Background
DNA methylation and histone modification are the two major epigenetic mechanisms catalyzed by DNMTs and HDACs, respectively [
1]. HDACs remove the acetyl groups from histones, whilst DNMTs catalyse the transfer of a methyl group from S-adenosylmethionine to the 5-carbon position of the cytosine pyrimidine ring, both leading to the condensation of chromatin to its inactive state [
2,
3]. In cancer cells, an abundance of hypo-acetylated histones is usually associated with DNA hyper-methylation and gene silencing [
4]. These findings are the basis for the development of HDAC and DNMT inhibitors as cancer therapeutics. Such compounds block the activity of HDACs and DNMTs, leading to increased expression of epigenetically silenced genes which mediate cellular and metabolic changes such as cell growth arrest, differentiation and apoptosis [
5‐
9].
Hydrophobic vorinostat (suberoylanilide hydroxamic acid, SAHA) and hydrophilic decitabine (5-aza-2
′-deoxycytidine, Dacogen) are US Food and Drug Administration (FDA) approved HDAC and DNMT inhibitors for the treatment of cutaneous T-cell lymphoma and myelodysplastic syndrome, respectively [
10,
11]. The combination of vorinostat and decitabine have been shown to have promising activity in patients with myelodysplastic syndrome without significant toxicity in a phase I clinical trial [
12]. Under neutral conditions, decitabine has a reported half-life of 7 days at 4°C or 21 hours at 37°C
in vitro[
13]. However, decitabine is degraded more rapidly
in vivo with a half-life of only 25 minutes [
13]. Such chemical instability of decitabine has led to its administration in the clinic as a cold and continuous intravenous infusion in an effort to reach the maximal-tolerated doses required to achieve clinical response [
14,
15].
The development of drug formulation using nanotechnology (e.g. liposomes) has been used to improve drug stability [
16,
17]. Despite the potential of improving the delivery of epigenetic drugs, the subsequent assessment of changes in their epigenetic activity is largely dependent on the availability of a suitable and rapid screening bioassay. A commonly used cell-based assay for both DNMT and HDAC inhibitors is the quantification of the re-expression of known epigenetically-silenced genes by reverse transcription polymerase chain reaction (RT-PCR) and western blot analysis [
5,
18]. However, this traditional approach is not high-throughput and may produce gene-specific results. Other assays that have been used include estimation of global DNA methylation using capillary electrophoresis, DNA digestion with methylation-sensitive restriction enzymes, or analysis of specific DNA methylation using bisulfite sequencing and methylation-specific PCR [
19]. However, these assay systems designated for assaying DNMT or HDAC inhibitors are time-consuming, cumbersome and subject to misinterpretation [
20‐
22]. Consequently, the rapid identification and validation of novel epigenetic drugs are hampered due to the lack of an efficient screening method.
In this study, a cell-based assay system was developed to compare the activity of different epigenetic drugs. This assay system is based on mammalian MCF10A cells expressing a fusion protein between red-fluorescent protein (RFP) and bacterial nitroreductase (TMnfsB) driven by CMV promoter. Epigenetic silencing has been shown to silence genes driven by CMV promoter in both stably transfected cells and transgenic pigs [
23,
24]. Silenced CMV promoter driven genes were shown to be reactivated after treatment with epigenetic drugs such as butyrate, trichostatin A and decitabine [
23]. Human cells expressing TMnfsB are able to metabolize the monofunctional alkylating prodrug CB1954 (5-(azaridin-1-yl)-2,4-dinitro-benzamide) to highly cytotoxic hydroxylamino- and amino-derivatives, which induce rapid cell death [
25]. Therefore, TMnfsB was utilized as a tool to obtain clones with inactivated CMV promoters. The
TMnfsB open reading frame has been codon optimized to increase the sensitivity of stable human cell lines to the prodrug CB1954 [
26]. An assay system for gene reactivation was developed by identifying clones where expression of RFP-TMnfsB was suppressed at the transcriptional level, but could be re-established by subsequent treatment with epigenetic drugs. Since RFP expression in these clones is low, it was used as a signal to evaluate the reactivation of gene expression by flow cytometry. Using this newly developed assay system, it was shown that decitabine which encapsulated in the liposomes has a higher gene restoring ability than pure decitabine, zebularine and RG108.
