So far, there is no clear pathogenesis and no cure for systemic lupus erythematosus (SLE). The therapeutic benefits of existing drug therapies are far from ideal. The proteome is a major source of therapeutic targets. Therefore, new drug targets for SLE need to be discovered. Based on the STROBE-Mendelian randomization (MR) checklist, we performed MR to explore potential drug targets for SLE, using genome-wide association study summary statistics of plasma and cerebrospinal fluid (CSF) and further replicated in the external validation. Bidirectional MR, reverse causality testing by Steiger filtering, Bayesian co-localization were used. In addition, protein-protein interaction networks (PPI) were performed to reveal potential associations between proteins and current SLE drugs. At false discovery rate (FDR) significance (PFDR <
.05), MR analysis revealed 8 proteins. Five proteins decreased the SLE risks, whereas the other 3 proteins increased the SLE risks. None of the 8 proteins had reverse causality except sICAM-1. Bayesian co-localization suggested that 5 proteins shared the same variant with SLE. PPI network suggested that intercellular adhesion molecular 1 (ICAM-1), Fc-gamma-RIIb (FCG2B) and N-terminal pro-B-type natriuretic peptide (N-terminal pro-BNP) interacted with targets of current SLE medications. Our integrative analysis revealed that SLE risk is causally associated with ICAM-1, FCG2B, and N-terminal pro-BNP. These 3 proteins have the potential to become drug targets of SLE, especially for ICAM-1 and FCG2B. More further studies are also warranted to support this finding.