New Zileuton-Hydroxycinnamic Acid Hybrids: Synthesis and Structure-Activity Relationship towards 5-Lipoxygenase Inhibition

A novel series of zileuton-hydroxycinnamic acid hybrids were synthesized and screened as 5-lipoxygenase (5-LO) inhibitors in stimulated HEK293 cells and polymorphonuclear leukocytes (PMNL). Zileuton’s (1) benzo[b]thiophene and hydroxyurea subunits combined with hydroxycinnamic acid esters’ ester linkage and phenolic acid moieties were investigated. Compound 28, bearing zileuton’s (1) benzo[b]thiophene and sinapic acid phenethyl ester’s (2) α,β-unsaturated phenolic acid moiety 28, was shown to be equipotent to zileuton (1), the only clinically approved 5-LO inhibitor, in stimulated HEK293 cells. Compound 28 was three times as active as zileuton (1) for the inhibition of 5-LO in PMNL. Compound 37, bearing the same sinapic acid (3,5-dimethoxy-4-hydroxy substitution) moiety as 28, combined with zileuton’s (1) hydroxyurea subunit was inactive. This result shows that the zileuton’s (1) benzo[b]thiophene moiety is essential for the inhibition of 5-LO product biosynthesis with our hydrids. Unlike zileuton (1), Compound 28 formed two π–π interactions with Phe177 and Phe421 as predicted when docked into 5-LO. Compound 28 was the only docked ligand that showed a π–π interaction with Phe177 which may play a part in product specificity as reported.

