Skip Navigation


Human Molecular Genetics Advance Access originally published online on July 26, 2006
Human Molecular Genetics 2006 15(17):2623-2635; doi:10.1093/hmg/ddl188
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrowOA All Versions of this Article:
15/17/2623    most recent
ddl188v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (7)
Google Scholar
Right arrow Articles by Shen, Y.
Right arrow Articles by Fan, G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shen, Y.
Right arrow Articles by Fan, G.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 2006 The Author(s)
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abnormal CpG island methylation occurs during in vitro differentiation of human embryonic stem cells

Yin Shen1, Janet Chow1, Zunde Wang2,{dagger} and Guoping Fan1,*

1 Department of Human Genetics, Institute of Stem Cell Biology and Medicine, David Geffen School of Medicine, UCLA, 695 Charles Young Drive South, Los Angeles, CA 90095, USA and 2 EpiGenX Pharmaceuticals Inc., 5385 Hollister Avenue, Santa Barbara, CA 93111, USA

* To whom correspondence should be addressed. Tel: +1 3102670439; Fax: +1 3107945446; Email: gfan{at}mednet.ucla.edu

Received June 7, 2006; Accepted July 21, 2006


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Directed differentiation of human embryonic stem cells (hESCs) into specific somatic cells holds great promise for cell replacement therapies. However, it is unclear if in vitro hESC differentiation causes any epigenetic abnormality such as hypermethylation of CpG islands. Using a differential methylation hybridization method, we identified 65 CpG islands (out of 4608 CpG islands or 1.4%) that exhibited increased DNA methylation during the conversion of hESCs into neural progenitor/stem cells (NPCs). These methylated CpG islands belong to genes in cell metabolism, signal transduction and cell differentiation, which are distinctively different from oncogenic CpG island hypermethylation observed in cancer-related genes during tumorigenesis. We further determined that methylation in these CpG islands, which is probably triggered by de novo DNA methyltransferase Dnmt3a, is abnormally higher in hESC–NPCs than in primary NPCs and astrocytes. Correlating with hypermethylation in promoter CpG islands of metabolic enzyme gene CPT1A and axoneme apparatus gene SPAG6, levels of CPT1A and SPAG6 mRNAs are significantly reduced in hESC–NPCs when compared with hESCs or primary neural cells. Because CPT1A is involved in lipid metabolism and CPT1A deficiency in human is associated with the hypoketotic hypoglycemia disorder, the reduced CPT1A expression in hESC–NPCs raises a potential concern for the suitability of these cells in cell transplantation. Collectively, our data show that abnormal CpG island methylation takes place in a subset of genes during the differentiation/expansion of hESC derivatives under current culture conditions, which may need to be monitored and corrected in future cell transplantation studies.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
DNA cytosine methylation is one of the major epigenetic factors in vertebrate animals that is involved in gene regulation, genomic imprinting, X-chromosome inactivation and genome stability (1). During embryonic development, DNA methylation is established by the coordinated actions of a family of DNA (cytosine-5) methyltransferases (Dnmts) including the maintenance enzyme Dnmt1 and de novo DNA methyltransferases Dnmt3a and Dnmt3b (2). DNA methylation is required for animal development because mutant mice lacking either the maintenance enzyme (Dnmt1) or both of the de novo DNA methyltransferases (Dnmt3a/3b) exhibit significant demethylation in the genome and die at embryonic day (E) 8–10 just after gastrulation (3,4). Alterations in DNA methylation have been associated with cancer and human genetic disorders (5). For example, DNMT3B deficiency in humans leads to significant demethylation in centromeric minor satellite repeats, and patients suffer a rare genetic disorder, namely ICF syndrome (Immunodeficiency, Centromere instability and Facial anomalies) (4,6). DNA methylation deficiency in mouse neural precursor/stem cells induces a precocious astrogliogenesis phenotype (7,8), suggesting that proper control of DNA methylation is essential for the cell differentiation program. In somatic cell nuclear transfer experiments, reprogramming the DNA methylation pattern is one of the essential epigenetic changes and its faulty reprogramming in cloned embryos is postulated to be one of the leading causes for the very low efficiency of animal cloning (9).

Pluripotent human embryonic stem cells (hESCs) have the capacity to differentiate into multi-lineage cell types and the potential to be a major source of donor cells for cell transplantation and tissue repair (10,11). For example, hESC-derived oligodendrocyte progenitor cells have been shown to remyelinate and restore locomotion after spinal cord injury in rodents (12). However, recent work suggests that genetic lesions and epigenetic instability including abnormal DNA methylation in cancer-related genes occur during long-term passage of hESCs (13). This raises concerns about whether in vitro differentiation of hESCs may also lead to abnormalities in DNA methylation patterns in comparison with normal somatic cells. In this study, we used a DNA microarray-based method to first assess if changes in DNA methylation occur in a library of 4608 CpG islands during the conversion of undifferentiated hESCs into neural progenitor/stem cells (NPCs). The majority of CpG islands are associated with the gene promoter/first exon and are normally protected from DNA methylation during embryogenesis (14–16). Our results indicate that only a small fraction of CpG islands are subject to de novo DNA methylation upon hESC differentiation. However, this subset of CpG islands is distinctively different from CpG islands methylated in cancer cells (17). Furthermore, levels of DNA methylation in these selected CpG islands in hESC-derived NPCs (hESC–NPCs) become abnormally high when compared with levels in primary human cells, leading to the silencing of these genes in hESC–NPCs.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
In vitro neural differentiation of hESCs
We have adopted a culture protocol for directed neural differentiation of hESCs, which has been successfully used to derive homogenous populations of NPCs from mouse and human ESCs (18,19) (Fig. 1A–I, see Materials and Methods). Through consecutive passages in serum-free DMEM/F12 medium with B27 supplement and bFGF treatment, homogenous populations of NPCs can be obtained from the third passage of NPCs (P3). Using antibodies against neuronal marker ßIII TUBULIN (TuJ1 antigen), astroglial marker glial fibrillary acidic protein (GFAP) and neural stem cell marker NESTIN (7), we found that P3 NPC cultures contained over 95% NESTIN- and SOX2-positive NPCs and a small number of spontaneously differentiated ßIII TUBULIN-positive neurons (Fig. 1E). In late passage, hESC-derived NPCs (≥P8), while virtually all cells are NESTIN- and SOX2-positive NPCs (Fig. 1F and I), a small number of cells (5–10%) are also GFAP-positive (data not shown), suggesting the presence of a small percentage of glial progenitors in these late passage hESC–NPCs. Upon withdrawal of bFGF treatment and treatment of neurotrophins, ~50% of P3 hESC–NPCs differentiate into post-mitotic neurons within a week. In contrast, most late passage hESC–NPCs are prone to differentiate into mixed population of neurons and astrocytes (Supplementary Material, Fig. S1), suggesting a switch of neurogenic to gliogenic competence over the time (20). The hESC–NPCs after long-term passaging (>P20) maintain a normal karyotype as assayed by standard G-band staining of metaphase chromosome spreads (data not shown).


