Theoretical and Experimental Study of New Dihydroorotate Dehydrogenase and Tryparedoxin Peroxidase Inhibitors: One More Step in the Study of Leishmaniasis Infection

In this study, the viability of new dihydroorotate dehydrogenase and tryparedoxin peroxidase inhibitors is reported. In vitro antileishmanial activity was evaluated using a Leishmania (V) panamensis strain, and the cytotoxicity of the compounds was assessed using U-937 cells. The in vivo therapeutic response was evaluated in golden hamsters (Mesocricetus auratus) experimentally infected with L. (V) panamensis and treated with a 1% topical formulation of compounds 4a–f. On the other hand, in silico studies considering the synthesized compounds were also carried out. All of the compounds showed promising in vitro activity, with mean EC 50 effective concentration values ranging from 3.8 µM to 19.3 µM. Likewise, treatment with compounds 4a–f produced improvement in most of the hamsters and cured some; in particular, those treated with compounds 4b, 4c, 4d, and 4f reacted the best. Molecular dynamics (MD) simulations, computational docking, and MM/GBSA studies indicate the promising bioavailability and absorption characteristics of the studied compounds, which are expected to be orally active. In addition, the studied 2-arylquinolines are absorbable at the blood–brain barrier, but not in the gastrointestinal tract. Finally, ADMET properties suggest that these molecules can be safely used as leishmaniasis inhibitors. represent the number and percentages of hamsters, according to the result at the end of the study. a At 3 months after the end of treatment. b n = 5 animals per group. c Via topical (40 mg) once/day/20 days. d MA: meglumine via IL (200 μg) twice/week/3 weeks. cure, 100% healing of the area and complete disappearance of the lesion; improvement, percentage of reduction around the lesion greater than 20%; failure, less than 20% of reduction of the area of lesion, or any increase in the size of the lesion. built for each model, were conned inside a periodic simulation box. The water model TIP3P 34 , with 12.552 molecules, was utilized as solvent. Moreover, Na + and Cl - ions were added, for neutralizing the systems and maintaining an ionic concentration of 0.15 mol·L -1 . Full geometric optimizations of the two molecules were done by density functional theory method by M05-2X 35 -D3 36 , in conjunction with the 6-31G(d,p) basis set. The compounds (4a–f) and avin mononucleotide (FMN) were parametrized using the LigParGen web server and the OPLS-AA/1.14*CM1A(-LBCC) force eld parameters were used for organic ligands 37–39 . The partial charges were determined for each ligand by the restrained electrostatic potential (RESP) model 40 . MD simulations were realized using the modelled CHARMM22 and CHARMM36 force elds 41,42 within the NAMD software 43 . 20,000 steps of conjugate-gradient energy minimization were included for each system, followed by simulations of 5 ns with xed protein backbone atoms, and a gradual release of the backbone over 100,000 ps with restraints from 10 to 0.0 kcal·mol -1 Å -2 . The total duration of simulation was ~70 ns per system. During the molecular dynamic’s simulations, motion equations were integrated with 2 femtoseconds time-steps in the NPT ensemble at a pressure of 1 atm. SHAKE algorithm was applied to all hydrogen atoms, and a 12Å Van der Waals cut-off was set. Temperature was set to 310 K, by the Nosé–Hoover thermostat method with a relaxation time of 1 ps. Pressure was controlled at 1 atm by using a Nosé–Hoover–Langevin piston. Long-range electrostatic forces considered by the particle-mesh Ewald approach. MD runs data was collected every 1 ps. Molecular visualization of the systems and trajectory analysis was done using VMD 44


Introduction
Leishmaniasis is a neglected tropical disease that, for decades, has represented a public health problem worldwide, mainly affecting the most impoverished populations. It is endemic in about 98 countries located in the tropics and sub-tropics. The complexity of its transmission, as well as the lack of effective health policies, are the main obstacles in the control of this disease. It is caused by more than 17 species of the genus Leishmania, transmitted to humans by different species of female sand ies of the genus Phlebotomus (in the Old World) and Lutzomyia (in the New World) [1][2][3] . According to the World Health Organization (WHO), a prevalence of approximately 12 million cases has been estimated in all forms of the infection, and around 350 million people live in areas at risk of contracting the infection [4]. Annually, about 1.5 to 2 million new cases of cutaneous leishmaniasis (CL) and mucosal leishmaniasis (ML) are o cially reported, as well as at least 0.5 million new cases of visceral leishmaniasis (VL) and about 60,000 deaths [4][5][6][7][8][9] .
