The histone deacetylase inhibitor trichostatin a decreases lymphangiogenesis by inducing apoptosis and cell cycle arrest via p21-dependent pathways

Abstract

Background: The formation of new lymphatic vessels provides an additional route for tumour cells to metastasize. Therefore, inhibiting lymphangiogenesis represents an interesting target in cancer therapy. First evidence suggests that histone deacetylase inhibitors (HDACi) may mediate part of their antitumor effects by interfering with lymphangiogenesis. However, the underlying mechanisms of HDACi induced anti-lymphangiogenic properties are not fully investigated so far and in part remain unknown.

Methods: Human lymphatic endothelial cells (LEC) were cultured in vitro and treated with or without HDACi. Effects of HDACi on proliferation and cell cycle progress were analysed by BrdU assay and flow cytometry. Apoptosis was measured by quantifying mono- and oligonucleosomes in the cytoplasmic fraction of cell lysates. In vitro lymphangiogenesis was investigated using the Matrigel short term lymphangiogenesis assay. The effects of TSA on cell cycle regulatory proteins and apoptosis-related proteins were examined by western blotting, immunofluorescence staining and semi-quantitative RT-PCR. Protein- and mRNA half-life of p21 were analysed by western blotting and quantitative RT-PCR. The activity of the p21 promoter was determined using a dual luciferase assay and DNA-binding activity of Sp1/3 was investigated using EMSA. Furthermore, siRNA assays were performed to analyse the role of p21 and p53 on TSA-mediated anti-lymphangiogenic effects.

Results: We found that HDACi inhibited cell proliferation and that the pan-HDACi TSA induced G0/G1 arrest in LEC. Cell cycle arrest was accompanied by up-regulation of p21, p27 and p53. Additionally, we observed that p21 protein accumulated in cellular nuclei after treatment with TSA. Moreover, we found that p21 mRNA was significantly up-regulated by TSA, while the protein and mRNA half-life remained largely unaffected. The promoter activity of p21 was enhanced by TSA indicating a transcriptional mechanism. Subsequent EMSA analyses showed increased constitutive Sp1/3-dependent DNA binding in response to HDACi. We demonstrated that p53 was not required for TSA induced p21 expression and growth inhibition of LECs. Interestingly, siRNA-mediated p21 depletion almost completely reversed the anti-proliferative effects of TSA in LEC. In addition, TSA induced apoptosis by cytochrome c release contributed to activating caspases-9, -7 and -3 and downregulating the anti-apoptotic proteins cIAP-1 and -2.

Conclusions: In conclusion, we demonstrate that TSA - a pan-HDACi - has distinct anti-lymphangiogenic effects in primary human lymphatic endothelial cells by activating intrinsic apoptotic pathway and cell cycle arrest via p21-dependent pathways.

Keywords: G0/G1 cell cycle arrest; Histon deacetylase inhibitors (HDACi); Lymphangiogenesis; intrinsic apoptotic pathway; p21.

Fig. 1

Effects of HDACi on primary human lymphatic endothelial cells. a, b Cells were exposed to increasing concentrations of trichostatin A (TSA), sodium butyrate (NaB) and valproic acid (VPA) alone and (c) in the presence or absence of 20 ng/ml VEGF-A and 100 ng/ml VEGF-C for 24 h as indicated. Cell proliferation and viability was measured using the BrdU (a, c) and Alamar blue assay (b). Average absorbance values (mean ± SE) from 4 wells per experimental condition are displayed; data are expressed as cell proliferation and vialibity in percentage (%) with regard to solvent controls (=100%; ethanol and H₂O). Results were confirmed in four independent sets of experiments. *p < 0.05 vs Ctrl. d Quantification of cytotoxicity. Cells were incubated with increasing concentrations of TSA for 24 h as indicated. Cytotoxicity was quantified by using the LDH assay. Average absorbance values (mean ± SE) from quadruplicate determinations per experimental condition were calculated; data are expressed as cytotoxicity in percentage (%). e A two-dimensional, short term in vitro Matrigel assay of LECs that were left untreated or were incubated with 400 nM TSA for 6 h on Matrigel. Results were confirmed in three independent sets of experiments

Fig. 2

TSA induces G0/G1 cell cycle arrest in LECs. a Cells were incubated with 400 nM TSA and ethanol as a control for 24 h after blocking the S-phase of the cell-cycle by treatment with a serum depleted-medium for 24 h. Analysis of cell cycle as assessed by FACS using propidium iodide-stained LECs. The figure shows the percentage of cells in G0/G1, S and G2/M with regard to solvent controls (Ethanol). Data displayed are representative of eight experiments that were performed revealing comparable results. *p < 0.05 vs Ctrl. b, c and d Representative western blot analyses of LECs that were left untreated (solvent only) or were treated with TSA in a concentration- and time-dependent manner. Cell cycle control related proteins, like p21, p27, p53, cyclin A/D1, CDK4, CDK6 and α-Tubulin as loading control were detected by enhanced chemiluminescence. Comparable results were obtained from three independent experiments

