Modifications on C6 and C7 positions of 3-phenylquinolone efflux

Subscriber access provided by EDINBURGH UNIVERSITY LIBRARY | @ http://www.lib.ed.ac.uk

Article

Modifications on C6 and C7 positions of 3-phenylquinolone efflux pump inhibitors led to potent and safe anti-mycobacterial treatment adjuvants Tommaso Felicetti, Diana Machado, Rolando Cannalire, Andrea Astolfi, Serena Massari, Oriana Tabarrini, Giuseppe Manfroni, Maria Letizia Barreca, Violetta Cecchetti, Miguel Viveiros, and Stefano Sabatini ACS Infect. Dis., Just Accepted Manuscript • Publication Date (Web): 25 Mar 2019 Downloaded from http://pubs.acs.org on March 25, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Infectious Diseases

Modifications on C6 and C7 positions of 3-phenylquinolone efflux pump inhibitors led to potent and safe anti-mycobacterial treatment adjuvants Tommaso Felicetti,1,

∥

Diana Machado,2,

∥

Rolando Cannalire,1, * Andrea Astolfi,1 Serena

Massari,1 Oriana Tabarrini,1 Giuseppe Manfroni,1 Maria Letizia Barreca,1 Violetta Cecchetti,1 Miguel Viveiros,2, * and Stefano Sabatini.1 1Department

of Pharmaceutical Sciences, University of Perugia, via del Liceo 1, 06123 Perugia,

Italy. 2Global

Health and Tropical Medicine, GHTM, Instituto de Higiene e Medicina Tropical, IHMT,

Universidade NOVA de Lisboa, UNL, Rua da Junqueira 100, 1349-008 Lisboa, Portugal *(R.C.) E-mail: [email protected]. Phone: +39 075 585 5146. Fax: +39 075 585 5115 *(M.V.) E-mail: [email protected]. Phone: +351 213 652 653. Fax: +351 213 632 105.

Nontuberculous mycobacteria (NTM) are ubiquitous microbes belonging to Mycobacterium genus. Among all NTM pathogens, M. avium is one of the most frequent agents causing pulmonary disease, especially in immunocompromised individuals and cystic fibrosis patients. Recently, we reported the first ad hoc designed M. avium efflux pump inhibitor (EPI) (1b) able to strongly boost clarithromycin (CLA) MIC against different M. avium strains. Since the 3phenylquinolone derivative 1b suffered from toxicity issues towards human macrophages, herein we report a two-pronged medicinal chemistry workflow for identifying new potent and safe NTM EPIs. Initially, we followed a computational approach exploiting our pharmacophore models to screen FDA approved drugs and in-house compounds to identify “ready-to-use” NTM EPIs and/or new scaffolds to be optimized in terms of EPI activity. Although nicardipine 2 was identified as new NTM EPI, all identified molecules still suffered from toxicity issues. Therefore, based on the promising NTM EPI activity of 1b, we undertook the design, synthesis and biological evaluation of new 3-phenylquinolones differently functionalized at the C6/C7 as well as N1 positions. Among the ACS Paragon Plus Environment

ACS Infectious Diseases 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 56

27 synthesized 3-phenylquinolone analogues, compounds 11b, 12b and 16a exerted excellent NTM EPI activity at concentrations below their CC50 on human cells, with derivative 16a being the most promising compound. Interestingly, 16a also showed a good activity in M. avium-infected macrophages both alone as well as in combination with CLA. The antimycobacterial activity observed for 16a only when tested in ex vivo model suggests a high therapeutic potential of EPIs against M. avium, which seems to need functional efflux pumps to establish intracellular infections.

Keywords: Nontuberculous mycobacteria, Mycobacterium avium, efflux pump inhibitors, antimicrobial resistance, 3-phenylquinolones, antibiotic synergy.