Figure 1881
View larger version (97K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. Characterization of hESCs and hESC-derived NPCs. (A–C) Undifferentiated hESCs in phase-contrast microscopic graph (A), or stained with antibodies against OCT4 (B) and SSEA-4 (C). (D–F) show immunostaining with NESTIN (NPC marker in red) and TuJ1 (neuronal marker in green) antibodies of hESCs (D) and P3 (E) and P14 (F) hESC–NPCs. The dotted line in (D) outlines a hESC colony that exhibits moderate NESTIN staining signals when compared with the strong staining in hESC–NPCs (E and F). Mouse MEF feeder cells also have a background staining of NESTIN, which is much weaker than hESC–NPCs. (G–I) SOX2 staining of undifferentiated hESCs and hESC–NPCs. (H) and (I) also show colocalization of SOX2 and NESTIN in hESC–NPCs. DAPI counterstaining (in blue) shows nuclear morphology.

 
Selective increase in CpG island methylation
We have used a modified method of differential DNA methylation hybridization (DMH) (21) to identify potential methylation changes in genomic DNA harvested from hESCs and P3 and P14 hESC–NPCs. Genomic DNA was cut with methyl-sensitive SmaI restriction enzyme at the site CCC{downarrow}GGG that is preferentially localized within CpG islands (22) and ligated to blunt adaptors for ligation-mediated PCR amplification. The amplified DNA fragments from hESCs, and P3 and P14 hESC–NPCs were then labeled with either Cy3 or Cy5 fluorescent dye and cross-hybridized to a custom-made DNA microarray containing 4608 CpG island fragments to identify differentially hybridized spots that resulted from different levels of DNA methylation. This CpG island library was constructed with presumably unmethylated CpG islands purified from SmaI-cut DNA fragments of adult human leukocyte DNA and a number of known CpG islands that become hypermethylated in cancer cells (Z.W. et al. in preparation, see Materials and Methods). Our first analysis of the microarray data show that there are two distinct groups of hybridized spots on each CpG island microarray among three pairs of comparisons (hESCs versus P3 hESC–NPCs, hESCs versus P14 hESC–NPCs and P3 hESC–NPCs versus P14 hESC–NPCs). In the first group of hybridized spots, the signal intensity is biased towards probes from DNA samples of hESCs and P3 hESC–NPCs when compared with P14 hESC–NPCs, suggesting that these CpG island clones could be genomic regions with increased DNA methylation in P14 hESC–NPCs. The second group of hybridized spots, representing the majority of hybridized signals with no significant difference or biased hybridization towards P14 hESC–NPCs. We experimentally checked methylation status of several clones in this second group and confirmed that these CpG clones did not exhibit changes in DNA methylation between hESCs and P14 hESC–NPCs (data not shown).

Among the first group of hybridized clones that exhibited stronger hybridized signals in hESCs samples (thus suggesting increased DNA methylation in differentiated hESC–NPCs), we first set up a 1.5-fold cut-off (hESC–NPCs/hESCs ratio <0.667) and identified 45 candidate clones. By sampling six microarray data sets between hESCs and P14 hESC–NPCs, or P3 and P14 hESC–NPCs, we further identified additional 20 clones that consistently showed a ratio <0.90 in all the arrays. By combining these two sets of clones together, we estimated that 1.4% of clones (or 65 out of 4608 probe sets on the CpG island microarray) that presumably have higher levels of DNA methylation in P14 hESC–NPCs than in undifferentiated hESCs or early passage P3 hESC–NPCs. It is worth noting that our current result is solely based on the SmaI site-based DMH method, which cannot detect increased methylation in CpG islands lacking the SmaI restriction enzyme site (CCCGGG). Thus, our result could be potentially an under-estimate of CpG island methylation across the genome during hESC differentiation.

To identify the nature of these genomic clones that exhibited increased DNA methylation during neural differentiation of hESCs and passages of hESC–NPCs, we randomly sequenced 56 clones and blasted them against the human genome sequence database. As annotated in Table 1, we found that 91.1% of clones (51/56) are located in designated CpG island regions (designated as GC%>50%, >200 bp and the ratio of observed/expected CpG >0.6) (23), confirming the high quality of this CpG island library. Among these CpG island clones, ~51% (26/51) are localized at promoter regions, 25% in exon and/or intron regions and the remaining 24% in intergenic regions or without any hit against the human sequence database (UCSC hg17 freeze May 2004). Gene ontology analysis of these annotated CpG islands (Table 1) suggests that these genes are involved in cell metabolism, signal transduction and cell differentiation and development. Among them are metabolic enzymes (e.g. carnitine palmitoyltransferase 1A or CPT1A), transcription regulators (e.g. zinc finger protein 451) and signaling molecules (e.g. BMP4 and brain adenylate cyclase 1), which could be important for the survival and differentiation of hESC–NPCs. To confirm the quality of DMH microarray results, we randomly picked 11 CpG islands and experimentally verified the increase in CpG island methylation in all these CpG islands in hESC–NPCs (Fig. 2 and Supplementary Material, Fig. S2). Bisulfite sequencing analysis (24) and COBRA assay (25) were used to quantify levels of DNA methylation of selected CpG sites in three gene promoters [CPT1A, ARMADILLO REPEAT CONTAINING 7 (ARMC7) and SPERM-ASSOCIATED ANTIGEN 6 ISOFORM 1 (SPAG6)] and in two coding exon regions [PROTOCADHERIN 17 (PCDH17) and STATHMIN 3 (STMN3)] (Fig. 2A–D). The CPT1A gene actually contains two alternative gene promoters: the exon1a promoter contains a 1.2 kb CpG island that was identified in our DMH array, and the exon1b promoter contains a 2.5 kb CpG island that is ~1.5 kb downstream of exon1a. We, therefore, analyzed if an increase in methylation in the CpG island occurs in both exon1a and 1b promoters (Fig. 2A) with the aim to see if methylation is regulated in a coordinated fashion on these two alternative CPT1A gene promoters. SPAG6 and ARMC7 gene promoters contain a single CpG island in the proximal gene promoter/first exon region (Fig. 2B and D). Our results confirmed that a significant increase in DNA methylation, albeit at various degrees, occurs in all these CpG islands during the conversion of hESCs into early passages of hESC–NPCs (e.g. P3). Moreover, all CpG islands reach the highest level of DNA methylation in late passage P14 hESC–NPCs, suggesting that levels of DNA methylation are further elevated during the prolonged passaging of hESC–NPCs.


