Kappa opioid receptor plasma levels are associated with sex and diagnosis of major depressive disorder but not response to ketamine

Abstract

Background:

Pre-clinical evidence indicates that the kappa opioid receptor (KOR)/dynorphin pathway is implicated in depressive-like behaviors. Ketamine is believed to partly exert its antidepressant effects by modulating the opioid system. This post-hoc study examined the following research questions: 1) at baseline, were there differences in KOR or dynorphin plasma levels between individuals with major depressive disorder (MDD) and healthy volunteers (HVs), or between men and women?; 2) in individuals with MDD, did KOR or dynorphin baseline plasma levels moderate ketamine’s therapeutic effects or side effects?; and 3) in individuals with MDD, were KOR or dynorphin plasma levels affected after treatment with ketamine compared to placebo?

Methods:

Thirty-nine unmedicated individuals with MDD (23 females) and 25 HVs (16 females) received intravenous ketamine (0.5 mg/kg) and placebo in a randomized, crossover, double-blind trial. Blood was obtained from all participants at baseline and at three post-infusion timepoints (230 minutes, Day 1, Day 3). Linear mixed model regressions were used.

Results:

At baseline, participants with MDD had lower KOR plasma levels than HVs (F_1,60=0.13, p<0.001), and women (MDD and HVs) had higher KOR plasma levels than men (F_1,60=4.98, p=0.03). Diagnosis and sex had no significant effects on baseline dynorphin levels. Neither KOR nor dynorphin levels consistently moderated ketamine’s therapeutic or side effects, nor were levels altered post-ketamine.

Conclusions:

In humans, diagnosis of MDD and biological sex are involved with changes in components of the KOR/dynorphin pathway. Neither KOR nor dynorphin levels consistently moderated ketamine’s therapeutic effects or side effects, nor were levels altered post-ketamine.

Introduction

Major depressive disorder (MDD) is the leading cause of disability worldwide¹. Unfortunately, only 30-40% of individuals with MDD enter remission after treatment with conventional antidepressants^2–4. These limited success rates are due in part to the heterogeneity of MDD, which has a multifactorial etiology involving diverse neurobiological pathways^5,6. In contrast, conventional antidepressants have a monothematic mechanism of action and exert virtually all their actions by modulating monoamine levels or their receptors^5–7.

Robust pre-clinical evidence indicates that the opioid system plays an important role in neurobiological processes intimately linked to MDD, such as regulating mood and motivational/consummatory behavior^8–11. Broadly, the opioid system comprises three types of receptors: kappa opioid (KOR), mu opioid (MOR), and delta opioid (DOR). Each of the three families of endogenous opioid peptides (dynorphins, β-endorphins, and enkephalins) has a greater affinity for a specific type of receptor; KORs predominantly interact with dynorphins, MORs with β-endorphins, and DORs with enkephalins^8–11. The different receptors/peptide pathways also have distinct functions in the nervous system (Table 1).

Table 1.

Summary of the main characteristics of the three types of opioid receptors

Characteristics	Kappa opioid receptor (KOR)	Mu opioid receptor (MOR)	Delta opioid receptor (DOR)
Endogenous ligand	Dynorphins	β-endorphins	Enkephalins
Situations associated with receptor activation	Exposure to stressful stimuli	Hedonic stimuli	Complex rewards
Signaling mechanism	Decreased intracellular calcium	Increased intracellular cAMP	Increased intracellular cAMP
Clinical effects	“Anti-reward” action	Euphoria / reward	Anxiolytic effects
	Dysphoria / anxiety	Antidepressant effects	Antidepressant effects
	Analgesic effects	Analgesic effects	Analgesic effects
	Sedation	Sedation	Suppressed Appetite
	Increased appetite	Suppressed appetite
Agonists	U50488	Morphine, fentanyl	AZD2327
Antagonists	Naloxone/naltrexone, buprenorphine, nor-BNI	Naloxone/naltrexone, samidorphan	Naloxone/naltrexone

Abbreviations: cAMP: cyclic adenosine monophosphate; nor-BNI: nor-binaltorphimine

