Structure–activity relationship studies of 3-substituted pyrazoles as novel allosteric inhibitors of MALT1 protease
Ken Nunettsu Asaba *, Yohei Adachi, Kazuyuki Tokumaru *, Akira Watanabe, Yasufumi Goto, Takumi Aoki
A B S T R A C T
We report the discovery of a novel series of 1,5-bisphenylpyrazoles as potent MALT1 inhibitors. Structur- e–activity relationship exploration of a hit compound led to a potent MALT1 inhibitor. Compound 33 showed strong activity against MALT1 (IC50: 0.49 μM), potent cellular activity (NF-κB inhibition and inhibition of IL2 production), and high selectivity against caspase-3, -8, and -9. The results of a kinetics study suggest that compound 33 is a non-competitive inhibitor of MALT1 protein.
Keywords:
MALT1
Paracaspase Allosteric inhibitors
Structure activity relationship
Introduction
The nuclear factor-κB (NF-κB) family of transcription factors is considered to play a central role in the immune systems, including immune-cell development, homeostasis, survival, and function.1–4 NF-κB is implicated in the pathogenesis of several autoimmune diseases, such as rheumatoid arthritis, systemic lupus erythematosus, type I dia- betes, multiple sclerosis, and inflammatory bowel disease.5–11 Although NF-κB is activated transiently during the normal immune response, it is chronically activated in tissues affected by autoimmune diseases and induces excessive inflammatory cytokines and chemokines, which leads to autoimmunity.
Mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1) is related to the activation of NF-κB signaling and the negative feedback loop required for this signaling.12–16 MALT1 acts as a scaffold protein, resulting in protein–protein interactions which lead to the activation of the IκB kinase complex and subsequent activation of NF-κB. In addition to this function, MALT1 is a protease. The para- caspase (cysteine protease) domain of MALT1 cleaves RelB, CYLD, and A20, all of which function as negative regulators of canonical NF-κB signaling. Therefore, MALT1 is an important therapeutic target for the treatment of immunomodulation disorders and lymphoma. In this study, we describe the discovery and structure–activity relationship (SAR) exploration of a series of allosteric small-molecule inhibitors of MALT1 protease.
Phenothiazine derivatives, including mepazine, thioridazine (1), and promazine, target the allosteric site of MALT1, preventing rearrange- ment of the protein from the inactive state to the activated state and inhibiting proteolytic cleavage of the substrate (Table 1).17,18 None of these phenothiazine derivatives significantly inhibits either caspase 3 or caspase 8, which are structurally the closest relatives of MALT1 in mammals.
We conducted molecular docking simulations with MALT1 to iden- tify selective small inhibitors of MALT1. The model was constructed from the crystal structure of MALT1 complexed with the allosteric in- hibitor thioridazine (PDB ID: 4I1R), and we used thioridazine (1) as the template compound for docking. Enzymatic biochemical assay of the top ranked compounds identified compound 2, which showed weak activity against MALT1 (Table 1). The calculated docking pose of 2 indicated that the 1,5-bisphenylpyrazole moiety is located similarly to the to give ethyl 1H-pyrazole-3-carboXmides (III). Compounds III were hydrolyzed to 1H-pyrazole-3-carboXylic acids (IV) using aqueous so- dium hydroXide solution in ethanol. Target compounds V were synthe- sized by the condensation of compounds IV with amines using amide coupling reagents such as carbonyldiimidazole and 1-[bis(dimethyla-mino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium fluorophosphate (HATU).
Substituents at the 5-position can be introduced by different methods, as shown in Scheme 2. 1-(4-Chlorophenyl)-5-hydroXy-1H- pyrazole-3-carboXylic acid (VIII) was synthesized using 4-chlorophenyl- hydrazine hydrochloride and dimethyl acetylenedicarboXylate. Next, VIII was converted to 1-(4-chlorophenyl)-3-(methylcarbamoyl)-1H- pyrazol-5-yl trifluoromethanesulfonate ester (X) via amidation and sulfonylation; then, X was coupled with boronic acid derivatives to obtain 1,5-bissubstituted pyrazole-3-carboXamides (V).
We first introduced various substituents at the 1-position of 1H-phenothiazine moiety of thioridazine (1) bound in the hydrophobic pocket of MALT1 (Fig. 1). The pose also showed that a hydrogen bond is formed between the N-methylamide nitrogen of 2 and glutamic acid E397, again similar to the crystal structure with thioridazine (1).17 We selected this 3-substituted pyrazole compound as a starting point as it has three sites for initial structure–activity studies, namely, two phenyl rings and a substituted amide.
