Mycoplasma pneumoniae: Current Knowledge on Macrolide Resistance and Treatment

Mycoplasma pneumoniae causes community-acquired respiratory tract infections, particularly in school-aged children and young adults. These infections occur both endemically and epidemically worldwide. M. pneumoniae lacks cell wall and is subsequently resistant to beta-lactams and to all antimicrobials targeting the cell wall. This mycoplasma is intrinsically susceptible to macrolides and related antibiotics, to tetracyclines and to fluoroquinolones. Macrolides and related antibiotics are the first-line treatment of M. pneumoniae respiratory tract infections mainly because of their low MIC against the bacteria, their low toxicity and the absence of contraindication in young children. The newer macrolides are now the preferred agents with a 7-to-14 day course of oral clarithromycin or a 5-day course of oral azithromycin for treatment of community-acquired pneumonia due to M. pneumoniae, according to the different guidelines worldwide. However, macrolide resistance has been spreading for 15 years worldwide, with prevalence now ranging between 0 and 15% in Europe and the USA, approximately 30% in Israel and up to 90–100% in Asia. This resistance is associated with point mutations in the peptidyl-transferase loop of the 23S rRNA and leads to high-level resistance to macrolides. Macrolide resistance-associated mutations can be detected using several molecular methods applicable directly from respiratory specimens. Because this resistance has clinical outcomes such as longer duration of fever, cough and hospital stay, alternative antibiotic treatment can be required, including tetracyclines such as doxycycline and minocycline or fluoroquinolones, primarily levofloxacin, during 7–14 days, even though fluoroquinolones and tetracyclines are contraindicated in all children and in children < 8 year-old, respectively. Acquired resistance to tetracyclines and fluoroquinolones has never been reported in M. pneumoniae clinical isolates but reduced susceptibility was reported in in vitro selected mutants. This article focuses on M. pneumoniae antibiotic susceptibility and on the development and the evolution of acquired resistance. Molecular detection of resistant mutants and therapeutic options in case of macrolide resistance will also be assessed.


INTRODUCTION
Mycoplasma pneumoniae is responsible for communityacquired respiratory tract infections, such as tracheobronchitis and pneumonia, particularly in school-aged children and young adults. These infections occur both endemically and epidemically at 3-to-7-year intervals worldwide (Atkinson et al., 2008). Numerous extra-respiratory manifestations of variable severity have also been associated with M. pneumoniae infections including dermatological manifestations and neurological complications. Before 2000, M. pneumoniae infections were easily treated using macrolides because only rare cases of resistance to macrolides had been reported in clinical isolates. Since 2000, macrolide resistance rates have been rising up to 90-100% in Asia, hindering the efficacy of common antibiotic regimens.
This mini-review focuses on M. pneumoniae intrinsic resistance, antibiotic susceptibility and on the development and the evolution of acquired macrolide resistance worldwide since the last published review . Methods for molecular detection of macrolide resistance-associated mutations and therapeutic options in case of infections with macrolide-resistant M. pneumoniae strains are also assessed.

ACTIVE ANTIBIOTICS AND INTRINSIC RESISTANCE
Like all microorganisms that lack cell wall, M. pneumoniae is intrinsically resistant to beta-lactams and to all antimicrobials targeting the cell wall, such as glycopeptides and fosfomycin. M. pneumoniae is also resistant to polymixins, sulfonamides, trimethoprim, rifampicin and linezolid (Bébéar and Kempf, 2005;Bébéar et al., 2011). Antibiotics with potential activity against M. pneumoniae that are used in clinical practice include macrolides, lincosamides, streptogramin combinations and ketolides (MLSK), tetracyclines and fluoroquinolones. These drugs achieve high intracellular concentration in mammalian cells and are thereby able to reach intracellular mycoplasmas. The MICs of the main antibiotics belonging to the MLSK group are the lowest against M. pneumoniae compared with those of the two other classes, except MIC of lincomycin that is high (see MIC of the sensitive reference strain M129 (ATCC 29342) in Table 1; Bébéar et al., 2011). MICs of tetracyclines and fluoroquinolones are about 10 times higher than those of MLSK, but newer fluoroquinolones such as levofloxacin and moxifloxacin show an enhanced activity against M. pneumoniae. Only fluoroquinolones and ketolides have a potential bactericidal action. Other antibiotics such as aminoglycosides and chloramphenicol show some activity against M. pneumoniae (MICs 2-10 µg/ml for chloramphenicol and MIC 4 µg/ml for gentamicin, Bébéar et al., 2011) but are not recommended for M. pneumoniae infections.