To date, pentavalent antimonies (Meglumine Antimoniate and Sodium Stibogluconate), together with Pentamidine Isethionate, Miltefosine, and Amphotericin B (AMB) are the drugs endorsed for the treatment of all clinical forms of leishmaniasis. However, these drugs have numerous limitations, such as the appearance of adverse effects; high toxicity in the pancreas, kidney, liver, and heart associated with prolonged therapeutic regimens with high doses; the appearance of strains with decreased sensitivity or even resistance; the high cost of the treatment; and variable e cacy, according to the strain of the parasite 8-12 . Therefore, it is necessary to continue the search for new drugs that are safe, effective, easy to administer, and inexpensive, in order to enable their general use and to contribute to the control of this disease.
Compounds with quinolinic nuclei belong to a group of nitrogenous heterocycles, which stand out for their great diversity of pharmacological functions (e.g., 10,13 ) have highlighted their activity against protozoan parasites, such as Plasmodium falciparum 14,15 , with Antitrypanosomal 5,16,17 and leishmanicidal activities; with respect to the latter, 8aminoquinoline (Sitamaquine) is currently used for the treatment of visceral Leishminiasis 5 . Likewise, 2styrylquinolines have shown activity against strains of Leishmania (V) panamensis 18, 19 ; likewise, formulations of quinoline derivatives have promising activity against strains of L. panamensis and L. braziliensis 20 . In this context, computational tools can offer useful information to interpret the trends and state structure-activity relationships, as well as allowing researchers to fully explore protein-drug interactions. For instance, molecular dynamics (MD) simulations allow us to identify the interactions occurring along the entire potential-energy curve, by applying interatomic potentials or molecular mechanics force elds 21,22 . Computational docking simulation can help to predict the best orientation and conformation of drug ligands, when bound to Leishmania protein targets. The relative binding energies of ligands can be exhaustively determined using MM/GBSA studies 23 . Computer-Aided Prediction of Pharmacokinetic (ADMET) Properties is a modern in silico technique that helps to characterize the bioavailability, oral absorption, clearance, and volume of distribution, as well as the penetration through the blood-brain barrier of the ligands 24 . Under this outline, the objective of this research is to synthesize 2-arylquinoline-type compounds and evaluate their leishmanicidal activity in vitro, in vivo, and in silico, such that they can become new, more effective, safe, and easy-to-administer treatment alternatives for CL and VL.

Chemistry
The reagents and solvents used were obtained commercially from national suppliers, Merck and Sigma Aldrich. To monitor the progress of the reaction, thin-layer chromatography was used on aluminium TLC silica gel sheets (60F 254 , Merck, Darmstadt, Germany), the Nuclear Magnetic Resonance spectra were obtained by NMR-one-dimensional 13  Synthesis of quinolinic derivatives 4a-f. For the chemical synthesis, a solution of quinaldine or 8-hydroxyquinaldine in acetic anhydride was used, the corresponding aromatic aldehyde was added. This solution was brought to re ux for 12 to 24 hours. Upon completion of the reaction, the mixture was allowed to cool to room temperature, then sodium bicarbonate was added. The mixture was extracted with a mixture of petroleum benzine/ethyl acetate. The organic phase was dried over anhydrous sodium sulphate, ltered, and concentrated under reduced pressure. Then, the crude product was puri ed by column chromatography (CC), using petroleum benzine/ethyl acetate with an increasing polarity gradient [17][18][19] as eluent. The chemical structures of the compounds were corroborated by NMR spectroscopic techniques: in one dimension ( 1 H, 13  ), 7.34 (dt, 1H, J = 6.1; 6.8 Hz, Ar -H 5" ), 7.14 (dd, 1H, J = 7.8; 7.9 Hz, Ar -H 6" ), 13 4 Hz, H 2' ), 6.96 (d, 1H, J = 7.9 Hz, Ar -H 4" ), 7.26 (t, 1H, J = 7.9 Hz, Ar -H 5" ), 7.43 (d, 1H, J = 8.0 Hz, Ar -H 6" ). 13  Then, four serial quadruple dilutions equivalent to 100, 25, 6.25, and 1.625 μg/mL were prepared. For amphotericin B, which was used as a control drug, four solutions were prepared from 1.0 µg/mL.
