BEZ235

Design, synthesis, and antitumor evaluation of quinoline-imidazole derivatives

Zhen Xiao1 | Fei Lei1 | Xiuying Chen2,3 | Xiaolei Wang1 | Lujie Cao1 |Kejun Ye1 | Wufu Zhu1 | Shan Xu1

Abstract

A series of compounds bearing quinoline-imidazole (8a–e, 9a–e, 10a–e, 11a–e, and 12a–e) not reported previously were designed and synthesized. The target compounds were evaluated for antitumor activity against A549, PC-3, HepG2, and MCF-7 cells by the MTT method, with NVP-BEZ235 being the positive control. Most compounds showed moderate activity and compound 12a showed the best activity against HepG2, A549, and PC-3 cells, with half-maximal inhibitory concentration (IC50) values of 2.42 ± 1.02 µM, 6.29 ± 0.99 µM, and 5.11 ± 1.00 µM, respectively, which was equal to NVP-BEZ235 (0.54 ± 0.13 µM, 0.36 ± 0.06 µM, 0.20 ± 0.01 µM). Besides, the IC50 value of 12a against the cell line WI-38 (human fetal lung fibroblasts) was 32.8 ± 1.23 µM, indicating that the target compounds were selective for cancer cells. So, 11a and 12a were evaluated against PI3Kα and mTOR to find out if the compounds acted through the PI3K-Akt-mTOR signal transduction pathway. The inhibition ratios to PI3Kα and mTOR were slightly lower than that of NVP-BEZ235, suggesting there may be some other mechanisms of action. The structure–activity relationships and docking study of 11a and 12a revealed that the latter was superior. Moreover, the target compounds showed better in vitro anticancer activity when the C-6 of the quinoline ring was replaced by a bromine atom.

K E Y W O R D S
cytotoxic activity, docking study, mTOR, PI3Kα, quinoline-imidazole

1 | INTRODUCTION

The signal pathway of PI3K (phosphoinositide 3-kinase)-Akt (protein kinase B)-mTOR (rapamycin target protein) is one of the hotspots in the study of therapeutic drugs. PI3K-Akt-mTOR[1] signal transduction pathway plays an important role in tumor cell cycle regulation, apoptosis, transcriptions, translations, metabolisms and angiogenesis, protein synthesis, and many other important biological processes. A number of studies have been reported on PI3K/mTOR dual inhibitors. In contrast to single target small molecule inhibitors, the inhibitors acting on both PI3K and mTOR targets inhibit cell proliferation and promote cell apoptosis, and have other advantages such as containing a small dose, high titer, and difficult to produce drug resistance.
Indeed, many compounds which contain quinoline-imidazolone such as NVP-BBD130, NVP-BEZ235 exhibit good biological activity, and these adduced compounds are representative ones that impede PI3K-AKT-mTOR signaling pathway (the structures of them are shown in Figure 1).[2–6] In addition, it is reported that NVP-BEZ235 inhibits mTOR (p70S6K) with an IC50 of 6 nM. And the IC50 against other serotonin kinases ATRs in the cells was 21 nM.[7,8] Nevertheless, inducing apoptosis function of NVP-BEZ23 is weak, and NVP-BEZ23 can trigger self-protection autophagy of tumor cells. As a result, NVPBEZ235 was often used in combination with other drugs for the treatment of cancer.[9,10] Another disadvantage is that NVP-BEZ235 inhibits the biological activity of other drugs in some cases. So there is an urgent need to develop compounds that overcome these shortcomings. Moreover, upadacitinib, which has a similar structure, is a Janus kinase (JAK) inhibitor being developed for the treatment of rheumatoid arthritis (RA) and other inflammatory diseases.[11] Upadacitinib is currently being evaluated in phase III trials for the treatment of RA and PsA, and in phase II studies in Crohn’s disease, atopic dermatitis, and ulcerative colitis. From these, we can see that the quinoline imidazole ring is a pharmacologically effective structure.
Based on the above findings, we planned to synthesize compounds that impede PI3K-AKT-mTOR signaling pathway, so we retained quinoline-imidazole moiety, the valid fragment of NVPBEZ235 in the derivative’s design. In the early stage of our experiment, we modified the C-6 position of the core quinoline ring by substituting larger molecular weight groups. However, the effects were poor in the anti-tumor cell activity test due to the poor solubility of the obtained compounds (molecular structure was too rigid). Hence, we designed the compounds 8a–e, in which methoxy group was introduced at the C-6 positions of the quinoline ring rather than quinoline group to increase density of the quinoline and reduce rigidity (Figure 2). Nevertheless, quinoline-imidazolos compounds 8a–e almost had no inhibitory activity against all test cell lines. At the beginning, inspired by upadacitinib, we held the view that the ketone on imidazole had an effect on the activity of the nitrogen atom. And the methyl group on the nitrogen atom of imidazole made the fixed electron pair of the nitrogen atom faced outwards and might not bind to the residue of the corresponding protein. So in order to enhance the activity of the nitrogen atom, on the one hand, we discarded the ketone and the methyl group on nitrogen atom. On the other hand, we added methyl at C-2 positions of imidazole to increase the electron cloud density of the imidazole. But the new acquired compounds 9a–e, 10a–e were still unsatisfactory in cytotoxic activity. In order to find out the reason of affecting the activity, we tried to decrease the electron cloud density of the quinoline ring. So the electron-donating methyl group at C-2 positions was removed and electron-withdrawing bromine atom instead of electron-donating methoxy groups was introduced at C-6 positions to get compounds 11a–e, 12a–e. In summary, we designed 25 compounds to study the effect of quinoline-imidazole derivatives on the anticancer activity. Synthetic routes are showed in Scheme 1.
Herein, we disclosed the synthesis and antitumor activity of all the target compounds against the A549 (human lung cancer), PC-3 (human prostatic cancer), HepG2 (human liver cancer), and MCF-7 (human breast cancer) cancer cell lines, as well as the action against PI3Kα and mTOR kinases. We also took experiment against healthy cell WI-38 to verify the selectivity for cancer cells. In addition, we introduced the docking studies results.

