ABSTRACT
Prevalence, trends, and treatment outcome estimates were generated for parC variants in macrolide-resistant Mycoplasma genitalium. Among 539 cases, the most common amino acid change was S83I, which increased from 13% in 2012 to 2013, to 23% in 2019 to 2020 (Ptrend = 0.046). From 381 moxifloxacin treatments, failure occurred in 58.7% (95% confidence interval [CI], 46.7 to 69.9) of cases with S83I. Other changes affecting S83 or D87 were uncommon and minor contributors to failure. The absence of S83I was highly predictive of moxifloxacin cure (96.4%; 95% CI, 93.7 to 98.2), highlighting diagnostic potential.
KEYWORDS: antibiotic resistance, Mycoplasma genitalium, fluoroquinolone, moxifloxacin, parC
INTRODUCTION
Rapidly escalating azithromycin resistance has been observed for Mycoplasma genitalium (1). An understanding of the mechanisms of azithromycin resistance has enhanced treatment outcomes through molecular assays that detect macrolide resistance mutations that are used in resistance-guided treatment strategies (2, 3). Notably, molecular tests are the only practical option for M. genitalium infection, as performing routine culture and antibiotic MIC analysis is not possible.
There is now increased dependence on second-line therapies, particularly the fluoroquinolone moxifloxacin. In contrast to azithromycin, the molecular basis of moxifloxacin resistance is poorly understood. Several single nucleotide polymorphisms (SNPs) of the parC gene affecting the serine amino acid at position 83 (S83) and aspartic acid at position 87 (D87) may contribute (4–8) (note that M. genitalium numbering is used). Due to the impracticalities of bacterial culture, in vitro MIC data for moxifloxacin resistance are very limited. Most information has been obtained from clinical studies examining treatment failures; however, these clinical data sets are generally too small to provide robust estimates of the impact of individual variants (4–6).
A better understanding of moxifloxacin resistance will assist the development of new molecular diagnostic tests to inform patient management. This study examined prevalence and trends of specific parC SNPs in macrolide-resistant M. genitalium clinical samples collected between 2012 and 2020 to further understand associations with moxifloxacin treatment outcomes.
Data were sourced from patients who were treated for macrolide-resistant M. genitalium infections and attended Melbourne Sexual Health Centre, Melbourne, Australia, between 2012 and 2020. These data represent three different groups, (i) patients attending from July 2012 to June 2013 (treatment with moxifloxacin [400 mg/day for 10 days] after failure of azithromycin [1-g single dose]) (6, 9), (ii) patients attending from June 2016 to June 2018 {sequential doxycycline (100 mg twice daily [bd] for 7 days) followed by moxifloxacin (400 mg daily for 7 days)} (2, 3, 7), and (iii) patients attending from August 2019 to December 2020 (sequential doxycycline [100 mg bd for 7 days] followed by combination doxycycline-moxifloxacin [doxycycline, 100 mg bd, and moxifloxacin, 400 mg daily, for 7 days]) (8). Microbiological failure was defined as a positive test of cure (recommended 14 to 28 days after the last dose). Demographics can be found in Table 1. Patients had not previously received moxifloxacin for the current infection and did not have a high risk of reinfection (e.g., participants who reported no sex since treatment, 100% condom use for sex with all partners, condomless sex with only new partners, or condomless sex with a fully treated partner). Detailed information and ethics approval are available in the parent studies (3, 6, 8, 9). Sequence data were obtained previously (6–8), and the nucleotide and amino acid sequence variations are referred to in comparison with the M. genitalium strain G37 (GenBank accession no. NC_000908).
TABLE 1.
