Design, synthesis and “in vitro” anti-leukemic evaluation of ferulic acid analogues as BCR-Abl inhibitors

Abstract

Semi-synthetic modification of ferulic acid isolated from Salicornia brachiata was performed and the derivatives showed gratifying in silico binding scores with the BCR-Abl protein, comparable with imatinib. Anti-proliferative activity against K562, U937 and Hep G2 cancer cell lines and the BCR-Abl kinase inhibitory activity using an ADP-Glo assay were investigated. Compounds 2i and 3j were potent BCR-Abl inhibitors and were also active against K562 cells.

---

The medicinal halophyte Salicornia brachiata, called “Umari keerai”, is a traditional Indian plant with enormous medicinal value.¹ Initially, a water extract of this plant was used in a study and the preliminary analysis showed the presence of phenylpropanoids especially ferulic acid in abundance (Fig. S1†). Ferulic acid (FA) has tremendous biological applications like antioxidant, antiinflammatory, anticancer, antiviral, antibacterial, antidiabetic, anti-aging and hepatoprotective activities. A very impressive fact about FA is that it also has significant activity against cardiovascular and Alzheimer’s diseases. It is also used in cosmetics and food preservatives.² Synthetically derived ferulic acid analogue moieties show many potent biological activities.³

In the present work, ferulic acid was semi-synthetically modified at its three labile sites – the phenolic –OH; the α,β-unsaturated carbonyl and the carboxylic acid, without much alteration of the parent ferulic acid motif (Fig. S2 and S3†). The chemical structures from the designed synthesis were firstly subjected to in silico docking studies against eight pro-apoptotic proteins that have major involvement in the leukemic pathway – P38, Bax, Bcl-2, anti-apoptotic proteins BCR-Abl, Akt, Erk and Jnk, and the cell signalling protein VEGFR.⁴ The fused protein BCR-Abl was chosen as a primary response for chronic myeloid leukemia. Translocated chromosomes called chromosome 9 and chromosome 22 result in the formation of a Philadelphia chromosome (BCR-Abl). This activates the RAS/MAP kinase and PI3K pathways to induce cell proliferation and inhibits apoptosis in chronic myeloid leukemia.⁵ Thus BCR-Abl was chosen as the target protein for the in silico studies. Imatinib was used as the standard drug for comparison of the binding ability of the ferulic acid derivatives against proteins involved in the leukemic pathway of BCR-Abl.⁶

In the present analysis, all the backbone atoms in the ligands are interacting with the ATP binding site residues of BCR-Abl (GLU286 & GLU282), by forming hydrogen bonds. In addition, the hydrophobic residues, such as Val, Ile, Leu and Phe, and charged residues are forming classical cationic–hydrophobic residue dyads. Among all the 39 compounds, (E)-4-(3-((4-fluorophenyl)thio)-3-oxoprop-1-en-1-yl)-2-methoxyphenyl acetate (2i) and methyl-(E)-3-(3-methoxy-4-(2-oxo-2-thiomorpholinoethoxy)phenyl)acrylate (3j) show a good glide score with BCR-Abl proteins (Table S1†). Fig. 1 clearly expresses the binding sites of imatinib, ferulic acid and the two ferulic acid analogues 2i and 3j for the BCR-Abl protein.

Fig. 1 Binding of (a) imatinib, (b) FA, and the designed ferulic acid derivatives (c) 2i and (d) 3j to the BCR-Abl protein.

