Introduction
Gram-negative bacteria resistant to last-resource antibiotics, for instance, several species in the Enterobacterales order, Acinetobacter baumannii or Pseudomonas aeruginosa, are listed by WHO Bacterial Priority Pathogens List 2024 as critical and high-priority pathogens due to the ability to transfer resistance genes and the severity of the infections they cause (WHO 2024).
Carbapenemases are a type of β-lactamases that confer resistance to virtually all β-lactams. They are classified into two groups according to the residues in the active site. Serin-β-lactamases (SBL) contain a serine residue that forms hydrogen bonds with β-lactams releasing inactive antibiotics and active β-lactamases.
Metallo-β-lactamases (MBL), with Zn++ ions at their active site, produce nucleophiles that attack the β-lactam ring, this resulting in the hydrolysis of β-lactam antibiotics (Zhang et al. 2024). MBL belong to Ambler class B β-lactamases, that consider both molecular structure and hydrolytic spectrum (Ambler 1980).
They can be divided in subclasses B1, B2, and B3 based on aminoacidic sequence and number of Zn ions. Subclass B1 enzymes are most prevalent and include IMP (Imipenemase), NDM (New Delhi MBL type), and VIM (Verona Integron-Encoded MBL) (Brem et al. 2016).
The dissemination of carbapenemases have been a source of concern since the second half of 1990s. In Europe, carbapenemases were reported mainly in P. aeruginosa but, after 2000s, a continuous increase of VIM and KPC (a type of SBL) produced by Klebsiella pneumoniae has been noticed (Hernández-García et al. 2020).
NDM and VIM enzymes are also highly prevalent in India and Middle East regions as well as in other parts of the world (ECDC 2025).
An effective strategy to overcome the resistance to β-lactams is combining them with certain molecules that inhibit the enzyme activity. Several SBL inhibitors have been approved for clinical use, such as clavulanic acid, avibactam, relebactam or tazobactam, combined with amino-penicillins, cephalosporins or carbapenems (Zhang et al. 2024).
However, to date, there is not any specific MBL inhibitor available, although a high number of molecules with different action mechanisms are under investigation (De Falco et al. 2025).
Interestingly, a few of the studied compounds are approved drugs that are currently used for treatment of different illnesses, such as EDTA or dimercaprol, employed to treat metal poisoning, the antiacid bismuth subcitrate for treatment of gastric ulcer, or captopril, used as antihypertensive (Palacios et al. 2020). The demonstration of utility of such medicines in infectious diseases as MBL inhibitors would be an example of drug repurposing, or drug repositioning, the latter being defined as the process demonstrating new clinical uses for known approved molecules There are very few successful examples of repurposed drugs for microbial infections, for example, the antifungal amphotericin B or the antibiotic paromomycin for treatment of leishmaniasis (Farha et al.
2019). But, as far as is known to the authors, no repurposed drug is used as an antibacterial medicine. In contrast, some antimicrobial compounds have been repurposed with non-infectious uses, e.g. the antiviral amantadine for Parkingson treatment, the antimalarial artemisin as an antitumoral drug (Pinzi et al.
2024), or eflornithine, an antitrypanosomal drug, repositioned for reduction of facial hair (Xue et al. 2018).
As MBL are Zn-containing metalloenzymes, molecules with thiol groups with capacity for binding or chelating Zn had been proposed as good candidates as MBL inhibitors (Hammond et al. 1999). Examples of assayed thiol-compounds are captopril and its fluorinated derivatives, dimercaprol, L-phenylalanine, mercapto-3-phenylpropionic acid, thioacetic acid, thiazol-carboxilic acid, thiomandelic acid, thiorphan or tiopronin (Klingler et al.
2015). Kinetic and dynamic properties of such compounds are known and the inhibition activity of some of them was described for several types of MBL (De Falco et al. 2025).
Among them, captopril has been widely studied as a MBL inhibitor and sometimes used as a model for newly synthesized derivatives and analogues (Muteeb et al. 2022) (Mouton and Vinks) or as a control for related compounds (Abdulaal et al. 2024).
Despite the well demonstrated inhibitory mechanism of captopril and the improvement of its chemical structure (Li et al. 2014, Wade et al. 2021), in vitro studies of the combination of captopril with carbapenems against bacterial strains are scarce.
