Advertisement

A neuronal cell-based reporter system for monitoring the activity of HDAC2

  • Kazuhiro Unemura
    Correspondence
    Corresponding author.
    Affiliations
    Medical innovation center, SK project, Kyoto University Graduate School of Medicine, Konoecho, Yoshida, Sakyo-ku, Kyoto, Kyoto 606-8315 Japan

    Shionogi & Co., Ltd. Laboratory for Drug Discovery and Disease Research, Shionogi Pharmaceutical Research Center, 3-1-1, Futaba-cho, Toyonaka-shi, Osaka, 561-0825, Japan
    Search for articles by this author
  • Masako Kawano
    Affiliations
    Medical innovation center, SK project, Kyoto University Graduate School of Medicine, Konoecho, Yoshida, Sakyo-ku, Kyoto, Kyoto 606-8315 Japan
    Search for articles by this author
  • Mai Takakura
    Affiliations
    Medical innovation center, SK project, Kyoto University Graduate School of Medicine, Konoecho, Yoshida, Sakyo-ku, Kyoto, Kyoto 606-8315 Japan
    Search for articles by this author
  • Ikuko Iwata
    Affiliations
    Medical innovation center, SK project, Kyoto University Graduate School of Medicine, Konoecho, Yoshida, Sakyo-ku, Kyoto, Kyoto 606-8315 Japan
    Search for articles by this author
  • Kana Hyakkoku
    Affiliations
    Shionogi & Co., Ltd. Laboratory for Drug Discovery and Disease Research, Shionogi Pharmaceutical Research Center, 3-1-1, Futaba-cho, Toyonaka-shi, Osaka, 561-0825, Japan
    Search for articles by this author
  • Naotaka Horiguchi
    Affiliations
    Shionogi & Co., Ltd. Laboratory for Drug Discovery and Disease Research, Shionogi Pharmaceutical Research Center, 3-1-1, Futaba-cho, Toyonaka-shi, Osaka, 561-0825, Japan
    Search for articles by this author
  • Tomohiko Okuda
    Affiliations
    Shionogi & Co., Ltd. Laboratory for Bio-Drug Discovery, Shionogi Pharmaceutical Research Center, 3-1-1, Futaba-cho, Toyonaka-shi, Osaka, 561-0825, Japan
    Search for articles by this author
  • Yukinori Hirano
    Correspondence
    Corresponding author.
    Affiliations
    Medical innovation center, SK project, Kyoto University Graduate School of Medicine, Konoecho, Yoshida, Sakyo-ku, Kyoto, Kyoto 606-8315 Japan

    Division of Life Science, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China
    Search for articles by this author
Open AccessPublished:October 10, 2022DOI:https://doi.org/10.1016/j.slasd.2022.10.001

      Abstract

      Given that histone acetylation via histone acetyltransferases (HATs) and histone deacetylases (HDACs) is significant in memory formation, HDAC2 has been thoroughly investigated as a potential therapeutic target for the treatment of cognitive dysfunction. Although HDAC inhibitors have been discovered through in vitro enzyme assay, off-target effects on other HDACs are common due to their conserved catalytic domains. Each HDAC could be regulated by specific intracellular molecular mechanisms, raising the possibility that a cell-based assay could identify selective inhibitors targeting specific HDACs through their regulatory mechanisms. Here, we propose a versatile, cell-based reporter system for screening HDAC2 inhibitors. Through RNA-sequencing from human cultured neuronal cells, we determined that expression of a transcriptional repressor, inhibitor of DNA binding 1 (ID1), is increased by knockdown of HDAC2. We also established the knock-in neuronal cell lines of a bioluminescence reporter gene to ID1. The knock-in cell lines showed significant reporter activity by known HDAC inhibitors and by HDAC2-knockdown but not by HDAC1-knockdown. Thus, our neuronal cell-based reporter system is a promising method for screening the specific inhibitors of HDAC2 but not HDAC1, by potentially targeting not only HDAC2, but also the regulatory mechanisms of HDAC2 in neurons.