2 | RESULTS AND DISCUSSION

2.1 | Chemistry

The preparation of target compounds 8a–e, 9a–e, 10a–e, 11a–e, and 12a–e is described in Scheme 1. In the beginning, we synthesized the key intermediates 7a–j, which were used as the starting material to yield FIGURE 2 The design of target compounds target compounds. Compounds 8a–e, 9a–e, 10a–e were synthesized from 7a–e, and 11a–e, 12a–e from 5f–j. Compounds 7a–e were methylated with methyl iodide under the action of sodium hydride to give the desired compounds 8a–e. And after cyclization reaction of 7a–j with triethylorthoformate heated at 120°C, compounds 9a–e and 12a– e were obtained after 2.5 h. The key intermediates 7a–e as raw material and 1,4-dioxane as the solvent, the mixture was heated at 120°C and refluxed with triethylorthoacetate. And the reaction was carried out for 2.5 h to obtain the compounds 10a–e; using the same method we acquired 11a–e from 7f–j.[12]

2.2 | Biological evaluation

2.2.1 | In vitro cytotoxic activity against cancer cell lines HepG2, A549, PC-3, and MCF-7

The cytotoxic activities of the target compounds were evaluated in parallel against four human cancer cell lines (HepG2, A549, PC-3, and MCF-7) by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Compounds NVP-BEZ235 and sorafenib were reference compounds. Results were summarized in Table 1 and the values were the average of at least two independent experiments.
As showed in Table 1, all the compounds showed higher inhibitory activity on cell line HepG2 and A549 than that of cell line PC-3 and MCF-7. Among these, the compounds 11a, 11b, 12a, and 12b showed good inhibition to cancer cell proliferation; 11a and 12a had the better inhibitory effect on the proliferation of human hepatoma cell line HepG2, and had a good effect on the proliferation of cell line A549 and cell line PC-3, but had no sensitivity to cell line MCF-7. Compound 12a represented the best cytotoxic activity against HepG2, A549, and PC3 cell lines, with IC50 values of 2.42 ± 1.02 µM, 6.29 ± 0.99 µM, and 5.11 ± 1.00 µM, respectively. Moreover, the different substituent groups had a significant impact on the activity. Compounds 8a–e almost had no inhibitory activity, probably because carbonyl ketone on the imidazole ring can affect the activity of the nitrogen atom. And the methyl group on the nitrogen atom of imidazole made the fixed electron pair of the nitrogen atom faced outwards and might not bind to the residue of the corresponding protein. Indeed, compounds 9a–e and 10a–e obtained through discarding the ketone and adding methyl group on imidazole to increase the electron cloud density showed better cytotoxic activity. But the result was still not satisfactory. We found that the electron-donating methoxy group at C-6 position and the electron-donating methyl group at C-2 position of quinoline ring had negative impact on activity. In addition, compound 11a–e and 12a–e in which introduced electron-withdrawing bromine atom instead of electron-donating methoxy group showed better activity than 9a–e and 10a–e, confirming this inference. In addition, the methyl group on the imidazole ring also affected cytotoxic activity passively. When the R group was cyano, the compounds exhibited selective inhibition of the proliferation to different cancer cell lines. Changing the cyano group to others like methoxy, bromine or other halogen groups was not a wise choice. In summary, the imidazole ring in these series of quinolino-imidazole derivatives was best kept in its original form without substituents on it, while the quinoline ring should have electron-withdrawing groups to increase its cytotoxic activity. At the same time, the R groups should probably be the same as the lead compound NVP-BEZ235, with cyano group being the best choice.

