mTOR-Dependent Synapse Formation Underlies the Rapid Antidepressant Effects of NMDA Antagonists

See allHide authors and affiliations

Science  20 Aug 2010:
Vol. 329, Issue 5994, pp. 959-964
DOI: 10.1126/science.1190287


The rapid antidepressant response after ketamine administration in treatment-resistant depressed patients suggests a possible new approach for treating mood disorders compared to the weeks or months required for standard medications. However, the mechanisms underlying this action of ketamine [a glutamate N-methyl-d-aspartic acid (NMDA) receptor antagonist] have not been identified. We observed that ketamine rapidly activated the mammalian target of rapamycin (mTOR) pathway, leading to increased synaptic signaling proteins and increased number and function of new spine synapses in the prefrontal cortex of rats. Moreover, blockade of mTOR signaling completely blocked ketamine induction of synaptogenesis and behavioral responses in models of depression. Our results demonstrate that these effects of ketamine are opposite to the synaptic deficits that result from exposure to stress and could contribute to the fast antidepressant actions of ketamine.

Major depressive disorder (MDD) afflicts about 17% of the population and is one of the leading causes of total disability and economic burden (1). Although there are medications that alleviate depressive symptoms, these have serious limitations. Most notably, available treatments require weeks or months to produce a therapeutic response, and only about one-third of patients respond to the first medication prescribed (2). In contrast, recent studies demonstrate that a single, low dose of a glutamate N-methyl-d-aspartic acid (NMDA) receptor antagonist produces a rapid (within hours) antidepressant response that lasts for up to 7 days (3, 4) and is effective in MDD patients who are resistant to traditional antidepressants (5). The mechanisms underlying rapid antidepressant actions are likely more complicated than simple NMDA receptor blockade and so far have not been identified. We carried out a series of studies to examine the cellular signaling pathways that mediate the behavioral actions of NMDA receptor blockade, focusing on signaling cascades known to rapidly influence synaptic plasticity (6).

The drug used for clinical trials is ketamine, a nonselective NMDA receptor antagonist (7). A low dose of ketamine (10 mg/kg), which is reported to have antidepressant actions in behavioral models of depression (7), rapidly activated the mammalian target of rapamycin (mTOR) signaling pathway in the prefrontal cortex (PFC) of rats (Fig. 1A). Activation of mTOR signaling was observed in a preparation enriched in synaptoneurosomes (fig. S1) and included increased levels of the phosphorylated and activated forms of eukaryotic initiation factor 4E binding protein 1 (4E-BP1), p70S6 kinase (p70S6K), and mTOR (Fig. 1A). Increased phosphorylation of 4E-BP1, p70S6K, and mTOR is transient, returning to basal levels by 2 hours after ketamine administration (Fig. 1A). In contrast, other antidepressants tested, including electroconvulsive seizure, imipramine, or fluoxetine, did not significantly influence mTOR signaling (fig. S2). Ketamine produced a similar rapid and transient increase in the phosphorylated and activated forms of extracellular signal–regulated kinase (ERK, including ERK1 and ERK2) and protein kinase B (PKB/Akt) (fig. S3A), growth factor signaling pathways that have been linked to activation of mTOR signaling (8). The activation of 4E-BP1, p70S6K, mTOR (Fig. 1B), ERK, and Akt (fig. S3B) was dose dependent, occurring at relatively low doses (5 to 10 mg/kg) that produce antidepressant behavioral actions but not at a higher anesthetic dose (7).

