejuni and C. coli (2.3% and 10.1%, respectively), but there was a decline over the following two years to 1.2% and 4.3% for these strains.⁵⁴ As of 2010 in the European Union, the highest rates of macrolide resistance for C. jejuni is in Malta (10%), and Italy for C. coli (33%).¹¹⁹ Unfortunately, rates are much higher in parts of Asia and Africa; for example, in Nigeria, nearly 80% of strains are macrolide-resistant.¹²⁰ Similar to observations made with FQ-resistance in South Africa, 88% of Campylobacter isolates from poultry raised commercially were erythromycin resistant vs. 0% for those isolates from small-scale family farms.⁶²

The multidrug efflux pump CmeABC also contributes to macrolide resistance^{36,37,101,107,113,121} and functions synergistically with 23S rRNA mutations to effect high-level macrolide resistance.^43,100,122 In mutants that are macrolide-resistant but lack 23S rRNA mutations, gene disruption of cmeB or antisense-mediated gene silencing of cmeA leads to inactivation of the CmeABC transporter and mediates reversion to a macrolide-susceptible phenotype.^100,123 The putative efflux pump CmeG may also contribute to macrolide resistance, as insertional mutagenesis of cmeG causes an 8-fold reduction in erythromycin resistance vs. the wild-type parent.⁴⁰ In addition, there is one study that suggests the existence of a second efflux system that contributes to low-level macrolide resistance, but it has not been further characterized.¹⁰⁷

A third mechanism of macrolide resistance involves altered membrane permeability mediated by expression of the major outer membrane porin (MOMP), chromosomally encoded by porA.^44,45 In Gram-negative bacteria, porins are outer membrane proteins that form transmembrane pores and allow the passive diffusion of hydrophilic molecules, including many antibiotics. Properties of the pore including its size and charge characteristics underlie the selectivity for what can pass through it. In C. jejuni and C. coli, MOMP forms a cation-selective pore that is smaller than pores typically found in E. coli,¹²⁴ and therefore should limit the entry of most antibiotics with a molecular weight greater than 360 such as the macrolides (MW > 700).⁴⁵ However, since macrolides are known to be very effective against Campylobacter, these drugs must be able to cross the outer and cytoplasmic membranes. Whereas porins provide an aqueous environment for the transport of hydrophilic molecules, the relatively hydrophobic macrolides are thought to gain access to the cytoplasm of Gram-negative bacteria via a “hydrophobic pathway”^125,126. This pathway seems to be promoted in E. coli and Salmonella strains bearing mutations in lipopolysaccharide (LPS) synthesis genes that yield truncated LPS (lacking hydrophilic O-antigen sugars). The outer membranes of these mutant strains are therefore relatively more hydrophobic than the parent strains, and exhibit increased susceptibility to hydrophobic antibiotics including macrolides.¹²⁶ Given that Campylobacter naturally expresses lipooligosaccharide (LOS), which lacks the hydrophilic sugars expressed by full-length LPS in other Gram-negative bacteria,^127,128 it is reasonable to speculate that this comparatively increased outer membrane hydrophobicity promotes the uptake of macrolides. This is supported by the observation that LOS truncation increases the susceptibility of C. jejuni to erythromycin by 8-fold, an effect that was doubled in C. jejuni mutants also carrying the A2074G mutation.⁴⁶

β-Lactam Resistance

The β-lactam antibiotics are a diverse class of compounds including penicillins, cephalosporins, carbapenems and monobactams, all of which contain the β-lactam ring required for antimicrobial activity. Individual members of this family are distinguished by various side chains that confer particular properties such as pharmacokinetics, resistance to stomach acid, hydrolysis by β lactamases, etc. By binding to and thereby inactivating the bacterial peptidoglycan transpeptidases (also known as penicillin-binding proteins) required to catalyze the final cross linking step, the resulting bacterial cell walls lack structural integrity and are subject to osmotic swelling and lysis. Exactly how this leads to bacterial cell death is not completely clear, but the unopposed action of autolysins, necessary for normal turnover and remodeling of peptidoglycan, may play a role.¹²⁹

Three mechanisms mediate β-lactam resistance in Campylobacter: (1) enzymatic inactivation by chromosomally-encoded β-lactamases, (2) reduced uptake due to alterations in outer membrane porins and (3) efflux.

