All experiments involving animals were performed in accordance with the animal protocol approved by the Institutional Animal Care and Use Committee at the Institute of Zoology, Chinese Academy of Sciences. Mice were housed in groups of 3–5 animals in a 12 h light/12 dark cycle, with standard mouse chow and water ad libitum. Utx^f/f (stock number 021926), transgenic Nestin-Cre (stock number 003771), and Emx1-Cre transgenic mice (stock number 005628) were bought from Jackson Lab. The conditional Utx knockout mice were generated by breeding Utx^f/f mice with either Nestin-Cre or Emx1-Cre transgenic mice.

All behavioral tests were performed during the light cycle between 07:00 and 19:00. Male mice at 2–3 months of age were used for all the behavioral tests. All the videos were analyzed by the Smart software (Pan Lab, Harvard Apparatus).

The open field test was conducted in a large box (72 × 72 × 36 cm) in a brightly lit room. Mice were placed in the center of the maze and were monitored from above by a video camera. The number of entries in the center zone (18 × 18 cm) of the maze was recorded over a 5-min trial to evaluate anxiety effects.

Light-dark box test was performed as previously described (Costall et al., 1993). An apparatus (45 × 27 × 27 cm) consisting of two chambers, a black chamber (18 × 27 cm) and a light chamber (27 × 27 cm), was used for the light/dark exploration test. Mice were placed into the dark box and allowed to move between the light box and dark box for 5 min. The total number of transitions and the time spent in each box were analyzed.

The elevated plus maze is a plus-shaped apparatus with four arms (two open, 62 × 8.5 cm; two closed, 62 × 8.5 × 30 cm), elevated 70 cm from the floor. Mice were placed at the junction of the four arms of the maze, and allowed to freely explore the maze for 5 min. The number of entries and duration in each arm were analyzed (Rodgers and Dalvi, 1997).

Morris water maze test was performed as described previously (Vorhees and Williams, 2006). A 120 cm diameter, 45 cm deep Morris water maze was filled with water to a depth of 25 cm. Target escape platform (diameter 13 cm) was hidden 1.5 cm beneath the surface of the water at the center of a given quadrant of the water tank. Four extra-maze cues, in different shapes, colors, and sizes, were uniformly located on the wall surrounding the water tank. The water temperature was adjusted to 21 ± 1°C. During training trials, mice were trained in four trials per day starting from different sites. The mice were allowed to swim for up to 1 min to locate the platform. If it failed to locate the platform within that time, escape was assisted. Mice were introduced gently to the hidden platform and allowed to rest on the platform for 15 s. For the probe trial, 24 h after the final training day the platform was removed and time spent and entry in each of the four quadrants were recorded.

Primary hippocampal neurons (1 × 10⁴ cells per well in a 24-well plate) were cultured from P0 Utx cKO and WT mice on plates coated with poly-D-lysine (100 μg/ml). The dissected hippocampus tissue was digested with trypsin-EDTA for 10 min at 37°C. The tissue was then washed three times with MEM plus 10% FBS and dissociated with the culture medium. Then neurons were grown in Neurobasal medium (Invitrogen) supplemented with 2% B27 (Invitrogen) and 2 mM GlutaMAX (Invitrogen) and penicillin/streptomycin. Primary hippocampal neurons were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol.

Stereotactic injections into the hippocampus (stereotaxic coordinates from Bregma: 2.0 mm caudal, 1.2 mm lateral, 2.0 mm ventral; 2.8 mm caudal, 2.0 mm lateral, 1.7 mm ventral) were performed as follows: 8-week-old Utx^flox/flox were bilaterally injected with 1 μl Adeno-associated virus (AAV) 2/8 (Heyuan, China; Titer: Control virus:6.24 × 10¹² V.G./ml; Cre virus: 5.04 × 10¹² V.G./ml) at a rate of 0.125 μl/min.

