Neuronal Activity Alters Bdnf–trkb Signaling Kinetics And

Transcript

ß 2014. Published by The Company of Biologists Ltd | Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 RESEARCH ARTICLE Neuronal activity alters BDNF–TrkB signaling kinetics and downstream functions ABSTRACT Differential kinetics of the same signaling pathway might elicit different cellular outcomes. Here, we show that high-frequency neuronal activity converts BDNF-induced TrkB (also known as NTRK2) signaling from a transient to a sustained mode. A prior depolarization (15 mM KCl, 1 hour) resulted in a long-lasting (.24 hours) activation of the TrkB receptor and its downstream signaling, which otherwise lasts less than an hour. The long-term potentiation (LTP)-inducing theta-burst stimulation but not the long-term depression (LTD)-inducing lowfrequency stimulation also induced sustained activation of TrkB. This sustained signaling facilitated dendritic branching and rescued neuronal apoptosis induced by glutamate. The change in TrkB signaling kinetics is mediated by Ca2+ elevation and CaMKII activation, leading to an increase in TrkB expression on the neuronal surface. Physical exercise also alters the kinetics of TrkB phosphorylation induced by exogenous BDNF. Sustained TrkB signaling might serve as a key mechanism underlying the synergistic effects of neuronal activity and BDNF. KEY WORDS: Signal transduction, Signaling kinetics, Neurotrophin, Neuronal activity, Neuronal surface receptor INTRODUCTION Brain-derived neurotrophic factor (BDNF), a member of the neurotrophin family, has multiple functions including neuronal survival, neuronal development and synaptic plasticity in the central nervous system (Huang and Reichardt, 2001; Lewin and Barde, 1996; Lu and Figurov, 1997; McAllister et al., 1999). For example, BDNF promotes dendritic outgrowth and branching, increases the density of dendritic spines and modulates spine morphological specializations (McAllister et al., 1995; Shimada et al., 1998). Many studies also suggest that BDNF plays a crucial role in hippocampal long-term potentiation (LTP) and long-term memory (Lu et al., 2008). Moreover, activation of BDNF signaling pathways has been considered as a disease-modifying strategy for neurological and psychiatric diseases (Martinowich et al., 2007; Zuccato and Cattaneo, 2009; Nagahara and Tuszynski, 2011; Lu et al., 2013). 1 Tsinghua-Peking Center for Life Sciences, Beijing, China. 2School of Medicine, Tsinghua University, 1 Qinghuayuan Road, Beijing, 100084, China. 3School of Life Sciences, Tsinghua University, 1 Qinghuayuan Road, Beijing, 100084, China. 4 GlaxoSmithKline, R&D China, Building 3, 898 Halei Road, Zhangjiang Hi-tech Park, Pudong, Shanghai, 201203, China. 5National Institute of Biological Sciences, Beijing, 102206, China. 6School of Life Sciences, Peking University, Beijing, 100871, China. *Author for correspondence ([email protected]) Received 4 August 2013; Accepted 11 February 2014 The interplay between neuronal activity and BDNF is a subject of broad interest. On one hand, the effects of BDNF on neuronal activity and synaptic transmission have been observed in cultured neurons and in slices by many laboratories. For example, the application of BDNF to neonatal hippocampal slices facilitates LTP, and this is mediated by an enhancement of synaptic response to tetanic stimulation (Figurov et al., 1996). On the other hand, the effects of BDNF are modulated by neuronal activity. For example, blockade of neuronal activity and synaptic transmission prevents the increase in dendritic arborization induced by BDNF (McAllister et al., 1996). Presynaptic depolarization greatly facilitates the BDNF-induced synaptic transmission at the neuromuscular junction of Xenopus (Boulanger and Poo, 1999). BDNF effectively regulates repetitive excitatory synaptic responses only when the synapses are stimulated at 100 Hz or higher, which induces severe synaptic fatigue (Gottschalk et al., 1998). It remains ambiguous how neuronal activity modulates BDNF– TrkB signaling kinetics and downstream functions. BDNF binds to TrkB (also known as NTRK2), its high-affinity receptor (Kaplan and Stephens, 1994; Squinto et al., 1991), leading to the activation of the downstream MAPK, PI3K and PLCc pathways. After its activation, the TrkB receptor undergoes endocytosis, recycling or degradation (Haapasalo et al., 2002; Ji et al., 2010; Sommerfeld et al., 2000). In cultured hippocampal neurons, the amount of TrkB receptor on the neuronal surface increases after a 15-second BDNF treatment and then declines rapidly (Haapasalo et al., 2002), whereas the total TrkB expression remains unchanged in the first few hours, followed by a reduction after several days through transcriptional regulation (Frank et al., 1996). It has been shown that the insertion and endocytosis of cell surface TrkB is acutely regulated by neuronal activity (MeyerFranke et al., 1998; Du et al., 2000; Du et al., 2003). In the present study, we asked whether and how neuronal activity alters the kinetics of BDNF signaling. Surprisingly, we found that neuronal stimulation converts transient TrkB activation and its downstream signals to a sustained mode of activation, leading to an enhancement of dendritic branching and attenuation of neuronal apoptosis induced by glutamate. This sustained TrkB activity is mediated by prolonged expression of surface TrkB induced by neuronal activity. Physical exercise known to enhance neuronal activity can also induce sustained TrkB signaling in vivo. These findings might provide new insights into BDNF signaling and function in vivo. RESULTS Priming by depolarization converts TrkB signaling from a transient to a sustained mode A number of methods, including high K+ concentrations and electrical stimulation, have been used to enhance neuronal activity (Boulanger and Poo, 1999; Du et al., 2000; Meyer-Franke 2249 Journal of Cell Science Wei Guo1,2,3, Yuanyuan Ji4, Shudan Wang1,3, Yun Sun1,5,6 and Bai Lu1,2,* et al., 1998). We used 15 mM KCl to depolarize cultured hippocampal neurons. Similar to previous reports (Grubb and Burrone, 2010; Tongiorgi et al., 1997), our Ca2+-imaging experiments showed that 15 mM KCl induced a sustained neuronal activation (supplementary material Fig. S1). Primary cultured hippocampal neurons [12–14 days in vitro (DIV)] were pretreated with 15 mM KCl or 15 mM NaCl (as a control for ionic strength) for 1 hour, followed by application of 1 nM BDNF. Proteins were collected at different time-points for quantitative western blotting. Acute application of BDNF (1 nM, 25 ng/ml) triggered a robust but transient increase in TrkB phosphorylation; the level of phosphorylated TrkB (pTrkB) reached its peak at about 15 minutes but declined rapidly and approached baseline by 4 hours after BDNF application (Ji et al., 2010). Interestingly, the same concentration of BDNF (1 nM) elicited a sustained TrkB phosphorylation when neurons were pretreated with 15 mM KCl for 1 hour (Fig. 1A, left). The levels of pTrkB at 2-hour and 4-hour time-points after BDNF application remained at ,80% of the maximum for the 15 mM KCl group, but declined to baseline for the 15 mM NaCl group (Fig. 1B). In absence of BDNF, TrkB phosphorylation was undetectable with 15 mM KCl alone (supplementary material Fig. S2A). Prior exposure of neurons to 15 mM NaCl (as a control for osmotic conditions and Cl2 concentration) resulted in a similar transient pattern of BDNF-induced TrkB phosphorylation (Fig. 1A, right). Depolarization-induced sustained TrkB signaling lasted as long as 24 hours, the longest time-point examined. Compared