FEBS Letters 589 (2015) 4053–4060
journal homepage: www.FEBSLetters.org
Global DNA hypomethylation coupled to cellular transformation and metastatic ability Soichiro Funaki a,b,1, Toshinobu Nakamura d,e,⇑,1, Tsunetoshi Nakatani a,1, Hiroki Umehara c, Hiroyuki Nakashima c, Meinoshin Okumura b, Keisuke Oboki f, Kenji Matsumoto f, Hirohisa Saito f, Toru Nakano a,e,c,⇑ a
Department of Pathology, Osaka University, Osaka 565-0871, Japan Department of General Thoracic Surgery, Graduate School of Medicine, Osaka University, Osaka 565-0871, Japan Graduate School of Frontier Biosciences, Osaka University, Osaka 565-0871, Japan d Nagahama Institute of Bio-Science and Technology, Shiga 526-0829, Japan e JST, CREST, Saitama 332-0012, Japan f Department of Allergy and Immunology, National Research Institute for Child Health and Development, Tokyo 157-8535, Japan b c
a r t i c l e
i n f o
Article history: Received 25 December 2014 Revised 15 November 2015 Accepted 16 November 2015 Available online 19 November 2015 Edited by Ned Mantei Keywords: DNA hypomethylation Transformation Metastasis Stella
a b s t r a c t Global DNA hypomethylation and DNA hypermethylation of promoter regions are frequently detected in human cancers. Although many studies have suggested a contribution to carcinogenesis, it is still unclear whether the aberrant DNA hypomethylation observed in tumors is a consequence or a cause of cancer. Here, we show that the enforced expression of Stella (also known as PGC7 and Dppa3) induced not only global DNA demethylation but also transformation of NIH3T3 cells. Furthermore, overexpression of Stella enhanced the metastatic ability of B16 melanoma cells, presumably through the induction of metastasis-related genes. These results provide new insights into the function of global DNA hypomethylation in carcinogenesis. Ó 2015 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction DNA methylation is important for epigenetic gene regulation and implicated in many biological processes, such as early embryonic development, cell differentiation, and zygotic and somatic cell reprogramming [1,2]. In addition, the pathogenesis of many types of disease, including cancer, has been suggested to be affected by DNA methylation status [2,3]. Circumstantial evidence concerning the DNA methylation status in malignant tumor cells is available; however, most is derived from clinical samples of established
Author contributions: T. Nakamura and T. Nakano conceived the project and wrote the manuscript. S. Funaki, T. Nakamura, T. Nakatani, H.U., H.N., and M.O. designed and performed the experiments and evaluated the results. K.O., K.M., and H.S. performed microarray analysis. ⇑ Corresponding authors at: JST, CREST, Saitama 332-0012, Japan. Fax: +81 749 64 8140 (T. Nakamura), +81 6 6879 3729 (T. Nakano). E-mail addresses: [email protected] (T. Nakamura), tnakano@ patho.med.osaka-u.ac.jp (T. Nakano). 1 These authors contributed equally to this work.
tumors. Little direct evidence of the precise pathogenic role of DNA methylation has been reported [4]. DNA methyltransferase 1 (DNMT1) mediates inheritance of DNA methylation patterns by daughter cells from parental cells [5]. It has been reported that Np95 is required for the recruitment of DNMT1 to hemi-methylated DNA generated at the replication fork during the S-phase [6], and that Np95 binds to hemimethylated DNA through its SET and RING finger-associated (SRA) domain [7]. We have investigated the function of Stella, a maternal factor essential for early embryogenesis. Stella preserves the DNA methylation status in zygotes by binding to dimethylated lysine 9 of histone H3 (H3K9me2) [8–10]. Recently, we reported that overexpression of Stella induced global DNA hypomethylation through binding to Np95 and thus the subsequent recruitment of DNMT1 to hemi-methylated DNA [11]. In the present study, we found that enforced expression of Stella in NIH3T3 cells not only induced global DNA demethylation but also induced transformation of the cells. In addition, we found that the metastatic activity of B16 melanoma cells was increased by the expression of Stella. Although this unique experimental sys-
http://dx.doi.org/10.1016/j.febslet.2015.11.020 0014-5793/Ó 2015 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
4054
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
tem is an artificial one, Stella-induced global DNA hypomethylation is a useful model that could facilitate investigation of the pathogenic role of DNA hypomethylation in cancer. 2. Materials and methods 2.1. Plasmids Stella and Ha-Ras cDNAs were inserted into the pMY-IRES-EGFP retroviral vector or pWPI lentiviral vector [12]. For generation of shRNA constructs, oligonucleotides were cloned into the pLVTH vector. Oligonucleotide sequences are shown in Supplementary Table 1. 2.2. Cell culture and viral infections NIH3T3 cells, B16F10 cells, 293T cells, and Plat-E packaging cells [13] were maintained in Dulbecco’s modified Eagle’s Minimum Essential Medium supplemented with 100 U/mL penicillin, streptomycin sulfate, and 10% fetal bovine serum (FBS). Retroviral and lentiviral infections were carried out as previously reported [11].
2.8. Microarray array analysis Microarray analysis was performed in NIH3T3 cells including two independent Stella-expressing NIH3T3 clones and a nonclonal Stella-expressing NIH3T3 cell population, and in B16-F10 cells including three independent Stella-expressing B16-F10 clones. Total RNA was extracted from cells using an RNeasy Total RNA Mini Kit (Qiagen). RNA was labeled and hybridized to GeneChip Mouse Genome 430 2.0 arrays (Affymetrix) according to the manufacturer’s instructions. Signal values and detection calls for all samples were determined using MAS5.0 (Affymetrix). Further analyses were performed with the Subio Platform version 1.15 plug-in software (Amami). To normalize the variations in staining intensity among chips, the signal values for all probes on a given chip were divided by the 75th percentile value for expression of all genes on the chip. To eliminate changes within the range of background noise, we used probes the raw signal values of which were >50 in at least any one sample. The microarray data have been deposited in the Gene Expression Omnibus database (GEO; http://www.ncbi.nlm.nih.gov/geo/) and given the series accession number GSE68837. 2.9. Genome-wide profiling of promoter methylation by the Microarray-based Integrated Analysis of Methylation (MIAMI) method
2.3. Soft agar assay NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells (45 days after infection), Stella-expressing NIH3T3 cells (45 days after infection), and EGFP-expressing NIH3T3 cells (45 days after infection) (3 103) were resuspended into single-cell suspensions in 3 mL of medium. The cell suspensions were mixed with 1 mL of 1.5% pre-warmed soft agar (BD Biosciences) and plated into 30-mm plates in triplicate. The colonies formed on the soft agar matrix were counted under a light microscope after 2 weeks. 2.4. Global DNA methylation analysis Global DNA methylation status was measured as described previously [11]. 2.5. In vivo growth assay NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells (45 days after infection), and Stella-expressing NIH3T3 cells (45 days after infection), (5 105) were injected subcutaneously into nude mice. After 2 weeks, the tumors were harvested and each tumor was weighed. 2.6. Wound healing assay B16-F10, EGFP-expressing B16-F10, and Stella-expressing B16F10 cells were grown to confluence in six-well culture plates. Cell layers were scraped with a sterile pipette tip and incubated at 37 °C with 5% CO2. Migration from the edge of the injured monolayer was quantified by measurement of the distance between the wound edges and the recovered edges. 2.7. Metastasis assay B16-F10, EGFP-expressing B16-F10, and Stella-expressing B16F10 cells (2 105) were suspended in 200 lL of PBS and injected in the lateral tail vein of C57BL/6 mice. Lungs were harvested 14 days post-injection and fixed in formalin. Metastases were counted and subjected to statistical analysis (one-way ANOVA).
Genome-wide profiling and promoter methylation status of Stella-expressing NIH3T3 cells were evaluated using Microarraybased Integrated Analysis of Methylation (MIAMI) method [14]. Briefly, genomic DNA was digested with methylation-sensitive HpaII or the methylation-insensitive isoschizomer MspI. The digested genomes were ligated with adaptor, and amplified by PCR with primers designed against the adapter sequences. The samples were then further digested with HpaII and MspI, and amplified again with the same primers. The amplified products were then labeled with Cy3 or Cy5 and co-hybridized to a microarray. After hybridization, the microarray was scanned, and the obtained fluorescence intensities were quantified and normalized. The same pooled samples were treated first with MspI instead of HpaII and analyzed on a duplicate array to correct for falsepositives caused by single nucleotide polymorphisms or incomplete digestion. Unchanged, DNA hyper-, and hypo-methylation of each genomic regions are indicated by ‘‘0”, ‘‘+1”, and ‘‘1” in column AB, respectively (Supplementary Table 2). 2.10. RT-PCR Total RNAs prepared from cells were treated with DNase I, and subjected to RT-PCR using the ThermoScript RT-PCR system (Invitrogen) and random hexamers as primers for cDNA synthesis. The primers used in this study are shown in Supplementary Table 1. PCR were carried out following condition: 5 min at 95 °C followed by 24 and 26 cycles (for IAP and Line 1), 35 cycles of (for Arhgap20, Smagp, and Notch3), or 30 cycles (for Gapdh) of PCR consisting of 30 s at 95 °C, 30 s at 60 °C, and 1 min at 72 °C. 2.11. Bisulfite sequence analysis B16 melanoma cells and Stella-expressing clones were treated with bisulfite using an EZ DNA Methylation-Direct Kit (Zymo Research). Sequences of the PCR primers are listed in Supplementary Table 1. PCR amplification of Arhgap20, Smagp, and Notch3 promoter regions was carried out using EpiTaq HS (Takara) under the following conditions: 2 min at 95 °C followed by 35 cycles of PCR consisting of 20 s at 98 °C, 30 s at 56 °C, and 30 s at 72 °C. The PCR products were purified using a QIAquick Gel Extraction Kit (Qiagen), cloned into the pGEM-T Easy Vector (Promega), and then
4055
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
sequenced using an (Applied Biosystems).