Total RNA were extracted from hippocampus tissues or cultured cells according to procedures using Trizol reagent (Invitrogen). Two micrograms of total RNA was reverse transcribed using either oligo (dT) primers or specific primers by using a Transcriptor First Strand cDNA Synthesis Kit (Roche). For real-time PCR analysis, according to the manufacturer's instructions, using a SYBR mix from Roche. 25 ng of cDNA and 0.5 mM primers were used in a final volume of 20 μl. The PCR steps were performed 30 s pre-denaturation at 95°C, followed by 45 cycles of 10 s denaturation at 94°C, 30 s annealing at 60°C, 30 s extension at 72°C. The analysis of RT-qPCR used the 2^−ΔΔC_T method. Each reaction was run in triplicate and analyzed following the ^ΔΔCt method using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a normalization control. All primers are listed in Table 1.

Hippocampal tissues and cultured neurons were lysed in buffer containing 25 mM HEPES at pH7.9, 150 mM NaCl, 1 mM PMSF, 20 mM NaF, 1 mM DTT, 0.1% NP40, and proteinase inhibitor cocktails (Roche). Protein concentrations were determined by Folin phenol method with bovine serum albumin as standard. Twenty micrograms of the protein was separated on 8–12% SDS-PAGE gels (Bio-Rad) and transferred to PVDF membranes (Millipore). The membranes were blocked in 5% BSA in TBS-T with 0.05% Tween-20 and incubated with primary antibodies at 4°C overnight. Dilutions of primary antibodies were 1:1000 for UTX (E409, Millipore), H3K27me3 (1:1,000, 07449, Millipore), PSD95 (ab2723, Abcam), Synapsin (ab8049, Abcam), and 1:10,000 for β-actin antibody (Sigma). As for the secondary antibodies, we used HRP-linked goat anti-mouse or HRP-Linked goat anti-rabbit at 1:500. Enhanced chemoluminescence (ECL, Pierce) was used for detection. Quantification of the blots was determined with Quantity One Ver.4.4.0 (BioRad, USA).

Adult mice were anesthetized, perfused with 4% PFA. Brain tissue was dissected out, equilibrated in 30% sucrose, and sectioned into 40 μm-thick serial sections. The brain sections were washed in PBS for 15 min three times, and then blocked in blocking solution (3% BSA+0.3%Triton X-100+0.2% sodium azide) at room temperature for 1 h. The primary antibodies we used are as follows: anti-H3K27me3 (1:1,000, 07449, Millipore), anti-Map2 (1:1,000, Mab3418, Millipore), anti-GFP (1:1,000, A10262, Life technology), anti-NeuN (1:1,000, millipore), anti-Doublecortin (1:500, millipore). After incubation in primary antibody solution at 4°C overnight, the brain sections were washed with TBS for 30 min three times and then incubated with the secondary antibodies conjugated to Alexa Fluor 488 or 594 with a concentration of 1:500 at room temperature. The sections were finally stained with DAPI and mounted using adhesion anti-fade medium.

Only transcripts that showed more than 1.5-fold differential expression compared to control were subjected to relevance network analysis. GO analysis was performed by Database for Annotation, Visualization and Integrated Discovery (DAVID v6.7) (Huang et al., 2009). Mouse phenotype and gene enrichment were analyzed by WEB-based Toolkit (https://toppgene.cchmc.org/).

Statistical analysis was performed using SPSS 17.0 software (SPSS Inc., Chicago, IL, USA) with one-way or two-way analysis of variance as specified in legend of each figure. Prior to all statistical analyses, data were examined for normality of variance using the Kolmogorov-Smirnov test. All data were presented as mean ± SEM, and statistically significant was defined as ^*p < 0.05; ^**p < 0.01; ^***p < 0.001.

To evaluate the function of Utx in the brain, we used Utx conditional knockout (cKO) mice that lack Utx expression in NSCs during development (Figure 1A). Utx cKO male mice could survive to adult, however, almost all Utx cKO female mice died within 3 weeks after birth. Utx deletion was then confirmed at both mRNA and protein levels in cKO mice (Figures 1B,D), and we did not observe any change of its Y chromosome homolog Uty at mRNA level in the cKO brain (Figure 1C). Since Utx is a histone demethylase and functions in removal of repressive trimethylation of H3K27 (Agger et al., 2007), we next examined the histone marks H3K27me3 in cKO mice. Consistent with our expectations, Utx deletion significantly resulted in increased expression of H3K27me3 in the hippocampus compared to that in the WT littermates (Figures 1E,F).

