In the O’Garra lab at The Francis Crick Institute, scientists study the signals that control our immune systems, to understand how they respond to diseases such as tuberculosis and what goes wrong when they damage healthy cells. Here, BSI members from the O'Garra lab introduce their new analytical tool - an open access app for researchers looking at how the immune response in the blood reflects the local response in the tissue, and vice versa. Through the app, researchers anywhere in the world can look up gene activity in the lungs and blood of mice infected with a range of pathogens, enabling rapid testing and generation of hypotheses by avoiding unnecessary additional mouse experiments.
Host immunity during infection and inflammation is complex, with a spectrum of responses having been reported across infections with parasites, viruses, bacteria, fungi or allergy.
These responses are often driven and dominated by specific groups of cytokines, with protective and/or damaging consequences for the host. In an attempt to understand host immune responses and identify genetic signatures associated with different diseases, transcriptomic approaches have been applied broadly to whole blood or peripheral blood mononuclear cells, which are readily obtained from patients.
However, little is known about how immune responses in blood are reflected at disease sites, from where human specimens are harder to obtain. Transcriptional studies of the tissue have been performed from different experimental models of disease, but this has been reported to a lesser extent for the blood. Additionally, transcriptional studies on the global immune responses spanning different experimental models of diseases across distinct types of immune responses, are scarce.
Building a framework
The O'Garra laboratory has been using transcriptomic approaches for more than a decade to advance the knowledge of the immune response and diagnosis in tuberculosis. Taking advantage of this expertise and in a major effort to provide a framework for discovery of pathways of gene regulation in disease, we worked with numerous collaborators from the UK and abroad to measure gene activity from RNASeq data across different diseases in both the whole tissue and blood. The goal? To see how the immune response in the blood reflects the local response in the tissue, and vice versa.
The data obtained from 10 different experimental mouse models of infectious and inflammatory diseases was integrated in a comprehensive resource of modular transcriptional signatures in blood and whole tissue. The aim was to identify commonalities and differences in the immune response to specific infections or challenges to aid the discovery of pathways in disease. Importantly, the data being accessible to anyone through an open access web app.
Using the new app, researchers anywhere in the world can look up gene activity in the lungs and blood of mice infected with a range of pathogens: the parasite Toxoplasma gondii, influenza A virus, respiratory syncytial virus (RSV), the bacterium Burkholderia pseudomallei, the fungus Candida albicans, or the allergen house dust mite (HDM). Blood transcriptional signatures from mice infected with Listeria monocytogenes, murine cytomegalovirus (MCMV), the malaria parasite Plasmodium chabaudi chabaudi, or a chronic infection with B. pseudomallei can also be interrogated through the app.
Developing the app
Using advanced and unbiased bioinformatics techniques, we clustered thousands of genes across the different disease models, based on their expression patterns and coregulation across all healthy and disease samples, into a biologically meaningful and visual form, which we refer to as modules. Thirty-eight lung and forty-one blood modules were derived as part of the study from samples from six mouse models of infection and inflammation, and annotated to determine their function and known physiological roles.
A broad range of immune responses was unveiled in the lung, where discrete modules were dominated by genes associated with type I or type II interferon (IFN), IL-17 or allergic responses (Figure1).
For example, of the 38 modules identified in the lung there is a discrete module associated with allergy containing over 100 genes that are overabundant only in the HDM allergic airway disease model. Indeed, each experimental model showed distinct immune responses. Type I and type II IFN-inducible genes were highly expressed in lungs of T. gondii, influenza A, and less so in RSV infected mice. Pathways driven by IL-17 were abundant only in the lungs of mice infected with B. pseudomallei and C. albicans, and a signature of Th2- type cytokines was abundant only in the lungs of mice challenged with HDM allergen.
Additionally, using raw RNAseq data from purified cells, obtained from the ImmGen consortium, we identified cell types associated with these modules. For example, there was a strong enrichment of mast cells within the allergy module that was observed in the HDM allergy model. These findings show that a broad spectrum of distinct immune response across different diseases can be demonstrated in whole tissue using bulk RNA-Sequencing approaches.
Type I IFN are known to be released in response to viruses, while type II IFN (IFN-g) activates phagocytes to kill intracellular pathogens, and IL-17 attracts neutrophils causing early inflammatory immune responses. Interestingly, IFN gene signatures were present in blood modules similarly to the lung, but IL-17 and allergy responses were not. Immune response genes associated with type I IFN, which are induced during viral infections, were highly active in both the lungs and blood of mice infected with the T. gondii parasite and the B. pseudomallei bacterium; this confirms previous studies suggesting that the type I IFN response is not restricted to viral infections.
We found that mice without the receptor for type I and/or type II IFN, and therefore without IFN signalling pathways, were less able to fight off T. gondii infection. This was observed for both type I and type II IFN, which have a complex relationship with each other. Our unbiased transcriptomic analyses revealed that, although genes known to be inducible by type I IFN were decreased in the absence of type I IFN receptor signalling as expected, they were completely abrogated in the absence of IFN-g signalling, revealing an advanced layer of regulation in an environment dominated by type II IFN resulting from T. gondii infection.
Use the app to find your favourite gene!
We found that both type I and type II IFN play a key role in protection against the parasite in part by controlling granulocyte responses which, when uncontrolled, can cause damage to the host. The data, published in Nature Communications last year1 and available through the online app, show the activity of more than 45,000 genes. By using the new app anyone can search individual genes and examine detailed changes in the lung and blood of mouse models associated with various infections and HDM allergen exposure.
Anyone can search individual genes and examine detailed changes in the lung and blood of mouse models associated with various infections.
The app is subdivided into five distinct pages that can be accessed through the tabs displayed on the top of the page, with a customised sidebar for user input on each page, and is very easy to use. The user can input any gene of interest to visualise its expression across the different mouse models of infection and inflammation in the ‘Gene expression’ tab or to find out which lung and blood module it belongs to in the ‘Gene lookup in modules’ tab.
As shown in the representative data from the app in Figure 2, Ifng (IFN-g) was most highly expressed in lungs of mice infected with T. gondii, as expected for this intracellular parasite; whereas the type I IFN signature gene Mx1 was highest during respiratory viral infection, with more modest induction in some of the other disease models. Expression of the IL-17 family cytokine gene Il17c was highest in lungs during acute B. pseudomallei infection, which drives a highly neutrophilic immune response, and to a lesser extent during C. albicans infection; whereas expression of the Th2 cytokine gene Il4 was restricted to the HDM allergic airway disease model.
In the ‘Lung modules’ and ‘Blood modules’ the user can visualise the expression of each lung or blood module across lung or blood samples (respectively) obtained from the different mouse models of infection and inflammation. The list of genes within each of the lung and blood modules, and the biological annotation of these modules, can be downloaded from the ‘Download data’ tab. The app is very interactive and detailed information for each sample point can be visualised by hovering over the data points. Additionally, all plots in the app can be adjusted by the user and downloaded as png files.
The app is very interactive and detailed information for each sample point can be visualised by hovering over the data points.
The app is already being used to guide ongoing research. Researchers in the lab of Professor Clare Lloyd at Imperial College London, who collaborated on the study, have used the app to check the expression of chemokine genes of interest identified in allergy in models of lung viral infection, generating ideas for future experiments in the Lloyd lab.
We hope that this tool will enable researchers everywhere to quickly test and generate hypotheses in this way, thus avoiding unnecessary additional mouse experiments.
Singhania et al. 2019 Nature Communications doi:10.1038/s41467-019-10601-6