4.1 Study Population
We recruited 1,584 patients from southern China, including 528 schizophrenia patients with a mean age of onset at 47.3 ± 13.4 (52.3% male), 528 bipolar disorder patients with a mean age of onset at 41.1 ± 12.8 (55.3% male) and 528 unrelated healthy individuals with a mean age of 42.72 ± 13.1 (43.6% male) as the control group. Schizophrenia and bipolar disorder patients were diagnosed on the basis of DSM-III-R criteria . Each patient was assessed by at least two psychiatrists independently according to the case records and interviews. A standard informed consent was signed by each participant and reviewed and approved by the Shanghai Ethical Committee of Human Genetic Resources.
4.2 SNP selection
Since SNPs on or near the pre-miRNAs are more likely to influence the generation and function of mature miRNAs, we chose to analyze the SNPs within the 300 bp sequence of mature miRNA. Candidate miRNA genes were identified from three different ways: (1) the databases of miRBase version 18.0 and Hapmap Phase III were downloaded; according to miRBase version 18.0 data, the genomic location of miRNA genes was determined; by searching SNPs in or near miRNA genomic region according to Hapmap Phase III data then, we found 457 SNPs on or near (within 300bp) 307 miRNA genes; (2) we utilized data of SNPedia miRNA project to find known SNPs near or on miRNAshttp://www.snpedia.com/index.php/SNPedia_microRNA_project; (3) we searched in PUBMED database by using the keywords "microRNA", "brain", “CNS”, "bipolar", "schizophrenia", “psychiatry” to find miRNAs which had been reported to be expressed in the brain and may play functional roles in psychiatric diseases. Finally, we chose 10 SNPs located in psychiatry susceptible loci, whose minor allele frequency was larger than 0.05 in Chinese Han population. (Table 1).
4.3 Genotyping and power calculation
Genomic DNA was extracted from peripheral blood lymphocytes using phenol-chloroform method. Selected SNPs were genotyped by TaqMan® SNP Genotyping Assays (Applied Biosystems, Foster City, CA) on ABI PRIM 7900 Sequence Detection Systems. The results were then analyzed by SDS 2.2 software (Applied Biosystems) for allelic discrimination. Allele deviations were assessed using Hardy Weinberg equilibrium (HWE) and the differences of allele and genotype frequencies between cases and controls were compared using SHEsis . Three models were constructed using R program. Homozygote (1/1) and heterozygote (1/0) risk allele were coded as 2 and 1, respectively and homozygote non-pathogenic allele (0/0) were coded as 0. The dominant model was defined as 1/1 + 1/0 versus 0/0 and the recessive model as 1/1 versus 1/0 + 0/0. Power calculations were post-hoc calculations performed on the G*Power 3.0 program. All reported P-values were two-tailed and statistical significance was defined as p < 0.05.
4.4 Plasmids and cells
LentiCRISPRv2 backbone plasmid was purchased from Addgene (#52961). sgRNA targeted to pre-miR-27a was designed (gtggctaagttccgcccccc) and cloned into LentiCRISPRv2 using Esp 3I (Thermofisher Scientific, Waltham, CA, USA) digestion as described  and this plasmid was referred to as lenti-CRISPR-miR-27a thereafter. For overexpression experiments, pri-miR-27a sequences containing T (wild type) and C (mutant) at rs895819 were cloned into pcDNA 3.1(+) under the control of CMV promoter.
U-251MG cell line was purchased from the BeNa Culture Collection and cultured in DMEM (Thermofisher Scientific) supplemented with 10% FBS (Thermofisher Scientific), 100 IU/mL of penicillin and 100 µg/mL of streptomycin at 37°C in a fully humidified atmosphere containing 5% CO2. U-251MG cells were verified prior to use by Saily Bio (Shanghai, China). Neural progenitor cells (NPCs) were purchased from ATCC (ACS-5003) and cultured in STEMdiff neural progenitor medium (STEMCELL Technologies, BC, Canada) at 37°C in a fully humidified atmosphere containing 5% CO2.
4.5 Cell line construction
Lentivirus carrying Cas9 and sgRNA was produced by transfecting 293T cells with lenti-CRISPR-miR-27a plasmid along with two helper plasmids pMd2.G (Addgene #12259) and psPAX2 (Addgene #12260) using lipofectamine 3000 (Invitrogen, CA, USA). At 72 h post transfection, the supernatant containing the lentivirus was harvested, filtered, and stored at -80 ºC for further application. U-251MG cells were infected with a MOI of 0.3 and selected using puromycin as described . At 7 days after puromycin selection, gene modification was verified by DNA sequencing. Single clones carrying edited miR-27a sequence were isolated by limited dilution and genotyped by DNA sequencing. The primers for the PCR amplification of miR-27a are listed in Supplementary Table 1.