Fig. 1

Ketamine transiently and dose-dependently activates mTOR signaling in rat PFC. Values represent mean ± SEM [n = 4 animals; *P < 0.05; **P < 0.01; analysis of variance (ANOVA)]. (A) Time course of ketamine [10 mg/kg, intraperitoneally (ip)]–induced mTOR signaling determined by Western blot analysis of phospho-mTOR (pmTOR), phospho-4E-BP1 (p4E-BP1), and phospho-p70S6K (pp70S6K) in synaptoneurosomes of PFC. Levels of total mTOR, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and p70S6K were also determined. (B) Dose-dependent activation, determined 1 hour after ketamine administration, of pmTOR, p4E-BP1, and pp70S6K. (C) Pretreatment (10 min) with NBQX (10 mg/kg, ip) blocked ketamine (10 mg/kg, ip) activation of pmTOR, p4E-BP1, and pp70S6K, as well as upstream signaling kinases phospho-ERK (pERK) and phospho-Akt (pAkt) (analyzed 1 hour after ketamine). Levels of pERK1 and pERK2 were similarly regulated and were combined for quantitative analysis. DMSO, dimethyl sulfoxide. (D) Pretreatment (30 min) with inhibitors of ERK (U0126, 20 nmol, ICV) or PI-3k/Akt (LY294002, 20 nmol, ICV) abolished ketamine (10 mg/kg, ip) activation of mTOR signaling proteins (analyzed 1 hour after ketamine administration).

The antidepressant actions of ketamine have been reported to require glutamate α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPA) receptors (7). In line with this, we found that pretreatment (10 min) with a selective AMPA receptor inhibitor, 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX), completely blocked the induction of phosphorylated 4E-PB1, p70S6K, and mTOR, as well as ERK and Akt in the PFC (Fig. 1C and fig. S4). We assessed the possible involvement of ERK and Akt pathways in the activation of mTOR by intracerebroventricular (ICV) infusion of specific pharmacological antagonists. Inhibition of either ERK (U0126) or Akt [via inhibition of the upstream Akt activator phosphatidylinositol (PI3)-kinase by LY294002] blocked ketamine induction of phosphorylated 4E-BP1, p70567, and mTOR (Fig. 1D and fig. S5).

Activation of mTOR has been functionally linked with local protein synthesis in synapses, resulting in the production of proteins required for the formation, maturation, and function of new spine synapses (8). To determine whether ketamine induction of mTOR produces similar effects, we analyzed levels of several synaptic proteins in synaptoneurosomes from PFC. Ketamine administration increased levels of the postsynaptic proteins PSD95 and GluR1, as well as the presynaptic protein synapsin I (Fig. 2, A and B). Ketamine-induced increases in these synaptic proteins was delayed relative to ERK, Akt, and mTOR, peaking at 2 to 6 hours, but remained elevated up to 72 hours. Ketamine also induced expression of activity-regulated cytoskeletal-associated protein (Arc), with an intermediate (peak at 1 to 2 hours) but transient time course. The functional connection between mTOR activation and synaptic protein synthesis was examined by ICV infusion of selective mTOR inhibitor rapamycin, which completely blocked the induction of PSD95, GluR1, and synapsin I (Fig. 2B and fig. S6).

Fig. 2

Ketamine rapidly increases synaptic proteins and spine number. (A) Time course for ketamine (10 mg/kg, ip) induction of synaptic proteins Arc, synapsin I, PSD95, and GluR1 in synaptoneurosomes of PFC and (B) blockade by pretreatment (30 min) with a selective mTOR inhibitor, rapamycin (0.2 nmol, ICV) (values represent mean ± SEM, n = 4 to 6 animals; *P < 0.05; **P < 0.01; ANOVA). (C) Ketamine-increased spine density in mPFC, analyzed by two-photon microscopy 24 hours after treatment. Representative images are shown of high-magnification Z-stack projections of apical tuft segments of neurobiotin-labeled layer V pyramidal cells (scale bar indicates 5 μm). (D) Results are the mean ± SEM (eight cells from four rats in each group; *P < 0.05; **P < 0.01, t test). (E) Ketamine enhanced mPFC layer V pyramidal cell EPSC responses. Sample whole-cell voltage-clamp recordings of 5-HT and hypocretin-induced EPSCs in slices (24 hours after Ketamine treatment). (F) Cumulative probability distributions showing significant increases in amplitude (P < 0.0001, Kolmogorov-Smirnov test-z value = 6.5 for 5-HT and 6.7 for Hcrt) and (G) frequency of 5-HT– and hypocretin-induced EPSCs (n = 12 neurons per group; *P < 0.05, t test). Ketamine induction of spine density and function were blocked by rapamycin infusions [(C) to (G)].

