|Home | About | Journals | Submit | Contact Us | Français|
There are few documented reports of antibiotic resistance in Chlamydia and no examples of natural and stable antibiotic resistance in strains collected from humans. While there are several reports of clinical isolates exhibiting resistance to antibiotics, these strains either lost their resistance phenotype in vitro, or lost viability altogether. Differences in procedures for chlamydial culture in the laboratory, low recovery rates of clinical isolates and the unknown significance of heterotypic resistance observed in culture may interfere with the recognition and interpretation of antibiotic resistance. Although antibiotic resistance has not emerged in chlamydiae pathogenic to humans, several lines of evidence suggest they are capable of expressing significant resistant phenotypes. The adept ability of chlamydiae to evolve to antibiotic resistance in vitro is demonstrated by contemporary examples of mutagenesis, recombination and genetic transformation. The isolation of tetracycline-resistant Chlamydia suis strains from pigs also emphasizes their adaptive ability to acquire antibiotic resistance genes when exposed to significant selective pressure.
Chlamydiae are a successful group of obligate intracellular pathogens that cause serious diseases in a wide range of hosts (Box 1). Chlamydial infection of cells is initiated by an infectious, but metabolically inactive, elementary body (EB) that subsequently differentiates into a metabolically active, but noninfectious, reticulate body (RB). All chlamydial development occurs within a membrane bound vacuole termed the inclusion (Figure 1A & D). Replication by Chlamydia trachomatis RBs is synchronous until approximately 18–24 h post-infection, at which point dedifferentiation to infectious EBs can first be observed . During infection, a subset of host-derived vesicles are trafficked to the inclusion, where Chlamydiae direct the modification of the inclusion membrane through secretion of proteins that facilitate vacuolar modification and manipulate host signaling pathways, via interactions with other chlamydial or host cell proteins [2–4]. Additional chlamydial proteins are secreted into the host cytosol, where they affect immune recognition and intracellular survival of the pathogen [4–7]. Most chlamydial developmental cycles are complete in 40–72 h when, in most cases, the host cell lyses and infectious progeny are released from the cell. Although chlamydia have a highly reduced genome of approximately 1 Mb , the inability to introduce gene-specific DNA or culture the organism in the absence of host cells imposes constraints on the experimental techniques available to study basic chlamydial biology. As a result, genome sequencing and comparative genomics are of primary importance in developing a clearer understanding of chlamydial biology.
It is estimated that 40 million individuals worldwide have active trachoma caused by singular or mixed infections of Chlamydia trachomatis, Chlamydia pneumoniae and Chlamydia psittaci [24,72,117]. An additional 8.2 million have trichiasis and 1.3 million are blind as a result of ocular infections caused by chlamydia. Particular strains of C. trachomatis that cause trachoma are hyperendemic to regions of sub-Saharan Africa, the Middle East, Asia and parts of South and Central America; however, the distribution and involvement of C. pneumoniae and C. psittaci strains in active trachoma cases around the world is currently unknown . Transmission occurs through both direct and indirect contact, and roughly 25% of all individuals infected are children under the age of 10 years. However, serious disease and blindness is found in older individuals caused by cumulative scarification left by untreated infections .
Sexually transmitted infections caused by C. trachomatis are the most prevalent bacterial cause of sexually transmitted infections worldwide, and around 92 million men and women are estimated to be infected . The majority of infections are asymptomatic in both men and women, but if left untreated can result in a variety of pathologies, including urethritis, cervicitis, salpingitis, pelvic inflammatory disease, ectopic pregnancy and infertility .
10% of community-acquired pneumonias are attributable to C. pneumoniae infection, and 50–80% of adults demonstrate antibody titers to the pathogen. Most primary infections are not serious, but secondary pathologies are associated with C. pneumoniae infections and they are a risk factor in the development of atherosclerosis, asthma, chronic bronchitis, chronic obstructive pulmonary disease and Alzheimer’s disease. The contribution of C. pneumoniae in these secondary sequelae can be controversial and challenging to assess because of the ubiquity of the pathogen [119–122].
Both respiratory and genital chlamydial pathogens are implicated in long-term infections that are resistant to antibiotic eradication. C. pneumoniae can disseminate via macrophages to deeper lung, arterial and joint tissues. C. trachomatis can infect mobile monocytes, which may carry the bacteria to joint tissues where they reside in an aberrant intracellular state and induce inflammation. Some studies have found a 50–80% prevalence of C. trachomatis DNA in patients with reactive arthritis .
Psittacosis outbreaks caused by C. psittaci were documented as early as 1879 and continue through to today, where they largely affect individuals that routinely contact psittacines (parrot-like birds), pigeons, turkeys, ducks and geese. Inhalation of contaminated aerosols from urine, feces or other secretions of birds represents the primary route of transmission. Untreated C. psittaci infections of the lung can involve severe respiratory distress and systemic organ involvement, leading to serious disease and death. In 2001, 11 reported cases were documented in the USA, although it is thought that many cases go unnoticed or undiagnosed .
