Bioorganic & Medicinal Chemistry Letters 30 (2020) 127171

Contents lists available at ScienceDirect

Bioorganic & Medicinal Chemistry Letters

journal homepage: www.elsevier.com/locate/bmcl

Influence of side-chain changes on histone deacetylase inhibitory and T

cytotoxicity activities of curcuminoid derivatives
La-or Somsakeesita, Thanaset Senawongb, Pakit Kumboonmac, Somprasong Saengleed,
Arunta Samankulb, Gulsiri Senawongb, Chavi Yenjaia, Chanokbhorn Phaosiria,
⁎

a
Natural Products Research Unit, Center of Excellence for Innovation in Chemistry, Ministry of Higher Education, Science, Research and Innovation (Implementation Unit-
IU, Khon Kaen University), Department of Chemistry, Faculty of Science, Khon Kaen University, Khon Kaen 40002, Thailand
b
Natural Products Research Unit, Department of Biochemistry, Faculty of Science, Khon Kaen University, Khon Kaen 40002, Thailand
c
Department of Applied Chemistry, Faculty of Science and Liberal Arts, Rajamangala University of Technology Isan, Nakhon Ratchasima 30000, Thailand
d
Ban Dong Subdistrict Administration Organization, Ubolratana District, Khon Kaen 40250, Thailand

ARTICLE INFO ABSTRACT

Keywords: Using curcuminoids as lead compounds, fifty-nine curcuminoid derivatives with different side chains at the
Curcumin phenolic moiety were synthesized. All compounds were investigated for their histone deacetylase (HDAC) in-
Turmeric hibitory activities. The potent pan-HDAC inhibitors were further tested against three human cancer cell lines
HDAC including Hela, HCT116 and MCF-7 with MTT-based assay. The bisethylamide 4z and the mono-sec-butyl de-
Anticancer
rivative 5j manifested good antiproliferative activities against HCT116 cancer cells with the IC50 values as
14.60 ± 1.19 μg/mL and 7.33 ± 0.98 μg/mL, respectively. Molecular docking study of both compounds with
Class I HDACs revealed that the compounds might bind tightly to the binding pocket of HDAC2. These findings
suggested that these compounds can be putative candidates for the development of anticancer drugs via in-
hibiting HDACs.

Histone acetyl transferases (HATs) and histone acetyl deacetylases which can cause several side effects. Although, the hydroxamic acid is
(HDACs) are crucial enzymes in the regulation of gene expression. In the potent zinc-chelating functional group, its drawbacks arise from the
normal cells, there is a balance between histone acetylation and dea- pharmacokinetics issues such as glucuronidation, sulfonation and en-
cetylation.1–3 The overexpression of HDACs have been found to cause zymatic hydrolysis resulting in a short in vivo half-life.18 It also has
various diseases including cancer, diabetes, cardiac and neurodegera- mutagenic properties.19 Therefore, several non-hydroxamic acid HDAC
tive diseases.4–8 To date, HDACs are divided into four groups. Class I inhibitors have been designed and evaluated.20–23 Following our in-
(HDACs 1, 2, 3 and 8), class II (HDACs 4, 5, 6, 7, 9 and 10) and class IV terest in this area, we have focused on investigating the non-toxic and
(HDAC 11) HDACs all belong to the zinc-dependent enzymes, whereas non-hydroxamic acid HDAC inhibitors with isoform selectivity. Hy-
class III (SIRT1-7) HDACs are NAD+-dependent enzymes. Various droxycapsaicin (2) and curcumin (3) are examples of HDAC inhibitors
HDAC isoforms are different in localizations, functions and tissue dis- derived from chili and turmeric, respectively. Although the HDAC in-
tributions. Class I HDACs play a crucial role in tumorigenesis through hibitory activity of both compounds are not comparable to that of
the formation of complexes with other proteins.9,10 Class II HDACs in- SAHA (1), they have isoform selectivity and low toxicity to non-cancer
volve in a variety of biological processes and play key roles in activity- cell line (Fig. 1).24–26
dependent gene regulation in response to neural activity.11,12 Currently, The natural derived curcumin (3) was received a lot of attention
four HDAC inhibitors including SAHA (vorinostat®, 1) from Merck, because of its variety of biological activities such as anti-inflammatory,
romidepsin from Gloucester Pharmaceuticals, belinostat and panobi- antioxidant, anticancer, anti-HIV, anti-angiogenic and antibacterial
nostat have been approved by USFDA for the treatment of cutaneous T- activities.27 However, the simple structural with two phenolic func-
cell lymphoma, peripheral T-cell lymphoma and multiple mye- tional groups connected by a conjugated β–diketone system of cur-
loma.13–16 Additionally, chidamide has been recently approved by cumin (3) is unsuitable for clinical use due to its poor solubility, me-
Chinese FDA for curing relapsed or refractory PTCL.17 Most of the hy- tabolic unstability and bioavailability.28 Interestingly, minor
droxamic acid based HDAC inhibitors such as SAHA are pan-inhibitors modifications of curcumin (3) lead to derivatives with improved

