REST-Governed Gene Expression Profiling in a Neuronal Cell Model Reveals Novel Direct and Indirect Processes of Repression and Up-Regulation

Abstract

The role of REST changes in neurons, including the rapid decrease of its level during differentiation and its fluctuations during many mature functions and diseases, is well established. However, identification of many thousand possible REST-target genes, mostly based on indirect criteria, and demonstration of their operative dependence on the repressor have been established for only a relatively small fraction. In the present study, starting from our recently published work, we have expanded the identification of REST-dependent genes, investigated in two clones of the PC12 line, a recognized neuronal cell model, spontaneously expressing different levels of REST: very low as in neurons and much higher as in most non-neural cells. The molecular, structural and functional differences of the two PC12 clones were shown to depend largely on their different REST level and the ensuing variable expression of some dependent genes. Comprehensive RNA-Seq analyses of the 13,700 genes expressed, validated by parallel RT-PCR and western analyses of mRNAs and encoded proteins, identified in the high-REST clone two groups of almost 900 repressed and up-regulated genes. Repression is often due to direct binding of REST to target genes; up-regulation to indirect mechanism(s) mostly mediated by REST repression of repressive transcription factors. Most, but not all, genes governing neurosecretion, excitability, and receptor channel signaling were repressed in the high REST clone. The genes governing expression of non-channel receptors (G protein-coupled and others), although variably affected, were often up-regulated together with the genes of intracellular kinases, small G proteins, cytoskeleton, cell adhesion, and extracellular matrix proteins. Expression of REST-dependent genes governing functions other than those mentioned so far were also identified. The results obtained by the parallel investigation of the two PC12 clones revealed the complexity of the REST molecular and functional role, deciphering new aspects of its participation in neuronal functions. The new findings could be relevant for further investigation and interpretation of physiological processes typical of neurons. Moreover, they could be employed as tools in the study of neuronal diseases recently shown to depend on REST for their development.

Keywords: PC12 clones; RNA-Seq; cooperative transcription factors; different REST levels; differential gene expression; gene repression and up-regulation.

Figure 1

Differential gene expression in wt and hrPC12 cells. (A) Density plot of log2 mean abundance (reads mapped to a gene) in wtPC12 (red) and hrPC12 (blue) cells after normalization. (B) log/log mean abundance plot showing the distribution of genes in the two clones. Points distant to the red line reveal differential mean values. The distribution of the abundance and differences between the two clones are those expected for a RNA-Seq investigation. (C) Volcano plot of changes. Red circles represent significant gene differences, repressed (left) and up-regulated (right) in the hrPC12 cells (adjusted p-values < 0.01).

Figure 2

RT-qPCR Validation. The Figure compares the wtPC12/hrPC12 values of 24 genes, obtained by RT-qPCR duplicate samples, to the values obtained by RNA-Seq. Panel (A) illustrates the close matching results of 19 genes (qPCR values are mean of log2 fold change ± SD; Pearson correlation coefficient, R² = 0.986). Panel (B) shows the PCR message results of 5 genes exhibiting in one clone values too little to be appropriately appreciated by RT-qPCR. The samples were only run on gels (lane 1 and 2 = hrPC12; 3 and 4 = wtPC12). The results confirm Ntrk2 (the TrkB receptor) to be repressed, and the other 4 genes to be up-regulated in the hrPC12 cells (RNA-Seq ratios: Ntrk2, −5.2; Fn1, 7.41; Nts, 9.44; Lamc3, 7.47; and Myod1, 8.18).

Figure 3

Proteins encoded by genes differentially expressed in hrPC12 and wtPC12 cells. Representative results of western blots of 11 proteins analyzed 2–4 times. Panel (A) shows the results of the repressed post-synaptic protein Shank3 and K⁺ channel Kcnk, with the up-regulated annexin1 (Anxal); controls with tubulin (Tub), and GAPDH. Panel (B) shows the results of the repressed Grin1, a subunit of the NMDA receptor, and the unchanged Slca17a7 Na⁺ transporter, with GAPDH control; Panels (C–E) show the repressed post-synaptic protein Shank2, transcription factor Ascl1, SNARE protein SNAP25, and MAGUK Dlg4, together with the moderately repressed FEV and the up-regulated synaptic protein Eps8. Controls with GAPDH.

Figure 4

Quantification of the western blot data of Figure 3. The proteins encoded by 9 out of 11 genes: 7 down-regulated, 1 unchanged, and 1 up-regulated, exhibit levels parallel to the RNA-Seq values. Annexin 1 (Anxal) exhibits a significant up-regulation, however less marked than the up-regulated gene expression shown by RNA-Seq. The transcription factor protein Fev appears unchanged whereas its gene expression is strongly repressed. Mean fold change values are shown ± standard deviation. Western blot replicates employed ranged from 2 to 4. Statistical analysis was carried out by a paired t-test. For the RNA-Seq analysis, the adjusted p-value of the negative binomial (DESeq) was used. (^*p < 0.05, ^**p < 0.01, ^***p < 0.001).

Figure 5

Mechanistic insights into the REST gene expression control of PC12 cells. (A) Shortest Path networks linking, through transcriptional regulation, the PRC 1, 2, and the REST complex proteins to genes repressed in hrPC12. The labeled circles represent the genes repressed in our dataset, shown in the Metacore™ databases and ENCODE dataset as putative targets of either the PRC 1-2 proteins (Suz12, EED, vEHZ2, and JARID2, blue circles) or the REST complex proteins (REST, Sin3B, Sin3A, and coREST, yellow circles). The putative targets of both REST and PRCs are shown in green. For repressed gene targets, the size of the node is proportional to the log2FC between hrPC12 and wtPC12 cells. The top 10-enriched Gene Ontology Biological Processes, carried out using the DAVID tool of enrichment, show genes under putative repression by REST only, PCR only or both. (B) Heatmap of the repressed transcription factor genes used in (C). Statistical significance: cut-off at log2FC > ±2, adjusted p-values < 0.01. Levels of expression are shown after variance stabilization transformation and color code as in the scale shown. (C) Shortest path network linking the repressed hrPC12 transcription factor genes (green) of B to genes excluded from the network of REST/PRC-governed genes of (A) (red). In addition to Rest (included in the network to highlight the transcription factors that are among its potential direct targets), the most relevant genes of factors are Gata-2, Ascl1, and c-Fos (AP-1). The size of the nodes is as in (A). (D) Top connected hubs of a direct interaction network containing the repressed and up-regulated genes. The vertical position and size of the node indicates the number of connections with modulated genes. The horizontal position refers to their log2FC behavior. The red/left nodes correspond to the genes repressed in hrPC12 cells; the green/right nodes to the up-regulated ones. Black arrows indicate interactions predicted with unspecified effects; red and green arrows indicate interactions with experimental evidence for inhibitory and activatory effects, respectively.