Results
Selective κ-ΟR agonists induce the autophagic flux in neuronal cells. It is known that neuronal autophagy is involved in cell growth, survival and synaptic plasticity (11, 13). Here, we investigated whether specific κ-OR agonists could trigger autophagy in neuronal cells and modulate synaptic organization. We thus treated Neuro-2A cells, stably expressing κ-ΟR, with U50,488H, a κ-OR-specific agonist, and monitored the levels of the lipidated LC3 (LC3-II), a reliable and specific marker of autophagosome formation located at the membrane of the autophagosome. As shown in Fig. 1A, increasing concentrations of U50,488H for 6 h caused a dose-dependent increase in LC3-II accumulation. Addition of the lysosomal inhibitor bafilomycin A1 (BafA1), which prevents the fusion of autophagosomes with lysosomes indicated a significant increase in LC3-II levels in cells exposed simultaneously to U50,488H and BafA1, relative to BafA1 alone (Fig. 1B), indicating that κ-OR activation upregulates the autophagic flux. Finally, LC3-II accumulation was markedly reversed upon treatment with the opioid antagonist, naloxone, prior to U50,488H exposure, further confirming the κ-ΟR-dependent autophagic activation (Fig. 1C).
To recapitulate U50,488H-mediated autophagy in a native neuronal milieu, we treated cortical neuronal cultures to U50,488H, that resulted in a significant increase of LC3-positive autophagosomes that appeared as puncta, compared with untreated controls (Fig. 1D, images a, e); an effect that was abrogated when cells were pre-treated with naloxone (images e, i). Consistent with these findings, immunoblot analysis of primary neuronal cultures showed that naloxone blocked the increase in LC3-II accumulation caused by U50, 488H exposure (Fig. 1E). Collectively, these results demonstrate that κ-OR activation induces autophagy in neuronal cells.
To deduce whether other κ-OR agonists exert similar effects on autophagy initiation, we exposed Neuro-2A cells to varying concentrations of the endogenous κ-OR neuropeptide dynorphin1-13, which also resulted in increased LC3-II and Beclin 1 levels (Fig. 2A, B). This effect was blocked by the selective κ-ΟR antagonist nor-BNI (Fig. 2C). Finally, dynorphin1-13-treated primary neuronal cultures indicated an increase of LC3-positive puncta compared to control neurons (Fig. 2D compare images a with e). These data suggest that κ-OR-induced autophagy is not selective to U50,488H, but can also be mediated by the endogenous κ-ΟR neuropeptide dynorphin.
Because a key initial event of the autophagosome biogenesis is the formation of the pre-autophagosomal structure (PAS), composed of ULK1, which is a complex of a serine/threonine protein, with the focal adhesion kinase (FIP200) and other proteins, we examined the timing of U50,488H-mediated early autophagic events. Treatment of primary neuronal cultures with U50,488H for 1, 6, and 24 h, triggered a marked increase of FIP200 and ULK1 protein levels reaching a peak at 6 h agonist exposure (Fig. 3A). In parallel, U50,488H treatment of Neuro-2A cells for various time intervals indicated a time-dependent increase of Beclin1, a key mediator of autophagosome formation. This increase peaked at 6 h following U50,488H administration (Fig. 3B). Similarly, as shown in Fig. 3C, exposure of Neuro-2A cells to U50,488H increased the protein levels of ATG5 and Beclin 1, with a concomitant decrease of p62 known to increase when autophagy is inhibited and decrease when autophagy is induced (21). Additionally, Becn1and Atg 5 mRNA levels were also elevated after 6 h U50,488H cell exposure (Fig. 3D). These results clearly demonstrate that κ-OR is involved in autophagosome biogenesis in neuronal cells.
Identification of the κ-OR signaling pathway that regulates the autophagic machinery. The κ-OR is coupled to pertussis toxin-sensitive Gi/o proteins to regulate a variety of effectors (1, 8). Το define the role of G proteins, we pretreated Neuro-2A cells with pertussis toxin (PTX), which ribosylates Gαi/o subunits. PTX blocked U50,488H-mediated increase of LC3-II and Beclin 1 levels (Fig. 4A, B), suggesting that Gi/o proteins are important players in κ-ΟR-mediated autophagy. To examine whether ERK1,2 is implicated in κ-OR- induction of autophagy the levels of ERK1,2 phosphorylation of Neuro-2A cells were assessed in the presence or absence of PTX. U50,488H enhanced ERK1,2 phosphorylation after 15 min and 6 h post-exposure and this effect was abolished by PTX exposure (Fig. 4C). Moreover, when the cells were pretreated with the ERK1,2 inhibitor PD98059 prior to agonist activation, U50,488H-mediated-ERK1,2 phosphorylation was abolished with a concomitant decrease of the LC3-II and Beclin1 levels, relative to the untreated cells (Fig. 4D). On the other hand, JNK activation cannot recapitulate the effects of ERK1,2 phosphorylation on κ-OR-mediated induction of the autophagic pathway. Indeed, no effects on LC3-II accumulation were detected in U50,488H-treated cells relative to the untreated ones upon pre-treatment with the JNK inhibitor SP600125 (Fig. 4E). These results suggest that ERK1,2 is implicated in κ-OR-mediated autophagy.
κ-OR regulates Beclin1 transcription via CREB activation.Because it is known that CREB regulates various autophagic genes (22) and that a consensus CRE binding site (TGACGTCA) exists in the mouse Becn1 promoter, we sought to determine if autophagic genes are regulated by p-CREB upon U50,488H exposure. CREB was phosphorylated in response to U50,488H cell exposure, and this effect was abolished by the ERK1,2 inhibitor PD98059 (Fig. 4F). Moreover, chromatin immunoprecipitation (ChIP) assay in isolated chromatin fragments of Neuro-2A cells indicated that CREB binding in theBecn1 promoter was greatly enhanced by U50,488H exposure relative to untreated cells (Fig. 4G). Therefore, κ-ΟR-mediated increase in Beclin1 levels appears to involve transcriptional activation of the beclin1 gene by ERK1,2-activated CREB.