⁎
Corresponding author.
E-mail address: [email protected] (C. Phaosiri).

https://doi.org/10.1016/j.bmcl.2020.127171
Received 7 January 2020; Received in revised form 26 March 2020; Accepted 3 April 2020
Available online 04 April 2020
0960-894X/ © 2020 Elsevier Ltd. All rights reserved.
L.-o. Somsakeesit, et al. Bioorganic & Medicinal Chemistry Letters 30 (2020) 127171

Fig. 1. Structures of known histone deacetylase inhibitors and curcuminoids.

Table 1
In vitro histone deacetylase inhibitory activity of the compounds at 100 μM.
Cpd R1 R2 %* Cpd R1 R2 %*

3 H H 33 ± 0.25 5c C2H5 H 7 ± 0.01

4a CH3 CH3 20 ± 0.48 5d H C2H5 8 ± 0.10
4b CH3 H 13 ± 0.28 5e C4H9 C4H9 9 ± 1.29
4c C2H5 C2H5 13 ± 0.19 5f C4H9 H 10 ± 0.06
4d C2H5 H 34 ± 0.63 5g H C4H9 8 ± 0.01
4e C4H9 C4H9 9 ± 0.04 5h sec-C4H9 sec-C4H9 7 ± 0.01
4f C4H9 H 8 ± 0.21 5i sec-C4H9 H 7 ± 0.04
4g sec-C4H9 sec-C4H9 7 ± 0.17 5j H sec-C4H9 6 ± 0.06
4h sec-C4H9 H 8 ± 0.17 5k allyl allyl 15 ± 0.28
4i allyl allyl 13 ± 0.26 5l propagyl propagyl 10 ± 0.31
4j allyl H 13 ± 0.35 5m benzyl benzyl 30 ± 0.45
4k propagyl propagyl 15 ± 0.24 5n benzyl H 19 ± 0.44
4l propagyl H 14 ± 0.29 5o H benzyl 15 ± 3.75
4m benzyl benzyl 17 ± 0.29 5p 4-fluoro benzyl 4-fluoro benzyl 26 ± 0.17
4n benzyl H 10 ± 0.13 5q H 4-fluoro benzyl 9 ± 0.23
4o 4-fluoro benzyl 4-fluoro benzyl 11 ± 0.26 5r 4-tert-butyl benzyl 4-tert-butyl benzyl 37 ± 0.10
4p 4-fluorobenzyl H 11 ± 0.12 5s CH2CO2CH3 CH2CO2CH3 14 ± 0.36
4q 4-tert-butyl benzyl 4-tert-butyl benzyl 20 ± 0.56 5t CH2CO2CH3 H 59 ± 1.43
4r 4-tert-butyl benzyl H 12 ± 0.30 5u H CH2CO2CH3 25 ± 1.19
4s CH2CO2CH3 CH2CO2CH3 14 ± 0.13 5v CH2COOH CH2COOH 8 ± 0.01
4t CH2CO2CH3 H 9 ± 0.31 6a CH2CO2CH3 CH2CO2CH3 30 ± 0.43
4u CH(CH3)CO2CH3 CH(CH3)CO2CH3 88 ± 1.47 6b CH2CO2CH3 H 20 ± 0.35
4v TBDMS TBDMS 9 ± 0.14 6c sec-C4H9 H 17 ± 0.28
4w TBDMS H 8 ± 0.17 6d CH2COOH CH2COOH 7 ± 0.12
4x CH2COOH CH2COOH 5 ± 0.02 6e CH2CONH2 CH2CONH2 17 ± 0.05
4y CH2CONH2 CH2CONH2 18 ± 0.02 7a CH3 – 51 ± 0.83
4z CH(CH3)CONH2 CH(CH3)CONH2 7 ± 0.04 7b benzyl – 31 ± 0.33
4zz CH2CONHNH2 CH2CONHNH2 12 ± 0.01 7c CH2CO2CH3 – 28 ± 0.30
5a CH3 CH3 15 ± 0.04 8a CH2CO2CH3 – 18 ± 0.46
5b C2H5 C2H5 10 ± 0.18 9 H H 85 ± 0.62