Moreover, the number of clinical isolates in most publications is quite low, sometimes only recombinant strain is used, and the experimental conditions or antibiotic assayed are heterogeneous (Brem et al. 2016, Tehrani and Martin 2017, Wade et al. 2021).
Evidence of activity of dimercaprol is even more limited, although it has been included as a candidate in several reviews (Doquier et al. 2018, De Falco et al. 2025).
Following an initiative of the European Medicines Agency (EMA), which prompts studies on drug repurposing (EMA 2019), the objective of this work is to study the activity of two thiol compounds, L-captopril and dimercaprol, on an elevated number of carbapenemase-expressing clinical isolates of several species of Enterobacterales and Pseudomonas, in combination with the three carbapenems more frequently used in treatment, ertapenem, imipenem, and meropenem. In addition to this, a broad literature search was undertaken to analyse together the results of previous studies, these being sometimes difficult to compare due to the diverse methodologies and approaches employed. Finally, specific pharmacological estimations are done to correlate the concentrations of inhibitors used in the laboratory with those employed for therapeutic uses.
Materials and methods
Bacterial strains and chemicals
Bacterial strains were isolated at the Hospital Clínico San Carlos (HCSC), Madrid, Spain. They were used anonymously and numbered randomly for this study. Isolates were identified using MALDI-TOFF-MS (Bruker Daltonik Maldi Biotyper) in accordance with the manufacturer’s instructions.
Enterococcus faecalis ATCC 29212, Escherichia coli ATCC 25922, and P. aeruginosa ATCC 27853 were included as reference strains as recommended by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) (EUCAST 2025).
Carbapenemase production was tested by an immunochromatographic (IC) assay that detects most prevalent β-lactamase enzymes: VIM, NDM, IMP (MBL) and KPC, OXA-48 (SBL) (RESIS-5 OKNVI Cores Bioconcept, Suiss). Genes encoding carbapenemases (blaOXA, blaKPC, blaNDM, blaVIM) were detected by multiplex-PCR as described previously (Poirel et al. 2011).
Müeller–Hinton Broth and Müeller–Hinton Agar (Scharlau, Spain) were used for strains cultivation. Antimicrobials and inhibitors were provided by the hospital pharmacy at HCSC: ertapenem (ETP) (Invanz, MSD, Spain), imipenem (IPM) (Aurovitas, Spain), meropenem (MEM) (Ranbaxy, India), L-captopril (Merck, MSD, Spain), and dimercaprol (Samarth, India). Antimicrobial susceptibility discs were provided by Beckton Dickinson (Spain).
Susceptibility testing and evaluation of L-captopril and dimercaprol as MBL inhibitors
Diffusion, microdilution and kill-curves methodologies were performed according to Clinical and Laboratory Standards Institute (CLSI) (CLSI 2025) and all the assays were performed in duplicate. EUCAST breakpoints were used for interpretation of the susceptibility results (EUCAST 2025).
In order to exclude the growth inhibiting activity of L-captopril and dimercaprol, blank discs were prepared with a range of concentrations of the compounds (100–1000 mg/l) and tested on M–H agar plates by diffusion methods on selected strains. For determining an adequate concentration of each compound for further studies, 10 µl of adequate solutions of the mentioned concentration range were added onto 10 µg antimicrobial discs of ertapenem, imipenem and meropenem. The bacterial inoculum was equivalent to no.
0.5 McFarland (c.a. 1.5 × 108 CFU/ml) and the agar plates were incubated at 37°C for 18–20 h.
Minimum Inhibitory Concentration (MIC) of the strains was determined by the microdilution method. Briefly, serial antibiotic dilutions were prepared in 96-U-well microtiter plates and the effect of L-captopril and dimercaprol was evaluated by adding the compounds, at the adequate concentration determined previously, to the concentration range of each carbapenem. Additionally, low concentrations of L-captopril (4 and 8 mg/l) were also tested.
In the assays, a growth control without antibiotic nor inhibitor, and a susceptibility control without inhibitor were included. Microplates were inoculated with 105 CFU/ml of each bacterial strain to obtain a final volume of 100 μl and incubated at 37°C for 18–20 h. MIC was determined as the lowest concentration at which visible growth was inhibited.