ACS Paragon Plus Environment

Page 3 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Nontuberculous mycobacteria (NTM) comprise a large group (>190 species) of environmental microorganisms present in soil and water. They include some opportunistic pathogenic subspecies that can infect airways thereby causing inflammatory lung disease in susceptible individuals, such as patients with chronic pulmonary diseases or immunodeficiency.1 Indeed, some NTM have a high impact on patients with cystic fibrosis (CF) disease with probably an underestimated prevalence around 10%.2 However, the prevalence of these infections is growing due to increased life expectancy of CF patients, improvement in diagnostic tools, and greater awareness of clinicians.3 Among NTM, M. avium and M. intracellulare (grouped in M. avium complex, MAC) and M. abscessus complex (MABSC) are the most common NTM isolated from CF patients. In particular, MAC accounts for nearly half of NTM infections and seems to be the predominant mycobacterial pathogen in the USA, especially targeting adults (>18 years old) with mild CF;4 conversely, MABSC is more frequent in Europe and in younger patients with a worst disease.4 Treatment of infections caused by NTM poses a great clinical challenge, because it is generally long, expensive, and significantly likely to fail.5 The most frustrating aspect of NTM disease management is the poor correlation between in vitro susceptibility testing and the in vivo response to antibiotics.6 The only evidence that supports such correlation is between in vitro macrolides susceptibility and in vivo response against MAC.6 No specific drugs are available while the treatment guidelines for MAC infections, recommended by the American Thoracic Society, indicate a combination of a macrolide, such as clarithromycin (CLA) or azithromycin (AZT), ethambutol and rifamycin lasting at least one year, until culture conversion.6 During the first 2-3 months, injectable amikacin is suggested for in vitro macrolide-resistant MAC. Worthy of note, MAC shows high intrinsic multidrug resistance (MDR) as a consequence of the decreased permeability of the mycobacteria cell wall and the drug efflux mediated by efflux pumps.7 As generally observed in the microbial world, overexpression of the efflux pumps is the first step in the development of high-level target-based drug resistance in mycobacteria.8 M. avium efflux pumps, able to extrude macrolides, give rise to macrolide-resistant phenotypes9 and can be ACS Paragon Plus Environment


Page 4 of 56

mainly associated to two superfamilies: i) ATP-binding cassette (e.g. MAV_3306 and MAV_1695) and ii) Major Facilitator Superfamily (e.g. MAV_1406).8 Efflux pumps cause a decrease in the intracellular antimicrobial concentration, with the subsequent increase of the MIC values.7,8,10 As a consequence, the microorganisms survive in presence of a sub-lethal drug concentration that may allow for point mutations on the macrolides target (i.e. the 23S ribosomal RNA) generating a high level target-based macrolide resistance.8 Due to the crucial role of efflux pumps in the generation of MDR bacterial strains, their inhibition is emerging as a valuable antimicrobial strategy based on the co-administration of the failing antibiotics with an efflux pump inhibitor (EPI) able to restore their initial antibacterial activity.11,12 The efflux pump inhibition offers a great opportunity to fight antimicrobial resistance renewing “old drugs” and preventing the evolutionary pressure on bacteria that evolve resistance only for compounds exerting bactericidal or bacteriostatic effects. Indeed, drug efflux is a primary and unspecific mechanism by which bacteria overcome chemotherapeutic interventions; a strategy able to reduce the insurgence of resistance could thus be more rewarding than both the modification of existing antibiotics and the identification of new antibacterials, which on the long run may be unfruitful as specific antimicrobial resistance could quickly arise. Few NTM EPIs such as phenothiazines and verapamil (VP) have been reported to date and they actually can only be considered pharmaceutical tools rather than molecules with a therapeutic potential.9,13,14 Recently, based on the likely similarity between efflux pumps belonging to the same family, we showed that some 2-phenylquinoline-based Staphylococcus aureus NorA EPIs also restored antibiotic activity against M. smegmatis and M. avium strains (wild-type and resistant).15 Subsequently, through a chemical optimization round, we identified a series of 3-phenylquinolone derivatives having an improved NTM EPI activity. The three hit compounds (1a-c – Figure 1) were able, at ½ and ¼ their MICs, to strongly boost macrolides and fluoroquinolones activity against M. avium strains overexpressing efflux pumps.16 The three compounds showed comparable EPI activity while differing for both the alkylamino chain at quinolone N-1 position and cytotoxicity. ACS Paragon Plus Environment

Page 5 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Among them, the N-ethylpiperidine 3-phenylquinolone analogue 1b (Figure 1) was the most promising. To date, the 3-phenylquinolone derivatives represent the first and sole specifically designed compounds exhibiting potent NTM EPI activity; however, they still suffer from significant toxicity towards human macrophages, thereby highlighting the challenging need to identify new and/or more potent and safer compounds. Herein, we report our efforts toward this aim applying a two-pronged strategy: i) a computational approach exploiting our in-house pharmacophore models in the attempt to identify new chemotypes of NTM EPIs; ii) the optimization of our 3phenylquinolone derivatives through the design, synthesis and biological evaluation of new analogues. Compd

O

OPr

R1

1a

Et N

1b

N

1c

N

Et

N R1

Figure 1. 3-Phenylquinolones 1a-c previously identified as potent NTM EPIs with 1b resulting the most promising.

