Inflammation, stress and depression: an exploration of ketamine’s therapeutic profile

Abstract

Well-established animal models of depression have described a proximal relationship between stress and central nervous system (CNS) inflammation – a relationship mirrored in the peripheral inflammatory biomarkers of individuals with depression. Evidence also suggests that stress-induced proinflammatory states can contribute to the neurobiology of treatment-resistant depression. Interestingly, ketamine, a rapid-acting antidepressant, can partially exert its therapeutic effects via anti-inflammatory actions on the hypothalamic-pituitary-adrenal (HPA) axis, the kynurenine pathway or by cytokine suppression. Further investigations into the relationship between ketamine, inflammation and stress could provide insight into ketamine’s unique therapeutic mechanisms and stimulate efforts to develop rapid-acting, anti-inflammatory-based antidepressants.

Introduction

Depression is the leading cause of disability worldwide, affecting 322 million people.¹ In the USA, research suggests that approximately one-third of sufferers have treatment-resistant depression (TRD), broadly defined as non-response to conventional antidepressants.² One of the primary obstacles to understanding depression is its characteristic heterogeneity in the course of illness, biomarkers, treatment response and genetic polymorphisms. As such, recent efforts within psychiatry have sought to establish clinically relevant biomarker and symptom-based subgroups under the umbrella of depressive disorders.

For over three decades, researchers have studied the relationship between depressive symptoms and inflammatory states.³ This interest began with the observation that chronic administration of interferon (IFN)-α – a proinflammatory cytokine used to treat hepatitis C and other malignancies – precipitated depressive symptoms that responded to standard antidepressant interventions.⁴ As first introduced by Dantzer and colleagues,⁵ the concept of ‘sickness-behavior’ has linked inflammation and depression through the hypothesis that constant activation of the peripheral immune system leads to circulating proinflammatory cytokines that increase immune signaling in the brain and subsequently worsen sickness behavior, predisposing a person to depression. Preclinical studies also found that peripheral immune system activation via systemic administration of lipopolysaccharide (LPS), an endotoxin, reliably triggered ‘depressive-like’ behaviors in rodents.⁶ Acute and chronic stressors that play an integral part in the etiology of depression⁷ also reliably trigger inflammatory responses.⁸

Indeed, emerging evidence from population-based studies supports the notion that chronic, low-grade inflammation – although not present in all individuals with depression – could nevertheless play a key part in the pathophysiology of depression for a subset of patients.⁹ Data from longitudinal studies suggest that dysregulation of the inflammatory response is associated with a more severe course of illness, higher recurrence of depressive symptoms and worse outcomes, including impaired brain connectivity within motivation and reward circuits,¹⁰ increased suicidality¹¹ and, notably, greater resistance to conventional therapies.¹² Reward circuits can also impact a hallmark symptom of TRD: anhedonia, which has also been consistently linked to inflammation.¹³ Other factors such as obesity and other conditions associated with chronic inflammation also appear to increase the development of inflammation-associated sickness and depressive symptoms, as well as their persistence.⁵

Subanesthetic doses of the glutamatergic modulator racemic ketamine, as well as its enantiomers, have consistently been shown to exert rapid-acting antidepressant effects in patients with TRD and treatment-resistant bipolar depression (reviewed in ¹⁴). Ketamine has also been found to successfully treat traditionally treatment-refractive symptom domains such as anhedonia, suicidality and amotivation.^15,16 Within the context of this review, it is important to note that, although researchers primarily attribute ketamine’s therapeutic effects to upregulated neuroplasticity induced via glutamatergic modulation,¹⁷ growing evidence suggests that it might also regulate acute and chronic inflammatory reactions and restore immune homeostasis.^18,19

This review of previous and emerging research discusses the links between depression, stress and inflammation, particularly inflammation as an potential indicator of TRD, and summarizes the preclinical and clinical evidence for ketamine’s anti-inflammatory and immunomodulatory properties in the context of its antidepressant effects. Potential mediators of the process – including the kynurenine pathway and the hypothalamic-pituitary-adrenal (HPA) axis – are also discussed, as is the hypothesis that ketamine’s unique ability to reduce depressive symptoms in TRD could in fact be caused by its ability to reduce stress-induced inflammation.

Chronic stress, depression and inflammation: an overview

Stress is an inherent physiologically or emotionally coordinated response that activates processes in the body to maintain homeostasis after threatening stimuli or, under acute stress conditions, helps anticipate challenges or response to dangerous situations. Chronic stress is loosely defined as a sustained threat lasting at least several weeks that is accompanied by a resulting negative emotional state and deleterious effects on body systems. Under chronic, prolonged stress conditions, the brain and body lose their ability to restore homeostasis. The link between inflammation and chronic stress probably results from an evolutionary adaptation.⁸ In prehistoric environments, this connection between the perception of danger and the risk of subsequent tissue injury or pathogen exposure was believed to be so reliable that evolution favored anticipatory inflammatory responses to many environmental stressors, including psychosocial stressors. In the context of the present review, chronic stress is known to be a major risk factor for depression.⁷

The relationship between chronic stress and depression holds true in preclinical models, where chronic stress protocols [e.g., social defeat, unpredictable mild stress and chronic corticosterone (CORT) administration] are the gold standard for producing depressive-like behaviors in animals, including symptom profiles such as learned helplessness and anhedonia.²⁰ In animal models, the upregulation of stress hormones was found to robustly increase inflammatory markers such as tumor necrosis factor (TNF)-α and interleukin (IL)-1β.²¹ Preclinical studies also found that chronic stress induces central nervous system (CNS) inflammation characterized by the secretion of cytokines and neuroinflammation.²² Interestingly, one study found that chronic social instability stress did not alter hippocampal proinflammatory cytokines; however, the study was conducted in females, suggesting potential gender differences in the links between chronic stress and inflammation.²³