Figure 1882
View larger version (47K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2. Bisulfite sequencing analysis and COBRA assay of selective CpG island methylation in hESCs, hESC–NPCs, primary NPCs and astrocytes, and normal cortex and leukocytes. Methylation patterns in CPT1A exon1a, exon1b (A), SPAG6 (B) gene promoters and PCDH17 (C) exon CpG island. Each filled black dot represents one methylated CpG site and an open dot an unmethylated CpG. Schematic diagram of promoter and CpG sites in vertical bars are presented for each gene. Arrows indicate the transcription initiation site of human promoters examined. CpG sites in the regions analyzed by bisulfite genomic sequencing are underlined by black bars. (D) COBRA methylation assay of DNA methylation in ARMC7 gene promoter and STMN3 exon/intron region. The CpG site analyzed is indicated by arrow.

 


View this table:
[in this window]
[in a new window]

  		Table 1. Candidate CpG islands that exhibit increased DNA methylation during hESC differentiation and NPC passage

 
Selective DNA hypermethylation in hESC–NPCs in comparison to primary human NPCs, astrocytes and leukocytes
To determine if methylation patterns of these CpG islands in hESC–NPCs (Table 1) are unique to hESC–NPCs or simply reflect neural cell-type specific methylation that could occur in primary neural cells during embryogenesis, we performed extensive bisulfite methylation analysis in four representative CpG islands within CPT1A and SPAG6 gene promoters and the PCDH17 exon region in hESC–NPCs and primary NPCs and astrocytes derived from human fetal brain (Fig. 2A–C). We also analyzed adult leukocytes which represent a compeletely different cell lineage. If CpG island methylation in hESC–NPCs belongs to cell-type specific methylation, we would expect to see a similar degree of CpG island methylation in hESC–NPCs when compared with primary NPCs and astrocytes. Overall, these select CpG islands remain unmethylated in cultured primary NPCs and astrocytes, as well as in fetal cortical tissue and adult leukocytes. Our results suggest that higher-than-normal levels of DNA methylation in these select CpG islands in hESC–NPCs represent an accumulation of aberrant DNA hypermethylation during the conversion of hESCs into NPCs and the expansion of hESC–NPCs in culture.

Hypermethylated CpG islands in hESC–NPCs are distinctively different from those in the promoter of cancer-related genes
To further determine whether hypermethylation in select gene loci is unique to hESC–NPCs or shares the same characteristics as hypermethylation found in cancer cells, we compared our list of hypermethylated loci to the list of known hypermethylated genes as reported for cancer cells (17). Abnormal methylation in tumor suppressor genes or other cell-cycle control genes is one of the major causes underlying cancer formation (26,27). For example, a recent comprehensive genome-wide search for hypermethylated CpG islands in colon cancer cell line Caco-2 and prostate cancer cell line PC-3 as well as colon tumor samples has identified a total of 367 hypermethylated CpG islands (17). Among the 26 methylated CpG island promoters identified in our study (Table 1), there is no overlap with any of the 367 CpG island promoters reported for cancer cells (17). Our data suggests that the population of methylated CpG islands in hESC–NPCs is distinctively different from that found in cancer cells. We further confirmed the unmethylated status of five CpG islands in promoter regions of cancer-related genes (VHL, THBS1, ERBIN, TMEFF2 and P16). As shown in Supplementary Material, Figure S3, CpG islands in these cancer-related genes (VHL, THBS1, ERBIN, TMEFF2 and P16) do not exhibit any increase in DNA methylation in hESC–NPCs. Furthermore, we also found that undifferentiated hESCs maintain unmethylated CpG islands in either CPT1A, SPAG6, PCDH17 genes, or these five cancer-related genes (VHL, THBS1, ERBIN, TMEFF2 and P16) at moderate passage P54 and extremely late passage P170 (data not shown), suggesting that this undifferentiated hESC line (HSF6) is epigenetically stable at these gene loci in our culture conditions.

DNA hypermethylation and gene silencing
To understand the significance of hypermethylation of CpG islands on gene promoter activity during hESC differentiation, we performed semi-quantitative RT–PCR analysis of CPT1A and SPAG6 transcripts. CPT1A is a mitochondrial protein involved in fatty acid metabolism, and mutation of this gene causes recurrent attacks of hypoketotic hypoglycemia (28,29). SPAG6 gene is a component of the axoneme central apparatus in sperm cells and may also be involved in the motility of ependymal cilia in the nervous system (30,31). We found that CPT1A and SPAG6 mRNA levels are high in hESCs, but significantly decreased or below detectable levels in P3 and P14 hESC–NPCs (Fig. 3A). This result is consistent with the notion that methylation of CpG islands in the promoter causes gene silencing. In contrast, when CpG methylation occurs in the coding exon of PCDH17, there is no correlation between the levels of CpG methylation and the level of PCDH17 mRNA (data not shown).


Figure 1883
View larger version (31K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3. Semi-quantitative RT–PCR analysis of mRNA levels of CPT1A and SPAG6 genes during hESC differentiation. (A) Down-regulation of mRNA transcripts from CPT1A exon1a and exon1b and SPAG6 mRNAs upon hESC differentiation and prolonged passaging of hESC–NPCs. Exon-specific primers for CPT1A gene are labeled next to the gel picture. (B) Comparison of CPT1A and SPAG6 mRNA levels among hESCs, hESC–NPCs, primary NPCs and astrocytes. (C) Treatment of AzaC (1 µM) to late passages of hESC–NPCs (P8–10) for 6 days led to an increase in CPT1A gene expression. Ratio of densitometry reading of CPT1A transcripts over ß-ACTIN is shown in bar graph to the right with P-value <0.01 (unpaired t-test).

 
Neural cells in the brain express both CPT1A and CPT1C isoforms (32). Indeed, we found that both CPT1A and CPT1C mRNA are detected in human fetal NPCs and astrocytes by RT–PCR (Fig. 3B and Supplementary Material, Fig. S4). In addition, the level of CPT1C isoform is significantly increased in hESC–NPCs, consistent with its abundant expression in differentiated neural lineage cells (32, Supplementary Material, Fig. S4). In contrast, the level of CPT1A in late passage P14 hESC–NPCs is significantly lower than in normal fetal NPCs and astrocytes (Fig. 3B), consistent with the possibility that DNA hypermethylation in CPT1A promoters could result in the imbalanced expression of CPT1 isoforms in hESC–NPCs. To test whether the inhibition of CPT1A gene expression in late passage hESC–NPCs can be reversed in culture, we treated late passage hESC–NPCs (P8–10) with the demethylating agent 5-azacytidine (AzaC) for 6 days in culture. AzaC treatment significantly increased mRNA levels of CPT1A, suggesting that DNA methylation-mediated gene silencing of CPT1A can be prevented in late passage hESC–NPCs by treating cultures with a DNA methyltransferase inhibitor such as AzaC (Fig. 3C).

