Anti-Mycobacterium tuberculosis Activity of Esters of Quinoxaline 1,4-Di-N-Oxide

Tuberculosis continues to be a public health problem in the world, and drug resistance has been a major obstacle in its treatment. Quinoxaline 1,4-di-N-oxide has been proposed as a scaffold to design new drugs to combat this disease. To examine the efficacy of this compound, this study evaluates methyl, ethyl, isopropyl, and n-propyl esters of quinoxaline 1,4-di-N-oxide derivatives in vitro against Mycobacterium tuberculosis (pansusceptible and monoresistant strains). Additionally, the inhibitory effect of esters of quinoxaline 1,4-di-N-oxide on M. tuberculosis gyrase supercoiling was examined, and a stability analysis by ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS) was also carried out. Results showed that eight compounds (T-007, T-018, T-011, T-069, T-070, T-072, T-085 and T-088) had an activity similar to that of the reference drug isoniazid (minimum inhibitory concentration (MIC) = 0.12 µg/mL) with an effect on nonreplicative cells and drug monoresistant strains. Structural activity relationship analysis showed that the steric effect of an ester group at 7-position is key to enhancing its biological effects. Additionally, T-069 showed a high stability after 24 h in human plasma at 37 °C.


Introduction
According to the World Health Organization (WHO), tuberculosis (TB) continues to be one of the leading infectious diseases in the world [1]. First-and second-line drugs are available for the treatment of the disease. However, the present therapy has been ineffective due to its long duration, as well as the emergence of resistance to these drugs [2]. In view of the significance of TB as an infectious disease