The KOR/dynorphin pathway is involved in the stress response, and its prolonged activation leads to several clinical effects, including dysphoric/depressive mood and reduced sensitivity to reward (anhedonia)^10,12,13, which are core symptoms of MDD¹⁴. Generally, KOR agonism results in dysphoria/depression while KOR antagonism alleviates these symptoms^11,15,16. The KOR/dynorphin pathway has thus been hypothesized to be hyperactivated in MDD, and the use of KOR antagonists thought to be potentially therapeutic, particularly for treatment-resistant cases^8,9,13. One randomized, placebo-controlled trial of 141 individuals with MDD compared buprenorphine/samidorphan, a combination that mostly antagonizes KORs, in two distinct doses (2 mg/2 mg and 8 mg/8 mg) versus placebo¹⁵. The 2 mg/2 mg arm (but not the 8 mg/8 mg arm) was superior to placebo in reducing depressive symptoms. An additional study (FORWARD-5) confirmed that 2 mg/2 mg of buprenorphine/samidorphan was superior to placebo in reducing depression scores¹⁷, which provides some support for the role of KOR antagonism in treating depression. However, this combination treatment was also associated with two negative studies (FORWARD-3)¹⁸ and (FORWARD-4)¹⁷.

Despite substantial pre-clinical evidence demonstrating that potentiation of the KOR/dynorphin pathway is associated with depressive-like behaviors^10,13, human data are limited and mixed. One study found that individuals with MDD had greater serum levels of KORs and dynorphins¹⁹, though all study participants were male, a significant limitation given the important sex differences in the KOR/dynorphin pathway^20–22. Two post-mortem studies found that MDD was linked to decreased expression of the gene that encodes prodynorphin, the dynorphin precursor, in areas of the amygdaloid complex—one in the periamygdaloid cortex²³ and one in the accessory basal and amygdalohippocampal nuclei²⁴. In contrast, one positron emission tomography (PET) study found no statistically significant association between MDD and KOR levels in four brain areas (amygdala, hippocampus, ventral striatum, and raphe nuclei)²⁵, and a postmortem analysis found no evidence of gene expression for the gene that encodes KORs in the cingulate or prefrontal cortices²⁶. In addition, almost all studies conducted in humans to date have significant limitations, including relatively small sample sizes, non-representative samples, and depressed participants receiving psychotropic medications (which may confound the relationship between MDD and KOR/dynorphin levels).

In recent years, racemic (R,S)-ketamine (hereafter referred to as ketamine) and (S)-ketamine (esketamine) have revolutionized the treatment of MDD. Ketamine’s rapid antidepressant effects manifest within a few hours as opposed to weeks with conventional antidepressants, and approximately half of individuals with treatment-resistant depression respond to a single ketamine infusion compared to the less than 15% who respond to conventional antidepressants²⁷. While most studies seeking to elucidate ketamine’s complex mechanism of action have focused on its ability to modulate the glutamatergic system^27,28, growing evidence suggests that ketamine could exert its antidepressant effects in part through the opioid system^11,29–31. However, the evidence regarding which family of opioid receptors (KORs, MORs, or DORs) is the most important for ketamine’s antidepressant actions is both limited and somewhat conflicting^11,29–31. Briefly, stronger evidence suggests that ketamine’s antidepressant effects are due to an interaction with KORs and MORs than interaction with DORs^11,29–32. Nonetheless, ketamine’s effects on opioid receptors is complex, and DORs have also been implicated in some of ketamine’s actions, particularly on perception of pain³³. Ketamine’s side effects may also be mediated through interactions with opioid receptor signaling³². Ketamine has significantly weaker binding affinity to the KOR (~25 uM) than to N-methyl-D-aspartate receptors (~1 uM)³⁴, with (S)-ketamine having greater affinity than (R)-ketamine^11,32. Presently, it remains unclear which specific pathway within the opioid system—KORs/dynorphins, MORs/β-endorphins, or DORs/enkephalins—is most closely linked to ketamine’s mechanism of action²⁹.

