Depression-like behaviors induced by defective PTPRT activity through dysregulated synaptic functions and neurogenesis

Depressive disorder is a major health problem with a high prevalence and a heavy socioeconomic burden (Levy et al., 2018; Berton and Nestler, 2006; Kessler and Bromet, 2013). Although many types of antidepressants have been developed so far, current antidepressants still show a delayed onset of action as well as a lack of efficacy (Berton and Nestler, 2006; Rush, 2007). At present, many antidepressant drugs are designed to target monoaminergic neurotransmission, but there is a growing consensus that altered monoaminergic transmission is not sufficient to explain the etiology of depressive disorder (Berton and Nestler, 2006; Chaudhury et al., 2015; Krishnan and Nestler, 2011; Luscher et al., 2011). Functional imaging studies on patients with depression show dysregulated neuronal connectivity, and rodent models of depression display neuronal atrophy, reduced synaptic density and cell loss (Zeng et al., 2012; Holmes et al., 2019; Duman and Aghajanian, 2012; Czéh et al., 2018). It has also been found that typical antidepressants increase neurotrophic factor expression and enhance synaptic plasticity. On the other hand, atypical antidepressants such as ketamine, an NMDA receptor antagonist, increase synaptic formation and produce rapid antidepressant responses in patients with treatment-resistant depression (Machado-Vieira et al., 2012; Jaso et al., 2017). The understanding of neuronal loss and synaptic dysfunction mechanisms in depressive disorder could thus lead to the identification of key factors for the development of fast-acting and effective antidepressants.

Stressful life events influence the onset and course of depression, but not all people who live a stressful life develop depressive disorder (Caspi et al., 2003). Both genetic and environmental determinants, as well as their interaction, are crucial factors causing depressive disorder (Dunn et al., 2015; Mandelli and Serretti, 2013). The heritability of major depressive disorder has been estimated at ∼30–40%, and depression is a polygenic trait influenced by many genetic variants (Howard et al., 2018). Recently, genome-wide association studies (GWAS) with large sample sizes (over 100,000 participants) combined with publically available genomic data have identified many gene variants and genetic loci significantly associated with depressive disorder (Howard et al., 2019, 2018; Wray et al., 2018; Hyde et al., 2016; Major Depressive Disorder Working Group of the Psychiatric GWAS Consortium, 2013). These mega-analyses reveal new druggable genes associated with depression that are not currently associated with antidepressant treatment. A meta-analysis has identified new genes and gene pathways associated with synaptic structure and neurotransmission as well as gene sets associated with depression (Howard et al., 2019).

PTPRT (protein tyrosine phosphatase receptor T; also known as RPTPρ) is mainly expressed in the central nervous system and has been known to regulate the synaptic formation and dendritic arborization of hippocampal neurons (Besco et al., 2006; Lim et al., 2009; Park et al., 2012; Lee, 2015). PTPRT has the characteristic extracellular ectodomain along with two intracellular catalytic domains with tyrosine phosphatase activity. PTPRT forms homophilic trans dimers through its extracellular ectodomain, and its expression both in pre- and post-synapses can strengthen synaptic formation. PTPRT appears to obtain easy accessibility to its synaptic substrates linked to adhesion molecules like neuroligin and neurexin by forming additional cis interactions. It has been reported that knock-in mice expressing inactive PTPRT-D1046A have higher social approach scores than wild-type mice (Rajamani et al., 2015). Recently, several studies have reported that deficits in expression of the PTPRT gene induce abnormal brain functions and neurological diseases including depressive disorder; however, the molecular and physiological mechanisms are not well understood (Schuurs-Hoeijmakers et al., 2013; Quintela et al., 2017; Allen-Brady et al., 2009; Aston et al., 2005; Hyde et al., 2016; Roberson-Nay et al., 2018 preprint).

When we examined the behaviors of PTPRT-null and PTPRT-D401A-mutant mice, they displayed enhanced depression-like behaviors compared to wild-type mice. PTPRT-deficient mice exhibited dysregulated synaptic functions and attenuated hippocampal neurogenesis. We investigated new physiological roles of PTPRT using a multidisciplinary approach, and suggest that PTPRT deficiency plays a crucial role in the depression-like behaviors of mice.

PTPRT^−/− null mice were generated by targeting exon 22 of PTPRT, which, as described previously, encodes the first catalytic domain (Zhao et al., 2010). When the expression of PTPRT protein was examined with western blotting of samples from the brains of wild-type (PTPRT^+/+), PTPRT^+/− heterozygote, and PTPRT^−/− null mice, PTPRT disappeared completely in the PTPRT-null mice (Fig. 1A). No remarkable differences were observed in the Nissl-stained brain sections of PTPRT-null mice compared with those of wild-type mice (Fig. S1). PTPRT-null mice did not display any aberrant phenotypes in the behavioral tests on motor functions, social interaction, anxiety, learning and memory (Fig. 1B–L). However, PTPRT-null mice displayed enhanced depression-like behaviors compared with wild-type mice on the tail suspension test (TST), forced swim test (FST) and sucrose preference test (Fig. 1M–O). Intraperitoneal injection of imipramine, an antidepressant, reduced depression-like behaviors in PTPRT-null mice (Fig. 1P,Q). These results suggest that a deficiency of the PTPRT gene enhances depression-like behaviors in mice.

To characterize the association between PTPRT and depression-like behaviors, the chronic restraint stress model, a well-validated model of depression, was used (Chiba et al., 2012). The mice subjected to chronic restraint stress for 14 days displayed significantly enhanced depression-like behaviors on the TST and FST, compared with control mice (Fig. 2A,B). Interestingly, the expression of PTPRT was reduced in the hippocampus of the mice exposed to chronic restraint stress compared with expression in control mice (Fig. 2C), whereas PTPRT expression was not significantly changed in other brain regions. Previously, we found that PTPRT is inactivated by the phosphorylation of tyrosine 912, which is located in the wedge of the catalytic domain (Lim et al., 2009). Here, an anti-pY912-PTPRT antibody was produced to detect the phosphorylation of tyrosine 912, and inactive phospho-PTPRT proteins were identified in the hippocampus of the mice. Inactive PTPRT increased in the hippocampus of those mice that were exposed to stress, compared with levels in control mice (Fig. 2D). These results show that chronic restraint stress that induces depression-like behaviors attenuated the expression of PTPRT and PTPRT activity.

To confirm whether deficiency of the PTPRT gene enhances depression-like behaviors, PTPRT was knocked down transiently, and depression-like behaviors were examined (Fig. 2E–G). When lentivirus containing shRNA targeting PTPRT (Lenti-PTPRT-shRNA) was injected into the dentate gyrus of the hippocampus, mice displayed significantly enhanced depression-like behaviors, compared with control mice (Lenti-GFP). When the same Lenti-PTPRT-shRNA was injected into the CA1 region of the hippocampus, mice did not display the enhanced depression-like behaviors (Fig. S2). On the other hand, when adeno-associated virus (AAV-GFP) containing the PTPRT gene (AAV-PTPRT) was injected into the dentate gyrus of PTPRT-null mice, the rescued expression of PTPRT relieved them from the depression-like behaviors effectively (Fig. 2H–J). These results show that depression-like behaviors were enhanced by the PTPRT deficiency in the dentate gyrus, and that depression-like behaviors of PTPRT-null mice were ameliorated by the PTPRT rescue.

