Pharmaceutical compositions for prevention and / or treatment of infections and antibacterial-induced dysfunctions

Novel pharmaceutical compositions leveraging synergistic drug pairs effectively target multidrug-resistant bacteria while preserving healthy intestinal flora, addressing antimicrobial resistance and its side effects.

JP2024160273APending Publication Date: 2024-11-13EURO LAB FUER MOLEKULARBIOLOGIE EMBL
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
JP2024125600
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2018-04-27
Filing Date
2024-08-01
Publication Date
2024-11-13

AI Technical Summary

Technical Problem

The development and spread of antimicrobial resistance pose a significant threat to public health, rendering known antimicrobial compounds ineffective against bacterial strains, and current antibiotic therapies often harm healthy intestinal flora, leading to bacterial overgrowth and resistance development.

Method used

Development of novel pharmaceutical compositions that systematically screen and evaluate drug-drug interactions in Gram-negative bacteria, utilizing synergistic pairs of compounds, including non-antibiotics like vanillin, to inhibit bacterial growth effectively while minimizing harm to commensal bacteria.

Benefits of technology

These compositions provide narrow-spectrum antimicrobial therapy effective against multidrug-resistant pathogens, reducing the selection for resistance and minimizing damage to the intestinal microbiome.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024160273000032
    Figure 2024160273000032
  • Figure 2024160273000033
    Figure 2024160273000033
  • Figure 2024160273000034
    Figure 2024160273000034
Patent Text Reader

Abstract

To provide pharmaceutical compositions for the prevention and / or treatment of bacterial infections and antibacterial-induced dysfunctions.SOLUTION: The compositions of the present invention demonstrate high species-specificity in inhibiting the growth of a small number of bacterial species, and are effective also against multi drug resistant (MDR) clinical isolate species. One of those combinations pairs a non-antibiotic drug, vanillin, with an antibiotic drug, spectinomycin, to demonstrate a strong inhibitory effect on the growth of clinically relevant Gram-negative pathogenic and multi-drug resistant E. coli isolates. A second set of compounds combines the polymyxin colistin with loperamide, rifamycin, or macrolide. This invention relates to combinations that enable narrow-spectrum antibacterial therapies, constituting a major effort of current and future drug development efforts in order to prevent major side effects of antibacterial strategies. This invention also relates to pharmaceutical combinations useful to prevent an adverse effect on the gut microbiome, induced by the use of antibacterial compounds.SELECTED DRAWING: None
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to the field of therapeutics, and more particularly to pharmaceutical compositions for the prevention and / or treatment of bacterial infections and antibiotic-induced dysfunctions. The compositions of the present invention show high species specificity in inhibiting the growth of a few bacterial species, and most importantly, are also effective against multidrug-resistant (MDR) clinical isolates. Interestingly, one of these combinations pairs the non-antibiotic vanillin with the antibiotic spectinomycin, which shows a surprisingly strong inhibitory effect against the growth of clinically relevant Gram-negative pathogens and multidrug-resistant Escherichia coli (E. coli) isolates. The second set of compounds combines the polymyxin colistin with loperamide, rifamycin or macrolides. Importantly, the present invention relates to combinations that allow narrow-spectrum antibiotic therapies, which are the main focus of current and future drug development efforts to prevent major side effects of antibacterial strategies. The present invention also relates to pharmaceutical combinations useful for preventing the adverse effects on the gut microbiome induced by the use of antibacterial compounds. [Background technology]

[0002] The spread of antimicrobial resistance has become a serious public health concern, making once treatable diseases deadly again and undermining the breakthroughs of modern medicine. The discovery of new antimicrobial therapies is essential, but new drug development takes years and unfortunately, antibiotic development has stalled over the past 30 years. As a result, only a handful of new antibiotic classes have entered the market since the 1990s, but none of these are active against the Gram-negative pathogens that currently pose the greatest threat to public health.

[0003] In general, antibiotics affect the risk of harming normal and healthy intestinal flora. Such inhibition can promote bacterial overgrowth and be the cause of the emergence of antibiotic resistance in microorganisms. By inhibiting the bacterial multiplication cycle, antibiotics rapidly select for resistant subpopulations. Thus, the proportion of nosocomial antibiotic-resistant opportunistic pathogens causing infections has more than doubled in the past decade, and antibiotic-resistant bacteria themselves can cause serious infections. As a result, an even more serious problem affects the possible transfer of resistance factors to other bacteria.

[0004] Drug combinations and repurposing of drugs can act as a first line of defense against the alarming rise in multidrug-resistant (MDR) bacterial infections, for example in clinically relevant Gram-negative pathogens such as Escherichia coli, Salmonella typhimurium, and Pseudomonas aeruginosa. Drug combinations exponentially increase the potential therapeutic solution space, and promising candidates can be rapidly translated into clinical applications when the individual compounds are already approved or in use.

[0005] Polypharmacy is common in many diseases, but has been little investigated in bacterial infections. Antibiotic combinations bring challenges as well as opportunities. One such challenge is that proper interaction assessment requires mixing several concentrations of each drug in a checkerboard format. This requires rigorous testing, making systematic interaction studies difficult. As a result, current knowledge is scarce and derived from many independent studies, each testing a small number of pairwise combinations with diverse assays and metrics.

[0006] To clarify general principles regarding antibacterial drug combinations, there is a need to systematically profile drug-drug interactions across different strains and species. Previous large-scale studies had to compromise either at the drug level or the species / strain level. With the primary goal of identifying antibiotic adjuvants, most studies profiled combinations of single antibiotics with 1000-2000 previously approved drugs or even larger chemical libraries.

[0007] Many bacterial species have been profiled in this way, including E. coli, Pseudomonas aeruginosa, and Staphylococcus aureus, but the screen design (examining synergy at single drug doses only) precludes comparative analysis. Smaller scale screens capable of quantifying both synergy and antagonism have been performed, but these have been limited to E. coli, testing pairwise combinations of 21 antibiotics at a single concentration. Larger scale screens exist for antifungal drugs, but these suffer from similar limitations; for example, the largest dose-dependent profiling of combinations was the profiling of approximately 200 drug pairs in Saccharomyces cerevisiae, whereas many more pairs have been tested in various fungal species in recent years, albeit at single drug doses. Summary of the Invention [Problem to be solved by the invention]

[0008] The emergence and spread of antibacterial resistance poses one of the most serious threats to public health, resulting in the development of bacterial strains against which known antibacterial compounds are ineffective.Therefore, there is a need to provide novel antibacterial pharmaceutical compositions that can overcome bacterial antibiotic resistance by being effective against selective bacterial strains.Such narrow-spectrum antibacterial therapy can overcome the main challenges of current antibacterial strategies, such as mitigating the spread of antibiotic resistance and side effects caused by the adverse effects of compounds on healthy microorganisms present in the patient's body.

[0009] It is therefore an object of the present invention to provide a novel antibacterial pharmaceutical composition. A further object of the present invention is to provide an antibacterial pharmaceutical composition that is effective against clinically relevant Gram-negative pathogens such as multidrug-resistant (MDR) bacterial strains, e.g., Escherichia coli, Salmonella typhimurium, and Pseudomonas aeruginosa. A further object of the present invention is the development of an antibacterial therapy that can prevent the development of antibiotic-resistant bacteria as well as the adverse effects of compounds on healthy microorganisms present in the patient's body, such as the intestinal flora. [Means for solving the problem]

[0010] The problem of the present invention is solved by providing a method for systematically screening and evaluating drug-drug interactions in various strains and species of clinically relevant Gram-negative bacteria. The results of the screening provide multiple synergistic pairs of compounds that exhibit strong antibacterial activity, or antagonistic pairs that can be used as antidote treatment strategies to prevent damage caused by antibiotic use, for example, in healthy microbiomes. Some of these compound combinations show great species specificity in inhibiting the growth of single species of bacteria, and most importantly, are also effective against MDR clinical isolates. Compared to broad spectrum antibacterial therapies, such selective pharmaceutical compositions are much less likely to select antimicrobial resistance and are less harmful to the patient's gut flora.

[0011] The inventors of the present invention have found that some of the pharmaceutical pairs identified by the methods described in this invention contain non-antibiotic compounds, such as food additives, similar to vanillin. Importantly, such combinations allow for narrow spectrum antibacterial therapy, which is a major challenge in current and future drug development to prevent the major side effects of antibacterial strategies. Other objects of the present invention will become apparent to those skilled in the art upon review of the detailed description of the present invention.

[0012] In a first aspect of the invention, the object of the invention is a method for identifying a synergistic antibacterial effect of at least two drug compounds against bacteria, comprising the steps of: a) providing the bacteria to be tested for antimicrobial effect; b) selecting said at least two drug compounds; At least one of the drug compounds is known to have an antibacterial effect against the bacteria of a), and at least one other of the drug compounds is (i) an antibiotic, or a pharma- ceutically acceptable salt thereof, known to belong to the same class as the selected first drug compound and / or to target the same bacterial cellular process as the selected first drug compound; and (ii) a human targeted drug, a food additive, or a pharma- ceutical acceptable salt thereof; Selected from; c) identifying a synergistic antibacterial effect of the at least two drug compounds on the bacteria; and d) selecting the at least two drug compounds identified in step c); and The problem is solved by providing a method, comprising:

[0013] As used herein, "synergy" or "synergistic effect" refers to the effect that occurs when at least two compounds interact to produce an overall effect that is greater than the sum of the individual effects of any of the aforementioned compounds used alone. Thus, the combination greatly enhances the antibacterial effect of one of the compounds used alone. In other words, synergistic effect means that the total antibacterial effect of the combination of two components against a bacterium or bacterial species is greater than the sum of the antibacterial effects of each component when measured separately.

[0014] As used herein, "antagonism" or "antagonistic effect" refers to the effect that occurs when at least two compounds interact to produce an overall effect that is less than the sum of the individual effects of any of them used alone. As used herein, an "antagonistic" combination may hinder the antibacterial effect of one of the compounds used alone. In other words, an antagonistic effect means that if at least one of the compounds has an antibacterial effect against a bacterium or bacterial species, this antibacterial effect can be countered by at least one other compound. Thus, the at least one other compound can be considered an "antidote" to the first compound described above. A combination of at least two compounds is said to have an "antagonistic effect" if the antibacterial effect of at least one of the two compounds is masked when used in combination with another compound.

[0015] "Synergistic" and "antagonistic" effects are scored by permutation p-values, as explained below in the "Methods" section of this specification.

[0016] In the context of the present invention, the term "human-targeted drug" refers to a compound intended for use in humans. Preferably, the mechanism of action (MOA) of said drug is known and may affect human cells intracellularly, extracellularly or within the human cell membrane. The use of said human-targeted drugs may suffer from the fact that they have side effects that harm human cells or organisms. Examples of such human-targeted drugs include, but are not limited to, antipsychotics, anesthetics, acid-reducing medications, chemotherapeutic drugs and blood-pressure medications. Conversely, the term "antibiotic" in the context of the present invention refers to a compound that is preferably microbiologically active, i.e., for use against pathogenic / undesirable microorganisms.

[0017] The object of the present invention is a method for identifying the antagonistic antibacterial effect of at least two pharmaceutical compounds on bacteria, comprising: a) providing a first bacterium to be tested for said antimicrobial effect; b) selecting said at least two drug compounds; wherein at least one of the drug compounds is known to have an antibacterial effect against a second bacterium and at least one other of the drug compounds is selected from an antibiotic, a human targeted drug, a food additive, or a pharma- ceutical acceptable salt thereof, preferably the at least one other of the drug compounds is known to belong to a different class than the first selected drug compound and / or to target a different bacterial cellular process than the first selected drug compound; c) identifying the antagonistic antibacterial effects of the at least two drug compounds on the first bacterium of a); and d) selecting the at least two drug compounds identified in step c); and The present invention is further resolved by providing a method, comprising:

[0018] Preferably, the first bacterium in a) above and the second bacterium in b) above belong to different bacterial genera and / or different bacterial species.

[0019] The bacteria are gram-positive or gram-negative bacteria, and examples of the bacteria include those belonging to the genus Enterobacter, Escherichia, Shigella, Serratia, Proteus, Pseudomonas, Acinetobacter, Staphylococcus, Streptococcus, Pseudomonas, Salmonella, Helicobacter, Helicobacter, Citrobacter, Treponema, Mycobacterium, Bordetella, Borrelia, Brucella, Corynebacteria, Fusobacterium, Leptospira, Listeria, Pasteurella, Rickettsia sia, Faecalibacteria, Eggerthella, Lactonifactor, Coliform, Bacillus, Franscicella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Bifidobacteria, Mobiluncus s), Enterococcus, Actinomyces, Neisseria, Chlamydia, Vibrio, Diplococcus, Lactobacillus, Kingella, Yersinia, Klebsiella, Bacteroides, Eubacterium, Alistipes,The genera Ruminococcus, Roseburia, Parabacteroides, Prevotella, Coprococcus, Dorea, Blautia, Odoribacter, Clostridium, Collinsella, and Bilophylla are all members of the genera. lophila, Akkermansia, Veillonella, Haemophilus, Desulfovibrio, Butyrivibrio, and / or Campylobacter, or any closely related genera thereof, and optionally said bacterium is an antibiotic-resistant bacteria, in particular a multi-drug resistant strain thereof.

[0020] More preferably, in the aforementioned method for identifying an antagonistic antibacterial effect of at least two pharmaceutical compounds against bacteria, said first bacterium to be tested for said antibacterial effect is a commensal and / or probiotic bacterium, such as Enterobacter, Staphylococcus, Streptococcus, Pseudomonas, Escherichia, Helicobacter, Neisseria, Campylobacter, Clostridium, Citrobacter, Vibrio, Treponema, Mycobacterium, Klebsiella, Mycobacterium ... Gella, Actinomycetes, Bacteroides, Bordetella, Brucella, Corynebacterium, Diplococcus, Fusobacterium, Leptospira, Pasteurella, Proteus, Rickettsia, Shigella, Parabacteroides, Odoribacter, Faecalibacterium, Collinsella, Egasella, Lactonifactory, Roseburia, Coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Kingella kingae, Haemophilus, Bifidobacterium, Mobiluncus, Prevotella, Akkermansia, Bylophila, Blautia, Coprococcus, Dorea, Eubacteria, Lactobacillus, Ruminococcus, Veillonella, and / or Enterococcus.

[0021] More preferably, the method as described above for identifying antagonistic antibacterial effects of at least two pharmaceutical compounds against bacteria, wherein said second bacterium is a pathogenic bacterium, such as Enterobacter, Escherichia, Shigella, Serratia, Proteus, Pseudomonas, Acinetobacter, Staphylococcus, Streptococcus, Pseudomonas, Salmonella, Helicobacter, Citrobacter, Treponema, Mycobacterium, Bordetella, Borrelia, Brucella, Corynebacterium, The bacteria is a member of the genera Fusobacterium, Leptospira, Listeria, Pasteurella, Rickettsia, Faecalis, Egasella, Lactonifactory, Coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Bifidobacterium, Mobiluncus, Enterococcus, Actinomyces, Neisseria, Chlamydia, Vibrio, Diplococcus, Lactobacillus, Kingella, Yersinia, and / or Klebsiella.

[0022] Importantly, the above-mentioned method for identifying the antagonistic antibacterial effect of at least two drug compounds on bacteria can be used to identify the antagonistic antibacterial effect of at least two drug compounds on a first bacterium, the first bacterium being a commensal and / or probiotic bacterium, while the at least two drug compounds have an antibacterial effect on a second bacterium, the second bacterium being a pathogenic bacterium. Thus, the at least two drug compounds have an antibacterial effect on pathogenic bacteria, while preventing and / or reducing collateral damage to commensal and / or probiotic bacteria. Thus, the use of at least two drug compounds is more advantageous than using only one of the compounds, since the combination of compounds prevents damage to the compounds that have an antibacterial effect on commensal and / or probiotic bacteria.

[0023] A further aspect of the invention then relates to a method for developing a targeted therapy for use in the prevention and / or treatment of a bacterial infection, comprising the steps of: a) carrying out the method as described above; b) identifying the selective antibacterial effect of the combination of at least two drug compounds against bacteria; c) selecting the at least two drug compounds identified in step b); and The present invention relates to a method comprising the steps of:

[0024] A further preferred embodiment of the present invention is a method for producing an antibacterial pharmaceutical composition comprising the steps of: a) carrying out the method as described above; b) formulating the selected combination into an antibacterial pharmaceutical composition; The method includes:

[0025] Yet another aspect of the present invention relates to an antibacterial pharmaceutical composition prepared by the above-described method.

[0026] The present invention also relates to a method for preventing adverse effects on the gut microbiome using the antibacterial pharmaceutical composition of the present invention, wherein at least one of the components of the composition has an antibacterial effect on at least a first and a second bacterial species, and at least one other component interferes with said antibacterial effect on said first bacterial species.