An increase in synapse-associated proteins indicates that ketamine administration rapidly increases synapse and spine formation. We therefore examined the influence of ketamine on spine number and morphology by two-photon imaging of the apical tuft of prelabeled layer V medial PFC (mPFC) pyramidal neurons. Analysis of the images shows that ketamine administration increased spine density in both distal and proximal segments of the apical tuft 24 hours after ketamine injection (Fig. 2, C and D). Ketamine also increased the population of mushroom (large-diameter) spines (fig. S7), an indication of increased spine maturation and synaptic strengthening (9). Synaptic strengthening is supported by analysis of the electrophysiological properties of the same cells, demonstrating that excitatory postsynaptic current (EPSC) frequency responses to serotonin [5-hydroxytryptamine (5-HT)], which are mediated predominantly by cortical-cortical synapses in the apical tuft (10), are significantly increased by ketamine administration (Fig. 2, E to G). Responses to hypocretin and orexin, which occur selectively at apical thalamocortical synapses (11), are also increased (Fig. 2, E to G). In both cases, there is a significant increase in EPSC amplitude (Fig. 2F), consistent with the elevation in mushroom spines (fig. S7) and GluR1. Ketamine induction of spines and 5-HT and hypocretin-stimulated excitatory postsynaptic potentials (EPSPs) are blocked by infusion of rapamycin (Fig. 2, C to G, and fig. S7).

We next determined whether activation of mTOR signaling is required for the antidepressant behavioral actions of ketamine. As previously reported, a low dose of ketamine produced a rapid antidepressant effect in two well-established behavioral models of despair, the forced swim test (FST) and learned helplessness (LH) in response to inescapable stress (IES), as well as a model of anxiety, the novelty suppressed feeding test (NSFT) (Fig. 3, A to C). Traditional antidepressants are effective in these tests after acute (1 day, FST), subchronic (6 days, LH), and chronic (21 days, NSFT) treatment, whereas a single dose of ketamine was effective in all three tests. Moreover, infusion (ICV) of rapamycin completely abolished the antidepressant actions of ketamine in all three tests (Fig. 3, A to C). The role of mTOR signaling in the actions of ketamine was further examined by local infusion of rapamycin into the mPFC, which resulted in a complete blockade of the behavioral effects of ketamine in FST and NSFT tests (Fig. 3, D and E) (location of rapamycin infusions into the mPFC is shown in Fig. 3F). The effects of inhibitors of ERK and PI3k/Akt, which mediate ketamine induction of mTOR, were also examined. Infusion (ICV) of selective ERK (U0126) or PI3k/Akt (LY294002) inhibitors blocked the antidepressant actions of ketamine in the FST and NSFT (Fig. 3, G and H). Lastly, the behavioral actions of ketamine in the FST occurred with a dose-response profile similar to that for induction of ERK, Akt, and mTOR (Fig. 3I). This addresses a major point of discussion in the field as to whether an anesthetic dose of ketamine could produce an antidepressant effect.

Fig. 3

Rapid behavioral actions of ketamine require mTOR signaling. Rapamycin was infused (0.2 nmol, ICV) 30 min before ketamine (10 mg/kg, ip), and responses in the FST (A), NSFT (B), or LH (C) paradigms were determined. Values represent mean ± SEM (n = 6 to 8 animals; *P < 0.05; **P < 0.01; ANOVA). Infusion of rapamycin (0.01 nmol) into the mPFC blocked the antidepressant actions of ketamine (10 mg/kg, ip) in the FST (D) and NSFT (E). (F) Location of rapamycin infusions in the PFC (infusion sites are the same on the right and left sides because of the use of bilateral cannulae). Pretreatment with inhibitors of ERK (U0126, 20 nmol, ICV) or PI3 kinase/Akt (LY294002, 20 nmol, ICV) blocked the behavioral effects of ketamine in FST (G) and NSFT (H). (I) Low (10 mg/kg) but not a high (80 mg/kg) anesthetic dose of ketamine produced an antidepressant action in the FST.

