Amyloid protein-mediated differential DNA methylation status regulates gene expression in Alzheimer’s disease model cell line

Biochemical and Biophysical Research Communications 414 (2011) 700–705

Contents lists available at SciVerse ScienceDirect

Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc

Amyloid protein-mediated differential DNA methylation status regulates gene expression in Alzheimer’s disease model cell line Hye Youn Sung a, Eun Nam Choi a, Sangmee Ahn Jo c, Seikwan Oh b, Jung-Hyuck Ahn a,⇑ a

Department of Biochemistry, School of Medicine, Ewha Womans University, 911-1 Mok-6-dong, Yangcheon-ku, Seoul 158-710, Republic of Korea Department of Neuroscience and TIDRC, School of Medicine, Ewha Womans University, 911-1 Mok-6-dong, Yangcheon-ku, Seoul 158-710, Republic of Korea c Department of Pharmacy, College of Pharmacy, Dankook University, San 29 Anseo-dong, Dongnam-gu, Cheonan-si, Chungnam 330-714, Republic of Korea b

a r t i c l e

i n f o

Article history: Received 20 September 2011 Available online 6 October 2011 Keywords: Alzheimer’s disease Amyloid precursor protein Swedish mutation DNA methylation

a b s t r a c t The Swedish mutation of amyloid precursor protein (APP-sw) has been reported to dramatically increase beta amyloid production through aberrant cleavage at the beta secretase site, causing early-onset Alzheimer’s disease (AD). DNA methylation has been reported to be associated with AD pathogenesis, but the underlying molecular mechanism of APP-sw-mediated epigenetic alterations in AD pathogenesis remains largely unknown. We analyzed genome-wide interplay between promoter CpG DNA methylation and gene expression in an APP-sw-expressing AD model cell line. To identify genes whose expression was regulated by DNA methylation status, we performed integrated analysis of CpG methylation and mRNA expression profiles, and identified three target genes of the APP-sw mutant; hypomethylated CTIF (CBP80/ CBP20-dependent translation initiation factor) and NXT2 (nuclear exporting factor 2), and hypermethylated DDR2 (discoidin domain receptor 2). Treatment with the demethylating agent 5-aza-20 -deoxycytidine restored mRNA expression of these three genes, implying methylation-dependent transcriptional regulation. The profound alteration in the methylation status was detected at the 435, 295, and 271 CpG sites of CTIF, and at the 505 to 341 region in the promoter of DDR2. In the promoter region of NXT2, only one CpG site located at 432 was differentially unmethylated in APP-sw cells. Thus, we demonstrated the effect of the APP-sw mutation on alteration of DNA methylation and subsequent gene expression. This epigenetic regulatory mechanism may contribute to the pathogenesis of AD. Ó 2011 Elsevier Inc. All rights reserved.

1. Introduction Alzheimer’s disease (AD) is the most common form of agedependent neurodegenerative disorders and is characterized by progressive memory loss and deterioration of cognitive ability. Extracellular deposition and aggregation of senile beta amyloid (Ab) plaques in the brains of patients are the hallmarks of AD. Ab is produced by abnormal sequential cleavage of amyloid precursor protein (APP) by b- and c-secretases [1]. The Swedish double mutant of APP (K670M/N671L) is one of the mutations leading to familial AD. The Swedish APP mutant leads to excess Ab production [2] and an early onset of AD [3]. Recently, global alterations in gene expression, including genes associated with AD pathogenesis, have Abbreviations: 5-aza-dC, 5-aza-20 -deoxycytidine; Ab, beta amyloid; AD, Alzheimer’s disease; APP, amyloid precursor protein; APP-sw, Swedish mutation of amyloid precursor protein; DMS, differentially methylated CpG site; MSP, methylation-specific PCR; qRT-PCR, quantitative real-time polymerase chain reaction; BSP, bisulfite sequencing PCR; CTIF, CBP80/CBP20-dependent translation initiation factor; NXT2, nuclear exporting factor 2; DDR, discoidin domain receptor; NBS, Nijmegen breakage syndrome protein. ⇑ Corresponding author. Fax: +82 2 2652 7846. E-mail address: [email protected] (J.-H. Ahn). 0006-291X/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2011.09.136

