Abstract

   microRNA-592 (miR-592) has been linked to neurogenesis, but the
   influence of miR-592 knockout in vivo remains unknown. Here, we report
   that miR-592 knockout represses IPC-to-mature neuron transition,
   impairs motor coordination and reduces social interaction. Combining
   the RNA-seq and tandem mass tagging-based quantitative proteomics
   analysis (TMT protein quantification) and luciferase reporter assays,
   we identified MeCP2 as the direct targetgene of miR-592 in the mouse
   cortex. In Tg(MECP2) mice, lipofection of miR-592 efficiently reduced
   MECP2 expression in the brains of Tg(MECP2) mice at E14.5. Furthermore,
   treatment with miR-592 partially ameliorated the autism-like phenotypes
   observed in adult Tg(MECP2) mice. The findings demonstrate that miR-592
   might play a novel role in treating the neurodevelopmental-associated
   disorder.

   Subject terms: Developmental neurogenesis, Autism spectrum disorders
     __________________________________________________________________

   graphic file with name 41419_2022_4721_Figa_HTML.jpg

Introduction

   Accumulating evidence has indicated that some miRNAs are directly
   related to neurodevelopment, and loss of miRNAs function may lead to
   neurological disorders [[42]1, [43]2]. miRNAs, 19- to
   25-nucleotide-long RNAs, can regulate gene expression at the
   post-transcriptional level and result in translational repression or
   degradation [[44]3]. miR-592 is in the untranslated exon region of
   glutamate metabotropic receptor 8 (Grm8), which is associated with ASD
   [[45]4, [46]5]. In a large-scale CNV/miRNA genes association autism
   research data, dysregulation of hsa-miR-592 expression has been found
   in ASD patients [[47]6]. We previously identified miR-592 as
   neural-enriched miRNA which could induce astrogliogenesis
   differentiation arrest or/and enhance neurogenesis in vitro [[48]7]. By
   microarray analysis, loss of miR-592 in embryonic stem cells (ESCs)
   influenced ESC pluripotency and neurotrophin signalling pathway
   [[49]8]. Moreover, the abnormal miR-592 expression has been found in
   glioma and Alzheimer’s disease [[50]9, [51]10]. Together, these
   findings suggest that miR-592 contributes to pathogenesis in
   neurodevelopmental disorders. To date, whether miR-592 deficiency leads
   directly to neurodevelopmental disorders is unclear.

   Here, we generated miR-592 knockout (miR-592^−/−) mice to investigate
   the impact of miR-592 in vivo. Adult miR-592^−/− mice were
   characterized as reduced motor coordination and social interaction.
   Combining the bioinformatics analysis confirmed that methyl-CpG-binding
   protein 2 (MeCP2) was a targetgene of miR-592. MeCP2, an X-linked gene
   encoding the methylcytosine binding protein, is known to be the
   therapeutic target of ASD [[52]11]. Overexpression of MECP2 causes
   MeCP2 duplication syndrome (MDS) and loss of MECP2 cause Rett syndrome
   (RTT), which is both accompanied by autism-like features [[53]12,
   [54]13]. Studies on restoring MECP2 in adult Tg(MECP2) mice have
   precedents. Notwithstanding, abnormal MECP2 expression of adult mice
   was still prominent 4 weeks after the initiation of treatment [[55]14].

   In our study, normalized MECP2 expression levels have been started to
   be observed in miR-592-treated developing mouse brains after 1 week.
   Behaviour experiments showed that miR-592 treatment partly ameliorated
   social interaction deficits. In summary, these findings highlight the
   novel roles of miR-592 in neurodevelopment.

Materials and methods

Mice

   All procedures were conducted in accordance with the guidelines of the
   National Health and Medical Research Council of China and were approved
   by the animal ethics review board of Tongji University.

   miR-592 knockout mice were generated using CRISPR-Cas9 targeting.
   Shaorong Gao Laboratory provided an established setup protocol
   [[56]15]. The sgRNAs are shown in Fig. [57]1d. In vitro transcription
   of customized sgRNAs was performed using MEGAshortscript T7 kit (Thermo
   Fisher Scientific, Waltham, MA, USA). The source of Cas9-mRNA was
   GeneArt CRISPR Nuclease mRNA (Thermo Fisher Scientific, Waltham, MA,
   USA). The knockout efficiency was checked by PCR using DNA extracted
   from toe tissue from 2-week-old founder mice. The PCR primers used for
   miR-592^−/− mouse genotyping is listed in Table [58]1.

