Modules, networks and systems medicine for understanding disease and aiding diagnosis

Abstract

Many common diseases, such as asthma, diabetes or obesity, involve altered interactions between thousands of genes. High-throughput techniques (omics) allow identification of such genes and their products, but functional understanding is a formidable challenge. Network-based analyses of omics data have identified modules of disease-associated genes that have been used to obtain both a systems level and a molecular understanding of disease mechanisms. For example, in allergy a module was used to find a novel candidate gene that was validated by functional and clinical studies. Such analyses play important roles in systems medicine. This is an emerging discipline that aims to gain a translational understanding of the complex mechanisms underlying common diseases. In this review, we will explain and provide examples of how network-based analyses of omics data, in combination with functional and clinical studies, are aiding our understanding of disease, as well as helping to prioritize diagnostic markers or therapeutic candidate genes. Such analyses involve significant problems and limitations, which will be discussed. We also highlight the steps needed for clinical implementation.

Figure 1

A single disease phenotype can be caused by multiple mechanisms. As an example, asthma can be triggered by allergens, microbes and other environmental factors, each of which may activate different disease mechanisms, which are depicted as shared (black) and specific (red) networks.

Figure 2

A disease module. (a) Conceptual model of how disease-associated genes (blue nodes), identified by high-throughput analysis, tend to co-localize in the human protein-protein interaction network (white nodes), forming a module (blue oval). The genes in the module are assumed to be more important for the disease than extramodular genes. (b) An actual disease module from allergic patients, showing extracellular proteins that were putatively co-regulated with IL13. Blue nodes are associated with cytokine activity, purple nodes are associated with hormone activity, and orange nodes are associated with growth factor activity according to Gene Ontology Molecular Function. The diagram in (b) is reproduced, with permission, from Bruhn et al. Science Translational Medicine 2014 [33].

Figure 3

A module-based approach to identify disease-relevant diagnostic and therapeutic candidate genes in allergy. (a) Twenty-five putative IL13-regulating transcription factors (TFs) were identified by combining data from mRNA microarrays, sequence-based predictions and the literature. (b) IL13-regulating TFs were validated by siRNA-mediated knockdown of the 25 TFs in human total CD4⁺ T cells polarized toward T_H2 using IL13 as a read-out. The target genes of the TFs were identified by combined siRNA knockdown of the positively screened TFs/known IL13-regulating TFs from literature and microarray analyses. This resulted in a module of genes that was co-regulated with IL13 in T_H2-polarized cells and significantly overlapped with differentially expressed genes from allergen-challenged T cells from allergic patients. For further validation experiments, the study focused on module genes that encoded secreted proteins and had not been previously associated with allergy. (c) Functional, diagnostic and therapeutic studies involving one of the module genes, S100A4, were performed in patients with seasonal allergic rhinitis, allergic dermatitis and a mouse model of allergy. (d) Model of S100A4-induced disease mechanisms. Allergic inflammation requires the sensitization of the immune system by allergens, resulting in the production of antigen-specific T cells. The interaction of dendritic cells (DC) in the draining lymph node with T cells is a critical step that is dependent on S100A4. B-cell maturation as a result of T cell-B cell crosstalk (for example, the release of T_H2 cytokines by T cells) leads to the production of IgE and IgG1 by plasma cells. Cytokines and chemokines released by T cells stimulate the migration of circulating granulocytes (for example, neutrophils and eosinophils) to the inflammatory site (skin). Differentiation of naïve T cells into CD8⁺ cytotoxic T cells will exacerbate the skin damage. Blue arrows indicate the flow of the allergic responses. Green arrows indicate the promotion of these processes by S100A4. GEM, gene expression microarray.

Figure 4

An idealized systems medical approach to personalized treatment. (a) All factors that influence a disease can potentially be described by networks. For example, symptoms and signs that tend to co-occur can be linked and form a module that corresponds to a disease (pink oval). That module may be linked to underlying modular protein changes (blue oval). Similarly, the disease module may be linked to co-occurring environmental factors (green oval). (b) Each of the modules in (a) can be further divided to represent different sublayers, from which (c) predictive markers from the different sublayers can be identified, and used for (d) personalized treatment. MLDM, multilayer disease module; nc-RNA, noncoding RNA; PPI, protein-protein interaction; SNPs, single-nucleotide polymorphisms.

Figure 5

Relationship between different disease modules on the protein-protein interaction network. (a) A hypothetical model of three different diseases mapped on the human protein-protein interaction network. The modules are dispersed in the network. (b) Instead, meta-analysis of mRNA microarray and genome-wide association study data show that disease modules partially overlap and form a shared module (grey) [39]. The shared module has important pathogenic, diagnostic and therapeutic implications.