Synthesis, structure–activity relationship and binding mode analysis of 4-thiazolidinone derivatives as novel inhibitors of human dihydroorotate dehydrogenase

A series of 4-thiazolidinone derivatives were synthesized and evaluated as novel human dihydroorotate dehydrogenase (hDHODH) inhibitors.

Abstract

A series of 4-thiazolidinone derivatives were synthesized and evaluated as novel human dihydroorotate dehydrogenase (hDHODH) inhibitors. Compounds 26 and 31 displayed IC₅₀ values of 1.75 and 1.12 μM, respectively. The structure–activity relationship was summarized. Further binding mode analysis revealed that compound 31 could form a hydrogen bond with Tyr38 and a water-mediated hydrogen bond with Ala55, which may be necessary for maintaining the bioactivities of the compounds in this series. Further structural optimization of the para- or meta-position of the phenyl group at R will lead to the identification of more potent hDHODH inhibitors.

Introduction

Human dihydroorotate dehydrogenase (hDHODH), a critical rate-limiting enzyme in the pyrimidine de novo biosynthesis pathway, plays a key role in the biosynthesis of nucleotide and cell proliferation.1,2 The significance of DHODH in rapidly proliferating cells such as tumor cells and active T and B lymphocytes makes it an ideal target for pharmacological intervention.3 Some inhibitors of DHODH have proven efficacy for the treatment of malaria,4,5 autoimmune diseases,6–8 cancer,9 psoriasis,10 virus proliferation11–13 and acute myeloid leukemia.14 Leflunomide (1) and brequinar (3) are two representative examples of such DHODH inhibitors (Fig. 1). Leflunomide and its active metabolite, teriflunomide, have been approved for the treatment of rheumatoid arthritis and multiple sclerosis.3,15,16 However, administration of leflunomide for an extended period would cause numerous adverse effects including gastrointestinal symptoms, liver toxicity, hypertension, interstitial lung disease and birth defects, preventing it from being utilized widely.17–19 Brequinar was used as an antitumor and immunosuppressive agent in phase II clinical trials but failed due to its narrow therapeutic window.20–22 Compound 4 has already exhibited promising results for inflammatory bowel disease in the phase IIa clinical trial, and it is now assessed for Crohn's disease in the phase IIb trial.23 Consequently, the intensive need to discover novel and potent hDHODH inhibitors for further development remains.

Fig. 1. Structures of representative inhibitors of hDHODH.

4-Thiazolidinones, a class of heterocycles including nitrogen and sulfur atoms, have been reported as biologically important scaffolds and exhibit broad biological activities, such as anticonvulsant activity, cardiovascular effects, antibacterial activity, anticancer activity, antihistaminic activity (H1-antagonist), anti-inflammatory activity etc.24–27 As part of our continuing studies on 4-thiazolidinones,28 compound 5 was found to have inhibitory activity against hDHODH by random screening. Then, a series of 4-thiazolidinones were synthesized and evaluated for their inhibitory activities against hDHODH.

Herein, we described the discovery and SAR study of 4-thiazolidinone derivatives as novel hDHODH inhibitors. The binding modes of compounds 21 and 31 were also studied by molecular docking to help further elucidate the SAR.

Results and discussion

Chemistry

The synthesis of the starting materials aryl isothiocyanates is shown in Scheme 1. Dithiocarbamate salts were prepared by the reaction of triethylenediamine and carbamodithioic acids which were formed by treatment of aromatic amines with carbon bisulfide. Isothiocyanates were then obtained using treated dithiocarbamate salts with BTC. 4-Isothiocyanatobenzoic acid was prepared by the reaction of 4-aminobenzoic acid with TCDI in the presence of TEA.

Scheme 1. Synthesis of intermediates. (i) DABCO, CS₂, acetone, r.t., 12 h; (ii) BTC, CHCl₃, r.t., 4 h, 70–95%; (iii) TEA, DCM, r.t., 90%.

4-Thiazolidinones and their analogs were prepared according to the routes depicted in Scheme 2. Aryl isothiocyanates were treated with active methylene compounds and potassium hydroxide in DMF to obtain ketene-N,S-acetal salts. These ketene-N,S-acetal salts were further reacted with 2-chloroacetyl chloride, 3-bromopropanoyl chloride or 1,2-dibromoethane to give compounds 5–11 and 13–31. Compound 12 was prepared by de-protection of compound 11 with TFA in DCM. Reaction of isothiocyanate with 2-mercaptoacetic acid afforded an intermediate, 2-(carbamothioylthio)acetic acid, which was subsequently cyclized to produce the additional analogue 32. All the final compounds were fully characterized by spectroscopic techniques.

