Susceptibility of Clinical Enterobacterales Isolates With Common and Rare Carbapenemases to Mecillinam

Purpose: To investigate the susceptibility of carbapenemase-producing Enterobacterales (CPE) to mecillinam based on the recently updated European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints for uncomplicated Urinary Tract Infection (uUTI). Methods: The challenge collection consisted of 105 molecularly characterized Enterobacterales [Klebsiella spp. (N = 49), Escherichia coli (N = 30), Enterobacter cloacae (n = 13), Citrobacter freundii (N = 9), Proteus mirabilis (N = 3), and Raoultella ornithinolytica (N = 1)]. Isolates produced OXA-48 (N = 18), OXA-48-like (N = 18), VIM (N = 22), NDM (N = 22), KPC (N = 12), IMI (N = 9), IMP (N = 6), GES (N = 1), OXA-58 (N = 2) or combinations thereof (N = 5). MICs of carbapenems were determined by agar gradient diffusion (AGD). MICs of mecillinam were assessed by agar dilution (reference method) and compared to disk diffusion (DD) and AGD. Results: Overall 23/105 CPE (21.9%) were susceptible to mecillinam. Susceptibility was observed in E. coli (N = 12), E. cloacae (N = 7), and Klebsiella pneumoniae (N = 4) producing IMI, OXA-48, OXA-48-like, and NDM-1 carbapenemases. MIC50 for mecillinam in all isolates was 128 mg/L while MIC50 for meropenem was 8 mg/L. Lower MICs for mecillinam were found in IMI (MIC50 8 mg/L) and OXA-48-like (MIC50 16 mg/L) producers. The comparison of the different susceptibility methods showed very major errors of 12.2% with AGD and 8.5% with disk diffusion when compared to the reference method. Conclusion: Mecillinam susceptibility was restricted to isolates producing IMI-, OXA-48-like, and NDM-1 carbapenemases and was documented despite high carbapenem MICs in some isolates. Mecillinam could be a promising oral antimicrobial in uUTI caused by E. coli and E. cloacae isolates carrying IMI- and OXA-48-like carbapenemases; however, susceptibility testing by AGD and disk diffusion remains problematic.


INTRODUCTION
Limited options for antimicrobial therapy in infections caused by carbapenemase-producing Enterobacterales (CPE) are a major global health concern with increasing morbidity and mortality (Nordmann and Poirel, 2014;Martin et al., 2018). In urinary tract infections (UTI) the prevalence of multidrug-resistant Enterobacterales is increasing in in-and outpatients (Mazzariol et al., 2017). In an observational study of over 20.000 communityonset culture-positive UTIs from the United States, carbapenemresistant Enterobacterales (CRE) were found in 3.0% of cases (Zilberberg et al., 2017). Repurposing previously approved antimicrobials according to their potential in these difficult to treat infections is an important strategy for future disease management (Theuretzbacher et al., 2015).
Pivmecillinam is the oral prodrug of mecillinam, a beta-lactam antibiotic that has been known since the 1970s (Lund and Tybring, 1972;Neu, 1976). Its activity against multidrug-resistant Enterobacterales has led to first-line recommendations for the treatment of uncomplicated UTI (uUTI) in several guidelines (e.g., the updated German guidelines or the international clinical practice guidelines of Infectious Diseases Society of America (IDSA) and European Society of Clinical Microbiology and Infectious Diseases (ESCMID; Gupta et al., 2011;Kranz et al., 2017). The European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoint is set at S ≤ 8 mg/L, R > 8 mg/L and can be applied to Escherichia coli, Klebsiella spp., Proteus mirabilis, Enterobacter spp., Citrobacter spp., and Raoultella spp. (EUCAST, 2020). CLSI has recently introduced a breakpoint for mecillinam, with S ≤ 8 mg/L, I = 16 mg/L, and R ≥ 32 mg/L, currently for E. coli only and not for other Enterobacterales (CLSI, 2020).
The aim of the present study was to assess the in vitro activity of mecillinam in all species for which EUCAST breakpoints have been defined and in isolates producing frequent and rare carbapenemases.

Comparison of Different Methods for Susceptibility Testing
The discrimination of susceptible (≤8 mg/L) and resistant (>8 mg/L) isolates was clear when using the reference method agar dilution; borderline cases were rare in the cohort. Categorical agreement of the reference method with AGD and disk diffusion was 90.4 Frontiers in Microbiology | www.frontiersin.org and 93.3%, respectively (Table 3). Compared to the high concordance of agar dilution results upon precision testing, performance of AGD and disk diffusion was more variable (Figure 1). Six isolates showed disagreement in both agar diffusion methods with agar dilution. In addition, AGD disagreed in an additional four isolates and disk diffusion in one isolate compared to agar dilution. Most major errors occurred in K. pneumoniae isolates (7/11; 63.6%); AGD showed more major errors (10/11) than disk diffusion (7/11). Overall AGD and disk diffusion yielded very major errors (false susceptibility) in 10/82 (12.2%) and 7/82 (8.5%) isolates, respectively. Supplementary Table S2 shows all isolates with mecillinam MICs defined as very major errors stratified by carbapenemases and MICs. No difficulties were documented in the interpretation of the carbapenemase-negative comparators, including the ATCC quality control-strain at any stage of the study.

