Introduction

Sixteen Mycoplasma species have been isolated from human respiratory and urogenital specimens [1]. One of these species, M. pneumoniae, is a main pathogen in respiratory tract infections (RTI) acquired in the community. In school-aged children and young adults with community-acquired pneumonia (CAP), M. pneumoniae accounts for as many as 10–30% of cases [2–8]. In distinction, M. pneumoniae pneumonia in younger children and in the elderly is infrequent but not rare [9–15].

M. pneumoniae pneumonia is diagnosed based on characteristic chest radiographic abnormalities, patient symptoms, and clinical laboratory data. Diagnosis ultimately is confirmed using serologic tests performed upon paired sera obtained during both acute and convalescent phases. Conventional culture using pleuropneumonia-like organism (PPLO) broth, which requires more than 2 weeks, has not been carried out routinely. As a consequence, antibiotic choice usually is empirical.

Development of molecular methods such as polymerase chain reaction (PCR) assays to be used in combination with conventional diagnostic tests using serology and culture have contributed to improved diagnosis and characterization of M. pneumoniae infection in pediatric and adult patients [16–28]. Macrolides (ML) such as clarithromycin (CAM) and azithromycin (AZM) generally are recognized as first-choice agents against M. pneumoniae [29–33]. However, in parallel with increased use of oral ML with a 14-membered ring (14-ML) and AZM for RTI, M. pneumoniae showing resistance to ML has been isolated increasingly in clinical samples from Japanese pediatric patients with CAP [31, 34, 35]. Prevalence of ML-resistant (MLr) M. pneumoniae in pediatric patients has increased rapidly [35, 36] and MLr M. pneumoniae is beginning to be isolated worldwide [37–41].

In this review, we summarize the emergence and increase of MLr M. pneumoniae isolates, resistance mechanisms, and clinical features of MLr M. pneumoniae infection.

Diagnosis of infection

M. pneumoniae infection tends to show cyclic epidemics every 3–5 years; these outbreaks are particularly likely to occur in summer or early fall [42–46]. Although the incidence of M. pneumoniae pneumonia is highest in children 5–9 years old, such pneumonia also occurs frequently in younger children and in the elderly. Prevalence of M. pneumoniae infection also may vary according to population and diagnostic methods used.

Isolation of M. pneumoniae by culture from clinical samples such as throat or nasopharyngeal swabs was considered standard for diagnosis several years ago. However, M. pneumoniae is difficult to grow in liquid culture; cultures using PPLO broth require at least 2 weeks for completion. Compared with serologic tests or molecular techniques, including PCR, sensitivity of cultures may be <60% or 70%, even in laboratories with expertise [6, 18, 20]. Therefore, culture methods are only used rarely for routine diagnosing M. pneumoniae infection.

Serologic methods are used most frequently to diagnose M. pneumoniae infection. Frequently used, widely available serologic tests for M. pneumoniae include complement fixation (CF), enzyme immunoassay (EIA), and particle agglutination (PA) assays. However, these ordinarily require paired sera obtained during acute and convalescent phases in order to demonstrate rises in antibody titers; fourfold increase is thought to be significant. The second sample should be obtained 7–14 days after the first, provided that M. pneumoniae infection still is suspected.

Although one alternative, the ImmunoCard based on an immunoglobulin M (IgM) assay, is rapid and easy to perform, reliable diagnosis of M. pneumoniae infection still cannot be made on the basis of single acute-phase sera; confirmation using convalescent-phase samples still is necessary, as otherwise, both false-positive and false-negative results are frequent [24, 47, 48]. Additionally, sensitivity of IgM assays tends to be low, particularly for adult patients who are known to have weak IgM responses during primary infection or reinfection [49, 50]. Neither culture nor serologic testing can provide timely information for guiding choice of chemotherapeutic agents to use for early intervention. Serologic methods generally have only low sensitivity during the acute phase of disease.

Development of molecular methods (e.g. PCR) has lessened the importance of culture as a way to detect M. pneumoniae directly. Two specific targets in the P1 adhesion gene and the 16S ribosomal RNA (rRNA) gene are the main ones used in PCR assays to detect M. pneumoniae DNA. Various PCR methods have been developed in several laboratories. M. pneumoniae detection by conventional PCR or real-time PCR has been examined as an approach to early diagnosis. Many studies have described application of multiplex PCR, hybridization assays, nucleic acid sequence-based amplification (NASBA), real-time PCR using a molecular beacon probe or TaqMan probe, or cycling probe, to the identification of M. pneumoniae DNA [16–28]. Among these methods, real-time PCR has both high sensitivity and high specificity and can detect pathogen DNA even when damaged by empirical administration of antibiotics. Sensitivity (60–100%) and specificity (96.7–100%) of real-time PCR are both higher than those of serologic assays for M. pneumoniae [18, 19, 22, 24–26]. Almost all PCR-positive cases (>90%) also were confirmed serologically [22, 24]. In addition, real-time PCR using a fluorescent probe allows continuous monitoring of in vitro DNA amplification, eliminating nonspecific amplification product using fluorescence; no gel electrophoresis is needed. This makes the method suitable for clinical laboratory settings. A molecular approach including real-time PCR is useful for confirming M. pneumoniae and providing rapid diagnosis of CAP, thus permitting rational choice of effective antimicrobial agents.