For kill-curves experiments, bottles containing 50 ml of freshly prepared Mueller–Hinton broth, supplemented with the carbapenem (0.5 × MIC) alone and in combination with L-captopril, were inoculated 105 CFU/ml of the strain and incubated at 37°C for 24 h. Samples were taken at 0, 2, 3, 6, and 24 h, and after adequate dilutions, five drops of 50 μl each were placed on Mueller–Hinton agar plates. Once the drops were dry, plates were incubated at 37°C for 20 h.
The time-kill curves were constructed by plotting mean colony counts (log10 CFU/ml) vs. time. Bactericidal activity was interpreted as a ≥ 3 log10 decrease in CFU/ml when compared with the values of a 0.5 × MIC curve.
Systematized literature review
A wide literature review was performed with a systematic scheme to acquire comparable experimental data on the activity of L-captopril and dimercaprol in combination with carbapenems on clinical strains. Recommended and common searching routes were followed (Both et al. 2016) and electronic databases WOS, Scopus and PubMed were used.
A preliminary scoping search considered reviews on “carbapenems”, “carbapenemases”, “β -lactamase inhibitors”, and “carbapenemase inhibitors”. For primary studies, Boolean combinations (AND, OR) were applied to terms “captopril”, “dimercaprol”, “lactamases”, “methallo-β-lactamases”, “methallo-β-lactamase inhibitors” “MBL”, and “MBL inhibitors”. Searching filters such as data of publication (2010–2025) and exclusion of reviews were used.
Duplicated publications were eliminated manually and references included in selected publications were also considered.
Pharmacological calculations
Pharmacological (pharmacodynamics and pharmacokinetic, PD/PK) parameters were used as defined in standard pharmacology manuals: Sd (Standard dose of the compound in clinical use) and Md (Maximum dose of the compound recommended for clinical use); Vd (Volume of distribution, proportionality constant that relates the total amount of drug in the body to the plasma concentration of the drug at a given time); Cp (Plasma concentration, concentration of drug in plasma); Pb/Tb (Plasma protein and Tissue binding, proportion of drug that binds to plasmatic proteins or tissues); Cf (Free concentration, concentration of drug in plasma not bound to proteins or tissues); D (Dose administered of drug). The main formula used for calculations were:
Data for L-captopril, imipenem, (used as example of carbapenems to be combined with L-captopril) and relebactam (as an example of approved BLM inhibitor in combination with imipenem) were obtained from European Medicines Agency (EMA), Food and Drugs Administration (FDA), or scientific publications. Plasma concentrations and doses were calculated according to formula (1) and (2) and are shown in Table 1 and in Supplementary material.
Two approaches were used to evaluate the possibility of translation of L-captopril in vitro active concentrations to clinical doses: a) Evaluation of the concentration of the inhibitor in plasma, and b) Correlation of concentrations ratio (R) of inhibitor and carbapenem in vitro and in plasma, according to the formula:
Results and discussion
Since 2021, the European Lab Network for Resistant Micro-organisms’ Surveillance (EARS-Net) analyses the data of the cabapenemase producing bacteria isolated by associated hospitals, such as Hospital Clínico San Carlos, Madrid, Spain. For this work, 49 Gram-negative clinical isolates were selected and identified as Citrobacter freundii (n = 2), Enterobacter cloacae complex (n = 6), E. coli (n = 1), K. pneumoniae (n = 9), and Serratia marcescens (n = 5) within the Enterobacterales order, together with 24 strains of P. aeruginosa and 2 P. putida.
Although all the Enterobacterales strains tested produced a MBL enzyme, the susceptibility to carbapenems was heterogenous, with wide ranges of MIC values of 0.5–128 mg/l (Table 2). Nevertheless, as a group when MIC90 values were calculated, they showed high resistance to ertapenem and imipenem (32 mg/l and 64 mg/l, respectively) and moderate resistance to meropenem (16 mg/l) (EUCAST breakpoints: ertapenem, R > 0.5 mg/l, imipenem R > 4 mg/l, meropenem R > 8 mg/l). Individual MIC values of enterobacterial isolates are shown in Table 3.
