Genomic epidemiology of Neisseria gonorrhoeae elucidating the gonococcal antimicrobial resistance and lineages/sublineages across Brazil, 2015–16

Abstract

Objectives

Neisseria gonorrhoeae antimicrobial resistance (AMR) surveillance is imperative internationally, but only eight (22.9%) countries in the WHO Region of the Americas reported complete AMR data to the WHO Global Gonococcal Antimicrobial Surveillance Program (WHO GASP) in 2016. Genomic studies are ideal for enhanced understanding of gonococcal populations, including the spread of AMR strains. To elucidate the circulating gonococcal lineages/sublineages, including their AMR determinants, and the baseline genomic diversity among gonococcal strains in Brazil, we conducted WGS on 548 isolates obtained in 2015–16 across all five macroregions in Brazil.

Methods

A total of 548 gonococcal isolates cultured across Brazil in 2015–16 were genome sequenced. AMR was determined using agar dilution and/or Etest. Genome sequences of isolates from Argentina (n = 158) and the 2016 WHO reference strains (n = 14) were included in the analysis.

Results

We found 302, 68 and 214 different NG-MAST, MLST and NG-STAR STs, respectively. The phylogenomic analysis identified one main antimicrobial-susceptible lineage and one AMR lineage, which was divided into two sublineages with different AMR profiles. Determination of NG-STAR networks of clonal complexes was shown as a new and valuable molecular epidemiological analysis. Several novel mosaic mtrD (and mtrR and mtrE) variants associated with azithromycin resistance were identified.

Conclusions

We describe the first genomic baseline data to support the Brazilian GASP. The high prevalence of resistance to ciprofloxacin, tetracycline and benzylpenicillin, and the high number of isolates with mosaic penA and azithromycin resistance mutations, should prompt continued and strengthened AMR surveillance, including WGS, of N. gonorrhoeae in Brazil.

Introduction

Gonorrhoea is the second most prevalent bacterial sexually transmitted infection (STI) globally. The WHO Region of the Americas has the second highest estimated gonorrhoea incidence worldwide; with 23 cases per 1000 women and 32 per 1000 men estimated in 2016. With the exception of North America, the number of reported cases in this WHO Region is limited due, in particular, to a lack of sufficient aetiological diagnostics.^1–3

Neisseria gonorrhoeae has acquired or developed antimicrobial resistance (AMR) to all antimicrobials introduced for empirical monotherapy.^4–6 This AMR development has been caused by many AMR mechanisms, e.g. penA, associated with resistance to penicillins and cephalosporins through changes in the penicillin-binding protein 2 (PBP2); ponA encoding PBP1 (L421P), associated with penicillin resistance; gyrA and parC, associated with resistance to fluoroquinolones; and 23S rRNA and 16S rRNA, associated with resistance to macrolides and spectinomycin, respectively.^1,5–8 These target mutations are often present in combination with mechanisms that decrease the susceptibility to many antimicrobials, e.g. by reducing the influx (porB1b mutations) and/or increasing the efflux of antimicrobials (e.g. overexpressed MtrC–MtrD– MtrE efflux pump).^1,5–8

The major concern is the emergence of resistance to the extended-spectrum cephalosporin (ESC) ceftriaxone, the last option for empirical monotherapy of gonorrhoea. Sporadic cases of high-level resistance to ceftriaxone were reported in Japan, France and Spain in 2009–11,^9–11 which resulted in considerable public health concern that gonorrhoea could become untreatable. Dual antimicrobial therapy (ceftriaxone 250–500 mg plus azithromycin 1–2 g) was introduced as the first-line treatment in many countries.^12–16 However, the first global gonorrhoea treatment failure with recommended dual therapy was reported in 2016 in the UK¹⁷ and the first gonococcal strain with ceftriaxone resistance and high-level resistance to azithromycin was described in 2018, also in the UK.¹⁸ Since 2015, a strain with the novel mosaic penA-60.001, causing ceftriaxone resistance, has been spreading internationally and has caused several ceftriaxone treatment failures.^19–26