2.2.2 | In vitro cytotoxic activity against healthy cell line WI-38

The activities against healthy cell line WI-38 of 11a, 11c, 12a–12c are expressed in Table 2. According to Table 2, the selected five compounds had almost no toxicity to healthy cell line WI-38. The SCHEME 1 Synthetic routes to the target compounds IC50 values of 11c, 12b, and 12c against cell line WI-38 were greater than 50 µM, the IC50 values of 11a, 12a were 44.6 ± 2.04 µM and 32.8 ± 1.23 µM, respectively, which indicated that cytotoxicity against cancer cell lines was not simply toxicity. And these quinoline-imidazole derivatives had selectivity for cancer cells.

2.2.3 | Activity against PI3Kα/mTOR kinase of compounds 11a and 12a

We further evaluated compounds 11a and 12a for the PI3K and mTOR kinase inhibitory activity. The results are expressed in Table 3. It can be seen from Table 3 that the activity of 12a was better than that of 11a, but the inhibitions of PI3Kα and mTOR were slightly lower than that of lead compound NVP-BEZ235, indicating that there may be other mechanism of action.

2.3 | Molecular docking study

In this paper, we selected compounds 11a and 12a as ligand examples, the docking of the compounds with the p110a protein molecule (PDB: 1E8X) and the mTOR protein molecule (PDB: 4JT6) were scored. The results of the docking scores are shown in Table 4. In the binding mode, the overall fractions of the compounds were lower than that of the lead compound NVP-BEZ235. Compound 12a scored slightly higher than 11a. The results of this docking were consistent with the antitumor activity in vitro.
At the beginning, we have searched out the binding mode of NVP-BEZ235 docked in the catalytic site of PI3Ka. In the structure of NVPBEZ235, the nitrogen atom of the core quinolone ring was involved in a key H-bond interaction with the backbone of the conserved PI3Kα Val851 (Val882 in g). The other two hydrogen bonds’ interactions (dotted lines) involved the backbone of the conserved Asp933 (Asp964 in PI3Kα) and the conserved Ser774 (Ser806 in PI3Kα) side chain. These H-bond interactions were complemented by favorable van der Waals contacts with conserved hydrophobic residues of the ATP catalytic site, including Met772 (Met804 of PI3Kα), Tyr836 TABLE 2 Activity against healthy cell line WI-38 of 11a, 11c, (Tyr837 of PI3Kα), and Met922 (Tyr953 of PI3Kα). To explore the binding modes of target compounds with the active site of PI3K and mTOR, molecular docking simulation studies were carried out by using AutoDock Vina v1.02 and analyzed by Discover Studio 2.5 Visualization. The molecular docking diagrams are shown in Figure 3.
It can be seen from the PI3K/mTOR protein kinase binding modes of both compounds 11a and 12a that although all designed synthetic derivatives were mated with the mTOR protein to form the same amount of hydrogen bonds, the amino acid species associated with their hydrogen bonds was different. The derivatives’ molecules formed hydrogen bonds with the valine residue Val2240, and the lead compound binds to the serine residue Ser2165 with hydrogen bonds. At the same time, compared 11a with 12a, the amino acid residues of hydrogen bond’s binding with the PI3K ligand protein were same, but the number of hydrogen bonds less than the lead compound. In this paper, the overall activity of target compounds was lower than that of the lead compound NVP-BEZ235, the reason may be related to the number of hydrogen bonds bound to the PI3K/mTOR protein and the type of amino acid residues. The results obtained by the abovementioned molecular docking and the docking pattern were consistent with the result of in vitro antitumor activity.