Scheme 2. Synthesis of compounds 5–32. (i) KOH, DMF, r.t. 15–30 min; ClCH₂COCl, 0 °C; r.t., 12 h, 68–80%. (ii) KOH, DMF, r.t. 15 min; BrCH₂CH₂COCl, 0 °C; r.t., 12 h, 60%. (iii) KOH, DMF, r.t. 15 min; BrCH₂CH₂Br, 0 °C; r.t., 12 h, 48%. (iv) TFA, DCM, r.t., 85%. (v) TEA, dioxane, reflux, 65%.

Inhibitory activities against hDHODH and SAR study

Table 1 highlights the SAR for the 4-thiazolidinone moiety. If 4-thiazolidinone was replaced by 1,3-thiazin-4-one, the inhibitory activity of compound 6 decreased dramatically. Replacement of the 4-thiazolidinone of 5 with thiazolidin produced 7, with apparently decreased activity again. We next investigated the effect of the R₂ group. Compounds 8–11 showed similar activity to 5, suggesting that the carbon chain length of the ester group probably had no apparent effect on the activity. When the ester structures transformed into their acid, the activity of compound 12 became very poor with an inhibitory rate of only 14.867% at 10 μM. Introducing a –CONH₂ group to the R₂ position also showed weak activity (compound 13), which implied that polar groups were not tolerated at the R₂ position. Compounds 14 and 15 with ester groups at R₁ also showed dramatically lower enzyme inhibitory activities than their cyano group-substituted counterparts (compounds 5 and 8). The above discussion indicated that the 4-thiazolidinone scaffold, the cyano substitution for R₁ and the ester structure for R₂ are key factors for potent hDHODH inhibitors.

Table 1. Structures and activities for 4-thiazolidinone analogs 5–15.

Compd	R1	R2	R3	n	Inhibitory rate at 10 μM (%)	hDHODH IC50 a (μM)
5	CN	CO2Me	O	1	34.866	>10
6	CN	CO2Me	O	2	15.690	>10
7	CN	CO2Me	H	1	11.063	>10
8	CN	CO2Et	O	1	38.194	>10
9	CN	CO2Pr	O	1	33.481	>10
10	CN	CO2Bu	O	1	28.770	>10
11	CN	CO2tBu	O	1	33.933	>10
12	CN	COOH	O	1	14.867	>10
13	CN	CONH2	O	1	11.757	>10
14	CO2Me	CO2Me	O	1	4.271	>10
15	CO2Et	CO2Et	O	1	8.713	>10
Brequinar						0.0084

^aAttempts to determine IC₅₀ values were made if the inhibition rate at 10 μM was higher than 50%.

In order to optimize the activity of lead compound 5, N-substituted derivatives were then explored (Table 2). Compound 16 with o-chlorophenyl or compound 17 with an o-iodophenyl group showed equal or lower enzyme inhibitory activities compared with the lead compound 5. When m-Cl, p-F and p-Br groups were introduced to the phenyl of the scaffold (compounds 19–21), the inhibitory activities were higher than those of compound 5, which are shown in Table 2. Among them, given the dramatically improved potency observed with 4-bromophenyl analog 21, further investigations were performed to study the effect of various para-substituted phenyls. Many para-substituted analogs displayed submicromolar potencies (compounds 20–24 and 26). The compounds with hydrophobic groups (–F, –Br, –NO₂, –CF₃, –OCH₃ and -t-Bu) at the para-position of phenyl showed better potencies. Compound 27 with a hydrophilic group (–COOH) displayed decreased potency. Pyridinyl derivatives (28 and 29) were not beneficial for increasing nhibitory activities. These results indicated that the introduction of bulky hydrophobic groups to R₁ might improve the activity. Based on this idea, naphthalenyl was introduced to the N-position to give analogs 30 and 31. Gratifyingly, compound 31 (2-naphthyl) was slightly more potent than 26 with an IC₅₀ value of 1.12 μM. The poor inhibitory activity of compound 32 further verified the key effect of 2-cyanoacetate substitution on potency.

Table 2. Structures and activities for 4-thiazolidinone analogs 16–32.

Compd	R	Inhibitory rate at 10 μM (%)	hDHODH IC50 a (μM)
16		40.010	>10
17		18.377	>10
18		34.109	>10
19		56.856	8.51
20		49.438	10.64
21		68.979	2.68
22		74.431	3.03
23		62.932	4.41
24		52.475	9.45
25		40.378	>10
26		79.074	1.75
27		15.064	>10
28		43.738	>10
29		45.360	>10
30		21.886	>10
31		80.327	1.12
32		25.632	>10

^aIC₅₀ values were determined from three independent tests, and attempts to determine IC₅₀ values were made if the inhibition rate at 10 μM was higher than 50%.