Expression of a penicillinase-type of β-lactamase in Campylobacter confers resistance to amoxicillin, ampicillin and ticarcillin, which can be overcome with the β-lactamase inhibitors tazobactam, clavulanic acid and sulbactam.¹⁰⁵ This enzyme does not affect susceptibility to the carbapenems or cephalosporins. More recently, a class D β-lactamase OXA-61, was identified in Campylobacter.¹³⁰ This enzyme shows similarity to other OXA-type genes in Fusobacterium, Acinetobacter and Pseudomonas, and mediates resistance to penicillin, oxacillin, ampicillin, amoxicillin-clavulanate, piperacillin and carbenicillin.^130,131 While OXA-61 is highly prevalent in the veterinary and human populations studied,¹³¹ pooled national data on the prevalence of β-lactam resistance in general is not available since the NARMS does not include the β-lactam class for Campylobacter ⁵⁴. However, it appears that the prevalence of β-lactamase varies widely in both poultry and human populations, but is usually greater than 20%.^105,131-134 Finally, two genes encoding a metallo-β-lactamase type of enzyme has been reported, although it is not yet clear if expression actually leads to β-lactamase resistance.^131,132,135

As with macrolides, the cation-selective MOMP in C. jejuni and C. coli tend to exclude most β-lactams with a molecular weight greater than 360 or which are anionic.⁴⁵ The partial positive charge and small size of imipenem (MW 299), ampicillin (MW 333) and cefpirome (MW 347) are consistent with passage through MOMP, and susceptibility to these antibiotics in the absence of a second mechanism such as β-lactamase production. Amoxicillin’s molecular weight of 365 would seem to preclude efficient passage through MOMP, although its partial positive charge might facilitate entry through MOMP; alternatively, a non-MOMP-dependent mechanisms may mediate its entry.⁴⁵

The CmeABC efflux pump may also contribute to β-lactam resistance. Insertional mutagenesis of cmeB in C. jejuni strain 81-176¹³⁶ and another strain led to a 32-fold increase in ampicillin susceptibility.³⁶ In another study using NCTC strain 11168, the cmeB mutant was 4 times more susceptible to ampicillin compared with the parent strain, and overexpression of cmeB lead to a 4-fold increase in ampicillin resistance.⁴⁴ Inactivation of the putative efflux pump CmeDEF by insertional mutagenesis of cmeF only led to a 2-fold increase in ampicillin resistance in strain 11168 and a 2-fold increase in cefotaxime resistance in the well-described, invasive, human outbreak strain 81-176.³⁹ Also, inactivation of the putative efflux pump CmeG did not affect cefotaxime resistance in C. jejuni strain 11168.⁴⁰ Therefore, at this point it seems that CmeABC is the most potent efflux pump with regard to β-lactams.

Tetracycline Resistance

The tetracyclines were discovered in the 1940s and have activity against Gram-negative and Gram-positive organisms. Due to their heavy use in the past for both human and veterinary indications, widespread resistance has somewhat limited their use today. Commonly used members of this class are tetracycline, doxycycline and minocycline. The tetracyclines are lipophilic protein synthesis inhibitors that likely use a combination of the hydrophobic pathway described for macrolides as well as outer membrane porins to gain access to the bacterial ribosome; exactly how each pathway contributes to tetracycline entry in Campylobacter is not completely clear. Known mechanisms of tetracycline resistance in Campylobacter are (1) alteration of tetracycline’s ribosomal target and (2) efflux.

In other Gram-negatives, tetracyclines form a complex with magnesium, which imparts a positive charge that facilitates passage of the complex through pores formed by OmpC and OmpE.¹³⁷ Although MOMP of Campylobacter shares an antigenically-related region to OmpC in E. coli,^138,139 it is not certain that the high molecular weight of tetracyclines (> 400) allows passage through the relatively small pores (MW exclusion ~360) imparted by MOMP.⁴⁵ Nevertheless, once inside the bacteria cytoplasm, tetracyclines reversibly bind to the 30S subunit of ribosomes and inhibit protein synthesis by preventing the attachment of charged aminoacyl-tRNA to the ribosomal A site.^140,141 The major mechanism of tetracycline resistance in Campylobacter as well as other Gram-negatives is protection of an unoccupied A site by the binding of bacterial protein TetO to that site.^142,143 TetO can be encoded on the chromosome,¹⁴⁴ or more commonly, on the plasmids pTet in C. jejuni¹⁴⁵ and pCC31 in C. coli.^146,147 According to 2010 NARMS data, 43% of C. jejuni and 49% of C. coli isolates are tetracycline-resistant,⁵⁴ making this class of antibiotic of little use in veterinary or human Campylobacter-mediated disease.^54,144,148

