Novel Specific Metallo-β-Lactamase Inhibitor ANT2681 Restores Meropenem Activity to Clinically Effective Levels against NDM-Positive Enterobacterales

ABSTRACT The global dissemination of metallo-β-lactamase (MBL)-producing carbapenem-resistant Enterobacterales (CRE) is a serious public health concern. Specifically, NDM (New Delhi MBL) has been a major cause of carbapenem therapy failures in recent years, particularly as effective treatments for serine-β-lactamase (SBL)-producing Enterobacterales are now commercially available. Since the NDM gene is carried on promiscuous plasmids encoding multiple additional resistance determinants, a large proportion of NDM-CREs are also resistant to many commonly used antibiotics, resulting in limited and suboptimal treatment options. ANT2681 is a specific, competitive inhibitor of MBLs with potent activity against NDM enzymes, progressing to clinical development in combination with meropenem (MEM). Susceptibility studies have been performed with MEM-ANT2681 against 1,687 MBL-positive Enterobacterales, including 1,108 NDM-CRE. The addition of ANT2681 at 8 μg/ml reduced the MEM MIC50/MIC90 from >32/>32 μg/ml to 0.25/8 μg/ml. Moreover, the combination of 8 μg/ml of both MEM and ANT2681 inhibited 74.9% of the Verona integron-encoded MBL (VIM)-positive and 85.7% of the imipenem hydrolyzing β-lactamase (IMP)-positive Enterobacterales tested. The antibacterial activity of MEM-ANT2681 against NDM-CRE compared very favorably to that of cefiderocol (FDC) and cefepime (FEP)-taniborbactam, which displayed MIC90 values of 8 μg/ml and 32 μg/ml, respectively, whereas aztreonam-avibactam (ATM-AVI) had a MIC90 of 0.5 μg/ml. Particularly striking was the activity of MEM-ANT2681 against NDM-positive Escherichia coli (MIC90 1 μg/ml), in contrast to ATM-AVI (MIC90 4 μg/ml), FDC (MIC90 >32 μg/ml), and FEP-taniborbactam (MIC90 >32 μg/ml), which were less effective due to the high incidence of resistant PBP3-insertion mutants. MEM-ANT2681 offers a potential new therapeutic option to treat serious infections caused by NDM-CRE.


RESULTS
MBL-positive isolates in a 2015 to 2018 global collection of Enterobacterales clinical isolates. A total of 218,521 Enterobacterales isolates were collected by IHMA from 2015 to 2018. Most of the isolates were collected in Europe (n = 94,132; 43.1%), followed by North America (n = 37,603; 17.2%), Asia/Pacific (n = 34,436; 15.8%), Latin America (n = 32,302; 14.8%), and Middle East/Africa (n = 20,048; 9.2%). Overall, 3.8% of isolates were nonsusceptible to MEM (MIC of $2 mg/ml). Approximately 75% of these (n = 6,072 isolates) were molecularly characterized for the presence of class A, B, and D carbapenemases. Although similar proportions of MEM nonsusceptible (MEM-NS) isolates were obtained over the 4-year period, 4%, 2.7%, 4%, and 4.3%, respectively (Fig. 1), a clear increase in the percentage of MBL-positive isolates could be observed within the MEM-NS population, from 21.6% in 2015 to 31.2% in 2018 (Fig. 1). This was driven by a statistically significant rise in NDM-positive Enterobacterales, which accounted for 15.7% of MBL-positive isolates in 2015 and for 26.6% in 2018 (P , 0.001), as well as the slight decrease in the percentage of the two other major MBLs, IMP (from 0.8% to 0.5%) and VIM (from 5.1% to 4.4%), over the same period (Fig. 2).
Restoration of MEM antibacterial activity with different ANT2681 concentrations. In vitro susceptibility testing against a small set of Gram-negative and Gram-positive organisms showed that ANT2681 had no intrinsic antibacterial activity (data not shown) and, therefore, its only role in the combination is to restore the activity of MEM to clinically efficacious levels.