All the strains (100%) were resistant to ertapenem and 65% to imipenem, while susceptibility to meropenem was heterogeneous (R, 30.4%; I, 34.8%; S, 34.8%) (Fig. 2A). The different degrees of resistance could be based on the level of expression of carbapenemases and the different affinity and activity against carbapenems, as well as the contribution of additional resistance mechanisms (Queeman and Bush 2007).
Pseudomonas group had very high MIC50 and MIC90 values (128 mg/l and 256–512 mg/l, respectively) for the three carbapenems (Table 2) when compared with EUCAST breakpoints (IPM R>4 mg/l; MEM R> 8 mg/l). Pseudomonas aeruginosa shows intrinsic resistance to β-lactams as the production of hydrolytic enzymes is combined with other resistance mechanisms, mainly low permeability of its outer membrane and active efflux systems (Livermore 2002).
A consistent increase in carbapenem resistant Gram-negative strains has been observed in many countries since 2019, although the distribution is not homogeneous in the different areas. In Europe, K. pneumoniae incidence is higher than E. coli and resistance in Pseudomonas is heterogeneous (ECDC 2025). In Spain K. pneumoniae resistant to carbapenems represents 77% of the isolates, followed by Enterobacter cloacae complex (16.7%) and E. coli (11.2%), and, regarding carbapenemase type, OXA-48 (SBL) is detected more frequently (67%), followed by VIM (47%), NDM (16%), and KPC (8%) (Redlab 2024 et al.
2024). All the strains of this study produced a VIM type carbapenemase, as assessed by immunochromatographic and genetic methods. Multiple carbapenemase IC tests are recommended among the phenotypic tests for carbapenemase screening in enterobacteria as they are rapid, easy to use and have comparable sensitivity and specificity to genetic methods (Tamma and Simmer 2018).
Pseudomonas aeruginosa is the most important known reservoir of VIM enzymes, but it has been reported also in Enterobacterales, mainly Enterobacter spp., E. coli, and K. pneumoniae, and also sporadically in Acinetobacter spp (Alvisi et al. 2025). Although NDM has become the most globally disseminated metallo-β-lactamase (European Committee on Antimicrobial Susceptibility Testing) (Li et al.
2014), VIM remains highly relevant in Southern Europe, particularly in Spain, where it continues to be frequently detected in P. aeruginosa and Enterobacterales from hospital settings. Its persistence is associated with endemic dissemination and repeated nosocomial outbreaks (ECDC 2025, Olivieri et al. 2025).
For this reason, VIM-producing clinical isolates were selected as a representative and clinically meaningful model to evaluate the inhibitory potential of L-captopril and dimercaprol against MBL-mediated carbapenem resistance.
Summary of publications on activity of L-captopril and dimercaprol as MBL inhibitors on Gram-negative bacterial strains
Due to the good inhibition properties of D-captopril stereoisomer and other thiols on different MBL, a considerably large number of structural, biochemical and kinetic studies have been conducted with captopril derivatives and analogues (Brem et al. 2016, Ma et al. 2021, Kondratieva et al.
2024). Additionally, different strategies for designing and synthesizing new molecules with better affinity have been carried out (Ezati et al. 2023).
However, the possibility of using the stereoisomer L-captopril (and not other related molecules) is somehow more attractive, as it is an approved drug in clinical use and its pharmacological properties and toxicity are already known. Consequently, it would be desirable to have robust experimental evidence of in vitro, or even in vivo, activity of the compound on clinical isolates. With this objective, we performed a wide and systematized literature search with the following results: reviews on β -lactamase inhibitors, 27; primary publications (not duplicated) on captopril/dimercaprol AND lactamases, 70, captopril/dimercaprol AND MBL inhibitors, 29.
From among them, only 10 studies used bacterial strains (Table 4). These publications were analyzed in detail for type of strains and MBL produced, captopril, dimercaprol, and antibiotics intervention details, methodology used, MIC values, and carbapenems’ susceptibility results. These data will be further compared with the obtained in the present work and a comparative summary of the previous publications is provided as Supplementary material.
