Advertisement

Histone methyltransferase SETDB1 regulates the development of cortical Htr3a positive interneurons and mood behaviors

  • Author Footnotes
    † JL and SZ contributed equally to this work
    Jiaqi Li
    Footnotes
    † JL and SZ contributed equally to this work
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    † JL and SZ contributed equally to this work
    Shenghui Zheng
    Footnotes
    † JL and SZ contributed equally to this work
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Yuhao Dong
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Hao Xu
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Yueyan Zhu
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Jie Weng
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Daijing Sun
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Shunying Wang
    Affiliations
    Huashan Hospital, Fudan University, 200040, Shanghai, China
    Search for articles by this author
  • Lei Xiao
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Yan Jiang
    Correspondence
    Correspondence: Yan Jiang,
    Affiliations
    Institutes of Brain Science, State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, 200032, Shanghai, China
    Search for articles by this author
  • Author Footnotes
    † JL and SZ contributed equally to this work
Open AccessPublished:August 30, 2022DOI:https://doi.org/10.1016/j.biopsych.2022.08.021

      Abstract

      Background

      GABAergic interneurons (INs) are highly heterogeneous, and 5-hydroxytryptamine (serotonin) receptor 3A (Htr3a) labels a subpopulation of cortical INs originating from the embryonic caudal ganglionic eminence (GE). SETDB1 is one of the histone H3K9 methyltransferases and plays an essential role in the excitatory neurons, but its role in regulating cortical inhibitory INs remains largely unknown.

      Methods

      In the current study, we generated transgenic mice with conditional knockout of Setdb1 in neural progenitor cells (NPCs) (Setdb1-NS-cKO) and GABAergic neurons (Setdb1-Gad2-cKO). In addition, we performed RNA-seq, ATAC-seq, ChIP-seq, luciferase assay, Chromatin Conformation Capture (3C), and CRISPR/dCas9 to study the epigenetic mechanism underlying SETDB1-mediated transcriptional regulation of Htr3a. We also performed in situ hybridization and whole-cell recording to evaluate the functional properties of cortical Htr3a+ INs and behavioral tests for mood.

      Results

      We detected a significant upregulation of Htr3a expression in the embryonic GE of Setdb1-NS-cKO and identified the endogenous retroviral sequence RMER21B as a new target of SETDB1. RMER21B showed enhancer activity and formed distal chromatin interaction with the promoter of Htr3a. In addition, we observed an increased number and enhanced excitability of Htr3a+ INs in the knockout cortex. Moreover, Setdb1-Gad2-cKO mice exhibit anxiety and depressive-like behaviors, which were partially reversed by the type-3 serotonin receptor (5-HT3R) antagonist.

      Conclusions

      These findings suggest that SETDB1 represses Htr3a transcription via RMER21B-mediated distal chromatin interaction in the embryonic GE and regulates the development of cortical Htr3a+ INs and mood behaviors.

      Keywords

      Introduction

      GABAergic interneurons (INs) originate from ganglionic eminences (GE) and reside in the neocortex via tangential migration during brain development. They are multiple polarized neurons and critical for maintaining the balance between synaptic excitation and inhibition (E/I) (
      • Yizhar O.
      • Fenno L.E.
      • Prigge M.
      • Schneider F.
      • Davidson T.J.
      • O’shea D.J.
      • et al.
      Neocortical excitation/inhibition balance in information processing and social dysfunction.
      ). Dysregulation of cortical INs is a potential pathogenesis for neuropsychiatric disorders (
      • Marín O.
      Interneuron dysfunction in psychiatric disorders.
      ).
      INs account for 20-30% of total neurons in the adult cortex (
      • Ma T.
      • Wang C.
      • Wang L.
      • Zhou X.
      • Tian M.
      • Zhang Q.
      • et al.
      Subcortical origins of human and monkey neocortical interneurons.
      ), and Htr3a labels the majority of INs derived from caudal ganglionic eminence (CGE) (
      • Tao G.
      • Li Z.
      • Wen Y.
      • Song X.
      • Wei S.
      • Du H.
      • et al.
      Transcription factors Sp8 and Sp9 regulate medial ganglionic Eminence-Derived cortical interneuron migration.
      ,

      Limoni G (2018): Specification and integration of interneuron subtypes in the neocortex: University of Geneva.

      ). Htr3a encodes the A subunit of type-3 serotonin receptor (5-HT3R) (
      • Walstab J.
      • Rappold G.
      • Niesler B.
      5-HT(3) receptors: role in disease and target of drugs.
      ), which is widely expressed in the brain and plays an essential role in neuropsychological behaviors (
      • Walstab J.
      • Rappold G.
      • Niesler B.
      5-HT(3) receptors: role in disease and target of drugs.
      ). Htr3a knockout mice displayed improved psychological functions, including anti-anxiety and anti-depressive-like behaviors (
      • Bhatnagar S.
      • Sun L.M.
      • Raber J.
      • Maren S.
      • Julius D.
      • Dallman M.F.
      Changes in anxiety-related behaviors and hypothalamic-pituitary-adrenal activity in mice lacking the 5-HT-3A receptor.
      ,
      • Martin V.
      • Riffaud A.
      • Marday T.
      • Brouillard C.
      • Franc B.
      • Tassin J.P.
      • et al.
      Response of Htr3a knockout mice to antidepressant treatment and chronic stress.
      ). Genetic studies reported the association of HTR3A variants with affective disorders (
      • Gatt J.M.
      • Williams L.M.
      • Schofield P.R.
      • Dobson-Stone C.
      • Paul R.H.
      • Grieve S.M.
      • et al.
      Impact of the HTR3A gene with early life trauma on emotional brain networks and depressed mood.
      ,
      • Niesler B.
      • Weiss B.
      • Fischer C.
      • Nöthen M.M.
      • Propping P.
      • Bondy B.
      • et al.
      Serotonin receptor gene HTR3A variants in schizophrenic and bipolar affective patients.
      ,
      • Kim H.W.
      • Kang J.I.
      • Lee S.H.
      • An S.K.
      • Sohn S.Y.
      • Hwang E.H.
      • et al.
      Common variants of HTR3 genes are associated with obsessive-compulsive disorder and its phenotypic expression.
      ).
      SETDB1 (SET domain, bifurcated 1) is one of the histone methyltransferases mediating H3K9 trimethylation (
      • Karimi M.M.
      • Goyal P.
      • Maksakova I.A.
      • Bilenky M.
      • Leung D.
      • Tang J.X.
      • et al.
      DNA methylation and SETDB1/H3K9me3 regulate predominantly distinct sets of genes, retroelements, and chimeric transcripts in mESCs.
      ). It primarily functions as a transcriptional repressor (
      • Schultz D.C.
      • Ayyanathan K.
      • Negorev D.
      • Maul G.G.
      • Rauscher 3rd, F.J.
      SETDB1: a novel KAP-1-associated histone H3, lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins.
      ,
      • Minkovsky A.
      • Sahakyan A.
      • Rankin-Gee E.
      • Bonora G.
      • Patel S.
      • Plath K.
      The Mbd1-Atf7ip-Setdb1 pathway contributes to the maintenance of X chromosome inactivation.
      ,
      • Matsui T.
      • Leung D.
      • Miyashita H.
      • Maksakova I.A.
      • Miyachi H.
      • Kimura H.
      • et al.
      Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET.
      ), which acts not only directly on the gene promoters but also through long-range chromatin interactions (
      • Zhu Y.
      • Sun D.
      • Jakovcevski M.
      • Jiang Y.
      Epigenetic mechanism of SETDB1 in brain: implications for neuropsychiatric disorders.
      ,
      • Chandrasekaran S.
      • Espeso-Gil S.
      • Loh Y.E.
      • Javidfar B.
      • Kassim B.
      • Zhu Y.
      • et al.
      Neuron-specific chromosomal megadomain organization is adaptive to recent retrotransposon expansions.
      ). SETDB1 is an essential gene during embryonic development and is highly expressed in the fetal brain (
      • Zhu Y.
      • Sun D.
      • Jakovcevski M.
      • Jiang Y.
      Epigenetic mechanism of SETDB1 in brain: implications for neuropsychiatric disorders.
      ,
      • Tan S.L.
      • Nishi M.
      • Ohtsuka T.
      • Matsui T.
      • Takemoto K.
      • Kamio-Miura A.
      • et al.
      Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development.
      ). Genetic ablation of Setdb1 in the developing neocortex resulted in premature death with impaired neurogenesis and promoted astrogenesis (
      • Tan S.L.
      • Nishi M.
      • Ohtsuka T.
      • Matsui T.
      • Takemoto K.
      • Kamio-Miura A.
      • et al.
      Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development.
      ). Although SETDB1 is low in the adult brain, it plays a vital role in the excitatory neurons and participates in cognition and mood behaviors (
      • Jiang Y.
      • Jakovcevski M.
      • Bharadwaj R.
      • Connor C.
      • Schroeder F.A.
      • Lin C.L.
      • et al.
      Setdb1 histone methyltransferase regulates mood-related behaviors and expression of the NMDA receptor subunit NR2B.
      ,

      Markouli M, Strepkos D, Chlamydas S, Piperi C (2020): Histone lysine methyltransferase SETDB1 as a novel target for central nervous system diseases. Prog Neurobiol.101968.