Many efforts have been made to develop anti-leukotriene drugs, given the involvement of these mediators in many pathologies. Approved by the FDA in 1996, zileuton ((Zyflo ® ) (1); Figure 1) was prescribed as the first 5-LO inhibitor, which inhibits leukotriene (LTB 4 , LTC 4 , LTD 4 , and LTE 4 ) biosynthesis. The original indication was to be used for prophylaxis and chronic treatment of asthma in adults and children over 12 years of age [14]. However, zileuton's (1) dosing regimen as well as requirements for liver transaminase monitoring have limited its clinical use [15,16]. The limited use of Previous work from our team has shown that caffeic acid phenethyl ester (CAPE), a bioactive component of honeybee propolis, was significantly more potent than zileuton (1) for the inhibition of LTs biosynthesis in human polymorphonuclear leukocytes (PMNL, IC50 = 0.52 μM) [17]. Several phenolic acid-based analogues have been developed based on modifications of the caffeic moiety or of the ester linkage, and some have shown improved inhibition of leukotriene biosynthesis [18][19][20][21]. Moreover, we have recently demonstrated that sinapic acid phenethyl ester ((SAPE) (2), Figure 1) was more potent than zileuton (1) as 5-LO inhibitor (PMNL, IC50 = 0.3 μM), which highlights the importance of methoxy and hydroxyl groups on the cinnamic acid moiety [21].
Zileuton's (1) benzo[b]thiophene moiety is a key pharmacophore that has been proven to act as an anti-inflammatory [22], antioxidant [23], anti-cancer [24], analgesic [22], and many more [23]. Therefore, several benzo[b]thiophene-based drugs have been developed over the years to treat numerous diseases [23]. N-hydroxyurea-Cetirizine hybrid, described by Lewis et al., showed histaminergic binding and 5-lipoxygenase inhibiting activities comparable to the corresponding Nhydroxyurea analog such as zileuton (1) [25]. Hydroxyurea alone is to date the only FDA approved drug for the treatment of adult patients with sickle cell disease through reduction of clinical complications [26].
Herein, we now report the synthesis, characterization, and the anti 5-LO activity of new zileutonhydroxycinnamic acid hybrids. Compounds that link zileuton's (1) benzo[b]thiophene subunit and hydroxycinnamic acid ester's two subunits were investigated ( Figure 2). Compounds that link zileuton's (1) hydroxyurea subunit and hydroxycinnamic acid ester's acid subunit were also investigated ( Figure 2). The influence of the methoxy and hydroxyl groups on the phenolic acid moiety was investigated by varying the number and position of the oxygen functions (hydroxy/methoxy groups). The influence of the α,β-unsaturation was also investigated.  Previous work from our team has shown that caffeic acid phenethyl ester (CAPE), a bioactive component of honeybee propolis, was significantly more potent than zileuton (1) for the inhibition of LTs biosynthesis in human polymorphonuclear leukocytes (PMNL, IC 50 = 0.52 µM) [17]. Several phenolic acid-based analogues have been developed based on modifications of the caffeic moiety or of the ester linkage, and some have shown improved inhibition of leukotriene biosynthesis [18][19][20][21]. Moreover, we have recently demonstrated that sinapic acid phenethyl ester ((SAPE) (2), Figure 1) was more potent than zileuton (1) as 5-LO inhibitor (PMNL, IC 50 = 0.3 µM), which highlights the importance of methoxy and hydroxyl groups on the cinnamic acid moiety [21].
Zileuton's (1) benzo[b]thiophene moiety is a key pharmacophore that has been proven to act as an anti-inflammatory [22], antioxidant [23], anti-cancer [24], analgesic [22], and many more [23]. Therefore, several benzo[b]thiophene-based drugs have been developed over the years to treat numerous diseases [23]. N-hydroxyurea-Cetirizine hybrid, described by Lewis et al., showed histaminergic binding and 5-lipoxygenase inhibiting activities comparable to the corresponding N-hydroxyurea analog such as zileuton (1) [25]. Hydroxyurea alone is to date the only FDA approved drug for the treatment of adult patients with sickle cell disease through reduction of clinical complications [26].
Herein, we now report the synthesis, characterization, and the anti 5-LO activity of new zileuton-hydroxycinnamic acid hybrids. Compounds that link zileuton's (1) benzo[b]thiophene subunit and hydroxycinnamic acid ester's two subunits were investigated ( Figure 2). Compounds that link zileuton's (1) hydroxyurea subunit and hydroxycinnamic acid ester's acid subunit were also investigated ( Figure 2). The influence of the methoxy and hydroxyl groups on the phenolic acid moiety was investigated by varying the number and position of the oxygen functions (hydroxy/methoxy groups). The influence of the α,β-unsaturation was also investigated.  Previous work from our team has shown that caffeic acid phenethyl ester (CAPE), a bioactive component of honeybee propolis, was significantly more potent than zileuton (1) for the inhibition of LTs biosynthesis in human polymorphonuclear leukocytes (PMNL, IC50 = 0.52 μM) [17]. Several phenolic acid-based analogues have been developed based on modifications of the caffeic moiety or of the ester linkage, and some have shown improved inhibition of leukotriene biosynthesis [18][19][20][21]. Moreover, we have recently demonstrated that sinapic acid phenethyl ester ((SAPE) (2), Figure 1) was more potent than zileuton (1) as 5-LO inhibitor (PMNL, IC50 = 0.3 μM), which highlights the importance of methoxy and hydroxyl groups on the cinnamic acid moiety [21].
Zileuton's (1) benzo[b]thiophene moiety is a key pharmacophore that has been proven to act as an anti-inflammatory [22], antioxidant [23], anti-cancer [24], analgesic [22], and many more [23]. Therefore, several benzo[b]thiophene-based drugs have been developed over the years to treat numerous diseases [23]. N-hydroxyurea-Cetirizine hybrid, described by Lewis et al., showed histaminergic binding and 5-lipoxygenase inhibiting activities comparable to the corresponding Nhydroxyurea analog such as zileuton (1) [25]. Hydroxyurea alone is to date the only FDA approved drug for the treatment of adult patients with sickle cell disease through reduction of clinical complications [26].
Herein, we now report the synthesis, characterization, and the anti 5-LO activity of new zileutonhydroxycinnamic acid hybrids. Compounds that link zileuton's (1) benzo[b]thiophene subunit and hydroxycinnamic acid ester's two subunits were investigated ( Figure 2). Compounds that link zileuton's (1) hydroxyurea subunit and hydroxycinnamic acid ester's acid subunit were also investigated ( Figure 2). The influence of the methoxy and hydroxyl groups on the phenolic acid moiety was investigated by varying the number and position of the oxygen functions (hydroxy/methoxy groups). The influence of the α,β-unsaturation was also investigated.

Synthesis
Esters 4-9, bearing the benzo[b]thiophene moiety, were obtained following a single step esterification of benzo[b]thiophene-2-carboxylic acid (3) with the appropriate alkyl bromide in the presence of sodium carbonate (Na 2 CO 3 ), potassium iodide (KI), and hexamethylphosphoramide (HMPA) as a solvent. As shown in Scheme 1, the length of the linker and the presence of a methoxy or hydroxyl group on the ester moiety was investigated. The 1 H-NMR analysis of this series shows the characteristic hydrogen of the benzo[b]thiophene (S-C=CH: 8 ppm) as well as the methylenes (4-2 ppm) of the linker ester.