with the levels of pTrkB at the 15-minute time-point, the relative levels of pTrkB at the 8-hour and 24-hour time-points after BDNF application were ,100% and ,50%, respectively, for the 15 mM KCl group, but were only ,20% at the same time-points for the 15 mM NaCl group (Fig. 1C,D). The results suggest that priming with neuronal depolarization could convert BDNF-induced TrkB activation from a transient to a sustained mode. In subsequent Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 experiments, we focused on kinetic changes in signaling within the first 4 hours, which exhibit a clear difference between these two modes. Next, we examined whether the sustained TrkB activation induced by high K+ could lead to the same sustained activation of the three major signaling pathways downstream of TrkB – MAPK, PI3K and PLCc. In hippocampal neurons, BDNFinduced activation of the MAPK pathway was detectable by using an antibody against phosphorylated Erk (pErk) on western blots. A sustained pattern of Erk phosphorylation was evident in cells pretreated with 15 mM KCl followed by BDNF application (Fig. 2). For comparison, in the presence of 15 mM NaCl, BDNF induced a marked increase in pErk, which peaked around 15 minutes, followed by a rapid decline to basal levels within the next 4 hours (Fig. 2). For the PI3K and PLCc pathways, we examined the phosphorylation of Akt and PLCc1 (pAkt and pPLC-c1). Similarly, sustained increases in pAkt and pPLC-c1 were induced by BDNF in the presence of high K+, whereas transient ones were induced in the presence of NaCl (Fig. 2). We then examined the phosphorylation of CREB, the transcription factor downstream of MAPK that is known to mediate BDNF-induced gene expression and synaptic modulation (Finkbeiner et al., 1997). As expected, BDNF triggered a sustained phosphorylation of CREB with 15 mM KCl pretreatment (supplementary material Fig. S3). It should be noted that treatment with 15 mM KCl alone (but not NaCl alone) could induce CREB activation, as shown at the ‘0’ timepoint for the KCl+BDNF group. This is because high K+ can also increase phosphorylation of CREB through Ca2+ influx (Bito et al., 1996). However, high K+ alone induced only transient activation of CREB (Sala et al., 2000). Taken together, these results suggest that neuronal activity (depolarization) could modulate the kinetics of BDNF-induced TrkB activation and its downstream signaling. Fig. 1. Depolarization by high K+ converts BDNF-induced TrkB activation from a transient to a sustained signaling mode. (A,B) Cultured neurons were treated with 15 mM NaCl or 15 mM KCl for 1 hour, followed by BDNF application (1 nM). Cells were washed, lysed and collected at different time-points (from 0 to 4 hours) after BDNF application. Representative western blots (A) and quantitative plots (B) are presented. (C,D) TrkB phosphorylation at long-term time-points (8 hours and 24 hours) in the presence or absence of high K+. Representative western blots (C) and quantitative plots (D) are presented. Ctrl, samples without any treatment. The pTrkB signals were measured by densitometry and normalized to total TrkB. In B, the dotted line indicates pTrkB level before BDNF application. For B and D, data are presented as the mean6s.e.m.; n54; *P,0.05, **P,0.01, ***P,0.001 (paired t-test). 2250 Journal of Cell Science RESEARCH ARTICLE RESEARCH ARTICLE Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 Fig. 2. Kinetics of BDNF-induced Erk, Akt and PLCc activation in the presence of 15 mM NaCl or 15 mM KCl. Representative western blots (A) and quantitative plots (B) are shown. The phosphorylation signals were measured by densitometry and normalized to the corresponding total protein. Dotted line, the level of phosphorylated protein before BDNF application. Data are presented as the mean6s.e.m.; n54; *P,0.05, **P,0.01 (paired t-test). Previously, we reported that neurite branching is regulated by sustained activation of TrkB induced by gradual stimulation with BDNF (Ji et al., 2010). Here, we found that sustained TrkB activation by depolarization coupled with BDNF stimulation also facilitated neurite branching (Fig. 3A). Hippocampal neurons (DIV3) treated with BDNF either in the presence of 15 mM NaCl or 15 mM KCl for 3 days were analyzed for dendritic complexity after immunostaining with an anti-MAP2 antibody, and dendritic morphology was analyzed. Treatment with BDNF alone increased both the number of primary neurites and the number of branch points (Fig. 3A, P,0.01, ANOVA). Treatment with high K+ alone had no effect on either parameter. However, high K+ coupled with BDNF selectively facilitated dendritic branching, without affecting the number of primary dendrites (Fig. 3A). The number of branch points (3.3560.17 per cell, 6s.e.m.) in neurons treated with BDNF+15 mM KCl was significantly higher than that in neurons treated with either 15 mM KCl alone (2.0560.14) or BDNF+15 mM NaCl (2.4660.16) (P,0.001, ANOVA). Neuronal depolarization therefore augmented the effect of BDNF on dendritic branching. BDNF has been shown to promote neuronal survival in vitro (Arancibia et al., 2008; Mattson et al., 1995). Here, we examined whether depolarization alters the neuronal survival function of BDNF in a model of glutamate-induced neuronal excitotoxicity (Almeida et al., 2005; Mattson et al., 1995). Hippocampal neurons were treated with 15 mM NaCl, 15 mM KCl or control medium, followed by incubation with BDNF (1 nM) for 6 hours. Subsequently, the cultures were treated with either vehicle or glutamate (125 mM) for 15 minutes and then returned to the original medium. After a further 7 hours, neuronal apoptosis, as reflected by intracellular ATP level, was measured (Fig. 3B). Application of glutamate generally reduced cell survival by ,50%. In either control cultures or cultures pretreated with NaCl, BDNF failed to reverse glutamate-induced cell death. By contrast, prior neuronal depolarization by exposure to high K+ significantly attenuated the toxic effect of glutamate (P,0.05, ANOVA), suggesting that neuronal depolarization also augments the neuroprotective function of BDNF. Electrical stimulation sustains BDNF-induced TrkB activation To determine the specific patterns of neuronal activity that are crucial for converting transient to sustained BDNF-induced TrkB activation, we applied electric-field stimulation to cultured hippocampal neurons using a protocol previously established to reliably induce action potentials in neurons throughout the culture dishes (Du et al., 2000; Nagappan et al., 2009). Cultured hippocampal neurons were electrically stimulated by platinum wires attached to the lids of culture dishes, using the LTP-inducing theta-burst stimulation (TBS, 20 pulses per 5 seconds) protocol for 60 minutes in Tyrode’s solution at room temperature. At the end of field stimulation, the cultures were treated with BDNF and processed for pTrkB measurements at different time-points (from 0 to 4 hours). We found that, similar to high K+ stimulation, field stimulation with TBS converted BDNF-induced TrkB phosphorylation from a transient to a sustained mode (Fig. 4A,B). In the absence of BDNF, TrkB phosphorylation was undetectable with TBS alone (supplementary material Fig. S2D). As a control, we employed another stimulation protocol, which has been often used to induce long-term depression (LTD). Remarkably, the LTDinducing stimulation (LIS, 15 minutes, 1 Hz) paradigm failed to alter the kinetics of TrkB phosphorylation (Fig. 4C,D). 