Scheme 1 illustrates the synthesis of 1,5-bissubstituted pyrazole-3- carboXamides (V). Commercially available ketones (I) were condensed with diethyl oXalate using sodium ethoXide in ethanol. The obtained ethyl 2,4-dioXobutanoates (II) were treated with hydrazines in ethanol pyrazole (Table 2). The benzyl (3) and cyclohexyl (4) derivatives exhibited potency similar to that of compound 2, whereas methyl de- rivative (5) was completely inactive, suggesting that large hydrophobic substituents such as phenyl groups are required to fill the hydrophobic pocket in MALT1. We performed chloride scanning to identify the optimal substitution position and identified 4-chlorophenyl (6) as the most potent of the chloride-substituted phenyls (7, 8). Substituents at the 5-position of 1H-pyrazole were explored by fiXing 4-chrolophenyl as the 1-position substituent, then introducing substituents at the 5-posi- tion to identify moieties more potent than phenyl (Table 3). In addi- tion, we examined the effect of removing the chloride or changing the substitution position of the chloride, and found that both attenuated MALT1 inhibition (9, 10, 11). Other p-substituted phenyl groups were explored. The potency of the 4-methoXyphenyl (12) derivative was similar to that of 2, whereas substitution with other p-substituted phenyl groups such as 4-fluorophenyl (13) or heteroaromatics such as 3-thienyl (19) provided inactive compounds. In particular, compounds with polar functional group substituents (15, 16, 20, 21, 22) had significantly less potency compared with 2, consistent with the docking result (Fig. 1) showing the phenyl ring located to the hydrophobic pocket of MALT1. Next, we explored substitutes at the 3-position (Table 4). Comparing secondary amide (6) to tertiary amides (23, 24), amide hydrogen is essential for increased potency. This result was consistent with the binding mode of compound 2 shown in Fig. 1, which implies that the amide hydrogen of compound 6 interacts with E397. The potency of the aminoethyl (25) and hydroXyethyl (27) derivatives was similar to or higher than that of 6, whereas the aminopropyl (26), hydroXypropyl (28), and aminocarbonylethyl (29) derivatives were less potent. The phenylethyl (30) and cyclohexylmethyl (31) were inactive. These results indicate that hydrophilic substituents such as amino and hydroxyl groups are favorable substituents at the 3-position. Therefore, we next focused on the amino groups. Introducing a methyl group into 25 pro- vided more potent compounds (32, 33), with N,N-dimethylaminoethyl (33) being 14-fold more potent than compound 6. The acetoamide (34) and sulfonylamide (35) derivatives exhibited less potencies compared to 25 or 33, showing that the basicity of the amine is a key factor to improving potency. Lengthening the carbon chain of compound 33 led to the slightly less potent compounds 36 and 37, suggesting that optimal carbon chain length may be important for improving inhibitory activity. For example, in a series of 1,5-bisphenylpyrazoles, the ethylene linker is advantageous (33 vs. 36 vs. 37; 25 vs. 26; 27 vs. 28).
Cyclized substituents were introduced into N,N-dimethylaminoethyl (33) (Table 5). N,N-Dimethylaminopiperidine (38) had weak potency compared with 33 but was more potent than pyrrolidine (24) lacking an N,N-dimethylamino moiety. Inhibitory potency against MALT1 was enhanced by introducing amino groups at appropriate positions, suggesting that amino group interact with an amino acid of MALT1. The potency of the N-methylpiperidin-2-ylmethyl derivative (40) was similar to that of 33, whereas other cyclized substituents resulted in decreased potency (39, 41). Lipophilic efficiency (LipE) is an important metric for optimizing potency and ADME (absorption, distribution, metabolism, and excretion) properties.19,20 To improve lipophilic efficiency, an oXygen atom was introduced into 40 to provide the 4-hy- droXy-1-methylpiperidin-2-ylmethyl derivative (43) and removal of the methyl group from 42 to provide the morpholin-3-ylmethyl deriv- ative (44). The LipE of 43 and 44 was improved compared to that of 33 and 6 (1.62, 1.50, 1.02, and 0.28 respectively).