PTPRT gene-targeted mutant mice carrying a missense mutation (D401A) in the extracellular FN domain were generated by screening of the ENU-driven mutant mouse genomic DNA library and subsequent in vitro fertilization (Sakuraba et al., 2005). Sequencing genomic DNA of PTPRT-D401A-mutant mice confirmed the replacement of the highly conserved aspartic acid residue with alanine (Fig. 3A). The expression of PTPRT protein in the brains of PTPRT-D401A-mutant mice was comparable to that in wild-type mice (Fig. 3B). It has been found that PTPRT regulates the synaptic formation of hippocampal neurons through the interaction between its extracellular ectodomains (Lim et al., 2009). Recombinant PTPRT protein with the D401A mutation displayed augmented interactions between the ectodomains compared with wild-type PTPRT (Fig. 3C). When PTPRT expressed in heterologous cells was immunoprecipitated with anti-pY912-PTPRT antibody, more PTPRT protein with the D401A mutation was phosphorylated compared with wild-type PTPRT (Fig. 3D). Similarly, the level of inactive phospho-PTPRT was increased in the brains of PTPRT-D401A-mutant mice compared with that in wild-type mice (Fig. 3E). In a previous study, overexpression of PTPRT was found to increase the synaptic formation of hippocampal neurons (Lim et al., 2009). When overexpressed in hippocampal neurons, PTPRT-D401A did not increase synaptic formation as much as wild-type PTPRT (Fig. 3F–H). These results show that the D401A mutation attenuates PTPRT activity through aberrantly regulated interactions between extracellular ectodomains. When behavioral tests were performed, PTPRT-D401A-mutant mice also displayed enhanced depression-like behaviors, compared with wild-type mice (Fig. 3I–K). No other aberrant phenotypes were observed in PTPRT-mutant mice (Fig. S3). Intraperitoneal injection of imipramine relieved the PTPRT-D401A-mutant mice from the depression-like behaviors (Fig. 3L,M). These results show that the depression-like behaviors observed in PTPRT-D401A-mutant mice are induced by defective PTPRT activity.

The numbers of excitatory and inhibitory synapses were decreased in the hippocampal neurons of PTPRT-null and D401A-mutant mice compared with those in wild-type mice (Fig. 4A,B). Neonatal PTPRT-null mice (P7, postnatal day 7) showed reduced expression of AMPA-type glutamate receptors, especially GluR2 (also known as GRIA2), in hippocampal tissues compared with wild-type mice (Fig. 4C). However, the expression of AMPA receptors was not altered in the hippocampus of juvenile PTPRT-null mice (P21, postnatal day 21) compared with wild-type mice (Fig. 4D). When the hippocampal neurons of PTPRT-null mice were immunostained in vitro at 15 days (DIV 15), puncta of GluR1 (also known as GRIA1) and GluR2 were significantly decreased in number and intensity compared with those in wild-type mice (Fig. 4E,F). These results suggest that the decrease in excitatory synapses in the hippocampal neurons of PTPRT-deficient mice could be relevant to the reduced expression of glutamate receptors.

Immunoprecipitation with anti-PTPRT antibody was performed in the rat synaptosome to characterize the interaction between PTPRT and AMPA receptors, and we observed that considerable amounts of GluR1 and GluR2 were recruited by PTPRT (Fig. 5A). In contrast, when expressed in heterologous cells, GluR2 interacted directly with PTPRT, but GluR1 did not (Fig. 5B,C). GluR2 interacted with D401A-mutant PTPRT strongly compared with wild-type PTPRT (Fig. 5D). Interestingly, the tyrosine residues of GluR2 were effectively phosphorylated by Fyn tyrosine kinase and dephosphorylated by PTPRT, whereas those of GluR1 were not (Fig. 5E,F). Then, immunoprecipitation with anti-phosphotyrosine antibody (4G10) was performed, followed by western blotting to detect AMPA receptors in the synaptosomes of PTPRT-deficient mice (Fig. 5G,H). It was evident that tyrosine phosphorylation of GluR2 was increased in PTPRT-deficient mice compared with the level of phosphorylation in wild-type mice. These results suggest that the tyrosine residues of GluR2 were dephosphorylated by PTPRT in the brain, and that GluR2 could be a neuronal substrate of PTPRT. A previous study has suggested that the trafficking of GluR2 is regulated by certain tyrosine phosphatases, but provided no additional information regarding the specific tyrosine phosphatases that might regulate GluR2 trafficking (Hayashi and Huganir, 2004; Moult et al., 2006). Here we examined whether the trafficking of glutamate receptors could be regulated by PTPRT (Fig. 5I,J). HA-tagged GluR1 and GluR2 were expressed in PTPRT-knockdown hippocampal neurons, and their distributions on the surface and the internal side were analyzed. The distributional ratio of GluR1 between the surface and internal side did not change, but the distributional ratio of GluR2 was altered significantly by the knockdown of PTPRT. These results suggest that the membrane trafficking of GluR2 could be regulated by PTPRT in hippocampal neurons.

Long-term potentiation (LTP) was measured in the hippocampus of PTPRT-deficient mice to examine whether synapse plasticity was changed by defective PTPRT activity. The magnitudes of LTP at the perforant–dentate gyrus synapse were significantly higher in both PTPRT-null and D401A-mutant mice than in wild-type mice (Fig. 6A,E), whereas the magnitudes of LTP at the Schaffer collateral–CA1 synapse did not differ between PTPRT-null and D401A-mutant mice and wild-type mice (Fig. 6B,F). To determine the basic properties of synaptic transmission in the PTPRT-deficient mice, we examined paired-pulse facilitation (PPF) ratios at the perforant–dentate gyrus synapse and the Schaffer collateral–CA1 synapse. PTPRT-null and D401A-mutant mice displayed significantly lower PPF ratios at the perforant–dentate gyrus synapse than wild-type mice (Fig. 6C,G). The PPF ratios at the Schaffer collateral–CA1 synapse of the PTPRT-null and D401A-mutant mice, however, showed no significant difference from the ratios observed in wild-type mice (Fig. 6D,H). These results suggest that defective PTPRT activity modulated the presynaptic release of glutamate, and that synaptic plasticity was augmented at the perforant–dentate gyrus synapse of PTPRT-deficient mice.