[0027] More preferably, in the aforementioned method for preventing adverse effects on the intestinal microbiome using the antibacterial pharmaceutical composition of the present invention, said first bacterium to be tested for said antibacterial effect is a commensal bacterium and / or a probiotic bacterium, such as Enterobacter, Staphylococcus, Streptococcus, Pseudomonas, Escherichia, Helicobacter, Neisseria, Campylobacter, Clostridium, Citrobacter, Vibrio, Treponema, Mycobacterium, Klebsiella, Actinomycetes, Bacteroides, Bordetella, Brucella, Corynebacterium, Diplococcus, Fusobacterium, Leptospira , Pasteurella, Proteus, Rickettsia, Shigella, Parabacteroides, Odoribacter, Faecalibacterium, Collinsella, Egasella, Lactonifactory, Roseburia, Coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Kingella kingae, Haemophilus, Bifidobacterium, Mobiluncus, Prevotella, Akkermansia, Bylophila, Blautia, Coprococcus, Dorea, Eubacterium, Lactobacillus, Ruminococcus, Veillonella, and / or Enterococcus.

[0028] The at least one first bacterial species may be one bacterial species or 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39 or 40 other bacterial species or any number of commensal and / or probiotic bacterial species.

[0029] Additionally, preferably, in the above-mentioned method for preventing adverse effects on the intestinal microbiome using the antibacterial pharmaceutical composition of the present invention, the above-mentioned second bacterium is a pathogenic bacterium, for example, Enterobacter, Escherichia, Shigella, Serratia, Proteus, Pseudomonas, Acinetobacter, Staphylococcus, Streptococcus, Pseudomonas, Salmonella, Helicobacter, Citrobacter, Treponema, Mycobacterium, Bordetella, Borrelia, Brucella, Corynebacterium, Lactobacillus, Lactobacillus casei ... In one embodiment, the bacterium is a member of the genera Escherichia, Fusobacterium, Leptospira, Listeria, Pasteurella, Rickettsia, Faecalis, Egasella, Lactonifactory, Coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Bifidobacterium, Mobiluncus, Enterococcus, Actinomyces, Neisseria, Chlamydia, Vibrio, Diplococcus, Lactobacillus, Kingella, Yersinia, and / or Klebsiella.

[0030] The problem is also solved by the present invention by providing a method for preventing the emergence and / or spread of antibacterial resistance using the antibacterial pharmaceutical composition of the present invention.

[0031] A further aspect of the invention is an antibacterial pharmaceutical composition comprising: (i) vanillin or a pharma- ceutically acceptable salt thereof, or a vanillin derivative or a pharma- ceutically acceptable salt thereof; (ii) at least one antibacterial compound or a pharma- ceutically acceptable salt thereof; The present invention relates to an antibacterial pharmaceutical composition comprising:

[0032] Preferably, in the antibacterial pharmaceutical composition as described above, the at least one antibacterial compound or a pharma- ceutical acceptable salt thereof is selected from the group consisting of an aminoglycoside, a macrolide, a penicillin, a tetracycline, a lincosamide, a quinolone, a fluoroquinolone, a β-lactam, a polymyxin, a monobactam, a glycylcycline, an ansamycin, a sulfonamide, an oxazolidinone, a carbacephem, a carbapenem, a cephalosporin, a streptogramin, a cephalospor ... streptogramin, glycopeptides, polypeptides, arsphenamine, chloramphenicol, clindamycin, lincomycin, daptomycin, trimethoprim, novobiocin, ethambutol, fosfomycin, fusidic acid, furazolidone, isoniazid, linezolid, metronidazole, mupirocin, nitrofurantoin, platensimycin, pyrazinamide, quinupristin the aminoglycosides selected from rifamycins such as quinupristin, dalfopristin, rifampin, rifampicin, rifabutin or rifaximin, tinidazole, viomycin and capreomycin, or pharma- ceutin , in particular streptomycin, dihydrostreptomycin, amikacin, apramycin, arbekacin, astromycin, bekanamycin, dibekacin, framycetin, gentamicin, hygromycin, isepamicin, kanamycin, neomycin, netilmicin, paromomycin, rhodostreptomycin, ribostamycin, sisomicin, spectinomycin, tobramycin, vancomycin and verdamycin, or pharma-ceutin, in particular spectinomy ...

[0033] The present invention also provides an antibacterial pharmaceutical composition, comprising: (i) a polymyxin selected from colistin and polymyxin B, or a pharma- ceutically acceptable salt thereof, preferably colistin or a pharma- ceutically acceptable salt of colistin; (ii) at least one other pharmaceutical compound, or a pharma- ceutically acceptable salt thereof, wherein the at least one other pharmaceutical compound, or a pharma- ceutically acceptable salt thereof, is selected from the group consisting of loperamide, rifamycin, such as rifampicin, rifabutin, or rifaximin, macrolides, aminoglycosides, penicillins, tetracyclines, lincosamides, quinolones, fluoroquinolones, β-lactams, polymyxins, monobactams, glycylcyclines, ansamycins, sulfonamides, oxazolidinones, carbacephems, carbapenems, cephalosporins, streptogramins, glycopeptides, polypeptides, arsphenamine, chloramphenicol, clindamycin, lincomycin, daptomycin, trimethoprim, novobio at least one other drug compound selected from the group consisting of rifamycin, ethambutol, fosfomycin, fusidic acid, furazolidone, isoniazid, linezolid, metronidazole, mupirocin, nitrofurantoin, platensimycin, pyrazinamide, quinupristin, dalfopristin, rifampin, tinidazole, viomycin and capreomycin, or a pharma- ceutically acceptable salt thereof, preferably wherein said at least one other drug compound is a rifamycin, such as loperamide, rifampicin, rifabutin or rifaximin, or a macrolide, such as erythromycin, azithromycin, clarithromycin or roxithromycin, or a pharma-ceutically acceptable salt thereof; The present invention relates to an antibacterial pharmaceutical composition comprising:

[0034] The present invention also provides an antibacterial pharmaceutical composition, comprising: (i) Aminoglycosides, macrolides, penicillins, tetracyclines, lincosamides, quinolones, fluoroquinolones, β-lactams, polymyxins, monobactams, glycylcyclines, ansamycins, sulfonamides, oxazolidinones, carbacephems, carbapenems, cephalosporins, streptogramins, glycopeptides, polypeptides, arsphenamines, chloramphenicol, clindamycin, lincomycin, daptomycin, trimethoprim ... an antibacterial compound selected from prim, novobiocin, ethambutol, fosfomycin, fusidic acid, furazolidone, isoniazid, linezolid, metronidazole, mupirocin, nitrofurantoin, rifamycin, such as rifampicin, rifabutin or rifaximin, platensimycin, pyrazinamide, quinupristin, dalfopristin, rifampin, tinidazole, viomycin and capreomycin, or a pharma- ceutically acceptable salt thereof; (ii) at least one other pharmaceutical compound selected from antibiotics, human targeting drugs, food additives, or pharma- ceutically acceptable salts thereof, preferably, said at least one other pharmaceutical compound is procaine, metformin, benzalkonium, berberine, erythromycin, clarithromycin, aztreonam, loperamide, pyocyanin, phenazine methosulfate, clindamycin, rifamycin such as rifampicin, rifabutin, or rifaximin, paraquat, trimethoprim, doxycycline, curcumin, vanillin, caffeine, acetylsalicylic acid, epigallocatechin gallate, CHIR090, minocycline, spectinomycin, and fosfomycin, or pharma-ceutically acceptable salts thereof; Including, Optionally, the problem is solved by providing an antibacterial pharmaceutical composition, wherein the antibacterial compound (i) has an antibacterial effect against at least a first bacterium and a second bacterium, the compound (ii) antagonizes the antibacterial effect of the compound (i) against at least one of the first bacteria, and the compound (ii) does not antagonize the antibacterial effect of the compound (i) against the second bacterium.

[0035] In the context of the present invention, "antagonizing" refers to having an "antagonizing effect." In other words, antagonistic effect means that if at least one of the compounds has an antibacterial effect against a bacterium or bacterial species, this antibacterial effect is countered by at least one other compound, i.e., at least one other compound is "antagonized."

[0036] More preferably, in the above-mentioned antibacterial pharmaceutical composition, the first bacterium is a commensal bacterium and / or a probiotic bacterium, for example, Enterobacter, Staphylococcus, Streptococcus, Pseudomonas, Escherichia, Helicobacter, Neisseria, Campylobacter, Clostridium, Citrobacter, Vibrio, Treponema, Mycobacterium, Klebsiella, Actinomyces, Bacteroides, Bordetella, Brucella, Corynebacterium, Diplococcus, Fusobacterium, Leptospira, Pasteurella, Proteus, Rickettsia, Shigella, Paramecium, or the like. The antibacterial pharmaceutical composition is an antibacterial pharmaceutical composition comprising a member of the genera Bacteroides, Odoribacter, Faecalibacterium, Collinsella, Egasella, Lactonifactor, Roseburia, Coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Kingella kingae, Haemophilus, Bifidobacterium, Mobiluncus, Prevotella, Akkermansia, Bylophila, Blautia, Coprococcus, Dorea, Eubacterium, Lactobacillus, Ruminococcus, Veillonella, and / or Enterococcus.

[0037] In addition, in the pharmaceutical composition described above, the second bacterium is preferably a pathogen, for example, Enterobacter, Escherichia, Shigella, Serratia, Proteus, Pseudomonas, Acinetobacter, Staphylococcus, Streptococcus, Pseudomonas, Salmonella, Helicobacter, Citrobacter, Treponema, Mycobacterium, Bordetella, Borrelia, Brucella, Corynebacterium, Fusobacterium, Leptospira, Lysobacterium, Leptospira ... In one embodiment, the pharmaceutical composition is a member of the genera Escherichia, Pasteurella, Rickettsia, Faecalibacterium, Egasella, Lactonifactory, Coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Bifidobacterium, Mobiluncus, Enterococcus, Actinomyces, Neisseria, Chlamydia, Vibrio, Diplococcus, Lactobacillus, Kingella, Yersinia, and / or Klebsiella.

[0038] Optionally, said (ii) compound antagonizes said antibacterial effect of said (i) compound against 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39 or 40 other bacterial species, or any number of other bacterial species.

[0039] The composition of compounds as described above can be used in general for a wide variety of medical applications, in particular for the prevention and / or treatment of bacterial infections and / or dysbiosis.Preferably, the composition as described above is for use in the prevention and / or treatment of bacterial infections, the bacterial infections being selected from gastrointestinal infections, genitourinary tract infections, upper and lower respiratory tract infections, rhinitis, tonsillitis, pharyngitis, bronchitis, pneumonia, visceral infections, nephritis, hepatitis, peritonitis, endocarditis, meningitis, osteomyelitis, eye infections, ear infections, skin infections, subcutaneous infections, post-burn infections, diarrhea, colitis, pseudomembranous colitis, skin disorders, toxic shock syndrome, bacteremia, septicemia, pelvic inflammatory disease, central nervous system infections, wound infections, intraperitoneal infections, intravascular infections, bone infections, joint infections, acute bacterial otitis media, pyelonephritis, deep abscesses and tuberculosis.

[0040] More preferably, it is a composition as described above, wherein the bacterial infection to be treated and / or prevented is caused by a Gram-negative bacterium, preferably said Gram-negative bacterium is a member of the family Enterobacteriaceae or Moraxellaceae, for example a Gammaproteobacteria such as a member of the genera Enterobacter, Escherichia, Salmonella, Klebsiella, Yersinia, Shigella, Serratia, Proteus, Pseudomonas, and / or Acinetobacter, or any closely related genera thereof, and optionally said Gram-negative bacterium is an antibiotic-resistant bacterium, in particular a multi-drug resistant strain thereof.

[0041] Importantly, many bacterial species present in the human or patient body are commensal or probiotic bacterial species. Compounds with antibiotic effects are often also effective against these commensal or probiotic bacterial species, thereby damaging, for example, a healthy intestinal microbiome. It is therefore important to prevent such damaging effects on commensal and / or probiotic bacteria, while allowing antibiotic effects against pathogenic bacteria that are the cause of infection and / or disease.

[0042] The term "microbiota" collectively refers to the entire set of microorganisms found associated with higher organisms, such as humans. Organisms belonging to the human microbiota can generally be classified into bacteria, archaea, yeasts, and unicellular eukaryotes, as well as viruses and various parasites.

[0043] The term "microbiome" collectively refers to the entire set of microorganisms, their genetic elements (genomes), and environmental interactions found associated with higher organisms such as humans.

[0044] The microbiome contains many symbiotic or probiotic bacterial strains. The term "commensal" refers to organisms that are usually harmless to the host and can even form a mutualistic relationship with the host. It has been shown that the human body contains approximately 100 trillion symbiotic organisms, ten times more numerous than human cells.

[0045] The term "probiotic" as used herein means a live microorganism that confers a health benefit to the host when administered in appropriate amounts. Probiotics are available in foods and dietary supplements, such as, but not limited to, capsules, tablets, and powders. Examples of foods that contain probiotics are yogurt, fermented and unfermented milk, miso, tempeh, and some juices and soy drinks. Some bacterial strains of the microbiome, such as Lactobacillus, Bifidobacterium, Enterococcus, Streptococcus, Pediococcus, Leuconostoc, Bacillus, Escherichia, and Lactococcus, are known to have probiotic functions.

[0046] The term "dysbiosis" (also called bacterial overgrowth) refers to any kind of microbiome imbalance. For example, species that are normally underrepresented in the healthy human microbiome become overrepresented in dysbiosis, whereas species that normally dominate in healthy humans become underrepresented in dysbiosis. Most often, dysbiosis is a gastrointestinal disorder, especially in the setting of small intestinal bacterial overgrowth (SIBO) or small intestinal fungal overgrowth (SIFO). Dysbiosis has been reported to be associated with diseases such as inflammatory bowel disease, bacterial vaginosis, and colitis.

[0047] Yet another embodiment of the present invention relates to a composition for the above mentioned uses, wherein the components of said composition are administered simultaneously, separately or sequentially to a subject, said subject being a mammal such as a human, preferably a human patient, and optionally said composition is in liquid, dry or semi-solid form, e.g. in the form of tablets, coated tablets, effervescent tablets, capsules, powders, granules, dragees, troches, pills, ampoules, drops, suppositories, emulsions, ointments, gels, tinctures, pastes, creams, moist compresses, mouthwashes, plant juices, nasal drops, inhalation mixtures, aerosols, mouthwashes, mouth sprays, nasal sprays or room sprays.

