Synthesis of new piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives as inhibitors of Candida albicans multidrug transporters by a Buchwald–Hartwig cross-coupling reaction

Two series of piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives were prepared via a Buchwald–Hartwig cross-coupling reaction and then evaluated for their ability to inhibit the drug efflux activity of CaCdr1p and CaMdr1p transporters of Candida albicans overexpressed in a Saccharomyces cerevisiae strain. In the initial screening of twenty-nine piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives, twenty-three compounds behaved as dual inhibitors of CaCdr1p and CaMdr1p. Only four compounds showed exclusive inhibition of CaCdr1p or CaMdr1p. Further biological investigations were developed and for example, their antifungal potential was evaluated by measuring the growth of control yeast cells (AD1-8u−) and efflux pump-overexpressing cells (AD-CDR1 and AD-MDR1) after exposition to variable concentrations of the tested compounds. The MIC80 values of nineteen compounds ranging from 100 to 901 μM for AD-CDR1 demonstrated that relative resistance index (RI) values were between 8 and 274. In comparison, only seven compounds had RI values superior to 4 in cells overexpressing Mdr1p. These results indicated substrate behavior for nineteen compounds for CaCdr1p and seven compounds for CaMdr1p, as these compounds were transported via MDR transporter overexpressing cells and not by the AD1-8u− cells. Finally, in a combination assay with fluconazole, two compounds (1d and 1f) have shown a synergistic effect (fractional inhibitory concentration index (FICI) values ≤ 0.5) at micromolar concentrations in the AD-MDR1 yeast strain overexpressing CaMdr1p-protein, indicating an excellent potency toward chemosensitization.


Introduction
Transplantation surgery, cancer chemotherapy, and HIV infections have led to a worldwide rise of the immunocompromised population, and hence also of bacterial and fungal opportunistic infections. 1 The fungal genera most oen associated with invasive fungal infections include Candida, Aspergillus, and Cryptococcus, 2 with opportunistic strains of Candida albicans accounting for approximately 50-60% causes of candidiasis, particularly in immunocompromised patients. The treatment of these Candida infections relies heavily on azole antifungal agents, 3 for which widespread and prolonged use has led to the rapid emergence of multidrug resistant (MDR) isolates of C. albicans as well as of non-albicans species. 4 Various mechanisms potentially contributing to the development of MDR have been identied, and the induction of genes encoding drugefflux pumps, like the primary ATP binding cassette (ABC) transporter genes CaCDR1 and CaCDR2 and the secondary major-facilitator superfamily (MFS) transporter gene CaMDR1, has been shown to play a prominent role in the development of resistance to antifungal drugs. [5][6][7] Overexpression of these pump proteins may lead to an increased efflux of drug substrates in MDR clinical isolates. 4,8,9 The potent modulators of multidrug transporter CaCdr1p such as the immunosuppressants cyclosporin, FK520 and FK506, the natural polyphenol curcumin, the quorum-sensing molecule farnesol, the antabuse drug disulram, the antibiotic milbemycin, some synthetic-D-octapeptides, the anti-inammatory drug ibuprofen and the antibacterial unnarmicins have been displayed to prevent drug extrusion and restore fungicidal synergism with the azoles and other drugs. [10][11][12][13][14] Unlike CaCdr1p, there is only a handful number of chemosensitizers in case of CaMdr1p such as verapamil and enniantin B. 15,16 Recently, a further screening from a library of synthetic aromatic compounds sharing a cyclobutene-dione moiety was investigated for the discovery of new inhibitors of MFS and ABC transporters of C. albicans. A few specic inhibitors of MFS transporter CaMdr1p were then identied. 17 Therefore, the search for novel inhibitors able to block the drug extrusion mediated by these efflux proteins represents an attractive approach to reverse MDR.