This post-hoc study sought to compare KOR and dynorphin plasma levels between unmedicated adults with MDD and healthy volunteers (HVs) from both sexes, examine the impact of KOR and dynorphin plasma levels on ketamine’s therapeutic and side effects, and study KOR and dynorphin plasma levels as possible therapeutic targets of ketamine. The three main questions of interest were:

1. At baseline, were there differences in KOR or dynorphin plasma levels between individuals with MDD and HVs, or between men and women?;

2. In individuals with MDD, did KOR or dynorphin baseline plasma levels moderate ketamine’s therapeutic effects or side effects; that is, did KOR or dynorphin baseline levels strengthen or weaken the relationship between ketamine treatment and improvement in psychiatric symptoms or side effects and, therefore, have predictive value in determining which individuals would experience robust benefit or harm post-ketamine?;

3. In individuals with MDD, were KOR or dynorphin plasma levels affected after treatment with ketamine compared to placebo?

Materials and Methods

Participants

Thirty-nine individuals with treatment-resistant MDD (23 females, 16 males) and 25 HVs (16 females, 9 males) were included in this post-hoc analysis of previously published data (NCT00088699). Thirty-five of the 39 participants with MDD in this study were part of that original clinical trial; demographic and recruitment information have previously been published³⁵. Inclusion criteria for participants with MDD were: 1) ages 18-65 years; 2) diagnosis of recurrent MDD without psychotic features according to the Structured Clinical Interview for Axis I DSM-IV Disorders (SCID)-Patient Version³⁶; 3) depressive symptoms of at least moderate severity, as assessed by a Montgomery-Åsberg Depression Rating Scale (MADRS)³⁷ score ≥ 20 at baseline and before each infusion; 4) history of non-response to at least one adequate antidepressant trial, as assessed by the Antidepressant Treatment History Form³⁸; and 5) a current major depressive episode lasting at least four weeks. Eligibility criteria for HVs included age between 18 and 65 years, absence of Axis I disorders according to the SCID Non-Patient Version, and absence of first-degree relatives with Axis I disorders.

Exclusion criteria for participants with MDD were: 1) use of drugs (except nicotine or caffeine) in the two weeks prior to screening; 2) positive drug urine test at screening (except for prescribed benzodiazepines); 3) presence of any medical illness likely to change brain physiology and/or morphology; 4) clinically significant abnormal laboratory tests; and 5) serious suicidal or homicidal risk, including a MADRS item 10 (suicidality) score higher than 4. Exclusion criteria for HVs were: 1) any medical illness likely to change brain physiology and/or morphology; 2) treatment with medications that affect central nervous system (CNS) function, perfusion, or metabolism in the two weeks prior to initiation of the study; 3) diagnosis of current or past alcohol or substance abuse or dependence (except for nicotine or caffeine); 4) female participants who were either pregnant or nursing; and 5) NIMH employee/staff or their immediate family members.

Both participants with MDD and HVs were in good physical health according to a comprehensive medical evaluation that included medical history, physical examination, blood tests, electrocardiogram, chest x-ray, urine toxicology, and urinalysis. Both groups were also free of psychotropic medications for at least two weeks (five weeks for fluoxetine and three weeks for aripiprazole).

This study was reviewed and approved by the National Institutes of Health (NIH) combined CNS Institutional Review Board (IRB). All participants provided written informed consent.

Procedures and measures

Participants with MDD and HVs were randomly assigned to an infusion order (ketamine-placebo or placebo-ketamine), and infusions were performed two weeks apart. Individuals were admitted to an inpatient psychiatric research unit at the NIMH and tapered off psychiatric medications before the first infusion. Subanesthetic-dose ketamine was administered intravenously (0.5 mg/kg over 40 minutes). Clinical measures were obtained at baseline (one hour before each infusion), at 40, 80, 120, and 230 minutes post-infusion, and at Days 1, 2, 3, 7, 10, and 11 post-infusion. Blood samples were obtained from individuals with MDD and HVs at baseline, 230 minutes, Day 1, and Day 3 post-infusion.