[0048] The invention will now be further described in the following examples and with reference to the accompanying drawings, in which the invention is not limited thereto. For the purposes of the present invention, all cited references are incorporated herein by reference in their entirety. [Brief description of the drawings]

[0049] [Figure 1-1]High-throughput profiling of paired drug mixtures in Gram-negative bacteria. a) Selection of drugs and strains for screening. The 79 drugs used for the combination screen are grouped into categories. Antimicrobials are grouped by target, except for antibiotic classes for which sufficient representatives (>2) were screened to constitute separate categories: β-lactams, macrolides, tetracyclines, fluoroquinolones, and aminoglycosides. Classification of human target drugs and food additives was not further refined, as the MoA is unclear in most cases. A subset of 62 sequence drugs was profiled against the full set of 79 drugs in 6 strains. Strains are color-coded according to species. b) Quantification of drug-drug interactions. Growth was profiled by measuring absorbance (OD595 nm) over time in the absence of drug, in the presence of a single drug, and in the presence of both drugs. Interactions were defined according to Bliss independence. Fitness significantly lower or higher than expected (fa x fq) indicates synergy or antagonism, respectively. Synergy and antagonism were assessed by growth on a 4 x 4 checkerboard. [Figure 1-2] Same as above [Figure 2-1]Figure 1 shows the principle of drug-drug interaction networks. a) Antagonism is prevalent compared to synergy. Observed ratio of detectable synergy and antagonism in 6 strains. We detected more antagonistic interactions (1354) than synergistic interactions (1230), but the power of detection of these antagonisms is lower, 12778 combinations compared to 16920 combinations. b) and d) Drug-drug interaction network in E. coli. Nodes represent either drug categories (b) or drugs (d) classified according to the common cellular process they target. The color of the nodes indicates the target cellular process according to Fig. 1a, and the size of the node reflects the number of drugs in the category / group. Edges represent synergy (blue) and antagonism (orange), and the thickness reflects the number of interactions between drugs in the respective category / group. Interactions between drugs of the same category or common cellular targets are represented by self-interaction edges. Conserved interactions, including weak interactions, are shown. c) and e) Antagonism occurs almost exclusively between drugs belonging to different categories (c) or targeting different cellular processes (e), whereas synergy is more prevalent between drugs in the same category (c) or targeting the same process (e). In b and d, quantification of drug-drug interactions in E. coli is shown. p-values ​​from chi-square tests are shown. [Figure 2-2] Same as above [Figure 3-1]Conservation of drug-drug interactions between strains and species. a) Drug-drug interactions are conserved in E. coli. Scatter plot of interaction scores for two E. coli strains; significant interactions for at least one strain are shown. Dark blue: conserved strong interactions in both strains, light blue: strong interactions in one strain and concordant action in the other (conserved weak interactions). Grey: interactions occurring exclusively in one strain or reciprocal interactions between strains (non-conserved interactions). R represents Pearson correlation and n represents the total number of interactions plotted. b) Drug-drug interactions are highly conserved within all three species. Significant interactions observed in at least one strain per species are shown. Colors are as in a, incomparable refers to combinations with significantly different single drug dose responses between strains. c) Drug interaction profiles are phylogenetically derived. Clustering of strains is based on Pearson correlation of drug interaction profiles (considering all pairwise drug mixtures). Strains of the same species cluster with two Enterobacteriaceae species, E. coli and S. Typhimurium, and act more similarly to each other than to the phylogenetically more distant P. aeruginosa. d) Drug-drug interactions are largely species-specific. Venn diagram shows the overlap of interactions between the three species, where n=total number of interactions, nc=opposing interactions between species (synergy in one species, antagonism in the other), not explained by the Venn diagram. e) Conserved drug-drug interaction network. Nodes represent individual drugs, classified and colored by target cellular process (as in Fig. 2d). Drug names are represented by three-letter codes. Dashed and full edges correspond to conserved interactions between two or three species, respectively. f) Synergy is more conserved than antagonism. Mosaic plots show quantification of synergy and antagonism between conserved and non-conserved interactions between bacterial species. Chi-squared p-values ​​are shown. [Figure 3-2] Same as above [Figure 3-3] Same as above [Figure 4-1] Vanillin induces a multiple antibiotic resistance (mar) phenotype. a) The drug-drug interaction profiles of vanillin and aspirin (acetylsalicylic acid) are similar (see FIG. 14), suggesting a common MoA. A schematic of the induction of the mar response via inactivation of the MarR repressor by salicylate / aspirin is shown. b) Vanillin increases AcrA protein levels in a marA-dependent manner. Representative immunoblots of exponentially growing cells untreated or treated with vanillin (150 μg / ml) or aspirin (500 μg / ml) are shown, with loading controlled by cell density and constitutively expressed RecA. Barplots represent quantification of AcrA protein levels, n=5-6. c) marA expression levels in response to vanillin (150 μg / ml) or aspirin (500 μg / ml) treatment are stronger in wild type than in the ΔmarR mutant. Expression was measured by RT-qPCR and normalized to no drug treatment in wild type, n=4. d) and e) Vanillin (150 μg / ml) and aspirin (500 μg / ml) increase the MIC of chloramphenicol (d) or ciprofloxacin (e). Antagonism is weaker and abolished in the ΔmarA and ΔacrA mutants, respectively, n=3. Error bars represent standard deviation (b-e). [Figure 4-2] Same as above [Figure 4-3] Same as above [Figure 5-1]Figure 1. Potent synergistic combinations against Gram-negative MDR clinical isolates. a) In vitro synergy shown as 8x8 checkerboards. The vanillin-spectinomycin combination is synergistic only against MDR E. coli strains. The drug pairs are the same for each row and are shown in the first checkerboard. The species for which interactions were detected in the screen are shown after the last checkerboard. Concentrations increase in equal steps for each drug (see key), and only the minimum and maximum concentrations are shown in μg / ml for the first strain of each species. The same concentration range was used, except for colistin. Higher colistin concentrations were used for colistin-resistant K. pneumoniae 929 as indicated. One of two biological replicates is shown. b) Drug synergy against the same MDR strains in a Galleria mellonella infection model. Larvae were infected with E. coli and K. pneumoniae MDR isolates (106 and 104 CFU, respectively) and left untreated or treated with single agents or combinations. Percent larval survival was monitored at the indicated intervals post-infection, n=10 larvae per treatment. Means of triplicate or quadruplicate biological samples are shown, error bars represent standard deviation. [Figure 5-2] Same as above [Figure 6-1]Figure 3 shows the data analysis pipeline. a) Flowchart of the data analysis pipeline. b) Estimation of single drug fitness of arrayed drugs. Because drug-drug interactions are rare, the slope of the best-fit line between gaq (growth with dual drugs) and gq (growth with query drug alone, estimated from the average of the top 5% growth wells between plates) between np query drugs (plates) corresponds to the proxy for the fitness of arrayed drugs alone, fa (Eq. 3). R represents the Pearson correlation coefficient between gaq and gq between np plates. As an example of arrayed drugs with several interactions, well A9 of E. coli BW25113 containing 3 μg / ml spectinomycin is shown, where several query drugs (plates) deviate from the expected fitness (light grey dots), so only half of the plate, corresponding to the interquartile range of fa, was used to estimate fa. c) Density distribution of the first, second and third quartiles of the Bliss score (ε) distribution of E. coli. Q1, Q2 and Q3 represent the medians of the first, second and third quartiles of the ε distribution, respectively. n represents the number of admixtures used. [Figure 6-2] Same as above [Figure 7-1]Figure 14. Data quality control. a) High replicate correlation for single and dual drug treatments. Transparent box plots contain Pearson correlation coefficients between plates of the same batch containing only array drugs (LB was used instead of the second drug). n represents the total number of correlations. Full box plots contain Pearson correlation coefficients between double drug replicate wells in the same plate across all plates. n represents the number of wells used for correlation, nmax = (62 drugs + 1 LB) x 3 concentrations = 189. Only wells with median growth above 20% were considered. b) Wells with lower median growth have lower replicate correlation. Dual drug correlation coefficients used to generate box plots from a are plotted as a function of median growth of all wells across all plates for E. coli iAi1. Wells with overall lower growth (due to strong inhibition of array drugs) are less reproducible due to a lower spread of growth values ​​combined with the sigmoidal nature of the drug dose-response curves. c) Drug interactions are rare. Density distribution of all Bliss scores (ε) determined per strain. d) The ability to detect synergy and antagonism varies with the effect of single drug treatments. Bliss scores (ε) are plotted as a function of expected fitness (fa×fq) for all drug concentration ratios of all combinations in E. coli BW (example). Box plots summarizing both variables are shown in addition to the axes (n=101322, center line corresponds to the median, whiskers cover 1.5 times the IQR). Blind spots for detecting antagonism and synergy are shown, both of which are based on expected fitness (see also Fig. 8c and d) and therefore dependent on the growth of strains with single drugs. The number of drugs used in the screen that do not inhibit E. coli alone increases the number of drug mixtures with antagonism blind spots. e) Scatter plot of the number of interactions per drug versus the minimum fitness of the drug alone (as determined in the screen). Strong and weak interactions are represented. n represents the total number of interactions and R is the Pearson correlation coefficient. Strains are color coded as panels a and c.f) Density distribution of the number of interactions per drug across all strains. [Figure 7-2] Same as above [Figure 7-3] Same as above [Figure 8-1]Figure 3. Benchmarking. a) The validation set is enriched for synergy and antagonism to better assess true and false positives. Comparison of interaction rates between screening and validation sets. Screening aggregation considers both strong and weak interactions (Fig. 3b). b) Number of benchmarked interactions per strain. c) and d) Sensitivity analysis of statistical thresholds for calling interactions. c) Total amount of interactions as a function of expected fitness (fa × fb) cutoff used to restrict ε distribution to suitable drug concentrations. Strong drug-drug interactions are classified into full distribution only (i.e., all expected fit wells), suitable wells only (i.e., all wells with fa × fb greater than the cutoff for synergy and all wells with fa × fb less than (1-cutoff) for antagonism), or both, depending on the ε distribution in which they were significant. Weak drug-drug interactions are assigned independently and are represented in white for completeness. We selected an expected fit cutoff of 0.2 because it resulted in the highest precision and recall (91% and 74%, respectively) along with the highest number of total interactions detected after benchmarking against the validation dataset. d) Receiver Operating Characteristic (ROC) curves of screening across different p-value thresholds (Wilcoxon rank sum permutation test) as unique criteria for assigning interactions. The p-value selected for the screening threshold (0.05) is indicated by a grey cross. Sensitivity to additional parameters for calling hits is shown, which allows interactions to be either antagonistic or synergistic but not both (unilateral interactions), i.e., strong and weak interaction thresholds. True and false positive rates were estimated based on the validation dataset. The precision and reproducibility of the final and best parameter run set is shown, with one-way interactions, p<0.05, fa×fb cutoff=0.2, and |ε|>0.1 for strong interactions and |ε|>0.06 for weak interactions.TP, TN, FP and FN stand for True Positives, True Negatives, False Positives and False Negatives, respectively. n indicates the total number of benchmarked drug mixtures. e) Synergy between β-lactams by Loewe additivity interaction model. 8x8 checkerboard results of three combinations between β-lactams of four strains are shown. The grey line in each plot represents the null hypothesis of the Loewe additivity model, whereas the black line corresponds to the IC50 isobol, which is estimated by fitting a logistic curve to the interpolated drug concentrations (colored dots). Since piperacillin did not reach 50% growth inhibition in E. coli, IC20 and IC40 isobols were used for the amoxicillin + piperacillin combination in E. coli BW and E. coli iAi1, respectively. [Figure 8-2] Same as above [Figure 8-3] Same as above [Figure 8-4] Same as above [Figure 8-5] Same as above [Figure 9-1]Benchmarking of irreparable drug-drug interactions. a) Bar plots show the division of benchmarked drug mixtures according to their conservation within species. Pie charts show the proportion of false positives and true positives (FP and TP) and false negatives and true negatives (FN and TN) within the irreparable drug-drug interactions. b) Amoxicillin and cefotaxime combination in P. aeruginosa: an example of an irreparable drug-drug interaction. The results of the screening are shown in the upper box. The BLISS scores as a function of expected fitness for both strains are shown on the left and the density distribution of the BLISS scores is shown on the right. n represents the total number of BLISS scores, Q1 and Q3 indicate the BLISS scores in the first and third quartiles, respectively. Antagonism was only detected for PAO1 (Q3>0.1). PA14 was highly resistant to both drugs (top left panel), which made the detection of antagonism impossible. Benchmark results show that the interaction is antagonistic in both strains, although weaker in PA14 and barely visible at higher concentrations (bottom box). The colors on the checkerboard reflect the goodness of fit, with black dots corresponding to drug ratios with Bliss scores above 0.1. [Figure 9-2] Same as above [Figure 9-3] Same as above [Figure 10-1]Benchmarking of conserved weak drug-drug interactions. a) Bar plot shows the partition of the benchmarked mixture as in Fig. 9a. Pie chart shows the proportion of false positives (FP) and true positives (TP) within the conserved weak interactions. b) Doxycycline and amikacin combination in Salmonella typhimurium: an example of a conserved weak drug-drug interaction. The results of the screening are shown in the upper box. The Bliss scores as a function of expected fitness for both strains are shown on the left and the density distribution of the Bliss scores is shown on the right. n represents the total number of Bliss scores, Q1 and Q3 indicate the Bliss scores in the first and third quartiles, respectively. A strong synergy was only detected for ST14028 (Q1<-0.1), after which a conserved weak synergy was assigned to ST LT2 (Q1<-0.06). The benchmark results shown in the lower box confirm that the interaction is synergistic in both strains. The colors on the checkerboard reflect the goodness of fit, with black dots corresponding to drug ratios with Bliss scores below -0.1. [Figure 10-2] Same as above [Figure 10-3] Same as above [Figure 11-1] Figure 1. Drug-drug interaction networks for Salmonella and Pseudomonas. a) and b) Drug category interaction network. Nodes represent drug categories according to Fig. 1a. Node color / size and edge color / thickness are plotted as in Fig. 2b. Here conserved interactions, including conserved weak interactions, are shown. c) and d) Drug-drug interactions across cellular processes. Here, represented similarly to a and b, but with classification of drug categories targeting the same general cellular process. e) Quantification of synergy and antagonism in the networks of a and b. p-values ​​of chi-square tests are shown. Similar to E. coli, antagonism occurs more frequently and almost exclusively than synergy between drugs belonging to different categories in S. typhimurium and P. aeruginosa. Very few interactions occur between drugs of the same category in P. aeruginosa. [Figure 11-2] Same as above [Figure 11-3] Same as above [Figure 11-4] Same as above [Figure 12-1]Figure 1: Drug antagonism is often due to a decrease in intracellular drug concentration. a) Schematic of MoA decreasing intracellular drug concentration (black) via decreased uptake or increased efflux by addition of a second drug (antagonist; blue). b) Various antagonists of gentamicin (red, 5 μg / ml) and ciprofloxacin (gold, 2.5 μg / ml) identified in our screen of E. coli BW also rescue the killing effect of the two bactericidal drugs in the same strain or its parent strain MG1655 (top right and top left panels, respectively). With the exception of clindamycin (against gentamicin) and curcumin (against ciprofloxacin), all other antagonists decrease the intracellular concentration of the interacting drug (bottom panel), gentamicin detected using a radiolabeled compound and ciprofloxacin detected by LC-MS / MS. The degree of rescue (upper panel) reflects the decrease in intracellular concentration (lower panel), meaning that most of these interactions are highly dependent on regulating the intracellular concentration of the antagonized drug. c) In the E. coli BW mutant, which lacks key components controlling the intracellular concentration of the antagonized drug, the antagonism dissipates. Aminoglycosides depend on PMF-energized uptake and thus on the respiratory complex, while ciprofloxacin is effluxed by AcrAB-TolC. For gentamicin, most interactions dissipate when respiration is impaired, even the interaction with clindamycin (which does not regulate intracellular gentamicin concentration, see panel (b)), probably because MoA and the uptake of aminoglycosides are linked in a positive feedback loop. For ciprofloxacin, the antagonism with paraquat and caffeine is abolished in the ΔacrA mutant, implying that both compounds induce the AcrAB-TolC pump (known for paraquat). In contrast, the interactions with curcumin, benzalkonium and doxycycline remain largely intact in the ΔacrA mutant. First, curcumin is expected since it does not modulate intracellular ciprofloxacin concentrations (see panel b).In the other two cases, other components (possibly more than one) in addition to AcrAB-TolC are likely responsible for the changes in ciprofloxacin uptake / efflux. Ciprofloxacin and gentamicin concentrations were adjusted in all strains according to their MICs (MICs of 70% and 100% for ciprofloxacin and gentamicin, respectively). Bliss interaction scores (ε) were calculated similarly to the screening and are represented by the mean and standard deviation among triplicate to eight replicate samples. d) Antagonism network of gentamicin and ciprofloxacin in E. coli BW. Nodes represent drugs colored according to the target cellular process (as in Fig. 1a). Solid and dashed edges represent antagonistic drug-drug interactions where intracellular antibiotic concentrations were measured and where intracellular antibiotic concentrations were not measured, respectively. Drug interactions resulting in a decrease in the intracellular concentration of the antagonized drug are represented by black edges. e) Quantification of antagonistic drug-drug interactions of the network in (d). The fluoroquinolones and aminoglycosides bars illustrate the presumption of antagonistic interactions against all other members of the two classes, assuming they exhibit the same activity as ciprofloxacin and gentamicin, respectively. [Figure 12-2] Same as above [Figure 12-3] Same as above [Figure 12-4] Same as above [Figure 13-1]Figure 2: Most drug-drug interactions are conserved within species and only partially driven by MoA. a) and b) Drug-drug interactions are conserved in Salmonella typhimurium (a) and Pseudomonas aeruginosa (b). Scatter plots of interaction scores for two strains of each species; only significant interactions for at least one strain are shown. Coloring and classification are as in Fig. 3a. R represents Pearson correlation and n represents the total number of plotted interactions. The low correlation for Pseudomonas aeruginosa is likely due to fewer and weaker interactions overall. c) Monochromaticity among all drug categories. The monochromaticity index (MI) reflects whether the interaction between the two categories of drugs is more synergistic (MI=-1) or antagonistic (MI=1) than the background ratio of synergy and antagonism. MI was calculated using all interactions of the six strains for all category pairs with at least two interactions. White cells in the heatmap correspond to category pairs where no interactions or an insufficient number of interactions were observed. d) Human targeted drugs, and LPS or PMF inhibitors, are strong and promiscuous adjuvants. The density distribution of MI per drug category in panel c is shown. n represents the amount of drugs in the category involved in i interactions. [Figure 13-2] Same as above [Figure 13-3] Same as above [Figure 13-4] Same as above [Figure 14-1]Figure 5: Active synergy against Gram-negative MDR clinical isolates. a) Additive mixed drugs against MDR E. coli and K. pneumoniae clinical isolates (related to Fig. 5a). Interactions are shown as 8x8 checkerboards, synergies have a thick black border. Drug pairs are the same per row and are shown in the first checkerboard. The species for which interactions were detected in the screen are shown after the last checkerboard. Concentrations increase in equal steps for each drug (see color tone), only the minimum and maximum concentrations are shown in μg / ml for the first strain of each species. Except for colistin, the same concentration range was used for all E. coli and K. pneumoniae MDR strains. One of two replicates is shown. [Figure 14-2] Same as above [Figure 14-3] Same as above [Figure 15-1]Figure 3: Mechanism of action of vanillin-spectinomycin synergy. a) The MIC of spectinomycin is decreased by the addition of 100 μg / ml vanillin in wild-type E. coli BW and in single gene knockouts of members of the AcrAB-TolC efflux pump or its MarA regulator. Thus, vanillin-spectinomycin synergy is independent of the effect of vanillin on AcrAB-TolC (Figure 4). b) The synergy is specific between vanillin-spectinomycin, as 500 μg / ml aspirin, a vanillin-related compound, antagonizes spectinomycin and increases the MIC by about 3-fold. c) Profiling of vanillin-spectinomycin combinations in the E. coli BW Keio collection to deconvolute MoA. A violin plot of the drug interaction score ε for all mutants (n=9216) for the combination of vanillin-spectinomycin (synergy) and for the control combination of vanillin and another aminoglycoside, amikacin (antagonism) is shown. The red dots show the interaction scores for the two mdfA deletion clones in the Keio library. Vanillin-spectinomycin synergy is lost in the absence of mdfA, whereas vanillin-amikacin antagonism is unaffected, indicating that vanillin-spectinomycin synergy is dependent on MdfA. d) Deletion of mdfA increases the MIC of spectinomycin and abolishes synergy with vanillin, independent of the presence or absence of AcrAB-TolC. Mild overexpression of mdfA from a plasmid (pmdfA) further enhances synergy with vanillin, decreasing the MIC of spectinomycin by approximately 2-fold (compared to the MIC of the combination in the wild type). Thus, MdfA levels directly correlate to the degree of spectinomycin-vanillin synergy. e) Overexpression of mdfA increases spectinomycin susceptibility even though the MIC is unchanged. Growth of E. coli BW and pmdfA was measured in two-fold serial dilutions of spectinomycin (OD595 nm after 8 hours) and normalized to the untreated growth of the corresponding strains (white and black dots; mean of n=3).The dose response of spectinomycin was calculated using a logistic fit of the averaged data points (note that the MIC is calculated by first fitting the individual replicates and then averaging). The fitted curves are represented by solid and dashed lines for pmdfA and E. coli BW, respectively. f) Vanillin induces intracellular accumulation of spectinomycin in an mdfA-dependent manner. Intracellular spectinomycin is measured using tritium-labeled compounds (n=4). For all MIC bar plots, error bars represent standard deviation, n=3-10. [Figure 15-2] Same as above DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0050] The term "infection" as used in the context of the present invention relates to the presence of bacteria, viruses, fungi, protozoa or other microorganisms in or on a subject, and to invasion by bacteria, viruses, fungi, protozoa or other microorganisms. Invasion includes the unwanted proliferation of pathogenic microorganisms in a host organism. More generally, a microbial infection can be any situation in which the presence of a microbial population(s) is harmful to the host animal. Thus, a microbial infection exists when there are too many microorganisms in or on a mammal, or when the effects of the presence of a microbial population(s) are damaging to cells or other tissues of the mammal. Thus, inhibiting the proliferation of such invasive microorganisms benefits a subject infected by a microbial population(s). Examples of bacterial infections are urinary tract infections (UTI), kidney infections (pyelonephritis), gynecological and obstetrical infections, respiratory tract infections (RTI), acute exacerbation of chronic bronchitis (AECB), community-acquired pneumonia (CAP), hospital-acquired pneumonia (HAP), ventilator-associated pneumonia (VAP), intra-abdominal pneumonia (IAI), acute otitis media, acute sinusitis, sepsis, catheter-associated sepsis, chancroid, chlamydia, skin infections, and bacteremia.

