Rapid analysis of heterogeneously methylated DNA using digital methylation-sensitive high resolution melting: application to the CDKN2B (p15) gene
© Candiloro et al; licensee BioMed Central Ltd. 2008
Received: 28 April 2008
Accepted: 03 November 2008
Published: 03 November 2008
Methylation-sensitive high resolution melting (MS-HRM) methodology is able to recognise heterogeneously methylated sequences by their characteristic melting profiles. To further analyse heterogeneously methylated sequences, we adopted a digital approach to MS-HRM (dMS-HRM) that involves the amplification of single templates after limiting dilution to quantify and to determine the degree of methylation. We used this approach to study methylation of the CDKN2B (p15) cell cycle progression inhibitor gene which is inactivated by DNA methylation in haematological malignancies of the myeloid lineage. Its promoter region usually shows heterogeneous methylation and is only rarely fully methylated. The methylation status of CDKN2B can be used as a biomarker of response to treatment. Therefore the accurate characterisation of its methylation is desirable.
MS-HRM was used to assess CDKN2B methylation in acute myeloid leukaemia (AML) samples. All the AML samples that were methylated at the CDKN2B promoter (40/93) showed varying degrees of heterogeneous methylation. Six representative samples were selected for further study. dMS-HRM was used to simultaneously count the methylated alleles and assess the degree of methylation. Direct sequencing of selected dMS-HRM products was used to determine the exact DNA methylation pattern and confirmed the degree of methylation estimated by dMS-HRM.
dMS-HRM is a powerful technique for the analysis of methylation in CDKN2B and other heterogeneously methylated genes. It eliminates both PCR and cloning bias towards either methylated or unmethylated DNA. Potentially complex information is simplified into a digital output, allowing counting of methylated and unmethylated alleles and providing an overall picture of methylation at the given locus. Downstream sequencing is minimised as dMS-HRM acts as a screen to select only methylated clones for further analysis.
Epigenetic mechanisms, in particular DNA methylation, play a major role in the modulation of gene activity in cancer. Methylation has been shown to silence a large number of genes in nearly every type of cancer [1, 2]. Whilst it is generally accepted that methylation of a promoter may be necessary for gene silencing, it is clear that in many cancers promoters are often heterogeneously methylated. This is an important issue both for the detection and quantitation of methylation.
The tumour suppressor gene CDKN2B (p15), which is silenced in a variety of haematological malignancies , is one such heterogeneously methylated locus. Silencing of CDKN2B expression can occur with only partial methylation of the promoter and many differently methylated CDKN2B alleles frequently co-exist [4–6].
Methylation of CDKN2B has been used as a biomarker for myelodysplastic syndrome and acute myeloid leukaemia (AML) , and as a prognostic indicator either alone or in combination with other loci [8, 9]. Monitoring changes in CDKN2B methylation over time would also prove useful in assessing residual disease. However, due to its heterogeneity, quantification of CDKN2B methylation is challenging.
Methylation-sensitive high resolution melting (MS-HRM) is a methodology that is particularly suitable for the rapid analysis of clinical samples . MS-HRM differentiates methylated and unmethylated templates on the basis of the marked difference in melting behaviour due to their different base compositions following bisulphite conversion. We have used MS-HRM to analyse methylated promoter regions in cancer [10, 11], and the H19/IGF2 imprinting centre in imprinting disorders . Here, we show that MS-HRM is an appropriate methodology for the detection of heterogeneously methylated cancer samples using the CDKN2B gene as an example.
Digital methylation-sensitive high resolution melting (dMS-HRM) was introduced as a methodology for counting methylated and unmethylated alleles of the BRCA1 gene . In that case, dMS-HRM was used to confirm the MS-HRM analysis. In this communication, we show that dMS-HRM enables detailed analysis of DNA methylation in complex heterogeneously methylated templates, eliminating the need for sequencing analysis in most cases. This can be done in a time- and cost-effective fashion as we have shown using the clinically important CDKN2B gene as an example.