Mechanisms of ML resistance in M. pneumoniae

Intrinsically, absence of cell walls in M. pneumoniae confers resistance to β-lactams, and indeed to all antibiotics that inhibit synthesis of cell walls. In M. pneumoniae infections, 14-ML and 15-membered ring ML (15-ML) usually are considered the first-line agents. M. pneumoniae ordinarily are susceptible to all ML, including ketolides.

Main mechanisms of microorganism resistance to ML identified in previous studies include modification of the target sites in 23S rRNA by methylation or mutation, as well as on efflux pump [51–53]. ML act on the 50S ribosomal subunit to inhibit protein synthesis. The site of peptide bond formation on the large 50S ribosomal subunit is associated with the central loop in domain V of 23S rRNA. Nucleotides present in the central loop are necessary constituents of the ML binding site, which implies that all ML bind to the same site (Fig. 1a). A loop of hairpin 35 in domain II also may be part of binding site of ketolides; mutations involving ribosomal proteins L4 and L22 may contribute as well [51–54].

Fig. 1
figure 1

Binding site of macrolides on the 23S ribosomal RNA. a Secondary structure of the large subunit of ribosomal RNA (Escherichia coli; http://www.rna.ccbb.utexas.edu/). b Central loop in domain V (Mycoplasma pneumoniae numbering) and the loop of hairpin 35 in domain II (E. coli numbering) of 23S rRNA as modified according to [34, 51]. Circles indicate positions of mutations associated with macrolide resistance in clinical isolates. The square shows the position of a substitution associated with 16-membered-ring macrolide resistance in a mutant selected in vitro

Full size image

ML inhibit protein synthesis mainly by binding to domain V of 23S rRNA at nucleotide positions 2063 and 2064 (M. pneumoniae numbering); these positions appear essential for binding (Fig. 1b) [55]. Mutations at A2063 or A2064 confer the highest resistance to these antimicrobials. A lower level of antibiotic resistance is caused by mutations at positions A2067 and C2617, which are near A2063 and A2064 in the secondary structure, although slightly present outside the central point of ML interaction. The A2067G mutation was reported to cause highest levels of resistance only for 16-membered-ring ML (16-ML), as it interferes with formation of the covalent bond specific to 16-ML [56].

Ketolides are a new class of antimicrobial agent characterized by the substitution at position 3 from l-cladinose to keto group. These antibiotics are active against Streptococcus pneumoniae strains resistant to ML because of interaction at both A2058/2059 (Escherichia coli numbering) in domain V and A752 in domain II [51, 54, 57]. Substitutions in ribosomal proteins L4 and L22 have been found to confer resistance in S. pneumoniae [58, 59], but mutations in domain II of 23S rRNA and in ribosomal proteins L4 and L22 are not involved in ML resistance of clinical M. pneumoniae isolates [60]. Furthermore, neither plasmids nor erm genes mediating ribosomal modification have been described with respect to M. pneumoniae.

Fluoroquinolone (FQ)-resistant M. pneumoniae mutants may be obtained in vitro by exposure to subinhibitory concentrations of FQ [61]. These mutants possess mutations in gyrA and gyrB genes, which express topoisomerase II, and in parC and parE genes, which express topoisomerase IV; both of these enzymes are essential for bacterial DNA replication. FQ often are chosen as empiric treatment of respiratory tract infection in adults, as they are active against the main pathogens, including S. pneumoniae, Haemophilus influenzae, and M. pneumoniae. In contrast to in vitro findings, no FQ-resistant strains have been observed in clinical settings.

Tetracyclines inhibit protein synthesis by preventing association of aminoacyl-tRNA to the acceptor (A) site of the ribosome [62, 63]. These antibiotics bind to the ribosomal 30S subunit involving 16S rRNA and several ribosomal proteins. Only the tetM determinant associated with conjugative transposon Tn916, which confers resistance to the tetracyclines in M. hominis, has been documented so far [64]. These agents are administered for treatment of M. pneumoniae infection in adults and in pediatric patients >8 years, but development of resistance to tetracyclines has not yet been reported.