Fig. 1. Expression pattern of miR-592 indicated an important role in brain
development.

   [59]Fig. 1
   [60]Open in a new tab

   a Distribution of miR-592 in mature neurons in cortical slices of
   8-week-old WT mice as revealed by co-immunofluorescence experiments
   with an antibody against Tuj1 (green) and FISH of miR-592 (red). The
   amplification images are shown at the bottom of each group. Scale
   bars = 200 μm (upper), 50 μm (lower). Experiments were carried out
   using three biological replicates. Each sample was observed in three
   fields of view. b RT-PCR analysis of miR-592 expression in the
   developing cerebral cortex. The data are presented as the fold change
   compared with the expression level at E14.5. All data are presented as
   mean ± S.E.M. Experiments were repeated nine times (three biological
   and three analytical repeats). c Distribution of miR-592 in sagittal
   sections of the developing cerebral cortex at E12.5 to P56. ISH
   experiments were carried out using three biological replicates. Each
   sample was observed in three fields of view. d Illustration of the
   sgRNAs designed for miR-592^−/− mice. The arrows mark the
   sgRNA-targeting sequences. e, f Genotyping (e) and Sanger-sequencing
   analysis (f) of miR-592^−/− mice. The expected fragment sizes were as
   follows: WT, 526 bp; miR-592^−/− mice, 369 bp. g, h In situ
   hybridization (g) and RT-PCR (h) showing effective knockdown of
   miR-592. All data are presented as mean ± S.E.M. ***p < 0.001 (unpaired
   two-tailed t test). Experiments were carried out using three biological
   replicates. Each sample was observed in three fields of view or three
   analytical repeats. i Body weight changes over time (days). All data
   are presented as mean ± S.D. **p < 0.01, ***p < 0.001 (Student’s t
   test). Experiments were carried out using ten biological replicates. j
   Nissl staining showings evidence of cortical developmental
   malformations. Experiments were carried out using three biological
   replicates. Each sample was observed in three fields of view.

Table 1.

   Primers.
   Gene Sense Antisense Figure
   miR-592 (genotyping) 5’-CCATGACAACCGACCCTTGA-3’
   5’-AGCAACAACCTATATCCCACGG-3’ Fig. [61]1e
   MeCP2 (genotyping) 5’-CGCTCCGCCCTATCTCTGA-3’ 5’-ACAGATCGGATAGAAGACTC-3’
   Data not shown
   MeCP2 (dual-Luciferase) 5’-TGTTCTTCCTGGTGACTCTG-3’
   5’-CCCTTGTCCTACTCTATGGT-3’ Fig. [62]5b
   U6 (RT-PCR) 5’-CGCTTCGGCAGCACATATAC-3’ 5’-AATTTGCGTGTCATCCTTGC-3’ Fig.
   [63]1b
   miR-592 (RT-PCR) 5’-AAGGGATTCTGATGTTGGTCACAC-3’
   5’-GCTGTCAACGATACGCTACGTAACG-3’ Fig. [64]1b
   Sox2 (RT-PCR) 5’-TCTCAAACTGTGCATAATGGAGTAA-3’
   5’-CCCTTTTATTTTCCGTAGTTGTATT-3’ Fig. [65]2d
   Map2 (RT-PCR) 5’-CAATCTTCACATTACCACCTCCA-3’ 5’-CTCTAAAGAACATCCGTCAC-3’
   Fig. [66]2d
   TBR1 (RT-PCR) 5’-GCTTCGTCACAGTTTCGATGG-3’ 5’-CCGTTGGTAATGACCGGGTG-3’
   Fig. [67]2d
   TBR2 (RT-PCR) 5’-GCAATAAGATGTACGTTCACCCA-3’
   5’-GCAGAGACTGCAACACTATCAT-3’ Fig. [68]2d
   CTIP2 (RT-PCR) 5’-CCCGACCCTGATCTACTCAC-3’ 5’-GGAGGTGGACTGCTCTTGT-3’
   Fig. [69]2d
   PAX6 (RT-PCR) 5’-TACCAGTGTCTACCAGCCAAT-3’ 5’-TGCACGAGTATGAGGAGGTCT-3’
   Fig. [70]2d
   MeCP2 (RT-PCR) 5’-TTCTATTCTGGGCTTTTGATTTGT-3’
   5’-CCCTTGTCCTACTCTATGGTTATCA-3’ Figs. [71]4g, h and [72]5c
   Pten (RT-PCR) 5’-TGGATTCGACTTAGACTTGACCT-3’
   5’-GCGGTGTCATAATGTCTCTCAG-3’ Fig. [73]4g
   Lrrc7 (RT-PCR) 5’-CAAGCTCTACGGAAACTAAGCA-3’
   5’-ACACCGTTTTTACTGATGTCGAG-3’ Fig. [74]4g
   FOXP1 (RT-PCR) 5’-CACCTCAGGTTATCACTCCTCA-3’ 5’-AGCTGCAACTGTTCCTGTTGT-3’
   Fig. [75]4g
   GAPDH (RT-PCR) 5’-GCTGTCAACGATACGCTACGTAACG-3’ 5’-TGAAGGGGTCGTTGATCG-3’
   Fig. [76]4g
   Grm8 (RT-PCR) 5’-TCCCTTCCCTCTCCAACCTAACATG-3’
   5’-CCACGCTCTTCCAATCCTCTTTCC-3’ Fig. [77]S1e
   BDNF (RT-PCR) 5’-TCATACTTCGGTTGCATGAAGG-3’ 5’-AGACCTCTCGAACCTGCCC-3’
   Fig. [78]S4h
   [79]Open in a new tab

