3.2. qPCR Validation of Candidate MHC II Genes Identified in Exon Array Data (Experiment 2)
The main effects for all qPCR analyses are shown in Table 6. Age and pH were selected as covariates, while diagnosis, brain region, and sex were main effects. The ANCOVA for age was significant (HLA-DRB1 and HLA-DPA1 p < 0.0005, while CD74 was nominally significant p = 0.05). Including pH in the ANCOVA was not significantly related to gene expression, and sex was significant for HLA-DRB1 expression (p = 0.001).
Highly significant main effects were found for HLA-DPA1 and CD74 (p-values < 10−20), and the resulting line plots are shown in Figure 5. For diagnosis effect, CD74 (p = 3.6 × 10−5) and HLA-DPA1 (p = 1.7 × 10−7) were significant, while HLA-DRB1 did not reach significance (p = 0.078). Only CD74 was significant for region by diagnosis interaction. Regional analyses were conducted and comparison results are shown in Table 7 A,B,C. The lack of findings for HLA-DRB1 in all brain regions by qPCR is consistent with the exon microarray findings. Three neuropsychiatric disorder groups (SZ, MDD, and BD) had a significantly lower expression of HLA-DPA1 in the hippocampus. Further, CD74 and HLA-DPA1 showed concordant expression at the brain region level, suggesting broad co-expression. SZ showed reduced expression in similar regions as CD74 and HLA-DPA1. HLA-DRB1 showed a variable relationship across diagnosis and brain regions, and, due to higher cycle numbers, might not be reliably detected in brain, at least the isoform that we measured by qPCR.