MECHANISMS OF M. PNEUMONIAE ACQUIRED RESISTANCE AND RESISTANCE MOLECULAR DETECTION METHODS
In M. pneumoniae, only antimicrobial target modifications by acquired mutations have been associated with antibiotic resistance . The high mutation rates and the small amount of genetic information dedicated to DNA repair in mycoplasmas (Rocha and Blanchard, 2002) may be associated with this single mode of antibiotic resistance.
Resistance through mutation was reported in in vitro-selected mutants for all three classes of antibiotics used to treat M. pneumoniae whereas to date, resistance in clinical isolates was only reported for the MLSK antibiotic class.

Acquired Resistance to Macrolides and Related Antibiotics
Macrolide resistance in the M. pneumoniae species, which harbors only one ribosomal operon, is defined by mutations in the ribosomal target of the antibiotic, i.e., the 23S rRNA and the ribosomal proteins L4 and L22 Bébéar et al., 2011). The A2058G (Escherichia coli numbering) transition in the peptidyltransferase loop of domain V of 23S rRNA is the most common mutation that is associated with macrolide resistance (Figure 1, Table 2). Other substitutions have been reported at position 2058 (A2058C, A2058T), at position 2059 (A2059G, A2059C), at position 2062 (A2062G) and at position 2611 (C2611G, C2611A). No mutation has been detected in domain II of 23S rRNA. Mutations in conserved regions of ribosomal L4 and L22 proteins such as single amino acid change, insertion and deletion of amino acids have also been associated with lowlevel macrolide resistance in in vitro selected mutants (Pereyre et al., 2004a). Rare mutations have been reported in vivo in ribosomal proteins L4 and L22 but were not associated with significant increased MICs of macrolides . Comparison of sequencing results with antimicrobial susceptibility testing confirmed that mutations A2058G and A2059G led to a high level resistance to 14-and 15membered macrolides and lincosamides (Xin et al., 2009;Cao et al., 2010;Akaike et al., 2012;Zhao et al., 2013b; Table 1). Whereas 16-membered macrolides were highly affected by the A2059G substitution, the A2058G mutation was associated with an intermediate level of resistance to these antibiotics. Mutations at position 2611 were associated with lowlevel of resistance to MLSK. Interestingly, the streptogramin combinations, quinupristin-dalfopristin and pristinamycin, and the ketolide solithromycin (CEM-101) retained activity on resistant mutants harboring mutations at position 2058, 2059, and 2611 (Pereyre et al., 2007;Waites et al., 2009; Table 1). However, an in vitro mutant selection study showed that the A2062G transition was associated with significant increased MICs of these two streptogramin combinations (Pereyre et al., 2004a). Cross-resistance was not observed between MLSK and other antibiotic families commonly used against M. pneumoniae because isolates with macrolide resistance-associated mutations remain susceptible to tetracyclines and fluoroquinolones ( Table 1).
Several molecular methods applicable directly on respiratory specimens were developed to detect macrolide resistance and to circumvent the fastidious, insensitive and time-consuming isolation of M. pneumoniae from clinical samples. Apart from the conventional amplification and sequencing of the hot spots of the 23S rRNA gene, macrolide resistance determination was achieved by PCR-restriction fragment lengh polymorphism (Matsuoka et al., 2004), real-time PCR and melting curve analysis (Peuchant et al., 2009), pyrosequencing (Spuesens et al., 2010(Spuesens et al., , 2012) and real-time PCR and high resolution melt (HRM) analysis (Wolff et al., 2008). A nested-PCR combined with single-strand conformation polymorphism and capillary electrophoresis (Lin et al., 2010) and a singe nucleotide polymorphism (SNP) PCR assay (Ji et al., 2014) were also developed to detect macrolide-resistant mutants directly from clinical specimens. Most of these in-house approaches allow resistance screening in M. pneumoniae-positive respiratory tract samples but the clinical sensitivity i.e., the proportion of M. pneumoniae-positive specimens capable of being resistance typed varies according to methods, ranging between 72.6 and 80.2% in the studies where it was calculated (Wolff et al., 2008;Peuchant et al., 2009;Spuesens et al., 2012). The need to perform such tests differs according to the prevalence of macrolide resistance in each country. In countries where the percentage of macrolide resistance is over 10%, it could be recommended that all M. pneumoniae detection be followed up with an assay capable of detecting macrolide resistanceassociated mutations. This strategy would allow a non-macrolide treatment to be promptly started in the event that a macrolideresistant genotype is detected in an individual patient. In contrast, in countries where macrolide resistance remains below 10%, this kind of test could be performed only in case of treatment failure.
Currently, this strategy is hampered by the lack of commercially available sensitive kits that detect macrolide resistance-associated mutations. However, such kits are currently in development and may soon become available. They will be useful for routine diagnostics in microbiology laboratories.

