Open Access

Epigenetics of gene expression in human hepatoma cells: expression profiling the response to inhibition of DNA methylation and histone deacetylation

BMC Genomics20067:181

DOI: 10.1186/1471-2164-7-181

Received: 24 March 2006

Accepted: 19 July 2006

Published: 19 July 2006

Abstract

Background

DNA methylation and histone deacetylation are epigenetic mechanisms that play major roles in eukaryotic gene regulation. We hypothesize that many genes in the human hepatoma cell line HepG2 are regulated by DNA methylation and histone deacetylation. Treatment with 5-aza-2'-deoxycytidine (5-aza-dC) to inhibit DNA methylation with and/or Trichostatin A (TSA) to inhibit histone deacetylation should allow us to identify genes that are regulated epigenetically in hepatoma cells.

Results

5-aza-dC had a much larger effect on gene expression in HepG2 cells than did TSA, as measured using Affymetrix® HG-U133 Plus 2.0 microarrays. The expression of 1504 probe sets was affected by 5-aza-dC (at p < 0.01), 535 probe sets by TSA, and 1929 probe sets by the combination of 5-aza-dC and TSA. 5-aza-dC treatment turned on the expression of 211 probe sets that were not detectably expressed in its absence. Expression of imprinted genes regulated by DNA methylation, such as H19 and NNAT, was turned on or greatly increased in response to 5-aza-dC. Genes involved in liver processes such as xenobiotic metabolism (CYP3A4, CYP3A5, and CYP3A7) and steroid biosynthesis (CYP17A1 and CYP19A1), and genes encoding CCAAT element-binding proteins (C/EBPα, C/EBPβ, and C/EBPγ) were affected by 5-aza-dC or the combination. Many of the genes that fall within these groups are also expressed in the developing fetal liver and adult liver. Quantitative real-time RT-PCR assays confirmed selected gene expression changes seen in microarray analyses.

Conclusion

Epigenetics play a role in regulating the expression of several genes involved in essential liver processes such as xenobiotic metabolism and steroid biosynthesis in HepG2 cells. Many genes whose expression is normally silenced in these hepatoma cells were re-expressed by 5-aza-dC treatment. DNA methylation may be a factor in restricting the expression of fetal genes during liver development and in shutting down expression in hepatoma cells.

Background

DNA methylation at the C5 position of a cytosine within a CpG dinucleotide is a primary mechanism in gene silencing [1]. Approximately 60% to 90% of the CpG dinucleotides in the genome of a vertebrate are methylated [2, 3]. 5-aza-2-deoxycytidine (5-aza-dC) exerts its demethylating function by sequestering DNA methyltransferase 1 (DNMT1) to 5-aza-dC-substituted DNA via the irreversible binding of the cysteine in the catalytic domain of DNMT1 to the 6 position of the cytidine ring (4,5). This decreases the cellular concentration of DNMT1, which leads to genomic DNA demethylation in the course of successive cell divisions [68]. 5-aza-dC has been shown to reactivate silenced genes in vitro [9, 10].

Regulation of gene expression in eukaryotes also involves the acetylation and deacetylation of histones, followed by chromatin remodeling [1114]. Histones H3 and H4 are acetylated on the ε-amino groups of lysines to a greater extent than H2A or H2B [15]. Histone acetylation increases the accessibility of nucleosomal DNA to transcription factors and other regulatory molecules [16, 17]. Turnover of the acetyl groups on histone molecules occurs very rapidly in cells [18, 19] and the level of acetylation is controlled by an equilibrium of histone acetyltransferase and histone deacetylase (HDAC) activities [2023]. Trichostatin A (TSA) is a reversible inhibitor of HDAC in vitro and in vivo, active at nanomolar concentrations, that causes a marked increase in histone acetylation [24].

DNA methylation and histone deacetylation can be mechanistically linked in gene silencing [25]. For example, the silenced rRNA genes can be re-expressed after either 5-aza-dC or TSA treatments [2629] but treatment with both drugs together does not result in additional re-expression, suggesting that DNA methylation and histone deacetylation cooperate to silence rRNA gene expression [26]. Microarray analyses using human colorectal cancer cells showed that HDAC inhibitors alone could re-express some genes, but not genes that contain hypermethylated CpG islands; genes with hypermethylated CpG islands were only re-expressed after 5-aza-dC treatment [30]. 5-aza-dC and TSA have been shown to act in a synergistic fashion to re-express silenced genes in colorectal carcinoma cells [31, 32].

The human hepatocellular carcinoma cell line HepG2 [33] synthesizes and secretes many of the major human plasma proteins, such as fibrinogen, plasminogen, and α2-macroglobulin [34]. These cells exhibit many cellular features of normal human hepatocytes [35] but also display characteristics resembling those of a fetal hepatocyte, because albumin and α-fetoprotein are both expressed; adult hepatocytes exclusively express albumin [36]. HepG2 cells also fail to express many of the hepatocyte-specific metabolic enzymes that characterize the neonatal and adult liver [34, 37]. These cells express very low levels of the alcohol dehydrogenase 1B (ADH1B) and ADH1C genes, both of which are regulated by epigenetic mechanisms in HepG2 cells [38].

We were interested in determining what genes are regulated by DNA methylation and histone deacetylation in HepG2 cells. Affymetrix® HG-U133 Plus 2.0 microarrays were used to measure the expression of approximately 54,000 transcripts in the presence and absence of 5-aza-dC and/or TSA. We report that the expression of many genes was altered upon 5-aza-dC treatment, and many fewer with TSA treatment. Groups of genes involved in xenobiotic metabolism, steroid biosynthesis, and CCAAT element binding were affected by 5-aza-dC (some exclusively with 5-aza-dC and TSA). Most of these genes are also expressed in the fetal and adult liver.

Results

Global gene expression in treated HepG2 cells

We treated HepG2 cells with 5-aza-dC to inhibit DNA methylation, TSA to inhibit histone deacetylation, or the combination of both 5-aza-dC and TSA, and measured changes in gene expression using Affymetrix® HG-U133 Plus 2.0 microarrays. To ensure that 5-aza-dC treatments resulted in DNA demethylation, we performed methylation-sensitive PCR on a site 2916 bp upstream of the ADH1B translational start site. If the site is methylated, it is resistant to Hpa II digestion and can be amplified by PCR; if it is demethylated, this site will be cleaved and can no longer be amplified. Figure 1A demonstrates that genomic DNA from untreated cells was methylated at the bp -2916 Hpa II site, as demonstrated by the presence of robust 319 bp PCR products. DNAs from 5-aza-dC-treated HepG2 cells had methylation levels less than one-fifth that of untreated cells, as evidenced by the absence of PCR products (Figure 1A and 1B).
https://static-content.springer.com/image/art%3A10.1186%2F1471-2164-7-181/MediaObjects/12864_2006_Article_564_Fig1_HTML.jpg
Figure 1

Treatment of HepG2 cells with 2.5 μM 5-aza-dC results in DNA demethylation. Demethylated DNA is cleaved by Hpa II digestion and cannot be amplified; therefore the amount of methylated DNA is reflected in the amount of PCR product. (A) Electropherogram of genomic DNA from untreated HepG2 cells or cells treated with 5-aza-dC; DNA was extracted and digested with Rsa I + Hpa II and subjected to PCR using primers that flank the endogenously methylated Hpa II site at bp -2916 upstream of the ADH1B gene. (B) Quantitation of the PCR products produced in (A) using 2 X PicoGreen® and an ABI PRISM® 7700 Sequence Detection System. The order of the quantitation is the same as the order of the PCR products seen in (A).