To further the study, semi-synthetic modification of FA was performed. Fig. 2 represents the schemes for the synthesis of compounds 2a–2k. Compounds 2a–2g were prepared using acid–amine coupling reactions of (E)-3-(4-acetoxy-3-methoxyphenyl)acrylic acid chloride with aliphatic and heterocyclic amines. Then the compounds (E)-2-methoxy-4-(3-oxo-3-(phenylthio)prop-1-en-1-yl)phenyl acetate (2h), (E)-4-(3-((4-fluorophenyl)thio)-3-oxoprop-1-en-1-yl)-2-methoxyphenyl acetate (2i), (E)-2-methoxy-4-(3-oxo-3-(p-tolylthio)prop-1-en-1-yl)phenyl acetate (2j) and phenyl-(E)-3-(4-acetoxy-3-methoxyphenyl)acrylate (2k) were prepared by the addition of 4-fluoro-thiophenol, thio-cresol, thio-phenol and phenol to the FA-acid chloride respectively. The above three kinds of reactions were performed with the same reaction conditions, using a triethylamine catalyst at a temperature of 10 °C for 3 h. The structures of the all compounds were confirmed with proton and carbon NMR studies.†

Fig. 2 Schemes for the synthesis of compounds 2a–2k.

Fig. 3 presents a schematic diagram of the reactions involved in the synthesis of compounds 3a–3x and 4a–4c. Compound 3, methyl ferulate, was prepared by the esterification of ferulic acid with methanol and sulfuric acid at 80 °C for 3 h. The formation of 3 was confirmed using proton NMR.† Compound 3 was used for the synthesis of derivatives 3a–3x. Synthesis of compounds 3a–3l was carried out in two steps. In the first step, the respective amine was converted into a halogenated N-phenyl acetamide derivative. Then to 3, the prepared N-phenyl acetamide was added, and refluxed with K₂CO₃ and CH₃CN (Fig. 3). Compounds 3m–3w were prepared by O-alkylation of alkyl bromides with 3. A reaction between compound 3 and an isocyanate provided compound 3x. In this case, DABCO was used as the catalyst and THF as the solvent. A type of Friedel–Crafts reaction of ferulic acid with 2-naphthol and resorcinol, using the Lewis acid AlCl₃, yielded 1,2-dihydro-3H-benzo[f]chromen-3-one derivatives. Fig. 3 presents the scheme for the synthesis of the 1,2-dihydro-3H-benzo[f]chromen-3-one compounds (4a–4c). The structure of the compounds and the formation of a lactone ring were confirmed with proton and carbon NMR spectroscopy, and unequivocally confirmed using single crystal XRD (Fig. S4†).

Fig. 3 Schemes for the synthesis of compounds 3a–3x and 4a–4c.

All the synthesized FA analogues were screened for anti-proliferative activity against K562 (human chronic myeloid leukemia), U937 (human acute monocytic myeloid leukemia) and Hep G2 (human liver cancer) cancer cell lines.† Cell viability was checked in a dose dependent manner for different concentrations of the FA analogues (0.1 to 100 μg mL⁻¹). Also the effect of the FA analogues on normal PBM (peripheral blood mononuclear) cells was analysed and they were found to be safe (Table S1†). Among all the tested compounds with K562 cells, 2i and 3j significantly (p < 0.05) decreased the cell viability. 2i and 3j had IC₅₀ values of 6.3 μg mL⁻¹ and 4 μg mL⁻¹ respectively. 2h and 2j also presented IC₅₀ values of 35 μg mL⁻¹ and 60 μg mL⁻¹ respectively. The rest of the compounds showed minor activity against K562 cells (with an IC₅₀ value more than 100 μg mL⁻¹) (Table S1†). The purity of the key compounds is depicted using HPLC-MS (Fig. S5–S8†). As the FA analogues were active against K562 cells, the synthesized FA derivatives were then subjected to a BCR-Abl kinase inhibitory activity investigation with the established ADP-Glo assay.⁷ 16 compounds exhibited BCR-Abl antiproliferative activity, compared to imatinib (Table 1).

