Towards Translation of PqsR Inverse Agonists: From In Vitro Efficacy Optimization to In Vivo Proof-of-Principle

RESEARCH ARTICLE www.advancedscience.com Towards Translation of PqsR Inverse Agonists: From In Vitro Efficacy Optimization to In Vivo Proof-of-Principle Mostafa M. Hamed, Ahmed S. Abdelsamie, Katharina Rox, Christian Schütz, Andreas M. Kany, Teresa Röhrig, Stefan Schmelz, Wulf Blankenfeldt, Alejandro Arce-Rodriguez, José Manuel Borrero-de Acuña, Dieter Jahn, Jessica Rademacher, Felix C. Ringshausen, Nina Cramer, Burkhard Tümmler, Anna K. H. Hirsch, Rolf W. Hartmann, and Martin Empting* Pseudomonas aeruginosa (PA) is an opportunistic human pathogen, which is involved in a wide range of dangerous infections. It develops alarming resistances toward antibiotic treatment. Therefore, alternative strategies, which suppress pathogenicity or synergize with antibiotic treatments are in great need to combat these infections more effectively. One promising approach is to disarm the bacteria by interfering with their quorum sensing (QS) system, which regulates the release of various virulence factors as well as biofilm formation. Herein, this work reports the rational design, optimization, and in-depth profiling of a new class of Pseudomonas quinolone signaling receptor (PqsR) inverse agonists. The resulting frontrunner compound features a pyrimidine-based scaffold, high in vitro and in vivo efficacy, favorable pharmacokinetics as well as clean safety pharmacology characteristics, which provide the basis for potential pulmonary as well as systemic routes of administration. An X-ray crystal structure in complex with PqsR facilitated further structure-guided lead optimization. The compound demonstrates potent pyocyanin suppression, synergizes with aminoglycoside antibiotic tobramycin against PA biofilms, and is active against a panel of clinical isolates from bronchiectasis patients. Importantly, this in vitro effect translated into in vivo efficacy in a neutropenic thigh infection model in mice providing a proof-of-principle for adjunctive treatment scenarios. M. M. Hamed, A. S. Abdelsamie, C. Schütz, A. M. Kany, T. Röhrig, A. K. H. Hirsch, R. W. Hartmann, M. Empting Helmholtz-Institute for Pharmaceutical Research Saarland (HIPS) Helmholtz Centre for Infection Research (HZI) Campus E8.1 66123 Saarbrücken, Germany E-mail: [email protected] The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/advs.202204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. DOI: 10.1002/advs.202204443 1. Introduction Pseudomonas aeruginosa (PA) is a Gramnegative opportunistic bacterium that is responsible for a wide range of acute and chronic infections in humans.[1,2] It is wellknown that the bacteriostatic or bactericidal action of antibiotics intrinsically leads to development of resistance resulting in the emergence of multi- or even pan-resistant PA strains.[3] In order to cope with this problem, there is a pressing need for new nontraditional and innovative therapy options beside the continuous search for “traditional” antibiotics.[4] Also enhancing or rescuing the efficacy of antibiotics in clinical use is of interest, helping to circumvent or reduce resistance.[5] One strategy, which could help to tackle PA infections via new modes-of-action while synergizing with antibiotics, is the so-called pathoblocker approach, which aims to disrupt the virulence of the bacteria without killing them.[5–8] Working either as stand-alone anti-infective agents or via adjunctive treatment together M. M. Hamed, A. S. Abdelsamie, K. Rox, C. Schütz, A. M. Kany, T. Röhrig, A. K. H. Hirsch, R. W. Hartmann, M. Empting German Centre for Infection Research (DZIF) Partner Site Hannover-Braunschweig Saarbrücken 66123 Saarbrücken, Germany A. S. Abdelsamie Department of Chemistry of Natural and Microbial Products Institute of Pharmaceutical and Drug Industries Research National Research Centre El-Buhouth St., Dokki, Cairo 12622, Egypt K. Rox Department of Chemical Biology (CBIO) Helmholtz Centre for Infection Research (HZI) Inhoffenstr. 7 Braunschweig, 38124 Saarbrücken, Germany S. Schmelz, W. Blankenfeldt Department of Structure and Function of Proteins (SFPR) Helmholtz Centre for Infection Research (HZI) Inhoffenstr. 