Schizophrenia and bipolar disorder (BPD) are believed to share clinical features, etiological factors, and disease pathologies (such as impaired cognitive functions and dendritic spine pathology). Meanwhile, there is growing evidence of shared genetic risk between schizophrenia and BPD, despite that our knowledge of the functional risk variations and biological mechanisms is still limited. Here, we conduct summary data-based Mendelian randomization (SMR) analyses through combining the statistical data from genome-wide association studies (GWAS) of both schizophrenia and BPD and multiple expression quantitative trait loci (eQTL) datasets of the human brain dorsolateral prefrontal cortex (DLPFC) tissues. These integrative investigations identify a lead risk locus at the chromosome 3p21.1 region, which contains numerous single-nucleotide polymorphisms (SNPs) in varied linkage disequilibrium (LD) and encompasses more than 20 genes. Further analyses suggest that many SNPs at 3p21.1 are significantly associated with both schizophrenia and BPD, and even depression, and the psychiatric risk alleles at 3p21.1 are correlated with mRNA expression of multiple genes such as NEK4, GNL3, and PBRM1. We also identify a 335-bp functional Alu polymorphism rs71052682 in significant LD with the psychiatric GWAS risk SNP rs2251219, and confirm the regulatory effects of this Alu polymorphism on transcription activities. We then explore the involvement of the 3p21.1 locus in the common clinical features and etiology of these illnesses. We reveal that psychiatric risk alleles at 3p21.1 in low-to-high LD consistently predict worse cognitive functions in humans, and manipulating the gene expression (NEK4, GNL3, and PBRM1) linked with higher genetic risk could reduce the density of mushroom dendritic spines in rat primary cortical neurons, mirroring the spine pathology in the prefrontal cortex of psychiatric patients. Our results find that, although the risk alleles at 3p21.1 are in low-to-moderate LD spanning a large genomic area, their underlying biological mechanisms in psychiatric disorders likely converge. These results provide essential insights into the neural mechanisms underlying the chromosome 3p21.1 risk locus in the shared pathological and etiological features of both schizophrenia and BPD.

精神分裂症和双相情感障碍 (BPD) 被认为具有共同的临床特征、病因和疾病病理（例如认知功能受损和树突棘病理）。与此同时，越来越多的证据表明精神分裂症和 BPD 之间存在共同的遗传风险，尽管我们对功能风险变异和生物学机制的了解仍然有限。在这里，我们通过结合来自精神分裂症和 BPD 的全基因组关联研究 (GWAS) 的统计数据和人脑背外侧前额叶皮层的多表达数量性状基因座 (eQTL) 数据集进行基于汇总数据的孟德尔随机化 (SMR) 分析(DLPFC) 组织。这些综合研究确定了染色体 3p21.1 区域的先导风险位点，它包含许多处于不同连锁不平衡（LD）中的单核苷酸多态性（SNP），并包含20多个基因。进一步分析表明，3p21.1 处的许多 SNP 与精神分裂症和 BPD 甚至抑郁症均显着相关，并且 3p21.1 处的精神病风险等位基因与多个基因的 mRNA 表达相关，例如NEK4、GNL3和PBRM1。我们还在具有精神病学 GWAS 风险 SNP rs2251219 的显着 LD 中鉴定了一个 335-bp 功能性Alu多态性 rs71052682，并证实了该Alu多态性对转录活性的调节作用。然后，我们探讨了 3p21.1 基因座在这些疾病的常见临床特征和病因中的作用。我们揭示了从低到高 LD 中 3p21.1 的精神病风险等位基因始终预测人类认知功能较差，并操纵基因表达（NEK4、GNL3和PBRM1) 与较高的遗传风险相关可以降低大鼠初级皮层神经元中蘑菇树突棘的密度，反映精神病患者前额叶皮层的脊柱病理学。我们的研究结果发现，尽管 3p21.1 的风险等位基因处于跨越大基因组区域的低到中度 LD，但它们在精神疾病中的潜在生物学机制可能会趋同。这些结果为精神分裂症和 BPD 的共同病理和病因学特征中染色体 3p21.1 风险基因座的神经机制提供了重要的见解。