Ketamine and Hydroxynorketamine: The complex balancing act within fear neural ensembles

Since its synthesis in 1962 as a rapid-acting general anesthetic, ketamine has had an unexpectedly multifaceted role within medicine and psychiatry. Ketamine was first approved by the Food and Drug Administration (FDA) in 1970 as a dissociative anesthetic for adult and pediatric surgical procedures. About 30 years later, an initial small study published in Biological Psychiatry showed that subanesthetic doses of IV ketamine can have significant rapid antidepressant effects in patients with major depressive disorder (MDD) (1). Since then, there has been an explosion of preclinical and clinical research to understand ketamine’s wide range of therapeutic actions. With intensive academic and pharmaceutical industrial investments into enantiomer-specific versions of ketamine and its analogs, intranasal esketamine (S-ketamine) was approved by FDA in 2019 for treatment-resistant depression (TRD). The development of this medication for psychiatric use has been a breakthrough for the field, spurring intense interest in developing other similarly rapid-acting psychiatric treatments within the psychedelics class.

Despite much attention and resources poured into translating ketamine use into clinical psychiatric use, we have yet to fully understand the mechanisms of action underlying the effects of ketamine and its active metabolites on the brain. This mechanistic understanding is a crucial and necessary step to developing safer, effective protocols for delivery of ketamine and related drugs to treat psychiatric disorders. Thus far, multiple mechanisms of action have been proposed and discussed in detail in other reviews (2,3). Briefly, there are two main mechanisms for ketamine’s acute mechanism of action. First is the disinhibition hypothesis, which posits that ketamine preferentially inhibits N-methyl-D-aspartate receptors (NMDARs) on cortical inhibitory GABAergic interneurons that lead to disinhibition of glutamatergic pyramidal neurons on the presynaptic side. Subsequently, there is a transient “glutamate burst”“ in the prefrontal cortex (PFC) and hippocampus. This then leads to AMPA receptor (AMPAR) activation, driving brain-derived neurotrophic factor (BDNF) release and its signaling through the TrkB receptor, eventually leading to increased mTOR-dependent synaptic protein synthesis and increased synaptic plasticity (2). A second complimentary mechanism of action involves the postsynaptic site, in which ketamine leads to inhibition of calcium-calmodulin-dependent kinase (eEF2 kinase) and allows for the rapid translation of postsynaptic BDNF mRNA, leading to BDNF release and subsequent rapid synapse potentiation (3). The main active ketamine metabolite, (2R,6R)-hydroxynorketamine (HNK), induces similar inhibition of eEF2 kinase activity, suggesting that this mechanism underlies the longer-lasting therapeutic effects of ketamine through its active metabolites. Additional mechanisms of action involve the activation of the mu-opioid receptor (MOR) system (4) or even direct binding and activation of TrkB (3). Despite these advances, the longer-lasting neuroplasticity effects of ketamine and its metabolites are still under active investigation and debate.

Notably, there are differential effects of the two ketamine enantiomers, with (S)-ketamine having ~4x higher NMDAR affinity than (R)-ketamine (2). However, (R)-ketamine shows comparable therapeutic efficacy with fewer dissociative side effects in animal models, suggesting that there are likely NMDAR-independent mechanisms that drive the antidepressant effects. Indeed, (2R,6R)-HNK demonstrates antidepressant-like effects in rodents by enhancing AMPAR signaling and activating BDNF/TrkB-mTOR pathways, promoting synaptogenesis in the PFC and ventral hippocampus (5). This effect is also sex-dependent, with more robust reversal of chronic stress-induced anhedonia in female rodents with (2R,6R)-HNK, with its effects requiring more intact AMPAR signaling (6). Despite these studies, it is notable that HNK alone cannot fully replicate ketamine’s behavioral or synaptic effects, implying that there is likely a spatiotemporal differential in the rapid effects of ketamine (likely more NMDAR-mediated) and the longer-lasting neural ensemble and network rewiring induced by ketamine’s metabolites (less NMDAR-dependent, more AMPAR dependent).