Basic research and clinical studies demonstrate that stress and depression can cause atrophy of PFC pyramidal neurons in rodent models and in postmortem tissue of depressed subjects (10, 12), and brain-imaging studies report a decrease in PFC volume in depressed patients (13). Exposure to stress in the LH paradigm decreased levels of synapatic proteins GluR1, PSD95, and synapsin I, and a single dose of ketamine reversed this deficit in a rapamycin-sensitive manner (fig. S8). These results indicate that ketamine induction of synapse-associated protein synthesis provides a mechanism for rapid reversal of stress- and/or depression-mediated deficits and could enhance PFC-mediated connectivity and function.

Ketamine is a psychotomimetic drug with abuse potential, and a more selective agent would be desirable for clinical antidepressant use. There have been reports that selective NMDA receptor subunit 2B (NR2B) antagonists produce antidepressant actions similar to ketamine in both preclinical models and clinical trials (7, 14). Therefore, we examined the influence of Ro 25-6981, a NR2B-selective compound, on mTOR signaling and rapamycin-sensitive behaviors. Ro25-6981, at a dose (10 mg/kg) identified in the FST to be optimal (fig. S9), produced a transient activation of mTOR signaling (increased phosphorylation of 4E-BP1, p70S6K, and mTOR), as well as ERK and Akt signaling in synaptoneurosomes from PFC (Fig. 4, A and B). Ro 25-6981 administration also increased levels of the synaptic proteins Arc, PSD95, GluR1, and synapsin I in synaptoneurosomes (Fig. 4, A and B). Induction of Arc was more rapid and transient, whereas PSD95, GluR1, and synapsin I were delayed, similar to the delay observed after ketamine administration. Lastly, Ro 25-6981 administration produced rapid antidepressant effects (24 hours) in the FST and NSFT, which were blocked by preinfusion (ICV) of rapamycin (Fig. 4, C and D).

Fig. 4

Selective NR2B antagonist Ro 25-6981 activates mTOR signaling and produces rapamycin-sensitive behavioral effects. (A and B) Effects of Ro 25-6981 (10 mg/kg, ip) on pmTOR, p4E-BP1, pp70S6K, pERK, pAkt, Arc, synapsin I, PSD95, and GluR1 in PFC. (C and D) Pretreatment with rapamycin (0.2 nmol, ICV) abolished the actions of Ro 25-6981 in FST (C) and NSFT (D). Values represent mean ± SEM (n = 6 to 8 animals; *P < 0.05; **P < 0.01; ANOVA).

Fast activation of mTOR signaling resulting in rapid and sustained elevation of synapse-associated proteins and spine number in the PFC represents a mechanism for the rapid antidepressant actions of ketamine. These effects result in elevated 5-HT neurotransmission, a primary target of traditional antidepressants, although the latter can take weeks or months to induce a therapeutic response (2). The mechanisms underlying the induction of mTOR signaling are unclear, but the requirement for glutamate-AMPA receptor activation is consistent with the hypothesis that there is a subset of NMDA receptors, possibly on γ-aminobutyric acid (GABA)–releasing interneurons, that when antagonized, block GABA release lead to disinhibition of glutamate signaling (15). Further characterization of these actions of NMDA receptor blockade and the pathways that stimulate mTOR signaling and mediate the rapid induction of synapses in the PFC will provide novel therapeutic targets for antidepressant drug development.

Supporting Online Material

Materials and Methods

Figs. S1 to S9


References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. This work is supported by U.S. Public Health Service grants MH45481 and 2P01 MH25642 and by the Connecticut Mental Health Center.
View Abstract

Stay Connected to Science

Navigate This Article