7 Braunschweig, 38124 Saarbrücken, Germany Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (1 of 14) www.advancedsciencenews.com www.advancedscience.com with a backbone antibiotic, this strategy would also help to decrease the development of resistance.[5–8] To be more precise, the pharmacokinetic/pharmacodynamic (PK/PD) relationship of the clinical response to standard-of-care (SoC) aminoglycosides in patients is typically characterized by the ratio of exposure (expressed, e.g., as Cmax) to in vitro efficacy (commonly expressed as the minimal inhibitory concentration, MIC).[9] Hence, boosting the antibacterial effect through potentiating agents has the potential to improve an efficient clinical response. In chronic infections, PA forms hard-to-eradicate biofilms, tremendously impairing antibiotic efficacy.[10] Here, instead of the MIC value, which is measured against planktonic bacteria, the minimal biofilmeradicatingconcentration(MBEC)shouldratherbeconsideredas a critical efficacy parameter.[11] In short, enhancing the antibioticmediated biofilm eradication should directly improve clinical response and, hence, most likely reduce the risk of resistance development. Our strategy to disrupt bacterial virulence is based on the interfere with quorum sensing (QS)—a cell-to-cell communication system, which coordinates the release of the various virulence factors and co-regulates biofilm formation (Figure 1).[2,12–14] A promising target for QS inhibitors (QSI) in PA is Pseudomonas W. Blankenfeldt Institute for Biochemistry Biotechnology and Bioinformatics Technische Universität Braunschweig Braunschweig Germany A. Arce-Rodriguez, J. M. Borrero-de Acuña, D. Jahn Institute of Microbiology Technische Universität Braunschweig 38106 Braunschweig, Germany J. M. Borrero-de Acuña, D. Jahn Braunschweig Integrated Centre of Systems Biology (BRICS) Technische Universität Braunschweig 38106 Braunschweig, Germany J. M. Borrero-de Acuña Departamento de Microbiología Facultad de Biología Universidad de Sevilla Av. de la Reina Mercedes, no. 6, Sevilla CP 41012, Spain J. Rademacher, F. C. Ringshausen Department for Respiratory Medicine Medizinische Hochschule Hannover Carl-Neuberg-Str. 1, 30625 Hannover, Germany J. Rademacher, F. C. Ringshausen, B. Tümmler Biomedical Research in Endstage and Obstructive Lung Disease (BREATH) German Center for Lung Research (DZL) 30625 Hannover, Germany F. C. Ringshausen European Reference Network on Rare and Complex Respiratory Diseases (ERNLUNG) Frankfurt Germany N. Cramer, B. Tümmler Department for Pediatric Pneumology Allergology and Neonatology Medizinische Hochschule Hannover Carl-Neuberg-Str. 1, 30625 Hannover, Germany A. K. H. Hirsch, R. W. Hartmann, M. Empting Department of Pharmacy Saarland University Campus E8.1 66123 Saarbrücken, Germany quinolone signaling receptor (PqsR), also called multiple virulence factor regulator (MvfR).[15] PqsR is a transcriptional regulator that controls the Pseudomonas quinolone signaling (PQS) system and is considered as one of the master regulators of PA virulence.[16] We and others have demonstrated the utility of PqsR as an attractive antivirulence target.[15,17,18–23] Previously, we have reported the first PqsR antagonists developed from the natural ligand HHQ,[18,19] but also the discovery of a structurally divergent hit scaffold as well as first successful lead-generation efforts.[20,21,23] Herein, we describe the rational hit-to-lead-to-candidate optimization of a new chemical class of potent PqsR antagonists, starting from the previously published optimized hit 1.[23] The frontrunner compound reported in this study potently impairs pyocyanin production in typical lab strains as well as clinical isolates from noncystic fibrosis bronchiectasis patients and enhances the eradication efficacy of SoC antibiotic tobramycin against PA biofilms. The binding mode to PqsR was elucidated by an X-ray cocrystal structure. Furthermore, we include extensive data on biological profiling regarding in vitro and in vivo drug metabolism and pharmacokinetics (DMPK) characterization studies. Notably, we conducted an in vivo proof-of- concept study demonstrating the efficacy of an adjunctive treatment regimen in a neutropenic thigh infection model in mice. 2. Results and Discussion 2.1. Medicinal Chemistry-Driven Optimization Recently, we have reported compound 1(Figure 1 and Table 1) as a hit compound obtained during our fragment-based drug discovery efforts.