      Introduction

      Memory is pivotal for daily life, as it allows us to communicate with people, guide our behavior, and participate in social activities. Dysfunction of memory significantly decreases the quality of life, as seen in patients with Alzheimer's disease. Given that the number of patients with Alzheimer's disease has been increasing with global population aging, understanding essential memory functions is necessary to identify effective therapeutic treatment options. Memory consolidation requires chromatin modification, including processes such as histone acetylation, which induces de novo gene expression. Neural activation during learning induces intracellular signaling [
      • Mews P.
      • Donahue G.
      • Drake A.M.
      • et al.
      Acetyl-CoA synthetase regulates histone acetylation and hippocampal memory.
      ], which then alters chromatin structure by employing epigenetic regulators significant for memory consolidation [
      • Nott A.
      • Watson P.M.
      • Robinson J.D.
      • et al.
      S-nitrosylation of histone deacetylase 2 induces chromatin remodelling in neurons.
      ,
      • Pao P.
      • Tsai L.
      Histone deacetylases 1 and 2 in memory function.
      ]. Previous studies have suggested that histone acetylation may play a key role in memory consolidation [
      • Levenson J.M.
      • O'Riordan K.J.
      • Brown K.D.
      • et al.
      Regulation of histone acetylation during memory formation in the hippocampus.
      ,
      • David S.J.
      Epigenetics and cognitive aging.
      ,
      • Peleg S.
      • Sananbenesi F.
      • Zovoilis A.
      • et al.
      Altered histone acetylation is associated with age-dependent memory impairment in mice.
      ] and have determined that it is regulated by histone acetyltransferase (HAT) and histone deacetylase (HDAC). Importantly, knockout of HDAC2, but not HDAC1, critically enhances memory [
      • Guan J.
      • Haggarty S.J.
      • Giacometti E.
      • et al.
      HDAC2 negatively regulates memory formation and synaptic plasticity.
      ,
      • Hanson J.E.
      • Deng L.
      • Hackos D.H.
      • et al.
      Histone deacetylase 2 cell autonomously suppresses excitatory and enhances inhibitory synaptic function in CA1 pyramidal neurons.
      ], demonstrating that epigenetic regulators, especially HDAC2, could be a potential target for the development of a drug to treat dementia.
      HDAC inhibitors have been developed for cancer therapy [
      • Eckschlager T.
      • Plch J.
      • Stiborova M.
      • et al.
      Histone deacetylase inhibitors as anticancer drugs.
      ], and recently considered as a potential target for the development of memory enhancement drugs [
      • Wagner F.F.
      • Zhang Y.
      • Fass D.M.
      • et al.
      Kinetically selective inhibitors of histone deacetylase 2 (HDAC2) as cognition enhancers.
      ,
      • Zhao W.
      • Ghosh B.
      • Tyler M.
      • et al.
      Class I histone deacetylase inhibition by tianeptinaline modulates neuroplasticity and enhances memory.
      ,
      • Fuller N.O.
      • Pirone A.
      • Lynch B.A.
      • et al.
      CoREST complex-selective histone deacetylase inhibitors show prosynaptic effects and an improved safety profile to enable treatment of synaptopathies.
      ], which could effectively treat dementia. There are four classes of HDAC family proteins. The development of the selective inhibitors for HDACs in each class have been unsuccessful due to the highly conserved catalytic domains [
      • Ganai S.A.
      • Abdullah E.
      • Rashid R.
      • et al.
      Combinatorial in silico strategy towards identifying potential hotspots during inhibition of structurally identical HDAC1 and HDAC2 enzymes for effective chemotherapy against neurological disorders.
      ,
      • Rodrigues D.A.
      • Pinheiro P.D.S.M.
      • Sagrillo F.S.
      • et al.
      Histone deacetylases as targets for the treatment of neurodegenerative disorders: Challenges and future opportunities.
      ]. Previously, a class I HDACs (HDAC1, 2, and 3) selective inhibitor, CI-994, and a pan-HDAC inhibitor, SAHA, have been reported to improve memory in rodent behavioral models such as fear conditioning or Morris water maze [
      • Benito E.
      • Urbanke H.;
      • Ramachandran B.
      • et al.
      HDAC inhibitor-dependent transcriptome and memory reinstatement in cognitive decline models.
      ,
      • Burns A.M.
      • Farinelli-Scharly M.
      • Hugues-Ascery S.
      • et al.
      The HDAC inhibitor CI-994 acts as a molecular memory aid by facilitating synaptic and intracellular communication after learning.
      ]. However, the use of CI-994 and SAHA must be limited as CI-994 causes thrombocytopenia, similar to the pan-HDAC inhibitor used in clinical studies [
      • Nemunaitis J.J.
      • Orr D.
      • Eager R.
      • et al.
      Phase I study of oral CI-994 in combination with gemcitabine in treatment of patients with advanced cancer.
      ,
      • Undevia S.D.
      • Kindler H.L.
      • Janisch L.
      • et al.
      A phase I study of the oral combination of CI-994, a putative histone deacetylase inhibitor, and capecitabine.
      ], most likely because non-specific HDAC inhibitors impair cell proliferation. Knockdown of both HDAC1 and HDAC2 suppresses cell proliferation of fibroblasts and liver cancer cell lines, unlike knockdown of either HDAC1 or HDAC2 [
      • Wilting R.H.
      • Yanover E.
      • Heideman M.R.
      • et al.
      Overlapping functions of Hdac1 and Hdac2 in cell cycle regulation and haematopoiesis.
      ,
      • Ler S.Y.
      • Leung C.H.W.
      • Khin L.W.
      • et al.
      HDAC1 and HDAC2 independently predict mortality in hepatocellular carcinoma by a competing risk regression model in a southeast asian population.
      ], suggesting their redundant functions in cell proliferation. Therefore, an HDAC2-specific inhibitor would be preferable to avoid peripheral side effects.
      Despite the highly conserved amino acid sequence between catalytic domains of HDAC1 and HDAC2 at 94%, HDAC2 is exclusively involved in memory function [
      • Guan J.
      • Haggarty S.J.
      • Giacometti E.
      • et al.
      HDAC2 negatively regulates memory formation and synaptic plasticity.
      ,
      • Hanson J.E.
      • Deng L.
      • Hackos D.H.
      • et al.
      Histone deacetylase 2 cell autonomously suppresses excitatory and enhances inhibitory synaptic function in CA1 pyramidal neurons.
      ], suggesting that HDAC2-specific regulation does not occur with HDAC1. Indeed, specific posttranslational modifications in HDAC2 and HDAC1, such as nitrosylation in HDAC2 and acetylation in HDAC1 [
      • Brunmeir R.
      • Lagger S.
      • Seiser C.
      Histone deacetylase 1 and 2-controlled embryonic development and cell differentiation.
      ,
      • Ma P.
      • Schultz R.M.
      HDAC1 and HDAC2 in mouse oocytes and preimplantation embryos: Specificity versus compensation.
      ] have been reported. Thus, HDAC2 selective inhibitors can be developed by targeting such exclusive regulatory mechanisms. However, due to the lack of an experimental system, we were unable to conduct screening for a drug targeting the HDAC2-specific regulatory mechanisms. To this end, we aimed to develop a cell-based reporter system that monitors HDAC2 activity. Using a human neuroblastoma cell line, we performed transcriptome analysis to determine the downstream genes of HDAC2, which can be utilized as a reporter gene. Knock-in of the bioluminescence reporter gene to the candidate gene demonstrated that the reporter expression is indeed increased by the known HDAC inhibitors. Thus, our cell-based system could be potentially used to screen the intracellular regulatory mechanisms for HDAC2.