* = Remaining HDAC activity.

stability and activity.29–34 In this work, we continued with the mod- were designed to improve the pharmacokinetic profiles from conjuga-
ifications of curcumin (3) to improve inhibitor-enzyme binding and tion of glucuronide and sulfate of the phenolic group in the body by
aimed to determine whether these compounds would be potent HDAC substitution of the protons at phenolic groups with various aliphatic
inhibitors and manifest antiproliferative activities against cancer cells and aromatic side chains. Besides improving stability, these substitu-
as well as maintaining their low toxicity. The curcumin derivatives tions were also designed to increase inhibitor-enzyme binding via van

2
L.-o. Somsakeesit, et al. Bioorganic & Medicinal Chemistry Letters 30 (2020) 127171

Fig. 2. Structures of curcuminoid derivatives.

Table 2
In silico histone deacetylase inhibitory activity of the selected compounds.
Cpd IC50 ** Class I HDACs

HDAC1 HDAC2 HDAC8

** **
ΔG* Ki ΔG* Ki ΔG* Ki**

3 116 ± 0.22 −5.86 50.79 −5.47 97.67 −5.71 64.98

4g 0.77 ± 0.46 −6.39 20.71 −6.99 7.47 −6.81 10.22
4x 0.92 ± 0.19 −7.79 1.96 −7.37 3.99 −6.61 14.24
4z 0.90 ± 0.23 −8.22 0.936 −8.54 0.55 −8.26 0.88
5j 0.99 ± 0.34 −7.64 2.50 −8.15 1.06 −7.43 3.60

* (kcal/mol), ** (μM).

der Waals, hydrogen bonds, hydrophobic, hydrophilic and ¶-¶ interac- consideration, the selected substitutions were introduced into de-
tions. methoxycurcumin (5) to provide compounds 5a-5v. The mono-sec-
Curcumin (3) as the major curcuminoid along with two minor butyl derivative 5j showed the least remaining activity of HDAC and the
curcuminoids, demethoxycurcumin (5, R1 = R2 = H) and bisde- dicarboxylic 5v also exhibited very low remaining activity. The bisa-
methoxycurcumin (6, R1 = R2 = H) were obtained as previously de- cetamide derivatives of demethoxycurcumin (5) were synthesized,
scribed.26 Derivatives 4a-4w were synthesized by reacting curcumin (3) however, they could not be obtained in good yields. For the minor
with K2CO3 and the corresponding alkyl, benzyl as well as silyl halides curcuminoid, bisdemethoxycurcumin (6), only selected derivatives
under reflux condition. Compound 4s was further hydrolyzed under a were synthesized based on the SAR of curcumin (3) and demethox-
basic condition to provide the dicarboxylic 4x. Trans-aminations of the ycurcumin (5). The dicarboxylic 6d still acted as the most potent in-
bismethylester 4s with ammonium hydroxide and hydrazine yielded hibitor, nevertheless the mono-sec-butyl derivative 6c showed a weaker
the bisacetamine 4y and 4zz, respectively. The bismethylester 4u was inhibitory activity than 6d. The bisacetamide 6e could be obtained in a
also reacted with ammonium hydroxide to obtain the bisethylamide 4z. satisfactory yield but it only showed the same order of inhibition with
The newly synthesized derivatives were fully characterized and the 6c. Interestingly, the Michael adduct 7a-7c were gained as by-products
spectroscopic data of the previously synthesized compounds were from alkylation and benzylation of curcumin (3). It might come from a
compared to the literature.30,35 The obtained compounds were initially trace amount of water in the reaction. The tetra-substituted methylester
screened for HDAC inhibitory activity using the Fluor-de-Lys HDAC 8 was also obtained as the by-product. However, all by-products
activity assay kit.26 The remaining HDAC activity is shown in Table 1. showed moderate inhibition against HDACs. The octahydrocurcumin
Except 4u, all derivatives showed good to excellent HDAC inhibitory (9) obtaining from the reduction of curcumin (3)36 exhibited only weak
activity at 100 μM. The di-sec-butyl derivative 4g showed the best in- inhibitory activity (Fig. 2). On the basic of these results, the most potent
hibition among the alkyl, benzyl and silyl derivatives of curcumin (3). derivatives 4g, 4x, 4z and 5j were further evaluated for their IC50 va-
The dicarboxylic 4x was the most potent HDAC inhibitor, whereas the lues as shown in Table 2. All the selected inhibitors exhibited the IC50
bisethylamide 4z showed the same order of inhibition with 4g. Taking values in a submicromolar range which was lower than that of the lead
the structure–activity relationship (SAR) of curcumin derivatives into a compound, curcumin (3).Table 3.