The pyrrolo [1,2-a]quinoxaline heterocyclic framework constitutes the basis of an important class of compounds possessing interesting biological activities. These compounds have been reported as key intermediates for the assembly of several heterocycles including antipsychotic agent, 18 anti-HIV agent, 19 adenosine A 3 receptor modulator, 20 antiparasitic agents, [21][22][23][24][25] and antitumor agents. [26][27][28][29][30][31] We also previously demonstrated that the pyrrolo [1,2-a]quinoxaline heterocyclic scaffold could lead to the preparation of bacterial multidrug resistance pump inhibitors. 32,33 In this context and as part of a programme on the development of new efflux pump inhibitors (EPIs), we decided to broaden the structural diversity and used the pyrrolo [1,2-a] quinoxaline moiety as a template for the design of new derivatives 1 and 2 in which a piperazine is incorporated in position 1, 4 or 9 of the heterocyclic core in analogy with the EPI pyrimidine and quinoline derivatives I-III, quinine and MS-209 used in the various multidrug resistance therapies (Fig. 1). [34][35][36][37][38] Results and discussion
Compounds 1e, 1g, 1h, 1i, 1m, 1p nally revealed a weak inhibitory activity in both cell lines overexpressing the two types of efflux pumps. Compound 1c and 1f showed exclusive but weak inhibition of CaCdr1p, whereas compounds 1d and 1q demonstrated their exclusive impact on CaMdr1p with efflux inhibition ranged from 29 to 50%. Compounds 1r and 2b are not active on both efflux pumps.
The antifungal potential of all piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives was also evaluated by measuring the growth of control yeast cells (AD1-8u À ) and efflux pumpoverexpressing cells (AD-CDR1 and AD-MDR1) when exposed to variable concentrations of the tested compounds for 48 h. The yeast growth in the absence of inhibitor was considered as 100%. The results were expressed as MIC 80 , the concentration needed to decrease 80% of cells growth (ESI Table S1 †). In case of the control yeast cells (AD1-8u À ), compounds did not reveal any signicant antifungal activity, as demonstrated by their  transported via MDR transporter overexpressing cells and not by the AD1-8u À cells. Interestingly, all the nineteen compounds were observed to inhibit the Nile Red transport from the AD-CDR1 cells and simultaneously behaved as substrate of CaCdr1p (Table 2). Then it could be suggested that the Nile Red and these compounds seem to share the same drug binding pocket of CDR1, undergoing the kinetics of competitive inhibition. By contrast, in the case of AD-MDR1 cells, the route of efflux transport for Nile Red and these compounds did not overlap as only seven compounds showed substrate behavior.
Here the results suggest the presence of an allosteric drug binding pocket for MDR1 and thus following the path of noncompetitive kinetics.
The ability of the compounds to sensitize yeast growth to the antifungal agent uconazole was evaluated by the checkerboard method. 43 In this assay, the control (AD1-8u À ) cells and the CaCdr1p-and CaMdr1p-overexpressing cells were grown in the presence of either uconazole alone or a combination therapy (efflux pump inhibitor plus uconazole). The results, expressed as the fractional inhibitory concentration index (FICI), are summarized in ESI Table S2. † FICI values # 0.5 indicate synergistic interaction between the inhibitor and the substrate. 44 It was observed that the two piperazinyl-pyrrolo[1,2a]quinoxaline derivatives 1d and 1f with FIC ¼ 0.0076 and 0.25, respectively, displayed strong synergistic effects (FICI ¼ 0.15 and 0.4, respectively) when both are combined with uconazole (FIC ¼ 0.15) in the AD-MDR1 yeast strain overexpressing the MFS CaMdr1p, decreasing 129-fold the MIC 80 of the antifungal agent. High FICI values ($1) were found for the remaining compounds (ESI Table S2 †). Similarly, high FICI values were also found in the AD-CDR1.
The effect of the compounds 1d and 1f was examined by confocal imaging of GFP-tagged Cdr1p and Mdr1p, and revealed the non-effect of these compounds on the intactness of the overexpressing strains AD-CDR1 and AD-MDR1 (Fig. 5).
Finally, compounds 1d and 1f were further evaluated for their ability to chemosensitize the azole-resistant clinical isolate (F5) of C. albicans together with the azole-susceptible strain This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 2915-2931 | 2919 (F2). 45,46 As can be observed in Table 3, when combined with uconazole, compounds 1d and 1f (FICI ¼ 0.6 and 0.78, respectively) were able to reduce the effective concentration of uconazole.
About the structure-activity relationships on both series of piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives, the three best dual inhibitors belonging to the series 2 (2c, 2d, 2f) has a common structural feature, namely the presence of a spacer (oxy-propan-2-ol or amino-propan-2-ol) between the tricyclic scaffold and the piperazinyl moiety. On the other hand, exclusive inhibitors of CaCdr1p (1c, 1f) or CaMdr1p (1d, 1q) have the piperazinyl moiety directly linked to the pyrroloquinoxaline. Four of these compounds (2c, 2f, 1d and 2d) have a benzhydryl substituent or related on the piperazine ring. Further pharmacomodulation works will be carried out to extend and deepen our knowledge.