U50,488H administration induces autophagy and promotes synaptic alterations in mouse hippocampus. We next sought to examine whether we could recapitulate U50,488H-mediated κ-ΟR autophagy in vivo and determine whether specific brain regions are involved. To this end, mice were injected with saline (vehicle) or U50,488H for 7 consecutive days and the levels of LC3-II and Beclin1 were measured in the hippocampus, cortex and striatum. U50,488H resulted in a significant increase of LC3-II and Beclin1 in the mouse hippocampus as compared to vehicle, but with no significant changes in cortical and striatal lysates (Fig. 5A-C). Collectively, these results suggest that κ-OR-mediated autophagy is detected specifically to the mouse hippocampus.
Autophagy contributes to synaptic plasticity by degrading specific proteins that are essential for synaptic function and spine remodeling (13). To elucidate whether U50,488H-mediated autophagy leads to synaptic alterations, initially the levels of proteins enriched in dendritic spines such as PSD-95 and spinophilin, were examined in the hippocampus, cortex and striatum of U50,488H-treated mice. As shown in Fig. 5D, spinophilin and PSD-95 in hippocampal lysates were significantly decreased in U50,488H-treated mice compared with saline-treated controls. However, no significant alterations for these proteins were detected in the cortex or striatum (Fig. 5E, F). This was further confirmed in isolated synaptosomes where a pronounced decrease of spinophilin, PSD-95, as well as SNAP25 was detected in U50,488H-injected mice, as compared to control ones (Fig. 5G). All these suggest that these synaptic proteins are degraded, possibly by being engulfed in the κ-ΟR-mediated autophagic cargo.
To test this hypothesis and in view of the known interaction of LC3 with autophagic cargos through the LC3-interacting regions (LIR) of various proteins, spinophilin, SNAP25 and PSD-95 among them (13, 23), we examined whether these proteins interact with LC3. Co-immunoprecipitation studies of hippocampal lysates using an LC3 antibody indicated that spinophilin, PSD-95 and SNAP25 do interact with LC3 (Fig. 6A). Moreover, to further define whether these synaptic protein alterations are indeed due to autophagy induction we measured their levels in the presence of BafA1. Treatment of Neuro-2A cells with U50,488H decreases the levels of spinophilin and PSD95. Inhibition of autophagy by BAF1 treatment did not alter these protein levels, suggesting that U50,488H-κ-ΟR activation indeed leads to degradation of these synaptosomal proteins (Fig. 6B). Finally, to verify whether these κ-OR-mediated effects are due to alterations in neuronal sprouting, the number of branches in U50,488H-treated hippocampal neuronal cultures were measured. U50,488H significantly reduced the number of branches relative to the controls (Fig. 6C), suggesting that κ-ΟR-induced autophagy modulates neuronal morphogenesis, possibly by degrading key synaptic proteins.
Activation of the endogenous κ-OR/dynorphin system upon stress upregulates autophagy in the hippocampus and results in synaptic alterations. It is well documented that the κ-ΟR/dynorphin system plays an important role in anxiety and stress-related behaviors and that κ-OR antagonists exhibit anxiolytic effects (24, 25). To examine whether stress-induced endogenous dynorphin release impacts on autophagy regulation, we examined the consequences of acute stress on autophagy in the hippocampus. To this end, mice injected with either vehicle or the κ-OR selective antagonist, nor-BNI, which is known to exert anxiolytic effects, were subjected to a two-day modified forced swim test (FST) (Fig. 7A). Male C57BL/6J mice were divided into 4 groups saline-control or saline-FST (stressed) and norBNI-not stressed or nor-BNI-FST (stressed). nor-BNI significantly decreased immobility time following the FST, compared to saline-treated mice, suggesting that nor-BNI attenuates stress-related behavior (Fig. 7B). Subsequently, the levels of the autophagic markers LC3-II and Beclin1 were measured in isolated hippocampal lysates and found to be significantly increased in stressed animals (FST) relative to vehicle injected-non-stressed ones (Fig. 7C). In contrast, this increase in autophagic markers was not detected in nor-BNI-treated mice under control or FST conditions when compared with the saline-control group (Fig. 7C). Moreover, as expected, no significant alterations of LC3-II levels were detected in the cortices of the same treatment groups (Fig. 7D), confirming that the hippocampus is the target region for κ-ΟR-induced autophagy under acute stress.
Finally, to confirm that dynorphin/κ-OR-induced autophagy-mediated changes in the structural reorganization of hippocampal synapses during stress may be rescued by nor-BNI, we measured the levels of synaptic proteins in hippocampi of stressed and naïve animals subjected to nor-BNI, or saline treatment. Our results demonstrated that spinophilin, PSD-95, and SNAP-25 protein levels were significantly reduced in stressed animals relative to the control ones (Fig. 7E). In contrast, the levels of these synaptic proteins in nor-BNI injected mice prior to FST were at the same levels as the control nor-BNI-injected ones devote of the stressor (Fig. 7E, lanes 5-8). Again, no significant alterations in cortical lysates of these proteins were detected, irrespective of the stress-related regime or nor-BNI administration status (Fig. 7F). Collectively, these results suggest that the endogenous dynorphin release due to the acute FST results in κ-ΟR-mediated induction of autophagy that in turn leads to aberrant hippocampal synaptic alterations.