      Materials and methods

      Plasmid construction

      MicroRNA expression vectors were constructed using BLOCK-iT™ Pol II miR RNAi Expression Vector Kits (K493600, ThermoFisher Scientific, San Jose, CA, USA). Oligo DNAs (Supplementary Table 3) were annealed and cloned into a pcDNA6.2-GW/EmGFP-miR vector, resulting in pmiR-HDAC2#1-5. CRISPR-based knock-in constructs included an sgRNA- and Cas9-expressing plasmid and a donor plasmid. To obtain an sgRNA-expressing plasmid, two gRNA target sequences were synthesized as previously described [
      • Ran F.A.
      • Hsu P.D.
      • Wright J.
      • et al.
      Genome engineering using the CRISPR-Cas9 system.
      ]. The obtained DNA fragment was cloned into Cas9-expressing plasmid, resulting in 2xID1-gRNAs-Cas9 plasmid. The sequences of gRNAs targeting nearby the start codon of ID1 were 5ʹ-GCCAAGAATCATGAAAGTCGCCAG-3ʹ and 5ʹ-GACTTTCATGATTCTTGGCGAC-3ʹ. To obtain a donor plasmid, the 1.0 kb genomic regions flanking the start codon of ID1 were amplified from the genomic DNA of SH-SY5Y cells. Substitutions of PAM sequences without changing amino acid sequence were included. The 5ʹ-1kb fragment was cloned into the XhoI-HindIII-digested pBluescript SK-(-) vector, and the 3ʹ-1kb fragment was cloned into the BamHI-SacI-digested resulting plasmid. The obtained plasmid was further digested by HindIII and BamHI and the PCR fragment containing Nanoluc::PEST and Hygr, which was amplified from pNL2.2[NlucP/Hygro] (N1071, Promega, Madison, WI, USA), resulting in the donor plasmid.

      Cell culture, transfection

      Human neuroblastoma SH-SY5Y cells were cultured in Nutrient Mixture F-12 Ham (N4888, Sigma, St. Louis, MO, USA)/E-MEM (M2279, Sigma, St. Louis, MO, USA) + 1% MEM Non-essential Amino Acid Solution (M7145, Sigma, St. Louis, MO, USA) + 2 mM L-Glutamine (G7513, Sigma, St. Louis, MO, USA) + 15% FBS (26140079, ThermoFisher Scientific, San Jose, CA, USA). Plasmid DNA of either miR-HDAC1, miR-HDAC2, or negative control miR (miR-neg, which does not target any vertebrate gene), were transfected to SH-SY5Y cells by electroporation using the NeonTM Transfection System (MPK5000, ThermoFisher Scientific, San Jose, CA, USA) with a pulse voltage at 1,300 V and pulse width of 30 ms, or NEPA21 (NEPAGENE, Chiba, Japan) with a poring pulse voltage at 185 V, pulse width of 2.5 ms for two times, and transfer pulse voltage of 20 V with a pulse width of 50 ms for five times, which achieves more than 90% transfection efficiency. The drugs used were CI-994 (HY-50934, MedChemExpress, New Jersey, USA), SAHA (S1047, Selleck Chemicals, Houston, TX, USA) and IOX-1 (ab144394, Abcam, Cambridge, UK).

      Quantification of transcripts (RT-qPCR)

      Total RNA was extracted from SH-SY5Y cells using a Nucleospin RNA plus kit (740984.50, Takara, Shiga, Japan). cDNA was synthesized with 0.2 μg of RNA using PrimeScript™ RT Master Mix (RR036A, Takara, Shiga, Japan), and analyzed via quantitative real-time PCR (BioRad Laboratories, Hercules, CA, USA or Applied Biosystems, Waltham, MA, USA) using TB Green® Premix Ex Taq™ II (RR820A, Takara, Shiga, Japan). Transcripts of GAPDH or YWHAZ were used for normalization.

      Quantification of protein

      Protein extracts were prepared from SH-SY5Y cells with 50 mM Tris-HCl (pH7.5), 120 mM NaCl, 0.5% NP-40 and protease inhibitor cocktail (4693124001, Sigma, St. Louis, MO, USA). Total protein contents were measured with BCA kit (23228, ThermoFisher Scientific, San Jose, CA, USA) and adjusted. HDAC2 protein was detected with HDAC2 antibody (ab12169, Abcam, Cambridge, UK) and its expression level was quantified by using Wes (ProteinSimple, San Jose, CA, USA).

      RNA-seq analysis

      Total RNA were processed via Oligo d(T)25 Magnetic Beads (New England Biolabs inc. Ipswich, MA, USA), according to the manufacturer's instructions. The obtained mRNA was used to generate a library using NEBNext Ultra II RNA Library Prep Kit for Illumina (New England Biolabs Inc. Ipswich, MA, USA), according to the manufacturer's instructions. Paired-end reads were generated in the HiSeq X Ten software (Illumina, San Diego, CA, USA).

      CRISPR knock-in

      The sgRNA-expressing plasmid and the donor plasmid were electroporated into SH-SY5Y cells. Transfected cells were selected by hygromycin B and single-cloned with limiting dilution. Genomic DNA was extracted from each clone and the insertion of Nanoluc::PEST was confirmed using PCR. Loss of DNA amplification from the endogenous ID1 locus was considered a homozygous knock-in; a heterozygous knock-in was characterized by the amplification of both the DNA of intact ID1 locus and ID1::Nanoluc-PEST. The copy number was further confirmed by quantitative PCR, by comparing amplification of NlucP gene to that of the endogenous ID1 gene (Figure 4B).

