Methylation levels of normal and chronic lymphocytic leukemia B lymphocytes: computer‐assisted quantitative analysis of anti‐5‐methylcytosine antibody binding to individual nuclei

Abstract

The role played by a consecutive series of DNA modifications in human cancer cells is being progressively uncovered. A mutator phenotype causing the observed higher rates of spontaneous mutations may be required for genomic instability and the onset of multistep carcinogenesis (1, 2). Hypermethylation of tumor suppressor genes and/or hypomethylation of specific growthpromoting proto-oncogenes may contribute to the selective advantage of neoplastic cells and thus represent one of the steps involved in the onset of malignancy. Hypomethylation of proto-oncogenes was observed in liver tumors and leukemias (3). In several reports reduced levels of global DNA methylation in tumor cell populations were described (3, 4). Hypomethylation has been found to be positively correlated with both malignant transformation and tumor progression (4). We have recently developed a computer-assisted method (5) based on the indirect immunolabeling by anti-5-MeC antibodies (6) to quantify eu-and heterochromatin methylation in individual nuclei. The reliability of the quantitative approach in the evaluation of methylation levels is further confirmed by the good agreement of molecular and cytological results obtained in experimentally hypermethylated mouse cell cultures (7).In the present report experiments were set up to assess the application of such technology to human primary lymphocyte subsets. Two sets of cytospins were prepared from Lymphoprep (Nycomed Pharma AS, Oslo, Norway) separated mononuclear cells of three normal individuals following removal of adherent cells. For each donor, in one set of cytospins the CLB-BLY1 (CD22) monoclonal antibody (Ortho Diagnostics, Raritan, NJ) was used to detect B cells, and in the other the UCHT1 (CD3) monoclonal antibody (Dako, Glostrup, Denmark) was used to recognize T lymphocytes. The immunocytochemical reaction was performed with the immunoperoxidase technique using Dako reagents as previously described (7). Diaminobenzidine (DAB, Sigma, Milan, Italy), which gave a deep brown staining of the labeled cells, was used as the substrate.All slides were then exposed to UV light for DNA denaturation, followed by indirect immunolabeling with the monoclonal anti-5-MeC antibodies (8) and subse-