Parasite. The L. (V) panamensis strain M / HOM / 87 / UA140 was used, transfected with the gene for green uorescent protein (UA140-EGFP) generated in a previous work 54 . Parasites were cultured as promastigotes in Novy-MacNeil-Nicolle (NNN) biphasic medium and phosphate buffered saline (PBS) with glucose (pH 6.9) as the liquid phase. The cultures were incubated at 26 ºC. To ensure a greater infection of macrophages in vitro, the L. (V) panamensis strain was maintained by successive passages in experimentally infected hamsters (Mesocricetus auratus), making periodic aspirations of hamster lesions using PBS and #26 needles. The aspirated samples were grown in NNN culture medium at 26 o C, until promastigotes were obtained, which were used to infect U937 macrophages as described below [54][55][56][57] .
Then, 100 µL of each of the corresponding concentrations of each compound (200, 50, 12.5, or 3.125 µg/mL) was added to each well. As a viability control (no cytotoxicity), cells incubated in complete RPMI-1640 medium were used, while cells exposed to doxorubicin (DOX) were used as a cytotoxicity control 54 .
Cytotoxicity was determined according to the percentage decrease in the number of living cells obtained for each concentration of compound or amphotericin B (AMB), according to the optical densities (OD) obtained in each experimental condition and, in comparison, with the OD obtained from cells not exposed to any compound. The decrease in cell viability (called inhibition of cell growth) was calculated using the OD values for each evaluated condition, through the following equation: % Viability = [OD cells exposed to the compound ÷ OD cells not exposed] × 100. The OD values obtained for cells in the absence of compounds corresponded to 100% viability 54,55 .
The cells in the presence of the different solutions of the compounds at the respective concentrations, as well as the cells exposed to amphotericin B and doxorubicin and those not exposed, were incubated at 37 °C in a 5% CO 2 atmosphere for 72 hours. After the incubation period, 10 µL/well of a MTT solution with a concentration of 5 µg/mL (Sigma) was added and the dishes were incubated at 37 °C for 3 hours. After this incubation period, 100 μL/well of a solution of 50% isopropanol (Merck Millipore) and 10% sodium dodecyl sulfate (SDS) (Merck Millipore) was added, in order to solubilize the formazan crystals formed. The plates were incubated for another 30 minutes and the production of formazan (which is proportional to the percentage of viable cells) was measured in a microplate reader (Benchmark Bio-Rad Hercules, CA, USA) at a wavelength of 570 nm 54 .
From the percentage of viability, the percentage of mortality was calculated, which corresponds to the reciprocal of viability. Finally, with the mortality percentages, the mean lethal concentration (LC 50 ) was calculated, using the dose response analysis method Probit with the SAS Data Analysis statistical program (SAS Institute Cary NC, USA). The tests were carried out twice, with three replicates for each concentration evaluated. The cytotoxicity of each compound was classi ed, according to the LC 50 values, using the following scale: high cytotoxicity, LC 50 < 100 µg/mL; moderate cytotoxicity, 100 < LC 50 < 200 μg/mL; and low cytotoxicity, LC 50 > 200 μg/mL 54 . In vitro antileishmanial activity of compounds 4a-f. The activity of the compounds was evaluated in intracellular amastigotes obtained after in vitro infection of macrophages U-937. For this, the U-937 cells maintained in suspension culture were centrifuged at 1,500 rpm for 10 minutes and, after discarding the supernatant, the button cells were resuspended at a concentration of 1×10 5 cells/mL of complete RPMI 1640 medium and 0.1 μg/mL phorbol myristate acetate (PMA; Sigma). In each well of a 24-well cell culture plate (Falcon), 1 mL of the cell suspension was dispersed and incubated at a temperature of 37 °C with a 5% CO 2 atmosphere. After 72 hours, cells were infected with promastigotes in stationary phase of growth at a ratio of 15:1 parasites/cell. The dishes were incubated at 34 °C under 5% CO 2 for 2 hours. Subsequently, two washes with PBS were carried out to eliminate free parasites, 1 mL of complete RPMI 1640 medium was added, and the cells were incubated again for 24 hours. After this, the infected cells  The effectiveness of each treatment was determined after comparing the lesion sizes prior to and after treatments.