been demonstrated in neuroglial cells expressing APP-Swedish mutant [4]. Epigenetic regulation such as DNA methylation and histone modification has been suggested to be a mechanism leading to aberrant changes in gene expression during progression of AD [1]. DNA methylation is the most studied epigenetic mechanism that alters gene expression and is catalyzed by specific DNA methyltransferases that transfer a methyl group to the cytosine of CpG dinucleotides using S-adenosyl methionine as the methyl donor [5]. The methylation of CpG sites in the promoters of specific genes physically obstructs the binding of transcriptional factors and recruits methyl–CpG-binding domain proteins and histone deacetylases that are associated with gene silencing and formation of inactive heterochromatin. Thus, DNA methylation is an important mechanism of gene silencing and gene inactivation, and the methylation status at promoter CpG sites plays a pivotal role in gene regulation [6]. In several AD brain studies, the loss of DNA methylation has important consequences. Furthermore, elevated levels of S-adenosylhomocysteine, a potential inhibitor of methyltransferases, have been detected in AD brains and have an inverse correlation with patient cognition, suggesting involvement of epigenetic changes during the progression of AD [7]. DNA hypomethylation

H.Y. Sung et al. / Biochemical and Biophysical Research Communications 414 (2011) 700–705

in promoter CpGs of AD-associated genes such as APP, Presenilin 1 (PS1), and b-site APP-cleaving enzyme1 (BACE1) is also observed in AD brains [8,9] and results in abnormal upregulation of these genes, leading to excess accumulation of Ab. In addition, an increased level of methylation at specific loci in the promoter region of certain genes such as Apolipoprotein E and methylenetetrahydrofolatereductase has also been found in the cortex of AD brains compared to controls [10]. This suggests that simultaneous occurrence of hyper/hypomethylation may contribute to the pathogenesis of AD. In this study, we investigated aberrant genome-wide CpG methylation patterns in human neuroglioblastoma H4 cells harboring the Swedish APP mutation and compared them to patterns in wild-type H4 cells. We further integrated CpG methylation and mRNA expression profiles to identify genes whose expression was regulated by the DNA methylation-dependent epigenetic mechanism that may contribute to the pathogenesis of AD. 2. Materials and methods 2.1. Cell culture Human glioblastoma H4 cells and the APP695-Swedish mutant (H4-sw; K595N/M596L)-expressing H4-sw cell line were grown in Dulbecco’s modified Eagle medium (Gibco/BRL) containing 10% fetal bovine serum (Gibco/BRL), 100 U/mL penicillin (Gibco/BRL), 100 lg/mL streptomycin (Gibco/BRL), and 2 mM L-glutamine (Gibco/BRL). For maintaining H4-sw cells, 500 lg/mL geneticine (Gibco/BRL) was also added to the growth medium. 2.2. Global DNA methylation analysis using microarrays Genomic DNA from H4 and H4-sw cells was extracted using a Qiagen QIAamp DNA Mini kit (Qiagen) according to the manufacturer’s instructions. For genome-wide screening of DNA methylation, the Illumina HumanMethylation27 BeadChip (Illumina) targeting 27,578 specific CpG sites located within the proximal promoter regions of transcription start sites of 14,475 consensus coding sequence in the National Center for Biotechnology Information database was used according to the manufacturer’s recommendations. DNA methylation values were described as b-values, which were calculated by subtracting the background of negative controls on the array and taking the ratio of the methylated signal intensity to the sum of both methylated and unmethylated signals. The b-values ranged from 0 (completely unmethylated) to 1 (fully methylated) on a continuous scale for each CpG site. Forty-three targets with detection p-values > 0.05 were excluded from the 27,578 targets, and thus, a total of 27,535 CpG targets was used the final analysis. To identify differentially methylated CpG sites in H4 and H4-sw cell lines, we used the difference in the mean b-value (Db; mean b-value in H4-sw  mean b-value in H4). If the absolute difference in the mean b-values (|Db|) between the two cell lines was >0.06 (0.06 means the standard deviation at identical CpG sites), the site was defined as a differentially methylated CpG site (DMS). We described hypermethylated CpG sites/ genes as those with Db greater than 0.06, and hypomethylated CpG sites/genes as those with Db less than 0.06. 2.3. RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR) Total RNA was extracted using an RNeasy Plus Mini kit (Qiagen) according to the manufacturer’s protocol. One microgram of total RNA was converted to complementary DNA using Superscript II reverse transcriptase (Invitrogen) and oligo-(dT)12–18 primers (Invitrogen) according to the manufacturer’s directions. qRT-PCR was