Male Utx cKO mice and their WT littermates were subjected to a battery of behavioral tests including tests for anxiety-like behaviors as well as learning and memory. We firstly conducted open field test and found that Utx cKO mice had smaller entries into the center of the arena (Figures 1G,I), whereas total distance traveled was comparable to that of WT littermates (Figure 1H). Increased anxiety-like behaviors of cKO mice were then confirmed in the light-dark box test and the elevated plus maze test. In the light-dark box test, Utx cKO mice showed a substantially decreased entries (Figures 1J,K) and time spent (Figure 1L) in the light box. Similarly, Utx cKO mice displayed decreased entries (Figures 1M,N) and spent less time in the open arms when compared with the WT littermates in the elevated plus maze test (Figure 1O).

Next, we determined whether Utx cKO mice have deficits in spatial learning and memory using the Morris water maze test. We found that cKO mice take a longer time to locate the hidden platform in the training trials (Figure 1P) and spend less time in the platform zone during the probe test (Figures 1Q,R), indicating that Utx cKO mice have deficits in hippocampus-dependent spatial learning and memory.

To further confirm the function of Utx in the brain, we crossed homozygous Utx^f/f mice with transgenic Emx1-Cre mice, in which the endogenous Emx1 locus directs expression of Cre recombinase specifically in the neocortex and hippocampus, to generate Emx1-Cre Utx cKO mice with Utx deletion in the forebrain (Figure 2A). Consistent with the observations in Nestin-Cre Utx cKO mice, these Emx1-Cre Utx cKO mice also displayed anxiety-like behaviors as indicated by decreased entry into the center zone in the open field test (Figures 2B–D), decreased entries (Figures 2E,F), and time spent (Figure 2G) in the light box in the light-dark box test. In the Morris water maze test, Emx1-Cre Utx cKO mice took prolonged time to locate the hidden platform during both training (Figure 2H) and probe trials (Figures 2I,J), indicating that impaired spatial learning and memory also exists in Emx1-Cre Utx cKO mice. These results suggest that Utx specific ablation in the forebrain phenocopies the behavioral deficits observed in Nestin-Cre Utx-cKO mice, further strengthening the critical role of Utx in regulation of cognitive behaviors.

The hippocampus is an important region in the brain in the information processing, consolidation, and storage, and responsible for cognitive ability and memory retention. It is also involved in emotional and cognitional functions (Eichenbaum, 2004). We then asked the question as to whether deletion of Utx in the hippocampus of adults could phenocopy the cognitive deficits exhibited in Utx cKO mice. To test this idea, we stereotactically injected AAV 2/8 virus expressing either GFP or Cre into the hippocampus of 8-week-old Utx^f/f mice (Figure 3A). As shown in Figure 3B, expression of Utx at the protein level was significantly downregulated along with Cre recombinase after virus injection. In particular, we observed that deletion of Utx in the hippocampus of adult Utx^f/f mice exhibited normal locomotion (Figure 3D), but showed fewer entries into the center zone (Figures 3C,E) in the open field test. This is consistent with the anxiety-like behavior and cognitional deficits in Utx cKO mice. Similarly, mice with Utx down-regulation in the hippocampus also showed decreased entries in the light chamber in the light-dark box test (Figures 3F,G,H), and fewer entries (Figures 3I,J), and time spent (Figures 3K) in the open arms in the elevated plus maze test. In the Morris water maze test, mice with Utx down-regulation in the hippocampus displayed the lower ability to locate the platform (Figures 3L,M,N), indicating impaired spatial learning and memory abilities. Taken together, these results indicate that deletion of Utx in the adult hippocampus indeed replicates the behavioral deficits observed in cKO mice in which deletion of Utx occurs during development, further supporting the idea that Utx plays critical roles in regulation of emotion and cognition in adults.

