Characterization of the novel DNA gyrase inhibitor AZD 0914 : Low resistance potential 1 and lack of cross - resistance in Neisseria gonorrhoeae . 2 3 4 5 6 Richard

36 37 The unmet medical need for novel intervention strategies to treat Neisseria gonorrhoeae 38 infections is significant, and is increasing as rapidly emerging resistance in this pathogen is 39 threatening to eliminate currently available treatment options. AZD0914 is a novel bacterial 40 gyrase inhibitor that possesses potent in vitro activity against high-level ciprofloxacin and 41 extended-spectrum cephalosporin resistant isolates and is currently in clinical development 42 for the treatment of N. gonorrhoeae infections. The propensity to develop resistance against 43 AZD0914 was examined in N. gonorrhoeae and found to be extremely low, a finding 44 supported by similar studies in Staphylococcus aureus. Genetic characterization of both first 45 step and second step mutants that exhibited decreased susceptibility to AZD0914 identified 46 substitutions in the conserved GyrB TOPRIM domain, confirming DNA gyrase as the primary 47 target of AZD0914 and providing differentiation from fluoroquinolones. Analysis of available 48 bacterial gyrase and topoisomerase IV structures including those bound to fluoroquinolone 49 and non-fluoroquinolone inhibitors has allowed the rationalization of the lack of cross50 resistance that AZD0914 shares with fluoroquinolones. Microbiological susceptibility data 51 also indicate that the topoisomerase inhibition mechanisms are subtly different between 52 N. gonorrhoeae and other bacterial species. Taken together, these data support the 53 progression of AZD0914 as a novel treatment option for the oral treatment of 54 N. gonorrhoeae infections. 55 56 57 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom INTRODUCTION 58 Neisseria gonorrhoeae is the obligate human pathogen responsible for the sexually 59 transmitted disease gonorrhea, an extremely prevalent infection, with global estimates by the 60 World Health Organization in 2012 exceeding 100 million infections per year (1). Although 61 N. gonorrhoeae infections were once relatively easily treated with sulfonamides or penicillin, 62 the emergence and spread of multi-drug resistant isolates has increased the difficulty of 63 effective treatment and novel therapeutic alternatives are needed (2). 64 65 Several fluoroquinolone drugs, notably ciprofloxacin, were used widely for the treatment of 66 gonorrhea starting in the early 1990s. These classes of antibacterial agents are potent DNA 67 synthesis inhibitors that target the bacterial Type II topoisomerases, DNA gyrase and 68 Topoisomerase IV, which are essential for bacterial DNA replication, recombination, and 69 separation of daughter chromosomes. However, N. gonorrhoeae isolates that were resistant 70 to ciprofloxacin arose quickly and treatment guidelines needed to be amended as effective 71 treatment of infections was being compromised. Indeed, the global burden of ciprofloxacin72 resistant N. gonorrhoeae isolates is now significant and is between 35% and >95% in many 73 countries (3-10) although the level in the USA is slightly lower at 15% (11). Ciprofloxacin 74 resistance in N. gonorrhoeae, like in other species, is typically manifested by point mutations 75 in the quinolone-resistance-determining region (QRDR) of the A subunits of the Type II 76 topoisomerases (GyrA and ParC) (5, 12-14). However, although in other species like 77 Staphylococcus aureus and Escherichia coli where low level fluoroquinolone resistance can 78 be manifested through point mutations in the topoisomerase B subunits, these mutations are 79 rarely identified in N. gonorrhoeae (15, 16). The high resistance to fluoroquinolones has 80 resulted in the removal of this class of drug from treatment guidelines to manage 81 N. gonorrhoeae infections. Intramuscular ceftriaxone, an expanded-spectrum cephalosporin 82 that targets the penicillin-binding proteins essential for bacterial cell wall synthesis, now 83 represents the last treatment option for the management of N. gonorrhoeae infections. 