Pigini Paolo, Xu Huilin, Ji Yan, Lindmeier Hannah, Saltzman Henry R, Yun Shuqi, Alves Christiano R R, Silva M Catarina, Gao Dadi, Morini Elisabetta
bioRxiv. 2025 Jun 27:2025.06.26.661736. doi: 10.1101/2025.06.26.661736.
Poison exons (PE) are highly conserved exon cassettes whose inclusion creates premature termination codons (PTCs) and triggers nonsense-mediated decay (NMD) of the mature transcript, therefore reducing protein expression. Despite their important role in post-transcriptional regulation, PEs remain poorly annotated due to the lack of systematic, transcriptome-wide approaches. In this study, we comprehensively investigated the function of PEs in the human brain and assessed the impact of pathogenic variants on their splicing. A comparative analysis of 957 eukaryotic transcriptomes revealed that humans exhibit the highest enrichment of NMD-targeted transcripts. To delineate the full landscape of PEs in the human transcriptome, we considered conserved (PhyloP score ≥ 20) alternatively spliced exons, less than 300 base pairs (bp) long, with the ability to introduce PTCs positioned more than 150 base pairs (bp) downstream of the transcription start site and outside the last two exons of the transcript. Our analysis identified 12,014 PEs in the human genome. For each PE, we calculated the percent-spliced-in (PSI) value across tissue types and developmental stages using GTEx and BrainSpan RNA-seq dataset, respectively. Overall, we identified 117 PEs uniquely found in the human brain, 1,214 PEs with brain-specific differential splicing compared to other tissues, and 1,610 PEs which splicing change during brain development. Integrating ClinVar and SpliceAI, we identified 1,877 annotated pathogenic variants predicted to affect the splicing of 891 PEs in the human brain. Notably, many of these variants were associated with neurodevelopmental and neurodegenerative disorders such as epilepsy, intellectual disability and frontotemporal dementia. We functionally validated the impact of selected variants using an optimized CRISPR prime-editing workflow in human cell lines. Our findings highlight PEs as pivotal regulators of gene expression in the human brain and establish a foundation for therapeutic strategies targeting PE mis-splicing in neurological diseases.
毒性外显子（PE）是高度保守的外显子盒，其包含会产生过早终止密码子（PTC）并触发成熟转录本的无义介导衰变（NMD），从而降低蛋白质表达。尽管它们在转录后调控中发挥着重要作用，但由于缺乏系统的全转录组方法，PE的注释仍然很少。在本研究中，我们全面研究了PE在人类大脑中的功能，并评估了致病变异对其剪接的影响。对957个真核转录组的比较分析表明，人类中NMD靶向转录本的富集程度最高。为了描绘人类转录组中PE的全貌，我们考虑了保守的（PhyloP评分≥20）可变剪接外显子，长度小于300个碱基对（bp），能够引入位于转录起始位点下游超过150个碱基对（bp）且在转录本最后两个外显子之外的PTC。我们的分析在人类基因组中鉴定出12,014个PE。对于每个PE，我们分别使用GTEx和BrainSpan RNA测序数据集计算了跨组织类型和发育阶段的剪接百分率（PSI）值。总体而言，我们在人类大脑中鉴定出117个独特的PE，与其他组织相比有1,214个具有大脑特异性差异剪接的PE，以及1,610个在大脑发育过程中剪接发生变化的PE。整合ClinVar和SpliceAI，我们鉴定出1,877个注释的致病变异，预计会影响人类大脑中891个PE的剪接。值得注意的是，这些变异中的许多与神经发育和神经退行性疾病有关，如癫痫、智力残疾和额颞叶痴呆。我们使用优化的CRISPR碱基编辑工作流程在人类细胞系中功能验证了所选变异的影响。我们的研究结果突出了PE作为人类大脑中基因表达的关键调节因子，并为针对神经系统疾病中PE错配剪接的治疗策略奠定了基础。