DISCUSSION
The increasing prevalence of multidrug-resistant Enterobacterales in UTI has complicated therapy and particularly, CPE constitutes a major concern. They leave very few options for antimicrobial therapy and expedite expansion of acquired carbapenem resistance between bacteria by mobile genetic elements. Data on new or rediscovered antimicrobials which could be effective in a particular type of carbapenemase is crucial for future disease management.
To date, usage of mecillinam is established only in a few European countries (Pinart et al., 2017). In 2020, breakpoints have been updated by EUCAST to include more species, now E. coli, Klebsiella spp., Citrobacter, spp., Enterobacter spp., Raoultella spp., and P. mirabilis. Prescription of mecillinam may also rise in the future due to increasing antimicrobial resistance, introduction of mecillinam in countries outside of Europe, or the evaluation of mecillinam for other indications such as complicated UTI or bloodstream infections (Theuretzbacher et al., 2015;Jansaker et al., 2018;Fuchs et al., 2019;Boel et al., 2021). In addition, the European Medicines Agency (EMA) warning issued on the use of fluoroquinolones further limits oral therapeutic options in UTI especially in multidrug-resistant (MDR) pathogens (EMA, 2019).
In the present study, we demonstrate susceptibility of mecillinam in OXA-48-like-, IMI-, and NDM-1-producing Enterobacterales with both high and low carbapenem MICs. Compared to the results of the other groups on CRE and mecillinam, our study demonstrates that not only NDM-1 and OXA-48-like isolates show susceptibility to mecillinam but also isolates producing IMI. Susceptibility of mecillinam in IMI-producers was present despite high MICs of carbapenems (median carbapenem MIC >32 mg/L in IMI-producers), highlighting that even if isolates express their carbapenemase enzymes to a high amount, they hydrolyze mecillinam to a lesser extent compared to other beta-lactams.
Since this is to the best of our knowledge the first study demonstrating mecillinam susceptibility in carbapenem resistant IMI-producers, we performed an additional agar dilution protocol with a 10 times higher inoculum in these isolates (data not shown) to minimize a confounding effect of the inoculum, which has been described previously (Thomas et al., 2006;Marrs et al., 2014).
In line with other studies, poor mecillinam activity was documented in KPC-, IMP, and VIM-producers, irrespective of species (Marrs et al., 2014). Data on P. mirabilis as multidrugresistant pathogen are limited, our study included three isolates [OXA-58 (N = 2) and NDM-1 (N = 1)], all of which had a mecillinam MIC of ≥256 mg/L.
Our study has some limitations. Isolates in this study were not only from urine but from different clinical specimens. Species other than E. coli, which causes most uUTI, were over-represented. Nevertheless, all CPE in the challenge collection have been reported as the cause of UTI, and we included a large variety of molecularly characterized CPE. However, for some rare carbapenemases only few isolates could be assessed (e.g., GES), and generalizability is limited for these variants. Furthermore, some isolates in our collection had low carbapenem MICs, indicating a low expression of the detected carbapenemase enzymes. Besides the level of gene expression, clinical isolates carrying carbapenemases can additionally have other mechanisms that might impact the susceptibility against mecillinam such as previous antimicrobial therapy or other beta-lactamases or porin modifications, which were not the focus of our study.
In general, in vivo data on the potential of mecillinam to achieve microbiological cure are still scarce in other species than E. coli, and treatment failure despite mecillinam susceptibility has been documented in UTI (Søraas et al., 2014). Especially for infections caused by CPE data on clinical outcome after mecillinam therapy is not available, and low MICs do not necessarily correlate with clinical success. Nevertheless, mecillinam might be considered for therapy in uUTI caused by E. coli and E. cloacae isolates carrying IMI-, OXA-48, and OXA-48-like-carbapenemases. Future studies are warranted on in vivo activity of mecillinam against infections caused by CPE.
To conclude, with the increasing prevalence of MDR pathogens in inpatient and outpatient UTI, oral treatment has become more and more difficult. In this large cohort of CPE, we showed that mecillinam susceptibility is restricted to isolates producing the carbapenemases IMI, OXA-48-like, and NDM-1 and is more frequent in E. coli and E. cloacae than in other species. Even in isolates with high carbapenem MICs, mecillinam can still be active. Further research is needed on appropriate susceptibility testing in clinical routine and the correlation of mecillinam therapy and clinical outcome in CPE infections.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
AH: conceptualization, supervision, validation, resources, and writing-review and editing. FF: project administration, FIGURE 1 | Susceptibility testing of mecillinam by agar diffusion methods in an isolate of K. pneumoniae producing NDM. Agar gradient diffusion and disk diffusion of the isolate with an inhibition zone diameter of 12 mm (resistant) and an MIC of 8 mg/L (susceptible) by gradient test, compared to 64 mg/L (resistant) by the reference method.