   Tg(MECP2) mice were generously gifted from Dr. Zilong Qiu and BTBR
   T + Itpr3tf/J (BTBR) mice (cat#.002282) were purchased from the Jackson
   Laboratory. The PCR primers used for Tg(MECP2) mouse genotyping are
   also listed in Table [80]1.

In situ hybridization (ISH)

   ISH and fluorescence in situ hybridization (FISH) combined with
   immunostaining were performed, with reference to a previous report
   [[81]16]. The brains of miR-592^−/− mice were dissected as soon as the
   animals were euthanized and fixed overnight in 4% paraformaldehyde and
   then embedded in paraffin. The brains were sectioned at a thickness of
   4 μm. Qiagen synthesized the Mmu-miR-592-5p miRCURY LNA miRNA Detection
   probe (5DiGN-CATCATCGCATATTGACACAAT-3DiG_N, Qiagen, Hilden, Germany)
   and at hybridization 40 °C for 2 h.

   For ISH, paraffin sections with the probe were incubated with
   anti-digoxigenin AP (1:400; Roche, Switzerland) at 37 °C for 50 min.
   Alkaline phosphatase activity was detected by developing the slides in
   BCIP/NBT. For FISH, a Cy3 tyramide signal amplification (TSA) kit
   (1:100; PerkinElmer, USA) allowed approximately 500-fold amplification
   of the original signal. The following antibody was used,
   anti-digoxigenin-POD (1:100, Roche, Switzerland). The sections were
   washed in DEPC water and mounted with DAPI or processed for
   immunohistochemistry. The results were checked with bright-field or
   fluorescence microscopy. The time course of expression was kept
   identical for a given experiment.

Histology

   Eight-week-old mice were used for histological analysis. Three mice
   were randomly selected from each group for histological examination.
   The hearts, livers, spleens, lungs, kidneys, testes, and brains of
   miR-592^−/− mice were dissected immediately after the animals were
   euthanized [[82]17]. Tissues from at least three mice of each genotype
   were stained with HE. For Nissl staining, slides were incubated with
   0.1% Nissl dye for 10 min.

Immunofluorescence (IF) staining

   IF staining was performed as previously reported in the literature
   [[83]18]. The primary antibodies used were anti-Pax6 (1:350, Abcam, UK,
   ab195045), anti-pH3 (1:1000, Abcam, UK, ab267372), anti-SOX2 (1:200,
   Abcam, UK, ab79351), anti-TBR2 (1:100, Abcam, UK, ab183991), anti-MAP2
   (1:500, Abcam, UK, ab32454), anti-Tuj1 (1:1000, Abcam, UK, ab7751),
   anti-GFAP (1:1000, Abcam, UK, ab7260), anti-Olig2 (1:1000, Abcam, UK,
   ab109186), anti-Iba1 (1:500, Abcam, UK, ab178846), anti-Syn1 (1:500,
   Abcam, UK, ab32532), anti-PSD95 (1:500, Abcam, UK, ab76115),anti-MeCP2
   (1:500, Abcam, UK, ab50005), and anti-BDNF (1:200, Abcam, UK,
   ab108319). The secondary antibodies used were an
   Alexa-Fluor®488-conjugated goat anti-rabbit antibody (1:500,
   Invitrogen, Carlsbad, USA, Carlsbad, USA) and an
   AlexaFluor®546-conjugated goat anti-mouse antibody (1:500, Invitrogen,
   Carlsbad, USA). For bromodeoxyuridine/5-bromo-2′-deoxyuridine (BrdU)
   labelling, E14.5 pregnant mice were intraperitoneally given BrdU
   (Sigma, USA) at 25 mg/kg body weight. Animals were sacrificed within
   30 min of injection. Cleaved caspase-3 staining was done using an
   anti-cleaved-caspase-3 antibody (1:100, Abcam, UK, ab32042). TUNEL
   staining was performed using TUNEL apoptosis detection kit.

