It has been proven that the mammalian brain continuously produces newborn neurons in the hippocampal dentate gyrus (DG) throughout adulthood (Alvarez-Buylla and Garcia-Verdugo, 2002). Adult hippocampal neurogenesis is a complex process that begins with stem cells proliferation and followed by neuronal lineage specification, maturation, migration and incorporation into the hippocampal circuitry (Ge et al., 2008). Deficits in adult hippocampal neurogenesis can be seen in the pathogenesis of neurological diseases, including Alzheimer’s disease (AD), Parkinson’s disease (PD), depression and epilepsy (Balu and Lucki, 2009). Neurite growth is an important process in adult hippocampal neurogenesis and this process is crucial for the newborn neurons to establish new synaptic connections with the existing hippocampal circuitry by extending axonal and dendritic projections (Zhao et al., 2006). Neurite growth is regulated by a specific range of the intracellular free calcium concentration ([Ca²⁺]_i) (Mattson, 1987; Zheng, 2007; Toth et al., 2016). Sustained elevation of [Ca²⁺]_i inhibits growth cone advancement, and maximal neurite outgrowth occurred within an optimal range of Ca²⁺ concentrations (Tang and Dent, 2003). Ca²⁺ influx through Ca²⁺-permeable ion channels on the cell membrane or release from the intracellular Ca²⁺ stores can increase [Ca²⁺]_i. In neurons, Ca²⁺ influx mainly occurs through voltage-gated calcium channels (VGCCs) and ligand-gated ion channels such as N-methyl-D-aspartate glutamate receptor (NMDAR), nicotinic acetylcholine receptor (nAChR), and transient receptor potential (TRP) channels (Pankratov and Lalo, 2014; Benemei et al., 2015). It has been reported that an L-type VGCC antagonist promotes axon outgrowth in dopaminergic brain slice co-cultures (Sygnecka et al., 2015). NMDAR has been proven to regulate neurite growth in the hippocampus (Nacher and McEwen, 2006). Activation of α7nAChR significantly reduces axon growth in hippocampal neurons and thereby contributes to synaptic development and plasticity in the hippocampus (Nordman et al., 2014). Therefore, Ca²⁺-permeable channels may provide effective targets for the regulation of neurite growth.

The TRP channel family consists of several non-selective cationic channels that are sensitive to multiple physicochemical stimuli and capable of coupling their activity to downstream modulations of intracellular signals (Benemei et al., 2015). TRP vanilloid 4 (TRPV4) is a member of the vanilloid TRP subfamily (White et al., 2016). Activation of TRPV4 induces Ca²⁺ influx, which increases [Ca²⁺]_i in various cell types. It has been reported that TRPV4 mediates neurotrophic factor-derived neuritogenesis in developing peripheral neurons (Jang et al., 2012). Some mutations in TRPV4 that augment Ca²⁺ influx have been shown to underlie the pathogenesis of TRPV4-linked axonal neuropathies (Fecto et al., 2011). In our recent study, activation of TRPV4 promoted the proliferation of stem cells in the adult hippocampal DG (Tian et al., 2016). Whether activation of TRPV4 participates in the modulation of the neurite growth of newborn neurons in the hippocampal DG remains unclear. Activation of AMP-activated protein kinase (AMPK) has been shown to regulate the hippocampal neuronal structure (Ramamurthy et al., 2014). In addition to playing a pivotal role in modulating neuronal survival and differentiation, protein kinase B (Akt) has recently emerged as an important regulator of neurite outgrowth (Read and Gorman, 2009). Activation of TRPV4 has been reported to modulate the AMPK and Akt signaling pathways in vitro (Hong et al., 2016). In the present study, we sought to determine whether TRPV4 activation affects the dendrite morphology of newborn neurons and further explored whether the AMPK or Akt signaling pathway is involved in TRPV4 action.