Many genes were affected by the treatment of 5-aza-dC and/or TSA in HepG2 cells (Table 1). The expression of 1504 probe sets was changed by 5-aza-dC (p-value of 0.01; FDR ≤ 0.17), 535 by TSA (FDR ≤ 0.46), and 1929 by the combination of 5-aza-dC and TSA (FDR ≤ 0.13) (Table 1). Hierarchical clustering of the microarrays based on a set of 355 genes that varied most across the experiment (chosen based upon the ratio of standard deviation to the mean ≥ 0.6, without regard for whether the variation was within or between conditions) showed a clear separation of cells that received 5-aza-dC from cells that did not (Figure 2). Within this major division, cells treated with TSA did not segregate from untreated cells. Thus, 5-aza-dC had a marked effect on gene expression while TSA had only a minimal effect. At a p-value of 0.01, 5-aza-dC treatment increased the expression of 883 probe sets while decreasing the expression of 620 probe sets. TSA increased 244 and decreased 291 while 5-aza-dC plus TSA increased 1078 and decreased 851. Additional file 1 shows genes for which any of the treatments significantly altered its expression (p ≤ 0.05).
Table 1

Number of probe sets changed by different treatments

p-/values1

5-aza-dC

TSA

5-aza-dC and TSA

0.05

3974 (.326)2

2161 (.569)

4927 (.257)

0.01

1504 (.172)

535 (.460)

1929 (.132)

0.001

346 (.075)

55 (.455)

468 (.053)

0.0001

73 (.041)

10 (.200)

126 (.024)

1Based on Welch's t-test log(signal).

2Number and (FDR) of probe sets significant at each p-value.

https://static-content.springer.com/image/art%3A10.1186%2F1471-2164-7-181/MediaObjects/12864_2006_Article_564_Fig2_HTML.jpg
Figure 2

Hierarchical cluster analysis of genes that vary across the experiment. Genes that varied across the experiment (CV ≥ 0.6) were selected and clustered [75] based upon Euclidean distance. Colors represent differences from the median expression of the gene across different conditions: the top 5% are yellow, then in declining order from red to green (median), then progressively more blue (below median) to blue (bottom 5%).

Gene re-expression in 5-aza-dC treated cells

DNA methylation is known to play a large role in silencing the expression of many genes, thus we wanted to examine which genes in HepG2 cells were re-expressed by treatment with 5-aza-dC. We defined re-expressed stringently, requiring that the gene be undetectable ("Absent" as defined by the Affymetrix MAS5 algorithm) on any of the 4 arrays before treatment and detected ("Present") on all 4 arrays after treatment [see Additional file 2]. Treatment of HepG2 cells with 5-aza-dC restored the expression of 211 probe sets (188 different genes). Only two genes were turned off by 5-aza-dC. The additional presence of TSA did not substantially change the number of probe sets (208) that were re-expressed. A relaxed definition of re-expression (allowing one of the 8 arrays in the comparison to be called "Marginal") resulted in an additional 45 probe sets being re-expressed by 5-aza-dC and 33 by 5-aza-dC and TSA. Several genes regulated by imprinting were re-expressed or significantly up-regulated (H19) by treatment with 5-aza-dC, including H19 (maternally expressed), NDN (paternal), NNAT (paternal), and MEG3 (maternal) [see Additional file 2]. The latter three genes were not detectably expressed in untreated cells; H19 was minimally detected in untreated cells but is clearly being re-expressed. Another set of re-expressed genes included TSPY1, DAZ2, and DAZ4; these genes are involved in spermatogenesis [see Additional file 2].

Tissue culture and cancer cells are known to have an increase in genome wide methylation of CpG islands [3942]. Thus, we examined the frequency of CpG islands within the promoters of 18 (10%, randomly selected) of the genes re-expressed by 5-aza-dC. Using CpG Island Finder [43, 44] we found that 13 of 18 genes contained at least one CpG island. Of the 10 genes whose expression was confirmed by quantitative real-time RT-PCR (below) we found that 8 contained at least one CpG island.

Only four genes were turned on by the combination of 5-aza-dC plus TSA but remained off after treatment with 5-aza-dC or TSA alone (Table 2). The most striking of these was chromogranin A (CHGA), which is a central regulator of catecholamine storage and release [45]; expression was not detected on any of the arrays for 5-aza-dC, TSA, or untreated cells but it was present on all 4 arrays from 5-aza-dC plus TSA-treated cells. We confirmed this synergism by quantitative real-time RT-PCR. The expression of CHGA was not reliably detected in untreated cells or in two samples of TSA-treated cells. Expression was detected at very low-levels in two other TSA-treated samples and in all three 5-aza-dC-treated cells (CHGA Ct - GAPDH Ct [ddCt] = 7.7, corresponding to 0.5% of the level of GAPDH). Cells treated with the combination of both 5-aza-dC and TSA showed greatly enhanced expression of CHGA, approximately 80-fold, illustrating the synergy (ddCt = 1.4; 38% of the level of GAPDH).
Table 2

Genes re-expressed by 5-aza-dC and TSA, but not by either drug alone

Gene Symbol

Unigene

Affymetrix ID

Fold Change

p-value

Control Signal

5-aza-dC + TSA Signal

5-aza-dC Signal

Description

CHGA

Hs.150793

204697_s_at

3.03

0.012

214

648

214

Chromogranin A (parathyroid secretory protein 1)

TKTL1

Hs.102866

214183_s_at

57.92

0.00004

49

2809

1276

Transketolase-like 1

BG1

Hs.250616

206465_at

6.16

0.00007

172

1057

630

Lipidosin

-

Hs.399852

1559213_at

43.98

0.0003

19

842

193

IMAGE:5394246, mRNA

Groups of genes involved in liver processes are affected by 5-aza-dC

Genes affected by 5-aza-dC treatment were further analyzed to determine whether the expression of particular groups of genes (as defined by Gene Ontology) was affected by DNA methylation. We were initially interested in identifying groups of genes related to liver processes since this cell line is of liver origin (hepatoma). Table 3 shows groups of genes involved in xenobiotic metabolism, steroid biosynthesis, and CCAAT element binding that are regulated by DNA methylation; some genes are regulated by both DNA methylation and histone deacetylation, but not solely methylation. Many genes within these groups are also expressed in the developing fetal liver and adult liver (Table 3).
Table 3

Groups of genes involved in liver processes that are affected by 5-aza-dC.