Treatment outcome at the end of study was recorded as cure (healing of 100% of the area and complete disappearance of the lesion), improvement (any percentage of reduction of lesion), failure (low decrease or an increase in the size of the lesion), or relapse (reactivation of lesion after initial cure). Toxicity of the cream formulations was determined according to changes in the body weight obtained during and after treatment, as well as The best interaction binding energy (kcal·mol -1 ) was selected for evaluation. Docking result 3D representations were used, from the Discovery Studio 68 molecular graphics system. In this context, and based on our past experience 73 , we demonstrated that, in LmDHODH, the S2 sub-site is essential for the activity of the LmDHODH enzyme, which contains active amino acids within the loop (a4-bA) 45

Results And Discussion
Compounds 4a-f, were obtained using acetic anhydride and high temperatures as a reaction condition, with yields between 37-90%. The synthesis strategy was based on the modi cation of the aromatic aldehyde and the quinolinic ring, to determine the change in biological activity of the compounds (Figure 1) 25 .

Biological activities
The results of in vitro cytotoxic and Leishmanicidal activity are summarized in Table 1. All compounds 4a-f were shown to be cytotoxic to U-937 cells. As expected, Amphotericin B (AMB) and Doxorubicin (DOX) showed high cytotoxicity, with LC 50 values of 0.05 µM and 0.02 µM, respectively. Despite the high cytotoxicity, all compounds 4a-f were also active against intracellular amastigotes of L. (V) panamensis, with inhibition percentages greater than 50% ( Table 1). The most active compounds were 4c, inhibiting 91.2%, followed by 4d and 4c, with inhibition of 88.4% and 87.8%, respectively. AMB, used as a control compound for antileishmanial activity, showed a percentage of inhibition of 69% (Table 1).
The dose-response relationship showed that all compounds 4a-f are highly active against intracellular amastigotes of L. (V) panamensis, with compound 4b being the most active, with an EC 50  Although toxicity is an important criterion in the development of new drugs, the criterion of speci c biological activity is even more important-in this case, antileishmanial activity-as cytotoxicity can be controlled or reduced by applying drug delivery systems, such as liposomes, and other types of nanoparticles that improve pharmacological activity, without loss of pharmacological potential. For example, AMB is a drug in current use for the treatment of cutaneous and visceral leishmaniasis which, in its free form-amphotericin B deoxycholate-is associated with high renal toxicity, such that treatment must be administered in hospitalized patients. In contrast, if amphotericin B is in a colloidal dispersion, lipid complexes, and liposomes, it has fewer adverse effects and lower renal toxicity 26,27 .