701

performed in a 20-ll reaction mixture containing 1 ll complementary DNA, 10 ll SYBR Premix EX Taq (Takara Bio), 0.4 ll Rox reference dye (50, Takara Bio), and 200 nM primers for each gene. The primer sequences were as follows: CTIF (forward), 50 -CTCGTGTGCCCCATCTACAC-30 ; CTIF (reverse), 50 -TCATCTCAGGCGCTGTTCC-30 ; NXT2 (forward), 50 -AGGGCTGGATGCCCTAAATA-30 ; NXT2 (reverse), 50 -TTTGGGA CTGAGTTGCTTGC-30 ; DDR2 (forward), 50 -GGGAGAGTATCTTGCTGGGC-30 ; DDR2 (reverse), 50 -ACAGCTGGGAATAGGGCTGT-30 ; APP (forward), 50 -GCGAAA CTCATCTTCACTG-30 ; APP (reverse), 50 -GGCAACACACAAACTCTAC-30 ; GAPDH (forward), 50 -AATCCCATCACCATCTTCCA-30 ; GAPDH (reverse), 50 -TGGACTCCACGACGTACTCA-30 . Reactions were run at 50 °C for 2 min and 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min, followed by a dissociation stage of 1 cycle at 95 °C for 15 s, 60 °C for 20 s, and 95 °C for 15 s on an ABI PRISM 7000 sequence detection system (Applied BioSystems). All PCR reactions were performed in triplicate, and the specificity of the reaction was detected by melting curve analysis at the dissociation stage. Comparative quantitation of each target gene was performed based on the cycle threshold (CT), which was normalized to GAPDH using the DDCT method. 2.4. Methylation-specific PCR (MSP) and 5-aza-20 -deoxycytidine (5-aza-dC) treatment Bisulfite treatment of genomic DNA was performed using an EpiTect Bisulfite kit (Qiagen). qRT-MSP was carried out with bisulfitemodified genomic DNA as a template using specific primer sequences for the methylated or unmethylated forms of each gene. The following methylated/unmethylated-specific primers were used: CTIF (M forward, 50 -TGGTATTTAGTGTAGTATTAATTAC30 ; U forward, 50 -TGGTATTTAGTGTAGTATTAATTAT-30 ; internal control forward, 50 -GATGTTTAAATGTTAATTATAATATATG-30 ; reverse, 50 -AAAATACAAAATCTATACTAATAAC AA-30 ), NXT2 (M forward, 50 -TTTGAGAAGTTTATAGTGTAGTTAC-30 ; U forward, 50 -TTTGAGAAGTTTATAGTGTAGTTAT-30 ; internal control forward, 50 -GGAGGAAATAATGAGTTTAAATTTT-30 ; reverse, 50 -TATACTAATAAATCTCAACAACCTA-30 ), DDR2 (M forward, 50 -GTTAAAATATTAAAGGTTAGTTTCGC-30 ; U forward, 50 -GTTAAAATATTAAAGGTTAGTTTTGT-30 ; internal control forward, 50 -GAGA GTTTAGTTTTAGGTTTTTAT-30 ; reverse, 50 -CCTACCCCACACCTCATTTT-30 ). Quantitation of the methylated/unmethylated-specific CpG sites was performed by normalizing the values to those of the corresponding internal control, and the ratio of the value of each methylated/unmethylated-specific CpG site was calculated by dividing the normalized value by the sum of both the methylated and unmethylated values. For demethylating methylated CpG sites, H4 and H4-sw cells were treated with 10 lM 5-aza-dC (Sigma–Aldrich) for 3 days, and the medium was replaced daily. 2.5. Bisulfite sequencing PCR (BSP) BSP of the interesting promoter regions of CTIF, NXT2, and DDR2 was carried out in a 50-ll reaction mixture containing 10 ng bisulfite-modified genomic DNA, 1.5 mM MgCl2, 200 lM dNTPs, 1 U Platinum Taq polymerase (Invitrogen), 1 Platinum Taq buffer, and 200 nM specific BSP forward and reverse primers for each gene. The BSP primers were designed using MethPrimer software. For CTIF, the BSP product was 407 bp (position in the human GRCh37/hg19 assembly: ch18 46,064,840–46,065,246) and contained 21 CpGs. Sequences for the BSP primers are: 50 TTTTAGTTTTTAGTTTGGT-30 (forward) and 50 -TAACTACCAAAACAACCCAATCT-30 (reverse). For NXT2, the BSP product was 382 bp (position in the human GRCh37/hg19 assembly: chX 108,778,766–108,779,147) and contained 10 CpGs. Sequences for