A variety of Chlamydia infect different animal species, causing a variety of diseases in many organ systems. Serious chlamydial disease is found in sheep and other ruminants, swine, marsupials, cats, birds and fish. Specific chlamydial species or strains are generally limited to the different animal host species .
Although Chlamydiae share many similarities with other Gram-negative bacteria, they constitute a distinct phylogenetic and genetic lineage. Their genomes are marked by a high degree of genetic conservation and very limited evidence of horizontally acquired foreign DNA. Chlamydiae can evolve in vitro resistance to antibiotic stressors through the accumulation of point mutations, and these resistance properties can be circulated among strains via horizontal gene transfer and homologous recombination [9–21]. Despite this ability to evolve in the laboratory, stable genetic antibiotic resistance in clinical settings has yet to be documented.
The primary frontline antichlamydial antibiotics, tetracyclines (TETs) and azithromycin (AZM), are highly effective in the treatment of uncomplicated chlamydial infections . However, accumulating data suggest that a break in the normal chlamydial developmental cycle can result in persistence and long-term infection that is refractory to antibiotic therapy. An understanding of this phenomenon is far from complete. Although 50% of genital C. trachomatis infections resolve spontaneously within 1 year of testing , a further understanding of long-term infections is important, because it is hypothesized that persistence can cause a cascade of potentially serious inflammatory-induced sequelae, such as pelvic inflammatory disease, infertility, blindness, arthritis, asthma and atherosclerosis [24–28].
Because Chlamydiae are widely distributed and often have high prevalence in human populations, these organisms are often present with other bacterial, viral and parasitic organisms [29–31]. For example, 15–60% of individuals with Neisseria gonorrhea genital tract infections are also infected with C. trachomatis, and concurrent infections with both Treponema pallidum and C. trachomatis also occur [29,32]. These co-infections have historically led to therapy complications. For example, β-lactam antibiotics have historically been the recommended drugs for both T. pallidum and N. gonorrheae [32–34], but treatment of chlamydial infections with these antibiotics induces chlamydia to become persistent. This persistence may exacerbate disease in the genital tract and lead to treatment failure and long-term complications (see later) [33,34]. For these and other reasons, carefully evaluated broad spectrum antibiotic therapies for bacterial genital tract infections are recommended, and this has been the case for many years . While significant antibiotic resistance is emerging in N. gonorrheae and T. pallidum, stable antibiotic resistance remains undetected in human chlamydial isolates, despite significant selective pressures. This lack of chlamydial antimicrobial resistance in clinical settings reinforces the relative resistance of Chlamydiae to alterations of genome structure, a subject that remains a significant barrier of progress for chlamydial research scientists.
In vitro or in vivo evidence of chlamydial persistence can be demonstrated in all chlamydia species, and can be routinely induced in the laboratory when infected cells are exposed to β-lactam antibiotics, IFN-γ or are deprived of iron supplements or amino acids [35,36]. Persistent or ‘aberrant’ RBs continue to synthesize proteins and replicate DNA, but they halt cell division. The resulting inclusions contain small numbers of very large aberrant RBs, and yield a prolonged infection caused by viable but nonculturable chlamydia (Figure 1B & E). Removal of the stressor results in septum formation, RB division and differentiation to EBs . Failure to respond to antibiotic treatment can follow establishment of chlamydial persistence in vitro, and it may be challenging in vivo to differentiate persistence from potential cases of antibiotic resistance. Although uncomplicated infections are quite responsive to antibiotics, unresolved genital, ocular and respiratory infections that fail to respond to antibiotic treatment are extensively documented [36–38]. It is possible that this is a function of poor therapeutic control of aberrant, persistent Chlamydiae in patients.
Both in vitro and in vivo evidence of penicillin treatment show that a dramatic change in the bacterial cell structure can suspend the developmental lifecycle and trigger a persistent state.
Several studies have identified possible ways that antibiotic therapy in clinical settings or long-term infection might lead to phenotypic resistance to antibiotics that are normally very effective in both C. trachomatis and Chlamydia (Chlamydophila) pneumoniae [19,32,39–41]. Examples include a study showing that persistent chlamydia became phenotypically resistant to AZM clearance after initial exposure to penicillin , and work showing that the macrolide erythromycin blocked EB to RB differentiation if added prior to infection, but caused RBs to enlarge and blocked differentiation to EBs when the antibiotic was added 18 or 24 h postinoculation .
The presence of chlamydial RNA and DNA in culture-negative patients showing evidence of chronic chlamydial disease provides support for some form of persistence in clinical settings [44–47]. Atypical RBs were found in cases of reactive arthritis (Reiter’s syndrome) and in chronic prostatitis cases caused by C. trachomatis after antibiotic treatment [25,48]. Morphologically aberrant RBs in macrophages from an aortic valve sample from chronic C. pneumoniae infection have been identified . These in vivo reports, along with in vitro experimental data, establish possible mechanisms for clinical treatment failures in chlamydial infections that might lead to erroneous conclusions regarding the antibiotic resistance of a clinical isolate.