In Brazil, syndromic management of STIs has been widely implemented since 1993. This has resulted in a lack of laboratory diagnostics, including culture, of gonococci, except for laboratory testing of high-risk groups in some settings.³ Consequently, prior to the Brazilian Gonococcal Antimicrobial Surveillance Program (GASP),³ internationally published gonococcal AMR data from Brazil have been very limited. Since 2017 and informed by the Brazilian GASP,³ dual therapy (ceftriaxone 500 mg plus azithromycin 1 g) has been recommended for first-line treatment of uncomplicated gonorrhoea in Brazil.²⁷

Genomic studies of gonococci are ideal for enhanced understanding of the evolution of the gonococcal population (antimicrobial-susceptible and resistant isolates), molecular epidemiology and to detect AMR determinants for prediction of AMR.^5,28–36 However, WGS data for very few gonococcal isolates from Central and South America have been published, i.e. only selected isolates with decreased susceptibility or resistance to ESCs from Argentina in 2011–16 (n = 158)^34,35 and isolates obtained in Rio de Janeiro, Brazil in 2006–15 (n = 116).³⁶ Despite successful AMR surveillance in Latin America in earlier years,^37,38 only 8 (22.9%) of the 35 countries in the entire WHO Region of the Americas reported resistance data for all key antimicrobials (ceftriaxone, azithromycin and ciprofloxacin) in the most recent (2016) WHO Global GASP report.³⁹ Nevertheless, in 2018, AMR data from the first nationwide GASP in Brazil were published.³

To elucidate the circulating gonococcal lineages/sublineages, including their AMR determinants, and the baseline genomic diversity among gonococcal strains in Brazil, we conducted WGS on 548 isolates obtained in 2015–16 across all five macroregions in Brazil.

Materials and methods

Neisseria gonorrhoeae isolates

Overall, 548 (99.6%; two isolates were non-viable) of the 550 gonococcal isolates from 2015–16 previously reported in the nationwide AMR study in Brazil³ were examined by WGS. The isolates were cultured from consecutive men aged ≥18 years with urethral discharge in all the five macroregions in Brazil:³ northern region (Manaus, Amazonas: n = 99); northeastern region (Salvador, Bahia: n = 104); central-western region (Brasilia, Federal District: n = 68); southeastern region (Belo Horizonte, Minas Gerais: n = 103; São Paulo, São Paulo: n = 28); and southern region (Florianópolis, Santa Catarina: n = 74; Porto Alegre, Rio Grande do Sul: n = 72). For comparison, we included previously genome-sequenced isolates with decreased susceptibility or resistance to ESCs from Argentina (n = 158)³⁴ and the 2016 WHO reference strains (n = 14).⁴⁰ MICs (mg/L) of all isolates were determined using agar dilution and/or Etest (bioMérieux, Marcy-l’Étoile, France), as previously described.^3,40 Only whole MIC dilutions were reported and the clinical breakpoints were according to the EUCAST (www.eucast.org/clinical_breakpoints v10.0). For azithromycin, no clinical breakpoints exist and the EUCAST azithromycin epidemiological cut-off (ECOFF) of MIC >1 mg/L (www.eucast.org/clinical_breakpoints) was used to indicate isolates with azithromycin resistance determinants (considered as azithromycin resistant). Decreased ESC susceptibility was defined as an MIC >0.064–0.125 mg/L, because isolates with such MICs have caused ESC treatment failures.^33,34,39,41

WGS, molecular epidemiology and antimicrobial resistance determinants

Genomic DNA was extracted using a customized QIAsymphony (QIAGEN GmbH, Hilden, Germany) protocol. Quality control and normalization of the extracted DNA and the Nextera XT DNA libraries (Illumina, Inc., San Diego, CA, USA) were performed using Qubit (Thermo Fischer Scientific, Waltham, MA, USA) and Tapestation (Agilent Technologies, Santa Clara, CA, USA).