ABI
PRISM
3100
Genetic
Analyzer
3. Results and discussion 3.1. Transformation of NIH3T3 cells by Stella
2.12. Ethics statement
Control
Ha-Ras
*
Stella Day12
Stella Day 40
Stella
Number of colonies
200
EGFP
*
100
50 N.D.
H
C
a-
on
R
tro
as
l
N.D.
E
D 109
1010 Control Ha-Ras Stella
108
Control Ha-Ras Stella
109 Number of cells
Number of cells
*
150
0
107 106 105
108 107 106 105
104
104 0
3
6
9
12
15
0
3
6
Days
F
*
*
FP
Ha-Ras
n.s.
la
Control
C
EG
B
St
A
We recently reported that Stella induced DNA hypomethylation in NIH3T3 cells. Quite unexpectedly, transformation of the NIH3T3 cells was observed during culture (Fig. 1A). Considering that DNA hypomethylation induces transformation, we carried out additional experiments. Treatment of NIH3T3 cells with high and low
el
Experiments were performed in accordance with the guidelines of the Osaka University Animal Care and Use Committee. All animal experiments were approved by Osaka University Animal Care and Use Committee.
Control
Stella
9
12
15
Days
Ha-Ras
G
500
*
*
n.s.
Tumor weight (mg)
400 300 200 100 N.D. as
la
R a-
St
el H
C
on
tro
l
0
Fig. 1. Transformation of NHI3T3 cells by Stella. (A) Morphology of NIH-3T3 cells, Stella-expressing NIH3T3 cells (days 12 and 40), and Ha-Ras-expressing NIH3T3 cells. (B, C) Anchorage-independent growth of NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells, Stella-expressing NIH3T3 cells, and EGFP-expressing NIH3T3 cells in soft agar over 14 days. Mean and standard deviation (n = 5) are shown. The number of colonies was significantly increased in Ha-Ras- and Stella-expressing NIH3T3 cells compared to the controls (*P < 0.00017, one-way ANOVA with Dunnet’s post hoc test). (D, E) Growth curves before (day 6 after retroviral infection) (D) and after (day 40 after retroviral infection) (E) transformation by Stella. Means and standard deviations (n = 3) are shown. Blue, red, and green lines indicate control NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells, and Stella-expressing NIH3T3 cells, respectively. (F) In vivo transplantation assay. Control NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells, and Stella-expressing NIH3T3 cells were subcutaneously injected into the backs of immunodeficient mice (KSN/SLC mice, n = 6). Data are from representative mice at 14 days after injection. (G) Weight of tumors derived from Ha-Ras-expressing NIH3T3 cells and Stella-expressing NIH3T3 cells. No tumors were formed from parental NIH3T3 cells. Tumor weight was significantly increased in Stella-expressing NIH3T3 cells compared to controls (*P < 0.00005, one-way ANOVA with Dunnet’s post hoc test).
4056
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
Stella-transformed cells were transplantable and their in vivo growth was essentially identical to that of Ha-Ras-transformed cells.
concentrations of 5-azacytidine (5-azaC), an inhibitor of DNMT, induced cell death and blocked transformation, respectively (Supplementary Fig. 1). Meanwhile, knockdown of Dnmt1 by shRNA significantly reduced the growth rate but did not induce transformation of NIH3T3 cells (Supplementary Fig. 1). Thus, we could not induce the transformation of NIH3T3 cells by conventional methods for reducing Dnmt1 activity. Unfortunately, it is rather difficult to determine whether these results are explained by the reduced level of DNA methylation or differences in the sites of DNA demethylation. Although transformation by oncogenic Ha-Ras commenced within 7 days, that induced by Stella was observed 3–4 weeks after introduction of the gene. There were significant differences not only in the duration but also the frequency of transformation. The vast majority of Ha-Ras-expressing cells were transformed, but less than 1% of Stella-expressing cells exhibited transformation. However, the morphology of NIH3T3 cells transformed by Stella was not significantly different from that of cells transformed by Ha-Ras. In addition, like Ha-Ras-transformed cells, Stellaexpressing transformed cells showed anchorage-independent colony formation in agar (Fig. 1B and C). Although the growth of Stella-expressing cells was significantly slower than that of the control cells before transformation, it became significantly higher than that of the control cells after transformation (Fig. 1D and E). Compared to Ha-Ras, Stella took significantly longer to induce transformation, and the frequency of transformation was lower. These observations strongly suggest that transformation is not a direct oncogenic effect of Stella. Thus, we considered it likely that hypomethylation induced by Stella is the major cause of the transformation. Next, we subcutaneously injected control, Stella-transformed, and Ha-Rastransformed cells into immunodeficient mice to examine the in vivo transplantability of the cells (Fig. 1F and G). The
3.2. Genome-wide DNA methylation analysis and gene expression profiling of Stella-transformed cells Genome-wide DNA methylation conditions in NIH3T3 cells and Stella-transformed NIH3T3 cells were analyzed by the Microarraybased Integrated Analysis of Methylation (MIAMI) method [14]. There were no hypermethylated regions in the genome of Stellatransformed cells (Supplementary Table 2, Supplementary Figs. 2 and 3). By contrast, DNA hypomethylation was detected throughout the genome of Stella-transformed NIH3T3 cells (Supplementary Table 2, Supplementary Figs. 2 and 3) and the hypomethylated regions were distributed almost randomly on all chromosomes (Fig. 2A). The percentages of the hypomethylated regions in individual promoters ranged from 9.6% to 18.6% (Supplementary Fig. 2). To examine alterations in gene expression in Stella-transformed NIH3T3 cells, we carried out microarray gene expression analysis. Two independent clones and non-clonal expansion cells were subjected to the analysis and no typical expression pattern was observed (Fig. 2B). The Stella-expressing clones showed >2-fold upregulation and downregulation of 432 and 482 genes, respectively (Fig. 2C and Supplementary Table 3). No tumor suppressor genes were down-regulated in all three samples. By contrast, the potent oncogene c-Jun [15] was among the significantly upregulated genes. However, knockdown of c-Jun expression in Stella-transformed NIH3T3 cells by shRNA was not associated with significant morphological or growth changes, indicating that the increase in c-Jun expression was not the major cause of
Hypomethylationed / Total (%)
A
B 20 2 up
15
10
5
0 1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 X 0
Chromosome
C Non-clonal
Non-clonal
210
-2 down
DOWN-regulated
tro
Clone #2
l n lo al ne # C lo 1 ne #2
337
C
UP-regulated
208
Clone #1
lo
122
Clone #2
on
72
Clone #1
-c
133
82 482
C
53
61
on
432
243 48
N
26
Fig. 2. Genome-wide analysis of DNA methylation status and gene expression profiling of Stella-transformed NIH3T3 cells. (A) Genome-wide DNA methylation status determined by the Microarray-based Integrated Analysis of Methylation (MIAMI) method. The percentages of hypomethylated regions in each chromosome are shown. (B) Heat map representing color-coded expression levels of differentially expressed genes (two-way ANOVA, up- or downregulated >twofold) in non-clonal-, clone #1-, clone #2-, Stella-expressing NIH3T3 cells. (C) Venn diagram indicating overlap of up- and downregulated genes among non-clonal-, clone #1-, clone #2-, Stella-expressing NIH3T3 cells.