RESULTS AND DISCUSSION In silico efforts to identify new NTM EPIs Recently we developed two pharmacophore models (named ModB and ModC) that enabled us to successfully repurpose three different approved drugs, including the antihypertensive nicardipine 2 (Figure 2), as inhibitors of the S. aureus efflux pump NorA.17 These common features pharmacophore-models were developed and validated using potent NorA inhibitors belonging to four different chemical families, such as the in-house 2-phenylquinolines. To be noted, in a previous study we demonstrated that 2-phenylquinoline NorA EPIs, cornerstone during the building of the pharmacophore models, were also endowed with good EPI



Page 6 of 56

activity against NTM, in particular against M. smegmatis and M. avium.15 Therefore, this EPI activity of 2-phenylquinolines against different microorganisms (i.e. S. aureus, M. smegmatis and M. avium) suggests the presence of common chemical requirements for both the NorA and NTM efflux pumps inhibition. In order to validate this hypothesis, herein, we performed two different set of experiments. First, we used our 3-phenylquinolone NTM EPIs 1a-c (Figure 1) to verify their matching to the pharmacophoric elements described in ModB and ModC. The pharmacophore mapping of compounds 1a-c on both models showed a good fitting with ModB (fitness values ranging from 1.84 to 1.87) suggesting that these compounds could also act as NorA EPIs. Then, we tested compounds 1a-c by ethidium bromide (EtBr) efflux inhibition assays against S. aureus SA-1199B (norA+/A116E GrlA mutation) to validate their predicted NorA inhibitory activity. All three compounds (1a-c) at 50 µM exhibited EtBr efflux inhibition higher than 70%, which is the threshold value used to distinguish active from inactive compounds in our previous work.17 In parallel, nicardipine (2 – Figure 2), which resulted our best NorA EPI among the repurposed molecules in the previous work,17 was evaluated in restoring CLA and EtBr MICs against M. smegmatis mc2155, a non-pathogenic fast-growing NTM commonly used in laboratory before proceeding on pathogenic strains. Compound 2 at 1/4 MIC (see Table 1 for MIC values) was able to significantly reduce CLA MIC (32-fold) and EtBr MIC (≥128-fold), an activity comparable to that of 3-phenylquinolone derivative 1b, included in the test as reference compound (Table 2). Interestingly, when compared with the two standard reference NTM EPIs, VP and thioridazine (TZ), nicardipine 2 displayed a higher synergism with CLA and EtBr (Table 2). Moreover, in order to demonstrate that the synergistic effect of 2 with CLA and EtBr was due to efflux pump inhibition, EtBr accumulation assays with/without glucose were carried out against M. smegmatis mc2155 in presence of 2 (compound 1b was still included for comparison) (Table 3). Since EtBr accumulation assays depend on the fluorescence existing only when EtBr is in the intracellular compartment, results are evaluated by the relative final fluorescence (RFF) (the higher the RFF ACS Paragon Plus Environment

Page 7 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


value, the greater is the EtBr intracellular accumulation due to efflux inhibition). Interestingly, nicardipine 2 exhibited RFF values comparable to that of 3-phenylquinolone 1b, thus emerging as a new M. smegmatis EPI. Worth noting, the obtained results supported our hypothesis of a common pharmacophore among EPIs directed against different microorganisms and highlighted the ability of ModB and ModC to describe these pharmacophoric elements. With these results as backdrop, we performed a pharmacophore-based virtual screening with the aim to identify new NTM EPIs. Toward this aim, we used two chemical libraries that could provide a different output: i) an in-house built library of FDA approved drugs for drug repurposing purpose and ii) the proprietary collection of compounds designed/synthesized in our group for a scaffold hopping approach. Among the more than 7000 in silico screened compounds, 98 molecules were kept based on their fitting to ModB and ModC. Two different criteria were used in the selection, retaining only molecules having a fitness ≥1.5 for both models or ≥1.8 for only one model. The 98 molecules were classified based on their chemical family and considering chemical diversity, availability and cost, six virtual hits were selected: 2 drugs (mepyramine (3), trimebutine (4)) and 4 proprietary compounds (5,18 6,18 719 and 819) (Figure 2 and Table 1 for fitness values). Thus, compounds 3-8 were evaluated for their intrinsic antimycobacterial activity (Table 1) and then at ¼ MICs in combination with CLA and EtBr against M. smegmatis mc2155 (Table 2). With the exception of trimebutine 4, all tested compounds showed a significant synergistic activity with CLA and/or EtBr. By defining the synergistic effect in terms of modulation factor (MF), namely the antibiotic MIC fold reduction in presence of the EPI, mepyramine 3 exhibited at 64 µg/mL a MF = 4 in combination with CLA and EtBr. Similarly, in house compounds 5-7 at ¼ MIC showed a MF = 4 in combination with CLA. On the other hand, when combined with EtBr, only compounds 5 and 6 retained significant synergistic activity showing MF of 4 and 8, respectively. Interestingly, compound 8 at 64 µg/mL showed a high synergistic effect with CLA exhibiting a MF of 64; unfortunately, this effect was lost when 8 was tested in combination with EtBr. ACS Paragon Plus Environment