2251 Journal of Cell Science High K+ facilitates BDNF regulation of survival and neurite branching RESEARCH ARTICLE Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 To verify whether sustained activation of TrkB is selectively induced by LTP-inducing stimuli, and to control the number of stimulation pulses, we employed two additional stimulation patterns – a classical LTP-inducing tetanic stimulation (100 Hz, 1 second, given every 10 minutes, 300 pulses in 30 minutes) and a sustained low-frequency stimulation with the same number of pulses (LFS, 300 pulses in 30 minutes). We found that tetanic stimulation could sustain BDNF-induced TrkB phosphorylation (Fig. 4E,F). By contrast, cultures exposed to LFS exhibited an almost identical transient pattern of TrkB phosphorylation compared with the unstimulated control (Fig. 4G,H). Activity-regulated pTrkB depends on intracellular Ca2+ elevation and CaMKII activation High K+ or electrical stimulation leads to Ca2+ influx and activation of CaMKII (also known as CAMK2G) (Berridge, 1998; Lisman et al., 2002; Obrietan et al., 2002). To delineate the molecular mechanisms involved in the activity-dependent regulation of sustained TrkB signaling, hippocampal neurons were pretreated with specific pharmacological inhibitors for 30 minutes before treatment with high K+ and BDNF. Pretreatment with the NMDA-receptor blocker MK801 (50 mM) significantly reduced the levels of pTrkB at 2-hour and 4hour time-points (Fig. 5A–C). Pretreatment with KN-93 (20 mM), a CaMKII inhibitor, also attenuated sustained TrkB phosphorylation (Fig. 5A–C). Because KN-93 might inhibit other Ca2+–calmodulin-dependent (CaM) kinases, we used 2252 autocamtide-2-related inhibitory peptide (AIP), a peptide known to selectively inhibit CaMKII (Du et al., 2000; Ishida et al., 1995). Hippocampal neurons were pretreated with 20 mM AIP for 3 hours, followed by treatment with 15 mM KCl for 1 hour, and then BDNF for 2 or 4 hours. Specific blockade of CaMKII by AIP reversed the sustained TrkB activation (supplementary material Fig. S4A–C). By contrast, inhibition of the MAP kinase pathway by using U0126 (20 mM) failed to alter the sustained kinetics of TrkB phosphorylation (Fig. 5A–C). To determine whether high-K+-mediated sustained activation of TrkB required the kinase activity of TrkB, cultured hippocampal neurons were treated with K252a (100 nM), a pan-Trk kinase inhibitor. Western blot analysis showed that BDNF-induced TrkB phosphorylation was completely blocked by K252a (Fig. 5A–C). To validate that it is neuronal activity, not the effect of high K+ per se, that confers sustained TrkB activation, we examined whether similar mechanisms are also involved in TBS-induced conversion of transient to sustained TrkB phosphorylation. Hippocampal neurons were pretreated with various inhibitors for 30 minutes, followed by TBS for 1 hour and then BDNF application. Similar to high K+ stimulation, sustained TrkB phosphorylation by TBS was blocked by the NMDA-receptor inhibitor MK801 and the CaMKII inhibitor KN-93, but not by the MAP kinase inhibitor U0126 (Fig. 5D–F). K252a also completely abolished TrkB phosphorylation. Taken together, these data suggest that Ca2+ influx through the NMDA receptor Journal of Cell Science Fig. 3. Effect of high K+ on BDNF-dependent regulation of dendritic growth and neuronal survival. (A) High K+ increases BDNF-induced dendritic branching of hippocampal neurons. Neurons (DIV 3) were treated with BDNF in the presence of 15 mM NaCl or KCl. After 3 days, neurons were fixed and stained with anti-MAP2 antibody. Representative images of MAP2-stained hippocampal neurons under different conditions are shown above the quantitative data. Ctrl, control. n.120, from three independent experiments. Data are presented as the mean6s.e.m. The complete ANOVA analysis is shown in supplementary material Table S1. (B) Attenuation of glutamate-induced toxicity by BDNF is facilitated by neuronal depolarization. Cultured hippocampal neurons were pretreated with either 15 mM NaCl or 15 mM KCl, and then BDNF (25 ng/ml) for 6 hours. The cultures were exposed to vehicle or glutamate (125 mM) for 15 minutes and then the medium was replaced with neurobasal medium with or without BDNF. Neuronal survival was determined by measuring intracellular ATP 7 hours later. BDNF significantly attenuated the toxicity in the high K+ group but not in the NaCl group. Three independent experiments were performed, and the same treatment was repeated six times per experiment. Data are presented as the mean6s.e.m.; n.s., nonsignificant; *P,0.05 (ANOVA). RESEARCH ARTICLE Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 and CaMKII activity are necessary for the sustained TrkB phosphorylation that is regulated by neuronal activity. It seemed that the inhibitors of the NMDA receptor and CaMKII only partially blocked the effect of neuronal activity. We next examined some other factors using a series of inhibitors. CdCl2 (a general Ca2+ channel blocker, 0.2 mM) almost completely blocked high-K+-induced sustained TrkB activation at 4 hours, and nimodipine (Nimo, a potent L-type Ca2+-channel antagonist, 10 mM) resulted in partial blockade thereof (Fig. 5G,H). CNQX (an AMPAR inhibitor, 50 mM) had no statistically significant effect on TrkB activation at 4 hours (Fig. 5G,H). Tetrodotoxin (TTX, a sodium channel blocker, 1 mM) had no effect on high-K+-induced sustained pTrkB, but interfered with TBS-induced sustained pTrkB (supplementary material Fig. S2C,E). Therefore, the sustained TrkB activation depends on the Ca2+ elevation induced by neuronal activation. Increased surface TrkB expression mediates sustained TrkB signaling Neuronal activity has been shown to increase TrkB insertion into the neuronal cell membrane in the absence of BDNF stimulation (Du et al., 2000; Meyer-Franke et al., 1998; Zhao et al., 2009). It 2253 Journal of Cell Science Fig. 4. TBS or tetanic stimulation sustains BDNF-induced TrkB phosphorylation. (A–D) Cultured hippocampal neurons were stimulated by two different stimulation patterns, TBS or LIS, for 1 hour. The neurons were then treated with 1 nM BDNF and proteins were collected at different time-points. Representative western blots (A,C) and quantitative plots (B,D) are presented. (E–H) Cultured hippocampal neurons were stimulated by tetanic stimulation or LFS. Then the neurons were treated with 1 nM BDNF and proteins were collected at different time-points (from 0 to 4 hours). Representative western blots (E,G) and quantitative plots (F,H) are presented. The dotted line indicates the pTrkB level before BDNF application. Data are presented as the mean6s.e.m.; n54; *P,0.05, **P,0.01 (paired t-test). RESEARCH ARTICLE Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 was unclear whether high-frequency activity could increase TrkB insertion and recycling, and whether such a mechanism could underlie the sustained TrkB activation induced by the combination of neuronal activity and BDNF. To examine this possibility, we performed surface biotinylation experiments using samples at different time-points after BDNF application. Cell surface proteins were biotinylated at the end of each time-point under non-permeable conditions, pulled down with streptavidin and analyzed by using western blotting. In general, the application of BDNF induced a marked decrease in the level of cell-surface-associated TrkB due to ligand-induced receptor endocytosis. When neurons were depolarized by 15 mM KCl before BDNF treatment, the surface levels of TrkB remained as stable as those of a control group, even at the 2- or 4-hour time point. (Fig. 6A,B). However, the surface levels of TrkB declined to 43%66.8% (6s.e.m.) of control (0 time-point) at 