Participant characteristics from the parent study groupsa
| Characteristic | Data for study group: |
||
|---|---|---|---|
| 2012–2013 (n = 54) | 2016–2018b (n = 321) | 2019–2020 (n = 164) | |
| Age (median [IQR] [yrs]) | 29.3 (24.9–33.3)c | 28.6 (25.1–34.4) | 30 (26–34) |
| Sex (no. [%]) | |||
| Female | 6 (11.1) | 51 (15.9) | 39 (23.8) |
| Male, heterosexual | 32 (59.3) | 45 (27.4) | |
| Male, MSM | 16 (29.6) | 78 (47.6) | |
| Male, not otherwise defined | 267 (83.2) | ||
| Other/unknown | 3 (0.9) | 2 (1.2) | |
| HIV status (no. [%]) | |||
| Positive | NA | NA | 8 (4.9) |
| Negative | NA | NA | 152 (92.7) |
| Unknown | NA | NA | 4 (2.4) |
| Symptomatic (no. [%]) | |||
| Yes | 52 (96.3) | NA | 106 (64.6) |
| No | 2 (3.7) | NA | 55 (33.5) |
| Unknown | 4 (1.8) | ||
| Infection site (no. [%]) | |||
| Cervix/vagina | 2 (3.7) | 45 (14.0) | 32 (19.5) |
| Urine/urethral | 48 (88.9) | 209 (65.1) | 106 (64.6) |
| Rectum | 4 (7.4) | 64 (19.9) | 25 (15.2) |
| Multiple sites | 1 (0.6) | ||
| Unknown | 3 (0.9) | ||
NA, not available; IQR, interquartile range; MSM, men who have sex with men.
The 2016 to 2018 study group comprises two subgroups treated with either moxifloxacin or sitafloxacin.
Age data only available for 27 out of 54 participants in this group.
Overall, the most common SNP at 18.7% (95% confidence interval [CI], 15.5 to 22.3) was G248T conferring amino acid change S83I, which increased in prevalence from 13% (95% CI, 5.4 to 24.9) in 2012 to 2013, to 23% (95% CI, 17.0 to 30.4) in 2019 to 2020 (Ptrend = 0.046) (Table 2). Other SNPs affecting ParC S83 (S83R, S83C, and S83N) were rarely detected, with individual overall prevalence ranging from 0.6 to 1.1% and no apparent changes across the time points (Fig. 1).
TABLE 2.
Prevalence of single nucleotide polymorphisms in parC in pretreatment samples collected from patients with macrolide-resistant Mycoplasma genitalium infection, Melbourne, Australia, from 2012 to 2020
| parC SNPa | Amino acid change | No. (%) of SNPs detected in study group: |
Overall SNP detection |
|||
|---|---|---|---|---|---|---|
| 2012–2013 (n = 54) | 2016–2018 (n = 321) | 2019–2020 (n = 164) | No. (%) (n = 539) | 95% CI | ||
| G241A | G81S | 0 | 1 (0.3) | 0 | 1 (0.2) | 0.01–1.0 |
| A247C/T249G | S83R | 2 (3.7) | 3 (0.9) | 1 (0.6) | 6 (1.1) | 0.4–2.4 |
| G248T | S83I | 7 (13.0) | 49 (15.3) | 38 (23.2) | 101 (18.7) | 15.5–22.3b |
| A247T | S83C | 0 | 2 (0.6) | 1 (0.6) | 3 (0.6) | 0.1–1.6 |
| G248A | S83N | 0 | 4 (1.2) | 0 | 4 (0.7) | 0.2–1.9 |
| G259A | D87N | 1 (1.9) | 14 (4.4) | 4 (2.4) | 22 (4.1) | 2.6–6.1 |
| A260G | D87G | 0 | 1 (0.3) | 0 | 1 (0.2) | 0.01–1.0 |
| G259T | D87Y | 0 | 2 (0.6) | 4 (2.4) | 6 (1.1) | 0.4–2.4 |
| No change | WTc at S83/D87 | 44 (81.5) | 246 (76.6) | 116 (70.7) | 406 (75.3) | 71.5–78.9 |
Additional variations were detected but are not included in the table, as they have not been shown to contribute to moxifloxacin failure. These include C184T (P62S), C234T (silent), T269A (I90N), and C324T (silent).
Trend analysis was significant for an increase in the prevalence of the G248T SNP (conferring S83I) (Ptrend = 0.046). All other parC SNPs did not demonstrate a trend.
WT, wild type.
FIG 1.
Prevalence of single nucleotide polymorphisms causing changes in key amino acids of the parC gene of M. genitalium. Samples were collected from patients with macrolide-resistant M. genitalium infection, Melbourne, Australia, from 2012 to 2020.