Susceptibility testing of all 360 MBL-producing Enterobacterales (not carrying KPC or OXA) collected in 2016 was performed with MEM and increasing concentrations of ANT2681 in order to determine the level of ANT2681 necessary to restore MEM activity (MIC 90 # 8 mg/ml). The data showed that ANT2681 concentrations of $8 mg/ml were required to reduce the MEM MIC 90 from .64 mg/ml to #8 mg/ml for all MBL-positive Enterobacterales and from .64 mg/ml to #4 mg/ml for NDM-producing CRE (Table 1). Further increasing the ANT2681 concentration provided little benefit and, therefore, ANT2681 was fixed at 8 mg/ml in the subsequent in vitro susceptibility studies.  Antibacterial activity of MEM and MEM-ANT2681 (8 lg/ml) against different subsets of organisms. Susceptibility studies were performed against 1,687 MBL-positive Enterobacterales and the results analyzed with regard to the presence of specific carbapenemases. The data confirmed that ANT2681 significantly potentiates MEM activity against a broad range of clinical isolates producing NDM as the sole carbapenemase (n = 1,108), reducing MEM MIC 90 from .32 mg/ml to 8 mg/ml (Table 2). ANT2681 also restored MEM activity against the five isolates carrying NDM plus VIM, but it was less active against VIM-positive isolates (n = 251) and inactive versus IMP-positive organisms (n = 49) ( Table 2). As expected, the effectiveness of the combination was substantially reduced if a serine-carbapenemase was also present in the organism. A slight MEM-potentiating effect was observed against Enterobacterales that carried NDM plus either KPC or OXA (n = 245), although not enough to be clinically useful.
In these studies, the combination of MEM and ANT2681, both at 8 mg/ml, inhibited the growth of 78.7% of the entire collection of Enterobacterales tested and of 89.2% of strains that carried only an MBL carbapenemase. Specifically, it inhibited 92.5% of the NDM-CRE isolates, 74.9% of the VIM-positive, and 85.7% of the IMP-positive Enterobacterales (Table 2). This concentration of MEM-ANT2681 also inhibited growth of 21.2% of NDM-positive strains in which a KPC or OXA carbapenemase was also present ( Table 2).
Similar MIC 90 values were obtained against isolates carrying NDM-4 (n = 31) and NDM-6 (n = 23), but a higher value was observed against the pool of organisms carrying all other NDM variants (Table 3). This is due to the fact that 9/10 of the NDM-9 isolates, representing a third of the strains in the pooled group, had MEM-ANT2681 MICs of $16 mg/ml. There is a high probability that these particular isolates are clonally related, as they were all identified as NDM-9/CTX-M-15/SHV-OSBL K. pneumoniae and all were collected in Guatemala, with 8/9 isolated in 2018 from the same hospital. The fact that MEM-ANT2681 displays a MIC of 0.25 mg/ml against the only non-Guatemalan NDM-9 K. pneumoniae isolate, collected in the Philippines, indicates that ANT2681 is active against this NDM variant and that some other determinant(s) accounts for the high MIC in the Guatemalan strains. This remains to be confirmed by additional studies. Comparative analysis of the antibacterial activity of MEM-ANT2681 and comparator antibiotics against NDM-CRE clinical isolates. Several comparator agents, either marketed drugs or combinations in clinical development with activity against MBL-CREs, were also tested against one or both of the 2017 and 2018 NDM-CRE collections ( Table  4). The activity of MEM-ANT2681 (MIC 90 , 8mg/ml; 92.5% inhibited at this concentration) compared favorably to that of other antibiotics tested, such as cefepime (FEP)-taniborbactam (MIC 90 , 32mg/ml), cefiderocol (FDC) (MIC 90 , 8mg/ml; 84.3% susceptible), colistin (CST) (MIC 90 , .16 mg/ml), and eravacycline (ERV) (MIC 90 , 2mg/ml; 52.2% susceptible) ( Table 4). The only other combination with clear potential for clinical activity against NDM-CRE was aztreonam-avibactam (ATM-AVI) (MIC 90 , 0.5 mg/ml) ( Table 4).