AML samples were obtained from patients referred to the Department of Haematology, Aarhus University Hospital, Denmark. DNA was re-dissolved in TE buffer (1×) at a final concentration of 5 ng/μl. Genomic DNA was extracted from peripheral blood of healthy controls using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. The investigation was approved by the Peter MacCallum Cancer Centre Ethics of Human Research Committee (Approval number 02/26). Whole genome amplification (WGA) was performed as described previously . The purpose of using DNA subjected to two rounds of WGA is to ensure that the DNA is completely unmethylated, as normal, healthy individuals may have low-level methylation at the CDKN2B promoter in peripheral blood cells .
200 ng of genomic DNA from the AML samples was subjected to bisulphite modification by using the EpiTect Bisulfite Kit (Qiagen) according to the manufacturer's instructions. DNA was eluted once in 20 μl of buffer EB. Universal Methylated DNA (Chemicon, Millipore, Billerica, MA) and WGA product were used as the controls. 500 ng of each was modified. The modified control DNA underwent a second elution in 30 μl of buffer EB.
PCR cycling and MS-HRM was performed on the Rotor-Gene 6000 (Corbett Research, Sydney, Australia), an HRM-enabled real time PCR instrument. Each sample was analysed in duplicate for MS-HRM.
Primers were designed according to the principles outlined in Wojdacz et al . Briefly, the primers should contain a limited number of CpG dinucleotides (usually one or two) which should be kept as far as possible from the 3' end of the primers. This allows the control of PCR bias by appropriate choice of annealing temperature.
The primers used to amplify bisulphite-treated DNA were CDKN2B-F, 5'-GTTAGGCGTTTTTTTTTAGAAGTAATTTAGG-3' and CDKN2B-R, 5'-TACGACTTAAAACCCCGTACAATAACC-3' and do not amplify unmodified genomic DNA (data not shown). The amplified region corresponds to [GenBank: AL449423] nucleotides 99845 to 99958, and encompasses nine CpG dinucleotides. PCR was performed in 100 μl PCR tubes (Corbett Research) with a final volume of 20 μl, containing 200 nmol/l of each primer, 200 μmol/l of each dNTP, 0.5 U of HotStarTaq DNA polymerase (Qiagen) in the supplied PCR buffer containing 2.5 mmol/l MgCl2, 5 μmol/l SYTO9 (Invitrogen, Carlsbad, CA), and 10 ng of bisulphite-treated DNA. The initial denaturation (95°C, 15 minutes) was followed by 50 cycles for MS-HRM of 10 seconds at 95°C, 30 seconds at 59°C, 30 seconds at 72°C; one cycle of 1 minute at 95°C, 72°C for 1.5 minutes and a HRM step from 65°C to 90°C rising at 0.2°C per second, and holding for 1 second after each stepwise increment. The annealing temperature of 59°C was chosen as it gave a near-proportional amplification of methylated and unmethylated templates.
The basis of dMS-HRM is the PCR amplification of single molecules following limiting dilution [15, 16]. To perform limiting dilution, the DNA concentration is determined. Depending on the starting concentration, a dilution series (in the order of 1:1000 to 1:2000 for 10 ng/μl) should be used. The dilution that performs best (that is, a substantial proportion of negative reactions) is chosen and used for the final experimental set-up.
The Poisson distribution is used as a guideline to determine the expected distribution of templates. For example, if the sample is diluted such that one amplifiable template is found on average per reaction, 36.8% of reactions will have no templates, 36.8% will have one template, and 26.4% will have more than one template. Amplification will occur when one or more template molecules are present in the PCR reaction.
