Ketamine and rapid-acting antidepressants: a new era in the battle against depression and suicide

Ketamine reverses the synaptic abnormalities caused by stress

Glutamate and NMDA receptors play an important role in cellular models of learning and memory, notably long-term potentiation (LTP), characterized by sustained synaptic strengthening in response to prior high-frequency stimulation. Actions of ketamine at NMDA receptors could influence NMDA function and synaptic plasticity in brain regions implicated in depression¹⁴. The possibility that synaptic changes are relevant to depression is supported by evidence that chronic stress, often used in rodent models of depression, causes significant loss of synapses and even retraction of apical dendrites in the prefrontal cortex (PFC) and hippocampus^15,16 (Figure 1). The relevance of stress-induced synaptic deficits in rodent models is in turn supported by evidence from brain imaging studies demonstrating decreased volume of PFC and hippocampus in depressed patients and post-mortem studies showing decreased synapse number¹⁷. In contrast to the effects of chronic stress, we found that a single dose of ketamine rapidly increases the number and function of spine synapses in layer V pyramidal neurons in the medial PFC (mPFC) and rapidly reverses the synaptic deficits of these neurons caused by 3 weeks of chronic stress exposure^18,19 (Figure 1). Increased levels of synaptic proteins, including levels of glutamate AMPA receptor GluA1, were observed after 2 hours, consistent with the onset of the therapeutic actions of ketamine. These studies demonstrate that ketamine rapidly increases synaptic function in the mPFC and that this reverses the synaptic pathophysiology of depression²⁰.

Figure 1. Schematic model for the initial cellular target sites of rapid-acting antidepressants and subsequent synaptic changes.

Stress and depression cause neuronal atrophy and decreased synapse number in the medial prefrontal cortex (PFC) and hippocampus that is associated with the depressive symptoms and behaviors in rodent models. Conversely, fast-acting antidepressants like ketamine rapidly increase synapse number and function and reverse the synaptic deficits caused by chronic stress. The synaptic actions of ketamine, as well as several other agents (that is, esketamine, [2R,6R]-hydroxynorketamine, mGluR2/3 antagonists LY341,495 and MGS0039, and scopolamine acting at acetylcholine muscarinic 1 receptors), are activity-dependent and are thought to result from a burst of glutamate via blockade of receptors on tonic firing GABA interneurons, resulting in disinhibition of glutamate transmission. This burst of glutamate causes activity-dependent release of brain-derived neurotrophic factor (BDNF), stimulation of TrkB-Akt and mechanistic target of rapamycin complex 1 (mTORC1) signaling, and rapid increases in synaptic protein synthesis that underlie new synapse formation. Negative allosteric modulators, including CP-101,606, CERC-301, and Ro 25-6981 agents like rapastinel, may increase synapse formation by enhancing NMDA function and thereby increasing BDNF release and downstream mTORC1 signaling. A role for mTORC1 is further supported by evidence that an agent that increases mTORC1 activity also produces synaptic and rapid antidepressant responses. In addition to these sites, there is evidence that the GABA_A-positive allosteric modulating agents brexanolone and SAGE-217 also produce rapid antidepressant responses. The intersection of these agents with the mechanisms underlying the rapid response to glutamatergic agents remains to be identified.

In addition to the effects of ketamine in the mPFC, there is a recent high-impact report demonstrating the effects of ketamine in the lateral habenula, a region that inhibits the major reward centers in the brain²¹. This study shows that depressive-like behavior in rodent models (rat congenital learned helplessness [LH] and mouse chronic restraint stress) is characterized by increased burst firing of neurons in the lateral habenula, which causes inhibition of the activity of the major reward and emotion pathways and the ventral tegmental dopamine system as well as the serotonin dorsal raphe neurons. This burst firing is driven by NMDA receptor activity as well as low-voltage-sensitive T-type calcium channels; ketamine, after either systemic or local intracerebral administration, is sufficient to block the burst firing and depressive behavior in the LH rats. It has been postulated that blockade of the anti-reward effects exerted by the lateral habenula could underlie the rapid antidepressant actions of ketamine and that the synaptic effects of ketamine in the mPFC could be more related to the sustained actions of ketamine, although further studies are needed to test this hypothesis.