Biological Activity
A common procedure to obtain quinoxaline 1,4-di-N-oxide derivatives is the use of benzofuroxane N-oxide as a principal reagent [15]. Additionally, benzofuroxane N-oxide is a heterocyclic compound with antitubercular activity [16]. Therefore, as a first step, benzofuroxane-N-oxide derivatives were used to obtain esters of quinoxaline 1,4-di-N-oxide, which were also evaluated in vitro on M. tuberculosis strain H37Rv and the NRP strain. Results are shown in Table 1. SAR analysis showed that the compound T-046 with a hydrogen atom at R 1 position on the benzofuroxane N-oxide ring had a lower MIC value (58.5 µg/mL). However, when a carboxylate group was added at R 1 -position, the biological activity was enhanced sevenfold (compound T-014, MIC = 7.6 µg/mL). This effect increased when a carboxylate group with an aliphatic linear substituent was incorporated at the same position on the benzofuroxane N-oxide ring (T-074, MIC = 0.87 µg/mL). These results confirm that the steric effect at R 1 -position modulates anti-M. tuberculosis activity. Additionally, a second substitution at R 2 -position drastically reduced the activity (T-036, MIC >100 µg/mL). These results show that the benzofuroxane-5-carboxylate N-oxide ring is a good scaffold to develop new antitubercular agents. Table 1. Biological activity of benzofuroxane N-oxide derivatives using the microplate Alamar blue assay (MABA) against Mycobacterium tuberculosis strain H37Rv and the low oxygen recovery assay (LORA) against the nonreplicating persistent strain.
Molecules 2018, 23, x 3 of 12 Following on our main objective (esters of quinoxaline 1-4-di-N-oxide derivatives), 18 compounds of methyl and ethyl ester derivatives were evaluated on M. tuberculosis strain H37Rv. Results are shown in Table 2. All compounds of the methyl ester series showed good anti-M. tuberculosis activity (MIC < 0.35 µ g/mL). SAR analysis of the methyl series, showed that compounds with a carboxylate (methyl and ethyl) group at R1-position and a methyl group at R2-position had high activity (MIC < 0.5 µ g/mL). However, acetyl, benzoyl, and arylcarboxamide groups at R1-position decreased the effect. On the other hand, quinoxaline 1,4-di-N-oxide derivatives with a trifluoromethyl group at R2-position also showed good anti-M. tuberculosis activity (MIC < 2.0 µ g/mL) in spite of the presence of a carbonyl group with aliphatic or aromatic substitutes at R1-position. A comparison between analog compounds (T-003 and T-018) with a methyl and trifluoromethyl group at R2-position, respectively, showed that the electronegative properties of trifluoromethyl enhance their biological activity. Compound T-018 showed the best biological activity (MIC = 0.15 µ g/mL), a value similar to the reference drug isoniazid (MIC = 0.12 µ g/mL). Also, this compound showed the best activity (low oxygen recovery assay (LORA) MIC = 0.34 µ g/mL) against the NRP M. tuberculosis strain. Analyzing all LORA results of quinoxaline 1,4-di-N-oxide derivatives, these compounds had the same biological behavior on M. tuberculosis strain H37Rv; therefore, substitutions at R1-and R2-position also affect biological activity on the NRP strain.
In the ethyl ester series in general, all quinoxaline 1,4-di-N-oxide derivatives showed good anti-M. tuberculosis activity (MIC < 2.5 µ g/mL) except compounds T-026 and T-045, which had ten times less activity (MIC > 22 µ g/mL). These compounds have an aliphatic linear substitution or a free amino group at R1-position with a trifluoromethyl and methyl group at R2-position, respectively.
Compound T-015 with a carboxyethyl and methyl group at R1-and R2-position, respectively, showed a MIC of 0.50 µ g/mL. In addition, compounds with a trifluoromethyl group at R2-position showed better anti-M. tuberculosis activity with an acetyl or aromatic group at R1-position.
Compounds T-007 and T-011 showed MIC values (0.14 and 0.10 µ g/mL) similar to that of isoniazid (MIC = 0.12 µ g/mL). These compounds have a thiophene and a naphthyl group, respectively. These groups could consider a bioisostere from a phenyl group present in the best compound (T-018) of the methyl ester series. Additionally, T-007, T-011, and T-018 showed the best effect on the NRP M. tuberculosis strain. An analysis between the methyl and ethyl series showed that a substitution at R3-position does not affect biological activity; for example, compounds T-008 and T-016 showed a similar MIC value. Additionally, the ethyl quinoxaline 1,4-di-N-oxide series confirm that substitutions at R1-and R2-position also affect the biological behavior on the NRP M. tuberculosis strain. Following on our main objective (esters of quinoxaline 1-4-di-N-oxide derivatives), 18 compounds of methyl and ethyl ester derivatives were evaluated on M. tuberculosis strain H37Rv. Results are shown in Table 2. All compounds of the methyl ester series showed good anti-M. tuberculosis activity (MIC < 0.35 µg/mL). SAR analysis of the methyl series, showed that compounds with a carboxylate (methyl and ethyl) group at R 1 -position and a methyl group at R 2 -position had high activity (MIC < 0.5 µg/mL). However, acetyl, benzoyl, and arylcarboxamide groups at R 1 -position decreased the effect. On the other hand, quinoxaline 1,4-di-N-oxide derivatives with a trifluoromethyl group at R 2 -position also showed good anti-M. tuberculosis activity (MIC < 2.0 µg/mL) in spite of the presence of a carbonyl group with aliphatic or aromatic substitutes at R 1 -position. A comparison between analog compounds (T-003 and T-018) with a methyl and trifluoromethyl group at R 2 -position, respectively, showed that the electronegative properties of trifluoromethyl enhance their biological activity. Compound T-018 showed the best biological activity (MIC = 0.15 µg/mL), a value similar to the reference drug isoniazid (MIC = 0.12 µg/mL). Also, this compound showed the best activity (low oxygen recovery assay (LORA) MIC = 0.34 µg/mL) against the NRP M. tuberculosis strain. Analyzing all LORA results of quinoxaline 1,4-di-N-oxide derivatives, these compounds had the same biological behavior on M. tuberculosis strain H37Rv; therefore, substitutions at R 1 -and R 2 -position also affect biological activity on the NRP strain.
In the ethyl ester series in general, all quinoxaline 1,4-di-N-oxide derivatives showed good anti-M. tuberculosis activity (MIC < 2.5 µg/mL) except compounds T-026 and T-045, which had ten times less activity (MIC > 22 µg/mL). These compounds have an aliphatic linear substitution or a free amino group at R 1 -position with a trifluoromethyl and methyl group at R 2 -position, respectively.
Compound T-015 with a carboxyethyl and methyl group at R 1 -and R 2 -position, respectively, showed a MIC of 0.50 µg/mL. In addition, compounds with a trifluoromethyl group at R 2 -position showed better anti-M. tuberculosis activity with an acetyl or aromatic group at R 1 -position. Compounds T-007 and T-011 showed MIC values (0.14 and 0.10 µg/mL) similar to that of isoniazid (MIC = 0.12 µg/mL). These compounds have a thiophene and a naphthyl group, respectively. These groups could consider a bioisostere from a phenyl group present in the best compound (T-018) of the methyl ester series. Additionally, T-007, T-011, and T-018 showed the best effect on the NRP M. tuberculosis strain. An analysis between the methyl and ethyl series showed that a substitution at R 3 -position does not affect biological activity; for example, compounds T-008 and T-016 showed a similar MIC value. Additionally, the ethyl quinoxaline 1,4-di-N-oxide series confirm that substitutions at R 1 -and R 2 -position also affect the biological behavior on the NRP M. tuberculosis strain.  Results shown in Table 2 confirm that methyl and ethyl ester groups at R3-position on the quinoxaline 1,4-di-N-oxide ring do not produce changes in anti-M. tuberculosis activity. However, these results (at R1-and R2-position) also suggest that steric effects modify biological activity. Therefore, we proposed obtaining two new series with an aliphatic substituent (isopropyl and n-propyl) at R3-position to confirm a positive, negative, or null effect on biological activity (Table 3). SAR analysis of the isopropyl series confirmed that a carboxylate (methyl and ethyl) group at R1-position enhanced anti-M. tuberculosis activity. As in the previous methyl and ethyl series, benzoyl, arylcarboxamide, or amine groups at R1-position reduced the activity with a methyl group at R2-position on the quinoxaline 1,4-di-N-oxide ring. Interestingly, an isopropyl group at R1-position drastically decreased biological activity. This confirms that steric effects are important in anti-M. tuberculosis activity. Also, in the isopropyl series, compounds with a trifluoromethyl group at R2-position showed the best activity. In particular, compound T-069 showed the best MIC (0.08 µ g/mL) of all the quinoxaline 1,4-di-N-oxide derivatives. This compound is an analogue of compound T-011 (ethyl series), although compound T-069 had a MIC value ten times higher. Therefore, the isopropyl group at R3-position is a key factor to enhance anti-M. tuberculosis activity. In the n-propyl series, the SAR partner is the same as the previous series (methyl, ethyl and isopropyl). The best compound in this series was T-089 (an analogue of T-011 and T-069) with a MIC of 0.12 µ g/mL on strain H37Rv and 0.15 µ g/mL on NRP M. tuberculosis strain. A SAR analysis of all esters of quinoxaline 1,4-di-N-oxide is shown in Figure 1. Results shown in Table 2 confirm that methyl and ethyl ester groups at R 3 -position on the quinoxaline 1,4-di-N-oxide ring do not produce changes in anti-M. tuberculosis activity. However, these results (at R 1 -and R 2 -position) also suggest that steric effects modify biological activity. Therefore, we proposed obtaining two new series with an aliphatic substituent (isopropyl and n-propyl) at R 3 -position to confirm a positive, negative, or null effect on biological activity (Table 3). SAR analysis of the isopropyl series confirmed that a carboxylate (methyl and ethyl) group at R 1 -position enhanced anti-M. tuberculosis activity. As in the previous methyl and ethyl series, benzoyl, arylcarboxamide, or amine groups at R 1 -position reduced the activity with a methyl group at R 2 -position on the quinoxaline 1,4-di-N-oxide ring. Interestingly, an isopropyl group at R 1 -position drastically decreased biological activity. This confirms that steric effects are important in anti-M. tuberculosis activity. Also, in the isopropyl series, compounds with a trifluoromethyl group at R 2 -position showed the best activity. In particular, compound T-069 showed the best MIC (0.08 µg/mL) of all the quinoxaline 1,4-di-N-oxide derivatives. This compound is an analogue of compound T-011 (ethyl series), although compound T-069 had a MIC value ten times higher. Therefore, the isopropyl group at R 3 -position is a key factor to enhance anti-M. tuberculosis activity. In the n-propyl series, the SAR partner is the same as the previous series (methyl, ethyl and isopropyl). The best compound in this series was T-089 (an analogue of T-011 and T-069) with a MIC of 0.12 µg/mL on strain H37Rv and 0.15 µg/mL on NRP M. tuberculosis strain. A SAR analysis of all esters of quinoxaline 1,4-di-N-oxide is shown in Figure 1. Table 3. Biological activity of isopropyl and n-propyl quinoxaline-7-carboxylate 1,4-di-N-oxide derivatives using the MABA against Mycobacterium tuberculosis strain H37Rv and the LORA against nonreplicating persistent strain.
Molecules 2018, 23, x 5 of 12 Table 3. Biological activity of isopropyl and n-propyl quinoxaline-7-carboxylate 1,4-di-N-oxide derivatives using the MABA against Mycobacterium tuberculosis strain H37Rv and the LORA against nonreplicating persistent strain.  Finally, eight compounds (MIC < 0.30 µ g/mL) from all esters of quinoxaline 1,4-di-N-oxide derivatives were selected to be tested against monoresistant M. tuberculosis and some  Table 3. Biological activity of isopropyl and n-propyl quinoxaline-7-carboxylate 1,4-di-N-oxide derivatives using the MABA against Mycobacterium tuberculosis strain H37Rv and the LORA against nonreplicating persistent strain.  Finally, eight compounds (MIC < 0.30 µ g/mL) from all esters of quinoxaline 1,4-di-N-oxide derivatives were selected to be tested against monoresistant M. tuberculosis and some Finally, eight compounds (MIC < 0.30 µg/mL) from all esters of quinoxaline 1,4-di-N-oxide derivatives were selected to be tested against monoresistant M. tuberculosis and some nontuberculous Mycobacterium (NTM) strains. Additionally, the half maximal cytotoxicity concentration (CC 50 ) on mammalian cell macrophages and the selectivity index (SI) were determined. The results are shown in Table 4. In general, all compounds had similar MICs values on M. tuberculosis and monoresistant strains. This suggests that the mechanism of action of esters of quinoxaline 1,4-di-N-oxide derivatives is different from reference drugs. Although our compounds had good antimycobacterial activity, none of the quinoxaline 1,4-di-N-oxide derivatives showed better MIC values than rifampicin in monoresistant strains, except against M. tuberculosis strain H37Rv RR. Compounds T-022 and T-088 showed the best biological activity on this strain and the best SI value (SI > 70). Activity against the NTM strains was not as good as with the tuberculosis strains; only compound T-085 was active against all NTM strains, and the slow grower NTM M. avium strain was susceptible to all compounds except T-088. These results confirm that these esters of quinoxaline 1,4-di-N-oxide can be considered for development of new pharmacological options for the treatment of TB susceptible or drug resistant.