The primary outcome measure was improvement in MADRS score. The MADRS is a 10-item clinician-administered rating scale where each item is scored from 0 to 6; higher scores indicate greater severity of depressive symptoms. Measures of anhedonia were also examined via the Snaith-Hamilton Pleasure Scale (SHAPS)³⁹ and the Temporal Experience of Pleasure Scale (TEPS)⁴⁰. The SHAPS is a 14-item self-administered instrument with total scores ranging from 0 to 14 where higher scores indicate greater anhedonia. The TEPS is an 18-item self-administered scale developed to measure two distinct components of anhedonia: anticipatory anhedonia (10 items) and consummatory anhedonia (8 items). The instrument provides an overall anhedonia score as well as specific subscale scores. Unlike the SHAPS, lower scores on the TEPS indicate greater anhedonia. Anxiety symptoms were also measured using the 14-item clinician-rated Hamilton Anxiety Rating Scale (HAM-A)⁴¹, and trauma-related symptoms were assessed with the 17-item PTSD Checklist Civilian Version (PCL-C)⁴².

Scores on the 23-item Clinician-Administered Dissociative States Scale (CADSS)⁴³ at 40-minutes post-infusion—the timepoint when ketamine’s side effects typically peak⁴⁴—were used to assess ketamine’s side effects. Although both the Brief Psychiatric Rating Scale (BPRS)⁴⁵ and the 11-item Young Mania Rating Scale (YMRS)⁴⁶ were explored as potential tools to assess the moderating effects of KORs/dynorphins on manic and psychotomimetic side effects, respectively, initial analyses suggested that ketamine had no effect on either measure (Supplemental Table S1).

Measuring biomarkers

Venous blood samples were drawn from MDD participants and HVs using Becton Dickinson (BD) vacutainer tubes supplied with heparin for anticoagulation. Blood was centrifuged at 3000 rpm (1500 x g) for 10 minutes, and plasma was then aliquoted into cryotubes and stored at −80 °C until thawed for assay.

Commercial ELISA kits were used to measure KORs (Human Kappa-Type Opioid ELISA Kit, MyBioSource, CA, USA) and dynorphins (Human Dynorphin ELISA Kit, Cusabio, China). Plasma samples were not diluted. Standards and samples were pipetted into the wells of pre-coated 96-well plates in duplicate and incubated at 37°C for two hours. After removing unbonded liquid, biotinylated specific antibody was added to each well and incubated for one hour at 37°C. After washing three times, avidin-HRP was added to the wells and incubated for one hour at 37°C. After washing five times, TMB substrate was then added and incubated for 15 minutes at 37°C followed by adding stop solution. Finally, the optical density of each well was measured at 450 nm with Synergy HTX Multi-mode Plate reader (BioTek, USA).

The sensitivities of the KOR and dynorphin assays were 0.019 ng/ml and 0.78 pg/mL, respectively, and the detection ranges were 0.05–15 ng/ml and 3.12–200 pg/ml, respectively. All procedures were conducted according to the manufacturer’s instructions. The inter-assay and intra-assay coefficients of variation (CVs) were <10%.

Statistical analysis

Because KORs and dynorphins were both right skewed, both were log-transformed prior to all analyses, and linear mixed model regression was used throughout. In order to test for baseline diagnostic group (MDD versus HV) and sex (female versus male) differences in circulating marker levels, the main effects of diagnosis and sex, respectively, were used; age and infusion (first or second) were included as covariates, and a random intercept per person was also included to account for within-person dependence resulting from the two period-specific baselines. A diagnosis*sex interaction was not significant and was not included in the final model for either KORs or dynorphins.

