1. Introduction

Aminopeptidases represent a class of zinc-containing metal loenzymes responsible for the cleavage of amino acids from the N-terminus of various peptides or proteins [1–3]. The relationship between aminopeptidases and cancer was first reported in the late 1950s [4, 5], which attracted the attention of many scientists in the following dozen years. The first aminopeptidase inhibitor bestatin was discovered in 1976 (Fig. 1) [6], and it was subsequently used as an effective anti-cancer agent. Notwithstanding many other aminopeptidase inhibitors in clinical trials, bestatin is so far the only clinically approved one in Japan used for anti-tumor treatment [7]. Recently, it is suggested that aminopeptidases play significant roles in regulating the supply of cellular free amino acids to maintain the survival and proliferation of cancer cells [8]. Especially, cancer cells are dependent on some specific amino acids and the deprivation of these amino acids may provide a simple, effective and non-genotoxic strategy for anticancer therapy [9]. Thus, the development of novel aminopeptidase inhibitors for the treatment of cancer is still of tremendous potential.

Five key aminopeptidases—aminopeptidase N (APN), leucine aminopeptidase (LAP), puromycin-sensitive aminopeptidase (PuSA), leukotriene A₄ hydrolase (LTA₄H) and endoplasmic reticulum aminopeptidase 1/2 (ERAP1/2) are now deemed to have close correlation with a wide range of cancers [8]. LTA₄H, a monomeric zinc-containing soluble protein, acts as a dual-function enzyme with both epoxide hydrolase and aminopeptidase activi ties [10, 11]. In the 5-lipoxygenase pathway of arachidonic acid metabolism, LTA₄H catalyzes the conversion of unstable epoxide Leukotriene A₄ to diol Leukotriene B₄ [12], which is associated with a variety of inflammatory diseases [13–15]. As an epoxide hydrolase, LTA₄H has long been a crucial target in anti inflammatory drug discovery. Besides, LTA₄H exhibits a significant sequence homology to some other zinc peptidases such as aminopeptidase M, aminopeptidase B, and thermolysin [3, 16]. On the basis of its zinc binding motif—HEXXHX18E [17, 18], LTA₄H was proved to possess the peptide-cleaving activity as a member of the M₁ family of zinc aminopeptidases [19]. To the best of our knowledge, there is currently rare effective aminopeptidase inhibitors specifically acting on LTA₄H. Therefore, the search for novel LTA₄H aminopeptidase inhibitors to obtain potential anticancer agents is valuable.

1, 2, 3-Triazin-4-one represents a privileged structure in medici nal chemistry, which serves as a unique building block for the construction of many compounds with anti-cancer activity (Fig. 1, 1–4). Compounds 2–4 were reported to have matrix metalloproteinase (MMP) inhibitory activity [20–22]. Among them, compound 3 showed good activity in vivo and had a potential for further development [22]. It is well-known that MMPs are also zinc-containing metalloenzymes and share a similar catalytic mechanism with LTA₄H. So, we consider that 1, 2, 3-triazin-4-one may be used as a possible pharmacophore in our design of novel LTA₄H aminopeptidase inhibitors. Furthermore, 4-thiazolidinone is another important motif found in a number of bioactive compounds [23]. For example, compound 5 could selectively kill both non-small cell lung cancer cell line H₄₆₀ and its paclitaxel resistant variant H460taxR [24], and compound 6 showed promising cytotoxicity against renal cancer cell lines (UO-31, TK-10, and ACHN) [25]. More importantly, due to LTA₄H has an L-shaped substrate binding pocket [26], we speculate that a flexible carbon chain of appropriate length may be necessary to make the entire molecule fit the binding pocket well. Thereupon, the above two fragments were combined by a flexible carbon chain to design our target molecule (Fig. 2). Thirty-three novel 1, 2, 3-benzotriazin-4-one derivatives with 4-thiazolidinone moieties were herein synthesized and their bioactivity evaluation against LTA₄H aminopeptidase was performed. The influence of linker length and substituent to inhibitory activity was also investigated.