Multiple clinical studies have reported elevated levels of proinflammatory cytokines in individuals with depression. For instance, meta-analyses found that elevated levels of C-reactive protein (CRP) – a common marker of inflammation – predicted subsequent depressive symptoms²⁴ and were strongly associated with a diagnosis of depression.²⁵ Another recent meta-analysis of individuals with depression found that a quarter of participants had low-grade inflammation (CRP >3 mg/l), and half had elevated CRP levels (CRP >1 mg/l).²⁶ Other meta-analyses reported cerebrospinal fluid (CSF) and peripheral elevations of other proinflammatory markers such as IL-6, IL-8 and TNF-α.^27,28 Supporting the notion that higher levels of inflammation play a causative part in depression, one longitudinal study found that participants with elevated IL-6 and CRP levels at age nine were more likely to be evaluated as depressed at age 18.²⁹ Nevertheless, many other studies have found no such association between increased levels of CRP and IL-6 (reviewed in ³⁰), suggesting that any putative relationship between depression and inflammation remains unclear. Such differences could arise from disparities in the chronicity of MDD, given that some evidence suggests that the early phases of MDD might not present with heightened inflammation, which can predict poor response to the initial prescription of antidepressants.^31,32 Baseline low-grade inflammation has been shown to mediate endothelial dysfunction, which in turn can predict persistent depressive symptoms and impact the chronicity of MDD.³³ In a sample from The Netherlands, higher baseline inflammatory markers predicted a subsequent chronic course of illness in women, and depressive severity predicted subsequent higher levels of IL-6 in men and women.³⁴ Thus, the presence of increased inflammatory markers could represent a distinct subgroup within the depressive diagnostic label, and the mixed results warrant further investigation.

Genetic differences and epigenetic changes are also major drivers of response to chronic stress. Genetically, studies have determined that ~40% of MDD is heritable.³⁵ The past two decades of research have also determined that chronic stress can produce significant epigenetic changes that contribute to the pathophysiology of depression.³⁶ As the major connection between genetics and the environment, epigenetic changes affect the availability of DNA to transcription factors through DNA methylation and histone modifications, including methylation, acetylation, phosphorylation and more.³⁷ Major findings on inflammation-promoting epigenetic changes after chronic stress include histone deacetylase inhibitors, which have been shown to ameliorate depressive-like behavior and microglial activation.^38–40 In addition, chronic corticosterone-induced expression of the gene Nfkb1 was shown to be accompanied by upregulation of TNF-α and IL-1β.⁴¹ Although an in-depth discussion of epigenetic mechanisms is beyond the scope of this review, these mechanisms were recently reviewed extensively (see ⁴²).

It is also important to acknowledge the potential role of gender in the interplay between chronic stress, depression and inflammation. Rates of depression have consistently been found to be two-to-three-fold higher in females,⁴³ and clinical research also suggests that women might be particularly vulnerable to the effects inflammation on depressive symptomatology.⁴⁴ By contrast, some preclinical models found that males were more vulnerable to developing depressive-like behaviors after inflammatory insult.⁴⁵ Throughout this review, gender differences will be discussed wherever possible.

The relationship between inflammation and TRD

As noted above, evidence suggests that individuals with MDD with heightened inflammatory markers can constitute a subpopulation uniquely associated with treatment-refractory symptoms.⁴⁶ Notably, the ability to model TRD in animals is essential for future contributions to preclinical research exploring the underlying mechanisms of inflammation in treatment-resistance; but this field is still in its infancy.⁴⁷ Proposed solutions include animal models that receive cyclic exposure to chronic stress and/or using animals that do not respond to conventional antidepressants, such as Wistar–Kyoto rats.⁴⁸ Another potential proxy would be the measurement of behaviors known to be more present in TRD, such as anhedonia, anergia and motivation.

In clinical research, body mass index (BMI)-corrected serum CRP levels were recently found to be significantly elevated in TRD participants relative to treatment-responsive MDD participants, unmedicated MDD participants and healthy volunteers.⁴⁹ These findings are complemented by two randomized, controlled trials that found that baseline CRP levels predicted lack of response to conventional antidepressants.^50,51 Another study found distinct results in whole-blood samples, with a significant upregulation in mRNA-indicated inflammasome activation and glucocorticoid resistance in the MDD population (untreated versus treatment-responsive versus TRD). Of the mRNAs identified, six (P2RX7, IL-1β, IL-6, TNF-α, CXCL12 and GR) differentiated between TRD and treatment-responsive subgroups.⁵² By contrast, another study found no evidence of large inflammatory differences in the peripheral blood mononuclear cells (PBMCs) of healthy volunteers versus MDD patients (untreated versus treatment-responsive versus TRD) but did find strong evidence of increased biological aging in the MDD sample.⁵³

One study of unmedicated MDD participants found that those who, on average, had failed to respond to three or more antidepressant trials had significantly higher levels of CRP, IL-6, TNF receptor 2 (sTNF-R2) and TNF-α than those who, on average, had failed to respond to less than one trial.⁴⁶ A meta-analysis found that higher baseline levels of inflammatory markers in general were associated with poor treatment response, and that high TNF-α levels in particular were associated with TRD.⁵⁴ An analysis of participants with MDD and bipolar depression who participated in a randomized, controlled trial of escitalopram versus nortriptyline found that cutoffs for absolute mRNA levels of IL-1β and macrophage migration inhibitory factor (MIF) in blood accurately predicted 100% of the non-responders in their study.⁵⁵ Interestingly, a randomized, controlled trial of the anti-inflammatory agent infliximab found that its antidepressant effects were specific to a subset of TRD participants with elevated baseline plasma CRP levels >5mg/l⁵⁶; because this impact was not consistent with results observed in individuals with bipolar I and II depression, it suggests a potential unique efficacy for TRD.⁵⁷ Finally, adjunctive use of the anti-glucocorticoid therapeutic metyrapone actually increased IL-6 levels in individuals diagnosed with TRD, an increase associated with poorer outcomes to treatment; this finding was hypothesized to result from potential glucocorticoid system overcompensation.⁵⁸