Gene Symbol

Unigene

Affymetrix ID

Fold Change

p-value

FDR

Control Signal

5-aza-dC Signal

Description

Xenobiotic Metabolism

ABHD6^

Hs.476454

221552_at

1.82

0.00394

0.12238

634

1151

Abhydrolase domain containing 6

AKR1C1^

Hs.460260

1555854_at

2.92

0.00027

0.04452

345

1007

Aldo-keto reductase family 1, member C2 (dihydrodiol dehydrogenase 2; bile acid binding protein; 3-a

CYP3A4^

Hs.442527

205999_x_at

1.78

0.01792

0.21785

370

660

Cytochrome P450, family 3, subfamily A, polypeptide 4

CYP3A5^

Hs.150276

214234_s_at

2.53

0.00159

0.08806

968

2447

Cytochrome P450, family 3, subfamily A, polypeptide 5

CYP3A7^

Hs.111944

211843_x_at

4.11

0.00387

0.12175

515

2120

Cytochrome P450, family 3, subfamily A, polypeptide 7

HAMP^

Hs.8821

220491_at

2.34

0.01066

0.17693

1076

2512

Hepcidin antimicrobial peptide

UGT1A6^

Hs.546397

215125_s_at

2.73

0.00318

0.11285

143

389

UDP glycosyltransferase 1 family, polypeptide A9

UGT2B15^

Hs.150207

207392_x_at

1.79

0.01965

0.22477

323

577

UDP glycosyltransferase 2 family, polypeptide B15

UGT2B28

Hs.137585

211682_x_at

-2.69

0.00131

0.08111

11036

4109

UDP glycosyltransferase 2 family, polypeptide B28

Steroid Biosynthesis

CYP17A1

Hs.438016

205502_at

1.92

0.00208

0.09647

325

622

Cytochrome P450, family 17, subfamily A, polypeptide 1

CYP19A1^

Hs.511367

203475_at

2.43

0.00356

0.11831

470

1144

Cytochrome P450, family 19, subfamily A, polypeptide 1

CYP51A1^

Hs.417077

202314_at

-1.31

0.00975

0.17073

11672

8910

Cytochrome P450, family 51, subfamily A, polypeptide 1

DHCR7^

Hs.503134

201791_s_at

-1.4

0.02202

0.23665

3628

2588

7-dehydrocholesterol reductase

FDFT1^

Hs.546253

208647_at

-1.37

0.00820

0.16076

24377

17858

Farnesyl-diphosphate farnesyltransferase 1

FDXR^

Hs.69745

207813_s_at

2.31

0.00046

0.05618

2066

4764

Ferredoxin reductase

HMGCR^

Hs.11899

202539_s_at

-1.34

0.03567

0.28427

10570

7864

3-hydroxy-3-methylglutaryl-Coenzyme A reductase

HSD17B1^

Hs.500159

205829_at

6.02

0.00853

0.16330

157

944

Hydroxysteroid (17-beta) dehydrogenase 1

HSD17B2^

Hs.162795

204818_at

2.68

0.00198

0.09502

1818

4869

Hydroxysteroid (17-beta) dehydrogenase 2

LSS^

Hs.517366

202245_at

-1.32

0.00717

0.15151

5751

4356

Lanosterol synthase (2,3-oxidosqualene-lanosterol cyclase)

PMVK

Hs.30954

203515_s_at

1.6

0.00661

0.14527

1046

1674

Phosphomevalonate kinase

RODH^

Hs.524513

37512_at

3.67

0.00046

0.05648

173

634

3-hydroxysteroid epimerase

SC5DL^

Hs.287749

211423_s_at

-1.19

0.01127

0.18081

10122

8494

Sterol-C5-desaturase (ERG3 delta-5-desaturase homolog, fungal)-like

SQLE^

Hs.71465

209218_at

-1.32

0.00603

0.14008

16131

12229

Squalene epoxidase

STARD4^

Hs.93842

226390_at

-1.22

0.01833

0.21953

4385

3593

START domain containing 4, sterol regulated

CCAAT element binding proteins

CEBPA*^

Hs.76171

204039_at

1.69

0.00049

0.04052

3106

5262

CCAAT/enhancer binding protein (C/EBP), alpha

CEBPB^

Hs.517106

212501_at

-1.95

0.01864

0.22082

28789

14762

CCAAT/enhancer binding protein (C/EBP), beta

CEBPG^

Hs.429666

225527_at

-2.15

0.04308

0.30614

3734

1741

CCAAT/enhancer binding protein (C/EBP), gamma

CEBPZ*^

Hs.135406

203341_at

-1.34

0.00286

0.08197

4880

3655

CCAAT/enhancer binding protein zeta

CUTL1*^

Hs.438974

214743_at

-1.57

0.00957

0.1289

1672

1067

Cut-like 1, CCAAT displacement protein (Drosophila)

NFIC*^

Hs.170131

226895_at

1.23

0.01727

0.16248

1452

1786

Nuclear factor I/C (CCAAT-binding transcription factor)

NFYC^

Hs.233458

202215_s_at

1.38

0.01163

0.18301

1612

2228

Nuclear transcription factor Y, gamma

* = Genes that are only significantly affected by both 5-aza-dC and TSA.

^ = Expressed in fetal and/or adult liver.

Verification of microarray results by quantitative real-time RT-PCR

Quantitative real-time RT-PCR assays were performed to confirm microarray gene expression changes (Table 4). Genes were selected based on their role in the liver, imprinting status, cell cycle function, or as important transcription factors. The pattern of differences detected by RT-PCR was consistent with those from the microarrays, but the fold-changes obtained by real-time RT-PCR assays were, in most cases, much greater than fold-changes seen in microarray analyses. This observation was particularly true for genes re-expressed by 5-aza-dC treatment; in that case, the denominator in the microarray experiment (baseline, control value) is essentially zero (the gene is called "Absent") so the exact fold change calculated from microarray data is not meaningful or accurate. RT-PCR assays allow one to obtain a lower baseline value, and therefore a higher fold-change. It should be noted that we have not determined the amplification efficiency of each assay so these calculated fold-changes are based on 100% efficiency; this can overstate the fold change.
Table 4

Validation of fold-changes from microarray analyses by real-time RT-PCR

  

5-aza-dC

TSA

5-aza-dC and TSA

 

Control

Array1

RT-PCR1

Array

RT-PCR

Array

RT-PCR

CDK10 3

300*

3.5 (.00013)

8.3 (.00000)

NT2

NT

3.8 (.00001)

7.2 (.00000)

CEBPA 3

3106

NT

NT

1.43 (.0038)

4.16 (.004)

1.7 (.00049)

10.4 (.00000)

CEBPB 3

28789

-2.0 (.019)

-2.6 (.0008)

NT

NT

-2.0 (.019)

-3.5 (.00009)

CYP3A7 3

515*

4.0 (.0043)

11 (.00000)

NT

NT

4.3 (.0043)

8.0 (.00000)

GSTP1 3

367**

8.5 (.00014)

18200 (.00000)

NT

NT

12 (.00005)

107000 (.00000)

H19 3

255

156 (.00000)

5080 (.00000)

NT

NT

141 (.00000)

330 (.00000)

NNAT

205*

8.0 (.0073)

38 (.00000)

NT

NT

10.8 (.0053)

33 (.00000)

PKIB

464*

7.4 (.0029)

70 (.00000)

NT

NT

6.6 (.0013)

35 (.00000)

S100A4 3

281*

8.8 (.0012)

27 (.00000)

3.0 (.00039)

4.9 (.00000)

8.2 (.00018)

11 (.00000)

TSPY1

11*

96 (.00002)

532 (.00000)

NT

NT

153 (.00000)

1290 (.00000)

*: Called "Absent" on all 4 control arrays; **: called "Absent" on 3 of 4 control arrays

1The fold-change for either the microarray analyses or real-time RT-PCR assays (relative fold-change). P-values (Welch's t-test) are shown in parentheses.

2NT = not tested, because it was not significantly affected in the microarray analyses.

3Expression is observed in the fetal liver.