L-captopril as an efficient MBL inhibitor in Enterobacterales strains, but not within Pseudomonas
L-captopril did not affect growth of clinical isolates or reference strains when tested at different concentrations (100–1000 mg/l). Similarly, in the literature, no growth inhibition was described at the tested concentrations (Table 4).
The activity of D-captopril stereoisomer was first reported at low concentrations (8 mg/l), while L-captopril showed no or little effect at this concentration (Hammond et al. 1999, Brem et al. 2016).
In this work, similar results were obtained in two selected strains (1 K. pneumoniae, 1 E. coli), for which 4 or 8 mg/l of L-captopril did not influence the activity of meropenem and ertapenem. An optimal concentration of 400 mg/l was selected as the minimum amount at which maximum inhibition was observed. In previous publications, different concentrations of L-captopril were used as adequate for MBL inhibition assays, ranging from 108 mg/l to 1000 mg/l.
The effect of L-captopril combined with carbapenems is shown as MIC50 and MIC90 values in Table 2. Within the Enterobacterales group, a remarkable decrease in MIC50 can be observed (from 4 to 0.125 mg/l for ertapenem and from 8 to 0.5 mg/l for imipenem and meropenem). The inhibitor causes a decrease of more than four-fold MIC (Table 3, marked with bold letters) in a high proportion of the isolates, 78.2% when it is combined with ertapenem and imipenem, and 65% with meropenem.
MIC shifts are represented in Fig. 1 and reveals more frequent decreases of 4–6 MIC-fold for the three carbapenems. Similar results of MIC diminution have been described previously (≥ four-fold MIC reduction in 66%–100% of strains) for imipenem and meropenem, while lower proportions (0%–36%) were reported by other authors MIC decreases (Table 4).
Moreover, the activity of carbapenems could be restored in most cases when the antibiotics were combined with L-captopril (Fig. 2A). A high percentage of resistant (R) or dose-dependent (I, dose dependent/susceptible at increased doses) isolates showed a susceptible phenotype (S) when the MBL inhibitor was present: ETP, 73.9%; IPM, 87%; MEM 82.6%.
In addition, kill-curves of a representative isolate of K. pneumoniae assessed the recuperation of activity of carbapenems when combined with L-captopril (Fig. 3). After 24 h of incubation, the growth of the strain was similar in the control curve (without antibiotic) and with a subinhibitory concentration of imipenem (4 mg/l: 0.5 x MIC), reaching c.a.
108 CFU/ml. Nevertheless, when this concentration (0.5 × MIC = one-fold MIC diminution) was combined with 400 mg/l of the inhibitor, a continuous diminution of number of bacterial cells was found, ≥ 103 CFU/ml, which indicates the complete restoration of imipenem activity, and the bactericidal capacity of the antibiotic recovered.
In previous studies heterogeneous results of activity of L-captopril were reported (Table 4). Despite very interesting physicochemical and biochemical properties of captopril were described on purified IMP, NDM and VIM enzymes, the activity of imipenem and meropenem (ertapenem was not tested) was not fully restored when tested on bacterial strains (Klinger et al. 2015, Brem et al.
2016, Tehrani and Martin 2017, Büttner et al. 2018, Wan et al. 2019, Rudresh et al.
2023, Kondratieva et al. 2024), possibly due to the low concentration of captopril tested in some studies. Our results agree with some of the previous studies, as the combination L-captopril (400 mg/l) with imipenem or meropenem, resensitized >80% of the clinical isolates.
Completely new data are provided in this work for ertapenem, as this antibiotic was not assayed previously, and resensitization of a high proportion of VIM-producing enterobacteria is demonstrated (Fig. 2A).
Regarding the type of MBL, better activity of L-captopril was found against certain type of enzymes in some publications (Table 4). The isolates assayed in the present study produced only VIM enzymes and the activity of L-captopril on this type correlates with that described by two authors (Brem et al. 2016, Bütner et al.
2018). However, the data are not homogeneous, and it is difficult to reach a clear conclusion. It has been described that the thiol moiety could not be the only parameter for MBL inhibition (Klingler et al.
2015, Tehrani and Martin 2017) and that the general mechanism of thiols, a strong binding to Zn, is not always required for maximum activity of each type of enzyme (Wade et al. 2021). Additionally, the high heterogeneity of active sites among different types of MBL and allocation of Zn in a shallow pocket makes the design of a single potent inhibitor difficult (Kang et al.