      ). However, no studies have reported the function of SETDB1 in the GABAergic neuronal lineage. In the current study, we studied SETDB1 in cortical GABAergic neurons and revealed its important roles in regulating Htr3a transcription and the development of Htr3a+ INs, with implications for mood behaviors.

      Methods and Materials

      All mouse work was approved by the Animal Care and Use Committee of Shanghai Medical College, Fudan University. Mice were housed under 21-24°C, 45-55% humidity with a 12-hour day/night cycle. All mice were allowed sterile water and food ad libitum. Both genders were used in the following experiments. Setdb1 mice were gifts kindly provided by Dr. Akbarian group in Icahn School of Medicine at Mount Sinai, New York (
      • Jiang Y.
      • Loh Y.E.
      • Rajarajan P.
      • Hirayama T.
      • Liao W.
      • Kassim B.S.
      • et al.
      The methyltransferase SETDB1 regulates a large neuron-specific topological chromatin domain.
      ).
      NPC sphere culture was generated using mouse GE tissue according to a published protocol (
      • Guo W.
      • Patzlaff N.E.
      • Jobe E.M.
      • Zhao X.
      Isolation of multipotent neural stem or progenitor cells from both the dentate gyrus and subventricular zone of a single adult mouse.
      ).
      RNA extraction and RT-PCR were performed using standard protocols. The primers were listed in Table S1.
      RNA-seq, ATAC-seq, ChIP-seq and data analysis. See Supplementary Material.
      Luciferase assay was performed using the Dual-Glo luciferase system (Promega, E2920).
      Chromosome Conformation Capture (3C) was performed using a standard protocol with minor modifications (
      • Jiang Y.
      • Loh Y.E.
      • Rajarajan P.
      • Hirayama T.
      • Liao W.
      • Kassim B.S.
      • et al.
      The methyltransferase SETDB1 regulates a large neuron-specific topological chromatin domain.
      ).
      In situ hybridization was performed using a published protocol (
      • Li J.
      • Wang C.
      • Zhang Z.
      • Wen Y.
      • An L.
      • Liang Q.
      • et al.
      Transcription Factors Sp8 and Sp9 Coordinately Regulate Olfactory Bulb Interneuron Development.
      ).
      RNAScope was performed according to the manufacturer's instructions (Advanced Cell Diagnostics, 320851).
      Whole-cell recording on brain slices was adapted from published procedures (
      • Xiao L.
      • Priest M.F.
      • Nasenbeny J.
      • Lu T.
      • Kozorovitskiy Y.
      Biased oxytocinergic modulation of midbrain dopamine systems.
      ).

      Behavioral assays

      Open field, forced swim, tail suspension, elevated plus maze, and three-chamber social tests were performed using standard protocols. All behavioral assays were performed using 3- to 5-month-old gender- and age-matched littermates.
      A complete description of methods is included in the Supplementary Material.

      Results

      Upregulation of Htr3a expression in GE from Setdb1-NS-cKOs

      To study the function of SETDB1 during the development of cortical GABAergic INs, we dissected the GE from Setdb1-NS-cKO (KO) and wildtype (WT) mice at E15.5 (Fig. S1A). qRT-PCR and WB confirmed the loss of SETDB1 in cultured neural progenitors (GE_NPCs) and GE_tissue (Fig. S1B-C). We performed RNA-seq on GE_NPCs and GE_tissue. Principal component analysis (PCA) separated all the samples by genotype (Fig. S2A). Differential expression analysis identified a total of 854 and 441 significantly altered genes in GE_NPCs and GE_tissue, respectively, of which around 60% were upregulated in KO (Fig. 1A-B, Table S2). GO analysis revealed that the upregulated genes in GE_NPCs were enriched in the GABAergic synaptic transmission pathways (Fig. S2B), suggesting promoted neuronal differentiation after the loss of Setdb1. Interestingly, we noticed multiple upregulated genes encoding cell markers for CGE-derived cortical INs, including Htr3a, Calbindin 2 (Calb2), and Cholecystokinin (Cck). There was no significant change for MGE or LGE-derived cortical INs marker genes, for example, somatostatin (Sst) or forkhead box P1 (Foxp1) (Fig. 1C-D, Fig. S2C). Parvalbumin (Pv) is a well-studied marker gene for MGE-derived INs, but was not detectable in GE_NPCs. Moreover, RNA-seq on GE_tissue detected a significant increase of Htr3a in KO, but not Calb2, Cck, Sst, or Foxp1 (Fig. 1C-D, Fig. S2C). All the above-mentioned changes were validated by real-time RT-PCR using two reference genes for normalization (Fig. 1E, Fig. S2D).
      Figure thumbnail gr1
      Figure 1Increase of Htr3a transcription in GE from Setdb1-NS-cKOs. (A-B) MA plots and clustered heatmaps show RNA-seq signal of differential expressed genes from Setdb1-NS-cKO (KO) compared with wildtype (WT) in (A) GE_NPCs and (B) GE_tissue at E15.5. P < 0.05. Red dot, "Up", up-regulated genes; blue dot, "Down", down-regulated genes; gray dot, "N.S.", non-significant genes. GE_NPCs, n = 3 WT/ 3 KO; GE tissue, n = 2 WT/3 KO. (C) Differential expression (RNA-seq, KO/WT) of Htr3a, Calb2, Cck, Sst, Foxp1, and Setdb1. Notice significant increases of Htr3a in GE_NPCs and GE_tissue. (D) IGV map tracks show RNA-seq signal of Htr3a in KO and WT from GE_NPCs and GE_tissue. (E) Real-time RT-PCR. Mann-Whitney U test. Mean ± SEM. *P < 0.05, **P < 0.01. GE_NPCs, n = 4-5 WT/6-8 KO; GE_tissue, n = 4-5 WT/6 KO.

      Epigenetic alterations at Htr3a locus in GE_NPCs from Setdb1-NS-cKOs

      Increased chromatin accessibility is considered a predisposition for gene activation, so we performed ATAC-seq in GE_NPCs from Setdb1-NS-cKOs. PCA and hierarchical clustering analysis showed a clear separation between WT and KO (Fig. S3A-B). Differential analysis identified a total of 21,562 significantly altered ATAC-seq peaks, of which 69.2% were upregulated in KO (Fig. S3C, Table S2). Most differential peaks were annotated to intergenic regions and gene bodies, and only around 16% were located at the gene promoters (Fig. S3D-E). We checked the ATAC-seq signal at the promoters of differentially expressed genes (DEGs) from RNA-seq. Surprisingly, there was no noticeable difference at the upregulated gene promoters between WT and KO, and only a subtle decrease for the downregulated genes in KO (Fig. S3F). We then performed a permutation test for the distance between differential ATAC-seq peaks and DEGs. Interestingly, the upregulated ATAC-seq peaks were significantly closer to the upregulated genes, while the downregulated peaks were further away from the downregulated genes (Fig. S3G). In summary, chromatin accessibility is massively altered after the loss of Setdb1 in GE_NPCs, but it may indirectly contribute to the regulation of gene transcription.
      We next focused on the differential ATAC-seq peaks at Htr3a locus. Seven ATAC-seq peaks were detected within the 32 Kb window surrounding the transcription site (TSS) of Htr3a, and three were significantly upregulated in KO (Fig. 2A-B). Peak #4 was located at Htr3a_TSS, but was not altered in KO. Peak #3 resided in the intronic sequence, and peaks #6 and #7 were located around 12 Kb upstream of Htr3a_TSS (Fig. 2A-B). Interestingly, peak #7 was overlapped with a DNA repeat element RMER21B (Fig. 2A), which belongs to the LTR/ERV1 family. SETDB1/KAP1 complex is critical for ERV silencing, and de-repression of ERVs due to the loss of SETDB1 has been reported in many cell types, including neurons (
      • Chandrasekaran S.
      • Espeso-Gil S.
      • Loh Y.E.
      • Javidfar B.
      • Kassim B.
      • Zhu Y.
      • et al.
      Neuron-specific chromosomal megadomain organization is adaptive to recent retrotransposon expansions.
      ,
      • Tan S.L.
      • Nishi M.
      • Ohtsuka T.
      • Matsui T.
      • Takemoto K.
      • Kamio-Miura A.
      • et al.
      Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development.
      ). Once activated, some ERVs behave as enhancers to affect nearby genes (
      • Karimi M.M.
      • Goyal P.
      • Maksakova I.A.
      • Bilenky M.
      • Leung D.
      • Tang J.X.
      • et al.
      DNA methylation and SETDB1/H3K9me3 regulate predominantly distinct sets of genes, retroelements, and chimeric transcripts in mESCs.
      ,
      • Matsui T.
      • Leung D.
      • Miyashita H.
      • Maksakova I.A.
      • Miyachi H.
      • Kimura H.
      • et al.
      Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET.
      ,
      • Tan S.L.
      • Nishi M.
      • Ohtsuka T.
      • Matsui T.
      • Takemoto K.
      • Kamio-Miura A.
      • et al.
      Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development.
      ,
      • Adoue V.
      • Binet B.
      • Malbec A.
      • Fourquet J.
      • Romagnoli P.
      • van Meerwijk J.P.
      • et al.
      The histone methyltransferase SETDB1 controls T helper cell lineage integrity by repressing endogenous retroviruses.
      ). We, therefore, speculate that RMER21B participates in SETDB1-mediated transcriptional regulation of Htr3a in the embryonic GE_NPCs.
      Figure thumbnail gr2
      Figure 2Epigenetic alterations at Htr3a locus in GE_NPC from Setdb1-NS-cKOs. (A) IGV map tracks show RNA-seq, ATAC-seq and H3K27ac ChIP-seq signal at Htr3a locus in GE_NPC from Setdb1-NS-cKO (KO) and wildtype (WT) mice. Green and yellow bars highlight the signal at Htr3a gene promoter and RMER21B (red), respectively. Arrows and numbers (
      • Yizhar O.
      • Fenno L.E.
      • Prigge M.
      • Schneider F.
      • Davidson T.J.
      • O’shea D.J.
      • et al.
      Neocortical excitation/inhibition balance in information processing and social dysfunction.
      ,
      • Marín O.
      Interneuron dysfunction in psychiatric disorders.
      ,
      • Ma T.
      • Wang C.
      • Wang L.
      • Zhou X.
      • Tian M.
      • Zhang Q.
      • et al.
      Subcortical origins of human and monkey neocortical interneurons.
      ,
      • Tao G.
      • Li Z.
      • Wen Y.
      • Song X.
      • Wei S.
      • Du H.
      • et al.
      Transcription factors Sp8 and Sp9 regulate medial ganglionic Eminence-Derived cortical interneuron migration.
      ,