Potential involvement of Dnmt3a with de novo methylation in CpG islands
We next examined which DNA methyltransferases carry out de novo DNA methylation activity in CpG islands during the conversion of hESCs into NPCs, and further passaging of hESC–NPCs. We first determined the expression of DNMT3A, DNMT3B and DNMT1 in hESCs and P3 and P14 hESC–NPCs by immunocytochemistry and RT–PCR analysis. DNMT1 mRNA is expressed at comparable levels in hESCs, and P3 and P14 hESC–NPCs (data not shown), consistent with its maintenance role in DNA methylation. Both DNMT3A and DNMT3B are highly expressed in hESCs, raising the possibility that both enzymes may be involved in de novo methylation of CpG islands during the conversion of hESCs into early passage hESC–NPCs. However, immunocytochemistry shows that DNMT3B protein is dramatically down-regulated in P3 and P14 hESC–NPCs (Fig. 4A), whereas DNMT3A enzyme is expressed at a similar level during all stages of cell differentiation. Quantification of DNMT3A and DNMT3B mRNAs levels by real-time RT–PCR shows a similar pattern for the stage-specific expression of DNMT3A and DNMT3B in hESCs, and P3 and P14 hESC–NPCs (Fig. 4B). In fact, DNMT3A mRNA level is transiently increased in P3 hESC–NPCs (Fig. 4B) and maintained at levels in P7–10 or P14 hESC–NPCs comparable to those in hESCs or in fetal astrocytes, respectively. The expression analysis of DNMTs suggests that DNMT3A could be the enzyme mediating the increase in DNA methylation in CpG islands in late passage hESC–NPCs. To determine if DNMT3A is directly associated with those CpG islands, we performed chromatin immunoprecipitation (ChIP) experiments with a rabbit antibody against DNMT3A in P3 and P14 hESC–NPCs. ChIP assays showed that gene promoters of CPT1A (both exon1a and exon1b promoters) and SPAG6 are associated with DNMT3A (Fig. 5). In contrast, there is no DNMT3A association with the ERBIN gene promoter, which is unmethylated throughout the entire period of cell culture. Taken together, our data suggests that selective targeting of DNMT3A to specific gene promoters could be a leading mechanism underlying the hypermethylation of these CpG island promoters during the prolonged culture of NPCs.


Figure 1884
View larger version (40K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 4. Analysis of mRNA and protein levels of DNMT3A and DNMT3B genes during hESC differentiation. (A) Immunostaining of hESCs, and P3 and P14 hESC–NPCs with antibodies against DNMT3A (red) and DNMT3B (green). DAPI counter-staining shows nuclear morphology. (B) Real-time PCR assay for expression levels of DNMT3A and DNMT3B. Relative DNMT3A and DNMT3B levels were normalized with GAPDH.

 


Figure 1885
View larger version (31K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 5. ChIP analysis of DNMT3A association with selective gene promoters in NPCs. ChIP primers were designed within the methylation assay region. ERBIN promoter, which displayed no methylation change during differentiation, was used as the negative control. DNMT3A was selectively associated with CPT1A exon1a and exon1b, and SPAG6 promoters in P3 and P14 hESC–NPCs. Note that PCR conditions were slightly different for CPT1A exon1a prmoter region (40 cycles) from CPT1A exon1b, SPAG6 and ERBIN promoters (35 cycles), potentially due to the different PCR amplication efficiencies at different promoter regions.

 
Methylation status of repeat elements and non-CpG island promoters of OCT4 and GFAP genes during hESC differentiation
The bulk of DNA methylation in the mammalian genome occurs in various types of repeat elements including Alu repeats, retrotransposons, long- or short- interspersed repetitive elements (LINE and SINE) and pericentromeric repeats, which constitute ~40% of the human genome sequence (33,34). We surveyed different classes of repeat elements in the genome which were not covered in our initial DMH screen. Except for a slight transient increase in the MER52C class of repeat elements in P3 hESC–NPCs, we did not find any significant changes in DNA methylation patterns among hESCs, and P3 and P14 hESC–NPCs. Importantly, the methylation patterns in these repeat elements in hESCs and hESC–NPCs are similar to those in normal leukocytes (Fig. 6), suggesting that DNA methylation patterns in the bulk of repeat elements are stable during in vitro hESC differentiation.


Figure 1886
View larger version (33K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 6. Bisulfite sequencing analysis of DNA methylation in repeat elements during hESC differentiation by (A) Alusg (SINE), (B) L1PA13 (LINE) and (C) MER52C (LTR). Each dot represents one CpG site with filled black dots as methylated and open dots as unmethylated. Chromosome locations of these repeat elements are shown in schematic diagrams. Analyzed regions are underlined by black bars, and CpG sites represented by vertical bars.

 
It is also noted that de novo methylation can occur in non-CpG island gene promoters. Upon neural induction, the pluripotent marker gene OCT4 is silenced in P3 hESC–NPCs. Indeed, our DNA methylation analyses show dramatic increases in DNA methylation, consistent with a previous report (35). Moreover, our DNMT3A ChIP assay demonstrates that the OCT4 gene promoter is targeted by DNMT3A for de novo DNA methylation during in vitro hESC differentiation (Supplementary Material, Fig. S5). We also examined methylation status in a non-CpG island promoter region (–1640 to –1215 bp) within the human GFAP (GFAP) gene. In both hESCs and hESC–NPCs, this region is heavily methylated (80–90%) and only a minor decrease in DNA methylation occurs in late passages of P14 hESC–NPCs when compared with hESCs and P3 hESC–NPCs (Supplementary Material, Fig. S6). In contrast, cortical NPCs exhibit a lower level of methylation (65.6%) in vivo and cultured primary NPCs only contain a minimal level of methylation (7.8%) in this region, suggesting that demethylation occurs rapidly in primary NPCs in vitro (Supplementary Material, Fig. S6). Our data suggest the DNA methylation in non-CpG island regions could also be differentially regulated between primary NPCs and hESC–NPCs during in vivo and in vitro differentiation.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 
Our present study provides the first glimpse of methylation changes in 4608 CpG islands during neural differentiation of hESCs. Using a DNA microarray-based methylation profiling technique, we demonstrate that ~1.4% of CpG islands undergo distinct de novo methylation during the conversion of hESCs into differentiated hESC–NPCs. One of the goals of this study is to determine if the current neural differentiation protocol will cause any abnormal methylation in cancer-related genes in hESC-derivatives. After comparing our list of hypermethylated CpG islands with the list of genes reported to become hypermethylated in different types of cancer cells (17), we did not find any overlap between these two groups of CpG island genes. Our results indicate that the event of de novo methylation during hESC differentiation is distinctively different from that in cancer cells. One caveat of the above comparison is that hESC–NPCs are distinctively different by lineage from colon and prostate cancer cells. It may be more meaningful to compare our results with the profile of hypermethylated CpG islands in brain tumors that are derived from neural progenitors in future studies.