2024).
The study of the potential activity of L-captopril as an MBL inhibitor is challenging for a drug repurposing aim, as has been mentioned by several authors. This was the objective of testing 11 approved thiols on 6 IMP and NDM-producing Gram-negative bacteria, and it was found that L-captopril showed a moderate MBL inhibitory activity at 128 mg/l, compared with thiorphan (used for diarrhoea treatment) or tiopronin (approved in the USA for preventing kidney stone formation) (Klingler et al. 2015).
The therapeutic utility of the compound combined with meropenem was also evaluated in vivo in animal models as the nematode Galleria mellonella. The larvae were experimentally infected with one IMP-expressing K. pneumoniae strain and, after treatment with meropenem + captopril, the survival rate was higher than when using meropenem only. Also, the bacterial load was lower than in the control group without an inhibitor (Zhao et al.
2021).
On the other hand, the effect of L-captopril on carbapenems susceptibility of Pseudonomas group was minimum. As shown in Table 2, MIC50 and MIC90 only decreased 1–3 folds, which is not considered of clinical significance (Mouton et al. 2018).
Only 4% of the imipenem and 31% of meropenem resistant strains partially resensitized at increased exposure (I) (Fig. 2B). Due to the intrinsic resistance of P. aeruginosa, ertapenem is not commonly used for treatment, and breakpoints for ertapenem are not defined by EUCAST (2025).
The calculated data included in Fig. 2B for this antibiotic refer to the proportion of the strains that were not inhibited at ≥16 mg/l (Livermore et al. 2005).
Similarly, captopril did not show any effect on six strains of P. aeruginosa reported previously in three studies (Brem et al. 2016, Büttner et al. 2018, Tehrani and Martin 2017).
The observed limited reduction in carbapenem MICs in Pseudomonas spp. compared to Enterobacterales despite the presence of L-captopril may be explained by the multifactorial nature of carbapenem resistance in this organism. Although all isolates in the study harboured VIM-type metallo-β-lactamases, additional resistance mechanisms likely play a major role in P. aeruginosa. The intrinsic low permeability of the outer membrane, often associated with downregulation or loss of the OprD porin, significantly limits carbapenem uptake.
Furthermore, the overexpression of efflux pumps such as MexAB-OprM and MexXY can actively extrude β-lactam antibiotics, reducing their intracellular accumulation and thereby diminishing the impact of enzyme inhibition alone. In addition, the inducible AmpC β-lactamase may contribute to residual resistance, and its activity is not affected by L-captopril (Livermore 2002). Overall, these mechanisms likely explain why inhibition of VIM enzymes alone results in only a modest decrease in MIC values in Pseudomonas, in contrast to the more pronounced effect observed in Enterobacterales, where carbapenem resistance appears to be more directly dependent on metallo-β-lactamase activity.
Dimercaprol does not have relevant activity as an inhibitor of MBL on Gram-negative bacteria
Dimercaprol has been proposed in different reviews as a suitable candidate among the thiol-containing compounds based on its capacity for binding several enzymes such as NDM (Quian et al. 2019, Docquier et al. 2018, Palacios et al.
2020, de Falco et al. 2025). Nevertheless, only two publications reported experimental data of its activity on bacterial strains with poor results (Table 4).
The compound was tested at 128 mg/l in combination with imipenem on five enterobacteria and one P. aeruginosa strains and showed a moderate activity (50% of the strains were resensitized) (Klingler et al. 2015), but no activity was found at 124 mg/l and meropenem on 1 E. coli strain (Wade et al. 2021).
In this work, dimercaprol showed also a moderate effect only on ertapenem activity, as 52.17% of the resistant enterobacterial strains reversed susceptible. Nevertheless, the compound had little effect when it was combined with imipenem or meropenem (<6% and c.a. 17% of resistant isolates became susceptible, respectively) (Fig.
2A). Within the Pseudomonas group, no activity was encountered in combination with imipenem and a low effect on meropenem activity, as only 23% of the resistant isolates turned susceptible at increased doses (Fig. 2B).