4057
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
genes analyzed were re-methylated after the elimination of Stella. It has been reported that Np95 binds to chromatin via histone H3K9me3 and that DNA methylation levels were restored in the Np95 KO ES cells by the forced expression of Np95 [16,17]. Considering that Stella did not affect the level of H3K9me3 [9], it is likely that the association between Np95 and H3K9me3 facilitates the remethylation of the original locus after the elimination of Stella. Three possible explanations for the role of DNA hypomethylation in cancer have been suggested: chromosomal aberrations, reactivation of transposable elements, and increases in expression of potential oncogenic genes such as oncogenes and some imprinted genes. Nonetheless, its precise function remains unclear. As shown in Fig. 3, elimination of Stella resulted in revertant cells that were similar to the original NIH3T3 cells, showing that permanent abnormalities—such as loss of heterozygosity (LOS)—were not the major cause of the transformation. Meanwhile, Stella expression did not affect the expression of retrotransposons in NHI3T3 cells (Supplementary Fig. 6). Taken together, these observations indicate that transformation by Stella cannot be explained by the three possibilities mentioned above. Although there have been many reports on global DNA hypomethylation in human malignant tumors, they were descriptive and did not determine the pathogenic role of hypomethylation [4]. It is difficult to determine whether DNA hypomethylation is the cause or result of malignant tumors. To address this question, studies were performed using mutant mice with reduced Dnmt1
transformation (Supplementary Fig. 4). We further confirmed that cancer associated genes were not altered in Stella expressing cells by gene ontology (GO) analysis (Supplementary Tables 4 and 5). Since cell adhesion-related genes were up- and down-regulated in the Stella-transformed NIH3T3 cells, it is possible that Stella facilitates anchorage independent cell growth. 3.3. Reversion of the transformed phenotype by elimination of Stella expression Oncogenic mutations may have been acquired during the relatively long period of culture before Stella-induced transformation. If this were the case, cellular transformation would have been maintained even after Stella expression was abolished. To exclude this possibility, we constructed a floxed-Stella lentivirus, infected the cells with the vector, and eliminated the gene using Cre recombinase after transformation. The morphology of the Stellatransformed cells reverted to that of NIH3T3 cells, growth of the cells was normalized, and the cells lost their in vivo tumorforming ability (Fig. 3A–C). It was further confirmed that the global DNA methylation level of these revertant cells was identical to that of control NIH3T3 cells (Fig. 3D). To examine whether restoration of DNA methylation occurs in the original locus after the elimination of Stella, we analyzed the DNA methylation status of the promoter regions of Oct3/4 and Nanog, and differential methylated regions of Peg1 and H19. As shown in Supplementary Fig. 5, all
A
B
109 Control
Stella elimination
Number of cells
Stella
108
Stella Stella elimination
107 106 105 104 0
5
10
15
20
Days
C
D
* 80
Stella elimination
CCmGG (%)
Stella
n.s.
100
*
60 40 20
tio n
el la
in a
St St
el la
el
im
C
on
tro l
0
Fig. 3. Reversion of phenotype by elimination of Stella. (A) Morphologies of Stella-expressing NIH3T3 (Stella) cells and Stella-eliminated NIH3T3 cells (Stella elimination). (B) Growth curves of control (NIH3T3 cells), Stella-expressing NIH3T3 cells, and Stella-eliminated NIH3T3 cells. Mean and standard deviation (n = 3) are shown. Blue, green, and red lines indicate control, Stella-expressing NIH3T3 cells, and Stella-eliminated NIH3T3 cells, respectively. (C) In vivo transplantation assay. 5 106 Stella-expressing NIH3T3 cells and Stella-eliminated NIH3T3 cells were injected subcutaneously into the backs of immunodeficient mice (KSN/SLC mice, n = 4). Tumors formed in all mice injected with Stella-expressing NIH3T3 cells and none with Stella-eliminated NIH3T3. Data are from representative mice at 14 days after injection. Arrow indicates tumor. (D) Global DNA methylation status of Stella-expressing NIH3T3 cells, and Stella-eliminated NIH3T3 cells. DNA methylation level was decreased significantly in Stella-expressing NIH3T3 cells compared to controls (*P < 0.015, one-way ANOVA with Dunnet’s post hoc test).
4058
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
was examined under conditions of DNA hypomethylation [19]. DNA hypomethylation increased the incidence of microadenomas with LOH at the APC locus, but strongly reduced the incidence of macroscopic intestinal tumors. Meanwhile, DNA hypomethylation caused development of multifocal liver tumors with LOH at the APC locus in these mice. Thus, simple reduction of Dnmt1 activity did not appear to induce malignant transformation. Another model involves disruption of the interactions between Dnmt1, PCNA, and Np95 by introduction of a fusion protein containing parts of Dnmt1 and Np95 [20]. Astrocyte and glial
activity. Mice with the hypomorphic allele of Dnmt1, with about 10% of the normal enzymatic activity, exhibited global DNA hypomethylation. T-cell lymphomas developed in the mice after a prolonged period, and a chromosome 15 abnormality was detected [18]. These data indicate that DNA hypomethylation has the potential to induce malignant tumor formation. However, the effects were not direct, being caused by DNA hypomethylationdependent chromosomal instability. Tumorigenesis in ApcMin/+ mice, which are predisposed to intestinal tumors by a mutation in the canonical Wnt pathway,
B *
C Number of nodules
60 EGFP
20 Stella
n.s.
60 40 20 0
l tro
la
on
el
0
C
St
EG
on
tro
l
FP
0
C
*
80
12hr
la
40
D
*
Control
80 CCmGG (%)
100
el
*
St
n.s.
FP
100
EG
A
F G #1
Control
24
2 up 29
47 39
EGFP
33
24
62
#2 Stella
#3
DOWN-regulated #1
100
0
* 80
4 39
*
323
n.s.
8
60
61
59
13
#3
#2 UP-regulated
40 20
#3
#2
el la St
EG
tro on C
FP
C
on
0
#1
tro
l
-2 down
l
Recovered area / wounded area (%)
E
Fig. 4. Stella-enhanced metastatic ability of melanoma cells. (A) Global DNA methylation status of B16-F10 cells, EGFP-expressing B16-F10 cells, and Stella-expressing B16F10 cells. Means and standard deviation (n = 3) are shown. DNA methylation level was decreased significantly in Stella-expressing B16-F10 cells compared to controls (*P < 0.0049, one-way ANOVA with Dunnet’s post hoc test). (B) Representative photographs of lung metastatic nodules. (C) The numbers of lung metastatic nodules at 14 days after transplantation of B16-F10 cells, EGFP-expressing B16-F10 cells, and Stella-expressing B16-F10 cells. Number of nodules was significantly increased in Stella-expressing B16-F10 cells compared to controls (*P < 0.014, one-way ANOVA with Dunnet’s post hoc test). (D, E) Wound healing assay of B16-F10 cells, EGFP-expressing B16-F10 cells, and Stella-expressing B16-F10 cells. Representative photographs at 12 h after scratching (D), and the percentages of the recovered area/wound area, respectively (E). Wound healing rate was significantly higher in Stella-expressing B16-F10 cells compared to controls (*P < 0.00039, one-way ANOVA with Dunnet’s post hoc test). (F) Heat map representing color-coded expression levels of differentially expressed genes (two-way ANOVA, up- or downregulated > twofold) in control, clone #1-, clone #2-, and clone#3Stella-expressing B16-F10 cells. (G) Venn diagram indicating overlap of up- and downregulated genes among clone #1-, clone #2-, and clone#3- Stella-expressing B16-F10 cells.
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
precursor cell lines expressing the fusion protein showed global DNA hypomethylation and exhibited a higher proliferation index and lower apoptosis than control cells. Chromosomal instability and the overexpression of oncogenes, such as PDGF and Ha-Ras, were induced by the fusion protein. This study indicated that the major causes of the malignant transformation were aberrations of specific oncogenes and tumor suppressor genes. The molecular mechanisms of these models seem to be different from that in Stella-transformed cells. 3.4. Stella enhanced metastatic ability of B16 melanoma cells We next assessed the effects of Stella on the metastatic ability of B16 melanoma cells. The global DNA methylation level of B16 melanoma cells was reduced by Stella expression as in the case of NIH3T3 cells (Fig. 4A). We intravenously injected wild-type B16 melanoma cells, EGFP-expressing B16 melanoma cells, and Stella-expressing B16 melanoma cells into C57BL/6 mice, and counted lung metastases 14 days later (Fig. 4B and C). Stellaexpressing cells produced significantly more lung metastases than did parental cells and control EGFP-expressing cells. Since the metastatic ability of malignant cells is generally well correlated with their migratory ability, we performed an in vitro wound healing assay to evaluate the effects of Stella on migration potential. As shown in Fig. 4D and E, Stella-expressing B16 melanoma cells showed significantly better wound healing properties compared to control B16 melanoma cells and EGFP-expressing B16 melanoma cells. These data clearly indicate that Stella expression enhanced the mobility of B16 melanoma cells. The relationship between DNA hypomethylation and metastasis is unclear. It was reported that DNA hypomethylation occurred at a very late stage in prostate cancer progression and became evident at metastasis [21]. 5-azaC was reported to increase the metastatic capacity of the human melanoma cell line DX3 after intravenous injection into immunodeficient nude mice [22]. In this study, 24h exposure of cells to 5-azaC resulted in a 40-fold increase in the number of lung tumor nodules. Moreover, the metastatic capacity increased through successive cycles of growth in vitro. Although Stella caused both transformation and metastasis, the underlying molecular mechanisms may differ, i.e., global DNA hypomethylation or the activation of some metastasis-related genes by DNA hypomethylation. 