The metallo-β-lactamases strike back: emergence of taniborbactam escape variants

ABSTRACT

Metallo-β-lactamases (MBLs) have evolved relatively rapidly to become an international public health threat. There are no clinically available β-lactamase inhibitors with activity against MBLs. This may change with the introduction of cefepime-taniborbactam. Herein, we review three manuscripts (S. I. Drusin, C. Le Terrier, L. Poirel, R. A. Bonomo, et al., Antimicrob Agents Chemother 68:e01168-23, 2024, https://doi.org/10.1128/aac.01168-23; C. Le Terrier, C. Viguier, P. Nordmann, A. J. Vila, and L. Poirel, Antimicrob Agents Chemother 68:e00991-23, 2024, https://doi.org/10.1128/aac.00991-23; D. Ono, M. F. Mojica, C. R. Bethel, Y. Ishii, et al., Antimicrob Agents Chemother 68:e01332-23, 2024, https://doi.org/10.1128/aac.01332-23) in which investigators describe elegant experiments to explore MBL/taniborbactam interactions and modifications to MBLs, in response, to reduce the affinity of taniborbactam. Challenges with MBL inhibition will not disappear; rather, they will evolve commensurate with advancements in medicinal chemistry.

COMMENTARY

Metallo-β-lactamases (MBLs) are adept at evolving and have developed a formidable international footprint (1). They are divided into three subclasses (i.e., B1, B2, and B3) based broadly on differences in amino acid sequences at active sites, loop architecture, substrate profiles, and zinc content and ligands (2, 3). However, MBLs are highly divergent even within subclasses (2). The B1 subclass of MBLs is arguably the most clinically relevant; B1 MBLs are largely plasmid borne and can readily be transferred between bacterial strains (2). New Delhi Metallo-β-lactamases (NDM), Verona Integron Metallo-β-lactamases (VIM), and Imipenemases (IMP) are the most problematic members of the B1 subclass.

MBLs present a major public health threat through their production by both Enterobacterales and non-fermenting Gram-negative organisms. This is exemplified by bla_NDM which was first identified in 2008 in Escherichia coli and Klebsiella pneumoniae isolates from a Swedish patient returning from New Delhi, India (4). Since then, NDMs have become widespread around the world—achieving a reach far beyond health care settings—and have spread to virtually every continent (5). Similarly, bla_VIM was first identified in 1997 in Verona, Italy and has also spread internationally. As an example, VIM-producing Pseudomonas aeruginosa comprises upward of 20% of carbapenem-resistant P. aeruginosa in Latin America—paralleling estimates from Italy (6).

Several attributes make MBLs particularly concerning when compared to the serine carbapenemases (e.g., Klebsiella pneumoniae carbapenemases). First, MBLs have the ability to readily hydrolyze virtually all β-lactam antibiotics, with the notable exception of aztreonam. Second, their spread has been largely unfettered—mostly due to inadequate infection prevention measures—but also in part because of the lack of clinically available β-lactamase inhibitors with activity against MBLs. Third, MBLs are identified in both environmental and nosocomial reservoirs, making them somewhat ubiquitous. Additionally, both MBL variants with increasing affinity for zinc, higher stability under zinc-limiting conditions, or lower zinc requirements are emerging, enabling them to thrive even in settings of relative zinc scarcity, which is common in states of human infection (7–11). To this point, the NDM-15 variant has evolved to function with a single zinc-binding site, rather than the more typical dual zinc-binding sites (8). Finally, new MBL variants are being described at an impressive frequency, highlighting the remarkable adaptability of these enzymes. To date, more than 60, 80, and 100 variants have been described in the NDM, VIM, and IMP families, respectively (12).

Recently, three intriguing manuscripts were published in Antimicrobial Agents and Chemotherapy, which describe the role of taniborbactam (13), a bicyclic boronate β-lactamase inhibitor, and the factors that appear to both promote and limit its ability to inactivate some of the most clinically relevant MBLs (14–16). The potential availability of taniborbactam in the near future is very welcome news as we are in dire need of MBL-active β-lactamase inhibitors. However, reports of NDM-variants (e.g., NDM-9) and VIM-variants (e.g., VIM-83) with single amino acid substitutions making them less susceptible to inhibition by taniborbactam temper some enthusiasm (15, 17). Regrettably, taniborbactam is also inefficient at inhibiting IMP-type MBLs, a significant contributor to carbapenemase production in the Far East and Australia (18).