Sequencing libraries were prepared using the Nextera XT DNA library preparation kit (Illumina) and WGS was performed using Illumina MiSeq (Illumina), as previously described.³¹ All reads were quality controlled and trimmed according to Phred quality score Q30 and tested for contamination using best matching kmer spectra against the 2016 WHO reference strains⁴⁰ with at least 95% mapped reads, and additional contamination tests were performed in the automatic array in Pathogenwatch (https://pathogen.watch/), which utilizes Mash.⁴² Assemblies, in silico detection of AMR determinants, typing schemes [N. gonorrhoeae Multi-Antigen Sequence Typing (NG-MAST) www.ng-mast.net; MLST (https://pubmlst.org/), and N. gonorrhoeae Sequence Typing for Antimicrobial Resistance (NG-STAR)] and the frequency of 23S rRNA gene mutations were obtained and identified using a customized CLC Genomics Workbench v12.0, as previously described.³³ To explore the correlations between different AMR determinants of isolates, NG-STAR alleles and STs were analysed using goeBURST,⁴³ and minimum spanning trees were generated using PHYLOViZ 2.0.⁴⁴ The available NG-STAR database (https://ngstar.canada.ca/) was analysed (28 March 2020) and the minimum spanning tree was used for the nomenclature of the clonal complexes (CCs). NG-STAR STs were assigned CCs when sharing ≥5 of the 7 NG-STAR loci, and the founder ST in the tree with the highest number of links was used for the nomenclature of the CCs. Finally, the separate minimum spanning tree with STs from Brazil was subdivided into goeBURST groups using the level 2 option in PHYLOViZ 2.0.⁴⁴

A phylogenomic tree was created using IQ-TREE v1.6.10⁴⁵ based on 42 244 SNP sites in the multiple sequence alignment (MSA). MSA was obtained by mapping all reads to the gonococcal strain FA1090 (GenBank: AE004969.1) using BWA-MEM v0.7.17,⁴⁶ SNPs were called and 6179 recombinant regions were removed using Gubbins v1.4.10 with a mean region size of 5252 bp.⁴⁷ Phylogeography and related metadata were visualized in Microreact.⁴⁸ PATRISTIC⁴⁹ was used to calculate the average distance in the tree with 5% outlier removal, and the average patristic distance of 8550 SNPs was used to define the lineages using RAMI.⁵⁰ The sublineages were defined as above with an average PATRISTIC distance of 5562 SNPs within lineage A.

All raw sequence reads for the Brazilian (PRJEB36607), Argentinian (PRJEB36608) and 2016 WHO reference panel (PRJEB14020) isolates are available through the ENA (https://www.ebi.ac.uk/ena), and the phylogenomic tree and metadata for all examined isolates (n = 720) are available through Microreact (https://microreact.org/project/Golparian_et_al/b3efffec).

Results

Antimicrobial susceptibility

For the Brazilian isolates (n = 548), the level of resistance to tetracycline, ciprofloxacin, benzylpenicillin, azithromycin, cefixime and ceftriaxone was 62.4%, 54.7%, 38.1%, 5.1%, 0.2% and 0%, respectively. The level of decreased susceptibility to ceftriaxone and cefixime was 0.4% and 6.9%, respectively.

Molecular epidemiology and antimicrobial resistance determinants

Analysis of molecular epidemiology was initially performed using the WGS data and three different traditional typing schemes, i.e. NG-MAST, MLST and NG-STAR (Table 1). In total, 302 different NG-MAST STs were found among the 548 isolates, and the most common STs were ST338 (n = 31), ST1407 (n = 29), ST1582 (n = 8), ST2992 (n = 8) and ST6066 (n = 8). Two hundred and seventeen STs were represented by a single isolate and 204 STs were new. Sixty-eight different MLST STs were identified, and the most prevalent were ST1901 (n = 94), ST1588 (n = 92), ST8161 (n = 30), ST1596 (n = 27) and ST13844 (n = 26). Twenty-three STs were represented by a single isolate and 21 STs were new. The NG-STAR classification⁵¹ assigned 214 different STs, with the most common being ST90 (n = 47), ST2217 (n = 11), ST2076 (n = 10), ST2079 (n = 10) and ST1560 (n = 8). One hundred and six NG-STAR STs were represented by a single isolate and 148 STs were new. The goeBURST analysis of the NG-STAR STs revealed 55 CCs and seven ungroupable isolates. Eight main CCs were found, and these included 58.4% of all isolates (Figure 1). The largest NG-STAR CCs were CC309 (n = 88) followed by CC90 (n = 79), CC42 (n = 36), CC63 (n = 36) and CC127 (n = 26). In total, 17 goeBURST groups at level 2 option in PHYLOViZ 2.0⁴⁴ were identified, of which seven groups included two or more NG-STAR STs [Figure 1; Microreact (https://microreact.org/project/Golparian_et_al/205cc441)]. Group one was clearly the largest (189 STs) and the minimum spanning tree was built based on 110 links of six identical loci and 78 links of five identical loci.