Conclusions
The chemical approach by Buchwald-Hartwig Pd-catalyzed amination of the 1-bromo-4-phenylpyrrolo[1,2-a]quinoxaline 7 or the 4-chloropyrrolo[1,2-a]quinoxalines 5a-e was successfully used to access to new piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives 1a-w. In parallel, six other derivatives containing as well a quinoxaline moiety (compounds 2a-f, previously synthetized 42 ) were added to this study. Then twenty-nine compounds have been selected on their potential to inhibit fungal multidrug resistance pumps as pyrrolo[1,2-a]quinoxaline template was already used and efficient for the inhibition of bacterial efflux pumps. 32,33 Currently, we broadened our horizon to look into the role of these compounds to inhibit the CaCdr1p and CaMdr1p transporters in pathogenic yeast C. albicans. Our study based on the biological assays corroborated with the piperazinyl-pyrrolo[1,2a]quinoxaline derivatives to be the putative and promising modulators of efflux pumps in the pathogenic yeast C. albicans. The results of this work have demonstrated that most of the compounds could inhibit the efflux of Nile Red mediated by both the ABC transporter CaCdr1p and the MFS pump CaMdr1p. Some compounds were able to inhibit specically the efflux of Nile Red without being themselves substrates of the efflux-pump proteins CaMdr1p and CaMdr1p. This assumption was corroborated by the relative resistance index values close to 1 obtained from the cytotoxicity assays, showing that the presence of efflux-pump proteins did not affect the growth and the viability of yeast cells. In S. cerevisiae cells expressing CaCdr1p and CaMdr1p, the greater inhibitory effect on Nile Red efflux was obtained with compounds 2c-f (5 < MIC 80 (mM) < 14) on CaMdr1p. For compounds 2c-e, the best three compounds (MIC 80 ¼ 5-6 mM), the main structural feature is the presence of a bulky group R (e.g. compound 2c with a diphenylmethyl moiety).
In the combination assay with uconazole, the two compounds 1d and 1f have shown a synergistic effect (FICI values # 0.5) at micromolar concentrations in the AD-MDR1 yeast strain overexpressing CaMdr1p-protein, indicating an excellent potency toward chemosensitization. Interestingly, compound 1d showed exclusive and maximum Nile Red efflux inhibition on AD-MDR1 strain and showed excellent chemosensitization in the presence of uconazole, whereas this was not observed with compound 1f. In this context, it is important to mention that each drug/compound may interact differently with different amino acid residues within the binding pocket of CaMdr1p, which could explain the different behavior of Nile Red and uconazole with these compounds. As no synergy has been found in the clinical isolate F5 overexpressing CaMdr1p, a signicant decreasing of the effective concentration of the antifungal agent was also observed, corroborating the results obtained in the AD-MDR1 strain. It is also interesting to note that compound 1d is the unique active compound bearing a benzhydryl moiety in the sub-series 1.
Finally, this study has shown that piperazinyl-pyrrolo[1,2-a] quinoxaline derivatives are able to reverse antifungal resistance, mediated by efflux pumps belonging to both ABC and MFS superfamilies of transporters of the pathogenic yeast C. albicans. Therefore, at non-inhibitory concentrations, these compounds stand as wise candidates chosen to be potential modulators in MDR reversal. For example, compound 1d could offer a new treatment strategy known as combo-therapy in the use of new azole antifungals recently designed. 47,48 Nevertheless further chemical modications will be carried out to synthetize a second generation of piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives designed specically as EPIs of the pathogenic yeast C. albicans. Once again, compound 1d will be investigated for assessing the structural importance of its benzhydryl moiety (e.g. nature and position of additional substituents). By similarity, a pharmacomodulation study around compound 2c will be also managed, using rational drug design tools such as 3D structural characteristics of efflux pumps 49 and recent chemical features 50 to design new piperazinyl-pyrrolo[1,2-a]quinoxaline derivatives as specic EPIs of C. albicans.