Discussion

DNA methylation has a larger role in the inhibition of gene expression in the human hepatoma cell line HepG2 than does histone deacetylation. This is demonstrated by the greater effect by the DNMT inhibitor 5-aza-dC on gene expression than of the HDAC inhibitor TSA, as evidenced by the numbers of probe sets with altered gene expression (Table 1). It is also demonstrated by hierarchical clustering (Figure 2) in which the cells treated with 5-aza-dC, either alone or with TSA, segregated from cells not treated with TSA. 5-aza-dC treatment re-expressed 211 probe sets (188 genes) that had been silent in HepG2 cells [see Additional file 2], and up regulated more genes than were down regulated. In the overall pattern of gene expression, we do not detect much synergy between 5-aza-dC and TSA on gene expression in these cells. A small number of genes (4) required both 5-aza-dC and TSA for re-expression; we confirmed the synergistic effects of 5-aza-dC and TSA on CHGA gene expression by performing quantitative real-time RT-PCR.

A subset of autosomal genes, known as imprinted genes, display expression exclusively from one of the parental alleles [46]. The imprinted genes NDN, NNAT, and MEG3 were turned on in 5-aza-dC-treated cells (Table 4 and see Additional file 2), demonstrating that the DNA demethylation was sufficient to re-express genes silenced by DNA methylation in the HepG2 cells. The expression of H19 (maternal) was greatly increased (156-fold [see Additional file 2]; essentially turned on) while the expression of IGF2 (paternal) was decreased (-4.2-fold, p = 0.02 [see Additional file 1]. Both genes lie together on chromosome 11p15, and the upstream region of H19 is the downstream region of IGF2; this intergenic region, also known as the differentially methylated region, is responsible for monoallelic expression of both genes [4749]. If this DMR is unmethylated, the transcription factor CTCF can bind and block the activity of an enhancer that activates IGF2 expression; this allows the enhancer to now interact with the H19 promoter [47]. Our results are consistent with this.

5-aza-dC also turned on the expression of several Y chromosome-specific genes [see Additional file 2]. Past studies using prostate cancer cell lines showed that DNA methylation regulates the expression of these genes and that 5-aza-dC can restore their expression [50]. The expression of TSPY1 is normally restricted to the testis where it plays a role in spermatogonial proliferation [51]. TSPY1 contains a putative CpG island in its promoter that is hypermethylated in melanoma cell lines; treatment with 5-aza-dC can restore TSPY1 expression [52]. The DAZ cluster of genes are thought to be involved in germ cell differentiation in humans and old world monkeys [53]; deletions in this region can cause infertility in males [54]. This finding is interesting given that we treated hepatoma cells, a cell type that normally does not express spermatogenesis genes. It confirms that DNA methylation plays a large role in shutting down the expression of many genes that have no role in contributing to the proper functioning of that cell type.

We further analyzed genes that were significantly affected by 5-aza-dC treatment and identified groups of genes that are involved in xenobiotic metabolism, steroid biosynthesis, and CCAAT element binding (Table 3). The most affected in the xenobiotic metabolism group, the CYP3A genes, play major roles in metabolic disposition of wide variety of drugs [55]. CYP3A4 and CYP3A5 comprise nearly 30%-40% of total hepatic cytochrome p450 [55] while CYP3A7 expression is restricted to the fetal liver [56]. CYP3A expression in untreated HepG2 cells is fairly low (signals ranging from 370 to 968), suggesting that their expression is reduced in these partially de-differentiated cells; however, 5-aza-dC greatly increased the expression of each CYP3A gene. This increase in expression was confirmed by real-time RT-PCR for CYP3A7 (Table 4).

Genes within the steroid biosynthesis group also included two CYP genes. The CYP19A1 gene, which was re-expressed by 5-aza-dC treatment [see Additional file 2], encodes the main estrogen biosynthesis enzyme, which converts androgen to estrogen [57]. CYP19A1 expression can be increased in endometriotic stromal cells upon in vivo down-regulation of C/EBPβ [58]. This finding is interesting given that C/EBPβ expression was decreased by 2-fold; this may contribute to CYP19A1 re-expression. The CYP17A1 enzyme (Table 3) functions in the biosynthesis of testosterone through the regulation of 17-hydroxylase and 17, 20-lyase [59]. CYP17A1 is not expressed in the fetal or adult liver; hence DNA methylation may be a major mechanism to transcriptionally silence its expression in the liver.

We examined seven genes that bind CCAAT DNA elements to regulate transcription of many important liver genes (Table 3). Most notable are the three genes encoding C/EBPα, C/EBPβ and C/EBPγ (Table 3; CEBPA, CEBPB, and CEBPG) that were differentially affected by 5-aza-dC or TSA and the combination of 5-aza-dC and TSA. Surprisingly, the expression of CEBPB and CEBPG were decreased by 5-aza-dC; because methylation generally inhibits gene expression [1], this down-regulation could be a secondary effect of the increase in an inhibitory factor. CEBPA was up-regulated by TSA and the combination of 5-aza-dC and TSA, but not with 5-aza-dC alone. The transcriptional state of many other hepatocyte-specific genes in these cells is likely altered based on the expression profile of these C/EBP genes. CEBPB can generate two isoforms from a single mRNA: the full-length protein termed liver activation protein (LAP) or the truncated liver inhibitory protein (LIP) isoform [6063]. LIP is a dominant-negative factor that can attenuate transcription by forming a heterodimer with LAP to negate the transactivation ability of LAP [62]. LIP is still able to attenuate the transcriptional activity of LAP even when present at sub-stoichiometric amounts (i.e. LAP/LIP ratio of 5:1) [62, 64]. C/EBPγ heterodimerization with C/EBPα and LAP can also attenuate transcription of C/EBP responsive promoters, suggesting dominant negative regulation by this factor [65]. Hence, the decreased expression of C/EBPβ and C/EBPγ transcripts may aid in increasing the expression of the hepatocyte-specific genes they regulate.

The CEBPA gene is regulated by histone deacetylation (and slightly increased with 5-aza-dC and TSA) in the HepG2 human hepatoma cell line. An increase in C/EBPα expression will increase the expression of several hepatocyte-specific genes that are regulated by this factor. However, a larger picture emerges from our data. HepG2 cells display many characteristics of fetal and adult hepatocytes. Many genes regulated by DNA methylation that function in liver processes (Table 3) are expressed in the fetal and adult liver. DNA demethylation may be, in part, responsible for silencing the expression of fetal liver genes. C/EBPα and C/EBPβ may play important roles in this. The haptoglobin (Hp) gene, which is important in liver development, is transcriptionally regulated by C/EBPα and C/EBPβ in the rat liver; Dinic et al. have shown that Hp transcription is regulated by C/EBPα during normal liver development while C/EBPβ regulates this gene during the acute phase response during the later phase of differentiation and in the adult [66]. Also, the suppression of C/EBPα expression may be a prerequisite to biliary cell differentiation in a hepatoblast population in the developing mouse liver [67].

Conclusion

In summary, we demonstrate that DNA methylation plays a much larger role than histone deacetylation in regulating gene expression in HepG2 human hepatoma cells. Many genes were in fact turned on in these cells by the DNMT inhibitor 5-aza-dC; which included imprinted genes and spermatogenesis genes. 5-aza-dC affected groups of genes involved in liver processes such as xenobiotic metabolism, steroid biosynthesis, and CCAAT element binding. A majority of the genes in these groups are expressed in the developing fetal and adult liver indicating that DNA methylation may play a role in restricting the expression of fetal liver genes and in shutting down expression in hepatoma cells.