Neurite growth is a critical process of neurogenesis. Newborn neurons extend axonal and dendritic projections and establish new synaptic connections to the existing hippocampal circuitry (Alvarez-Buylla and Garcia-Verdugo, 2002; Zhao et al., 2006). Ca²⁺ has been proven to be an important regulator of neurite growth. Axon outgrowth and dendritic development occur within the optimal [Ca²⁺]_i level, and this process may be inhibited when the [Ca²⁺]_i level is below or above this optimal [Ca²⁺]_i level (Mattson, 1987; Redmond and Ghosh, 2005; Zheng, 2007). As Ca²⁺-permeable channels, some TRP family members induce Ca²⁺ influx, leading to increases in [Ca²⁺]_i. Accordingly, these channels have been shown to play a role in modulating neurite growth (Benemei et al., 2015). For example, TRPC5 participates in controling hippocampal neurite length and growth cone morphology, and activation of TRPC5 inhibits neurite extension as well as a rapid vesicular trafficking mechanism regulated by growth factors (Greka et al., 2003; Hui et al., 2006). Inhibition of TRPM2 markedly increases axonal growth, whereas over-expression of it inhibits axonal growth (Jang et al., 2014). Blockage of TRPV1 promotes neurite outgrowth after sciatic nerve injury (Ren et al., 2015). However, activation of TRPV2 leads to axon outgrowth in the developing dorsal root ganglion and motor neurons (Cohen et al., 2015). A gain-of-function of TRPV4 mutation increased its Ca²⁺ channel activity, which underlies the pathogenesis of TRPV4-linked axonal neuropathies (Fecto et al., 2011). Activation of TRPV4 may increase the [Ca²⁺]_i in hippocampal neurons (Shibasaki et al., 2007). In this study, DCX⁺ fibers was expressed as DCX⁺ fibers/DCX⁺cell and DCX⁺ fibers in Mm and Mo subregions were reduced by TRPV4 agonist (Figure 1), suggesting that the density and length of dendrite fibers per newborn neuron were impaired by TRPV4 activation. We previously reported that application of GSK1016790A did not affect the number of DCX⁺ fibers (Tian et al., 2016). In our previous study, GSK1016790A was consecutively injected for 5 days and DCX staining was performed 14 days after the last GSK1016790A injection (Tian et al., 2016). In this case, GSK1016790A was supposed to be absent 7 days after injection and the result that GSK1016790A had no effect on DCX⁺ fibers may not truly reflect the direct effect of TRPV4 activation on the dendritic arborization of newborn neurons. In this study, DCX staining was performed on the same day of the last GSK1016790A injection, and therefore, this inconsistency about the effect of TRPV4 activation on the DCX⁺ fibers is likely due to the difference in drug application protocols between the present and previous studies (Tian et al., 2016).

Akt is a key mediator of several aspects of neurite outgrowth, including elongation, branching and caliber (Read and Gorman, 2009). Previous reports from our group showed that activation of TRPV4 reduced the phosphorylation of Akt (Jie et al., 2015, 2016; Hong et al., 2016). AMPK is a serine/threonine kinase that functions as an important sensor in energy homeostasis in mammalian cells. In addition to an increased ratio of intracellular AMP to ATP, Ca²⁺/calmodulin-dependent protein kinase kinase-β can also activate AMPK in response to an increased [Ca²⁺]_i (Hawley et al., 2005). We recently reported that acute application of a TRPV4 agonist caused an increase in the phosphorylation of AMPK, which was related to the TRPV4-mediated Ca²⁺ influx (Hong et al., 2016). AMPK has been shown to inhibit neuronal development at multiple stages, during both axon outgrowth and dendrite growth and arborization (Ramamurthy et al., 2014). Activation of AMPK pathway can suppress axon initiation and neuronal polarization through interfering with PI3K signaling (Amato et al., 2011; Amato and Man, 2012). Application of AMPK antagonist CC promotes the dopaminergic dendrite outgrowth (Wakita et al., 2014). These reports prompted us to study whether the AMPK and Akt signaling pathways were involved in the TRPV4-induced inhibition of the dendritic arborization of newborn neurons. Here, we first examined the phosphorylation of AMPK and Akt at different time points after GSK1016790A treatment. The present results showed that the AMPK activity increased within 2 h and then decreased from 1 to 5 days after the beginning of the experiment, whereas the Akt activity consistently decreased from 30 min to 5 days after the beginning of the experiment. Notably, the decrease in the p-Akt protein level in GSK1016790A-injected mice was markedly blocked by the AMPK antagonist CC, but it was unaffected by the AMPK activator AICAR (Figure 2). Therefore, during chronic TRPV4 activation, the initial activation of AMPK is probably responsible for the inhibition of Akt signaling. We also found that the GSK1016790A-induced decrease of DCX⁺ cells and DCX⁺ fibers was markedly blocked by the PI3K agonist 740 Y-P or by AMPK antagonist CC. By contrast, the application of AMPK agonist AICAR did not have a similar effect in GSK1016790A-injected mice (Figure 5). Akt signaling also plays an important role in neuronal survival and previous studies show that activation of TRPV4 results in neuronal death, which is mediated, at least partially, by inhibiting Akt signaling (Luo et al., 2003; Jie et al., 2015, 2016). Therefore, our findings suggest that AMPK activation and the related down-regulation of Akt signaling are involved in the TRPV4-induced decrease of newborn neurons and the dendritic arborization of single neuron. Nevertheless, it was still unclear how AMPK activity was subsequently inhibited when TRPV4 was chronically activated. A previous study showed that activation of extracellular signal-regulated kinase 1/2 (ERK1/2) is responsible for dephosphorylation of AMPK at Thr¹⁷² caused by α- melanocyte-stimulating hormone (Damm et al., 2012). Of note, activation of TRPV4 has been proven to increase ERK1/2 activity (Jie et al., 2015; Tian et al., 2016). More experiments are needed to clarify whether the enhanced ERK1/2 is involved in the decrease in AMPK activity during the later stage of TRPV4 activation. The present data showed that the TRPV4-induced decrease in the p-Akt protein level was not affected by AICAR, indicating that other factors contribute to the inhibition of Akt signaling during chronic TRPV4 activation.