The cellular activities of three compounds (thioridazine (1) and compounds 6 and 33) were evaluated (Table 6). The inhibition of NF-κB activity and IL2 production by compound 33 (82%inh, 94%inh) was stronger than that of compound 6 (25%inh, 54%inh) and similar to that of thioridazine (1) (85%inh, 94%inh). Additionally, to confirm the direct inhibition of MALT1 substrate cleavage, degradation of CYLD in Jurkat cells was measured by western blotting (Fig. 2). Compound 33 dose-dependently decreased CYLD degradation product (Frag-CYLD) induced by PMA/Ionomycin stimulation. The value of relative Frag- CYLD/Full-CYLD ratio of 33 at 10 µM was same as that of unstimu- lated cells without MALT1 activation.
We investigated the selectivity of 33 against caspase-3, 8, and 9, which are structurally related to paracaspase, using a caspase drug screening kit (Promokine) according to the manufacturer’s protocol (Fig. 3). The pan-caspase inhibitor Z-VAD-fmk was used as a positive control for each experiment. Compound 33 exhibited little activity against caspase-3, 8, and 9, even at concentrations up to 100 µM, showing it is a selective MALT1 paracaspase inhibitor.
A kinetics study was conducted to investigate the mode of action of 33 (Fig. 4). Compound 33 at a concentration around the IC50 (0.5 µM) strongly decreased Vmax, from 35,612 to 19,979 RFU/s, while Km remained essentially unchanged (246.9 to 234.2 µM), showing that compound 33 is a non-competitive inhibitor. This suggests that 33 does not bind to the paracaspase active site but rather to an allosteric pocket. In summary, SAR exploration of hit compound 2 identified by screening led to the discovery of 1,5-bisphenylpyrazoles as novel, potent, and selective MALT1 inhibitor s. Our exploration showed that phenyl rings are required to maintain potency, and the secondary amide with an aminoethyl moiety is an essential substructure for improved potency. Derivatization of the 3-position of 1H-pyrazoles led to com- pound 33, which exhibits cellular activity similar to that of thioridazine (1). Investigation of the binding mode of 33 indicated a non-competitive inhibition mechanism against MALT1 protein. Further studies on structural optimization and identification of the binding site of this series of compounds is in progress.
References
1 DiDonato JA, Mercurio F, Karin M. Immun Rev. 2012;246:379–400.
2 Hayden Matthew S, Ghosh Sankar. Cell Res. 2011;21:223–244.
3 Hayde MS, West AP, Ghosh S. Oncogene. 2006;25:6758–6780.
4 Liang Yan, Zhou Yang, Shen Pingping. Cell Mol Immunol. 2004;1(5):343–350.
5 Sun S-C, Chang J-H, Jin J. Trends Immunol. 2013;34(6):282–289.
6 Herrington Felicity D, Carmody Ruaidhrí J, Goodyear Carl S. J Biomol Screen. 2016; 21:223–242.
7 Pai S, Thomas R. J Autoimmun. 2008;31:245–251.
8 Greve B, Weissert R, Hamdi N, et al. J Immunol. 2007;179:179–185.
9 Han Z, Boyle DL, Manning AM, Firestein GS. Autoimmunity. 1998;28:197–208.
10 Isabelle Cleynen, Peter Jüni, GeertruidaBekkering E, et al. PLoS One. 2011;6(9), e24106.
11 Liu Ting, Zhang Lingyun, Joo Donghyun, et al. Signal Transduct Target Ther. 2017;2: 17023.
12 Klein Theo, Fung Shan-Yu, Renner Florian, et al. Nat Commun. 2015;6:8777.
13 Rosebeck Shaun, Rehman Aasia O, Lucas Peter C, et al. Cell Cycle. 2011;10(15): 2485–2496.
14 Afonina Inna S, Elton Lynn, Carpentier Isabelle, et al. FEBS J. 2015;282(17): 3286–3297.
15 Coornaert Beatrice, Baens Mathijs, Heyninck Karen, et al. Nat Immunol. 2008;9:263–271.
16 Hailfinger Stephan, Nogai Hendrik, Pelzer Christiane, et al. Proc Natl Acad Sci U S A. 2011;108(35):14596–14601.
17 Schlauderer Florian, Lammens Katja, Nagel Daniel, et al. Angew Chem Int Ed Engl. 2013;52:10384–10387.
18 Nagel Daniel, Spranger Stefani, Vincendeau Michelle, et al. Cancer Cell. 2012;22: 825–837.
19 Ryckmans Thomas, Edwards Martin P, Horne Val A, et al. Bioorg Med Chem Lett. 2009;19:4406–4409.
20 Johnson Ted W, Gallego Rebecca A, Edwards Martin P. J Med Chem. 2018;61: 6401–6420.