Because of the altered synaptic plasticity at the perforant–dentate gyrus synapse, synaptic transmission was measured in the dentate gyrus of PTPRT-deficient mice. The frequency of miniature inhibitory postsynaptic currents (mIPSCs) was decreased in the dentate gyrus granule cells of both PTPRT-null and D401A-mutant mice (Fig. 7B,D; Fig. S4), whereas miniature excitatory postsynaptic currents (mEPSCs) did not change significantly compared with those in wild-type (Fig. 7A,C). Neurotransmitter levels were then analyzed in the hippocampal tissues of PTPRT-deficient mice (Fig. 7E). Although serotonin and catecholamines did not show changes, GABA levels were significantly reduced in the hippocampus of PTPRT-null and D401A-mutant mice, compared with levels in wild-type. Glutamate was considerably reduced in D401A-mutant mice, but not in PTPRT-null mice. When hippocampal slices were immunostained with anti-GABA antibody, GABA intensity was observed to be decreased significantly in the dentate gyrus of PTPRT-null mice compared with the intensity in wild-type mice (Fig. 7F). In the hippocampal CA3 region of PTPRT-null mice, GABA intensity was decreased compared with that in wild-type mice, but the difference was not significant (Fig. 7G). When the interneurons releasing GABA were examined in the dentate gyrus of PTPRT-null mice, they showed no reduction in cell numbers compared with those of wild-type mice (Fig. S5) (Czéh et al., 2018). The expression levels of glutamic acid decarboxylase (GAD), which produces GABA, and vesicular GABA transporter (vGAT), which transports GABA, were not altered in the hippocampus of PTPRT-null and D401A-mutant mice (Fig. S6). However, expression of GAT3 (also known as SLC6A11), a GABAergic transporter, was significantly reduced in the hippocampal tissues of PTPRT-null and D401A-mutant mice, whereas the expression of GAT1 (also known as SLC6A1) was comparable to that in wild-type mice (Fig. 7H,I). GAT3 interacted with PTPRT in heterologous cells as well as in the rat synaptosome (Fig. 7J,K). Immunostaining of hippocampal slices displayed the colocalization of PTPRT and GAT3 in the dentate gyrus (Fig. S7). GAT3 pulled down less D401A-mutant PTPRT than wild-type PTPRT when immunoprecipitation was performed with the lysate of heterologous cells (Fig. 7L). Interestingly, tyrosine phosphorylation of GAT3 was increased in the synaptosome of PTPRT-null mice compared with that in wild-type mice (Fig. 7M). These results suggest that the synaptic function of GAT3 associated with the release of GABA could be regulated by PTPRT through tyrosine dephosphorylation. When allopregnanolone (3α,5α-tetrahydroprogesterone, also known as 3α,5α-THP), a potent positive allosteric modulator of the action of GABA_A receptor, was applied, the inhibitory transmissions in the dentate gyrus granule cells of PTPRT-null mice were rescued to a level comparable to that in wild-type mice (Fig. 7N). In addition, the depression-like behaviors of PTPRT-null mice were ameliorated by intraperitoneal injection of allopregnanolone (Fig. 7O). These results indicate that attenuated GABAergic synaptic function of the dentate gyrus granule cells could be associated with the depression-like behaviors of PTPRT-deficient mice.

Attenuated neurogenesis has been shown in the dentate gyrus of patients with depression (Boldrini et al., 2009; Lucassen et al., 2010; Boldrini et al., 2013; Yun et al., 2016). When hippocampal neurogenesis was investigated in PTPRT-deficient mice after intraperitoneal injection of 5-bromo-2′-deoxyuridine (BrdU) followed by brain sampling 2 days later, there was no difference in the number of BrdU-positive dividing nerve cells in the dentate gyrus of PTPRT-null and wild-type mice (Fig. 8A). However, the number of newborn granule cells expressing doublecortin (DCX) was reduced, and the neurite length of the cells decreased in the dentate gyrus of PTPRT-null and D401A-mutant mice (Fig. 8B,C). These results show that the defective PTPRT activity attenuated neurogenesis as well as the maturation of newborn granule cells in the dentate gyrus of PTPRT-deficient mice. The proliferation and multi-lineage differentiation of neural stem cells (NSCs) were then examined in cultured embryonic NSCs prepared from the hippocampus of PTPRT-null mice. The results showed a decrease in the number of PTPRT-null NSCs derived from secondary neurospheres compared with the number of wild-type NSCs derived in the same manner (Fig. 8D). When neurogenic ability was examined by counting the number of Tuj1-positive neurons (indicating expression of neuron-specific β-tubulin) differentiated from PTPRT-null NSCs, a significant reduction was observed compared with the number differentiated from wild-type NSCs (Fig. 8E). However, the number of GFAP-positive astrocytes differentiated from PTPRT-null NSCs displayed no difference from wild type. In addition, the neurons derived from PTPRT-null NSCs showed reduced neurite length and attenuated dendritic arborization compared with those derived from wild-type NSCs (Fig. 8F). These results suggest that defective PTPRT activity attenuates the proliferation and the neurogenic ability of embryonic NSCs in the hippocampus of PTPRT-deficient mice.

Brain-derived neurotrophic factor (BDNF) plays significant roles in neurogenesis and promotes the differentiation and proliferation of NSCs through activation of Akt (Bartkowska et al., 2007; Bath et al., 2012; Goncalves et al., 2016). In the brain synaptosome, PTPRT interacted with BDNF receptor Ntrk2 as well as with p75NTR (NGFR), PIK3CA, PDPK1 and Akt1, which are involved in the neurotrophin signal pathway (Fig. 8G). When PTPRT and Ntrk2 were expressed in heterologous cells, PTPRT was pulled down effectively by Ntrk2 (Fig. 8H). Interestingly, the phosphorylation level of activity-dead PTPRT (PTPRT-2CS) was higher than that of wild-type PTPRT (PTPRT-WT) because of auto-dephosphorylation (Fig. 8I) (Lim et al., 2009). These results suggest that PTPRT can interact with Ntrk2 in the brain, and that the tyrosine residues of PTPRT are phosphorylated by Ntrk2 tyrosine kinase. The phosphorylations of Akt1 at Ser124 and PDPK1 at Ser241 were decreased significantly in the hippocampal tissues of PTPRT-null mice compared with their phosphorylation in wild-type mice (Fig. 8J,K). PTPRT-null NSCs also showed decreased phosphorylation of Akt1 at Ser473 and S6 at Ser240/244 compared with levels in wild-type NSCs (Fig. 8L). The expression of Ntrk2 was not altered in the hippocampal tissue or the NSCs of PTPRT-null mice compared with wild type (Fig. S8). Although the regulation mechanism is not clear so far, PTPRT appears to play pivotal roles in neurogenesis in the dentate gyrus and regulation of the proliferation and differentiation of hippocampal NSCs.

In this study, we found that defective PTPRT activity induced depression-like behaviors in PTPRT-null and D401A-mutant mice through a dysregulation in synaptic functions and aberrant hippocampal neurogenesis. PTPRT regulates the synaptic formation as well as the dendritic arborization of hippocampal neurons (Lim et al., 2009; Park et al., 2012; Lee, 2015). Several genetic studies have suggested that deficits of the PTPRT gene induce abnormal brain functions and neurological diseases. Cases of a point mutation and copy number gain of the PTPRT gene were suspected as pathogenic in a family and a child with intellectual disability, respectively (Schuurs-Hoeijmakers et al., 2013; Quintela et al., 2017). In addition, a genome-wide linkage analysis in a six-generation family has mapped the PTPRT gene within a region significant for autism spectrum disorder (ASD) (Allen-Brady et al., 2009). On the other hand, a clinical case study has reported that patients with depression display altered expression of the PTPRT gene in the temporal cortex (Aston et al., 2005). A GWAS analysis with a large sample size of patients with depression revealed that the PTPRT gene is included in the loci associated with a risk of major depression (Hyde et al., 2016). A recent study analyzed patients with early-onset major depression and reported that the PTPRT gene is included in regions of DNA methylation enrichment (Roberson-Nay et al., 2018 preprint).