      Limoni G (2018): Specification and integration of interneuron subtypes in the neocortex: University of Geneva.

      ,
      • Walstab J.
      • Rappold G.
      • Niesler B.
      5-HT(3) receptors: role in disease and target of drugs.
      ,
      • Bhatnagar S.
      • Sun L.M.
      • Raber J.
      • Maren S.
      • Julius D.
      • Dallman M.F.
      Changes in anxiety-related behaviors and hypothalamic-pituitary-adrenal activity in mice lacking the 5-HT-3A receptor.
      ) indicate seven ATAC-seq peaks. "†", up-regulated ATAC-seq peaks; "$", up-regulated H3K27ac ChIP-seq peak. KO/WT, FDR < 0.05. LTR, long terminal repeat retrotransposons. Notice increased ATAC-seq and H3K27ac ChIP-seq signal at RMER21B. (B) Differential analysis for seven peaks labeled in panel A. (C-D) ChIP qPCRs show (C) H3K27ac and (D) H3K9me3 signal at RMER21B and Htr3a promoter in WT and KO. MuLV-int is used as a positive control. Hbb and L1Md_F2 are used as negative controls. N = 3-4 WT/4-5 KO. Mean ± SEM. Student's t-test. *P < 0.05, **P < 0.01.
      Histone modification H3K27ac is recognized as one of the enhancer marks, so we performed H3K27ac ChIP-seq in GE_NPCs. PCA separated all the samples by genotype (Fig. S4A). Differential analysis identified 83 significantly altered peaks, all of which were upregulated in KO (Fig. S4B, Table S2). These upregulated peaks were primarily annotated to intergenic regions (45.8%) and gene promoters (30.1%) (Fig. S4C). We checked the H3K27ac signal at the promoters of DEGs and observed a visible increase at the upregulated gene promoters in KO, and a slight decrease for the downregulated genes (Fig. S4D). In addition, permutation test showed that upregulated H3K27ac peaks were significantly closer to the upregulated genes (Fig. S4E). We further checked the H3K27ac signal at Htr3a locus and indeed, there was a strong peak at RMER21B, which was significantly increased in KO (Fig. 2A-B). ChIP_qPCR confirmed the increase of H3K27ac at RMER21B, but not Htr3a promoter (Fig. 2C).
      SETDB1 is an H3K9me3 histone methyltransferase, so we performed H3K9me3 ChIP-seq in GE_NPCs. As expected, there was a massive downregulation of H3K9me3 in KO (Fig. S5A), including many genomic loci for zinc finger genes (Fig. S5B), known targets of SETDB1/KAP1 complex (
      • O'Geen H.
      • Squazzo S.L.
      • Iyengar S.
      • Blahnik K.
      • Rinn J.L.
      • Chang H.Y.
      • et al.
      Genome-wide analysis of KAP1 binding suggests autoregulation of KRAB-ZNFs.
      ,
      • Frietze S.
      • O'Geen H.
      • Blahnik K.R.
      • Jin V.X.
      • Farnham P.J.
      ZNF274 recruits the histone methyltransferase SETDB1 to the 3' ends of ZNF genes.
      ). These differential loci were primarily annotated to the intergenic regions (Fig. S5C). Signal profiling showed a slight decrease of H3K9me3 on all the RMER21Bs in the mouse genome in KO (Fig. S5D). Moreover, ChIP_qPCR confirmed the reduction of H3K9me3 on the RMER21B at Htr3a locus, but not the promoter of Htr3a (Fig. 2D).
      RMER21B belongs to the ERV1 family and have 1,252 copies in the mouse genome. We analyzed the distance of all the RMER21Bs to their closest genes. Only 141 loci located on the gene promoters (< 3 Kb to TSS) (Fig. S6A). Seven RMER21Bs were overlapped with the upregulated ATAC-seq peaks, but there was no transcriptional activation of adjacent genes except Htr3a (Fig. S6B-C). Moreover, only the RMER21B in the current study showed upregulation of H3K27ac in KO (Fig. S6B), highlighting the uniqueness of the RMER21B at Htr3a locus. Together, epigenetic analyses suggest that RMER21B functions as an enhancer at Htr3a locus, and SETDB1 inhibits its activity in GE_NPCs.

      Transcriptional activation activity of RMER21B in vitro

      We continued to evaluate the transcriptional regulatory activity of RMER21B using a luciferase reporter system. We analyzed the sequence of the RMER21B at Htr3a locus (chr9: 48,922,965-48,923,351, mm10) and found it can be divided into two fragments (F1, F2), in which F2 consists of five highly homologous repetitive elements (r1-r5) (Fig. 3A, Fig. S7B). Moreover, Homer motif search predicated multiple transcription factors (TFs) binding sites in F2 (Fig. 3A, Fig. S7A). We first cloned the RMER21B, F1, and F2 into the luciferase reporter vector carrying minimal-promoter (pGL4.26) and assayed in mouse N2A and human 293T cell lines (Fig. 3B). In both cells, the luciferase signal was significantly higher in the RMER21B groups than that in the control groups. Interestingly, the luciferase signal in the F2 groups was similar to that in the RMER21B groups, but there was no difference between the F1 and control groups, suggesting the transcriptional activation activity of RMER21B was solely contributed by F2.
      Figure thumbnail gr3
      Figure 3Transcriptional activation activity of RMER21B in mouse and human cells. (A) Schematic illustration of the sequence features in RMER21B. F1, fragment 1; F2, fragment 2; r1-5, five repetitive elements in F2; ZNF692, Srebp2, E2F3, predicted transcription factor binding motifs. (B-E) Luciferase assay in N2a and 293T cells. (B) Transcriptional activity of RMER21B, F1, and F2. Notice RMER21B and F2 induce luciferase activity. Notice F1 has no effect. N = 3 per group. Student's t-test. *P < 0.05, **P < 0.01, #P < 0.001 (C) Transcriptional activity of r1-r5. N = 3 per group. Student's t-test. All groups were compared to the control group. *P < 0.05, **P < 0.01, #P < 0.001. Notice mild but significant activation effects of individual DNA repeat, except r5. (D) Transcriptional activity of different copies of DNA repeats: r1, r1+2, r1+2+3, r1+2+3+4 and r1+2+3+4+5 (F2). N = 4 per group. One-way ANOVA with Dunnett's multiple comparisons test. All groups were compared to the control group. #P-adjusted < 0.001. Notice the fragment with the first three DNA repeats ("r1+2+3") show the highest activation effects. (E) Transcriptional activity of RMER21B and RMER21BΔZNF in response to SETDB1. RMER21BΔZNF, RMER21B with mutations on ZNF692 motif. Setdb1, mouse full-length Setdb1 cDNA. N = 3 per group. Two-way ANOVA with Sidak multiple comparisons. *P-adjusted < 0.05. All data are expressed as Mean ± SEM.
      Next, we tested the five short repeats (r1-r5) individually and found each has a very mild transcriptional activation activity, except r5 (Fig. 3C). Among r1 to r4, r3 has the most substantial effect, but much smaller compare to F2 (Fig. 3C). We speculated that there might be a superimposed effect of short repeats for the activation activity of RMER21B, so we continued to assay for fragments containing different numbers of repeats, including r1, r1-2, r1-3, r1-4, and r1-5 (F2) (Fig. 3D). Indeed, the luciferase signal was dramatically induced by fragments containing ≥ two copies of repeats, and r1-3 displayed the highest activity in both N2a and 293T cells (Fig. 3D). In line with these results, r3 showed the highest conservation among the sequence alignment of all five repeats, while r5 was the lowest (Fig. S7B). Moreover, each of the r1, r2, and r3, but not r4 or r5, contains a binding motif for Srebp2 (Sterol regulatory element binding protein-2) (Fig. 3A, Fig. S7A), a potent activator for cholesterol synthesis (
      • Madison B.B.
      Srebp2: A master regulator of sterol and fatty acid synthesis.
      ).
      Zinc-finger proteins are key factors mediating the repressive activity of the SETDB1/KAP1 complex (
      • Sripathy S.P.
      • Stevens J.
      • Schultz D.C.
      The KAP1 corepressor functions to coordinate the assembly of de novo HP1-demarcated microenvironments of heterochromatin required for KRAB zinc finger protein-mediated transcriptional repression.
      ). Homer motif search predicated a binding motif for zinc-finger protein 692 (ZNF692) on the RMER21B (Fig. 3A, Fig S7A). We deleted the ZNF692 binding site (RMER21BΔZNF) and assayed for its activity with or without the co-transfection of the expressing plasmid containing mouse full-length Setdb1 cDNA (pCDH-Setdb1). In both N2a and 239T cells, the transcriptional activation activities of RMER21B and RMER21BΔZNF were inhibited by SETDB1 (Fig. 3E). However, two-way ANOVA with Sidak multiple comparisons tests revealed a "mutation" effect in N2a cells, as the activity of "RMER21BΔZNF" (P-adj < 0.05) or "RMER21BΔZNF+Setdb1" (P-adj = 0.09) was higher compared to its corresponding control group (Fig. 3E). Moreover, there was an interactive effect ("mutation" x "Setdb1") in 293T cells, as "RMER21BΔZNF" was significantly less affected by SETDB1 compared to "RMER21B" (P < 0.05, Two-way ANOVA with Sidak multiple comparisons) (Fig. 3E). These data suggested that ZNF69 may partially contribute to SETDB1-mediated repression on RMER21B.