There are only a few reports describing the isolation of antibiotic-resistant C. trachomatis strains from patients [50–55]. Although 11 of the 15 reportedly resistant isolates were associated with clinical treatment failure, all of the isolates screened displayed characteristics of ‘heterotypic resistance’, a form of phenotypic resistance in which a small proportion of an infecting microbial species is capable of expressing resistance at any one time. This phenomenon has also been described in Staphylococcus spp. [56,57], and parallel observations of similar phenotypic resistant states can be referred to in the literature as drug indifference, persistence, tolerance and, in some cases, as properties of biofilms [58,59]. It is possible that these descriptors of bacterial interactions with antibiotics can be associated with chlamydial aberrancy and phenotypic antibiotic resistance in Chlamydiae. For example, tolerance is often specific to antibiotics that affect cell wall synthesis, as is shown in the penicillin persistence model of Chlamydiae [58,59].
In each case of clinical resistance reported, only a small portion of the population (<1–10%) expressed resistance, and those that did also displayed altered inclusion morphology. In addition, the isolates could not survive long-term passage (in the presence or absence of antibiotics) or lost their resistance upon passage. In some cases, heterotypic resistance was observed when a large inoculum was infected on to cells, but a smaller inoculum was not resistant under the same conditions [50,60]. Many of these characteristics suggest that a form of phenotypic resistance is responsible for the sustained presence of small populations of clinical strains of C. trachomatis under antibiotic stress and may be an adaptive behavior that influences the survival of bacteria within communities rather than stable genetic resistance mechanisms employed by singular cells.
A distinct characteristic of chlamydial growth is the asynchronous differentiation of RBs to EBs that begins relatively early and continues throughout the developmental cycle. A midstage inclusion will harbor actively dividing RBs as well as nondividing EBs. It is plausible that multistage development is an evolved trait that can ensure the survival of a subset of the population regardless of the timing of antibiotic or metabolic stress. AZM, clarithromycin, levofloxacin and ofloxacin approach 100% inhibition in synchronized assays, but when used in a continuous model of C. pneumoniae infection, none of these antibiotics eliminated the organism, even in the presence of concentrations greater than four-times their minimum inhibitory concentrations (MICs) [39,40,61]. A continuous model may more accurately reflect in vivo infections because inclusions of varying developmental stages will be present at any given time. The standard MIC assay synchronizes the infection and applies antibiotics within 1–2 h post infection, long before EB differentiation can be observed. Perhaps chlamydia are most vulnerable in the log-phase of growth prior to EB differentiation, and are capable of expressing phenotypic resistance when both replicating and nonreplicating forms are present. This principle is corroborated by other studies, in particular one in which ciprofloxacin and ofloxacin failed to eradicate C. trachomatis in infected cells and induced persistence when antibiotics were applied to established infections (2–3 days post infection) [41,43]. Although it is assumed that the inclusion is a nutrient-rich environment, it is unknown whether adequate nutrient levels can support replication and sustain active metabolism, or whether toxic byproducts accumulate, particularly in the late stages of the developmental cycle when several hundred bacteria occupy a single inclusion. These factors may also contribute to the onset of phenotypic or heterotypic resistance observed both in vivo and in vitro.
It is challenging to distinguish persistence from issues of treatment compliance, re-infection of treated patients and actual antibiotic resistance in Chlamydiae. It remains even more challenging to assess the relevance of heterotypic resistance when it is observed in strains isolated from patients with clinical treatment failure. In the absence of true genetic differences, it is challenging to find a way to study antibiotic resistance that arises only under certain conditions in approximately 1% of the population and which often does not appear to manifest itself following expansion of the bacteria.
All antimicrobial susceptibility assays in the chlamydial system involve isolating and expanding clinical isolates and then culturing chlamydial progeny in cells with media containing different dilutions of antibiotics. There remains no universal testing methodology for these assays, and the techniques themselves are technically challenging and time-consuming . Many different cell lines and techniques are used in different diagnostic laboratories, which presents significant challenges in monitoring and evaluating potential emergence of antibiotic resistance. The following can all influence the outcome of an antibiotic susceptibility analysis: cell line; passage number of both host cells and Chlamydiae; multiplicity of chlamydial infection; developmental stage when antibiotic is added to the infected cells; and presence or absence of cycloheximide (used to slow growth of the host cells) . Additional attributes of chlamydial growth as well as cellular uptake of antibiotics by host cells can vary substantially in different models of polarized and nonpolarized cells [42,62–65]. For example, different cell lines permit differential growth of chlamydia strains when exposed to the same concentration of AZM. The relevance of this observation is highlighted in a report of AZM-resistant isolates from patients with relapsing infections that were characterized using a cell line permissive to chlamydial growth in the presence of inhibitory concentrations of AZM [55,66]. These observations highlight the challenges that arise when small differences in methodological approaches can further complicate the interpretation of in vitro resistance and its association with clinical relevance.