[23] This compound showed a half maximal inhibitory concentration (IC50) of 2.3 ×10−6m toward PqsR in a heterologous lacZ reporter gene assay in E. coli and served as our starting point in the optimization process.[23] Activation of PqsR by its native agonist PQS leads to the direct and indirect transcriptional modulation of genes responsible for environmental adaptation, virulence factor expression, iron acquisition, redox signaling, antibiotic tolerance, cytotoxicity, and immune modulation/evasion.[16,24] One prominent phenotypic effect of disrupting PqsR-dependent QS is the blockade of pyocyanin production as part of the phenazine biosynthesis (Figure 1).[16] Therefore, all compounds synthesized were tested for the inhibition of this important virulence factor as well as activity in the reporter gene assay. Generally, a reasonable correlation was observed between inverse agonistic activity against PqsR (E. coli reporter-gene assay) and pyocyanin inhibition in PA (Tables 1– 3). Compound optimization was further guided by early in vitro ADME properties (Absorption, Distribution, Metabolism, Excretion),focusingparticularlyonkineticsolubilityandmetabolicstability (against mouse liver microsomes; MLM). As a first step, we evaluated the importance of the amino substituent on the head group. Compound 2lacking the amino group showed similar activity to that of 1, indicating that the amino group does not contribute significantly to binding. The position of the CF3and the type of substituent was previously evaluated during our initial fragment growing efforts.[23] For further optimization, we kept the pyridine head lacking the amino group due to better chemical accessibility, enabling faster design-make- test cycles. For initial structure activity relationship (SAR) consid- Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (2 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Figure 1. Generation of pathoblockers against PA by inhibition of quorum sensing via inverse agonists of the transcriptional regulator PqsR (MvfR). Multiple virulence mechanisms such as phenazine biosynthesis, AQ biosynthesis, antibiotic tolerance, iron uptake, T2SS, and T6SS as well as biofilm formation are under the control of PqsR.[13,14] The medicinal chemistry driven optimization strategy applied in this study is depicted in the inset in the top right corner. Table 1. Pseudomonas quinolone signaling receptor (PqsR) inverse agonistic activity and pyocyanin inhibitory activity as well as kinetic solubility and metabolic stability of compounds 1–6. RIC 50 PqsR [×10−9m] IC50 Pyocyanin [×10−6m] Solubility [×10−6m] MLM t1/2 [min]a) 1 – 2309 107 22207 ±4038 176 ±47.4 >200 54 31376 ±735 19.0 ±1.2 >200 44 4861 ±162 14.4 ±1.6 >200 61 54955 ±4850 71.4 ±27.2 >200 20 6>10000 126 ±10.6 >200 4 Importance of the NH2 group and identification of growth vector. a) Half-life against mouse liver microsomes (MLM). erations we compared the potency based on the E. coli reportergene assay. Replacing the 4-fluoro substituent in the Eastern ring with Cl (3)andthenBr(4) resulted in further improvement in activity. Varying the position of the Br to meta (5)orortho (6)positions was detrimental for activity suggesting that the para position was the most suitable direction for further elongation. Replacing the p-Br in 4with a p-chlorophenyl (7) showed a significant boost in activity, clearly indicating the utility of installing a linear bi-aromatic motif in this position (Table 2). As a next step, we aimed to optimize this biphenyl ring system by introducing different heteroaryls. Firstly, we substituted the middle phenyl ring with different pyridines 8and 9, which resulted in variable Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (3 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Table 2. Pseudomonas quinolone signaling receptor (PqsR) inverse agonistic activity and pyocyanin inhibitory activity as well as kinetic solubility and metabolic stability of compounds 7–13. RIC 50 PqsR [×10−9m] IC50 Pyocyanin [×10−6m] Solubility [×10−6m] MLM t1/2 [min]a) 7101 ±37.3 0.449±0.074 10 >40 8210 ±125 3.3 ±1.3 