Mammalian target of rapamycin and GSK3β, two major substrates that are downstream of Akt, have been identified to play key roles in Akt-mediated neurite growth (Read and Gorman, 2009). Activation of the PI3K-Akt-mTOR signaling pathway can phosphorylate p70S6k, which plays a fundamental role in axonal and dendritic growth. In cultured hippocampal neurons, the dendritic complexity was reduced by inhibition of PI3K and knockdown of mTOR or p70S6k (Jaworski et al., 2005). In AD model mice, the neurite growth of newborn neurons in the hippocampal DG is impaired with reduced phosphorylation of Akt and mTOR (Li L. et al., 2010). Here, decreased p-mTOR and p-p70S6k protein levels were found in GSK1016790A-injected mice, and this inhibitory effect caused by TRPV4 activation was markedly blocked by 740 Y-P or CC (Figure 3). Together with the above discussion, these results indicate that activation of TRPV4 may increase AMPK activity and then lead to inhibition of PI3K-Akt-mTOR-p70S6k signaling pathway. GSK3β has been proven to play an important role in neuronal morphogenesis. GSK3β activity is inversely correlated with PI3K-Akt signaling activity (Cross et al., 1995). One of the downstream targets of GSK3β is MAP2, which is important for microtubule polymerization and dendrite elongation (Sánchez et al., 2000; Lim and Walikonis, 2008). Phosphorylation of MAP2 causes it to lose its ability to effectively associate with the microtubules. Akt activation was shown to inactivate GSK3β and consequently reduce phosphorylation of MAP2, which can promote microtubule polymerization and dendrite elongation (Lim and Walikonis, 2008). Another important downstream target of GSK3β is CRMP-2 that is critical for promoting axon specification and neurite extension through reorganization of actin filaments, regulation of microtubule assembly, endocytosis of adhesion molecules and also axonal protein trafficking (Yoshimura et al., 2005, 2006). GSK3β phosphorylates CRMP-2 to decreases its tubulin heterodimer binding ability (Yoshimura et al., 2005, 2006). In the present study, the p-GSK3β (Y²¹⁶) protein level increased significantly in GSK1016790A-injected mice, indicating higher activity of GSK3β upon TRPV4 activation (Bhat et al., 2000). In addition to the increased p-GSK3β protein level, the p-MAP2 and p-CRMP-2 protein levels were increased by GSK1016790A treatment, indicating a decrease in the MAP2 and CRMP-2 function upon TRPV4 activation (Figure 4). Here, the increase in the p-GSK3β (Y²¹⁶), p-MAP2 and p-CRMP-2 protein levels was markedly lower in the mice that were co-injected with GSK1016790A and 740 Y-P or and CC (Figure 4); this finding indicates that activation of TRPV4 may increase AMPK activity to down-regulate PI3K-Akt signaling, then increase GSK3β activity, and ultimately decrease MAP2 and CRMP-2 function.

The present results suggest that activation of AMPK and inhibition of Akt signaling are likely responsible for TRPV4-induced impairment of dendritic arborization of newborn neurons, which is mediated by downstream substrates of Akt, including mTOR and GSK3β (Figure 6). In this study, we did not provide the direct evidence for TRPV4 activation-induced modulation of axon outgrowth. Concerning the important role of AMPK, Akt and the related substrates in the modulation of axonogenesis, the activation of AMPK and the down-regulation of Akt signaling provide a possibility that the axon outgrowth of newborn neurons may be impaired by TRPV4 activation. More experiments are needed to prove this speculation. Here it is noted that besides the decrease of DCX⁺ fibers, the number of DCX⁺ cells was also reduced by activation of TRPV4. It has been reported that apoptosis of newborn cells occurs during adult neurogenesis (Sierra et al., 2010). In our previous study, the inhibited Akt signaling is involved in TRPV4 activation-induced apoptosis and therefore this action may be a reason for the present decreased DCX⁺ cells (Jie et al., 2015). Neurite growth is important for the survival of newborn neurons and impaired neurite growth may result in a decline of newborn neurons in the pathological condition (Li L. et al., 2010). In the present study, we did not examine the dying process of DCX⁺ cells upon TRPV4 activation, but the present reduced DCX⁺ fibers indicates that these newborn neurons were already “unhealthy” cells and their subsequent survival might be impaired. Collectively, TRPV4 activation-induced impairment of dendrite morphology may result from the decreased newborn neurons and the impairment of dendritic arborization of single neuron. As a Ca²⁺-permeable channel, TRPV4 may be a new potential target for modulating neurite growth. The present study also provides more evidence that TRPV4 activation probably plays an important role in the modulation of several processes of adult hippocampal neurogenesis.

YT, MQ, ZW, and SS performed experiments; CW, ZS, and YL analyzed data; LeC designed experiments and wrote the article; YD and LiC revised the manuscript and all authors approved the final version.

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