PTPRT-null and D401A-mutant mice in this study displayed normal phenotypes on behavioral tests of social interaction, anxiety, learning and memory, but had enhanced depression-like behaviors (Figs 1, 3). In the hippocampus of mice subjected to chronic stress, a well-validated model of depression, PTPRT expression was reduced and inactive phospho-PTPRT expression increased (Fig. 2C,D). Basic and clinical studies have shown that depressive disorder is associated with attenuated neuronal synaptogenesis (Duman and Aghajanian, 2012). Here, hippocampal neurons of PTPRT-deficient mice displayed a decrease in excitatory and inhibitory synapses (Fig. 4A,B). NMDA receptor antagonists have been suggested to exert rapid antidepressant effects by enhancing AMPA receptor functions, and the US FDA have approved esketamine [S(+) enantiomer of ketamine] for patients with treatment-resistant depression (Maeng et al., 2008; Jaso et al., 2017). We found that neonatal PTPRT-null mice exhibited reduced expression levels of AMPA receptor GluR2, compared with levels in wild-type mice (Fig. 4C). Hippocampal neurons prepared from PTPRT-null mice also displayed decreased AMPA receptor levels (Fig. 4E,F). PTPRT effectively interacted with GluR2 and dephosphorylated tyrosine residues of GluR2 (Fig. 5C,F), and tyrosine phosphorylation of GluR2 was increased in the brains of PTPRT-deficient mice (Fig. 5G,H). These results suggest that GluR2 could be an endogenous substrate of PTPRT in the brain, and that the expression of GluR2 is regulated by PTPRT. A previous study has suggested that the trafficking of GluR2 is regulated by tyrosine phosphatases, but the authors did not specify which tyrosine phosphatases might be involved in this (Hayashi and Huganir, 2004; Moult et al., 2006). The distribution of GluR2 between the surface and internal side was altered by PTPRT knockdown, and PTPRT could regulate the membrane trafficking of GluR2 (Fig. 5I,J). Defective PTPRT activity induced dysfunctions of AMPA receptors and attenuated the formation of excitatory synapses in PTPRT-deficient mice. It has been shown previously that GluR2-null mice exhibit enhanced LTP similar to PTPRT-deficient mice (Jia et al., 1996), and here we found that dysregulated GluR2 in PTPRT-deficient mice augmented LTP (Fig. 6A,E). Although PTPRT-deficient mice exhibited altered hippocampal synaptic plasticity, they did not display reduced learning and memory in the behavioral tests (Fig. 1H–J; Fig. S3N–Q). A previous study reported that juvenile rats exposed to chronic stress have impaired recognition memory through the reduced expression of glutamate receptors and attenuated glutamatergic transmission in prefrontal cortex (PFC), but not in hippocampus (Yuen et al., 2012). The authors suggested that PFC has higher sensitivity to chronic stress than hippocampus regarding cognitive impairment. We observed that PTPRT expression was changed in hippocampus by chronic stress, but was not significantly changed in cortex (Fig. 2C). Our results suggest that the alteration of glutamatergic function in the hippocampus of PTPRT-deficient mice could induce depressive behaviors, but has little effect on working memory.

The transient knockdown of PTPRT in the dentate gyrus induced depression-like behaviors, but PTPRT knockdown in the CA1 region did not (Fig. 2F,G; Fig. S2). Augmented LTP was observed at the perforant–dentate gyrus synapse, but not at the Schaffer collateral–CA1 synapse, of PTPRT-deficient mice (Fig. 6). Inhibitory synaptic transmission was decreased in the dentate gyrus granule cells of PTPRT-deficient mice (Fig. 7B,D). The dentate gyrus is well known for generating new neurons during adulthood. Here, PTPRT-deficient mice displayed aberrant development of newborn granule cells in their dentate gyrus (Fig. 8B,C). Adult hippocampal neurogenesis has been evaluated as a candidate mechanism for the etiology of depressive disorder (Yun et al., 2016; Sheline et al., 1996; Bremner et al., 2000; Geuze et al., 2005). In an earlier study using an animal model of depression, hippocampal neurogenesis was found to buffer stress responses and ameliorate depressive behaviors (Snyder et al., 2011). Most antidepressants and antidepressant-like behavioral effects stimulate adult hippocampal neurogenesis (Boldrini et al., 2009; Sahay and Hen, 2007; David et al., 2009). It has also been suggested that adult neurogenesis could play a role in the structural plasticity of the mature brain by generating and adding newborn neurons into the existing brain circuitry (Eliwa et al., 2017; Bond et al., 2015). In a recent study, it was observed that young adult-born granule cells in the dentate gyrus activate local GABAergic interneurons to evoke strong inhibitory input into mature granule cells (Drew et al., 2016), suggesting that attenuated inhibitory synaptic transmission in the granule cells of PTPRT-deficient mice could be induced by aberrant neurogenesis in the dentate gyrus (Fig. 7B,D). It has been reported that signal pathways of neurotrophic factors control hippocampal neurogenesis and neuronal processes (Krishnan and Nestler, 2008). Clinical studies have shown that serum levels of BDNF are decreased in major depressive patients, and the administration of BDNF produces antidepressant-like activity through the modification of neurogenesis in an animal model of depression (Hashimoto et al., 2004; Siuciak et al., 1997; Vithlani et al., 2013). Chronic treatment with antidepressants increases BDNF expression and upregulates BDNF signaling that promotes adaptive neuronal plasticity (Nibuya et al., 1995; Russo-Neustadt et al., 2000; Alme et al., 2007). Here, PTPRT interacted with BDNF receptor Ntrk2 and signaling molecules in the BDNF pathway (Fig. 8G). In the hippocampal tissues and NSCs of PTPRT-null mice, the phosphorylation of Akt1, PDPK1 and S6 was attenuated (Fig. 8J–L). It has been shown previously that the activation of Akt1 by phosphoinositide 3-kinase (PI3K) promotes the differentiation of NSCs into GABAergic neurons (Chang et al., 2016; Oishi et al., 2009). Hippocampal NSCs of PTPRT-null mice displayed aberrant proliferation and multi-lineage differentiation (Fig. 8D,E). Attenuated hippocampal neurogenesis could thus induce dysregulation of inhibitory synapses in PTPRT-deficient mice.

Many clinical studies have reported that GABA concentrations are reduced in patients with depression, compared with concentrations in controls ( Luscher et al., 2011; Gabbay et al., 2012; Smiley et al., 2016; Romeo et al., 2018). On the other hand, it has been shown that the expression and function of GAD, the crucial enzyme producing GABA, is not changed in patients with depression (Pehrson and Sanchez, 2015). In the current study, PTPRT-deficient mice also displayed reduced hippocampal GABA concentrations, but not altered expressions of GAD and vGAT GABA transporter (Fig. 7E–G; Fig. S6). The expression of GAT3 GABA transporter was significantly decreased in the hippocampus of PTPRT-deficient mice (Fig. 7H). Earlier studies have shown that GAT3 is reduced in helpless-rat models of depression, whereas the expression levels of GAT1, GAD and vGAT do not change (Zink et al., 2009). PTPRT interacted with GAT3, and tyrosine phosphorylation of GAT3 was increased in PTPRT-null mice (Fig. 7J–M). The expression of GAT3 appeared to be regulated by PTPRT in a manner similar to GAT1, whose function is regulated through tyrosine dephosphorylation (Whitworth and Quick, 2001). In another recent study, the disinhibition of somatostatin-positive GABAergic interneurons was seen to alleviate the depressive behaviors of an animal model of depression (Fuchs et al., 2017). Many studies have suggested that enhancing GABAergic inhibitory synaptic inputs could be an effective strategy for antidepressant therapy. Recently, the US FDA announced allopregnanolone (also known as brexanolone), a potent positive allosteric modulator of GABA_A receptors, as the first-ever approved drug to treat postpartum depression (Kanes et al., 2017; Gee et al., 1995; Lambert et al., 2003; Reddy, 2010). Here, treatment with allopregnanolone rescued inhibitory synaptic transmission and ameliorated the depression-like behaviors of PTPRT-deficient mice (Fig. 7N,O). These results suggest that the attenuation of GABAergic inhibitory synaptic functions induced the depression-like behaviors of PTPRT-deficient mice.

