2026 ECNP Congress

PS01-1047

Multimodal CSF immune profiling in first-episode psychosis reveals blood–CSF barrier dysfunction and synaptic autoantibody enrichment

Vasishta Polisetty — Y. Mizuno, O. Kriel, L. He, T. Pollak, O. Howes, K. Schmack · The Francis Crick Institute, Neural Circuits and Immunity in Psychosis Lab, London, United Kingdom

2026 ECNP Congress

Background

Accumulating evidence implicates autoimmune mechanisms in the pathogenesis of primary psychotic disorders. Genetic studies implicate immune-related loci, biomarker studies indicate elevated inflammation, epidemiological studies show associations with prior infections and autoimmune conditions, and clinically, psychosis is a well-recognised manifestation of autoimmune encephalitides. [1,2] However, the pathological significance of autoimmunity in first-episode psychosis (FEP) remains inconclusive, owing to methodological variability and bias in assay techniques, reliance on serum over cerebrospinal fluid (CSF) and a lack of robust causal data. [3]

Aims

This study aims to systematically investigate autoimmune processes in FEP through multimodal investigations of CSF and serum, combining classical biochemical indices of barrier function, assays for brain-reactive antibodies with unbiased antigen discovery.

Methods

Paired CSF and serum samples from 41 FEP patients and 23 controls were analysed for total protein, albumin, and IgG, with blood–CSF barrier integrity assessed via Reiber diagram using age-adjusted albumin quotient cut-offs. Brain-reactive antibodies were evaluated using an unbiased rodent tissue-based assay, with fluorescence intensity extracted from 13 brain regions and PCA applied to quantify staining patterns. Antigen discovery was performed using a HuProt microarray (~23,000 full-length human proteins). Following within-sample z-score normalisation, case-enriched hits were identified through three distinct, complementary hit-calling methods. Consensus proteins were assessed for synaptic enrichment using SynGO over- representation analysis.

Results

Reiber analysis revealed blood–CSF barrier dysfunction in 9.7% of cases. Group-level comparisons of albumin and IgG quotients did not reach significance; however, absolute CSF IgG concentrations were significantly elevated in cases relative to controls (p < 0.05, Mann–Whitney). PCA of TBA staining patterns identified five outliers that all happened to be cases. Three of these outliers also exhibited barrier dysfunction on Reiber analysis. CSF IgG correlated with PC1 scores (r = 0.57, p < 0.001, Pearson correlation), alongside total protein and albumin, suggesting staining patterns reflect increased CNS antibody exposure secondary to barrier permeability.

HuProt screening revealed a significant global elevation in mean log-transformed fluorescence intensity in cases relative to controls (t = 6.05, p < 0.001, Cohen’s d = 1.58). Notably, buffer spot fluorescence correlated strongly with HuProt PC1 (r = −0.83, p < 0.001), suggesting a global increase in non-specific binding.

For autoantigen discovery, cross-method hit-calling comparison identified 30 case-enriched proteins across at least two strategies and an eight-protein high-confidence consensus across all three. SynGO analysis revealed significant synaptic enrichment in the 30-protein set (7/30 mapped, 23.3%, synapse CC: adj. p = 0.036), which strengthened markedly in the eight-protein consensus (3/8 mapped, 37.5%, synapse CC: adj. p = 0.024). This progressive increase suggests the high-confidence consensus is preferentially enriched for synaptic autoantigens.

Conclusions

Our findings provide novel evidence for autoimmune involvement in FEP patients. The identified blood–CSF barrier dysfunction in a subset of cases suggests a route for CNS autoantibody exposure, pointing to a potential biomarker for stratification for future immunotherapy trials. Our unbiased autoantibody screening identifies synaptic proteins as candidate antigenic targets, providing a foundation for future causal investigations into autoimmunity in psychosis.

References

  1. Pathmanandavel, Karrnan, Jean Starling, Russell C. Dale, and Fabienne Brilot. 2013. ‘Autoantibodies and the Immune Hypothesis in Psychotic Brain Diseases: Challenges and Perspectives’. Clinical and Developmental Immunology 2013: 257184. https://doi.org/10.1155/2013/257184.
  2. Al-Diwani, Adam A. J., Thomas A. Pollak, Sarosh R. Irani, and Belinda R. Lennox. 2017. ‘Psychosis: An Autoimmune Disease?’ Immunology 152 (3): 388–401. https://doi.org/10.1111/imm.12795.
  3. Guasp, Mar, and Josep Dalmau. 2023. ‘Searching for Neuronal Antibodies in Psychiatric Diseases: Uncertain Findings and Implications’. Neurology 101 (15): 656–60. https://doi.org/10.1212/WNL.0000000000207486.