4.6 RNA-Seq analyses
WT and miR-27a knockout U-251MG cells were harvested with three biological replicates in each group. Total RNA was extracted using Trizol reagent (Invitrogen, CA, USA). Whole-transcriptome sequencing was performed and analyzed by Vazyme Biotech (Nanjing, Jiangsu, China). RNA-Seq short reads were aligned to the human genome (GRCh38) using HISAT  with a maximum of two mismatches. On average, approximately 55.5 million reads across all samples were aligned to the reference, accounting for 96.5% of the total reads. Gene expression was counted as the number of short reads fully or partially aligned to the annotated gene model and was presented as expected number of Fragments Per Kilobase of transcript sequence per Millions base pairs sequenced (FPKM). Expressed genes were defined as genes with more than ten reads in total mapped in all samples and at least two of three replicates having more than two reads each. In total, 15,406 genes met the criteria and were defined as expressed in both WT and miR-27a KO U-251MG libraries. Differentially expressed genes (DEGs) were identified using the CuffDiff module in CuffLink program  and are normalized to the library size between samples. P-values were adjusted for multiple testing using false discovery rates . Significant DEGs were identified with an FDR ≤ 0.05 and a log2 (fold change) ≥ 1. Gene Ontology (GO) enrichment analyses was performed using Gorilla  by comparing the up/down-regulated DEGs to a list of all expressed genes. Significant GO terms with FDR ≤ 0.05 were reported.
4.7 RT-qPCR analyses of candidate miR-27a targets
The total RNA of WT and miR-27a KO U-251MG was extracted using Trizol reagent (Invitrogen, CA, USA) according to the manufacturer's instructions. One microgram of RNA was transcribed into cDNA using random primers and the SuperScript™ III First-Strand Synthesis System (Invitrogen, Massachusetts, USA). RT-qPCR was performed using the resulting cDNA templates and the GoTaq® qPCR Master Mix (Promega) in an Applied Biosystems 7900 Real-Time PCR Cycler (Applied Biosystems). Primers for RT-qPCR were designed by Beacon Designer 7.01 (www.premierbiosoft.com). The RT-qPCR data were analyzed using SDS software (Applied Biosystems). GAPDH or U6 small nuclear RNA (snRNA) were used as internal control for mRNA and microRNA, respectively.
MiR-27a mimic/inhibitor and their negative control (NC) were all purchased from Shanghai GenePharma Co. Ltd (Shanghai, China). We utilized Lipofectamine RNAiMAX reagent (Invitrogen, USA) to transfect miR-27a mimic/inhibitor or their corresponding control into U-251MG cells according to the protocol provided by the manufacturer. MiR-27a mimic and inhibitor at two doses (30pmol, 90pmol) were used to up-regulate or down-regulate miR-27a expression. Then we investigated the relative mRNA expression of candidate target genes compared with negative control using QPCR technology.
4.4 Western Blotting
The miR-27a knockout mutant constructed on the U-251MG cell line and wild-type were plated in 6-well plates at a density of 3 × 105 cells/mL and U-251MG were treated with mir-27a mimics and inhibitor using RNAiMAX at the indicated concentrations. Cells were collected after 72h, washed with cold 1× PBS, and lysed in 1× SDS buffer containing protease inhibitor cocktails (HY-K0010, MCE). Protein in cell lysate was quantified by detergent compatible Bradford assay kit (no. 23246, Thermo). Primary antibodies used in this study include ICAM1 antibody (ab53013, abcam) and β-actin (#4970, Cell Signaling Technology). The Millipore immobilon Western chemiluminescence substrate was used for signal development. Blots were imaged in an Amersham Imager 600 (GE Healthcare). The gray values of the bands were quantified by ImageJ software. The expression of β-actin was used as an internal control.
4.10 Overexpression of WT and mutant miR-27a in NPCs
The NPCs were counted and seeded on to 6 cm dish to reach a confluency of 80% on the day of transfection. The medium was replaced with Opti-MEM medium (Gibco) at 1 h before transfection. Transfection was performed with 5 µg miR-27a overexpression plasmid for each well using lipofectamine 3000 (Invitrogen). At 4 h after transfection, the medium was replaced with complete medium and incubated in 37°C. After 20 h, the NPCs were detached by ACCUTASE (STEMCELL Technologies) and collected by centrifugation at 1000 rpm for 5 min for further analyses.
4.11 Statistical analyses
All experimental data were presented as means ± standard deviation (S.D.). Two-tailed Student’s t tests (T-test) were used to evaluate differences between groups. ANNOVA tests were used to evaluate the significance of differences among three or more groups All data were representative of an average of three independent experiments and significant differences were indicated as * p < 0.05;** p < 0.01;*** p < 0.001.