The rationale of the cytokine microarray project

The aim of this project is to develop, apply and adapt DNA microarray technology for quantitative and kinetic measurements of cytokine-induced gene expression in cell culture systems, transgenic mice, animal models of disease and in human patients that are studied by members of the SFB566.

In the first phase of the project, we decided to design, to produce and to characterize our own DNA microarrays. To utilize the microarray in mouse and man a gene list was selected with a maximal overlap of functionally identical genes. Initially this microarray covered mainly genes induced during inflammation. To enable many experiments the microarray had to be produced at low cost; hence we limited the number of genes initially to about 100 highly regulated inflammatory genes (including many cytokines and cytokine receptors; see overview). To keep the microarray versatile and flexible we chose MWG Biotech's oligonucleotide platform and also established a close cooperation with MWG Biotech and later with the company Ocimum. To achieve maximal sensitivity and for highly reproducible results we chose to design three oligonucleotide probes per gene which were selected by algorithms developed at MWG Biotech as well as by extensive data bank comparisons. To validate the probes the microarray was then tested in a large variety of biological systems in man and mouse and a strategy was developed to select the best probe per gene (see poster of our data presented at the International Cytokine Society Meeting 2003 in Dublin gif 349 kB). Final validation of the human and murine probes was performed in 214 and 87 hybridizations, respectively. The underlying material for the validation experiments was derived from the participating projects of the SFB566 and external collaborators. Altogether more than 25 different cell types (view list) and more than 30 different stimuli (view list) were used.

Since 2006, the low density microarray platform was largely replaced by using the entire Agilent microarray work flow. Price reductions and the very good performance of their microarrays (see link: Available microarrays) as well as their accompanying hardware has resulted in the substitution of low density microarrays by high-density microarrays and by genome-wide analyses of gene expression in most of our studies. The latter are now being performed in a highly standardized way on a routine basis.

In addition to providing the basic services of a central microarray facility, special care is taken to optimize the experimental design of the individual microarray studies by offering as much advice as possible to the collaborating research groups. All microarray experiments are conducted in a highly standardized fashion by our microarray laboratory. All raw data are thoroughly analyzed, discussed with the cooperation partners and then stored in a data base (called CytoBASE) that is accessible via a WEB-based interface to all users. There is free access to a large set of validation experiments and to many of the already published microarray experiments. Hence, we have been establishing a continuously growing repository of mRNA expression data. We have also explored and utilized multiple ways of visualizing results from data analyses. This experience has proven to be an indispensable aspect of our work to assist researchers who use the microarray lab in drawing meaningful conclusions from their experiments.

Ultimately this data base forms the nucleus to comprehensively follow the changes in cytokine-networks and cytokine-dependent gene expression in health and disease. As a long term goal, by bioinformatics analysis all data will be integrated electronically into models of cytokine networks and the signaling circuits regulating them. On the basis of this information key proteins in strategically prominent locations can then be targeted by RNAi, dominant-negative mutants, and, eventually, in vivo by genetic strategies developed within the SFB566 to experimentally perturb the response of targeted cells and thus verify/modify the circuit model accordingly. The accuracy of this model will be evaluated further using in vivo models of cytokine-dependent diseases such as chronic inflammation or infection.

We also try to collaborate with as many research groups as possible who work in related areas. We are also interested in obtaining samples from patients or animal models. 

Please contact us if you are interested.

Michael Kracht and Oliver Dittrich-Breiholz
(project leaders responsible for the content of these sites).