Methods

Drug treatments

5-aza-dC (Sigma, St. Louis, MO) and TSA (Sigma) were prepared in dimethyl sulfoxide (DMSO) (Sigma). HepG2 cells (HB-8065; ATCC, Manassas, VA) (5 × 105 cells) were plated on 100 mm culture dishes and grown in minimum essential medium (Sigma) with 10% fetal bovine serum (Life Technologies, Rockville, MD) and 2 mM glutamine in a 5% CO2 atmosphere. A 2 × 2 experimental design was used: cells were treated with 2.5 μM 5-aza-dC, 500 nM TSA, both 2.5 μM 5-aza-dC and 500 nM TSA, or DMSO (untreated). Plates were seeded and harvested at the same time; appropriate dishes were treated with 5-aza-dC 48 h after cell seeding and their media and drugs were replaced every 24 h. TSA was incubated with appropriate dishes for the last 24 h before harvesting. DNA and RNA were extracted 4 days after the initial 5-aza-dC treatment using TRIzol® (Invitrogen, Carlsbad, CA).

DNA methylation analysis

Methylation-sensitive PCR was modified from a protocol [68] and quantitation of PCR products were previously described by Dannenberg et al. [38]. This assay is based on the fact that Hpa II cannot digest a methylated CCGG site; thus fragments containing methylated CCGG sites remain intact and can be amplified by PCR, whereas unmethylated CCGG sites are cut and cannot be amplified. The region upstream of ADH1B from bp -2956 to -2638 was analyzed using primers HE1741/1742 (Table 5).
Table 5

Primer sequences

Primers

Sequences

HE1741

CCAGGGATTAGGAGTGGACC

HE1742

GGAGGGGAAGAGCAGTTGTC

HE2251

GATTGGACCTCATTACATAAGGG

HE2252

TCTTTTGGGGTTAGCAAGTTG

HE2261

AGGGTTGTGTGAAGACAGCG

HE2262

TTGTTGCGGTGGTGGGGAAAG

HE2245

GGACCCTCAGCCTTGTTTG

HE2246

AACCCCCCTCACCTCATTGG

HE2255

TGTCCAAACCAACCGCAC

HE2256

AACAGCAACAAGCCCGTAG

HE2253

CCTGATGTCCAGCAGAAAGTGC

HE2254

GTTAGAAGAAGTCCTCCAAAGCG

HE2219

TACCAGTCCAATACCATCCTGCG

HE2220

TGCCTTCACATAGTCATCCTTGC

HE2241

ACTGGTTGGAGTTGTGGAGACG

HE2242

TGGAATGCTTGAAGGCTGCTC

HE2369

AATCAAAACACCGCACCAG

HE2370

ATCAGTGAGGGGCAAGGGGGGTTC

HE2243

CAGACGGAACCTCAGATTTGCC

HE2244

CAGTGCGATTTGGAAGTAACCTG

HE2265

GCAAAGAGGGTGACAAGTTCAAGC

HE2266

GCCAGGGTGGAAAAAAAAAAGTGC

HE2249

CTCTGACCACAACTTCGCAG

HE2250

CCACATTTACCCCCTCTTCC

HE3106

CTACCCCGAGGAGAAGAAAGAG

HE3107

CTGTGTTCAGAGAAGGAATCGTC

Microarray analysis

Total RNA prepared from HepG2 cells was treated with 1 U of DNase I, RNase-Free (Roche, Indianapolis, IN) per microgram of RNA. An RNeasy® Mini Kit (Qiagen, Valencia, CA) was used to purify RNA. RNA quality was then tested using a Bioanalyzer (Agilent Technologies, Palo Alto, CA) and by measuring absorbance from 210 to 350 nm. RNA preparations from four independent dishes from each experimental condition (5-aza-dC, TSA, both 5-aza-dC and TSA, and untreated) were separately processed and analyzed. Ten micrograms of total RNA was synthesized into cDNA for each sample, and then biotinylated cRNA was generated by in vitro transcription, following the standard Affymetrix protocols [69]. Biotinylated cRNAs were fragmented, and each sample was hybridized to an Affymetrix® HG-U133 Plus 2.0 gene chip (54,675 probe sets) at 42°C for 17 h. Chips were washed, stained, and scanned following the standard Affymetrix protocol [69].

Statistical analysis

Expression data generated by the Affymetrix Microarray Suite®, version 5.0, were exported and analyzed using Microarray Data Portal (MDP) [70]. Data have been deposited into GEO [71] under Series accession number GSE5230. Probe sets that were not called present on at least 50% of the arrays in at least one-treatment group were eliminated before further analysis; these primarily represent noise [72, 73]. Log-transformed signals were then compared using Welch's approximate t-test, which allows for unequal variance. False discovery rate (FDR) was calculated according to Benjamini and Hochberg [74]. Genes were considered turned from "off to on" or "on to off" if the difference of the fraction present was +/- 1; that is, if genes went from "Absent" (according to the MAS5 algorithm) in all 4 arrays to "Present" in all four arrays, or vice versa and the log signal differed significantly (p ≤ 0.05). Genes turned on by the combination of 5-aza-dC plus TSA (as defined above) but not by 5-aza-dC alone were also selected, based on a p value for both vs. 5-aza-dC <0.05 and difference in fraction present ≥ 0.5. Gene groupings were created using Gene Ontology on MDP; the CCAAT-binding protein genes were grouped manually. Fetal liver expression was determined by using the Expression Profile Viewer (NCBI). Genes are considered as expressed in the fetal liver if they are expressed in the liver and also expressed in the embryo.

Putative CpG islands were identified using the CpG Island Finder [43, 44]. We selected a random 10% of the genes that were re-expressed [see Additional file 2] and examined their promoters from -1500 to +500 bp. We also examined 10 genes that were confirmed by RT-PCR (Table 4). CpG Island Finder uses a sliding window 201 bp in length to calculate the CpG dinucleotide percentage for each window where it defines the maximum of these CpG percentages as the CpG score and the corresponding window as the CpG window [43]. Assuming a bimodal normal distribution and an overlap of 10% (5% from each group), 6.5 as a cutoff value is often used because it satisfies the condition μ+2σ < 6.5 < μ-2σ [43]. However, we used the default value of 6%, set by the CpG Island Finder [44], to identify promoter CpG islands.

Hierarchical clustering

A set of 355 genes that varied across the experiment (without consideration of whether variation was within or between conditions, to avoid bias) was selected based upon a coefficient of variation (CV; ratio of standard deviation to mean) of 0.6 or greater. Hierarchical clustering of the microarrays [75] was performed using MDP.

Real-time quantitative RT-PCR

The following human mRNA sequences: CDK10, CEBPA, CEBPB, CHGA, CYP3A7, GSTP1, H19, NNAT, PKIB, S100A4, and TSPY1 were used to design gene-specific primers for real-time RT-PCR using MacVector (Accelrys, San Diego, CA). Primer sequences were tested for uniqueness by aligning them against the human genome using NCBI BLAST [76]. The RT-PCR products and Accession numbers (GenBank) for each mRNA sequence are listed in Table 6; the sequences of the primers used are listed in Table 5.
Table 6

Real-time RT-PCR products. Positions are relative to the translational start site.

Gene

Accession_No.