Association of ML resistance and mutation site in 23S rRNA from MLr M. pneumoniae

Specific sites of the 23S rRNA mutation contribute to ML resistance in M. pneumoniae. In actuality, based on our results, clinical isolates with mutations at positions A2063 and A2064 show high resistance to 14-ML and 15-ML, with minimal inhibitory concentration (MIC) ≥ 32 μg/ml (Table 1). Especially, the A2063G mutation conveys the highest level of resistance to 14-ML, 15-ML, and telithromycin (TEL) but intermediate resistance to josamycin and rokitamycin belong to 16-ML. The A2063C and A2064G mutations result in a high level of resistance to all ML [36, 65]. The ketolide TEL is affected by both mutations, although A2063G was associated with MICs higher than those associated with A2064G.

Table 1 Antimicrobial activity of macrolides, fluoroquinolones, and minocycline against Mycoplasma pneumoniae strains isolated from clinical samples
Full size table

MICs for minocycline (MINO) and FQ in MLr strains were equivalent to those in susceptible strains. No strains having resistance to MINO and FQ have been observed among clinical isolates.

Several strains with the C2617 to G or A mutations show slight increases in MIC of ML compared with those with mutations at positions 2063 and 2064. A2063G in domain V of 23S rRNA is the most frequent mutation association with ML resistance, followed by A2064G; others (A2063C, C2617G, and C2617A) are rare.

Ribosomal protein L4 and L22 mutants with a few amino acids deleted or inserted have been described, but roles of these mutations in resistance to ML are uncertain [60]. No mutation has been detected in domain II of 23S rRNA.

Emergence and increase of MLr strains

MLr M. pneumoniae isolates possessing a nucleotide mutation in 23S rRNA first were isolated from pediatric patients with CAP, as reported by Okazaki and colleagues in 2001 [34].

Together with growing use of ML in Japan, resistant strains increased rapidly year by year (Fig. 2) [36]. Along with greater overall prevalence of M. pneumoniae infection in 2006, MLr strains increasingly were isolated and identified in diverse regions across Japan.

Fig. 2
figure 2

Increases in numbers of macrolide-resistant Mycoplasma pneumoniae strains from 2002 to 2008 in Japan according to our analyses. Asterisks indicate macrolide-resistant strains isolated from adult patients. The dotted line can be expressed by the equation y = 7x − 8.62 (r = 0.9605)

Full size image

In Japanese adult CAP patients, MLr strains first were isolated in 2007; since then, occasional isolations have continued [66]. An increase in MLr strains among adult patients, las in pediatric patients, may occur in the future. However, FQ are recommended for CAP treatment in adults, and no FQ-resistant strains have been observed among isolates from adult patients with CAP.

Emergence of MLr isolates having either an A2063G or an A2064G mutation has been reported not only in Japan but also in other countries, including France, the USA, Denmark, and China [37–41]. A notably high isolation rate (92%) of MLr strains was reported in China [38, 40].

The emergence and increase of MLr M. pneumoniae has attracted the attention of microbiologists worldwide. Several recent studies demonstrated direct detection of MLr isolates using real-time PCR [37, 41, 67].

DNA profiles of MLr- and ML-susceptible (MLs) strains obtained by pulsed-field gel electrophoresis (PFGE) have suggested classification into two groups [36, 68]. Significant differences in these DNA patterns were not recognized between strains, geographic locations, or years.

PCR-based restriction fragment length polymorphism (RFLP) types of the P1 adhesin gene also were used to divide strains into two types. MLr isolates were present in both types [38, 39, 69, 70], showing no predominance.

Clinical aspects of MLr M. pneumoniae infection

The clinical course of patients with MLr M. pneumoniae infection appears to be prolonged [36, 66, 71, 72]. Patients infected with MLr strains mostly had been treated with ML previously [36, 40]. Among CAP patients infected with MLr, treatment frequently was changed from ML to MINO or levofloxacin (LVFX) because of either persistent symptoms (i.e., fever and cough) or unresolved or worsening chest radiographic abnormalities. In our observations, treatment was changed from ML to MINO or LVFX in 8.2% of patients with MLs strains and in 39.7% of patients with MLr strains, representing a significant difference between the two patient groups (P < 0.01). ML also was more frequently discontinued in favor of MINO among MLr patients than MLs patients in a report by Suzuki et al. [71]. Neither LVFX nor MINO are recommended for pediatric patients. However, when ML are ineffective against M. pneumoniae infection, pediatricians have little choice of antimicrobials except for MINO. In adult inpatients infected with MLr strains, antibiotic agents were changed from CAM to intravenous pazufloxacin when symptoms did not improve [66]. Orally administered respiratory FQ such as moxifloxacin, sitafloxacin, and garenoxacin may be selected for adult outpatients with MLr strains.