2. Results and discussion 2.1. Synthesis

The synthetic procedure of the target compounds Ⅳ-1–32 is illustrated in Scheme 1. Using various substituted anthranilamides Ⅰ-1–11 as starting materials, substituted 1, 2, 3-benzotriazin-4-ones Ⅱ-1–11 could be easily prepared in one pot via diazotization, nucleophilic addition and cyclization [27] in yields of above 80%. Subsequently, in the presence of potassium carbonate, Ⅱ-1–11 reacted with dibromoalkane (2–6 carbons) in acetone to obtain 3-bromoalkane-1, 2, 3-benzotriazin-4-ones [28] Ⅲ-1–33 in yields of 46%–62%. As an alkylation agent, dibromoalkane exhibits two reaction sites, but we obtained the N-alkylation of compound Ⅲ-1-33, N-3 alkylated ones were the main products. Clark et al. predicted 1, 2, 3-benzotriazin-4(3H)-one to be the most stable of its three tautomers [27]. To confirm the structures of N-3 alkylated products, 2D NMR (HMBC) was performed. From the HMBC spectrum (500 MHz) of 3-(3-bromopropyl)-7-chlorobenzo[d] [1, 2, 3]triazin-4(3H)-one, it can be seen that the methylene (H-5) is strongly related to the carbonyl group (C-4) of benzotriazinone but not related to the benzene ring (The corresponding spectra are deposited in Supporting information). Because dibromoalkane can readily react with NH at the 3-position of 1, 2, 3-benzotriazin-4-one to afford a N3, N3'-(alkane-1, 3-diyl)bis(benzo[d][1, 2, 3]triazin-4 (3H)-one), to minimize the incidence of this side reaction, different amounts of dibromoalkane were tried. It was found that when dibromoalkane was of excess up to 5 equiv., the generation of the N3, N3'-(alkane-1, 3-diyl)bis(benzo[d][1, 2, 3]triazin-4(3H)-one)

could be well inhibited. Then the target compounds Ⅳ-1–33 were synthesized via the reaction of 3-bromoalkane-1, 2, 3-benzotriazin-4-ones Ⅲ-1–33 with 2-cyanoimino-4-oxothiazolidine potassium salt. The addition of potassium iodide was profitable to the above nucleophilic substitution owing to better leaving ability of iodide ion than bromide ion. The chemical structures of Ⅳ-1–33 were characterized by ¹H NMR, ¹³C NMR, ¹⁹F NMR and HRMS (see Supporting information).

2.2. Biological activity

The LTA₄H aminopeptidase inhibitory activities of compounds Ⅳ-1–33 were evaluated in vitro at a con-centration of 10 μmol/L by the method reported in the literature [29]. Bestatin was used as an assay positive control. The assay results are summarized in Table 1. It was found that all compounds exhibited certain inhibitory activities against LTA₄H aminopeptidase. Many compounds showed > 50% inhibition rates and their IC₅₀ values were given.

As shown in Table 1, some compounds with a 2-carbon chain showed only < 20% inhibitory activities (e.g., Ⅳ-1, Ⅳ-3 and Ⅳ-5), which were much lower than those of their corresponding compounds with a 3-carbon chain (e.g., Ⅳ-1 vs. Ⅳ-2, Ⅳ-3 vs. Ⅳ-4, Ⅳ-5 vs. Ⅳ-6). It seemed that a shorter linker such as a 2-carbon chain was not beneficial to inhibitory activity. When the linker was fixed as a 3-carbon chain, it was found that the introduction of electron-withdrawing groups (CF₃, NO₂), electron-donating groups (CH₃) and halogen (F, Cl, Br) into the benzotriazinone ring was profitable to inhibitory activity. For example, compounds Ⅳ-4 and Ⅳ-6–9 were more active than Ⅳ-2 with inhibition rates of above 30.2%. Among them, Ⅳ-6 (7-Cl) was the most potent with an IC₅₀ of 7.39 ± 0.19 μmol/L. Considering the L-shaped pocket of LTA₄H, the influence of linker length on inhibitory activity was investigated. When the length of linker was changed from 2 to 6 carbons, it was found that the activities did not increase following a certain regularity (e.g., Ⅳ-3, Ⅳ-4, Ⅳ-11–13, Ⅳ-5, Ⅳ-6, Ⅳ-14–16). However, when the length of linker was changed from 4 to 6 carbons, the majority of compounds with different substituents showed the increasing inhibitory potencies with the increasing of linker length except compounds with 7-Br (e.g., Ⅳ-11–13, Ⅳ-14– 16, Ⅳ-20–22, Ⅳ-23–25, Ⅳ-26–28). For compounds with the same substituents at the same positions on the benzotriazinone ring, a 6-carbon chain length accompanied higher activities than other chain lengths except compounds Ⅳ-19 and Ⅳ-28. For example, compounds Ⅳ-13, Ⅳ-16 and Ⅳ-25 with a 6-carbon chain had > 50% inhibitory activities with the IC₅₀ values of 7.70 ± 1.65, 1.30 ± 0.20, and 7.19 ± 1.30 μmol/L, respectively.

