1181: Clinical Relevance of Tumor Necrosis Factor Superfamily Cytokines

Background/Purpose: The tumor necrosis factor superfamily (TNFSF) includes 48 proteins involved in immune activation, inflammation, and cell death and represent an area of intense focus in current drug development efforts. However, the proteins are difficult to measure in large clinical populations, thus the clinical relevance of many of these proteins in human disease is not well characterized and the potential effects of drugs targeting them is largely unknown. The ability to associate genetically driven variability in protein levels with phenotype risk provides an opportunity to determine the potential relevance of TNFSF proteins using large populations with both genotype and clinical information available.

Methods: To define the clinical manifestations associated with TNFSF members, we examined genetic models for 38 TNFSF proteins available in the UK Biobank pQTL web portal (http://uk-ppp.gwas.eu). We estimated genetically determined levels of TNFSF proteins in participants of European ancestry in the Vanderbilt University Medical Center biobank (BioVU) and in the All of Us Research Program (AoU). We performed a phenome-wide association study (PheWAS) in each biobank to define clinical associations for TNFSF levels adjusted for sex, age, and 5 principal components; and significant associations (P-value≤5x10^-5) in BioVU and AoU were combined in a meta-analysis. We examined proteins with significant associations in the meta-analysis (P-value ≤5x10^-5) using a Mendelian randomization (MR) approach. We used genetic predictors of TNFSF proteins as instrumental variables and summary statistics for clinical diagnoses as outcome in an inverse-variance weighted regression (IVWR). We tested horizontal pleiotropy using Egger regression. A P-value< 0.05 was considered significant in the IVWR and an intercept P-value ≤0.05 suggested pleiotropy.

Results: Genetically predicted levels of twelve proteins were significantly associated with at least one clinical diagnosis in either biobank. In the meta-analysis, higher levels for TNFRSF4, TNFRS9, TNFRSF8, and TNFRSF1B increased the risk of several thyroid and autoimmune diseases including lupus; higher levels of TNFSF3 were associated with a reduced risk of both ankylosing spondylitis and Reiter’s disease (Fig. 1A). TNFSF13 increased the risk for dyslipidemias, TNFRSF6B for diverticulosis and TNFRSF6 for non-alcoholic liver disease (Fig. 1B). In the IVWR analyses, TNFRSF4 (P-value =1.5x10^-5) and TNFRSF8 (P-value=0.02) were associated with autoimmune thyroiditis without evidence of horizontal pleiotropy (intercept P-value >0.05 for both proteins).

Conclusion: Genetic predisposition to higher levels of several TNFSF proteins is associated with increased risk of autoimmune, metabolic, hepatic, and gastrointestinal disorders and may identify potential therapeutic targets.