84 However, organisms with decreased susceptibility to these cephalosporins are now 85 emerging with several reports of high-level resistance to these drugs (5, 17-19), essentially 86 removing the last option for empirical therapy. Furthermore, even though drugs like 87 ciprofloxacin are no longer widely used to treat gonorrhea, resistance to this drug persists in 88 these multi-drug-resistant pathogens and has resulted in patients having fewer effective 89 treatment options (3-10). Taken together, this has created an urgent need for the discovery 90 and development of novel, first-in-class oral agents that can successfully treat infections 91 caused by N. gonorrhoeae carrying pre-existing resistance caused by long term 92 fluoroquinolone and cephalosporin usage. 93 94 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom AZD0914 is a novel spiropyrimidinetrione that is active against N. gonorrhoeae via inhibition 95 of DNA synthesis (20) and is equally effective against isolates that are both resistant and 96 susceptible to existing agents, including international refernce strains and the 2008 WHO 97 reference panel (16). Given the propensity for N. gonorrhoeae to develop resistance, in this 98 study we examined the ability of AZD0914 to suppress the emergence of resistance in vitro 99 in N. gonorrhoeae isolates. Finally, we also evaluated the propensity of resistance 100 emergence in isolates that carried resistance to fluoroquinolones and cephalosporins to 101 understand if there was a variation in these populations that could impact the utility of this 102 drug in the future. 103 104 MATERIALS AND METHODS 105 Bacterial Strains. The recent N. gonorrhoeae clinical isolates (ARC4673, ARC4676, and 106 ARC4677) were obtained from a clinic in Nanjing, China. The high-level ceftriaxone107 resistant isolate, first described by Camara et al. (17) was kindly provided by Dr. Ardanuy 108 from the L'Hospitalet de Llobregat in Barcelona, Spain and is referred to as ARC4682. The 109 three Chinese clinical isolates all contained the same ’A’ nucleotide deletion in the mtrR 110 promoter region as has been described for the ceftriaxone-resistant isolate (17). The 111 reference isolate ATCC 49226 was obtained from the American Type Culture Collection, and 112 was found not to contain a deletion in the mtrR promoter region. 113 Antimicrobial susceptibility testing. The minimum inhibitory concentration (MIC) against 114 each isolate was determined using agar dilution methodology according to Clinical 115 Laboratory Standards Institute (CLSI) documents M07-A9 and M100-S23 (21, 22). An 116 appropriate reference compound was included in each test, and N. gonorrhoeae ATCC 117 49226 was used as the quality control reference isolate (The CLSI approved QC range for 118 AZD0914 has been determined to be 0.06 – 0.5 μg/mL). AZD0914 was dissolved at high 119 concentrations in dimethyl suloxide (DMSO) and diluted. Ciprofloxacin was obtained from 120 MP Biomedicals (Santa Ana, CA). 121 Frequency of resistance studies. A suspension of cells (~OD600 3.2) was made from a 122 freshly grown (24hr) chocolate agar plate. Dilutions were plated on compound-free plates to 123 determine the CFU/mL of the suspension. First step spontaneous resistance frequencies 124 were determined after plating 100μL aliquots onto 1x GC agar (BD, Franklin Lakes, NJ) plus 125 Isovitalex (BD, Franklin Lakes, NJ) containing AZD0914 at 2-fold increasing concentrations 126 of the agar dilution MIC. Plates were incubated at 36C  1 C in 5%CO2 for 48hr. The 127 number of colonies that grew were used to calculate resistance frequencies. If no colonies 128 were observed the frequency of resistance was expressed as less than the frequency of 129 obtaining one resistant variant. Subsequent selection experiments were performed on GC 130 agar supplemented with 1% hemoglobin (BD, Franklin Lakes, NJ) as a higher burden on 131 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom bacteria could be evaluated and were repeated on multiple test occasions to ensure that 132 mutants that were obtained represented independent events and were not siblings. 