Molecular Assay for Detection of Ciprofloxacin Resistance in Neisseria gonorrhoeae Isolates from Cultures and Clinical Nucleic Acid Amplification Test Specimens

Neisseria gonorrhoeae, the causative agent of gonorrhea infection, has the second highest reported rate of bacterial sexually transmitted infections in Canada, with >12,000 reported cases in 2012 (1). Strains of N. gonorrhoeae have acquired resistance to all of the antibiotics commonly used for treatment (2). In Canada, 27.9% of the N. gonorrhoeae isolates obtained from 2008 to 2012 were resistant to ciprofloxacin (3). Fluoroquinolone resistance in N. gonorrhoeae is conferred by mutations in subunit A of DNA gyrase (gyrA) and the ParC subunit of topoisomerase IV (parC) (4).

Beginning in the early 2000s, an increasing number of gonococcal infections have been detected by nucleic acid amplification tests (NAATs), and a decreasing number of laboratories across Canada are culturing N. gonorrhoeae. In fact, >70% of the gonococcal infections in Canada are now detected using NAATs, and antimicrobial susceptibility data are not available for these cases (3).

In this study, real-time PCR assays were developed to detect single nucleotide polymorphisms (SNPs) in genes associated with ciprofloxacin resistance. Three SNPs associated with ciprofloxacin resistance (gyrA S91, gyrA D95, and parC D86/S87/S88) were selected and evaluated using N. gonorrhoeae cultures. As a proof of principle, all SNPs were also evaluated using clinical specimens tested by the APTIMA Combo 2 assay on the Tigris platform (Hologic, Bedford, MA), which also had a matched culture isolate.

Two hundred fifty-two N. gonorrhoeae isolates, 58 clinical Hologic APTIMA Combo 2 CT/NG NAAT specimens (10 N. gonorrhoeae-positive APTIMA specimens, 24 N. gonorrhoeae-positive APTIMA specimens with corresponding culture isolates [from the same anatomical site, collected on the same day], and 24 N. gonorrhoeae-negative APTIMA specimens), and 50 different nongonococcal strains were tested to evaluate efficacy of the assay, as previously described (5). N. gonorrhoeae isolates were selected to represent a range of ciprofloxacin MICs and a diverse group of N. gonorrhoeae multiantigen sequence types (NG-MAST) and temporal and geographic distribution. The isolates also represent a range of cephalosporin MICs and included isolates with decreased susceptibility to ceftriaxone (MIC, ≥0.125 μg/ml; n = 55) and cefixime (MIC, ≥0.25 μg/ml; n = 32). The MICs were determined using the agar dilution method, as previously described (6). Sensitivities and specificities were calculated using the ciprofloxacin resistance breakpoint of 1 μg/ml, according to the CLSI guidelines (7).

Three SNP targets associated with ciprofloxacin resistance were chosen: gyrA S91, gyrA D95, and parC D86/S87/S88. Oligonucleotide primers and probes were chosen for each target region using Primer Express software version 3.0 (Life Technologies), with one probe to detect the wild-type (WT) allele and one positive-control probe that should produce a positive signal in all N. gonorrhoeae isolates. Both gyrA assays were performed using primers gyrA_F1 (TACGCGATGCACGAGCTG) and gyrA_R1 (AGTTGCCCTGTCCGTCTATCAG), along with the control probe gyrA_CTRL (VIC-CTGGAATGCCGCCTAC) and either the WT probe gyrA_S91wt (FAM-ACGGCGATTCCGCA) (FAM, 6-carboxyfluorescein) or gyrA_D95wt (FAM-CAGTTTACGACACCATCG). Assays of the parC gene were performed using the primers parC_F1 (GCGCGATATGGGTTTGACG) and parC_R1 (GGTAAAATCCTGAGCCATGCG) and the probes parC_wt (FAM-CGACAGTTCCGCCTAT) and parC_CTRL (VIC-CGTGGTCGGCGAGAT). DNA extraction, preparation, real-time PCR, and analysis of the results were performed as previously described (5). The results were considered positive if they had a quantification cycle (Cq) value of <40 (8) and a relative fluorescence of the probe minus the baseline (ΔRn) value of >0.7. The limit of detection (LOD) for the PCRs was determined using serial dilutions of DNA from N. gonorrhoeae control strains, as previously described (5). SNPs detected by the assay were validated by comparison with aligned gene sequences obtained through whole-genome sequencing (9).

