Transcriptional Vulnerability of Immune Genes Underlying Cortical Abnormalities across Psychiatric Disorders

Technical Summary: Transcriptional Vulnerability of Immune Genes Underlying Cortical Abnormalities across Psychiatric Disorders

Problem Statement
Psychiatric disorders are characterized by complex symptom overlaps and comorbidities that challenge the current symptom-based diagnostic paradigm (DSM-5). While previous research has established that cortical thickness abnormalities are linked to genetic risk and regional vulnerability across disorders, the specific immunological mechanisms driving these regional vulnerabilities remain unclear. Specifically, there is a lack of systematic characterization regarding how immune-related gene transcriptomes contribute to the commonalities and specificities of cortical thickness alterations across different psychiatric disorders. This study addresses the gap in understanding the neuroimmune mechanisms underlying the regional vulnerability of the cortex in psychiatric conditions.

Methodology
The study employed a multi-modal integrative approach combining neuroimaging data with transcriptomic profiles:

1. Data Sources:

   • Cortical Abnormality Maps: Data were sourced from the ENIGMA consortium, encompassing 15,000 patients and over 22,000 controls across six disorders: Attention-Deficit/Hyperactivity Disorder (ADHD), Autism Spectrum Disorder (ASD), Bipolar Disorder (BD), Major Depressive Disorder (MDD), Obsessive-Compulsive Disorder (OCD), and Schizophrenia (SCZ). Cortical thickness was measured at 34 regions of interest (ROIs) using the Desikan-Killiany parcellation.
   • Gene Expression Data: Immune-related gene expression data were derived from the Allen Human Brain Atlas (AHBA) and curated immune databases (GeneOntology, KEGG, ImmPort, ImmuneSigDB). The analysis focused on the left hemisphere (due to data availability in AHBA) and utilized a matrix of 976 immune-related genes across 34 brain regions.

2. Statistical Analysis:

   • Partial Least Squares (PLS) Regression: Used to identify latent variables representing the maximum covariance between the immune gene expression matrix and the cortical abnormality matrix. This allowed for the extraction of orthogonal factors representing the best covarying patterns.
   • Bootstrap Resampling: Performed 10,000 times to estimate the reliability of gene weights and determine statistical significance (confidence intervals excluding zero).
   • Spatial Null Models: "Spin tests" were utilized to preserve spatial autocorrelation, ensuring that observed associations were not merely artifacts of spatial proximity.
   • Gene Set Enrichment Analysis (GSEA): Conducted using GSEA-P software to identify biological pathways (Gene Ontology) significantly enriched in genes with high positive or negative loadings.

3. Grouping Strategy: Disorders were clustered based on cortical abnormality patterns into three groups: ADHD, ASD, and a "non-developmental" group (BD, MDD, OCD, SCZ).

Key Results

• Cross-Disorder Covariation:

  • PLS analysis revealed a significant latent variable linking immune gene expression to cortical abnormalities across all six disorders ($P_{spin} = 0.0190$).
  • A spatial gradient was observed: the prefrontal lobe showed negative scores (associated with cortical thinning and specific immune pathways), while the occipital lobe showed positive scores (associated with cortical thickening).
  • Two Immune-Related Vulnerability Clusters:
    • Negative Vulnerability Cluster: Enriched in innate immunity pathways, specifically microglia activation and cytokine secretion. These were strongly associated with cortical thinning in frontal and temporal regions.
    • Positive Vulnerability Cluster: Enriched in adaptive immunity pathways, including lymph node development and T cell differentiation regulation. These were associated with the occipital regions.

• Disorder-Specific Patterns:

  • ADHD: Showed a significant covarying pattern similar to the common model. Enrichment analysis highlighted innate immunity pathways, including microglial cell activation, neutrophil activation, and degranulation.
  • ASD: Showed a non-significant covarying pattern overall, but when analyzing the whole-brain transcriptome, innate immunity pathways (specifically positive regulation of macrophage activation) remained a primary negative pathway.
  • Non-Developmental Disorders (BD, MDD, OCD, SCZ): This group exhibited a distinct pattern where the positive pathways included cytokine secretion and microglial cell activation. Notably, when validating against the whole-brain transcriptome, this group did not show a major involvement of immune gene pathways, suggesting a potential shift toward synaptic mechanisms.

• Validation:

  • Validation against the whole-brain gene transcriptome confirmed that for ADHD and ASD, immune pathways remained major negative contributors (accounting for ~22-24% of pathways).
  • Neither innate nor adaptive immunity showed a clear bias in the overall immune pathways for ADHD and ASD, indicating a complex interplay.

Significance and Claims
The authors claim that these findings offer new insights into the mechanisms of vulnerability to psychiatric disorders from a neuroimmune perspective. The study suggests that:

1. Shared Mechanisms: There is a shared vulnerability across psychiatric disorders linked to microglia activation and cytokine secretion, particularly in the context of the "second hit" model where environmental stressors trigger abnormal immune responses.
2. Distinct Mechanisms: There are distinct immune-related pathways for neurodevelopmental disorders (ADHD, ASD) versus non-developmental disorders. ADHD and ASD show strong associations with innate immunity and cortical thickness abnormalities, whereas the non-developmental group's vulnerability may be more closely tied to synaptic pruning and network wiring rather than direct immune gene expression in the cortical thickness model.
3. Clinical Implications: The results provide a basis for specific preventive approaches tailored to different categories of psychiatric disorders, moving beyond a one-size-fits-all diagnostic model.

Limitations Acknowledged by the Authors
The paper explicitly notes several limitations:

• Data Scope: The analysis relied on post-mortem data from the left hemisphere only, with a small sample size for gene expression (six donors), limiting generalizability.
• Individual Specificity: The study used group-level data and did not account for individual patient specificity.
• Multifactorial Nature: The authors emphasize that immune vulnerability is the result of a complex combination of genetic, immune, environmental, and developmental factors, and cortical thickness changes dynamically over a lifespan.

In conclusion, the study provides a neuroimmune perspective on the commonalities and specificities of cortical abnormalities, highlighting the role of microglia and cytokine secretion in shared vulnerability while distinguishing the immune profiles of neurodevelopmental disorders from other psychiatric conditions.