Identification and characterization of a novel small-molecule inhibitor of β-catenin signaling

Abstract

Hepatocellular carcinoma (HCC), the third most common cause of cancer-related deaths worldwide, lacks effective medical therapy. Large subsets of HCC demonstrate Wnt/β-catenin activation, making this an attractive therapeutic target. We report strategy and characterization of a novel small-molecule inhibitor, ICG-001, known to affect Wnt signaling by disrupting β-catenin-CREB binding protein interactions. We queried the ZINC online database for structural similarity to ICG-001 and identified PMED-1 as the lead compound, with ≥70% similarity to ICG-001. PMED-1 significantly reduced β-catenin activity in hepatoblastoma and several HCC cells, as determined by TOPflash reporter assay, with an IC50 ranging from 4.87 to 32 μmol/L. Although no toxicity was observed in primary human hepatocytes, PMED-1 inhibited Wnt target expression in HCC cells, including those with CTNNB1 mutations, and impaired cell proliferation and viability. PMED-1 treatment decreased β-catenin-CREB binding protein interactions without affecting total β-catenin levels or activity of other common kinases. PMED-1 treatment of Tg(OTM:d2EGFP) zebrafish expressing GFP under the β-catenin/Tcf reporter led to a notable decrease in β-catenin activity. The PMED effect on β-catenin signaling lasted from 12 to 24 hours in vitro and 6 to 15 hours in vivo. Thus, using a rapid and cost-effective computational methodology, we have identified a novel and specific small-molecule inhibitor of Wnt signaling that may have implications for HCC treatment.

Figure 1

Computational approach to identifying SMs that inhibit Wnt signaling to affect primary human hepatocytes and hepatic tumor cells. A: A search of the ZINC version 8.0 database lead-like subset identified 13,399 SMs structurally similar to ICG-001. A similarity score of ≥70% restricted the hits to 24 molecules, of which 20 passed in silico ADME/tox filtering. B: Chemical structure of ICG-001, PMED-1, and PMED-2. C: HepG2 cells treated with 200 μmol/L PMED-1 or PMED-2 for 24 hours exhibited a significant decrease in TOPflash luciferase reporter activity, compared with 2% DMSO. D: Primary hepatocytes treated with 2% DMSO or with 200 μmol/L PMED-1 or PMED-2 for 18 or 48 hours exhibited a significant decrease in viability by MTT assay only after PMED-2 treatment. E: HepG2 cells treated with 200 μmol/L PMED-1 for 24 hours monitored by live-cell imaging exhibit notable morphological changes reminiscent of cell death as early as 3 hours after treatment, compared with the 2% DMSO-treated controls. F: HepG2 cells treated with 100 μmol/L PMED-1 exhibited a significant decrease in cell viability as assessed by MTT assay, compared with 1% DMSO control. Data are expressed as means ± SD. ^∗P < 0.05, ^∗∗P < 0.01. Original magnification, ×200 (E).

Figure 2

PMED-1 targets β-catenin signaling to affect cell proliferation in HepG2 cells. A: HepG2 cells treated with 25 μmol/L PMED-1 for 24 hours (white bars) exhibited a significant decrease in TOPflash reporter activity, but less so at 48 hours (black bars). B: Decrease in Ki-67⁺ cells (red) was observed after treatment of HepG2 cells with 25 μmol/L PMED-1 for 24 hours, compared with 0.2% DMSO. C: DNA synthesis in HepG2 cells as measured by [³H]thymidine incorporation exhibited a decrease in response to 24 (white bars) or 72 (gray bars) hours of treatment with 25 μmol/L of PMED-1, compared with 0.2% DMSO. D: Western blots using HepG2 cell lysates reveal decreased β-catenin targets after 25 μmol/L PMED-1 treatment. Data are expressed as means ± SD. ^∗P < 0.05, ^∗∗P < 0.01. Original magnification, ×100 (B).