There were few SNPs detected at amino acid position D87. The second most common variant after S83I was SNP G259A (D87N), with an overall prevalence of 4.1% (95% CI, 2.6 to 6.1). The prevalence of this and other D87 SNPs (D87G and D87Y) did not change over the study period (Fig. 1). The SNP conferring ParC G81S was only detected in one sample (0.3%; 95% CI, 0.01 to 1.7). SNPs were detected outside the quinolone resistance determinant region (which corresponds to amino acids 90 to 100) but have not been shown to contribute to moxifloxacin failure (6). There was a decline in the prevalence of infections that were wild type at both S83 and D87 from 81.5% (95% CI, 68.6 to 90.8) in 2012 to 2013 to 70.7% (95% CI, 63.1 to 77.6) in 2019 to 2020, although the trend did not reach significance (Ptrend = 0.085).
The association between specific parC SNPs and moxifloxacin treatment outcomes was examined (Table 3). Of the 75 cases with ParC S83I, 44 failed moxifloxacin (58.7%; 95% CI, 46.7 to 69.9). Of the six infections with ParC S83R, three failed moxifloxacin (50.0%; 95% CI, 11.8 to 88.2). Of the 17 infections with D87N, only two failed (11.8%; 95% CI, 1.5 to 36.4), and of the four cases with D87Y, one failed (25%, 95% CI, 0.6 to 80.6). While moxifloxacin treatment failures were not observed for patients with infections with the ParC variant S83C, failures have previously been reported for sitafloxacin treatment (7).
TABLE 3.
Summary of moxifloxacin treatment outcomes by single nucleotide polymorphisms in parC
| parC SNP | Amino acid change | No. of cases with specific SNP who had treatment failure/total no. of cases with each SNP treated with moxifloxacin for study group: |
Overall analysis |
|||
|---|---|---|---|---|---|---|
| 2012–2013 (n = 54) | 2016–2018 (n = 201)a,b,c | 2019–2020 (n = 164) | No. of cases with specific SNP who had treatment failure/total no. of cases with each SNP treated with moxifloxacin (%) (n = 381) | 95% CI | ||
| A247C/T249G | S83R | 2/2 | 0/3 | 1/1 | 3/6 (50.0) | 11.8–88.2 |
| G248T | S83I | 4/7 | 16/30 | 24/38 | 44/75 (58.7) | 46.7–69.9 |
| A247T | S83C | 0/1 | 0/1 | 0/2 (0) | ||
| G248A | S83N | 0/3 | 0/3 (0) | |||
| G259A | D87N | 0/1 | 1/12 | 1/4 | 2/17 (11.8) | 1.5–36.4 |
| A260G | D87G | 0/1 | 0/1 (0) | |||
| G259T | D87Y | 1/4 | 1/4 (25.0) | 0.6–80.6 | ||
| No change | Absence of S83/D87 (WT) | 0/44 | 4/151 | 2/78 | 6/273 (2.2) | 0.8–4.7 |
| Absence of G248T | Absence of S83I | 2/47 | 5/171 | 5/88 | 11/306 (3.6) | 1.9–6.4 |
A total of 321 samples were sequenced from the 2016 to 2018 group, but only 201 patients were treated with moxifloxacin (the remainder were treated with sitafloxacin). SNP, single nucleotide polymorphism.
For this study group, pretreatment samples were not available for 2 patients, so posttreatment sequences were used.
Additionally, there was one patient with a parC G241A/G81S variant in the 2016 to 2018 time period; this was successfully treated with moxifloxacin.
Of the infections that were wild type at both S83/D87, only 6 out of 273 infections failed moxifloxacin (2.2%; 95% CI, 0.8 to 4.7), and wild-type status was highly predictive of cure (97.8%; 95% CI, 95.3 to 99.2). The absence of S83I alone was associated with 96.4% cure (95% CI, 93.7 to 98.2).