Clinical isolates can be extremely fastidious and often have much slower growth rates, increased potential for cytotoxicity or persistence, or can be present in very low numbers relative to the reference strains that are used as controls in MIC/minimum chlamydiacidal concentration assays. Although specific percentages vary, there is a significant fraction of nucleic acid amplification test (NAAT)-positive cases that are not detected by culture. Culture recovery rate is particularly low from rectal samples and is modestly better from urine, cervical, oropharyngeal and other sites [67–69]. Despite the marginal sensitivity of culture-based diagnostics, their specificity approaches 100%. By contrast, even though the sensitivity of NAATs is 20–30% better than culture and other non-NAATs, the occurrence of false positives and a lack of reproducibility between different NAATs has led the US Centers for Disease Control and Prevention to recommend confirmatory testing in certain cases [70,71]. Additionally, molecular-based diagnostics are limited by the inherent bias in our understanding of particular strains, species or even ecology of infectious organisms implicated in disease pathologies. This is particularly true in regards to the etiology of Chlamydia ocular infections, where it was recently shown that many individuals may carry mixed infections with C. trachomatis, C. pneumoniae and/or C. psittaci. These Chlamydiae are not routinely screened for, and have likely been missed using the routine species-specific NAATs . Each of these issues highlights the challenges associated with the diagnosis of chlamydial infections. Although NAATs are expensive, culture is certainly more so. Culture-based methods have also become a less attractive tool because of sensitivity issues, and the time and technical expertise required for their completion. This means many laboratories are not positioned to perform routine culture and are thus ill-equipped to conduct routine antibiotic screening of chlamydial isolates. This may hinder timely and accurate assessment of antibiotic resistance of clinical chlamydial isolates, even if such isolates are present in patients.
Chlamydiae are known to acquire resistance through mutations to six major classes of antibiotics. Both naturally acquired and laboratory-generated resistance found in selected chlamydial strains have facilitated the study of conserved biological pathways, such as peptidoglycan synthesis, folate synthesis and methionine synthesis, which cannot be approached directly in the chlamydial system [17,73,74]. The ability to generate resistant mutants has supported new experimental methods that facilitate recombination and transformation in or between Chlamydiae in vitro (Table 1). The following sections will describe resistance phenotypes that are stably expressed by Chlamydiae in cell culture systems.
Tetracyclines block bacterial protein synthesis by preventing aminoacyl tRNAs from interacting with ribosomes. TETs are widely used in both human and veterinary medicine because of their relatively low cost, broadspectrum of activity and excellent tissue distribution. Gram-negative, Gram-positive, atypical bacteria (Chlamydia, Mycoplasma, Rickettsia) and even some protozoa respond to therapeutic doses of TET. TET and its derivatives are often well-absorbed, have low toxicity and are relatively inexpensive . Since their initial discovery several decades ago, a large percentage of the total volume of antibiotics used in veterinary medicine for both therapy and growth promotion were TET [76,77,201]. In many bacterial systems, TET resistance is quite common . Resistance to TET was first discovered in a Shigella dysenteriae isolate in 1953, only 7 years after the initial discovery of the drug . As of 2005, 38 genes that encode TET efflux pumps, ribosomal protection proteins or inactivating enzymes were known . Several of the resistance genes encode proteins that function against a broad range of TET derivatives. Many of the factors associated with chlamydial ecology in the veterinary system led to speculation that this might be the first antibiotic in which Chlamydiae would show clinically relevant stable antimicrobial resistance, and this speculation has proved true. In the mid-1990s, the first stably resistant Chlamydia suis strains were isolated from diseased and normal pigs in the Midwestern USA . Eight independent strains were identified, and each exhibited high level resistance to TET (Figure 1D & F). Six of eight strains also exhibited a stable, but currently uninvestigated, resistance to sulfadiazine. Unlike previous reports of TET resistance in chlamydial strains, these strains were passaged up to 15 times in antibiotic-free media, and survived in media containing antibiotics without showing obvious signs of morphological abnormalities . Genetic characterization of the isolates revealed the presence of foreign genomic islands (ranging in size from 6 to 13.5 kb) that had integrated into the chlamydial chromosome . Each island carries genes encoding a TET efflux pump and a regulatory repressor (tet[C] and tetR, respectively), a unique insertion sequence (IScs605) plus three to ten additional genes involved in plasmid replication and mobilization. This TET resistance allele is identical to the tet(C) gene in the cloning vector pSC101 and a wide range of other vectors. In 2008, a report identified 14 additional C. suis strains collected in Italy that shared 100% nucleotide identity with the tet(C) gene from the original US strains . Of the 14 strains, 12 of these isolates grew in the presence of TET, whereas two could not. It is not clear why these two strains, which did contain the resistance gene, were unable to grow in the presence of TET.