Expperimental data of activity of dimercaprol on the individual isolates can be seen in Supplementary material.
Dimercaprol is a very simple molecule clinically used for heavy metal poisoning processes, by chelating metal ions and, similarly, its activity as MBL inhibitor is based on a simple Zn stripping mechanism. As shown, the potency of dimercaprol is very reduced, probably due to the lack of a carboxylate group, present in different thiol compounds, that contributes to an adequate interaction and binding to the active site of the enzyme (Rosi et al. 2021).
In summary, the combined results of our work and of all previous ones, show that: (i) L-captopril at concentration ≥ 260 mg/l (medium value over a range of 64–1000 mg/l) inhibits different types of MBL and restores the activity of carbapenems ETP, IPM or MEM against > 80% of Enterobacterales strains (over 82 tested strains); (ii) the combination of 200 mg/l of L-captopril (average concentration) with carbapenems has no clinical utility on Pseudomonas spp, as the activity of the antibiotics was restored in <3% of the strains, over 32 isolates tested; and (iii) the combination of dimercaprol and carbapenems is not satisfactorily effective against carbapenemase-producing Gram-negative bacteria of Enterobacterales and Pseudomonas groups.
Analysis of PD/PK parameters and estimated doses of L-captopril did not confirm the possibility of repositioning it as MBL inhibitor
Major advantages of drug repositioning strategies are complete knowledge of pharmacokinetic and pharmacodynamic characteristics and the fact that compounds have passed all clinical tests in Phase I, II, and III, so their safety has been confirmed. Consequently, the risk of failure, and the required time and investment are significatively reduced (Xue et al. 2018).
However, active in vitro concentrations of the compound for new usage might not be similar to those confirmed as being non-toxic previously. Therefore, at least an estimation of plasmatic concentrations and dosage corresponding to assayed ones in the laboratory should be performed for comparison with standard/maximum doses in clinical use.
For treatment of infectious diseases, MIC values are used as reference to adjust therapeutical doses of antibiotics, although pharmacodynamic and pharmacokinetic parameters, depending on the biodisponibility and the bactericidal or bacteriostatic activity of the molecule, should be considered, because measurement of single MIC data are not sufficiently accurate (Mouton et al. 2018). Maximum killing activity of β-lactams occurs at concentrations ≥4 times the MIC, and this concentration over-MIC should be maintained at least 40%–50% of the dosing interval (Craig 2014).
When a β-lactamase inhibitor is combined, a sufficient concentration should be supplied to render bacterial β-lactamases inactive (Dudley 1995). The approved combination imipenem-relebactam was used as reference for the calculations: (a) 4 × 2 mg/l (imipenem MIC breakpoint) = 8 mg/l is close to plasma concentration of free compound (Cf 16.5 mg/l) for standard dose (500 mg); (b) relebactam Cf, 10.2 mg/l at standard dose (250 mg) is 2.5 times its optimal concentration assessed in vitro, 4 mg/l (Lob et al. 2017).
With the same approach, calculated plasma concentrations of L-captopril were (a) 1600 mg/l (4 × 400 mg/l—optimum in vitro concentration) or (b) 1000 mg/l (2.5 × 400 mg/l), which are in both cases superior to the concentration of free drug (Cf, 0.8 mg/l) at standard dose (50 mg). These calculated values are also higher than the concentration at the maximum recommended dose (150 mg, Cf 2.3 mg/l) and the amount measured in plasma for a case of death due to L-captopril overdose (60.4 mg/l) (Park et al. 1990).
At present, dosage adjustment of β-lactam/β-lactamase inhibitor combinations is commonly based on considering similar PD/PK characteristics of both compounds, and, although these parameters are not defined for all the inhibitors in clinical use, the objective should be to ensure adequate concentrations of the two agents at the site of infection (Abodakpi 2019 et al. 2019). Optimal in vivo activity of a certain drug combination might be achieved if their delivery to target cells occurs at defined ratios as well as optimal concentrations (Harasym et al.
2004). Based on this, the ratio of in vitro concentrations of L-captopril and imipenem was R = 400/2 = 200, that results, after using the formula (1) and (2), in a theoretical dose of the inhibitor D = 202.1 g, far beyond the therapeutic concentration in clinical use as antihypertensive. Two authors had suggested that the assayed concentrations of L-captopril that restored the activity of carbapenems in vitro could not be reached after usual dosage (Klingler et al.