3 | CONCLUSION

In this paper, we designed and synthesized quinoline-imidazole derivatives, a total of 25 compounds, and the anticancer activity of all the compounds were evaluated by MTT method. Most compounds showed moderate activity, and compounds 11a and 12a with good antitumor activity were screened out. The anticancer activity of the compounds 11a and 12a was equal to the leading compound NVPBEZ235. Analysis of the test results revealed that the introduction of the electron donor methoxy group at the C-6 position of the quinoline ring showed a significant decrease in the antitumor activity compared to the introduction of the electron withdrawing bromine atom. The introduction of different substituents on the quinoline ring C-4 aniline had a certain effect on the activity of the compounds. And the antihealthy cell line WI-38 experiment showed that target compounds had almost no effect on healthy cells. These compounds were selective for cancer cells. In order to further study its mechanism, we used NVCEUltra enzyme activity evaluation method, NVP-BEZ235, PI103 as positive control drugs, to test the inhibition activity of compounds 11a and 12a to PI3Kα and mTOR enzyme. The inhibition of PI3Kα and mTOR was lower than that of lead compound NVP-BEZ235 and positive control drug PI103, indicating that there may be some other mechanism of action. Further studies on the binding of compounds 11a and 12a to PI3K/mTOR protein kinases were carried out. The best compound in vitro antitumor activity 12a scored higher than 11a. The results predicted that the in vitro antitumor activity of the molecule was related to the hydrogen bond formed by the docking of the molecules and the type of amino acid residue.

4 | EXPERIMENTAL

4.1 | Chemistry

4.1.1 | General

All melting points were obtained on a Büchi Melting Point B-540 apparatus (Büchi Labortechnik, Flawil, Switzerland) and were uncorrected. NMR spectra were performed using Bruker 400 MHz spectrometers (Bruker Bioscience, Billerica, MA, USA) with TMS as an internal standard. Mass spectra (MS) were taken in electrospray ionization (ESI) mode on Agilent 1100 liquid chromatography-mass spectrometry (LC-MS) (Agilent, Palo Alto, CA, USA). Thin layer chromatography (TLC) analysis was carried out on silica gel plates GF254 (Qindao Haiyang Chemical, Qingdao, China). All materials were obtained from commercial suppliers and used without purification, unless otherwise specified. Yields were optimized. The InChI codes of the investigated compounds together with some biological activity data are provided as Supporting Information.

4.2 | Biological evaluation

4.2.1 | Cytotoxicity assay in vitro

The target compounds (8a–e, 9a–e, 10a–e, 11a–e, and 12a–e) were evaluated for cytotoxic activity against A549 (human lung cancer cell), PC-3 (human prostate cell), HepG2 (human hepatocellular carcinoma cell) by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) method, with NVP-BEZ235 being the positive control according to the reported procedures.[13]
The cells were seeded evenly in 96-well plates and 24 h later 25 target compounds were added to 96-well plates at five concentrations (1000, 333, 111, 37, 1 µg/mL). Plates were incubated at 37°C in a 5% CO2 incubator for 72 h. After 72 h, the culture medium in the well plate was removed and the prepared MTT solution (MTT powder was dissolved in PBS at pH 7 to a concentration of 5 mg/mL, filtered through a 0.22 µm membrane filter) was added to 20 µL per well. After 3.5 h, the fluid was removed and 150 µL DMSO was added per well. Placed plates in a magnetic shaker shock for 3 min using dual wavelength (492 and 630 nm) crystals completely dissolved in the enzyme immunoassay. Measured the absorbance of each well. The IC50 value for each drug was calculated using the Bliss method based on absorbance.

