|Home | About | Journals | Submit | Contact Us | Français|
Surveillance of gonococcal antimicrobial resistance and the molecular characterization of the mechanisms underlying these resistance phenotypes are essential in order to establish correct empirical therapies, as well as to describe the emergence of new mechanisms in local bacterial populations. To address these goals, 149 isolates were collected over a 1-month period (October-November 2008) at the Ontario Public Health Laboratory, Toronto, Canada, and susceptibility profiles (8 antibiotics) were examined. Mutations in previously identified targets or the presence of some enzymes related to resistance (r), nonsusceptibility (ns) (resistant plus intermediate categories), or reduced susceptibility (rs) to the antibiotics tested were also studied. A significant proportion of nonsusceptibility to penicillin (PEN) (89.2%), tetracycline (TET) (72.3%), ciprofloxacin (CIP) (29%), and macrolides (erythromycin [ERY] and azithromycin; 22.3%) was found in these strains. Multidrug resistance was observed in 18.8% of the collection. Although all the strains were susceptible to spectinomycin and extended-spectrum cephalosporins (ESC) (ceftriaxone and cefixime), 9.4% of them displayed reduced susceptibility to extended-spectrum cephalosporins. PBP 2 mosaic structures were found in all of these ESCrs isolates. Alterations in the mtrR promoter, MtrR repressor (TETr, PENns, ESCrs, and ERYns), porin PIB (TETr and PENns), and ribosomal protein S10 (TETr) and double mutations in gyrA and parC quinolone resistance-determining regions (QRDRs) (CIPr) were associated with and presumably responsible for the resistance phenotypes observed. This is the first description of ESCrs in Canada. The detection of this phenotype indicates a change in the epidemiology of this resistance and highlights the importance of continued surveillance to preserve the last antimicrobial options available.
The reported incidence of gonococcal infections has increased in Canada over the last 11 years (1997 to 2008), from 15 cases to 38 cases per 100,000 (41). Coupled with this increase, antimicrobial resistance in Neisseria gonorrhoeae has continued to emerge worldwide, limiting empirical treatment regimens for the disease. Since the 1980s, a rapid rise in resistance to penicillin (PEN) and tetracycline (TET) led to the discontinuation of their use for the treatment of gonococcal infections. More recently, the emergence of clinical isolates of N. gonorrhoeae resistant to ciprofloxacin (CIP) has been reported globally (2, 4, 13, 57, 58). During 2006, 28% of N. gonorrhoeae isolates were resistant to ciprofloxacin in Ontario, Canada (38). Clinical strains displaying intermediate susceptibility or resistance to azithromycin (AZM) are now emerging (9, 18, 19, 32, 40, 46, 60).
Extended-spectrum cephalosporins (ESC) (ceftriaxone and cefixime) are recommended as the first-line treatment of gonococcal infections in Canada and the United States; azithromycin, spectinomycin, and the fluoroquinolones are second-line options (6, 7, 42). While resistance of N. gonorrhoeae to extended-spectrum cephalosporins has not been described, reduced susceptibility (ESCrs) has led to considerable concern, particularly in Japan (24, 25, 28, 50, 59). Clinical failures associated with reduced susceptibility to the extended-spectrum cephalosporins have been described (1, 54).
Resistance to penicillin (PENr) in N. gonorrhoeae can be mediated by the plasmid-encoded TEM-1 β-lactamase or by mutations in chromosomal genes. A recent study summarizes the stepwise transfer of PENr from a resistant to a susceptible strain (62). In addition to reduced affinity of PBP 2, increased efflux pump activity (MtrCDE, by mutations in the promoter region or in its repressor, MtrR), and impermeability (outer membrane proteins PIA and PIB), mutations in PBP 1 and PilQ secretin have also been involved in the acquisition of high-level chromosomally mediated PENr (44, 61). In addition to these mutations, ESCrs strains have mosaic structures in the transpeptidase-encoding domain of PBP 2 (3, 26, 27, 28, 35, 48, 56, 62). Some of these genetic modifications (derepressed efflux and impermeability), as well as target mutations (ribosomal protein S10, encoded by the rpsJ gene), reduce the susceptibility of the bacteria to tetracycline, whereas their combination and/or the expression of the tet(M) gene confer high-level TETr (21, 22, 33). Ciprofloxacin resistance in gonococci has been attributed mainly to chromosomal mutations in the target of the compound, in a region named the quinolone resistance-determining region (QRDR) of the parC and gyrA genes, which encode the ParC and GyrA subunits of topoisomerase IV and DNA gyrase, respectively (5, 51). Reduced intracellular drug accumulation was also suggested for CIPr isolates (49, 51). Spectinomycin resistance (SPTr) was associated with point mutations in the 16S rRNA genes (20). There is limited information describing the molecular bases of macrolide resistance in gonococci. Although macrolide resistance mediated by ribosomal methylation or mutations has been described, increased activity of the MtrCDE efflux pump is the most common mechanism responsible for the phenotype (8, 11, 12, 34, 43, 53, 55, 60).