Mosaic penA alleles, PBP2 A501 amino acid substitutions, mtrR and porB1b mutations, which combined decrease the ESC susceptibility, were found in 16.6%, 4.6%, 30.7% and 36.3% of isolates, respectively (Table 2). Forty penA alleles (9 new: 4 non-mosaic and 5 mosaic alleles) were found, of which 9 were mosaic alleles. The most common penA alleles were penA-19.001 (n = 116), mosaic penA-34.001 (n = 78) and penA-14.001 (n = 69). Mosaic penA [34.001 (n = 53), 34.018 (n = 2), 93.001 (n = 2), 133.001 (n = 1) and 134.001 (n = 1)] in combination with mtrR (-35 A-deletion and/or G45D) and porB1b mutations were present in 59 isolates (10.8%), of which 26 (44.1%) and 2 (3.4%) had decreased susceptibility or resistance to cefixime and ceftriaxone, respectively. Of these 59 isolates, 29 (49.2%), 52 (88.1%) and 56 (94.9%) belonged to NG-MAST ST1407, MLST ST1901 and NG-STAR CC90, respectively (Figures 1 and 2). The A501V and A501T mutation was found in 2.4% and 2.2% of isolates, respectively.

The 168 (30.7%) isolates with mtrR mutations, resulting in overexpressed MtrC–MtrD–MtrE efflux pump, had a −35 A-deletion in the mtrR promoter (25.5%), mosaic mtrR (2.7%), MtrR G45D substitution (2.4%), a −35 A-deletion + mtr₁₂₀ (0.9%) and A→C SNP in the A-repeat of the mtrR promoter (0.7%). Isolates with mosaic mtrR promoters (n = 15; azithromycin MICs of 1–4 mg/L) also had mosaic sequences in mtrD (100%), mtrC (66.7%) and mtrE (33.3%). In total, 10.9% (n = 60), 9.1% (n = 50) and 6.8% (n = 37) of isolates contained an mtrC mosaic/disruption, an mtrD mosaic and an mtrE mosaic, respectively. An mtrC GC hexarepeat deletion,⁵² potentially contributing to increased antimicrobial susceptibility, was found in 42 (7.7%) isolates with azithromycin MICs of 0.032–4 mg/L; azithromycin-resistant isolates (n = 4; MIC > 1 mg/L) had mosaic mtrD variant J (Figure 3). Overall, 20 mtrD alleles were found and characterized according to the 2016 WHO reference strains where possible. Of these, 10 mosaic mtrD sequences (variants A–E, G, I, J, L and WHO P) in isolates with azithromycin MICs of 0.064–8 mg/L (Figure 3) were found and the two isolates with an azithromycin MIC of 0.064 mg/L also had an mtrC GC hexarepeat deletion. Additionally, seven non-mosaic mtrD variants (variants H, K, M, WHO M, WHO N, WHO O and WHO U) were found with minor differences, and WHO F, WHO L and WHO Z had unique non-mosaic variants (Figure 3). The azithromycin-resistant isolates (n = 28) contained a 23S rRNA C2611T mutation in all four alleles (n = 10; MIC = 4–16 mg/L), mosaic mtrR-mtrC-mtrD (n = 5; MIC = 2 mg/L), mosaic mtrR-mtrC-mtrD-mtrE (n = 4; MIC = 2–4 mg/L), mosaic mtrR-mtrD (n = 4; MIC = 2–4 mg/L), mosaic mtrD-mtrE plus MtrR G45D (n = 2; MIC = 2 mg/L), −35 A-deletion in the mtrR promoter (n = 2; MIC = 2 mg/L) or mosaic mtrD-mtrE (n = 1; MIC = 2 mg/L) (Figure 3). Isolates with a 23S rRNA C2611T mutation (n = 10) belonged to nine NG-MAST STs and seven NG-STAR STs, which were in different parts of the minimum spanning tree (Figure 1). Notably, one isolate belonging to NG-STAR CC90 (ST91) was azithromycin resistant (MIC = 16 mg/L), and had decreased cefixime susceptibility (MIC = 0.125 mg/L) and elevated ceftriaxone MICs (MIC = 0.064 mg/L).