Of interest, while several SNPs clearly contribute to moxifloxacin treatment failure, none are particularly strong predictors of failure compared to 23S rRNA gene SNPs and azithromycin treatment (9). This suggests that variations in parC are less “potent” contributors to resistance, and this is consistent with MIC data (10, 11). Isolates with and without 23S rRNA mutations have 10,000-fold differential azithromycin MIC, with MICs in resistant bacteria elevated to >16 mg/L. However, isolates with/without a single ParC S83I change have a much lower moxifloxacin MIC difference (around 100-fold), and MICs in “resistant” isolates are lower (1 to 8 mg/L). Of note, additional SNPs outside parC (e.g., changes in gyrA, not reported in this study) may also confer resistance and have an additive effect (7).
There is increasing reliance on moxifloxacin in the face of rapidly diminishing efficacy of macrolides. The increasing prevalence of infections with the ParC S83I variation suggests that treatment practices may induce or select for this variation, although we have not identified evidence of selection within moxifloxacin-treated individuals in our data sets (6–8). The low prevalence of other parC SNPs affecting S83 and D87 suggests that these are not being strongly selected for through treatment practices or that any contribution to antibiotic resistance is offset by an impact on bacterial fitness. S83R is a good example of such a variant: a relative high failure rate (50%, 3/6) was observed (noting the limitations of small numbers), but this variant did not show an increasing trend in prevalence as seen for S83I.
This study has limitations. Some SNPs are rare, impacting estimates on contribution to treatment failure, although rarity may also reflect limited clinical relevance. Samples were drawn from a single sexual health clinic, which may limit generalizability; however, Melbourne Sexual Health Centre is the only public sexually transmitted infection (STI) service for a city of more than 5 million people, and a substantial portion of clients are recent migrants and international students visiting Australia. Each of the studies used slightly different treatment regimens incorporating moxifloxacin, which may impact the associations with cure. Data were derived from macrolide-resistant infections, but it should be noted that parC mutations are likely to be rarer in macrolide-susceptible infections due to a lack of selective pressure from fluoroquinolone treatments. SNPs in gyrA also contribute to treatment failure but less significantly than parC (7); gyrA sequence data were not available for all groups, so they are not presented here.
In conclusion, the most common M. genitalium parC amino acid change, S83I (parC SNP G248T), has increased in prevalence over the last 8 years, and almost two-thirds of infections with this SNP failed moxifloxacin treatment. The take-home message is that macrolide-resistant M. genitalium infections without the ParC S83I variant have a 96.4% probability of moxifloxacin cure. Infections that are wild type at both S83 and D87 have an extremely high likelihood of cure at 97.8%. The predictive value of parC resistance SNPs for moxifloxacin (and sitafloxacin) failure is much lower than that observed for macrolide resistance mutations and azithromycin. These findings inform the use of assays that report parC SNPs (either individually or pooled) and highlight the value of wild-type targets in clinical care. Assays targeting common clinically relevant variants such as ParC S83I and wild-type targets for ParC S83 (and D87) are likely to confer the greatest gains in individualized therapy, first-line cure, and antimicrobial stewardship and could enable clinicians to dispense with tests of cures in specific subgroups (8, 12).
ACKNOWLEDGMENTS
This work is supported by the Australian Research Council Research Hub to Combat Antimicrobial Resistance (project ID IH190100021 to G.L.M., C.S.B., and D.M.W.) and National Health and Medical Research Council Investigator Grant APP1197951 (to S.M.G.).