Synthesis
Esters 4-9, bearing the benzo[b]thiophene moiety, were obtained following a single step esterification of benzo[b]thiophene-2-carboxylic acid (3) with the appropriate alkyl bromide in the presence of sodium carbonate (Na2CO3), potassium iodide (KI), and hexamethylphosphoramide (HMPA) as a solvent. As shown in Scheme 1, the length of the linker and the presence of a methoxy or hydroxyl group on the ester moiety was investigated. The 1 H-NMR analysis of this series shows the characteristic hydrogen of the benzo[b]thiophene (S-C=CH: 8 ppm) as well as the methylenes (4-2 ppm) of the linker ester. Compounds 11-31 having benzo[b]thiophene and α,β-unsaturated carbonyl moieties were synthesized through a base-catalyzed aldol condensation by reacting the appropriate aldehyde with 2-acetylbenzothiophene (10) (Scheme 2). Our best yields were obtained by using pyrrolidine as a base. All our attempts to obtain compounds 11 and 12 using the same method in the presence of pyrrolidine failed which can be explained by the instability/non-reactivity of benzaldehyde and 3methoxybenzaldehyde under such reaction conditions. Therefore, an alternative synthesis route for the aldol condensation using a stoichiometric quantity of sodium ethoxide in ethanol was used to obtain compounds 11 and 12 (Scheme 2). Unfortunately, all our attempts, whether with the first or the second method, failed to obtain dihydroxyl analogs. From the least substituted compound (11: R 2 -R 5 = H) to compounds variously substituted with hydroxyl and methoxy (12-31), a total of 21 hybrids were obtained in this subseries (Scheme 2). Structures of compounds  were confirmed by NMR spectroscopy. In all compounds, 1 H-NMR spectra showed the characteristic hydrogen of the benzo[b]thiophene (S-C=CH: 8 ppm). The characteristic α,β-unsaturation peaks were observed at 6-7 ppm. The value of the coupling constant (J = 16 Hz) confirms the trans stereochemistry of α,βunsaturation. Further, 13 C-NMR spectra also confirmed the presence of a carbonyl group by showing a peak near to 180 ppm.
An adaptation of the method described by Pariente-Cohen et al. [27] with our carboxylic acids allowed us to obtain the hydroxyurea hybrids 33-40 (Scheme 3). As described by Pariente-Cohen et al., N-acylation occurs through the reaction of hydroxylurea with our activated carboxylic acids. The acylation took place on the secondary nitrogen rather than on the hydroxyl group or the primary nitrogen. X-ray crystallography analysis of hydroxyurea reveals that the C-NH2 bond is shorter than the C-NHOH bond (1.328°A versus 1.347°A) [27,28] which suggest that the N of the -NHOH group is more nucleophilic since the C-NH2 contributes the dominant resonance structure and might have more sp2 hybridization. Compounds 11-31 having benzo[b]thiophene and α,β-unsaturated carbonyl moieties were synthesized through a base-catalyzed aldol condensation by reacting the appropriate aldehyde with 2-acetylbenzothiophene (10) (Scheme 2). Our best yields were obtained by using pyrrolidine as a base. All our attempts to obtain compounds 11 and 12 using the same method in the presence of pyrrolidine failed which can be explained by the instability/non-reactivity of benzaldehyde and 3-methoxybenzaldehyde under such reaction conditions. Therefore, an alternative synthesis route for the aldol condensation using a stoichiometric quantity of sodium ethoxide in ethanol was used to obtain compounds 11 and 12 (Scheme 2). Unfortunately, all our attempts, whether with the first or the second method, failed to obtain dihydroxyl analogs. From the least substituted compound (11: R 2 -R 5 = H) to compounds variously substituted with hydroxyl and methoxy (12-31), a total of 21 hybrids were obtained in this subseries (Scheme 2). Structures of compounds  were confirmed by NMR spectroscopy. In all compounds, 1 H-NMR spectra showed the characteristic hydrogen of the benzo[b]thiophene (S-C=CH: 8 ppm). The characteristic α,β-unsaturation peaks were observed at 6-7 ppm. The value of the coupling constant (J = 16 Hz) confirms the trans stereochemistry of α,β-unsaturation. Further, 13 C-NMR spectra also confirmed the presence of a carbonyl group by showing a peak near to 180 ppm.
An adaptation of the method described by Pariente-Cohen et al. [27] with our carboxylic acids allowed us to obtain the hydroxyurea hybrids 33-40 (Scheme 3). As described by Pariente-Cohen et al., N-acylation occurs through the reaction of hydroxylurea with our activated carboxylic acids. The acylation took place on the secondary nitrogen rather than on the hydroxyl group or the primary nitrogen. X-ray crystallography analysis of hydroxyurea reveals that the C-NH 2 bond is shorter than the C-NHOH bond (1.328 • A versus 1.347 • A) [27,28] which suggest that the N of the -NHOH group is more nucleophilic since the C-NH 2 contributes the dominant resonance structure and might have more sp2 hybridization.  To avoid any side reactions that would be due to the presence of the hydroxyls of some of our carboxylic acids, the hydroxyurea hybrids 34-37 were obtained following activation in the presence of BOP. The benzoyl analog (38), obtained by Pariente-Cohen et al. starting with benzoic anhydride [27], was obtained with acyl chloride with a moderate yield. With the same procedure, phenylacetyl chloride and hydroxycinnamoyl chloride gave the corresponding hydroxyurea analogs 39 and 40 (Scheme 3).   To avoid any side reactions that would be due to the presence of the hydroxyls of some of our carboxylic acids, the hydroxyurea hybrids 34-37 were obtained following activation in the presence of BOP. The benzoyl analog (38), obtained by Pariente-Cohen et al. starting with benzoic anhydride [27], was obtained with acyl chloride with a moderate yield. With the same procedure, phenylacetyl chloride and hydroxycinnamoyl chloride gave the corresponding hydroxyurea analogs 39 and 40 (Scheme 3). Whether activated by a peptide coupling agent (benzotriazol-1-yloxy)tris(dimethylamino) phosphonium hexafluorophosphate (BOP) in this case, or by transformation into acyl chloride, cinnamic acid gave the corresponding hydroxy urea analog 33 (Scheme 3).
To avoid any side reactions that would be due to the presence of the hydroxyls of some of our carboxylic acids, the hydroxyurea hybrids 34-37 were obtained following activation in the presence of BOP. The benzoyl analog (38), obtained by Pariente-Cohen et al. starting with benzoic anhydride [27], was obtained with acyl chloride with a moderate yield. With the same procedure, phenylacetyl chloride and hydroxycinnamoyl chloride gave the corresponding hydroxyurea analogs 39 and 40 (Scheme 3).