Although high-level resistance to tetracyclines can be mediated by TetO alone, the contribution of efflux to tetracycline resistance is demonstrated by the increase in tetracycline MIC when efflux pumps are genetically inactivated. For example, disruption of the putative efflux pump cmeG rendered the mutant strain 4-fold more susceptible to tetracycline compared with the wild-type strain.⁴⁰ Also, inactivation of the CmeABC efflux pump by disruption of cmeB led to an 8-fold decrease in the tetracycline MIC in a TetO-minus poultry isolate,³⁶ and similar findings were described with the NCTC isolate 11168 when cmeB was disrupted.³⁷ In a different study of other C. jejuni strains including 81-176, cmeB disruption rendered strains 16- to 64-fold more susceptible to tetracycline compared with the parent strains.¹¹² These studies also suggested that when both CmeABC and TetO are functional, the impact on tetracycline resistance is synergistic.^36,37,112

Aminoglycoside Resistance

Aminoglycosides are protein synthesis inhibitors of many Gram-positive and Gram-negative organisms. They contain amino-modified sugars, are positively charged, water-soluble and have molecular weights ranging from 445 to 600.^149,150 Commonly used members of this group include gentamicin, kanamycin, amikacin, neomycin, tobramycin and streptomycin. The initial binding of aminoglycosides to negatively charged bacterial membranes is electrostatic in nature and relatively slow compared with the second phase of rapid but reversible binding to the 30S segment of the ribosome.¹⁵¹ Transfer of aminoglycosides across the bacterial cytoplasmic membranes requires oxygen, an intact electron transport system and ATP.^150,152,153 According to 2010 NARMS data, > 99% of C. jejuni and 88% of C. coli isolates are susceptible to aminoglycosides.⁵⁴ These data suggest that despite Campylobacter’s microaerophilic nature, sufficient oxygen is present for the uptake of aminoglycosides.

There are two major means by which aminoglycosides exert antimicrobial activity: (1) interference with the translocation of the nascent peptide chain from the ribosomal A site to the P site leading to premature termination, and (2) interference with proof-reading, leading to incorporation of incorrect amino acids and dysfunctional protein.¹⁵⁴ The main mechanism of aminoglycoside resistance in C. jejuni is via aminoglycoside modifying enzymes, which are usually plasmid-borne.

Aminoglycoside resistance was first detected in C. coli and was mediated by a 3′-aminoglycoside phosphotransferase (encoded by aphA-3) that had been previously described as conferring kanamycin resistance in Streptococcus and Staphylococcus.²⁹ This aphA-3 gene remains the most common source of aminoglycoside resistance in Campylobacter. In some strains, aphA-3 is located downstream of an insertion sequence (IS607*) bearing similarity to IS607 found in H. pylori.²⁷ In other strains, aphA-3 is found with genes encoding streptomycin resistance (encoded by aadE, a 6′-adenylyl transferase) and streptothricin resistance (encoded by sat, an acetyl transferase).²⁷ The existence of a similar resistance cluster in Enterococcus suggests that Campylobacter acquired these genes via horizontal transfer.²⁷ Other Campylobacter strains harbor mosaic plasmids that contain various aminoglycoside resistance genes and insertion or transposon sequences from Gram-negative (i.e., H. pylori, E. coli and Salmonella) and Gram-positive sources (i.e., Enterococcus), along with tetO.^{24,25,27,28,30-32,34,35} Acquisition of such plasmids by susceptible C. jejuni confers a multi-drug-resistant phenotype and that can present a clinical challenge in both the veterinary and human populations.