38.9 50 9179 ±103 2.9 ±3.9 26.6 57 10 435 ±198 35.1 ±27.5 >200 75 11 131 ±9.3 7.2 ±4.6 160.7 >480 12 818 ±436 >20 48.1 74 13 158 ±36.1 3.1 ±0.9 42.3 6 Bi-aryl core optimization. a) Half-life against mouse liver microsomes (MLM). potencies with the pyridine substitution pattern in 9being the most potent. Accordingly, this pyridine core was kept for modification of the terminal phenyl ring (10 and 11). Compound 11 was the most potent and the positions of the nitrogen atoms in this compound seemed to be optimal among the other derivatives 8–11. Although, the biphenyl compound 7was more active than 11, we continued our optimization with 11 due to better solubility and metabolic stability. We then tested the effect of the linker atom by replacing the nitrogen atom with oxygen (12)and sulfur (13). The oxygen linker (12) led to a significant loss in activity, while the sulfur (13) showed slightly lower activity as well as impaired solubility and metabolic stability compared to 11. We used compound 11, with the most favorable nitrogen linker, as a scaffold for further optimization of the Eastern part of the molecule (Table 3). Replacing the Cl with a OMe (14) resulted in slight loss in activity. Adding an extra meta substituent like a Me (15)orCF 3(16) did not show any improvement, while adding asecondCl(17) resulted in a significant boost in activity. Com- Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (4 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Table 3. Pseudomonas quinolone signaling receptor (PqsR) inverse agonistic activity and pyocyanin inhibitory activity as well as kinetic solubility and metabolic stability of compounds 14–25. RIC 50 PqsR [×10−9m] IC50 Pyocyanin [×10−6m] Solubility [×10−6m] MLM t1/2 [min]a) 14 235 ±13 2.9 ±0.8 >200 77 15 154 ±79.8 1.9 ±0.3 112.5 <5 16 176 ±9.8 2.1 ±0.4 100.9 384 17 53.9 ±11.8 0.802 ±0.2 76.2 45 18 96.8 ±12.4 1.5 ±0.2 42.4 2 19 159 ±13.6 4.2 ±1.1 25.6 35 20 39.4 ±2.6 0.234 ±0.057 16.6 61 21 60.5 ±11.4 0.32 ±0.11 10.9 32 (Continued) Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (5 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Table 3. (Continued). RIC 50 PqsR [×10−9m] IC50 Pyocyanin [×10−6m] Solubility [×10−6m] MLM t1/2 [min]a) 22 14.2 ±4.0 0.521 ±0.3 41.7 >85 23 61.4 ±34.9 0.591 ±0.07 16.5 3 24 27.5 ±4.7 0.245 ±0.04 47.2 19 25 43.7 ±12.7 0.523 ±0.05 37.9 >480 Identification of the pyrimidine core. a) Half-life against mouse liver microsomes (MLM). bining the p-OMe and the m-Cl (18) showed comparable activity to monosubstituted p-Cl (11) being around half as active than the dichloro motif (17). An effort to further elongate the molecule with a phenoxy moiety (19) did not show further improvement in activity compared to the chloro-bearing compound 11 while being more potent than the methoxy derivative 14. By looking at the energy-minimized conformation of 19 (Figure 2), we concluded that the favored conformation of this bi-aryl pyridine system was most likely in a tilted and nonplanar one. Clearly, this would not only impact the orientation of the terminal meta substituents but also influence the steric requirements and pi-stacking characteristics of the scaffold. To test whether a planar system was beneficial for activity, we designed and synthesized pyrimidine analogues 20 and 21 of the two potent pyridine derivatives 17 and 19, respectively) so far. By this means, we aimed to remove the steric ortho-effect of the corresponding phenyl hydrogen atoms and, thereby, favor the planar conformation. Indeed, pyrimidines 20 and 21 turned out to be more potent than their dipyridine counterparts 17 and 19. Therefore, this motif was kept constant for further optimization efforts. Replacing the terminal phenoxy group by an isopropoxy substituent (22) was beneficial for potency in the reporter-gene assay in contrast to installing a cyclopropyl (23) in this position hinting at the importance of the oxygen linker for target interaction. Adding an extra Cl besides the para-isopropoxy 24 improved the antivirulence efficacy in PA despite being slightly less active in the reporter-gene assay (cf. 22). Finally, introducing an additional nitrogen in the disubstituted phenyl ring of 24 leading to pyridine-bearing 25 was detrimental for activity. In conclusion, we were able to identify and optimize a novel bi-aryl scaffold resulting in nanomolar activity. The most potent compounds were pyrimidines 20 and 24. However, compound 24 was chosen as our frontrunner as it showed the best compromise in the biological activities together with suitable solubility. 