Characterization of Dendritic Cell Subsets Through Gene Expression Analysis

Dendritic cells (DCs) are immune sentinels of the body and play a key role in the orchestration of the communication between the innate and the adaptive immune systems. DCs can polarize innate and adaptive immunity toward a variety of functions, sometimes with opposite roles in the overall control of immune responses (e.g., tolerance or immunosuppression versus immunity) or in the balance between various defense mechanisms promoting the control of different types of pathogens (e.g., antiviral versus antibacterial versus anti-worm immunity). These multiple DC functions result both from the plasticity of individual DC to exert different activities and from the existence of various DC subsets specialized in distinct functions. Functional genomics represents a powerful, unbiased, approach to better characterize these two levels of DC plasticity and to decipher its molecular regulation. Indeed, more and more experimental immunologists are generating high-throughput data in order to better characterize different states of DC based, for example, on their belonging to a specific subpopulation and/or on their exposure to specific stimuli and/or on their ability to exert a specific function. However, the interpretation of this wealth of data is severely hampered by the bottleneck of their bioinformatics analysis. Indeed, most experimental immunologists lack advanced computational or bioinformatics expertise and do not know how to translate raw gene expression data into potential biological meaning. Moreover, subcontracting such analyses is generally disappointing or financially not sustainable, since companies generally propose canonical analysis pipelines that are often unadapted for the structure of the data to analyze or for the precise type of questions asked. Hence, there is an important need of democratization of the bioinformatics analyses of gene expression profiling studies, in order to accelerate interpretation of the results by the researchers at the origin of the research project, of the data and who know best the underlying biology. This chapter will focus on the analysis of DC subset transcriptomes as measured by microarrays. We will show that simple bioinformatics procedures, applied one after the other in the framework of a pipeline, can lead to the characterization of DC subsets. We will develop two tutorials based on the reanalysis of public gene expression data. The first tutorial aims at illustrating a strategy for establishing the identity of DC subsets studied in a novel context, here their in vitro generation in cultures of human CD34(+) hematopoietic progenitors. The second tutorial aims at illustrating how to perform a posteriori bioinformatics analyses in order to evaluate the risk of contamination or of improper identification of DC subsets during preparation of biological samples, such that this information is taken into account in the final interpretation of the data and can eventually help to redesign the sampling strategy.