Neurologic complications in herpes simplex encephalitis: clinical, immunological and genetic studies.

Patients with herpes simplex virus (HSV) encephalitis (HSE) often develop neuronal autoantibody-associated encephalitis (AE) post-infection. Risk factors of AE are unknown. We tested the hypotheses that predisposition for AE post-HSE may be involved, including genetic variants at specific loci, human leucocyte (HLA) haplotypes, or the blood innate immune response against HSV, including type I interferon (IFN) immunity. Patients of all ages with HSE diagnosed between 1 January 2014 and 31 December 2021 were included in one of two cohorts depending on whether the recruitment was at HSE onset (Spanish Cohort A) or by the time of new neurological manifestations (international Cohort B). Patients were assessed for the type of neurological syndromes; HLA haplotypes; blood type I-IFN signature [RNA quantification of 6 or 28 IFN-response genes (IRG)] and toll-like receptor (TLR3)-type I IFN-related gene mutations. Overall, 190 patients (52% male) were recruited, 93 in Cohort A and 97 in Cohort B. Thirty-nine (42%) patients from Cohort A developed neuronal autoantibodies, and 21 (54%) of them developed AE. Three syndromes (choreoathetosis, anti-NMDAR-like encephalitis and behavioural-psychiatric) showed a high (≥95% cases) association with neuronal autoantibodies. Patients who developed AE post-HSE were less likely to carry the allele HLA-A*02 (4/21, 19%) than those who did not develop AE (42/65, 65%, P = 0.0003) or the Spanish general population (2005/4335, 46%, P = 0.0145). Blood IFN signatures using 6 or 28 IRG were positive in 19/21 (91%) and 18/21 (86%) patients at HSE onset, and rapidly decreased during follow-up. At Day 21 after HSE onset, patients who later developed AE had higher median IFN signature compared with those who did not develop AE [median Zs-6-IRG 1.4 (0.6; 2.0) versus 0.2 (-0.4; 0.8), P = 0.03]. However, a very high median Zs-6-IRG (>4) or persistently increased IFN signature associated with uncontrolled viral infection. Whole exome sequencing showed that the percentage of TLR3-IFN-related mutations in patients who developed AE was not different from those who did not develop AE [3/37 (8%) versus 2/57 (4%), P = 0.379]. Multivariate logistic regression showed that a moderate increase of the blood IFN signature at Day 21 (median Zs-6-IRG >1.5 but <4) was the most important predictor of AE post-HSE [odds ratio 34.8, interquartile ratio (1.7-691.9)]. Altogether, these findings show that most AE post-HSE manifest with three distinct syndromes, and HLA-A*02, but not TLR3-IFN-related mutations, confer protection from developing AE. In addition to neuronal autoantibodies, the blood IFN signature in the context of HSE may be potentially useful for the diagnosis and monitoring of HSE complications.