In conclusion, we observed enhanced depression-like behaviors in PTPRT-deficient mice and examined the physiological roles of PTPRT that might be related to depressive disorder. PTPRT regulated the expression of AMPA receptors as well as membrane trafficking of GluR2 and, in addition, controlled GABAergic synaptic functions and neurogenesis in the dentate gyrus. Our results suggest that PTPRT deficiency could play a crucial role in the etiology of depressive disorder, and that PTPRT-deficient mice should be used for understanding the physiological mechanisms underlying depressive disorder.

This study was performed in accordance with the relevant guidelines under Korean law. All animal use was approved by the Animal Use and Care Committee of Korea Research Institute of Bioscience and Biotechnology (Permit Number: KRIBB-AEC-17152) and the Institutional Animal Care and Use Committee of Eulji University (Permit Number: EUIACUC 11-12).

PTPRT^−/− null mice were generated by targeting exon 22 of PTPRT, which encodes the first catalytic domain, as described previously (Zhao et al., 2010). PTPRT-D401A-mutant mice were generated by screening the N-ethyl-N-nitrosourea (ENU)-driven mutant mouse genomic DNA library and subsequent in vitro fertilization (Sakuraba et al., 2005). Since then, PTPRT-null and D401A-mutant mice were backcrossed for more than 10 generations onto the C57BL/6J genetic background. Animals were housed in a temperature-controlled (24°C) environment under a 12-h light/dark cycle (lights on at 7 am) with food and water ad libitum.

8-week-old male mice were used for behavioral tests.

The forced swim test (FST) was performed as previously described (Kim and Han, 2006). Briefly, each mouse was habituated to the test room and were placed in a transparent Plexiglas cylinder containing water for 6 min. The immobility time during the last 5 min was recorded and scored by a blinded investigator.

The tail suspension test (TST) was performed as previously described (Kim and Han, 2006). Briefly, after habituation, each mouse was suspended by the tail using adhesive tape for 6 min. The immobility time during the last 5 min was recorded and scored by a blinded investigator.

Mice were habituated to sucrose over a 3 day period by replacing water bottles with bottles containing sucrose solution (1%). Mice were given free access to two bottles, one with tap water and the other with 1% sucrose solution, after water deprivation of 16 h. After 1 h, the weights of bottles were measured for calculation of fluid consumption. Sucrose preference was obtained by calculating the sucrose intake as a percentage of total intake: ⁠.

Locomotor activity was assessed by measuring the distance of movement in an open field that was made of white Plexiglas (45×45×40 cm). Each mouse was habituated the test room for ∼30 min. After that, the mice were placed individually in the arena for 60 min. The horizontal locomotion of mice in an open field was measured using a computerized video tracking system, SMART (Panlab, Barcelona, Spain).

The rotarod test was performed as follows: for evaluation of coordination and motor learning the accelerating rotarod (Daejong Instrument Industry Co., Seoul, Korea) test was performed on three consecutive days. The mice were given three trials a day with an intertrial interval of 30 min. Acceleration speed from 4 to 40 rpm over a 5 min period was chosen. The latency to fall off was the measure of motor coordination and improvement across trials was the measure of motor learning. The cut-off time was set at 5 min.

The cylinder test was performed as previously described (Park et al., 2016). Sensorimotor function was evaluated using the cylinder test. Each mouse was placed into a transparent acrylic cylinder (diameter, 20 cm) and recorded for 5 min. The number of times that the right and left forelimbs came into contact with the wall were counted by a blinded observer.

The light-dark box test was performed as follows: the light box (20×20×20 cm) and dark box (10×20×20 cm) were made of white Foamex, with a shuttle door (5×7 cm) between the two at floor level. The light box was open at the top and illuminated with an 800 lux light at the bottom. The dark box was painted black and had a removable black lid. The test started by placing a mouse into the dark box, with the shuttle door closed initially. After 10 s, the shuttle door was opened. The total numbers of transitions between the dark and light compartments, crossings, and the time spent in the light box during a 5-min period were measured.

The elevated plus-maze (EPM) test was performed as follows: the EPM was made of gray Formax. The apparatus consisted of four arms (30×7 cm), which were elevated 50 cm above the floor and placed at right angles to each other. Two of the arms had 20-cm high walls (closed arms), while the other two had no walls (open arms). Each mouse was placed at the center of the platform and left to explore the arms for 5 min. The number of entries into the open and enclosed arms and the time spent in each arm and center area were recorded. Entry into each arm was scored as an event if the animal placed all paws into the corresponding arm.

Hyponeophagia refers to the inhibition of feeding produced by exposure to novelty. Mice were trained to drink sweetened condensed milk for three consecutive days. Mice were presented with diluted (1:3 ratio of milk:water) sweetened condensed milk (Carnation, NC, USA) for 30 min each day. Milk was presented in 10-ml serological pipettes with sippers attached with parafilm. Pipettes were closed with rubber stoppers and positioned through wire cage lids. In home-cage testing, the latency to drink and the volume consumed were recorded every 5 min for 30 min. Home-cage testing occurred under dim lighting (∼50 lx). Then, in novel-cage testing, mice were placed into new clean cages of the same dimensions but without shavings, with pipettes containing the milk positioned. Novel-cage testing occurred under bright lighting (∼1200 lx), with white paper placed under cages to enhance aversiveness.

Novel object recognition is a validated and widely used test for assessing recognition memory (Park et al., 2017). Mice were placed individually in a 40×20×20 cm testing chamber for 10 min with two identical objects (familiar, acquisition session). Mice were returned to home cages and placed back into the testing chamber 24 h later, in the presence of one of the original objects and one novel object (novel, recognition session) for 10 min. The original objects consisted of cylindrical wooden blocks, which had a height and diameter of 10 cm and 2 cm, respectively. The novel object was a rectangular wooden block that was 10×2.5×2 cm. The acquisition and recognition sessions were video recorded and a blinded observer scored the time spent exploring the objects and the number of times the objects were touched. Exploration was defined as sniffing and touching the object with the nose and/or forepaws. The duration and number of times spent exploring both objects were calculated. A discrimination index was calculated for each animal and expressed using the following formula, using values from day 2: ⁠.

The Y-maze alternation test was conducted as previously described (Park et al., 2017). A Y-shaped maze with three transparent arms of 50×10×12 cm was used. Each mouse was placed at the end of one arm facing the center and allowed to explore the maze for a period of 10 min. The sessions were video recorded and scored for entries into arms. The percentage of spontaneous alternation was calculated as the ratio of the actual to possible alternations (defined as the total number of arm entries minus 2), which was multiplied by 100(Conrad et al., 1996):

⁠.

The social interaction test was conducted as previously described (Felix-Ortiz and Tye, 2014). The social interaction chamber consisted of white acryl-walled box (40×20×20 cm). A single mouse was placed in the test box for 3 min for habituation. After that, a novel male mouse was introduced to the test cage and allowed to explore freely for 3 min. All behaviors were video recorded, and social interaction, such as body sniffing, anogenital sniffing and direct contact, was analyzed for 3 min.

The three-chamber social interaction test box composed of one center chamber and two side chambers (each chamber was 30×19 cm) (Sauer et al., 2015). Between each chamber, there was a door (6×6 cm). Stranger boxes consisted of wire cups with an 8.5 cm diameter. The test consisted of four steps. First, a mouse was habituated in the chamber for 10 min with the doors open. Next, the mouse was gently placed in the center chamber, and the doors were closed. A stranger box was then installed in each of the two side chambers. In the second step, the doors were then opened for 10 min; this was the social box habituation step. Then, the mouse was gently placed in the center chamber, the doors were closed and a same gender novel mouse (novel mouse 1) was placed in the novel box on one side. The doors were then opened for 10 min; this third step was the social interaction step. Finally, the doors were closed and a same gender novel mouse was placed in the empty chamber [novel mouse 1 (familiar) versus novel mouse 2 (novel)]. The doors were then opened for 10 min in the last step of the test, the social novelty step. The social interaction test was measured by using a computerized video tracking system, SMART (Panlab, Barcelona, Spain).