Page 8 of 56

Considering the overall promising results, we decided to investigate the cytotoxicity of the best compounds 2, 3, 5, 6, 8 against human macrophages by evaluating their CC50 values (Table 1). Unfortunately, the two drugs 2 and 3 showed very high toxicity at concentrations (CC50 of 6.29 and 56.5 µM, respectively) many fold lower than those needed for EPI activity on M. smegmatis, thereby ruling out our idea of “ready-to-use” compounds as NTM EPIs in therapy. Similarly, the inhouse compounds 5, 6 and 8 exhibited high level of macrophages toxicity (CC50 of 4.17, 4.43 and 4.86, respectively). Therefore, although the pharmacophore models were able to identify new compounds as M. smegmatis EPIs, their intrinsic toxicity towards human macrophages limited their developability within a medicinal chemistry strategy. Design of new 3-phenylquinolones Coming back to the starting goal of identifying more potent and safer NTM EPIs, we shifted our attention towards a classical medicinal chemistry approach to optimize our 3-phenylquinolone derivatives (1a-c), being to date the best NTM EPIs reported in literature. In view of a wider structure-activity relationship study, we moved the alkylamino chains from N1 (derivatives 1a-c) to the oxygen at the C-4 of the quinolone core to afford two explorative Oalkylated derivatives 9a and 9b (Figure 2), thus providing information about the best position for alkylamino functionalization. Of note, these compounds showed good fitness values higher than 1.5 with ModB (1.8 for both derivatives) and ModC (1.6 for both derivatives). Although both derivatives retained good EPI activity against M. smegmatis, they showed greater toxicity against human macrophages in comparison to N-alkylated counterparts 1a and 1b. For this reason, we revisited the N-alkylated derivatives (Tables 1 and 2). Therefore, on the basis of the improvement in potency and safety obtained by C-6 and C-7 functionalization of the 2-phenylquinoline core for our S. aureus EPIs,20,21 we planned for the introduction of different substituents on these positions of the 3-phenylquinolone core. In the design of the new compounds, N,N-diethylaminoethyl (a), ethylpiperidine (b) and ethylazepane (c) were retained at the N-1 position (Figure 2), being the three


Page 9 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


alkylamino chains that provided the starting hit compounds 1a-c. Then, the C-6 and the C-7 positions were explored (Figure 2). In particular, given the excellent results obtained from the – OMe group on the 2-phenylquinoline scaffold to obtain NorA EPIs, we thought to introduce the – OMe group also on the 3-phenylquinolone scaffold both at C-6 (derivatives 10a-c) and C-7 (derivatives 11a-c) positions. Subsequently, the functionalization with either electron-withdrawing or electron-donating differently sized substituents was investigated. Hence, we also designed C-6 halogen 12a-c and 13a-c (fluorine and chlorine, respectively), C-6/C-7 dichloro 14a-c, C-6 trifluoromethyl 15a-c, -iso-propyl 16a-c and nitrile 17a-c derivatives (Figure 2).



