Quantitative Microscopy to Measure the Nuclear Architecture

Advances in light microscopy lead to breakthroughs in biology. To further unravel the mysteries of life and the mechanisms behind diseases, better microscope techniques are needed to validate biological hypotheses. This thesis presents how integration of optics and computing leads to better images for quantitative microscopy and how quantitative microscopy is used to quantify the nuclear reorganization in Hodgkin's lymphoma, to relate chromosomal proximity to translocation frequency, and how the morphology of the nuclear lamina deteriorates when cells age, become senescent or become apoptotic. Structured illumination microscopy (SIM) is a superresolution technique in which the fundamental limits of resolution and optical sectioning can be circumvented by ``coding'' the information using patterned illumination followed by ``decoding'' using computational reconstruction methods. Current SIM reconstruction methods consist of a patchwork of individual steps and still yield image artifacts that are suppressed in an ad-hoc manner. This thesis presents a unified computational framework for image reconstruction and image filtering in 2D and 3D SIM. The method optimizes a trade-off between resolution and non-negativity. This is performed by incorporating Lukosz-bound apodization in Tikhonov-Miller filtering. The new method greatly reduces image artifacts such as ringing and negative intensities while preserving the resolution gain. This allows for a better interpretation of SIM data, because ringing artifacts could otherwise be interpreted as actual structure. The nuclear reorganization of DNA has been studied at both the chromosome and the individual gene level. Visualization and quantification of the clustering of all DNA in the cell, and the changes thereof in cancer cells, has been made possible with structured illumination microscopy. Understanding the architectural changes in cancer provides insight in the nature of the disease. This thesis presents a method to measure the size distribution of DNA structure and the structure of the DNA-free space as well as the presence of upstream binding factor (UBF), a protein associated with nucleoli. The progressive nature of Hodgkin's lymphoma (HL) is quantified by studying the DNA structure in normal, Hodgkin and multi-nucleated Reed-Sternberg cells. The chromosomal translocation frequency in cells is related to the cells' propensity to form tumors. This thesis will study the relation between translocation frequency and chromosomal proximity for c-myc-activating translocations in murine cells. We found a clear relation between chromosomal proximity and translocation frequency. Nuclear material is separated from the rest of the cell by the nuclear lamina. The nuclear lamina deteriorates when cells age, mutate, become senescent or apoptotic. A measurement of the morphology of the lamina enables the assessment of the state of cells. This thesis will describe how to identify the cell state by measuring the morphology and intensity over the lamina for a range of different cells. The robustness of the method is illustrated by the identification of three distinct groups in a collection of 9 different cell types