In addition, this study addresses whether in certain cases de novo methylation of CpG islands in hESC–NPCs are a result of a normal cell differentiation program that occurs in vivo or represents abnormal hypermethylation that takes place under our current differentiation/culture conditions. Bisulfite sequencing analysis in hESC–NPCs, primary fetal NPCs and astrocytes, and leukocytes indicates that levels of DNA methylation in CPT1A and SPAG6 promoters are higher in hESC–NPCs than in primary cells, indicating that methylation levels in these two CpG islands are abnormally high. In fact, such an increase in promoter methylation can lead to transcriptional silencing (Fig. 3), which may affect cell physiology. Indeed, deficiency of CPT1A has been associated with hypoketotic hypoglycemia in humans (28,29), suggesting that reduced expression of CPT1A in hESC–NPCs can potentially render these cells less efficient in lipid metabolism. Our results indicate that monitoring methylation changes in hESC derivatives may be a necessary step with the current differentiation/culture conditions of hESCs. Preventing de novo DNA methylation at these specific gene loci may be required if expanded cultures of NPCs are to be used in therapeutic applications in the future.

It has been recently reported that although the DNA methylation pattern is relatively stable in imprinting loci during long-term culture of hESCs (41), an increase in methylation in the promoters of certain tumor suppressor genes is observed in several lines of hESCs (13). In our studies, we did not observe any increased methylation in five cancer-related genes in HSF6 hES cell line over a period of 3 years with 170 passages, suggesting that HSF6 cell line is epigenetically quite stable under current hESC culture conditions.

De novo DNA methyltransferases must be involved in CpG island methylation during hESC differentiation, as has been postulated previously for methylation of CpG islands in mouse ES cells (36). In our study, we found that expression of de novo methyltransferase DNMT3A is much higher than DNMT3B expression in hESC–NPCs, suggesting that DNMT3A could be the major enzyme for de novo CpG island methylation. This possibility was confirmed by our ChIP assays which demonstrated the association of DNMT3A with CpG island DNA sequences in hESC–NPCs. Therefore, DNTM3A could be a prime target for preventing DNA methylation in CpG islands through biochemical and pharmacological approaches (37,38).

Our current study is limited to profiling DNA methylation in a CpG island library containing 4608 fragments during neural differentiation of one cell line of HSF6 hESCs. It will be of interest to verify and extend our findings with additional hESC cell lines and different differentiation paradigms such as lymphocyte and myocyte differentiation. Furthermore, dynamic DNA methylation changes could also occur in non-CpG island regions in cell lineage differentiation genes (7,8). Thus, with the advent of whole-genome tiling arrays (39) and new methods for genome-wide DNA methylation profiling such as DNA immunoprecipitation with 5' methyl-cytosine antibodies (17,40), it will be feasible to analyze the step-by-step changes in the DNA methylation pattern at all genomic regions during hESC differentiation. Because DNA methylation can directly or indirectly influence gene expression, it will be necessary to couple the genome-wide analysis of DNA methylation patterns with gene expression patterns. Comparing genome-wide methylation patterns with gene expression changes will allow us to determine if DNA methylation is causally linked to lineage-specific gene expression during development. This information will be extremely useful for developing novel strategies to control DNA methylation and gene expression during hESC differentiation for the ultimate goal of obtaining the most suitable hESC derivatives for cell therapy in the future.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
SUPPLEMENTARY MATERIAL
 REFERENCES
 
Cultures of hESCs and fetal astrocytes
Undifferentiated hESCs (HSF6 cell line with NIH code UC 06) were maintained on gamma-irradiated mouse feeder layer in DMEM high glucose supplement with 20% Knockout Serum Replacer (Invitrogen), 1 mM glutamine, 1% non-essential amino acids and 0.1 mM ß-mercaptoethanol. hESCs were passaged every 3–4 days using Dispase and Collagenase IV (Invitrogen) at a concentration of 1 mg/ml in KSR medium. Medium was changed everyday for optimal cell growth. Fetal astrocytes were purchased from AllCells LLC., and cultures were expanded as suggested by the vendor. Fetal brain tissue was obtained from the Surgical Pathology Laboratory at UCLA (Dr. Michael Teitell). Human primary NPCs were derived from primary cultures of fetal cortical tissues at the 11-week gestation stage using a protocol for culturing mouse NPCs (7). The use of hESCs and fetal brain tissues was approved by UCLA IRB.

Directed neural differentiation of hESCs in vitro
hESCs were passaged onto gamma-irradiated mouse feeder cell layer using the standard culture conditions as described (10). To induce neural differentiation, hESCs between passages 50–90 were first preplated for 1–2 h to get rid of most feeder cells, and then plated onto matrigel-coated plates with hESC medium and bFGF treatment for two consecutive passages. In the middle of the second passage, hESC culture medium was switched to DMEM/F12 serum-free medium/B27 supplement, which is suitable for the expansion of NPCs. Subsequently, partially differentiated hESCs were dissociated and plated onto poly-ornothine/fibronectin substrates with the serum-free medium and bFGF treatment (10 ng/ml) and continuously passaged to be converted into homogenous hESC-derived NPCs. In the first passage of hESC–NPCs (P1), we observed significant cell death in our cultures, suggesting a selection pressure for many non-NPC cell types. However, cell death was reduced to a minimum in the following two passages (P2 and P3) with enriched populations of hESC–NPCs. After NPCs became 80% confluent, they were sequentially passaged by scraping them off the plate or with light trypsin–EDTA treatment with serum-free DMEM/F12–B27 medium. Sequentially passaged cells were fixed with 4% paraformaldehyde (PFA)/PBS for immunostaining or harvested for RNA and DNA purification. For AzaC treatment, late passaged hESC–NPCs (P8–10) were treated with AzaC at the concentration of 1 µM for 6 days.