702

H.Y. Sung et al. / Biochemical and Biophysical Research Communications 414 (2011) 700–705

the BSP primers are: 50 -TTGGGAGAATATAAAAGTTTG-30 (forward) and 50 -ATCTCCCTAAAACCAATAAC-30 (reverse). For DDR2, the BSP product was 531 bp (position in the human GRCh37/hg19 assembly: ch1 162,601,562–162,602,092) and contained seven CpGs. Sequences for the BSP primers are: 50 -TAATTTAGAAAGTTTTTATAAATGT-30 (forward) and 50 -AAACTATAAACAAAACTATAATTTA-30 (reverse). The reaction was run at 95 °C for 5 min, followed by 30 cycles at 95 °C for 30 s, 50–55 °C for 30 s, and 72 °C for 30 s, and then a final elongation step at 72 °C for 5 min. BSP products were purified using a QIAquick Gel Extraction kit (Qiagen) according to the manufacturer’s protocols and ligated into the yT&A cloning vector (Yeastern Biotech). The ligation products were used to transform competent DH5a Escherichia coli cells (RBC Bioscience) using standard procedures. BSP product-positive clones were confirmed with colony PCR using BSP primers to verify the size of the insert. Plasmid DNA was then extracted from at least 15 insert-positive clones using a QIAprep Spin Miniprep kit (Qiagen) and sequenced using the M13 primer to analyze the methylation status at specific CpG sites.

Table 2 List of candidate genes whose expression was regulated by the DNA methylation status in H4-sw cells. Hypomethylated/up-regulated genes

2.6. Beta amyloid (Ab) quantitation (enzyme-linked immunosorbent assay) Ab1–40 and Ab1–42 in the culture media were quantitated as described [11] using an enzyme-linked immunosorbent assay kit (Biosource International) according to the manufacturer’s instructions.

a b

Hypermethylated/down-regulated genes

Gene

Relative expressiona

Db b

Gene

Relative expressiona

Db b

HOXD4 FAM58A SCMH1 CTIF PCDHB15 DKK2 MRTO4 PDPN LHPP CRYAB PPP4C L3MBTL SLC38A3 OGDHL DNAJC7 WT1 SLC6A11 RERG RBM47 TMEM180 CYP26A1 COLEC10 NXT2

2.74 1.526 1.606 1.838 4.878 4.545 1.743 33.33 2.19 4.651 1.61 1.657 25 20.83 1.642 6.667 13.33 5.556 1.887 2.099 37.5 1.974 1.638

0.31 0.31 0.31 0.31 0.31 0.31 0.32 0.32 0.32 0.33 0.33 0.33 0.34 0.35 0.35 0.35 0.36 0.37 0.37 0.38 0.38 0.39 0.47

DDR2 CRMP1 ACAD11 QDPR PLAC8 DLK2

0.612 0.542 0.738 0.717 0.167 0.693

0.4 0.37 0.36 0.34 0.32 0.31

Expression fold change >1.5. |A0| > 0.3.

3. Results 3.1. Differential DNA methylation between H4 and H4-sw cell lines The Swedish double mutant-harboring H4 cell line (H4-sw) expresses around 2.6-fold higher levels of APP and secretes 1.2– 1.7 ng/mL Ab1–40 and 147–243 pg/mL Ab1–42. None of these were detected in the normal H4 cell line (Supplementary Fig. 1). To investigate epigenetic abnormalities in AD, we performed global DNA methylation profiling on normal and H4-sw cells using the human methylation 27 bead chip containing 27,578 specific CpG sites located within the proximal promoter regions of transcription start sites of 14,475 consensus coding sequence in the National Center for Biotechnology Information database. Among the total 27,535 analyzed CpG sites (43 targets were excluded because of a detection p-value > 0.05), 4937 sites were hypomethylated, and 1359 sites were hypermethylated in H4-sw cells compared to normal H4 cells (Table 1). The distribution of b-values at the differentially methylated CpG sites between the two cell lines was significantly lower in H4-sw cells than in H4 cells (Supplementary Fig. 2). 3.2. Selection of candidate genes whose expression was regulated by DNA methylation status To identify genes whose expression was regulated by epigenetic alteration in H4-sw cells expressing the APP Swedish mutations,

we integrated the global DNA methylation profiling data with previously reported transcriptome RNA sequencing data [4] using stringent selection criteria (|Db| > 0.3, expression fold change >1.5). Expression of candidate genes was considered to be upregulated (fold change >1.5) by hypomethylation (Db < 0.3) at promoter CpG sites and downregulated (fold change >1.5) by hypermethylation (Db > 0.3) at promoter CpG sites in H4-sw cells compared to normal H4 cells. The candidate genes are listed in Table 2.