      Bioinformatic analysis

      Read quality was first assessed using FastQC (version 0.11.4) (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), and proceeded to adaptor trimming using trimmomatic [
      • Bolger A.M.
      • Lohse M.
      • Usadel B.
      Trimmomatic: A flexible trimmer for illumina sequence data.
      ], followed by mapping to the Human reference genome, GRCh38 from UCSC using STAR [
      • Dobin A.
      • Davis C.A.
      • Schlesinger F.
      • et al.
      STAR: Ultrafast universal RNA-seq aligner.
      ]. The filtered reads (4.3–8.9 million reads) were analyzed using HTSeq-count [
      • Anders S.
      • Pyl P.T.
      • Huber W.
      HTSeq-A python framework to work with high-throughput sequencing data.
      ]. The obtained numbers of the reads mapped on exons were analyzed in R ver. 4.0.3 using the DESeq2 package, and the resulting q-values and fold changes were plotted using the EnhancedVolcano package in R ver. 4.0.3 (Kevin Blighe, Sharmila Rana and Myles Lewis (2020)). Principal component analysis (PCA) was performed in R ver. 4.0.3 using Maptools (Roger Bivand and Nicholas Lewin-Koh (2022).). A heat map based on rpkm was generated using Treeview [
      • Saldanha A.J.
      Java treeview - extensible visualization of microarray data.
      ] (Fig. 1J). Finally, GO enriched groups were determined using WebGestalt (WEB-based Gene SeT AnaLysis Toolkit) [
      • Liao Y.
      • Wang J.
      • Jaehnig E.J.
      • et al.
      WebGestalt 2019: Gene set analysis toolkit with revamped UIs and APIs.
      ].
      Fig 1
      Fig. 1Genes regulated by HDAC2 in SH-SY5Y cells. A and B, Efficacy of HDAC2 knockdown. SH-SY5Y cells were transfected by plasmids expressing five different miRNAs for HDAC2 or carrying the control sequence (miR-neg). After 4 days, RNA was extracted and analyzed using RT-qPCR. Protein extracts were analyzed with anti-HDAC2 antibody. A one-way ANOVA indicated a statistically significant difference in (A) (P < 0.0001; n = 3). C–J, RNA-seq analysis. RNA prepared in panel A was subjected to RNA-seq analysis. C, Principal component analysis delineating the separation between the samples. D and E, Volcano plot indicating q-value and log2 fold change of genes with q-values < 0.05. F and G, Venn diagram showing overlap in increased or decreased genes on account of miR-HDAC2#2 and #5. H and I, Biological process-associated GO terms significantly enriched in 257 or 247 genes, which are increased or decreased by miR-HDAC2, respectively. J, Heatmap of the log2 fold change in expression of 504 genes significantly altered at > 1.5-fold by knockdown of HDAC2. Increased genes, 257; decreased genes, 247.

      Luciferase assay

      SH-SY5Y cells were plated on 96-well plates at the density of 20,000 cells/well (Fig. 4C-F) or on 384-well plates at the density of 5,000 cells/well (Fig. 4G-I). After incubation as indicated, cell number was first estimated using CellTiter-Fluor™ Cell Viability Assay (G6080, Promega, Madison, WI, USA). CellTiter-Fluor Reagent was added to the cell culture medium at a 1:4 ratio, mixed, and incubated for 30 min at 37 °C. Fluorescence was detected GloMax® Discover Microplate Reader (GM3000, Promega, Madison, WI, USA) or ARVO X (PerkinElmer, Waltham, MA, USA), and cell density was obtained according to the manufactural instruction. Then, the Nano-Glo® Luciferase Assay System (N1110, Promega, Madison, WI, USA) was used to measure Nanoluc expression, carried out according to the manufacturer's instruction. Briefly, the Nano-Glo Luciferase assay reagent with the substrate was added to the cell culture medium at a 1:1 ratio. After 5 min incubation, luminescence was detected using the GloMax® Discover Microplate Reader, ARVO X (PerkinElmer, Waltham, MA, USA) or PHERAstar FSX (BMG LABTECH, Ortenberg, Germany). The obtained values of luminescence were normalized by the cell density, and relative luminescence was calculated relative to the control sample.

      Statistical analysis

      No statistical calculations were used to predetermine sample sizes. Our sample sizes were similar to those generally used in this field of research. Statistical analyses were performed using Prism version 5.0. An unpaired two-tailed t-test was used for comparisons between two groups, and a one-way ANOVA, followed by Dunnet multiple comparisons was used for comparisons among multiple groups. P values < 0.05 were regarded as statistically significant. All data are presented as the mean ± s.e.m.

      Results

      Identification of the candidates of HDAC2-downstream target genes

      In this study, we aimed to establish a reporter cell line monitoring HDAC2 activity. A gene encoding a bioluminescence protein, luciferase, can be knocked-in to an HDAC2-target gene, whose expression would be increased by HDAC2 inhibition, thereby tracking HDAC2 activity via luminescence. Therefore, we performed transcriptome analysis to identify genes whose expression is increased by knockdown of HDAC2. We utilized human neuroblastoma SH-SY5Y cell line, which is proneuronal cell line and widely employed as an in vitro neuronal model for neurodegeneration [
      • Taylor-Whiteley T.R.
      • Le Maitre C.L.
      • Duce J.A.
      • et al.
      Recapitulating Parkinson's disease pathology in a three-dimensional human neural cell culture model.
      ], neurotrauma [
      • Skotak M.
      • Wang F.
      • Chandra N.
      An in vitro injury model for SH-SY5Y neuroblastoma cells: Effect of strain and strain rate.
      ], developmental neurotoxicity [
      • Attoff K.
      • Johansson Y.
      • Cediel-Ulloa A.
      • et al.
      Acrylamide alters CREB and retinoic acid signalling pathways during differentiation of the human neuroblastoma SH-SY5Y cell line.
      ] and neurite outgrowth [
      • Schikora J.
      • Kiwatrowski N.
      • Förster N.
      • et al.
      A propagated skeleton approach to high throughput screening of neurite outgrowth for in vitro parkinson's disease modelling.
      ]. Expression of miRNA targeting HDAC2 coding regions for 4 days in SH-SY5Y resulted in nearly 80% decrease in mRNA expression (Fig. 1A) and 30-40% decrease in protein expression of HDAC2 (Fig. 1B), compared to the one transfected with a control miRNA with a random sequence that does not target any vertebrate gene (miR-neg). The same RNA samples expressing miR-neg, as well as miR-HDAC2#2 and #5, were subjected to RNA-seq analysis. The principal component analysis demonstrated the clear segregation of gene expression profiles across the samples (Fig. 1C). We determined the differentially expressed genes by expressing miR-HDAC2 using DESeq2 (Fig. 1D and E). The number of significantly increased genes by miR-HDAC2#2 and #5 were 1,022 and 885, respectively, wherein genes showing >1.5-fold increase were 380 and 433 in number, respectively (Fig. 1F, Supplementary Table 1). Significantly decreased genes by miR-HDAC2#2 and #5 were 1,302 and 1,089, respectively, wherein genes showing > 1.5-fold decrease were 329 and 442 in number, respectively (Fig. 1G, Supplementary Table 2). The number of overlapping genes identified from the two independent analyses using miR-HDAC2#2 and #5 included 257 genes significantly increased by >1.5-fold and 247 genes significantly decreased by >1.5-fold (Fig. 1F and G). Gene ontology (GO) analysis showed enrichment of developmental genes in the increased genes (Fig. 1H), and enrichment of neural function-related genes in the decreased genes (Fig. 1I). The genes showing a robust increase in their expression included adrenoceptor alpha 2A (ADRA2A), TNF receptor superfamily member 12A (TNFRSF12A), SMAD family member 6 (SMAD6), and inhibitor of DNA binding 1 (ID1) inhibiting transcriptional activity of helix-loop-helix DNA binding proteins (Fig. 1J).