A

O

MeO

N+

O-

O

N

Me

MeO

HN

O Me

Me N Me

Page 10 of 56

MeO

N

N Me O

MeO

Me OMe

N Me

O

Me

O

3 (mepyramine)

2 (nicardipine)

4 (trimebutine) F

O N

H N

N N

O S CF3

O F

N O

OH 7

NH O

N

O

N H

N

OH 8

R' C-6 and C-7 positions preferred for improving the biological profile of our 2-phenylquinoline S. aureus EPIs

O

6

S O

N

S

S

5

B

Me N Me

H N

N N

O O

OR4

N OPr

9a

N Et

9b

N

Et

OR4

R6 = MeO H F Cl Cl CF3 iPr CN R 7 = H MeO H H Cl H H H 10 11 12 13 14 15 16 17

OPr

O

N

R6 R7

Best N-1 alkylamino chains for 3-phenylquinolones

OPr

N R1 N Et

a

Et

N N

c

b

Figure 2. Two-pronged approach herein applied: A) in silico identified FDA drugs 2-4 (top) and in-house compounds 5-8 (bottom) tested as new potential NTM EPIs; B) design of new compounds 9a, 9b, 10a-c, 11a-c, 12a-c, 13a-c, 14a-c, 15a-c, 16a-c and 17a-c.


Page 11 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Chemistry Compounds 9a, 9b, 10a-c, 11a-c, 12a-c, 13a-c, 14a-c, 15a-c, 16a-c and 17a-c were synthesized following the procedure depicted in Scheme 1. Starting from the aldehyde derivative 18, obtained as previously reported by us,16 the reaction with properly functionalized anilines gave the enamine derivatives 19-26 that, owing to their poor stability, were immediately cyclized in pre-heated 240 °C Dowtherm A to give the key synthones 27-34. By a similar procedure, the unsubstituted 3phenylquinolone 35, previously reported by us,16 was obtained and next reacted with POCl3 to give the chloro derivative 36 that after reaction with alkylamino alcohol chains in presence of NaH afforded the desired O-alkylated analogues 9a and 9b. 3-Phenylquinolones 27-34 were functionalized at N-1 position by reaction with the three selected chloroalkylamines using K2CO3 as base in dry DMF at 80 °C to give compounds 10a-c, 11a-c, 12a-c, 13a-c, 14a-c, 15a-c, 16a-c and 17a-c (structures in Figure 2). Scheme 1a Cl

OPr

OR

OPr

iv N 9a and 9b

N 36 iii OPr OPr EtO2C CHO

i

EtO2C

ii

R

R7

HN

18

O

R7 R6 19 R6 = -OMe; R7 = -H 20 R6 = -H; R7 = -OMe 21 R6 = -F; R7 = -H 22 R6 = -Cl; R7 = -H 23 R6 = R7 = -Cl 24 R6 = -CF3; R7 = -H 25 R6 = -iPr; R7 = -H 26 R6 = -CN; R7 = -H

OPr

O R

6

OPr

6

v N H

27 R6 = -OMe; R7 = -H 28 R6 = -H; R7 = -OMe 29 R6 = -F; R7 = -H 30 R6 = -Cl; R7 = -H 31 R6 = R7 = -Cl 32 R6 = -CF3; R7 = -H 33 R6 = -iPr; R7 = -H 34 R6 = -CN; R7 = -H 35 R6 = R7 = -H

aReagents

R7

N R' 10a-c 11a-c 12a-c 13a-c 14a-c 15a-c 16a-c 17a-c

and conditions: i) properly functionalized anilines, EtOH, 90 min-3 h, 70-80 °C; ii) Dowtherm A, 15 min-4 h, 250 °C, 17-56%; iii) POCl3, reflux, 2 h, 100%; iv) alkylamino alcohol, NaH, dry DMF, rt, 3h, 14-55%; v) chloroalkylamines, K2CO3, dry DMF, 80 °C, 90 min-4 h, 10-55%. ACS Paragon Plus Environment