133 Whole genome sequencing and analysis. Genomic DNA purified on the Maxwell 16 134 platform (Promega, Madison WI) and quantitated using the Qubit fluormeter (Invitrogen Life 135 Technologies, Grand Island, NY) was used as input material for library construction. DNA 136 libraries were prepared using the Nextera library construction protocol (Illumina, San Diego, 137 CA, USA) following the manufacturer’s instructions and sequenced on a MiSeq Sequencer 138 (Illumina, San Diego, CA, USA). For each isolate, approximately 2.5 million 150 bp paired 139 end sequence reads were de novo assembled and analyzed using the CLCBio suite of 140 software tools (Cambridge, MA, USA). The Multi Locus Sequence Type (MLST) profiles of 141 the isolates were determined by comparison of the seven allele sequences against the 142 curated public database available at http://pubmlst.org/neisseria and they were also typed 143 using the NG-MAST database available at http://www.ng-mast.net 144 Molecular analysis and transformation. The gyrB gene was amplified from strain 4676145 D1 using the For (5’-CCACAAGGTGCGGACTGTTTTGGC-3’) and 146 Rev (5’-AAAGGAGACGCGGCAAGCTGCCCG-3’) oligonucleotides (Eurofins MWG Biotech, 147 Huntsville, AL) in a standard polymerase chain reaction (PCR) using the High Fidelity PCR 148 mix (Roche, Nutley, NJ). Cycling conditions were a 30 second denaturation at 94°C, a 30 149 second annealing step at 55°C, followed by a 3 minute extension at 72°C. The polymerase 150 chain reaction (PCR) product was purified using a QIAquick PCR Purification kit (Qiagen, 151 Valencia, CA). Transformation in N. gonorrhoeae was performed by spotting 20 μL of DNA 152 in water (representing between 1 μg and 10 μg DNA) onto a GC agar plate (BD) and, after 153 allowing the DNA solution to dry onto the plate, streaking the recipient N. gonorrhoeae 154 strain, ATCC 49226, through the dried DNA spot and incubating at 36°C in 5% CO2. Cells 155 that grew on the area of the dried DNA were then streaked onto GC agar containing 156 AZD0914 at concentrations that were 2x – 4x MIC for the recipient strain. Single colonies 157 were then purified and characterized by whole genome sequencing as described above. 158 Molecular modeling 159 Crystal structures that are available in the public domain were used for the structural 160 analyses. These were the structures of the catalytic core of S. aureus DNA gyrase, in the 161 apo form (PDB:2XCO) and in complex with ciprofloxacin (PDB:2XCT), the catalytic core of 162 Streptococcus pneumoniae topoisomerase IV bound with levofloxacin (PDB:3K9F) and the 163 quinazolinedione PD 0305970 (PDB:3LTN) as well as the structure of Acinetobacter 164 baumannii topoisomerase IV in complex with moxifloxacin (PDB:2XKK). The structures 165 were visualized and the mutations identified in this study were mapped using the graphics 166 program Pymol (Schrödinger, Portland, OR; www.pymol.org). 167 168 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom RESULTS AND DISCUSSION 169 Characterization of N. gonorrhoeae isolates and frequencies of spontaneous 170 resistance. The susceptibility of AZD0914 (Figure 1) to four fluoroquinolone-resistant 171 N. gonorrhoeae isolates is shown in Table 1. Despite high-level resistance to ciprofloxacin, 172 the MIC of AZD0914 was 0.03 – 0.06 μg/mL confirming that no cross-resistance with 173 ciprofloxacin exists. Importantly, AZD0914 was equally potent against the high-level 174 ceftriaxone resistant isolate (Table 1). Sequence analysis of the four Type II topoisomerase 175 subunit proteins demonstrated that all four isolates encoded highly similar GyrA proteins. All 176 carried the Ser91Phe and Asp95Ala/Gly double mutation known to commonly confer 