5-LO Product Biosynthesis Assays
All compounds, including known inhibitors zileuton (1) and SAPE (2), were assayed at 1 µM ( Figure 1) for inhibition of LTs biosynthesis in stimulated HEK293 cells. These cells are stably transfected with human 5-LO and serves as a highly reproducible model of 5-LO product biosynthesis in which compounds can be preliminarily screened for anti-LTs activity before moving on to more complex systems [17,29]. As shown in Figure 3, analogs bearing zileuton's (1) benzo[b]thiophene subunit combined with a phen-alkyl moiety through an ester linkage were inactive. Esterification of benzo[b]thiophene-2-carboxylic acid to obtain benzyl to phenbutyl analogs (4-7) had no effect on the inhibitory activity of 5-LO. All four esters were inactive. Even the addition of a hydroxyl (8) or methoxy (9) at para-position of the phenethyl group has no effect on the activity.

5-LO Product Biosynthesis Assays
All compounds, including known inhibitors zileuton (1) and SAPE (2), were assayed at 1 µ M ( Figure 1) for inhibition of LTs biosynthesis in stimulated HEK293 cells. These cells are stably transfected with human 5-LO and serves as a highly reproducible model of 5-LO product biosynthesis in which compounds can be preliminarily screened for anti-LTs activity before moving on to more complex systems [17,29]. As shown in Figure 3, analogs bearing zileuton's (1) benzo[b]thiophene subunit combined with a phen-alkyl moiety through an ester linkage were inactive. Esterification of benzo[b]thiophene-2-carboxylic acid to obtain benzyl to phenbutyl analogs (4-7) had no effect on the inhibitory activity of 5-LO. All four esters were inactive. Even the addition of a hydroxyl (8) or methoxy (9) at para-position of the phenethyl group has no effect on the activity. Analogs bearing zileuton's (1) benzo[b]thiophene subunit combined with a phenolic acid moiety through a α,β-unsaturated ketone had interesting results as shown in Figure 4a,b. The number and position of the oxygen functions (hydroxy/methoxy groups) seem to be critical for inhibition of 5-LO product biosynthesis in stimulated HEK293 cells (Figure 4a) Unsubstituted compound (11) and ortho, meta, and para-monosubstituted (12-16) analogs were less active than SAPE (2) (Figure 4a). Compounds (11), 14 (ortho-), and 16 (para-) were nonsignificantly different than Zileuton (1) (Figure 4a).
Assays for the inhibition of 5-LO product biosynthesis by hybrids (33-40) combining zileuton's (1) hydroxyurea subunit and hydroxycinnamic acid moieties show that the hydroxyurea subunit seems to not be critical for the inhibition of the biosynthesis of 5-LO products in stimulated HEK293 cells. As shown in Figure 5, the presence of hydroxyl and methoxy functions (OH and OCH 3 ) does not seem to have an effect. Analogues obtained by condensation of hydroxyurea with cinnamic (compound 33), ferulic (compound 34), isoferulic (compound 35), and dimethoxy cinnamic (compound 36) acids were all inactive. Even compound 37 obtained with sinapic acid (3,5-dimethoxy-4-hydroxy substitution) was inactive. Of note, compound 37 had the same sinapic acid (3,5-dimethoxy-4-hydroxy substitution) moiety as the best inhibitor, compound 28, obtained in series 2. This result shows that the zileuton's (1) benzo[b]thiophene moiety is essential for the inhibition of the biosynthesis of 5-LO products with our hybrids. The linker between the hydroxylated nitrogen and the phenyl does not seem to be crucial for the inhibition of 5-LO products since analogs bearing the hydroxycinnamic acid (40), acetyl benzoic acid (39), or benzoic acid (38) moieties were inactive ( Figure 5). Measuring the inhibition capacity of tri-substituted hybrids (26)(27)(28)(29) shows that the number and position of the oxygen functions (hydroxy/methoxy groups) seem to be critical for the inhibition 5-LO product biosynthesis in HEK293 cells (Figure 4b). While compound 27, bearing three methoxy moieties (positions 2, 4, and 6), was equipotent to Zileuton (1) (Figure 4b), compound 26 having three methoxy moieties (positions 3, 4, and 5) was inactive.