2.2. Structure of the Target–Inhibitor Complex We solved the crystal structure of 24 in complex with PqsR at a resolution of 2.74 Å (Figure 3). Overall, the observed binding mode was in accordance with previously reported QSI bearing the trifluoromethylpyridine headgroup.[20,21,23] Notably, the detected electron density confirmed a more planar conformation of the bi-aryl ring system in 24. Upon binding to PqsR, both rings tilted about 24°as observed by the corresponding dihedral angle (see the inset in Figure 3A). In comparison to the values shown Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (6 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Figure 2. A) Energy-minimized conformation of compound 19 showing the nonplanar biphenyl ring. b) Energy-minimized conformation of compound 21 show the planar biphenyl ring. C) Overlay of compounds 19 and 21. in Figure 2, this would translate to an angle of 156°between both ring systems, which is closer to 180°(full planarity) than 119° (see Figure 3A). Investigating the complex more closely revealed that the meta-chloro substituent of the Eastern ring fills a small pocket near Ile186, while the isopropoxy group protrudes into a tunnel-like cavity framed by Ile186 and Tyr258. We conclude that the planar conformation promotes this space-filling mode- of-interaction, rendering the pyrimidine scaffold beneficial for achieving high affinity. 2.3. In-Depth In Vitro Activity Assessment 2.3.1. Abolishing Alkylquiniolone Biosynthesis Biosynthesis of the alkylquinolone (AQ) autoinducers PQS and HHQ, as well as related derivatives HQNO and 2-AA is mediated by PqsR-dependent pqsA–E operon expression. These individual autoinducers are known to mediate further virulence traits such as persister-cell formation, autolysis, eDNA release, etc.[25] Inhibition of PqsR is expected to have a downregulating effect on the pqsA–E operon leading to decreased concentrations of AQs. In order to test whether compound 24 has an inhibitory Figure 3. X-ray cocrystal structure of 24 in complex with PqsR. A) 3D rendering of the ligand–receptor complex. Protein surface and carbons are shown in cyan, carbons of compound 24 are shown in gray, nitrogen is blue, oxygen is red, fluorine is green, and chlorine is orange. Hydrogens left out for clarity. The inset shows a close-up of the bi-aryl ring system, highlighting the decreased angle of the indicated dihedral (white markings). B) 2D interaction scheme of compound 24. Table 4. IC50 values of compound 24 on the biosynthesis of individual alkylquinolones (AQs) Pseudomonas aeruginosa strains PA14 and RP73. Alkylquinolone IC50 PA14 [×10−9m] IC50 RP73 [×10−9m] PQS 349 160 HHQ 233 453 HQNO 347 230 2-AA 386 213 effect on this operon, two PA strains, PA14 (typical lab strain) and RP73 (cystic-fibrosis clinical isolate),[26] were used to evaluate concentration-dependentautoinducer suppression by compound 24. These strains produce different AQ base levels in vitro (absolute concentrations PA14 <RP73). Incubation with compound 24 showed a strong dose-dependent inhibition with IC50 values between 160 and 453 ×10−9m for individual AQs (Table 4,Figure 4A) with no considerable differences between both strains. 2.3.2. Impact on PA Biofilms As a next step, we investigated the performance of our frontrunner 24 on PA biofilms. We previously reported on the impact of Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (7 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Figure 4. Detailed efficacy assessment of compound 24 in cell-based assay systems. A) Dose-dependent inhibition of the QS autoinducer molecule PQS in PA strains PA14 and RP73. B) Dose-dependent reduction of PA14 eDNA release as determined via propidium iodine staining. IC50 0.346 ×10−6 m (95%CI: 0.173–0.630). Mean ±SEM of at least three independent measurements are depicted. C) Combination of 24 and tobramycin (TOB) in the Calgary biofilm device. Biofilms were grown for 24 h in presence of 24/DMSO, followed by