ABSTRACT: Treatment of PBMC with 0.6 pM cytokine from 0-24h Abstract: Background Interferons are key modulators of the immune system, and are central to the control of many diseases. The response of immune cells to stimuli in complex populations is the product of direct and indirect effects, homotypic and heterotypic cell interactions. Dissecting the global transcriptional profiles of immune cell populations may provide insights into this regulatory interplay. The host transcriptional response may also be useful in discriminating between disease states, and in understanding pathophysiology. The transcriptional programs of cell populations in health therefore provide a paradigm for deconvoluting disease-associated gene expression profiles. Results: We used human cDNA microarrays to (1) compare the gene expression programs in human peripheral blood mononuclear cells (PBMCs) elicited by 6 major mediators of the immune response: interferons a, b, w and g, IL12 and TNFa; and (2) characterise the transcriptional responses of purified immune cell populations (CD4+ and CD8+ T cells, B cells, NK cells and monocytes) to IFNg stimulation. We defined a highly stereotyped response to type I interferons, while responses to IFNg and IL12 were largely restricted to a subset of type I interferon-inducible genes. TNFa stimulation resulted in a distinct pattern of gene expression. Cell type-specific transcriptional programs were identified, highlighting the pronounced response of monocytes to IFNg, and emergent properties associated with IFN-mediated activation of mixed cell populations. Conclusions: This information provides a detailed view of cellular activation by immune mediators, and contributes an interpretive framework for the detection and definition of host immune responses in a variety of disease settings. PBMC extraction and stimulation Human primary peripheral blood mononuclear cells (PBMCs) were purified from whole blood of healthy donors using Ficoll-Paque PLUS (GE Healthcare) according to manufacturers instructions. PBMCs were incubated at 1.5-2.0 x 106 cells/well for 24 h before stimulation in RPMI 1640 medium supplemented with 10% heat-inactivated foetal calf serum and 2 mM L-glutamine (Invitrogen) at 37C, 5% CO2. Cells were treated with 0.006 pM, 0.6 pM or 60 pM recombinant IFNg (R&D Systems), and sampled at time intervals from 0.5 h to 12 h after stimulation. Additionally, cells were treated with 0.1 % BSA/PBS alone and used for untreated (mock) control time courses. RNA extraction and amplification Total RNA was extracted in TRIzol LS (Invitrogen) followed by standard chloroform purification and isopropanol precipitation. RNA was resuspended in RNase-free water, quantitated on the NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies) and stored at -80C. 500 ng total RNA was amplified using the MessageAmp modified Eberwine linear amplification procedure (Applied Biosystems). All samples to be compared were extracted and amplified together. Microarray analysis 4 ug amplified RNA was labelled with Cy5-dUTP (GE Healthcare) and combined with 3 ug of Cy3-labelled reference cDNA derived from a pool of RNA derived from a panel of 11 human cell lines (Stratagene Universal Human Reference RNA). The samples were washed and concentrated using MinElute columns (Qiagen) and competitively hybridised to custom printed cDNA microarrays containing 37,632 elements for approximately 18,000 unique human genes. The hybridised slides were scanned using a GenePix 4000A microarray scanner (Axon Instruments). Comparative spot intensities were calculated from the images, and areas of poor quality excluded from further analysis using GenePix Pro 6.0 (Axon Instruments). Data were deposited in the Stanford Microarray Database. Analysis was restricted to cDNA elements with a regression correlation of > 0.6, fluorescence intensities of > 2.5 fold signal/background in Cy3 or Cy5 channels and a minimum signal intensity of > 100 in both channels for at least 80% of the arrays. The expression ratios were normalised for array variation, and the data zero-transformed using a custom-designed Microsoft Excel macro (C. Liu, Stanford). The statistical package SAM (Significance Analysis of Microarrays, version 1.15) was used to identify genes significantly differentially expressed in the normalised data sets by pairwise comparison with a minimum 2 fold cutoff at a false discovery rate of < 1% of the median. The transformed datasets were then hierarchically clustered using Cluster 2.11 and the results displayed using Treeview 1.60.
