Genetic susceptibility to schizophrenia through neuroinflammatory pathways is associated with retinal thinning: Findings from the UK-Biobank

Abstract

The human retina is part of the central nervous system and can be easily and non-invasively imaged with optical coherence tomography. While imaging the retina may provide insights on central nervous system-related disorders such as schizophrenia, a typical challenge are confounders often present in schizophrenia which may negatively impact retinal health. Here, we therefore aimed to investigate retinal changes in the context of common genetic variations conveying a risk of schizophrenia as measured by polygenic risk scores. We used population data from the UK Biobank, including White British and Irish individuals without diagnosed schizophrenia, and estimated a polygenic risk score for schizophrenia based on the newest genome-wide association study (PGC release 2022). We hypothesized that greater genetic susceptibility to schizophrenia is associated with retinal thinning, especially within the macula. To gain additional mechanistic insights, we conducted pathway-specific polygenic risk score associations analyses, focusing on gene pathways that are related to schizophrenia. Of 65484 individuals recruited, 48208 participants with available matching imaging-genetic data were included in the analysis of whom 22427 (53.48%) were female and 25781 (46.52%) were male. Our robust principal component regression results showed that polygenic risk scores for schizophrenia were associated with retinal thinning while controlling for confounding factors (b = -0.03, p = 0.007, pFWER = 0.01). Similarly, we found that polygenic risk for schizophrenia specific to neuroinflammation gene sets revealed significant associations with retinal thinning (b = -0.03, self-contained p = 0.041 (reflecting the level of association), competitive p = 0.05 (reflecting the level of enrichment)). These results go beyond previous studies suggesting a relationship between manifested schizophrenia and retinal phenotypes. They indicate that the retina is a mirror reflecting the genetic complexities of schizophrenia and that alterations observed in the retina of individuals with schizophrenia may be connected to an inherent genetic predisposition to neurodegenerative aspects of the condition. These associations also suggest the potential involvement of the neuroinflammatory pathway, with indications of genetic overlap with specific retinal phenotypes. The findings further indicate that this gene pathway in individuals with a high polygenic risk for schizophrenia could contribute through acute-phase proteins to structural changes in the retina.

Figure 1.

a. Illustration of the anatomy of the eye and retina, displayed as the cross section of the eye, and macular section of the retina. Optical coherence tomography measurements divide the macula in 9 subfields as indicated by white crosshairs. Macular thickness is measured in micrometers (μm). Structure of different retinal layers is displayed on the right. b. Association between polygenic risk score for schizophrenia and overall macular thickness in each eye respectively. Solid lines represent the regression estimates, while complementary shaded areas correspond to 95 percent confidence intervals. c. Individual macular subfields associated with polygenic risk scores for schizophrenia for both left eye (left) and right eye (right). Color coding corresponds to coefficients (b) from robust linear regression model. Higher b values correspond to greater thinning with increasing polygenic risk scores for schizophrenia. The sizes of subfields are slightly inflated for visualization purposes only. CS: Center subfield, IS: Inner Superior, OS: Outer Superior, IN: Inner Nasal, ON: Outer Nasal, II: Inner Inferior, OI: Outer Inferior, IT: Inner Temporal, OT: Outer Temporal.

Figure 2.

Diagram of inclusion and exclusion of participants in population. QC = Quality control. ICD = International Classification of Diseases. ICD-10, F20–29 categorization includes individuals with diagnosed schizophrenia, schizotypal and delusional disorders.

Figure 3.

Association between polygenic risk for schizophrenia enriched for multiple gene pathways and first retinal principal component at p-value threshold 1. Confidence intervals (95 percent) are visualized by grey error bars. Microvasculature pathways: 1. HP abnormal retinal vascular morphology (M43559) 2. HP premature coronary artery atherosclerosis (M36658). Inflammatory pathways: 1. GOBP acute inflammatory response (M6557) 2. GOBP neuroinflammatory response (M24927) 3. Biocarta TGFB pathway (M18933) 4. GOBP chronic inflammatory response (M15140). Signaling pathways influencing neuronal development: 1. GOBP Wnt signaling pathway involved in midbrain dopaminergic neuron differentiation (M25305) 2. GOBP positive regulation of dopamine receptor signaling pathway (M24111) 3. ST Wnt beta catenin pathway (M17761).

Figure 4.

Mediation analysis: C-reactive protein related to associability acting as a partial mediator in the link between polygenic risk scores specific to neuroinflammation for schizophrenia and macular phenotypes represented by PC1. Coef = beta coefficient, PRSSZ = polygenic risk for schizophrenia, CRP = C-reactive protein. ** = p < 0.01, *** = p < 0.001