Methods
Plasmids
The mammalianized nitroreductase gene B (
TMnfsB) vector was generated by subcloning the nitroreductase open reading frame from existing constructs kindly provided by Grohmann et al. [
26] into the pDsRED-C1-monomer vector at a
XhoI/
BamHI site. A retroviral plasmid pLNCX2-
RFP-TMnfsB expressing RFP-TMnfsB fusion was generated by subcloning the
RFP-TMnfsB coding fragment from the existing construct pDsRED-
TMnfsB (
SnaBI/
BamHI) into the pLNCX2 vector (
SnaBI/
BglII). All constructs were confirmed by sequencing using appropriate primers (Additional file
1).
Cell culture
All human cell lines were purchased from the American Type Culture Collection (ATCC) except the Phoenix retrovirus producer cell line which was kindly provided by Prof. Garry Nolan of Stanford University (United States). All cell lines were grown in the ATCC recommended media.
Reagents
CB1954 (soluble to 2 mg/mL in aqueous solution), decitabine (soluble to 50 mg/mL in aqueous solution), 2(1H)-pyrimidinone riboside (zebularine; soluble to 16 mg/mL in DMSO) and RG108 (soluble to 10 mg/mL in DMSO) were purchased from Sigma. RG108 is known to be an ineffective DNMT inhibitor [
27] and was used as a negative control. Vorinostat (10 mM) was kindly supplied by Dr. Lisa Butler of The University of Adelaide (South Australia). All drugs were dissolved in DMSO except decitabine, which was prepared in water for liposomal formulation. The synthetic lipids 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] sodium salt (DOPG), 1,2 distearoyl-sn-glycero-3-phosphocholine (DSPC), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-2000] ammonium salt (DSPE-PEG2000) and natural cholesterol lipid were purchased from Avanti Polar Lipids.
Generation of stable cell line and clonal selection
Recombinant retrovirus encoding RFP-TMnfsB was produced using the Phoenix packaging cell line transfected with Lipofectamine 2000 (Invitrogen) according to the recommended protocol. Stable cell lines expressing RFP-TMnfsB were generated by G418 selection of MCF10A cells transduced with retrovirus expressing RFP-TMnfsB for approximately 2 months. G418-resistant MCF10A cells were grown into colonies in 10 cm dishes and potential clones where TMnfsB was spontaneously silenced were isolated by treating these colonies with 5 μM of CB1954 for 72 hours. Surviving colonies, which were potentially epigenetically silenced, were isolated as CB1954-resistant clones. The integrity of RFP-TMnfsB in CB1954-resistant clones was determined by screening using RT-PCR. Finally, colonies with silenced RFP-TMnfsB insert were identified by assessing TMnfsB and RFP expression using RT-PCR and flow cytometry, respectively, after treatment with epigenetic drugs.
Real-time polymerase chain reaction (RT-PCR)
RNA and DNA from the cells were extracted using the RNeasy plant mini kit (Qiagen) and the DNeasy Blood and Tissue Kit (Qiagen), respectively. cDNA was generated using random primers and 20 U of reverse transcriptase (Promega).
TXNIP,
TMnfsB and
RFP-TMnfsB expression were determined by qRT-PCR using IQ™ SYBR green supermix (Biorad) and primers listed in Additional file
1. Cycling conditions were: 10 min at 95°C followed by 40 repeats of 95°C for 10 s, annealing at appropriate temperature for 15 s and extension at 72°C for 10 s. β-actin expression was used for normalization of target gene expression.
Western blotting
Western blot analysis of RFP-TMnfsB fusion protein expressed in MCF10A cells was performed using a rabbit polyclonal anti-RFP antibody (Invitrogen) or mouse anti-β-actin antibody (Sigma-Aldrich), and a secondary donkey anti-rabbit IgG-HRP (GE Healthcare) or a sheep anti-mouse IgG-HRP (GE Healthcare) [
28]. Total cellular proteins were extracted as described previously [
29] and visualized by an Enhanced Chemiluminescence Detection Kit (Amersham Biosciences).
Flow cytometry
The reactivation of silenced RFP-TMnfsB was determined by flow cytometry. Cells were plated at 40% 24 hours prior to treatment. The approximate doubling time of the cells is 48 hours. Cells were treated with each drug (decitabine 1, 5, 10, 30 and 50 μM; zebularine 50, 100, 250 and 500 μM; RG108 10 and 100 μM; vorinostat 1 and 2 μM) for 48 or 72 hours in triplicate. The red-fluorescence of cells was analyzed at a log scale of geometric mean of FL3-H using FACSCalibur flow cytometer (BD). Data were processed using WinMDI v2.8 software.