Table 1 In silico docking and BCR-Abl kinase inhibitory studies of the ferulic acid derivatives

S. no.	Compound code	Docking score against the BCR-Abl protein	BCR-Abl kinase inhibition^a^,^b (IC50, μM)
a Mean value.b Boldface: IC50 ≤ imatinib.
1	1	−6.29	0.024
2	2	−8.17	0.255
3	2a	−6.21	0.012
4	2b	−6.08	2.25
5	2c	−6.38	0.019
6	2d	−6.98	0.44
7	2e	−7.02	>500
8	2f	NB	1.69
9	2h	−7.63	0.012
10	2i	−8.86	0.019
11	2j	−7.03	0.44
12	2k	−6.66	>500
13	3	−6.48	1.69
14	3a	−7.89	0.44
15	3b	−6.81	0.32
16	3d	−7.16	>500
17	3e	NB	1.69
18	3f	−7.52	0.024
19	3g	−7.38	0.89
20	3h	−7.65	0.017
21	3i	−7.64	58
22	3j	−8.26	0.024
23	3k	−7.79	16.2
24	3l	−7.12	0.012
25	3m	−7.72	0.024
26	3n	−5.98	2.65
27	3o	−6.54	2.58
28	3p	−5.8	0.015
29	3q	−6.52	1.69
30	3r	−6.15	1.44
31	3s	−5.6	>500
32	3t	−6.42	0.019
33	3u	−6.81	0.032
34	3v	−5.98	0.15
35	3w	−7.52	0.24
36	3x	−7.2	0.18
37	4a	−8.25	0.24
38	4b	−8.22	0.022
39	4c	−6.8	0.015
40	Imatinib	−7.78	0.074

---

To find the mechanism of apoptosis, ROS generation studies,⁸ investigation of the biochemical parameters and DNA fragmentation studies were carried out for 2i and 3j. Furthermore, annexin V-FITC/PI dual staining followed by FACS analysis was carried out to evaluate the extent of the apoptosis.

Reactive oxygen species (ROS) are formed as a natural by-product of normal metabolism and play crucial roles in cell signalling and homeostasis.⁹ However, an increase in the intracellular ROS may result in significant damage to the cell morphology, which ultimately leads to cell death.¹⁰ DCFH2-DA is widely used as an oxidant-sensitizer; it is non-fluorescent, and switches to a highly fluorescent DCF when oxidized by ROS and other peroxides. After addition of DCFH2-DA to the cells treated with FA, 2i and 3j showed strong fluorescence at 520 nm, when compared with control untreated cells, indicating drug induced ROS generation in the cells. There was a 2 fold, 5 fold and 4 fold increase in the ROS generation observed for the cells treated with FA, 2i and 3j respectively, when compared with the control untreated cells (Fig. 4A). Furthermore, FA, 2i and 3j induced ROS generation in K562 cells was confirmed by analysing the cells in a fluorescence activated cell sorter (FACS, BD) (Fig. 4B) and observing the cells under a fluorescence microscope (Fig. 4C).

Fig. 4 FA, 2i and 3j induced ROS generation by estimation of the DCF fluorescence using (A) spectrofluorometry, (B) FACS analysis and (C) fluorescence microscopy. *p < 0.05 and **p < 0.01 indicate significant differences compared to the untreated cells.

Malondialdehyde (MDA) is used for estimation of the damage caused by reactive oxygen species. MDA is a very reactive aldehyde resulting from the peroxidation of biological membranes.¹¹ The MDA level was significantly (p < 0.01) increased in K562 cells treated with FA, 2i and 3j, by 100%, 156% and 176% respectively, compared to their control group. Glutathione is an important antioxidant in the cellular system. So, to determine the glutathione level, we measured both the reduced and oxidized form of glutathione. The reduced glutathione (GSH) level was decreased (26.67%) significantly (p < 0.05) in the FA treated K562 cells and even more significantly (p < 0.01) in the cells treated with 2i (73%) and 3j (60%), compared to the control group. The oxidized glutathione (GSSG) level was increased (150%) significantly (p < 0.05) in the FA treated cells and even more significantly (p < 0.01) in the cells treated with 2i (270%) and 3j (230%), compared to the control group. The redox ratio was decreased significantly (p < 0.05) in the K562 cells treated with 2i (12 fold) and 3j (7.8 fold), compared to the control group (Fig. 5).