3.5. Genome-wide gene expression profiling of Stella-expressed melanoma cells Three independent clones were subjected to microarray gene expression analysis (Fig. 4F). Of these, 323 and 39 genes were > 2-fold up- and down-regulated, respectively, in the Stellaexpressing clones (Fig. 4G and Supplementary Table 6). Gene ontology analysis revealed that the GO term ‘‘metastasis” was not identified (Supplementary Tables 7 and 8). However, Arhgap20 [23], Smagp [24], and Notch3 [25], all of which are metastasisrelated genes, were upregulated and the increased expression levels of these genes were confirmed by RT-PCR (Supplementary Fig. 7). Bisulfite sequencing analysis revealed that Stella expression induced demethylation of Arhgap20 and Notch3 promoter regions (Supplementary Fig. 8). These data suggest that expression of Stella enhanced the metastatic ability of B16 melanoma cells through the induction of metastasis-related genes, such as Arhgap20 and Notch3, an effect mediated by alteration of their DNA methylation status. Alteration of DNA hypomethylation is associated with several aspects of gene regulatory mechanisms. One typical feature is activation of gene regulation through hypomethylation of promoter regions of many kinds of genes. Germline-specific tumor antigens
4059
such as MAGE, which are silenced by DNA methylation in normal tissues, have been reported to be upregulated by DNA hypomethylation of their promoter regions [26]. Although the reasons are unclear, some other genes, such as the synuclein gamma and clausin-4 genes, are hypomethylated in cancers [27,28]. Deregulation of retrotransposon genes by DNA hypomethylation has been suggested to play a role in genomic instability [29]. However, in Stella-transformed NIH3T3 cells, neither aberrant expression of retrotransposons nor specific activation of cancer-associated genes was detected by microarray analysis. Cancer metastasis is the spread of tumor cells from the original neoplasm to other organs through angiogenesis, invasion, colonization, and proliferation [30]. 5-azaC treatment was reported to upregulate the prometastatic genes urokinase plasminogen activator, matrix metalloproteinase 2, metastasis-associated gene 1, and CXC chemokine receptor 4, which stimulate cancer cell invasiveness and metastasis, in the non-invasive breast cancer cell line MCF-7 [31]. However, in Stella-expressing B16 melanoma cell clones, their expression was not changed (Supplementary Table 6). Instead, expression of the metastasis-related genes Arhgap20, Smagp, and Notch3 was upregulated in Stella-expressing B16 melanoma cell clones (Supplementary Table 6). Therefore, we postulate that Stella enhanced the metastatic ability of B16 melanoma cells by inducing metastasis-related genes through alteration of their DNA methylation status. It is uncertain whether increased expression of just these three genes is sufficient for transformation. However, we would like to emphasize that these three genes were up-regulated in three independent Stella-expressing clones. 4. Conclusions and perspective In our previous study, we unexpectedly found that a maternal factor, Stella, which preserves the DNA methylation status in early embryos, induced global DNA hypomethylation in somatic cells. In the present study, we found that global DNA hypomethylation was itself critical for the transformation of NIH3T3 cells. By contrast, it is likely that the metastatic ability of B16 melanoma cells was enhanced by activation of metastasis-related genes. Although we cannot exclude the possibility that some unknown function(s) of Stella affect the transformation of NIH3T3 cells, the Stellamediated artificial global DNA hypomethylation system will give new insights into the roles of aberrant DNA methylation in various diseases. Acknowledgements We thank Dr. D. Trono for providing the lentiviral vectors. We also thank Ms. N. Asada for assistance, and Ms. A. Mizokami and Ms. M. Imaizumi for secretarial assistance. This work was supported in part by grants from the Uehara Memorial Foundation, the Ministry of Education, Science, Sports, Culture, and Technology of Japan, and by JST, CREST. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.febslet.2015.11. 020. References [1] Feng, S., Jacobsen, S.E. and Reik, W. (2010) Epigenetic reprogramming in plant and animal development. Science 330, 622–627. [2] Smith, Z.D. and Meissner, A. (2013) DNA methylation: roles in mammalian development. Nat. Rev. Genet. 14, 204–220. [3] Robertson, K.D. (2005) DNA methylation and human disease. Nat. Rev. Genet. 6, 597–610.
4060
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
[4] Ehrlich, M. (2009) DNA hypomethylation in cancer cells. Epigenomics 1, 239– 259. [5] Goll, M.G. and Bestor, T.H. (2005) Eukaryotic cytosine methyltransferases. Annu. Rev. Biochem. 74, 481–514. [6] Sharif, J. et al. (2007) The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature 450, 908–912. [7] Bostick, M., Kim, J.K., Esteve, P.O., Clark, A., Pradhan, S. and Jacobsen, S.E. (2007) UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science 317, 1760–1764. [8] Nakamura, T. et al. (2007) PGC7/Stella protects against DNA demethylation in early embryogenesis. Nat. Cell Biol. 9, 64–71. [9] Nakamura, T. et al. (2012) PGC7 binds histone H3K9me2 to protect against conversion of 5mC to 5hmC in early embryos. Nature 486, 415–419. [10] Bian, C. and Yu, X. (2014) PGC7 suppresses TET3 for protecting DNA methylation. Nucleic Acids Res. 42, 2893–2905. [11] Funaki, S., Nakamura, T., Nakatani, T., Umehara, H., Nakashima, H. and Nakano, T. (2014) Inhibition of maintenance DNA methylation by Stella. Biochem. Biophys. Res. Commun. 453, 455–460. [12] Wiznerowicz, M. and Trono, D. (2003) Conditional suppression of cellular genes: lentivirus vector-mediated drug-inducible RNA interference. J. Virol. 77, 8957–8961. [13] Morita, S., Kojima, T. and Kitamura, T. (2000) Plat-E: an efficient and stable system for transient packaging of retroviruses. Gene Ther. 7, 1063–1066. [14] Hatada, I. et al. (2006) Genome-wide profiling of promoter methylation in human. Oncogene 25, 3059–3064. [15] Bohmann, D., Bos, T.J., Admon, A., Nishimura, T., Vogt, P.K. and Tjian, R. (1987) Human proto-oncogene c-jun encodes a DNA binding protein with structural and functional properties of transcription factor AP-1. Science 238, 1386– 1392. [16] Rothbart, S.B. et al. (2012) Association of UHRF1 with methylated H3K9 directs the maintenance of DNA methylation. Nat. Struct. Mol. Biol. 19, 1155–1160. [17] Liu, X., Gao, Q., Li, P., Zhao, Q., Zhang, J., Li, J., Koseki, H. and Wong, J. (2013) UHRF1 targets DNMT1 for DNA methylation through cooperative binding of hemi-methylated DNA and methylated H3K9. Nat. Commun. 4, 1563. [18] Gaudet, F., Hodgson, J.G., Eden, A., Jackson-Grusby, L., Dausman, J., Gray, J.W., Leonhardt, H. and Jaenisch, R. (2003) Induction of tumors in mice by genomic hypomethylation. Science 300, 489–492. [19] Eads, C.A., Nickel, A.E. and Laird, P.W. (2002) Complete genetic suppression of polyp formation and reduction of CpG-island hypermethylation in Apc(Min/+) Dnmt1-hypomorphic mice. Cancer Res. 62, 1296–1299.
[20] Hervouet, E. et al. (2010) Disruption of Dnmt1/PCNA/UHRF1 interactions promotes tumorigenesis from human and mice glial cells. PLoS One 5, e11333. [21] Yegnasubramanian, S. et al. (2008) DNA hypomethylation arises later in prostate cancer progression than CpG island hypermethylation and contributes to metastatic tumor heterogeneity. Cancer Res. 68, 8954–8967. [22] Ormerod, E.J., Everett, C.A. and Hart, I.R. (1986) Enhanced experimental metastatic capacity of a human tumor line following treatment with 5azacytidine. Cancer Res. 46, 884–890. [23] Herold, T. et al. (2011) Expression analysis of genes located in the minimally deleted regions of 13q14 and 11q22-23 in chronic lymphocytic leukemiaunexpected expression pattern of the RHO GTPase activator ARHGAP20. Genes Chromosomes Cancer 50, 546–558. [24] Tarbe, N.G., Rio, M.C., Hummel, S., Weidle, U.H. and Zoller, M. (2005) Overexpression of the small transmembrane and glycosylated protein SMAGP supports metastasis formation of a rat pancreatic adenocarcinoma line. Int. J. Cancer 117, 913–922. [25] Zhang, Z., Wang, H., Ikeda, S., Fahey, F., Bielenberg, D., Smits, P. and Hauschka, P.V. (2010) Notch3 in human breast cancer cell lines regulates osteoblastcancer cell interactions and osteolytic bone metastasis. Am. J. Pathol. 177, 1459–1469. [26] De Smet, C., Loriot, A. and Boon, T. (2004) Promoter-dependent mechanism leading to selective hypomethylation within the 50 region of gene MAGE-A1 in tumor cells. Mol. Cell. Biol. 24, 4781–4790. [27] Czekierdowski, A., Czekierdowska, S., Wielgos, M., Smolen, A., Kaminski, P. and Kotarski, J. (2006) The role of CpG islands hypomethylation and abnormal expression of neuronal protein synuclein-gamma (SNCG) in ovarian cancer. Neuro Endocrinol. Lett. 27, 381–386. [28] Gupta, A., Godwin, A.K., Vanderveer, L., Lu, A. and Liu, J. (2003) Hypomethylation of the synuclein gamma gene CpG island promotes its aberrant expression in breast carcinoma and ovarian carcinoma. Cancer Res. 63, 664–673. [29] Howard, G., Eiges, R., Gaudet, F., Jaenisch, R. and Eden, A. (2008) Activation and transposition of endogenous retroviral elements in hypomethylation induced tumors in mice. Oncogene 27, 404–408. [30] Cheishvili, D., Boureau, L. and Szyf, M. (2014) DNA demethylation and invasive cancer: implications for therapeutics. Br. J. Pharmacol.. [31] Chik, F. and Szyf, M. (2011) Effects of specific DNMT gene depletion on cancer cell transformation and breast cancer cell invasion; toward selective DNMT inhibitors. Carcinogenesis 32, 224–232.