Furthermore, the influence of substituted position on inhibitory activity was also investigated. When the linker length was increased from three carbons to a certain extent, Cl or Br located at 7-position on the benzotriazinone ring always seemed to provide higher potency. Notably, compound Ⅳ-16 (7-Cl) with a 6-carbon chain displayed 80.6% inhibitory activity with an IC₅₀ of 1.30 ± 0.20 μmol/L, a little higher than Ⅳ-18 (IC₅₀ = 1.71 ±0.18 μmol/L). It suggested that Cl group at 7-position might lead to higher activity. To verify this, the linker was fixed as a 6-carbon chain, and compounds with electron-withdrawing NO₂, and electron-donating OCH₃ at 7-postion were synthesized. As expected, compound Ⅳ-16 (7-Cl) was still the most potent. In addition, compounds with Cl at other three positions (5-, 6-, and 8-position) were also synthesized to examine the relationship between substituted position and inhibitory activity. The data showed that the inhibition rates of compounds with 5-and 8-Cl (Ⅳ-30, Ⅳ-31) were lower than 50%, and the IC₅₀ of the compound with 6-Cl (Ⅳ-29) reached to 3.87 ± 0.78 mM, still lower than that of the compound with 7-Cl (IC₅₀ = 1.30 ± 0.20 μmol/L). The com pound with 7-Cl exhibited the highest inhibitory activity among four compounds with a 6-carbon chain and differently positioned Cl (e.g., Ⅳ-16, Ⅳ-29–31).

The compounds which LTA₄H aminopeptidase inhibitory IC₅₀ better than 10 μmol/L were selected to test the proliferation inhibitory activities in THP1 human AML cell line. MTS assay showed the best three compounds in LTA₄H aminopeptidase inhibition also had the better cell proliferation inhibitory activities, compound Ⅳ-16 and Ⅳ-18 were the potent compounds with 80.6% and 78.8% inhibitory activity (IC₅₀ = 1.30 ± 0.20 μmol/L, 1.71 ±0.18 μmol/L, respectively), which both showed similar proliferation inhibitory activity as bestatin in THP1 human AML cell line.

Docking study was used to explore the binding mode between the LTA₄H and compound Ⅳ-16. The result of docking analysis showed that Ⅳ-16 was well docked in the active grooves of this enzyme (Fig. 3). The compound interactions with the residues in the active pocket of LTA₄H mainly depending on hydrogen bonds. For example, compound Ⅳ-16 could form hydrogen bonds with the polar residues Tyr267, Ala377, Leu365 and Asp312 around the active site, which indicated that the 1, 2, 3-triazin-4-one and 4-thiazolidinone scaffolds are necessary to the inhibitory properties of compounds. And the linker of 6-carbon chain also contribute to compound Ⅳ-16 more flexible and close into the active pocket. These results are consistent with our original design.

From the results above, it can be concluded that 7-Cl and a longer linker length were indeed profitable to the inhibitory activity, and compound Ⅳ-16 was promising for further structural modification. Besides, a 6-carbon chain length possibly made the entire ligand molecule competitively bind the L-shaped pocket of LTA₄H well and finally inhibited the aminopeptidase activity of LTA₄H. Further study is now in progress.

3. Conclusion

In conclusion, a series of novel 1, 2, 3-benzotriazin-4-one derivatives were designed, synthesized and characterized. Their LTA₄H aminopeptidase inhibitory activities were evaluated in vitro at the concentration of 10 μmol/L. Many compounds exhibited moderate to good inhibitory activities. In particular, compound Ⅳ-16 was the most potent with 80.6% inhibitory activity (IC₅₀ = 1.30 ± 0.20 μmol/L), which showed similar proliferation inhibitory activity as bestatin in THP1 human AML cell line. Our results indicated that 1, 2, 3-benzotriazin-4-one scaffold was promising to study further as a lead structure for the development of LTA₄H aminopeptidase inhibitors.

4. Experimental 4.1. Chemicals and instruments

All reagents and solvents were of AR grade orpurifiedby standard methods before use. All melting points (mp) were determined on a Büchi melting point B540 apparatus (Büchi Labortechnik AG, Flawil, Switzerland) and are uncorrected. ¹H NMR, ¹³C NMR and ¹⁹F NMR spectra were recorded in DMSO-d₆ on a Bruker AM-400 spectrome ter at ambient temperature (¹H 400 MHz; ¹³C 100 MHz; ¹⁹F 376 MHz). Chemical shifts are reported in ppm values with tetramethylsilane (TMS) as the internal standard. Coupling con stants (J) are reported in hertz. High-resolution mass spectra (HR MS) were recorded under electro-spray ionization condition on a WatersMicromass LC-TOF spectrometer. Reactions weremonitored by thin-layer chromatography (TLC) performed on silica plates (silica gel 60 F254) using UV-light.

