The Genome in Three Dimensions: From Basics to Diagnostics

Misteli T. Spatial positioning: A new dimension in genome function. Cell 119: 153–6, 2004.

hen we hear about genomes, we usually think of genome sequences. However, if genome sequencing projects have taught us one thing, it is that the sequence of a genome is not everything. The answers to some of the most fundamental questions, such as why humans have only four times as many genes as yeast and how the right genes are turned on at the right time in the right place, will likely not come from sequence information. Clearly, regulated genome function involves more than sequence. One regulatory factor in genome function that is gaining increasing recognition is the spatial organization of genomes within the interphase cell nucleus.

All non-chloroplastic, non-mitochondrial eukaryotic genomes exist and function in the cell nucleus. Advanced imaging methods have revealed that the interior of the human nucleus is divided into distinct neighborhoods where various functions such as transcription and RNA processing occur in spatially separate subcompartments. Given the existence of such functionally specialized regions, it seems obvious to consider that the placement of chromosomes and genes within the nucleus might contribute to their proper function and regulation. Yet, the study of how genomes are spatially organized and what this organization means for function is only in its infancy.

What we have learned is that genomes are indeed non-randomly organized. Some chromosomes have a tendency to localize toward the center of the nucleus, whereas others preferentially associate with the edges of the nucleus (Figure 1). This in turn places some chromosomes closer to others to form pairs and clusters, leading to the creation of defined genome neighborhoods. Since genes are located on chromosomes, it is not surprising to find that the position of genes is similarly non-random. The patterns of genome organization differ among cell types and tissues and might be related to the differential sets of genes, which are expressed in cells at various times during development and differentiation.

Figure 1. High-throughput genome imaging. Genomes are non-randomly positioned within the cell nucleus. Automated high-throughput imaging systems and mining of positioning information allows exploitation of chromosome- and gene-positioning patterns for basic discovery and diagnostic applications.

Efforts are currently under way in many laboratories to describe genome organization patterns and to ask how positioning is linked to gene function. The general approach in these studies is to correlate the expression pattern of a gene with its position relative to nuclear landmarks, such as the periphery, heterochromatin regions, or other genes. For example, we have shown that during T-cell differentiation, the CD4 locus moves from a peripheral position to a more internal nuclear position in correlation with its activity. This type of analysis can now be performed routinely for single genes; the next step is to map the positions of sets of genes (for example, those that have been identified as co-regulated) by use of microarray analysis.

There is also reason to believe that spatial genome organization plays a role in cancer. Recent experimental data from several laboratories demonstrate that in many tumors, including leukemias and liver tumors, the most frequent translocations occur between chromosomes that are generally in close proximity. This means that the non-random, spatial arrangement of the genome might predispose cells to particular translocation events.

The non-random position of genomic regions is of great interest as a fundamental mechanism in gene regulation, but analysis of spatial genome organization also has direct applications as a novel strategy in diagnostics. A disease-causing gene may be repositioned as its activity becomes aberrant. Similarly, gene markers in pre-malignant or pre-metastatic cells might already be misplaced prior to the cells becoming neoplastic. Spatial positioning might be a better early indicator than gene activity, because changes in positioning patterns often occur prior to changes in a gene’s activity. A particular advantage of such an interphase genome-positioning mapping method is its applicability to solid tumor samples, whose genomes currently cannot be easily analyzed. Since such solid tumors constitute the majority of all human tumors, positioning diagnostics would fill a significant gap in our diagnostic repertoire.

The full exploration of the spatial organization of the genome and the development of three-dimensional diagnostic methods requires technology development. The strategy is to implement high-throughput microscopy systems that can acquire large amounts of positioning data (Figure 1). These images will be automatically processed and analyzed using dedicated three-dimensional positioning software tools. The resulting distributions will then be analyzed by advanced pattern-recognition tools to correlate expression with position and to define patterns that are characteristic of a particular physiological state. Efforts to create such systems are currently under way at the NCI. These systems will eventually be used to generate extensive three-dimensional maps of genomes, to follow the changes in genome organization patterns during differentiation, development, and disease progression. Most importantly, they have a high potential to translate what we learn about fundamental genome organization to disease-relevant therapeutic applications.

Tom Misteli, PhD
Senior Investigator
Laboratory of Receptor Biology and Gene Expression
NCI-Bethesda, Bldg. 41/Rm. B610
Tel: 301-402-3959
Fax: 301-496-4951
mistelit@mail.nih.gov