[0051] Said infections to be prevented and / or treated by the compositions of the present invention are preferably caused by Gram-negative bacteria, said Gram-negative bacteria being Gammaproteobacteria, e.g., members of the Enterobacteriaceae or Moraxellaceae families, such as members of the genera Enterobacter, Escherichia, Salmonella, Klebsiella, Yersinia, Shigella, Serratia, Proteus, Pseudomonas, and / or Acinetobacter, and optionally said bacteria are antibiotic-resistant bacteria, in particular multi-drug resistant strains thereof.

[0052] The term "antibiotic" as used herein refers to a chemical that, at low concentrations, kills or prevents the growth of certain microorganisms, usually bacteria, although some antibiotics are also used to treat fungal or protozoan infections. Antibiotics are used in human, animal or horticultural medicine to treat infections caused by microorganisms. Antibiotics included in the present invention include aminoglycoside antibiotics, polymyxins, oxazolidinones, streptogramins, ansamycins, carbacefems, carbapenems, cephalosporins, glycopeptides, glycylcyclines, macrolides, monobactams, penicillins, polypeptides, quinolones, fluoroquinolones, sulfonamides, β-lactams, tetracyclines, as well as vancomycin, daptomycin, trimethoprim, and the like. , novobiocin, arsphenamine, chloramphenicol, clindamycin, lincomycin, ethambutol, fosfomycin, fusidic acid, furazolidone, isoniazid, metronidazole, mupirocin, nitrofurantoin, platensimycin, pyrazinamide, polymyxins, quinupristin / dalfopristin, rifamycins, e.g., rifampicin, rifabutin or rifaximin, tinidazole, viomycin, and others such as capreomycin.

[0053] The term "pharmaceutical composition" refers to a preparation that is present in such a manner as to effect the biological activity of the active ingredient contained therein, and does not contain additional ingredients that are unacceptably toxic to the subject to which the composition is administered. The pharmaceutical composition of the present invention can be administered by a variety of methods known in the art. As will be appreciated by those skilled in the art, the route and / or mode of administration will vary depending on the desired results. Pharmaceutically acceptable diluents include saline and aqueous buffer solutions. "Pharmaceutically acceptable carrier" refers to ingredients in a pharmaceutical formulation that are non-toxic to a subject, other than the active ingredient. Pharmaceutically acceptable carriers include any solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like that are physiologically compatible. The carrier may be suitable for intravenous, intramuscular, subcutaneous, parenteral, spinal, or epidermal administration (e.g., by injection or infusion).

[0054] The pharmaceutical compositions according to the present invention may also contain auxiliary agents such as preservatives, wetting agents, emulsifying agents, and dispersing agents. Prevention of the presence of microorganisms can be ensured both by the above sterilization methods and by the inclusion of various antibacterial and antifungal agents, such as parabens, chlorobutanol, phenol, sorbic acid, and the like. It may also be desirable to include isotonic agents, such as sugars, sodium chloride, and the like, in the compositions. Furthermore, prolonged absorption of the injectable pharmaceutical form can be brought about by the inclusion of agents that delay absorption, such as aluminum monostearate and gelatin.

[0055] The phrases "parenteral administration" and "administered parenterally" as used herein refer to modes of administration other than enteral and topical administration, usually by injection, including, but not limited to, intravenous, intramuscular, intraarterial, intrathecal, intracapsular, intraorbital, intracardiac, intradermal, intraperitoneal, transtracheal, subcutaneous, subcuticular, intraarticular, subcapsular, subarachnoid, intraspinal, epidural, and intrasternal injection and infusion. Regardless of the route of administration selected, the compounds of the present invention, which may be used in a suitable hydrated form, and / or the pharmaceutical compositions of the present invention, are formulated into pharma- ceutical acceptable dosage forms by conventional methods known to those skilled in the art. The actual dosage level of the active ingredient in the pharmaceutical compositions of the present invention may be varied to obtain an amount of the active ingredient that is effective, but not toxic to the patient, to achieve the desired therapeutic response for a particular patient, composition, and mode of administration. The selected dosage level will depend on a variety of pharmacokinetic factors, including the activity of the particular composition of the invention being used, the route of administration, the time of administration, the rate of excretion of the particular compound being used, the duration of treatment, other drugs, compounds, and / or materials used in combination with the particular composition being used, the age, sex, weight, condition, overall health, and medical history of the patient being treated, as well as factors well known in the medical arts.

[0056] The composition must be sterile and fluid to the extent that the composition is deliverable by syringe. Besides water, in one embodiment, the carrier is an isotonic buffered saline solution. The proper fluidity can be maintained, for example, by the use of a coating such as lecithin, by maintaining the desired particle size in the case of dispersions, and by the use of surfactants. In many cases, it is preferable to include isotonic agents, such as sugars, polyalcohols such as mannitol or sorbitol, and sodium chloride in the composition.

[0057] The dosage regimen will be determined by the attending physician and clinical factors. As is well known in the medical field, the dosage for any one patient depends on many factors, including the patient's size, body surface area, age, the particular compound to be administered, sex, time and route of administration, overall health, and other drugs administered in combination. A typical dose may range, for example, from 0.001 μg to 1000 μg, although doses below or above this exemplary range are envisioned, especially considering the factors mentioned above. In general, the dosage regimen for normal administration of the pharmaceutical composition should be in the range of 1 μg to 10 mg units per day. If the dosage regimen is a continuous infusion, it should also be in the range of 1 μg to 10 mg units per kilogram of body weight per minute, respectively. Progress can be monitored by periodic evaluation. The compositions of the present invention can be administered locally or systemically. Administration is generally parenteral, e.g., intravenously, and pharmaceutical compositions can also be administered directly to a target site, e.g., by biolistic delivery to an internal or external target site, or by a catheter to a site in an artery. Preparations for parenteral administration include sterile aqueous or non-aqueous solutions, suspensions, and emulsions. Examples of non-aqueous solvents are propylene glycol, polyethylene glycol, vegetable oils, e.g., olive oil, and injectable organic esters, e.g., ethyl oleate. Aqueous carriers include water, alcoholic / aqueous solutions, emulsions, or suspensions, including saline and buffered media. Parenteral vehicles include sodium chloride solution, Ringer's dextrose, dextrose, and sodium chloride, lactated Ringer's, or fixed oils. Intravenous vehicles include fluid and nutrient replenishers, electrolyte replenishers (e.g., Ringer's dextrose-based electrolyte replenishers), and the like. Preservatives and other additives may be present, such as, for example, antimicrobial agents, antioxidants, chelating agents, and inert gases and the like. Additionally, the pharmaceutical compositions of the present invention may contain additional agents such as interleukins or interferons depending on the intended use of the pharmaceutical composition.

[0058] The term "subject" as used in certain embodiments in connection with the present invention preferably refers to a mammal, such as a mouse, rat, guinea pig, rabbit, cat, dog, monkey, or preferably a human. The term "patient" preferably refers to a mammal, such as a mouse, rat, guinea pig, rabbit, horse, cattle, cow, cat, dog, monkey, or preferably a human, such as a human patient, for whom diagnosis, prognosis or treatment is desired. The subject of the present invention may be at risk of suffering from diseases such as bacterial infections, viral infections, fungal infections, and parasitic infections. A more detailed description of medical applications related to the present invention is provided elsewhere in this specification.

[0059] As used herein, the term "treating" means stabilizing or reducing adverse symptoms associated with a disease, reducing the severity of a disease symptom, slowing the rate of progression of a disease, inhibiting or stabilizing the progression of a disease condition, or altering a metric associated with a disease state in a desired manner. EXAMPLES

[0060] To consistently and systematically evaluate drug-drug interactions in clinically relevant Gram-negative bacteria, we used a 4x4 checkerboard assay to screen 79 drugs alone and in pairwise combinations in 6 strains of 3 different species. We obtained approximately 3000 drug mixtures per strain, which allowed us to detect many general principles of drug-drug interactions. These included: i) most interactions are species-specific, while synergy is less frequent but more conserved, and ii) antagonism occurs exclusively between drugs targeting different processes, whereas synergy is common for drugs targeting the same process. We further demonstrate that antagonism is often caused by a decrease in intracellular drug concentration, and that many synergies are effective even against multidrug-resistant and extensively drug-resistant (MDR and XDR) clinical isolates. Using these data, the inventors then investigated the interaction mechanism of vanillin, which has no antibacterial activity on its own but antagonizes many drugs when used in combination, and in particular acts synergistically with spectinomycin against E. coli.

[0061] We dose-dependently profiled approximately 3000 drug pair combinations in three different Gram-negative bacteria, all associated with difficult-to-treat infections. Overall, we identified over 2500 synergistic and antagonistic interactions in the six strains tested, representing approximately 15% of the combinations screened. This quantitative and comprehensive dataset enabled us to derive general principles behind drug-drug interactions, address conservation across bacterial species, and discover potent synergistic interactions that are also effective against MDR clinical isolates of the same or closely related species.

[0062] Three general principles emerge from our data. First, drug-drug interactions are highly species-specific, even if individual drugs have the same cellular targets across bacterial species. This is likely because the underlying mechanisms behind drug-drug interactions are not conserved. Such mechanisms rely on both intracellular wiring between targeted processes and on the regulation of combined drug uptake / efflux. The wiring between processes appears to be poorly conserved even among closely related microbial species, and both uptake and drug efflux rely on the most diverse parts of the bacterial cell: its envelope, harboring redundant transport systems, and assembly machinery / enzymes. There are multiple consequences for the species specificity of drug-drug interactions. This means that for antibacterial drugs, narrow-spectrum therapy, the main endeavor of current and future drug development, can result from synergistic combinations of already approved drugs. On the other hand, species-specific antagonism can be used to mitigate the collateral damage to the gut microbiota caused by antibiotic therapy. Such antagonism may be a more general counterstrategy to minimize the adverse effects of drugs on the human gut microbiota, since non-antibiotics also cause significant damage to the resident gut flora.

[0063] Second, antagonism and synergy have clearly separable properties. Antagonism strictly occurs between drugs targeting different processes, whereas synergy is more likely to occur for drugs targeting the same process. This difference has a clear mechanistic basis at the drug target level. Inhibiting processes chemically or genetically at different stages is known to result in synergistic effects between organisms. Some of the most robust antibacterial monotherapies come from multi-targeted drugs that inhibit the same or directly related processes. On the other hand, combining drugs that target different core processes can help the organism reach a more stable equilibrium, as is the case with DNA and protein synthesis inhibitors. Consistently, genetic interactions are more commonly mitigated when genes are part of different functional processes in yeast.

[0064] Third, antagonism is more common than synergy, indicating that if a random or empirical mixture of drugs has an effect, this is most likely a reduction in the efficacy of the individual drugs. Even commonly used drug mixtures in the clinic, e.g., linezolid and meropenem in septic patients, can show strong antagonistic effects against some pathogens. Although antagonistic interactions raise questions of efficacy and potential toxicity in the clinic, their use can counter-select resistant isolates. On the other hand, synergy is more conserved than antagonism between pathogen species, which facilitates the clinical use of combinations.

[0065] And while antibacterial drugs of the same class had similar interactions with other drugs, most antagonism we tested was due to modulation of intracellular drug concentrations. This suggests that drug-drug interactions are only partially induced by MoA and should not be automatically interpreted as direct functional interactions of their primary targets. This is likely the reason for the low conservation of drug-drug interactions across bacterial species, while their primary targets are highly conserved. Furthermore, many antibiotic classes showed members with different subdivisions, or outlier behaviors. This exposes the danger of drawing general conclusions for the whole class by examining one of its members. Similarly, we did not observe exclusive synergy or antagonism between bactericidal drugs and oxidative stress, suggesting that the interrelationships of these various classes of antibiotics with reactive oxygen species may be more complex than previously thought. The interactions we report here are at the growth inhibition level. Although we did not systematically investigate, 16 / 16 drug-drug interactions were also detectable at the killing level. In the future, more systematic profiling will be required to evaluate how the outcomes of drug-drug interactions are related at different levels (inhibition, killing, persistent formation).

[0066] Beyond elucidating general principles, our study provides an unparalleled number of drug-drug interactions in Gram-negative species. Taking advantage of established insect infection models, we have demonstrated the efficacy of several synergistic pairs against MDR clinical isolates in vitro and, for two of these, in vivo. Many more drug pairs remain to be uncovered within our dataset. Interestingly, human target drugs were among the most frequent antibiotic adjuvants in our screens, and although we included only four food additives, we identified 64 synergistic interactions, one of which inhibited the growth of MDR E. coli isolates. Notably, in this case, vanillin acted synergistically with spectinomycin, as it increased its intracellular concentration via the specific enterobacterial transporter MdfA. Thus, profiling of more human target drugs and food additives in future combination screens will not only increase the potential room for resolution, but may also lead to efficient therapeutic strategies against MDR pathogens. Such adjunctive strategies are particularly relevant because many more human targeted drugs inhibit bacterial growth than previously recognized.