4.2. General synthetic procedure for target compounds

To a mixture of 3-bromoalkyl-1, 2, 3-benzotriazin-4-one (Ⅲ, 2 mmol), and potassium iodide (0.332 g, 2 mmol) in DMF (15 mL) was added 2-cyanoimino-4-oxothiazolidine potassium salt (0.358 g, 2 mmol), and the mixture was stirred at 100 ℃. After the reaction was completed (monitored by TLC), the solvent was removed under reduced pressure to afford a dark brown oil, which was washed with water (50 mL) and extracted with CH₂Cl₂ (40 mL × 2). The organic layer was dried over anhydrous Na₂SO₄, concentrated and purified by flash chromatography on silica gel, eluting with petroleum ether (60–90 ℃)/EtOAc to afford the target compound Ⅳ. The general procedures for the synthesis of intermediates Ⅱ, Ⅲ and the spectra data of compounds Ⅳ-1–33 are listed in Supporting information.

N-(3-(4-(7-Bromo-4-oxobenzo[d][1, 2, 3]triazin-3(4H)-yl)bu tyl)-4-oxothiazolidin-2-ylidene)cyanamide (Ⅳ-16) (7-Br, 4C): yield, 10%; mp 168.5–168.8 ℃; ¹H NMR (400 MHz, DMSO-d₆): δ 8.46 (s, 1H), 8.16 (d, 1H, J = 8.4 Hz), 8.09 (d, 1H, J = 8.4 Hz), 4.38 (t, 2H, J = 6.4 Hz), 4.30 (s, 2H), 3.64 (t, 2H, J = 6.4 Hz), 1.89–1.75 (m, 2H), 1.69–1.56 (m, 2H); ¹³C NMR (100 MHz, DMSO-d₆): δ 179.7, 173.1, 154.9, 145.0, 136.3, 130.7, 129.2, 127.3, 118.9, 114.2, 49.2, 42.8, 36.0, 25.8, 24.0. HRMS (ESI) calcd. for C₁₅H₁₂N₆O₂S⁷⁹Br (M-H)^-, 418.9926, found, 418.9929; calcd. for C₁₅H₁₂N₆O₂S⁸¹Br (M-H)^-, 420.9905, found, 420.9900.

4.3. LTA₄H aminopeptidase assay

The LTA₄H aminopeptidase inhibitor activity assay wasmeasured by monitoring the hydrolysis of Ala-p-NA as previously described [30]. Briefly, indicated compounds were incubated in reaction buffer (50 mmol/L Tris, 100 mmol/L NaCl, pH 7.5) plus LTA₄H protein (500 ng)for10 minin96-wellplatesat roomtemperature.Inthenext step, Ala-p-NA was added to initiate the reaction and recorded the the formation of p-NA in the absorption at 405 nm keeping about 10 min using Perkin Elmer Multimode Plate Reader (Evision). The IC₅₀ values were determined from the results of at least three independent tests and calculated by the GraphPad5.

4.4. MTS assay

The antiproliferative activity of indicated compounds was measured by MTS assay as previously described [31]. The THP1 human AML cell line (5.0 ×10³) were seeded into 96-well plates in 200 mL RPMI-1640 medium supplemented with 10% FBS and penicillin/streptomycin. Subsequently, different concentra tions of indicated compounds were added and incubated for another 48 h at 37 ℃ in a 5% CO₂-containing atmosphere. After that, 20 mL of MTS Solution was added and incubated for 4 h, the optical density was measured at 490 nm using Perkin Elmer Multimode Plate Reader. The experiment was carried out at least three times.

4.5. Molecular docking

To better understand the structural basis for the inhibitory activities of the inhibitors against LTA₄H, we select the most potency inhibitor (Ⅳ-16) to study the binding model with LTA₄H enzyme by molecular docking. The crystal structure of 4L2L was downloaded from the Protein Data Bank, then all the bound ligands and water molecules of 4L2L were removed from the complex crystal structure. Subsequently, Hydrogen atoms and charges were added using the 'AutoDockTools' and the indicated compound was located into the active pockets of LTA₄H. The docking was conducted according to the instruction of AutoDock Vina described previously [32], then to select the best pose according to the scoring function. The result of docking was showed by PyMol.

Acknowledgments

This work was financially supported by the Na tional Natural Science Foundation of China (No. 21272071), and the Fundamental Research Funds for the Central Universities. The authors also appreciate the partial support from Shanghai Foundation of Science and Technology (No. 15431902100).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cclet.2016.12.014.