Reconstruction of gene regulatory modules from RNA silencing of IFN-α modulators: experimental set-up and inference method

Abstract

BackgroundInference of gene regulation from expression data may help to unravel regulatory mechanisms involved in complex diseases or in the action of specific drugs. A challenging task for many researchers working in the field of systems biology is to build up an experiment with a limited budget and produce a dataset suitable to reconstruct putative regulatory modules worth of biological validation.ResultsHere, we focus on small-scale gene expression screens and we introduce a novel experimental set-up and a customized method of analysis to make inference on regulatory modules starting from genetic perturbation data, e.g. knockdown and overexpression data. To illustrate the utility of our strategy, it was applied to produce and analyze a dataset of quantitative real-time RT-PCR data, in which interferon-α (IFN-α) transcriptional response in endothelial cells is investigated by RNA silencing of two candidate IFN-α modulators, STAT1 and IFIH1. A putative regulatory module was reconstructed by our method, revealing an intriguing feed-forward loop, in which STAT1 regulates IFIH1 and they both negatively regulate IFNAR1. STAT1 regulation on IFNAR1 was object of experimental validation at the protein level.ConclusionsDetailed description of the experimental set-up and of the analysis procedure is reported, with the intent to be of inspiration for other scientists who want to realize similar experiments to reconstruct gene regulatory modules starting from perturbations of possible regulators. Application of our approach to the study of IFN-α transcriptional response modulators in endothelial cells has led to many interesting novel findings and new biological hypotheses worth of validation.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-016-2525-5) contains supplementary material, which is available to authorized users.

Genetic perturbation of IFN-α transcriptional modulators in human endothelial cells uncovers pivotal regulators of angiogenesis

“…RNAi-mediated knockdown of seven established (IRF1, IRF7, STAT1, USP18) or candidate (GBP1, OAS2, IFIH1) IFN-α modulators [29] , [31] , [32] , [33] , [34] was achieved using Stealth siRNAs and a panel of core IFN-α regulated genes was monitored by TaqMan Array Cards (see Materials and Methods). Statistical analysis showed STAT1 and IFIH1 as mainly positive regulators, and USP18, GBP1 and IRF1 as mainly negative regulators of IFN-α transcriptional response.…”

“…Cells were stimulated with human recombinant IFN-α2 (Merck & Co., White House Station, NJ, USA) using a final concentration of 1,000 U/ml. This concentration was chosen based on our previous studies [22] , [29] . Where indicated, HUVECs were treated with the JAK inhibitor ruxolitinib (2 µM, Selleckchem, Munich, Germany).…”

“…The criteria used for the pre-selection of transcripts were described in our previous work, Grassi et al . [29] . qRT-PCR reactions were run on an ABI Prism 7900HT Sequence Detection System (Life Technologies) and raw data were extracted using the SDS v2.4 software package (Life Technologies).…”

“…The effects of each siRNA were evaluated with respect to a calibrator siRNA using the comparative threshold cycle method ( method). The significance analysis, developed to elicit significant regulations induced by each perturbation, and the inference method, used to reconstruct a regulatory subnetwork, were originally described in [29] . Briefly, for the significance analysis we proposed a two-stage selection procedure that first filters observations based on the variance distribution, and then performs a variable-by-variable statistical test procedure, which uses the biological variance estimated through a measurement error model, to assign a p-value to each modulation.…”

“…If a gene was significantly modulated during both stimulation and wash-out phase, only the regulation in the stimulation phase was reported in the reconstructed network. We refer to [29] for the complete description of the inference method to reconstruct putative multi-output feed-forward loop (FFL) subnetworks.…”