DNA hybridization arrays: a powerful technology for nutritional and obesity research.

Obesity is a serious and growing public health problem throughout much of the world. In fact, obesity has reached epidemic proportions in many developed countries. Increased food intake and decreased energy expenditure due to the sedentary lifestyle of our developed societies have contributed to the widespread development of obesity. However, these environmental factors cannot totally explain the development of the disorder. In fact, genetic predisposition for obesity may underlie the increased tendency for weight gain in some individuals (Levin, 2000; Marti et al. 2000). Recent research has identified over 200 genes, markers and chromosomal regions associated with or linked to human obesity (Chagnon et al. 2000). The contribution of genetic background to obese phenotypes is a complex question yet to be answered. The major aim, therefore, is to study those combinations of genes and mutations that are implicated in the development of obesity in human subjects. In order to be able to control the spread of obesity, it is necessary to establish how environmental factors affect the onset of obesity in individuals with a specific genetic background. With potentially thousands of genes examined in a single hybridization, DNA hybridization array technology constitutes a powerful tool to identify and compare patterns of gene expression (Lander, 1999; Yang et al. 1999). Thousands of DNA spots are robotically deposited on a solid-state matrix (nylon membranes, glass microscope slides, silicon or ceramic chips). These matrixes are then hybridized with labelled cDNA representing total RNA pools from test and reference samples. The differential expression of genes between both test and reference samples are detected based on the fact that the higher the expression of a gene, the higher the intensity of the output labelled signal (Schena et al. 1995, 1996; Duggan et al. 1999). An array starts with the selection of the ‘probes’ to be printed in the array. In many cases these probes are chosen directly from databases like Genbank (Benson et al. 1998), dbEST (Boguski et al. 1993) and UniGene (Schuler et al. 1996), the resource backbones of the array technologies (Duggan et al. 1999). There are a variety of options for making arrays (Bowtell, 1999; Cheung et al. 1999), with four different array types developed according to the different characteristics of the matrix, probe number and density, array size and type of label. These include macroarrays, microarrays, high-density oligonucleotide arrays and microelectronic arrays (Freeman et al. 2000). In the macroarrays the DNA probes are immobilized on a membrane-based matrix, the targets are radioactively labelled, and the different samples are hybridized on individual separate arrays. Phosphoimagers are then used to detect the amount of bound labelled target. Microarrays are produced by spotting up to 10 000 polymerase chain reaction products, or more, representing specific genes, onto a glass or plastic slide matrix (Schena et al. 1995). Following purification and quality control, a few nanolitres of the polymerase chain reaction product are printed on coated glass microscope slides using a computer-controlled high-speed robot. Total RNA from both the test and reference samples is labelled with fluorescent tags (for example, red and green dye) by a single round of reverse transcription. The matrix is then simultaneously hybridized with the two resulting fluorescently-labelled cDNA in a competitive manner, and fluorescence scanners are used for detection (Schena et al. 1995, 1996; Fig. 1). The highdensity oligonucleotide arrays differ from the microarrays in that the DNA probe is generated in situ on the surface of the matrix by a method called photolitography. These arrays (Genechipsw; Affymetrix, Santa Clara, CA, USA) can contain between 40 000 and 60 000 probes, providing the highest density of probes of any array. However, the probes are limited in length and therefore specificity, and a mismatch detection scheme has to be used to determine specific hybridization (Lipshutz et al. 1999; Freeman et al. 2000). The microelectronic arrays, the newest hybridization array, consist of sets of electrodes covered by a thin layer of agarose coupled with an affinity moiety which provides controlled electrophoretic fields (Cheng et al. 1998; Heller et al. 2000). Analysis and handling of the array data is one of the most difficult aspects in the utilization of the technology. The images of the scanner are imported into software which has to transform the fluorescence data into information about the clones, such as gene name, clone identifier, intensity values, intensity ratios, normalization constant and CI (Duggan et al. 1999). Clustering algorithms (hierarchical and nonhierarchical) have been the most used tools for analysing array data, and make it possible to find groups of genes or clusters with similar behaviour (Brazma & Vilo, 2000; Celis et al. 2000). Due to the adaptable nature of the fabrication and hybridization methods, these techniques have been widely applied. Hybridization arrays have been applied to the discovery of sets of genes that have roles in diseases such as cancer (DeRisi et al. 1996; Kononen et al. 1998; Cole et al. 1999; Marx, 2000; Sallinen et al. 2000), insulin resistance (Aitman et al. 1999), hypertension (Lee et al. 2000), rheumatoid arthritis and inflammatory bowel disease (Heller et al. 1997). British Journal of Nutrition (2001), 86, 119–122 DOI: 10.1079/BJN2001410 q The Author 2001