177 fluoroquinolone resistance (5, 13). Three of the isolates also carried a Ser87Arg substitution 178 in the quinolone-resistance determining region of ParC, whereas ARC4676 carried two 179 distinct substitutions, Gly85Cys and Ala89Thr and ARC4673 had an additional Ser88Pro (Table 180 2). The ParE ATPase subunit of Topoisomerase IV was identical between all four isolates, 181 and the equivalent subunit in DNA gyrase (GyrB) in two isolates was identical to the 182 N. gonorrhoeae FA1090 query sequence, whereas ARC4673 and ARC4682 carried 183 substitutions in the N-terminal and C-terminal regions, outside of the QRDR, respectively 184 (Table 2). Despite the similarity of the sequences of the Type II topoisomerase subunits, 185 especially GyrA and ParE, MLST analysis confirmed that these isolates were unrelated 186 sharing only 4-5 alleles in common (Table 2). The ceftriaxone resistance in ARC4682 is 187 mediated by the mosaic penA allele which is the mechanism of high-level cephalosporin 188 resistance in N. gonorrhoeae (17, 19). Overall, the data shows that AZD0914 clearly lacks 189 cross-resistance to these pre-existing fluoroquinolone and cephalosporin resistance 190 mechanisms. 191 192 To study the frequency of spontaneous resistance, a fresh suspension of each 193 N. gonorrhoeae isolate as well as the fluoroquinolone-susceptible reference isolate (ATCC 194 49226) was plated onto agar containing increasing concentrations of AZD0914. The 195 frequencies of spontaneous resistance to AZD0914 in the five N. gonorrhoeae isolates were 196 all low, with no resistant variants being isolated in three of the isolates. The frequencies 197 ranged from 1.5 x 10 to <5.2x 10 at 4-fold the MIC, and stably resistant variants could only 198 be isolated from one isolate, ARC4682 (Table 2). At 8-fold the MIC, no variants could be 199 obtained in any isolate and the frequencies ranged from <3.3 x 10 to <2.1 x 10 (Table 2). 200 In contrast, the induction of resistance emergence by ciprofloxacin in N. gonorrhoeae has 201 been shown to be higher, with 3 of 4 isolates generating mutants at 8-fold MIC at 202 frequencies in the 10 range (23). At concentrations of AZD0914 at 2-fold the MIC, isolates 203 ARC4682 and ARC4676 both had a small number of colonies which were characterized 204 further (see below). This low level of resistance development against AZD0914 was not 205 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom solely limited to N. gonorrhoeae. When tested against five Staphylococcus aureus isolates, 206 the frequency of resistance to AZD0914 was measured below the limit of detection in all 207 isolates (ranging from <1.1x10 to <2.2x10) with no variants being isolated at 4-fold MIC 208 and higher although three S. aureus ATCC 33591 variants were obtained at 2-fold MIC. 209 210 Characterization of first step resistant mutants. Selected colonies from the 211 N. gonorrhoeae ARC4682 and ARC4676 plates were tested for changes in MIC values 212 compared to the parent isolate. The MIC values for AZD0914 against both parental isolates 213 tested one doubling dilution higher that on the plates prepared for the selection experiment, 214 at 0.125 μg/mL (Table 3). After a minimum of six independent replicates, 0.125 μg/mL 215 represents the modal MIC value for these isolates. The two ARC4676 variants that grew on 216 the 2-fold MIC plate (4676-D1 and 4676-D2) both had an AZD0914 MIC value of 1 μg/mL, 217 which represented an 8-fold increase over the parent isolate (Table 3). Given the similarity 218 in target with AZD0914, the susceptibility of these variants was also tested against 219 ciprofloxacin. Both mutants notably demonstrated significant reduction in the ciprofloxacin 220 MIC, by 32 to 64-fold, bringing the variants into the susceptible range for ciprofloxacin (Table 221 3). 