Amplifications from single templates can be readily identified when the melting curves are analysed. They show a smooth and sharp single signal. Melting curves from two (or more) templates generally result in signals showing two peaks if heteroduplexes are not formed, or more complex patterns when heteroduplexes are formed. Digital MS-HRM was performed using 60 cycles of amplification. Sixty replicates of diluted template were analysed per sample.
dMS-HRM products were cleaned up with the PCR-M clean-up kit (Viogene, Taipei, Taiwan), according to the manufacturer's instructions, further processed with ExoSapIT (GE Healthcare, Little Chalfont, England), followed by the sequencing reaction using Big Dye Terminator v3.1 chemistry (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. Sequencing was performed in both directions using the PCR primers given above as sequencing primers. The initial denaturation (95°C, 1 minute) was followed by 30 cycles of 10 seconds at 95°C, 30 seconds at 59°C and 3 minutes at 72°C. The sequencing products were purified by ethanol precipitation and separated on a 3100 Genetic Analyser (Applied Biosystems). The sequencing data for the dMS-HRM products were analysed using BiQ Analyzer software (Max-Planck-Institut für Informatik, Saarbrücken, Germany) .
MS-HRM of AML samples
dMS-HRM of AML samples
In dMS-HRM, an increase in the melting temperature of the products relative to the controls can be seen (Figure 4a). However, sequencing of digital 'clones' obtained from the controls showed that their sequences were directly comparable to those from sample 9164 (Figure 4b). dMS-HRM analysis of control DNA was replicated with the template DNA being diluted into an equivalent amount of background DNA used for MS-HRM (fish sperm and unmodified genomic control DNA), and the same result was obtained (data not shown).
Incomplete conversion was detected on one occasion as the peak had a higher melting temperature than fully methylated 'clones' (Figure 4a). This was shown by direct sequencing to have been fully methylated and to have an additional unconverted cytosine (Figure 4c). Previous reports show that conversion of cytosine to uracil is 97.0 to 99.8% complete, whilst still leaving 5-methylcytosines intact [18, 19]. Such a low non-conversion rate is not expected to be a major problem for dMS-HRM.
In our work on developing MS-HRM , we found that many promoter regions were heterogeneously methylated in cancer specimens (Candiloro et al, unpublished results). Heterogeneously methylated DNA samples result in multiple products after PCR. Not only will there be many different homoduplexes, but multiple heteroduplexes will also form between similarly methylated sequences, giving rise to distinctive melting curves. These often do not lie within the range defined by fully methylated and unmethylated DNA controls due to the earlier melting of heteroduplexes relative to homoduplexes. Thus MS-HRM allows ready identification of heterogeneously methylated samples (Figure 2). Depending on the amount of DNA methylation (and the amount of normal cells present in the tumour sample), the resulting melting curve is flattened and exhibits a complex melting pattern.
Quantification of the degree of heterogeneous methylation by MS-HRM remains problematic. MS-HRM shows great sensitivity when the samples contains a mixture of fully methylated and fully unmethylated templates  as it has been empirically observed that the sequence differences between these are too great for heteroduplexes to form. However, the sensitivity of MS-HRM is considerably diminished for heterogeneously methylated samples due to multiple heteroduplex formation creating a melting profile that cannot be directly compared with methylated and unmethylated controls.
To use a PCR-based technique, PCR bias needs to be eliminated or at least minimised. Given that the DNA methylation pattern in the entire interrogated sequence, and not just the primer binding sites, influences PCR bias , eliminating bias in a heterogeneous population of alleles becomes nearly impossible. It must also be kept in mind that cloning from a heterogeneous pool may introduce a cloning bias [21, 22].
Digital PCR from single templates eliminates these problems . When single molecules are amplified, there is no bias as there is no competition between different molecules within a single reaction.
We previously introduced dMS-HRM as a tool to count the number of methylated and unmethylated alleles for the BRCA1 promoter . In this communication, we have adapted this methodology to examine the complexity and the degree of methylation of the CDKN2B promoter. Cameron et al determined that about 40% of the CpG sites within the CDKN2B CpG island need to be methylated to achieve complete silencing of the CDKN2B gene, regardless of the exact CpG methylation pattern . Detection of DNA methylation at the CDKN2B promoter has been problematic as many methodologies are not well suited to the study of heterogeneously methylated regions.