journal homepage: www.FEBSLetters.org
Global DNA hypomethylation coupled to cellular transformation and metastatic ability Soichiro Funaki a,b,1, Toshinobu Nakamura d,e,⇑,1, Tsunetoshi Nakatani a,1, Hiroki Umehara c, Hiroyuki Nakashima c, Meinoshin Okumura b, Keisuke Oboki f, Kenji Matsumoto f, Hirohisa Saito f, Toru Nakano a,e,c,⇑ a
Department of Pathology, Osaka University, Osaka 565-0871, Japan Department of General Thoracic Surgery, Graduate School of Medicine, Osaka University, Osaka 565-0871, Japan Graduate School of Frontier Biosciences, Osaka University, Osaka 565-0871, Japan d Nagahama Institute of Bio-Science and Technology, Shiga 526-0829, Japan e JST, CREST, Saitama 332-0012, Japan f Department of Allergy and Immunology, National Research Institute for Child Health and Development, Tokyo 157-8535, Japan b c
a r t i c l e
i n f o
Article history: Received 25 December 2014 Revised 15 November 2015 Accepted 16 November 2015 Available online 19 November 2015 Edited by Ned Mantei Keywords: DNA hypomethylation Transformation Metastasis Stella
a b s t r a c t Global DNA hypomethylation and DNA hypermethylation of promoter regions are frequently detected in human cancers. Although many studies have suggested a contribution to carcinogenesis, it is still unclear whether the aberrant DNA hypomethylation observed in tumors is a consequence or a cause of cancer. Here, we show that the enforced expression of Stella (also known as PGC7 and Dppa3) induced not only global DNA demethylation but also transformation of NIH3T3 cells. Furthermore, overexpression of Stella enhanced the metastatic ability of B16 melanoma cells, presumably through the induction of metastasis-related genes. These results provide new insights into the function of global DNA hypomethylation in carcinogenesis. Ó 2015 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction DNA methylation is important for epigenetic gene regulation and implicated in many biological processes, such as early embryonic development, cell differentiation, and zygotic and somatic cell reprogramming [1,2]. In addition, the pathogenesis of many types of disease, including cancer, has been suggested to be affected by DNA methylation status [2,3]. Circumstantial evidence concerning the DNA methylation status in malignant tumor cells is available; however, most is derived from clinical samples of established
Author contributions: T. Nakamura and T. Nakano conceived the project and wrote the manuscript. S. Funaki, T. Nakamura, T. Nakatani, H.U., H.N., and M.O. designed and performed the experiments and evaluated the results. K.O., K.M., and H.S. performed microarray analysis. ⇑ Corresponding authors at: JST, CREST, Saitama 332-0012, Japan. Fax: +81 749 64 8140 (T. Nakamura), +81 6 6879 3729 (T. Nakano). E-mail addresses: [email protected] (T. Nakamura), tnakano@ patho.med.osaka-u.ac.jp (T. Nakano). 1 These authors contributed equally to this work.
tumors. Little direct evidence of the precise pathogenic role of DNA methylation has been reported [4]. DNA methyltransferase 1 (DNMT1) mediates inheritance of DNA methylation patterns by daughter cells from parental cells [5]. It has been reported that Np95 is required for the recruitment of DNMT1 to hemi-methylated DNA generated at the replication fork during the S-phase [6], and that Np95 binds to hemimethylated DNA through its SET and RING finger-associated (SRA) domain [7]. We have investigated the function of Stella, a maternal factor essential for early embryogenesis. Stella preserves the DNA methylation status in zygotes by binding to dimethylated lysine 9 of histone H3 (H3K9me2) [8–10]. Recently, we reported that overexpression of Stella induced global DNA hypomethylation through binding to Np95 and thus the subsequent recruitment of DNMT1 to hemi-methylated DNA [11]. In the present study, we found that enforced expression of Stella in NIH3T3 cells not only induced global DNA demethylation but also induced transformation of the cells. In addition, we found that the metastatic activity of B16 melanoma cells was increased by the expression of Stella. Although this unique experimental sys-
http://dx.doi.org/10.1016/j.febslet.2015.11.020 0014-5793/Ó 2015 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
4054
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
tem is an artificial one, Stella-induced global DNA hypomethylation is a useful model that could facilitate investigation of the pathogenic role of DNA hypomethylation in cancer. 2. Materials and methods 2.1. Plasmids Stella and Ha-Ras cDNAs were inserted into the pMY-IRES-EGFP retroviral vector or pWPI lentiviral vector [12]. For generation of shRNA constructs, oligonucleotides were cloned into the pLVTH vector. Oligonucleotide sequences are shown in Supplementary Table 1. 2.2. Cell culture and viral infections NIH3T3 cells, B16F10 cells, 293T cells, and Plat-E packaging cells [13] were maintained in Dulbecco’s modified Eagle’s Minimum Essential Medium supplemented with 100 U/mL penicillin, streptomycin sulfate, and 10% fetal bovine serum (FBS). Retroviral and lentiviral infections were carried out as previously reported [11].
2.8. Microarray array analysis Microarray analysis was performed in NIH3T3 cells including two independent Stella-expressing NIH3T3 clones and a nonclonal Stella-expressing NIH3T3 cell population, and in B16-F10 cells including three independent Stella-expressing B16-F10 clones. Total RNA was extracted from cells using an RNeasy Total RNA Mini Kit (Qiagen). RNA was labeled and hybridized to GeneChip Mouse Genome 430 2.0 arrays (Affymetrix) according to the manufacturer’s instructions. Signal values and detection calls for all samples were determined using MAS5.0 (Affymetrix). Further analyses were performed with the Subio Platform version 1.15 plug-in software (Amami). To normalize the variations in staining intensity among chips, the signal values for all probes on a given chip were divided by the 75th percentile value for expression of all genes on the chip. To eliminate changes within the range of background noise, we used probes the raw signal values of which were >50 in at least any one sample. The microarray data have been deposited in the Gene Expression Omnibus database (GEO; http://www.ncbi.nlm.nih.gov/geo/) and given the series accession number GSE68837. 2.9. Genome-wide profiling of promoter methylation by the Microarray-based Integrated Analysis of Methylation (MIAMI) method
2.3. Soft agar assay NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells (45 days after infection), Stella-expressing NIH3T3 cells (45 days after infection), and EGFP-expressing NIH3T3 cells (45 days after infection) (3 103) were resuspended into single-cell suspensions in 3 mL of medium. The cell suspensions were mixed with 1 mL of 1.5% pre-warmed soft agar (BD Biosciences) and plated into 30-mm plates in triplicate. The colonies formed on the soft agar matrix were counted under a light microscope after 2 weeks. 2.4. Global DNA methylation analysis Global DNA methylation status was measured as described previously [11]. 2.5. In vivo growth assay NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells (45 days after infection), and Stella-expressing NIH3T3 cells (45 days after infection), (5 105) were injected subcutaneously into nude mice. After 2 weeks, the tumors were harvested and each tumor was weighed. 2.6. Wound healing assay B16-F10, EGFP-expressing B16-F10, and Stella-expressing B16F10 cells were grown to confluence in six-well culture plates. Cell layers were scraped with a sterile pipette tip and incubated at 37 °C with 5% CO2. Migration from the edge of the injured monolayer was quantified by measurement of the distance between the wound edges and the recovered edges. 2.7. Metastasis assay B16-F10, EGFP-expressing B16-F10, and Stella-expressing B16F10 cells (2 105) were suspended in 200 lL of PBS and injected in the lateral tail vein of C57BL/6 mice. Lungs were harvested 14 days post-injection and fixed in formalin. Metastases were counted and subjected to statistical analysis (one-way ANOVA).
Genome-wide profiling and promoter methylation status of Stella-expressing NIH3T3 cells were evaluated using Microarraybased Integrated Analysis of Methylation (MIAMI) method [14]. Briefly, genomic DNA was digested with methylation-sensitive HpaII or the methylation-insensitive isoschizomer MspI. The digested genomes were ligated with adaptor, and amplified by PCR with primers designed against the adapter sequences. The samples were then further digested with HpaII and MspI, and amplified again with the same primers. The amplified products were then labeled with Cy3 or Cy5 and co-hybridized to a microarray. After hybridization, the microarray was scanned, and the obtained fluorescence intensities were quantified and normalized. The same pooled samples were treated first with MspI instead of HpaII and analyzed on a duplicate array to correct for falsepositives caused by single nucleotide polymorphisms or incomplete digestion. Unchanged, DNA hyper-, and hypo-methylation of each genomic regions are indicated by ‘‘0”, ‘‘+1”, and ‘‘1” in column AB, respectively (Supplementary Table 2). 2.10. RT-PCR Total RNAs prepared from cells were treated with DNase I, and subjected to RT-PCR using the ThermoScript RT-PCR system (Invitrogen) and random hexamers as primers for cDNA synthesis. The primers used in this study are shown in Supplementary Table 1. PCR were carried out following condition: 5 min at 95 °C followed by 24 and 26 cycles (for IAP and Line 1), 35 cycles of (for Arhgap20, Smagp, and Notch3), or 30 cycles (for Gapdh) of PCR consisting of 30 s at 95 °C, 30 s at 60 °C, and 1 min at 72 °C. 2.11. Bisulfite sequence analysis B16 melanoma cells and Stella-expressing clones were treated with bisulfite using an EZ DNA Methylation-Direct Kit (Zymo Research). Sequences of the PCR primers are listed in Supplementary Table 1. PCR amplification of Arhgap20, Smagp, and Notch3 promoter regions was carried out using EpiTaq HS (Takara) under the following conditions: 2 min at 95 °C followed by 35 cycles of PCR consisting of 20 s at 98 °C, 30 s at 56 °C, and 30 s at 72 °C. The PCR products were purified using a QIAquick Gel Extraction Kit (Qiagen), cloned into the pGEM-T Easy Vector (Promega), and then
4055
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
sequenced using an (Applied Biosystems).