2 hours and 29%66.5% at 4 hours after BDNF application in the group treated with 15 mM NaCl. A 15-second BDNF treatment (the 2254 positive control) rapidly increased the surface levels of TrkB (Fig. 6A,B) as previously reported (Haapasalo et al., 2002). Treatment with 15 mM KCl alone induced virtually no increase in surface TrkB expression (supplementary material Fig. S4D,E). We then investigated whether a similar mechanism underlies TBS-induced sustained TrkB activation. Cultured hippocampal neurons were stimulated with TBS or LIS, or not stimulated, and then treated with BDNF. Biotinylation experiments were performed to measure the surface levels of TrkB at different time-points. BDNF induced a reduction in surface TrkB levels, to 51%616% of pretreatment levels at 2 hours and 37%61.8% at 4 hours. Prior treatment with TBS maintained surface TrkB expression after 2 or 4 hours of BDNF treatment (Fig. 6C,D). Similar to the high K+ experiment, TBS increased surface TrkB without altering total TrkB levels in these neurons (Fig. 6A,B). At 4 hours after BDNF treatment, surface TrkB in the TBS group recovered from a slight decline that was evident at the 2-hour time-point and increased to a much higher level. However, in the Journal of Cell Science Fig. 5. The role of glutamate receptors, CaMKII and Ca2+ channels in activity-dependent BDNF-induced sustained TrkB activation. (A–C) Hippocampal neurons were pretreated with 100 nM K252a, 50 mM MK801, 20 mM KN-93 or 20 mM U0126 for 30 minutes before incubation with KCl (15 mM, 1 hour) and then 1 nM BDNF. Representative western blots (A) and quantitative plots of the 2-hour time-point (B) and 4-hour time-point (C) are presented. Note that the pTrkB level at the 2-hour or 4-hour time-point in each condition was compared with that of the DMSO + 15 mM KCl + 1 nM BDNF group. (D–F) Hippocampal neurons were pretreated with 100 nM K252a, 50 mM MK801, 20 mM KN-93 or 20 mM U0126 for 30 minutes before TBS stimulation and then 1 nM BDNF treatment. Representative western blots (D) and quantitative plots of the 2-hour time-point (E) and 4-hour time-point (F) are presented. Note that the pTrkB level at the 2-hour or 4-hour time-point in each condition was compared with that of the DMSO + TBS + 1 nM BDNF group. (G,H) Hippocampal neurons were pretreated with the Ca2+-channel blockers CdCl2, nimodipine or the AMPAR antagonist CNQX for 30 minutes before stimulation with 15 mM KCl (1 hour). Then 1 nM BDNF was added to the cultured neurons, which were collected 4 hours later for pTrkB measurement. Representative western blots (G) and quantitative plots (H) are presented. Ctrl, control. Data show the mean6s.e.m.; n54; *P,0.05, **P,0.01, ***P,0.001 (ANOVA). RESEARCH ARTICLE Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 LIS group, the surface TrkB level declined to 44%69.0% of control (0 time-point of the nonstimulated group) after 4 hours of BDNF treatment. Lastly, we examined whether the NMDA receptor and CaMKII are involved in maintaining the increased surface TrkB expression induced by neuronal activity. Hippocampal neurons were pretreated with various inhibitors for 30 minutes, then treated with 15 mM KCl or TBS for 1 hour, followed by BDNF application. Biotinylation experiments were performed to measure surface levels of TrkB at the 4-hour time-point. Increased surface TrkB expression was blocked by the NMDAreceptor inhibitor MK801 and the CaMKII inhibitor KN-93, but not by the MAP kinase inhibitor U0126 (Fig. 7). Thus, sustained TrkB activation could be explained, at least in part, by an increase in the surface expression of TrkB induced by neuronal activity, Ca2+ influx through the NMDA receptor and activation of CaMKII. Running prolongs hippocampal TrkB signaling in vivo We also investigated whether an increase in neuronal activity could lead to sustained BDNF signaling in vivo. It has been shown that exercise dramatically enhances neuronal activity in the hippocampus (Vanderwolf, 1969; Kuo et al., 2011; Oladehin and Waters, 2001). Here, we used a treadmill to enforce running and examined whether it induced sustained TrkB phosphorylation when BDNF was injected into the hippocampus after running. The hippocampal CA1 area was selected to study the effect of running on TrkB signaling in vivo, because exercise elicits the strongest effects on theta activity and c-fos induction in this area (Buzsa´ki, 2002; Oladehin and Waters, 2001). At 1 week before the test, rats were implanted with guide cannulas into both the right and left dorsal hippocampal CA1 areas. After 0.5 hours of running on the treadmill and 0.5 hours of rest, rats were anesthetized briefly (for ,2–3 minutes) by using isoflurane, and 0.2 mg of BDNF was injected into the right, but not the left, hippocampus. At 1-hour and 4-hour time-points, ,1 mg of tissue around the injection site was dissected out. We found that running prolonged BDNF-induced TrkB phosphorylation. There was a very low level of TrkB phosphorylation in the left PBS-injected hippocampus, possibly induced by a low level of endogenous constitutively secreted BDNF, and running elicited a small 2255 Journal of Cell Science Fig. 6. Neuronal activity attenuates BDNF-induced downregulation of TrkB expression on the neuronal surface. Hippocampal neurons were treated with 15 mM KCl, 15 mM NaCl, TBS or LIS for 1 hour before application of 1 nM BDNF. Cell surface proteins were biotinylated at different time-points after BDNF application, and collected using immobilized streptavidin. Equal amounts of biotinylated proteins were loaded to analyze expression of TrkB at the cell surface. Representative western blots (A,C) and quantitative plots (B,D) are presented. Note that surface TrkB expression in each condition was compared with that of the control (0 time-point of 1 nM BDNF treatment). Data show the mean6s.e.m.; n54; *P,0.05, **P,0.01 (ANOVA). RESEARCH ARTICLE Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 Fig. 7. Prolonged expression of TrkB at the cell surface induced by neuronal activity depends on the NMDA receptor and CaMKII. Hippocampal neurons were pretreated with 50 mM MK801, 20 mM KN-93 or 20 mM U0126 for 30 minutes before incubation with 15 mM KCl, 15 mM NaCl or stimulation with TBS for 1 hour, and then treatment with 1 nM BDNF for 4 hours. Cell surface proteins were biotinylated and collected with immobilized streptavidin. Equal amounts of biotinylated proteins were loaded to analyze surface TrkB expression. Representative western blots (A,C) and quantitative plots (B,D) are presented. Data show the mean6s.e.m.; n54; *P,0.05, **P,0.01 (ANOVA). increase in pTrkB at the 4-hour time-point (Fig. 8A, right four lanes). In the right hippocampus, injection of exogenous BDNF elicited a marked increase in pTrkB at 1 hour, but this declined significantly 4 hours after injection in the absence of prior running. By contrast, injection of BDNF into the right hippocampus in rats that experienced prior running elicited a sustained TrkB phosphorylation (Fig. 8A, left four lanes). To calculate the TrkB activation induced by injected BDNF, we subtracted the pTrkB:TrkB ratio of the left (PBS-injected) hippocampus from that of the right (BDNF-injected) hippocampus. The pTrkB:TrkB ratio of the non-running group at 4 hours was 45%613% (6s.e.m.) of that at 1 hour, whereas that of the running group was 89%611% (Fig. 8B). Thus, physical exercise, a potential way to increase hippocampal neuronal activity (Vanderwolf, 1969; Kuo et al., 2011; Oladehin and Waters, 2001), significantly prolonged BDNF-induced TrkB signaling in the hippocampus in vivo. Fig. 8. Running on the treadmill prolonged TrkB activation induced by exogenous BDNF in vivo. Rats were forced to run on the treadmill for 0.5 hours and rest for another 0.5 hours. BDNF was injected into the right, but not left, hippocampal CA1. Approximately 1 mg of tissue from around the injection site was dissected out at 1-hour and 4-hour time-points and was processed for the analysis of TrkB phosphorylation by western blotting. (A) Representative western blot showing the effect of running on TrkB phosphorylation induced by BDNF injection to the right hippocampus. As a control, the left hippocampus (PBS injection) was processed in the same way (right four lanes). (B) Quantification of the effect of running on TrkB phosphorylation kinetics. The values were derived from the formula shown on the right. Data are presented as the mean6s.e.m.; n54; *P,0.05 (paired t-test). 