Inhibitory Assay
DNA gyrases are enzymes within the class of type II topoisomerase, which form a heterotetramer composed of four subunits encoded by the gyrA and gyrB gene [17]. This enzyme catalyzes the unwinding of double-stranded DNA through the introduction of negative supercoils. This process is blocked by fluoroquinolones causing the formation of covalent enzyme-DNA adducts, which leads to chromosome fragmentation and cell death in bacteria. Thus, DNA gyrases are known targets for fluoroquinolones. However, some studies have hypothesized that quinoxaline derivatives could inhibit DNA gyrases from M. tuberculosis [14,18]. Keeping in pace with these conjectures, we analyzed the inhibitory effect of esters of quinoxaline 1,4-di-N-oxide derivatives on M. tuberculosis  [19]. Based on these results, esters of quinoxaline 1,4-di-N-oxide derivatives may not be M. tuberculosis gyrase supercoiling inhibitors, suggesting that these esters of quinoxaline 1,4-di-N-oxide derivatives have another mechanism of action that needs to be explored. quinoxaline 1,4-di-N-oxide derivatives is different from reference drugs. Although our compounds had good antimycobacterial activity, none of the quinoxaline 1,4-di-N-oxide derivatives showed better MIC values than rifampicin in monoresistant strains, except against M. tuberculosis strain H37Rv RR. Compounds T-022 and T-088 showed the best biological activity on this strain and the best SI value (SI > 70). Activity against the NTM strains was not as good as with the tuberculosis strains; only compound T-085 was active against all NTM strains, and the slow grower NTM M. avium strain was susceptible to all compounds except T-088. These results confirm that these esters of quinoxaline 1,4-di-N-oxide can be considered for development of new pharmacological options for the treatment of TB susceptible or drug resistant.