Preparation of liposomal decitabine
Liposomal formulations were prepared according to the method developed by Sunoqrot and colleagues with minor modifications [
30]. Briefly, 5 mg (32.5 mol%) DOPG, 4.9 mg (32.1 mol%) DSPC, 1.8 mg (3.3 mol%) DSPE-PEG2000 and 2.4 mg (32.1 mol%) cholesterol were dissolved in 5 mL of chloroform. Thin lipid films were generated after removing the solvent in a rotary evaporator for 2 hours at room temperature. Liposomes were formed when thin lipid films (4 mM) were hydrated in 5 mL of water or 0.88 mM decitabine dissolved in water for 1 hour at room temperature and stored at 4°C. The samples were extruded ten times using 200 and 400 nm polycarbonate membranes to obtain unilamellar liposomes.
HPLC (Shimadzu LC-10AT) analysis was done using a XTerraTM C8 analytical column at 254 nm, using MiliQ water as mobile phase and a flow rate of 0.8 mL/min. The limit of quantification of decitabine is 10 ng/mL [
31,
32].
Liposomes characterization
The size and zeta potential of liposomes were characterized by dynamic laser light scattering (Malvern Zetasizer Nanoseries). Data are expressed as the mean plus standard deviation of three technical repetitive measurements. For determination of encapsulation efficiency, free decitabine in the supernatant was collected after centrifugation at 82,508 xg for 30 minutes at 4°C and measured by HPLC. The encapsulation efficacy of decitabine was defined as the mass ratio between the amount of drugs incorporated in liposomes and that used in the liposome preparation.
Controlled release study of liposomes formulated decitabine
A controlled release study was performed using dialysis tubing (regenerated cellulose tubing, Mw cut-off 12000, 43 mm flat width, Crown Scientific, Australia) incubated in phosphate buffered saline (PBS) at 37°C. A 0.25 mL decitabine liposome suspension was added to the dialysis tubing immersed in a beaker with 10 mL of PBS as the release medium. Aliquots of 0.1 mL were collected from the solution outside the dialysis tubing at different time points. The volume of PBS was maintained by addition of 0.1 mL PBS after each withdrawal. The concentration of decitabine in each sample was determined using HPLC.
Statistical analysis
Data were analyzed by GraphPad Prism (GraphPad Software, Inc.) using unpaired two-tailed t-tests, and linear and nonlinear regression.
Discussion
EPISSAY, a cell-based assay system for screening of epigenetic drugs was developed based on the human non-malignant breast epithelial cell line MCF10A expressing the well-characterized CMV promoter driving RFP fused with a mammalianized version of the bacterial nitroreductase
nfs gene. The
nfs gene has been used in gene-directed enzyme prodrug therapy [
37] since treatment of mammalian cells expressing
nfs with CB1954 results in its chemical reduction to cytotoxic metabolites. Exposure of the derivative MCF10A with CB1954 was used as a strategy for the selection of cell lines with silenced
nfs genes.
The EPISSAY was verified by treatment with the known epigenetic drugs decitabine, zebularine or vorinostat; all of which resulted in increased red-fluorescence due to reactivation of the CMV promoter. There was a linear relationship between
nfs expression and the red-fluorescent signal confirming that levels of gene message and translated protein are directly related. The response was further confirmed by measuring expression levels of known independent endogenous genes
TXNIP[
5] and
ANKRD11[
34].
EPISSAY could be a time-saving assay for screening compounds with gene reactivating activity. Standard methodologies used to assess epigenetic compounds are based on quantitative real-time PCR and western blot analysis of genes known to be silenced in a particular cell line. For example, quantification of the re-expression of an endogenous gene p16 in human T24 bladder carcinoma cell line was previously used [
38]. Such approaches are time-consuming as they require cell collection for RNA and protein extractions prior to analysis. Other cell-based assay systems which use exogenous expression of genes (e.g.
Escherichia coli β-D-galactosidase gene with and green fluorescent reporter) have previously been investigated for their potential in screening epigenetic drugs by using fluorescent microscopy and plate readers. However, these other systems have limitations such as the non-quantitative data obtained and/or additional sample treatments required (e.g Paraformaldehyde fixing, the addition fluorogenic compounds) prior to screening [
18‐
20] (Additional file
4). EPISSAY requires limited sample preparation, may be formatted for multi-well plates, and rapid results can be generated from RFP reading using flow cytometry to obtain quantitative data.