Position

Primer pair

CDK10

NM_003674.2

+436 to +946

HE2261/62

CEBPA

NM_004364.2

+1900 to +2254

HE2245/46

CEBPB

NM_005194.2

+1216 to +1483

HE2255/56

CHGA

NM_001275

+1629 to +2146

HE3106/07

CYP3A7

NM_000765.2

+972 to +1453

HE2253/54

GSTP1

NM_000852.2

+189 to +365

HE2219/20

H19

BC_053637

+1342 to +1546

HE2241/42

NNAT

NM_005386

+613 to +967

HE2369/70

PKIB

NM_181795.1

+115 to +454

HE2243/44

S100A4

NM_002961

+61 to +387

HE2265/66

TSPY1

NM_003308.2

+782 to +1080

HE2249/50

Four 500 ng aliquots of TSA-treated and untreated RNA preparations used in the microarray analyses and three 500 ng aliquots of 5-aza-dC-treated and 5-aza-dC and TSA-treated (only three used for these conditions due to little or no total RNA available after microarray analyses) were reverse transcribed using the SuperScript First Strand Synthesis System (Invitrogen). Resulting cDNAs were diluted 1:25, then amplified using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). GAPDH (glyceraldehyde-3-phosphate dehydrogenase) expression was quantified using TaqMan® GAPDH Control Reagents (Applied Biosystems) as an internal control to normalize each gene against (note: GAPDH amplification not run in parallel with each gene due to limiting amounts of cDNA). PCR products were quantified in real-time using an ABI PRISM® 7700 Sequence Detection System (Applied Biosystems). PCR conditions were 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 sec and 51–60°C for 20 sec. Ct values for each gene and condition were normalized against GAPDH and then each condition was normalized against the untreated cells to obtain a fold-change. Welch's t-tests were performed to analyze the real-time RT-PCR data. We did not determine the amplification efficiency of each assay; the fold-changes are calculated based on 100% efficiency.

Declarations

Acknowledgements

This research was supported in part by R37AA06460 from the National Institute of Alcohol Abuse and Alcoholism. We thank Tiffeny Cross for technical assistance with cell culture and Ron Jerome and Chunxiao Zhu for preparation of labeled cRNA and handling of microarrays. Microarray experiments were carried out using the facilities of the Center for Medical Genomics at Indiana University School of Medicine, which is supported in part by a grant to HJE from the Indiana 21st Century Research and Technology Fund and by the Indiana Genomics Initiative (INGEN). INGEN is supported in part by the Lilly Endowment, Inc.

Authors’ Affiliations

(1)
Department of Biochemistry and Molecular Biology, Indiana University School of Medicine

