Novel oligoamine analogues inhibit lysine-specific demethylase 1 and induce reexpression of epigenetically silenced genes

Abstract

Purpose: Abnormal DNA CpG island hypermethylation and transcriptionally repressive histone modifications are associated with the aberrant silencing of tumor suppressor genes. Lysine methylation is a dynamic, enzymatically controlled process. Lysine-specific demethylase 1 (LSD1) has recently been identified as a histone lysine demethylase. LSD1 specifically catalyzes demethylation of mono- and dimethyl-lysine 4 of histone 3 (H3K4), key positive chromatin marks associated with transcriptional activation. We hypothesized that a novel class of oligoamine analogues would effectively inhibit LSD1 and thus cause the reexpression of aberrantly silenced genes.

Experimental design: Human colorectal cancer cells were treated with the oligoamines and changes in mono- and dimethyl-H3K4 and other chromatin marks were monitored. In addition, treated cells were evaluated for the reexpression of the aberrantly silenced secreted frizzled-related proteins (SFRP) Wnt signaling pathway antagonist genes. Finally, the effects of the LSD1 inhibitors were evaluated in an in vivo xenograft model.

Results: Treatment of HCT116 human colon adenocarcinoma cells in vitro resulted in increased H3K4 methylation and reexpression of silenced SFRP genes. This reexpression is also accompanied by a decrease in H3K9me2 repressive mark. Importantly, cotreatment with low doses of oligoamines and a DNA methyltransferase inhibitor highly induces the reexpression of the aberrantly silenced SFRP2 gene and results in significant inhibition of the growth of established tumors in a human colon tumor model in vivo.

Conclusions: The use of LSD1-inhibiting oligoamine analogues in combination with DNA methyltransferase inhibitors represents a highly promising and novel approach for epigenetic therapy of cancer.

Figure 1

Inhibition of LSD1 by polyamine analogues. (A) Structures of natural polyamines and members of the oligoamine family of polyamine analogues. (B) 0.038 μg/μl of purified Δ184 LSD1 protein were incubated with 5 μM H3K4me2 (1-21 aa) as substrate in the presence of 10 μM of the indicated analogue. Effect of oligoamines on enzymatic activity of LSD1 was examined using luminol-dependent chemiluminescence to measure the production of H₂O₂. The integral values were calibrated against standards containing known concentrations of H₂O₂, and the activities expressed as picomoles of H₂O₂ per milligram of protein per minute. Shown are means ± SD of independent experiments performed in triplicate.

Figure 2

Inhibition of recombinant human LSD1 by oligoamines. (A) 0.05 μg/ul of purified bulk histones were incubated with or without 0.01 μg/μl purified full-length LSD1 for 3 hr at 37°C. Effects of LSD1 enzymatic activity were analyzed by Western blotting using antibodies that specifically recognize the dimethyl group of H3K4. Relative levels were determined by quantitative Western analysis using the Odyssey infrared detection system. Total H3 was used as loading control. The results represent the mean of three determinations ± SD. (B) 0.038 μg/μl of purified Δ184 LSD1 protein were incubated with 5 μM H3K4me2 (1-21 aa) as substrate in the presence of increasing concentrations of PG-11144 and PG-11150. Effect of oligoamines on enzymatic activity of LSD1 was examined using luminol-dependent chemiluminescence to measure the production of H₂O₂. The vertical bars indicate mean ± SD. (C) The effects of increasing concentrations of PG-11144 on LSD1 activity in the presence of increasing substrate concentrations. Double reciprocal plots indicate inhibition of LSD1 by PG-11144 to be competitive. Shown are means ± SD of independent experiments performed in triplicate.

Figure 3

Inhibition of LSD1 by oligoamines increases global H3K4me2, resulting in re-expression of aberrantly silenced genes in treated human colon cancer cells. (A) HCT116 cells and RKO cells were exposed to increasing concentrations of PG-11144 or PG-11150 for 24 h or 48 hr, and 30 μg of nuclear protein/lane were analyzed for expression of LSD1, H3K4me1, H3K4me2, and H3 as a loading control. (B) HCT116 cells were treated for 24 h with the indicated concentrations of oligoamines and DAC. Total RNA was extracted for RT-PCR analysis of SFRP1 and SFRP2 expression. GAPDH is included as an internal control. (C) The levels of re-expression of SFRP1 and SFRP2 induced by oligoamines in HCT116 cells were compared to levels resulting from DAC treatment using SYBR Green/qPCR with GAPDH as an internal control. (D) HCT116 cells were treated for 24 h with the indicated concentrations of oligoamines or DAC. Total RNA was extracted for RT-PCR analysis of SFRP4 and GATA5 expression. GAPDH is included as an internal control.

Figure 4

Inhibition of LSD1 by oligoamines increases activating H3K4me2 and H3K4me1 marks and decreases repressive H3K9me2 marks at the promoters of re-expressed genes. (A) HCT116 cells were treated with 10 μM of PG-11144 for 24 hr. Quantitative chromatin immunoprecipitation analysis was used to determine the occupancy of the indicated promoters by multiple activating and repressive marks. (B) Composite graph showing enrichment of H3K4me2 in the proximal promoter region of SFRP2, spanning ∼-1000 bp to +300 bp relative to the transcriptional start site (TSS). The quantified results are the means of three independent experiments with S.D. as indicated.

Figure 5

Synergy of oligoamines and DNA methyltransferase inhibitor in the re-expression of an aberrantly silenced gene. HCT116 cells were treated for 24 hr with the indicated doses of DAC and oligoamines alone or simultaneously. RNA was extracted and cDNA was synthesized and subjected to quantitative real time PCR for the SFRP2 gene using the DNA-intercalating SYBR green reagent. GAPDH expression was used as an internal standard. The quantified results are the mean of three independent experiments performed in triplicate in the MyiQ single color real-time PCR machine (Bio-Rad). S.D. is indicated by the error bars.

Figure 6

Effects of polyamine analogues on LSD1 and tumor growth in nude mice bearing HCT116 xenografts. Mice were randomized into treatment (n = 5) and control (n = 5) groups by 20 days post-implantation of HCT116 xenograft. The treatment groups given as single agents were: vehicle (saline control at qdx5), PG-11144 (10 mg/kg, 2× wk), 2d (10 mg/kg qdx5), and 5-Aza (2 mg/kg qdx5). The combination treatment groups were: PG-11144 (10 mg/kg, 2× wk)/5-Aza (2 mg/kg qdx5), and 2d (10 mg/kg qdx5)/5-Aza (2 mg/kg qdx5). (A) Tumor sample of HCT116 xenografts treated with vehicle (saline control at qdx5), PG-11144 (10 mg/kg, 2× wk), or 2d (10 mg/kg qdx5) for 37 days. 30 μg of nuclear protein/lane were analyzed for expression of H3K4me2, and H3/H4 as a loading control. (B) Tumor volumes of mice measured twice weekly. The vertical bars indicate mean tumor size (in mm³) ± SE. (C) Weights of mice were measured weekly. The vertical bars indicate mean mouse weight (g) ± SE.