Assays for the inhibition of 5-LO product biosynthesis by hybrids (33-40) combining zileuton's (1) hydroxyurea subunit and hydroxycinnamic acid moieties show that the hydroxyurea subunit seems to not be critical for the inhibition of the biosynthesis of 5-LO products in stimulated HEK293 cells. As shown in Figure 5, the presence of hydroxyl and methoxy functions (OH and OCH3) does not seem to have an effect. Analogues obtained by condensation of hydroxyurea with cinnamic (compound 33), ferulic (compound 34), isoferulic (compound 35), and dimethoxy cinnamic (compound 36) acids were all inactive. Even compound 37 obtained with sinapic acid (3,5-dimethoxy-4-hydroxy substitution) was inactive. Of note, compound 37 had the same sinapic acid (3,5dimethoxy-4-hydroxy substitution) moiety as the best inhibitor, compound 28, obtained in series 2. This result shows that the zileuton's (1) benzo[b]thiophene moiety is essential for the inhibition of the biosynthesis of 5-LO products with our hybrids. The linker between the hydroxylated nitrogen and the phenyl does not seem to be crucial for the inhibition of 5-LO products since analogs bearing the hydroxycinnamic acid (40), acetyl benzoic acid (39), or benzoic acid (38) moieties were inactive ( Figure 5).  To further probe the inhibitory activity of compounds that approached the inhibitory activity of zileuton (1) and SAPE (2), compounds were selected for further screening in PMNL ( Figure 6). 5-LO is highly expressed in PMNL and these cells are important physiological producers of LTB 4 [17]. Compound 28 was shown to be the best inhibitor exceeding zileuton's (1) inhibitory capacity in PMNL by more than three times. The calculated IC 50 value for the inhibition of 5-LO product biosynthesis in PMNL for compound 28 was 0.37 µM (95% confidence interval: 0.25-0.53 µM). The presence of the hydroxyl group in position 4 and the methoxy groups in positions 3 and 5 seems to be an ideal combination for inhibition of 5-LO ( Figure 6). These structural changes may influence the availability, cell permeation, or even the stability of inhibitors in PMNL versus HEK293 cells. A comparison of the two monosubstituted analogs, compounds 14 and 16, reveals that hydroxyl at position 4 (16) seems to be more favorable for the inhibition of 5-LO, although neither compound showed significant inhibition compared to controls. As in HEK293 cells, disubstituted compounds 17 and 18 were less active than zileuton (1) and SAPE (2) (Figure 6). To further probe the inhibitory activity of compounds that approached the inhibitory activity of zileuton (1) and SAPE (2), compounds were selected for further screening in PMNL ( Figure 6). 5-LO is highly expressed in PMNL and these cells are important physiological producers of LTB4 [17]. Compound 28 was shown to be the best inhibitor exceeding zileuton's (1) inhibitory capacity in PMNL by more than three times. The calculated IC50 value for the inhibition of 5-LO product biosynthesis in PMNL for compound 28 was 0.37 μM (95% confidence interval: 0.25-0.53 μM). The presence of the hydroxyl group in position 4 and the methoxy groups in positions 3 and 5 seems to be an ideal combination for inhibition of 5-LO ( Figure 6). These structural changes may influence the availability, cell permeation, or even the stability of inhibitors in PMNL versus HEK293 cells. A comparison of the two monosubstituted analogs, compounds 14 and 16, reveals that hydroxyl at position 4 (16) seems to be more favorable for the inhibition of 5-LO, although neither compound showed significant inhibition compared to controls. As in HEK293 cells, disubstituted compounds 17 and 18 were less active than zileuton (1) and SAPE (2) (Figure 6). A position interchange between the hydroxyl at position 4 and a methoxy at position 5 of compound 28, as in compound 31, leads to a total loss of activity ( Figure 6). The substitution of the hydroxyl at position 4 of compound 28 by a methoxy, as in compound 27, also leads to a complete loss of the 5-LO inhibition ( Figure 6). The inactivity of hybrid 37, as in HEK293 cells, shows that the zileuton's (1) benzo[b]thiophene moiety is essential for the inhibition of 5-LO product biosynthesis with our hydrids.