In vivo therapeutic response of the topical formulation of compounds 4a-f The evolution of ulcers was monitored for 90 days after the end of the treatment. When the treatment was effective, the ulcerative lesions gradually regressed to complete healing (0.0 mm 2 ) or reduced in size. On the contrary, when the treatment did not work, the size of the lesion increased. Treatment with 1% cream containing compound 4d was the most effective, managing to cure of 40% of hamsters and producing improvement in the remaining of 60% the hamsters in the group (Table 2). For their part, 1% creams containing compounds 4b or 4c managed to cure0f 20% hamsters in each group and produced improvements in the 80% hamsters, while treatment with compound 4f cured 20% hamsters and failed in the remaining in the 80% hamsters. On the other hand, creams containing compounds 4a or 4e produced improvement in the 80% hamsters and failures in the remaining hamster. As expected, MA treatment produced the highest percentage of cures, with thr80% of the hamsters cured (   Table 3. With compound 4d, the percentages of improvement in the three hamsters that were not cured were close to 80%, while in the hamsters treated with compounds 4b and 4c, the improvement percentages in the four hamsters were not cured were between 40.6-90.3%. In the group treated with compound 4a, the percentages of improvement ranged between 58.9-88.7% (see Table 3). With compounds 4a and 4f, there were failures in the treatment, as evidenced by percentages of reduction of the lesion less than 10% and even negative values, corresponding to an increase in the size of the lesion (see Table 3). The values are the reduction percentages, comparing the size at the end of the study with respect to the size before the treatment. Negative values indicate an increase in the size of the lesion at the end of the study. The appearance of the lesions before treatment and at the end of the study in a representative hamster in each treatment with compounds 4a-f are shown in Figure S1 in supplementary information.

Effect of treatment with cream formulation 4a-f on the body weight of hamsters with cutaneous leishmaniasis.
No loss in average body weight in hamsters was observed during the study; therefore, no detrimental effect on hamster weight or toxic effects could be attributed to compound treatment. According to the weight of the animals at the beginning and during the study, no signi cant differences were observed in the groups of hamsters treated with compounds or MA. (Figure S2 in supplementary information) Likewise, no alterations were observed in the levels of ALT, BUN, and serum creatinine, measured 8 days after treatment with compounds 4a-f or MA, which suggests that liver and kidney function were not affected in a way that could be attributed to these treatments.

Molecular docking simulation
According to the obtained results shown in Table 4, Figure 2, and Tables S1-S2 (see in supplementary information). The molecular docking experiments showed more favorable interactions, as well as ligand e ciency with LmDHODH target. In general, the low K d values indicate strong binding of the molecule to the protein. Therefore, compounds 4af exhibited promising activity against intracellular amastigotes of L. (V) panamensis, as these compounds formed a stable complex with each target studied. Table 4 also shows that 4b and 4d presented better interaction energies in 2arylquinoline-LmDHODH interactions. According to the experimental data, compounds 4b-d were able to produce healing and improvement of the lesions in the hamsters after being treated. In Tables 5 and 6, the interacting residues for both targets are summarized. Herein, it is possible to observe some residue differences in the binding modes of the active compounds.

ADMET Properties
The goal of calculating ADMET pro les is to provide, with reasonable accuracy, a preliminary prediction of the in vivo behaviour of a compound, in order to assess its potential to become a drug 28 . The molecules used in this study were submitted to the calculation of their absorption, distribution, metabolism, excretion, and toxicological properties (ADMET). Furthermore, physicochemical properties, such as molecular hydrogen bond acceptor (HBA), hydrogen bond donor (HBD), weight (MW), topological polar surface area (TPSA), rotatable bond count (RB), octanol/water partition coe cient (LogP), and Molar Refractivity (MR) were calculated, using the SwissADME webserver 29 . Compound toxicological properties were analysed taking into account the Lipinski, Ghose, Veber, and P zer toxicity empirical rules.