A cost- and time-effective quantitative method is desirable for analysis of both research and clinical specimens where it is useful to know both the number of methylated templates and the degree of their methylation. Bisulphite sequencing of individual clones is an effective research approach but is too laborious and expensive for routine clinical use [4, 5]. Our results show that dMS-HRM can be used as a more rapid alternative approach. In addition, dMS-HRM products may be directly sequenced. As dMS-HRM identifies the methylated templates, the need to sequence unmethylated templates is eliminated. Sequencing of the individual dMS-HRM amplicons enabled the identification of the exact methylation pattern. Sequencing confirmed that amplicons of similar Tm had similar numbers of methylated CpGs, although the actual positions varied (Figure 4).
MS-HRM and dMS-HRM can be compared with two other readily performed approaches that can recognise heterogeneous methylation. Denaturing gradient gel electrophoresis [6, 24] is an effective way of visualising the complexity of heterogeneous methylation, but the methodology has not become widely used. Pyrosequencing is a comparatively rapid and quantitative method for methylation analysis [25–27]. Its quantitative approach and the characterisation of methylation pattern make it a useful method in a clinical setting, as shown for CDKN2B [7, 28] and MGMT . Heterogeneously methylated templates can be recognised if there is marked variation in the degree of methylation for the interrogated CpG positions. The great advantage of MS-HRM and dMS-HRM is that they are in-tube methodologies that are rapidly performed and do not need post-PCR processing as the melting analysis is carried out in the same reaction vessel as the amplification.
MS-HRM can readily identify heterogeneously methylated templates by their characteristic melting patterns. This paper introduces dMS-HRM as an analytical method for the rapid study of heterogeneous methylation using the clinically important CDKN2B (p15) gene as an example. Although MS-HRM will often be sufficient to assess the methylation status of the CDKN2B gene, dMS-HRM will provide a more readily interpretable visualisation which may be quantified by scoring the individual peaks. dMS-HRM can also rapidly generate clonal templates for sequencing or pyrosequencing, eliminating the need for conventional cloning. Moreover, dMS-HRM will significantly reduce the amount of sequencing required as only clearly methylated amplicons will need to be sequenced.
List of abbreviations used
acute myeloid leukaemia
digital methylation-sensitive high resolution melting
methylation-sensitive high resolution melting
whole genome amplification.
We would like to thank Lasse Kristensen for critical reading of the manuscript. This work was funded by a grant from the Leukaemia Foundation of Victoria.
- Esteller M: Epigenetics in Cancer. N Engl J Med. 2008, 358: 1148-1159. 10.1056/NEJMra072067.View ArticlePubMedGoogle Scholar
- Jones PA, Baylin SB: The Epigenomics of Cancer. Cell. 2007, 128: 683-692. 10.1016/j.cell.2007.01.029.PubMed CentralView ArticlePubMedGoogle Scholar
- Herman JG, Civin CI, Issa J-PJ, Collector MI, Sharkis SJ, Baylin SB: Distinct patterns of inactivation of p15INK4B and p16INK4A characterize the major types of hematological malignancies. Cancer Res. 1997, 57: 837-841.PubMedGoogle Scholar