Free Radical Scavenging Activity Assay
A mechanism by which 5-LO activity can be inhibited is through reductive inhibition of the ferric non-heme iron of the enzyme. Interactions with stable free radicals provides an evaluation of the reducing ability of the test compounds. Radical scavenging activities of compounds tested in stimulated PMNL were assayed using 2,2-Diphenyl-1-picrylhydrazyl (DPPH) as a stable radical and are expressed as IC50 concentrations in Table 1. Mono-and di-substituted analogs (14, 16, 17, and 18), which were non-significantly different than zileuton (1) in stimulated HEK293 cells (Figure 4a) had a weak anti-free radical activity compared to that of SAPE (2) or vitamin C but more important than A position interchange between the hydroxyl at position 4 and a methoxy at position 5 of compound 28, as in compound 31, leads to a total loss of activity ( Figure 6). The substitution of the hydroxyl at position 4 of compound 28 by a methoxy, as in compound 27, also leads to a complete loss of the 5-LO inhibition ( Figure 6). The inactivity of hybrid 37, as in HEK293 cells, shows that the zileuton's (1) benzo[b]thiophene moiety is essential for the inhibition of 5-LO product biosynthesis with our hydrids.

Free Radical Scavenging Activity Assay
A mechanism by which 5-LO activity can be inhibited is through reductive inhibition of the ferric non-heme iron of the enzyme. Interactions with stable free radicals provides an evaluation of the reducing ability of the test compounds. Radical scavenging activities of compounds tested in stimulated PMNL were assayed using 2,2-Diphenyl-1-picrylhydrazyl (DPPH) as a stable radical and are expressed as IC 50 concentrations in Table 1. Mono-and di-substituted analogs (14, 16, 17, and 18), which were non-significantly different than zileuton (1) in stimulated HEK293 cells (Figure 4a) had a weak anti-free radical activity compared to that of SAPE (2) or vitamin C but more important than that of zileuton (1). Among the three tri-substituted hybrids tested, compounds 27 and 31 were the most active with IC 50 of 6.0 µM and 6.6 µM, respectively (Table 1). Anti-radical activity of selected compounds indicates that this ability does not appear to be crucial for 5-LO inhibition by the tested compounds. Indeed, even though it was the best inhibitor of the whole series for the inhibition of 5-LO product biosynthesis in stimulated HEK293 and human PMNL cells, compound 28 had weak anti-free radical activity compared to SAPE (2) and vitamin C ( Table 1). Compounds 31 and 37, bearing zileuton's (1) hydroxyurea subunit instead of benzo[b]thiophene subunit as compound 28, were among the best compounds with antiradical activity (Table 1) while inactive for the inhibition of the biosynthesis of 5-LO products in both stimulated HEK293 and human PMNL cells.

