Lower serum annexin A1 and neurogranin levels in chronic schizophrenia: a case-control study

Synaptic dysfunction and immune mechanisms are both implicated in schizophrenia. Annexin A1 (ANXA1) is a glucocorticoid-induced pro-resolving protein and neurogranin (Ng) is a post-synaptic, calmodulin-binding protein studied as a peripheral correlate of synaptic biology. Circulating ANXA1 has not previously been measured in schizophrenia, and serum Ng is sparsely characterised, so no directional hypothesis was pre-specified.

In a single-centre case-control study, 35 patients meeting DSM-5 criteria for schizophrenia (diagnosed with the SCID-5) and 35 sex-comparable healthy controls were enrolled. Fasting serum ANXA1 and Ng were measured in duplicate by sandwich enzyme-linked immunosorbent assay. Symptom severity was rated with the PANSS and CGI-S, and antipsychotic treatment was documented including chlorpromazine equivalents. The pre-specified primary analysis was the Mann-Whitney U test for each biomarker, with Bonferroni and false-discovery-rate correction; group differences were then adjusted for age, sex and body-mass index by analysis of covariance. All other analyses were exploratory.

Both biomarkers were lower in patients: ANXA1 median 6.21 (IQR 5.12-8.34) vs. 13.06 (8.82-23.66) ng/mL (U = 309, p = 3.6 × 10⁻⁴; rank-biserial r = - 0.50) and Ng 174.2 (149.1-281.0) vs. 380.0 (237.1-547.1) pg/mL (U = 302, p = 2.7 × 10⁻⁴; r = - 0.51). Both remained significant after correction for multiple testing and in ANCOVA adjusted for age, sex and body-mass index (adjusted difference - 6.71 ng/mL, p = 0.011; -182.3 pg/mL, p = 0.0008). Neither biomarker was associated with chlorpromazine-equivalent dose, clozapine use, antipsychotic polypharmacy, long-acting injectable use or lifetime electroconvulsive therapy. The two proteins correlated strongly in patients (r = 0.900) and controls (r = 0.844), with no significant between-group difference in the correlation (Fisher r-to-z, p = 0.339), and neither tracked PANSS or CGI-S. Sex-stratified differences reached significance in women only, but the group × sex interaction was not significant.

Serum ANXA1 and neurogranin were lower in chronic schizophrenia and showed no statistically significant association with the treatment variables measured, although medication effects cannot be excluded in a uniformly medicated sample. Because no inflammatory, central or longitudinal measures were obtained, these exploratory findings cannot demonstrate impaired resolution of inflammation, altered central synaptic integrity, or trait status; they identify two candidate peripheral markers that require replication.