Although the incidence of gonococcal infections is increasing, the use of nucleic acid amplification testing reduced the number of samples collected for culture, limiting the availability of bacterial isolates for phenotypic studies (e.g., antimicrobial resistance testing). However, urgent monitoring of local resistance patterns in N. gonorrhoeae and the underlying mechanisms is required to effectively guide treatment recommendations. Since there are no updated surveillance data about antimicrobial resistance in N. gonorrhoeae from Ontario, Canada, the objectives of this study were (i) to determine the susceptibility profiles to all the relevant antigonococcal antimicrobial agents in clinical strains isolated across the province and (ii) to characterize the molecular mechanisms producing resistance phenotypes. Particular attention was paid to those conferring reduced susceptibility to extended-spectrum cephalosporins and multidrug resistance (MDR). In addition, the clonal relationship between these isolates was established.
(Preliminary data were presented in abstract form at the 18th International Society for Sexually Transmitted Diseases Research/British Association of Sexual Health and HIV, London, United Kingdom, 2009; the 26th International Congress of Chemotherapy and Infection/Association of Medical Microbiology and Infectious Disease Canada-Canadian Association for Clinical Microbiology and Infectious Diseases Annual Conference, Toronto, ON, Canada, 2009; and the 50th Interscience Conference on Antimicrobial Agents and Chemotherapy, Boston, MA, 2010.)
The Ontario Public Health Laboratories (PHL) serve as the provincial microbiological reference service for the province of Ontario, Canada, with a population of 13,210,700 in 2010 (Census Canada [http://www40.statcan.gc.ca/l01/cst01/demo02a-eng.htm]). PHL is the only laboratory in Ontario that provides routine clinical susceptibility testing for culture isolates of N. gonorrhoeae. Considering that all private and hospital laboratories across the province must submit all isolates of N. gonorhoeae to PHL for confirmation and susceptibility testing, a collection of clinical isolates obtained during a 1-month period would be a representative sample of N. gonorrhoeae cultures in Ontario. To accomplish this, all consecutive, unique patient N. gonorrhoeae isolates received and confirmed from 15 October to 15 November 2008 at the PHL were included in this study (n = 149). Patient data were not included because they are not routinely available at the PHL except on a case-by-case basis. The collection size was chosen in order to provide a detailed analysis of the common mechanisms of antibiotic resistance in N. gonorrhoeae isolates throughout the province. Primary specimens and isolates received for confirmation of N. gonorrhoeae were subcultured on New York City medium and incubated for 24 to 72 h in 5% CO2 at 35 to 37°C. Gram stain, oxidase, and carbohydrate utilization tests (glucose, maltose, sucrose, and O-nitrophenyl-β-d-galactopyranoside [ONPG]) were performed. Identification of N. gonorrhoeae also included testing growth on blood agar at 22°C and on nutrient agar at 35.5°C. A GenProbe Accuprobe N. gonorrhoeae Culture Identification kit was used as a second confirmatory method for N. gonorrhoeae. All isolates were cultured on GC agar and 1% defined growth supplement (10) at 37°C in 5% CO2 for 20 to 24 h and stored at −86°C. Each sample was subcultured twice before antimicrobial testing and DNA extraction were performed. The MICs of penicillin, ceftriaxone, cefixime, ciprofloxacin, tetracycline, erythromycin, and spectinomycin were determined by the agar dilution method and Etest, according to the Clinical and Laboratory Standards Institute guidelines (10). The MICs of azithromycin were tested by Etest in all the isolates with reduced susceptibility to erythromycin (ERYrs) (MICs ≥ 2 μg/ml). N. gonorrhoeae ATCC 49226 and strains F, G, K, L, N, O, and P of the 2008 WHO N. gonorrhoeae reference strain panel were used as quality control strains (52). The presence of β-lactamase activity was tested using nitrocefin.