Behavioural studies

   Mice were placed in white plastic cages with freely accessible water
   and food in an SPF animal house at 22 ± 2 °C with a 12 h light/dark
   cycle. All behavioural experiments were performed using 25–30 g male
   mice. All mice were age- and sex-matched and acclimated to the
   environment for at least 1 week. Eight mice were randomly selected from
   each group for behavioural examination. Behavioural experiments were
   performed from 9:00 to 17:00. The animals were habituated to the test
   room for two hours before starting the behavioural experiments. The
   video of the experiment was rated by two participants who were blinded
   to the treatment group after the behavioural evaluation.

   The open-field test (OFT) was carried out as the methods in previous
   studies [[84]19]. Mice were video recorded for 5 min as soon as placed
   individually in the middle of the squared box (42 × 42 × 42 cm).
   Between each experiment, the square box was cleaned with 75% alcohol
   and wiped dry, to remove odours. The recording and behaviour analyses
   were carried out using EthoVision XT version 13 (Noldus Information
   Technology, Wageningen, the Netherlands).

   The rotarod test (RT) was conducted according to the methods in a
   previous study [[85]20]. The rod initially rotated at 4 rpm, gradually
   increasing to a maximum of 40 rpm over a 5 min period. Two days before
   the experiment, mice were trained on the apparatus in two or three
   trials, with a 1 min break between trials. The mice were placed on the
   rotarod, and the latency to fall (cm) was recorded.

   An elevated plus-maze (EPM) test was conducted according to the methods
   in a previous study [[86]21]. The EPM had two closed arms (60×20×40 cm)
   and two open arms (arms not enclosed by walls). The mice were allowed
   to explore the maze freely for 5 min. The number of times each arm was
   accessed and the time spent were recorded.

   For the light–dark transition (LDT) test, a cage (50×30×30 cm) was
   separated into two parts: a light and a dark chamber [[87]22]. The mice
   were first placed in the dark chamber and allowed to freely explore
   between the two chambers for 5 min. The time spent in the chamber was
   documented.

   For a self-grooming test, mice were left alone in a box
   (26 × 26 × 26 cm) for 10 min [[88]19]. The total time all body parts
   were groomed was added up from a video record.

   For the novel object preference (NOP) test [[89]23], an opaque walled
   box(26 × 26 × 26 cm) was placed into a soundproof box. The test was
   performed in two steps, in the adaptive step, the mice were given
   10 min to explore two familiar objects in the opaque walled box. Then
   one object was replaced with a different one, mice were given another
   10 min to explore two objects automatically. The time spent in
   exploring the novel object was recorded.

   For the three-chambered test, wire mesh cylinders were placed into a
   three-chambered apparatus (60 × 40 × 21 cm). In the adaptation task,
   the mice were given 10 min for exploring the three-chambered apparatus.
   In the sociability test session, an unfamiliar same-sex mouse called
   stranger1 had been placed in the wire mesh cylinders. The test mouse
   was placed in the central chamber and allowed to freely explore these
   three chambers for 10 min. In the social novelty preference test, a
   second novel mouse, stranger2, was placed in the opposite chamber,
   which was empty in the previous session. The test mouse could freely
   explore the chambers for 10 min in this session. The video was recorded
   for 10 min, and the time of interaction between two mice was recorded
   [[90]24].

   For the Morris water maze (MWM) task, a circular tank (188 cm in
   diameter and 50 cm in height) was filled with water to a depth of 40 cm
   [[91]20]. A platform (d = 10 cm) was submerged 1.5 cm below the water’s
   surface in the northeast quadrant of the maze, and mice were placed
   into the tank from the far end of the maze. If a mouse failed to find
   the platform within 1 min, it was manually guided to the platform. The
   mice were trained each day for five days before the final test. The
   task performance measures, including average velocity, duration, and
   the distance travelled in the centre, were recorded for analysis.