treatment of indicated samples with TOB for another 24 h. Viable cells were determined after 48 h via colony-forming units (CFU) count. Means and SD of three independent measurements are depicted (*p< 0.05, ***p<0.001 according to Tukey’s range test). D) Inhibition of pyocyanin production in 14 PA clinical isolates from bronchiectasis patients and the laboratory strains PA14 and PAO1 (dashed lines) by compound 24. Each strain was grown in LB-medium for 16 h in presence of either 2 ×10−6 m24/DMSO or DMSO as control. Subsequently, pyocyanin was chloroform-extracted from the supernatants and spectrophotometrically quantified by its absorbance at 520 nm. Means and SD of at least three independent measurements are depicted (*p<0.05, **p<0.01, ***p<0.001 according to unpaired Student’s ttests). QSIs on the important biofilm-matrix component eDNA (extracellular DNA).[20] In line with these findings, 24 showed dosedependent inhibition of eDNA release in PA14, with a submicromolar IC50 of 0.346 ×10−6m (Figure 4B). Intrigued by the potency of 24 to reduce signal molecule release as well as eDNA in PA14, we investigated whether these effects would result in potentiation of antibiotic-mediated biofilm eradication. Therefore, we employed the Calgary biofilm device,[27] which allows straightforward analysis of biofilm susceptibility toward different agents. Briefly, biofilms were grown on pegs in presence/absence of 24 for 48 h. After that, biofilms were disrupted using sonication, followed by colony-forming units (CFU) determination, which gave access to the amount of viable cells in the biofilm.[27,28] We observed that 24 significantly attenuates PA14 biofilm formation compared to a DMSO control when present in the growth medium from the beginning (Figure S1A, Supporting Information) and 8 h after the onset of biofilm growth (Figure S1B, Supporting Information). In a next step, we investigated the impact of this effect on antibiotic treatment of the biofilms. To this end, we used the SoC aminoglycoside tobramycin as the antibiotic backbone[29] and analyzed its activity in the peg lid model after pretreating biofilms with different concentrations of 24 (Figure 4; see also Figure S1D,E, Supporting Information). Tobramycin was used at a concentration of 0.5 μgmL −1and exerted only minor effects on untreated biofilms (p=0.03). However, biofilms grown in presence of 1 and 10 ×10−6mof24 were much more susceptible to tobramycin treatment than untreated biofilms (p<0.001). Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (8 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License www.advancedsciencenews.com www.advancedscience.com Strikingly, combination treatment caused a >3-fold higher CFU reduction compared to single treatment with the antibiotic. Likewise, CFU reduction by 1 and 10 ×10−6mof24 was significantly higher in combination with tobramycin (p<0.001). It is particularly intriguing that the subefficacious pathoblocker dose of 1 ×10−6m(p=0.93) still benefits tobramycin activity. This could be related to the submicromolar efficacy on eDNA release, which is a key mediator of tobramycin resistance.[29] The 0.1 × 10−6m24 was not sufficient to do so (p=0.21), demonstrating dose-dependency of the observed effect. This effect could further be translated to the aminoglycoside amikacin, whose activity on PAO1 was also boosted in presence of 10 ×10−6m24 (see Figure S1C, Supporting Information). In total, these findings highlight the potential of combination treatment to enhance the efficiency of P. aeruginosa biofilm eradication by conventional antibiotics. 2.3.3. Antivirulence Efficacy Against Clinical Isolates of PA from Bronchiectasis Patients In order to test the efficacy of 24 to inhibit PqsR of other non-CF P. aeruginosa strains, a set of 14 clinical isolates obtained from bronchiectasis patients were treated with the compound and the corresponding inhibition of pyocyanin production was quantified (Figure 4D). The collection contains representative isolates of the major clones in the P. aeruginosa population and pairs of clone B421 isolates from the same respiratory secretion (Table S2, Supporting Information). Moreover, the analyzed 14 P. aeruginosa strains revealed a high degree of variability in important phenotypical