3.3. Validation of CpG methylation and mRNA expression To validate the epigenetic regulation of the candidate genes, we confirmed their gene expression with qRT-PCR and their methylation status at the specific CpG sites, which are the same sites tested in the methylation microarray, with MSP (Fig. 1). Except for genes whose expression was detected after more than 35 cycles of the CT value, the transcriptional expression of three genes (CTIF, NXT2, DDR2) was more than 1.5-fold different between the two cell lines, in agreement with our mRNA transcriptome sequencing data (Fig. 1A). The specific CpG site of DDR2 (located at 420 from its transcription start site) was highly hypermethylated, and the CpG sites of CTIF (located at 435 from its transcription start site) and NXT2 (located at 432 from its transcription start site) were

Table 1 DNA hyper/hypomethylation patterns of H4-sw cells. Pattern

Hyperb Hypoc a b c

CpG sites no.

Genes no.

CpG site no.

Non-promoter CpGIa

Non-promoter CpG

Promoter CpGIa

Promoter CpG

Gene no.

Non-promoter CpGIa

Non-promoter CpG

Promoter CpGIa

Promoter CpG

1359 4937

341 973

199 1040

281 1495

515 1300

1266 4011

321 809

187 845

254 1224

485 1095

CpG island. Hypermethylated CpG site/gene. Hypomethylated CpG ste/gene.

703

H.Y. Sung et al. / Biochemical and Biophysical Research Communications 414 (2011) 700–705

A

5-aza dC treatment

20

H4 H4-sw

15

A

40

H4

-5-Aza +5-Aza

H4-SW

Relative expression

Relative expression

35 10 5 2.0 1.5 1.0

30 25 6 4 2

0.5 0

0.0

CTIF

0.8

NXT2

DDR2

NXT2

CTIF

NXT2

DDR2

DDR2

B methylation index

B

1.0

methylation index

CTIF

0.6 0.4 0.2

1.0

CTIF

NXT2

DDR2

0.8 0.6 0.4 0.2 0.0

0.0

-5AZA

H4

H4-sw

H4

H4-sw

H4

H4-sw

methylated unmethylated Fig. 1. mRNA expression and the DNA methylation status at the specific CpG sites of CTIF, NXT2, and DDR2. mRNA expression of each gene was measured with qRTPCR, (A) and the DNA methylation status at specific CpG sites was analyzed using quantitative MSP (B) Data are the means ± standard error of the mean of triplicate experiments. Statistical analyses were performed using a t-test (⁄p < 0.05).

significantly hypomethylated in H4-sw cells. After treatment of H4 cells with the demethylating agent 5-aza-dC, transcriptional expression of CTIF and NXT2 was dramatically increased (Fig. 2A) due to decreased methylation activity at the specific CpG residue of each gene (Fig. 2B). In addition, DDR2 expression was significantly upregulated (Fig. 2A), and the methylation activity at the specific CpG site of the gene was reduced in H4-sw cells (Fig. 2B) after treatment with 5-aza-dC. 3.4. Methylation status analysis of the identified genes in H4 and H4-sw cell lines The methylation status of CpG sites in the promoter region of the three genes was determined with BSP. The CpG sites in the 435 to 271 region in the CTIF promoter, especially the CpGs located at 435, 295, and 271 from the transcription site, were mainly demethylated in H4-sw cells (Fig. 3). In the promoter region of NXT2, only one CpG site, located at 432 from the transcription start site, was unmethylated in H4-sw cells compared to H4 cells, indicating that methylation of this specific CpG site is critical for epigenetic regulation of NXT2 (Fig. 3). Obvious aberrant hypermethylation was observed at the 505 to 314 region in the promoter of DDR2 in H4-sw cells (Fig. 3). 4. Discussion In this study, we performed genome-wide analysis of CpG methylation to examine the global epigenetic abnormality in AD

+5aZA

-5AZA

+5aZA

-5AZA

+5aZA

methylated unmethylated Fig. 2. Expression change of CTIF, NXT2, and DDR2 following demethylation in H4 and H4-sw cells. The cells were treated with 10 lM 5-aza-20 -deoxycytidine (5-azadC) for 3 days. Gene expression was measured with qRT-PCR, (A) and the DNA methylation status at the specific CpG sites was analyzed using quantitative MSP. (B) Data are the means ± standard error of the mean (n = 3).