Inhibitory Assay
DNA gyrases are enzymes within the class of type II topoisomerase, which form a heterotetramer composed of four subunits encoded by the gyrA and gyrB gene [17]. This enzyme catalyzes the unwinding of double-stranded DNA through the introduction of negative supercoils. This process is blocked by fluoroquinolones causing the formation of covalent enzyme-DNA adducts, which leads to chromosome fragmentation and cell death in bacteria. Thus, DNA gyrases are known targets for fluoroquinolones. However, some studies have hypothesized that quinoxaline derivatives could inhibit DNA gyrases from M. tuberculosis [14,18]. Keeping in pace with these conjectures, we analyzed the inhibitory effect of esters of quinoxaline 1, 4- [19]. Based on these results, esters of quinoxaline 1,4-di-N-oxide derivatives may not be M. tuberculosis gyrase supercoiling inhibitors, suggesting that these esters of quinoxaline 1,4-di-N-oxide derivatives have another mechanism of action that needs to be explored. M. tuberculosis and monoresistant strains. This suggests that the mechanism of action of esters of quinoxaline 1,4-di-N-oxide derivatives is different from reference drugs. Although our compounds had good antimycobacterial activity, none of the quinoxaline 1,4-di-N-oxide derivatives showed better MIC values than rifampicin in monoresistant strains, except against M. tuberculosis strain H37Rv RR. Compounds T-022 and T-088 showed the best biological activity on this strain and the best SI value (SI > 70). Activity against the NTM strains was not as good as with the tuberculosis strains; only compound T-085 was active against all NTM strains, and the slow grower NTM M. avium strain was susceptible to all compounds except T-088. These results confirm that these esters of quinoxaline 1,4-di-N-oxide can be considered for development of new pharmacological options for the treatment of TB susceptible or drug resistant.