In order to assess whether molecules can be selected as potential 4a-f inhibitors, we calculated some pharmacokinetic properties (Table 5). These results were contrasted against Lipinski 30 , Ghose 31 , Veber 32 , and P zer 33 rules. If any of the compounds only satis ed two of the rules of Lipinski and Ghose, we took that compound as precautionary; if it satis ed only one rule, then this molecule is not a good candidate. Following Veber's rules, if a compound does not meet any of these parameters, then it is not a good drug candidate. P zer's toxicity rules were also taken into account-if any of our ligands did not meet these parameters, then it was not considered a good drug candidate. According to Table 5, it is observable that the candidates were within the range of expected values for the Lipinski's and Gelovani's parameters. Therefore, their bioavailability and absorption are not poor, and they were expected to be orally active. Additionally, the Boiled-egg model (see Figure S5 in supplementary information) was used to calculate the lipophilicity and polarity of these molecules. The results showed that all of the studied 2arylquinolines are highly absorbable at the blood-brain barrier, while not being absorbable in the gastrointestinal tract. Finally, the ADMET properties suggested that these compounds may be safe compounds for use as leishmaniasis inhibitors (see Figure S4 in supplementary information). The molecular docking results allowed us to recognize that the LmDHODH-ligand complexes showed more favourable interactions and ligand e ciency. Therefore, the nally selected position of the ligands, based on the docking score and predicted binding energy, was studied to describe the molecular interactions of LmDHODH with the bound ligands over time. In this context, molecular dynamics (MD) simulation showed the dynamic behaviour of the LmDHODH-ligand molecular system, assessing the stability of the complex. The most highly stable conformations for the LmDHODH system were subjected to the study of molecular dynamics with the CHARMM force eld. Thus, six LmDHODH complexes were built for each model, that were con ned inside a periodic simulation box.
The water model TIP3P 34 , with 12.552 molecules, was utilized as solvent. Moreover, Na + and Clions were added, for neutralizing the systems and maintaining an ionic concentration of 0.15 mol·L -1 . Full geometric optimizations of the two molecules were done by density functional theory method by M05-2X 35 -D3 36 , in conjunction with the 6-31G(d,p) basis set. The compounds (4a-f) and avin mononucleotide (FMN) were parametrized using the LigParGen web server and the OPLS-AA/1.14*CM1A(-LBCC) force eld parameters were used for organic ligands [37][38][39] . The partial charges were determined for each ligand by the restrained electrostatic potential (RESP) model 40 . MD simulations were realized using the modelled CHARMM22 and CHARMM36 force elds 41,42 within the NAMD software 43 . 20,000 steps of conjugate-gradient energy minimization were included for each system, followed by simulations of 5 ns with xed protein backbone atoms, and a gradual release of the backbone over 100,000 ps with restraints from 10 to 0.0 kcal·mol -1 Å -2 . The total duration of simulation was ~70 ns per system. During the molecular dynamic's simulations, motion equations were integrated with 2 femtoseconds time-steps in the NPT ensemble at a pressure of 1 atm.
SHAKE algorithm was applied to all hydrogen atoms, and a 12Å Van der Waals cut-off was set. Temperature was set to 310 K, by the Nosé-Hoover thermostat method with a relaxation time of 1 ps. Pressure was controlled at 1 atm by using a Nosé-Hoover-Langevin piston. Long-range electrostatic forces were considered by the particle-mesh Ewald approach. MD runs data was collected every 1 ps. Molecular visualization of the systems and trajectory analysis was done using the VMD software package 44 . The nal snapshots of the molecular dynamics simulations are illustrated in Figure S6 (see supplementary information) Previous studies have shown that the FMN molecule plays a functional role in the active site of LmDHODH, acting as a stabilizing cofactor of the active site and forming an aromatic box whose function is to stabilize the ligand [45][46][47] .
The molecular simulation results showed differences in the binding and interaction of compounds 4a, 4b, 4c, 4d, and 4f with the main binding site (BP1); see Figures S6 and S7 in supplementary information) The results also show that, throughout the simulation trajectory, compound 4a remained stable in this original binding site (where the catalytic function of LmDHODH is found, regions S1-S5); interacting in a stable way with FMN and representing a 10% interaction throughout the molecular simulation (see Figure S8 in supplementary information).
