Medical Device Link.

Originally Published IVD Technology November/December 2004

Figure 1. An effective method of methylation determination. In this example of the Pyrosequencing method, dCTP (the nucleotide added) has a base that is complementary to the next unpaired nucleotide on the template (G) and is incorporated by a DNA polymerase. The pyrophosphate (PPi) released is converted to ATP and then to a light signal via an enzyme cascade, including ATP-sulphurylase and luciferase. Excess nucleotide and ATP are degraded by apyrase before addition of the next nucleotide (click to enlarge).

Variation in gene expression plays a major role in the development of many diseases. Investigation into gene expression processes may generate information that can be used to develop effective IVDs. One factor that controls gene transcription is the methylation level of gene promoter regions.

Many laboratories have investigated the relationship between gene transcription and methylation levels. Such methylation-driven modifications of the genome are known as an “epigenetic” (Greek for “upon” genetic) event. Epigenetic events have been associated with the development and progression of many physiological changes.

This article assesses the value of methylation analysis for clinical research into a number of diseases (primarily cancer) and includes a brief discussion of the methods available for performing such research.^1–3 The article also discusses a proprietary technology that can be used to provide quality assurance when conducting methylation analysis studies.

Methylation and Physiology

Portions of the human genome are referred to as CpG islands because those areas are rich in CpG dinucleotides (i.e., cytosine-deoxyribose phosphates followed immediately by a guanine-deoxyribose phosphate) when compared with their surrounding regions. Cytosines of CpG islands in adult tissues, of which 60% are found in gene promoter regions, are normally nonmethylated, whereas those of CpG groups outside CpG islands are methylated by the activity of DNA methyltransferases (DNMTs).

Over the life span of an organism, DNA methylation serves a number of functions in the body. DNA methylation is important to the control of gene expression during the development of an organism. In some cases, methylation controls normal gene expression in adults (e.g., inactivation of the X chromosome in females). Various disease states are associated with CpG methylation, including Fragile X syndrome (i.e., expansion and methylation of the CGG repeat in the 5' untranslated region of the FMR1 gene, as well as promoter methylation), and Prader-Willi and Angelman syndromes (i.e., aberrations in methylation-dependent transcriptional silencing of imprinted genes in maternal or paternal chromosomes).

Most research in epigenetics is focused on cancer development. The development of cancer frequently involves an increase in the methylation of CpG islands, leading to gene inactivation. One example of this methylation is the silencing of the mismatch-repair gene MLH1 by promoter hypermethylation. This silencing appears to be the mechanism underlying the microsatellite-instability phenotype in endometrial cancers. Another example is the methylation of the DAP kinase gene. This process appears to be strongly linked to the development of non-small-cell lung carcinoma.⁴

While hypermethylation may lead to decreased expression of tumor-suppressor and DNA-repair genes, hypomethylation may permit expression of oncogenes and even cause chromosome instability and activation of retrotransposons (transposable elements that perform a retrovirus-like process of reverse transcription). The study of hypomethylation is now undergoing a renaissance.

For example, strong evidence implicates hypomethylation-mediated gene activation as a contributor to cancer development. In a number of genes, this activation contributes to a variety of cancers (e.g., the cyclin D2 and maspin genes in the development of gastric carcinoma, S100A4 colon cancer, and human papillomavirus 16 [i.e., cervical cancer]).⁵ The mechanisms by which DNA methylation modulates gene expression have been described in detail in the literature.^4,6

Changes in methylation patterns indicate changes in gene expression. Conventionally, changes in gene expression are studied via expression arrays or proteomics. However, DNA methylation studies offer a number of advantages for studying gene expression. First, the methylation pattern in a DNA molecule is relatively stable, in contrast to RNA transcripts. Second, methylation measurements can be made with absolute reference points. Finally, changes in methylation patterns may be both qualitative and quantitative, leading to assays with high specificity and sensitivity. In addition, such assays are more general than those for individual mutations and are localized to promoter regions in contrast to mutations that can be spread out in the gene.^7,8

Methylation level determinations offer myriad oncology-related clinical applications.