222 223 N. gonorrhoeae isolate ARC4682 had variants emerge at both 2-fold and at 4-fold the MIC to 224 AZD0914 and three representative colonies were selected from each concentration for 225 further characterization. All three of the variants from the 2-fold MIC plate (4682-D1, D2 and 226 D3) and two variants from the 4-fold MIC plate (4682-D4 and D5) had AZD0914 MIC values 227 of 2 μg/mL, which represents a 16-fold increase over the parent isolate (Table 3). The final 228 variant from the 4-fold MIC plate (4682-D6) had an AZD0914 MIC value of 1 μg/mL (Table 229 3). Similar to the mutants from N. gonorrhoeae ARC4676, the MIC value for ciprofloxacin 230 was consistently lower, but only by 4-fold and remained in the non-susceptible range (Table 231 3). 232 233 The genome sequences of the resistant variants were obtained, and the sequences reads 234 were assembled and mapped against the parent isolate to identify any genetic changes that 235 could be attributable to the decrease in susceptibility to AZD0914. Very few differences 236 were observed between the sequences of the parental isolates and the mutants. All of the 237 ARC4682 variants, irrespective of the selection plate from which they were obtained, 238 contained an Asp429Asn substitution in the GyrB polypeptide (Table 3; Figure 2A). In 239 contrast, the two ARC4676 variants both contained a Lys450Thr substitution in GyrB (Table 3; 240 Figure 2A). Additional replicate selection experiments using ARC4676 and ARC4682 were 241 performed on rich media that supported optimal growth of the N. gonorrhoeae isolates. 242 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom Even on supplemented media, the frequency of obtaining resistant variants against 243 AZD0914 was equally low. Representative variants from these selections were 244 characterized by whole-genome sequencing and the Asp429Asn substitution in GyrB, with a 245 concomitant elevation of the AZD0914 MIC value, was independently observed on multiple 246 occasions in both isolates. Interestingly however, the acquisition of this mutation in the 247 ARC4676 background (represented by 4676-D3), while manifesting an equivalent 16-fold 248 elevation in the AZD0914 MIC value, maintained the same susceptibility to ciprofloxacin, 249 which is in contrast to the 4-fold reduction in MIC consistently observed with the Asp429Asn in 250 mutants obtained in the ARC4682 background on multiple test occasions (Table 3). 251 252 A sequence alignment of the Type II topoisomerase B subunits (GyrB and ParE) from other 253 bacterial species shows that the Asp429 residue is exquisitely conserved (Figure 2A) and 254 forms part of the EGDSA motif that has been previously implicated in resistance to certain 255 fluoroquinolones in other species (24). As further supportive evidence of the importance of 256 this conserved Asp residue in the potency of AZD0914, one of the three S. aureus ATCC 257 33591 mutants selected against AZD0914 that gave a four-fold decrease in susceptibility 258 (MICparent = 0.25 μg/mL increasing to 1 μg/mL) carried an equivalent Asp437Asn substitution. 259 The remaining two S. aureus ATCC 33591 mutants, which also exhibited a similar four-fold 260 decrease in AZD0914 susceptibility contained a Ser442Pro substitution in GyrB just distal to 261 the conserved Asp residue. All of these S. aureus variants also showed a modest 2-fold 262 increase in the fluoroquinolone MIC (MICparent = 0.25 μg/mL increasing to 0.5 μg/mL). In 263 addition, a related compound, QPT-1, also selected for substitutions at this conserved Asp 264 residue in S. aureus and a Asp437Asn substitution resulted in elevation of the MIC values of 265 QPT-1 and ciprofloxacin by 8 to 16-fold and 2 to 4-fold, respectively (25). 