On the other hand, compounds 4b, 4c, 4d, and 4f were stably located in a position close to the binding site of LmDHODH (BP2 site), with a low percentage of participation of the residues in the regions S1-S5 (see Figure 10); this uctuation space made a null interaction with the FMN cofactor, (see Figures S6 and S7  Molecular dynamics simulations showed that LmDHODH residues directly interacted with the ligands (4a-f). The most frequent LmDHODH residues are illustrated in Figure S8 (see in supplementary information). Additionally, the potential inhibitors evaluated here interacted with the before-mentioned pockets (BP1 and BP2) through electrostatic and hydrophobic interactions. In the case of compound 4a, it showed interactions with the S1-S4 regions, with the residues gly71, leu72, ser100, ser130, gln139, asn128, asn195, phe218, gly198, ser196, and ile197 (see Figure  S8 in supplementary information). These coincide with those previously reported by the scienti c community, which are the S1, S2, S3, and S4 sites, thus validating the protocol used in this work 45,[47][48][49] . For compounds 4b, 4c, 4d, and 4f, a low percentage was shown, with the only interaction region being S2, involving residues leu102, leu72, val140, asn107, cys150, asn107, val140, and ser100; the other residue interactions were distributed at the BP2 binding site. In the case of compound 4e, there was no stable interaction at the main binding site, causing the escape of this compound into the solvated medium during the molecular simulation. These results document that compounds 4a-f are reversible inhibitors of LmDHODH.

Free Energy Calculation
The molecular MM/GBSA method was employed, in order to estimate the binding free energy of the LmDHODH complexes. For calculations from a total of 70 ns of MD, the last 50 ns were extracted for analysis, and the explicit water molecules and ions were removed. The MM/GBSA analysis was performed on three subsets of each system: the protein alone, the ligand alone, and the complex (protein-ligand). For each of these subsets, the total free energy (ΔG tot ) was calculated as follows: where E MM is the bonded and Lennard-Jones energy terms, G solv is the polar contribution of solvation energy and non-polar contribution to the solvation energy, T is the temperature, and ΔS conf corresponds to the conformational entropy 50 . Both E MM and G solv were calculated using the NAMD software with the generalized Born implicit solvent model 51,52 . ΔG tot was calculated as a linear function of the solvent-accessible surface area, which was calculated with a probe radius of 1.4 Å 53 . The binding free energy of LmDHODH and ligand complexes (ΔG bind ) were calculated by the difference, where ΔG tot values are the averages over the simulation: The binding free energy (MM/GBSA) was computed after the MD simulation, considering the last 70 ns for all of the complexes; the results are given in Table 4. Compound 4d had a binding free energy of -30.05 kcal·mol -1 with the LmDHODH enzyme, while compound 4b showed a comparable binding free energy of -25.73 kcal·mol -1 . In the case of compounds 4a and 4c, they showed relatively higher binding energy, with values of -22.32 kcal·mol -1 and -21.13 kcal·mol -1 , respectively. Compounds 4e and 4f had the highest binding energy values (-11.42 kcal·mol -1 and -18.59 kcal·mol -1 , respectively), indicating the low stability of these compounds at the LmDHODH binding site; (see Figures  S7 and S8 in supplementary information). The results obtained from MM/GBSA (see Table 6) calculations also demonstrated that compound 4a had a higher binding energy than compound 4d, with an absolute difference of 7.73 kcal·mol -1 . This difference was due to the interaction with the cofactor avin mononucleotide. In particular, the 4a compound had better activity at both the experimental and in silico levels.

Conclusion
Compounds 4a-f showed promising in vitro activity against intracellular amastigotes of L. (V) panamensis. A cream formulation containing 1% of compounds 4b-d was able to produce healing and improvement of lesions in hamsters after treatment. It is noteworthy that there was a correlation between the results of the leishmanicidal activity in vitro and in vivo, demonstrating the promising activity of the compounds; likewise, the administration of the compounds during the study did not generate obvious signs of toxicity or signi cant weight loss in the treated animals. These results suggest that the compounds synthesized in this work, especially 4b-d, are promising and could be considered therapeutic targets to be evaluated in future studies for the treatment of cutaneous Leishmaniasis. Molecular dynamics (MD) simulations and MM/GBSA studies revealed that the compound 4a has a preferential interaction with the cofactor avin mononucleotide, suggesting better activity at the in silico level, which was con rmed by our experimental results. A computational docking study showed that molecules 4b and 4d present better interaction energies in 2-arylquinoline-LmDHODH interactions, in agreement with the experimental data regarding the treatment of hamster lesions. Finally, we predicted that the synthesized 2-arylquinolines are absorbable at the blood-brain barrier, but they have no action in the gastrointestinal tract.

Declarations
Con ict of interest