Previously described targets related to resistance or reduced susceptibility to the antimicrobials tested were studied (Table (Table1).1). DNA extracts of each isolate were prepared by boiling a 1-μl loop of each isolate in lysis buffer (1% Triton X-100, 0.5% Tween 20, 1 mM EDTA, 10 mM Tris-HCl, pH 8). After centrifugation, 2 μl of these supernatants was used per PCR. The primers used for PCR amplification are detailed in Table Table1.1. PCRs were performed using standard conditions or conditions previously described (Table (Table1).1). DNA samples were stored at −20°C. All the amplicons were sequenced with the same specific primers used for PCR under the BigDye terminator methodology in a model 3130xl DNA sequence analyzer (Applied Biosystem/Perkin-Elmer, Foster City, CA). The nucleotide and deduced amino acid sequences were analyzed with the Vector NTI analysis software package (version 10.3.0; Invitrogen Corp.). Searches of sequences were performed with the BLAST program, available at the National Center for Biotechnology Information website (http://www.ncbi.nim.nih.gov/). Multiple-sequence alignments were performed with the ClustalX program, available at the European Bioinformatics Institute website (http://www.ebi.ac.uk/).
N. gonorrhoeae multiantigen sequence typing (NG-MAST) was performed on all isolates by sequencing internal fragments of two highly polymorphic loci, porA/B and tbpB (31). The edited and trimmed sequences were uploaded into a publicly accessible database on the NG-MAST website (http://www.ng-mast.net) to obtain the allele number and the sequence type (ST).
The PBP 2 nucleotide sequences associated with reduced susceptibility to extended-spectrum cephalosporins identified in this study have been deposited in the GenBank database under accession numbers HQ204552 to HQ204565.
The susceptibility profiles of 149 consecutive, nonduplicate N. gonorrhoeae isolates, received and confirmed from 15 October to 15 November 2008 at the PHL, Toronto, Canada, are presented in Table Table2.2. MDR (defined as resistance to 3 or more different antimicrobials; ERYrs was included in this definition) was observed in 28 isolates (18.8%): 19 were resistant to 3 antimicrobials, and 9 were resistant to 4 of them (Fig. (Fig.1).1). While all the strains were susceptible to spectinomycin, ceftriaxone, and cefixime, 14 isolates demonstrated reduced susceptibility to extended-spectrum cephalosporins, with MICs from 0.125 to 0.25 μg/ml.
PENr was observed in 18 isolates (12.2%; MICs from 2 to 6 μg/ml). However, 115 isolates (77.2%) displayed intermediate resistance (0.12 to 1 μg/ml) (Table (Table2).2). Only 5 PENr isolates showed penicillinase activity, and they displayed the highest MICs. Seventeen out of 18 PENr isolates were also CIPr and TETr, and 10 of them were also ERYrs. Reduced susceptibility to extended-spectrum cephalosporins (ESCrs) was observed in 14 isolates (9.4%; MICs, 0.125 to 0.25 μg/ml for cefixime and 0.032 to 0.125 μg/ml for ceftriaxone); all of them were CIPr, and most were TETr (13 isolates) and ERYrs (11 isolates), but only 4 were PENr (9 with intermediate resistance to penicillin [PENi]).
To characterize the molecular backgrounds of these phenotypes, all the PENr and 12 PENi isolates (n = 30), as well as all the ESCrs isolates (n = 14), were studied. Amino acid changes deduced from the nucleotide sequences of ponA (PBP 1), penA (PBP 2), porA/B (outer membrane proteins PIA and PIB), pilQ (PilQ, a pilus secretin protein), and mtrR (MtrR, a transcriptional repressor of the operon mtrCDE encoding an efflux pump), as well as mutations in the mtrR promoter and the presence of blaTEM-1, were analyzed (Tables (Tables33 and and44).