The Morris water maze (MWM) is 122 cm in diameter and is constructed of a stainless steel tank. A Perspex platform (14 cm in diameter and 49 cm in height) was submerged 1.5 cm below the surface of the water in the middle of a quadrant. White tempera paint was added to the water. The water temperature was maintained at 26°C. The pool was surrounded by curtains, and the mice underwent 8 days of training with four training trials of 90 s (inter-trial interval = 60 s) per day. The platform was consistently placed in the same spatial location of the pool throughout the acquisition period. During each trial, a mouse was released into the water facing the wall of the pool. After climbing the platform, the mouse was allowed to stay on it for 30 s. Recording and analysis were performed with an Ethovision video tracking system (Noldus Information Technology, the Netherlands). The escape latency was analyzed by measuring the time taken for the mouse to reach the platform during an acquisition trial.

To provide restraint, an 8-week-old mouse weighing 22–23 g was first placed in a ventilated 50 ml conical tube, plugged with a 3-cm-long intermediate tube and finally placed into a 50 ml tube (Chiba et al., 2012). The mouse could not move forward or backward on this device. This restraint stress was delivered to the animal for 2 h at the set time every day starting at 10:00 am. A control mouse remained in the original cage and remained intact in this home environment. After binding stress management, the arrested animals were put back together and returned to their normal home environment. This procedure was repeated for 14 days.

Immunoblotting was performed as previously described (Kim et al., 2016). Briefly, mice were sacrificed and the brain tissue was quickly removed and homogenized in a homogenization buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS and 0.1% sodium deoxycholate) containing protease inhibitor cocktail (Roche, Mannheim, Germany). NSCs were lysed with RIPA buffer, and 30 µg of NSC protein lysates was used for each analysis. Protein samples were resolved using SDS–PAGE and then transferred onto a polyvinylidene fluoride membrane (Bio-Rad, CA, USA). Blots were incubated with primary and secondary antibodies followed by visualization using an Enhanced Chemiluminescence kit (Atto Corp., Japan). Immunoblot images were quantified using Quantity One 1-D analysis version 4.6.1 software (Bio-Rad Laboratories, Inc., CA, USA) or ImageJ software (NIH, Bethesda, MD, USA).

Primary cultured hippocampal neurons were prepared from P1 mice brains as described previously (Cheon et al., 2017). Briefly, dissected hippocampi were dissected with trypsin and plated on coverslips coated with poly-L-lysine in Neurobasal medium (Thermo Fisher Scientific) supplemented with B27 (Thermo Fisher Scientific). After 2–3 h of incubation, the plating medium was changed with a growth medium (plating medium and glutamate). Cultured hippocampal neurons (day in vitro, DIV 7–8) were transfected by the calcium phosphate method. For immunofluorescence staining, after 7–9 days of transfection, cultured hippocampal neurons were fixed in 4% (v/v) formaldehyde and 4% (w/v) sucrose, and permeabilized with 0.2% (v/v) Triton X-100 in phosphate-buffered saline (PBS) followed by incubation with primary antibodies and fluorophore-conjugated secondary antibodies. Images captured by confocal microscopy (LSM 810, Zeiss) were analyzed blindly using MetaMorph software. The density of dendritic spine and synaptic protein clusters were measured from ∼27–32 dendrites of ∼7–9 neurons.

The mice were deeply anesthetized and transcardially perfused with saline followed by 4% paraformaldehyde in PBS. Brains were removed, postfixed overnight, and then cut into 40-μm coronal sections with a vibratome (Vibratome VT1200S, Leica Microsystems GmbH, Wetzlar, Germany). The sections were stained with 1% cresyl violet: sections were washed, dehydrated, cleared, mounted and coverslipped using mounting medium (Canada Balsam, Sigma-Aldrich). The slides were dried, after which photomicrographs were taken using a compound light microscope (Olympus, Japan) with 1.25× and 4× objectives (total magnification of 12.5× and 40×).

Full-length human PTPRT (NM_133170, amino acids 1-1460) was subcloned into GW1-CMV and the ectodomain of PTPRT was fused with human Fc domain at the C-terminus, as described previously (Lim et al., 2009). The PTPRT-D401A and PTPRT-D401A-ecto-Fc plasmids expressing D401A-mutated full-length PTPRT and D401A-mutated human-Fc-fused PTPRT ectodomain were generated using the QuickChange Site-directed Mutagenesis kit (Agilent Technologies, CA, USA).

Lentivirus containing shRNA for knockdown of PTPRT (GGAACCATGATAAGAATC), and AAV containing full-length PTPRT were produced using GFP-expression shuttle vectors (Lim et al., 2009). To achieve the expression of PTPRT shRNA and PTPRT in dentate gyrus, we bilaterally microinjected 1 μl of lentivirus or AAV at the dentate gyrus of three-week-old wild-type and knockout mice (−1.8 mm AP, ±1.5 mm ML, −2.0 mm DV). GFP-expressing neurons were shown in the dentate gyrus area (Fig. 2F,I). After five weeks, behavioral experiments were performed.

Three-week-old male mice were anaesthetized with intraperitoneal injections of ketamine and xylazine (61.8 mg/kg and 6.13 mg/kg, respectively) and mounted on stereotaxic apparatus (Stoelting Europe, Dublin, Ireland). Injections were made with a 10 µl syringe, connected to the injector by a polyethylene tubing, and controlled by an injection pump at 0.5 µl/min.

Cultured hippocampal neurons transfected with HA-tagged GluR1 and GluR2 (generously provided by Dr Sheng, Broad Institute of MIT and Harvard, USA) and with PTPRT-shRNA-pSup.gfp (Lim et al., 2009) were fixed with 4% (v/v) formaldehyde and 4% (w/v) sucrose for 8 min at room temperature, and incubated with mouse anti-HA antibody at 4°C overnight. The next day, neurons were washed and incubated with Cy3-conjugated anti-mouse secondary antibody. Then neurons were fixed with ice-cold methanol at −20°C for 1 m 30 s, and washed, followed by incubation with rabbit anti-HA antibody and guinea pig anti-GFP antibody at room temperature for 3 h. Neurons were then washed and incubated with Cy5-conjugated anti-rabbit and FITC-conjugated anti-guinea pig secondary antibodies. Images captured by confocal microscopy (LSM 810, Zeiss) were analyzed blindly using MetaMorph software.

5-bromo-2-deoxyuridine (BrdU; 150 mg/kg/day) was intraperitoneally injected for three consecutive days. After 2 days, mice were deeply anesthetized and transcardially perfused with saline followed by 4% paraformaldehyde in PBS. Brains were removed, postfixed overnight, and then cut into 40-μm coronal sections with a vibratome (Vibratome VT1200S, Leica Microsystems GmbH, Wetzlar, Germany). An immunohistochemical assay was performed with a primary anti-BrdU antibody (AbD Serotec, CA, USA).

