Viral perturbations of host networks reflect disease etiology

Abstract

Many human diseases, arising from mutations of disease susceptibility genes (genetic diseases), are also associated with viral infections (virally implicated diseases), either in a directly causal manner or by indirect associations. Here we examine whether viral perturbations of host interactome may underlie such virally implicated disease relationships. Using as models two different human viruses, Epstein-Barr virus (EBV) and human papillomavirus (HPV), we find that host targets of viral proteins reside in network proximity to products of disease susceptibility genes. Expression changes in virally implicated disease tissues and comorbidity patterns cluster significantly in the network vicinity of viral targets. The topological proximity found between cellular targets of viral proteins and disease genes was exploited to uncover a novel pathway linking HPV to Fanconi anemia.

Figure 1. Linking a viral proteome to virally implicated diseases through the host interactome.

A, Viral proteins (virome) interact with host proteins (viral targets) in the host interactome, which in turn are linked to various human diseases (phenome) through mutations in particular disease susceptibility genes (variome). B, Determining topological proximity between viral targets and genes associated with virally implicated diseases by measuring the shortest path lengths between them. For each disease, the minimum number of hops of interactions needed to connect any of its associated genes to any viral targets is designated as the shortest path. C,D, The average shortest path for either EBV (C) or HPV16 (D) was significantly shorter than random expectation.

Figure 2. Virally implicated diseases associated with genes in the neighborhoods of viral targets.

A,B, The number of virally implicated diseases in the neighborhoods was higher than randomly expected for EBV (A) and HPV16 (B). C,D, The number of differentially expressed genes in the neighborhood of viral targets of either EBV (C) or HPV16 (D) was significantly higher compared to that in the neighborhood of randomly sampled host genes. The total number of genes regulated by EBV and HPV targets is 109 and 122, respectively. Expression level was measured in tissues of two virally implicated diseases respectively, Burkitt's lymphoma (EBV) and cervical cancer (HPV), and compared to normal tissues. E,F, Known virally implicated diseases in the vicinity of viral targets for EBV (E) and HPV16 (F). Examples of paths that are known to correspond to disease mechanism are highlighted in grey and listed individually underneath.

Figure 3. Viral disease networks.

A,C, The neighborhoods of viral targets in the host interactome, along with their disease associations, represent “viral disease networks”. Diseases associated with genes in the neighborhood of EBV (A) or HPV16 (C) targets that are not yet characterized with viral implications are shown as grey nodes. Node size is proportional to the degree of a node (number of neighbors it has) in the viral disease network. B,D, Diseases associated with genes in the neighborhoods of randomly generated viral targets of EBV (B) or HPV16 (D) are significantly sparser than the neighborhoods of actual viral targets. E, Benchmarking the prioritization using relative risk with virally infected patients showed that the higher-ranked diseases in the prioritization are more often associated with viral infection. F, Differentially expressed genes in E6 or E7 induced IMR90 and HFK cell populations with their associated diseases. If a gene is regulated by a specific viral protein target, it is also almost always differentially expressed in the cell population where that specific viral protein is induced. For example, EDN1 is regulated by FOS, an E7 target, and EDN1 is differentially expressed in E7 induced cell populations. Large grey nodes: diseases with high relative risk among HPV patients.