Of the 252 gonococcal isolates tested, 97 isolates (38.4%) were susceptible, with ciprofloxacin MICs of <1 μg/ml, while the other 155 (61.6%) were resistant to ciprofloxacin (MIC, ≥1 μg/ml). The assay concordance for each genetic marker (percentage of isolates called correctly as WT or SNP according to sequencing data) for the 252 N. gonorrhoeae isolates was 100% for gyrA S91 and parC D86/S87/S88 and 99.6% for gyrA D95. One isolate was positive for the gyrA D95 SNP using the real-time PCR assay, but sequencing identified a different SNP in the probe region, which caused the false-positive SNP assay result. The LODs were 5 pg/reaction for both gyrA SNPs and 5 fg/reaction for parC. The sensitivities (percentage of resistant isolates containing the SNP of interest) and specificities (percentage of sensitive isolates containing a WT allele) are shown in Table 1.

Of the 24 N. gonorrhoeae-positive APTIMA specimens with matched cultures, the assay concordance for each genetic marker was 100% for parC and 83.3% for both gyrA assays (Table 2). One rectal swab specimen gave a false SNP result for gyrA D95. The lower genetic marker performances for the gyrA APTIMA NAAT specimens were due to false-negative results, likely because of the higher LOD of the gyrA assays. Of the 24 N. gonorrhoeae-negative APTIMA specimens, two pharyngeal swabs showed cross-reactivity with the parC assay (WT result). The isolation sites for the APTIMA specimens are shown in Table 3.

Of the 50 nongonococcal isolates tested, 38 gave negative results for all three assays. Cross-reactive species are shown in Table 4. Cross-reactivity with similar species is expected when using highly conserved genes, such as gyrA and parC. This is a limitation when using DNA from a NAAT specimen, as false-positive results can occur in specimens of mixed culture, such as pharyngeal swabs, due to the possible presence of nongonococcal Neisseria species (10). However, none of the N. gonorrhoeae-negative APTIMA specimens gave false-positive results for gyrA, and only two showed a cross-reaction for parC, both of which were pharyngeal swabs. As only a limited number of N. gonorrhoeae-negative genital specimens were tested, additional testing needs to be performed to thoroughly assess the risk of cross-reactivity. Balashov et al. (11) observed that molecular assays may not be applicable to extragenital specimens due to the prevalence of non-gonorrhoeae Neisseria species in these specimens.

All ciprofloxacin-resistant isolates contained ≥2 SNPs, while 97.9% of the susceptible isolates had ≤1 SNP present. One hundred percent of the isolates containing all three SNPs were resistant to ciprofloxacin. Based on these values, the presence of any 2 of these SNPs is sufficient to determine if an isolate is resistant to ciprofloxacin. The high specificity and sensitivity of this assay make it useful for predicting ciprofloxacin resistance in a clinical sample. This is proof of principle that the detection of these SNPs could be useful in a test for targeted therapy with ciprofloxacin, as ∼70% of Canadian N. gonorrhoeae isolates are currently susceptible to this antibiotic (3).

While the elevated prevalence of ciprofloxacin resistance prevents the effective use of fluoroquinolones for empiric gonorrhea treatment, it is important to monitor resistance prevalence, as resistant strains often have elevated MICs to other antibiotics, including third-generation cephalosporins (3). As the use of molecular methods for laboratory diagnosis of gonorrhea becomes increasingly widespread and cultures become less readily available, molecular methods for detecting antimicrobial resistance genes provide an alternative to culture-based antimicrobial susceptibility testing (12). Although the results of this study highlight the utility of this molecular method to determine ciprofloxacin resistance in the absence of N. gonorrhoeae culture isolates, caution should be advised, as cross-reactivity has been observed in other Neisseria species.

We thank Gary Liu, Pam Sawatzky, and Anton Kowalski from the Streptococcus and Sexually Transmitted Diseases Unit for their laboratory technical assistance.