4.2.2 | PI3K and mTOR kinase assay

The inhibitory activities against PI3K and mTOR kinase assay were evaluated through the mobility shift assay or Lantha screen assay together with reference compound NVP-BEZ235b according to the reported procedures.[11] PI3K alpha and mTOR enzymes were purchased from Millipore and Lance Ultra evaluation kit was purchased from PE. The PI3Kα and mTOR enzymes were dissolved in BEZ235 kinase buffer to a concentration of 2.5 nM; the Ulight-4E-BP1 peptide and ATP were dissolved in kinase buffer; the EDTA was dissolved in kinase buffer to a concentration of 8 nM. The Eu-antiphospho-4E-BP1 antibody was dissolved in kinase buffer and adjusted to a concentration of 2 nM. In accordance with the previous orifice plate dosing arrangement, added 2.5 µL kinase solution to each well, added 2.5 µL reaction solution to the control wells, shook the plate to mix the added solution; added 2.5 µL to each well according to the plate dosing setting. Then 2.5 µL test compounds or positive control drug solution was added to each well according to the plate dosing pattern. Add 5 µL basal solution to each well and shake the plate to mix the solution. The well plates were placed at room temperature for 1 h. Then, 10 µL of Eu-antiphospho-4E-BP1 and EDTA reaction solution was added, centrifuged, and left at room temperature for 1 h. Envision was used to read the test data. The 665 nM Lance signal was read and then converted to inhibition (%) according to the formula: (Lance signal value − min)/(max−min) × 100. The IC50 value was obtained by MS Excel and Graphpa 5.0 analysis.

REFERENCES

[1] J. LoPiccolo, G. M. Blumenthal, W. B. Bernstein, P. A. Dennis, Drug Resist. Updat. 2008, 11(1), 32.
[2] X. Q. Wang, L. X. Liu, Y. Li, C. J. Sun, W. Chen, L. Li, H. B. Zhang, X. D. Yang, Eur. J. Med. Chem. 2013, 62, 111.
[3] B. Zhou, Z. F. Liu, G. G. Deng, W. Chen, M. Y. Li, L. J. Yang, Y. Li, X. D. Yang, H. B. Zhang, Org. Biomol. Chem. 2016, 14(39), 9423.
[4] A. Kim, S. Park, J. E. Lee, W. S. Jang, S. J. Lee, H. J. Kang, S. S. Lee, Leukemia Res. 2012, 36(7), 912.
[5] N. Awasthi, P. L. Yen, M. A. Schwarz, R. E. Schwarz, J. Cell. Biochem. 2012, 113(3), 784.
[6] D. Sarker, J. E. Ang, R. Baird, R. Kristeleit, K. Shah, V. Moreno, P. A. Clarke, F. I. Raynaud, G. Levy, J. A. Ware, K. Mazina, R. Lin, J. Wu, J. Fredrickson, J. M. Spoerke, M. R. Lackner, Y. Yan, L. S. Friedman, S. B. Kaye, M. K. Derynck, Clin. Cancer Res. 2015, 21(1), 77
[7] D. Roulin, L. Waselle, A. Dormond-Meuwly, M. Dufour, N. Demartines, O. Dormond, Mol. Cancer 2011, 10(2), 90.
[8] M. C. Manara, G. Nicoletti, D. Zambelli, S. Ventura, C. Guerzoni, L. Landuzzi, P.-L. Lollini, S.-M. Maira, C. García-Echeverría, M. Mercuri, Clin. Cancer Res. 2010, 16(2), 531.
[9] J. A. Sznol, L. B. Jilaveanu, H. M. Kluger, Curr. Cancer Drug Targets 2012, 13(2), 165.
[10] E. Jifu, J. Xing, H. Gong, J. He, W. Zhang, Tumour Biol. 2015, 36(2), 1091.
[11] B. Klünder, M. E. F. Mohamed, A. A. Othman, Clin. Pharmacokinet. 2017, (6937), 1.
[12] T. S. Xu, F. Lei, H. Tao, J. Q. Li, Z. W. Zhang, L. X. Wang, S. Xu, P. W. Zheng, Basic Clin. Pharmacol. Toxicol. 2016, 119, 31.
[13] F. Lei, C. Sun, S. Xu, Q. Wang, Y. OuYang, C. Chen, H. Xia, L. Wang, P. Zheng, W. Zhu, Eur. J. Med. Chem. 2016, 116, 27