Fig. 5 Biochemical estimation of (A) GSH, (B) GSSG, (C) the GSH/GSSG ratio, and (D) lipid peroxidation in K562 cells after treatment with FA, 2i, 3j and imatinib. *p < 0.05 and **p < 0.01 indicate significant differences compared to the untreated cells.

Induction of apoptosis is the most important key event, and also the most studied, for anticancer strategies. In the present study, the induction of apoptosis in leukemic cells by the ferulic acid derivatives was monitored through DNA fragmentation¹², Et–Br/AO staining¹³ and annexin V-FITC/propidium iodide staining.¹⁴

Significant (p < 0.01) DNA fragmentation was observed for the cells treated with 2i and 3j, compared to the control group (Fig. S9†). Et–Br/AO staining images showed that AO was taken up by both viable and nonviable cells and it emits green fluorescence, whereas Et–Br was taken up only by the non-viable cells and it emits orange-red fluorescence because of intercalation into DNA. It is clear that the K562 cells treated with FA, 2i and 3j showed orange-red fluorescence, suggesting the presence of non-viable cells¹⁵ (Fig. S9C†).

To evaluate the extent of the apoptosis, annexin V-FITC/PI dual staining followed by FACS analysis was carried out (Fig. 6). The results showed that 4.23% of the untreated K562 cells and 12.54% of the FA treated K562 cells undergo apoptosis, and that 48.56% of the 2i treated and 42.49% of the 3j treated K562 cells experience apoptosis. The results from annexin V-FITC and PI-staining of the K562 cells indicate late stage apoptosis. Thus, these results confirm the pivotal role of 2i and 3j in inducing apoptosis in K562 cells. As BCR-Abl kinase activity is dependent on its mRNA expression,¹⁶ a study of the BCR-Abl kinase domain expression using conventional RT-PCR and real time PCR in the K562 cells after treatment with FA, 2i and 3j, in comparison with imatinib, was carried out (Fig. 7). It was found that the BCR-Abl kinase domain expression significantly (p < 0.01) decreased upon treatment with 2i and 3j, and was comparable with results using the known BCR-Abl inhibitor, imatinib. Plausibly, the decrease in BCR-Abl kinase domain expression can be linked with and correlated to the activity of 2i and 3j in K562 cells.

Fig. 6 Flow cytometric detection of apoptosis using annexin V/PI staining of the K562 cells with FA, 2i, 3j and imatinib.

Fig. 7 FA, 2i, 3j and imatinib decreased the BCR-Abl kinase domain mRNA expression. Results using (A) conventional RT-PCR and (B) quantitative real-time PCR. **p < 0.01 indicates significant differences compared to the untreated cells.

Based on the anti-proliferative studies of the ferulic acid derivatives, a plausible anti-leukemic pathway for 2i and 3j is presented in Fig. 8. When the compounds bind to the receptor site of the BCR-Abl protein, this inhibits the ATP signalling, followed by inhibition of the cancer cell survival and proliferation signals. This leads to apoptosis and the proliferation of the CML cells will be controlled.

Fig. 8 Proposed mechanism for 2i and 3j with the BCR-Abl protein.

Conclusions

In conclusion, a series of ferulic acid derivatives was synthesized and evaluated for anti-proliferative activity against K562, U937 and Hep G2 cells. The compounds were further tested for BCR-Abl kinase inhibition. In silico studies of the BCR-Abl protein with these ferulic acid analogues demonstrated good docking scores, comparable with the drug imatinib. Among the compounds tested for in vitro anti-leukemic activity against K562 cells, compounds 2i and 3j showed excellent activity, demonstrating that further development of anti-leukemic drugs from the derivatives of ferulic acid deserves attention.

Acknowledgements

The authors would like to thank the management of SASTRA UNIVERISTY for providing the necessary facilities and the TRR Funding. The financial support of SERB-DST, Government of India, under the EMR scheme (EMR/2016/000326) is earnestly acknowledged.