Bisulfite genomic sequencing analysis and COBRA assay
Genomic DNA samples were obtained from hESCs and early (P3 hESC–NPCs) and late (P14 hESC–NPCs) passages of hESC–NPCs. DNA was digested with BglII overnight and treated with sodium bisulfite for 15 h as described previously (24). Converted DNA samples were cleaned using Wizard® DNA Clean Up Kit (Promega) and were amplified by PCR reactions. PCR products were either directly sequenced or cloned into the Topo Vector 4.0 (Invitrogen) for sequencing of allelic methylation patterns. The sequence and primer information are available upon request. The COBRA assay followed the previously published protocol (25). MethPrimer program (http://www.urogene.org/methprimer/index1.html) was used to design primers for bisulfite sequencing PCR or restriction PCR for STMN3 (6F5) and ARMC7 (33A11) gene promoter CpG islands. The primers for STMN3 are 5'-TTTTAGGGTATTTTTAATATAATGAATTTA and 5'-ACAAAACTAACAACAAACAAAAAAC, which amplify the bisulfite-treated DNA into a 247 bp fragment with a HpyCH4IV site. Expected fragment sizes are 143 and 104 bp if the CpG site is methylated. The primers for ARMC7 gene promoter CpG island are 5'-TTTTTTGTAAAAGGGGTAAAGGATT and 5'-CAACAAAATCTAACAACACCCATATTA, which amplify the bisulfite-treated DNA into a 169 bp fragment restricted by TaqI into 132 and 37 bp fragments if the CpG site is methylated. The PCR reactions were performed with AmpliTaq Gold DNA polymerase from Applied Biosciences for 35 cycles with an annealing temperature of 55°C.

RT–PCR assays
RNA samples were extracted from hESCs, hESC–NPCs and primary astrocytes using TRIzol® (Invitrogen), and converted to cDNA with Omniscript® RT Kit (Qiagen) or iScript cDNA Synthesis Kit (Bio-Rad) for real-time PCR. PCR primers were designed across the exons for the genes of interest. PCR was carried out to assay gene expression and ß-ACTIN was used as a control. Real-time PCR was carried out with Bio-Rad iCycler real-time PCR machine using iQ SYBR Green Supermix (Bio-Rad). Primers were designed across the exons, and each primer pair was tested for PCR efficiency using a standard curve and melting curve of series diluted cDNA. Relative gene expression levels were calculated after they were normalized with expression levels of GAPDH.

Immunohistochemistry
For SSEA4 (Chemicon) staining, hESCs on coverslips were transferred to 24-well plates and live-stained with primary antibody directly diluted in culture medium for 30 min at 37°C. The cells were washed three times with PBS and incubated with Cy2- or Cy3-labeled anti-mouse IgG (1:500) secondary antibody (Jackson Immunoresearch, West Grove, PA, USA) for 1 h at 37°C. Stained cells were then fixed with 4% PFA/PBS and counter-stained with DAPI. For OCT4, NESTIN, SOX2, TuJ1, DNMT3A and DNMT3B immunostaining, cultured cells were first fixed with 4% PFA/PBS for 20 min at room temperature and washed with PBS. Cells were then permeablized with 0.4% Triton-X in TBST for 20 min and blocked in 10% milk and 1% normal goat serum for 1 h. Coverslips were incubated overnight at 4°C in primary antibodies diluted in 3% BSA in TBST [monoclonal mouse anti-OCT4 (1:20, Santa Cruz), polyclonal rabbit anti-NESTIN (1:200), polyclonal goat anti-SOX2 (1:200, Santa Cruz), monoclonal mouse anti-TuJ1 (1:2000, a gift from Dr. Frankfurter at University of Virginia), monoclonal mouse anti-Dnmt3a (1:400, Imgenex) and polyclonal rabbit anti-Dnmt3b (1:1000, a gift from Dr. En Li at Novartis Institute)]. After being washed three times with PBS, coverslips were incubated in fluorochrome-conjugated secondary antibodies for 1 h at room temperature with protection from light. DAPI or Hoechst dye #33342 was used to label cell nuclei.

Chromatin immunoprecipitation
ChIP assays were performed according to Upstate protocol. Cells were crosslinked with 1% formaldehyde in PBS for 20 min at room temperature and scraped off the plate following three washes with cold PBS. After that, cells were lysed with SDS lysis buffer with proteinase inhibitors and 1 mM dithiothretol (DTT). Cell lysates were sonicated to break down chromatin into 200–500 base pairs in length. Chromatin was 1:10 diluted, and precleared with protein A agarose beads (Upstate) for 1 h at 4°C. Approximately, 3% of the chromatin from each sample was saved as input. The rest of the chromatin was used for immunoprecipitation with either polyclonal rabbit anti-Dnmt3a (a gift from Dr. En Li at Novartis Institute) or no antibody as a control at 4°C overnight. Immunoprecipitated antibody–protein–DNA complexes were collected by protein A agarose beads at 4°C for 1 h and extensively washed. The complex was eluted and reverse crosslinked at 65°C for 6 h and proteinase K was added for further incubation at 45°C for 1 h. DNA samples were extracted using phenol:chloroform method. ChIP primers were designed within the corresponding methylation assay regions and PCR was performed for 35 (CPT1A exon1b, SPAG6, ERBIN) or 40 cycles (CPT1A exon1a) with the following cycling condition (94°C, 30 s; 56°C, 30 s; 72°C, 1 min).

Microarray-based methylation profiling
A genomic library of human unmethylated CpG island fragments was constructed by size fraction of SmaI cut DNA samples pooled from healthy human leukocytes (Z. Wang, unpublished data). Individual clones were cultured in 96-well microplates and inserts from 4226 random clones (44 plates) were PCR amplified with M13 vector primers. The quality of this CpG island fragment library was first assessed by sequencing analysis of 58 randomly picked clones. The majority of these 58 clones from this SmaI genomic fragment library were from CpG islands: of the 63 SmaI fragments (five clones have double inserts), 45 (71%) were from gene promoter region CpG islands; eight were from rRNA gene repeats; and the remaining were repetitive sequences (three inserts) and non-CGI sequences (seven inserts). Additional 384 CpG island fragments from 320 cancer-related genes were directly amplified from human genomic DNA. The total of 4608 PCR-amplified fragments were purified and spotted onto polylysine-coated microscope slides with an Affymetrix pin spotting machine.

For microarray hybridization, genomic DNA samples were treated with the methylation-sensitive restriction enzyme SmaI at 25°C overnight. A partially double-stranded linker with a blunt end was used to ligate the SmaI-digested fragments. For linker preparation, oligos (100 µM each) are mixed and incubated at 75°C for 2 min and cool to room temperature for 30–60 min. The oligo sequences for linker are:

  • 5'-AGCACTCTCCAGCCTCTCACCGAC-3'
  • 3'-AGAGTGGCTG-5'

Linkers were ligated to digested genomic DNA samples with T4 DNA ligase at 16°C overnight. The liagted genomic DNA were subjected to PCR amplification. PCR conditions are: 72°C, 5 min to fill the bottom stand of the linker, then 25 cycles of 95°C for 40 s, 72°C for 2 min using standard PCR reagents with the addition of betaine and DMSO. PCR products were run on agarose gel to check for amplification efficiency before labeling.