Page 12 of 56

Biological evaluation Cytotoxicity determination on human macrophages Considering that the decrease in cytotoxicity of new derivatives was one of the main goals and that mycobacteria are intracellular pathogens, CC50 values against human monocyte-derived macrophages (3 days exposure), which are the preferred host cells for M. avium, were firstly assessed for compounds 10b, 10c, 11a, 11b, 12b, 12c, 13a-c, 14a-c, 15a-c, 16a-c and 17a-c (Table 1). Table 1. CC50 values against human macrophages and MIC evaluation against M. smegmatis and M. avium strains for compounds 2-8, 9a, 9b, 10b, 10c, 11a, 11b, 12b, 12c, 13a-c, 14a-c, 15a-c, 16a-c, 17a-c and starting hit 1b. Fitness values for compounds 2-8 on pharmacophore models ModB and ModC. MIC (µg/mL)b Fitness values CC50 M. avium Compd. M. (µM)a smegmatis 104 104CLA3 104CLA4 ModB ModC 2 6.29 >256 -c -c -c 1.8 NFd 3 56.5 256 -c -c -c 2.2 1.2 4 -c >256 -c -c -c 1.9 1.1 5 4.17 >256 -c -c -c NFd 1.8 6 4.43 128 -c -c -c NFd 1.9 7 -c 256 -c -c -c 1.5 1.8 8 4.86 256 -c -c -c 1.6 1.6 9a 9.02 128 -c -c -c 1.8 1.6 9b 13.73 128 -c -c -c 1.8 1.6 10b 26.12 128 -c -c -c -c -c c c c c 10c 18.80 64 -c c 11a 30.37 256 256 256 256 -c c 11b 25.54 128 128 128 128 -c c 12b 49.02 64 128 256 256 -c c c c c 12c 25.82 64 -c c c c c 13a 20.24 64 -c c c c c 13b 14.91 32 -c c 13c 12.41 32 >256 >256 >256 -c c 14a 169.90 >256 >256 >256 >256 -c 14b 106.20 >256 -c -c -c -c -c 14c 37.48 256 -c -c -c -c -c 15a 13.39 256 -c -c -c -c -c 15b 18.72 256 -c -c -c -c -c c c c c 15c 13.39 256 -c c 16a 56.32 32 128 128 128 -c c 16b 12.57 32 128 128 64 -c c 16c 8.22 16 256 128 128 -c c c c c 17a 11.42 256 -c c c c c 17b 14.94 64 -c c c c c 17c 4.49 128 -c d 1b 82.25 128 256 256 256 1.8 NF aCC b 50 = 50% cytotoxic effect of the compounds on human monocyte-derived macrophages MIC = minimum inhibitory concentration determined by performing at least two biological replicates and the final value is given as the result of two concordant values. c- = Not determined. dNF = compounds did not fit the corresponding pharmacophore model.


Page 13 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Derivatives 10a, 11c, and 12a were excluded from this and additional assays due to solubility issue. In this experiment as well as in the further biological assays, we included as internal reference compound the starting hit 1b, i.e. the derivative showing the most favourable balance between EPI activity/cytotoxicity within the first series.16 Regardless of the alkylamino chains, many compounds (10b, 10c, 11a, 11b, 12b, 12c, 13a-c, 14c, 15a-c, 16a-c, 17a-c and, of course, 9a and 9b) showed lower CC50 values than starting hit 1b. On the other hand, dichloro derivatives 14a and 14b were the less cytotoxic. Preliminary EPI activity assessment against M. smegmatis To define a safety profile, MIC and synergistic activity at ¼ MIC with CLA, CPX and EtBr were evaluated for all compounds against M. smegmatis. We performed the initial screening against M. smegmatis, which is a widely employed fast-growing non-pathogenic mycobacterium offering advantages on safety and data acquisition; indeed, it provides reliable data to quickly select compounds to be tested against the slow-growing opportunistic pathogenic M. avium. All 3-phenylquinolone analogues exhibited low antimycobacterial activity (MIC ≥32.0 µg/mL), thus resulting suitable EPIs (Table 1). However, when combined at ¼ MIC with antibiotics and EtBr, they showed moderate to good synergistic effects against M. smegmatis (Table 2). To better comprehend the synergistic effect and potency of the compounds, it is important to consider MIC values that influence the concentrations used (¼ MIC) to obtain the corresponding MF values. In particular, iso-propyl derivatives 16b and 16c (¼ MIC) displayed a MF ≥64 with CLA against M. smegmatis in a comparable manner to the starting hit 1b.