ABI
PRISM
3100
Genetic
Analyzer
3. Results and discussion 3.1. Transformation of NIH3T3 cells by Stella
2.12. Ethics statement
Control
Ha-Ras
*
Stella Day12
Stella Day 40
Stella
Number of colonies
200
EGFP
*
100
50 N.D.
H
C
a-
on
R
tro
as
l
N.D.
E
D 109
1010 Control Ha-Ras Stella
108
Control Ha-Ras Stella
109 Number of cells
Number of cells
*
150
0
107 106 105
108 107 106 105
104
104 0
3
6
9
12
15
0
3
6
Days
F
*
*
FP
Ha-Ras
n.s.
la
Control
C
EG
B
St
A
We recently reported that Stella induced DNA hypomethylation in NIH3T3 cells. Quite unexpectedly, transformation of the NIH3T3 cells was observed during culture (Fig. 1A). Considering that DNA hypomethylation induces transformation, we carried out additional experiments. Treatment of NIH3T3 cells with high and low
el
Experiments were performed in accordance with the guidelines of the Osaka University Animal Care and Use Committee. All animal experiments were approved by Osaka University Animal Care and Use Committee.
Control
Stella
9
12
15
Days
Ha-Ras
G
500
*
*
n.s.
Tumor weight (mg)
400 300 200 100 N.D. as
la
R a-
St
el H
C
on
tro
l
0
Fig. 1. Transformation of NHI3T3 cells by Stella. (A) Morphology of NIH-3T3 cells, Stella-expressing NIH3T3 cells (days 12 and 40), and Ha-Ras-expressing NIH3T3 cells. (B, C) Anchorage-independent growth of NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells, Stella-expressing NIH3T3 cells, and EGFP-expressing NIH3T3 cells in soft agar over 14 days. Mean and standard deviation (n = 5) are shown. The number of colonies was significantly increased in Ha-Ras- and Stella-expressing NIH3T3 cells compared to the controls (*P < 0.00017, one-way ANOVA with Dunnet’s post hoc test). (D, E) Growth curves before (day 6 after retroviral infection) (D) and after (day 40 after retroviral infection) (E) transformation by Stella. Means and standard deviations (n = 3) are shown. Blue, red, and green lines indicate control NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells, and Stella-expressing NIH3T3 cells, respectively. (F) In vivo transplantation assay. Control NIH3T3 cells, Ha-Ras-expressing NIH3T3 cells, and Stella-expressing NIH3T3 cells were subcutaneously injected into the backs of immunodeficient mice (KSN/SLC mice, n = 6). Data are from representative mice at 14 days after injection. (G) Weight of tumors derived from Ha-Ras-expressing NIH3T3 cells and Stella-expressing NIH3T3 cells. No tumors were formed from parental NIH3T3 cells. Tumor weight was significantly increased in Stella-expressing NIH3T3 cells compared to controls (*P < 0.00005, one-way ANOVA with Dunnet’s post hoc test).
4056
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
Stella-transformed cells were transplantable and their in vivo growth was essentially identical to that of Ha-Ras-transformed cells.
concentrations of 5-azacytidine (5-azaC), an inhibitor of DNMT, induced cell death and blocked transformation, respectively (Supplementary Fig. 1). Meanwhile, knockdown of Dnmt1 by shRNA significantly reduced the growth rate but did not induce transformation of NIH3T3 cells (Supplementary Fig. 1). Thus, we could not induce the transformation of NIH3T3 cells by conventional methods for reducing Dnmt1 activity. Unfortunately, it is rather difficult to determine whether these results are explained by the reduced level of DNA methylation or differences in the sites of DNA demethylation. Although transformation by oncogenic Ha-Ras commenced within 7 days, that induced by Stella was observed 3–4 weeks after introduction of the gene. There were significant differences not only in the duration but also the frequency of transformation. The vast majority of Ha-Ras-expressing cells were transformed, but less than 1% of Stella-expressing cells exhibited transformation. However, the morphology of NIH3T3 cells transformed by Stella was not significantly different from that of cells transformed by Ha-Ras. In addition, like Ha-Ras-transformed cells, Stellaexpressing transformed cells showed anchorage-independent colony formation in agar (Fig. 1B and C). Although the growth of Stella-expressing cells was significantly slower than that of the control cells before transformation, it became significantly higher than that of the control cells after transformation (Fig. 1D and E). Compared to Ha-Ras, Stella took significantly longer to induce transformation, and the frequency of transformation was lower. These observations strongly suggest that transformation is not a direct oncogenic effect of Stella. Thus, we considered it likely that hypomethylation induced by Stella is the major cause of the transformation. Next, we subcutaneously injected control, Stella-transformed, and Ha-Rastransformed cells into immunodeficient mice to examine the in vivo transplantability of the cells (Fig. 1F and G). The
3.2. Genome-wide DNA methylation analysis and gene expression profiling of Stella-transformed cells Genome-wide DNA methylation conditions in NIH3T3 cells and Stella-transformed NIH3T3 cells were analyzed by the Microarraybased Integrated Analysis of Methylation (MIAMI) method [14]. There were no hypermethylated regions in the genome of Stellatransformed cells (Supplementary Table 2, Supplementary Figs. 2 and 3). By contrast, DNA hypomethylation was detected throughout the genome of Stella-transformed NIH3T3 cells (Supplementary Table 2, Supplementary Figs. 2 and 3) and the hypomethylated regions were distributed almost randomly on all chromosomes (Fig. 2A). The percentages of the hypomethylated regions in individual promoters ranged from 9.6% to 18.6% (Supplementary Fig. 2). To examine alterations in gene expression in Stella-transformed NIH3T3 cells, we carried out microarray gene expression analysis. Two independent clones and non-clonal expansion cells were subjected to the analysis and no typical expression pattern was observed (Fig. 2B). The Stella-expressing clones showed >2-fold upregulation and downregulation of 432 and 482 genes, respectively (Fig. 2C and Supplementary Table 3). No tumor suppressor genes were down-regulated in all three samples. By contrast, the potent oncogene c-Jun [15] was among the significantly upregulated genes. However, knockdown of c-Jun expression in Stella-transformed NIH3T3 cells by shRNA was not associated with significant morphological or growth changes, indicating that the increase in c-Jun expression was not the major cause of
Hypomethylationed / Total (%)
A
B 20 2 up
15
10
5
0 1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 X 0
Chromosome
C Non-clonal
Non-clonal
210
-2 down
DOWN-regulated
tro
Clone #2
l n lo al ne # C lo 1 ne #2
337
C
UP-regulated
208
Clone #1
lo
122
Clone #2
on
72
Clone #1
-c
133
82 482
C
53
61
on
432
243 48
N
26
Fig. 2. Genome-wide analysis of DNA methylation status and gene expression profiling of Stella-transformed NIH3T3 cells. (A) Genome-wide DNA methylation status determined by the Microarray-based Integrated Analysis of Methylation (MIAMI) method. The percentages of hypomethylated regions in each chromosome are shown. (B) Heat map representing color-coded expression levels of differentially expressed genes (two-way ANOVA, up- or downregulated >twofold) in non-clonal-, clone #1-, clone #2-, Stella-expressing NIH3T3 cells. (C) Venn diagram indicating overlap of up- and downregulated genes among non-clonal-, clone #1-, clone #2-, Stella-expressing NIH3T3 cells.