Molecular Docking
To predict, at a molecular level, the possible interactions between 5-LO and the best ligand reported in this study, compound 28 and its related analogs 31 and 37 were docked into 5-LO (PDB ID: 3O8Y [30]). The unmodified apo structure of 5-LO, 3O8Y. "Stable-5-LO" is mutated mostly on the non-catalytic domain with a small exception of a short 3 residues sequence being mutated on the catalytic domain [30]. While these mutations may affect the structure, test results showed that catalytic fidelity was conserved [30]. This crystallized structure was used by many groups as well as our group for molecular modeling [19][20][21]31]. Zileuton (1) and SAPE (2) were docked as references. All docked compounds had better affinity for the active site than zileuton (1) and SAPE (2). As reported by De Lucia et al. [32] and demonstrated by our group recently [21], zileuton (1) was located in the inner part of the binding pocket and is stabilized by hydrogen bonds with Leu420, Ala424, Phe421, and Asn425 (Table 2). SAPE (2) was stabilized by the formation of a π-π interaction with His372 [21]. With a lower predicted binding energy than zileuton (1), compound 28 had a better affinity for 5-LO than zileuton (1) ( Table 2). Compound 28, as with SAPE (2), does not form any hydrogen bonds with the 5-LO pocket, unlike to zileuton (1). However, compound 28 forms two π-π interactions with Phe177 and Phe421, which can explain its superiority as an inhibitor compared to zileuton (1). Compound 28 was the only docked ligand that showed a π-π interaction with Phe177 which may play a part in product specificity [33].
Interestingly, zileuton's (1) benzo[b]thiophene moiety of compound 28, the most active inhibitor in human PMNL cells, is in close proximity to Leu420, Ala424, and Asn425 (Figure 7). This binding site has been reported previously for fungal 5-LO inhibitors [34]. Unlike compound 28, compound 37 forms three hydrogen bonds with Tyr181, Gln363, His367, and Asn425. The last amino acid was also involved in a hydrogen bond with zileuton (1). Given its weak inhibitory activity of 5-LO in stimulated HEK293 and human PMNL cells, this implies that these interactions are not favorable for an optimal inhibition. With a lower predicted binding energy than zileuton (1), compound 28 had a better affinity for 5-LO than zileuton (1) ( Table 2). Compound 28, as with SAPE (2), does not form any hydrogen bonds with the 5-LO pocket, unlike to zileuton (1). However, compound 28 forms two π-π interactions with Phe177 and Phe421, which can explain its superiority as an inhibitor compared to zileuton (1). Compound 28 was the only docked ligand that showed a π-π interaction with Phe177 which may play a part in product specificity [33].
Interestingly, zileuton's (1) benzo[b]thiophene moiety of compound 28, the most active inhibitor in human PMNL cells, is in close proximity to Leu420, Ala424, and Asn425 (Figure 7). This binding site has been reported previously for fungal 5-LO inhibitors [34]. Unlike compound 28, compound 37 forms three hydrogen bonds with Tyr181, Gln363, His367, and Asn425. The last amino acid was also involved in a hydrogen bond with zileuton (1). Given its weak inhibitory activity of 5-LO in stimulated HEK293 and human PMNL cells, this implies that these interactions are not favorable for an optimal inhibition.

General Synthetic Experimental Procedures
Chemicals were purchased from Sigma-Aldrich and Alfa Aesar. The synthesized compounds were purified by flash chromatography (Isco, Inc. CombiFlash ® Sg100c). Thin layer chromatography (TLC) was carried out on silica gel-coated aluminum sheets (EMD Millipore TM TLC Silica Gel 60) with detection by UV light (245 nm, UVP ® UVS-14 EL Series UV lamp). Melting points were obtained with a melting point apparatus (MELTEMP ® 1001D). NMR spectra were recorded on Bruker ® Avance III 400 MHz spectrometer with TMS as an internal standard. High-resolution mass measurements were performed on Bruker ® Doltonics' micrOTOF instrument in positive or negative electrospray. Analytical high-performance liquid chromatography (HPLC) was performed on an Agilent Technologies system (Agilent1100 Series) with an ACE C18 column (150 mm