Decitabine is a demethylating agent that is FDA approved as an anti-cancer agent [
13]. Since decitabine is degraded
in vivo with a half-life of only 25 minutes, daily treatments are required to maintain appropriate drug levels both
in vitro and
in vivo[
39]. To improve the stability and bioavailability of decitabine, the drug was encapsulated in PEGylated liposomes, as liposomes are known to protect drugs from degradation and allow controlled release of drug into the environment [
40]. This formulation achieved an encapsulation efficiency of ~50%. Only 3.3 mol% of PEG 2000 was used in this study as a higher PEG content is known to reduce adsorption of liposomes onto cells [
41].
Liposomes were extruded through filters with defined pore size (200 nm and 400 nm) to obtain unilamellar liposomes. Although extrusion does not affect the encapsulation efficiency [
42], it narrowed the size distribution of the liposomes from 1 μm to approximately 150 nm. The smaller size of the drug-loaded liposomes has been reported to passively targeting disease tissues due to their enhanced angiogenesis [
43].
We used the EPISSAY system to determine if liposomal encapsulation enhanced the gene reactivating activity of decitabine. Following 72 hours of treatment, decitabine encapsulated in unilamellar liposomes showed 50% more potency than pure decitabine, suggesting that decitabine was protected in the liposomes and slowly released into the media. These results were supported by a controlled release study comparing the drug release of decitabine from unilamellar and multilamellar liposomes. This showed that the release rate of decitabine from unilamellar liposomes was slower, suggesting unilamellar liposomal formulation may decrease the rate of degradation of decitabine by providing protection to the drug. In addition, the liposomal formulation and the presence of phospholipids in the cell media could also contribute to the enhancement of decitabine activity [
44,
45].
Collectively, the liposomal decitabine that was synthesised here was validated as a more potent epigenetic drug. However, we have only confirmed this
in vitro. An
in vivo study of liposomal decitabine is recommended to assess its applicability for clinical use, and to confirm if the present limitations of decitabine use in the clinic could be overcome by this formulation. The use of liposomes/PEG to encapsulate drugs to improve their bio-availability and stability is now gaining momentum with a number of drugs eg doxorubicin [
17], rhenium radionuclides [
46] and dexamethasone phosphate [
47], liposome-encapsulated doxorubicin now having FDA approval.
Conclusions
In this pilot study, we have constructed and evaluated a novel bioassay for epigenetic compounds. The readout of the EPISSAY system is red-fluorescence, which may allow the adaptation of the assay system to a multi-well format allowing high throughput, rapid, and cheap bioassay in the future. EPISSAY was successful in providing evaluation of different liposomal formulations of decitabine. The EPISSAY can detect the gene reactivating effects of decitabine, zebularine or vorinostat. Linear correlation between the message of an endogenous gene
ANKRD11 and red-fluorescent reading has been shown in the EPISSAY cells treated with pure decitabine and unilamellar liposomes-formulated decitabine (Additional file
5).
Using SEQUENOM MassARRAY EpiTYPER, no major changes in methylation of the CMV promoter was detected in the EPISSAY cells before and after treatments with decitabine (Additional file
6 and
7). Although vorinostat is known as a HDAC inhibitor to activate gene expression, zebularine and decitabine are usually considered to function as demethylating agents or DNMT inhibitors [
48]. However, there are now multiple studies that show these agents can also function as HDAC inhibitors [
49‐
51]. This suggests that the
TMnfsB gene was most likely silenced by histone modification rather than direct methylation of the CMV promoter. There is a potential of adopting this assay as a high throughput, rapid and low cost epigenetic drug screening platform are unique aspects of the EPISSAY system. We conclude that our EPISSAY bioassay system provides a novel and rapid system to screen the efficiencies of epigenetic and newly formulated drugs for gene reactivation.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SPL carried out all the experimental work and drafted the manuscript. RK carried out the molecular biology studies, participated in the experimental design and contributed to drafting and editing of the manuscript. YA and WW participated in the study of nanotechnology. KH, PMN and DJW contributed to the molecular biology studies. RJS was involved in the design of the study, performed the statistical analysis and edited the manuscript. CP reviewed the study and participated in the nanotechnology work. DFC supervised the study, and contributed to its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.