4057
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
genes analyzed were re-methylated after the elimination of Stella. It has been reported that Np95 binds to chromatin via histone H3K9me3 and that DNA methylation levels were restored in the Np95 KO ES cells by the forced expression of Np95 [16,17]. Considering that Stella did not affect the level of H3K9me3 [9], it is likely that the association between Np95 and H3K9me3 facilitates the remethylation of the original locus after the elimination of Stella. Three possible explanations for the role of DNA hypomethylation in cancer have been suggested: chromosomal aberrations, reactivation of transposable elements, and increases in expression of potential oncogenic genes such as oncogenes and some imprinted genes. Nonetheless, its precise function remains unclear. As shown in Fig. 3, elimination of Stella resulted in revertant cells that were similar to the original NIH3T3 cells, showing that permanent abnormalities—such as loss of heterozygosity (LOS)—were not the major cause of the transformation. Meanwhile, Stella expression did not affect the expression of retrotransposons in NHI3T3 cells (Supplementary Fig. 6). Taken together, these observations indicate that transformation by Stella cannot be explained by the three possibilities mentioned above. Although there have been many reports on global DNA hypomethylation in human malignant tumors, they were descriptive and did not determine the pathogenic role of hypomethylation [4]. It is difficult to determine whether DNA hypomethylation is the cause or result of malignant tumors. To address this question, studies were performed using mutant mice with reduced Dnmt1
transformation (Supplementary Fig. 4). We further confirmed that cancer associated genes were not altered in Stella expressing cells by gene ontology (GO) analysis (Supplementary Tables 4 and 5). Since cell adhesion-related genes were up- and down-regulated in the Stella-transformed NIH3T3 cells, it is possible that Stella facilitates anchorage independent cell growth. 3.3. Reversion of the transformed phenotype by elimination of Stella expression Oncogenic mutations may have been acquired during the relatively long period of culture before Stella-induced transformation. If this were the case, cellular transformation would have been maintained even after Stella expression was abolished. To exclude this possibility, we constructed a floxed-Stella lentivirus, infected the cells with the vector, and eliminated the gene using Cre recombinase after transformation. The morphology of the Stellatransformed cells reverted to that of NIH3T3 cells, growth of the cells was normalized, and the cells lost their in vivo tumorforming ability (Fig. 3A–C). It was further confirmed that the global DNA methylation level of these revertant cells was identical to that of control NIH3T3 cells (Fig. 3D). To examine whether restoration of DNA methylation occurs in the original locus after the elimination of Stella, we analyzed the DNA methylation status of the promoter regions of Oct3/4 and Nanog, and differential methylated regions of Peg1 and H19. As shown in Supplementary Fig. 5, all
A
B
109 Control
Stella elimination
Number of cells
Stella
108
Stella Stella elimination
107 106 105 104 0
5
10
15
20
Days
C
D
* 80
Stella elimination
CCmGG (%)
Stella
n.s.
100
*
60 40 20
tio n
el la
in a
St St
el la
el
im
C
on
tro l
0
Fig. 3. Reversion of phenotype by elimination of Stella. (A) Morphologies of Stella-expressing NIH3T3 (Stella) cells and Stella-eliminated NIH3T3 cells (Stella elimination). (B) Growth curves of control (NIH3T3 cells), Stella-expressing NIH3T3 cells, and Stella-eliminated NIH3T3 cells. Mean and standard deviation (n = 3) are shown. Blue, green, and red lines indicate control, Stella-expressing NIH3T3 cells, and Stella-eliminated NIH3T3 cells, respectively. (C) In vivo transplantation assay. 5 106 Stella-expressing NIH3T3 cells and Stella-eliminated NIH3T3 cells were injected subcutaneously into the backs of immunodeficient mice (KSN/SLC mice, n = 4). Tumors formed in all mice injected with Stella-expressing NIH3T3 cells and none with Stella-eliminated NIH3T3. Data are from representative mice at 14 days after injection. Arrow indicates tumor. (D) Global DNA methylation status of Stella-expressing NIH3T3 cells, and Stella-eliminated NIH3T3 cells. DNA methylation level was decreased significantly in Stella-expressing NIH3T3 cells compared to controls (*P < 0.015, one-way ANOVA with Dunnet’s post hoc test).
4058
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
was examined under conditions of DNA hypomethylation [19]. DNA hypomethylation increased the incidence of microadenomas with LOH at the APC locus, but strongly reduced the incidence of macroscopic intestinal tumors. Meanwhile, DNA hypomethylation caused development of multifocal liver tumors with LOH at the APC locus in these mice. Thus, simple reduction of Dnmt1 activity did not appear to induce malignant transformation. Another model involves disruption of the interactions between Dnmt1, PCNA, and Np95 by introduction of a fusion protein containing parts of Dnmt1 and Np95 [20]. Astrocyte and glial
activity. Mice with the hypomorphic allele of Dnmt1, with about 10% of the normal enzymatic activity, exhibited global DNA hypomethylation. T-cell lymphomas developed in the mice after a prolonged period, and a chromosome 15 abnormality was detected [18]. These data indicate that DNA hypomethylation has the potential to induce malignant tumor formation. However, the effects were not direct, being caused by DNA hypomethylationdependent chromosomal instability. Tumorigenesis in ApcMin/+ mice, which are predisposed to intestinal tumors by a mutation in the canonical Wnt pathway,
B *
C Number of nodules
60 EGFP
20 Stella
n.s.
60 40 20 0
l tro
la
on
el
0
C
St
EG
on
tro
l
FP
0
C
*
80
12hr
la
40
D
*
Control
80 CCmGG (%)
100
el
*
St
n.s.
FP
100
EG
A
F G #1
Control
24
2 up 29
47 39
EGFP
33
24
62
#2 Stella
#3
DOWN-regulated #1
100
0
* 80
4 39
*
323
n.s.
8
60
61
59
13
#3
#2 UP-regulated
40 20
#3
#2
el la St
EG
tro on C
FP
C
on
0
#1
tro
l
-2 down
l
Recovered area / wounded area (%)
E
Fig. 4. Stella-enhanced metastatic ability of melanoma cells. (A) Global DNA methylation status of B16-F10 cells, EGFP-expressing B16-F10 cells, and Stella-expressing B16F10 cells. Means and standard deviation (n = 3) are shown. DNA methylation level was decreased significantly in Stella-expressing B16-F10 cells compared to controls (*P < 0.0049, one-way ANOVA with Dunnet’s post hoc test). (B) Representative photographs of lung metastatic nodules. (C) The numbers of lung metastatic nodules at 14 days after transplantation of B16-F10 cells, EGFP-expressing B16-F10 cells, and Stella-expressing B16-F10 cells. Number of nodules was significantly increased in Stella-expressing B16-F10 cells compared to controls (*P < 0.014, one-way ANOVA with Dunnet’s post hoc test). (D, E) Wound healing assay of B16-F10 cells, EGFP-expressing B16-F10 cells, and Stella-expressing B16-F10 cells. Representative photographs at 12 h after scratching (D), and the percentages of the recovered area/wound area, respectively (E). Wound healing rate was significantly higher in Stella-expressing B16-F10 cells compared to controls (*P < 0.00039, one-way ANOVA with Dunnet’s post hoc test). (F) Heat map representing color-coded expression levels of differentially expressed genes (two-way ANOVA, up- or downregulated > twofold) in control, clone #1-, clone #2-, and clone#3Stella-expressing B16-F10 cells. (G) Venn diagram indicating overlap of up- and downregulated genes among clone #1-, clone #2-, and clone#3- Stella-expressing B16-F10 cells.
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
precursor cell lines expressing the fusion protein showed global DNA hypomethylation and exhibited a higher proliferation index and lower apoptosis than control cells. Chromosomal instability and the overexpression of oncogenes, such as PDGF and Ha-Ras, were induced by the fusion protein. This study indicated that the major causes of the malignant transformation were aberrations of specific oncogenes and tumor suppressor genes. The molecular mechanisms of these models seem to be different from that in Stella-transformed cells. 3.4. Stella enhanced metastatic ability of B16 melanoma cells We next assessed the effects of Stella on the metastatic ability of B16 melanoma cells. The global DNA methylation level of B16 melanoma cells was reduced by Stella expression as in the case of NIH3T3 cells (Fig. 4A). We intravenously injected wild-type B16 melanoma cells, EGFP-expressing B16 melanoma cells, and Stella-expressing B16 melanoma cells into C57BL/6 mice, and counted lung metastases 14 days later (Fig. 4B and C). Stellaexpressing cells produced significantly more lung metastases than did parental cells and control EGFP-expressing cells. Since the metastatic ability of malignant cells is generally well correlated with their migratory ability, we performed an in vitro wound healing assay to evaluate the effects of Stella on migration potential. As shown in Fig. 4D and E, Stella-expressing B16 melanoma cells showed significantly better wound healing properties compared to control B16 melanoma cells and EGFP-expressing B16 melanoma cells. These data clearly indicate that Stella expression enhanced the mobility of B16 melanoma cells. The relationship between DNA hypomethylation and metastasis is unclear. It was reported that DNA hypomethylation occurred at a very late stage in prostate cancer progression and became evident at metastasis [21]. 5-azaC was reported to increase the metastatic capacity of the human melanoma cell line DX3 after intravenous injection into immunodeficient nude mice [22]. In this study, 24h exposure of cells to 5-azaC resulted in a 40-fold increase in the number of lung tumor nodules. Moreover, the metastatic capacity increased through successive cycles of growth in vitro. Although Stella caused both transformation and metastasis, the underlying molecular mechanisms may differ, i.e., global DNA hypomethylation or the activation of some metastasis-related genes by DNA hypomethylation. 