and virulence-associated traits including the production of virulence factors (pyocyanin and pyoverdine), biofilm formation, and motility (swimming and swarming) (Figure S2, Supporting Information). The high strain variability observed between distinct clones and within clones from separate and matching habitats allowed for the parallel testing of high or low pyocyanin producers. Overall, treating the 14 clinical isolates with compound 24 caused a dramatic reduction in the production of pyocyanin by all strains, especially clear-cut results were observed for P. aeruginosa strains that naturally produced high amounts of pyocyanin (Bron 8, 27, 46, 59, 76; Figure 4D). In summary, these results confirmed the broad activity of 24 in efficiently targeting the PQS signaling system of not only P. aeruginosa laboratory strains, but also of bronchiectasis-derived clinical isolates. In view of the promising in vitro efficacy data, frontrunner 24 was further profiled in terms of in vitro pharmacokinetics and safety. 2.4. In Vitro ADME Profiling 2.4.1. Metabolism Compound 24 showed acceptable metabolic stability in MLM (Table 3). Hence, it was further profiled using mouse liver S9 fractions (t1/2 50 min) as well as hepatocytes (t1/2 9 min, Table 5). The results prompted us to investigate different routes of administration in order to assess the impact of drug metabolism on systemic compound availability. Table 5. Primary in vitro pharmacokinetic (PK) data for 24. Parameter Value Kinetic solubility [×10−6m] 47.2 Calu-3 (Papp) [10–6 cm s−1]3.1±0.28 Mouse liver S9 t½[min] 50 Mouse hepatocytes t½[min] 9 Plasma binding human [%] >99 Plasma binding murine [%] >99 2.4.2. Plasma Protein Binding Plasma protein binding (PPB) was measured against human and murine plasma and determined to be very high (>99%) in both cases. High plasma protein binding can be favorable in terms of systemic exposure, as it can prolong circulation times. However, very high PPB might lead to disadvantages such as decreased availability of free drug for interaction with its target, unfavorably long elimination rates, and risk of drug accumulation. In order to understand whether the observed parameters are suitable for in vivo efficacy and reasonable clearance we conducted an in vivo PK study (vide infra). 2.4.3. Permeability Across Lung Epithelial Cell Line Calu-3 Transepithelial apparent permeability of compound 24 was assayed in the lung epithelial cell line Calu-3, which were grown at an air–liquid interface until a stable transepithelial electrical resistance (TEER) was reached. Wells with a TEER <300 Ω×cm2 were excluded from analysis. Reference compound ciprofloxacin exhibited an intermediate apparent permeability with 2.5 ×10−6 ±8.9 ×10−7cm ×s−1at 1 ×10−6m, which is in accordance with literature.[30] Compound 24 showed intermediate permeability with an apparent permeability value of 3.1 ×10−6±2.8 ×10−7 cm ×s−1at 1 ×10−6m, which was comparable to ciprofloxacin (Figure S2, Supporting Information). 2.5. In Vitro Safety Pharmacology In order to assess potential safety risks of 24, we subjected it to primary safety-pharmacology assays, including hERG inhibition, CYP inhibition, AhR activation, Mini-AMES, as well as a CEREP- 44 off-target panel, which were conducted at the CROs Cyprotex and Eurofins (Table 6). A functional electrophysiology hERG assay revealed an IC50 value of 15.8 ×10−6m. In the context of the observed hERG activity, the aforementioned high PPB might lead to a suitable safety window for application humans as clinical outcome is correlating well with the margin provided by hERG IC50 and the peak free plasma exposure.[31] With the aim to further explore potential ion-channel inhibition by 24, 13 additional ion channels were investigated at Charles River Laboratories. Functional inhibition was observed only for hKv1.5 (IC50 ≈30 ×10−6m), while for all other targets tested IC50 values were >30 ×10−6m (see Supporting Information). Adv. Sci. 2023,10, 2204443 © 2023 The Authors. Advanced Science published by Wiley-VCH GmbH 2204443 (9 of 14) 21983844, 2023, 5, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/advs.202204443 by Universidad De Sevilla, Wiley Online Library on [14/03/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License