Inhibitory Assay
DNA gyrases are enzymes within the class of type II topoisomerase, which form a heterotetramer composed of four subunits encoded by the gyrA and gyrB gene [17]. This enzyme catalyzes the unwinding of double-stranded DNA through the introduction of negative supercoils. This process is blocked by fluoroquinolones causing the formation of covalent enzyme-DNA adducts, which leads to chromosome fragmentation and cell death in bacteria. Thus, DNA gyrases are known targets for fluoroquinolones. However, some studies have hypothesized that quinoxaline derivatives could inhibit DNA gyrases from M. tuberculosis [14,18]. Keeping in pace with these conjectures, we analyzed the inhibitory effect of esters of quinoxaline 1, 4- [19]. Based on these results, esters of quinoxaline 1,4-di-N-oxide derivatives may not be M. tuberculosis gyrase supercoiling inhibitors, suggesting that these esters of quinoxaline 1,4-di-N-oxide derivatives have another mechanism of action that needs to be explored. quinoxaline 1,4-di-N-oxide derivatives is different from reference drugs. Although ou had good antimycobacterial activity, none of the quinoxaline 1,4-di-N-oxide deriva better MIC values than rifampicin in monoresistant strains, except against M. tube H37Rv RR. Compounds T-022 and T-088 showed the best biological activity on this best SI value (SI > 70). Activity against the NTM strains was not as good as with th strains; only compound T-085 was active against all NTM strains, and the slow gro avium strain was susceptible to all compounds except T-088. These results confirm th of quinoxaline 1,4-di-N-oxide can be considered for development of new pharmacol for the treatment of TB susceptible or drug resistant.

Inhibitory Assay
DNA gyrases are enzymes within the class of type II topoisomerase, w heterotetramer composed of four subunits encoded by the gyrA and gyrB gene [17]. catalyzes the unwinding of double-stranded DNA through the introduction of negati This process is blocked by fluoroquinolones causing the formation of covalent adducts, which leads to chromosome fragmentation and cell death in bacteria. Thus, are known targets for fluoroquinolones. However, some studies have hypothesized th derivatives could inhibit DNA gyrases from M. tuberculosis [14,18]

Stability Analysis by UPLC-MS
The presence of ester groups at 7-position on the quinoxaline 1,4-di-N-oxide ring could suggest that these compounds will not be stable when tested in animal models. Therefore, a simple assay to test the stability of compounds T-018 and T-069-methyl and isopropyl ester derivatives, respectively-was performed using ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS). The results are shown in Figure 2. Figure 2A, shows only one peak with a time retention of 1.26 min corresponding to compound T-018 (m/z = 393.02). After 12 h in human plasma, two new low intensity peaks are seen (time retention= 0.72 and 2.05 min) (Figure 2a). These peaks, which increased in intensity after 24 h (Figure 2b) represented 4.82 and 4.12%, respectively, suggesting a low degradation of compound T-018. In Figure 2B, only one peak was observed with a time retention of 1.51 min corresponding to compound T-069 (m/z = 427.02). After 12 h, a second peak is seen at 0.62 min (Figure 2c). Finally, after 24 h, another peak occurs at 2.72 min (Figure 2d). Both peaks at 24 h represent 0.83% and 2.20%, respectively, suggesting that T-069 had a lower degradation. These results show that esters of quinoxaline 1,4-di-N-oxide are stable in human plasma after 24 h.

