2026 ECNP Congress

PS04-4264

Distinct epitope-derived humoral immunotypes define the immune landscape of treatment-resistant schizophrenia

Vladislav Yakimov — A. Rähni, E. Boudriot, J. Moussiopoulou, P. Falkai, A. Schmitt, K. Palm, F. Raabe · University Hospital LMU Munich, Department of Psychiatry and Psychotherapy, Munich, Germany

2026 ECNP Congress

Background and aim of the study: Immune dysregulation contributes to the pathophysiology in a substantial subset of individuals with schizophrenia spectrum disorders (SSDs) [1], including treatment-resistant schizophrenia (TRS) [2], yet serological markers that enable clinically useful stratification are lacking. Epidemiological and serological data link infection-related humoral immunity to increased schizophrenia risk [3], suggesting that stable epitope-level antibody imprints may capture etiologically relevant immune pathways amenable to stratification. We aimed to derive data-driven immune modules and immunotypes from high-dimensional epitope features and to test whether TRS is concentrated in specific humoral immune configurations. We hypothesized that serological epitope imprints are only sparsely shared between SSDs and TRS.

Methods

In our study, we profiled the humoral immune response using Mimotope Variation Analysis (MVA), a next-generation phage display technology [4], in sera from 462 participants with SSDs and healthy controls (HCs). The full epitope repertoire (113,400 features) was screened in two univariate discovery–validation analyses contrasting (i) SSDs vs HCs and (ii) TRS vs non-TRS within SSDs. The cohort was randomly split into discovery and validation sets (80/20): epitopes were screened in the discovery set and only candidate epitopes were re-tested in the held-out validation set. Associated epitopes underwent principal component analysis (PCA) followed by density-based clustering (DBSCAN) in the first 30 principal components to derive immune “modules”. Subject-level module scores (mean z-scored epitope signal per module) entered Gaussian mixture models (GMMs) to define data-driven immunotypes. Associations with SSD diagnosis and TRS status were tested using logistic regression, controlling for age and sex.

Results

In the SSD vs HC contrast, 13,266 epitopes were associated with SSDs, organizing into nine immune modules (5,425 epitopes) and nine immunotypes. These immunotypes did not differ in SSD probability, suggesting largely shared immune architectures between SSDs and HCs. In the TRS vs non-TRS contrast, 1,941 epitopes were associated with TRS, forming five modules (925 epitopes) and five TRS-related immunotypes. Modules were well-separated in t-distributed Stochastic Neighbor Embedding (t-SNE) space and dominated by epitopes depleted in TRS, indicating a relative lack of specific humoral responses rather than generalized activation. TRS probability differed significantly across immunotypes, ranging from ~0.12 in the most non-TRS–like to ~0.65 in the most TRS-enriched immunotype, corresponding to an ~13.6-fold difference in odds. Only 291 epitopes overlapped between SSD- and TRS-associated sets (2.2% of SSD-associated and 15.0% of TRS-associated features), indicating that TRS-related immune signatures only partially recapitulate the broader SSD signal.

Conclusions

Cross-sectional epitope-based immunotyping suggests that treatment-resistant schizophrenia is concentrated in specific humoral immune configurations rather than being randomly distributed across the SSD immune landscape. These TRS-enriched immunotypes represent promising candidates for prognostic stratification, but require biological annotation (putative pathogen-derived and human protein targets) and prospective validation in longitudinal cohorts [5], including evaluation of confounding by comorbidity.

References

    • Khandaker GM, Cousins L, Deakin J, Lennox BR, Yolken R, Jones PB. Inflammation and immunity in schizophrenia: implications for pathophysiology and treatment. Lancet Psychiatry. 2015 Mar;2(3):258-270. doi: 10.1016/S2215-0366(14)00122-9. Epub 2015 Feb 25. PMID: 26359903; PMCID: PMC4595998.
    • Chen W, Tian Y, Gou M, Wang L, Tong J, Zhou Y, Feng W, Li Y, Chen S, Liu Y, Wang Z, Pan S, Zhang P, Huang J, Yang X, Li CR, Tian L, Hong LE, Tan Y. Role of the immune-kynurenine pathway in treatment-resistant schizophrenia. Prog Neuropsychopharmacol Biol Psychiatry. 2024 Mar 2;130:110926. doi: 10.1016/j.pnpbp.2023.110926. Epub 2023 Dec 24. PMID: 38147973.
    • Leucht S, Siafis S, McGrath JJ, McGorry P, Howes OD, Tamminga C, Carr R, Bighelli I, Schneider-Thoma J, Priller J, Davis JM. Schizophrenia. Nat Rev Dis Primers. 2025 Nov 27;11(1):83. doi: 10.1038/s41572-025-00667-6. PMID: 41309720.
    • Rähni A, Jaago M, Sadam H, Pupina N, Pihlak A, Tuvikene J, Annuk M, Mägi A, Timmusk T, Ghaemmaghami AM, Palm K. Melanoma-specific antigen-associated antitumor antibody reactivity as an immune-related biomarker for targeted immunotherapies. Commun Med (Lond). 2022 May 11;2:48. doi: 10.1038/s43856-022-00114-7. PMID: 35603273; PMCID: PMC9095616.
    • Yakimov V, Neuwinger L, Weber MM, Brantl M, Maurus I, Sautner J, John M, Karslı B, Hasanaj G, Bungard A, Hasan A, Wagner E, Fischer L, Steiner P, Schworm B, Priglinger S, Papiol S, Falkai P, Schmitt A, Raabe FJ, Keeser D, Roell L, Moussiopoulou J, Boudriot E; CDP-Working Group. Clinical deep phenotyping of treatment response in schizophrenia (CDP-STAR): design and methodology of a prospective multimodal observational study. Eur Arch Psychiatry Clin Neurosci. 2025 Sep 8. doi: 10.1007/s00406-025-02100-1. Epub ahead of print. PMID: 40924103.