3.5. Genome-wide gene expression profiling of Stella-expressed melanoma cells Three independent clones were subjected to microarray gene expression analysis (Fig. 4F). Of these, 323 and 39 genes were > 2-fold up- and down-regulated, respectively, in the Stellaexpressing clones (Fig. 4G and Supplementary Table 6). Gene ontology analysis revealed that the GO term ‘‘metastasis” was not identified (Supplementary Tables 7 and 8). However, Arhgap20 [23], Smagp [24], and Notch3 [25], all of which are metastasisrelated genes, were upregulated and the increased expression levels of these genes were confirmed by RT-PCR (Supplementary Fig. 7). Bisulfite sequencing analysis revealed that Stella expression induced demethylation of Arhgap20 and Notch3 promoter regions (Supplementary Fig. 8). These data suggest that expression of Stella enhanced the metastatic ability of B16 melanoma cells through the induction of metastasis-related genes, such as Arhgap20 and Notch3, an effect mediated by alteration of their DNA methylation status. Alteration of DNA hypomethylation is associated with several aspects of gene regulatory mechanisms. One typical feature is activation of gene regulation through hypomethylation of promoter regions of many kinds of genes. Germline-specific tumor antigens
4059
such as MAGE, which are silenced by DNA methylation in normal tissues, have been reported to be upregulated by DNA hypomethylation of their promoter regions [26]. Although the reasons are unclear, some other genes, such as the synuclein gamma and clausin-4 genes, are hypomethylated in cancers [27,28]. Deregulation of retrotransposon genes by DNA hypomethylation has been suggested to play a role in genomic instability [29]. However, in Stella-transformed NIH3T3 cells, neither aberrant expression of retrotransposons nor specific activation of cancer-associated genes was detected by microarray analysis. Cancer metastasis is the spread of tumor cells from the original neoplasm to other organs through angiogenesis, invasion, colonization, and proliferation [30]. 5-azaC treatment was reported to upregulate the prometastatic genes urokinase plasminogen activator, matrix metalloproteinase 2, metastasis-associated gene 1, and CXC chemokine receptor 4, which stimulate cancer cell invasiveness and metastasis, in the non-invasive breast cancer cell line MCF-7 [31]. However, in Stella-expressing B16 melanoma cell clones, their expression was not changed (Supplementary Table 6). Instead, expression of the metastasis-related genes Arhgap20, Smagp, and Notch3 was upregulated in Stella-expressing B16 melanoma cell clones (Supplementary Table 6). Therefore, we postulate that Stella enhanced the metastatic ability of B16 melanoma cells by inducing metastasis-related genes through alteration of their DNA methylation status. It is uncertain whether increased expression of just these three genes is sufficient for transformation. However, we would like to emphasize that these three genes were up-regulated in three independent Stella-expressing clones. 4. Conclusions and perspective In our previous study, we unexpectedly found that a maternal factor, Stella, which preserves the DNA methylation status in early embryos, induced global DNA hypomethylation in somatic cells. In the present study, we found that global DNA hypomethylation was itself critical for the transformation of NIH3T3 cells. By contrast, it is likely that the metastatic ability of B16 melanoma cells was enhanced by activation of metastasis-related genes. Although we cannot exclude the possibility that some unknown function(s) of Stella affect the transformation of NIH3T3 cells, the Stellamediated artificial global DNA hypomethylation system will give new insights into the roles of aberrant DNA methylation in various diseases. Acknowledgements We thank Dr. D. Trono for providing the lentiviral vectors. We also thank Ms. N. Asada for assistance, and Ms. A. Mizokami and Ms. M. Imaizumi for secretarial assistance. This work was supported in part by grants from the Uehara Memorial Foundation, the Ministry of Education, Science, Sports, Culture, and Technology of Japan, and by JST, CREST. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.febslet.2015.11. 020. References [1] Feng, S., Jacobsen, S.E. and Reik, W. (2010) Epigenetic reprogramming in plant and animal development. Science 330, 622–627. [2] Smith, Z.D. and Meissner, A. (2013) DNA methylation: roles in mammalian development. Nat. Rev. Genet. 14, 204–220. [3] Robertson, K.D. (2005) DNA methylation and human disease. Nat. Rev. Genet. 6, 597–610.
4060
S. Funaki et al. / FEBS Letters 589 (2015) 4053–4060
[4] Ehrlich, M. (2009) DNA hypomethylation in cancer cells. Epigenomics 1, 239– 259. [5] Goll, M.G. and Bestor, T.H. (2005) Eukaryotic cytosine methyltransferases. Annu. Rev. Biochem. 74, 481–514. [6] Sharif, J. et al. (2007) The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature 450, 908–912. [7] Bostick, M., Kim, J.K., Esteve, P.O., Clark, A., Pradhan, S. and Jacobsen, S.E. (2007) UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science 317, 1760–1764. [8] Nakamura, T. et al. (2007) PGC7/Stella protects against DNA demethylation in early embryogenesis. Nat. Cell Biol. 9, 64–71. [9] Nakamura, T. et al. (2012) PGC7 binds histone H3K9me2 to protect against conversion of 5mC to 5hmC in early embryos. Nature 486, 415–419. [10] Bian, C. and Yu, X. (2014) PGC7 suppresses TET3 for protecting DNA methylation. Nucleic Acids Res. 42, 2893–2905. [11] Funaki, S., Nakamura, T., Nakatani, T., Umehara, H., Nakashima, H. and Nakano, T. (2014) Inhibition of maintenance DNA methylation by Stella. Biochem. Biophys. Res. Commun. 453, 455–460. [12] Wiznerowicz, M. and Trono, D. (2003) Conditional suppression of cellular genes: lentivirus vector-mediated drug-inducible RNA interference. J. Virol. 77, 8957–8961. [13] Morita, S., Kojima, T. and Kitamura, T. (2000) Plat-E: an efficient and stable system for transient packaging of retroviruses. Gene Ther. 7, 1063–1066. [14] Hatada, I. et al. (2006) Genome-wide profiling of promoter methylation in human. Oncogene 25, 3059–3064. [15] Bohmann, D., Bos, T.J., Admon, A., Nishimura, T., Vogt, P.K. and Tjian, R. (1987) Human proto-oncogene c-jun encodes a DNA binding protein with structural and functional properties of transcription factor AP-1. Science 238, 1386– 1392. [16] Rothbart, S.B. et al. (2012) Association of UHRF1 with methylated H3K9 directs the maintenance of DNA methylation. Nat. Struct. Mol. Biol. 19, 1155–1160. [17] Liu, X., Gao, Q., Li, P., Zhao, Q., Zhang, J., Li, J., Koseki, H. and Wong, J. (2013) UHRF1 targets DNMT1 for DNA methylation through cooperative binding of hemi-methylated DNA and methylated H3K9. Nat. Commun. 4, 1563. [18] Gaudet, F., Hodgson, J.G., Eden, A., Jackson-Grusby, L., Dausman, J., Gray, J.W., Leonhardt, H. and Jaenisch, R. (2003) Induction of tumors in mice by genomic hypomethylation. Science 300, 489–492. [19] Eads, C.A., Nickel, A.E. and Laird, P.W. (2002) Complete genetic suppression of polyp formation and reduction of CpG-island hypermethylation in Apc(Min/+) Dnmt1-hypomorphic mice. Cancer Res. 62, 1296–1299.
[20] Hervouet, E. et al. (2010) Disruption of Dnmt1/PCNA/UHRF1 interactions promotes tumorigenesis from human and mice glial cells. PLoS One 5, e11333. [21] Yegnasubramanian, S. et al. (2008) DNA hypomethylation arises later in prostate cancer progression than CpG island hypermethylation and contributes to metastatic tumor heterogeneity. Cancer Res. 68, 8954–8967. [22] Ormerod, E.J., Everett, C.A. and Hart, I.R. (1986) Enhanced experimental metastatic capacity of a human tumor line following treatment with 5azacytidine. Cancer Res. 46, 884–890. [23] Herold, T. et al. (2011) Expression analysis of genes located in the minimally deleted regions of 13q14 and 11q22-23 in chronic lymphocytic leukemiaunexpected expression pattern of the RHO GTPase activator ARHGAP20. Genes Chromosomes Cancer 50, 546–558. [24] Tarbe, N.G., Rio, M.C., Hummel, S., Weidle, U.H. and Zoller, M. (2005) Overexpression of the small transmembrane and glycosylated protein SMAGP supports metastasis formation of a rat pancreatic adenocarcinoma line. Int. J. Cancer 117, 913–922. [25] Zhang, Z., Wang, H., Ikeda, S., Fahey, F., Bielenberg, D., Smits, P. and Hauschka, P.V. (2010) Notch3 in human breast cancer cell lines regulates osteoblastcancer cell interactions and osteolytic bone metastasis. Am. J. Pathol. 177, 1459–1469. [26] De Smet, C., Loriot, A. and Boon, T. (2004) Promoter-dependent mechanism leading to selective hypomethylation within the 50 region of gene MAGE-A1 in tumor cells. Mol. Cell. Biol. 24, 4781–4790. [27] Czekierdowski, A., Czekierdowska, S., Wielgos, M., Smolen, A., Kaminski, P. and Kotarski, J. (2006) The role of CpG islands hypomethylation and abnormal expression of neuronal protein synuclein-gamma (SNCG) in ovarian cancer. Neuro Endocrinol. Lett. 27, 381–386. [28] Gupta, A., Godwin, A.K., Vanderveer, L., Lu, A. and Liu, J. (2003) Hypomethylation of the synuclein gamma gene CpG island promotes its aberrant expression in breast carcinoma and ovarian carcinoma. Cancer Res. 63, 664–673. [29] Howard, G., Eiges, R., Gaudet, F., Jaenisch, R. and Eden, A. (2008) Activation and transposition of endogenous retroviral elements in hypomethylation induced tumors in mice. Oncogene 27, 404–408. [30] Cheishvili, D., Boureau, L. and Szyf, M. (2014) DNA demethylation and invasive cancer: implications for therapeutics. Br. J. Pharmacol.. [31] Chik, F. and Szyf, M. (2011) Effects of specific DNMT gene depletion on cancer cell transformation and breast cancer cell invasion; toward selective DNMT inhibitors. Carcinogenesis 32, 224–232.