Between the PENns and ESCrs isolates analyzed, no mutations in the pilQ gene were found. blaTEM-1 was detected in only 5 PENr isolates, but not in any ESCrs isolates. Most of the PENns (n = 23) and 7 of the ESCrs isolates contained a single amino acid change in PBP 1 (L421P), previously related to 3- to 4-fold reduction in the rate of acylation of the enzyme to β-lactams (44). Single (position A121) and double (positions G120 and A121) substitutions were observed in porins PIA and PIB (Tables (Tables33 and and4).4). Mutations in these two positions increased resistance to penicillin and tetracycline (36). However, additional mutations affecting the mtrR determinant are required for full PENr and TETr (37). MtrR represses the expression of the mtrCDE operon that encodes an efflux pump member of the resistance-nodulation-division family. Amino acid changes in MtrR (A39T, G45D, and the double mutant A39T/R44H) were found in 2, 6, and 3 PENns isolates, respectively, and 1 ESCrs isolate (G45D). These substitutions impede the binding of MtrR to its DNA target, increasing the expression of the MtrCDE efflux pump (55). Similar effects (overexpression of the efflux pump) were described in strains with mutations in the mtrR promoter gene (29, 60). Nineteen out of the 30 PENns and 12/14 ESCrs isolates showed a single adenine deletion in the 13-bp inverted-repeat located in the mtrR promoter gene; 2 PENi isolates and 1 ESCrs isolate with a single thymine insertion in the promoter were detected. Mutations affecting both the MtrR repressor and mtrR promoter were observed in only 2 PENns and 1 ESCrs isolates.
PBP 2 remained wild type in 29 PENns isolates. A single PENr isolate with multiple substitutions in penA in a mosaic-like structure was found, but that isolate did not show ESCrs. Most of the PENr isolates without penicillinase activity (n = 13/18; MIC = 2 μg/ml) showed substitutions in PIB and PBP 1 and a 1-adenine deletion in the mtrR promoter (Table (Table3).3). Six additional PENi isolates that were not TEM-1 producers harbored mutations affecting PBP 1, permeability, and efflux pump expression, as well, suggesting that the combination of these different mutations is responsible for the PENns phenotype. All isolates with ESCrs displayed a mosaic PBP 2 structure. Of these, 5 PBP 2 amino acid patterns were found, designated XXXIV to XXXVIII according to the nomenclature used previously (Table (Table44 and Fig. Fig.22 ) (25, 26, 27, 35, 48, 56). Two of them were new PBP 2 mosaic types (XXXVII and XXXVIII). The most common PBP 2 type found in this study was the sequence pattern XXXVIII (9/13 isolates, 5 of them belonging to ST3158). This PBP2 type was previously described in the United States and is closely related to pattern XXXII (1 substitution, L551P) (35, 39). The 14 ESCrs isolates with PBP 2 mosaic structure mostly showed mutations in the mtrR promoter gene, the porin PIB, and PBP 1 (Table (Table4),4), supporting the association of these mutations with the reduced susceptibility to cefixime and ceftriaxone previously described by Lindberg et al. (27).
TETr was observed in 38 isolates (25.5%, MICs from 2 to 64 μg/ml). Most of these TETr isolates were CIPr (n = 32), and just 17 were PENr. As in the case of penicillin, a high number of isolates (n = 70; 47%) displayed intermediate resistance to the antibiotic (0.5 to 1 μg/ml) (Table (Table22).
Seven isolates were positive for the tet(M) gene by PCR (Table (Table5).5). The same mutations affecting PIA/B found in PENns isolates were observed in TETr isolates (3 single mutants in position A121 and 29 double mutants in positions G120/A121), as well as amino acid changes in MtrR (2 A39T and 4 G45D single mutants and 1 A39T/R44H double mutant). Twenty-seven isolates with a single adenine deletion and 3 with a single thymine insertion in the mtrR promoter gene were also detected. Although mutations in porins PIA/B and the derepression of the MtrCDE efflux pump are common resistance mechanisms for both penicillin and tetracycline, only 17 out of 38 TETr isolates were PENr, as well. Most of these TETr isolates showed single amino acid mutations in the ribosomal protein S10 (V57M; n = 31). This substitution, which can specifically reduce the affinity of tetracycline for its target, was described as increasing the MIC of tetracycline 3- to 4-fold independently of the bacterial genetic background (22). The combination of mutations affecting the tetracycline target (ribosomal protein S10), antibiotic influx (porins PIA/B), and efflux (derepression of the MtrCDE efflux pump) are the most important mechanisms of TETr in this collection.