using the AD model cell line H4-sw, which harbors the Swedish mutation. H4-sw cells produced high levels of the toxic form of Ab (Supplementary Fig. 1). In addition to secreting excess Ab, expression of genes associated with AD pathogenesis such as Presenilin 2, AQP1, glycogen synthase kinase 3, and cyclin-dependent kinase 5 is regulated by the Swedish mutation, implicating a link between this mutation and progression of AD pathogenesis [4]. Among the total 27,535 CpG sites we analyzed, 6296 CpG sites were differentially methylated in H4-sw cells relative to normal H4 cells. In particular, 23% of DMS were hypermethylated, and 78% of DMS were hypomethylated in H4-sw cells (Table 1). The distribution of the methylation b-values in H4-sw cells was clearly different than that in H4 cells, indicating that global DNA methylation was significantly reduced in H4-sw cells (Supplementary Fig. 2). Consistent with our results, Chen et al. have reported that a high concentration (approximately 25 lM) of Ab, especially Ab1–40, can itself induce global DNA hypomethylation in cerebral endothelial cell culture [12]. An association between global DNA hypomethylation and AD has been demonstrated in several AD brain studies. In the entorhinal cortex of patients with AD, global DNA hypomethylation was observed as assayed by immunoreactivity for 10 epigenetic markers and factors [13]. Another study has also shown epigenetic differences in cortical neurons from a pair of monozygotic twins that were discordant for AD. The twin suffering from AD showed significantly reduced levels of DNA methylation when compared to the non-demented twin [14]. DNA hypomethylation in promoter CpGs of AD-associated genes such as APP, PS1, and BACE1, which results in abnormal upregulation of these genes, was also observed in AD brains [8,9]. In our current study, how-

704

H.Y. Sung et al. / Biochemical and Biophysical Research Communications 414 (2011) 700–705

A

B

Fig. 3. Methylation status analysis of the promoter-associated CpG residues of CTIF, NXT2, and DDR2 in H4 and H4-sw cells. Methylation status analysis was determined with bisulfite sequencing PCR. The location of each CpG site is shown by the position from the transcription start site of each gene. Filled circles, >75% methylation; half-filled circles, 25–75% methylation; empty circles, <25% methylation (n = 15–20). The specific sites tested in the methylation array are marked with triangles, and differentially methylated sites between H4 and H4-sw cells are in boxes.

ever, differences in the methylation status at the promoter CpG sites of those genes were not observed in the genome-wide methylation analysis of the two cell lines, H4 and H4-sw. To identify the genes whose expression was regulated by the promoter-related CpG methylation status in H4-sw cells, we integrated the global DNA methylation profiling data with the previously reported transcriptome sequencing data [4]. The candidate genes were selected using stringent selection criteria (|Db| > 0.3, expression fold change >1.5) as listed in Table 2. After validating the integrated data with qRT-PCR and MSP at the specific promoter CpG sites, we finally selected three potential genes, two that were hypomethylated (CTIF, NXT2) and one that was hypermethylated (DDR2) in H4-sw cells compared to H4 cells (Fig. 1). The results of treatment with the demethylating agent 5-aza-dC revealed that alteration of expression of these genes in H4-sw cells was clearly associated with methylation-dependent transcriptional regulation (Fig. 2). According to the BSP map of H4-sw cells, partially differentially demethylated patterns relative to H4 cells were observed in CTIF, whereas almost all CpG sites in the analyzed promoter region of

DDR2 were hypomethylated (Fig. 3). Interestingly, only one CpG site located at 432 from the transcription site was hypomethylated in NXT2 (Fig. 3), providing evidence that the methylation status of this specific CpG site is critical for regulation of expression of NXT2. CTIF is a translation initiation factor containing a MIF4G domain [15]. This protein directly interacts with CBP80 and forms the CBP80/CBP20 translation initiation complex, which is involved in CBP80/CBP20-mediated mRNA translation [15]. NXT2 is a member of the family of nuclear exporting factors and is likely homologous to another nuclear exporting factor protein, NXT1, which is involved in exporting nuclear RNA by forming heterodimers with NXF1 (nuclear export factor 1) through an NTF2like domain in eukaryotes [16]. The biological function of NXT2 is unknown in vertebrates. Only one functional study has demonstrated that a splice variant of Nxt2 is required for myocardial cell differentiation and formation of the cardiac valve at the atrioventricular boundary in zebrafish [17]. In yeast, Nxt2 co-precipitates with Mog1, implying involvement in nuclear transport, RNA metabolism, and signal transduction [18]. In another study, global