The C. suis tet(C) islands share more than 99% identity to sections of the resistance plasmid pRAS3.2 isolated from Aeromonas salmonicida , a Gram-negative pathogen of salmon and trout that grows poorly and becomes avirulent at mammalian body temperatures . However, this plasmid lacked the IScs605 sequences. More recently, a TET-resistant element isolated from another aquatic associated Gram-negative bacteria, Laribacter hongkongensis, shared 100% nucleotide identity to IScs605 of C. suis . L. hongkongensis is an emerging cause of community-acquired gastroenteritis and travelers’ diarrhea in humans. The significance of either Aeromonas or Laribacter to the acquisition by C. suis of the tet(C) island is not clear. The discovery of the tet(C) islands represents the first identification of antibiotic resistance acquired through horizontal gene transfer in any obligate intracellular bacteria. Developing hypotheses to explain the acquisition of tet(C) islands isolated by C. suis is challenging and currently relegated to scientifically supported speculation. However, as more data accumulate, clues surface that may start to piece it together. Two of the genes identified in the islands were part of a novel insertion element (IScs605) related to other insertion sequence elements found in Helicobacter spp. IScs605 mediated site-specific transposition and integration in a heterologous system where the transposed DNA localized adjacent to a conserved pentameric sequence (5′-TTCAA) in 36 of 38 sequenced clones. Each island integrated next to a TTCAA sequence within the inv homolog of C. suis . These data, from both the transposition assays in a heterologous system, and the sequence specificity surrounding the integration site of each island, suggest the insertion sequence element mediated transposition into each of the TET-resistant C. suis genomes. Prior to this discovery, no other insertion sequence had been identified in Chlamydiae. From the work on C. suis and the singular L. hongkongensis isolate, there are now two sources that link an aquatic organism to the islands found in C. suis. L. hongkongensis is the only known organism to harbor the IScs605 insertion element, while the plasmid in A. salmonicida is the only species that shares identity with the remaining sequences of the tet(C) island. The feeding and rearing practices in the pig industry that rely heavily upon the prophylactic delivery of TET and the use of fish as a significant feed source may have promoted the ideal environment for the acquisition of DNA by chlamydia that commonly infect the porcine intestinal epithelia [Andersen AA, Pers. Comm.]. Many questions remain regarding how the tet(C) island was delivered to bacteria that grow within vacuoles inside cells, and how the island was incorporated into the C. suis genome. No conjugative machinery or competence genes have been identified in the genomes of Chlamydia spp. and the absence of a practical genetic system renders these questions very challenging.
Rifamycins, represented in most studies by rifampin (RIF), are bactericidal antibiotics that specifically interact with the β-subunit of RNA polymerase to inhibit bacterial transcription. These are not primary drugs of choice for treating chlamydial infections, although they do possess strong in vitro activity and are a therapeutic option in the treatment of clinical infections. Rapid emergence of resistance in vitro has been demonstrated in C. trachomatis, C. pneumoniae, C. caviae, C. psittaci, C. suis and C. muridarum after exposure to subinhibitory concentrations of drug [10–12,14,19,20,88]. Amino acid substitutions in the RNA polymerase (RNAP) β-subunit decrease the binding capacity of RNAP to RIF, which allows bacterial survival even under high concentrations of drug. Many bacterial species develop resistance through nucleotide changes in the RNAP β-subunit gene, rpoB. Similar to these bacteria, RIF-resistant Chlamydiae carry a variety of conserved and unique nucleotide changes in the central region of rpoB. A singular amino acid substitution leads to low-level resistance, but the acquisition of an additional substitution increases the MIC several fold. Single mutations increased the MIC from 0.008 μg/ml to between 0.5 and 64 μg/ml in C. trachomatis serovar D, and to between 4 and 64 μg/ml in serovar K. The nucleotide at position 471 of rpoB (Escherichia coli position 526) was the most common site mutated in resistant clones of C. trachomatis serovars D and K. When this nucleotide change was found in combination with one additional mutation, the MIC increased from 64 to 512 μg/ml for a serovar D isolate, and from 64 to 256 μg/ml for a serovar K isolate [19,88].
Work by Kutlin et al.  and Rothstein et al.  led to the development of RIF-resistant C. pneumoniae strains, but increases in resistance were modest and often took repeated passage for success. In most cases, resistance was associated with mutations in rpoB; however, of the two C. pneumoniae strains evaluated, only one strain (TW-183) developed resistance and carried the rpoB mutations . Rifalazil (RZL), a semisynthetic rifamycin derivative, has high efficacy against C. trachomatis infections in clinical trials and is effective in vitro against C. pneumoniae. Both C. trachomatis and C. pneumoniae strain TW-183 develop resistance to RZL when passaged in subinhibitory concentrations of the drug and acquire mutations in rpoB; however, C. pneumoniae strain CWL-029 did not develop such resistance . As seen with RIF, strains of C. pneumoniae can be selected for resistance to low concentrations of RZL and require more passages to develop resistance . Interestingly, RZL maintains activity against both RIF-resistant C. trachomatis and C. pneumoniae mutants [88,90]. Although clinical resistance to rifamycins in chlamydia has not been documented, the ability of these organisms to quickly accumulate mutations in vitro raises concern about the use of these drugs in treating infections.