Acquired Resistance to Tetracyclines and Fluoroquinolones
To date, no tetracycline or fluoroquinolone resistance has been reported in M. pneumoniae clinical isolates. However, resistant strains have been selected in vitro for both classes of drugs. Target mutations were identified in the 16S rRNA gene of tetracyclineresistant mutants selected with subinhibitory concentrations of doxycycline. Mutations were associated with reduced susceptibility to tetracycline, doxycycline and minocycline with MICs remaining below ≤2 µg/ml (Degrange et al., 2008). Mutations within conserved regions of the gyrA, gyrB, parC, and parE genes referred to as the quinolone resistancedetermining regions were reported for fluoroquinolone-resistant mutants selected with different fluoroquinolones and were associated with MICs of ciprofloxacin, levofloxacin and moxifloxacin up to 32, 16, and 4 µg/ml, respectively (Gruson et al., 2005). Mutations rates were low for levofloxacin and moxifloxacin, ranging from 1.3 × 10 −6 to 7 × 10 −9 (Gruson et al., 2005).

PREVALENCE OF MACROLIDE RESISTANCE IN M. PNEUMONIAE
Recent rates of macrolide resistance in M. pneumoniae clinical isolates in countries in which publications have been released since the last review  are presented in Table 2. Prior to the year 2000, very few M. pneumoniae clinical isolates were resistant to macrolides. Rare strains resistant to erythromycin were reported in the literature between 1968 and 1999 in Japan, Israel, Finland, USA and France (Niitu et al., 1970;Stopler and Branski, 1986;Critchley et al., 2002;Pereyre et al., 2007). By contrast, several Japanese studies have reported a significant and constant increase in macrolide resistance rates since 2000, reaching 30% in 2006, around 60% in 2009and up to 89% in 2010(Morozumi et al., 2008Okada et al., 2012;Matsuda et al., 2013). However, regional differences in rates of macrolide-resistant M. pneumoniae were recently reported in Japan, for example in Hokkaido island, where rates ranged from 0 to 100% according to regions (Ishiguro et al., 2015). The situation is worse in China where a dozen of articles have reported a prevalence of macrolide resistance   Table 2). It should be noted that most reports regarding macrolide resistance relate on hospitalized patients. It cannot be excluded that the macrolide resistant rate in M. pneumoniae may be higher in hospitalized patients in whom the resistant population may be concentrated than in outpatients. However, comprehensive studies on outpatients are not easily achievable because many M. pneumoniae infections such as mild tracheobronchitis are often undiagnosed.
The high macrolide resistance rates in these countries are certainly associated with antibiotic selective pressure because of extensive macrolide use. This is supported by the highest macrolide resistance rates being reported in countries with extensive macrolide use such as Japan (Okada et al., 2012). In addition, macrolide resistance was often associated with recent receipt of macrolides, suggesting that a resistant subpopulation may develop or expand during the course of macrolide therapy within an individual patient (Averbuch et al., 2011;Cardinale et al., 2011;Chironna et al., 2011;Hantz et al., 2012;Dumke et al., 2014). Acquisition of resistance has first been documented in patients receiving macrolides (Averbuch et al., 2011;Cardinale et al., 2011) then confirmed using typing methods such as adhesin P1 typing and multi-locus variable-number tandemrepeat analysis (MLVA) in patients receiving macrolides (Hantz et al., 2012;Dumke et al., 2014).