      Chromatin interaction between RMER21B and Htr3a promoter

      It has been reported that de-repression of ERVs often affects the transcription of proximity genes by producing chimeric transcripts (
      • Karimi M.M.
      • Goyal P.
      • Maksakova I.A.
      • Bilenky M.
      • Leung D.
      • Tang J.X.
      • et al.
      DNA methylation and SETDB1/H3K9me3 regulate predominantly distinct sets of genes, retroelements, and chimeric transcripts in mESCs.
      ,
      • Matsui T.
      • Leung D.
      • Miyashita H.
      • Maksakova I.A.
      • Miyachi H.
      • Kimura H.
      • et al.
      Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET.
      ,
      • Tan S.L.
      • Nishi M.
      • Ohtsuka T.
      • Matsui T.
      • Takemoto K.
      • Kamio-Miura A.
      • et al.
      Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development.
      ,
      • Adoue V.
      • Binet B.
      • Malbec A.
      • Fourquet J.
      • Romagnoli P.
      • van Meerwijk J.P.
      • et al.
      The histone methyltransferase SETDB1 controls T helper cell lineage integrity by repressing endogenous retroviruses.
      ). In the current study, many ERVs were transcriptionally activated in the knockout GE_NPCs, most of which belong to the ERVK family (Fig. S8A-B, Table S2). Indeed, some of these ERV loci exhibited activation effects on nearby genes. For example, the transcription of MuLV-int (chr8:123,426,327-123,434,006, mm10) and its neighboring gene Tubb3 were significantly increased in KO (Fig. S8C). However, there was no transcriptional activation of any ERV element (including RMER21B) at Htr3a locus (Fig. S8D).
      Distal chromatin interaction between enhancer and promoter is an important epigenetic mechanism for regulating gene transcription (
      • Jiang Y.
      • Jakovcevski M.
      • Bharadwaj R.
      • Connor C.
      • Schroeder F.A.
      • Lin C.L.
      • et al.
      Setdb1 histone methyltransferase regulates mood-related behaviors and expression of the NMDA receptor subunit NR2B.
      ,
      • Schoenfelder S.
      • Fraser P.
      Long-range enhancer-promoter contacts in gene expression control.
      ). RMER21B is located around 12 Kb upstream of the TSS of Htr3a, and our data from ATAC-seq, H3K27ac ChIP-seq, and luciferase assay indicated its enhancer properties. Therefore, we performed Chromosome Conformation Capture (3C) on GE tissue to test whether there is a chromatin interaction between RMER21B and Htr3a promoter. As illustrated, four types of chimeras can be formed among the accessible ends of a chromatin loop after cut-and-ligation in the 3C library (Fig. S9A). Accordingly, we designed 12 sets of 3C primers and obtained expected PCR products from 2 different primer pairs (Fig. 4A, Fig. S9B). All 3C_PCR products were sequencing verified, confirming the chimeric junctions of the chromatin interaction between RMER21B and Htr3a promoter (Fig. 4B). Moreover, we compared the strength of chromatin interaction between KO and WT and only detected a trend of slight increase in KO (P = 0.1) (Fig. 4C). We speculate that the formation of RMER21B/Htr3a-promoter interaction is mediated by proteins other than SETDB1, and SETDB1 is recruited to repress the enhancer activity of RMER21B in the embryonic GE. To test this hypothesis, we constructed CRISPR/dCas9-KRAB system and introduced a repressive effector at RMER21B using sgRNAs targeting RMER21B ("sgRMER21B") in GE_NPCs and then checked Htr3a expression. As expected, qRT-PCR detected significant increases of Htr3a in KO compared to WT. Moreover, in the KO cells, there was a significant decrease in the "sgRMER21B" group compared to the "Scramble" control (Fig. 4D), indicating that RMER21B participates in SETDB1-mediated transcriptional repression of Htr3a.
      Figure thumbnail gr4
      Figure 4RMER21B-mediated distal-regulation of Htr3a transcription. (A-C) Chromatin Conformation Capture (3C) assay. (A) Schematic diagram shows two sets of 3C primers (P-1/E-1, P-2/E-2) for detecting chromatin interaction between Htr3a promoter (green) and RMER21B (orange). P, Htr3a promoter; E, enhancer (RMER21B). Arrowheads indicate the location and orientation of 3C primers. Short red lines indicate MboI sites. (B) Representative images show 3C-PCR products and sequencing data. PE1, PCR product using P-1/E-1; PE2, PCR product using P-2/E-2. Notice the chimeric 3C products containing a re-ligated MboI site flanked by DNA sequences from Htr3a promoter (green) and RMER21B element (orange). (C) (Left) Representative gel images and (Right) bar graph show 3C-PCR using GE_tissue from wildtype (WT) and Setdb1-NS-cKO (KO) mice. B2m for normalization. L, ligase; NL, no ligase. N = 3/group. Mean ± SEM. Mann-Whitney U test. P = 0.10 (D) (Left) Schematic diagram of CRISPR/dCas9-KRAB epigenomic editing system. (Right) Realtime RT-PCR shows Htr3a expression after the treatment of CRISPR/dCas9-KRAB in GE_NPCs from wildtype (WT) and Setdb1-NS-cKO (KO). "Scramble", scrambled sgRNA; sgRMER21B, sgRNAs targeting RMER21B. WT, n = 11 Scramble/8 sgRMER21B; KO, n = 6 Scramble/7 sgRMER21B. Mean ± SEM. Student's t-test, one-tailed. *P < 0.05. (E) Schematic diagram of RMER21B participating in SETDB1-mediated transcriptional repression of Htr3a via a distal chromatin loop interaction in mouse GE_NPC. (F) Distal chromatin loop interaction at human HTR3A locus (from "Double Elite" dataset). Bars, GeneHancer Regulatory Elements; Arc, chromatin interaction from Clustered interactions of GeneHancer Regulatory Elements and genes.
      Together, these data suggeste that RMER21B participates in SETDB1-mediated transcriptional regulation of Htr3a via distal chromatin interactions. The enhancer activity of RMER21B is repressed by SETDB1 to maintain a low level of Htr3a in the embryonic GE_NPCs. Conversely, the loss of SETDB1 releases the repressive control and leads to the upregulation of Htr3a in the knockout cells (Fig. 4E).
      RMER21Bs are rodent-specific DNA repeats, and the RMER21B at Htr3a locus is highly conserved among 16 mice strains and rat (Fig. S10). There is no homologous sequence in the human genome. However, a putative enhancer element is located at ∼20 Kb upstream of the human HTR3A_TSS, and Capture-C showed a chromatin loop connecting it with HTR3A promoter ("Double Elite" dataset) (Fig. 4F), suggesting functional conservation for the distal regulation of Htr3a/HTR3A transcription between mouse and human.

      SETDB1 regulates the development of cortical Htr3a+ INs

      Htr3a encodes the subunit A of 5-HT3R and is considered a marker gene for CGE-derived cortical INs (

      Vucurovic K, Gallopin T, Ferezou I, Rancillac A, Chameau P, van Hooft JA, et al. (2010): Serotonin 3A receptor subtype as an early and protracted marker of cortical interneuron subpopulations. Cerebral cortex (New York, NY : 1991). 20:2333-2347.

      ). It was reported that HTR3A participated in regulating the development of CGE-derived INs (
      • Murthy S.
      • Niquille M.
      • Hurni N.
      • Limoni G.
      • Frazer S.
      • Chameau P.
      • et al.
      Serotonin receptor 3A controls interneuron migration into the neocortex.
      ). We performed in situ hybridization to label Htr3a+ cells, and consistent with the literature (

      Vucurovic K, Gallopin T, Ferezou I, Rancillac A, Chameau P, van Hooft JA, et al. (2010): Serotonin 3A receptor subtype as an early and protracted marker of cortical interneuron subpopulations. Cerebral cortex (New York, NY : 1991). 20:2333-2347.