Imaging studies are also beginning to confirm that this peripheral inflammation is mirrored in the brain itself. Positron emission tomography (PET) studies of translocator protein 18 kDa (TSPO) – a biomarker of neuroinflammation – have typically reported greater TSPO binding in the prefrontal cortex (PFC) and anterior cingulate cortex (ACC) of individuals experiencing a major depressive episode.⁵⁹ In an open-label trial of TRD participants who received the anti-inflammatory agent celecoxib, investigators plotted the reduction in Hamilton Depression Rating Scale (HAM-D) score against baseline TSPO volume in the PFC and ACC and found that HAM-D scores rapidly dropped post-treatment as baseline TSPO distribution volume decreased.⁶⁰ A recent parallel study measured the impact of minocycline, a tetracycline antibiotic with anti-inflammatory properties, on TRD participants experiencing a major depressive episode; minocycline did not significantly impact TSPO binding,⁶¹ although another study found that it significantly decreased HAM-D scores in participants with elevated CRP levels (CRP ≥3mg/ml).⁶² Other studies investigating minocycline as an adjunctive treatment for TRD found no significant change in depressive symptoms.⁶³ Finally, increases in immune factors after ex vivo LPS stimulation of PBMCs were associated with reduced reward anticipation in the ventral striatum, as measured via functional magnetic resonance imaging (fMRI).⁶⁴ This builds on previous research that found that endotoxin administration to healthy volunteers significantly increased depressed mood over time and reduced ventral striatum responses to reward.⁶⁵ This effect could also be gender-dependent, given that females demonstrated greater reductions in ventral striatum activity in response to reward.⁶⁶ Inflammation can mediate motivational behavioral responses by dampening dopamine activity within reward circuits, resulting in disrupted frontostriatal functional connectivity.⁶⁷ Inflammatory processes are therefore well-situated to influence the neural circuits underlying motivational symptoms related to anhedonia. This is particularly important because behavioral responses to reward and social stimuli in patients with anhedonia have been associated with suicidality⁶⁸ and treatment resistance.⁶⁹ Interestingly, depressive symptoms such as reduced motivation and anhedonia correlate significantly with central IL-6 soluble receptor (IL-6sr)⁷⁰ as well as peripheral CRP levels.^10,70 A resting-state fMRI study of depressed participants found that plasma CRP levels correlated with decreased connectivity between the ventral striatum and ventromedial PFC (vmPFC), and that this change in connectivity was itself correlated with the severity of anhedonia.¹⁰ Consistent with this finding, administration of IFN-α for 4–6 weeks in 14 individuals with hepatitis C not only induced anhedonia but also reduced bilateral activation of the ventral striatum in an fMRI reward task⁷¹; change in striatal activity again correlated with anhedonia scores.

Relatedly, reduced motivation has been correlated with central levels of TNF-α.⁷⁰ Anhedonia, anergia and amotivation all fall under the symptom interest–activity dimension of depression; in the large Genome-Based Therapeutic Drugs for Depression (GENDEP) (n = 811) and Sequenced Treatment Alternatives to Relieve Depression (STAR*D) (n = 3637) studies, this dimension was shown to best predict poor antidepressant response.⁷²

Glutamate can also modulate the interplay between inflammation and depression. Higher glutamate release from microglial cells appears to increase concentrations of extracellular glutamate, promote maladaptive glutamate metabolism, contribute to loss of synaptic fidelity and decrease the specificity of neurotransmission – all of which can worsen depressive-like behaviors and increase circuit dysfunction.⁷³ Although most research in this area has focused on chronic stress, acute traumatic stress can similarly provoke glutamatergic signaling dysfunction.⁷⁴ Administration of the proinflammatory cytokine IFN-α increased glutamate concentrations in the dorsal anterior cingulate cortex and basal ganglia.^75,76 Notably, individuals with depression who also had high concentrations of plasma CRP and high levels of basal ganglia glutamate were significantly more likely to have more-severe symptom presentations of anhedonia and cognitive slowing.⁷⁷ In analyses of postmortem tissue, glutamate was also found to be increased in the frontal cortex of those diagnosed with MDD or bipolar disorder.⁷⁸ Increased mGluR2/3 expression was also found in the PFC of individuals with MDD, a finding paralleled in a Rhesus monkey model of depression.⁷⁹ By contrast, mGluR5 expression was found to be decreased in the PFC of postmortem MDD tissue, and participants with MDD had significantly lower levels of mGluR5 binding in multiple regions, as ascertained through PET imaging.⁸⁰ mRNA alterations related to glutamate signaling pathways have also been found in the locus coeruleus and hippocampus of individuals with MDD.^81,82

Another important indicator of glutamatergic activity implicated in depression is the glutamate–glutamine cycle. Glutamate is synthesized from glutamine in neurons and, after release into the synapse, is taken up by sodium-dependent glutamate transporters (EAAT1 and EEAT2) located on astrocytes. This scavenged glutamate is then converted back to glutamine and transported back to neurons by neutral amino acid transporters (SNAT1 and SNAT2), continuing the cycle. In the CSF, glutamine levels have been found to be upregulated in individuals with MDD.⁸³ In addition, increases in the ratio of glutamine to glutamate have been found in the CSF of individuals with MDD compared with healthy volunteers – a ratio that correlated with the severity of depression in a three-year follow-up.⁸⁴ Changes in this cycle have also been implicated in suicide; specifically, differential changes in neuronal and astrocytic components of the cycle were observed in postmortem tissue obtained from healthy volunteers, individuals with MDD and those who died by suicide.⁸⁵