Among clinical features of MLr M. pneumoniae infection, fever duration has been significantly longer than for MLs-strain infection. In our data, fever persisted after ML initiation for 1.6 ± 0.8 days in MLs M. pneumoniae infection and for 4.1 ± 2.3 days in MLr-strain infection, showing a significant difference (P < 0.01). In other studies, febrile days during ML administration also were significantly greater in MLr-infected patients than in MLs-infected patients (3.5–4.0 days vs. 1.0–1.5 days) [71, 72]. Furthermore, the mean duration of persistent cough after ML administration was 7.0 days in MLs-infected patients and 11.4 days in MLr-infected patients [72]. Finally, the efficacy rate of ML therapy was 91.5% and 22.7% in MLs- and MLr-strain infections, respectively [72].

Fever may have resolved spontaneously in some patients with continuing ML treatment for MLr infections; M. pneumoniae infection is associated with occasional spontaneous symptomatic recovery. According to previous reports, fever duration was 1 week in symptomatically treated M. pneumoniae infection [73, 74]. In another study, mean fever duration in hospitalized patients receiving AZM or EM was 2.1 days, whereas fever persisted for about 1 week in patients not receiving ML [75]. Similarly, mean cough duration was reported as 8.5 days [76]. Although treatment failure or serious illness has not yet been attributed to MLr-strain infection, the clinical course of patients infected with MLr strains may be prolonged.

A brief consideration of the host response is helpful here. M. pneumoniae resides on the surface of ciliated human respiratory epithelial cell. Several membrane proteins in M. pneumoniae have high affinity for various surface receptors on host cells [77]. Through interaction with recognition receptors, including toll-like receptors 2 and 6, mycoplasma membrane lipoproteins are able to induce host immune responses [77, 78]. Adherence of M. pneumoniae to host cells in the respiratory tract is mediated by the P1 adhesin protein and accessory proteins [79–82]. Following adherence, macrophages become activated and release cytokines, and a mononuclear cell inflammatory response is established. Human lung epithelial cells infected with M. pneumoniae induce expression of interleukin (IL)-8, which is both a potent chemoattractant and an activator of neutrophils, monocytes, and T lymphocytes [83–85]. In particular, significant production of IL-8 and IL-18, a cytokine that induces interferon gamma (IFN-γ) production and promotes a type 1 cytokine response, is associated with the severity of M. pneumoniae pneumonia in children [86–88] and adults [89].

As previously described, the inflammatory response to M. pneumoniae infection is considered to play a crucial role in the pathogenesis of the ensuing clinical disease. ML possess antimicrobial activity against M. pneumoniae, and also have anti-inflammatory effects. For example, CAM is able to suppress IL-8 induction, even in MLr M. pneumoniae infection [78, 86, 90]. Thus, clinical efficacy of ML in treating M. pneumoniae infection may reflect not only direct antimicrobial activity but also anti-inflammatory effects of inhibition of production of cytokines, including IL-8.

Symptoms and severity of illness due to M. pneumoniae were similar in younger and older patients, and the mortality rate was low, even in the elderly [13, 91, 92]. This disease ordinarily is mild, and reinfection may occur in adults who already experienced M. pneumoniae infection, as protective immunity usually is not established following initial infection [93]. However, symptoms appear to be prolonged in certain clinical situations regardless of antimicrobial resistance [15, 94]. Takahashi and colleagues reported M. pneumoniae DNA in sputum after 2 weeks of MINO therapy. Elderly persons with M. pneumoniae pneumonia may have a longer clinical course than younger persons. Further observations from various patients infected with M. pneumoniae are necessary.

In the future, alternative treatment strategies, such as promptly initiated steroid therapy, might be considered for patients with MLr strains as a symptomatic measure [95–97].

Conclusions

In pediatric patients with CAP, MLr M. pneumoniae possessing a 23S-rRNA mutation was first isolated in 2000; numbers of these isolates have increased rapidly year by year in Japan. In other countries, MLr M. pneumoniae also is emerging. Symptoms often are relatively prolonged in patients with MLr M. pneumoniae infection. MLr M. pneumoniae should be considered an emerging issue for not only pediatric patients but also adult patients with CAP.

No FQ-resistant M. pneumoniae strain has been observed thus far among specimens from adult patients with CAP. However, emergence of infections with FQ-resistant M. pneumoniae might occur, considering the increasing FQ prescription rate among adult patients. Further studies are needed to clarify the prevalence of MLr M. pneumoniae in pediatric and adult patients with RTI and to establish clinical guidelines regarding the most appropriate antimicrobial agents to use against these strains. Interventions to prevent emergence and increases of novel antibiotic-resistant strains are needed, including rapid identification of pathogens by methods such as real-time PCR, as well as antimicrobial treatments capable of eliminating the organisms.