Recent studies have shown that most genetic factors predisposing to schizophrenia have only a modest effect. GWA studies alone seem insufficient to identify the majority of these genetic factors. Expression level is an index of function of genes and may be useful for identifying risk genes for schizophrenia at the transciptomic level. In this study, we took advantage of recently available next generation sequencing technologies (i.e., RNA-Seq) to sequence poly-A tailed mRNAs from blood samples of 6 individuals and 2 pools of schizophrenia patients and controls. In the 6 individually sequenced samples, we found 218 genes showing differentially expression between cases and controls. Among these genes, 19 were nominally significant at the expression level in the 2 pooled samples. In our IPA analysis, we found that MIF regulation of innate immunity and TREM1 signaling were highly enriched in these 19 genes. Furthermore, of the 218 DEGs, 21 reached nominal significance in gene-based association analysis of the MGSGWAS dataset. Nineteen of these 21 genes are directly involved in immune response/diseases, or have been studied for candidates for schizophrenia and other neuronal diseases. Two genes, S100A8 and TYROBP, showed the same direction of expression changes in the individual and pooled sequencing datasets, and they also reached nominal significance in gene-based association analysis.
S100A8, also called MRP-8, encodes a calcium binding protein involved in inflammatory responses. It has been implicated in rheumatoid arthritis , systemic lupus erythematosus  and cancers [45, 46]. Intriguingly, rheumatoid arthritis may be correlated with schizophrenia [47, 48]. TYROBP, also known as DAP12, encodes an immunoreceptor adaptor protein that plays a key role in osteoclast differentiation and maturation [49, 50]. Mutations in this gene lead to the Nasu-Hakola disease [51–54], a rare autosomal recessive disorder characterized by bone cyst and presenile dementia. In addition to their functions in the immune system, both genes are expressed in human brain. S100A8 shows elevated expression in cerebral ischemia  and posttraumatic brain injuries . In a mouse model study, S100A8 expression increases significantly after chronic treatment with the antipsychotic drug olanzapine, which is used primarily to treat schizophrenia and bipolar disorder patients. TYROBP is implicated in the developmental neuronal death in hippocampus , impaired glutamatergic synaptic functions  and brain myelination . All of these factors have been suspected to be involved in schizophrenia. TYROBP knockout mouse studies reveal deficits in cognitive functions and prepulse inhibition , symptoms that have been manifested in many schizophrenia patients. However, neither gene has been studied directly for schizophrenia. They may be novel candidates for the disease.
Glatt et al  applied microarray techniques to compare gene expression of peripheral blood cells (PBCs) and the dorsolateral prefrontal cortex (DLPFC) of the brain to identify risk factors for schizophrenia. They detected 123 differentially expressed genes in the blood samples. Among our 218 DEGs, 13 genes had the same direction of expression changes as reported by Glatt et al. Specifically, eight genes were down-regulated (CD74, FCN1, FGR, HLA-DPA1, HLA-DRB1, IL10RA, PSAP, and ZFP36L2) and five were up-regulated (GOS2, HBA1, HBA2, HBB, and IL8). The overlap of 13 genes with same direction of expression change is unlikely by chance considering they were selected from a genome-wide gene pool (P-value = 5.36 × 10-6). Interestingly, among the eight down-regulated genes, CD74 was consistently found down- regulated in three gene expression data sets (our individual sample, our pooled sample, and PBC sample in the Glatt et al study). Gene CD74 encodes a protein in MHC and is located in a region implicated by genome-wide linkage meta-analyses [61, 62]. Additionally, the MHC locus on chromosome 6p was the most consistent finding from GWA studies [1–3]. Another gene, HLA-DRB1, also located in the MHC locus, was found to be differentially expressed in three data sets (the individual RNA-Seq dataset in this study, the PBC and DLPFC datasets in the Glatt et al study). HLA-DRB1 has been reported for positive association with schizophrenia [63, 64]. We also found that SNPs influencing the expression of HLA-DRB1 (expression quantitative trait loci - eQTLs) were significantly associated with schizophrenia in the CATIE and MGS datasets (unpublished data). This result provides empirical evidence that a combination of GWA data and eQTL analysis may be effective to identify risk genes.