      ), these cells originate from CGE and migrate along the tangential migration stream towards the cortical plate and hippocampus during brain development (Fig. 5). We compared the in situ signal between KO and WT and found at E15.5, Htr3a signal was significantly increased at CGE in KO (Fig. 5A-B). At E18.5 (Fig. 5C-D) and P1 (Fig. 5E-F), the migration of Htr3a+ INs were promoted in KO, reflected by significant increases of positive cells in the cortical plate and hippocampus. We checked the expression of CALB2 (MGE cell marker) and FOXP1 (LGE cell marker) by immunofluorescence at E15.5 and found no difference between WT and KO (Fig. S11).
      Figure thumbnail gr5
      Figure 5Increase of Htr3a+ cells due to the loss of Setdb1 in the developing brain. (A, C, E) Representative images and (B, D, F) bar graphs show Htr3a in situ hybridization in brains from wildtype (WT) and Setdb1-NS-cKO (KO) mice at (A, B) E15.5, (C, D) E18.5, and (E, F) P1. CGE, caudal ganglionic eminences; Cp, cortical plate; Hip, hippocampus. Scale bar, 200 μm. E15.5, n = 12 WT/11KO; E18.5, n = 11 WT/13-14 KO; P1, n = 22 WT/22 KO. Mean ± SEM. Mann-Whitney U test. *P < 0.05, **P < 0.01, #P < 0.001.
      Setdb1-NS-cKO mice exhibited premature death around P10. We, therefore, generated Setdb1-Gad2-cKOs to study the effect of SETDB1 on cortical Htr3a+ INs in adulthood. We performed in situ hybridization in the adult brain and found Htr3a+ cells were primarily located in the superficial cortical layer L1/2, with a few in the deep layer L6 (Fig. 6). Consistent with data from the developing brain, there was a significant increase of Htr3a+ cells in KO, particularly in the prefrontal cortex (PFC) (Fig. 6), a critical brain region for emotion and cognitive functions. We arbitrarily divided PFC into three layers ("upper", "middle", and "deep") and found significant increases of Htr3a+ cells in all three layers in KO (Fig. 6A-E). In addition, we performed in situ hybridization for Gad1, Pv, and Sst, respectively, and there was no difference in the number of positive cells in cortex between KO and WT (Fig. S12), suggesting that other types of cortical INs were not altered due to the loss of Setdb1. We also introduced a Cre-dependent Ribo-GFP transgene into the Setdb1-Gad2-cKO mice to label all the GABAergic neurons (Setdb1-Gad2-cKO/Ribo-GFP). We then performed RNAScope for Htr3a and found all Htr3a+ cells were GFP+, confirming their GABAergic lineage (Fig. 6F). Furthermore, the distribution of Htr3a+ cells in the cortex was similar to what was observed with in situ hybridization. Double-label RNAScope showed no overlap between Htr3a+ and Pv+ or Sst+ cells in cortex (Fig. S13), indicating no aberrant expression of Htr3a in other types of INs after the loss of Setdb1.
      Figure thumbnail gr6
      Figure 6Increase of Htr3a+ cells in adult Setdb1-Gad2-cKO cortex. (A) (Left) Representative images show Htr3a in situ hybridization signal in adult cortex from wildtype (WT) and Setdb1-Gad2-cKO (KO) mice. Cx, cortex; PFC, prefrontal cortex; Str, Striatum; cc, corpus callosum; Ac, anterior commissure. PFC is highlighted in red lines. Scale bar, 1 mm. (Right) High-resolution images of PFC, which is arbitrarily divided into "Upper", "Middle", and "Deep" layers. Scale bar, 200 μm. (B-E) Bar graphs show Htr3a+ cell density in (B) total, (C) "Upper", (D) "Middle", and (E) "Deep" layers of PFC. N = 20 WT/18 KO. Mean ± SEM. Student's t-test. *P < 0.05, #P < 0.001. (F) Representative images show RNAscope signal of Htr3a in cortex from Setdb1-Gad2-cKO/Ribo-GFP mice. Red, Htr3a; Green, Gad2-GFP signal; Blue, DAPI. Arrows point to Htr3a+ cells. Scale bar, 100 μm. Notice all Htr3a+ (red) are Gad2-GFP+ (green), indicating Htr3a+ cells are GABAergic neurons.

      Increased excitability of Htr3a+ INs in the cortex of Setdb1-Gad2-cKOs

      5-HT3R is a type of inotropic receptor and triggers rapid depolarization after transient Na+ and K+ influx (
      • Maricq A.V.
      • Peterson A.S.
      • Brake A.J.
      • Myers R.M.
      • Julius D.
      Primary structure and functional expression of the 5HT3 receptor, a serotonin-gated ion channel.
      ). It has been reported that activation of Htr3a+ INs enhances the GABAergic neurotransmission in the neocortex (
      • Lee S.
      • Hjerling-Leffler J.
      • Zagha E.
      • Fishell G.
      • Rudy B.
      The largest group of superficial neocortical GABAergic interneurons expresses ionotropic serotonin receptors.
      ). To check whether increased Htr3a expression affected the electrophysiological properties of cortical Htr3a+ INs, we prepared brain sections from two-week-old Setdb1-Gad2-cKO/Ribo-GFP mice and performed whole-cell recording on GFP+ neurons in L1/2 of PFC (Fig. 7, Fig. S14). For all the GFP+ cells recorded, there was no significant difference in resting membrane potential or membrane resistance between KO and WT (Fig. 7A). We identified Htr3a+ cells by applying the 5-HT3R selective agonist mCPBG and revealed significant increases in membrane resistance and action potential firings in KO (Fig. 7B), indicating increased excitability of Htr3a+ INs due to the loss of Setdb1. INs are essential components of local circuits regulating the E/I balance in the cortex. We also recorded the excitatory pyramid neurons (PNs) (Fig. S15) and found no difference in membrane resistance between KO and WT (Fig. S15C). Surprisingly, there were slight but significant increases in resting potential and action potential firings in KO (Fig. S15D-F). Previous study on Htr3a knockout mice reported increased excitability of hippocampal Pv+ INs (
      • Huang L.
      • Wang J.
      • Liang G.
      • Gao Y.
      • Jin S.Y.
      • Hu J.
      • et al.
      Upregulated NMDAR-mediated GABAergic transmission underlies autistic-like deficits in Htr3a knockout mice.
      ), the major type of inhibitory neurons innervating PNs (

      Delevich K, Tucciarone J, Huang ZJ, Li B (2015): The mediodorsal thalamus drives feedforward inhibition in the anterior cingulate cortex via parvalbumin interneurons. The Journal of neuroscience : the official journal of the Society for Neuroscience. 35:5743-5753.

      ,
      • Udakis M.
      • Pedrosa V.
      • Chamberlain S.E.L.
      • Clopath C.
      • Mellor J.R.
      Interneuron-specific plasticity at parvalbumin and somatostatin inhibitory synapses onto CA1 pyramidal neurons shapes hippocampal output.
      ). We, therefore, speculate that Htr3a+ INs may innervate Pv+ INs and de-repress its inhibition on PNs, which may explain the increased excitability of cortical PNs in Setdb1-Gad2-cKOs.
      Figure thumbnail gr7
      Figure 7Altered electrophysiological properties of Htr3a+ neurons in adult Setdb1-Gad2-cKO cortex. Whole-cell recording on brain section from wildtype (WT) and Setdb1-Gad2-cKO/Ribo-GFP (KO) mice: (Left) resting potential, (Middle) membrane resistance, (Right) action potential firings induced by different level of current injection. GFP+ cells on cortical layer 1 and 2 are selected. (A) Total: GFP+ cells. N = 18 WT/15 KO. (B) Htr3a+: GFP+ cells response to 5-HT3 receptor agonist mCPBG. N = 4 WT/7 KO. Notice increased membrane resistance and action potential firing in Htr3a+ KO cells. Bar graphs, Mann-Whitney U test. Dot-line, mix-effects analysis with Bonferroni post-test. Mean ± SEM. *P-adjusted < 0.05, **P-adjusted < 0.01, #P-adjusted < 0.001.