References

1. J. Seralathan, P. Stevenson, S. Subramaniam, R. Raghavan, B. Pemaiah, A. Sivasubramanian and A. Veerappan, Spectrochim. Acta, Part A, 2014, 118, 349–355.
2. (a) N. Kumar and V. Pruthi, Biotech Reports, 2014, 4, 86–91; (b) C. Mancuso and R. Santangelo, Food Chem. Toxicol., 2014, 65, 185–195; (c) A. Sgarbossa, D. Giacomazza and M. Carlo, Nutrients, 2015, 7, 5764–5782.
3. (a) N. Kumar, S. Kumar, S. Abbat, K. Nikhil, S. M. Sondhi, P. V. Bharatam, P. Roy and V. Pruthi, Med. Chem. Res., 2016, 1–18; (b) L. C. Lopes, B. Roman, M. A. Medeiros, A. Mukhopadhyay, P. Utrilla, J. Galvez, S. G. Maurino, V. Moltiva, A. Lourenco and A. S. Feliciano, Tetrahedron Lett., 2011, 52, 6392–6395; (c) W. Li, N. Li, Y. Tang, B. Li, L. Liu, X. Zhang, H. Fu and J. Duan, Bioorg. Med. Chem. Lett., 2012, 22, 6085–6088; (d) S. Oliver, O. Vittorio, G. Cirillo and C. Boyer, Polym. Chem., 2016, 7, 1529–1544; (e) A. Piazzon, U. Vrhovsek, D. Masuero, F. Mattivi, F. Mandoj and M. Nardini, J. Agric. Food Chem., 2012, 60, 12312–12323.
4. M. W. Deininger, J. M. Goldman and J. V. Melo, Blood, 2000, 96, 3343–3358.
5. (a) R. Ren, Reviews, 2005, 5, 172–183; (b) S. Salesse and C. M. Verfaillie, Oncogene, 2002, 21, 8547–8559.
6. M. Deininger, E. Buchdunger and B. J. Druker, Blood, 2005, 105, 2640–2650.
7. X. Pan, J. Dong, R. Shao, P. Su, Y. Shi, J. Wang and L. He, Bioorg. Med. Chem. Lett., 2015, 25, 4164–4168.
8. D. Wu and P. Yotnda, J. Visualized Exp., 2011, 57, 1–4.
9. A. Bhattacharyya, R. Chattopadhyay, S. Mitra and S. E. Crowe, Physiol. Rev., 2014, 94, 329–354.
10. S. Fulda, A. M. Gorman, O. Hori and A. Samali, Int. J. Cell Biol., 2010, 214074.
11. D. R. Janero, Free Radical Biol. Med., 1990, 9, 515–540.
12. K. A. E. Diab, A. I. Elmakwy, O. M. Elmoneim and H. A. Sharaf, Free Radical Biol. Med., 2012, 1, 3–5.
13. M. G. Curcic, M. S. Stankovic, E. M. Mrkalic, Z. D. Matovic, D. D. Bankovic, D. M. Cvetkovic, D. S. Dacic and S. D. Markovic, Int. J. Mol. Sci., 2012, 13, 2521–2534.
14. K. B. A. Ahmed, S. K. Mahapatra, M. R. C. Raja, S. Subramanian, M. Sengan, N. Rajendran, S. K. Das, K. Haldar, S. Roy, A. Sivasubramanian and V. Anbazhagan, RSC Adv., 2016, 6, 18980–18989.
15. S. K. Mahapatra, S. P. Chakraborty, S. Das and S. Roy, Oxid. Med. Cell. Longevity, 2009, 2, 222–230.
16. Y. Dai, M. Rahmani, S. J. Corey, P. Dent and S. Grant, J. Biol. Chem., 2004, 279, 34227–34239.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, LC-MS data, DNA fragmentation, table of docking scores, and NMR spectra are presented. CCDC 1449366 and 1454627. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c6ra10106b

---