Stability Analysis by UPLC-MS
The presence of ester groups at 7-position on the quinoxaline 1,4-di-N-oxide ring could suggest that these compounds will not be stable when tested in animal models. Therefore, a simple assay to test the stability of compounds T-018 and T-069-methyl and isopropyl ester derivatives, respectively-was performed using ultra performance liquid chromatography-tandem mass spectrometry (UPLC-MS). The results are shown in Figure 2. Figure 2A, shows only one peak with a time retention of 1.26 min corresponding to compound T-018 (m/z = 393.02). After 12 h in human plasma, two new low intensity peaks are seen (time retention= 0.72 and 2.05 min) (Figure 2a). These peaks, which increased in intensity after 24 h (Figure 2b) represented 4.82 and 4.12%, respectively, suggesting a low degradation of compound T-018. In Figure 2B, only one peak was observed with a time retention of 1.51 min corresponding to compound T-069 (m/z = 427.02). After 12 h, a second peak is seen at 0.62 min (Figure 2c). Finally, after 24 h, another peak occurs at 2.72 min (Figure 2d). Both peaks at 24 h represent 0.83% and 2.20%, respectively, suggesting that T-069 had a lower degradation. These results show that esters of quinoxaline 1,4-di-N-oxide are stable in human plasma after 24 h.

Chemical Synthesis
All compounds from the methyl, ethyl, isopropyl, and n-propyl ester series of quinoxaline 1,4-N-oxide were obtained following the procedure previously reported by Gomez-Caro et al. [20]. These compounds were analyzed by IR, 1 H-NMR and elemental analysis and have been previously reported [21][22][23][24].

Antitubercular Assays
The antitubercular activity (MIC values) was assessed in vitro against Mycobacterium tuberculosis strain H37Rv ATCC 27294 according to a modified microplate Alamar blue assay (MABA) [25]. The assays were performed in triplicate independent experiments. The standard M. tuberculosis strain H37Rv was tested with known reference drugs rifampicin and isoniazid. The lowest drug concentration effecting an inhibition of 90% was considered as MIC. Additionally, an in vitro LORA test was done following the procedure previously reported by Cho et al. [26]. Antituberculosis activity testing against drug monoresistant strains (ATCC35822, ATCC35838, ATCC35837, and ATCC35820) or NTM clinical isolates was performed as described earlier by Luna-Herrera et al. [27].

Selectivity Index
The

Chemical Synthesis
All compounds from the methyl, ethyl, isopropyl, and n-propyl ester series of quinoxaline 1,4-N-oxide were obtained following the procedure previously reported by Gomez-Caro et al. [20]. These compounds were analyzed by IR, 1 H-NMR and elemental analysis and have been previously reported [21][22][23][24].

Antitubercular Assays
The antitubercular activity (MIC values) was assessed in vitro against Mycobacterium tuberculosis strain H37Rv ATCC 27294 according to a modified microplate Alamar blue assay (MABA) [25]. The assays were performed in triplicate independent experiments. The standard M. tuberculosis strain H37Rv was tested with known reference drugs rifampicin and isoniazid. The lowest drug concentration effecting an inhibition of 90% was considered as MIC. Additionally, an in vitro LORA test was done following the procedure previously reported by Cho et al. [26]. Antituberculosis activity testing against drug monoresistant strains (ATCC35822, ATCC35838, ATCC35837, and ATCC35820) or NTM clinical isolates was performed as described earlier by Luna-Herrera et al. [27].