Other genes which confer kanamycin resistance include aphA-1 and aphA-7, which were detected on plasmids in C. jejuni.^155-157 Unlike aphA-3 and aphA-1, which are thought have been horizontally acquired, aphA-7 has a similar G-C content to C. jejuni chromosomal DNA, suggesting it is intrinsic in Campylobacter.¹⁵⁸ Finally, there is a single report of a mutation in ribosomal protein S12 (encoded by rpsL) in C. coli that confers streptomycin resistance, but a similar mutation has not yet been described in C. jejuni.¹⁵⁹

The contribution of efflux to aminoglycoside resistance is less clear. In one study, the putative efflux pump inhibitors phenyl-arginine-β-naphthylamide and 1-(1-naphthylmethyl)-piperazine did not affect the MIC of kanamycin in 5 C. jejuni strains.¹⁶⁰ Another study which directly measured the effect of the putative efflux pump CmeG by insertional mutagenesis and comparison of MICs of various antibiotics found 16-fold reduction in gentamicin resistance in the mutated CmeG strain vs. the wild-type parent.⁴⁰ However, no effect on streptomycin resistance was noted, and overexpression did not lead to increased aminoglycoside resistance.⁴⁰ Therefore, the contribution of efflux to aminoglycoside resistance in Campylobacter is not completely clear but is likely to be less important than the plasmid-borne drug-modifying enzymes described previously.

Conclusions

Antibiotic resistance in C. jejuni is an increasing problem, as it is in many other microorganisms. Due to Campylobacter’s natural competence and hypervariable genomic sequences,⁴¹ there is considerable genomic plasticity that supports the emergence of resistant mutants. Because Campylobacter is a commensal of many animal species that are exposed to veterinary antimicrobials, ample opportunity exists for Campylobacter to continue to evolve additional resistance mechanisms. Furthermore, the over-use of antibiotics in the human population is an additional important source of selective pressure. Both scenarios contribute to the current problem of FQ resistance in Campylobacter. In this regard, the lack of a fitness cost (and perhaps even a fitness advantage⁷¹) of FQ-resistant C. jejuni is an issue that must be remembered when future resistance mutations arise in C. jejuni against other antimicrobials. Of greatest clinical concern would be the emergence of widespread macrolide resistance, since this class is the current treatment of choice for campylobacteriosis. A better understanding of the mechanism of macrolide entry (possibly via the hydrophobic pathway) may be useful in eventually mitigating the impact of acquired macrolide resistance. Also, the contribution of efflux pumps to antibiotic resistance warrants further study, since this mechanism acts synergistically with other mechanisms of antibiotic resistance to confer high-level resistance in many instances. Furthermore, genome sequencing predicts 14 potential efflux pumps,⁴¹ but only CmeABC and CmeG have been studied functionally,^40,44,161 making this a fertile area for future research.

The antibiotic resistance mechanisms discussed herein are summarized in Figure 1.

Figure 1. Summary of major antibiotic resistance mechanisms in Campylobacter. The ribosome, shown in blue at the left, is the site of two major resistance mechanisms. Binding of the TetO protein (shown in brown) to the A site (shown in dark purple) prevents tetracycline from occupying that site but stills allows access of the aminoacyl tRNA so that protein synthesis continues. Point mutations in 23S rRNA in the domain V region (shown in black) at position 2,075 principally and less often at position 2,074 (indicated by red stars) decrease the binding affinity for macrolides and lead to resistance. The major outer membrane protein (MOMP, shown in green), limits the entry of most antibiotics that are negatively charged or with a molecular weight larger than 360 kDa; the structure of the 552 kDa, dianionic antibiotic ceftriaxone is shown as an example. The Thr-86-Ile substitution in DNA gyrase (shown in light purple), is the main means of fluoroquinolone resistance, and this single mutation also confers high level resistance to this antibiotic class. The multi-drug efflux pump CmeABC (shown as stacked blue squares) contributes to resistance against fluoroquinolones, macrolides, β-lactams and tetracyclines, and works synergistically with other resistance mechanisms, often leading to high-level resistance. Aminoglycoside-modifying enzymes (AME; shown as the multi-colored star burst), principally of the aminoglycoside phosphotransferase family, are the main means of aminoglycoside resistance. Finally, β-lactamases (shown as the orange star burst) of the penicillinase type as well as the Ambler class D OXA-61 contribute to β-lactam resistance.

Glossary

Abbreviations:

FQ

fluoroquinolone

LOS

lipooligosaccharide

LPS

lipopolysaccharide

MIC

minimum inhibitory concentration

MOMP

major outer membrane protein

NARMS

National Antimicrobial Resistance Monitoring System

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.