Figure 3

IC₅₀ of PMED-1 in various HCC cells and the effect of PMED-1 on β-catenin signaling and biology. A–D: IC₅₀ was determined by response of TOPflash reporter activity to PMED-1 treatment for 24 hours in Snu-449 (A), Snu-398 (B), Huh-7 (C), and Hep3B (D) cells. E: Relative decrease in TOPflash reporter activity after 25 μmol/L PMED-1 treatment of various HCC cells for 24 hours. F: The most robust decrease in viability as measured by MTT assay was observed in Snu-398 cells after 24 hours of 25 μmol/L PMED-1 treatment. G: DNA synthesis as measured by [³H]thymidine incorporation was significantly reduced in Snu-398 cells treated for 24 (white bars) or 72 (gray bars) hours with 25 μmol/L PMED-1, compared with DMSO. H: Western blots using whole-cell lysates from Snu-398, Snu-449, and Hep3B cells treated with 25 μmol/L PMED-1 reveal notable decreases in β-catenin downstream targets. Data are expressed as means ± SD. ^∗P < 0.05, ^∗∗P < 0.01.

Figure 4

PMED-1 affects β-catenin–CBP interactions but does not affect activity of other common kinases. A: Western blot analysis of whole-cell lysates from untreated HepG2 cells or HepG2 cells treated for 24 hours with either 2% DMSO, 25 μmol/L PMED-1, or 25 μmol/L ICG-001 used for immunoprecipitation revealed significantly less CBP pull-down with anti–β-catenin antibody in cells treated with PMED-1 or ICG-001, compared with DMSO. Densitometry values (arbitrary units) for CBP are provided under the bands. B–K: Effect of H89 and PMED-1 on activity of kinases as assessed by IMAP-based FP assay (AKT and PKD) or TR-Fret assay (CAK, Plk1, and Plk2). Three replicates are shown. B: H89 decreases CAK activity in a dose-dependent manner. C: PMED-1 does not affect AKT activity. D: H89 decreases CAK activity. E: PMED-1 does not affect CAK activity. F: H89 decreases PKD activity at higher concentrations. G: PMED-1 does not affect PKD activity. H: H89 decreases Plk1 activity in a largely dose-dependent manner. I: PMED-1 had an inhibitory effect on Plk1 up to −1 μmol/L on the log scale (0.097 μmol/L); however, no decrease was evident at higher concentrations. J: H89 decreases Plk2 activity in a dose-dependent manner. K: PMED-1 does not affect Plk2 activity. IP, immunoprecipitation.

Figure 5

Efficacy and duration of the effect of PMED-1 effect on β-catenin activity in vivo and in vitro. A: GFP⁺ embryos obtained from outcrossing Tg(OTM:d2EGFP) zebrafish were treated with DMSO, XAV939, or PMED-1 for 15 hours (40 to 55 hpf), and GFP expression was assessed under a fluorescence microscope. Shown are bright-field images (top row), GFP fluorescence images (middle row), and merged images (bottom row). B: Stably transfected Hep3B cells with S33Y-β-catenin, treated with a single dose of PMED-1 at 25 μmol/L, exhibited a gradual decrease in TOPflash reporter activity; the decrease reached significance at 18 and 24 hours after treatment, with rebound at 36 and 48 hours. C: Epifluorescence images of GFP expression in Tg(OTM:d2EGFP) embryos. Embryos strongly GFP⁺ [obtained from incrossing Tg(OTM:d2EGFP) zebrafish] were treated with DMSO, XAV939, or PMED-1 from 48 hpf. Embryos were harvested 6, 12, or 24 hours later, and GFP expression was assessed under a fluorescence microscope. ^∗P < 0.05. The proportion of embryos (n/N) exhibiting the GFP expression is indicated at the lower left of each image. All images are dorsal view, with anterior to the left. Scale bars: 100 μm (A and C). li, liver; ov, otic vesicle; pf, pectoral fin; re, retina; te, tectum.