RNA-seq and TMT-based quantitative proteomic analysis

   Total RNAs were extracted from cortical tissues of eight-week-WT mice
   and miR-592^−/− mice were extracted with a MiRNeasy Mini Kit (Qiagen,
   Hilden, Germany). Nine RNA samples from three individuals of each
   genotype (WT mice, heterozygote, miR-592^−/− mice) were performed by
   Berry Genomics Corporation, Beijing, China. Only six RNA samples (WT
   and miR-592^−/− mice) were used in the subsequent analysis. Around 44
   million to 60 million paired-end reads were obtained for each sample
   after quality control. The RNA-Seq raw data files have been uploaded in
   NCBI (BioProject accession no. PRJNA801399; BioSample accession nos.
   SAMN25342029). For TMT-based quantitative proteomic analysis,
   8-week-mouse cortical tissues were quantified BCA (Thermo Fisher
   Scientific, Waltham, MA USA), dissected and extracted by RIPA buffer
   for the whole proteins. Protein integrity was evaluated by SDS-PAGE. As
   RNA-seq, six protein samples from three individuals of each genotype
   were performed by Wayen Biotechnologies. The TMT-based quantitative
   proteomic analysis raw data files have been uploaded in iProX
   ([92]https://www.iprox.cn/page/PSV023.html;?url=1643453114835XUyi,
   Program ID IPX0004057000). GO functional enrichment analysis was
   performed using Metascape software. Kyoto Encyclopedia of Genes and
   Genomes (KEGG) analysis was applied using the KEGG pathway database
   ([93]https://www.genome.jp/kegg/pathway.html) and the DAVID
   ([94]https://david.ncifcrf.gov/).

MeCP2 reporter system construction

   The predicted regulation of MeCP2 by miR-592 was investigated using a
   pmirGLO Dual-Luciferase miRNA Target Expression Vector (Promega,
   Madison, WI, USA). The PCR primers for the MeCP2 3’-UTR are listed in
   Table [95]1. They were used to generate the mouse MeCP2 3’-UTR sequence
   containing the miR-592 target site. The DNA fragments were purified in
   a 1.5% agarose gel using a MinElute Gel Extraction Kit (Qiagen, Hilden,
   Germany) and then inserted downstream to the luciferase gene at XbaI
   enzyme-digested vector pGL3-Control (Promega, Madison, WI, USA). 293T
   cells were seeded at a density of 5 × 10^5 cells per well in 6-well
   plates. When the cells reached 80 to 90% confluence, the 293T cells
   were co-transfected with wild type MeCP2 3’-UTR (MeCP2 3’-UTR) or
   mutant MeCP2 3’-UTR (MT MeCP2 3’-UTR), and with miR-592 LNA mimic
   (Qiagen, Hilden, Germany), or miRNA mimic negative control (Qiagen,
   Hilden, Germany). Xfect Transfection Reagent (Clontech, Takara Bio USA)
   was used for transfection.

Neural progenitor cell (NPC) and neuron culture

   NPCs were isolated from the cortices of E14.5 miR-592^−/− mice and WT
   mice as described previously [[96]7]. The brains were removed from the
   skulls and the cortices were minced and incubated with 0.05% trypsin at
   37 °C for 25 min. For NPCs culture, cells were cultured in neurobasal
   medium containing 2% B27 (Thermo Fisher Scientific, Waltham, MA USA),
   1% N2, 20 ng/ml EGF, 20 ng/ml FGF, and penicillin/streptomycin. The
   cells were fed every 2 days. The neural spheres were counted under
   microscopy after 7 days. For neuron culture, the transfected NPCs were
   switched to neuronal differentiation medium, and differentiation was
   assessed 4 days later. Neuronal differentiation medium contained 0.5 mM
   glutamine and penicillin/streptomycin. Neurons were plated in
   poly-D-lysine (Sigma, USA) coated 6-well plates (1 × 10^6).

Real-time PCR (RT-PCR)

   Total RNA was extracted using TRIzol. A MiRNeasy Mini Kit (Qiagen,
   Hilden, Germany) was used to extract miRNA and total RNA from tissues
   and cells. Then, RT-PCR was conducted using TB Green® Advantage® RT-PCR
   Premix (Takara Bio Inc, Shiga, Japan). The results were quantified by
   using the 2−ΔΔCq method. U6 and GAPDH were used as housekeeping genes.
   The primer sequences used are also listed in Table [97]1.

