Discovery of substituted N′-(2-oxoindolin-3-ylidene)benzohydrazides as new apoptosis inducers using a cell- and caspase-based HTS assay

Graphical abstract

The discovery and SAR studies of substituted N′-(2-oxoindolin-3-ylidene)benzohydrazides as novel and potent apoptosis inducers are reported.

Abstract

We report the discovery of a series of substituted N′-(2-oxoindolin-3-ylidene)benzohydrazides as inducers of apoptosis using our proprietary cell- and caspase-based ASAP HTS assay. Through SAR studies, N′-(4-bromo-5-methyl-2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazide (3g) was identified as a potent apoptosis inducer with an EC₅₀ value of 0.24 μM in human colorectal carcinoma HCT116 cells, more than a 40-fold increase in potency from the initial screening hit N′-(5-bromo-2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazide (2a). Compound 3g also was found to be highly active in a growth inhibition assay with a GI₅₀ value of 0.056 μM in HCT116 cells. A group of potentially more aqueous soluble analogs were prepared and found to be highly active. Among them, compound 4e incorporating a methyl piperazine moiety was found to have EC₅₀ values of 0.17, 0.088 and 0.14 μM in human colorectal carcinoma cells HCT116, hepatocellular carcinoma cancer SNU398 cells and human colon cancer RKO cells, respectively. Compounds 3g and 4e were found to function as inhibitors of tubulin polymerization.

Apoptosis, or program cell death, plays a crucial role in normal cell development and tissue homeostasis.¹ Apoptosis is used by organisms to control their cell numbers and to eliminate unneeded or damaged cells.² Inappropriate apoptosis induction is known to result in excessive cell death, and could cause degenerative diseases.³ Inadequate apoptosis, however, could lead to over proliferation of cells and cancer.⁴ In addition, it is known that the anti-tumor efficacy of many chemotherapeutical agents is correlated to their apoptosis inducing ability.⁵ Identification of compounds that promote or induce apoptosis in cancer cells, therefore, is an attractive approach for anticancer research.⁶

We have been interested in the discovery and development of apoptosis inducers as potential anticancer agents.⁷ Applying our novel caspase-3 substrates,⁸ we have developed a caspase- and cell-based, high throughput screening technology, termed Apoptosis Screening and AntiCancer Platform (ASAP), for the identification of apoptosis inducers.⁹ We have reported the discovery of several novel series of apoptosis inducers, including 4-aryl-4H-chromenes (1a),¹⁰ gambogic acid (1b),¹¹ 3-aryl-5-aryl-1,2,4-oxadiazoles (1c),¹² N-phenyl-1H-pyrazolo[3,4-b]quinolin-4-amines (1d),¹³ 4-anilinoquinazolines (1e)14, 15 and 4-aryl-3-(3-aryl-1-oxo-2-propenyl)-2(1H)-quinolinones (1f)¹⁶ (Chart 1 ). Herein we report the discovery of substituted N′-(2-oxoindolin-3-ylidene)-benzohydrazide (2a), an isatin derivative, as an apoptosis inducer using our HTS assay. SAR study of 2a led to the discovery of N′-(4-bromo-5-methyl-2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazide (3g) and analogs as potent apoptosis inducers.

Chart 1.

Many isatins and isatin derivatives have been synthesized and reported to have a variety of biological activities, including as SARS coronavirus 3C-like protease inhibitors,¹⁷ caspase-3 inhibitors,¹⁸ and as inhibitors of Src homology-2 domain containing protein tyrosine phosphatase-2.¹⁹ More recently, N-alkyl isatin acylhydrazone derivatives such as 7a (Chart 2 ) have been reported to be potent and selective cannabinoid receptor 2 inverse agonists for the potential treatment of neuropathic pain.²⁰ In addition, N-substituted isatins such as 7b have been reported to be cytotoxic with a mode of action that includes inhibition of tubulin polymerization, induction of G₂/M cell cycle arrest and activation of caspase-3 and -7.²¹

Chart 2.

Substituted N′-(5-bromo-2-oxoindolin-3-ylidene)-benzohydrazides 2a–2f were obtained from ChemDiv and Asinex, and their structures were confirmed by ¹H NMR and MS. Substituted N′-(2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazides 3a–3m were prepared from condensation of the corresponding substituted isatin (5)¹⁹ with substituted 3,4,5-trimethoxybenzohydrazide (6) according to reported procedures.20, 22 The N-substituted analogs 4a–4h were prepared from condensation of 2a, 3a, 3g and 3h with formaldehyde and an amine following literature procedures (Scheme 1 ).23, 24

Scheme 1.