CIPr was observed in 42 isolates (28.2%, MICs ranging from 3 to >32 μg/ml) (Table (Table2).2). Just 2 isolates presented intermediate MIC values (CIPi). Thirty-two out of 44 isolates (72.7%) were also TETr, and the remaining 12 isolates displayed a TETi phenotype. Just 17 isolates were also PENr (38.6%), and the remainder were PENi. All 44 of these isolates were studied at the molecular level.
The CIPr phenotype was mainly due to simultaneous chromosomal mutations in gyrA and parC QRDRs (n = 39; 92.3% of them were GyrAS91F-D95G ParCS87R) (Table (Table6).6). One CIPr (MIC, 6 μg/ml) and 1 CIPi isolates presented a double mutation in the gyrA QRDR (GyrAS91F-D95G), whereas the other CIPi isolate was wild type for both genes. Unexpectedly, 2 CIPr isolates (MICs, 12 and 32 μg/ml) were wild type for gyrA and parC QRDRs, as well as the mtrR gene/promoter and the porin PIB. These isolates remain under study.
Thirty-four isolates were identified with MICs of ≥2 μg/ml for ERY (22.8%). These isolates were designated ERYrs (Table (Table2).2). Low azithromycin MIC values (0.25 to 0.5 μg/ml) were observed in all the ERYrs isolates (data not shown). However, considering the European Committee of Antimicrobial Resistance Testing breakpoints for azithromycin (susceptible, ≤0.25; resistant, >0.5) and previous studies suggesting that MICs for azithromycin of 0.25 to 1 μg/ml indicate reduced susceptibility to the macrolide, all the ERYrs isolates were designated AZMrs, as well (14, 15, 43, 60).
Three different mechanisms involved in macrolide resistance were described in N. gonorrhoeae: efflux systems (30, 45), modification of the ribosomal target by methylases (43), and ribosomal modification by point mutations in the macrolide targets (18, 19, 34). Here, the presence of 23S rRNA methylases (ermA, ermB, ermC, and ermF genes), a plasmid-mediated efflux pump (mefA/E genes), macrolide 2′-phosphotransferase (mphA gene), ERY esterases (ereA and ereB genes), and possible mutations in the mtrR gene/promoter and riboproteins L4 and L22 (rplD and rplV genes) were tested in all the isolates with ERYrs. Moreover, mutations in the 4 copies of the 23S rRNA genes identified in N. gonorrhoeae were analyzed, as well, using the strategy described by Farrell and colleagues (16, 17). Briefly, a common primer inside the 4 copies of the 23S rRNA genes and specific external primers close to each of them were designed from the complete N. gonorrhoeae NCCP11945 genome (accession number NC_011035). Using the amplicons obtained in these first PCRs as DNA templates, nested PCRs were performed using common primers internal to the 23S rRNA gene (including the peptidyltransferase loop in domain V) to identify point mutations associated with macrolide resistance (Table (Table11).
Of all the genes searched by PCR, only ermB was detected in 1 isolate. PCR and sequencing analyses did not detect mutations in the riboproteins L4 and L22. A unique point mutation (transition C2403U by N. gonorrhoeae numbering [C2417U by E. coli numbering]) in domain V of the 23S ribosomal subunit was found in 6 isolates, 4 of them with the same mutation in 3 of the 4 copies of the 23S rRNA genes. However, this mutation is located far from the positions where single mutations were previously associated with macrolide resistance, suggesting that it is not involved in the development of the ERYrs phenotype. Most of the ERYrs isolates had mutations in the mtrR promoter (n = 29) or in MtrR (n = 3), suggesting that the overexpression of the MtrCDE efflux pump was the most important macrolide resistance mechanism observed in this study (Table (Table7).7). Two ERYrs isolates were wild type for all the genes tested and remain under study.