Despite these intriguing findings, the role of glutamate in depression remains unclear. As an example, magnetic resonance spectroscopy (MRS)⁸⁶ studies found decreased levels of glutamate, or no differences at all,⁸⁷ in individuals with depression. Other studies found that the glutamatergic neurons of individuals with depression exhibited decreased mitochondrial energy production.⁸⁸ One important caveat is that glutamate is often measured using Glx – a composite measure that includes glutamate and glutamine. At least one study that separated these measures found no significant differences in glutamate levels in participants with depression.⁸⁹

Preliminary research is also investigating the response of inflammatory proteins to psychological therapy. In one study, poor response to treatment was associated with higher baseline levels of TNF-α, IL-6 and soluble intracellular adhesion molecule-1 and with higher post-therapeutic levels of CRP, thymus and activation-regulated chemokine, and macrophage chemoattractant protein-4.⁹⁰ At least one review of the literature also reported a general reduction in inflammation after cognitive behavioral therapy for depression.⁹¹

Taken together, inflammatory markers seem to cause bona fide alterations in brain network activity that can, in turn, cause depressive symptoms. Thus, the evidence suggests that inflammation contributes to depressive pathology in at least some cases and that determining potential mediators of the stress response can inform the development of therapeutic interventions.

Potential mediators between depression and inflammation

HPA axis

HPA axis hyperactivity is one of the most consistent findings in studies exploring the underlying pathophysiology of depression. In healthy states, the HPA axis is activated by acute stress, stimulating the release of corticotropin-releasing hormone (CRH) and vasopressin (AVP) from the hypothalamus. This, in turn, stimulates the release of adrenocorticotrophic hormone (ACTH) and glucocorticoids (primarily cortisol in humans and corticosterone in rodents). After an acute stressor, these glucocorticoids interact with their widely expressed receptors (either mineralocorticoid or glucocorticoid receptors), some of which interact with the hypothalamus to form a negative feedback loop to shut off HPA axis activity. Chronic stress disrupts this feedback loop, causing a downregulation of glucocorticoid receptors that impairs the ability to shut off the HPA axis, leading to dysfunctional hyperactivity.⁹²

Many individuals with depression exhibit HPA axis dysfunction, such as continuously elevated levels of cortisol and CRH.⁹³ This hypersecretion can cause hypercortisolism and, as a result, decreased dopaminergic reward-system responsivity.⁹⁴ In females, increased hair cortisol concentrations were associated with poor performance on measures of cognition and memory, an association that appeared to be mediated by CRP levels.⁹⁵ Early-life adversity has also been shown to increase vulnerability to acute social stress, an effect mediated by HPA-axis and immune activation,⁹⁶ and this was also found to impact later diurnal HPA axis functioning in adulthood.⁹⁷

One of the most compelling theories regarding the clinical relevance of inflammation in depression is that inflammation can differentiate depressive subtypes and mediate specific symptoms. For example, a recent study found that biomarkers of HPA axis activity and subsequent inflammation (such as cortisol and CRP) were more strongly associated with the presence of somatic symptoms rather than cognitive-affective symptoms.⁹⁸ For instance, a recent review found that cancer patients – who are significantly more likely to have depressive symptoms and a worsened symptom profile – can exhibit increased depressive-like behaviors owing to hyperactivity of the HPA axis caused by cancer and anticancer treatments.⁹⁹ In a CORT-injected mouse model, the antidepressant-like effects of catalpol, an iridoid glucoside, also appeared to be mediated through the HPA axis, suppressing levels of CORT, ACTH and CRH.¹⁰⁰

The kynurenine pathway

One hypothesis of inflammation-mediated depressive pathogenesis is that stress and inflammatory cytokines promote kynurenine pathway signaling.¹⁰¹ Tryptophan, a precursor for serotonin synthesis, is competitively consumed by the kynurenine pathway. One of the rate-limiting enzymes of this pathway: indoleamine-2,3-dioxygenase (IDO), is expressed mainly in immune and neuronal cells and induced by cytokines, cortisol and LPS, generally indicating a proinflammatory state.¹⁰² Tryptophan-2,3-dioxygenase (TDO), the other main enzyme in the kynurenine pathway that catalyzes tryptophan catabolism, is also induced under proinflammatory states.¹⁰³ Thus, increased cytokine and cortisol levels can reduce serotonin levels via tryptophan depletion, a process that has been experimentally shown to induce depressive symptoms in vulnerable persons,¹⁰⁴ although these findings have not always been consistent.¹⁰⁵ Tryptophan–kynurenine metabolism can also provide a link between the gut–brain axis and inflammatory bowel disease and depression, two disorders that are strongly associated with one another.¹⁰⁶ Acute and chronic stress also impact the rate-limiting enzymes involved in the tryptophan–kynurenine balance (Figure 1). For example, IFN-γ and IL-1β are potent inducers of IDO and TDO, which are highly impacted by immune activation in the brain. In addition, stress-induced corticosterone release and the consequent cascading activation of hepatic TDO to tryptophan metabolism ultimately lead to the production of kynurenine, which provokes a depression-related behavioral phenotype.¹⁰⁷ IDO and TDO can therefore represent promising targets for the treatment of depression associated with stress-related disorders marked by kynurenine pathway activation (Figure 1).

Figure 1.