In North America, Europe, and Australia, rates of macrolide resistance dramatically contrast with those in reports from Asia. In the USA and Canada, rates have been recently reported between 3.5 and 13.2% ( Table 2). In Europe, rates have remained below 10% except in Italy were a rate of 26% was observed on a small number of M. pneumoniae-positive specimens collected during an outbreak .
All over the world, the A2058G transition largely predominates over the A2059G substitution and mutations at position 2611 and 2062 are rare ( Table 2). Nevertheless, the rarely reported A2058T transversion was found in 47% of macrolide-resistant M. pneumoniae strains infecting children during an outbreak in Fukuoka, Japan (Matsuda et al., 2013). Despite the high proportion of the A2058G transition, no association was reported between MLVA types and macrolide resistance in several studies (Dégrange et al., 2009;Benitez et al., 2012;Liu et al., 2012;Zhao et al., 2013a,b;Dumke et al., 2015;Diaz et al., 2015a,b) indicating that macrolide resistance is a result of the spread of multiple resistant clones. A possible correlation was reported in Jerusalem, Israel, between the MLVA type Z (7-4-5-7-2) and the A2058G-associated macrolide resistance but the number of cases was limited (Pereyre et al., 2012). Recently, an association between macrolide resistant M. pneumoniae isolates and the MLVA type 4-5-7-2 was suggested in China and Hong-Kong (Ho et al., 2015;Yan et al., 2015). However the prevalence of this MLVA type was high in these countries and the deletion of the unstable MPN1 marker from the MLVA method  led to a too weakly discriminant typing method to draw accurate conclusions.

CLINICAL RELEVANCE OF M. PNEUMONIAE MACROLIDE RESISTANCE
Regarding clinical presentation, no difference was observed between patients infected by macrolide-resistant and macrolidesensitive M. pneumoniae. Clinical symptoms, pneumonia severity, laboratory results, radiographic findings and prognostic factors were similar regardless of the M. pneumoniae susceptibility to macrolides (Matsubara et al., 2009;Cardinale et al., 2013;Miyashita et al., 2013;Wu et al., 2013;Diaz et al., 2015a). Most infections with macrolide-resistant M. pneumoniae have been reported in children because M. pneumoniae infections are more frequent in this population. Nevertheless, several adults have also been evaluated Ferguson et al., 2013;Ho et al., 2015;Diaz et al., 2015a). To date, no difference has been found in disease manifestations between children and adults infected by macrolide-resistant M. pneumoniae.
As expected, the efficacy of macrolide treatment was shown to be lower in patients infected with macrolideresistant isolates than in patients infected with macrolidesensitive isolates. Despite macrolide administration, the duration of fever and cough, the duration of hospitalization and antibiotic administration were significantly longer in patients with macrolide-resistant M. pneumoniae infections. Moreover, the persistence of symptoms led to change of antibiotic prescription more often (Suzuki et al., 2006;Morozumi et al., 2008;Matsubara et al., 2009;Cardinale et al., 2013;Wu et al., 2013;Zhou et al., 2014). However, the clinical relevance of macrolide resistance in patients was usually limited to prolonging symptoms of the disease and not increasing the risk of complications. Only a single study has reported that the incidence of extrapulmonary complications was higher in children with macrolide-resistant isolates and that the radiological findings were more serious (Zhou et al., 2014).