Immunohistochemistry was performed as described previously (Go et al., 2018). Briefly, mice were deeply anesthetized and transcardially perfused with saline followed by 4% paraformaldehyde in PBS. Brains were removed, postfixed overnight, and then cut into 40-μm coronal sections with a vibratome (Vibratome VT1200S, Leica Microsystems GmbH, Wetzlar, Germany). Free-floating sections were incubated in PBS containing 3% H₂O₂ (v/v), rinsed three times in PBS and blocked with 5% horse serum or goat serum for 1 h at room temperature. Sections were incubated overnight at 4°C with primary antibody. After washing, the sections were incubated with biotinylated secondary anti-rabbit IgG (Vector Laboratories, Inc., Burlingame, CA, USA) followed by incubation with avidin-biotinylated peroxidase complex (ABC kit, Vector Laboratories, Inc.) and 3,3′-diaminobenzidine (Sigma-Aldrich Co. LLC). Immunofluorescence staining was performed with an Alexa Fluor 488 goat anti-rabbit IgG antibody (secondary antibody; 1:200; Life Technologies Corporation, Grand Island, NY, USA). Immunoreactive cells in the hippocampus were counted under a microscope (Olympus Corporation, Tokyo, Japan) by a blinded observer.

Neural stem cells (NSCs) were dissociated from the hippocampus of embryonic mouse brain (embryonic day ∼15–16) by trypsinization (Kim et al., 2018). Cells were plated into an ultra-low attachment culture dish (Corning, Corning, NY) for neurosphere formation and maintained with NSC medium containing DMEM/Neurobasal medium (Thermo Fisher Scientific), 1% N2, 2% B27 (Thermo Fisher Scientific), 20 ng/ml FGF (R&D systems, Minneapolis, MN) and 20 ng/ml EGF (Thermo Fisher Scientific).

To measure the proliferation of cultured NSCs, neurospheres were dissociated into single cells by trypsinization, and dissociated cells (1000 cells/well) were plated into ultra-low attachment 24-well plates. After 6–10 days, secondary neurospheres were trypsinized, and the dissociated cell number was counted. To measure the multi-lineage differentiation ability of NSCs, cultured neurospheres were dissociated by trypsinization and transferred into 24-well plates coated with 50 µg/ml poly-D-lysine (Sigma, St. Louis, MO) and 10 µg/ml laminin (Sigma). NSC differentiation was induced by incubating the cells with differentiation medium containing DMEM/F12 with N2 and B27 without FGF or EGF. After 5–6 days, cells were fixed with 4% paraformaldehyde solution. Differentiated neurons and astrocytes were visualized by immunofluorescence staining with anti-Tuj1 (Biolegend, San Diego, CA) and anti-GFAP (Thermo Fisher Scientific, Waltham, MA) antibodies, respectively. The length of neuronal axons and dendrites and neuronal arborization were measured using ImageJ software. Tuj1- and GFAP-positive cells were also counted using ImageJ software.

Mice were decapitated under deep enflurane anesthesia, and the brains were quickly removed and transferred to ice-cold dissection buffer containing sucrose (212.7 mM), KCl (2.6 mM), NaH₂PO₄ (1.23 mM), NaHCO₃ (26 mM), dextrose (10 mM), MgCl₂ (10 mM) and CaCl₂ (0.5 mM). Coronal whole brain sections (300 µm in thickness) prepared using a vibratome (Vibratome 1000 plus, The Vibratome Company, St Louis, MO, USA) were placed into dissection buffer that was continuously bubbled with 5% CO₂/95% O₂ (v/v). The hippocampal slices were mainly used for patch clamp. The slices were held at 35°C for 1 h in a chamber filled with continuously oxygenated artificial cerebrospinal fluid (ACSF) with the following composition: NaCl (124 mM), KCl (5 mM), NaH₂PO₄ (1.25 mM), NaHCO₃ (26 mM), dextrose (10 mM), MgCl₂ (1.5 mM) and CaCl₂ (2.5 mM). The slices were then transferred to an open submersion-type recording chamber, maintained at 35˚C, and perfused with oxygenated ACSF at a flow rate of 2 ml/min.

A bipolar stimulating electrode was inserted into Schaffer Collaterals or perforant path to activate CA1 pyramidal cells or dentate gyrus granule cells in the hippocampus, respectively. A glass micropipette filled with ACSF was inserted into CA1 stratum radiatum or the granule cell layer of dentate gyrus to record field potentials (FPs). FPs were evoked by stimulating the Schaffer Collaterals with an electrical pulse 0.2 ms in duration, delivered through concentric bipolar stimulating electrodes (FHC, Bowdoinham, ME). The initial slope of extracellular FPs was recorded in the CA1 stratum radiatum. Baseline responses were obtained upon application of 50% of the maximal stimulation at 0.033 Hz. Long-term potentiation (LTP) of CA1 was induced using a conventional stimulation paradigm; the theta burst stimulation (TBS) protocol consisted of eight bursts, each of four 100 Hz pulses, administered at 200 ms intervals. For LTP of dentate gyrus, four TBSs were applied in the presence of bicuculline. The stimulus intensity during TBS was identical to that of the test pulse. All measurements were expressed as percentages of the average values calculated 20 min prior to LTP induction. Significant differences between groups were sought via evaluation of average LTP values 58–60 min after LTP. Low-frequency stimulation consisting of 900 pulses at 1 Hz for 15 min was applied to induce long-term depression (LTD). To measure paired-pulse facilitation (PPF), we used inter-stimulus intervals (ISIs) of 25 ms, 50 ms, 100 ms, 200 ms, 400 ms, 1000 ms and 2000 ms.

The ventral hippocampal slices were perfused with ACSF at a ∼2 ml/min flow. Data was recorded at a holding potential of −70 mV using an axopatch 700B amplifier (Axon Instruments, Foster City, CA, USA), and ACSF temperature was maintained at 35°C. Current output of excitatory postsynaptic currents (EPSCs) was filtered at 1 kHz and digitized at 10 kHz. For recording of spontaneous EPSCs (s-EPSCs), the recording pipette was filled with internal solution containing 120 mM CsMeSO₃, 5 mM MgCl₂, 8 mM NaCl, 1 mM EGTA, 10 mM HEPES, 0.001 mM QX-314, 0.5 mM Na₃ GTP and 2 mM Mg ATP (pH 7.2–7.3, 280–290 mOsm), and 20 µM 1(s),9(R)-(−)-Bicuculline methiodide (Bicuculline; Sigma-Aldrich, St. Louis, MO, USA) was added to ACSF. To record the miniature EPSCs (mEPSCs), we added additional 1 µM tetrodotoxin (TTX; Tocris Bioscience, Ellisville, MO, USA) to ACSF. Current output of inhibitory postsynaptic currents (IPSCs) was filtered at 4 kHz and digitized at 10 kHz. For recording of spontaneous IPSCs (sIPSCs), the recording pipette was filled with internal solution containing 140 mM KCl, 0.5 mM CaCl₂, 20 mM HEPES, 5 mM EGTA and 5 mM Mg ATP (pH 7.2–7.3, 310–315 mOsm), and 50 µM DL-2-Amino-5-phosphonopentanoic acid (AP5; Sigma-Aldrich, St. Louis, MO, USA) and 20 µM 6-cyano-7nitroquinoxaline-2,3-dione (CNQX; Tocris Bioscience, Ellisville, MO, USA) were added to ACSF. To record the miniature IPSCs (mIPSCs), we added additional 1 µM TTX to ACSF. All experiments were executed in 3–5 MΩ resistance of patch electrodes and access resistance <25 MΩ. Data were acquired using Clampex software (Axon instruments, Union City, CA) at gain of 1 for EPSCs and IPSCs.