Mechanisms underlying the synaptic actions of ketamine: disinhibition hypothesis

How might ketamine, an NMDA receptor channel blocker, cause a rapid increase in the number and function of spine synapses? There are different theories, but the one that has received the most attention is that ketamine actually increases glutamate transmission and causes an LTP-like enhancement of synapse formation in the mPFC¹⁴. This idea was first formulated when it was discovered that ketamine rapidly (30 minutes after systemic administration) increases extracellular glutamate in the mPFC of rodents²²; this effect was observed only at low subanesthetic doses of ketamine, whereas high anesthetic doses had no effect. Elevation of extracellular glutamate led to the hypothesis that low doses of ketamine selectively block NMDA receptors on GABA interneurons that inhibit glutamate transmission (Figure 1). This selectivity is based on the fact that GABA interneurons are tonic firing, which results in removal of the Mg²⁺ block of the NMDA receptor channel, allowing ketamine to enter, bind, and block the channel. Direct evidence for the disinhibition hypothesis has been reported recently in slice studies of hippocampus, which show that low concentrations of ketamine decrease inhibitory input onto pyramidal neurons and thereby lead to an increase in the synaptic drive of excitatory pyramidal neurons in the hippocampus²³. There is also an opposing “direct” hypothesis that the synaptic actions of ketamine are mediated by blockade of NMDA receptors on excitatory neurons, which leads to homeostatic control of synaptic activity²⁴. This homeostatic direct hypothesis is difficult to rationalize in the mPFC where there is a glutamate burst but could occur in other brain regions where there is no increase in glutamate.

Studies of the cellular signaling mechanisms underlying the synaptic actions of ketamine demonstrate a requirement for brain-derived neurotrophic factor (BDNF) and activation of pathways that increase the synthesis of synaptic proteins (Figure 1). The antidepressant actions of ketamine are blocked in conditional BDNF-knockout mice²⁴ and in mice with a knockin of the BDNF Val66Met allele, which blocks the processing- and activity-dependent release of BDNF²⁵. Further evidence that activity-dependent release of BDNF is required comes from studies demonstrating that infusion of a function-blocking antibody into the mPFC blocks the antidepressant actions of ketamine²⁶. Increased synthesis of synaptic proteins occurs via regulation of the mechanistic target of rapamycin complex 1 (mTORC1) as well as eukaryotic elongation factor 2 kinase (eEF2K). We have found that ketamine rapidly increases the phosphorylation of mTOR and downstream signaling proteins that stimulate synaptic protein synthesis and that the behavioral actions of ketamine are blocked by infusion of rapamycin, a selective inhibitor of mTORC1¹⁸ (Figure 1). Further evidence for mTORC1 is provided by recent studies demonstrating that direct stimulation of mTORC1 also produces rapid synaptic and antidepressant behavioral responses²⁷. Ketamine stimulation of mTORC1 signaling has been replicated in multiple laboratories^{18,19,28–42}, although there is a preliminary report of non-significant effects, raising a concern about the suitability of this target for drug screening⁴³. An alternative hypothesis is that ketamine blockade of NMDA receptors at rest results in deactivation of eEF2K and de-suppression of translation, thereby increasing BDNF expression in pyramidal neurons²⁴. There has also been evidence that ketamine influences the opioid system and inflammation processes, effects that could contribute to the actions of ketamine^44,45.