2256 BDNF–TrkB is a key signaling pathway involved in activitydependent processes such as synaptic plasticity. However, most, if not all, studies of BDNF signaling so far have been performed in cultures where neurons are either silent or fire action potentials spontaneously and infrequently. This is in marked contrast to the situation in the brain in vivo, where many neurons are constantly firing either intrinsically or in response to external stimulation. For example, upon stimulation of the retina by light, neurons in the visual cortex are activated. The prefrontal cortex neurons can also be activated when the visual stimulation is coupled to a working-memory task. Activation of the hippocampus follows if the visual image is processed for a long-term memory (Jeneson and Squire, 2012). An important but largely neglected question in neuronal cell biology is how exactly neuronal firing affects BDNF signal transduction and functions. In this study, we compared BDNF signaling and functions in the presence or absence of neuronal activity, using a simple but efficient in vitro culture model to mimic the in vivo situation. Typically, an acute application of BDNF triggers a robust but transient activation of TrkB (pTrkB) in cultured hippocampal neurons. Remarkably, when neurons are actively firing, the application of BDNF leads to a sustained TrkB activation (sustained pTrkB). A significant but unexpected finding is that sustained pTrkB signaling elicits cellular functions that are distinct from those elicited by transient TrkB activation. For example, several studies have shown that application of BDNF induces an increase in the number of primary dendrites (McAllister et al., 1995; Cheung et al., 2007; Ji et al., 2010; Ji et al., 2005). However, we found Journal of Cell Science DISCUSSION that application of BDNF together with neuronal depolarization promotes dendritic branching but not changes in the number of primary dendrites. Consistent with this finding, inhibition of neuronal activity prevents the increase in dendritic arborization induced by BDNF treatment of slices from the developing visual cortex (McAllister et al., 1996). These findings raise questions as to how relevant are the culture studies without neuronal activity with respect to BDNF functions under physiological conditions. Another interesting discovery is that BDNF is more efficacious in protecting neurons against glutamate-induced neuronal apoptosis when neurons were activated, although BDNF or high K+ (50 mM) alone has also been reported to induce neuronal survival (Ghosh et al., 1994; Arancibia et al., 2008; Mattson et al., 1995). Thus, manipulations that lead to sustained TrkB might be a viable therapeutic strategy for neurodegenerative diseases, such as Alzheimer’s disease. It is important to note that several previous studies have also pointed to the synergistic effects of increased neuronal activity and BDNF in the brain. For example, transcranial direct-current stimulation (tDCS) enhances cortical activity in humans (Nitsche and Paulus, 2000). It has been shown recently that tDCS induces a novel form of LTP through DCS-enhanced BDNF secretion in the M1 motor cortex (Fritsch et al., 2010), facilitating motor-skill learning in humans (Reis et al., 2009). Exercise, such as voluntary wheel-running, enhances BDNF expression and elicits a persistent firing pattern (theta rhythm) in the hippocampus (Vanderwolf, 1969). It has been shown that running on the treadmill facilitates hippocampal theta activity (Kuo et al., 2011), triggers c-fos activation in hippocampal neurons and enhances hippocampal LTP and cognition (Cotman and Berchtold, 2002; Oladehin and Waters, 2001). Although it remains challenging to demonstrate that enhanced neuronal activity, either by exercise or tDCS, could facilitate BDNF functions in vivo, we attempted to perform a simple experiment to test whether neuronal activity regulates BDNF–TrkB signaling in vivo. Our results showed that forced running converted transient TrkB signaling to sustained signaling induced by exogenous BDNF. Although it is difficult to test whether the function of physical exercise was mediated by increasing neuronal activity experimentally, it is conceivable that high frequency neuronal activity could prolong TrkB signaling by endogenous BDNF in vivo as well. It should be pointed out that forced running on the treadmill could induce the release of stress hormones (Yanagita et al., 2007), which might also affect BDNF signaling. Exercise is known to promote neural repair and facilitate brain functions (Cotman and Berchtold, 2002). Future experiments should examine whether the beneficial effects of exercise are mediated by sustained BDNF signaling. The present study established a simple culture system to investigate the role of neuronal activity in BDNF–TrkB signaling. Using this system, we demonstrated an increase in TrkB surface expression or, more accurately, a decrease in BDNF-induced downregulation of TrkB on the neuronal surface, resulting from the combination of BDNF and neuronal activity. Taken together, these results support a model in which neuronal activity drives sustained BDNF–TrkB signaling by a prolonged increase in TrkB expression on the neuronal surface. It is important to distinguish the present result from previous findings on activity-dependent increases in TrkB surface expression (Du et al., 2000; MeyerFranke et al., 1998). Meyer-Franke and colleagues first reported that increasing neuronal activity by cAMP or by high K+ facilitates the insertion of TrkB into the membranes of retinal ganglion neurons (Meyer-Franke et al., 1998). Subsequently, Du Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 and colleagues showed that high-frequency, but not low frequency or high K+, rapidly increased TrkB insertion into the surface of hippocampal neurons, which was detected immediately after the termination of stimulation in the absence of BDNF (Du et al., 2000). By contrast, the present study was designed to investigate how neuronal activity alters BDNF-induced downregulation of surface TrkB expression over a long period of time. In general, the application of BDNF to neurons would induce a time-dependent decrease in the levels of surface TrkB due to ligand-induced receptor endocytosis, which is known to be crucial for BDNF signaling. In this study, we found that prior neuronal activation, either by 15 mM K+ or high-frequency (but not low-frequency) stimulation, can prevent BDNF-induced downregulation of surface TrkB expression. Remarkably, after the termination of electrical stimulation, the TrkB on the surface of hippocampal neurons was maintained at a high level for at least 4 hours in the presence of BDNF. Therefore, it is suggested