To explore the role of Utx loss-of-function in hippocampal synaptic transmission and plasticity, we recorded hippocampal LTP, a cellular model for learning and memory, in acute hippocampal slices prepared from 6 to 7 week-old Utx cKO and WT mice. As shown in Figures 4B,C, LTP at Schaffer collateral synapses in the CA1 region (Figure 4A) was significantly impaired in cKO mice when compared with that in WT littermates. There was a significant difference in the amplitude of LTP at 55–60 min after induction between cKO and WT mice (Figures 4B,C). These results indicated that Utx deficiency impairs hippocampal long-term synaptic plasticity, which may underlie the behavioral deficits displayed by cKO mice.

To further determine the cause of the impaired hippocampal synaptic plasticity in Utx cKO mice, we recorded AMPA receptor (AMPAR)-mediated miniature excitatory postsynaptic currents (mEPSCs) in CA1 pyramidal cells using the whole-cell patchclamp technique. AMPAR mediated mEPSCs were pharmacologically isolated by bath application of the GABAA and NMDA receptor antagonists bicuculline (10 μM) and D-AP5 (25 μM) in the presence of TTX (0.5 μM). At a holding potential of −70 mV, mEPSCs displayed fast inward currents and sensitive to 20 μM DNQX, indicating AMPAR-mediated currents. As shown in Figure 4D, the amplitude, but not the frequency, of mEPSCs was significant decreased in slices from cKO mice when compared with that from WT mice. The results from electrophysiological recordings further support our speculation that Utx deletion impairs hippocampal synaptic transmission.

Impaired LTP and decreased amplitude of mEPSCs are most likely due to the deficits in postsynaptic function and/or reduction of functional synapses (Liao et al., 1995; Xiao et al., 2007). To test this assumption, we examined the immunoreactivity of synaptic-related proteins in the hippocampus of both WT and cKO mice. As shown in Figures 4E,F, expression of postsynaptic density protein PSD-95, a postsynaptic marker, was significantly decreased, while synapsin, a presynaptic marker, remained unchanged in cKO mice when compared to that in WT littermates. These results suggest that decreased expression of PSD-95 may lead to a deficit of postsynaptic plasticity in Utx cKO mice.

Neuronal signal integration as well as synaptic transmission and plasticity highly depend on the morphology of dendrites and their spines (Hering and Sheng, 2001). The impairment of hippocampal synaptic plasticity and the reduced PSD95 expression in Utx cKO mice prompted us to postulate that Utx deficiency may cause abnormal neuronal morphology. To determine whether disruption of Utx affects neuronal morphology, we traced Golgi-stained hippocampal neurons in the CA1 and dentate gyrus (DG). Utx cKO mice exhibited a significant reduction in dendritic length and number of branches in neurons in both the CA1 (Figures 5A–D) and DG regions (Figures 5E–G). Moreover, a reduction of spine density in hippocampal CA1 pyramidal neurons was also found in Utx deletion mice (Figures 5H,I). These data, together with the reduction in the postsynaptic PSD95 protein (Figure 4E), suggest an essential role for UTX in synaptic formation and dendritic development in hippocampal neurons, which in turn affect spatial learning and memory.

We next used a well-established in vitro primary neuron culture system to examine the effects of UTX deletion on the morphology of hippocampal neurons. Indeed, Utx cKO hippocampus neurons displayed a significant reduced in total dendritic length and dendritic complexity compared with control neurons isolated from littermate newborn pups (Figures 6A–C) at Day 7 in vitro (DIV7). At DIV19, cKO hippocampal neurons showed significant reductions in spine density (Figures 6D,E), which is consistent with the Golgi staining observations in the CA1 pyramidal neurons in vivo (Figures 5H,I). To further determine whether UTX regulates neuronal morphology, we acutely manipulated UTX expression by infecting in newborn hippocampus neurons isolated from Utx^f/f pups with lenti-cre viruses to delete UTX expression. As expected, we found that acute knockdown of UTX in hippocampus neurons also led to a significant decrease in total dendritic length and dendritic complexity (Figures 6F–H). These loss-of-function data in primary neurons further support our in vivo observation that deletion of UTX inhibits neuronal dendritic development.

