The schizophrenia brain is differentiated from the normal brain by subtle changes, with significant overlap in measures between normal and disease states. For the past 25 years, schizophrenia has increasingly been considered a neurodevelopmental disorder. This frame of reference challenges biological researchers to consider how pathological changes identified in adult brain tissue can be accounted for by aberrant developmental processes occurring during fetal, childhood, or adolescent periods. To place schizophrenia neuropathology in a neurodevelopmental context requires solid, scrutinized evidence of changes occurring during normal development of the human brain, particularly in the cortex; however, too often data on normative developmental change are selectively referenced. This paper focuses on the development of the prefrontal cortex and charts major molecular, cellular, and behavioral events on a similar time line. We first consider the time at which human cognitive abilities such as selective attention, working memory, and inhibitory control mature, emphasizing that attainment of full adult potential is a process requiring decades. We review the timing of neurogenesis, neuronal migration, white matter changes (myelination), and synapse development. We consider how molecular changes in neurotransmitter signaling pathways are altered throughout life and how they may be concomitant with cellular and cognitive changes. We end with a consideration of how the response to drugs of abuse changes with age. We conclude that the concepts around the timing of cortical neuronal migration, interneuron maturation, and synaptic regression in humans may need revision and include greater emphasis on the protracted and dynamic changes occurring in adolescence. Updating our current understanding of post-natal neurodevelopment should aid researchers in interpreting gray matter changes and derailed neurodevelopmental processes that could underlie emergence of psychosis.

Excitement and controversy have followed neuregulin (NRG1) since its discovery as a putative schizophrenia susceptibility gene; however, the mechanism of action of the associated risk haplotype (HapICE) has not been identified, and specific genetic variations, which may increase risk to schizophrenia have remained elusive. Using a postmortem brain cohort from 37 schizophrenia cases and 37 controls, we resequenced upstream of the type I–IV promoters, and the HapICE repeat regions in intron 1. Relative abundance of seven NRG1 mRNA transcripts in the prefrontal cortex were determined and compared across diagnostic and genotypic groups. We identified 26 novel DNA variants and showed an increased novel variant load in cases compared with controls (χ2=7.815; P=0.05). The average nucleotide diversity (θ=10.0 × 10−4) was approximately twofold higher than that previously reported for BDNF, indicating that NRG1 may be particularly prone to genetic change. A greater nucleotide diversity was observed in the HapICE linkage disequilibrium block in schizophrenia cases (θ(case)=13.2 × 10−4; θ(control)=10.0 × 10−4). The specific HapICE risk haplotype was associated with increased type III mRNA (F=3.76, P=0.028), which in turn, was correlated with an earlier age of onset (r=−0.343, P=0.038). We found a novel intronic five-SNP haplotype ∼730 kb upstream of the type I promoter and determined that this region functions as transcriptional enhancer that is suppressed by SRY. We propose that the HapICE risk haplotype increases expression of the most brain-abundant form of NRG1, which in turn, elicits an earlier clinical presentation, thus providing a novel mechanism through which this genetic association may increase risk of schizophrenia.

We previously identified a significant bipolar spectrum disorder linkage peak on 15q25-26 using 35 extended families with a broad clinical phenotype, including bipolar disorder (types I and II), recurrent unipolar depression and schizoaffective disorder. However, the specific gene(s) contributing to this signal had not been identified. By a fine mapping association study in an Australian case-control cohort (n = 385), we find that the sialyltransferase 8B (ST8SIA2) gene, coding for an enzyme that glycosylates proteins involved in neuronal plasticity which has previously shown association to both schizophrenia and autism, is associated with increased risk to bipolar spectrum disorder. Nominal single point association was observed with SNPs in ST8SIA2 (rs4586379, P = 0.0043; rs2168351, P = 0.0045), and a specific risk haplotype was identified (frequency: bipolar vs controls = 0.41 vs 0.31; χ2 = 6.46, P = 0.011, OR = 1.47). Over-representation of the specific risk haplotype was also observed in an Australian schizophrenia case-control cohort (n = 256) (χ2 = 8.41, P = 0.004, OR = 1.82). Using GWAS data from the NIMH bipolar disorder (n = 2055) and NIMH schizophrenia (n = 2550) cohorts, the equivalent haplotype was significantly over-represented in bipolar disorder (χ2 = 5.91, P = 0.015, OR = 1.29), with the same direction of effect in schizophrenia, albeit non-significant (χ2 = 2.3, P = 0.129, OR = 1.09). We demonstrate marked down-regulation of ST8SIA2 gene expression across human brain development and show a significant haplotype×diagnosis effect on ST8SIA2 mRNA levels in adult cortex (ANOVA: F(1,87) = 6.031, P = 0.016). These findings suggest that variation the ST8SIA2 gene is associated with increased risk to mental illness, acting to restrict neuronal plasticity and disrupt early neuronal network formation, rendering the developing and adult brain more vulnerable to secondary genetic or environmental insults.