REFERENCES
- 1.Machalek DA, Tao Y, Shilling H, Jensen JS, Unemo M, Murray G, Chow EPF, Low N, Garland SM, Vodstrcil LA, Fairley CK, Hocking JS, Zhang L, Bradshaw CS. 2020. Prevalence of mutations associated with resistance to macrolides and fluoroquinolones in Mycoplasma genitalium: a systematic review and meta-analysis. Lancet Infect Dis 20:1302–1314. 10.1016/S1473-3099(20)30154-7. [DOI] [PubMed] [Google Scholar]
- 2.Read TRH, Fairley CK, Murray GL, Jensen JS, Danielewski J, Worthington K, Doyle M, Mokany E, Tan L, Chow EPF, Garland SM, Bradshaw CS. 2019. Outcomes of resistance-guided sequential treatment of Mycoplasma genitalium infections: a prospective evaluation. Clin Infect Dis 68:554–560. 10.1093/cid/ciy477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Durukan D, Read T, Murray GL, Doyle M, Chow EP, Vodstrcil LA, Fairley CK, Aguirre I, Mokany E, Tan LY, Chen MY, Bradshaw CS. 2020. Resistance-guided antimicrobial therapy using doxycycline-moxifloxacin and doxycycline-2.5g azithromycin for the treatment of Mycoplasma genitalium infection: efficacy and tolerability. J Clin Microbiol 71:1461–1468. 10.1093/cid/ciz1031. [DOI] [PubMed] [Google Scholar]
- 4.Deguchi T, Maeda S, Tamaki M, Yoshida T, Ishiko H, Ito M, Yokoi S, Takahashi Y, Ishihara S. 2001. Analysis of the gyrA and parC genes of Mycoplasma genitalium detected in first-pass urine of men with non-gonococcal urethritis before and after fluoroquinolone treatment. J Antimicrob Chemother 48:742–744. 10.1093/jac/48.5.742. [DOI] [PubMed] [Google Scholar]
- 5.Couldwell DL, Tagg KA, Jeoffreys NJ, Gilbert GL. 2013. Failure of moxifloxacin treatment in Mycoplasma genitalium infections due to macrolide and fluoroquinolone resistance. Int J STD AIDS 24:822–828. 10.1177/0956462413502008. [DOI] [PubMed] [Google Scholar]
- 6.Murray GL, Bradshaw CS, Bissessor M, Danielewski J, Garland SM, Jensen JS, Fairley CK, Tabrizi SN. 2017. Increasing macrolide and fluoroquinolone resistance in Mycoplasma genitalium. Emerg Infect Dis 23:809–812. 10.3201/eid2305.161745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Murray GL, Bodiyabadu K, Danielewski J, Garland SM, Machalek DA, Fairley CK, Jensen JS, Williamson DA, Tan LY, Mokany E, Durukan D, Bradshaw CS. 2020. Moxifloxacin and sitafloxacin treatment failure in Mycoplasma genitalium infection: association with parC mutation G248T (S83I) and concurrent gyrA mutations. J Infect Dis 221:1017–1024. 10.1093/infdis/jiz550. [DOI] [PubMed] [Google Scholar]
- 8.Vodstrcil L, Plummer EL, Doyle M, Murray GL, Bodiyabadu K, Jensen JS, Whiley D, Sweeney EL, Williamson D, Chow EP, Fairley CK, Bradshaw CS. 2021. Combination therapy for Mycoplasma genitalium, and new insights into the utility of parC mutant detection to improve cure. Clin Infect Dis ciab1058. 10.1093/cid/ciab1058. [DOI] [PubMed] [Google Scholar]
- 9.Bissessor M, Tabrizi SN, Twin J, Abdo H, Fairley CK, Chen MY, Vodstrcil LA, Jensen JS, Hocking JS, Garland SM, Bradshaw CS. 2015. Macrolide resistance and azithromycin failure in a Mycoplasma genitalium-infected cohort and response of azithromycin failures to alternative antibiotic regimens. Clin Infect Dis 60:1228–1236. 10.1093/cid/ciu1162. [DOI] [PubMed] [Google Scholar]
- 10.Hamasuna R, Hanzawa H, Moritomo A, Matsumoto M, Aono H, Tomisaki I, Akasaka T, Fujimoto N, Jensen JS. 2022. Analysis of fluoroquinolone-resistance using MIC determination and homology modelling of ParC of contemporary Mycoplasma genitalium strains. J Infect Chemother 28:377–383. 10.1016/j.jiac.2021.11.011. [DOI] [PubMed] [Google Scholar]
- 11.Hamasuna R, Le PT, Kutsuna S, Furubayashi K, Matsumoto M, Ohmagari N, Fujimoto N, Matsumoto T, Jensen JS. 2018. Mutations in ParC and GyrA of moxifloxacin-resistant and susceptible Mycoplasma genitalium strains. PLoS One 13:e0198355. 10.1371/journal.pone.0198355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sweeney EL, Bradshaw CS, Murray GL, Whiley DM. 2022. Individualised treatment of Mycoplasma genitalium infection – it is time to incorporate fluoroquinolone resistance testing into clinical care. Lancet Infect Dis. 10.1016/S1473-3099(21)00629-0. [DOI] [PubMed] [Google Scholar]