      Expression of ID1 is acutely and selectively increased by HDAC2 inhibition

      RNA-seq analyses were performed at 4 days after expressing miRNA for HDAC2. The differentially expressed genes included the direct target of HDAC2, as well as an indirect target; for instance, the transcription factor expressed by HDAC2-knockdown can lead to the expression of additional sets of genes. We intended to exclude such secondary gene expression, by time course experiments using the known HDAC inhibitor, CI-994 (IC50 = 0.9 µM). Administration of CI-994 at 3 µM and 10 µM was sufficient to induce ID1 mRNA with acute treatment for 24 h (Fig. 2A). Treatment with 10 µM CI-994 for 24 h significantly increased mRNA expression of ID1 and TNFRSF12A but not that of SMAD6 or ADRA2A (Fig. 2B), suggesting that ID1 and TNFRSF12A expressions are acutely altered by HDAC inhibition. Treatment with 3 µM CI-994 for 96 h, a concentration that is presumably less toxic comparing to 10 µM, further increased mRNA expression of SMAD6, but not ADRA2A (Fig. 2C). No increase in ADRA2A mRNA expression could be attributed to cross-reactions with other HDACs that eventually suppress increase in ADRA2A mRNA expression. To test the specificity of ID1 and TNFRSF12A expression to HDAC inhibition, IOX-1, a potent and broad-spectrum inhibitor targeting another class of epigenetic regulators, the JmjC demethylases, was also examined. IOX-1 inhibits demethylases at IC50 of 0.1-2.3 µM [
      • Schiller R.
      • Scozzafava G.
      • Tumber A.
      • et al.
      A cell-permeable ester derivative of the JmjC histone demethylase inhibitor IOX1.
      ]. Treatment of IOX-1 at 0.2 mM for 24 h significantly increased expression of TNFRSF12A mRNA, but not ID1 mRNA (Fig. 2D). Thus, ID1 could be a selective HDAC2-downstream target gene whose expression is acutely induced by HDAC2 inhibition.
      Fig 2
      Fig. 2Expression of candidates for HDAC2 target genes upon inhibitor treatment. A–C, Expression of HDAC2 target genes after treatment of CI-994. SH-SY5Y cells were treated with 3 or 10 μM of CI-994, or with a vehicle containing the same amount of DMSO as 10 μM of CI-994 (A). After 24 h (A, B) or 96 h (C), RNA was extracted and analyzed using RT-qPCR. (A) A one-way ANOVA indicated a statistically significant difference in relative expression (P < 0.0001; n = 3). (B) t-test; P = 0.000029, 0.00074, 0.011, 0.0031. (C) t-test; P = 0.023, 0.023, 0.024, 0.46. D, Expression of HDAC2 target genes after treatment of IOX-1. SH-SY5Y cells were similarly analyzed after overnight-treatment with 0.2 or 0.4 mM of IOX-1. RNA was extracted and analyzed using RT-qPCR. A one-way ANOVA indicated a significant difference in expression of both genes (P < 0.0001; n = 3). Data are represented as mean ± s.e.m. n.s., not significant P > 0.05; *, P < 0.05; **, P < 0.01; ***, P < 0.001.

      Expression of ID1 is not affected by HDAC1 inhibition

      To further test the selectivity of ID1 expression to HDAC2 inhibition, we examined the effect of HDAC1-knockdown. HDAC1-knockdown was induced by expressing miRNA targeting HDAC1 coding regions (Fig. 3A). Four days post-transfection, five miRNAs examined showed more than a 60% decrease in HDAC1 expression compared to a control miRNA (miR-neg). Expression of ID1 mRNA was not increased by any miRNA targeting HDAC1 (Fig. 3B), rather it was reduced by one miRNA (#1, Fig. 3B). This was not due to a compensatory effect of an increased level of HDAC2 expression by HDAC1-knockdown (Fig. 3C), suggesting that knockdown of HDAC1 could indirectly decrease ID1 expression. These results demonstrate that ID1 is a selective reporter gene discriminating HDAC1 and HDAC2 inhibition.
      Fig 3
      Fig. 3ID1 expression is not controlled by HDAC1. A, Efficacy of HDAC1-knockdown. SH-SY5Y cells were transfected by plasmids expressing five different miRNAs for HDAC1 or carrying a control sequence (miR-neg). After 4 days, RNA was extracted and analyzed using RT-qPCR. A one-way ANOVA indicated a statistically significant difference in relative expression (P < 0.0001; n = 3-4). B and C, Expression of ID1 (B) and HDAC2 (C) mRNAs after HDAC1-knockdown. RNAs were analyzed using RT-qPCR. A one-way ANOVA indicated a statistically significant difference in relative expression in panel B (P = 0.0212; n = 3-4), but not in panel C (P = 0.5264; n = 3-4). Data are represented as mean ± s.e.m. *, P < 0.05; ***, P < 0.001.

