3.1. DNA METHYLATION AS A MARKER OF MALIGNANT DISEASE Tumors show characteristic profiles of methylated markers that are specific for given types and subtypes of cancer (32,97). Thus, there is considerable potential for using methylation as a tumor-specific marker (98). Methylation markers are likely to be the best DNA-based markers because there are no useful high-frequency, single-site mutations known for most cancers. With the right choice of high-frequency methylated genes, tumor cells or DNA can be identified without knowledge of which genes are methylated in the primary tumor.
Currently, the best example of a tumor-specific methylation marker is the GSTP1 gene, which is methylated in 95% of prostate cancers but not in benign prostatic hyperplasia (99). For other cancers, markers showing this degree of specificity remain to be identified. An alternative approach is to use a small panel of recurrently methylated markers rather than a single marker. By assembling a panel of four to five loci, it might be possible to cover close to 100% of tumors. It is critical to verify that the markers are generally not methylated to any significant degree in normal tissue.
Methylation markers can be used for early detection of cancer or for monitoring response to treatment. Sensitive techniques can be used to detect methylation in biological fluids such as (1) blood for the detection of disseminated carcinoma cells or tumor-derived DNA, or (2) various lumenal fluids such as urine for the detection of cells from bladder, kidney, and prostate cancers, or (3) washings such as bronchioalveolar and ductal lavages for the detection of early malignant cells in lung and breast cancer, respectively.
The detection of methylated sequences in DNA derived from the plasma or serum is a potentially important early detection and monitoring method for cancer (100). DNA alterations have been detected in patients with small or even in situ lesions, indicating that tumor DNA is shed into circulation early in the disease. This DNA probably derives from necrosis and apopto-sis of the tumor cells. In one study, plasma DNA concentration in normal individuals averaged less than 4 ng/mL, whereas the plasma DNA concentration in breast cancer patients ranged from 20 to 360 ng/mL (101).
The utility of methylation as a tumor marker has been demonstrated by numerous studies, including the following examples. Methylation of the p16INK4A tumor suppressor gene, which was shown to be an early change in lung cancer, was detected in the sputum of 3/7 patients with squamous cell carcinoma and 5/26 high-risk cancer-free individuals (102). In patients with resectable non-small-cell lung cancer, p16 methy-lation was detected in the in bronchioalveolar lavage fluid of 12/19 patients whose tumor showed p16 methylation (103). In patients with prostate cancer , GSTP1 promoter hypermethyla-tion was detected in 72% of plasma or serum samples, 50% of ejaculates, and 36% of urine samples (after prostate massage) (104). In ductal lavage specimens, methylated alleles of the cyclin D2 (CCND2) and RARfi2 genes were detected in fluid from patients with endoscopically detected carcinomas and ductal carcinoma in situ but rarely in fluid from healthy ducts (105).
3.2. DETERMINATION OF CLONALITY The clonality of tumors in females can be determined using methods that utilize methylation differences between the active and inactive X chromosomes (106). The X chromosomes are genetically distinguished using single-nucleotide polymorphisms (107) or by variation in copy number of a trinucleotide repeat of the androgen receptor locus (108). The technique uses a methylation-sensitive restriction enzyme to determine which allele of the polymorphic locus on the X chromosome is methylated. The unmethylated allele is digested and, consequently, is not ampli-fiable (107). Clonal cell populations, which all have the same allele of the X chromosome inactivated, will show preferential loss of one allele after restriction digestion and PCR amplification. One problem in using the clonality determination approach is that there might be significant variation in the balance or skewing of X chromosome inactivation in normal individuals. Appropriate normal tissue controls from the same individuals should always be employed as a control (109).
3.3. DIAGNOSIS OF IMPRINTING DISORDERS The majority of Prader-Willi syndrome and Angelman syndrome cases have lesions that effectively remove one of the copies of a region in chromosome 15q11-q13 (e.g., deletions or uniparental disomy). Prader-Willi syndrome arises from loss of the paternally derived copy of this region, whereas Angelman syndrome arises from loss of the maternally derived copy of this region. Both disorders are characterized at the molecular level by abnormal methylation of imprinted genes at 15q11-q13, including the small nuclear ribonucleoprotein N gene (SNRPN). The promoter region CpG island of the SNRPN gene is heavily methylated in the maternally derived allele and unmethylated in the paternally derived allele. In patients with Prader-Willi syndrome, only the methylated allele is present, whereas in patients with Angelman syndrome, only the unmethylated allele is present. Several MSP-based assays have been designed to evaluate the methylation status of the SNRPN CpG island for rapid diagnosis of these two syndromes. Bisulfite-modified DNA from patients with Prader-Willi syndrome only amplify with the methylation-specific pair, whereas modified DNA from patients with Angelman syndrome only amplify with the primers specific for the unmethylated sequence. Modified DNA from normal individuals amplifies with both primer pairs (110-112).
More recently, DHPLC-based approaches have been used to differentiate the two alleles. Both the methylated and unmethy-lated alleles are amplified at the same time and separated by DHPLC. The sensitivity and semiquantitative properties of DHPLC are able to detect mosaicism more readily than MSP-based approaches. The DHPLC approach has been also applied to other imprinted regions where a methylation difference is also present (e.g., in the Beckwith-Wiedemann syndrome) (113,114).