• Changes in the regulation of DNA methylation are an early signal in tumor development.
• By characterizing methylation processes, practitioners will be able to classify tumors.
• DNMT inhibitors are being tested as anticancer agents, with the associated monitoring of surrogate tissues or, in the case of leukemia, repeat samples. There are, however, concerns that long-term exposure to such agents may lead to chromosomal instability.
• Determining the methylation levels in DNA of cells in bodily fluids offers the possibility of obtaining information on gene expression through noninvasive sampling.8
Current methods for the analysis of methylation in bodily fluid samples have fairly low sensitivity but excellent specificity, making them valuable in population screening where the clinical follow-up of false-positives can be costly and invasive. In population screening, methylation markers can be used as a supplemental tool in risk assessment or disease detection. Such markers can enhance the specificity of existing screening methods with low specificity (e.g., prostate-specific antigen screening for prostate cancer).⁷

Methylation Determination Methods

Figure 2. Pyrosequencing is used to analyze methylation sites in exon 1 of the SNRPN gene. Genomic DNA contains C or methylated C (mC) in CpG dinucleotides (shown underlined and in bold). One C, in the boxed area, is not in a CpG dinucleotide. It is therefore not methylated and can be used as a control (shown in bold only). C and mC are converted by bisulfite treatment to U and finally T after amplification, while mC becomes C. The CpG positions (Y) and the control (T) in the amplified sequence are checked as C/T polymorphisms (click to enlarge).

A plethora of methods are available for analysis and quantification of global methylation and methylation profiles or patterns. Most methods for determining the methylation status of specific sites are based on modification of DNA by exposure to bisulfite.⁹

Methylated cytosines in an original DNA sample remain unaffected by bisulfite treatment, whereas unmethylated cytosines are converted to uracil. Polymerase chain reaction (PCR) can then be used to amplify this DNA such that methylated cytosine is copied to cytosine, and uracil is copied to thymine. Thus, the retention of cytosine in a specific position indicates methylation, whereas the appearance of thymine in a position that normally contains cytosine indicates the presence of unmethylated cytosine in the original DNA sample.

The bisulfite method has been one of the most significant developments in methylation analysis. However, exploitation of this method to its full potential places extra demands on the analytical technique.

Quality assurance is important when using bisulfite treatment in methylation studies because the products of such studies are prone to reaction artifacts. Treatment of genomic DNA with sodium bisulfite leads to effective deamination of unmethylated cytosines to uracil, while 5-methylcytosine is deaminated at a much slower rate.
The most commonly encountered artifact is incomplete conversion of cytosine to uracil. In addition, a small portion of 5-methylcytosines may be converted to thymine. Furthermore, treatment with bisulfite may lead to DNA degradation due to partial acid-catalyzed depurination.

Commercially available kits can improve the ease and robustness of bisulfite treatment (e.g., EZ DNA methylation kit by Biocompare Inc. [South San Francisco, CA] and Zymco Research Corp. [Orange, CA] CpGenome modification kit by Chemicon International [Temecula, CA] and Intergen Co. [Burlington, MA]. Nevertheless, when using bisulfite treatment, reaction conditions must be well controlled and controls must ensure that the treatment is complete but has not resulted in excessive degradation of DNA.

Furthermore, in the case of non- methylation-specific PCR, altering the base composition of the template molecule may complicate primer placement as primers must not be positioned over CpG sites where the level of methylation (and, thus, the sequence variation after bisulfite treatment) may be unknown.

Commercial and noncommercial software can help design of PCR primers for bisulfite-treated genomic DNA (e.g., CpGWARE primer design software by Chemicon International and Methprimer by the University of California, San Francisco Department of Urology, [available on-line at www.urogene.org/methprimer/]). Finally, the amplification of bisulfite-treated DNA may result in preferential amplification of either the methylated or unmethylated alleles.

The modified DNA strands that result from bisulfite treatment can be analyzed using a number of methods. These methods rely on oligonucleotide primers that are targeted to regions containing potentially methylated CpG sites. Changes in the methylation of such sites alter the base composition of the bisulfite-treated DNA (i.e., unmethylated cytosine converts to uracil, whereas methylated cytosine is retained). The resulting variations in priming efficiency or primer extension can then be used to determine the methylation level.