Experimental section
Chemistry General information. Commercially reagents were used as received without additional purication. Melting points were determined with an SM-LUX-POL Leitz hot-stage microscope and are uncorrected. NMR spectra were recorded with tetramethylsilane as an internal standard using a Bruker Avance 300 spectrometer. Splitting patterns have been designated as follows: For all compounds 1a-w, all NMR spectra are available in the ESI (Fig. S1-S45 †). Analytical TLC were carried out on 0.25 precoated silica gel plates (POLYGRAM SIL G/UV254) and visualization of compounds aer UV light irradiation. Silica gel 60 (70-230  The reaction mixture was then heated at 100 C during 15 h. Aer cooling, the mixture was diluted with methylene chloride. The reaction mixture was then ltered on Celite and then diluted with water. The organic layer was separated and the aqueous layer was extracted with methylene chloride (20 mL). The organic layers were collected, dried over magnesium sulfate, ltered and evaporated to dryness. Column chromatography of the residue on silica gel using ethyl acetate-methanol (8/2) as eluent gave the nal product 1.       compound or uconazole was observed and the false negatives were ruled out as we compare our experimental data with negative control that is AD1-8u À strain (empty vector strain). 15% glycerol stocks of these strains were maintained in À80 C storage that were freshly revived on YEPD before use. Reagents and media. Nile Red (>98%), curcumin (purity $ 99.5%), and verapamil (purity $ 99%) were obtained from Sigma Chemical Co. Fluconazole (>98%) was obtained from Ranbaxy and [3H]-uconazole (20 Ci mmol À1 ) was provided by Moravek Biochemicals and Radiochemicals. All routine chemicals were obtained from HiMedia and were of analytical grade.
Statistical analysis. Data are the means AE SD from duplicate samples of at least three independent experiments. Differences between the mean values were analyzed by Student's t test (GraphPad QuickCalcs: t, test calculator), and the results were considered as signicant when p < 0.05.
Transport assays. Transport assays were implemented by following the accumulation of Nile Red by ow cytometry with a FACsort ow cytometer (Becton-Dickinson Immunocytometry Systems) in cells overexpressing MDR transporters CaCdr1p (AD-CDR1) or CaMdr1p (AD-MDR1). Briey, the cells with an OD 600 of 0.1 were inoculated, which were allowed to grow at 30 C with shaking, until the OD 600 reached 0.25. The cells were then harvested and resuspended as a 5% cell suspension in diluted medium (containing one part of YEPD and two parts of water). Nile Red was added to a nal concentration of 7 mM, and the cells were incubated at 30 C for 30 min in absence or presence of each inhibitor at a concentration 10-fold higher than substrate (70 mM). The cells were then harvested and 10 000 cells were analyzed in the acquisition. The analysis was performed using the CellQuest soware (Becton Dickinson Immunocytometry Systems). Efflux of 100% was attributed to the cells not exposed to Nile Red and normalized with the efflux mediated via MDR transporters.
Confocal microscopy. Confocal imaging of GFP-tagged Cdr1p and Mdr1p was performed with a Bio-Rad confocal microscope (MRC 1024) with a 100Â oil immersion. The cells were washed and resuspended in an appropriate volume of 50 mM HEPES (pH 7.0). The cells were placed on the glass slides, and a drop of antifade reagent (Fluoroguard highperformance antifade reagent, Bio-Rad, Hercules, CA, USA) was added to prevent photobleaching. 55 Cytotoxicity and fractional inhibitory concentration index (FICI) determination. Yeast cells (10 4 ) were seeded into 96-well plates in either absence or presence of varying concentrations of inhibitors (3-800 mM), and grown for 48 h at 30 C. The optical density of each strain at 600 nm was measured for the cell growth. The growth in the absence of any inhibitor was considered as 100%, and the concentration producing 80% of cell growth inhibition was taken as the MIC 80 value; the resistance index (RI) was calculated as the ratio between the MIC 80 values determined for the strain overexpressing either CaCdr1p (AD-CDR1) or CaMdr1p (AD-MDR1) relative to that of the control strain (AD1-8u À ). The interaction of the respective inhibitors with uconazole was evaluated by the checkerboard method 43 and was expressed as FICI. The ranges of concentrations used were 1.25-65 mM for uconazole, and 3-800 mM for the inhibitors. FICI values were calculated as the sum of the FICs of each agent (uconazole and inhibitors). The FIC of each agent was calculated as the MIC 80 of the agent in combination divided by the MIC 80 of the agent alone.

Conflicts of interest
There are no conicts to declare.