Fluoroquinolones are bactericidal antibiotics that inhibit DNA gyrase and DNA topoisomerase IV . C. trachomatis, C. muridarum and C. suis can each develop quinolone resistance in vitro when exposed to subinhibitory concentrations of antibiotic [10–12,18,21,92]. After only four passages in 0.5 μg/ml of ofloxacin, the C. trachomatis MIC increased from 1 to 64 μg/ml. A similar result was achieved after four passages in the presence of 0.015 μg/ml of sparfloxacin. Two additional studies identified similar mutations associated with passage of C. trachomatis in the presence of quinolones, but the number of passages required to select for resistant mutants varied between four and 24 [18,21]. Quinolone-resistant strains were resistant to multiple derivatives and carried the same point mutation in the quinolone-resistance determining region of gyrA. Although, attempts to generate fluoroquinolone-resistant C. pneumoniae were unsuccessful for one group , a different group was able to cultivate moxifloxacin-resistant C. pneumoniae that carried an amino acid substitution at the same nucleotide position of gyrA as other fluoroquinolone-resistant C. trachomatis isolates .
There is also evidence for natural quinolone resistance, via mutations in the quinolone-resistance determining region of gyrA, in C. muridarum and the distantly related Chlamydiae-like bacteria, including Parachlamydia acanthamoebae, Neochlamydia hartmannellae, Simkania negevensis and Waddlia chondrophila. Some of this latter group have been associated with respiratory disease in humans, and this natural resistance is important to note as quinolones are often prescribed for the treatment of generalized respiratory disease [94–96].
Aminoglycosides interfere with translation initiation by interacting with the 30S ribosome. These antibiotics have poor penetration into mammalian cells, leading to MIC values for Chlamydiae that are extremely high (~1 mg/ml). Kasugamycin (KSM) and spectinomycin (SPC) are antibiotics used to generate aminoglycoside-resistant chlamydial strains in the laboratory. Passage of infected cells in concentrations greater than the MIC led to selection for C. psittaci 6BC at a frequency of approximately 2.3 × 10−5. Resistant strains carried mutations in the 16S rRNA gene at the KSM binding site [9,17], and resistance was present against all tested aminoglycosides.
Chlamydia trachomatis strains resistant to KSM were selected for using culture in sub-inhibitory concentrations of the antibiotic. Strains of C. trachomatis that were resistant to KSM did not have a mutation in the 16S rRNA, but did carry a two-nucleotide insertion in ksgA, which encodes a protein (KsgA) that is responsible for post-transcriptional methylation of ribosomal adenosine residues in other bacteria. The resistant C. psittaci strain was stable and grew comparable to wild-type strains. By contrast, the C. trachomatis KSM mutant was severely impaired for growth and was sensitive to high concentrations of antibiotic [9,17].
Similar to KSM, in vitro-generated and naturally occurring resistance to SPC is associated with mutations in the 16S rRNA gene. Exposure to subinhibitory concentrations of SPC selected for stable resistance in C. psittaci 6BC, and resistant mutants were recovered at a frequency of 1 × 10−6. Four different SPC-resistant mutants carried unique 16S mutations and varied in their fitness in competition assays with wild-type C. psittaci. One out of four of the mutations had no significant fitness cost to the bacteria; however, the other three mutations at adjacent nucleotides reduced bacterial fitness significantly. Some of these mutant genes conferred resistance to SPC in E. coli. The mutations identified in these studies were used to create an electroporation vector that was important in demonstrating the first stable transformation via electroporation of any Chlamydiae .
Spectinomycin-resistant C. trachomatis L2 mutants have not yet been generated. The inability to produce these mutants is likely due to the duplicity of rRNA operons and drug target sites [9,14,15]. For antibiotics that target ribosomal machinery, a single mutation in an organism encoding more than one rRNA operon is typically recessive, and the frequency at which resistant mutants can be recovered correlates with the number of ribosomal operons encoded in the genome. C. trachomatis encodes two nearly identical copies of the operon, whereas C. psittaci 6BC only encodes one. Simultaneous complementary mutations in two rRNA operons would arise at very low frequencies in vitro, possibly explaining why aminoglycoside-resistant strains containing mutations in 16S rRNA genes were not recovered in C. trachomatis.
Sulfonamide (SFM) and trimethoprim antibiotics interfere with bacterial folate synthesis, which is critical for DNA synthesis, repair and methylation. Stable trimethoprim-resistant mutants were reported to arise at very low frequencies (<5 × 10−10) in C. trachomatis cultured in vitro in subinhibitory concentrations of the antibiotic . C. trachomatis L2, C. psit-taci 6BC and C. suis (with the exception of some TET-resistant isolates) are all sensitive to SFM, while C. pneumoniae and all other tested strains of C. psittaci are naturally resistant. SFM resistance in other bacteria can be conferred through horizontal acquisition of mobile elements, but can also arise from mutations in the folate synthesis genes targeted by the drug. Specific insertions, repeats and point mutations in the folP gene (dihydropteroate synthase) can confer stable resistance to sulfa drugs, while mutations in the folA gene (dihydrofolate reductase) can confer resistance to trimethoprim . Iclaprim is a new dihydrofolate reductase inhibitor currently in development; however, this antibiotic maintains activity against both C. trachomatis and C. pneumoniae in vitro .