All isolates with a gyrA S91F mutation (n = 300) also had a gyrA D95 mutation (D95A/G/N; n = 301) and showed resistance to ciprofloxacin. One isolate had only a gyrA D95N mutation, and a ciprofloxacin MIC of 0.064 mg/L (susceptible, increased exposure).

The majority of isolates (62.6%) were resistant to tetracycline, mainly due to tetM (73.8%) or rpsJ V57M mutation (99.4%), which was present in nearly all (96.7%) isolates. Of the isolates with resistance to benzylpenicillin (38.1%), most contained β-lactamase (75.6%) and/or ponA L421P mutation (83.7%). No mutations associated with spectinomycin resistance, i.e. in 16S rRNA or rpsE, were found.

Two main WGS lineages and two sublineages were distinguished in the five macroregions of Brazil

The phylogenetic tree of isolates from Brazil (n = 548), Argentina (n = 158) and the 2016 WHO reference strains (n = 14) is shown in Figure 2 (https://microreact.org/project/Golparian_et_al/205cc441). The phylogenetic tree is divided into two main lineages, namely lineage A (n = 506) (https://microreact.org/project/Golparian_et_al/5b338e9e) and B (n = 214) (https://microreact.org/project/Golparian_et_al/5b338e9e) using RAMI with an average PATRISTIC distance between the clusters of 2727.45 SNPs. Lineage B is more diverse than lineage A, with an average PATRISTIC distance between the clones in the lineage of 7413.04 compared with 5539.06 SNPs. Furthermore, lineage B has an average PATRISTIC distance within the cluster of 117.23 SNPs between the nodes and the adjacent ancestral node, while lineage A has an average of 57.73 SNPs from the adjacent ancestral node. Lineage A included the majority of isolates from Brazil (336; 61.3%), Argentina (157; 99.4%) and the non-WT 2016 WHO strains (n = 13). Lineage B consisted of 212 (38.7%) isolates from Brazil, one (0.6%) isolate from Argentina and the wild-type WHO F reference strain. These two lineages were associated with the internationally described antimicrobial-resistant and susceptible lineages, respectively.^5,28,53 The main differences between these lineages included the AMR and AMR determinants to ESCs, azithromycin, ciprofloxacin and benzylpenicillin, which predominated in lineage A. Lineage A was subdivided into sublineages A1 (n = 186) (https://microreact.org/project/Golparian_et_al/3913e80e) and A2 (n = 314) (https://microreact.org/project/Golparian_et_al/ffa72b0b); the only isolates with mtr₁₂₀ (WHO L and five Brazilian isolates) were included in lineage A but not in any of the two sublineages (https://microreact.org/project/Golparian_et_al/a3091c1a). Sublineage A2 contained 155 isolates from Brazil with 82, 41, 113 and 92 isolates with mosaic penA (90.1% of all isolates with mosaic penA), mosaic mtrD (82.0% of all mosaics), mutations in mtrR (67.3% of all mtrR mutations) and porB1b (46.2% of all porB1b mutations), respectively. In sublineage A1, Brazilian isolates (n = 176) with plasmid-mediated resistance to penicillins and tetracycline [β-lactamase (71.0%) and tetM (63.6%), respectively] predominated. Worryingly, in sublineage A2, 77 isolates from Brazil had a close relationship with a separate cluster including 133 isolates with decreased susceptibility or resistance to ESCs from Argentina (https://microreact.org/project/Golparian_et_al/d201c4c4). One isolate each from Brasilia, Salvador and São Paulo (all NG-STAR CC90) and seven isolates (NG-STAR CC90) from Florianópolis in southern Brazil, a frequent holiday location for Argentinians, were within this Argentinian cluster in sublineage A2 (all except one CC90). We found no other main clustering of Brazilian isolates belonging to a geographic region, besides 34 isolates from Manaus that grouped in lineage B (https://microreact.org/project/Golparian_et_al/225c7500).