that prior neuronal activity facilitates exocytosis and/or recycling of TrkB receptors to counteract the ligand-induced endocytosis, thereby increasing the turnover of TrkB at the plasma membrane. We have performed several experiments to address a number of limitations in this culture system. First, what is the effect of elevated neuronal activity on endogenous BDNF, which could also contribute to the sustained TrkB activation? Both LTPinducing stimuli and back-propagating action potentials have been shown to elicit BDNF secretion in cultured neurons (Ga¨rtner et al., 2002; Kuczewski et al., 2008). It is likely that high K+ and TBS alone can also induce endogenous BDNF secretion. However, we found neither high K+ nor TBS alone can induce visible TrkB phosphorylation (supplementary material Fig. S2A,D). Furthermore, blocking endogenous BDNF with TrkB– Fc during stimulation also did not affect activity-dependent sustained TrkB activation (supplementary material Fig. S2B,E). Second, what is the role of spontaneous firing of cultured neurons, albeit at a very low frequency? We used TTX to block spontaneous neuronal activity and found that prior treatment with TTX for 30 minutes had no effect on high-K+-induced enhancement of TrkB phosphorylation (supplementary material Fig. S2C). Finally, one question that the present study did not and could not address is whether neuronal activity regulates TrkB signaling and kinetics that are induced by the endogenously released BDNF. There are several technical obstacles that prevent us from addressing this question. First, there is no method that is reliable and sensitive enough to measure endogenously released BDNF. We found that even with the very strong stimulation such as 15 mM K+ or TBS, the amount of endogenous BDNF released was too low to induce pTrkB. Second, all signaling studies require precise timing of initiation as well as monitoring over time of the signaling events. There is no method, as far as we are aware, that could induce and measure the release of a known amount of endogenous BDNF with precise timing. Last, but not least, it is difficult to separate the effects of stimulation on BDNF release and those on surface TrkB expression. Application of exogenous BDNF seems to overcome these obstacles. MATERIALS AND METHODS Antibody and reagents Antibodies were from Cell Signaling Technology (Danvers, MA) unless indicated otherwise. The antibodies used in this study were against: phospho-Akt (Ser 473), phospho-TrkA (490), Erk, phospho-Erk, CREB, phospho-CREB, PLCc, phospho-PLCc, TrkB, Akt (Santa Cruz 2257 Journal of Cell Science RESEARCH ARTICLE Biotechnology, Santa Cruz, CA) and MAP2 (EMD Millipore Corporation, Billerica, MA). AIP, MK801, KN-93, U0126 and K252a were all purchased from Tocris (Bristol, UK). TTX was purchased from Alomone labs (Jerusalem, Israel). TrkB–Fc was from Sino Biological (Beijing, China). CdCl2, nimodipine and CNQX were purchased from Sigma-Aldrich (St Louis, MO). Sulfo-NHS-LC–Biotin and ImmunoPure Immobilized Streptavidin were from Pierce (Rockford, IL). Cell culture and treatment Culture of hippocampal neurons from embryonic day (E)18 rat embryos was performed as described previously (Ji et al., 2005). Neurons were dissociated and plated on 100 ng/ml poly-D-lysine-coated 12-well plates at 300,000 cells/well. Neurons were cultured for 12–14 days and then processed for western blotting. Dendritic morphology was studied with cultures grown for only 3 days on coverslips at 5000 cells/coverslip. For BDNF treatments, neurons were treated with 15 mM NaCl or KCl in culture medium (neurobasal medium supplemented with B27 and GlutaMax, Gibco) for 1 hour before incubation with 1 nM BDNF. Neurons were collected at different time-points (from 0–4 hours) and lysed for biochemical experiments. All experiments involving animals were approved by Tsinghua University Committees on Animal Care. Ca2+ imaging After culturing for 12–14 days, rat hippocampal neurons on coverslips were loaded with the ratiometric Ca2+ indicator fura-2-AM (Molecular Probe; 2.5 mg/ml) in Tyrode’s buffer for 30 minutes at room temperature, followed by a wash step and 30 minutes in culture medium at 37 ˚C. Coverslips were mounted in the recording chamber. The cells were then maintained in a steady gravity-fed flow of Phenol-Red-free neurobasal medium (Gibco; 34–36 ˚C maintained with an in-line heater, Warner instruments). Capturing regions were selected on the cell bodies of several neurons. For wash-in of 15 mM KCl, liquid flow was switched to Phenol-Red-free neurobasal medium with 15 mM KCl. Images of selected regions were captured at both 340-nm and 380-nm excitation wavelengths at 1 Hz using a Nikon TiE microscope and Roper HQ2 camera coupled with Metafluor software. The data were imported into an Excel file. Fura-2 340:380 ratios were then calculated for 200 seconds before and 400 seconds during 15 mM KCl wash-in. Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 luminescence that could be recorded and quantified. This study was conducted after review by the Institutional Animal Care and Use Committee at GlaxoSmithKline (GSK) and in accordance with the GSK Policy on the Care, Welfare and Treatment of Laboratory Animals. Analysis of neurite complexity Cultured hippocampal neurons at 2 days in vitro were treated with BDNF in the presence of 15 mM NaCl or KCl and grown for an additional 3 days. The neurons were fixed and stained with a mouse antibody against MAP2. In a double-blind manner, images were acquired with an Olympus microscope (206). The number of primary neurites and the number of branch points per cell were analyzed. Typically, images of 25 neurons per condition were captured for each experiment, and four independent experiments were performed. Electrical stimulation of cultured hippocampal neurons On cultured hippocampal neurons, we used theta-burst stimulation (TBS) as described previously (Du et al., 2000). Electric-field stimulation was applied across a 12-well plate through a custom-made lid, which contained platinum wires that contacted the medium in each well. Cultured hippocampal neurons were pre-incubated in Tyrode’s solution (129 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 30 mM glucose, 25 mM HEPES pH 7.3, with an osmolarity of 31362 mOsm) for 1–2 hours and then stimulated with one of four kinds of electrical stimulation at room temperature. Each stimulation pulse (1 msecond, 5 V) was sufficient to elicit action potentials in these cultured neurons. The four programs of electrical stimulation were: (1) TBS; each episode consisted of four bursts, each with five biphasic pulses at 100 Hz (10msecond intervals), separated by an interburst interval of 200 mseconds. One episode was given every 5 seconds throughout the whole incubation period, (2) LTD-inducing electrical stimulation (LIS); 1 Hz, 15 minutes, (3) LTP-inducing tetanic stimulation; 100 Hz, 1 second, given every 10 minutes for 30 minutes, (4) Low frequency stimulation (LFS); 300 pulses in 30 minutes. Surface biotinylation analysis of TrkB Cells were lysed in the lysis buffer (20 mM HEPES pH 7.4, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 2 mM EGTA and a phospho-protease inhibitor cocktail purchased from Roche). Samples were solubilized with the loading buffer, incubated at 100 ˚C for 5 minutes and resolved by SDS-PAGE. Equal amounts of samples were loaded into discontinuous gels (10% acrylamide tested under reducing conditions). Proteins were transferred to PVDF membrane for 1.5 hours at 100 V. The membrane was blocked with 5% bovine serum albumin (BSA) for 1 hour at room