H.Y. Sung et al. / Biochemical and Biophysical Research Communications 414 (2011) 700–705

gene expression analysis showed highly upregulated NXT2 expression in Nijmegen breakage syndrome (NBS) protein-deficient cells compared to normal NBS cells, suggesting that NBS controls the expression of NXT2 [19]. A recent neuronal study demonstrated an important role for NBS in neuronal proliferation, survival, and differentiation [20]. However, the biological function of NTX2 in neurons and the brain has not been clarified. DDR2 is one of two discoidin domain receptors (DDR1 and DDR2), which are unusual receptor tyrosine kinases that contain a discoidin homology region in their extracellular domain and respond to collagen rather than growth factors [21]. DDR2 is mainly expressed in vascular and mesenchymal cells such as fibroblasts and smooth muscle cells and is implicated in many fundamental cellular processes. Recently, an association between DDR2 and the inflammatory response was reported, showing that DDR2 mediates the collagen-dependent release of cytokines through various signaling pathways such as mitogen-activated protein kinase and nuclear factor-kappa B [22,23]. These findings suggest that DDR2 may contribute to the pathogenesis of AD, which is consistent with a previous study showing epigenetic modification of several inflammatory-related genes (iNOS, IL-1, and TNFa) in AD brains [24]. In our current study, we clearly showed that the expression of CTIF, NXT2, and DDR2 was aberrantly regulated by methylationdependent epigenetic control in an AD model cell line. Further studies are needed to characterize the functional role of these genes and the molecular pathways involved in the pathogenesis of AD. Here, we provide evidence that progression of AD may be associated with global epigenetic dysregulation. Using integrated analysis of CpG methylation and mRNA expression profiles, we identified three methylation-regulated genes, CTIF, NXT2, and DDR2, which may be implicated in the pathogenesis of AD. Acknowledgments This research was supported by RP-Grant 2010 of Ewha Womans University and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MEST) (MRC, 2010-0029355). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbrc.2011.09.136. References [1] P. Seubert, T. Oltersdorf, M.G. Lee, R. Barbour, C. Blomquist, D.L. Davis, K. Bryant, L.C. Fritz, D. Galasko, L.J. Thal, et al., Secretion of beta-amyloid precursor protein cleaved at the amino terminus of the beta-amyloid peptide, Nature 361 (1993) 260–263. [2] Y.H. Suh, F. Checler, Amyloid precursor protein, presenilins, and alphasynuclein: molecular pathogenesis and pharmacological applications in Alzheimer’s disease, Pharmacological Reviews 54 (2002) 469–525. [3] J.C. Lambert, F. Wavrant-De Vrieze, P. Amouyel, M.C. Chartier-Harlin, Association at LRP gene locus with sporadic late-onset Alzheimer’s disease, Lancet 351 (1998) 1787–1788.