Page 14 of 56

Table 2. Synergistic activity of compounds 2-8, 9a, 9b, 10b, 10c, 11a, 11b, 12b, 12c, 13ac, 14a-c, 15a-c, 16a-c, 17a-c, starting hit 1b, and reference compounds VP and TZ in combination with CLA, CPX and EtBr against M. smegmatis mc2155. M. smegmatis Compd. MICa (µg/mL)[MF]b CLA CPX EtBr No EPI 8 0.5 25 2 0.25[32] -b ≤0.195[≥128] b 3 2[4] 0.625[4] -b 4 4[2] 12.5[2] b 5 2[4] 0.625[4] b 6 2[4] 3.125[8] -b 12.5[2] 7 2[4] b 12.5[2] 8 ≤0.125[≥64] 9a 2[4] 0.25[2] 3.125[8] 9b ≤0.125[≥64] 0.5 ≤0.195[≥128] 10b 2[4] 0.25[2] 1.562[16] 10c 2[4] 0.25[2] 3.125[8] 11a 0.25[32] 0.25[2] ≤0.195[≥128] 11b 0.25[32] 0.25[2] 0.39[64] 12b 2[4] ≤0.0156[≥32] 1.562[16] 12c 2[4] 0.25[2] 0.78[32] 13a 2[4] ≤0.0156[≥32] ≤0.195[≥128] 13b 4[2] ≤0.0156[≥32] 1.562[16] 13c 2[4] ≤0.0156[≥32] 0.39[64] 14a 2[4] ≤0.0156[≥32] 0.78[32] 14b 4[2] ≤0.0156[≥32] 0.39[64] 14c 4[2] ≤0.0156[≥32] ≤0.195[≥128] 15a 2[4] 0.5 ≤0.195[≥128] 15b 2[4] 0.5 ≤0.195[≥128] 15c 4[2] 0.5 0.39[64] 16a 2[4] 0.25[2] ≤0.195[≥128] 16b ≤0.125[≥64] 0.25[2] ≤0.195[≥128] 16c ≤0.125[≥64] 0.5 ≤0.195[≥128] 17a 0.5[16] 0.25[2] ≤0.195[≥128] 17b 4[2] 0.25[2] 3.125[8] 17c 0.5[16] 0.5 ≤0.195[≥128] 1b 0.25[32] 0.25[2] ≤0.195[≥128] VP 2[4] 0.25[2] ≤0.39[≥64] TZ ≤0.25[≥32] 0.25[2] ≤0.39[≥64] aMICs of antibiotics and EtBr were determined in presence of EPIs at ¼ MIC. bMF (i.e. modulation factor) as ratio of MIC in absence of EPI on MIC in presence of EPI. b- = Not determined.

Interestingly, 16b and 16c exerted this effect at a concentration of 8 and 4 µg/mL, respectively, that are concentrations about 32-fold lower than that of the starting hit 1b. Conversely, the same degree of CLA MIC potentiation was observed also for compound 9b, 11a and 11b but at concentrations close to that of the starting hit 1b. ACS Paragon Plus Environment

Page 15 of 56 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 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


When considering synergistic activity with CPX against M. smegmatis, regardless of the concentrations, compounds 12b, 13a-c and 14a-c showed always good MFs (≥32), definitely higher than 1b (MF = 2). Shifting attention on synergistic activity with EtBr, all compounds displayed good synergistic activities with MF values ranging from 8 to ≥128, thereby highlighting their capability to inhibit efflux pumps. Table 3. EtBr accumulation assays (RFF values) for compounds 2, 10b, 10c, 11a, 11b, 12b, 12c, 13a-c, 14a-c, 15a-c, 16a-c and 17a-c in comparison to starting hits 1b on M. smegmatis mc2155. M. smegmatis Compd. RFFa w/o glucose w/ glucose 2 1.10±0.06 2.75±0.24 10b 1.56±0.21 4.97±0.00 10c 1.62±0.08 5.14±0.03 11a 0.99±0.27 3.65±0.14 11b 1.33±0.21 5.30±0.20 12b 1.65±0.31 5.78±0.92 12c 1.44±0.37 3.91±1.63 13a 1.25±0.32 3.96±0.75 13b 1.11±0.18 3.70±0.16 13c 1.12±0.12 3.45±0.36 14a 0.68±0.02 1.74±0.09 14b 0.83±0.09 3.46±1.20 14c 1.09±0.05 2.92±1.05 15a 1.43±0.02 2.71±0.11 15b 1.19±0.12 2.71±0.00 15c 0.93±0.05 2.73±0.50 16a 1.23±0.07 2.80±0.19 16b 1.31±0.01 3.65±0.32 16c 1.64±0.12 3.17±0.71 17a 1.18±0.01 3.87±0.04 17b 0.98±0.09 3.42±0.02 17c 1.15±0.12 3.23±0.04 1b 1.23±0.11 3.45±0.34 aRFF values correspond to mean value ± standard deviation (SD) derived from three replicates.