      A reporter cell line targeting ID1 shows robust response to known HDAC inhibitors

      Among luciferase genes, Nanoluc tagged with PEST sequence (NlucP), which induces rapid degradation and therefore reduces background signals, has been widely used for reporter gene assays [
      • Berry S.L.
      • Hameed H.;
      • Thomason A.
      • et al.
      Development of NanoLuc-PEST expressing leishmania mexicana as a new drug discovery tool for axenic- and intramacrophage-based assays.
      ,
      • Neefjes M.
      • Housmans B.A.C.
      • van den Akker G.G.H.
      • et al.
      Reporter gene comparison demonstrates interference of complex body fluids with secreted luciferase activity.
      ]. NlucP was knocked-in to the ID1 endogenous gene locus of SH-SY5Y cells using CRISPR/Cas9. Positive clones with NlucP-Hygr insertion at the ID1 gene locus were determined using PCR amplification of genomic DNA with primers, Fw and Rev1 (Fig. 4A). A homozygous knock-in clone was further selected by PCR amplification of genomic DNA with primers, Fw and Rev2, which produced no amplification. We further confirmed the relative copy numbers of NlucP relative to the endogenous ID1 gene in each cell line, indicating that the established cell lines included both homozygous (cell line # 26, ID1NlucP/NlucP) and heterozygous (cell line # 11 and 16, ID1NlucP/+) knock-in cell lines (Fig. 4B). Hereafter, cell lines #16 and #26 were used as heterozygous (ID1NlucP/+) and homozygous (ID1NlucP/NlucP) knock-in cell lines, respectively. ID1NlucP/+ cells treated with SAHA and CI-994 for 24 h, but not with IOX-1, showed a significant increase in luminescence (Fig. 4C). Similarly, ID1NlucP/NlucP cells showed an increase in luminescence via the treatment of SAHA and CI-994 for 24 h, but not with IOX-1 (Fig. 4D). Knockdown of HDAC2 for 4 days increased luminescence reporter activity in ID1NlucP/NlucP cells, while knockdown of HDAC1 did not (Fig. 4E and F). Knockdown of HDAC2 for 4 days did not lead to a robust increase in luminescence compared to that resulting from treatment with SAHA and CI-994, suggesting that the ID1 reporter may be appropriate for acute inhibition of HDAC2. To test the feasibility of high-throughput screening (HTS) using this reporter cell line, reporter cells were seeded on a 384-well plate at the density of 5,000 cells/well and treated with CI-994 at 10 µM for 24 h (Fig. 4G). We repeated this experiment four times and found that the reporter activity was increased at the fold change of 4.5 ± 0.10, with Z’ > 0.5 (0.78 ± 0.027), which is suitable for HTS. We further examined the potential background from dead cells. Reporter cells were similarly seeded to a 384-well plate and treated with CI-994 up to 30 µM for 24 h. Incubation with CI-994 at 30 uM resulted in 35% reduction in survival (Fig. 4H), whereas luminescence signals normalized to the cell number didn't show increase (Fig. 4I), suggesting that false-positive luminescence is not generated due to dead cells. Thus, NlucP knock-in cell lines to ID1 could be used as an intracellular reporter system to evaluate potential HDAC2 inhibitors.
      Fig 4
      Fig. 4Reporter cell lines for ID1 expression exclusively monitors activity of HDAC2. A, Schematic representation of knock-in of the reporter gene. Nanoluc fused to PEST sequence (NlucP) was knocked-in to the ID1 gene locus with a selection marker, hygromycin B resistance gene (Hygr). B, Copy number of NlucP in the reporter cell lines. The clones showing hygromycin B resistance (#11, 16 and 26) were subjected to genomic DNA purification. Relative copy number of NlucP was analyzed using qPCR and normalized by ID1. C and D, NlucP knock-in cell lines induced luminescence by HDAC inhibitors. The cell lines carrying heterozygous (C, NlucP/+) or homozygous knock-in (D, NlucP/NlucP) of NlucP were treated with SAHA, CI-994, IOX-1, or vehicle which contains same amount of DMSO as 10 μM of CI-994 for 24 h, and luminescence was measured. A one-way ANOVA indicated a significant difference across all panels (P < 0.0001; n = 2-4). E and F, NlucP knock-in cell lines induced luminescence by knockdown of HDAC2, but not HDAC1. The homozygous knock-in cell lines (NlucP/NlucP) were transfected by plasmids expressing miRNAs for HDAC2 (E), HDAC1 (F), or control sequence (miR-neg). After 4 days, luminescence was measured. A one-way ANOVA indicated a significant difference in all panels (P < 0.0001; n = 32). G-I, NlucP knock-in cell lines induced luminescence by CI-994. The homozygous knock-in cell lines (NlucP/NlucP) were treated with CI-994 or vehicle which contains same amount of DMSO (0.3%) for 24 h, and luminescence was measured (G and I). Cell number was estimated in each well after 24 h incubation. (G) t-test, P < 0.0001; n = 32, 16. (H and I) A one-way ANOVA indicated a significant difference across all panels (P < 0.0001, n = 128, 8, 8, 8, 8, 8, 8, 8, 8). Data are represented as mean ± s.e.m. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