General Procedure for the Aldol Condensation of 2-acetylbenzothiophene in Presence of Sodium Ethoxide
To a solution of 2-acetylbenzothiophene (10) (1 eq) and the appropriate aldehyde (1.2 eq) in ethanol (15 mL) was added sodium ethoxide (1 eq). The mixture was then stirred at room temperature for 2 h. After concentration under vacuum, EtOAc (50 mL) was added. The resulting solution was stirred with sodium bisulfite solution (100 mL, 120 g/500 mL) for 30 min to remove any traces of starting aldehyde. After separation, the combined organic phase was washed with brine, decolorized with activated charcoal, dried over MgSO 4 , filtered, and concentrated under vacuum. The resulting crud product was purified by flash chromatography. To a stirred solution of 2-acetylbenzothiophene (10) (1.2 eq) and the appropriate aldehyde (1 eq) in 10 mL of THF was added 100 µL of pyrrolidine and a catalytic amount of acetic acid (4 drops). The solution was refluxed under argon atmosphere and monitored by TLC. Solvent was removed under vacuum, Water (50 mL) was added followed by extraction with EtOAc (3 × 25 mL). The combined organic fractions were then combined and stirred with sodium bisulfite solution (100 mL, 120 g/500 mL) for 30 min to remove any traces of starting aldehyde. After separation, the combined organic fractions were washed with brine, dried over MgSO 4 , decolorized with activated charcoal, filtered, and concentrated under vacuum. The resulting crud product was purified by flash chromatography.
Each mixture was then shaken vigorously and held in the dark for 30 min at room temperature. The absorbance of DPPH at 517 nm was then measured. The radical scavenging activity was expressed in terms of % inhibition of DPPH absorbance: where A control is the absorbance of the control (DPPH solution without test compounds) and A test is the absorbance of the test sample (DPPH solution plus compounds). Ascorbic acid was used as a positive control. Data were expressed as mean ± SEM of three independent experiments, each performed in triplicate. IC 50 values were calculated from a sigmoidal concentration-response curve-fitting model with a variable slope on GraphPad Prism software.

Molecular Docking
Docking simulations were performed with AutoDock Vina 1.1.2 [37] on 5-LO (PDB ID: 3O8Y [30]) according to the reported procedure [32]. The AutoDock Vina parameters were set as follows; box size: 30 × 20 × 30 Å, the center of box: x = 4.049, y = 21.347, z = −0.284, the exhaustiveness: 8, and the other parameters were left unchanged. The structures of all the compounds were optimized using the MMFF94 force field. The calculated geometries were ranked in terms of free energy of binding and the best pose was selected for further analysis. Molecular visualization was performed by with Maestro 11.7 [38] and LigPlot + 2.1 [39].

Statistical Analysis
Statistical analyses and graph design were performed using GraphPad Prism 5 software (GraphPad Software, San Diego, CA, USA).

Conclusions
Thirty-five zileuton-hydroxycinnamic acid hybrids were synthesized and evaluated for their 5-LO inhibition activities in stimulated HEK293 cells. To further probe the inhibitory activity of compounds that approached the inhibitory activity of zileuton (1) and SAPE (2), eight compounds were selected for further screening in PMNL. Compounds bearing zileuton's (1) benzo[b]thiophene subunit combined with a phenolic acid moiety through a α,β-unsaturated ketone had interesting results. The number and position of the oxygen functions (hydroxy/methoxy groups) seem to be critical for inhibition of 5-LO product biosynthesis. Compound 28, having zileuton's (1) benzo[b]thiophene and SAPE's (2) α,β-unsaturated 3,5-dimethoxy-4-hydroxyphenolic acid moieties, was three times as active as Zileuton (1) for the inhibition of 5-LO in PMNL. Even with a low antiradical potency, compound 28 was found to be the best 5-LO inhibitor.
The 3,5-dimethoxy-4-hydroxy tri-substitution seems to be the most favorable for good inhibition. Indeed, neither the 2,6-dimethoxy-4-hydroxy or the 3,4-dimethoxy-5-hydroxy substitution appears to be as effective. Compounds 30 and 31 having these substitution patterns were less active than compound 28. Unlike zileuton (1), compound 28 forms π-π interactions with both Phe177 and Phe421 as predicted when docked into 5-LO. Such interactions, in particular with Phe177, may explain the activity of this molecule compared to zileuton (1) or SAPE (2). Further investigations will be necessary to elucidate the mode of action and confirm the implication of this interaction.

Conflicts of Interest:
The authors declare no conflict of interest.