266 267 The GyrB Lys450 residue forms part of the PL(R/K)GK motif that is conserved in both 268 bacterial and eukaryotic topoisomerases, underlying its key functional role in the catalysis of 269 DNA cleavage (26). While the presence of a mutation in a resistant variant with no other 270 changes observed by whole genome sequencing strongly suggests that a substitution is 271 responsible for the elevated MIC values, the Lys450Thr substitution had not been implicated 272 in decreased susceptibility to this chemical series and had only been isolated in the initial 273 selection experiment. Accordingly, a back-cross experiment was performed to confirm that 274 this mutation was responsible and sufficient to confer AZD0914 resistance. The mutated 275 gyrB carrying the Lys450Thr substitution was amplified from the 4676-D1 mutant and 276 transformed back into the susceptible ATCC49226 N. gonorrhoeae strain. The resulting 277 transformants, confirmed to contain the 4676-D1 gyrB gene by genome sequencing, had an 278 AZD0914 MIC of 1μg/mL (8-fold above the parent), confirming the importance of this residue 279 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom in AZD0914 susceptibility. Further, as was noted with the original mutant in a ciprofloxacin280 resistant background, the introduction of the Lys450Thr mutation into a ciprofloxacin281 susceptible strain resulted in hyper-susceptibility and an 8-fold reduction in MIC (Table 3). 282 283 Selection of second step mutants. The 8-16 fold increase in MIC of AZD0914 observed 284 with the first step mutants was larger than expected, based on studies using this compound 285 in S. aureus, where the equivalent Asp437Asn substitution in GyrB only resulted in a 4-fold 286 increase. Even though the frequencies of resistance emergence to AZD0914 were low, the 287 subsequent decrease in susceptibility needs to be considered with the emerging clinical 288 pharmacokinetic data to determine whether these first-step mutants are likely to be 289 pharmacodynamically covered. Fluoroquinolone resistance in N. gonorrhoeae and other 290 species typically occurs in relatively small increments as the molecules target both DNA 291 gyrase and Topoisomerase IV, and the stepwise accumulation of mutations in both targets in 292 a back-and-forth pattern results in high level resistance. The preference for an initial target 293 depends on a combination of the specific bacterial species and the fluoroquinolone molecule 294 being tested. 295 296 Given the level of conservation between the B subunits of DNA gyrase (GyrB) and 297 Topoisomerase IV (ParE) in N. gonorrhoeae in the region surrounding Asp429 and Lys450 298 (Figure 2B) there seemed no a priori reason why AZD0914 would not target both Type II 299 topoisomerase enzymes in N. gonorrhoeae. Therefore, the first step Asp429Asn (4676-D3 300 and 4682-D1) and Lys450Thr (4676-D1) mutants were used in a selection experiment to try 301 and obtain second step mutants to gain some insight into whether this molecule targets both 302 Type II topoisomerases in N. gonorrhoeae. 303 304 The selection experiment was performed on GC agar enriched with 1% haemoglobin which 305 supported more robust growth, and after 24 hr or 48-hr incubation, no resistant colonies 306 were evident for any of the three isolates at 2-fold MIC or higher concentrations, with 307 calculated resistance frequencies being below the limit of detection, ranging from <1.1 x 10 308 to <7.9 x 10. After 72 hr incubation, a number of small colonies were evident on the 2x 309 MIC and 4x MIC plates. Representative colonies from these extended incubation plates 310 were selected and after several passages on drug free media, which helped increase the 311 robustness of their growth, were checked for any changes in MIC values compared to both 312 the original parent isolate and the corresponding first step variant. The variant of 4676-D1 313 (Lys450Thr) exhibited an AZD0914 MIC value of 2 μg/mL irrespective of whether they were 314 isolated from the 2x or 4x MIC plate which represents a modest 2-fold elevation of the MIC 315 over the first step variant. The first step variants 4676-D3 and 4682-D1 (Asp429Asn) had an 316 on Jne 4, 2017 by gest httpaac.asm .rg/ D ow nladed fom AZD0914 MIC value of 2 μg/mL and the second selection variants characterized from the 4x 317 MIC plates exhibited a further 4-fold increase in MIC (Table 3). 318