3
L.-o. Somsakeesit, et al. Bioorganic & Medicinal Chemistry Letters 30 (2020) 127171

Molecular docking study was conducted with ClassI HDACs which

13.89 ± 0.67

30.71 ± 1.10
18.08 ± 1.44
2.03 ± 0.16
mainly involved in development of cancer.37 The available crystal
structures of HDAC1, HDAC2 and HDAC8 were obtained from the

> 100
Protein Data Bank. The molecular modeling program Hyperchem 8.0
72 h was used to build the ligands. These ligands were optimized with the
AM1 level. Molecular docking studies were performed for 50 runs using
42.11 ± 1.15

38.33 ± 2.56
46.28 ± 2.08
AutoDockTools 1.5.4 (ADT) and AutoDock 4.2 programs and La-
2.08 ± 0.12

> 100 marckian genetic algorithm search.38,39 The results in Table 2 indicate
that all compounds show better binding affinity with all HDAC isoforms
48 h

than curcumin (3). However, there were only 2–3 fold selectivity
among the isoforms. The bisethylamide 4z showed the lowest Ki values
for all Class I HDACs. Fig. 3A) shows that the strong inhibitor-enzyme
51.40 ± 1.51
2.94 ± 0.46
MCF-7 cells

interactions consist of four hydrogen bonds between 4z and GLY32,

GLU103, TYR209 as well as LEU276 of HDAC2. The coordination of
> 100
> 100

> 100
24 h

amide functional group to the Zn2+ ion can also be observed. The most
potent compounds, 4z and 5j, against HDAC2 in silico were further
specifically investigated in vitro. The IC50 values of 4z and 5j against
14.60 ± 1.19
2.40 ± 0.15
7.89 ± 0.19

7.33 ± 0.98

HDAC2 were 1.04 ± 0.29 μM and 1.13 ± 0.21 μM, respectively. To

complete the evaluation of these potent HDAC inhibitors, the anti-
> 100

proliferative activities were determined in human cervical cancer

72 h

(HeLa), human colon cancer (HCT116) and human breast adenocarci-

noma cancer (MCF-7) cells. All compounds were less cytotoxic against
10.82 ± 1.44

23.68 ± 2.66

cancer cells and appeared to be less toxic to normal cells than curcumin
2.67 ± 0.20

8.53 ± 0.51

(3). Even though compound 4z showed the high antiproliferative ac-

> 100

tivity among the derivatives, it appeared to be toxic towards the non-

48 h

cancer cells (Vero cells). Interestingly, compounds 4g and 5j performed

the better cytotoxicity against all cancer cells than compound 4z.
Moreover, both compounds also exhibited lower toxicity towards non-
28.68 ± 2.65

