Synthesis and evaluation of derivatives of the proteasome deubiquitinase inhibitor b-AP15

Abstract

The ubiquitin-proteasome system (UPS) is increasingly recognized as a therapeutic target for the development of anticancer therapies. The success of the 20S proteasome core particle (20S CP) inhibitor bortezomib in the clinical management of multiple myeloma has raised the possibility of identifying other UPS components for therapeutic intervention. We previously identified the small molecule b-AP15 as an inhibitor of 19S proteasome deubiquitinase (DUB) activity. Building upon our previous data, we performed a structure-activity relationship (SAR) study on b-AP15 and identified VLX1570 as an analog with promising properties, including enhanced potency and improved solubility in aqueous solution. In silico modeling was consistent with interaction of VLX1570 with key cysteine residues located at the active sites of the proteasome DUBs USP14 and UCHL5. VLX1570 was found to inhibit proteasome deubiquitinase activity in vitro in a manner consistent with competitive inhibition. Furthermore, using active-site-directed probes, VLX1570 also inhibited proteasome DUB activity in exposed cells. Importantly, VLX1570 did not show inhibitory activity on a panel of recombinant non-proteasome DUBs, on recombinant kinases, or on caspase-3 activity, suggesting that VLX1570 is not an overtly reactive general enzyme inhibitor. Taken together, our data shows the chemical and biological properties of VLX1570 as an optimized proteasome DUB inhibitor.

Keywords: chalcone; deubiquitinase; in silico modeling; inhibitor; lead optimization; proteasome.

Figure 1

Structure–activity relationships for b-AP15. (A) Structure of the DUB inhibitor lead b-AP15 (VLX1500); also shown is the IC₅₀ value on HCT116 cells (CI: 95%). Cells were continuously exposed for 72 h and viability determined by FMCA. (B) Analogs containing hydroxyls on the side aryls or carboxyls in the piperidine ring. IC₅₀ values from HCT116 cells (72-h exposure). (C) Pairwise comparisons of piperidine- and azepane-containing compounds. Shown are three pairs of compounds with azepane or piperidine central rings and the in vitro antiproliferative activities associated with these compounds. IC₅₀ values from HCT116 cells (72-h exposure). (D) Effect of variations of the position on the nitro-group on the side aryls and varying the decoration of the nitrogen of the azepane ring. IC₅₀ values from HCT116 cells (72-h exposure).

Figure 2

In silico docking of VLX1570 with USP14 and UCHL5. (A) Precovalent docked poses of VLX1570 with UCHL5. Electrostatic colored surface for UCHL5 is shown to illustrate the crevice–groove, which VLX1570 occupies prior to catalytic linkage at the –SH thiol. (B) Covalent docked pose of VLX1570 with UCHL5. A molecular model for the predicted linkage between Cys88 and the VLX1570 β-carbon is shown. Important UCHL5 protein amino acid side chains with interaction with the VLX1570 molecule are labeled. Dashed lines indicate favorable interactions with drug and protein (VdW, H-bonds, or electrostatic). (C) Precovalent docked poses of VLX1570 with USP14. Electrostatic colored surface for USP14 is shown to illustrate the crevice–groove that VLX1570 occupies prior to catalytic linkage at the –SH thiol. (D) Covalent docked pose of VLX1570 with USP14. A molecular model for the predicted linkage between Cys114 and the VLX1570 β-carbon is shown. Important USP14 protein amino acid side chains with interactions with VLX1570 are labeled.

Figure 3

DTT affects the cytotoxicity of VLX1570. MelJuSo-Ub^G76V-YFP cells were exposed to different concentrations of VLX1570 in the presence or absence of DTT (dithiothreitol) for 72 h. Cell viability was assessed using the MTT assay.

Figure 4

(A) Metabolic pathways of VLX1570 in human cryopreserved hepatocytes. Hepatocytes were exposed to VLX1570 for 2 h, and metabolites were determined by LC/MS (see Materials and Methods). (B) Pharmacokinetics of VLX1570 in rats. VLX1570 was injected into Sprague Dawley rats (5 mg/kg) during 10 min. Plasma was collected after different times, and VLX1570 levels were determined by MS.

Figure 5

(A) Inhibition of active-site-directed labeling of proteasome deubiquitinases. Purified 19S proteasomes (5 nm) were pre-treated with DMSO, VLX1570, and b-AP15 (50 μm) for 10 min at room temperature, followed by labeling with HA-UbVS and immunoblotting. (B) Inhibition of Ub-VS labeling of USP14 by VLX1570 in exposed cells. KMS-11 myeloma cells were treated with 1 μm VLX1570 for 6 h and 25 μg protein from whole-cell lysates were subsequently labeled with Ub-VS (1 μm), followed by SDS gel electrophoresis and immunoblotting with an USP14 antibody. Shown are triplicate samples of cells exposed to solvent or VLX1570. The lower band represents inactive USP14 enzyme. (C) Induction of polyubiquitin conjugates in KMS-11 myeloma cells. Cells were exposed to DMSO, 1 μm VLX1570, 100 nm bortezomib, 100 μm IU1, 5 μm WP1130, or 1 μm b-AP15 for 6 and 18 h, extracts prepared and subjected to immunoblotting with the indicated antibodies.

Figure 6

(A) Determination of deubiquitinase inhibition by VLX1570. The following recombinant deubiquitinases were tested with regard to inhibition of activity by 20 μm. The tested DUBs were: 1. USP1; 2. USP2; 3. USP4; 4. USP5; 5. USP5 (+Ub@Kd); 6. USP5 (+Ub@Bmax); 7. USP6; 8. USP7; 9. USP8; 10. USP9x; 11. USP11; 12. USP14 (PS-VS@Kd); 13. USP15; 14. USP16; 15. USP19; 16. USP20; 17. USP21; 18. USP25; 19. USP28; 20. USP30; 21. USP35; 22. USP36; 23. USP45; 24. CYLD; 25. UCHL1; 26. UCHL3; 27. UCHL5; 28. BAP1; 29. OTU1; 30. OTUB2; 31. OTUD3; 32. OTUD5; 33. OTUD6A; 34. OTUD6B; 35. Cezanne; 36. AMSH-LP; 37. AMSH-LP (+Zn); 38. Ataxin3; 39. Ataxin3L; 40. JOSD1; 41. JOSD2. Note that USP14 was not inhibited under conditions where the enzyme was reconstituted with Ub-vinyl sulfone-treated proteasomes. Experiment was performed by Ubiquigent (http://www.ubiquigent.com). (B) In vitro inhibition of caspase-3 activity. VLX1570 (40 μm) or DEVD-CHO (10 μm) was incubated with caspase-3 prior to the addition of Rh110 caspase-3/7 substrate, and the fluorometric signal was measured every minute for 5 min. (C) Determination of kinase inhibition by VLX1570. The enzymatic activities of 211 different kinases were determined in the presence of absence of 10 μm VLX1570. The 5 most strongly inhibited kinases were CDK4 (cyclin-dependent kinase 4) (77% inhibition), CaMK4 (48%), FLT-4 (VEGFR3) (47%), NEK4 (43%), and NDR1 (35%). Experiment was performed by CEREP (http://www.cerep.fr).

Figure 7

MelJuSo-YFP and HCT116 cells were exposed to 5 μm WP1130, 1 μm VLX1570, or 1 μm b-AP15 for 4 h, cell extracts prepared and subjected to immunoblotting with Mcl-1 and β-actin antibodies.