Characterization of SALL2 gene isoforms and its targets across cell types reveals highly conserved networks. Characterization of SALL2 gene isoforms and its targets across cell types reveals highly conserved networks

The SALL2 transcription factor (TF), an evolutionarily conserved gene through vertebrates, is involved in normal development and neuronal differentiation and considered as a tumor suppressor in certain human cancers. Several transcriptional targets of SALL2 are identified, these include the p21CDKN1A and p16INK4A cyclin-dependent kinase inhibitors, and the PMAIP1 and Bax pro-apoptotic genes, among others in various cell types. The human and mouse SALL2 gene loci contain two promoters, each one controlling the expression of a different protein isoform (namely E1 and E1A). However, several improvements on the human genome assembly and gene annotation through next-generation sequencing technologies over time reveal correction and annotation of additional isoforms, obscuring dissection of SALL2 isoform-specific transcriptional targets. We here integrated current data of normal/tumor gene expression databases along with ChIP-seq binding profiles to analyze SALL2 isoforms expression distribution and infer isoform-specific SALL2 targets. We found that the canonical SALL2 isoform (E1) is one of the lowest expressed, while isoform E1A is highly predominant across cell types. To dissect SALL2 isoform-specific targets, we analyzed publicly available ChIP-seq data from glioblastoma multiforme (GBM) and in-house ChIP-seq datasets performed in SALL2 wild-type and isoform E1A knockout HEK293 cells. Another available ChIP-seq data in HEK293 cells (ENCODE Consortium Phase III) overexpressing a non-canonical SALL2 isoform (herein named short_E1A) was analyzed but not included in the final analysis because we demonstrated that short_E1A is mostly localized in the cytoplasm, making impractical to dissect its direct transcriptional targets in this cell model. Regardless of cell type, our analysis reveals a highly conserved network of brain-specific TFs (i.e., SALL3, POU3F2, and NPAS3) and PODXL as a gene that is likely regulated by SALL2 across cell types. Our data integration identified a conserved molecular network in which SALL2 regulates target genes and encourages validation of publicly available ChIP-seq datasets for assessing transcriptional targets of a specific gene/isoform. Financial support: This work was supported by Fondecyt Regular Grants #1151031, #1191172 to Roxana Pincheira Fondecyt Regular Grant #120 to Ariel Castro, Postdoctorate Fondecyt Grant #3160129 and Fondecyt de Iniciacion # 1119028 to Matias I.Hepp. Overall design: ChIP-seq of endogenous SALL2 performed in wild-type HEK293 cells, SALL2 E1A knockout isoforms in HEK293 cells and SALL2 total knockout in HEK293 cells

SALL2转录因子（transcription factor, TF）是一类在脊椎动物中进化保守的基因，参与正常发育与神经元分化过程，并在部分人类癌症中被认定为肿瘤抑制因子。目前已在多种细胞类型中鉴定出SALL2的多个转录靶标，包括p21CDKN1A、p16INK4A这类细胞周期蛋白依赖性激酶抑制剂，以及PMAIP1、Bax等促凋亡基因。人类和小鼠的SALL2基因座包含两个启动子，分别调控不同的蛋白质异构体（即E1与E1A）。然而，随着下一代测序技术对人类基因组组装与基因注释的不断优化，后续研究发现了更多的异构体并修正了原有注释，这使得SALL2异构体特异性转录靶标的精准解析变得困难。本研究整合了正常/肿瘤基因表达数据库的现有数据，结合染色质免疫共沉淀测序（Chromatin Immunoprecipitation sequencing，ChIP-seq）结合峰谱，对SALL2异构体的表达分布进行分析，并推断异构体特异性的SALL2靶标。研究发现，经典SALL2异构体（E1）的表达量处于极低水平之一，而异构体E1A在各类细胞中均占主导地位。为解析SALL2异构体特异性靶标，我们分析了公开获取的多形性胶质母细胞瘤（glioblastoma multiforme, GBM）中的ChIP-seq数据，以及在SALL2野生型与E1A异构体敲除的HEK293细胞中获得的自有ChIP-seq数据集。另有一份公开的HEK293细胞ChIP-seq数据（ENCODE联盟第三期），该样本过表达非经典SALL2异构体（本文命名为short_E1A），但未纳入最终分析——因为我们证实short_E1A主要定位于细胞质，使得在该细胞模型中解析其直接转录靶标变得不可行。无论细胞类型如何，本研究的分析均揭示了一个高度保守的脑特异性转录因子网络（即SALL3、POU3F2与NPAS3），以及PODXL这一可能在各类细胞中均受SALL2调控的基因。本研究的数据整合工作确定了一个保守的分子网络，其中SALL2可调控其靶标基因，并提示可通过公开的ChIP-seq数据集验证特定基因/异构体的转录靶标。资助信息：本研究获得的资助包括面向Roxana Pincheira的Fondecyt Regular项目编号#1151031、#1191172，面向Ariel Castro的Fondecyt Regular项目编号#120，以及面向Matias I. Hepp的Fondecyt博士后项目#3160129与Fondecyt启动项目#1119028。实验整体设计：在野生型HEK293细胞、SALL2 E1A异构体敲除的HEK293细胞以及全SALL2敲除的HEK293细胞中，对内源SALL2进行染色质免疫共沉淀测序（ChIP-seq）检测。