27.91 ± 1.66
43.59 ± 5.55
3.83 ± 0.07
HCT116 cells

cancer cells than the lead compound. After considering the binding
modes of 4g and 5j with HDAC2, it was clear that 5j showed more
> 100
24 h

interactions with HDAC2 than 4g. These major interactions of 5j and

HDAC2 consisted of the hydrogen bond to GLU103, ¶-¶ interaction with
HIS33, coordination to the Zn2+ ion and hydrophobic interaction of the
10.20 ± 0.20

19.56 ± 1.14
3.56 ± 0.47

9.33 ± 0.38

sec-butyl group as shown in Fig. 3B. The western blot analysis of

compounds 4z and 5j was conducted with HCT116 cells. The results in
> 100

Fig. 4 show accumulation of acetylated H3. The significant accumula-

72 h

tion could be observed with compound 4z at 40 μg/mL.

In conclusion, curcuminoids were modified at the phenolic groups
17.51 ± 2.04

25.80 ± 0.75
14.04 ± 1.39

to obtained fifty-nine derivatives which showed HDAC inhibitory ac-

3.57 ± 0.50

tivity. At least, four derivatives were identified as potent HDAC in-

> 100

hibitors with antiproliferative activity. Both non-polar and polar func-

48 h
Antiproliferative activities of the potent HDAC inhibitors against cancer cell lines.

tional groups contributed to the increased HDAC-inhibitor binding. This

work brought to light problems with the metabolic stability of natural
compound with potent biological activities.
32.31 ± 4.23
55.22 ± 1.39
5.39 ± 0.20
HeLa cells

> 100
> 100
24 h

Declaration of Competing Interest

The authors declare that they have no known competing financial

10.97 ± 0.80

18.66 ± 0.76

interests or personal relationships that could have appeared to influ-

ence the work reported in this paper.
> 100
> 100

> 100
72 h
IC50 values (mean ± SD; n = 3 (μg/mL)

Acknowledgments
17.87 ± 2.65

25.56 ± 5.47

Khon Kaen University is gratefully acknowledged for the financial

> 100
> 100

> 100

support of this work (Grant Number 6200014002). We also would like

48 h

to thank Mr. Kittisak Poopasith for the excellent NMR data. We are
thankful to the Postdoctoral Training Program, Graduate School, Khon
25.00 ± 1.11

Kaen University for providing a fellowship to Dr. Somprasong Saenglee.

A graduate fellowship given La-or Somsakeesit is supported by
Vero cells

> 100
> 100
> 100
> 100

Rajamangala University of Technology Isan (RMUTI). Center of

24 h

Excellence for Innovation in Chemistry, Ministry of Higher Education,

Science, Research and Innovation is also acknowledged.
Table 3

Cpd

4g
4x
4z
5j
3

4
L.-o. Somsakeesit, et al. Bioorganic & Medicinal Chemistry Letters 30 (2020) 127171

Fig. 3. The interaction modes of A) 4z and B) 5j in the active site of HDAC2.

Fig. 4. Western blot analysis. A. HCT116 cells were treated with absolute ethanol and DMSO (1:1 v/v) as control, 5j of 20, 40 and 60 μg/mL and 4z of 10, 20 and
40 μg/mL for 24 h. The cells were harvested and proteins were extracted and separated on SDS-PAGE. Expression of Ac-H3 were detected by immunoblotting. ERK1/
2 was used as a loading control in western blot analysis. The intensity of protein bans was quantitated by densitometric analysis. B. The bar graphs represent relative
fold of protein expression in HCT116 cells compared with internal control.