Based on the analysis of porA/B and tbpB alleles, 98 different STs were obtained for all the isolates included in this study, most of them (n = 73) represented by only a single isolate. Fifty-six of these 98 STs (57%) were found to be novel, indicating high genetic variability in the bacterial collection studied. ST2 was the most prevalent (n = 12; 8.1%), followed by ST51 (n = 6; 4%); ST3158 (n = 6; 4%); ST1105, ST3550, and ST3554 (n = 4; 2.7%); ST25, ST576, and ST865 (n = 3; 2%); and another 16 different STs with 2 isolates each (1.3%). However, all the STs represented by 3 or more isolates (except ST3158 and ST3550 [see below]) were not related to resistance to any antimicrobial. Thirty-one different STs were related to MDR: 21 were associated with resistance to 3 antimicrobials (9 STs were CIPr ERYrs TETr, 7 were PENr CIPr TETr, and 5 were PENr CIPr ERYrs), and 10 were resistant to 4 antimicrobials (PENr CIPr TETr ERYrs). Because of the high diversity of STs, no prevalent correlation between STs and resistance phenotypes was evident. However, some STs were clearly associated with MDR: 5/6 isolates with ST3158 (all CIPr TETr ERYrs and ESCrs), 2/2 isolates with ST1407 (1 PENr CIPr TETr ERYrs and 1 CIPr TETr ERYrs), and 2/4 isolates with ST3550 (PENr CIPr TETr ERYrs). ST3158 and ST1407 are closely related, sharing the same tbpB allele (allele 110) and just 2 point mutations of difference between their porA/B alleles (allele 1914 for ST3158 and allele 908 for ST1407). ST3550 is not related to ST3158 and ST1407.
Analyzing by alleles inside each ST, all 6 isolates harboring porB allele 908 (2 isolates with ST1407 and 1 with ST3387, ST3561, ST3587, and ST3588) were associated with resistance to 2 or more different antimicrobials (5 of them displaying MDR). A similar situation was observed in isolates harboring porB allele 4 (9/11 isolates belonging to 6 different STs), allele 1914 (7/8 isolates from 4 different STs), and allele 1884 (5/6 isolates from 4 different STs). However, the tbpB alleles in each ST were not related. On the other hand, 4/5 isolates harboring tbpB allele 35 (1 isolate with ST546, ST1634, ST3551, and ST3570) were associated with resistance to 2 or more different antimicrobials (ST546 and ST3570 are closely related, with 6 point mutations of difference between their porA/B alleles). Resistance to 2 or more antimicrobials was also observed in isolates harboring tbpB allele 49 (4/8 isolates belonging to ST3550, ST3569, and ST3590; the first 2 are closely related, with just 2 point mutations of difference between their porA/B alleles), allele 110 (8/14 isolates from ST2569, ST3553, and the previously mentioned ST1407 and ST3158), and allele 801 (2/2 isolates belonging to ST3561 and ST3563, both closely related, with just 2 point mutations of difference between their porA/B alleles).
Resistance to first-line antimicrobials is continuously emerging in N. gonorrhoeae. Resistance to ciprofloxacin has been reported in different countries, with rates leading to its discontinuation as the first-line option in gonorrhea therapy (e.g., 28% in Ontario, Canada) (38). This study demonstrated high genetic variability in the gonococcal population studied, with a significant proportion of N. gonorrhoeae isolates from Ontario with nonsusceptibility (resistant plus intermediate categories) to penicillin (89.2%), tetracycline (72.3%), ciprofloxacin (29%), and macrolides (22.3%). This is the first description of reduced susceptibility to extended-spectrum cephalosporins in Canada, all of them associated with PBP 2 mosaic structures. Alterations in the mtrR promoter, MtrR repressor (PENns, TETr, and ERYns), the porin PIB (PENns and TETr), and ribosomal protein S10 (TETr) and double mutations in gyrA and parC QRDRs (CIPr) were associated with and presumably responsible for the MDR phenotypes observed. Resistance to extended-spectrum cephalosporins, azithromycin, and spectinomycin were not detected, and the drugs remain therapeutic options for gonorrhea treatment in Ontario. However, the detection of reduced susceptibility to extended-spectrum cephalosporins associated with a multidrug resistance phenotype indicates a change in the epidemiology of this resistance and highlights the importance of continued surveillance to preserve the last antimicrobial options currently available.
Published ahead of print on 22 November 2010.