References

  1. Ng HH, Bird A: DNA methylation and chromatin modification. Curr Opin Genet Dev. 1999, 9: 158-163.PubMedView ArticleGoogle Scholar
  2. Kass SU, Pruss D, Wolffe AP: How does DNA methylation repress transcription?. Trends Genet. 1997, 13: 444-449.PubMedView ArticleGoogle Scholar
  3. Riggs AD, Porter TN: Overview of epigenetic mechanisms. Epigenetic mechanisms of gene regulation. Edited by: Russo VEA, Martienssen RA, Riggs AD. 1996, Cold Spring Harbor: Cold Spring Harbor Laboratory Press, 29-45.Google Scholar
  4. Juttermann R, Li E, Jaenisch R: Toxicity of 5-aza-2'-deoxycytidine to mammalian cells is mediated primarily by covalent trapping of DNA methyltransferase rather than DNA demethylation. Proc Natl Acad Sci USA. 1994, 91: 11797-11801.PubMedPubMed CentralView ArticleGoogle Scholar
  5. Ferguson AT, Vertino PM, Spitzner JR, Baylin SB, Muller MT, Davidson NE: Role of estrogen receptor gene demethylation and DNA methyltransferase. DNA adduct formation in 5-aza-2'-deoxycytidine-induced cytotoxicity in human breast cancer cells. J Biol Chem. 1997, 272: 32260-32266.PubMedView ArticleGoogle Scholar
  6. Creusot F, Acs G, Christman JK: Inhibition of DNA methyltransferase and induction of Friend erythroleukemia cell differentiation by 5-azacytidine and 5-aza-2'-deoxycytidine. J Biol Chem. 1982, 257: 2041-2048.PubMedGoogle Scholar
  7. Michalowsky LA, Jones PA: Differential nuclear protein binding to 5-azacytosine-containing DNA as a potential mechanism for 5-aza-2'-deoxycytidine resistance. Mol Cell Biol. 1987, 7: 3076-3083.PubMedPubMed CentralView ArticleGoogle Scholar
  8. Santi DV, Norment A, Garrett CE: Covalent bond formation between a DNA-cytosine methyltransferase and DNA containing 5-azacytosine. Proc Natl Acad Sci USA. 1984, 81: 6993-6997.PubMedPubMed CentralView ArticleGoogle Scholar
  9. Herman JG, Merlo A, Mao L, Lapidus RR, Issa J-P, Davidson NE, Sidransky D, Baylin SB: Inactivation of the CDKN2/p16MTS1 gene is frequently associated with aberrant DNA methylation in all common human cancers. Cancer Res. 1995, 55: 4525-4530.PubMedGoogle Scholar
  10. Otterson GA, Khleif SN, Chen W, Coxon AB, Kaye FJ: CDKN2 gene silencing in lung cancer by DNA hypermethylation and kinetics of p16INK4 protein induction by 5-aza 2' deoxycytidine. Oncogene. 1995, 11: 1211-1216.PubMedGoogle Scholar
  11. Kingston RE, Narlikar GJ: ATP-dependent remodeling and acetylation as regulators of chromatin fluidity. Genes Dev. 1999, 13: 2339-2352.PubMedView ArticleGoogle Scholar
  12. Kornberg RD, Lorch Y: Chromatin-modifying and -remodeling complexes. Curr Opin Genet Dev. 1999, 9: 148-151.PubMedView ArticleGoogle Scholar
  13. Vignali M, Hassan AH, Workman JL: ATP-dependent chromatin-remodeling complexes. Mol Cell Biol. 2000, 20: 2167-2175.View ArticleGoogle Scholar
  14. Wu J, Grunstein M: 25 years after the nucleosome model: chromatin modifications. Trends Biochem Sci. 2000, 25: 619-623.PubMedView ArticleGoogle Scholar
  15. Spencer VA, Davie JR: Role of covalent modifications of histones in regulating gene expression. Gene. 1999, 240: 1-12.PubMedView ArticleGoogle Scholar
  16. Lee DY, Hayes JJ, Pruss D, Wolffe AP: A positive role for histone acetylation in transcription factor access to nucleosomal DNA. Cell. 1993, 72: 73-84.PubMedView ArticleGoogle Scholar
  17. Vettesse-Dadey M, Grant PA, Hebbes TR, Crane-Robinson C, Allis CD, Workman JL: Acetylation of histone H4 plays a primary role in enhancing transcription factor binding to nucleosomal DNA in vitro. EMBO J. 1996, 15: 2508-2518.Google Scholar
  18. Sanders LA, Schechter NM, McCarty KS: A comparative study of histone acetylation, histone deacetylation, and ribonucleic acid synthesis in avian reticulocytes and erythrocytes. Biochemistry. 1973, 12: 783-91.PubMedView ArticleGoogle Scholar
  19. Jackson V, Shires A, Chalkley R, Granner DK: Studies on highly metabolically active acetylation and phosphorylation of histones. J Biol Chem. 1975, 250: 4856-4863.PubMedGoogle Scholar
  20. Csordas A: On the biological role of histone acetylation. Biochem J. 1990, 265: 23-38.PubMedPubMed CentralView ArticleGoogle Scholar
  21. Inoue A, Fujimoto D: Histone deacetylase from calf thymus. Biochim Biophys Acta. 1970, 220: 307-316.PubMedView ArticleGoogle Scholar
  22. Kikuchi H, Fujimoto D: Multiplicity of histone deacetylase from calf thymus. FEBS Lett. 1973, 29: 280-282.PubMedView ArticleGoogle Scholar
  23. Vidali G, Boffa LC, Allfrey VG: Properties of an acidic histone-binding protein fraction from cell nuclei. Selective precipitation and deacetylation of histones F2A1 and F3. J Biol Chem. 1972, 247: 7365-7373.PubMedGoogle Scholar
  24. Yoshida M, Kijima M, Akita M, Beppu T: Potent and specific inhibition of mammalian histone deacetylase both in vivo and in vitro by Trichostatin A. J Biol Chem. 1990, 265: 17174-17179.PubMedGoogle Scholar
  25. Bird A: DNA methylation patterns and epigenetic memory. Genes Dev. 2002, 16: 6-21.PubMedView ArticleGoogle Scholar
  26. Chen ZJ, Pikaard CS: Epigenetic silencing of RNA polymerase I transcription: a role for DNA methylation and histone modification in nucleolar dominance. Genes Dev. 1997, 11: 2124-2136.PubMedPubMed CentralView ArticleGoogle Scholar
  27. Lawrence RJ, Earley K, Pontes O, Silva M, Chen ZJ, Neves N, Viegas W, Pikaard CS: A concerted DNA methylation/histone methylation switch regulates rRNA gene dosage control and nucleolar dominance. Mol Cell. 2004, 13: 599-609.PubMedView ArticleGoogle Scholar
  28. Santoro R, Grummt I: Molecular mechanisms mediating methylation-dependent silencing of ribosomal gene transcription. Mol Cell. 2001, 8: 719-725.PubMedView ArticleGoogle Scholar
  29. Zhou Y, Santoro R, Grummt I: The chromatin remodeling complex NoRC targets HDAC1 to the ribosomal gene promoter and represses RNA polymerase I transcription. EMBO J. 2002, 21: 4632-4640.PubMedPubMed CentralView ArticleGoogle Scholar
  30. Suzuki H, Gabrielson E, Chen W, Anbazhagan R, Van Engeland M, Weijenberg MP, Herman JG, Baylin SB: A genomic screen for genes upregulated by demethylation and histone deacetylase inhibition in human colorectal cancer. Nat Genet. 2002, 31: 141-149.PubMedView ArticleGoogle Scholar
  31. Cameron EE, Bachman KE, Myohanen S, Herman JG, Baylin SB: Synergy of demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer. Nat Genet. 1999, 21: 103-107.PubMedView ArticleGoogle Scholar
  32. Zhu WG, Otterson GA: The interaction of histone deacetylase inhibitors and DNA methyltransferase inhibitors in the treatment of human cancer cells. Curr Med Chem Anti-Canc Agents. 2003, 3: 187-199.View ArticleGoogle Scholar
  33. Aden DP, Fogel A, Plotkin S, Damjanov I, Knowles BB: Controlled synthesis of HBsAg in a differentiated human liver carcinoma-derived cell line. Nature (London). 1979, 282: 615-616.View ArticleGoogle Scholar
  34. Knowles BB, Howe CC, Aden DP: Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen. Science. 1980, 209: 497-499.PubMedView ArticleGoogle Scholar
  35. Bouma ME, Rogier E, Verthier N, Labarre C, Feldmann G: Further cellular investigation of the human hepatoblastoma-derived cell line HepG2: morphology and immunocytochemical studies of hepatic-secreted proteins. In Vitro Cell Dev Biol. 1989, 25: 267-75.PubMedView ArticleGoogle Scholar
  36. Kunnath L, Locker J: Developmental changes in the methylation of the rat albumin and alpha-fetoprotein genes. EMBO J. 1983, 2: 317-324.PubMedPubMed CentralGoogle Scholar
  37. Cassio D, Weiss MC: Expression of fetal and neonatal hepatic functions by mouse hepatoma-rat hepatoma hybrids. Somatic Cell Genet. 1979, 5: 719-738.PubMedView ArticleGoogle Scholar
  38. Dannenberg LO, Chen H-J, Tian H, Edenberg HJ: Differential regulation of the alcohol dehydrogenase 1B (ADH1B) and ADH1C genes by DNA methylation and histone deacetylation. Alcohol Clin Exp Res. 2006, 30: 928-937.PubMedView ArticleGoogle Scholar
  39. Jones PA, Wolkowicz MJ, Rideout WM, Gonzales FA, Marziasz CM, Coetzee GA, Tapscott SJ: De novo methylation of the MyoD1 CpG island during the establishment of immortal cell lines. Proc Natl Acad Sci USA. 1990, 87: 6117-6121.PubMedPubMed CentralView ArticleGoogle Scholar