5
L.-o. Somsakeesit, et al. Bioorganic & Medicinal Chemistry Letters 30 (2020) 127171

Appendix A. Supplementary data 18. Kelly WK, O’Connor OA, Krug LM, et al. J Clin Oncol 2005, 23, 3923.
19. Shen S, Kozikowski AP. ChemMedChem. 2016;11:15–21.
20. Jiao P, Jin P, Li C, et al. Bioorg Med Chem Lett. 2016;26:4679.
Supplementary data to this article can be found online at https:// 21. Lu A, Luo H, Shi M, et al. Bioorg Med Chem Lett. 2011;21:4924.
doi.org/10.1016/j.bmcl.2020.127171. 22. Seidel C, Schnekenburger M, Zwergel C, et al. Bioorg Med Chem Lett. 2014;24:3797.
23. Botta CB, Cabri W, Cini E, et al. J Med Chem. 2011;54:2165.
24. Bora-Tatar G, Erden DD, Demir AS, Dalkara S, Yelekci K, Yurter HE. Bioorg Med
References Chem. 2009;17:5219.
25. Senawong T, Wongphakham P, Saiwichai T, Phaosiri C, Kumboonma P. Turk J Biol.
1. Bertrand P. Eur J Med Chem. 2010;45:2095. 2015;39:1.
2. Miller TA, Witter DJ, Belvedere S. J Med Chem. 2003;46:5097. 26. Kumboonma P, Senawong T, Saenglee S, et al. Med Chem Res. 2019;28:1773.
3. Micelli C, Rastelli G. Drug Discovery Today. 2015;20:718. 27. Anand P, Thomas SG, Kunnumakkara AB, et al. Biochem Pharmacol. 2008;76:1590.
4. Manal M, Chandrasekar MJN, Priya JG, Nanjan MJ. Bioorg Chem. 2016;67:18. 28. Masters A, Multhaup CL, Simms G, Pottgiesser J, Martins RN, Beyreuther K, EMBO J
5. Sharma S, Taliyan R. Pharmacol Res. 2016;113:320. 1985, 4, 2757.
6. Abend A, Kehat I. Pharmacol Ther. 2015;147:55. 29. Changtam C, Hongmanee P, Suksamrarn A. Eur J Med Chem. 2010;45:4446.
7. Hahnen E, Hauke J, Trankle C, Eyupoglu IY, Wirth B, Blumcke I. Expert Opin Invest 30. Changtam C, Koning HP, Ibrahim H, Sajid MS, Gould MK, Suksamrarn A. Eur J Med
Drugs. 2008;17:169. Chem. 2010;45:941.
8. Hu J, An B, Pan T, Li Z, Huang L, Li X. Bioorg Med Chem. 2018;26:5718. 31. Banuppriya G, Sribalan R, Padmini V, Shanmugaiah V. Bioorg Med Chem Lett.
9. Banerjee S, Adhikari N, Amin SA, Jha T. Eur J Med Chem. 2019;164:214. 2016;26:1655.
10. Suzuki T. Chem Pharm Bull. 2009;57:897. 32. Okuda M, Hijikuro I, Fujita Y, et al. Bioorg Med Chem Lett. 2016;26:5024.
11. Bürli RW, Luckhurst CA, Aziz O, et al. J Med Chem. 2013;56:9934. 33. Konno H, Endo H, Ise S, et al. Bioorg Med Chem Lett. 2014;24:685.
12. Chang S, Young BD, Li S, Qi X, Richardson JA, Olsen EN. Cell. 2006;126:321. 34. Das J, Pany S, Panchal S, Majhi A, Rahman GM. Bioorg Med Chem. 2011;19:6196.
13. Mann BS, Johnson JR, Cohen MH, Justice R, Pazdur R. Oncologist. 2007;12:1247. 35. Sribalan R, Kirubavathi M, Banuppriya G, Padmini V. Bioorg Med Chem Lett.
14. Deal watch: Celgene acquires Gloucester Pharmaceuticals, gaining approved HDAC 2015;25:4282.
inhibitor. Nat Rev Drug Disc 2010, 9, 94. 36. Feng JY, Liu ZQ. J Agric Food Chem. 2009;57:11041.
15. Campbell P, Thomas CM. J Oncol Pharm Pract. 2017;23:143. 37. Roche J, Bertrand P. Eur J Med Chem. 2016;121:451.
16. Greig SL. Target Oncol. 2016;11:107. 38. Sanner MF. J Mol Graph Model. 1999;17:57.
17. Lu X, Ning Z, Li Z, Cao H, Wang X. Intractable Rare Dis Res. 2016;5:185. 39. Morris GM, Huey R, Lindstrom W, et al. J Comp Chem. 2009;30:2785–2791.