[0067] In summary, we have produced a comprehensive resource of paired drug mixtures in Gram-negative bacteria that reveals important principles of drug-drug interactions and provides a framework for evaluating their conservation across organisms or individuals. Such information may be useful as a basis for comparable screening in other microorganisms, studies investigating the underlying mechanisms of paired drug mixtures, and computational prediction of these outcomes. Furthermore, some of the general principles of drug-drug interactions may apply beyond infectious disease drugs and microorganisms. For antimicrobial therapy, our study highlights the promise that non-antibiotics hold as adjuncts, providing new directions for narrow-spectrum therapy.

[0068] method Strains, strain cultures and drugs For each of the three Gram-negative species profiled in this study, we used two widely used and sequenced strains: E. coli K-12 BW25113 and O8 IAI1, Salmonella enterica serovar Typhimurium LT2 and 14028s, and P. aeruginosa PAO1 and PA14. To validate the selected synergistic effects, we profiled six MDR clinical Enterobacteriaceae isolates recovered from human patient specimens: E. coli 124, 1027, 1334, and Klebsiella pneumoniae 718, 929, and 980.

[0069] All mutant strains used in this study were generated using the E. coli Keio knockout collection after PCR confirmation and retransducing mutations into wild-type BW25113 with P1 phage. When necessary, the kanamycin resistance cassette was excised using plasmid pCP20. Plasmids used for overexpression of mdfA were obtained from a mobile E. coli ORF library.

[0070] Drugs used in the present invention were purchased from Sigma Aldrich, except for metformin hydrochloride (TCI Chemicals), clindamycin and bleomycin (Applichem), CHIR-090 (MedChemtronica), and vanillin (Roth). Stocks were prepared according to the supplier's recommendations (preferably dissolved in water) and stored in the dark at -30°C until arrayed in plates. For all mixed drug experiments, drugs were diluted to appropriate working concentrations in clear 384-well plates (Greiner BioOne GmbH) and each well contained a total volume of 30 μl of Lysogeny Broth medium. After drug addition, cells were inoculated from overnight cultures at an initial OD595 nm of 0.01. The same starting OD was used for all strains. All liquid handling (drug addition, cell mixing) was performed using a Biomek FX liquid handler (Beckman Coulter). Plates were sealed with a breathable membrane (Breathe-Easy™) and incubated at 37°C in a humidity-saturated incubator (Cytomat2, Thermo Scientific) with continuous shaking and without the lid to avoid condensation. OD595 nm was measured every 40 min for 12 h with a Filtermax F5 multimode plate reader (Molecular Devices).

[0071] Calculation of minimum inhibitory concentration (MIC) We defined the MIC as the lowest concentration required to inhibit microbial growth after 8 h of incubation at 37° C. in lysogeny broth with shaking (384-well plates, starting OD595 nm 0.01). MICs for all drugs were calculated using a logistic fit of growth (OD595 nm at 8 h) against two-fold serial dilutions of antibiotic concentrations for all strains used in high-throughput screening and follow-up experiments.

[0072] High-throughput screening of pairwise drug interactions For all mixed drug experiments, drugs were diluted to the appropriate working concentration in clear 384-well plates (Greiner Bio-One) and each well received a total volume of 30 μl of LB. After drug addition, cells were inoculated from overnight cultures at an initial OD595 nm of approximately 0.01. The same inoculum size was used for all strains. All liquid handling (drug addition, cell mixing) was performed using a Biomek FX liquid handler (Beckman Coulter). Plates were sealed with a breathable membrane (Breathe-Easy™) and incubated in a humidity-saturated incubator (Cytomat2, Thermo Scientific) at 37°C with continuous shaking and without the lid to avoid condensation. OD595 nm was measured every 40 min for 12 h by a Filtermax F5 multimode plate reader (Molecular Devices). A flow chart of the experimental and analysis pipeline is shown in Fig. 6a. Data analysis was performed using R and networks were created using Cytoscape.

[0073] Experimental Pipeline Screening of drug-drug interactions was performed using a 4x4 checkerboard. 62 drugs were arrayed in duplicates in 384-well plates with various concentrations (array drugs). Each plate contained 12 randomly distributed wells without array drugs, 9 wells received only the query drug, and 3 wells did not receive any drug. One query drug at a single concentration was added to all wells of the 384-well plate except for 3 control wells. All drugs were queried once and sometimes twice per concentration. We used 78 drugs in E. coli and S. typhimurium, and 76 drugs in P. aeruginosa as queries. A total of 79 query drugs were screened, of which 75 drugs were common to all three species. The 62 array drugs were a subset of the 79 query drugs. The same drug concentration was used in both the query drug and the array drug. Three drug concentrations (2-fold dilution series) were selected based on the strain- and drug-tailored MIC curves: for each single drug, the highest drug concentration (as close as possible to the MIC) and the lowest goodness of fit were determined.

[0074] For drugs that did not inhibit growth alone, we selected concentrations according to the susceptibility of other strains / species or according to their use in the clinic or for research. E. coli and S. typhimurium showed similar single drug dosage responses within the species, so the same drug concentrations were used for both strains of each species. In contrast, MICs for P. aeruginosa often differed by several fold, and therefore drug concentrations were adjusted between the two strains.

[0075] Screening of drug-drug interactions was performed using a 4x4 checkerboard. Sixty-two drugs were arrayed in duplicate in 384-well plates with various concentrations (array drugs). Each plate contained 12 randomly distributed wells without array drugs, nine wells received only the query drug, and three wells did not receive any drug. One query drug at a single concentration was added to all wells of the 384-well plate, except for three control wells. All drugs were queried once and sometimes twice per concentration. We used 78 drugs in E. coli and S. typhimurium, and 76 drugs in P. aeruginosa as queries. A total of 79 query drugs were screened, of which 75 drugs were common to all three species. The 62 array drugs were a subset of the 79 query drugs. The same drug concentration was used in both the query drug and the array drug.

[0076] Growth curve smoothing and analysis A Gompertz model was fitted to all growth curves (when growth was observed) by using the R package grofit version 1.1.1-1 for noise reduction. The quality of the fit was assessed by Pearson correlation (R), which was higher than 0.95 for about 95% of all 505 growth curves. An R below 0.95 indicated either a non-sigmoidal shaped growth curve, characteristic of some drugs such as fosfomycin, or very noisy data. In the first case, the original data were kept for further analysis. In the second case, the noisy data were removed. Plate effects were corrected by fitting a polynomial to the median of each row and column. The LB background signal was removed by subtracting the median curve of non-growing wells from the same plate. These were wells where either single or double drug treatment completely inhibited growth, and each plate contained at least three such wells. To correct for systematic effects, the data were processed per strain and per batch.

[0077] Goodness of fit estimation To assess fitness, we used single time point OD595 nm measurements (growth), which correspond to the transition of unperturbed grown cells to stationary phase and allow us to understand the effect of drugs on lag phase, growth rate or maximum growth. To this end, we used OD595 nm at 8 hours for both E. coli BW25113 and P. aeruginosa strains, 7 hours for the fast-growers E. coli iAi1 and S. typhimurium 14028s, and 9 hours for the slow-growing S. typhimurium LT2.

[0078] According to the Bliss independence model, assuming that drug-drug interactions are rare, for most drug mixtures, the fitness of the array drugs (fa) is equal to the fitness in the presence of both drugs (faq) divided by the fitness of the query drug alone (fq).

number

number

number

[0079] For wells containing drugs that had many interactions (Pearson correlation r between gaq and gq less than 0.7), we restricted the query drug points (minimum 18) to improve the correlation and estimate fa. Wells where r was still less than 0.7 even after restricting the number of plates were removed from further analysis due to high noise (~2%). For wells that showed no growth in more than 75% of the plates in the batch, fa was considered to be 0.

[0080] Interaction score Bliss Independence Bliss scores (ε) were calculated for each well as described above (Equation 1). At least 3 × 3 drug concentrations × 2 (duplicates) × 2 (query drug and array drug) = 36 or 18 (query drug only) scores were determined for each drug pair. Drug-drug interactions were inferred in three stages based on the Bliss independence model: a) strong interactions based on the full ε distribution, b) strong interactions based on the ε distribution restricted to the relevant drug concentration, and c) conserved weak interactions within species. All drug-drug interactions were included in the inter-species comparisons, drug-drug interaction networks, and monochromatic analysis presented in this study.

[0081] a) Strong drug-drug interactions based on complete ε distribution 10,000 permutations of the Wilcoxon rank sum test (per drug pair, per strain) were performed. For every permutation, the ε distribution for a given combination was compared to an ε distribution of the same size randomly sampled from the complete ε set for a given strain. Permutation p-values ​​were calculated as follows:

number

[0082] b) Strong drug-drug interactions based on ε distribution restricted to the appropriate drug concentrations. Because drug interactions are concentration dependent, the same statistical procedure was repeated after restricting the drug concentration ratios to those appropriate for either synergy or antagonism. This constraint was imposed by excluding blind spots for both interactions, ε values ​​corresponding to concentration ratios whose expected fit (product of the fits of the single drugs, fa × fb) is less than 0.2 for synergy and greater than 0.8 for antagonism (Fig. 7d). These interactions are described by their p-values ​​and ε, determined using the restricted drug concentration ratios. Most interactions were detected based on both the full ε distribution and the restricted ε distribution, but each of the various methods had uniquely identified interactions (Fig. 8c). Using an expected fit cutoff of 0.2, we identified the maximum number of strong interactions (1950), with 90 uniquely identified interactions from the full ε distribution and 379 uniquely identified interactions from the restricted ε distribution (see also sensitivity analysis).

[0083] Furthermore, by restricting the ε value based on the expected fit, it is possible to define whether synergy or antagonism is detectable for any given drug pair. For drug pairs with less than 5 ε scores in the appropriate expected fit space, no significant p-values ​​were found due to insufficient sample sizes for these. Synergy and antagonism could not be detected for 1% and 25% of all interactions, respectively.

[0084] c) Conserved weak drug-drug interactions within bacterial species For drug pairs showing strong drug-drug interactions in only one of the two strains per species, the criterion for assigning the interaction to the second strain was relaxed to |ε second strain| > 0.06 if the interaction sign was the same. Interactions assigned using this approach are referred to as conserved weak interactions.

[0085] Loewe Additive For combinations between β-lactams where a high-resolution 8×8 checkerboard with sufficient growth inhibition was available in the validation data set, interactions were identified using Loewe additivity. Drug-drug interactions were inferred from the shape of isoboles (lines of equal growth) in two-dimensional drug concentration plots. Unless otherwise stated, all isoboles correspond to 50% growth inhibition (IC50) and were obtained using lines representing isoboles and IC50 interpolated concentrations by fitting to a logistic model. To interpolate IC50 concentrations (or other ICn%), a logistic model was used to fit the growth for each concentration of the first drug across various concentrations of the second drug. The null hypothesis of this model is represented by a line of additivity, where the linear isoboles connect the individual ICs that are equal by the two drugs.

[0086] Sensitivity analysis We confirmed the validity of the main statistical parameters used to assign interactions by performing sensitivity analyses. Several expected fit (fa × fb) cutoffs were tested while keeping other parameters constant (Fig. 8c). The added value of restricting the ε distribution to the appropriate drug concentrations (based on expected fit) was strongly supported by the proportion of strong drug-drug interactions found exclusively using this criterion (about 19% for our chosen cutoff). The chosen cutoff (0.2; ignoring wells with fa × fb below 0.2 for synergy and above 0.8 for antagonism) resulted in the largest number of total interactions assigned, as well as the highest precision (91%) and reproducibility (74%) after benchmarking on the validation dataset (Fig. 8c).

[0087] The suitability of the applied thresholds to define strong (|ε| > 0.1) and weak (|ε| > 0.06) interactions was assessed by their impact on the true and false positive rates (TPR and FPR, respectively, Fig. 8d). A threshold of |ε| > 0.1 is useful because it imposes a minimum strength for assigning interactions. 0.1 corresponds to approximately three times the median of the first and third quartiles of all ε distributions (Fig. 6c). Lowering this threshold reduces the TPR, as some drug pairs are reassigned to neutral due to ambiguity in calling the interaction. Increasing this threshold reduces the TPR, as only very strong interactions are assigned (Fig. 8d). Drug-drug interactions are highly conserved within species, showing the high correlation of ε observed across all species (Fig. 3a, and Fig. 13a and b). This motivated us to initially relax the interaction strength threshold for the second strain if the interaction score |ε| exceeded 0.1, dubbing these interactions weak and conserved. Including the conserved weak interactions in our analysis increased the TPR by 15%. Adding a weak interaction threshold of |ε| > 0.06 (approximately twice the median of the first and third quartiles of all ε distributions) was critical to maintain an adequate FPR (Figure 8d).

[0088] Benchmark and clinical isolate checkerboard assays For validation experiments after screening and to test selected synergy against MDR clinical isolates, an 8x8 checkerboard assay was performed (Figure 5). As in screening, growth was assessed based on OD595 nm in early stationary phase for no-drug controls. The time points used in screening were again used for screening strains, but 8 hours was used for all E. coli and K. pneumoniae MDR isolates. Goodness of fit was calculated for each individual checkerboard by dividing the OD595 nm after single or double drug treatment by that without drug treatment. Bliss scores (ε) were calculated for all concentration ratios tested for each drug mixture, resulting in 49 ε values ​​for each drug pair. Drug mixtures were analyzed based on ε distribution after excluding wells where only one of the drugs and the subsequent combination with the second drug completely inhibited growth. Antagonism was called if the median of the ε distribution was above 0.1 or the Q3 was above 0.15. All experiments were performed with biological duplicates, and interactions were considered valid when duplicates were concordant.

[0089] Similarly, synergy was called if the median of the ε distribution was less than -0.1 or Q1 was less than -0.15. All interactions were then manually inspected, and 10 drug pairs that interacted only over a small concentration range but still had medians and quartiles just below the cutoffs were retrieved and assigned the appropriate drug interaction.

[0090] Evaluation of drug-drug interaction stability Conservation of drug-drug interactions between strains of the same species was assessed by Pearson correlation of the interaction score ε. For potentially non-conserved drug-drug interactions, the expected fitness distribution of the two strains was considered. Drug pairs were considered not comparable between the two strains when the two distributions were significantly different according to the Wilcoxon rank sum test (p-value < 0.05 after BH correction for multiple testing).

[0091] To evaluate the interspecies conservation of drug-drug interactions, the inventors considered only the drug pairs investigated in all three species. A drug-drug interaction was defined as being detected within a species when it was detected in at least one of the two strains and no change in the interaction signature was observed in the other strain. The interactions were then compared across the three species. Cases where the signatures of the drug-drug interactions changed across species (opposite interactions; approximately 7% of all interactions) were excluded from the comparative "interspecies" Venn diagram (Figure 3d). It should be noted that in the current analysis, a given drug-drug interaction may be conserved across species but not within a species.

[0092] Conservation at the single-drug level was defined based on shared resistance and susceptibility. A strain was considered sensitive to a given drug if one of the drug concentrations inhibited growth by at least 30%. According to the conservation of drug-drug interactions across species, a single-drug response is conserved across species when at least one strain of both species has the same signature (sensitivity or resistance).

[0093] Monochromaticity index The monochromaticity index (MI) between drug pairs was defined as follows according to Szappanos et al.

Equation

[0094] Evaluation of drug combinations in a Galleria mellonella infection model To assess the efficacy of drug combinations, last instar larvae of the greater wax moth (Gallium mellonella) were used as an in vivo model. Larvae were purchased from UK Waxworms (Sheffield, UK) and TZ-Terraristik (Cloppenburg, Germany). Stock solutions of vanillin (20% in DMSO), spectinomycin (distilled water), colistin (distilled water), and clarithromycin (20% DMSO / 0.01% glacial acetic acid) were freshly prepared and diluted to the required concentrations in PBS. Drug and bacterial suspensions were administered by injecting 10 μL aliquots into the hemocoel via the last left (drug) and right (antibiotic) prolegs using a Hamilton precision syringe. Controls included both uninfected larvae and larvae injected with the drug's solvent into both last remaining prolegs. Drug toxicity was preliminarily assessed by injection of serial dilutions of either single drugs or drug mixtures, with drugs used at doses that produced little / no toxicity. Similarly, to identify the optimal inoculum, 10 μl of serially diluted bacterial suspensions (1 × 10 2 Colony forming units (CFU) ~ 1 x 10 7 Time-kill curves were generated by inoculating larvae with 10 ...