Protein extraction and western blot analysis

   Brain tissue was lysed with RIPA buffer (Thermo Fisher Scientific,
   Waltham, MA, USA) with protease inhibitors and phosphatase inhibitors
   (Thermo Fisher Scientific, Waltham, MA, USA). The total soluble protein
   was quantified with a BCA Protein Assay Kit (Thermo Fisher Scientific,
   Waltham, MA, USA). Protein samples were loaded onto 8% or 10% SDS-PAGE
   gels, electrophoresis, and then transformed to PVDF membrane after
   fixation. The following antibodies were applied: anti-MeCP2 (1:5000,
   Abcam, UK, ab50005), anti-BDNF (1:1000, Abcam, UK, ab108319), anti-AKT
   (1:1000, Cell Signaling, 9272S), anti-Phospho-Akt (Ser473) (1:1000,
   Absin, abs130002), anti-mGluR8 (1:1000, Affinity, DF7121), anti-GAPDH
   (1:1000, Abcam, UK, ab8245), goat anti-rabbit IgG-HRP and
   Cy3-conjugated goat anti-mouse (1:5000, CST, Beverly, MA, USA), and
   Alexa Fluor®555 goat anti-rabbit IgG(H + L) (1:5000, Invitrogen,
   Carlsbad, USA). The membranes were treated following the ECL WB
   Protocol (Bio-Rad, Milan, Italy). The original images were recorded and
   analysed with Odyssey Infrared Imaging System.

LNA lipofection in vivo experiments

   LNA lipofection in vivo experiments were conducted with reference to
   previous reports [[98]25–[99]27]. In brief, miR-592 mimic solution
   (concentration) was mixed with 8 μl of Invivofectamine® 3.0 Reagent
   (Invitrogen, Carlsbad, USA) and incubated at 50 °C for 30 min. For the
   positive control okadaic acid (OKA) was dissolved in a final
   concentration of 2% DMSO. At E14.5, pregnant mice were anaesthetized
   with isoflurane. The uterine horns were exposed. A lateral ventricle of
   each embryo was injected using pulled glass capillaries (World
   Precision Instruments Inc, China) with Fast Green (BBI, Shanghai,
   China) combined with the LNA lipofection mix or OKA at 1 µl each.
   Finally, 20 pmol miR-592 mimic or the NC with a total volume of 1 μl
   were injected period into the one side lateral ventricle of the mouse.
   Since Lipofectamine mediated the transfection, no electric field was
   applied following the injection into telencephalic vesicles [[100]28,
   [101]29]. The mice were euthanized at various time points after
   lipofection, and the brains of the mice were dissected for
   immunofluorescence staining.

Statistical analysis

   Statistical analysis was performed with GraphPad Prism software. The
   data were presented as the mean ± S.D. or ±S.E.M. as indicated in the
   figure legends. P values <0.05 were considered to indicate a
   statistically significant difference.

Results

Expression pattern of miR-592 indicated an important role in brain
development

   We first examined the expression of miR-592 in the central nervous
   system (CNS). To investigate the location of miR-592 in CNS, we first
   performed FISH on adult mice brains. Colocalization of miR-592 with
   Tuj1 revealed that miR-592 was expressed in adult mouse neurons of the
   cortex (Fig. [102]1a). Accordingly, the expression patterns of miR-592
   in WT mice were detected by RT-PCR. The expression of miR-592 increased
   from E12.5 to E14.5, reached a peak at E16.5, then decreased to a low
   level around P1 and persisted through adulthood (Fig. [103]1b). The
   peak suggests that miR-592 plays an essential role in the early
   neurogenesis of mice [[104]30]. In ISH, miR-592 was detected mainly in
   the cell bodies of the ventricular zone (VZ) and subventricular zone
   (SVZ) at E12.5 and E14.5, while E16.5 to P7, miR-592 was detected most
   abundant in the intermediate zone (IZ) and cortical plate (CP). miR-592
   levels were declined precipitously after P14 (Fig. [105]1c and Fig.
   [106]S1a). Given this, miR-592 knockout mouse line was generated by
   CRISPR/Cas9 with three sgRNAs leading to a 157-bp deletion in Grm8
   (Fig. [107]1d). PCR genotyping, ISH and RT-PCR results showed effective
   knockout in miR-592^−/− mice (Fig. [108]1e–h). Isoform-sequencing
   (Iso-seq) was used to monitor structural variants of Grm8 transcripts
   and results indicated that Grm8 was not affected by knocking out
   miR-592 (Fig. [109]S1c–g). miR-592^−/− mice were viable and fertile,
   thereby, miR-592^−/− mice were mated with WT mice. The heterozygous
   strains were then cross-bred and used in subsequent trials to reduce
   the possibility of off-target events [[110]31]. However, an abnormal
   Mendelian ratio and sex ratio were observed in the progeny of
   heterozygote outcrosses (Fig. [111]S1b). Further results showed that
   deletion of the miR-592 in mice caused a 13% arrest of embryonic
   development (Fig. [112]S1h). We then examined mouse weights at
   different time points. miR-592^−/− mice weighed less than WT mice at
   the early time points but dramatically increased afterward (Fig.
   [113]1i). Nissl staining obtained from aborted miR-592^−/− mice
   displayed neurodevelopmental delay (Fig. [114]S1i, j). In the surviving
   miR-592^−/− mice, there were also irregular processes in the cortices
   (Fig. [115]1j). The above results indicate that abnormal developmental
   symptoms are exhibited after miR-592 knockout.