705

[4] J. Shin, S.B. Yu, U.Y. Yu, S.A. Jo, J.H. Ahn, Swedish mutation within amyloid precursor protein modulates global gene expression towards the pathogenesis of Alzheimer’s disease, BMB Reports 43 (2010) 704–709. [5] A. Bird, Does DNA methylation control transposition of selfish elements in the germline?, Trends in Genetics 13 (1997) 469–472 [6] M.M. Suzuki, A. Bird, DNA methylation landscapes: provocative insights from epigenomics, Nature Reviews Genetics 9 (2008) 465–476. [7] B.P. Kennedy, T. Bottiglieri, E. Arning, M.G. Ziegler, L.A. Hansen, E. Masliah, Elevated S-adenosylhomocysteine in Alzheimer brain: influence on methyltransferases and cognitive function, Journal of Neural Transmission 111 (2004) 547–567. [8] R.L. West, J.M. Lee, L.E. Maroun, Hypomethylation of the amyloid precursor protein gene in the brain of an Alzheimer’s disease patient, Journal of Molecular Neuroscience 6 (1995) 141–146. [9] A. Fuso, L. Seminara, R.A. Cavallaro, F. D’Anselmi, S. Scarpa, Sadenosylmethionine/homocysteine cycle alterations modify DNA methylation status with consequent deregulation of PS1 and BACE and betaamyloid production, Molecular and Cellular Neurosciences 28 (2005) 195–204. [10] S.C. Wang, B. Oelze, A. Schumacher, Age-specific epigenetic drift in late-onset Alzheimer’s disease, PLoS ONE 3 (2008) e2698. [11] M.J. Kim, S.S. Chae, Y.H. Koh, S.K. Lee, S.A. Jo, Glutamate carboxypeptidase II: an amyloid peptide-degrading enzyme with physiological function in the brain, The FASEB Journal 24 (2010) 4491–4502. [12] K.L. Chen, S.S. Wang, Y.Y. Yang, R.Y. Yuan, R.M. Chen, C.J. Hu, The epigenetic effects of amyloid-beta(1–40) on global DNA and neprilysin genes in murine cerebral endothelial cells, Biochemical and Biophysical Research Communications 378 (2009) 57–61. [13] D. Mastroeni, A. Grover, E. Delvaux, C. Whiteside, P.D. Coleman, J. Rogers, Epigenetic changes in Alzheimer’s disease: decrements in DNA methylation, Neurobiology of Aging 31 (2010) 2025–2037. [14] D. Mastroeni, A. McKee, A. Grover, J. Rogers, P.D. Coleman, Epigenetic differences in cortical neurons from a pair of monozygotic twins discordant for Alzheimer’s disease, PLoS ONE 4 (2009) e6617. [15] K.M. Kim, H. Cho, K. Choi, J. Kim, B.W. Kim, Y.G. Ko, S.K. Jang, Y.K. Kim, A new MIF4G domain-containing protein, CTIF, directs nuclear cap-binding protein CBP80/20-dependent translation, Genes and Development 23 (2009) 2033– 2045. [16] A. Herold, M. Suyama, J.P. Rodrigues, I.C. Braun, U. Kutay, M. Carmo-Fonseca, P. Bork, E. Izaurralde, TAP (NXF1) belongs to a multigene family of putative RNA export factors with a conserved modular architecture, Molecular and Cellular Biology 20 (2000) 8996–9008. [17] H. Huang, B. Zhang, P.A. Hartenstein, J.N. Chen, S. Lin, NXT2 is required for embryonic heart development in zebrafish, BMC Developmental Biology 5 (2005) 7. [18] M. Oki, L. Ma, Y. Wang, A. Hatanaka, C. Miyazato, K. Tatebayashi, H. Nishitani, H. Uchida, T. Nishimoto, Identification of novel suppressors for Mog1 implies its involvement in RNA metabolism, lipid metabolism and signal transduction, Gene 400 (2007) 114–121. [19] E.R. Jang, J.H. Lee, D.S. Lim, J.S. Lee, Analysis of ataxia-telangiectasia mutated (ATM)- and Nijmegen breakage syndrome (NBS)-regulated gene expression patterns, Journal of Cancer Research and Clinical Oncology 130 (2004) 225– 234. [20] W.T. Lee, W.H. Chang, C.H. Huang, K.J. Wu, NBS1, the Nijmegen breakage syndrome gene product, regulates neuronal proliferation and differentiation, Journal of Neurochemistry 102 (2007) 141–152. [21] F. Alves, W. Vogel, K. Mossie, B. Millauer, H. Hofler, A. Ullrich, Distinct structural characteristics of discoidin I subfamily receptor tyrosine kinases and complementary expression in human cancer, Oncogene 10 (1995) 609–618. [22] T.K. Mao, Y. Kimura, T.P. Kenny, A. Branchi, R.G. Gishi, J. Van de Water, H.J. Kung, S.L. Friedman, M.E. Gershwin, Elevated expression of tyrosine kinase DDR2 in primary biliary cirrhosis, Autoimmunity 35 (2002) 521–529. [23] A.R. Klatt, D. Zech, G. Kuhn, B. Paul-Klausch, G. Klinger, J.H. Renno, J. Schmidt, G. Malchau, K. Wielckens, Discoidin domain receptor 2 mediates the collagen II-dependent release of interleukin-6 in primary human chondrocytes, Journal of Pathology 218 (2009) 241–247. [24] H. Akiyama, S. Barger, S. Barnum, B. Bradt, J. Bauer, G.M. Cole, N.R. Cooper, P. Eikelenboom, M. Emmerling, B.L. Fiebich, C.E. Finch, S. Frautschy, W.S. Griffin, H. Hampel, M. Hull, G. Landreth, L. Lue, R. Mrak, I.R. Mackenzie, P.L. McGeer, M.K. O’Banion, J. Pachter, G. Pasinetti, C. Plata-Salaman, J. Rogers, R. Rydel, Y. Shen, W. Streit, R. Strohmeyer, I. Tooyoma, F.L. Van Muiswinkel, R. Veerhuis, D. Walker, S. Webster, B. Wegrzyniak, G. Wenk, T. Wyss-Coray, Inflammation and Alzheimer’s disease, Neurobiology of Aging 21 (2000) 383–421.