Selectivity Index
The determination of the half maximal cytotoxicity concentration (CC 50 ) of the compounds  T-006, T-011, T-018, T-022, T-069, T-085, T-088, and T-089 was carried out on the mouse macrophage  cell line J774A.1 (ATCC TIB-67). Briefly, cell monolayers were prepared in 96-well plates with 10,000 cells per well in Ham's F-12 medium supplemented with 10% Fetal bovine serum (FBS, By products, Guadalajara, Mexico) and antibiotics (penicillin and gentamicin). Five concentrations from 100 to 0.5 µg/mL of each of the compounds were tested. The cells were incubated for a period of 6 h ( − ) or 96 h ( + ). Before completing this period, 20 µL of Alamar blue solution was added to each well, quantifying the relative fluorescence units in a fluorometer (Fluoroskan Ascent FL, Labsystems). The percentage of cytotoxicity at each concentration was determined by comparing the values against the control of cells without treatment. The CC 50 was determined with the Probit regression analysis (MedCalc Statistical Software version 18.5, Ostend, Belgium). Selectivity index was calculated as the ratio of the CC 50 on the macrophage cell line J774A.1 and the MIC value against M. tuberculosis strain H37Rv [28].

Assay Set Up
The activity of the enzyme was determined prior to the testing of the quinoxaline 1,4-di-N-oxide derivatives and 1 U was defined as the amount of enzyme required to just fully supercoil the substrate under test conditions. Compounds were tested at 50 µM. Final DMSO concentration in the assays was 1% (v/v). Each reaction was stopped by the addition of 30 µL chloroform/iso-amyl alcohol (26:1) and 20 µL Stop Dye (40% sucrose, 100 mM Tris.HCl (pH 7.5), 10 mM EDTA, 0.5 µg/mL bromophenol blue), before being loaded on a 1.0% TAE (Tris.acetate 0.04 mM, EDTA 0.002 mM) gel run at 80 V for 2 h.
Raw gel data (fluorescent band volumes) collected from Syngene's GeneTools gel analysis software were converted to a % of the 100% control (the supercoiled enzyme plus DMSO control DNA band). These were analyzed using SigmaPlot Version 13.0 (Systat Software Inc., London, UK; 2016).

Chromatographic Analysis
Compounds T-018 and T-069 were selected to analyze their stability in human plasma. A total of 1 mg of each of the two compounds was dissolved in 1 mL of dichlorometane. Then, 0.1 mL was added to 0.9 mL of 0.1% formic acid in acetonitrile for analysis by UPLC with an ACQUITY QDa mass detector from Waters (Milford, MA, USA) under the following conditions: column: ACQUITY UPLC ® BEH C 18 1.7 µm, 2.1 × 100 mm; mobile phase: 0.1% formic acid/acetonitrile; run time: 5 min; flow rate: 0.3 mL/min; injection volume: 0.5 µL; temperature: 40 • C. After that, 3 mL of human plasma was obtained from a volunteer under medical supervision and following the standard procedures. Then, 1 mg of T-018 and T-069 were dissolved with DMSO 2% and added to 1 mL of human plasma.
Samples were incubated at 37 • C for 12 and 24 h. T-018 and T-069 were obtained by liquid extraction with dichloromethane (10 mL) three times and the solvent was eliminated for vacuum pressure. After that, 1 mL of dichlorometane was added and 0.1 mL was added to 0.9 mL of 0.1% formic acid in acetonitrile and analyzed by UPLC-MS. Percentage of degradation was calculated considering the area under the curve of each peak/the area under the curve of all peaks × 100% at 24 h only.

Conclusions
In this study, our results confirmed that esters of quinoxaline 1,4-di-N-oxide have good anti-M. tuberculosis activity (MIC < 5 µg/mL) except for five compounds (T-045, ethyl series; T-066 and T-108, isopropyl series; T-091 and T-0125, n-propyl series). Additionally, the best compounds (T-007, T-011, T-018, T-069, T-070, T-072, T-085, and T-088) with a MIC value of ≤0.15 µg/mL showed biological activity on both the nonreplicative cells and M. tuberculosis monoresistant strains. Structure-activity analysis showed that the compounds with better biological activity were obtained when carboxylate, trifluoromethyl, and isopropyl groups were present at R 1 , R 2 and R 3 -position, respectively, on the quinoxaline 1,4-di-N-oxide ring. An enzymatic assay showed that these compounds cannot inhibit M. tuberculosis gyrase supercoiling. Also, these compounds showed a low degradation in human plasma. Therefore, esters of quinoxaline 1,4-di-N-oxide are a good option for developing new antitubercular agents.