Methods based almost entirely on PCR (so-called methylation-specific PCR, or MS-PCR) may involve end-point analysis. MS-PCR–based kits can be used to determine methylation in the promoter regions of specific genes (e.g., CpG Wiz gene-specific amplification kits by Chemicon International and gene-specific promoter methylation detection kits by Panomics Inc. [Redwood City, CA]).

Recently, real-time quantitative PCR (Q-PCR) has been applied to the analysis of CpG methylation. For example, this technique was used to develop the Methylight method by researchers at the Norris Comprehensive Cancer Center at the University of Southern California School of Medicine (Los Angeles).10 This method relies on relative quantification against controls that are methylation insensitive, as well as a standard curve based on serial dilution of a control sample.

Other analytical methods include the use of methylation-specific restriction enzymes, methylation-specific single-base primer extension, arrays, or Sanger sequencing. The latter can be applied as direct genomic sequencing for semiquantitative analysis of multiple adjacent CpG sites, setting it apart from most methods, which generate data for an individual position per assay (i.e., primer). Alternatively, the sequencing of large numbers of cloned fragments provides quantitative information for multiple methylation sites, but is a tedious and cumbersome method.

Methylation Analysis via Pyrosequencing

Most analytical methods employ some sort of control to indicate the success of bisulfite treatment. Analysis technologies that both permit flexibility in the positioning of primers around methylation sites and also include controls that confirm the correct placement of the primer would be advantageous. Several research groups have realized such advantages when using Pyrosequencing technology by Biotage AB (Uppsala, Sweden), which features built-in controls for quality assurance of the final result, for the quantitative analysis of multiple adjacent CpG sites.

Pyrosequencing technology is a method for sequencing-by-synthesis in real time.¹¹ It is based on an indirect bioluminometric assay of the pyrophosphate (PPi) that is released from each deoxynucleotide (dNTP) upon DNA-chain elongation.

This method presents a DNA template- primer complex with a dNTP in the presence of exonuclease-deficient Klenow DNA polymerase (see Figure 1). The four nucleotides are sequentially added to the reaction mix in a predetermined order. If the nucleotide is complementary to the template base and thus incorporated, PPi is released. PPi is then used as a substrate, together with adenosine 5'-phosphosulphate (APS), by ATP sulphurylase, which results in the formation of adenosine triphosphate (ATP).

Luciferase then converts the ATP (together with luciferin) to oxyluciferin, AMP, PPi, and visible light that is detected by either a luminometer or a charge-coupled device. The light produced is proportional to the number of nucleotides added to the extended primer chain, and results in a peak indicating the number and type of nucleotide present in the form of a pyrogram.

Excess nucleotide is digested by apyrase present in the reaction mixture before the addition of the next nucleotide. Pyrosequencing technology enables DNA to be analyzed when either bound to solid support or in solution. Several DNA purification methods have been reviewed in the literature.¹²

Pyrosequencing Applications

Figure 3. Pyrosequencing technology is applied to the determination of methylation levels in six positions (M5–M10) in a region of the CpG island at the transcription start of the GSTP1 gene. The pyrograms (intensity readouts) show results from methylated DNA mixed with increasing amounts of unmethylated DNA (% methylated indicated), with C/T polymorphisms derived from CpG dinucleotides shaded. The C peaks decrease and T peaks increase in the shaded areas as the % methylated DNA decreases. Figure reproduced with permission from the authors and Eaton Publishing, USA.16 (click to enlarge).

The technology has been commercialized via a range of systems, software, and reagent kits for assay design and processing of up to 96 post-PCR samples in parallel using solid-phase carriers. The kits include methods for rapid sample preparation.¹³ The technology lends itself to automation as it involves a homogeneous assay.

The software designed for Pyrosequencing applications supports multiplexing of single nucleotide polymorphisms (SNPs), or mutation detection in different templates or positions. The software also facilitates detection of closely placed (and even contiguous) SNPs in one template, analysis of insertions and deletions, allele frequency quantification, and sequencing of short stretches of DNA.