[0095] Cell viability assay and intracellular antibiotic concentration Ciprofloxacin Overnight cultures of E. coli BW25113 were diluted 1:1000 into 50 ml LB and grown at 37° C. to an OD595 nm of approximately 0.5. Paraquat (50 μg / ml), vanillin (150 μg / ml), benzalkonium (5 μg / ml), caffeine (200 μg / ml), doxycycline (0.5 μg / ml), rifampicin (5 μg / ml), trimethoprim (5 μg / ml), or curcumin (100 μg / ml) were added to the cultures and incubated at 37° C. for 30 min before adding ciprofloxacin to a final concentration of 2.5 μg / ml. Cultures were incubated for 1 h at 37° C. in the presence of both drugs. Cell viability was determined by counting CFUs after 16 h incubation of washed cell pellets plated on drug-free agar Petri dishes. Intracellular ciprofloxacin was quantified using liquid chromatography coupled to tandem mass spectrometry (LC-MS / MS) as previously described. Unwashed cell pellets were directly frozen and dissolved in 350 μl acetonitrile, followed by three freeze-thaw cycles (thawing was performed in an ultrasonic bath for 5 min). Cell debris was pelleted at 16000 g and the supernatant was filtered through a 0.22 μm syringe filter before injection. Chromatographic separation was performed on a Waters BEH C18 column (2.1×50 mm; 1.7 μm) at 40° C. with a 2 min gradient at a flow rate of 0.5 mL / min as follows: (i) 0 min to 0.5 min, 1% mobile phase B; (ii) 0.5 min to 1.2 min, linear gradient of 1% to 95% mobile phase B; (iii) 1.2 min to 1.6 min, 95% mobile phase B; and (iv) 1.6 min to 1.7 min, return to initial conditions (mobile phase A consisted of 0.1% formic acid in water, mobile phase B consisted of 0.1% formic acid in acetonitrile). Samples were kept at 4 °C until analysis. Sample injection volume was 5 μL. Detection of ciprofloxacin was performed on a Waters Q-Tof Premier instrument using electrospray ionization in positive mode. The transition 332>314 was monitored, and the cone voltage was set to 8 and the collision energy was set to 20.At the time of addition of ciprofloxacin, intracellular ciprofloxacin was normalized to CFU.

[0096] Gentamicin Intracellular gentamicin was quantified by measuring [3H]-gentamicin (1 mCi / ml; Hartmann Analytic Corp.) as previously described. An overnight culture of E. coli MG1655 (the parental strain of BW25113) was diluted 1:1000 into 5 ml of LB and grown to an OD595 nm of approximately 0.1. [3H]-gentamicin was diluted in chilled gentamicin to a stock solution of 5 mg / ml (0.1 mCi / ml) and then added to the cultures at a final concentration of 5 μg / ml (0.1 μCi / ml) simultaneously with the second drug berberine (200 μg / ml), erythromycin (15 μg / ml), metformin (13000 μg / ml), procaine (6000 μg / ml), loperamide (400 μg / ml), benzalkonium (5 μg / ml), rifampicin (5 μg / ml), or clindamycin (200 μg / ml). Cultures were then incubated at 37° C. on a rotary shaker. At 0, 0.5, 1, 1.5 and 2 hours, 500 μl aliquots were removed and applied to 0.45 μm pore size HAWP membrane filters (Millipore) pretreated with 1 ml unlabeled gentamicin (250 μg / ml). The filters were washed with 10 ml 1.5% NaCl, placed in counting vials and dried at 52° C. for 30 minutes. 8 ml of liquid scintillation was then added to the dried filters and the vials were incubated overnight at room temperature before counting for 5 minutes. Gentamicin uptake efficiency was expressed as total accumulated ng of gentamicin per 108 cells and is plotted here for the final time point (2 hours) for simplicity. Cell viability was determined by CFU.

[0097] Spectinomycin Intracellular spectinomycin was quantified by measuring [3H]-spectinomycin (1 μCi / mg; Hartmann Analytic). Overnight cultures of E. coli BW25113 were diluted 1:1000 in 1 ml LB with and without vanillin (150 μg / ml) and grown to an OD595 nm of approximately 0.5. 50 μg / ml [3H]-spectinomycin:spectinomycin 1:100 was added and the cultures were incubated for 1 h. Cultures were pelleted, washed twice with PBS containing 50 μg / ml unlabeled spectinomycin, resuspended in 1% SDS, and incubated at 85°C for 20 min. Lysates were mixed with 8 ml of liquid scintillation (Perkin Elmer ULTIMA Gold) and counted for 1 min using a Perkin Elmer Tri-Carb 2800TR. The measured radioactivity was normalized to cell number as determined by OD595 nm.

[0098] RNA isolation, cDNA preparation, and quantitative RT-PCR Overnight cultures of E. coli BW25113 and the marR deletion mutant were diluted 1:2000 in 20 ml LB and grown at 37°C to an OD595 nm of approximately 0.2. Aspirin or vanillin was added to the cultures to final concentrations of 500 μg / ml and 150 μg / ml, respectively (DMSO was added as a control), followed by incubation for 30 min at 37°C with agitation. Cells were harvested and RNA extraction was performed using the RNeasy Protect Bacteria Mini Kit (Qiagen) according to the manufacturer's instructions. cDNA was purified using the SuperScript (商標)III Reverse Transcriptase (Thermo Fisher Scientific) was used to prepare the sequences for qRT-PCR. marA and mdfA expression levels were estimated by quantitative RT-PCR using SYBR™ Green PCR Master Mix according to the manufacturer's instructions (Thermo Fisher Scientific). Primer sequences for marA and recA have been previously described. All experiments were performed in at least triplicate biological samples, and relative expression levels were estimated according to Livak et al. using recA expression as a reference.

[0099] Immunoblot analysis for protein quantification Overnight cultures of E. coli BW25113 and the ΔmarA mutant were diluted 1:1000 in 50 ml LB containing 500 μg / ml aspirin, 150 μg / ml vanillin or DMSO (drug solvent control) and subsequently grown at 37°C with agitation to an OD595 nm of approximately 0.5. Cells were washed in PBS containing the corresponding drug or DMSO and then resuspended to an OD595 nm of 1. Cell pellets were resuspended in Laemmli buffer and heated to 95°C for 3 min followed by immunoblot analysis using anti-AcrA (a-AcrA) polyclonal antiserum (kind gift from KM Pos) at a 1:200000 dilution. Primary antiserum was detected with anti-rabbit HRP (A0545, Sigma) at a 1:5000 dilution. Pixel density of bands was quantified using ImageJ. At least five different biological samples were blotted and summarized by their mean and standard deviation.

[0100] Screening the E. coli Keio knockout collection to identify MoAs of drug interactions The E. coli Keio knockout collection (two independent clones per mutant) was arrayed in 1536 format in LB agar plates using a Rotor HDA (Singer Instruments) as previously described. Growth of each mutant was estimated from colony opacity after 13 h of incubation at 37°C in the absence and presence of vanillin (200 μg / ml), spectinomycin (4 μg / ml), and their combinations. All plates were imaged under controlled lighting conditions (spImager, S&P Robotics) using an 18-megapixel Canon Rebel T3i (Canon). Experiments were performed in triplicate. Fitness for each mutant was calculated by dividing growth in the condition (vanillin, spectinomycin, or both) by growth in LB after correcting for outer-frame plate effects. Bliss scores were calculated for each replicate according to Equation 1 and then averaged.

[0101] result High-throughput profiling of drug mixtures in Gram-negative bacteria Gram-negative bacteria cause some of the most difficult-to-treat infections in humans. We selected three Gammaproteobacterial species, Escherichia coli, Salmonella enterica serogroup Typhimurium, and Pseudomonas aeruginosa, all of which belong to the highest risk groups according to the World Health Organization, to study drug-drug interactions and their conservation among closely related species.

[0102] Because drug responses can vary between strains of the same species, we selected two strains per species (Fig. 1a). We surveyed each strain with up to 79 compounds alone and in pairwise combinations. Drugs consisted of 59% antibiotics from all major drug classes, 23% human target drugs and food additives, most of which had reported antibacterial and / or adjuvant activity, and 18% of other compounds, e.g., proton motive force (PMF) inhibitors or oxidative stress inducers, had known bacterial target or genotoxic effects due to antibiotic activity and / or potential uptake-related potential (Fig. 1a). Overall, we profiled up to 2883 pairwise drug mixtures in each of the six strains.

[0103] Drugs were pre-tested in all strains to select appropriate strain-tailored concentrations for combination screening. We selected three subinhibitory concentrations for each drug: near complete, moderate, and mild / no growth inhibition, corresponding on average to a minimum inhibitory concentration (MIC) of 50%-100%, 25%-50%, and 0%-25%, respectively. Along with no drug controls and single drug controls, we evaluated each drug mix in a 4x4 dose matrix using absorbance as the growth readout and calculated goodness of fit as the growth ratio between drug-treated and non-drug treated cells (Figure 1, Figure 6). All experiments were performed at least twice, on average four times, and all experiments had excellent replicate correlation (average Pearson correlation=0.93; Figure 7a and b).

[0104] We quantified all drug-drug interactions using the Bliss independence model (Figure 1b). In contrast to Loewe additivity, an alternative model for evaluating combination therapy, the Bliss model can be adapted to drugs that are ineffective alone but enhance the activity of other drugs (adjuvants). This feature is particularly relevant for our screens, which investigate intrinsically antibiotic-resistant microorganisms (P. aeruginosa and MDR clinical isolates) as well as human-targeted drugs or food additives that lack antibacterial activity. Consistent with the null hypothesis of the model, the Bliss scores are centered at 0 for all bacterial species (Figure 7c). From all Bliss scores (ε) determined for each combination (4 × 4 dose matrix), we derived a single interaction score ε ranging from -1 to 1. This score ε reflected the first and third quartiles of all Bliss scores within a drug pair. Synergy and antagonism were considered significant if the p-value was less than 0.05 (after Benjamini-Hochberg correction of 10,000 permutations of the Wilcoxon rank-sum test). A strong interaction had an additive effect size requirement of |ε| > 0.1, whereas a weak interaction was able to meet the effect size threshold for one of two strains of the same species but just below it for the other strain (|ε| > 0.06; Fig. 3a).

[0105] Overall, we detected about 19% interactions (strong / weak synergy and antagonism) for E. coli, about 16% for S. typhimurium, and about 11% for P. aeruginosa. This is between a hit rate of over 70% for the limited set of antibiotics tested in E. coli and a hit rate of less than 2% for a larger set of antifungal drugs tested in various fungi. The discrepancies are likely due to: (i) drug selection bias, (ii) single drug concentrations used in previous studies (which can greatly increase false negative and false positive rates), and (iii) different data analysis and parameter settings. For example, we observed that drugs that do not inhibit growth alone were involved in fewer interactions in our screens (Figure 7e). Of the 79 drugs tested herein, 69 had at least one interaction in each strain, with the median interactions for E. coli being 12–13 per drug, for S. typhimurium being 11, and for P. aeruginosa being 5–6 (Fig. 7f).

[0106] Since mixed drugs have not been systematically investigated in various bacteria before, we lacked a truth basis on which to benchmark these datasets. To overcome this limitation, we selected 242 combinations across 6 strains and created a validation set using the higher accuracy 8x8 checkerboard assay (Fig. 8a and b). We used this validation set to evaluate the performance of our interaction identification approach and to benchmark our screen (Fig. 8c and d). Overall, our precision and reproducibility were 91% and 74%, respectively. The slightly lower reproducibility (caused by false negatives) can be partially explained by the broad coverage of the drug concentration range in the validation experiments, which improves our ability to detect interactions (Fig. 9). Moreover, we confirmed 90% (n=46; Fig. 10) of all weak interactions tested, proving the rationale of our interaction identification approach to be true. Indeed, the inclusion of weak interactions in our hits contributes to higher reproducibility (Fig. 8d). For the handful of synergies observed between antibiotics of the same class (β-lactams), we confirmed the interactions using the Loewe additivity model, which is more suitable for assessing interactions between drugs with the same target (Fig. 8e).

[0107] Overall, we have generated a large, high-quality dataset of drug-drug interactions in Gram-negative bacteria, examining 17,050 mixed drugs in a dose-dependent manner. With this rich dataset in hand, we explored general principles that influence drug-drug interactions.

[0108] Antagonism and synergy have distinct selectivities We detected 1354 antagonistic drug-drug interactions and 1230 synergistic drug-drug interactions across the six strains, suggesting that these two interactions occur with similar frequency. However, when correcting for the power to detect both types of interactions, antagonism is about 50% more prevalent than synergy (Fig. 2a). This is because we could detect antagonism only for 75% of the combinations (those drug pairs in which at least one individual drug inhibits growth; Fig. 7d), whereas synergy is detectable in almost all combinations (99%). A higher incidence of antagonism has also been reported for antifungal drugs.

[0109] Surprisingly, antagonism and synergy showed a clear dichotomy in our data. Antagonism occurred almost exclusively between drugs targeting different cellular processes for all bacterial species, whereas synergy was more prevalent for drugs of the same class or targeting the same process (Fig. 2b-e and Fig. 11). Mechanistically, antagonism can be explained by interactions at the drug target level, where two inhibitors help the cell buffer another process that has been perturbed. In bacteria, DNA and protein synthesis inhibitors act in this way (Fig. 2b). Consistent with this being a more widespread phenomenon, genome-wide genetic interaction studies in yeast show that mitigating interactions (antagonism) are enhanced between essential genes (targets of infectious disease drugs) that are part of different functional processes. However, antagonism can also arise from interference effects, e.g., one drug inducing the expression of a pump necessary to remove the second drug from the cell. To investigate to what extent antagonism is induced by modulation of intracellular drug concentrations, we tested 16 antagonistic interactions of various drugs with gentamicin or ciprofloxacin in E. coli (Fig. 12a). Although initially detected at the growth inhibition level, all antagonism remained true at the killing level, with 14 / 16 decreasing intracellular gentamicin or ciprofloxacin concentrations (Fig. 12b). In some cases tested, this was likely because the second drug reduced PMF-activated uptake of gentamicin or increased efflux of ciprofloxacin by inducing expression of AcrAB-TolC, the main efflux pump in enterobacteria, and the antagonism was neutralized in the respective mutant backgrounds (Fig. 12c). Overall, our results suggest that most antagonism is due to modulation of intracellular drug concentrations rather than direct interaction of the main drug target (Figs. 12d and e).

[0110] Unlike antagonistic interactions, synergy often occurred between drugs targeting the same cellular process (Fig. 2b-e and Fig. 11). Indeed, synergy is significantly enhanced within drugs of the same category across all three species, given the approximately 15-fold higher number of possible drug mixtures between drug categories in our dataset (p-values ​​< 10-16, Fisher's exact test). Mechanistically, redundancy can be teased apart by targeting different stages of the same functional process. For example, β-lactams have different affinities for multiple and often overlapping penicillin-binding proteins (PBPs), which likely explains much of the synergy between them (Fig. 2b, Fig. 8e, and Fig. 11a and b).

[0111] Like antagonism, synergy can also occur through modulation of intracellular drug concentrations. Consistent with the general permeabilizing role of membrane-targeted compounds in many organisms, and with drug uptake being a major bottleneck for Gram-negative pathogens, a quarter of all detected synergies involved at least one of the eight membrane-targeted drugs in our screen (Wilcoxon rank-sum test, p-value = 0.044). For example, hydrophobic macrolides have difficulty passing through the negatively charged face of the outer membrane (OM), but can be enhanced by polymyxins, which bind to lipopolysaccharide (LPS) and disrupt the OM. However, membrane-targeted compounds also accounted for about 16% of antagonism, suggesting that disruption of membrane integrity may also reduce intracellular drug concentrations. Benzalkonium consistently reduces the intracellular concentrations of both gentamicin and ciprofloxacin (Figures 12b and c), likely due to interference with their import into the cells. One of the best known and widely used synergies is that between aminoglycosides and β-lactams. Consistent with its use against P. aeruginosa in the clinic, we detected multiple strong synergies between specific members of the two antibiotic classes in P. aeruginosa, with fewer interactions in the other two species (Figure 2b, Figures 11a and b).

[0112] Drug-drug interactions are largely species specific We examined the conservation of drug-drug interactions. Interactions within species were highly correlated (Fig. 3a, and Fig. 13a and b), with conservation ranging from 53% to 76% and differing between species (Fig. 3b). If we ignore non-comparable interactions where the concentration range tested prevents detection of synergy or antagonism for both strains, conservation is actually higher (68% to 87%, and an average of 80%) (Fig. 3b and Fig. 7d). The high conservation of drug-drug interactions within species is consistent with the finding that such interactions are generally insensitive to simple genetic perturbations. Despite the overall high degree of conservation within species, 13% to 32% of interactions were strain specific, the majority of which were neutral in the second strain. Very few mixed drugs were synergistic with one strain and antagonistic with the other (16 interactions), but such strain differences persisted in our validation set.