The hypothesized impact of ketamine on stress and inflammatory pathways. Chronic stress leads to overactivation of the hypothalamic-pituitary-adrenal (HPA) axis, which increases levels of corticotropin-releasing hormone (CRH) and cortisol while decreasing expression of glucocorticoid receptors (GRs). This decrease in GR expression prevents the shut-off of the HPA axis, leading to prolonged activation that can have negative consequences. Ketamine, an N-methyl-d-aspartate receptor (NMDAR) antagonist, appears to mediate this stress response by increasing the number of GRs. Studies examining ketamine’s effect on cortisol levels have yielded mixed results. Under chronic stress conditions, the kynurenine pathway, another potential mediator between stress and inflammation, demonstrates increased levels of indoleamine-2,3-dioxygenase (IDO), tryptophan-2,3-dioxygenase (TDO) and quinolinic acid, as well as decreased levels of kynurenic acid. Ketamine decreases IDO levels and the ratio of kynurenine:tryptophan through indirect mechanisms while blocking the action of quinolinic acid through direct NMDAR antagonism. Ketamine also decreases proinflammatory cytokine levels (increased by chronic stress) through the NLPR3-inflammasome pathway, decreasing microglial activation via TLR/p38 signaling, P2X7 receptors and signal transducer and activator of transcription 3 (STAT3) activation, as well as switching macrophages to the anti-inflammatory M2 phenotype. Figure created using Biorender.

Kynurenine pathway products are biologically active. Kynurenic acid (KA) is considered to be neuroprotective¹⁰⁸ and a potential therapeutic target for drug development in mood disorders. Another product, quinolinic acid (QA), is an endogenous neurotoxin that generates free radicals¹⁰⁹ and causes excitotoxicity by inducing the release and inhibiting the reuptake of glutamate.¹¹⁰ One major component of the kynurenine pathway is its ability to affect the glutamatergic system, where it directly and indirectly influences ionotropic and metabotropic glutamate receptors and vesicular glutamate transport.¹¹¹ These effects are hypothesized to act as a main link between chronic stress, depression and inflammation.¹¹² For instance, QA directly activates N-methyl-d-aspartate receptors (NMDARs), increases synaptosomal glutamate release and inhibits glutamate uptake, making it uniquely placed to mediate interactions between ketamine and inflammation.¹¹³ KA and QA are metabolized from kynurenine by astrocytes and microglia, respectively, and evidence suggests that individuals with MDD have reduced astrocyte density¹¹⁴ and function¹¹⁵ along with increased microglial activation and number.¹¹⁶

Supporting the clinical relevance of this pathway, studies have reported higher ratios of kynurenine to tryptophan levels,^117,118 lower levels of KA¹¹⁹ and lower KA:QA ratios¹²⁰ at baseline in MDD participants. In addition, QA elevations were found in the CSF of recent suicide attempters,¹²¹ and more QA-positive cells were found in the brain of suicide decedents.¹²² Finally, altered peripheral ratios of KA:QA levels were shown to correlate with increased anhedonia in MDD participants¹²³ as well as with depression and fatigue in cancer patients.¹²⁴

Inflammation-mediated tryptophan metabolism in the gut microbiome has also been implicated in depression.¹²⁵ Intestinal inflammation, particularly via increased IFN-γ, was able to induce IDO, shifting tryptophan metabolism toward the production of kynurenine rather than serotonin.^126,127 Kynurenine is able to cross the blood–brain barrier,¹²⁸ suggesting that alterations from the gut microbiome could enter the blood circulation and impact levels of kynurenine and kynurenic metabolites in the brain. Fecal microbiota transplantation from participants with depression into a rat model induced depressive-like behaviors that were associated with increased levels of inflammatory markers (IL-6, TNF-α, CRP) as well as an increased kynurenine:tryptophan ratio.¹²⁹ This finding was paralleled in another fecal microbiota transplant from chronically stressed mice to control mice that increased IDO1 expression and proinflammatory cytokine levels.¹³⁰ For a recent summary of studies exploring tryptophan metabolism, gut microbiota and depression, we refer the interested reader to a recent summary by Lukic and colleagues.¹³¹ Taken together, this evidence suggests that the kynurenine pathway could play a key part in mediating the links between inflammation and depression.

Ketamine and inflammation

The NMDAR antagonist ketamine is uniquely effective for treating TRD, with a response rate ranging from 25% to 85% at 24 h post-infusion.¹³² It has also been shown to effectively reduce suicidal ideation and anhedonia,^15,16 as well as fatigue and amotivation symptoms.^133,134 Recent clinical and preclinical evidence indicates that, at antidepressant doses, ketamine can exert these unique therapeutic effects in part by modulating inflammation.^19,117 It is important to note that most of the studies described below reflect acute, not chronic, ketamine administration, which could affect the interpretation of results.

Although considerable volumes of research, and clinical studies in particular, have focused on the (S)-ketamine enantiomer, significant preclinical work has begun to explore the mechanisms behind ketamine’s (R)-enantiomer and various ketamine metabolites. Briefly, (R)-ketamine binds with around fourfold less affinity or potency to inhibit NMDARs than the (S)-enantiomer and has shown promise in terms of its antidepressant effects and reduced adverse side effect profile in animal models for depression¹³⁵ and in early clinical trials.¹³⁶ In addition, a recent open-label study found that (R)-ketamine had rapid and sustained antidepressant effects,¹³⁷ and further clinical trials are currently underway by multiple companies. Two major ketamine metabolites [norketamine and hydroxynorketamine (HNK)], which have different binding capacities to NMDARs,¹³⁸ have also recently been studied in animal models of depression and early-phase clinical trials. Thus far, these metabolites appear to have potential antidepressant-like effects when administered directly.^139–141 In addition, varying levels of ketamine metabolites in plasma and CSF have been correlated with clinical response to ketamine.¹⁴² Although an in-depth discussion of the hypothesized distinct mechanisms underlying the effects of ketamine, its enantiomers and its metabolites is beyond the scope of this review, these have all recently been reviewed extensively.¹⁴³ In particular, differences between enantiomers and metabolites should be considered when defining the role that ketamine has in inflammatory processes.