      Anxiety and depressive-like behaviors in Setdb1-Gad2-cKO mice

      Serotonin is one of the monoamine neurotransmitters vital for regulating emotional behaviors (
      • Mohammad-Zadeh L.F.
      • Moses L.
      • Gwaltney-Brant S.M.
      Serotonin: a review.
      ). To study whether dysregulation of Htr3a+ INs led to emotion-related behavioral deficits, we applied a battery of behavioral paradigms to evaluate anxiety, depressive-like, and social behaviors in Setdb1-Gad2-cKO mice. Gender effect is one of the most important confounding factors for emotional behaviors (
      • He T.
      • Guo C.
      • Wang C.
      • Hu C.
      • Chen H.
      Effect of early life stress on anxiety and depressive behaviors in adolescent mice.
      ,
      • Huang Q.
      • Zhou Y.
      • Liu L.Y.
      Effect of post-weaning isolation on anxiety- and depressive-like behaviors of C57BL/6J mice.
      ); therefore, we performed behavioral tests on both genders separately. We observed increased body weight in Setdb1-Gad2-cKOs, but only the females reached statistical significance (Fig. 8A). Open field test showed no difference in the total travel distance between KO and WT; however, for both genders, the knockout mice showed significantly less time staying in the central zone and decreased number of center entries (Fig. 8B), suggesting an increased level of anxiety. Elevated plus-maze test showed no difference between WT and KO (Fig. 8C). Only female knockouts, but not males, showed increases in the total immobility time in forced swim (P < 0.05) and tail suspension (P = 0.0532) tests (Fig. 8D-E), indicating an increased level of behavior despair. In the three-chamber test, although both WT and KO demonstrated preference for an animal over an object, there was a significant "genotype" effect (P < 0.05, Two-way ANOVA) suggesting a mild social deficit in KO (Fig. 8F).
      Figure thumbnail gr8
      Figure 8Anxiety and depressive-like behaviors in Setdb1-Gad2-cKO mice. (A) Bodyweight. Female, n = 16 WT/13 KO. Male, n = 12 WT/10 KO. Student's t-test. (B) Open field: total distance, time in center, and the number of center entries. Female, n = 29 WT/25 KO. Male, n = 26 WT/27 KO. Student's t-test. (C) Elevated plus-maze: time in open and closed arms. Female, n = 20 WT/ 22 KO. Male, n = 19 WT/18 KO. Student's t-test. (D) Forced swim: latency to the first freezing (Mann-Whitney U test) and total immobility time (Student's t-test). Female, n = 14-16 WT/15-16 KO. Male, n = 10-11 WT/11-12 KO. (E) Tail suspension: latency to the first freezing (Mann-Whitney U test) and total immobility time (Student's t-test). Female, n = 18-21 WT/ 20-21 KO. Male, n = 19-20 WT/20 KO. (F) Three-chamber social test: time spent in exploration zones with animal and object. Female, n = 20 WT/17 KO. Male, n = 19 WT/18 KO. Two-way ANOVA with Bonferroni post-test. **P-adjusted < 0.01, #P-adjusted < 0.001. (G) Open field after 5-HT3R selective antagonist granisetron: time in center and the number of center entries. Female, n = 16 WT/12 KO. Male, n = 12 WT/10 KO. Student's t-test. (H) Forced swim after granisetron: latency to the first freezing (Mann-Whitney U test) and total immobility time (Student's t-test). Female, n = 14-16 WT/12-14 KO. Male, n = 12 WT/10-12 KO. All data expressed as Mean ± SEM. *P < 0.05, **P < 0.01, #P < 0.001.
      To test whether HTR3A directly contributes to the observed mood deficits, we applied seven days of 5-HT3R selective antagonist, once per day, i.p., in wildtype and Setdb1-Gad2-cKO mice and performed open filed and forced swim tests 24 hours after the last dose. We chose granisetron because it can cross the blood-brain barrier and is well tolerated in clinical trials (
      • Huang C.T.
      • Chen K.C.
      • Chen C.F.
      • Tsai T.H.
      Simultaneous measurement of blood and brain microdialysates of granisetron in rat by high-performance liquid chromatography with fluorescence detection.
      ,
      • Saito M.
      • Aogi K.
      • Sekine I.
      • Yoshizawa H.
      • Yanagita Y.
      • Sakai H.
      • et al.
      Palonosetron plus dexamethasone versus granisetron plus dexamethasone for prevention of nausea and vomiting during chemotherapy: a double-blind, double-dummy, randomised, comparative phase III trial.
      ). After the treatment of granisetron, there was no significant difference (time in center, center entries) between WT and KO in the open field test (Fig. 8G), suggesting the blockage of 5-HT3R ameliorated anxiety deficits in Setdb1-Gad2-cKOs. However, no rescue effect was observed in the forced swim test, as decreased latency and increased total immobility time were still observed in the female Setdb1-Gad2-cKOs after granisetron (Fig. 8H).

      Discussion

      In the current study, we reported an important role of SETDB1 in regulating cortical Htr3a+ INs. Epigenomic profiling, luciferase assay, and 3C showed that SETDB1 repressed the enhancer activity of RMER21B, which formed a distal chromatin interaction with Htr3a promoter. In situ hybridization showed increased number and promoted migration of Htr3a+ cells after the loss of Setdb1. Moreover, whole-cell recording detected increased excitability of Htr3a+ INs in the adult knockout cortex. Finally, multiple behavioral tests indicated anxiety and depressive-like phenotype of Setdb1-GAD2-cKOs, partially reversed by the 5-HT3Rs selective antagonist.
      It has been reported that Htr3a+ cells account for 20-30% of total INs in the adult cortex (
      • Tremblay R.
      • Lee S.
      • Rudy B.
      GABAergic Interneurons in the Neocortex: From Cellular Properties to Circuits.
      ). Previous studies using Htr3a-GFP BAC-transgenic mice showed that Htr3a-GFP+ cells were present in all the cortical layers, with preference for superficial layers (
      • Wei S.
      • Du H.
      • Li Z.
      • Tao G.
      • Xu Z.
      • Song X.
      • et al.
      Transcription factors Sp8 and Sp9 regulate the development of caudal ganglionic eminence-derived cortical interneurons.
      ,
      • Chittajallu R.
      • Craig M.T.
      • McFarland A.
      • Yuan X.
      • Gerfen S.
      • Tricoire L.
      • et al.
      Dual origins of functionally distinct O-LM interneurons revealed by differential 5-HT(3A)R expression.
      ,
      • Niquille M.
      • Limoni G.
      • Markopoulos F.
      • Cadilhac C.
      • Prados J.
      • Holtmaat A.
      • et al.
      Neurogliaform cortical interneurons derive from cells in the preoptic area.
      ). However, in the current study, Htr3a+ cells in the cortex were very sparse and located almost exclusively in L1/2 and L6, which were confirmed by using multiple in situ hybridization probes and RNAScope. We speculate that the BAC-Htr3a-GFP transgene is expressed with less constraint compared to the endogenous Htr3a. Single-cell RNA-seq (scRNA-seq) is a powerful method to check the detailed cell compositions from tissue homogenates, therefore, it might serve as an alternative method to check the changes of Htr3a+ cells upon Setdb1 deletion. However, we checked the published scRNA-seq datasets (GSE188528) and found that Htr3a, as well as Calb2, Cck and Sst, were only detected in very few cells at CGE. Plus, there were also Htr3a+ cells detected in MGE and LGE (Fig. S16). This might be due to the limited sequencing depth of scRNA-seq and the lack of clear anatomical boundaries for subregions inside GE. In the future study, transgenic mouse with higher-specificity for Htr3a+ neurons will be a useful tool for studying this unique population of cortical INs.
      5-HT3 receptors participate in a wide range of neurophysiological functions (
      • Yohn C.N.
      • Gergues M.M.
      • Samuels B.A.
      The role of 5-HT receptors in depression.
      ,
      • Bhatt S.
      • Devadoss T.
      • Manjula S.N.
      • Rajangam J.
      5-HT3 receptor antagonism a potential therapeutic approach for the treatment of depression and other disorders.
      ). Previous studies reported the association between HTR3A variants with bipolar disorder (
      • Hammer C.
      • Cichon S.
      • Mühleisen T.W.
      • Haenisch B.
      • Degenhardt F.
      • Mattheisen M.
      • et al.
      Replication of functional serotonin receptor type 3A and B variants in bipolar affective disorder: a European multicenter study.
      ,
      • Niesler B.
      • Flohr T.
      • Nöthen M.
      • Fischer C.
      • Rietschel M.
      • Franzek E.
      • et al.
      Association between the 5' UTR variant C178T of the serotonin receptor gene HTR3A and bipolar affective disorder.
      ). Interestingly, the affected SNP (C178T, c. -42C>T) is located on the 5' UTR of HTR3A, a noncoding sequence that commonly affected by epigenetic modifications. Indeed, both clinical and preclinical studies have reported the association of DNA methylation and histone acetylation at the promoter region of HTR3A/Htr3a with adverse environmental exposure and neuropathological behaviors (

      Hjort L, Rushiti F, Wang SJ, Fransquet P, S PK, S IC, et al. (2021): Intergenerational effects of maternal post-traumatic stress disorder on offspring epigenetic patterns and cortisol levels. Epigenomics. 13:967-980.