[0113] Conservation is relatively high within species but very low between species (Fig. 3c and d). The majority of interactions (70%) occurred in one of the three species, with only 5% of drug-drug interactions conserved across all three phylogenetically closely related species. Conservation is much higher at the single drug level in the three species (shared resistance / susceptibility to 73% of drugs), indicating that drug mixtures can confer species specificity to drug action. Such specificity may be useful to generate narrow-spectrum therapies with less collateral damage by exploiting pathogen-specific synergies and common commensal-specific antagonisms.

[0114] We then investigated the conserved drug-drug interaction network at the individual drug level for all three species (Fig. 3e). This not only revealed conserved weaknesses of Gram-negative bacteria, such as the strong synergy between colistin and macrolides, but also revealed that known antibiotic classes often act heterogeneously. For example, known synergy between β-lactams and aminoglycosides was limited to the potent aminoglycosides (amikacin and tobramycin) used in our screen, as well as β-lactams (piperacillin, aztreonam, cefotaxime) that specifically target cell division-related PBPs, which is consistent with previous reports. Furthermore, many human targeted drugs, such as loperamide, verapamil and procaine, show general potentiation effects similar to membrane targeted drugs, suggesting that they may also enhance drug uptake or impair efflux, consistent with previous reports of the role of loperamide in Escherichia coli and verapamil in Mycobacterium tuberculosis.

[0115] And, despite the lack of synergy (Fig. 2a), we found that synergy was significantly more conserved than antagonism (Fig. 3f), because i) synergy is enhanced between drugs of the same category and interactions within functional processes have been shown to be conserved through evolution, ii) membrane-targeted drugs have a general enhancing role for Gram-negative bacteria, facilitating drug passage into the OM, and iii) antagonism often relies on drug uptake / incorporation (Fig. 12), which is controlled by less conserved envelope mechanisms.

[0116] Drug-drug interaction profiles reveal the mechanism of action and chemical properties of drugs Pairwise drug interactions are believed to be mechanism of action (MoA) driven, with drug classes interacting with each other in purely synergistic or antagonistic ways. Because drug members of the same category exhibited different interactions in our conserved drug-drug interaction network (Fig. 3e), we decided to address this more systematically by calculating the monochromaticity index (MI) for all drug category pairs between all bacterial species, regardless of whether the interaction is conserved or not. When the proportion of synergy and antagonism in the interactions between two drug categories is the same as for all interactions, the MI is 0. For highly monochromatic category pairs, the MI approaches 1 and -1 for antagonism and synergy, respectively. The MI is high overall, and especially among well-defined antibiotic classes. Furthermore, many antibiotics, including β-lactams, tetracyclines, and macrolides, exhibit a mixture of antagonism and synergy with other antibiotic classes (Fig. 13c). Although β-lactams show diverse affinities for their multiple PBP targets (potentially explaining the intermixed interactions with other classes), the same is not true for protein synthesis inhibitors with specific targets. In this case, the heterogeneous class action may be due to different chemical properties of the class members and thus different dependencies on uptake and efflux systems. Aggregating the MI by drug category reinforced the view that broader categories show more discordant interactions (Fig. 13d). Interestingly, human targeted drugs were the largest category that showed mostly synergistic effects, supporting the hypothesis that many human targeted drugs can act as adjuvants.

[0117] Among the drugs tested, we selected the flavoring compound vanillin, which clusters with the structurally related acetylsalicylic acid (aspirin). Salicylate and aspirin induce expression of AcrAB-TolC, the main efflux pump in enterobacteria, via binding and inactivation of the transcriptional repressor MarR (Fig. 4a). Consistent with a similar action, vanillin treatment increased AcrA protein levels in E. coli due to overexpression of marA (Fig. 4b and c). Higher AcrA levels with vanillin or aspirin treatment resulted in higher chloramphenicol and ciprofloxacin MICs (Fig. 4d and e). As previously reported for salicylate, vanillin exerts an additive small effect on drug resistance independent of MarR / A, possibly via the MarA homologue Rob (Fig. 4c-e).

[0118] Overall, our data suggest that drug-drug interactions can be used to identify MoAs, although the interactions depend not only on the drug target but also on drug uptake and efflux, which are closely related to the chemical properties of the drugs, which is consistent with studies on the mechanisms of drug-drug interactions or computational prediction of these outcomes.

[0119] Effective drug synergy against MDR clinical isolates To test whether the interactions we detected are relevant for resistant isolates, we selected seven conserved strong synergists, including antibiotics, human targeted drugs, or food additives, and evaluated their efficacy against six MDR and XDR E. coli and K. pneumoniae clinical isolates. All these strains were recovered from infected patients, harbored extended spectrum β-lactamase (ESBL) resistance, and belonged to a well-represented clonal lineage with a wide variety of carbapenemases. One K. pneumoniae strain (929) was also resistant to the last resort antibiotic colistin. All drug pairs acted synergistically in most of the strains tested (Figure 5a and Figure 14). We further tested two of these synergies, colistin-clarithromycin and spectinomycin-vanillin, in an established infection model for assessing antibacterial activity, the greater wax moth ( Galleria mellonella ), and both combinations also acted synergistically in vivo by protecting Galleria mellonella against these MDR strains (Figure 5b).

[0120] The strongest of these synergies is between colistin and various macrolides (Figures 5 and 15). Although other polymyxins are known to facilitate the passage of macrolides into the OM of Gram-negative bacteria, this particular synergy occurs at low colistin concentrations (<0.3 μg / ml) and is active even against colistin-resistant strains (Figure 5, K. pneumoniae 929), implying that macrolides may also potentiate the action of colistin through an as yet unknown mechanism, and that macrolides may resensitize colistin-resistant pathogens to colistin. In addition to antibiotic pairs, combinations of antibiotics with human targeted drugs or food additives were also effective against MDR isolates. For example, the local anesthetic procaine resensitized E. coli and K. pneumoniae MDR isolates to fusidic acid and doxycycline (Figure 14), but did not inhibit bacterial growth alone in our screens.

[0121] Loperamide potentiated both doxycycline and colistin. It is currently unclear how drugs that inhibit ribosomes or PMF7, as well as combinations of macrolides and colistin, can enhance the activity of colistin, which acts primarily on OM. Furthermore, the local anesthetic procaine resensitized E. coli and K. pneumoniae MDR isolates to fusidic acid and doxycycline, but was unable to inhibit bacterial growth alone in our screens. And the commonly used food additive vanillin enhanced the activity of spectinomycin in E. coli MDR isolates. This narrow-spectrum, strong interaction opens up the possibility of repurposing a largely ignored antibiotic. Small amounts of vanillin (65 μg / ml) sensitized largely resistant E. coli to spectinomycin, bringing the MIC from >30 μg / ml to approximately 15 μg / ml. This synergy highlights the importance of investigating the role of food additives in antimicrobial therapy.

[0122] The strong enhancement of spectinomycin activity in MDR E. coli isolates by vanillin was intriguing because vanillin antagonizes many other drugs, including other aminoglycosides. We confirmed that this interaction was specific to spectinomycin and vanillin, but not other aminoglycosides or aspirin, and therefore independent of the effect of vanillin on AcrAB-TolC (Figure 15a-c). We then surveyed a genome-wide E. coli gene knockout library to identify mutants that suppress the vanillin-spectinomycin interaction but do not affect the amikacin (another aminoglycoside)-vanillin interaction. One of the top hits was mdfA, which encodes a Major Facilitator Superfamily (MFS) transporter that effluxes both electrogenic and electroneutral compounds (Figure 15c). Consistent with MdfA being involved in spectinomycin uptake, ΔmdfA cells were more resistant to spectinomycin and insensitive to vanillin (Figure S15d), whereas cells overexpressing mdfA were more sensitive to spectinomycin (Figure S15e, not visible at the MIC levels in Figure S15d), as previously reported, and vanillin further exacerbated this effect (Figure S15d). Addition of vanillin increased intracellular spectinomycin concentrations in an mdfA-dependent manner (Figure S15e). At this point, it is unclear how MdfA, known to efflux compounds outside the cell, promotes spectinomycin uptake inside the cell. However, the presence or absence of mdfA is consistent with species specificity of this interaction, since we detected synergy in E. coli and S. typhimurium isolates from our screen and subsequent MDR E. coli isolates, but not in the phylogenetically more distant P. aeruginosa and K. pneumoniae isolates that lack mdfA (Figure 5a). Moreover, this narrow and strong interaction opens up the possibility of repurposing largely neglected antibiotics.Small amounts of vanillin (65 μg / ml) sensitize mostly resistant E. coli to spectinomycin, reducing the MIC from >30 μg / ml to approximately 15 μg / ml, which is similar to the MIC of spectinomycin for Neisseria gonorrhoeae, and spectinomycin is still used clinically. This synergy highlights the importance of investigating the role of food additives in combination therapy, especially since some of the food additives in combination therapy have modest antibacterial activity.

[0123] Table 1: Drug-drug interactions detected in the screen. Abbreviations: "Anta" = antagonism, "Syn" = synergy, "NA" = data not available. [Table 1-A-1] [Table 1-A-2] [Table 1-A-3] [Table 1-A-4] [Table 1-A-5] [Table 1-A-6] [Table 1-A-7] [Table 1-A-8] [Table 1-B-1] [Table 1-B-2] [Table 1-B-3]

Table 1-B-4

Table 1-B-5

Table 1-B-6

Table 1-B-7

Table 1-B-8

Table 1-B-9

Table 1-C-1

Table 1-C-2

Table 1-C-3

Table 1-C-4

Table 1-C-5

Table 1-C-6

Table 1-C-7

Table 1-C-8

Table 1-C-9

[0124] English notation in tables Mixed soup Signs of interaction Drug 1 Drug 2 Escherichia coli BW25113 E. coli iAi1 Bacitracin Rifampin Benzalkonium Pseudomonic acid Fusidic acid Polymyxin B Loperamide Novobiocin Cefaclor Clarithromycin Minocycline Chlorhexidine erythromycin Mecillinam Levofloxacin Cycloserine D Pyocyanin Colistin Meropenem Metformin Amoxicillin Nitrofurantoin Piperacillin Doxycycline Penicillin G Verapamil Spectinomycin Bleomycin Cerulenin Streptozotocin Aztreonam Phleomycin Vanillin Imipenem Tobramycin Procaine Amikacin paraquat Trimethoprim Fosfomycin Azithromycin Moxifloxacin Acetylsalicylic acid Sulfamonomethoxine Cefsulodin Puromycin Ciprofloxacin Chloramphenicol Gentamicin Mitomycin C Caffeine Oxacillin Spiramycin Clindamycin Curcumin Triclosan Cefotaxime Berberine Flucytosine Cephalexin Linezolid Phenformin Interaction Score p-permutations [Explanation of symbols]

[0125] Drawing translation Figure 1a 79 Drugs Cell wall Human targeted Protein synthesis Membrane RNA polymerase other DNA other cell wall β-lactam human drug food additive food additive Antifungal Macrolide Aminoglycoside Tetracycline Other protein synthesis inhibitors Oxidative stress inducer fatty acid biosynthesis fatty acid biosynthesis other membrane Multiple folic acid biosynthesis folic acid biosynthesis Fluoroquinolone Interaction scores Antagonism Synergy ~3000 drug pairs / strain 62 drugs x 6 strains Escherichia coli Escherichia coli Salmonella Typhimurium Pseudomonas aeruginosa Figure 1b Growth profiling Synergy Drugs 1 & 2 Drug 1 No Drug Drug 2 Antagonism Expected growth Bliss independence model Fitness Expected fitness Drug 1 + 2 4x4 dosage matrix Figure 2a Interactions - as fraction of detectable Antagonism Synergy Figure 2b Drug category Aminoglycoside Oxidative stress other DNA RNA polymerase folic acid biosynthesis folic acid biosynthesis Fluoroquinolone Multiple other cell wall β-lactam other membrane fatty acid biosynthesis fatty acid biosynthesis food additive food additive human drug Other protein synthesis inhibitors Tetracycline Macrolide Antagonism Synergy n=455 interactions n=455 interactions Figure 2c Drug category Interactions Within Category Across Categories Figure 2d Drug general target cell wall cell wall human protein synthesis Oxidative stress membrane Figure 2e Drug general target Interactions Within Category Across categories Figure 3a EC iAi1 interaction score EC iAi1 interaction score EC BW interaction score EC BW interaction score Strong & conserved Strong interactions Weak & conserved Conserved weak interactions Non-conserved interactions Figure 3b nr interactions Number of interactions E. coli Escherichia coli S. Typhimurium P. aeruginosa Non-comparable interactions Non-conserved interactions Weak & conserved Conserved weak interactions Strong & conserved Strong interactions Figure 3c Pearson correlation coefficient Figure 3d E. coli Escherichia coli S. Typhimurium P. aeruginosa Figure 3e Human targeted Membrane & LPS Protein synthesis Cell wall Figure 3f Partially conserved interactions Synergy Antagonism Conserved Interactions Nonconserved interactions Fully conserved interactions Figure 4a 75 Drugs Interactions across all strains Similar drug interaction profile across species Vanillin Acetylsalicylic acid Acetylsalicylic acid mode of action Drug resistance Salicylate Acetylsalicylic acid MarR-inactive MarR-inactive Figure 4b E. coli BW E. coli BW Vanillin Aspirin Protein levels Control Vanillin Figure 4c marA expression (fold change) E. coli BW E. coli BW Control Vanillin Aspirin Figure 4d Chloramphenicol MIC (μg / ml) E. coli BW E. coli BW Control Vanillin Aspirin Figure 4e Ciprofloxacin MIC (μg / ml) Ciprofloxacin MIC (μg / ml) E. coli BW E. coli BW Control Vanillin Aspirin Figure 5a In vitro activity Clarithromycin E. coli Escherichia coli Colistin Vanillin Spectinomycin K. pneumoniae Not tested Figure 5b G. mellonella infections E. coli Escherichia coli K. pneumoniae Drug A: Colistin Drug B: Clarithromycin Drug A: Spectinomycin Drug B: Vanillin Legend % survival survival% Time post infection (h) No drug Drug A Drug B Drug A + Drug B fitness S. typhimurium P. aeruginosa Concentration drug A (μg / ml) Concentration drug B (μg / ml) Figure 6a Experimental and analytical pipeline MIC curves & Selection of drug concentrations 79 drugs (drug conc. (μg / ml)) (drug concentration (μg / ml)) Pairwise drug combinations screen Query drugs (79) No query drug Array drugs (62) No drug wells No array drug Growth curve smoothing & Correction for plate effects Time Double and single fitness calculation Growth double drug Growth single query drug Growth single array drug Single and double fitness replicate correlation Replicate Bliss scores Drug Interaction scores 3 species, 6 strains Density Single interaction scores Restriction to relevant drug concentrations Expected fitness Blind spot synergy Blind spot antagonism Comparison of drug interactions strain within species across species Selection of combinations for MRD clinical isolates Figure 6b Well A9 Well A9 Array drug: spectinomycin 3 μg / ml Growth double drug Growth single query drug 41 plates (query drugs) proxy Array drugs Top 5% growing wells across plates Figure 6c Density Distributions of Qx of Bliss scores (ε) distributions n=2883 combinations n=2883 combinations Figure 7a Pearson correlation Single drug replicate plates Double drug replicate wells per plate Figure 7b Pearson correlation Median growth across all plates n=181 wells n=181 wells Figure 7c Density Bliss scores Figure 7d Bliss interaction score Bliss expected fitness Antagonistic interactions No interaction Synergistic interactions Antagonism blind spot Synergy blind spot Figure 7e Nr interactions per drug Number of interactions per drug Lowest single drug fitness in screen p-val<10-16 p-value<10-16 Figure 7f Density Interactions per drug Strain Number interactions Figure 8a Screen Benchmarking n=17050 drug combinations Figure 8b nr benchmarked combinations Number of benchmarked combinations Figure 8c Recall (%) Reproducibility (%) Precision (%) #interactions Number of interactions Expected fitness cutoff Interactions found with: Interactions found with Both criteria only all expected fitness wells only relevant expected fitness wells Figure 8d True positive rate False positive rate weak interaction cutoff Recall=0.74 Reproducibility=0.74 Precision=0.91 p-value < 0.05 p-value < 0.05 1-sided interactions strong interaction threshold Figure 8e Cefotaxime (μg / ml) Amoxicillin (μg / ml) Cefsulodin (μg / ml) Penicillin G (μg / ml) Penicillin G (μg / ml) Piperacillin (μg / ml) fitness Figure 9a Neutral interactions Strong & conserved Strong interactions Weak & conserved Conserved weak interactions Non-comparable interactions Non-conserved interactions # interactions Number of interactions Figure 9b Amoxicillin + Cefotaxime High throughput screen Bliss scores Expected fitness Density n=36 wells n=36 wells n=30 wells n=30 wells Benchmarking Cefotaxime (μg / ml) Amoxicillin (μg / ml) fitness Replicate n=49 wells n=49 wells Figure 10a Neutral interactions Strong & conserved Strong interactions Weak & conserved Conserved weak interactions Non-comparable interactions Non-conserved interactions Figure 10b Doxycycline + Amikacin High throughput screen Bliss scores Expected fitness Density n=66 wells n=66 wells n=54 wells n=54 wells Benchmarking Amikacin (μg / ml) Doxycycline (μg / ml) fitness Replicate n=25 wells n=25 wells n=37 wells n=37 wells n=49 wells n=49 wells Figure 11a S. Typhimurium Antagonism Synergy Aminoglycoside Oxidative stress RNA polymerase other DNA Multiple Fluoroquinolone folic acid biosynthesis folic acid biosynthesis beta-lactam other cell wall fatty acid biosynthesis fatty acid biosynthesis other membrane food additive food additive human drug Tetracycline Other protein synthesis inhibitors Macrolide n=374 interactions n=374 interactions Figure 11b P. aeruginosa Antagonism Synergy Macrolide Oxidative stress folic acid biosynthesis folic acid biosynthesis RNA polymerase Multiple Fluoroquinolone other DNA beta-lactam other cell wall Antifungal other membrane fatty acid biosynthesis fatty acid biosynthesis food additive food additive human drug Other protein synthesis inhibitors Tetracycline Aminoglycoside Figure 11c S. Typhimurium membrane cell wall cell wall human protein synthesis Oxidative stress Figure 11d P. aeruginosa membrane cell wall cell wall human protein synthesis Oxidative stress Figure 11e S. Typhimurium P. aeruginosa # interactions Number of interactions Within Category Across Categories Figure 12a Drug Antagonist Outer membrane Periplasm Inner membrane Cytoplasm Respiratory chain Porins Efflux pumps Figure 12b Gentamicin Ciprofloxacin (CFU+ / - antagonist) Intracellular concentration ratio (+ / - antagonist) Control Procaine Metformin Benzalkonium Berberine Erythromycin Loperamide Clindamycin Rifampicin Paraquat Trimethoprim Doxycycline Curcumin Vanillin Caffeine Figure 12c Interaction score Procaine Benzalkonium Clindamycin E. coli BW E. coli BW Paraquat Doxyxycline Doxycycline (Correct: Doxycycline) Curcumin Caffeine Figure 12e # Antagonisms Number of antagonisms Gentamicin Ciprofloxacin Aminoglycosides Fluoroquinolones Antibiotic intracellular concentration: Decreased Not decreased Not tested Figure 12d Gentamicin Pyocyanin Benzalkonium Procaine Loperamide Metformin Berberine Rifampicin Aztreonam Aztreonam Clindamycin Clarithromycin Erythromycin Ciprofloxacin Fosfomycin Spectinomycin Minocycline Doxycycline Paraquat Acetylsalicylic acid Curcumin Caffeine Vanillin Trimethoprim Antibiotic intracellular concentration: Decreased Not decreased Not tested Figure 13a Strain comparison ST ST14028s interaction score STLT2 interaction score Strong & conserved Strong interactions Weak & conserved Figure 13b Strain comparison PA PA14 interaction score PAO1 interaction score PAO1 interaction score Strong & conserved Strong interactions Weak & conserved Figure 13c Monochromaticity index (MI) Monochromaticity index (MI) beta-lactam other cell wall other DNA RNA polymerase Fluoroquinolone folic acid biosynthesis folic acid biosynthesis Multiple other membrane fatty acid biosynthesis fatty acid biosynthesis Oxidative stress other protein synthesis Tetracycline Aminoglycoside Macrolide Antifungal food additive food additive human drug Figure 13d β-lactams β-lactams other cell wall Oxidative stress fatty acid biosynthesis fatty acid biosynthesis other membrane food additives food additives human drugs Fluoroquinolone folic acid biosynthesis folic acid biosynthesis Multiple other DNA RNA polymerase Tetracyclines Aminoglycosides Macrolides other protein synthesis Figure 14 Loperamide Erythromycin E. coli Escherichia coli Colistin K. pneumoniae Detected Azithromycin Procaine Fusidic acid Not tested for all species Doxycycline Legend fitness Concentration drug A (μg / ml) Concentration drug B (μg / ml) S. typhimurium P. aeruginosa Figure 15a Spectinomycin MIC (μg / ml) Spectinomycin MIC (μg / ml) E. coli BW E. coli BW Control Vanillin Figure 15b Spectinomycin MIC (μg / ml) Spectinomycin MIC (μg / ml) E. coli BW E. coli BW Control Vanillin Aspirin Figure 15c Interaction score Amikacin + Vanillin Spectinomycin + Vanillin Figure 15d Spectinomycin MIC (μg / ml) Spectinomycin MIC (μg / ml) E. coli BW E. coli BW Control Vanillin Figure 15e growth normalized to no drug log10 Spectinomycin conc. Spectinomycin concentration log10 E. coli BW E. coli BW Fig. 15f Fold change Intracellular spectinomycin normalised to E. coli BW E. coli BW E. coli BW Control Vanillin