Binding mode analysis

To further study the action mechanism of this series of derivatives, the binding modes of compounds 21 and 31 were simulated by molecular docking (Fig. 2). Fig. 2A tells that the carbonyl group of 4-thiazolidinone forms a hydrogen bond with Tyr38. When 4-thiazolidinone was replaced by 1,3-thiazin-4-one (compound 6) or thiazolidin (compound 7), the hydrogen bond mentioned above would be destroyed, leading to dramatically decreased inhibitory activity. The methyl ester of 21 is located at the entrance of the binding site, interacting with hydrophobic residues like Leu42, Leu46, Leu58 and Phe62. However, it should be noticed that the methyl moiety of the ester in 21 is exposed to the bulk solvent (Fig. 2B), thus elongation of the carbon chain length of the ester group may have little contribution to the bioactivity, which is why compounds 8–11 showed similar activity to 5. Because of their relatively hydrophobic profile, hydrophilic substituents like –COOH and –CONH₂ (compounds 12 and 13) are not favored at the entrance of the binding site. Fig. 2A also shows that the cyano group of 21 orients itself towards the inner side of the binding site, forming a water-mediated hydrogen bond with Ala55. Limited space is left around the cyano group for substitution. Larger groups like –CO₂Me and –CO₂Et may clash with residues Leu58 and Ala59, leading to sharply decreased bioactivity (14 and 15). Additionally, the 4-bromophenyl group of 21 generates beneficial van der Waals (vdW) interactions with the hydrophobic subsite formed by residues Met43, Ala59, Leu68, Phe98, Met111, Leu359 and Pro364. Obviously, hydrophilic groups like –COOH are disfavored at this hydrophobic subsite, thus compound 27 displayed seriously decreased bioactivity against hDHODH compared with 21. The binding mode of the phenyl group of 21 informs us that little space remained for substitution at the ortho-position of the phenyl, while larger hydrophobic groups are preferred at the para- or meta-position. This is consistent with the decreased inhibitory activities of 16–17, 25 and 30, the increased inhibitory activities of 18–19, and the improved inhibitory activities of 20–24, 26 and 31. As for compound 31, the concave surface accommodates the large naphthalenyl group well (Fig. 2B) and it displayed the most potent bioactivity of this series. Collectively, the 4-thiazolidinone scaffold, the cyano substitution for R₁, the ester structure for R₂ and the hydrophobic substitutions at the para- or meta-positions of the phenyl group for R are beneficial for maintaining the bioactivities of the inhibitors in this series.

Fig. 2. The proposed binding modes for representative compounds 21 and 31. The X-ray crystal structure of hDHODH (PDB ID: ; 4LS1) is shown as a transparent purple cartoon (A) or solid surface (B), and the docked inhibitors are represented as sticks with a transparent surface. Compound 21 is shown as pink sticks, while compound 31 is displayed as green sticks. For small molecules, oxygen atoms are colored red, nitrogen atoms are colored blue, and the sulfur atoms are colored yellow. Key residues (thin sticks) in the binding site are colored purple. Potential intermolecular hydrogen bonds are shown as black dashed lines. Water molecule W592 is depicted as a red ball.

Compared with ligand 3X2 in ; 4LS1 occupying both the hydrophobic and hydrophilic sections of the ubiquinone-binding site of DHODH (Fig. S1‡), compound 31 mainly occupies the hydrophobic subsite with its naphthalenyl group in contact with residues Met43, Ala59, Leu68, Phe98, Met111, Leu359 and Pro364 through favorable apolar interactions. This obviously distinct binding mode of compounds 31 and 3X2 informed us that a polar group, such as carboxyl, formyl and hydroxyl, which is connected to a suitable linker that is substituted for the naphthalenyl group of 31, would generate hydrogen bond or salt bridge interactions with residue Arg136 in the hydrophilic subsite of the ubiquinone-binding site, and thus may give us compounds with improved binding affinity against DHODH. This structural optimization work is undertaken.

Conclusions

In conclusion, a series of 4-thiazolidinone derivatives have been identified as hDHODH inhibitors. Several compounds exhibited moderate activities against hDHODH, especially compounds 26 and 31 with IC₅₀ values of 1.75 and 1.12 μM, respectively. The SAR study and binding mode investigation demonstrate that the 4-thiazolidinone scaffold, the cyano substitution for R₁, the ester structure for R₂ and the hydrophobic substitutions at para- or meta-positions of the phenyl group for R are favorable for improving inhibitory activity. For this series of 4-thiazolidinone derivatives, the hydrogen bond with Tyr38 and the water-mediated hydrogen bond with Ala55 were proposed to be indispensable for maintaining the inhibitory activity against hDHODH. The presented SAR indicates that further decoration of the phenyl group at R may provide us more potent hDHODH inhibitors.