miR-592 deletion impairs cortical neurogenesis and adult lineage choice

   All other organs examined in adult miR-592^−/− mice were histologically
   unremarkable (Fig. [116]S2a). Therefore, we focused mainly on
   miR-592^−/− mouse brains. We first examined the NPC levels in the
   developing cortices of miR-592^−/− mice. We used p-H3 and Pax6 to label
   mitotic cells and apical neural progenitor cells (APCs) in the VZ/SVZ
   of E14.5 mouse brains [[117]32]. Interestingly, ectopic expression of
   Pax6 was induced by miR-592 knockout in E14.5 mouse brains. The
   percentage of p-H3^+ cells were significantly decreased in miR-592^−/−
   mice compared with WT (Fig. [118]2a). Then we carried out BrdU
   labelling studies at E14.5. The results showed a significant decrease
   in the percentage of BrdU^+ cells out of total cells in the VZ/SVZ
   regions of E14.5 miR-592^−/− brains, suggesting a decreased cell
   proliferation rate in the miR-592^−/− cerebral cortex (Fig. [119]2b).
   Next, we examined the intermediate progenitor cells (IPCs) markered by
   TBR2 [[120]33]. Remarkably, we observed that nearly complete
   colocalization was observed between TBR2^+ and SOX2^+ cells in
   miR-592^−/− mice at E14.5 and TBR2^+ IPCs were reduced during
   neurogenesis (Fig. [121]2c, d). SOX2 labels both APCs and IPCs
   [[122]34–[123]36]. It suggests that miR-592 deletion may affect
   quiescent NPC activation and their lineage choice in the developing
   cortices.

Fig. 2. miR-592 deletion impairs cortical neurogenesis and adult lineage
choice.

   [124]Fig. 2
   [125]Open in a new tab

   a Double immunofluorescence staining. Pax6 staining (red
   immunofluorescence), p-H3 staining (green immunofluorescence), and DAPI
   staining (blue fluorescence). Unpaired two-tailed t test. Error bars
   represent S.E.M. Experiments were carried out using three biological
   replicates. Each sample was observed in three fields of view. b BrdU
   staining (red immunofluorescence) and DAPI staining (blue
   fluorescence). Unpaired two-tailed t test. Error bars represent S.E.M.
   Experiments were carried out using three biological replicates. Each
   sample was observed in three fields of view. c E14.5–P7 sagittal
   sections were immunostained for the radial glial marker SOX2 (pink),
   the IPCs marker TBR2 (green), and the neuronal marker MAP2 (red).
   Nuclei were labelled with DAPI. d E14.5–P7 RT-PCR showed that TBR2^+
   cells were dysregulated in the SVZ in miR-592^−/− mice beginning at
   E14.5. VZ/SVZ ventricular/subventricular zone (PAX6 and SOX2), IZ the
   intermediate zone (TBR2), CP cortical plate (TBR1 and CITP2), MZ
   marginal zone (MAP2). e–h Double immunofluorescence staining. e Tuj1
   staining (red immunofluorescence), GFAP staining (green
   immunofluorescence), and DAPI staining (blue fluorescence). f Tuj1
   staining (red immunofluorescence), Iba1 staining (green
   immunofluorescence), and DAPI staining (blue fluorescence). g Tuj1
   staining (red immunofluorescence), Olig2 staining (green
   immunofluorescence), and DAPI staining (blue fluorescence). h Syn1
   staining (red immunofluorescence), PSD95 staining (green
   immunofluorescence), and DAPI staining (blue fluorescence). All data
   were analysed by two-way ANOVA followed by Tukeys multiple comparison
   test. Error bars represent S.E.M. *p < 0.05, **p < 0.01, ***p < 0.001.
   The averages of nine different fields of view were calculated for each
   animal (counts from three fields of three sagittal slices).