Next, we investigated the effects of Utx loss-of-function on hippocampal gene expression, in order to identify potential mechanisms that might be associated with the synaptic plasticity and cognitive dysfunction in Utx cKO mice. To delineate the molecular pathways regulated by Utx, we conducted gene expression profiling (RNA-seq) using total RNA from the hippocampus of WT and cKO mice. One hundred forty seven down-regulated and 87 up-regulated genes were identified in cKO mice when compared with WT mice (Figures 7A–C and Table 2). To uncover the genes involved in the phenotypes of Utx loss-of-function, we performed Gene Ontology (GO) enrichment analysis of dysregulated genes in Utx cKO mice and found several enriched GO terms for biological processes, including cognition, learning and memory, anxiety, and synaptic transmission (Figure 7C). We then focused on the 147 genes with decreased expression in cKO mice, as their downregulation likely resulted from a loss of Utx. Most of those down-regulated genes are involved in regulation of Ion transport, calcium ion binding, neurite elongation, dendritic, and synaptic formation (Table 3), suggesting an essential role of Utx in Cognition, Learning and Memory, Anxiety, Synaptic Formation, and Function (Figures 7C,D and Table 3).

We then further examined some of those down-regulated genes relevant for synaptic plasticity and cognition found to be altered by RNA-seq. Using RT-qPCR, we verified that several genes critical for synaptic plasticity, and/or dendrite development including nitric oxide synthase 1 (Nos1), actin filament cross-linker α-Actinin-2 (Actn2), zinc finger transcription factor Egr3, transforming growth factor-β2 (Tgfb2), and Wnt4 were down-regulated in the hippocampus of Utx cKO mice (Figure 7D). Several other essential genes (Gpr88 and Edn1) involved in mental disorder, learning, and memory were also suppressed in the hippocampus of Utx cKO mice (Figure 7D). Amongst the down-regulated genes, we were particularly interested in Htr5b, a functionally unknown serotonin receptor. Serotonin and its receptors have been implicated in dendrite morphology, synaptic transmission and cognition (Wirth et al., 2016). To validate the results from RNA-seq analysis, we conducted qPCR and western blotting analyses and observed that Utx deletion significantly decreased expression of Htr5b in the hippocampus of cKO mice (Figures 7D–F). To further verify whether H3K27me3 enriches on the promoter of Htr5b, we performed ChIP analysis and found that H3K27me3 was enriched in the promoter region of Htr5b (Figure 7G), suggesting that Utx mediated H3K27me3 demethylase activity is required for the alteration in Htr5b in the hippocampus.

The serotonergic system is involved in many aspects of neural development, including neurite outgrowth, somatic morphology regulation, synaptogenesis, and control of dendritic spine shape and density (Wirth et al., 2016). To examine a functional relationship between Utx and Htr5b in mediating neural development, newborn hippocampal neurons from WT mice were transfected with lentivirus expressing both GFP and Htr5b shRNA sequences to suppress Htr5b at both mRNA and protein levels (Figures 8A,B). Knockdown of Htr5b resulted in a significant reduction in total dendritic length (Figures 8C,D) and spine density (Figures 8H,I). Given the fact that the expression level of Htr5b is decreased in Utx cKO mice that displayed neuronal morphological abnormalities (Figures 7E,F), we reasoned whether Htr5b gain-of-function could rescue neural development deficits in Utx cKO neurons. To test this assumption, exogenous Htr5b was expressed in Utx cKO newborn hippocampal neurons. Interestingly, Htr5b gain-of-function in cKO neurons was sufficient to restore the dendritic length (Figures 8E–G) and spine density (Figures 8J,K). These results suggest that Htr5b is a functional downstream target of Utx in modulating neuronal growth and morphology.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.