Claims

1. 1. A method for identifying a synergistic antibacterial effect of at least two drug compounds against bacteria, comprising: a) providing the bacterium to be tested for its antimicrobial effect; b) selecting said at least two drug compounds; At least one of the drug compounds is known to have an antibacterial effect against the bacteria of a), and at least one other of the drug compounds is (i) an antibiotic, or a pharma- ceutically acceptable salt thereof, known to belong to the same class as the selected first drug compound and / or to target the same bacterial cellular process as the selected first drug compound; and (ii) a human targeted drug, a food additive, or a pharma- ceutical acceptable salt thereof; Selected from: c) identifying a synergistic antibacterial effect of the at least two drug compounds on the bacteria; and d) selecting the at least two drug compounds identified in step c); and A method comprising:

2. 1. A method for identifying antagonistic antibacterial effects of at least two drug compounds against bacteria, comprising: a) providing a first bacterium to be tested for said antimicrobial effect; b) selecting said at least two drug compounds; wherein at least one of the drug compounds is known to have an antibacterial effect against a second bacterium and at least one other of the drug compounds is selected from an antibiotic, a human targeted drug, a food additive, or a pharma- ceutical acceptable salt thereof, preferably the at least one other of the drug compounds is known to belong to a different class than the first selected drug compound and / or to target a different bacterial cellular process than the first selected drug compound; c) identifying the antagonistic antibacterial effects of the at least two drug compounds on the first bacterium of a); and d) selecting the at least two drug compounds identified in step c); and A method comprising:

3. The bacteria are gram-positive or gram-negative bacteria, and examples of the bacteria include Enterobacter, Escherichia, Shigella, Serratia, Proteus, Pseudomonas, Acinetobacter, Staphylococcus, Streptococcus, Pseudomonas, Salmonella, Helicobacter, Citrobacter, Treponema, Mycobacterium, Bordetella, Borrelia, Brucella, Corynebacterium, Fusobacterium, Leptospira, Listeria, Pasteurella, Rickettsia, Faecalis, Egasella, Lactonifactory, coliform bacteria, Bacillus, Francisella, Acinetobacter, Legionella, Actinobacillus, Coxiella, Bifidobacterium 3. The method of claim 1 or 2, wherein the bacterium is a member of the genera Bacterium, Mobiluncus, Enterococcus, Actinomyces, Neisseria, Chlamydia, Vibrio, Diplococcus, Lactobacillus, Kingella, Yersinia, Klebsiella, Bacteroides, Eubacterium, Alistipes, Ruminococcus, Roseburia, Parabacteroides, Prevotella, Coprococcus, Dorea, Blautia, Odoribacter, Clostridium, Collinsella, Bylophila, Akkermansia, Veillonella, Haemophilus, Desulfovibrio, Butyrivibrio and / or Campylobacter, and optionally the bacterium is an antibiotic resistant bacteria, in particular a multidrug resistant strain thereof.

4. 1. A method for developing a targeted therapy for use in preventing and / or treating a bacterial infection, comprising: a) carrying out the method according to any one of claims 1 to 3; b) identifying the selective antibacterial effect of the combination of at least two drug compounds against bacteria; c) selecting the at least two drug compounds identified in step b); and A method comprising:

5. 1. A method for producing an antibacterial pharmaceutical composition, comprising: a) carrying out the method according to any one of claims 1 to 4; b) formulating the selected combination into an antibacterial pharmaceutical composition; A method comprising:

6. 6. An antibacterial pharmaceutical composition produced by the method according to claim 5.

7. 7. A method for preventing adverse effects on the gut microbiome using the antibacterial pharmaceutical composition of claim 6, wherein at least one of the components of the composition has an antibacterial effect against at least a first bacterial species and a second bacterial species, and at least one other component interferes with the antibacterial effect against the first bacterial species.

8. A method for preventing the emergence and / or spread of antibacterial resistance using the antibacterial pharmaceutical composition according to claim 6.

9. An antibacterial pharmaceutical composition comprising: (i) vanillin or a pharma- ceutically acceptable salt thereof, or a vanillin derivative or a pharma- ceutically acceptable salt thereof; (ii) at least one antibacterial compound or a pharma- ceutically acceptable salt thereof; 1. An antibacterial pharmaceutical composition comprising:

10. The at least one antibacterial compound or a pharma- ceutically acceptable salt thereof may be selected from the group consisting of an aminoglycoside, a macrolide, a penicillin, a tetracycline, a lincosamide, a quinolone, a fluoroquinolone, a β-lactam, a polymyxin, a monobactam, a glycylcycline, an ansamycin, a sulfonamide, an oxazolidinone, a carbacephem, a carbapenem, a cephalosporin, a streptogramin, a glycopeptide, a polypeptide, an arsphenamine, a chloramphenicol, a clindamycin, a lincomycin, a daptomycin, a trimethoprim, a novobiocin, an ethambutol, a fosfomycin, a fusidic acid, a furazolidone, an isoniazid, a linezolid, a metronidazole, a mupirocin, a nitrofurantoin, a platensimycin, a pyrazinamide, a quinupristin, a dalfopristin, a rifampin ...

10. The pharmaceutical composition according to claim 9, wherein the aminoglycoside is selected from rifamycins such as ampicin, rifabutin or rifaximin, tinidazole, viomycin and capreomycin, or a pharmaceutically acceptable salt thereof, in particular from streptomycin, dihydrostreptomycin, amikacin, apramycin, arbekacin, astromycin, bekanamycin, dibekacin, framycetin, gentamicin, hygromycin, isepamicin, kanamycin, neomycin, netilmicin, paromomycin, rhodostreptomycin, ribostamycin, sisomicin, spectinomycin, tobramycin, vancomycin and verdamycin, or a pharmaceutically acceptable salt thereof, preferably spectinomycin, or a pharmaceutically acceptable salt of spectinomycin.

11. An antibacterial pharmaceutical composition comprising: (i) a polymyxin selected from colistin and polymyxin B, or a pharma- ceutically acceptable salt thereof, preferably colistin or a pharma- ceutically acceptable salt of colistin; (ii) at least one other pharmaceutical compound, or a pharma- ceutically acceptable salt thereof, wherein the at least one other pharmaceutical compound, or a pharma- ceutically acceptable salt thereof, is selected from the group consisting of loperamide, rifampicin, rifabutin, or rifaximin, macrolides, aminoglycosides, penicillins, tetracyclines, lincosamides, quinolones, fluoroquinolones, β-lactams, polymyxins, monobactams, glycylcyclines, ansamycins, sulfonamides, oxazolidinones, carbacephems, carbapenems, cephalosporins, streptogramins, glycopeptides, polypeptides, arsphenamine, chloramphenicol, clindamycin, lincomycin, daptomycin, trimethoprim, novobio at least one other drug compound selected from the group consisting of rifamycin, ethambutol, fosfomycin, fusidic acid, furazolidone, isoniazid, linezolid, metronidazole, mupirocin, nitrofurantoin, platensimycin, pyrazinamide, quinupristin, dalfopristin, rifampin, tinidazole, viomycin and capreomycin, or a pharma- ceutically acceptable salt thereof, preferably wherein said at least one other drug compound is a rifamycin, such as loperamide, rifampicin, rifabutin or rifaximin, or a macrolide, such as erythromycin, azithromycin, clarithromycin or roxithromycin, or a pharma-ceutically acceptable salt thereof; 1. An antibacterial pharmaceutical composition comprising:

12. An antibacterial pharmaceutical composition comprising: (i) Aminoglycosides, macrolides, penicillins, tetracyclines, lincosamides, quinolones, fluoroquinolones, β-lactams, polymyxins, monobactams, glycylcyclines, ansamycins, sulfonamides, oxazolidinones, carbacephems, carbapenems, cephalosporins, streptogramins, glycopeptides, polypeptides, arsphenamines, chloramphenicol, clindamycin, lincomycin, daptomycin, trimethoprim, an antibacterial compound selected from prim, novobiocin, ethambutol, fosfomycin, fusidic acid, furazolidone, isoniazid, linezolid, metronidazole, mupirocin, nitrofurantoin, rifamycin, such as rifampicin, rifabutin or rifaximin, platensimycin, pyrazinamide, quinupristin, dalfopristin, rifampin, tinidazole, viomycin and capreomycin, or a pharma- ceutically acceptable salt thereof; (ii) at least one other pharmaceutical compound selected from antibiotics, human targeted drugs, food additives, or pharma- ceutically acceptable salts thereof, preferably, said at least one other pharmaceutical compound is procaine, metformin, benzalkonium, berberine, erythromycin, clarithromycin, aztreonam, loperamide, pyocyanin, phenazine methosulfate, clindamycin, rifamycin such as rifampicin, rifabutin or rifaximin, paraquat, trimethoprim, doxycycline, curcumin, vanillin, caffeine, acetylsalicylic acid, epigallocatechin gallate, CHIR090, minocycline, spectinomycin, and fosfomycin, or a pharma-ceutically acceptable salt thereof; Including, Optionally, the antibacterial compound (i) has an antibacterial effect against at least a first bacterium and a second bacterium, and the compound (ii) antagonizes the antibacterial effect of the compound (i) against at least one of the first bacteria, and the compound (ii) does not antagonize the antibacterial effect of the compound (i) against the second bacterium.

13. 13. The pharmaceutical composition according to any one of claims 6 or 9 to 12, wherein the composition is for use in the prevention and / or treatment of a bacterial infection selected from gastrointestinal infections, genitourinary tract infections, upper and lower respiratory tract infections, rhinitis, tonsillitis, pharyngitis, bronchitis, pneumonia, visceral infections, nephritis, hepatitis, peritonitis, endocarditis, meningitis, osteomyelitis, eye infections, ear infections, skin infections, subcutaneous infections, post-burn infections, diarrhea, colitis, pseudomembranous colitis, skin disorders, toxic shock syndrome, bacteremia, septicemia, pelvic inflammatory disease, central nervous system infections, wound infections, intraperitoneal infections, intravascular infections, bone infections, joint infections, acute bacterial otitis media, pyelonephritis, deep abscesses and tuberculosis.

14. 14. The pharmaceutical composition for use according to claim 13, wherein the bacterial infection is caused by a Gram-negative bacterium, preferably the Gram-negative bacterium is a member of the family Enterobacteriaceae or Moraxellaceae, for example a Gammaproteobacteria such as a member of the genera Enterobacter, Escherichia, Salmonella, Klebsiella, Yersinia, Shigella, Serratia, Proteus, Pseudomonas and / or Acinetobacter, optionally the Gram-negative bacterium is an antibiotic-resistant bacterium, in particular a multidrug-resistant strain thereof.

15. 15. A pharmaceutical composition for use according to any one of claims 6 or 9 to 14, wherein the components of the pharmaceutical composition are administered simultaneously, separately or sequentially to a subject, said subject being a mammal such as a human, preferably a human patient, and optionally the composition is in liquid, dry or semi-solid form, such as in the form of tablets, coated tablets, effervescent tablets, capsules, powders, granules, dragees, troches, pills, ampoules, drops, suppositories, emulsions, ointments, gels, tinctures, pastes, creams, moist compresses, mouthwashes, plant juices, nasal drops, inhalation mixtures, aerosols, mouthwashes, mouth sprays, nasal sprays or room sprays.