Pyrosequencing technology has a number of valuable attributes for methylation studies. The method provides internal controls to confirm the completeness of bisulfite treatment together with the ability to analyze several individual neighboring methylation sites. In addition, the surrounding sequence generated around the methylated areas provides control peaks that can be used to confirm the correct placement of the primer on the modified DNA template, which is now rich in adenine (A) and thymine (T).

Pyrosequencing can be used for the analysis of methylation sites in exon 1 of the SNRPN gene, a gene that has been implicated in Prader-Willi and Angelman syndromes (see Figure 2). In such analyses, bisulfite treatment converts the nonmethylated cytosines (C) to uracil (U), and PCR converts the uracil to thymine (T). Pyrosequencing is then used to determine the relative levels of T and C (corresponding to unmethylated C, and methylated C, respectively) in the original DNA sample, for each position in the PCR product.

In the example shown, four cytosines are in CpG dinucleotides and may, therefore, be at least partially methylated (mC), which protects them from bisulfite treatment. These positions can thus contain either mC or U after treatment (as shown by Y), which results in C or T (as shown by Y) after amplification by PCR. These positions yielded C and T peaks with a comethylation level ranging from 38.6 to 42.7%.

In addition, one cytosine is not in a CpG dinucleotide (therefore, it is not methylated) and can be used as a control for bisulfite treatment (see Figure 2). The pyrogram shows no peak for C in the control position, whereas it shows a full peak for T. This confirms that the bisulfite treatment has fully converted the unmethylated C to U, and hence to T via the PCR amplification. The surrounding sequence confirmed the correct placement of the sequencing primer.

A number of researchers have applied the technology to the study of CpG methylation.^14-19In the first study published, Pyrosequencing and a method based on single-base primer extension were applied to the analysis of a single CpG dinucleotide that is consistently hypomethylated in pilocytic astrocytomas but not in other gliomas.14 Pyrosequencing generated less scattering than the other method.

The technology was used for the determination of methylation levels in six positions in a region of the CpG island at the transcription start of the GSTP1 gene (see Figure 3).16 This gene codes for glutathione-S-transferase š and is unmethylated in normal tissue but highly methylated in prostate tumors.

A 140-bp region containing 17 potential methylation sites was amplified and 15 of the sites were analyzed using four sequencing primers. The linearity of the method was checked using mixtures with known levels of methylation. The six positions that were analyzed by a single sequencing primer are shaded (see Figure 3). The relative levels of T and C, corresponding to unmethylated C and methylated C, respectively, were checked for each position.

The method revealed a progressive decrease in the C peaks, with a corresponding increase in the T peaks, as the degree of methylation decreased. In this case, a generally high level of comethylation gave very good linearity, in the range 0–100% methylation.¹⁶ The method has been recently improved by including the use of single-stranded DNA binding protein to give high-quality data that demonstrated clear variation in the methylation level between several CpG sites. In addition, the data obtained from a few sites was confirmed using an independent method.¹⁸

The Pyrosequencing method was then further developed to include the use of universal biotinylated primers to reduce costs. The improved method compares favorably with combined bisulfite restriction analysis (COBRA), a method based on methylation-specific restriction enzymes, in the analysis of methylation sites in the promoter region of the CDKN2A gene, which codes for p16.15

Studies using Pyrosequencing along with primers indicated that methylation of neighboring CpG sites can be variable.18 In addition, such variation may lead to change in gene expression, suggesting that methods facilitating analysis of individual, multiple methylation sites could be of particular value.

Pyrosequencing technology has also been used to determine the methylation status of five CpG sites around the NotI restriction site of the T-cell leukemia 1 oncogene (TCL-1) promoter.¹⁷

A novel application of Pyrosequencing is its functionality in determining DNA hypomethylation of large numbers of loci in repetitive elements, which could be an alternative to laborious methods for quantifying global methylation levels.¹⁹

Conclusion

The determination of methylation patterns requires dependable methods for the detailed analysis of multiple, individual sites. This determination also requires well-defined controls to ensure that the data generated can be relied upon. Such methods will provide powerful tools in unraveling the complex nature of diseases where epigenetic events are implicated.

