Identification of ryuvidine as a KDM5A inhibitor

Abstract

KDM5 family members (A, B, C and D) that demethylate H3K4me3 have been shown to be involved in human cancers. Here we performed screening for KDM5A inhibitors from chemical libraries using the AlphaScreen method and identified a battery of screening hits that inhibited recombinant KDM5A. These compounds were further subjected to cell-based screening using a reporter gene that responded to KDM5A inhibition and 6 compounds were obtained as candidate inhibitors. When further confirmation of their inhibition activity on cellular KDM5A was made by immunostaining H3K4me3 in KDM5A-overexpressing cells, ryuvidine clearly repressed H3K4me3 demethylation. Ryuvidine prevented generation of gefitinib-tolerant human small-cell lung cancer PC9 cells and also inhibited the growth of the drug-tolerant cells at concentrations that did not affect the growth of parental PC9 cells. Ryuvidine inhibited not only KDM5A but also recombinant KDM5B and C; KDM5B was the most sensitive to the inhibitor. These results warrant that ryuvidine may serve as a lead compound for KDM5 targeted therapeutics.

Figure 1

Knockdown of KDM5A in HEK293 cells promotes expression of TFPI-2. RNA and protein were extracted at 48 h after introduction of siRNAs but those of TFPI-2 were extracted at 96 h after siRNA introduction. (A) siRNA85 and 86 significantly reduced KDM5A mRNA compared with negative control (NC) siRNA. GAPDH mRNA was used for normalization. (B) TFPI-2 mRNA was significantly increased upon KDM5A knockdown. (C) KDM5A protein was decreased upon siRNA-mediated knockdown. (D) TFPI-2 protein was increased upon KDM5A knockdown by siRNA. Data are shown as means ± SD (n = 3).

Figure 2

Reporter assay with HEK293TFPI-2-Luc cells. A HEK293 cell line was established that expressed the luciferase gene driven by the TFPI-2 promoter. (A) Knockdown of KMD5A and LSD1 increases activity of the luciferase reporter driven by the TFPI-2 promoter. The HEK293 stable cell line was transfected with the indicated siRNAs, and luciferase activity was assayed 72 h later. Luciferase levels in control transfected cells were set as 100%. (B) PBIT, a KDM5A inhibitor, increases reporter activity. Luciferase activity was assayed at 48 h after addition of PBIT. (C) NCL-1, a LSD1 inhibitor, increases reporter activity. Luciferase activity was assayed at 72 h after addition of NCL-1. Error bars show the means ± SD (n = 3). *

Figure 3

Ryuvidine inhibition of KDM5A was measured by MALDI-TOF/MS methods. Ryuvidine inhibition of KDM5A in the presence of 2-OG was evaluated. Levels of H3K4me3 (me3), H3K4me2 (me2) and H3K4me1 (me1) were measured by MALDI-TOF/MS. H3K4me3 at reaction time 0 min was defined as 100%. Data are presented as means ± SD (n = 3).

Figure 4

Ryuvidine repressed H3K4 demethylation in HEK293 cells expressing Flag-tagged KDM5A. Flag-tagged KDM5A was transiently overexpressed in HEK293 cells and cells were treated with ryuvidineat 2 uM for 48 h. Flag and H3K4me3 were detected by immunostaining; DAPI staining indicates nuclei. Transfection of empty vector did not produce Flag-positive cells and H3K4me3 was clearly detected. H3K4me3 was markedly reduced in Flag-positive cells in which KDM5A was overexpressed. Ryuvidine repressed the reduction of H3K4me3.

Figure 5

Effect of ryuvidine on gefitinib-tolerant PC9 cells (DTEPs). (A) Ryuvidine inhibits growth of gefitinib-tolerant PC9 cells (DTEPs) but not parental PC9 cells. MTT values in the presence of DMSO alone were defined as 100%. Error bars show the means ± SD (n = 3). (B) Ryuvidine represses generation of DTEPs.