Investigation of epigenetics in cell populations using three-dimensional quantitative DNA methylation imaging

Aberrations in DNA methylation patterns have been identified between “healthy” and “diseased” cells, especially in cancer cells and other cells exhibiting complex human syndromes, such as Fragile X syndrome, Praeder-Willi syndrome, ICF syndrome, etc. Improvements in recognizing these different methylation patterns could lead to breakthroughs in relevant disease diagnostics and prognostics. A great majority of methods currently available for assessing DNA methylation patterns focuses on molecular profiling of large volumes of cells and, thus, is not ideal for use in pathology, where understanding variations in methylation profiles of individual cells within a population is important. Although single cell profiling may be possible in newer generation microarray-based assays, sequencing of methylation patterns remains highly time-consuming and very costly, when applied in cell-by-cell fashion. Towards the ultimate goal of creating a high-throughput assay that could be used in translational medicine, imaging-based, quantitative, and cytometric analysis of DNA methylation patterns entitled “3-Dimensional Quantitative DNA Methylation Imaging (3D-qDMI)” has been developed. This dissertation investigates the feasibility of using such assay to delineate and quantify in situ global DNA methylation features, such as nuclear load and spatial distribution of methylated cytosine (MeC). ❧ 3D-qDMI entails highly specific immunofluorescence methods, high-resolution imaging via confocal laser-scanning microscopy, and dedicated software package designed to analyze fluorescence intensity patterns at both the cellular and population levels. Immunofluorescence staining of specimens is performed by monoclonal anti-MeC primary antibody, which has been shown to possess both the desired specificity and sensitivity needed for a DNA methylation assay. The specimens were counter-stained with DAPI, a non-specific dye, to visualize the inhomogeneous distribution of genomic DNA. The counter-stain establishes the baseline of spatial genomic density, which was then used to qualify and quantify MeC fluorescence. The stained specimens were imaged by high-resolution confocal laser scanning microscopy, which has a lateral resolution of ~ 150 nm and an axial resolution of ~ 500 nm for the conditions used in this dissertation. The images obtained were analyzed with automated and high-throughput software specifically developed to meet the needs of 3D-qDMI without compromising data reliability. The image analysis methods involve a pre-processing step to delineate nuclear regions of interest (ROIs) and three different analytical modules to assess both cellular-level topological co-distribution of DAPI and MeC signals and population-level homogeneity. Each of the novel analytical modules was validated using sets of images with explicitly known characteristics. These analytical modules work independent from each other and have been designed to be flexible, so that the end user has the options to choose which specific modules to run. ❧ 3D-qDMI was then tested for its ability to distinguish and discriminate cells based on their DNA methylation patterns, in cells with different proliferative capacities. Primary cells were cultured until replicative senescence was reached, while cancer cells were grown to provide sufficient samples, as necessary, to identify their growth rates. Separately, stress-induced premature senescence was studied via application of H₂O₂. Results from these studies showed that as the proliferative capacity of cells decreased, global hypomethylation and chromatin re-organization were observed. Global hypomethylation could also be identified in cancer cells, but distinct differences could be observed between cancer cells and primary cells reaching senescence for various topological parameters used in 3D-qDMI. These results suggest that these differences could be utilized in discriminating primary cells populations from cancer cell populations. Overall, the results from these feasibility studies confirmed that an imaging-based assay can be utilized to discern DNA methylation patterns of different cells. ❧ This dissertation demonstrates that an advanced optical imaging-based assay can be used to assess DNA methylation patterns at both single cell-level and population-level. The results found through 3D-qDMI were on par with previously published findings from molecular-based assays. 3D-qDMI is not developed to replace molecular-based assays: the current methods as presented in this dissertation cannot identify which DNA sequences are methylated nor determine the absolute number of MeC residues in the genome. Instead, 3D-qDMI was developed to address the need for a high-throughput, cost-effective assay for DNA methylation for use in pathology. Unlike the commonly available cellular phenotyping methods for tumor cells in biopsy samples, which rely on a static, morphology-based evaluation by a trained physician, 3D-qDMI delivers a dynamic and activity-based picture of cells. Further studies and development of the methods presented in this dissertation should allow for the ability to distinguish “healthy” cells from “diseased” cells, which could drastically change how clinical pathology is current performed.