Preclinical evidence of ketamine’s anti-inflammatory effects

Substantial preclinical evidence suggests that ketamine reduces inflammation by regulating the immune system. In vitro ketamine application to rodent glial cells¹⁴⁴ and macrophages¹⁴⁵ attenuated markers and mediators of LPS-induced inflammatory responses, such as TNF-α, IL-1β, high mobility group box 1 (HMGB1), nitric oxide (NO), inducible nitric oxide synthase (iNOS) and prostaglandin E-2.

In Wistar–Kyoto rats, a proposed model of TRD,¹⁴⁶ ketamine showed rapid-acting antidepressant effects by inhibiting the NLR family pyrin domain containing 3 (NLRP3) inflammasome pathway, an effect blocked by application of an autophagy inhibitor.¹⁴⁷ In addition, low-dose ketamine effectively mediated the gut microbiome bacterial population associated with inflammation in Wistar–Kyoto rats.¹⁴⁸

In other animal models, administration of intraperitoneal ketamine had prophylactic effects against LPS- and chronic-stress-induced depressive behaviors,^149–151 an effect that appears unique to ketamine versus other NMDAR antagonists.¹⁵² For example, ketamine administered one week before stressor application most effectively reduced freezing behavior in a contextual fear conditioning task^153,154 and improved performance on the tail suspension and splash tests.¹⁵⁵

Ketamine’s prophylactic effects appear to be, at least in part, mediated via inflammatory signaling, such as the NLRP3 inflammasome pathway after LPS or TNF-α adminstration.¹⁵⁶ Cyclooxygenase-2 (COX-2) expression was also shown to be reduced in a gender-specific manner after ketamine administration.¹⁵³ (R)-Ketamine has also been shown to have prophylactic effects in mice exposed to chronic stress, improving behavioral and biological outcomes, such as altered gene expression of Bdnf and Mecp2 mediated through miR-132-5p activity.¹⁵⁷ Interestingly, this prophylactic effect did not generalize to its metabolites [(2R,6R)-HNK and (R)-norketamine]; however, the prophylactic effects of different metabolites could be gender-specific, because ovarian hormones are required for the protective effects of ketamine and (2R,6R)-HNK in female mice.¹⁵⁸ These prophylactic effects of (R)-ketamine in chronic stress and inflammation-based models appear to be mediated through microRNA-149 and nuclear factor of activated T cells 4 (NFATc4), which play a part in mediating cytokine expression.^159,160 Together, these results emphasize the importance of enantiomer-specific research when discussing the impact of ketamine on inflammatory signaling.

Concurrent with the aforementioned behavioral changes, ketamine also attenuated plasma cytokine elevations¹⁶¹ and cytokine expression in rodent tissue samples from the PFC, hippocampus, cerebellum and spinal cord.^151,162 Differences in ketamine’s prophylactic efficacy can arise from different experimental paradigms, such as timing of ketamine administration, which should be considered carefully when designing future experiments. For instance, most, but not all,¹⁵⁰ studies focusing on ketamine’s prophylactic effects have administered this agent one week before the stressor to ascertain more long-term effects rather than immediate effects. Age and gender should also be considered in future clinical applications, given that some preclinical research found greater prophylactic efficacy in adolescents¹⁵³ as well as differences in female response.¹⁵⁸

Interestingly, ketamine appears to act directly on immune cells. For instance, S-ketamine – the S-enantiomer of racemic ketamine – decreased microglial activity levels in the CNS after chronic stress exposure.¹⁵¹ An immunohistochemistry study performed on rodent hippocampus samples found that ketamine reversed stress-induced activation of microglia caused by chronic restraint by downregulating Toll-like receptor (TLR)/p38 pathway activation and P2X7 receptors.¹⁶² In chronically stressed mice, pharmacological inhibition of TGF-β1 signaling in microglia eliminated (R)-ketamine’s antidepressant effects.¹⁶³ In addition, ketamine and its antidepressant metabolites altered the localization of signal transducer and activation of transcription 3 (STAT3) in human microglial cells to regulate the ‘response to interferon I’ inflammatory pathway.¹⁶⁴ Lastly, a recent study found that the antidepressant-like effects of (R)-ketamine were blocked by microglial depletion in chronic-stress-sensitive mice. (R)-Ketamine was also able to induce brain-derived neurotrophic factor (BDNF) transcription by inhibiting MeCP2 and increasing the expression of nuclear-receptor-binding protein 1 in microglial cultures.¹⁶⁵

More recently, acute intraperitoneal ketamine administration in mice was shown to skew the distribution of macrophage populations away from proinflammatory and cytokine-inducing M1 phenotypes toward tissue-supporting M2 phenotypes.¹⁶⁶ This finding was also observed in vitro in human monocyte cultures, where the effect could be abolished by inhibiting mammalian target of rapamycin (mTOR), a key protein implicated in ketamine’s antidepressant effects.¹⁶⁷ PBMC samples collected from healthy volunteers and participants with MDD after a suicide attempt or with active suicidal ideation found that macrophages in MDD participants also skewed toward the inflammatory M1 phenotype.¹⁶⁶