Discussion

The published gonococcal AMR data and particularly WGS-based molecular epidemiological data are highly limited in Central and South America. Brazil is the largest country in South America and this first national genomic study analysing gonococcal isolates from all macroregions in the country represents the baseline for future molecular epidemiological studies. We have shown that resistance and AMR determinants to ciprofloxacin, tetracycline and benzylpenicillin were widespread in all seven surveyed cities representing all five macroregions, and they should not be used as empirical first-line monotherapy in Brazil, which is in accordance with previous Brazilian studies and the international situation.^{3,5,6,28,33,34,39,54–59}

Although the recommended dual therapy in Brazil²⁷ appears to be highly effective, continued and enhanced national AMR surveillance, ideally including WGS, is imperative because decreased susceptibility to cefixime (6.9%) and ceftriaxone (0.4%), mainly due to mosaic penA-34 combined with mtr and porB1b AMR determinants, is present in Brazil and 5.1% of isolates were azithromycin resistant. Interestingly, 50 Brazilian isolates had a mosaic mtrD, including nine different mtrD variants (variants A–E, G, I, J and L), but only seven of these variants caused resistance to azithromycin (variants B, C, D, E, G, I and J) (Figure 3). The WHO P reference strain had a unique mosaic mtrD (variant WHO P) that in combination with the mtrR promoter and coding sequence mutations and a mosaic mtrC probably cause the resistance to azithromycin (4 mg/L). We found no unique amino acid alterations in all the mosaic MtrD variants causing azithromycin resistance, i.e. of previously described gain-of-function MtrD mutations,⁸ S821A was present in all variants except P and I, and K823E/D was present in all variants except I. In some gonococcal strains, the MtrD mosaic structures in combination with the mosaic structure of other efflux pump subunits can contribute to elevated MICs and resistance to azithromycin, also indicating that the mosaic structure of the pump genes frequently occurs in a single recombination event from related species. In general, it is essential to continue to survey for and subsequently verify new AMR determinants and, when relevant and feasible, also update tools and schemes for AMR prediction and/or typing.

There were a large number of different (n = 302) and new (n = 204) NG-MAST STs in the Brazilian gonococcal population, which was expected due to the limited previous molecular epidemiological data in the region, and ST338 (n = 31) was the most common. ST338 has also previously been associated with high-level tetracycline-resistant (TetM-positive) isolates in MSM in Europe;⁶⁰ similarly, ST2992 (n = 8) has been associated with MSM in Europe.³⁰ Worryingly, the second most common NG-MAST, ST1407 (n = 29), has frequently been associated with MDR including decreased susceptibility or resistance to ESCs, and caused multiple treatment failures with cefixime and ceftriaxone. This clone mostly harbours a mosaic penA, typically penA-34.001 or more rarely penA-10.001.^{1,6,10,11,30,32,33,41} In general, decreased susceptibility and resistance to ESCs internationally have most frequently been caused by a mosaic penA, and the most common NG-STAR ST90 (n = 47) among the Brazilian isolates contained a mosaic penA-34.001. Accordingly, it is crucial to continuously monitor this ST for public health purposes in combination with WGS data that often provide superior resolution, to track any changes in AMR lineages. In the present study, we applied a new approach inspired by MLST eBURST on the seven NG-STAR alleles to obtain information on related STs and define NG-STAR CCs. The minimum spanning tree (Figure 1) shows that CC90 was exceedingly common and had mosaic penA with elevated MICs of ESCs, resistance to ciprofloxacin, tetracycline and benzylpenicillin, and differed from CC63, CC42 and CC2092 which were susceptible to ESCs and ciprofloxacin. In contrast, CC63 had the highest median azithromycin MIC, probably because the majority (91.7%) of these isolates had a mosaic mtrD, an AMR determinant that is not included in the NG-STAR scheme, but these isolates could still be distinguished by the goeBURST approach. This also applies to CC42 and CC2092, where most isolates (94.4% and 100%, respectively) were TetM positive and consequently had a high median tetracycline MIC. The CC1674 consisted of isolates with PorB1a (82.4%) and were mainly susceptible to ESCs and azithromycin, but had high-level resistance to benzylpenicillin and tetracycline due to β-lactamase and TetM, respectively. As the relationship of different CCs gets closer to CC90 they become more MDR and less diverse. The NG-STAR goeBURST approach is a valuable additional tool for gonococcal AMR surveillance, with little in the way of computational requirements, to quickly get a visual snapshot of the circulating strains in any population, predict AMR and generate a simple nomenclature.

WGS phylogenomics showed, similarly to previous gonococcal WGS studies,^{5,28,30,33,53} that the Brazilian isolates form two main lineages, lineage A (n = 336) and lineage B (n = 212), with limited geographical structure related to the macroregions in Brazil. Lineage A and lineage B are associated with the AMR and antimicrobial-susceptible lineages, respectively, described in global gonococcal populations.^{5,28,30,33,53} Furthermore, running RAMI with the average PATRISTIC distance showed that lineage A is less diverse than lineage B. Isolates with decreased susceptibility or resistance to ESCs grouped in sublineage A2 (https://microreact.org/project/Golparian_et_al/ffa72b0b) including NG-STAR CC90 (https://microreact.org/project/Golparian_et_al/9f08a12b). Isolates with 23S rRNA C2611T mutation were found in several clades and did not appear to be due to any clonal expansions. Finally, the Brazilian isolates belonging to the internationally spread MDR clone NG-MAST ST1407,^{1,6,10,11,30,33,34,41} MLST ST1901 and NG-STAR ST90 were mainly clustered in sublineage A2, closely related to the isolates from Argentina with decreased susceptibility or resistance to ESCs.³⁴

Possible limitations of the Brazilian GASP,³ which are addressed in the ongoing second round of isolate collection, include representativeness (only men with urethral discharge and no extra-genital sites were sampled) and lack of clinical and epidemiological data of the patients.

In conclusion, this is the first study using WGS on gonococcal isolates from all five macroregions in Brazil, cultured in the Brazilian national GASP in 2015–16,³ and forms the genomic baseline for future studies in Brazil and internationally. The high prevalence of AMR and AMR determinants for ciprofloxacin, tetracycline and benzylpenicillin, and the high number of isolates with mosaic penA and azithromycin resistance mutations, prompt continued and strengthened AMR surveillance, including WGS, of N. gonorrhoeae in Brazil. It is also imperative to set up and/or substantially strengthen the culture-based gonococcal AMR surveillance, ideally supported by WGS, in other countries in South and Central America, and the Caribbean. National and international leadership and commitment (political and financial) are essential to achieve this.^3,39,61

Acknowledgements

We are very grateful to all the members of the Brazilian-GASP Network. The study was carried out with data provided by the Department of Chronic Conditions and Sexually Transmitted Infection, of the Secretariat of Health Surveillance of the Brazilian Ministry of Health.

Members of the Brazilian-GASP Network

Felipe de Rocco, Marcos André Schörner, Thais Mattos dos Santos, Jéssica Motta Martins, Hanalydia de Melo Machado, Ligia Maria Bedeschi Costa, Maria Rita Rabelo Costa, Simone Veloso Faria de Carvalho, Luciane Guimarães Dias, Waldemara de Souza Vasconcelos, Jairo de Souza Gomes, Maria de Fátima Pinto da Silva, Maria da Purificação Pereira da Silva, Rosana Barboza de Matos, Roberto José Carvalho da Silva, Cláudio Campos do Porto, Lidiane da Fonseca Andrade, Lúcia de Fátima Mendes Pereira, Leonor Henriette de Lannoy, Letícia Eidt, Guilherme Henrique de Oliveira Arnhold, Chayane Ariel Souza Coelho Muniz, Loeci Natalina Timm, Cassia Maria Zoccoli, Maria Luiza Bazzo, Lisléia Golfetto, Mauro Cunha Ramos and William Antunes Ferreira.

Funding

The study was supported by the Örebro County Council Research Committee and the Foundation for Medical Research at Örebro University Hospital, Örebro, Sweden, and the Brazilian Ministry of Health, through its Secretariat for Health Surveillance and its Department of Chronic Conditions and Sexually Transmitted Infection.

Transparency declarations

None to declare.

References

Author notes