2015) or were superior to the therapeutic safety levels (Rudresh et al. 2023). Nevertheless, they did not provide any calculations or data to demonstrate these points.
In this work, calculations were made for those published concentrations (64, 128, and 256 mg/l) and, effectively, dosages values were higher than standard or maximum recommended doses (32, 64, and 128 g, respectively). (Calculations and complete data are available in Supplementary material). It has been suggested that the lack or reduced activity of L-captopril and other thiol compounds, at low or moderate concentrations, is due to the high lipophilicity, which limits the transport through the Gram-negative outermembrane (Rotter et al.
2023). This would explain the high concentrations of the compounds needed to assess accumulation in the periplasmic space were MBL are located.
Given the results, it should be concluded that L-captopril cannot be used as MBL inhibitor in combination with carbapenems for treatment of infectious diseases, because the active concentrations in vitro could not be extrapolated to non-toxic doses in humans, despite the very promising experimental results obtained in this work and other previous studies. Considering the encouraging objective of drug repurposing of our and other works (Klingler et al. 2015, Dutta and Mishra 2018, Abdulaal et al.
2024), our conclusion could be certainly disappointing. Notwithstanding, negative results are not without value since research is often based on previously published work, it can be very valuable to know which experiments have not worked in addition to those that have worked. In doing so, duplication of efforts, biased meta-analyses, and ethical concerns regarding animal and human studies could be dismissed (Bik 2024).
As a counterbalance, L-captopril is under research as a new antimicrobial agent, due to its capacity of enzymatic inhibition. Two examples are the inhibitory effect on DapE (an amidohydrolase within the lysine biosynthetic pathway of bacteria) assayed on Haemophilus influenzae and Neisseria meningitidis (Starus et al. 2016) and inhibition of LdDPD (a dipeptidyl carboxypeptidase expressed in apicomplexan parasite stage differentiation) studied in Leishmania and Trypanosoma (Garsi et al.
2024). The results of this last publication are promising, and the active concentrations are acceptably low. Also, modulation in the severity of COVID-19 disease by L-captopril is investigated, as SARS-CoV-2 endocytosis depends on viral spike protein binding to ACE2, the same enzyme inactivated during hypertensive treatment (Santos et al.
2020). Finally, it has been described that the vasculo-protective effect induced by L-captopril, and other first line antihypertensive drugs, involves gut microbiota reshaping and gut-immune system communication that could ameliorate the dysbiosis state in hypertension (González-Correa et al. 2024).
Some strategies are currently proposed to overcome the problem of low access of thiol-derivatives to MBL location, such as chemical modification of the molecules to adequate their physicochemical properties and augment the hydrophobicity and electrostatic interactions (Rosi et al. 2021), development of nanoparticle-based systems for better delivery at the action site or conjugating them with molecules similar to natural substrates, such as siderophores (de Falco et al. 2025).
More recently, it was demonstrated that some of the SBL inhibitors show a partial activity against some MBL enzymes. This is the case of boronate derivatives that form complexes at the initial phase of interaction which is similar for SBL and MBL (Cahill et al. 2017), and the combinations taniborbactam-cefepime and xeruborbactam-meropenem are in phase III and I of clinical trials, respectively (Katsarou et al.
2025). Nevertheless, despite the extensive research developed, a main difficulty in designing effective MBL inhibitors, and also more specific SBL inhibitors, is the lack of uniformity in the laboratory approaches, i.e. traditional or computational methods for compound’s screening, biochemical or structural analysis, in vitro/in vivo assays of activity with clinical isolates and animal models or toxicity and PD/PK characterization (Reddy et al. 2025).
Consequently, many challenges arise in comparing the results across the various research groups to complete a robust pre-clinical evaluation. Some proposals for standardization include unified MIC breakpoint guideline, validated strains expressing different MBL types as reference, or availability of PD/PK profiles of β-lactamase inhibitors. Aligning research methodologies could help develop effective therapeutics to combat carbapenem-resistant bacteria, which, as mentioned, are of critical global health concern.
Supplementary Material