      Discussion

      HDAC family proteins share highly conserved domains, which can be exemplified in the class I HDACs, including HDAC1 and HDAC2, which share 94% identity in their catalytic domains. Given that HDAC2 is specifically involved in memory functions, selective HDAC2 inhibition is required to maximize potential positive effects as inhibition of both HDAC1 and HDAC2 results in cellular toxicity. In this study, we performed a cell-based assay, which is sensitive to HDAC2 inhibition but not HDAC1 inhibition. This cellular reporter assay is thus affected by the activity of HDAC2 itself, and also by the regulatory mechanisms upstream of HDAC2, suggesting that our assay system could target multiple mechanisms related to HDAC2 function. Thus, our cellular reporter system offers an entry point for screening of HDAC2-specific inhibitors that would be functional in vivo.
      Upon drug screening using our assay system, there will be two types of unanticipated compound to be discovered, rather than HDAC2-specific inhibitors. The first type will be a non-specific HDAC inhibitor that could also inhibit HDAC1. If a hit compound does not affect cell proliferation, it is likely that the compound does not strongly inhibit both HDAC2 and HDAC1, considering the toxicity of both HDAC2- and HDAC1-knockdown. However, a weaker effect may be exerted on HDAC1. To ensure no potential effect on HDAC1, an HDAC1-specific reporter system could be additionally incorporated to our reporter cell lines. The second type will be inhibitors for transcriptional regulation of ID1 rather than HDAC2. To test whether obtained inhibitors target HDAC2, histone acetylation which largely dependent on HDAC2, such as H4K12 acetylation [
      • Guan J.
      • Haggarty S.J.
      • Giacometti E.
      • et al.
      HDAC2 negatively regulates memory formation and synaptic plasticity.
      ], at the ID1 gene locus should be investigated to support the specificity of the hit compounds on the HDAC2 pathway. Our cellular assay provides a novel method for screening HDAC2-specific inhibitors, though it should be further verified in additional studies according to the aforementioned criteria. The future high-throughput screening using our reporter SH-SY5Y cell line could provide candidate compounds which potentially target neuronal HDAC2 functions. Whether the obtained compounds functionally intervene the memory functions has to be verified using in vivo model, which will be the secondary screening.
      We utilized the neuronal SH-SY5Y cell line to identify ID1 as one of the HDAC2-downstream genes, which expression is correlated to cognitive function in mice [
      • Lee J.
      • Ki I.H.
      • Cho J.H.
      • et al.
      Vanillin improves scopolamine-induced memory impairment through restoration of ID1 expression in the mouse hippocampus.
      ]. Among four top-listed genes in our RNA-seq analysis, ID1 only showed acute and specific increase in expression by HDAC inhibitor, suggesting that ID1 could be a direct target of HDAC2. Although the cell type is different, ChIP-seq analysis using microglia in mice indicated the direct binding of HDAC2 to the promoter of ID1 [
      • Guo X.
      • Chen D.
      • An S.
      • et al.
      ChIP-seq profiling identifies histone deacetylase 2 targeting genes involved in immune and inflammatory regulation induced by calcitonin gene-related peptide in microglial cells.
      ]. Trichostatin A, a pan-HDAC inhibitor, induces ID1 expression also in the acute myeloid leukemia (AML) cell lines [
      • Yu W.P.
      • Scott S.A.
      • Dong W.F.
      Induction of ID1 expression and apoptosis by the histone deacetylase inhibitor (trichostatin A) in human acute myeloid leukaemic cells.
      ]. Given that HDAC2 acts as a general transcriptional repressor, ID1-reporter SH-SY5Y cell line could be widely applied to screen HDAC2 activity itself. However, if one would target the cell type-specific regulation of HDAC2, but not HDAC2 activity itself, the appropriate cell lines with the reporter gene knock-in have to be established. This cell-type specific strategy will be significant to restrict the action of the compounds to the specific cell types and reduce the risk of secondary effect on cells or tissues other than the targeted ones.
      In vitro drug screening was deemed adequate if the target protein did not share a conserved sequence or structure with other proteins. In the case of the epigenetic regulators, each family protein shows highly conserved domains, including catalytic domains, which hampers the development of specific inhibitors. The specific inhibition is critical to reduce the chance of unanticipated side effects. Here, we successfully established a reporter cell line that discriminates HDAC1 and HDAC2 functions. We believe that this work provides a versatile reporter system for screening HDAC2 inhibitors, and demonstrates a future direction of cell-based assay to distinguish closely related molecular targets. Finally, our results may facilitate the discovery of inhibitors with high specificity and fewer unanticipated side effects.

      Funding

      This study was supported by Shionogi & Co., Ltd. (to Y.H.).

      Declaration of Conflicting Interests

      The authors declare no competing interests. Four authors (K.U., K.H., N.H. and T.O.) are employed by Shionogi & Co., Ltd.

      Acknowledgments

      We thank the members of SK project for helpful discussions.