  40. Antequera F, Boyes J, Bird A: High levels of de novo methylation and altered chromatin structure at CpG islands in cell lines. Cell. 1990, 62: 503-514.PubMedView ArticleGoogle Scholar
  41. Jones PA: DNA methylation errors and cancer. Cancer Res. 1996, 56: 2463-2467.PubMedGoogle Scholar
  42. Baylin SB, Herman JG, Graff JR, Vertino PM, Issa JP: Alterations in DNA methylation: a fundamental aspect of neoplasia. Adv Cancer Res. 1998, 72: 141-196.PubMedView ArticleGoogle Scholar
  43. Davuluri RV, Grosse I, Zhang MQ: Computational identification of promoters and first exons in the human genome. Nat Genet. 2001, 29: 412-417.PubMedView ArticleGoogle Scholar
  44. CpG Island Finder: [http://bioinformatics.med.ohio-state.edu/CpGisland/]
  45. Taupenot L, Harper KL, O'Connor DT: The chromogranin-secretogranin family. N Eng J Med. 2003, 348: 1134-1149.View ArticleGoogle Scholar
  46. Bartolomei MS, Tilghman SM: Genomic imprinting in mammals. Annu Rev Genet. 1997, 31: 493-525.PubMedView ArticleGoogle Scholar
  47. Bell AC, Felsenfeld G: Methylation of a CTCF-dependent boundary controls imprinted expression of the igf2 gene. Nature. 2000, 405: 482-485.PubMedView ArticleGoogle Scholar
  48. Pant V, Mariano P, Kanduri C, Mattson A, Lobanenkov VV, Heuchel R, Ohlsson R: The nucleotides responsible for the direct physical contact between the chromatin insulator protein CTCF and the H19 imprinting control region manifest parent of origin-specific long-distance insulation and methylation-free domains. Gene Dev. 2003, 17: 586-590.PubMedPubMed CentralView ArticleGoogle Scholar
  49. Thorvaldsen JL, Duran KL, Bartolomei MS: Deletion of the H19 differentially methylated domain results in loss of imprinted expression of H19 and Igf2. Genes Dev. 1998, 12: 3693-3702.PubMedPubMed CentralView ArticleGoogle Scholar
  50. Dasari VK, Deng D, Perinchery G, Yeh CC, Dahiya R: DNA methylation regulates the expression of Y chromosome specific genes in prostate cancer. J Urol. 2002, 167: 335-338.PubMedView ArticleGoogle Scholar
  51. Schnieders F, Dork T, Arnemann J, Vogel T, Werner M, Schmidtke J: Testis-specific protein, Y-encoded (TSPY) expression in testicular tissues. Hum Mol Genet. 1996, 5: 1801-1807.PubMedView ArticleGoogle Scholar
  52. Gallagher WM, Bergin OE, Rafferty M, Kelly ZD, Nolan IM, Fox EJ, Culhane AC, McArdle L, Fraga MF, Hughes L, Currid CA, O'Mahony F, Byrne A, Murphy AA, Moss C, McDonnell S, Stallings RL, Plumb JA, Esteller M, Brown R, Dervan PA, Easty DJ: Multiple markers for melanoma progression regulated by DNA methylation: insights from transcriptome studies. Carcinogenesis. 2005, 26: 1856-1867.PubMedView ArticleGoogle Scholar
  53. Reijo R, Lee TY, Salo P, Alagappan R, Brown LG, Rosenber M, Rozen F, Jaffe T, Straus D, Hovatta O: Diverse spermatogenic defects in humans caused by Y chromosome deletions encompassing a novel RNA-binding protein gene. Nat Genet. 1995, 10: 383-393.PubMedView ArticleGoogle Scholar
  54. Vogt PH, Edelmann A, Kirsch S, Henegariu O, Hirschmann P, Kiesewetter F, Kohn FM, Schill WB, Farah S, Ramos C, Hartmann M, Hartschuh W, Meschede D, Behre HM, Castel A, Nieschlag E, Weidner W, Grone HJ, Jung A, Engel W, Haidl G: Human Y chromosome azoospermia factors (AZF) mapped to different subregions in Yq11. Hum Mol Genet. 1996, 5: 933-943.PubMedView ArticleGoogle Scholar
  55. Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J, Watkins PB, Daly A, Wrighton SA, Hall SD: Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression. Nat Genet. 2001, 27: 383-391.PubMedView ArticleGoogle Scholar
  56. Wrighton SA, Vandenbranden M: Isolation and characterization of human fetal liver cytochrome P450HLp2: a third member of the P450III gene family. Arch Biochem Biophys. 1989, 268: 144-151.PubMedView ArticleGoogle Scholar
  57. Voigt KD, Bartsch W: Intratissular androgens in benign prostatic hyperplasia and prostatic cancer. J Steroid Biochem. 1986, 25: 749-57.PubMedView ArticleGoogle Scholar
  58. Yang S, Fang Z, Suzuki T, Sasano H, Zhou J, Gurates B, Tamura M, Ferrer K, Bulun S: Regulation of aromatase P450 expression in endometriotic and endometrial stromal cells by CCAAT/enhancer binding proteins (C/EBPs): decreased C/EBPbeta in endometriosis is associated with overexpression of aromatase. J Clin Endocrinol Metab. 2002, 87: 2336-2345.PubMedGoogle Scholar
  59. Picado-Leonard J, Miller WL: Cloning and sequence of the human gene for P450c17 (steroid 17 -hydroxylase/17,20 lyase): similarity with the gene for P450c21. DNA. 1987, 6: 439-48.PubMedView ArticleGoogle Scholar
  60. Baer M, Williams SC, Dillner A, Schwartz RC, Johnson PF: Autocrine signals control CCAAT/enhancer binding protein β expression, localization, and activity in macrophages. Blood. 1998, 92: 4353-4365.PubMedGoogle Scholar
  61. Buck M, Turler H, Chojkier M: LAP (NF-IL-6), a tissue-specific transcriptional activator, is an inhibitor of hepatoma cell proliferation. EMBO J. 1994, 13: 851-860.PubMedPubMed CentralGoogle Scholar
  62. Descombes P, Schibler U: A liver-enriched transcriptional activator protein, LAP, and a transcriptional inhibitory protein, LIP, are translated from the same mRNA. Cell. 1991, 67: 569-579.PubMedView ArticleGoogle Scholar
  63. Ossipow V, Descombes P, Schibler U: CCAAT/enhancer-binding protein mRNA is translated into multiple proteins with different transcription activation potentials. Proc Natl Acad Sci USA. 1993, 90: 8219-8223.PubMedPubMed CentralView ArticleGoogle Scholar
  64. Dearth LR, Hutt J, Sattler A, Gigliotti A, DeWille J: Expression and function of CCAAT/enhancer binding protein β (C/EBP β) LAP and LIP isoforms in mouse mammary gland, tumors and cultured mammary epithelial cells. J Cell Biochem. 2001, 82: 357-370.PubMedView ArticleGoogle Scholar
  65. Cooper C, Henderson A, Artandi S, Avitahi N, Calame K: Ig/EBP (C/EBP gamma) is a transdominant negative inhibitor of C/EBP family transcriptional activators. Nucleic Acids Res. 1995, 23: 4371-4377.PubMedPubMed CentralView ArticleGoogle Scholar
  66. Dinic S, Bogojevic D, Petrovic M, Poznanovic G, Ivanovic-Matic S, Mihailovic M: C/EBP alpha and C/EBP beta regulate haptoglobin gene expression during rat liver development and the acute-phase response. Mol Biol Rep. 2005, 32: 141-147.PubMedView ArticleGoogle Scholar
  67. Shiojiri N, Takeshita K, Yamasaki H, Iwata T: Suppression of C/EBP alpha expression in biliary cell differentiation from hepatoblasts during mouse liver development. J Hepatol. 2004, 41: 790-798.PubMedView ArticleGoogle Scholar
  68. Liang G, Gonzalgo ML, Salem C, Jones PA: Identification of DNA methylation differences during tumorigenesis by methylation-sensitive arbitrarily primed polymerase chain reaction. Methods. 2002, 27: 150-155.PubMedView ArticleGoogle Scholar
  69. Affymetrix support: [http://www.affymetrix.com/support/technical/manual/expression_manual.affx]
  70. Breese MR, Stephens MJ, McClintick JN, Grow MW, Edenberg HJ: Labrat LIMS: An extensible framework for developing laboratory information management, analysis, and bioinformatics solutions for microarrays [abstract]. SAC. 2003, 103-108.Google Scholar
  71. Gene Expression Omnibus: [http://www.ncbi.nlm.nih.gov/projects/geo/]
  72. McClintick JN, Edenberg HJ: Effects of filtering by present call on analysis of microarray experiments. BMC Bioinformatics. 2006, 7: 49-PubMedPubMed CentralView ArticleGoogle Scholar
  73. McClintick JN, Jerome RE, Nicholson CR, Crabb DW, Edenberg HJ: Reproducibility of oligonucleotide arrays using small samples. BMC Genomics. 2003, 4: 4-PubMedPubMed CentralView ArticleGoogle Scholar
  74. Benjamini Y, Hochberg Y: Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Roy Stat Soc B. 1995, 57: 289-300.Google Scholar
  75. Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA. 1998, 95: 14863-14868.PubMedPubMed CentralView ArticleGoogle Scholar
  76. NCBI BLAST: [http://www.ncbi.nlm.nih.gov/BLAST/]

Copyright

© Dannenberg and Edenberg; licensee BioMed Central Ltd. 2006

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Advertisement