TREATMENT OF M. PNEUMONIAE RESPIRATORY INFECTIONS
Macrolides and related antibiotics are the first-line treatment of M. pneumoniae respiratory tract infections mainly because of their low MIC against the bacteria, their low toxicity and the absence of contraindication in young children. The agent of first choice differs from country to country according to different published guidelines and owing to the fact that not all agents are available in all countries (Mandell et al., 2007;Bradley et al., 2011;Harris et al., 2011;Woodhead et al., 2011;Waites and Bébéar, 2013). The newer macrolides are now often the preferred agents with a 7-to-14 day course of oral clarithromycin or a 5-day course of oral azithromycin for treatment of community-acquired pneumonia due to M. pneumoniae (Waites and Bébéar, 2013). An appropriate antimicrobial therapy usually shortens the symptomatic period of M. pneumoniae infections, and hastens radiological resolution and recovery. However, using real-time PCR, it has been shown that the median time for carriage of M. pneumoniae DNA was 7 weeks after disease onset and that an adequate antibiotic treatment did not shorten the period of persistence of M. pneumoniae DNA in patient specimens (Nilsson et al., 2008). No treatment recommendation is available for extrapulmonary manifestations. In a few published case reports, macrolides and fluoroquinolones, mainly levofloxacin, have successfully been used (Scapini et al., 2008;Atkinson et al., 2011;Esposito et al., 2011;Meyer Sauteur et al., 2012;Godron et al., 2013).
In cases of macrolide-resistant M. pneumoniae strains, alternative antibiotic treatment can be required, including tetracyclines such as doxycycline and minocycline, or fluoroquinolones, primarily levofloxacin, even though fluoroquinolones and tetracyclines are contraindicated in all children and in children <8 year-old, respectively. Treatment lengths usually range between 7 and 14 days. As expected, fluoroquinolone and tetracycline regimens were shown to be more effective than macrolide regimens in patients infected by macrolide-resistant M. pneumoniae (Kawai et al., 2013;Miyashita et al., 2013). However, macrolides appear clinically effective in some patients infected by macrolide-resistant strains (Suzuki et al., 2006;Matsubara et al., 2009;Cardinale et al., 2013). This observation can be explained by the fact that M. pneumoniae infections are often self-limited diseases and that the antiinflammatory effects of macrolides may improve clinical symptoms.
In Europe, Oceania, and America, where the prevalence of macrolide-resistant strains remains low, macrolides are the drug of choice in children with M. pneumoniae respiratory infections. Nevertheless, in these continents, clinicians should be vigilant for macrolide treatment failure and consider using alternative drugs if symptoms persist or if there are signs of clinical deteriorations. In countries in which the prevalence of macrolide-resistant strains is high, the replacement of macrolides as the first choice treatment by tetracyclines or fluoroquinolones was considered. However, surprisingly, in Japan, according to the 2013 recommendations of the Japanese Pediatric Society, macrolides remain the first-line treatment despite macrolide resistance rates over 80%. In this country, the efficacy of macrolides has to be evaluated by defervescence 48-72 h following the administration of these antimicrobials. In pneumonia cases in which the initial macrolide therapy resulted in failure, administration of alternative antimicrobial treatment, either respiratory fluoroquinolones or tetracyclines, must be considered. In contrast to Europe and to the United States, oral tosufloxacin, a fluoroquinolone antibiotic, was approved in Japan for pediatric use as a second line treatment in patients with community-acquired pneumonia. Indeed, in one study performed for the registration application of tosufloxacin in Japan, the occurrence of joint paint was only 0.85% (2/235) and there was no magnetic resonance imaging abnormal finding on joints (data given by Dr T. Oishi, Japan). Another study on 83 pediatric patients with M. pneumoniae pneumonia treated with tosufloxacin reported that side effects included mild diarrhea, but that no patients had joint symptoms (Sakata, 2012). Although, tosufloxacin was les effective than minocycline or doxycycline in achieving defervescence within 24 h and in decreasing the DNA load of M. pneumoniae (Okada et al., 2012;Kawai et al., 2013), its use is accepted in children under 8-year old. In countries where tosufloxacin is not available, other available respiratory fluoroquinolones might be chosen in severe cases despite contraindication. In children over 8year old and adults, minocycline can be used as second-line treatment.
Although, no tetracycline or fluoroquinolone resistance has been reported in clinical isolates to date, resistant strains have been selected in vitro for both classes of drugs with target mutations identified in mutants (Gruson et al., 2005;Degrange et al., 2008). Thus, the risk of emergence of resistance in clinical isolates exists, especially for fluoroquinolones, if these antibiotics are inappropriately used. It should be noted that clinical resistance to fluoroquinolones has already been reported already in Mycoplasma genitalium, a urogenital mycoplasma phylogenetically close to M. pneumoniae, in which macrolide resistance mechanisms are similar to that of M. pneumoniae (Couldwell et al., 2013;Bissessor et al., 2015).
Consequently to macrolide resistance in M. pneumoniae, reevaluation of existing classes using and investigation of new classes of antimicrobials may be required to get additional treatment alternative beyond tetracyclines and fluoroquinolones, especially in children under 8 year-old. Randomized therapeutic trials will be necessary to establish guidelines regarding the most appropriate molecule, dose and length of treatment to use against the resistant strains. In the future, it will also be interesting to evaluate the activity of streptogramin combinations, such as oral pristinamycin, which has been shown to retain activity against 23S rRNA M. pneumoniae in in vitro mutants and in a few clinical isolates (Pereyre et al., 2004a(Pereyre et al., , 2007. Indeed, pristinamycin was reported to be active on a few cases of genital infections by macrolide-resistant fluoroquinolone-resistant M. genitalium isolates (Bissessor et al., 2015). Although, additional studies on a large number of strains are required, pristinamycin could become an alternative antibiotic treatment in countries where this antibiotic is available (Bebear, 2012).

CONCLUSION
Nowadays, M. pneumoniae macrolide resistance rates are extremely high in Asia and remain moderate in Europe and North America. Macrolide resistance detection using accurate molecular methods should be considered in all M. pneumoniae-positive specimens since it has both a direct application in clinical practice and an epidemiological surveillance interest. At the individual level, a rapid detection of resistance-associated mutations would enable the prompt prescription of an alternative antimicrobial regimen, especially in case of persistent or recurrent M. pneumoniae infection. At the community level, the high prevalence of macrolideresistant M. pneumoniae isolates in Asia underscore the potential for rapid emergence of macrolide resistance within M. pneumoniae in other parts of the world. Thus, further epidemiological studies are needed in Europe and the USA to monitor macrolide resistance rates. Moreover, macrolide stewardship may be needed for restricting the use of these antibiotics, reduce unnecessary antibiotic prescribing, especially in countries with remaining low macrolide resistant rates. In Asia, the epidemiological surveillance of antibiotic resistance would also be of interest to early detect potential selections of fluoroquinolone-and tetracycline-resistant clinical isolates associated with the increasing use of these classes of antibiotics.

AUTHOR CONTRIBUTIONS
All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication.