The pre-antibiotic era

Until the introduction of antibiotics, treatment consisted of rest, abstention from alcohol and sexual activity, systemic treatment with one or other of the balsams and urethral irrigations.5 Acute urethritis was treated by anterior urethral irrigation with a 1/3000 dilution of warm potassium permanganate for 2 weeks. Acute retention of urine was usually relieved by the administration of a soap and water enema, oral laudanum and hot baths; catheterisation was performed in those unresponsive to conservative management. Many clinicians performed urethroscopy on male patients after resolution of symptoms to detect persistent glandular infections and to dilate the urethra in cases of anterior and posterior urethral strictures. During the first world war, soldiers received prophylactic packets containing condoms, calomel ointment and Argyrol. Post-coital treatment centres, comprising a urethral irrigation facility, were also set up in many barracks.

Sulphonamides

The discovery of sulphanilamide by Gerhard Domagk in 1935 was a major breakthrough.5 Subsequent studies in Germany, UK and the USA showed that sulphanilamide could cure 80–90% of gonorrhoea cases.6 However, by 1944, many gonococci were clinically resistant and, by the end of the decade, over 90% of N gonorrhoeae isolates were microbiologically resistant to sulphonamides.6 7 Subsequent combination treatment with trimethoprim and sulfamethoxazole (TMP-SMX) was used with success in the 1970s, although with the disadvantage that patients required multidose treatment.8 9

Sulphonamide resistance in gonococci results from either oversynthesis of p-aminobenzoic acid, which effectively dilutes the antimicrobial agent, or mutations in the genes encoding dihydropteroate synthetase, which produce a mutant enzyme with reduced affinity for sulphonamide.10 Given that gonococcal dihydrofolate reductase has a low affinity for trimethoprim, which can be further reduced by genetic mutations, the TMP-SMX combination offered little advantage over sulphonamides alone.

Penicillins

Cecil Paine, a microbiologist working in Sheffield, was the first person to use a crude extract from Penicillium notatum to cure gonococcal infection in an infant with ophthalmia in 1930.11 Penicillin was first used to treat gonococcal urethritis in 1943. Although initially reserved for sulphonamide-resistant cases, penicillin subsequently became the drug of choice for gonorrhoea.12–14 Initially, cure rates with penicillin exceeded 95% and total doses as low as 45 mg were prescribed owing to scarcity and the high cost of the drug.13 Over time, gonococci became less susceptible as penicillin minimum inhibitory concentrations (MICs) rose, requiring higher doses to effect cure.15–17 Early treatment failure due to chromosomal resistance was eventually described and results from serial changes in the structure of penicillin-binding proteins (PBPs) and/or outer membrane permeability. Several genetic loci are involved, including the penA, penB, ponA, mtr and pem genes.

Mutation of the penA gene results in four- to eightfold increases in the penicillin MIC as a result of lowered PBP-2 affinity consequent to insertion of asparagine at amino acid position 345.18 19 The likely effect of this insertion is to alter a hydrogen-bonding network involving Asp-346 and the SXN triad motif, composed of Ser-362, Ser-363 and Asn-364, at the active site.19 Further mutations have been described near the C terminus of PBP-2, which lower the rate of acylation for penicillin fivefold relative to wild type through a small disordering of residues in the active site.19

The MtrC-MtrD-MtrE multiple transferable resistance efflux pump expels penicillin, tetracycline and macrolides.20 Mutation of the mtrR locus, which encodes a transcriptional repressor of the mtrCDE operon, increases gonococcal resistance to these agents.21 In the presence of mtr mutations, penB mutations reduce outer membrane permeability and raise the penicillin MIC fourfold, due to adjacent amino acid mutations in porin PorB1b—namely, Gly-120-to-Lys (G120K) or Gly-120-to-Asp (G120D) and Ala-121-to-Asp (A121D).22 23 Combined mutations in the penA, penB and mtr genes can raise the penicillin MIC 120-fold and pem mutations may further modify the expression of these three genes.10 24 Chromosomal resistance may be dependent upon a point mutation in the PBP-1 encoding ponA gene (ponA1), which results in a lower rate of acylation by penicillin.25 In addition, in the laboratory setting, decreased susceptibility to penicillin results from a penC (pilQ2) mutation in strains also containing the mtrR and penB resistance determinants.26 A Glu-666-to-Lys (E666K) mis-sense mutation in the pilQ gene interferes with the formation of the PilQ secretin complex, a major gonococcal outer membrane porin and a key requirement for type IV pilus formation.26 This mutation is not seen in clinical isolates, probably because it compromises type IV pilus function and thus pathogenic potential.

The emergence of high-level, plasmid-mediated resistance to penicillin in 1976 marked the effective end of penicillin as a therapeutic agent for gonorrhoea in many areas of the world.27 28 Penicillinase-producing N gonorrhoeae (PPNG) contained plasmids encoding a TEM-1-type β-lactamase. The plasmids were readily transferable between gonococci and PPNG spread rapidly. Several types of penicillinase-producing plasmids have been described, including the Asia, Africa, Toronto, Rio, Nîmes and New Zealand plasmids.29–31 The Asia, Africa and Toronto types have been associated with epidemic outbreaks.29 These β-lactamase plasmids may be characterised as either deletion derivates of the Asia plasmid (Africa, Toronto and Rio) or insertion derivatives of either the Asia (New Zealand) or Africa (Nîmes) plasmids.32

Tetracyclines

Tetracyclines were discovered in the late 1940s and were used initially to treat gonorrhoea in patients allergic to penicillin. Tetracyclines were effective in post-gonococcal urethritis and are today often given as additional treatment for suspected or proven chlamydial co-infection. Gonococci became less susceptible to tetracycline over time and chromosomal resistance soon appeared.16 The main mechanism of tetracycline resistance in gonococci involves ribosomal protection, whereby binding of tetracyclines to the bacterial ribosome is impaired. It is due to a Val-57-to-Met (V57M) point mutation in the rpsJ (tet-2) gene encoding ribosomal protein S10.33 Additional combined mutations in the penB, penC and mtr loci, modified by mutations in the tem gene, may further decrease N gonorrhoeae susceptibility by increasing efflux and bacterial cell entry of tetracyclines.24 26

Plasmid-mediated tetracycline-resistant N gonorrhoeae (TRNG) were first reported in the USA and shortly afterwards in the Netherlands.34 35 The 25.2 MDa (40.6 kb) TRNG plasmid carries the streptococcal tetM determinant responsible for tetracycline resistance.36 The TetM protein reduces the susceptibility of ribosomes to tetracyclines by tetracycline release, competes with elongation factor EF-G and has ribosome-dependent GTPase activity.37 Restriction endonuclease mapping and DNA sequencing demonstrated that the American-type and Dutch-type TRNG plasmids were different.38 39 American-type TRNG were initially described in the UK and Africa, Dutch-type TRNG isolates were reported in Asia and South America, and variants based on restriction endonuclease analysis have also been described in Uruguay and South Africa.40–42 TRNG are now widespread, possibly as a result of widespread use of tetracyclines in STI management.43

Spectinomycin and aminoglycosides

Spectinomycin was developed and marketed specifically for the treatment of gonorrhoea in the early 1960s. It was useful for treating gonorrhoea caused by PPNG although is not effective in treating pharyngeal gonorrhoea.44 45 Spectinomycin resistance was first reported in a penicillin susceptible gonococcus in the Netherlands in 1967, and subsequently in a PPNG isolate acquired in the Philippines in 1981.46 47 Spectinomycin was introduced as first-line treatment for US military personnel in the former Republic of Korea in 1981 but, after only 4 years, an 8.2% treatment failure rate was reported, which resulted in the replacement of spectinomycin with ceftriaxone.48 49

Kanamycin is cheap and sometimes used to treat gonorrhoea in Indonesia.50 51 Gentamicin is the national first-line treatment for gonorrhoea in Malawi and has also been used to treat gonococcal infections in Mongolia.51–53 Data on clinical resistance are few and the microbiological resistance breakpoints uncertain. Studies in Malawi confirm that gonococci remain susceptible after 14 years of first-line use.54

Although it is known that aminoglycosides bind to the bacterial ribosome and inhibit protein synthesis, their precise mode of action remains unclear.55 Several mechanisms of aminoglycoside resistance have been described among bacteria, including decreased drug uptake and accumulation, modification of the ribosomal target, drug efflux and enzymatic drug modification.55 Among gonococci, it appears that resistance is drug-specific and due to single-step mutations in various loci rather than enzymatic inactivation.10 56 The loci responsible for spectinomycin (spc) and streptomycin (str) resistance are genetically linked. The kanamycin resistance locus (kan) is located between the spc and str loci and also linked to the spc, str and rif loci. These loci are thought to represent a cluster of genes encoding for ribosomal proteins.10

Macrolides

Clinical and laboratory susceptibility data have demonstrated that erythromycin is of limited use in gonorrhoea treatment.57 58 Azithromycin is more active against gonococci and has the advantage of single-dose oral treatment. Azithromycin is not recommended by the WHO as a first-line treatment option for gonorrhoea owing to concerns over resistance and side effects of the 2 g dose.59–61

Macrolides act by binding to the 50S ribosomal subunit and inhibit the elongation of peptide chains. Bacterial resistance to these antimicrobial agents may result from drug efflux and/or modification of the ribosomal target, by either methylase-associated modification of 23S rRNA or genetic mutations in 23S rRNA.62 The MtrR repressor-regulated MtrC-MtrD-MtrE efflux system of N gonorrhoeae exports macrolides. Increased efflux may occur by deletion or insertional inactivation of either the mtrR gene or the mtrR promoter.20 63 64 Another macrolide efflux pump, encoded by mef, has been detected in some isolates, although its contribution to gonococcal macrolide resistance remains unclear.65 Expression of several 23S RNA methylases, encoded by the ermB, ermC and ermF genes, is responsible for modification of the gonococcal ribosomal target.66 67 These methylase genes are associated with conjugative transposons which facilitate interbacterial spread.66 Mutations in the peptidyltransferase loop of domain V of 23S rRNA also confer gonococcal resistance to macrolides.68

In vitro resistance to azithromycin is now emerging, perhaps as a consequence of the widespread use of this drug in STI clinics. The earliest reports of decreased susceptibility and resistance came from Latin America in the mid–late 1990s.69 A cluster of azithromycin-resistant gonococci (MICs 2–4 mg/l), due to a 153 base-pair insertion sequence located between the mtrR/mtrC promoter and the mtrC gene, was identified in patients in Kansas City in 1999.70 Most azithromycin-resistant gonococci isolated in Canada between 1997 and 1999 exhibited increased efflux, owing to the previously described single-base-pair (A) deletion in the 13 bp inverted repeat of the mtrR promoter region.63 68 Two isolates had a novel C2599T mutation (N gonorrhoeae numbering) in the 23S rRNA rrl gene. A high-level azithromycin-resistant gonococcus (MIC>2048 mg/l), isolated in Argentina in 2001, was recently reported to possess a different A2143G 23S rRNA rrl mutation.71 72 Recently, a cluster of N gonorrhoeae isolates with high azithromycin MICs (4096 mg/l) was reported in the UK.73

Quinolones

Fluoroquinolones, such as ciprofloxacin and ofloxacin, were widely used to treat gonorrhoea from the mid-1980s onwards. Although not suitable for pregnant women or children, ciprofloxacin has the advantage of minimal side effects, single-dose administration and excellent efficacy at all anatomical sites, including the oro-pharynx. Initially, low doses of 250 mg were prescribed but, by 1990, treatment failure was reported.74 The recommended treatment dose was raised to 500 mg but, with time, clinical resistance emerged, initially in the Asia-Pacific Region.75–77 Quinolones were abandoned as first-line gonococcal treatment in most Asia-Pacific countries in the mid-to-late 1990s, and subsequently in the USA, Europe and parts of Africa.78 79 In the USA and UK, quinolone-resistant gonococci appear particularly prevalent among men-who-have-sex-with-men.80 81

Quinolones interfere with bacterial DNA metabolism by the inhibition of two enzymes, DNA gyrase and topoisomerase IV. In Gram-negative bacteria, DNA gyrase is the primary target, whereas topoisomerase IV appears to be more important in Gram-positive bacteria. DNA gyrase, which catalyses supercoiling of linear DNA, is composed of GyrA and GyrB subunits, encoded by gyrA and gyrB, respectively. The function of topoisomerase IV, encoded by parC, is poorly understood. Characteristic point mutations within quinolone resistance determining regions of gyrA and parC genes are associated with resistance. Resistance is normally seen in gonococci with gyrA mutations alone or with mutations in both gyrA and parC genes.82–84

Cephalosporins

Cephalosporins were first discovered in 1945 and modern variants are chemical modifications of the prototypic molecule.85 Intramuscular ceftriaxone and various oral third-generation agents, such as cefixime and cefpodoxime, have been used to treat gonorrhoea following the demise of fluoroquinolones. Among oral cephalosporins, only cefixime has met the criterion on the lower bound of the 95% CI of the WHO recommended cure rate of 95% or greater.86 While most countries have opted for single-dose cefixime, other oral cephalosporins have been used when cefixime was not available—for example, cefpodoxime in the USA, ceftibuten in Hong Kong, cefditoren and celdinir in Japan, and cefuroxime in the UK.85 87–89

Following widespread use of oral cephalosporins in Japan, gonococci with decreased susceptibility soon emerged and treatment failures have now been reported.90 91 Many of these isolates were also resistant to fluoroquinolones, tetracycline and penicillin. Intravenous ceftriaxone (1 g) is now recommended as first-line treatment for urogenital and pharyngeal gonorrhoea in Japan.92 93 Gonococci with reduced cephalosporin susceptibility are spreading within the Asia-Pacific and other regions of the world.85 94–97 UK and USA-based surveillance programmes are now reporting ‘MIC creep’ to both oral cephalosporins and ceftriaxone, mirroring the situation observed with penicillin in the 1940–50s.4 Although there are, as yet, no confirmed cases of ceftriaxone-resistant urogenital gonorrhoea, two cases of pharyngeal gonorrhoea recently failed intramuscular ceftriaxone treatment at a dose of 250 mg.98 The oro-pharynx is thought to be a key site for the development of novel mechanisms of gonococcal resistance as a result of genetic exchange between commensal bacteria and N gonorrhoeae.85

In gonococci, reduced susceptibility and resistance to cephalosporins is chromosomally mediated. Japanese researchers identified mosaic penA genes in those gonococci with reduced susceptibility to oral cephalosporins.99 100 The gonococcal mosaic PBP-2 contains fragments of the PBP-2 proteins of commensal Neisseria species commonly found in the oro-pharynx, specifically Neisseria cinerea, Neisseria flavescens, Neisseria perflava, Neisseria polysaccharea and Neisseria meningitidis.99–102 Pulse-field gel electrophoresis of mosaic PBP-2 isolates initially highlighted their genetic similarity, but subsequent geographical and temporal variability has been reported.97 99 The Gly-545-to Ser (G545S), Ile-312-to-Met (I312M) and Val-316-to-Thr (V316T) mutations have been identified as the key PBP-2 amino acid substitutions responsible for reduced cephalosporin susceptibility in mosaic PBP-2 strains.103 Recombination of a mosaic penA gene from parent strain 35/02 into the cephalosporin susceptible N gonorrhoeae strain FA19 increased the cefixime MIC 100-fold and the ceftriaxone MIC 20-fold.104 However, the cefixime and ceftriaxone MICs of the transformed strain were substantially lower than the mosaic penA parent, indicating that additional genetic mechanisms were required to produce the higher cephalosporin MICs seen in some clinical isolates.

It is now clear that mosaic PBP-2 genes are not present in all isolates exhibiting decreased susceptibility to cephalosporins, and that a small number of susceptible isolates may also possess such genes.105 Osaka et al reported mosaic and non-mosaic PBP-2 amino acid sequences from gonococci with reduced cephalosporin susceptibility.102 The authors detected a previously described Ala-501-Val (A501V) mutation in the PBP-2 protein of the non-mosaic isolates, which raises cephalosporin MICs by up to fourfold.103 A variation of this mutation, Ala-501-to-Thr (A501T), has been reported from Australia and it is expected that other PBP-2 mutations will be identified in the future.106 Modelling of gonococcal mosaic and non-mosaic PBP-2 proteins suggests that decreased cephalosporin susceptibility is due to conformational alteration of the β-lactam-binding pocket.102

Additional mutations in the mtrR and penB genes have been described among gonococci with the greatest reduction in cephalosporin susceptibility.104 107 108 These include the previously reported single-base-pair (A) deletion in the mtrR promoter, a Gly-45-to-Asp (G45D) substitution in the DNA binding motif of MtrR, and penB mutations resulting in alterations in amino acid residues 101 and 102 in the putative loop 3 of PorB1b.107 Although it is generally thought that ponA, which encodes for PBP-1, is not essential for cephalosporin resistance, Takahata et al reported that a Leu-421-to-Pro (L421P) mutation was detected among all the gonococcal strains they studied with reduced susceptibility to cephalosporins.103 Working with gonococci containing the mosaic penA allele, Zhao et al demonstrated that reversion of the L421P ponA mutation back to wild type reduced the level of penicillin resistance twofold but had no effect on cefixime or ceftriaxone resistance.104 These data support the hypothesis that reduced cephalosporin susceptibility arose in circulating chromosomally mediated penicillin-resistant strains with existing ponA mutations.

The above genetic mechanisms appear to be of varying importance for generating reduced susceptibility to oral cephalosporins and ceftriaxone. Recombinant PBP-2 binding assays show that the mosaic PBP-2 elements inhibit oral cephalosporin, but not ceftriaxone, binding.109 Reduced ceftriaxone susceptibility is more dependent on the A501V PBP-2 mutation, as well as mtr and penB mutations.104 106 To date, there has been no reported plasmid-mediated cephalosporin resistance in N gonorrhoeae.

The future

The global increase in antimicrobial-resistant gonorrhoea is occurring within and across different classes of antimicrobial agents. New definitions for multidrug-resistant and extensively drug-resistant Neisseria gonorrhoeae have recently been published to assist surveillance efforts.4 Given the lack of new therapeutic drugs in the drug delivery pipeline, gonorrhoea may eventually become untreatable. The new parenteral antimicrobial agents, ertapenem and tigecycline, are active against gonococci but require further testing against multidrug-resistant strains.110 111 The wisdom of continued single-dose treatment requires challenging and alternative treatment strategies such as extended treatment, multidrug treatment and cycling of antimicrobial agents should be considered. It is unlikely that national treatment guideline recommendations will be sufficient to contain the emergence of multidrug resistance owing to the wider clinical use of antibiotics. The acquisition by N gonorrhoeae of extended-spectrum β-lactamases and carbapenemases, as has occurred in other Gram-negative bacteria, remains the ultimate threat to gonococcal control programmes.112 113 The impact of untreatable gonorrhoea on HIV transmission could be enormous in countries with a high HIV prevalence. The main public health challenge is to reduce the global burden of gonorrhoea, and hence dependence on antimicrobial agents in the longer term.4 89 The gonococcus continues to fight back, through evolutionary driven acquisition of antimicrobial resistance determinants. Although we have not yet had the ‘knockout’ punch, the gonococcus appears to be winning on points.