Azithromycin is a bacterial protein synthesis inhibitor and front-line drug for the treatment of chlamydia infections. High-level resistance to AZM was selected for in C. psittaci 6BC and C. caviae GPIC, while a C. trachomatis L2 strain was selected for in lower concentrations of AZM [13,16]. Cultivation of resistance was unsuccessful in C. pneumoniae clinical isolates with elevated MICs to AZM [37,99]. AZM-resistant C. psittaci strains were also resistant to other macrolides as well as a lincosamide, which share similar 23S rRNA target sites. Resistant strains were stable and survived passage in the presence and absence of these drugs. Similar to observations of resistance to KSM, the AZM resistant strains were isolated after exposure to inhibitory concentrations of AZM, while the modestly resistant (AZM tolerant) C. trachomatis strain was isolated only after exposure to subinhibitory concentrations of antibiotic. The AZM-tolerant C. trachomatis strain harbored a mutation in rplD that encodes the ribosomal protein L4.
Although some antibiotic-resistant mutations resulted in no overall effect on the physiology of the bacteria, in vitro AZM resistance imposes a competitive defect. The resistant C. psittaci strains were delayed in their differentiation from EB to RB compared with wild-type strains, and also had a slower doubling rate, produced significantly smaller plaques and were outcompeted in the absence of selection by the wild-type parent strain. The drug-tolerant C. trachomatis strain did not grow well in the absence of antibiotics, formed smaller plaques and produced fewer infectious particles than wild-type parent strains. The C. caviae AZM resistant strains carried mutations in the 23S rRNA of their single rRNA operon, produced fewer infectious particles in vitro and were less fit in vivo, compared with the wild-type strain .
Lincomycin is a bacteriostatic protein synthesis inhibitor that causes premature dissociation of peptidyl-tRNA from the ribosome . There is a single report of in vitro-generated lincomycin-resistant C. trachomatis mutants. These mutants were recovered at very low frequencies (<5 × 10−10) by growing and passaging infected cells in subinhibitory concentrations of antibiotic. The resistant mutants carried mutations in both 23S rRNA genes, corresponding to sites in E. coli that conferred similar resistance .
Extensive comparative analyses of DNA sequence data (including a broad range of clinical and laboratory isolates) support the conclusion that Chlamydiae are highly recombinogenic [46,84,101–115]. Discordant rates of genetic mutation between polymorphic loci and the rest of the genome are linked to evidence supporting genetic recombination as a source of genetic variation and genome maintenance and repair. This conclusion is challenged by the fact that chlamydial genomes are highly syntenous and encode only a very few extrachromosomal elements or genomic islands . In the absence of host-free growth or a genetic system, experiments addressing recombination or genetic exchange have been very difficult. Recent research by different laboratories are making progress in this area. Studies by Demars and colleagues used laboratory-generated antibiotic-resistant strains to develop an artificial system that screened for in vitro lateral gene transfer and recombination [10,11]. Co-infecting two resistant parental C. trachomatis isolates facilitated the selection of doubly resistant progeny strains with the antibiotic-resistant genetic markers from each parental strain. In one experiment, an ofloxacin (OF)-resistant C. trachomatis L1 strain harboring a mutation in gyrA (T249->G) and a RIF-resistant C. trachomatis D/UW-3/CX strain harboring a mutation in rpoB (C1400->T) were co-infected and grown in the presence of RIF and OF. A RIF- and OF-resistant C. trachomatis D strain was isolated that carried both the gyrA (T249->G) and the rpoB (C1400->T) mutations. Quantitative evidence supported lateral gene transfer, as opposed to spontaneous mutation, as the source of resistant phenotypes in the selected strains. Antibiotic resistance was stable in recombinants grown in the presence and absence of antibiotic selection. Sequencing of highly polymorphic loci (ompA, murA, pmpC, trpA, incA, ribF and recF), in addition to the mutated genes (rpoB and gyrA), was used to coarsely map several cloned recombinant genomes to estimate the length and composition of the transferred DNA. Remarkably, the authors provide evidence that large fractions of the genome were exchanged; between 123 and 790 kb was estimated to have transferred to the recipient. Although it is likely that very large fragments of the chromosome were exchanged in these crosses, it is difficult to genuinely determine the recipient and the donor strain in these experiments.