All of the Enterobacter cloacae isolates, 99.0% of Escherichia coli, and 90.3% of K. pneumoniae strains from the 2017 and 2018 collections had MEM-ANT2681 MICs of #8 mg/ml (   Table 4). Activity of MEM-ANT2681 versus NDM-E. coli isolates carrying an insertion in PBP3. ATM-AVI, FDC, and FEP-taniborbactam displayed noticeably higher MIC 90 values for NDM-E. coli isolates than for other NDM-positive species. It was speculated that this may be due to the presence of mutations in PBP3, the primary penicillin-binding protein target for several b-lactams, insertions in which have been reported to contribute to reduced susceptibility to ATM, FDC, and FEP (19)(20)(21). Hence, the ftsI gene (which encodes PBP3) was amplified and sequenced from 30 E. coli isolates that displayed ATM-AVI MICs $ 4 mg/ml (5 collected in 2017) or FDC MICs $ 8 mg/ml (25 collected in 2018). All these E. coli isolates also displayed FEP-taniborbactam MICs $ 16 mg/ml. Sequencing data revealed that all isolates had a four amino acid insertion at position 333 of the PBP3 protein (29 isolates having YRIN and 1 isolate having YRIK). Recently, Sadek et al. (21) have shown that the presence of this modified PBP3 is not sufficient to confer resistance to ATM-AVI and, in fact, all resistant isolates investigated in their study also carried CMY-42. CMY enzymes were also identified in the five E. coli that displayed ATM-AVI MICs $ 4 mg/ml. Although these PBP3 mutations do not affect activity of carbapenems, the presence of NDM in these particular isolates resulted in MEM MICs of $32 mg/ml. Addition of ANT2681 decreased MEM MICs against these E. coli isolates to between 0.06 and 4 mg/ml.

DISCUSSION
Antimicrobial resistance is seriously compromising modern medicine and our way of life. Among antibiotic-resistant bacteria, CRE has been classified as a priority 1 pathogen by WHO (22) and as an urgent threat by the CDC (23), and there is a global consensus that novel therapies are needed to treat infections caused by these Gram-negative pathogens.
The first case of KPC-producing K. pneumoniae emerged in the US in 1996 and was reported in 2001 (24). Ten years later, this resistance mechanism had disseminated worldwide and had been detected in all clinically relevant Enterobacterales (25). Fortunately, by the time this carbapenemase emerged, the b-lactamase inhibition potential of the diazabicyclooctanes (DBO) and boronic acid derivatives had already been discovered (26,27) and development of novel b-lactamase inhibitors with anti-KPC activity proceeded rapidly, resulting in the launch of ceftazidime-avibactam    However, in contrast to the situation when KPC first emerged, there were no preexisting MBL inhibitors in development and the difficulty in identifying novel and specific MBL inhibitors, together with the currently limited commercial potential in major pharmaceutical markets, has made progress in this area slow. Hence, safe and effective treatment options for serious infections caused by NDM-CRE are still required. ANT2681 is the culmination of a chemistry program designed to deliver a potent and selective NDM inhibitor. This compound shows no P450 inhibition, hERG inhibition, mammalian cytotoxicity, or genotoxicity, is selective toward other metalloenzymes (14), and was well tolerated in safety toxicology/pharmacology studies at exposures well above those necessary for efficacy (15). As the area under the concentration curve from 0 to 24 h (AUC 0-24 ) is the pharmacodynamic index driving ANT2681 efficacy (15) it should be feasible to administer an efficacious dose in combination with 2 g MEM, administered as a 3-h infusion every 8 h, which would provide coverage of CREs with MEM MICs of up to 8 mg/ml. ANT2681 significantly potentiates MEM activity against a broad range of NDM-producing clinical isolates (including all NDM variants present in this study), reducing MEM MIC 50 /MIC 90 from .32/>32 mg/ml to 0.25/8 mg/ml. In addition, the combination of MEM and ANT2681, both at 8 mg/ml, inhibited the growth of 74.9% of the VIM-positive and 85.7% of the IMP-positive Enterobacterales tested.
MEM-ANT2681 inhibits growth of 92.5% of NDM-CRE isolates at the anticipated susceptibility breakpoint (BP) of #8 mg/ml MEM and compares favorably to the activity of comparator drugs, such as FDC (84% susceptibility at the CLSI BP of #4 mg/ml) or FEP-taniborbactam, a combination being evaluated in phase 3 clinical trials (30) (78.3% susceptibility at the CLSI BP for FEP of #8 mg/ml). Of the additional comparator drugs tested (ATM-AVI, CST, and ERV), ATM-AVI was the only one that appears to have the clinical potential to treat NDM-CRE. While ATM is a monobactam and is not hydrolyzed by MBLs, MBL-positive isolates often carry ESBLs which do hydrolyze it (31). Since AVI is an effective inhibitor of ESBLs, the combination ATM-AVI could be an effective treatment option against MBL-producing CREs (32) and is currently undergoing phase 3 clinical trials. In the meantime, clinicians are already administering this combination by using two drugs already approved by the FDA, namely, ATM and CAZ-AVI (33)(34)(35). This practice provides a very clear indication of the shortness of drugs available to treat these deadly pathogens.