      Appendix. Supplementary materials

      References

        • Mews P.
        • Donahue G.
        • Drake A.M.
        • et al.
        Acetyl-CoA synthetase regulates histone acetylation and hippocampal memory.
        Nature. 2017; 546: 381-386
        • Nott A.
        • Watson P.M.
        • Robinson J.D.
        • et al.
        S-nitrosylation of histone deacetylase 2 induces chromatin remodelling in neurons.
        Nature. 2008; 455: 411-415
        • Pao P.
        • Tsai L.
        Histone deacetylases 1 and 2 in memory function.
        ACS Chem Neurosci. 2022; 13: 848-858
        • Levenson J.M.
        • O'Riordan K.J.
        • Brown K.D.
        • et al.
        Regulation of histone acetylation during memory formation in the hippocampus.
        J Biol Chem. 2004; 279: 40545-40559
        • David S.J.
        Epigenetics and cognitive aging.
        Science. 2010; 328: 701-702
        • Peleg S.
        • Sananbenesi F.
        • Zovoilis A.
        • et al.
        Altered histone acetylation is associated with age-dependent memory impairment in mice.
        Science. 2010; 328: 753-756
        • Guan J.
        • Haggarty S.J.
        • Giacometti E.
        • et al.
        HDAC2 negatively regulates memory formation and synaptic plasticity.
        Nature. 2009; 459: 55-60
        • Hanson J.E.
        • Deng L.
        • Hackos D.H.
        • et al.
        Histone deacetylase 2 cell autonomously suppresses excitatory and enhances inhibitory synaptic function in CA1 pyramidal neurons.
        J Neurosci. 2013; 33: 5924-5929
        • Eckschlager T.
        • Plch J.
        • Stiborova M.
        • et al.
        Histone deacetylase inhibitors as anticancer drugs.
        Int J Mol Sci. 2017; 18: 1-25
        • Wagner F.F.
        • Zhang Y.
        • Fass D.M.
        • et al.
        Kinetically selective inhibitors of histone deacetylase 2 (HDAC2) as cognition enhancers.
        Chem Sci. 2015; 6: 804-815
        • Zhao W.
        • Ghosh B.
        • Tyler M.
        • et al.
        Class I histone deacetylase inhibition by tianeptinaline modulates neuroplasticity and enhances memory.
        ACS Chem Neurosci. 2018; 9: 2262-2273
        • Fuller N.O.
        • Pirone A.
        • Lynch B.A.
        • et al.
        CoREST complex-selective histone deacetylase inhibitors show prosynaptic effects and an improved safety profile to enable treatment of synaptopathies.
        ACS Chem Neurosci. 2019; 10: 1729-1743
        • Ganai S.A.
        • Abdullah E.
        • Rashid R.
        • et al.
        Combinatorial in silico strategy towards identifying potential hotspots during inhibition of structurally identical HDAC1 and HDAC2 enzymes for effective chemotherapy against neurological disorders.
        Front Mol Neurosci. 2017; 10: 1-17
        • Rodrigues D.A.
        • Pinheiro P.D.S.M.
        • Sagrillo F.S.
        • et al.
        Histone deacetylases as targets for the treatment of neurodegenerative disorders: Challenges and future opportunities.
        Med Res Rev. 2020; 40: 2177-2211
        • Benito E.
        • Urbanke H.;
        • Ramachandran B.
        • et al.
        HDAC inhibitor-dependent transcriptome and memory reinstatement in cognitive decline models.
        J Clin Invest. 2015; 125: 3572-3584
        • Burns A.M.
        • Farinelli-Scharly M.
        • Hugues-Ascery S.
        • et al.
        The HDAC inhibitor CI-994 acts as a molecular memory aid by facilitating synaptic and intracellular communication after learning.
        Proc Natl Acad Sci U S A. 2022; 119: 1-12
        • Nemunaitis J.J.
        • Orr D.
        • Eager R.
        • et al.
        Phase I study of oral CI-994 in combination with gemcitabine in treatment of patients with advanced cancer.
        Cancer J. 2003; 9: 58-66
        • Undevia S.D.
        • Kindler H.L.
        • Janisch L.
        • et al.
        A phase I study of the oral combination of CI-994, a putative histone deacetylase inhibitor, and capecitabine.
        Ann Oncol. 2004; 15: 1705-1711
        • Wilting R.H.
        • Yanover E.
        • Heideman M.R.
        • et al.
        Overlapping functions of Hdac1 and Hdac2 in cell cycle regulation and haematopoiesis.
        EMBO J. 2010; 29: 2586-2597
        • Ler S.Y.
        • Leung C.H.W.
        • Khin L.W.
        • et al.
        HDAC1 and HDAC2 independently predict mortality in hepatocellular carcinoma by a competing risk regression model in a southeast asian population.
        Oncol Rep. 2015; 34: 2238-2250
        • Brunmeir R.
        • Lagger S.
        • Seiser C.
        Histone deacetylase 1 and 2-controlled embryonic development and cell differentiation.
        Int J Dev Biol. 2009; 53: 275-289
        • Ma P.
        • Schultz R.M.
        HDAC1 and HDAC2 in mouse oocytes and preimplantation embryos: Specificity versus compensation.
        Cell Death Differ. 2016; 23: 1119-1127
        • Ran F.A.
        • Hsu P.D.
        • Wright J.
        • et al.
        Genome engineering using the CRISPR-Cas9 system.
        Nat Protoc. 2013; 8: 2281-2308
        • Bolger A.M.
        • Lohse M.
        • Usadel B.
        Trimmomatic: A flexible trimmer for illumina sequence data.
        Bioinformatics. 2014; 30: 2114-2120
        • Dobin A.
        • Davis C.A.
        • Schlesinger F.
        • et al.
        STAR: Ultrafast universal RNA-seq aligner.
        Bioinformatics. 2013; 29: 15-21
        • Anders S.
        • Pyl P.T.
        • Huber W.
        HTSeq-A python framework to work with high-throughput sequencing data.
        Bioinformatics. 2015; 31: 166-169
        • Saldanha A.J.
        Java treeview - extensible visualization of microarray data.
        Bioinformatics. 2004; 20: 3246-3248
        • Liao Y.
        • Wang J.
        • Jaehnig E.J.
        • et al.
        WebGestalt 2019: Gene set analysis toolkit with revamped UIs and APIs.
        Nucleic Acids Res. 2019; 47: 199-205
        • Taylor-Whiteley T.R.
        • Le Maitre C.L.
        • Duce J.A.
        • et al.
        Recapitulating Parkinson's disease pathology in a three-dimensional human neural cell culture model.
        Dis Model Mech. 2019; 12: 1-10
        • Skotak M.
        • Wang F.
        • Chandra N.
        An in vitro injury model for SH-SY5Y neuroblastoma cells: Effect of strain and strain rate.
        J Neurosci Methods. 2012; 205: 159-168
        • Attoff K.
        • Johansson Y.
        • Cediel-Ulloa A.
        • et al.
        Acrylamide alters CREB and retinoic acid signalling pathways during differentiation of the human neuroblastoma SH-SY5Y cell line.
        Sci Rep. 2020; 10: 16714
        • Schikora J.
        • Kiwatrowski N.
        • Förster N.
        • et al.
        A propagated skeleton approach to high throughput screening of neurite outgrowth for in vitro parkinson's disease modelling.
        Cells. 2021; 10: 931
        • Schiller R.
        • Scozzafava G.
        • Tumber A.
        • et al.
        A cell-permeable ester derivative of the JmjC histone demethylase inhibitor IOX1.
        ChemMedChem. 2014; 9: 566-571
        • Berry S.L.
        • Hameed H.;
        • Thomason A.
        • et al.
        Development of NanoLuc-PEST expressing leishmania mexicana as a new drug discovery tool for axenic- and intramacrophage-based assays.
        PLoS Negl Trop Dis. 2018; 12: 1-20
        • Neefjes M.
        • Housmans B.A.C.
        • van den Akker G.G.H.
        • et al.
        Reporter gene comparison demonstrates interference of complex body fluids with secreted luciferase activity.
        Sci Rep. 2021; 11: 1-13
        • Lee J.
        • Ki I.H.
        • Cho J.H.
        • et al.
        Vanillin improves scopolamine-induced memory impairment through restoration of ID1 expression in the mouse hippocampus.
        Mol Med Rep. 2018; 17: 4399-4405
        • Guo X.
        • Chen D.
        • An S.
        • et al.
        ChIP-seq profiling identifies histone deacetylase 2 targeting genes involved in immune and inflammatory regulation induced by calcitonin gene-related peptide in microglial cells.
        J Immunol Res. 2020; 4384696
        • Yu W.P.
        • Scott S.A.
        • Dong W.F.
        Induction of ID1 expression and apoptosis by the histone deacetylase inhibitor (trichostatin A) in human acute myeloid leukaemic cells.
        Cell Prolif. 2008; 41: 86-97