3.4. DNA METHYLATION AS A PREDICTIVE MARKER Whereas de novo methylation of the promoter regions of DNA repair genes can lead to accelerated carcinogenesis by increasing the mutation rate, this loss of DNA repair capacity can also be the Achille's heel for the tumor during chemotherapy and or radiotherapy. Many therapies targeting cancer cells are effective because they cause DNA damage. Knowledge of the affected pathways can thus lead to rational choice of therapeutic agent and allow the prediction of "responders" and "nonre-sponders."
A compelling example is methylation of the O6 methylgua-nine DNA methyltransferase (MGMT) gene, which removes small alkyl groups from the O6 position of guanine. MGMT is methylated in a variety of cancers (115). In gliomas and diffuse large B-cell lymphomas, MGMT promoter methylation has been shown to be associated with response to chemotherapy with alkylating agents, which are now much more toxic to the deficient cells (116,117). Thus, loss of activity of MGMT, which initially favors tumor progression, is now responsible for the tumor's exquisite sensitivity to alkylating agents.
Methylation of MLH1, which leads to deficient mismatch repair in sporadic colon tumors, is another example of how knowledge of the lesion can suggest specific therapy. Mismatch repair (MMR) acts to recognize and process not only single-basepair mismatches and insertion-deletion loops that occur during DNA replication but also DNA adducts such as those resulting from treatment with cancer chemotherapy agents (118). However, whereas base mismatches and insertion-deletion loops are repaired by MMR, MMR-mediated recognition and processing of chemotherapy-induced adducts in DNA results in apoptosis.
The inability of MMR-deficient cells to recognize chemotherapy-induced adducts in DNA leads to resistance to alkylating agents. Interestingly, MMR-positive tumors are more sensitive to standard chemotherapies for colorectal cancer using 5-fluorouracil (119,120). This validates the standard chemotherapy for colon cancer in this group of tumors but suggests that the distinct group with MGMT methylation (121) might benefit from therapy with alkylating agents.
3.5. DNA METHYLATION AND THE EVALUATION OF HNPCC A particular problem in the study of hereditary nonpolyposis colorectal cancer (HNPCC) is deciding which patients to screen for mutations in one or more of the MLH1, MSH2, and MSH6 genes. Pedigrees first need to meet strict criteria such as the Amsterdam criteria to establish a high likelihood of hereditary cancer. A tumor from an affected member of the pedigree is then screened by examination of a panel of mononucleotide and dinucleotide repeats to determine microsatellite instability (MSI) and by immunohisto-chemistry for the MLH1, MSH2, and MSH6 enzymes. If the tumor is scored as MSI high and/or immunohistochemistry for one or more enzymes is negative, mutation screening is undertaken.
The MLH1 locus undergoes frequent methylation in sporadic colorectal tumors (37,38). Both MLH1-methylated tumors and tumors arising from patients with a germline MLH1 mutation have the same phenotype: MSI high and negative immunohistochemistry for the MLH1 protein. However, the identification of MLH1 methylation in a tumor does not disprove its hereditary origin, as methylation can occur as the second event in a patient with a germline mutation. The absence of MLH1 methylation in tumors of this phenotype is associated with a very strong likelihood of there being a germline mutation. The decision whether to proceed with germline testing should be based on all the known clinical and pathological information (122).
The second case when methylation status can be used in the context of HNPCC screening is that of MSI low tumors. These present somewhat of a diagnostic dilemma because it has been suggested that many tumors identified by the National Cancer Institute panel of five markers (123) as MSI high (defined as those tumors that show MSI for more than one marker) are, in fact, MSI low (defined as those tumors that show MSI for only one marker). Certain dinucleotide and tetranucleotide markers show instability at higher than expected frequencies in non-MSI high cancers (e.g., D2S123, MYCL, and D17S250). The same markers have been recommended for use in the NCI panel. Should two of these markers be mutated in the same cancer, then that cancer can be classified as MSI high. However, such cancers do not show immunohistochemical loss of DNA
MMR proteins, instability in mononucleotide markers, or the clinical and pathological features that have been associated with MSI high cancers (122). Therefore, these cancers should probably be included with MSI-low and not MSI-high cancers. It is considered that in some cases that methylation of the MGMT locus can give rise to the MSI-low phenotype (121). In this case, identification of MGMT methylation in a MSI-low tumor might confirm the decision not to go ahead with mutation screening.
3.6. QUALITY CONTROL ISSUES All of the quality control issues that apply to PCR-based tests are applicable to the methods described here. As with all PCR, utmost care must be taken to ensure that there is no carryover of PCR amplicons into the PCR setup areas. In addition, a sample of unmodified DNA should always be run to ensure that no amplification is observed under the conditions being used. This is particularly important to avoid false positives when MSP is being performed. The appropriate methylated and unmethylated controls should be used. It is convenient to use cell lines with known methylation status for this. If no positive control cell line is available, Sssl methylase can be used to methylate all CpG residues in any available DNA. Normal human DNA is preferable to human cancer cell line DNA as a substrate for Sssl methylation, as there is always the possibility that a cell line contains homozygous deletions for the regions of interest.
A second set of issues arises when quantitation is important. Preferential amplification of either methylated or unmethylated sequences (PCR bias) could occur (124). Usually, methylated sequences amplify less readily than unmethylated sequences, because the higher C+G content of the PCR products leads to their more inefficient denaturation. Mixing known ratios of methylated to unmethylated targets to construct a standard curve can be used to both assess and to compensate for this problem.
Was this article helpful?
Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.