Fig. 3

TSA treatment induces the intrinsic apoptotic pathway in LECs. a Cells were exposed to 400 nM TSA for 24 h as indicated. For apoptosis determination, we used a colorimetric assay that quantified histone complexed DNA fragments. Average absorbance values (mean ± SE) from 3 wells per experimental condition are displayed; data are expressed as apoptotic cells in percentage (%) with regard to solvent controls (=100%; ethanol). *p < 0.05 vs Ctrl. b and (c) Representative western blot analyses of LECs that were left untreated (solvent only) or were treated with TSA in a concentration- and time-dependent manner. Apoptosis related proteins, like cleaved caspases 9, 3 and 7, cIAP-1/2 and α-Tubulin as loading control were detected by enhanced chemiluminescence. Comparable results were obtained from three independent experiments. d LECs were left untreated (solvent only, Ethanol) or were treated with 400 nM TSA for 24 h. The release of cyctochrome c to cytosol was measured using the cytochrome c ELISA. Mean values from four independent experiments are depicted ± SE (error bars). *p < 0.05 vs Ctrl

Fig. 4

TSA mediates nuclear accumulation of p21 protein in LECs. Subconfluent LECs were either left untreated (solvent only) or treated with 400 nM TSA for 24 h. Cells were fixed and stained with anti-p21 antibody. Blue: Hoechst stained nuclei; red: p21 protein stained with specific antibody and secondary antibody conjugated with Alexa Fluor 546. Comparable results were obtained from three independent experiments

Fig. 5

p21, but not p27 and p53, mRNA expression is upregulated by TSA. a Semiquantitative RT-PCR analyses for p21, p27, p53 and β-Actin as loading control were performed on total RNA extracted from subconfluent LECs. LECs were left untreated (solvent only, ethanol), or were treated with TSA (at 400 nM) for the indicated times. Results were confirmed in three independent sets of experiments. b LECs were incubated in the presence or absence of TSA (400 nM) for 3 h and incubated for 0, 1, 3 and 6 h with fresh media containing actinomycin D (10 μg/ml). Then, p21 mRNA levels were quantified using quantitative PCR. Comparable results were obtained from three independent experiments. c LECs were incubated in the presence or absence of TSA (400 nM) for 3 h and incubated for 0, 1, 3 and 6 h with fresh media containing cycloheximide (10 μg/ml). p21 and α-tubulin protein were detected by enhanced chemiluminescence. p21 protein remaining at these time points was determined by densitometric scanning, the results of which were normalized to amounts of α-tubulin. Comparable results were obtained from three independent experiments

Fig. 6

TSA induces p21 promoter activity via increased Sp1/3-dependent binding to the −100/–81 bp sequence. a Analyses of 5’-deletional p21 promoter-based luciferase (Luc) constructs in LECs. Schematic representation of the respective reporter gene constructs on the right and the relative Luc activities (expressed as % basal activity of the −2325/+8 construct) in graphic format on the left. Black bars, untreated controls (Ctrl.); grey bars, TSA-treated cells (mean ± SE of three independent duplicate assays). *p < 0.05 vs Ctrl. b Representative EMSA using nuclear extracts of untreated (Ethanol only) and TSA-treated (at 400 nM, 1 h) LECs. The DNA-binding activity of Sp1/3 was measured by using nuclear extracts and biotin-labeled DNA probes with or without a competitive cold DNA probe. Supershift experiments were carried out by incubating nuclear extracts with Sp1/3 antibodies. The formation of Sp-dependent binding complexes is indicated by arrows to the left

Fig. 7

The pan-HDACi TSA mediates its anti-proliferative effects on LECs through p21 dependent pathways. a Cells were treated with siRNA against p21, p53 (b) and control siRNA and were exposed to 400 nM TSA or solvent only (=Ethanol; 100 %) for 24 h as indicated. Cell proliferation was measured using the BrdU assay. Average absorbance values (mean ± SE) from 3 wells per experimental condition are displayed; data are expressed as cell proliferation in percentage (%) with regard to solvent controls (=100%; ethanol). *^,**p < 0.05 vs Ctrl. c Representative Western blot analyses of LECs that were incubated with siRNA against p53 or control siRNA and treated with TSA (at 400 nM) or solvent only (Ethanol) as indicated for 24 h. p53, p21 and α-Tubulin protein as loading control were detected by enhanced chemiluminescence. Comparable results were obtained from three independent experiments