The PCR products were then purified by Qiagen PCR cleaning kit, labeled with either Cy3 or Cy5-dCTP, mixed and hybridized to the CpG island fragment microarray. The hybridized arrays were scanned with an Axon 4000B Microarray Scanner and the images were analyzed with GenPix Pro software. Three pairs of array results (hESCs versus P3 hESC–NPCs, hESCs versus P14 hESC–NPCs and P3 hESC–NPCs versus P14 hESC–NPCs) were normalized and the ratios of hybridization signals were averaged. For identifying CpG island clones exhibiting increased methylation, we initially set up a threshold of averaged ratio of hESC–NPCs probe intensity over hESCs probe intensity at <0.667 and hybridization signals must be present on at least three out of six slides (hESCs versus P3 NPCs and hESCs versus P14 NPCs). Additional candidate clones were included that are consistently below a ratio of 0.90 across all the arrays examined. Candidate clones were subsequently amplified and sequenced and their chromosomal locations were determined with UCSC BLAT genome browser (genome.ucsc.edu). All the sequences of CpG islands described in this article are available upon request.


   SUPPLEMENTARY MATERIAL
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
REFERENCES
 
Supplementary Material is available at HMG Online.


   ACKNOWLEDGEMENTS
 
We thank Mike Teitell, Yi Sun, Karen Reue and Peyman Golshani for critically reading the article. We also thank Yan Hu for technical help in the initial phase of this study. Y.S. conducted most of the experiments and participated in paper writing. J.C. assisted in experiments and in paper revision. Z.W. assisted in microarray experiments and data analysis. G.F. conceived and designed experiments and wrote the article. This work was supported by NIH Grant RO1 NS44405 and a research contract grant from HealthBanks Biotech Co., at Taiwan to G.F. Funding to pay the Open Access publication charges for this article was provided by UCLA

Conflict of Interest statement. The authors have declared that no competing interests exist.


   FOOTNOTES
 
{dagger} Present address: Division of Biology, Beckman Research Institute, City of Hope, Duarte, CA 91010, USA. Back


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 RESULTS
 DISCUSSION
 MATERIALS AND METHODS
 SUPPLEMENTARY MATERIAL
 REFERENCES
 

  1. Jaenisch R. and Bird A. (2000) Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat. Genet. 33:suppl., 245–254.

  2. Grace-Goll M. and Bestor T.H. (2005) Eukaryotic cytosine methyltransferases. Annu. Rev. Biochem. 74:481–514.[CrossRef][ISI][Medline]

  3. Li E., Bestor T.H., Jaenisch R. (1992) Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69:915–926.[CrossRef][ISI][Medline]

  4. Okano M., Bell D.W., Haber D.A., Li E. (1999) DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99:247–257.[CrossRef][ISI][Medline]

  5. Robertson K.D. (2005) DNA methylation and human disease. Nat. Rev. Genet. 6:597–610.[CrossRef][ISI][Medline]

  6. Xu G.L., Bestor T.H., Bourc'his D., Hsieh C.L., Tommerup N., Bugge M., Hulten M., Qu X., Russo J.J., Viegas-Pequignot E. (1999) Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature 402:187–191.[CrossRef][Medline]

  7. Fan G., Martinowich K., Chin M.H., He F., Fouse S.D., Hutnick L., Hattori D., Ge W., Shen Y., Wu H., et al. (2005) DNA methylation controls the timing of astrogliogenesis through regulation of JAK-STAT signaling. Development 132:3345–3356.[Abstract/Free Full Text]

  8. Takizawa T., Nakashima K., Namihira M., Ochiai W., Uemura A., Yanagisawa M., Fujita N., Nakao M., Taga T. (2001) DNA methylation is a critical cell-intrinsic determinant of astrocyte differentiation in the fetal brain. Dev. Cell 1:749–758.[CrossRef][ISI][Medline]

  9. Hochedlinger K. and Jaenisch R. (2003) Nuclear transplantation, embryonic stem cells, and the potential for cell therapy. N. Engl. J. Med. 349:275–286.[Free Full Text]

  10. Thomson J.A., Itskovitz-Eldor J., Shapiro S.S., Waknitz M.A., Swiergiel J.J., Marshall V.S., Jones J.M. (1998) Embryonic stem cell lines derived from human blastocysts. Science 282:1145–1147.[Abstract/Free Full Text]

  11. Keller G. (2005) Embryonic stem cell differentiation: emergence of a new era in biology and medicine. Genes Dev. 19:1129–1155.[Abstract/Free Full Text]

  12. Nistor G.I., Totoiu M.O., Haque N., Carpenter M.K., Keirstead H.S. (2005) Human embryonic stem cells differentiate into oligodendrocytes in high purity and myelinate after spinal cord transplantation. Glia 49:385–396.[CrossRef][ISI][Medline]

  13. Maitra A., Arking D.E., Shivapurkar N., Ikeda M., Stastny V., Kassauei K., Sui G., Cutler D.J., Liu Y., Brimble S.N., et al. (2005) Genomic alterations in cultured human embryonic stem cells. Nat. Genet. 37:1099–1103.[CrossRef][ISI][Medline]

  14. Bestor T.H., Gundersen G., Kolsto A.B., Prydz H. (1992) CpG islands in mammalian gene promoters are inherently resistant to de novo methylation. Genet. Anal. Tech. Appl. 9:48–53.[Medline]

  15. Bestor T.H. (2000) The DNA methyltransferases of mammals. Hum. Mol. Genet. 9:2395–2402.[Abstract/Free Full Text]

  16. Li E. (2002) Chromatin modification and epigenetic reprogramming in mammalian development. Nat. Rev. Genet. 3:662–673.[CrossRef][ISI][Medline]

  17. Keshet I., Schlesinger Y., Farkash S., Rand E., Hecht M., Segal E., Pikarski E., Young R.A., Niveleau A., Cedar H., Simon I. (2006) Evidence for an instructive mechanism of de novo methylation in cancer cells. Nat. Genet. 38:149–153.[CrossRef][ISI][Medline]

  18. Ying Q.L., Stavridis M., Griffiths D., Li M., Smith A. (2003) Conversion of embryonic stem cells into neuroectodermal precursors in adherent monoculture. Nat. Biotechnol. 21:183–186.[CrossRef][ISI][Medline]

  19. Conti L., Pollard S.M., Gorba T., Reitano E., Toselli M., Biella G., Sun Y., Sanzone S., Ying Q.L., Cattaneo E., Smith A. (2005) Niche-independent symmetrical self-renewal of a mammalian tissue stem cell. PLoS Biol. 3:e283.[CrossRef][Medline]

  20. Qian X., Shen Q., Goderie S.K., He W., Capela A., Davis A.A., Temple S. (2000) Timing of CNS cell generation: a programmed sequence of neuron and glial cell production from isolated murine cortical stem cells. Neuron 28:69–80.[CrossRef][ISI][Medline]