- Cameron EE, Baylin SB, Herman JG: p15INK4B CpG island methylation in primary acute leukemia is heterogeneous and suggests density as a critical factor for transcriptional silencing. Blood. 1999, 94: 2445-2451.PubMedGoogle Scholar
- Dodge JE, List AF, Futscher BW: Selective variegated methylation of the p15 CpG island in acute myeloid leukemia. Int J Cancer. 1998, 78: 561-567. 10.1002/(SICI)1097-0215(19981123)78:5<561::AID-IJC6>3.0.CO;2-R.View ArticlePubMedGoogle Scholar
- Aggerholm A, Guldberg P, Hokland M, Hokland P: Extensive intra- and interindividual heterogeneity of p15 INK4B methylation in acute myeloid leukemia. Cancer Res. 1999, 59: 436-441.PubMedGoogle Scholar
- Brakensiek K, Wingen LU, Langer F, Kreipe H, Lehmann U: quantitative high-resolution CpG island mapping with pyrosequencingTM reveals disease-specific methylation patterns of the CDKN2B gene in myelodysplastic syndrome and myeloid leukemia. Clin Chem. 2007, 53: 17-23. 10.1373/clinchem.2007.072629.View ArticlePubMedGoogle Scholar
- Aggerholm A, Holm MS, Guldberg P, Olesen LH, Hokland P: Promoter hypermethylation of p15INK4B, HIC1, CDH1, and ER is frequent in myelodysplastic syndrome and predicts poor prognosis in early-stage patients. Eur J Haematol. 2006, 76: 23-32. 10.1111/j.1600-0609.2005.00559.x.View ArticlePubMedGoogle Scholar
- Grövdal M, Khan R, Aggerholm A, Antunovic P, Astermark J, Bernell P, Engström LM, Kjeldsen L, Linder O, Nilsson L, Olsson A, Wallvik J, Tangen JM, Oberg G, Jacobsen SE, Hokland P, Porwit A, Hellström-Lindberg E: Negative effect of DNA hypermethylation on the outcome of intensive chemotherapy in older patients with high-risk myelodysplastic syndromes and acute myeloid leukemia following myelodysplastic syndrome. Clin Cancer Res. 2007, 13: 7107-7112. 10.1158/1078-0432.CCR-07-1193.View ArticlePubMedGoogle Scholar
- Wojdacz TK, Dobrovic A: Methylation-sensitive high resolution melting (MS-HRM): a new approach for sensitive and high-throughput assessment of methylation. Nucl Acids Res. 2007, 35: e41-10.1093/nar/gkm013.PubMed CentralView ArticlePubMedGoogle Scholar
- Snell C, Krypuy M, Wong E, kConFab , Loughrey M, Dobrovic A: BRCA1 promoter methylation in peripheral blood DNA of mutation negative familial breast cancer patients with a BRCA1 tumour phenotype. Breast Cancer Res. 2008, 10: R12-10.1186/bcr1858.PubMed CentralView ArticlePubMedGoogle Scholar
- Wojdacz TK, Dobrovic A, Algar EM: Rapid detection of methylation change at H19 in human imprinting disorders using methylation-sensitive high-resolution melting. Hum Mutat. 2008Google Scholar
- Kristensen LS, Mikeska T, Krypuy M, Dobrovic A: Sensitive melting analysis after real time – methylation specific PCR (SMART-MSP): high-throughput and probe-free quantitative DNA methylation detection. Nucl Acids Res. 2008, 36: e42-10.1093/nar/gkn113.PubMed CentralView ArticlePubMedGoogle Scholar
- Wojdacz TK, Hansen LL, Dobrovic A: A new approach to primer design for the control of PCR bias in methylation studies. BMC Res Notes. 2008, 1: 54-10.1186/1756-0500-1-54.PubMed CentralView ArticlePubMedGoogle Scholar
- Pohl G, Shih I-M: Principle and applications of digital PCR. Expert Rev Mol Diagn. 2004, 4: 41-47. 10.1586/14737126.96.36.199.View ArticlePubMedGoogle Scholar