While Pyrosequencing technology is currently not intended for diagnostic purposes, it has a promising future in the clinical research laboratory as a tool for linking aberrant CpG methylation to diseases such as cancer.

References

1. MF Fraga and M Esteller, “DNA Methylation: A Profile of Methods and Applications,” Biotechniques 33 (2002): 632, 634, 636–649.

2. M Shiraishi, AJ Oates, and T Sekiy, “An Overview of the Analysis of DNA Methylation in Mammalian Genomes,” Biological Chemistry 383 (2002): 893–906.

3. C Dahl and P Guldberg, “DNA Methylation Analysis Techniques,” Biogerontology 4 (2003): 233–250.

4. G Strathdee and R Brown, “Aberrant DNA Methylation in Cancer: Potential Clinical Interventions,” Expert Review of Molecular Medicine (2002): 1–17 [accessed 11 October 2004]; available from Internet: www.expertreviews.org/02004222h.htm.

5. AP Feinberg and B Tycko, “The History of Cancer Epigenetics,” Nature Reviews Cancer 4 (2004): 143–153.

6. PA Jones and SB Baylin, “The Fundamental Role of Epigenetic Events in Cancer,” Nature Reviews Genetics 3 (2002): 415–428.

7. PW Laird, “The Power and the Promise of DNA Methylation Markers,” Nature Reviews Cancer 3 (2003): 253–266.

8. M Widschwendter and PA Jones, “DNA Methylation and Breast Carcinogenesis,” Oncogene 21 (2002): 5462–5482.

9. SJ Clark et al., “High Sensitivity Mapping of Methylated Cytosines,” Nucleic Acids Research 22 (1994): 2990–2997.
10. CA Eads et al., “MethyLight: A High-Throughput Assay to Measure DNA Methylation,” Nucleic Acids Research 28 (2000): E32.

11. M Ronaghi, M Uhlen, and P Nyren, “A Sequencing Method Based on Real-Time Pyrophosphate,” Science 281 (1998): 363, 365.

12. H Fakhrai-Rad, N Pourmand, and M Ronaghi, “Pyrosequencing: An Accurate Detection Platform for Single Nucleotide Polymorphisms,” Human Mutation 19 (2002): 479–485.

13. J Dunker et al., “Parallel DNA Template Preparation Using a Vacuum Filtration Sample Transfer Device,” Biotechniques 34 (2003): 862–866, 868.

14. K Uhlmann et al., “Evaluation of a Potential Epigenetic Biomarker by Quantitative Methyl-Single Nucleotide Polymorphism Analysis,” Electrophoresis 23 (2002): 4072–4079.

15. S Colella et al., “Sensitive and Quantitative Universal Pyrosequencing Methylation Analysis of CpG Sites,” Biotechniques 35 (2003): 146–150.

16. J Tost, J Dunker, and IG Gut, “Analysis and Quantification of Multiple Methylation Variable Positions in CpG Islands by Pyrosequencing,” Biotechniques 35 (2003): 152–156.

17. C Kiss et al., “T Cell Leukemia I Oncogene Expression Depends on the Presence of Epstein-Barr Virus in the Virus-Carrying Burkitt Lymphoma Lines,” Proceedings of the National Academy of Sciences, U.S.A. 100 (2003): 4813–4818.

18. JM Dupont et al., “De Novo Quantitative Bisulfite Sequencing Using the Pyrosequencing Technology,” Analytical Biochemistry 333 (2004): 119–127.

19. AS Yang et al., “A Simple Method for Estimating Global DNA Methylation Using Bisulfite PCR of Repetitive DNA Elements,” Nucleic Acids Research 32 (2004): e38.

Nigel Tooke, PhD, and Monica Pettersson, MSc, are senior scientists at Biotage AB (Uppsala, Sweden). They work with the development of molecular biology methods and applications for Pyrosequencing and can be contacted at nigel.tooke@biotage.se and monica.pettersson@biotage.se, respectively.

Copyright ©2004 IVD Technology