Stereoisomers and metabolites of ketamine

Efforts have also continued to develop and identify agents with the rapid and efficacious actions of ketamine but with fewer or no side effects. There is evidence from rodent studies that the (R)-stereoisomer of ketamine also produces rapid antidepressant effects in rodent models but without the ketamine-like side effects on sensory motor gating or conditioned place preference⁴⁹. The reduced side effects may be related to the decreased affinity of the (R)-isomer for the NMDA receptor. In addition, (S)- but not (R)-ketamine is reported to stimulate mTORC1 signaling⁴². There is also evidence that the ketamine metabolite (2R,6R)-hydroxynorketamine—(2R,6R)-HNK—produces rapid and sustained antidepressant actions in several rodent models, again without influencing sensory motor gating or conditioned place preference⁵⁰. This study found that (2R,6R)-HNK increased AMPA-GluA1 receptor expression and activity and increased levels of BDNF and phosphorylation of eEF2K; we have also found that the behavioral actions of (2R,6R)-HNK are blocked in BDNF Val66Met mice or by infusion of a function-blocking antibody and by infusion of rapamycin into the mPFC⁵¹. However, others have reported that the antidepressant actions of (2R,6R)-HNK are less potent and stable compared with (R)-ketamine⁵² or do not produce antidepressant actions in a rat LH model⁵³. Zanos and colleagues⁵⁰ report that they could find no evidence that (2R,6R)-HNK blocks NMDA receptors, including binding to known sites on the receptor complex or NMDA activity in hippocampal slices, so the initial cellular trigger remains to be identified. Nevertheless, another study reported that a high concentration of (2R,6R)-HNK (50 µM) produces partial blockade of synaptic NMDA receptors and downstream signaling compared with the same concentration of ketamine⁵⁴.

Other glutamatergic agents: GluN2B and mGluR2/3 antagonists and rapastinel

There is also evidence from rodent studies and preliminary clinical trials that antagonists that are selective for the GluN2B subunit of the NMDA receptor have antidepressant efficacy (Figure 2). Clinical studies first reported that a single dose of the selective GluN2B antagonist CP101,606 produces antidepressant actions in depressed patients, although significant effects were not observed until 5 days later⁵⁵. Recent clinical studies with another GluN2B-selective agent, CERC-301, have been mixed; early studies reported positive effects, and a more recent report was negative, although this could be because of the lower dose used for the latter study⁵⁶. Rodent studies demonstrate that a single dose of the GluN2B antagonist Ro 25-6981 produces antidepressant actions in several different antidepressant models, including the forced swim test (FST), LH, and novelty suppressed feeding test (NSFT), as well as in the chronic unpredictable stress-anhedonia model^18,57. Additional studies are needed to further test the efficacy and rapid actions of GluN2B-selective agents.

Another agent of interest is rapastinel, also referred to as GLYX-13, which was initially thought to act as an NMDA glycine site partial agonist (Figure 2). More recent studies indicate that rapastinel has activity that is similar to that of a glycine site partial agonist but acts via an allosteric site⁵⁸. Preclinical studies demonstrate that a single dose of rapastinel results in antidepressant actions in several rodent models, including the FST, NSFT, LH, chronic mild stress/anhedonia, and social defeat stress^27,59. In contrast to ketamine, rapastinel does not influence sensory motor gating or conditioned place preference. A single dose of rapastinel also increases synapse number and function in the mPFC and requires BDNF release and mTORC1 activation^27,41. However, one study reports that the antidepressant actions of rapastinel in a social defeat model as well as its effects on BDNF-TrkB signaling are not as long-lasting as those of (R)-ketamine⁶⁰. A double-blind randomized clinical trial reported that a single dose of rapastinel produces antidepressant actions that persisted for about 7 days⁶¹. Rapastinel has received breakthrough classification and currently is in phase III clinical trials. Another agent that works via the glycine site is AV-101, a prodrug (L-4-chlorokynurenine) that is transported into the brain where it is converted to 7-chlorokynurenic acid, a potent antagonist of the glycine B coagonist site of the NMDA receptor⁶²; AV-101 is in phase II clinical trials and has received fast-track status from the FDA.

Based on evidence that the actions of ketamine occur via a “glutamate burst” and the subsequent enhancement of synaptic function, other approaches to increase glutamate neurotransmission have been tested. Most notable are the metabotropic GluR2/3 antagonists, which increase glutamate activity by blocking presynaptic autoreceptors (Figure 1). Preclinical studies have reported that mGluR2/3 antagonists, notably LY341,495 and MGS0039, produce antidepressant effects in a number of rodent models, including the FST, NSFT, and a chronic unpredictable stress-anhedonia model, and also require mTORC1 signaling^29,63,64. Clinical trials are needed to determine whether blockade of mGluR2/3 receptors produces a therapeutic response in patients and to determine the side effect profile of these agents.