Abstract
Background/Objectives: The complement system (CS), including classical (CP), alternative (AP) and lectin (LP) pathways, is a key component of innate immunity. CS is putatively activated by SARS-CoV-2 infection, but such causal relation was never demonstrated in general population studies. We tested the causality of SARS-CoV-2 infection and COVID-19 severity on CP, AP, and LP activation, using two-sample Mendelian Randomization (MR) analyses.
Methods: We selected instrumental variables (IV) from the COVID-19 Host Genetics Initiative European ancestry genome-wide association study (GWAS) release 7 (2021), using susceptibility to infection (ncases=122,616; ncontrols=2,475,240) and hospitalization (ncases=16,512; ncontrols=71,321) as exposures. Genetic associations with CP, LP and AP activation were retrieved from a recent GWAS of the Cooperative Health Research in South Tyrol (CHRIS) study (n=4990). We performed inverse-variance weighted fixed effect (IVW-FE) meta-analyses of the Wald estimator statistics, using the IVW-random effect meta-analysis, weighted median, weighted mode, and Egger regression as sensitivity analyses.
Results: We identified a causal effect of susceptibility to infection on LP activation (P=4.5×10-9), with evidence of heterogeneity and pleiotropy driven by a single IV in the ABO gene (I2=87.4%, Pheterogeneity=9.8×10-25). Pleiotropy-robust methods showed direction consistent estimates. Hospitalization was causally associated with lower CP activation (P=0.029).
Conclusion: COVID-19 severity is causally associated with lower CP activation. The evidence of pleiotropy prevents claims of causality of SARS-CoV-2 infection on LP activation, recently demonstrated by in vitro studies. This raises methodological questions on MR assumptions, which may prevent true causal effect identification.
Grants: Autonomous Province of Bolzano LG14 grant no. D52F20000770003 (PACE).