All of the animals were deeply anesthetized with isoflurane or sodium pentobarbital (80 mg/kg, intraperitoneal). The brain was rapidly removed from the cranium, and brain areas were dissected out on an ice-cold plate. The hippocampus was separated, weighed and placed in 1.5-ml microcentrifuge tubes. For glutamine, glutamate and GABA, the samples from both hemispheres were removed and pooled in 500 µl of deionized water. The samples were homogenized for 20 s with a sonic dismembrator (Thermo Fisher Scientific, USA) and centrifuged (Micromax RF/IEC, Thermo Fisher Scientific, USA) at 6000 g at 3°C for 20 min. Supernatants were filtered through 0.22 µm hydrophilic filters (Shleicher & Schuell, Germany), and 10 µl of brain filtrate was diluted in 990 µl of deionized water for amino acid analysis. Finally, samples were stored at approximately −20°C until analysis. An amino acid analyzer (AAA; HITACHI L-8900, Japan) used with autosampling injector 20 µl loop and an ultraviolet (UV) variable wavelength detector was used for monoamine assays where the samples were injected directly into a HITACHI HPLC Packed Post-column (#2622PF; 4.6×60 mm) packed with ion-exchange resin, flow rate 0.5 ml/min, UV 570 nm. The AAA utilizes a lithium citrate buffer system and analyses using the ninhydrin reaction method. The supernatant was obtained by centrifugation, diluted by tenfold using 0.02 N HCl, mixed with 5% (v/v) trichloroacetic acid at ratio of 1:4 in order to obtain free amino acids. This supernatant was filtered and injected into an ion-exchange column. Glutamate (Glu) and GABA were separated after 10 min. The resulting chromatogram identified each monoamine position and concentration from the sample as compared to that of the standard, and finally, the calculation of the content of each monoamine as µg/l in brain tissue was made according to established procedures (Pagel et al., 2000). Physiological solution determination could be carried out fully automatically in just 110 min.

Anti-phosphotyrosine antibody (4G10) agarose conjugate (Millipore, Germany) was used for immunoprecipitation to analyze the level of tyrosine phosphorylation of synaptic molecules. 4G10 agarose conjugate was incubated with synaptosome at 4°C for 4 h, and washed followed by boiling with SDS sample buffer. Protein samples were resolved using SDS–PAGE and immunoblotted with primary antibodies.

PTPRT expressed in heterologous cells was recruited by the Fc-fused PTPRT-ecto (RT-Fc) as described previously (Lim et al., 2009). Briefly, for trans interactions, purified RT-Fc pre-bound on protein A-beads was used for pulldown of PTPRT expressed in heterologous cells. For trans+cis interactions, purified RT-Fc was mixed with PTPRT expressed in heterologous cells, and then protein A-beads were added for pulldown. For cis interactions, RT-Fc was co-expressed with PTPRT in heterologous cells, and then a pulldown assay was performed with protein A-beads.

GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, USA) was used to perform all statistical analyses. Two-sample comparisons were conducted with Student's t-tests, while multiple comparisons were performed with a one-way analysis of variance (ANOVA) followed by Tukey post hoc test or a two-way ANOVA followed by Bonferroni post hoc tests. All results are presented as the mean±s.e.m. Differences with a P value less than 0.05 were considered to be statistically significant.

The primary antibodies used were mouse anti-PTPRT (Abfrontier, Cat# LF-MA0304; 1:2000), mouse anti-PSD-95 (Thermo Fisher Scientific, Cat# MA1-045; 1:200), rabbit anti-vGlut1 (SYSY, Cat# MA1-045; 1:500), mouse anti-gephyrin (SYSY, Cat# 147 111; 1:250), rabbit anti-vGAT (SYSY, Cat# 131 002; 1:250), rabbit anti-GAT1 (Abcam, Cat# ab426; 1:1000), rabbit anti-GAT3 (Abcam, Cat# ab181783; 1:2000), rabbit anti-Ntrk (Abcam, Cat# ab181560; 1:2000), rabbit anti-Ntrk1 (Cell Signaling Technology, Cat#2505; 1:500), rabbit anti-Ntrk2 (Abcam, Cat# ab18987; 1:1000), rabbit anti-Ntrk2 (ecto) (Abcam, Cat#ab33655; 1:1000), rabbit anti-p75NTR (Abcam, Cat# ab52987; 1:2000), rabbit anti-PIK3CA (Cell Signaling Technology, Cat# #4249; 1:1000), rabbit anti-PDPK1 (Cell Signaling Technology, Cat# 3062; 1:1000), rabbit anti-phospho-PDPK1 (Ser241) (Cell Signaling Technology, Cat# 3438; 1:1000), rabbit anti-Akt (Cell Signaling Technology, Cat# 9272; 1:2000), rabbit anti-phospho-Akt1 (Ser124) (Abcam, Cat# ab183556; 1:1000), rabbit anti-phospho-Akt1 (Ser473) (Cell Signaling Technology, Cat# 4060; 1:1000); rabbit anti-GAPDH (Cell Signaling Technology, Cat# 2118; 1:1000), mouse anti-S6 ribosomal protein (Cell Signaling Technology, Cat# 2317; 1:1000), rabbit anti-phospho-S6 ribosomal protein (Ser240/244) (Cell Signaling Technology, Cat# 5364; 1:1000), goat anti-doublecortin (Santa Cruz Biotechnology, Cat# sc-8666; 1:1000), rat anti-5-bromo-2-deoxyuridine (BrdU) (AbD Serotec, Cat# MCA2060GA; 1:1000), rabbit anti-calretinin (Millipore, Cat# ab22683; 1:2000), mouse anti-calreticulin (Abcam, Cat# ab22683; 1:200), rabbit anti-calbindin (Abcam, Cat# ab11426; 1:3000), mouse anti-parvalbumin (Millipore, Cat# MAB1572; 1:2000), rabbit anti-somatostatin (Thermo Fisher Scientific, Cat# PA5-16253; 1:200), rabbit anti-neuropeptide Y (NPY) (Abcam, Cat# ab30914; 1:200), rabbit anti-cholecystokinin (CCK) (Sigma-Aldrich, Cat# SAB1402711; 1:200), mouse anti-Tuj1 (Biolegend, Cat# 801201; 1:1000), rabbit anti-GluR1 and anti-GluR2 (generously provided by Professor Eujoon Kim, KAIST, Korea; 1:2000), mouse anti-HA (Roche, Cat# 11 583 816 001; 1:200), rabbit anti-HA (Santa Cruz Biotechnology, Cat# sc-805; 1:200) and rat anti-GFAP (Thermo Fisher Scientific, Cat# 13-0300; 1:200). The Cy3-, Cy5-, FITC- and HRP-conjugated secondary antibodies were obtained from Jackson ImmunoResearch. Biotinylated secondary anti-rabbit IgG were from Vector Laboratories, Inc.

Imipramine hydrochloride (20 mg/kg, intraperitoneal) was purchased from Sigma-Aldrich Co. LLC (St Louis, MO, USA) and diluted in saline. Allopregnanolone (3α,5α-tetrahydroprogesterone, 3α,5α-THP; 10 mg/kg, intraperitoneal) was purchased from Calbiochem (Cat# 127100), dissolved in ethanol and diluted in saline.

For all experiments, mice, cultured neurons and NSCs were randomly allocated to experimental groups, and all data collected throughout these studies were included in the analysis. No data were excluded. No sample-size estimates were conducted due to technical limitations on sample collection. All attempts were made to use maximal sample size in each experiment whenever possible.

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.243972.reviewer-comments.pdf