temperature and incubated with the primary antibodies in blocking solutions at 4 ˚C overnight before detection with HRP-conjugated secondary antibodies. Chemiluminescence was detected with ECL solution. Surface TrkB receptors were measured by biotinylation followed by western blotting using an anti-TrkB antibody as described previously (Du et al., 2000; Meyer-Franke et al., 1998). After various treatments, the cultured hippocampal neurons were quickly rinsed twice in ice-cold PBSCa2+-Mg2+ (PBS pH 7.4, containing 1 mM CaCl2 and 0.5 mM MgCl2). Cell surface proteins were biotinylated for 1 hour with Sulfo-NHS-L– Biotin (0.25 mg/ml) diluted in PBS-Ca2+-Mg2+. Biotinylation was stopped by removing the solution and incubating the cells in 10 mM ice-cold glycine in PBS-Ca2+-Mg2+ for 20 minutes. Neurons were then washed three times with cold PBS-Ca2+-Mg2+ and lysed with RIPA buffer. Approximately 20% of the biotinylated proteins were denatured by heating in SDS sample buffer to analyze total TrkB. The remaining lysate was precipitated with ImmunoPure Immobilized Streptavidin overnight at 4 ˚C with constant mixing. The biotinylated protein precipitates were washed with RIPA buffer and processed for western blot analysis. Neuronal survival assay BDNF signaling in vivo Hippocampal neurons were cultured for 8–10 days. After the neurons were treated with 15 mM NaCl or 15 mM KCl or control medium for 15 minutes, the medium was changed to the original medium, followed by 1 nM BDNF treatment for 6 hours. Subsequently, the cultures were treated with vehicle or glutamate (125 mM) for 15 minutes. The cultures were changed back into original medium for an additional 7 hours, and cell viability was measured by using the Cell Titer-Glo Luminescence Kit (Promega). The cell viability assay is a method to determine the number of viable cells in culture based on quantification of the ATP present, which signals the presence of metabolically active cells. Briefly, UltraGlo luciferase mixed with its substrate was added to cultured neurons with fresh medium. After shaking for 2 minutes to induce cell lysis, the plates were incubated at room temperature for 10 minutes to stabilize luminescence signals. The ATP levels were proportional to the Two-month-old male Sprague Dawley rats raised in 12-hour light/dark cycles were anesthetized and implanted with guide cannulas into the dorsal hippocampal CA1 region at the following stereotaxic coordinates; 23.6 mm posterior to bregma, 62.2 mm lateral, 2.6 mm dorsal-ventral, ventral to outer skull surface. The guide cannula was fixed in place by using dental acrylic around the protruding exterior of the cannula and two anchoring screws. At the completion of surgery, a sterile stainless-steel dummy cannula was inserted into the guide cannula to maintain the cannula lumen. Animals were allowed to recover from surgery for 1 week. After 0.5 hours of running (14 meter/min) on a forced-running treadmill and 0.5 hours of rest, the rats were anesthetized briefly (,2– 3 minutes) using isoflurane and were injected with 0.2 mg of BDNF. About 1 mg of tissue from around the injection site was dissected for analysis by western blotting. Western blot analysis 2258 Journal of Cell Science RESEARCH ARTICLE Data analysis Western blot data acquisition and analysis were performed using FluorChem HD2 Gel Imaging System and software (Alpha Innotech, San Leandro, CA). Neuronal viability was measured using a microplate reader system (BioTek, Synergy 4, Winooski, VT). All counts were processed with Microsoft Excel and are presented as the mean6s.e.m. Statistical significance was determined using a Student’s two-tailed t-test for two-group comparison or ANOVA for multiple comparison. Data plotting, statistical analysis and figure preparation were performed with Excel, GraphPad Prism, Adobe Photoshop and Illustrator CS. Acknowledgements We thank Guhan Nagappan (GlaxoSmithKline, R&D, Shanghai, China) for his help with critical review and editing of this manuscript. We thank the National Institute of Biological Sciences in Beijing for technical support for in vivo experiments, and Bailong Xiao, Xiaoling Liu and Yan Jiang (all Tsinghua University, Beijing, China) for their help with Ca2+ imaging. Competing interests Y.J. is an employee of GlaxoSmithKline, R&D, Shanghai, China. Author contributions W.G. co-designed and conducted most of the experiments and contributed to writing the manuscript. Y.J. conducted the neuronal survival experiments and contributed to writing the manuscript. S.W. conducted some of the biochemical experiments. Y.S. assisted with in vivo experiments. B.L. supervised the project, designed the experiments and wrote the paper. Funding This work was supported by funds from the Tsinghua University–Peking University Center for Life Sciences and GlaxoSmithKline R&D, China. W.G., S.W., Y.S. and B.L. were supported by funds from the Tsinghua University-Peking University Center for Life Sciences. Y.J. was supported by GlaxoSmithKline R&D, China. Supplementary material Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.139964/-/DC1 References Almeida, R. D., Manadas, B. J., Melo, C. V., Gomes, J. R., Mendes, C. S., Gra˜os, M. M., Carvalho, R. F., Carvalho, A. P. and Duarte, C. B. (2005). Neuroprotection by BDNF against glutamate-induced apoptotic cell death is mediated by ERK and PI3-kinase pathways. Cell Death Differ. 12, 13291343. Arancibia, S., Silhol, M., Moulie`re, F., Meffre, J., Ho¨llinger, I., Maurice, T. and Tapia-Arancibia, L. (2008). Protective effect of BDNF against beta-amyloid induced neurotoxicity in vitro and in vivo in rats. Neurobiol. Dis. 31, 316326. Berridge, M. J. (1998). Neuronal Ca2+ signaling. Neuron 21, 13-26. Bito, H., Deisseroth, K. and Tsien, R. W. (1996). CREB phosphorylation and dephosphorylation: a Ca(2+)- and stimulus duration-dependent switch for hippocampal gene expression. Cell 87, 1203-1214. Boulanger, L. M. and Poo, M. M. (1999). Presynaptic depolarization facilitates neurotrophin-induced synaptic potentiation. Nat. Neurosci. 2, 346-351. Buzsa´ki, G. (2002). Theta oscillations in the hippocampus. Neuron 33, 325340. Cheung, Z. H., Chin, W. H., Chen, Y., Ng, Y. P. and Ip, N. Y. (2007). Cdk5 is involved in BDNF-stimulated dendritic growth in hippocampal neurons. PLoS Biol. 5, e63. Cotman, C. W. and Berchtold, N. C. (2002). Exercise: a behavioral intervention to enhance brain health and plasticity. Trends Neurosci. 25, 295-301. Du, J., Feng, L., Yang, F. and Lu, B. (2000). Activity- and Ca(2+)-dependent modulation of surface expression of brain-derived neurotrophic factor receptors in hippocampal neurons. J. Cell Biol. 150, 1423-1434. Du, J., Feng, L., Zaitsev, E., Je, H. S., Liu, X. W. and Lu, B. (2003). Regulation of TrkB receptor tyrosine kinase and its internalization by neuronal activity and Ca2+ influx. J. Cell Biol. 163, 385-395. Figurov, A., Pozzo-Miller, L. D., Olafsson, P., Wang, T. and Lu, B. (1996). Regulation of synaptic responses to high-frequency stimulation and LTP by neurotrophins in the hippocampus. Nature 381, 706-709. Finkbeiner, S., Tavazoie, S. F., Maloratsky, A., Jacobs, K. M., Harris, K. M. and Greenberg, M. E. (1997). CREB: a major mediator of neuronal neurotrophin responses. Neuron 19, 1031-1047. Frank, L., Ventimiglia, R., Anderson, K., Lindsay, R. M. and Rudge, J. S. (1996). BDNF down-regulates neurotrophin responsiveness, TrkB protein and Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 TrkB mRNA levels in cultured rat hippocampal neurons. Eur. J. Neurosci. 