Decoding NMDAR encephalitis: proteomic markers and computational identification of potential therapeutic pathways

Abstract

Background

The proteome is a valuable resource for pinpointing therapeutic targets. Therefore, we conducted a proteome-wide Mendelian randomization (MR) study aimed at identifying potential protein markers and therapeutic targets for Anti-N-Methyl-D-Aspartate Receptor Encephalitis (NMDAR-E).

Methods

Protein quantitative trait loci (pQTLs) were obtained from seven published genome-wide association studies (GWASs) focusing on the plasma proteome, resulting in summary-level data for 734 circulating protein markers. Genetic associations with NMDAR-E were determined via a large meta-analysis encompassing 323 cases and 1,519 controls. To confirm the causal roles of candidate proteins, we performed sequential colocalization. Subsequently, we experimentally validated prioritized proteins through two complementary approaches: (1) stimulation of HMC3 microglial cells with patient-derived anti-NMDAR-IgG versus control human IgG to assess antibody-induced protein expression dynamics; (2) anti-NMDAR-IgG was stereotactically injected into the lateral ventricle of mice to establish the passive immunization NMDAR-E model, with comparative analysis of lateralized protein expression 7 days post-injection. Additionally, single cell-type expression analysis, protein–protein interaction (PPI) assessments, and evaluations of druggability were conducted to pinpoint enriched cell types and possible therapeutic targets.

Results

In total, genetically predicted levels of 37 proteins showed associations with NMDAR-E risk. Elevated levels of three proteins (SIRPA, LGALS3, CASP3) and decreased levels of two proteins (TREM2, IL1RN) were correlated with an increased risk of NMDAR-E. The identified protein-coding genes were predominantly expressed in CD20+ B cells, with comparably elevated expression also observed in mast cells and CD16+ monocytes within the peripheral blood mononuclear cells (PBMCs) of NMDAR-E patients. Anti-NMDAR-IgG induced upregulation of four proteins in human microglial cells and four laterally upregulated proteins in murine brains, with three overlapping responders. Furthermore, proteins TREM2, LGALS3, SIRPA, IL1RN and CASP3, which were initially targeted for drug development in cancer and autoimmune conditions, may also represent therapeutic options for NMDAR-E.

Conclusions

This study integrates genetic epidemiology with functional validation, successfully identifying protein biomarkers associated with NMDAR-E risk. The convergence of MR-predicted causal proteins with antibody-driven expression changes in human microglia and mouse models underscores their pathological relevance, providing actionable insights for biomarker screening and therapeutic development.