  21. Huang T.H., Perry M.R., Laux D.E. (1999) Methylation profiling of CpG islands in human breast cancer cells. Hum. Mol. Genet. 8:459–470.[Abstract/Free Full Text]

  22. Toyota M., Ho C., Ahuja N., Jair K.W., Li Q., Ohe-Toyota M., Baylin S.B., Issa J.P. (1999) Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res. 59:2307–2312.[Abstract/Free Full Text]

  23. Gardiner-Garden M. and Frommer M. (1987) CpG islands in vertebrate genomes. J. Mol. Biol. 196:261–282.[CrossRef][ISI][Medline]

  24. Clark S.J., Harrison J., Paul C.L., Frommer M. (1994) High sensitivity mapping of methylated cytosines. Nucleic Acids Res. 22:2990–2997.[Abstract/Free Full Text]

  25. Xiong Z. and Laird P.W. (1997) COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res. 25:2532–2534.[Abstract/Free Full Text]

  26. Jones P.A. and Baylin S.B. (2002) The fundamental role of epigenetic events in cancer. Nat. Rev. Genet. 3:415–428.[ISI][Medline]

  27. Reya T., Morrison S.J., Clarke M.F., Weissman I.L. (2001) Stem cells, cancer, and cancer stem cells. Nature 414:105–111.[CrossRef][Medline]

  28. Gobin S., Thuillier L., Jogl G., Faye A., Tong L., Chi M., Bonnefont J.P., Girard J., Prip-Buus C. (2003) Functional and structural basis of carnitine palmitoyltransferase 1A deficiency. J. Biol. Chem. 278:50428–50434.[Abstract/Free Full Text]

  29. Prip-Buus C., Thuillier L., Abadi N., Prasad C., Dilling L., Klasing J., Demaugre F., Greenberg C.R., Haworth J.C., Droin V., et al. (2001) Molecular and enzymatic characterization of a unique carnitine palmitoyltransferase 1A mutation in the Hutterite community. Mol. Genet. Metab. 73:46–54.[CrossRef][ISI][Medline]

  30. Neilson L.I., Schneider P.A., Van Deerlin P.G., Kiriakidou M., Driscoll D.A., Pellegrini M.C., Millinder S., Yamamoto K.K., French C.K., Strauss J.F. III. (1999) cDNA cloning and characterization of a human sperm antigen (SPAG6) with homology to the product of the Chlamydomonas PF16 locus. Genomics 60:272–280.[CrossRef][ISI][Medline]

  31. Zhang Z., Sapiro R., Kapfhamer D., Bucan M., Bray J., Chennathukuzhi V., McNamara P., Curtis A., Zhang M., Blanchette-Mackie E.J., Strauss J.F. III. (2002) A sperm-associated WD repeat protein orthologous to Chlamydomonas PF20 associates with Spag6, the mammalian orthologue of Chlamydomonas PF16. Mol. Cell. Biol. 22:7993–8004.[Abstract/Free Full Text]

  32. Price N., van der Leij F., Jackson V., Corstorphine C., Thomson R., Sorensen A., Zammit V. (2002) A novel brain-expressed protein related to carnitine palmitoyltransferase I. Genomics 80:433–442.[CrossRef][ISI][Medline]

  33. Lander E.S., Linton L.M., Birren B., Nusbaum C., Zody M.C., Baldwin J., Devon K., Dewar K., Doyle M., FitzHugh W., et al. (2001) Initial sequencing and analysis of the human genome. Nature 409:860–921.[CrossRef][Medline]

  34. Venter J.C., Adams M.D., Myers E.W., Li P.W., Mural R.J., Sutton G.G., Smith H.O., Yandell M., Evans C.A., Holt R.A., et al. (2001) The sequence of the human genome. Science 291:1304–1351.[Abstract/Free Full Text]

  35. Cowan C.A., Atienza J., Melton D.A., Eggan K. (2005) Nuclear reprogramming of somatic cells after fusion with human embryonic stem cells. Science 309:1369–1373.[Abstract/Free Full Text]

  36. Hattori N., Abe T., Suzuki M., Matsuyama T., Yoshida S., Li E., Shiota K. (2004) Preference of DNA methyltransferases for CpG islands in mouse embryonic stem cells. Genome Res. 14:1733–1740.[Abstract/Free Full Text]

  37. Brueckner B. and Lyko F. (2004) DNA methyltransferase inhibitors: old and new drugs for an epigenetic cancer therapy. Trends Pharmacol. Sci. 25:551–554.[CrossRef][Medline]

  38. Schneider-Stock R., Diab-Assef M., Rohrbeck A., Foltzer-Jourdainne C., Boltze C., Hartig R., Schonfeld P., Roessner A., Gali-Muhtasib H. (2005) 5-Aza-cytidine is a potent inhibitor of DNA methyltransferase 3a and induces apoptosis in HCT-116 colon cancer cells via Gadd45- and p53-dependent mechanisms. J. Pharmacol. Exp. Ther. 312:525–536.[Abstract/Free Full Text]

  39. Kim T.H., Barrera L.O., Zheng M., Qu C., Singer M.A., Richmond T.A., Wu Y., Green R.D., Ren B. (2005) A high-resolution map of active promoters in the human genome. Nature 436:876–880.[CrossRef][Medline]

  40. Schumacher A., Kapranov P., Kaminsky Z., Flanagan J., Assadzadeh A., Yau P., Virtanen C., Winegarden N., Cheng J., Gingeras T., Petronis A. (2006) Microarray-based DNA methylation profiling: technology and applications. Nucleic Acids Res. 34:528–542.[Abstract/Free Full Text]

  41. Rugg-Gunn P.J., Ferguson-Smith A.C., Pedersen R.A. (2005) Epigenetic status of human embryonic stem cells. Nat. Genet. 37:585–587.[CrossRef][ISI][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Proc. Natl. Acad. Sci. USAHome page
Y. Shen, Y. Matsuno, S. D. Fouse, N. Rao, S. Root, R. Xu, M. Pellegrini, A. D. Riggs, and G. Fan
X-inactivation in female human embryonic stem cells is in a nonrandom pattern and prone to epigenetic alterations
PNAS, March 25, 2008; 105(12): 4709 - 4714.
[Abstract] [Full Text] [PDF]

Home page
 Hum Mol GenetHome page
C. Allegrucci, Y.-Z. Wu, A. Thurston, C. N. Denning, H. Priddle, C. L. Mummery, D. Ward-van Oostwaard, P. W. Andrews, M. Stojkovic, N. Smith, et al.
Restriction landmark genome scanning identifies culture-induced DNA methylation instability in the human embryonic stem cell epigenome
Hum. Mol. Genet., May 15, 2007; 16(10): 1253 - 1268.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Supplementary Data
Right arrowOA All Versions of this Article:
15/17/2623    most recent
ddl188v1
Right arrow