Work in our laboratory confirmed the fundamental discoveries of Demars et al. and demonstrated lateral gene transfer between several different Chlamydia spp. . Studies of Suchland et al. used the horizontally acquired TET-resistant marker from C. suis R19 as a primary tool of selection, and mapped recombination sites in cloned recombinant progeny using genome sequencing. These studies showed routine transfer of TET resistance from C. suis to any of several C. trachomatis strains, as well as the mouse-tropic C. muridarum. Transfer of TET resistance involved the insertion of between 40 and 100 kb of C. suis DNA into recipient strains. Progeny from primary crosses were used as donors in subsequent crosses, and recent clinical strains readily acquired TET resistance via recombination. Some recombinants are marked by substantial genome rearrangements and/or genetic mosaicism, while other recombined sequences are relatively short products from classical double-crossover recombination events. Additional in vitro-generated interstrain recombinants made from parental strains carrying OF or RIF resistance harbor a unique genetic patchwork of parental DNA across the entire genome . The recombinant genomes include the expected antibiotic resistance genes from parental strains, but somewhat unexpectedly, regions unrelated to the loci that conferred antibiotic resistance used for selection. The three Chlamydiae used in these experiments (C. suis, C. trachomatis and C. muridarum) are unique in their ability to form fusogenic inclusions when occupying the same cell, and it was initially hypothesized that sharing the same vacuole might be a requisite for the exchange of DNA. However, extensive work [Rockey DD, Unpublished Data] by our laboratory has shown that nonfusogenic strains of C. trachomatis that lack IncA, an important protein involved in homotypic inclusion fusion, still recombine in vitro when subjected to the same selection parameters as IncA-positive, fusogenic strains. New studies using in vitro transformation and classical genetic techniques are beginning to tease out genetic regions specific to unique phenotypes and growth characteristics of different clinical and laboratory strains . These techniques are currently being used by our group, and others, to serve as a rudimentary genetic system until more direct methods are developed.
Although these results were initially surprising and controversial, it is now acknowledged that multiple antibiotic-resistance genes can be readily recombined between Chlamydia spp. It is likely that recombination occurs naturally, and therefore, clinical resistance might spread rapidly in patients, following an initial, perhaps rate-limiting, introduction of an exogenous resistance gene into the chlamydial population.
The first successful transformation via electroporation in chlamydia was recently published and highlights again the importance and utility of antibiotic resistance in the study of chlamydial genetics . These individuals developed and characterized several antibiotic-resistant strains and associated mutation frequencies, using SPC and KSM as the selecting antibiotics. Several plasmids were produced that carried a C. psittaci 16S rRNA gene with mutations that conferred resistance to both antibiotics. These constructs were amplified in a methylase-deficient E. coli strain and electroporated into C. psittaci. The chlamydia were infected onto host cells, grown in the presence of selecting antibiotics and transformants were plaque purified. Transformants were stable and could survive several passages in the presence and absence of antibiotics, and sequencing of progeny 16S rRNA confirmed that resistance was derived from the electroporated plasmid DNA. These authors used homologous sequences ranging from approximately 1.5–8 kb, and estimated that regions of recombination were between 0.4 and 1 kb.
Although the transformation frequency using electroporation is low compared with the frequency of doubly resistant strains recovered from recombination (10−6 vs 10−3), the work of Binet and Maurelli has provided an important proof of concept for introduction of foreign DNA into chlamydia. This technology is currently limited to altering only a single locus in selected chlamydial strains, and is not a generally applicable method for introducing or inactivating genes in these bacteria. Translation of these techniques into standard methods to introduce or inactivate chlamydial genes remains a significant challenge in this system.
Although there is no genetic evidence of antibiotic resistance leading to treatment failures in humans, the C. suis strains resistant to TET and the in vitro results with the other described antibiotics indicate that clinicians should be vigilant for the possibility in the future. Current culture and diagnostic methods may not be sufficient to detect emerging antibiotic-resistant strains due to these persistent states, the low recovery rate in culture from various infection sites, potential instability of resistant isolates in vitro, or the unknown significance of heterotypic resistance in treatment failure. Little is known about the heterotypic-resistant phenotype observed in the MIC assays, and whether or not it has biological relevance to in vivo conditions or can be correlated with cases of treatment failure. Although there is currently no evidence for heritable antibiotic resistance in human clinical settings, results discussed in this review indicate that several antibiotic-resistance genotypes can be generated and transferred to most C. trachomatis, C. suis or C. muridarum isolates, and that chlamydial antibiotic resistance will be an important laboratory tool for researchers.
The characterization of TET-resistant C. suis isolates, spontaneous mutants and recombination strains demonstrate that resistance can emerge and disseminate amongst chlamydia, and in some cases with relative ease. Major biological barriers have probably prevented the acquisition of DNA or mutations that promote antibiotic resistance in clinical and veterinary isolates. If resistance genes were to enter the chlamydial population from other species, we expect that this resistance could spread rapidly within species via recombination.
The use of antibiotic-resistant strains in chlamydial research has led to significant understanding of chlamydial recombination in vitro and the resistant strains will likely continue to serve as an essential tool to further our understanding of chlamydial genetics. By utilizing differently resistant strains with specific phenotypes, unique inter- and intrastrain crossing may generate chimeric chlamydial strains that display targeted phenotypes, while carrying a unique and unnatural set of genetic markers. These approaches may be useful for correlating unidentified genes with known chlamydial phenotypes, and for exploring the role of individual chlamydial genes in infection and disease.
Sara Weeks and Bob Suchland are gratefully acknowledged for their critical review of the manuscript. We wish to dedicate this work to our recently deceased colleague Walter E Stamm. Dr Stamm was a wonderful individual who contributed in so many ways to our personal and professional lives, and we miss him dearly.
For reprint orders, please contact: moc.enicidemerutuf@stnirper
Financial & competing interests disclosure
Daniel Rockey receives grant support from the NIH and the US Department of Defense (MT/DTRA). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Papers of special note have been highlighted as:
of considerable interest