Some of ketamine’s anti-inflammatory effects can also be mediated via apoptosis of various cell types. In neuronal cell types, anesthetic concentrations of (S)-ketamine attenuated expression of Bax, a proapoptotic protein, after cerebral ischemia.¹⁶⁸ By contrast, lower concentrations of ketamine have been shown to promote apoptosis of T lymphocytes via mitochondrial signaling in various cell lines, thus potentially inhibiting downstream cytokine production.¹⁶⁹ Accumulation of Th17 cells (a proinflammatory T cell subtype) and an imbalance of Th17:T_reg cells have also been associated with depression,¹⁷⁰ and ketamine was able to suppress differentiation of this cell subtype, although this process was not mediated by apoptosis.¹⁷¹

Ketamine can also indirectly affect inflammation by mediating HPA axis function. In chronically stressed mice, acute ketamine administration restored hippocampal glucocorticoid receptor expression, counteracting the negative feedback associated with HPA overactivation.¹⁷² In mice injected with LPS, ketamine significantly reduced corticosterone and ACTH production six hours later.¹⁷³ Similarly, single and repeated 7-day ketamine administration reduced corticosterone and ACTH levels in mice that had undergone 40 days of chronic mild stress.¹⁷⁴

Ketamine and kynurenine appear to converge during stress conditions to affect brain and behavior. One study found that, although ketamine did not affect QA production after LPS administration, it mediated the effects of QA by blocking NMDARs, where QA generally binds to contribute to inflammation.^113,150 In a chronic unpredictable mild stress model, ketamine decreased the KYN:tryptophan ratio in addition to other measures of inflammation.¹⁷⁵ To more closely mimic TRD, future studies with preclinical models should assess the impact of ketamine on inflammation after multiple cycles of chronic stress.

Clinical evidence of ketamine’s anti-inflammatory effects

Multiple inflammatory markers have been linked to ketamine’s clinical therapeutic efficacy. In a recent randomized, controlled trial, subanesthetic-dose ketamine (0.5 mg/kg) acutely decreased TNF-α levels in TRD patients, and these decreases correlated with reductions in Montgomery–Åsberg Depression Rating Scale (MADRS) scores.¹⁸ A smaller study of individuals with TRD similarly found that higher baseline levels of IL-6 were associated with antidepressant response to ketamine.¹⁷⁶ In an open-label trial, ketamine robustly reduced peripheral levels of multiple cytokines elevated at baseline in TRD participants but these levels returned to baseline within 24 h and did not correlate with antidepressant response.¹⁷⁷ In addition, a recent study in remitted depressed participants found significant decreases and time x treatment interactions for multiple cytokines.¹⁷⁸ However, other studies obtained mixed results. For example, a post hoc analysis of three ketamine randomized, controlled trials of participants with TRD and treatment-resistant bipolar depression found that ketamine decreased levels of soluble tumor necrosis factor receptor 1 (sTNFR1) but increased peripheral levels of IL-6 and TNF-α.¹⁷⁹ Interestingly, a recent open-label ketamine trial found that IL-8 did not predict antidepressant response to ketamine but that there was a trend toward prediction in females, suggesting a potential gender-specific effect.¹⁸⁰ Rapamycin, an mTORC1 inhibitor, was also found to prolong ketamine’s antidepressant effects, which could be at least partly due to its immunosuppressive actions.¹⁸¹

The effects of ketamine on the HPA system are less clear. One case study found that cortisol levels – as measured by the dexamethasone suppression test – normalized in a TRD participant who received three standard ketamine infusions; cortisol levels rose to baseline a week later as depressive symptoms returned.¹⁸² By contrast, a randomized, controlled trial of 12 healthy volunteers who received two back-to-back ketamine infusions (0.29 mg/kg for 1 h, then 0.57 mg/kg for 1 h) reported doubled plasma cortisol levels 200 min later.¹⁸³ Furthermore, another randomized, placebo-controlled trial of healthy volunteers found that the post-ketamine increase in cortisol was specific to ketamine, because the NMDA antagonist memantine caused no such effect.¹⁸⁴ For now, the dearth of properly-powered studies examining potential HPA biomarkers post-ketamine treatment in TRD participants makes it difficult to draw firm conclusions.

Echoing preclinical findings, modulation of the kynurenine pathway might be involved in ketamine’s anti-inflammatory effects. A randomized, controlled trial of TRD participants found that those who responded to ketamine had significantly lower plasma kynurenine:tryptophan ratios as well as lower kynurenine levels 230 min and 24 h post-ketamine administration.¹¹⁷ Furthermore, among participants with TRD and treatment-resistant bipolar depression who received six ketamine infusions over 12 days, those who responded had higher levels of serum KA, both absolute and relative to kynurenine, on Days 1 and 13.¹¹⁸ Moreover, at 24 h, both of these metrics correlated with MADRS score reductions at Days 1, 13 and 26. Finally, a recent randomized, controlled trial of individuals with bipolar depression reported that one ketamine infusion increased KA levels one and three days later and decreased IDO levels from 230 min post-infusion to three days later.¹⁸⁵ Despite these promising findings, it should be noted that another study found only trend-level decreases in serum kynurenine after repeated ketamine infusions and no change in cortisol-awakening response.¹¹⁹

There is also indirect evidence of ketamine’s anti-inflammatory effects. One post hoc analysis of four randomized, controlled trials (n = 108) found that greater BMI predicted antidepressant response to ketamine in individuals with MDD or bipolar depression,¹⁸⁶ which could be linked to the finding that proinflammatory agents are often deposited in adipose tissue.¹⁸⁷ Subsequently, researchers examined adipokine levels and found that ketamine reduced plasma levels of resistin, and that low baseline levels of adiponectin predicted antidepressant response.¹⁸⁸ These findings are congruent with anti-inflammatory effects; resistin is a potent proinflammatory agent¹⁸⁹ associated with obesity, whereas adiponectin is an anti-inflammatory molecule.¹⁹⁰ Another study of medication-free TRD participants found that ketamine decreased the expression of receptor activator of nuclear factor kappa-B ligand (RANKL), a downstream inflammatory mediator.¹⁹¹ In TRD participants, gene expression signatures related to IFN signaling pathway activation were upregulated in comparison to healthy volunteers; but this did not mediate a response to ketamine.¹⁹²