      ,
      • Jahn K.
      • Kurz B.
      • Sinke C.
      • Kneer J.
      • Riemer O.
      • Ponseti J.
      • et al.
      Serotonin system-associated genetic and epigenetic changes in pedophilia and child sexual offending.
      ,
      • Schechter D.S.
      • Moser D.A.
      • Pointet V.C.
      • Aue T.
      • Stenz L.
      • Paoloni-Giacobino A.
      • et al.
      The association of serotonin receptor 3A methylation with maternal violence exposure, neural activity, and child aggression.
      ,
      • Barker J.M.
      • Zhang H.
      • Villafane J.J.
      • Wang T.L.
      • Torregrossa M.M.
      • Taylor J.R.
      Epigenetic and pharmacological regulation of 5HT3 receptors controls compulsive ethanol seeking in mice.
      ,
      • Xu Y.
      • Liu X.
      • Zhang X.
      • Zhang G.
      • Zhang R.
      • Liu T.
      • et al.
      Histone acetylation of the htr3a gene in the prefrontal cortex of Wistar rats regulates ethanol-seeking behavior.
      ,
      • Zhang H.
      • Herman A.I.
      • Kranzler H.R.
      • Anton R.F.
      • Zhao H.
      • Zheng W.
      • et al.
      Array-based profiling of DNA methylation changes associated with alcohol dependence.
      ). Therefore, HTR3A variants may disrupt the epigenetic modifications in response to adverse environmental insults, thus contribute to the dysfunction of HTR3A and its related behaviors. In the current study, SETDB1-mediated epigenetic changes happen on the enhancer element RMER21B, which formed a distal chromatin interaction with Htr3a promoter. Although there is no homolog sequence of RMER21B in human, loop formation is also detected at HTR3A locus. Future studies on genomic variations of distal regulatory elements will offer new insight into the etiology of affective disorders.
      Blockage of 5-HT3 receptor activity is considered a potential therapeutic approach in treating anxiety and depression (
      • Bhatt S.
      • Devadoss T.
      • Manjula S.N.
      • Rajangam J.
      5-HT3 receptor antagonism a potential therapeutic approach for the treatment of depression and other disorders.
      ). MC4R neurons in the paraventricular nucleus express Htr3a and exhibit hyperexcitability in a high-fat-diet induced anxiety and depression model, which can be effectively rescued by the infusion of granisetron (
      • Xia G.
      • Han Y.
      • Meng F.
      • He Y.
      • Srisai D.
      • Farias M.
      • et al.
      Reciprocal control of obesity and anxiety-depressive disorder via a GABA and serotonin neural circuit.
      ). However, there is no study to date reporting the activity of cortical Htr3a+ INs in the model of affective disorders. Future study on the causal effects of Htr3a+ INs in mood behaviors will provide more evidence supporting 5-HT3A receptor as a potential therapeutic target for affective disorders.

      Data availability

      The raw and processed deep sequencing data were submitted to GEO (GSE186806, GSE200726). Secure tokens are available for review: ylczyegwjdszjov (RNA-seq, ATAC-seq), czgxwuiqtvaljex (ChIP-seq)

      Acknowledgment

      This work was supported by the National Natural Science Foundation of China (No.32170601, No. 81971272); the National Key R&D Program of China (2021ZD0203001); the Science and Technology Commission of Shanghai Municipality (No. 19ZR1405400); the Shanghai Municipal Science and Technology Major Project (No. 2018SHZDZX01), ZJ Lab, and the Shanghai Center for Brain Science and Brain-Inspired Technology.

      Supplementary Material

      Reference

        • Yizhar O.
        • Fenno L.E.
        • Prigge M.
        • Schneider F.
        • Davidson T.J.
        • O’shea D.J.
        • et al.
        Neocortical excitation/inhibition balance in information processing and social dysfunction.
        Nature. 2011; 477: 171-178
        • Marín O.
        Interneuron dysfunction in psychiatric disorders.
        Nat Rev Neurosci. 2012; 13: 107-120
        • Ma T.
        • Wang C.
        • Wang L.
        • Zhou X.
        • Tian M.
        • Zhang Q.
        • et al.
        Subcortical origins of human and monkey neocortical interneurons.
        Nat Neurosci. 2013; 16: 1588-1597
        • Tao G.
        • Li Z.
        • Wen Y.
        • Song X.
        • Wei S.
        • Du H.
        • et al.
        Transcription factors Sp8 and Sp9 regulate medial ganglionic Eminence-Derived cortical interneuron migration.
        Frontiers in molecular neuroscience. 2019; 12: 75
      1. Limoni G (2018): Specification and integration of interneuron subtypes in the neocortex: University of Geneva.

        • Walstab J.
        • Rappold G.
        • Niesler B.
        5-HT(3) receptors: role in disease and target of drugs.
        Pharmacol Ther. 2010; 128: 146-169
        • Bhatnagar S.
        • Sun L.M.
        • Raber J.
        • Maren S.
        • Julius D.
        • Dallman M.F.
        Changes in anxiety-related behaviors and hypothalamic-pituitary-adrenal activity in mice lacking the 5-HT-3A receptor.
        Physiol Behav. 2004; 81: 545-555
        • Martin V.
        • Riffaud A.
        • Marday T.
        • Brouillard C.
        • Franc B.
        • Tassin J.P.
        • et al.
        Response of Htr3a knockout mice to antidepressant treatment and chronic stress.
        British journal of pharmacology. 2017; 174: 2471-2483
        • Gatt J.M.
        • Williams L.M.
        • Schofield P.R.
        • Dobson-Stone C.
        • Paul R.H.
        • Grieve S.M.
        • et al.
        Impact of the HTR3A gene with early life trauma on emotional brain networks and depressed mood.
        Depression and anxiety. 2010; 27: 752-759
        • Niesler B.
        • Weiss B.
        • Fischer C.
        • Nöthen M.M.
        • Propping P.
        • Bondy B.
        • et al.
        Serotonin receptor gene HTR3A variants in schizophrenic and bipolar affective patients.
        Pharmacogenetics. 2001; 11: 21-27
        • Kim H.W.
        • Kang J.I.
        • Lee S.H.
        • An S.K.
        • Sohn S.Y.
        • Hwang E.H.
        • et al.
        Common variants of HTR3 genes are associated with obsessive-compulsive disorder and its phenotypic expression.
        Scientific reports. 2016; 632564
        • Karimi M.M.
        • Goyal P.
        • Maksakova I.A.
        • Bilenky M.
        • Leung D.
        • Tang J.X.
        • et al.
        DNA methylation and SETDB1/H3K9me3 regulate predominantly distinct sets of genes, retroelements, and chimeric transcripts in mESCs.
        Cell stem cell. 2011; 8: 676-687
        • Schultz D.C.
        • Ayyanathan K.
        • Negorev D.
        • Maul G.G.
        • Rauscher 3rd, F.J.
        SETDB1: a novel KAP-1-associated histone H3, lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins.
        Genes Dev. 2002; 16: 919-932
        • Minkovsky A.
        • Sahakyan A.
        • Rankin-Gee E.
        • Bonora G.
        • Patel S.
        • Plath K.
        The Mbd1-Atf7ip-Setdb1 pathway contributes to the maintenance of X chromosome inactivation.
        Epigenetics & chromatin. 2014; 7: 1-15
        • Matsui T.
        • Leung D.
        • Miyashita H.
        • Maksakova I.A.
        • Miyachi H.
        • Kimura H.
        • et al.
        Proviral silencing in embryonic stem cells requires the histone methyltransferase ESET.
        Nature. 2010; 464: 927-931
        • Zhu Y.
        • Sun D.
        • Jakovcevski M.
        • Jiang Y.
        Epigenetic mechanism of SETDB1 in brain: implications for neuropsychiatric disorders.
        Transl Psychiatry. 2020; 10: 115
        • Chandrasekaran S.
        • Espeso-Gil S.
        • Loh Y.E.
        • Javidfar B.
        • Kassim B.
        • Zhu Y.
        • et al.
        Neuron-specific chromosomal megadomain organization is adaptive to recent retrotransposon expansions.
        Nat Commun. 2021; 12: 7243
        • Tan S.L.
        • Nishi M.
        • Ohtsuka T.
        • Matsui T.
        • Takemoto K.
        • Kamio-Miura A.
        • et al.
        Essential roles of the histone methyltransferase ESET in the epigenetic control of neural progenitor cells during development.
        Development. 2012; 139: 3806-3816
        • Jiang Y.
        • Jakovcevski M.
        • Bharadwaj R.
        • Connor C.
        • Schroeder F.A.
        • Lin C.L.
        • et al.
        Setdb1 histone methyltransferase regulates mood-related behaviors and expression of the NMDA receptor subunit NR2B.
        J Neurosci. 2010; 30: 7152-7167
      2. Markouli M, Strepkos D, Chlamydas S, Piperi C (2020): Histone lysine methyltransferase SETDB1 as a novel target for central nervous system diseases. Prog Neurobiol.101968.