8, 1220-1230. Fritsch, B., Reis, J., Martinowich, K., Schambra, H. M., Ji, Y., Cohen, L. G. and Lu, B. (2010). Direct current stimulation promotes BDNF-dependent synaptic plasticity: potential implications for motor learning. Neuron 66, 198204. Ga¨rtner, A. and Staiger, V. (2002). Neurotrophin secretion from hippocampal neurons evoked by long-term-potentiation-inducing electrical stimulation patterns. Proc. Natl. Acad. Sci. USA 99, 6386-6391. Ghosh, A., Carnahan, J. and Greenberg, M. E. (1994). Requirement for BDNF in activity-dependent survival of cortical neurons. Science 263, 16181623. Gottschalk, W., Pozzo-Miller, L. D., Figurov, A. and Lu, B. (1998). Presynaptic modulation of synaptic transmission and plasticity by brain-derived neurotrophic factor in the developing hippocampus. J. Neurosci. 18, 6830-6839. Grubb, M. S. and Burrone, J. (2010). Activity-dependent relocation of the axon initial segment fine-tunes neuronal excitability. Nature 465, 1070-1074. ˚ kerman, K. E. O., Stoilov, P., Stamm, Haapasalo, A., Sipola, I., Larsson, K., A S., Wong, G. and Castre´n, E. (2002). Regulation of TRKB surface expression by brain-derived neurotrophic factor and truncated TRKB isoforms. J. Biol. Chem. 277, 43160-43167. Huang, E. J. and Reichardt, L. F. (2001). Neurotrophins: roles in neuronal development and function. Annu. Rev. Neurosci. 24, 677-736. Ishida, A., Kameshita, I., Okuno, S., Kitani, T. and Fujisawa, H. (1995). A novel highly specific and potent inhibitor of calmodulin-dependent protein kinase II. Biochem. Biophys. Res. Commun. 212, 806-812. Jeneson, A. and Squire, L. R. (2012). Working memory, long-term memory, and medial temporal lobe function. Learn. Mem. 19, 15-25. Ji, Y., Pang, P. T., Feng, L. and Lu, B. (2005). Cyclic AMP controls BDNF-induced TrkB phosphorylation and dendritic spine formation in mature hippocampal neurons. Nat. Neurosci. 8, 164-172. Ji, Y., Lu, Y., Yang, F., Shen, W., Tang, T. T. T., Feng, L., Duan, S. and Lu, B. (2010). Acute and gradual increases in BDNF concentration elicit distinct signaling and functions in neurons. Nat. Neurosci. 13, 302-309. Kaplan, D. R. and Stephens, R. M. (1994). Neurotrophin signal transduction by the Trk receptor. J. Neurobiol. 25, 1404-1417. Kuczewski, N., Porcher, C., Ferrand, N., Fiorentino, H., Pellegrino, C., Kolarow, R., Lessmann, V., Medina, I. and Gaiarsa, J.-L. (2008). Backpropagating action potentials trigger dendritic release of BDNF during spontaneous network activity. J. Neurosci. 28, 7013-7023. Kuo, T. B. J., Li, J. Y., Chen, C. Y. and Yang, C. C. H. (2011). Changes in hippocampal h activity during initiation and maintenance of running in the rat. Neuroscience 194, 27-35. Lewin, G. R. and Barde, Y. A. (1996). Physiology of the neurotrophins. Annu. Rev. Neurosci. 19, 289-317. Lisman, J., Schulman, H. and Cline, H. (2002). The molecular basis of CaMKII function in synaptic and behavioural memory. Nat. Rev. Neurosci. 3, 175-190. Lu, B. and Figurov, A. (1997). Role of neurotrophins in synapse development and plasticity. Rev. Neurosci. 8, 1-12. Lu, Y., Christian, K. and Lu, B. (2008). BDNF: a key regulator for protein synthesis-dependent LTP and long-term memory? Neurobiol. Learn. Mem. 89, 312-323. Lu, B., Nagappan, G., Guan, X., Nathan, P. J. and Wren, P. (2013). BDNF-based synaptic repair as a disease-modifying strategy for neurodegenerative diseases. Nat. Rev. Neurosci. 14, 401-416. Martinowich, K., Manji, H. and Lu, B. (2007). New insights into BDNF function in depression and anxiety. Nat. Neurosci. 10, 1089-1093. Mattson, M. P., Lovell, M. A., Furukawa, K. and Markesbery, W. R. (1995). Neurotrophic factors attenuate glutamate-induced accumulation of peroxides, elevation of intracellular Ca2+ concentration, and neurotoxicity and increase antioxidant enzyme activities in hippocampal neurons. J. Neurochem. 65, 17401751. McAllister, A. K., Lo, D. C. and Katz, L. C. (1995). Neurotrophins regulate dendritic growth in developing visual cortex. Neuron 15, 791-803. McAllister, A. K., Katz, L. C. and Lo, D. C. (1996). Neurotrophin regulation of cortical dendritic growth requires activity. Neuron 17, 1057-1064. McAllister, A. K., Katz, L. C. and Lo, D. C. (1999). Neurotrophins and synaptic plasticity. Annu. Rev. Neurosci. 22, 295-318. Meyer-Franke, A., Wilkinson, G. A., Kruttgen, A., Hu, M., Munro, E., Hanson, M. G., Jr, Reichardt, L. F. and Barres, B. A. (1998). Depolarization and cAMP elevation rapidly recruit TrkB to the plasma membrane of CNS neurons. Neuron 21, 681-693. Nagahara, A. H. and Tuszynski, M. H. (2011). Potential therapeutic uses of BDNF in neurological and psychiatric disorders. Nat. Rev. Drug Discov. 10, 209219. Nagappan, G., Zaitsev, E., Senatorov, V. V., Jr, Yang, J., Hempstead, B. L. and Lu, B. (2009). Control of extracellular cleavage of ProBDNF by high frequency neuronal activity. Proc. Natl. Acad. Sci. USA 106, 1267-1272. Nitsche, M. A. and Paulus, W. (2000). Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. J. Physiol. 527, 633-639. Obrietan, K., Gao, X.-B. and Van Den Pol, A. N. (2002). Excitatory actions of GABA increase BDNF expression via a MAPK-CREB-dependent mechanism – a positive feedback circuit in developing neurons. J. Neurophysiol. 88, 10051015. 2259 Journal of Cell Science RESEARCH ARTICLE RESEARCH ARTICLE Squinto, S. P., Stitt, T. N., Aldrich, T. H., Davis, S., Bianco, S. M., Radziejewski, C., Glass, D. J., Masiakowski, P., Furth, M. E., Valenzuela, D. M. et al. (1991). trkB encodes a functional receptor for brain-derived neurotrophic factor and neurotrophin-3 but not nerve growth factor. Cell 65, 885-893. Tongiorgi, E., Righi, M. and Cattaneo, A. (1997). Activity-dependent dendritic targeting of BDNF and TrkB mRNAs in hippocampal neurons. J. Neurosci. 17, 9492-9505. Vanderwolf, C. H. (1969). Hippocampal electrical activity and voluntary movement in the rat. Electroencephalogr. Clin. Neurophysiol. 26, 407-418. Yanagita, S., Amemiya, S., Suzuki, S. and Kita, I. (2007). Effects of spontaneous and forced running on activation of hypothalamic corticotropin-releasing hormone neurons in rats. Life Sci. 80, 356-363. Zhao, L., Sheng, A.-L., Huang, S.-H., Yin, Y.-X., Chen, B., Li, X.-Z., Zhang, Y. and Chen, Z.-Y. (2009). Mechanism underlying activity-dependent insertion of TrkB into the neuronal surface. J. Cell Sci. 122, 3123-3136. Zuccato, C. and Cattaneo, E. (2009). Brain-derived neurotrophic factor in neurodegenerative diseases. Nat. Rev. Neurol. 5, 311-322. Journal of Cell Science Oladehin, A. and Waters, R. S. (2001). Location and distribution of Fos protein expression in rat hippocampus following acute moderate aerobic exercise. Exp. Brain Res. 137, 26-35. Reis, J., Schambra, H. M., Cohen, L. G., Buch, E. R., Fritsch, B., Zarahn, E., Celnik, P. A. and Krakauer, J. W. (2009). Noninvasive cortical stimulation enhances motor skill acquisition over multiple days through an effect on consolidation. Proc. Natl. Acad. Sci. USA 106, 1590-1595. Sala, C., Rudolph-Correia, S. and Sheng, M. (2000). Developmentally regulated NMDA receptor-dependent dephosphorylation of cAMP response elementbinding protein (CREB) in hippocampal neurons. J. Neurosci. 20, 3529-3536. Shimada, A., Mason, C. A. and Morrison, M. E. (1998). TrkB signaling modulates spine density and morphology independent of dendrite structure in cultured neonatal Purkinje cells. J. Neurosci. 18, 8559-8570. Sommerfeld, M. T., Schweigreiter, R., Barde, Y.-A. and Hoppe, E. (2000). Down-regulation of the neurotrophin receptor TrkB following ligand binding. Evidence for an involvement of the proteasome and differential regulation of TrkA and TrkB. J. Biol. Chem. 275, 8982-8990. Journal of Cell Science (2014) 127, 2249–2260 doi:10.1242/jcs.139964 2260