Despite these promising findings, it is clear that more research is necessary to clarify ketamine’s effects on inflammation in general and on clinical depression subtypes linked to inflammation in particular. The mixed results suggest that future studies should compare acute versus chronic ketamine administration as well as the short- and long-term effects of ketamine, given that some of the aforementioned studies observed an immediate increase in inflammatory indicators post-ketamine administration that decreased with time. It is also important to note that some clinical studies used ketamine adjunctively with current antidepressant therapies, and that administration of concomitant medications could directly impact results on inflammation in comparison to ketamine administered alone. Nevertheless, promising preclinical evidence and strong associations between TRD and inflammation warrant further investigation into the mechanisms by which ketamine can either directly or indirectly mediate an inflammatory response.

Concluding remarks

In this era of personalized medicine, the quest to identify subpopulations of individuals with MDD based on pathophysiology, symptom dimensions and prognostic biomarkers of treatment efficacy holds considerable promise for improving the thus far inadequate therapeutic response associated with many currently available pharmacotherapies. This review presents evidence that chronic-stress-induced, systemic, proinflammatory states can constitute a pathogenic factor that can negatively impact treatment-responsiveness in depression. Meanwhile, preliminary but growing evidence suggests that ketamine’s unique efficacy in treating these same treatment-refractory symptoms could partly be the result of its anti-inflammatory effects, perhaps by directly counteracting the inflammatory consequences of chronic stress; these unique effects are not associated with conventional antidepressants. These effects can occur via some combination of cytokine suppression, alteration of the kynurenine pathway or HPA axis, direct actions on microglia and other monocytes, and additional mechanisms not discussed here. In addition, research that differentiates between the impact of ketamine’s enantiomers [(R)- and (S)-ketamine) and its metabolites [e.g., (2R,6R)-HNK and norketamine] is essential for properly outlining ketamine’s anti-inflammatory effects. Although promising preclinical and early-phase clinical research has been conducted with (R)-ketamine and (2R,6R)-HNK, further clinical studies are needed to verify these initial results.

In this context, the need to verify ketamine’s anti-inflammatory properties with rigorous, prospective, clinical research is clear, as is the need to use preclinical models to elucidate the molecular and cellular basis underlying these effects. The effects of gender must also be considered, given the mixed results regarding gender differences in inflammation. This is particularly important because few preclinical or clinical studies have explored the links between TRD, ketamine and inflammation in a female population. Even less attention has been paid to the effects of race on the links between chronic stress, inflammation and therapeutic response, despite initial correlational studies showing different relationships between inflammation and depression by race.^193–195 In addition, the generalizability of ketamine’s therapeutic efficacy has not been determined for multiple ethnic groups. Further research should prioritize determining the effects of race and ethnicity on inflammation-related outcomes and on ketamine’s therapeutic effects. Nevertheless, future research efforts in this area are likely to be complicated by several challenges. First, depressive symptoms that can derive from inflammation and respond to ketamine are neither universal nor specific to any one diagnostic category. Thus, advances in psychiatric nosology are probably needed to replicate research with greater inter-study validity. For example, one crucial issue is a fuller differentiation between unipolar and bipolar depression, because ketamine has been successfully used to treat both these neuropsychiatric disorders despite the fact that the underlying inflammation-mediated mechanisms are probably different. Second, immune system dysregulation has a multitude of other consequences that span multiple systems and that could be further confounded by other factors such as gender and BMI. A more complete understanding of these complex interactions, combined with improved identification of the heterogeneous etiologies of depressive symptoms, are needed to move this field forward.

Regardless, further systematic research into the connections between inflammation, treatment-resistant symptom severity and response to ketamine is warranted. Ideally, such investigations should measure central levels of inflammatory markers and products of related pathways such as the HPA and kynurenine pathways and correlate these with suicidal ideation, anhedonia and other hallmark symptoms of TRD.

Highlights:

• Chronic stress, depression and inflammation have been consistently linked

• Stress-induced inflammatory states can contribute to treatment resistance

• Ketamine could be uniquely placed to target inflammation

• Ketamine’s effects can be mediated through the HPA axis or kynurenine pathway

• Anti-inflammatory antidepressant therapies could improve treatment resistance

Biographies

Jenessa Johnston

Jenessa Johnston is a PhD student at the University of Victoria and a Pre-Doctoral Fellow at the National Institute of Mental Health (NIMH). Her research focuses on the underlying molecular mechanisms of novel antidepressants as well as the discovery of diagnostic and therapeutic biomarkers for various neuropsychiatric disorders.

Max Greenwald

Max Greenwald is an MD/PhD student specializing in neuroscience and psychiatry at the Yale School of Medicine. His research focuses on psychoactive drugs and their therapeutic pathophysiological mechanisms in psychological disorders

Bashkim Kadriu

Bashkim Kadriu, MD, is a board-certified psychiatrist. He is currently Executive Director of Translational and Experimental Medicine, Neuroscience, Jazz Pharmaceuticals. He previously worked as a Clinical Leader at Janssen Pharmaceuticals and as a neuroscientist at the Experimental Therapeutics and Pathophysiology Branch (ETPB), NIMH. His research interests include early drug development as well as the neurobiological correlates of mood disorders, with a particular emphasis on discovering biosignatures that guide novel therapies and their mechanisms of action.