- Vogelstein B, Kinzler KW: Digital PCR. Proc Nat Acad Sci USA. 1999, 96: 9236-9241. 10.1073/pnas.96.16.9236.PubMed CentralView ArticlePubMedGoogle Scholar
- Bock C, Reither S, Mikeska T, Paulsen M, Walter J, Lengauer T: BiQ Analyzer: visualization and quality control for DNA methylation data from bisulfite sequencing. Bioinformatics. 2005, 21: 4067-4068. 10.1093/bioinformatics/bti652.View ArticlePubMedGoogle Scholar
- Clark SJ, Harrison J, Frommer M: CpNpG methylation in mammalian cells. Nat Genet. 1995, 10: 20-10.1038/ng0595-20.View ArticlePubMedGoogle Scholar
- Grunau C, Clark SJ, Rosenthal A: Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucl Acids Res. 2001, 29: e65-5. 10.1093/nar/29.13.e65.PubMed CentralView ArticlePubMedGoogle Scholar
- Warnecke PM, Stirzaker C, Melki JR, Millar DS, Paul CL, Clark SJ: Detection and measurement of PCR bias in quantitative methylation analysis of bisulphite-treated DNA. Nucl Acids Res. 1997, 25: 4422-4426. 10.1093/nar/25.21.4422.PubMed CentralView ArticlePubMedGoogle Scholar
- Chhibber A, Schroeder BG: Single molecule PCR reduces bias: application to DNA methylation analysis by bisulfite sequencing. Anal Biochem. 2008Google Scholar
- Warnecke PM, Stirzaker C, Song J, Grunau C, Melki JR, Clark SJ: Identification and resolution of artifacts in bisulfite sequencing. Methods. 2002, 27: 101-10.1016/S1046-2023(02)00060-9.View ArticlePubMedGoogle Scholar
- Kraytsberg Y, Khrapko K: Single-molecule PCR: an artifact-free PCR approach for the analysis of somatic mutations. Expert Rev Mol Diagn. 2005, 5: 809-15. 10.1586/14737188.8.131.529.View ArticlePubMedGoogle Scholar
- Daskalakis M, Nguyen TT, Nguyen C, Guldberg P, Kohler G, Wijermans P, Jones PA, Lubbert M: Demethylation of a hypermethylated P15/INK4B gene in patients with myelodysplastic syndrome by 5-Aza-2'-deoxycytidine (decitabine) treatment. Blood. 2002, 100: 2957-2964. 10.1182/blood.V100.8.2957.View ArticlePubMedGoogle Scholar
- Colella S, Shen L, Baggerly KA, Issa J-PJ, Krahe R: Sensitive and quantitative universal Pyrosequencing methylation analysis of CpG sites. Biotechniques. 2003, 35: 146-150.PubMedGoogle Scholar
- Tost J, Dunker J, Gut I: Analysis and quantification of multiple methylation variable positions in CpG islands by pyrosequencing. Biotechniques. 2003, 35: 152-156.PubMedGoogle Scholar
- Uhlmann K, Brinckmann A, Toliat MR, Ritter H, Nrnberg P: Evaluation of a potential epigenetic biomarker by quantitative methyl-single nucleotide polymorphism analysis. Electrophoresis. 2002, 23: 4072-4079. 10.1002/elps.200290023.View ArticlePubMedGoogle Scholar
- Yang AS, Doshi KD, Choi SW, Mason JB, Mannari RK, Gharybian V, Luna R, Rashid A, Shen L, Estecio MR, Kantarjian HM, Garcia-Manero G, Issa JP: DNA methylation changes after 5-aza-2'-deoxycytidine therapy in patients with leukemia. Cancer Res. 2006, 66: 5495-5503. 10.1158/0008-5472.CAN-05-2385.View ArticlePubMedGoogle Scholar
- Mikeska T, Bock C, El-Maarri O, Hübner A, Ehrentraut D, Schramm J, Felsberg J, Kahl P, Büttner R, Pietsch T, Waha A: Optimization of quantitative MGMT promoter methylation analysis using pyrosequencing and combined bisulfite restriction analysis. J Mol Diagn. 2007, 9: 368-381. 10.2353/jmoldx.2007.060167.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.