There are other combinations at different stages of clinical development with reported activity against NDM-positive CRE. Meropenem-nacubactam (36) and cefepime-zidebactam (37) rely on the direct antibacterial activity of the BLi for their effect against MBL-CRE. QPX7728 is a dual SBL and MBL inhibitor, reported to be developed as a partner to different b-lactam antibiotics, which shows a very good microbiological profile when combined with MEM (38).
Worryingly, several recent publications report the emergence of E. coli isolates harboring a four-amino-acid insertion in PBP3, either YRIN or YRIK after residue 333 (19)(20)(21), or TIPY after residue 344 (39), which contribute to the reduced susceptibility of these isolates to b-lactam antibiotics. These strains, which have evolved through mutation and recombination and have been selected by b-lactam use, have started to disseminate globally and seem to preferentially acquire carbapenemase genes (40); they represent a new emerging clinical threat, as most novel b-lactam and b-lactam/BLI combinations, including FDC, FEP-taniborbactam, and ATM-AVI, are not active against them. In fact, MIC 90 values for these drugs against the subset of NDM-E. coli tested in this study were higher than those obtained against the entire collection of NDM-CRE, and molecular characterization of 30 of these isolates with reduced susceptibility confirmed the presence of the YRIN/YRIK insertion at position 333 of PBP3 in all cases. As the lack of MEM activity against these E. coli isolates is due to the presence of NDM, MEM-ANT2681 effectively restores the carbapenem activity, lowering MEM MICs to #4 mg/ml, and is thus a very promising therapeutic option for this important emerging threat.
In summary, ANT2681 restores MEM activity to clinically effective levels against NDM-CRE and, therefore, the MEM-ANT2681 combination represents a novel and potentially effective therapy to treat serious or life-threatening infections caused by these organisms, for which there are few treatment options that offer both efficacy and a favorable safety and tolerability profile. The availability of targeted narrow-spectrum drugs such as MEM-ANT2681 would support antibiotic stewardship efforts by helping to conserve broad-spectrum drugs for medical situations that require immediate empirical therapy or for which no other treatment option is available.
Bacterial strains. Enterobacterales tested in this study (n = 1,687) were selected from the IHMA Europe Sàrl, (Monthey, Switzerland) 2015 to 2018 global surveillance studies frozen stock culture collection based on their MEM nonsusceptible phenotype (MEM MIC of $2 mg/ml) (41) and the presence of an MBL gene. All Enterobacterales were originally isolated from specimens of patients (one isolate per patient) with a documented intraabdominal, urinary tract, skin and soft tissue, lower respiratory tract, or bloodstream infection. The identities of all isolates were confirmed by IHMA using matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (Bruker Daltonics, Billerica, MA).
The presence of the following clinically relevant b-lactamase genes had been evaluated by PCR as described previously (42) Antimicrobial susceptibility testing. Antimicrobial susceptibility testing of MEM-ANT2681 and comparator agents was performed by broth microdilution following Clinical and Laboratory Standards Institute (CLSI) guidelines using custom panels prepared by IHMA (41,43) in cation-adjusted Mueller-Hinton broth (CAMHB) (Becton, Dickinson). Iron-depleted CAMHB was used to test FDC. The panels were incubated at 35°C for 16 to 20 h in ambient air before MIC endpoints were read visually. MIC values corresponded to the first well with no visible growth.
Quality control (QC) testing was performed each day of testing as specified by CLSI using the following isolates: E. coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, and K. pneumoniae ATCC 700603. MICs were interpreted using either CLSI or EUCAST (European Committee on Antimicrobial Susceptibility Testing) guidelines or both when available (41,44).
MEM and MEM-ANT2681 were tested each year over a period of 4 years against all MBL-positive MEM-NS isolates collected by IHMA during the previous year as part of their global surveillance studies. These