The apoptosis inducing activity of substituted N′-(2-oxoindolin-3-ylidene)-benzohydrazides was measured using our cell- and caspase-based HTS assay⁷ in human colorectal carcinoma cells HCT116, hepatocellular carcinoma cancer SNU398 cells and human colon cancer RKO cells, and the results are summarized in Table 1, Table 2, Table 3 . Compound 2a was found to have EC₅₀ values of 4–10 μM in the three cell lines tested. By maintaining the 5-bromo group in the isatin, we explored replacement of the 3,4,5-trimethoxy groups in the benzoyl group of 2a by other groups. Table 1 showed that, except for compound 2b, all these compounds (2c–2f) were inactive up to 20 μM in all the three cell lines, indicating that the 3,4,5-trimethoxy group is preferred.

Table 1.

Activity of substituted N′-(5-bromo-2-oxoindolin-3-ylidene)-benzohydrazides in the caspase activation assay

Compound #	R1	R2	R3	R4	EC50a (μM)
					HCT116	SNU398	RKO
2a	H	OMe	OMe	OMe	10.7 ± 0.5	8.9 ± 0.2	4.4 ± 0.5
2b	H	OCH2O		H	>20	>20	9.7 ± 0.6
2c	OMe	H	H	H	>20	>20	>20
2d	H	F	H	H	>20	>20	>20
2e	H	Br	H	H	>20	>20	>20
2f	H	H	NO2	H	>20	>20	>20

^aCells were treated with the test compounds for 48 h, and data are the mean of three or more experiments and are reported as mean ± standard error of the mean (SEM).

Table 2.

Activity of substituted N′-(2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazides in the caspase activation assay

Compound #	R1	R2	R3	R4	EC50 (μM)a
					HCT116	SNU398	RKO
3a	H	OMe	H	H	6.6 ± 0.8	4.1 ± 0.8	4.0 ± 0.5
3b	H	I	H	H	10.5 ± 0.2	4.1 ± 0.1	3.6 ± 0.4
3c	H	OCF3	H	H	>20	>20	8.3 ± 1.5
3d	H	NH2	H	H	>20	>20	>20
3e	H	AcNH	H	H	>20	>20	>20
3f	Cl	H	H	H	3.8 ± 0.3	2.1 ± 0.2	2.0 ± 0.2
3g	Br	Me	H	H	0.24 ± 0.03	0.13 ± 0.01	0.23 ± 0.03
3h	Cl	Cl	H	H	0.64 ± 0.02	0.57 ± 0.11	0.73 ± 0.12
3i	Ph	Me	H	H	>20	>20	>20
3j	H	Me	Br	H	8.4 ± 1.1	3.7 ± 0.7	4.6 ± 0.8
3k	H	OMe	Br	H	>20	>20	>20
3l	H	Cl	H	Cl	>20	>20	>20
3m	H	Me	H	Br	10.6 ± 0.1	>20	4.9 ± 0.4

^aCells were treated with the test compounds for 48 h, and data are the mean of three or more experiments and are reported as mean ± standard error of the mean (SEM).

Table 3.

Activity of N-substituted N′-(2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazides in the caspase activation assay

Compound #	R1	R2	R3	EC50a (μM)
				HCT116	SNU398	RKO
4a	H	Br		>20	>20	5.1 ± 0.8
4b	H	OMe		>20	>20	>20
4c	Cl	Cl		0.49 ± 0.09	0.47 ± 0.11	0.66 ± 0.10
4d	Br	Me		0.31 ± 0.04	0.12 ± 0.01	0.20 ± 0.02
4e	Br	Me		0.17 ± 0.02	0.088 ± 0.010	0.13 ± 0.02
4f	Br	Me		0.17 ± 0.03	0.087 ± 0.016	0.14 ± 0.04
4g	Br	Me		0.25 ± 0.01	0.14 ± 0.01	0.21 ± 0.03
4h	Br	Me		0.25 ± 0.02	0.13 ± 0.01	0.23 ± 0.03

^aCells were treated with the test compounds for 48 h, and data are the mean of three or more experiments and are reported as mean ± standard error of the mean (SEM).