While distinct from its rapid antidepressant properties, ketamine’s role as a potential prophylactic agent has significant clinical implications. In this issue of Biological Psychiatry, Mastrodonato and colleagues (7) investigated how the prophylactic use of ketamine and (2S,6S)-hydroxynorketamine in male mice attenuated the learned fear response and more interestingly, affect brain regions and neural networks differentially. This work is preceded by prior studies showing that ketamine and/or HNK can be used as a prophylactic or as a protective treatment for stress and fear memory consolidation. In a key study, Chen and colleagues showed that prophylactic (2S,6S)-HNK, but not the (2R,6R)-HNK, led to buffering of learned fear only in male mice (8). This was an important finding, as it hinted that the (2S,6S)-HNK may work inherently different mechanistically, and that there is a sex-dependent response to specific HNK enantiomers. Mastrodonado et al. similarly found that prophylactic systemic injection of racemic ketamine and (2S,6S)-HNK in male mice 1 week before fear conditioning led to decreased learning of contextual fear, as shown by decreased freezing during re-exposure to the fearful contextual environment 5 days after the initial conditioning. They then used expression of c-fos as a measure of regional neuronal activity and correlated this expression pattern with the percentage of freezing during the contextual re-exposure. Modeled after another recent wide-scale microscopy studies to map cell-specific, region-specific activity by using c-fos expression post-acute injections of psychedelics, they also used whole brain imaging and various correlation analyses between brain regions to measure changes in connectivity with either ketamine or (2S,6S)-HNK prophylaxis.

Through this large-scale whole brain imaging analyses, Mastrodonado and colleagues found that although both drugs attenuated learned fear responses, the corresponding neural ensemble activities were not necessarily the same between the two drugs. With racemic ketamine, there was an increased connectivity between cortical and subcortical regions, with increased hippocampal activity correlated with freezing activity with contextual fear recall. This effect was specifically not seen with (2S,6S)-HNK. Furthermore, there were negative correlations between select cortical and subcortical regions with (2S,6S)-HNK, such as the amygdala and the prefrontal cortex (PFC), which were not present with racemic ketamine. Importantly, in addition to the known brain regions of prophylactic efficacy of ketamine and HNK, this wide-scale imaging and network analyses revealed several potential new nodes that mediate learned fear responses, such as retrosplenial cortex (RSP), insular cortex (IC), periaqueductal gray (PAG), nucleus reuniens (NR), and piriform cortex (PiR). For example, racemic ketamine and (2S,6S)-HNK affected the IC in opposite manners, which may lead to insights into the cognitive processing of fear-related emotions.

Translationally, these insights have implications for the clinical use of IV ketamine or HNK, which is currently still in phase 2 clinical trials as treatment for MDD. Clinically, the role of ketamine alone in treatment of PTSD has varied. In one study of civilian patients with chronic posttraumatic disorder (PTSD), 2 weeks of repeated IV ketamine administration (3 times per week) effectively ameliorated PTSD symptoms (9). However, with a larger and longer study in military veterans with PTSD, ketamine alone without concurrent behavioral interventions did not have a significant effect on PTSD symptoms, although it did have rapid antidepressant effects (10). Given these varied clinical effects, the Mastrodonado et al. study gives us some insight into how different forms and timing of ketamine can lead to differential effects. For example, understanding how the IC is differentially activated specifically with racemic ketamine or (2S,6S)-HNK across time may help refine which drug would be more efficacious to use prophylactically during trauma-focused psychotherapy to harness this region to improve cognitive control over PTSD symptoms. Based on prior clinical studies and the Mastrodonado et al. study, ketamine may be more useful for acute PTSD and mood-related symptomatic relief while (2S,6S)-HNK may be more helpful in facilitating the cognitive re-processing of traumarelated memories when these drugs are given prophylactically before trauma-focused psychotherapy. Although there are some initial studies looking at how ketamine can ameliorate mood and anxiety-related symptoms after an acute stressor, there are currently no definitive clinical studies done on prophylactic use of ketamine to prevent the progression to PTSD after an acute stressor.

While the study by Mastrodonado et al. represents an important step forward in understanding some of the spatiotemporal complexities of ketamine’s mechanism of action, there are still unresolved aspects. Given that racemic ketamine’s rapid efficacy hinges on NMDAR-driven glutamate dynamics and subsequent AMPAR/TrkB-dependent synaptogenesis, while HNK sustains effects via metabolite-specific AMPAR modulation, additional investigations to bridge these molecular mechanisms to systems-level functional restoration continues to be crucial. Aptly, this study highlights the importance of this wide-scale imaging and network analyses, as it has led to identification of previously underappreciated brain regions involved in contextual fear memory recall. Future work must also dissect sex-specific ketamine and its metabolite pharmacokinetics. Although Mastrodonado et al. appropriately mostly focused on male mice, given that prophylactic effects of HNK for learned fear is for the (2S,6S)-HNK enantiomer (8), it would have been helpful to see the network analyses in female mice, given there are clear sex differences in drug responses. In addition, network level analyses with (2R,6R)-HNK would have been helpful to have as a comparison to the both racemic ketamine and (2S,6S)-HNK, as it is the one HNK enantiomer that does not lead to attenuation of learned fear. Finally, whether ketamine or HNK-induced connectivity changes reflect transient excitation or enduring circuit reorganization remains unclear. Nevertheless, this study is a significant advance towards refining cross-species network biomarkers that leverage computational models to optimize the next-generation of rapid-acting antidepressants.