   The sagittal section of adult miR-592^−/− mice and WT mice were
   immunostained for Nestin, Ki67, Tuj1, GFAP, Iba-1, Olig2, and CD31 to
   test whether the irregular lineage choice affected terminal
   differentiation (Fig. [126]2e–h and Fig. [127]S2b, f). Nestin^+ and
   Ki67^+ double-positive cells were also significantly decreased in adult
   miR-592^−/− mice compared with WT (Fig. [128]S2b). Remarkably, an
   imbalance in neuronal versus glial lineage differentiation was
   detected. The numbers of astrocytes (GFAP^+) or microglia (Iba-1^+)
   were markedly increased in the cortices of adult miR-592^−/− mice,
   whereas those of neuron (Tuj1^+) were relatively decreased (Fig.
   [129]2e, f). We next investigated whether this change in miR-592^−/−
   mice showed a similar trend on both sides of the brain. P56 coronal
   section results of WT mice and miR-592^−/− mice showed that both the
   left and right brains hemispheres exhibited increased astrocytosis
   (Fig. [130]S2c). These results indicated that miR-592 may be also
   involved in the activation of astrocytes and microglia. Since
   neuroinflammation may be associated with neuronal apoptosis, we next
   examined whether miR-592 deletion was able to reduce neuronal apoptosis
   in the cortex. As expected, cleaved caspase 3 and NeuN-double-positive
   cells levels increased in the cortices of miR-592^−/− mice (Fig.
   [131]S2d). This result was confirmed by TUNEL^+ cells which
   significantly increased in the cortices of miR-592^−/− mice compared
   with that seen in WT mice (Fig. [132]S2e). Meanwhile, Glial cells have
   been associated with synaptic pruning and modulation of
   neurotransmission [[133]37–[134]39]. PSD95 and Syn1
   immunohistochemistry were used to determine whether miR-592 knockout
   affected dendritogenesis and synapse development. PSD95 immunostaining
   demonstrated significantly increased intensities (Fig. [135]2h). In
   addition, no noticeable change in CD31^+ endothelial cells was observed
   in miR-592^−/− mice (Fig. [136]S2f). We conclude that miR-592 knockout
   impairs cortical neurogenesis and adult lineage choice.

Genetic deletion of miR-592 reduces motor coordination and social interaction

   To further verify the effects of miR-592 on the functions of cortical
   neurons, we next subjected adult animals to a battery of behavioural
   tests and investigated its impact on neurobehaviour (Fig. [137]3a). We
   found that miR-592^−/− mice arrived at the centre zone significantly
   less frequently than WT mice in the OFT (Fig. [138]3b–d). Motor
   coordination and balance deficits in miR-592^−/− mice were confirmed in
   RT, in which the time to fall from a rot rod was tested (Fig. [139]3e).
   Compared to the WT mice the miR-592^−/− mice spent less time in the
   open arms of the EPM (Fig. [140]3f–h). Similarly, miR-592^−/− mice
   showed limited curiosity, spending less time in the light chamber in
   LDT than the WT mice (Fig. [141]3i). We further investigated whether
   learning and cognition were affected in miR-592^−/− mice. No
   significant differences were detected between the WT group and the
   miR-592^−/− group in the distance travelled in the centre and time
   spent in the centre area in the MWM (Fig. [142]3j–m). Therefore, we
   investigate whether miR-592 knockout impairs motor coordination and/or
   reduces social interaction in mice. BTBR mice were used as a positive
   control to adopt a standard measure of social interaction deficits
   [[143]40–[144]42]. Compared to WT littermates, miR-592^−/− mice showed
   more prolonged bouts of self-grooming (Fig. [145]3o). We next assessed
   the novelty preference of miR-592^−/− mice by NOP test and evaluated
   sociability using the three-chambered test. In the NOP test,
   miR-592^−/− mice presented limited preference for a new object (Fig.
   [146]3p and Fig. [147]S5h). For the three-chambered test, in the
   sociability test session, we found that WT C57 mice have a clear
   tendency to be curious about stranger mice, while miR-592^−/− mice
   lacked interest in interaction with stranger mice (Fig. [148]3q). In
   the social novelty preference test, miR-592^−/− mice and BTBR mice show
   less preferences for stranger2 than WT mice (Fig. [149]3r). These data