In one of the three aforementioned manuscripts, Drusin and colleagues sought to better elucidate the molecular basis by which certain MBLs (including IMP-1) are not inhibited by taniborbactam (14). Using elegant docking simulations and electrostatic surface calculations, they advanced the notion that “MBL escape variants” can arise from changes in electrostatic features due to single amino acid substitutions in MBL active site loops, leading to ineffective binding of taniborbactam (14). For example, the replacement of glutamic acid for a positively charged lysine in position 149 in NDM-9 decreases the negative charge that interacts with the amine group in the side chain of taniborbactam, reducing the affinity of taniborbactam for NDM-9 (14).

In the second manuscript, Ono et al. investigated the role of amino acid substitutions in K224 (16). K224 appears to serve as an anchoring residue by forming two hydrogen bonds with taniborbactam (19, 20). Through construction of a library of NDM-1 variants at position 224 by site-saturation mutagenesis, they found that a specific variant, K224I, conferred a comparable level of resistance to cefepime-taniborbactam as to cefepime, indicating that the inhibitory activities of taniborbactam were unable to rescue cefepime from this variant NDM (16). The microbiological and biochemical findings were supported by structural modeling and docking simulations. Their simulations reveal how specific single amino acid substitutions in NDM-1 decrease the inhibitory activity of taniborbactam despite only a modest increase in the inhibitor constant Ki.

Finally, Le Terrier and colleagues assessed the inhibitory potential of taniborbactam against a broad range of MBLs from all three subclasses through measurements of minimum inhibitory concentrations of taniborbactam-β-lactam associations, hydrolytic inhibitory concentrations of taniborbactam, and site-directed mutagenesis (15). Fortunately, taniborbactam was able to inhibit the activity of most members of the B1 subclass; notable exceptions being SIM-1 (i.e., Seoul Imipenemase 1) and a few specific NDM and VIM variants (i.e., NDM-9, NDM-30, and VIM-83), which all differ from NDM-1 or VIM-1 by single amino acid substitutions. Additionally, these investigations bolster previous findings, indicating that members of the B3 subclass appear to be able to escape the activity of taniborbactam, and taniborbactam activity against the B2 subclass is variable (19).

The findings of these three investigations highlight the critical need to support the development of new β-lactamase inhibitors with activity against MBLs and the necessity of laboratory-directed studies targeted to understand structure activity relationships as new inhibitors are being developed. However, the quest to identify effective MBL inhibitors poses challenges for several reasons. In contrast to serine β-lactamases, the active site in MBLs is in a shallow groove with only a few contact points to bind inhibitors (21). Moreover, the large structural diversity across MBLs with relatively low homology among active site residues poses challenges (21). Additionally, MBLs belong to a much larger superfamily of metalloproteins with diverse biological functions beyond β-lactam hydrolysis; members share a common protein fold resulting in similarities at active sites (22). Therefore, achieving the fine balance of developing effective broad spectrum MBL inhibitors that remain selective enough to avoid toxicity though inhibition of off-target enzymes remains challenging.

Based upon the available preclinical and clinical data and the favorable results of a clinical trial of 436 adults with urinary tract infections (23), cefepime-taniborbactam will likely be the first clinically available β-lactam-β-lactamase inhibitor with targeted activity against the most common MBLs in the B1 subgroup. Of note, patients with MBL-producing infections were excluded from the trial. Taniborbactam may be soon followed by xeruborbactam, another bicyclic boronate (24). Presumably, distinct amino acid substitutions will be identified that will challenge xeruborbactam and other inhibitors in development. Only through meticulous experiments focused on exploring substrate/inhibitor interactions will we be able to define the precise niche of new inhibitor combinations as they are introduced into the clinic. The challenges of overcoming MBL inhibition will not disappear but will evolve commensurate with advancements in medicinal chemistry. We applaud the investigators of the aforementioned studies in Antimicrobial Agents and Chemotherapy for their innovative work in advancing the science of MBLs and effective inhibitors.