By maintaining the 3,4,5-trimethoxybenzoyl group, we then explored substitutions in the 4- to 7-positions of isatin ring (Table 2). The 5-methoxy (3a) and 5-iodo (3b) analogs were similar or slightly more active than 2a. The 5-trifluoromethoxy (3c), 5-amino (3d) and 5-acetylamino (3e) analogs were less active or inactive in the three cell lines. These data suggested that strong electron withdrawing, or hydrophilic, or large groups are not favored at the 5-position. The 4-chloro analog 3f was >2-fold more potent than 2a in all the three cell lines, indicating that a small group at the 4-postion might increase potency. Combination of substitutions at both the 4- and 5-positions led to the 4-bromo-5-methyl analog 3g that was highly potent with an EC₅₀ value of 0.24 μM in HCT116 cells, >40-fold more potent than 2a. The 4,5-dichloro analog 3h also was highly active. Interestingly, 5-methyl-4-phenyl analog 3i was inactive up to 20 μM, indicating that a large group is not tolerated at 4-position. 5,6-Di-substituted analogs (3j and 3k) and 5,7-di-substituted analogs (3l and 3m) were found to have low activity or inactive, suggesting that substitutions at the 6- and 7-positions may not be preferred.

To explore the SAR further and to improve the aqueous solubility, we introduced an N-morpholinomethyl group into the nitrogen of the isatin of compounds 2a, 3a, 3g and 3h. Table 3 showed that compounds 4a–4d had activities similar to the corresponding non-N-substituted analogs. Several compounds with various aminomethyl groups were prepared from 3g and found to be highly active. Compounds 4e and 4f were the most potent ones, both with EC₅₀ values of 0.17 μM. These data indicated that substitution at the nitrogen of isatin is tolerated and it could be used to introduce aqueous solubility enhancing groups.

Overall, the apoptosis inducing activities of these compounds in human colon cancer HCT116 cells were similar to that observed in hepatocellular carcinoma cancer SNU398 cells and human colon cancer RKO cells (Table 1, Table 2, Table 3). Compound 3g, 4e and 4f, three of the most active compounds in HCT116 cells, also were the most active ones in SNU398 and RKO cells, suggesting that these compounds most probably will be broadly active against many cancer cell lines.

Representative compounds were assayed in a traditional growth inhibition (GI₅₀) assay to confirm that the active compounds in the caspase induction assay also inhibit tumor cell growth. The growth inhibition assays in T47D, HCT116 and SNU398 cells were run in a 96-well microtiter plate as described previously⁷ and the data are summarized in Table 4 . Compound 3g had GI₅₀ values of 0.056, 0.022 and 0.019 μM in HCT116, SNU398 and RKO cells, respectively, which are >100-fold more active than the original hit compound 2a. Compounds 4d–4f and 4h were also highly active with GI₅₀ values similar to that of 3g. Compound 3h was less active than 3g. These data confirmed that the cell-based caspase activation HTS assay is not only useful for the identification of inducers of apoptosis, but also for subsequent optimization and SAR studies.

Table 4.

Growth inhibition activity of substituted N′-(2-oxoindolin-3-ylidene)-3,4,5-trimethoxybenzohydrazides

Compound #	GI50a (μM)
	HCT116	SNU398	RKO
2a	7.1 ± 1.1	2.7 ± 0.1	2.7 ± 1.2
3g	0.056 ± 0.015	0.022 ± 0.008	0.019 ± 0.003
3h	0.32 ± 0.14	0.14 ± 0.04	0.054 ± 0.012
4d	0.062 ± 0.008	0.026 ± 0.011	0.019 ± 0.003
4e	0.051 ± 0.018	0.023 ± 0.008	0.018 ± 0.004
4f	0.088 ± 0.043	0.024 ± 0.006	0.036 ± 0.012
4h	0.086 ± 0.008	0.029 ± 0.014	0.043 ± 0.016

^aCells were treated with the test compounds for 48 h, and data are the mean of three experiments and are reported as mean ± standard error of the mean (SEM).

The potent compounds in this series of substituted N′-(2-oxoindolin-3-ylidene)-benzohydrazides, such as compounds 3g and 4e, were tested by cell cycle analysis¹⁰ and found to arrest HCT116 cells in G₂/M followed by apoptosis, which is similar to what was reported for compound 7b (Chart 2).²¹ We suspected that compounds 3g and 4e might be tubulin inhibitors. In a tubulin polymerization assay,²⁵ compounds 3g and 4e were found to inhibit tubulin polymerization with IC₅₀ values of 0.97 and 0.19 μM. In comparison, the IC₅₀ values for vinblastine and colchicine were 0.5 μM. These data suggest that inhibition of tubulin polymerization might be the main mechanism of action for these compounds as apoptosis inducers.

In conclusion, a series of substituted N′-(2-oxoindolin-3-ylidene)-benzohydrazides were identified as apoptosis inducers utilizing our ASAP assay. Through SAR studies, potent compound 3g and significantly more aqueous soluble compound 4e were identified with EC₅₀ values in HCT116 cells >40-fold more potent than the screening hit 2a. The mode of action for the potent compounds 3g and 4e was found to be inhibition of tubulin polymerization.