Ghiassian, Susan PhD; Withers, Johanna PhD; Akmaev, Viatcheslav PhD
The American Journal of Gastroenterology: October 2021 – Volume 116 – Issue – p S431-S432
Introduction: The map of human disease biology called the human interactome (HI) is a map of experimentally determined protein-protein interactions. Understanding the topology and dynamics of patient molecular data through analyses of the HI can reveal proteins and biological processes underlying disease development and progression. Such studies have enabled discovery and reprioritization of drug targets, insights into drug-drug interactions and identification of biomarkers for patient stratification. Application of this technology to autoimmune diseases (ADs) is particularly useful. An estimated 25-35% of patients with an ADs may develop one or more additional ADs. This points to a molecular background of immune system disfunction that could be targeted therapeutically to ameliorate a range of ADs.
Methods: A network biology platform was used to generate disease modules, which defined the HI network neighborhood and proteins relevant to the underlying biology of ten diseases. Each protein in the HI was then ranked based on its network-based similarity and connectivity to the disease module to generate a list of proteins with high relevance to disease and were considered candidate protein targets for therapeutic development. Top-ranked proteins in each disease were enriched for protein targets of approved therapies.
Results: The pairwise separation of 10 disease modules on the HI is illustrated in Figure 1. Among the autoimmune diseases studied, ulcerative colitis and Crohn’s disease displayed the lowest network separation. Rheumatoid arthritis was closest to multiple sclerosis and psoriatic arthritis. Disease pairs with small separation on the network are expected to have higher biological similarity in terms of common symptoms, comorbidities, and treatment opportunities. This prediction is supported by our discovery of numerous candidate drug targets that were common across all the autoimmune diseases studied. In contrast, proteins that could be targeted to influence neurological conditions such as Parkinson’s disease, Alzheimer’s disease and amyotrophic lateral sclerosis displayed a minimal network relationship to those of the autoimmune diseases.
Conclusion: The network relationship between disease-related proteins of seven autoimmune diseases suggested the existence of one or more molecular mechanisms underlying human autoimmunity that traverse traditional medical disease diagnoses. Future work may delineate molecular pathways and patterns that manifest in different autoimmune disease phenotypes.