Of note, fluorometric EtBr accumulation assays provided further confirmation of the compound ability to inhibit M. smegmatis efflux pumps showing (at ¼ MIC) RFF values high as that for starting hit 1b (Table 3). Glucose (at 0.4%) was included in the assays as the source of metabolic energy to provide the cells the ideal conditions for efflux. Therefore, compounds able to promote EtBr accumulation regardless of the presence of glucose can be considered more effective EPIs. Interestingly, most of our compounds worked even in presence of glucose thus showing a potent EPI activity.15



Page 16 of 56

EPI activity assessment against M. avium strains By comparing the MF values of the tested compounds in the preliminary evaluation on M. smegmatis and combining the concentrations needed to exert the synergistic effect with their CC50 values, eight compounds (11a, 11b, 12b, 13c, 14a and 16a-c) were advanced to the evaluation against three M. avium strains: 104 and the isogenic CLA resistant strains 104CLA3 and 104CLA4.15 Both resistant strains were obtained through several passages in presence of sub-optimal concentrations of the antibiotic and overexpress the MFS pump MAV_1406 and the ABC pump MAV_1695; in addition, 104CLA4 harbours a mutation on the macrolide target (Table 9). We determined the MICs, the synergistic activity with CLA, CPX and EtBr, and the RFF values in the EtBr accumulation assays against the three strains. As expected, all compounds did not exhibit any intrinsic antimycobacterial activity with MIC values ≥64.0 µg/mL (Table 1). From synergistic assays of 11a, 11b, 12b, 13c, 14a and 16a-c at ¼ MIC in combination with scalar concentrations of CLA, CPX and EtBr against M. avium strains (104, 104CLA3 and 104CLA4), all but one (13c - data not shown) compounds were selected for checkerboard assays with antibiotics and EtBr. Data displayed promising synergistic activities for different 3-phenylquinolone derivatives (Table 4). Focusing attention on the M. avium 104 wild-type strain, four compounds (11b and 16a-c) at low concentrations (1/32 MICs – 4-8 µg/mL) retained a relevant synergistic activity with CLA (MF = 4), thereby lowering CLA MIC down to 2 µg/mL. Very interestingly, 16a was able to reduce CLA MIC at a concentration (4 µg/mL = 8.75 µM) about 6-fold lower than its CC50 against human monocyte-derived macrophages (56.32 µM). The best synergistic activity with CPX was shown from the C-7 –OMe derivative 12b that at 4 µg/mL (1/32 MIC) reduced CPX MIC by 4-fold, which corresponds to a MIC of 4 µg/mL. As a proof that synergistic effect was mainly due to EPI activity, the compounds generally showed significant reduction of EtBr MIC against M. avium 104.


Page 17 of 56 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 40 41 42 43 44 45 46


Table 4. Checkerboard assays of compounds 11a, 11b, 12b, 14a and 16a-c in combination with CLA, CPX and EtBr against M. avium 104, 104CLA3 and 104CLA4 strains. M. avium 104 104CLA3 104CLA4 Compd. MICa (µg/mL)[MF]b CLA CPX EtBr CLA CPX EtBr CLA CPX EtBr No EPI 8 16 25 >512 16 25 >512 16 25 11a 1/4 2[4] 2[8] 6.25[4] 512[≥2] 2[8] 6.25[4] 512[≥2] 2[8] 6.25[4] 11a 1/8 4[2] 4[4] 6.25[4] 512[≥2] 4[4] 12.5[2] 512[≥2] 4[4] 6.25[4] 11a 1/16 4[2] 8[2] 6.25[4] 512[≥2] 4[4] 12.5[2] 512[≥2] 4[4] 12.5[2] 11a 1/32 4[2] 8[2] 12.5[2] 512[≥2] 4[4] 12.5[2] 512[≥2] 4[4] 12.5[2] 11b 1/4 2[4] 4[4] 6.25[4] 512[≥2] 2[8] 12.5[2] 512[≥2] 2[8] 6.25[4] 11b 1/8 2[4] 4[4] 12.5[2] 512[≥2] 8[2] 12.5[2] 512[≥2] 4[4] 6.25[4] 11b 1/16 2[4] 8[2] 12.5[2] 512[≥2] 8[2] 12.5[2] 512[≥2] 4[4] 12.5[2] 11b 1/32 2[4] 8[2] 12.5[2] 512[≥2] 8[2] 12.5[2] 512[≥2] 4[4] 12.5[2] 12b 1/4 0.5[16] 1[16] 0.78[32]

Modifications on C6 and C7 positions of 3-phenylquinolone efflux

Recommend Documents