        • Jiang Y.
        • Loh Y.E.
        • Rajarajan P.
        • Hirayama T.
        • Liao W.
        • Kassim B.S.
        • et al.
        The methyltransferase SETDB1 regulates a large neuron-specific topological chromatin domain.
        Nat Genet. 2017; 49: 1239-1250
        • Guo W.
        • Patzlaff N.E.
        • Jobe E.M.
        • Zhao X.
        Isolation of multipotent neural stem or progenitor cells from both the dentate gyrus and subventricular zone of a single adult mouse.
        Nature protocols. 2012; 7: 2005-2012
        • Li J.
        • Wang C.
        • Zhang Z.
        • Wen Y.
        • An L.
        • Liang Q.
        • et al.
        Transcription Factors Sp8 and Sp9 Coordinately Regulate Olfactory Bulb Interneuron Development.
        Cereb Cortex. 2018; 28: 3278-3294
        • Xiao L.
        • Priest M.F.
        • Nasenbeny J.
        • Lu T.
        • Kozorovitskiy Y.
        Biased oxytocinergic modulation of midbrain dopamine systems.
        Neuron. 2017; 95 (e365): 368-384
        • Adoue V.
        • Binet B.
        • Malbec A.
        • Fourquet J.
        • Romagnoli P.
        • van Meerwijk J.P.
        • et al.
        The histone methyltransferase SETDB1 controls T helper cell lineage integrity by repressing endogenous retroviruses.
        Immunity. 2019; 50 (e628): 629-644
        • O'Geen H.
        • Squazzo S.L.
        • Iyengar S.
        • Blahnik K.
        • Rinn J.L.
        • Chang H.Y.
        • et al.
        Genome-wide analysis of KAP1 binding suggests autoregulation of KRAB-ZNFs.
        PLoS Genet. 2007; 3: e89
        • Frietze S.
        • O'Geen H.
        • Blahnik K.R.
        • Jin V.X.
        • Farnham P.J.
        ZNF274 recruits the histone methyltransferase SETDB1 to the 3' ends of ZNF genes.
        PLoS One. 2010; 5e15082
        • Madison B.B.
        Srebp2: A master regulator of sterol and fatty acid synthesis.
        J Lipid Res. 2016; 57: 333-335
        • Sripathy S.P.
        • Stevens J.
        • Schultz D.C.
        The KAP1 corepressor functions to coordinate the assembly of de novo HP1-demarcated microenvironments of heterochromatin required for KRAB zinc finger protein-mediated transcriptional repression.
        Mol Cell Biol. 2006; 26: 8623-8638
        • Schoenfelder S.
        • Fraser P.
        Long-range enhancer-promoter contacts in gene expression control.
        Nat Rev Genet. 2019; 20: 437-455
      3. Vucurovic K, Gallopin T, Ferezou I, Rancillac A, Chameau P, van Hooft JA, et al. (2010): Serotonin 3A receptor subtype as an early and protracted marker of cortical interneuron subpopulations. Cerebral cortex (New York, NY : 1991). 20:2333-2347.

        • Murthy S.
        • Niquille M.
        • Hurni N.
        • Limoni G.
        • Frazer S.
        • Chameau P.
        • et al.
        Serotonin receptor 3A controls interneuron migration into the neocortex.
        Nat Commun. 2014; 5: 5524
        • Maricq A.V.
        • Peterson A.S.
        • Brake A.J.
        • Myers R.M.
        • Julius D.
        Primary structure and functional expression of the 5HT3 receptor, a serotonin-gated ion channel.
        Science (New York, NY). 1991; 254: 432-437
        • Lee S.
        • Hjerling-Leffler J.
        • Zagha E.
        • Fishell G.
        • Rudy B.
        The largest group of superficial neocortical GABAergic interneurons expresses ionotropic serotonin receptors.
        J Neurosci. 2010; 30: 16796-16808
        • Huang L.
        • Wang J.
        • Liang G.
        • Gao Y.
        • Jin S.Y.
        • Hu J.
        • et al.
        Upregulated NMDAR-mediated GABAergic transmission underlies autistic-like deficits in Htr3a knockout mice.
        Theranostics. 2021; 11: 9296-9310
      4. Delevich K, Tucciarone J, Huang ZJ, Li B (2015): The mediodorsal thalamus drives feedforward inhibition in the anterior cingulate cortex via parvalbumin interneurons. The Journal of neuroscience : the official journal of the Society for Neuroscience. 35:5743-5753.

        • Udakis M.
        • Pedrosa V.
        • Chamberlain S.E.L.
        • Clopath C.
        • Mellor J.R.
        Interneuron-specific plasticity at parvalbumin and somatostatin inhibitory synapses onto CA1 pyramidal neurons shapes hippocampal output.
        Nature communications. 2020; 11: 4395
        • Mohammad-Zadeh L.F.
        • Moses L.
        • Gwaltney-Brant S.M.
        Serotonin: a review.
        Journal of veterinary pharmacology and therapeutics. 2008; 31: 187-199
        • He T.
        • Guo C.
        • Wang C.
        • Hu C.
        • Chen H.
        Effect of early life stress on anxiety and depressive behaviors in adolescent mice.
        Brain and behavior. 2020; 10e01526
        • Huang Q.
        • Zhou Y.
        • Liu L.Y.
        Effect of post-weaning isolation on anxiety- and depressive-like behaviors of C57BL/6J mice.
        Experimental brain research. 2017; 235: 2893-2899
        • Huang C.T.
        • Chen K.C.
        • Chen C.F.
        • Tsai T.H.
        Simultaneous measurement of blood and brain microdialysates of granisetron in rat by high-performance liquid chromatography with fluorescence detection.
        Journal of chromatography B, Biomedical sciences and applications. 1998; 716: 251-255
        • Saito M.
        • Aogi K.
        • Sekine I.
        • Yoshizawa H.
        • Yanagita Y.
        • Sakai H.
        • et al.
        Palonosetron plus dexamethasone versus granisetron plus dexamethasone for prevention of nausea and vomiting during chemotherapy: a double-blind, double-dummy, randomised, comparative phase III trial.
        The Lancet Oncology. 2009; 10: 115-124
        • Tremblay R.
        • Lee S.
        • Rudy B.
        GABAergic Interneurons in the Neocortex: From Cellular Properties to Circuits.
        Neuron. 2016; 91: 260-292
        • Wei S.
        • Du H.
        • Li Z.
        • Tao G.
        • Xu Z.
        • Song X.
        • et al.
        Transcription factors Sp8 and Sp9 regulate the development of caudal ganglionic eminence-derived cortical interneurons.
        J Comp Neurol. 2019; 527: 2860-2874
        • Chittajallu R.
        • Craig M.T.
        • McFarland A.
        • Yuan X.
        • Gerfen S.
        • Tricoire L.
        • et al.
        Dual origins of functionally distinct O-LM interneurons revealed by differential 5-HT(3A)R expression.
        Nat Neurosci. 2013; 16: 1598-1607
        • Niquille M.
        • Limoni G.
        • Markopoulos F.
        • Cadilhac C.
        • Prados J.
        • Holtmaat A.
        • et al.
        Neurogliaform cortical interneurons derive from cells in the preoptic area.
        Elife. 2018; 7
        • Yohn C.N.
        • Gergues M.M.
        • Samuels B.A.
        The role of 5-HT receptors in depression.
        Mol Brain. 2017; 10: 28
        • Bhatt S.
        • Devadoss T.
        • Manjula S.N.
        • Rajangam J.
        5-HT3 receptor antagonism a potential therapeutic approach for the treatment of depression and other disorders.
        Curr Neuropharmacol. 2021; 19: 1545-1559
        • Hammer C.
        • Cichon S.
        • Mühleisen T.W.
        • Haenisch B.
        • Degenhardt F.
        • Mattheisen M.
        • et al.
        Replication of functional serotonin receptor type 3A and B variants in bipolar affective disorder: a European multicenter study.
        Transl Psychiatry. 2012; 2e103
        • Niesler B.
        • Flohr T.
        • Nöthen M.
        • Fischer C.
        • Rietschel M.
        • Franzek E.
        • et al.
        Association between the 5' UTR variant C178T of the serotonin receptor gene HTR3A and bipolar affective disorder.
        Pharmacogenetics. 2001; 11: 471-475
      5. Hjort L, Rushiti F, Wang SJ, Fransquet P, S PK, S IC, et al. (2021): Intergenerational effects of maternal post-traumatic stress disorder on offspring epigenetic patterns and cortisol levels. Epigenomics. 13:967-980.

        • Jahn K.
        • Kurz B.
        • Sinke C.
        • Kneer J.
        • Riemer O.
        • Ponseti J.
        • et al.
        Serotonin system-associated genetic and epigenetic changes in pedophilia and child sexual offending.
        J Psychiatr Res. 2021; 145: 60-69
        • Schechter D.S.
        • Moser D.A.
        • Pointet V.C.
        • Aue T.
        • Stenz L.
        • Paoloni-Giacobino A.
        • et al.
        The association of serotonin receptor 3A methylation with maternal violence exposure, neural activity, and child aggression.
        Behav Brain Res. 2017; 325: 268-277
        • Barker J.M.
        • Zhang H.
        • Villafane J.J.
        • Wang T.L.
        • Torregrossa M.M.
        • Taylor J.R.
        Epigenetic and pharmacological regulation of 5HT3 receptors controls compulsive ethanol seeking in mice.
        Eur J Neurosci. 2014; 39: 999-1008
        • Xu Y.
        • Liu X.
        • Zhang X.
        • Zhang G.
        • Zhang R.
        • Liu T.
        • et al.
        Histone acetylation of the htr3a gene in the prefrontal cortex of Wistar rats regulates ethanol-seeking behavior.
        Neural Regen Res. 2012; 7: 1021-1028
        • Zhang H.
        • Herman A.I.
        • Kranzler H.R.
        • Anton R.F.
        • Zhao H.
        • Zheng W.
        • et al.
        Array-based profiling of DNA methylation changes associated with alcohol dependence.
        Alcohol Clin Exp Res. 2013; 37: E108-115
        • Xia G.
        • Han Y.
        • Meng F.
        • He Y.
        • Srisai D.
        • Farias M.
        • et al.
        Reciprocal control of obesity and anxiety-depressive disorder via a GABA and serotonin neural circuit.
        Mol Psychiatry. 2021; 26: 2837-2853