To test whether baseline KOR or dynorphin levels moderated ketamine’s therapeutic effect in the MDD group, for each clinical outcome the model included baseline marker (KORs or dynorphins), drug (ketamine, placebo), time (230 minutes, Day 1, and Day 3 post-infusion), and all two- and three-way interactions. If the drug*time*baseline marker level interaction was not significant, the drug*baseline marker interaction was used to test whether the relationship between baseline marker level and outcome varied by drug collapsed over all timepoints. Models also included age, infusion, sex, baseline outcome, and average baseline outcome (per person) as covariates (as recommended by⁴⁷), and a random intercept per person. A similar approach was used to determine whether baseline levels of KORs or dynorphins moderated the strength of ketamine’s side effects (compared to placebo) at 40 minutes post-infusion.

To test for drug effects (ketamine versus placebo) on post-infusion KOR and dynorphin levels, models included fixed effects of drug, time, and their interaction. If the drug*time interaction was not significant, the main effect of drug was used to test whether ketamine’s effect on post-infusion marker levels differed from placebo. Covariates included order of infusion, sex, age, period-specific baseline marker level, and average baseline marker level (as recommended by⁴⁷), and a random intercept was used per person.

For each analysis with multiple timepoints, the residual variance/covariance matrix was allowed to vary by drug, and if an unstructured covariance matrix had model convergence problems, a first-order autoregressive covariance structure was used while continuing to allow the variances to differ by drug. For all models, residual plots (histograms, fitted versus predicted, and quantile-quantile (q-q)) were visually inspected to check model assumptions. If data points with high leverage were observed, sensitivity analyses were conducted to assess the robustness of the findings⁴⁸. A statistical significance level of p=0.05 was used for all analyses, except for moderation analyses where Bonferroni correction was used due to multiple comparisons and, as a result, the significance level was set at 0.006 (0.05 divided by 9) for moderation analyses. SAS (9.4) PROC MIXED was used to calculate all models, and R was used for plotting and descriptive statistics.

Results

The main baseline characteristics of the participants are displayed in Table 2. As expected, psychiatric symptoms and disorders were negligible in HVs. Previously published findings using a sample that largely overlaps with the current study³⁵ similarly found that, compared to placebo, ketamine had significant therapeutic effects in individuals with MDD; these included improving depressive, anhedonia, anxiety, and trauma-related symptoms over time (Supplemental Table S1). Compared to placebo, ketamine also caused greater dissociative side effects (Supplemental Table S1).

Table 2.

Baseline characteristics of healthy volunteers and individuals with major depressive disorder (MDD) (n=64)

Variables	Healthy volunteers (n=25) Mean (SD) [min; max] or n (%)	MDD (n=39) Mean (SD) [min; max] or n (%)
Demographics
Age	34 (10.6) [20.0; 56.0]	36.3 (10.1) [20.0, 66.0]
Sex (female)	16 (64%)	23 (59%)
Race (Caucasian)	14 (56%)	32 (82%)
Body mass index	27.4 (4.5) [19.4; 34.7]	27.6 (6.8) [19.7; 50.4]
Clinical variables
Montgomery-Åsberg Depression Rating Scale (MADRS)	1.5 (1.6) [0.0; 6.0]	33.4 (4.5) [24.0; 41.0]
Snaith-Hamilton Pleasure Scale (SHAPS) (n=45)	18.5 (4.3) [14.0; 28.0]	38.8 (4.3) [29.0; 47.0]
Temporal Experience of Pleasure Scale (TEPS) (n=45)	86.4 (9.3) [67.0; 100.0]	46.4 (9.9) [30.0; 70.0]
Previous suicide attempt (n=63)	0 (0%)	24 (63%)
Age of onset (n=35)	NA	15.8 (6.5) [4.0; 33.0]
Duration of current episode (months) (n=33)	NA	41.4 (68.0) [1.0; 264.0]

Values for the MADRS, SHAPS, and TEPS correspond to scores before the first infusion

Abbreviations: max = maximum; min = minimum; NA = not applicable; SD = standard deviation

Impact of diagnostic group (MDD vs HV) and sex on baseline KOR and dynorphin levels

At baseline, the MDD group had lower KOR levels than the HV group (F_1,60=0.13, p<0.001) (Figure 1), and women (analyses including both diagnostic categories) had higher KOR levels than men (F_1,60=4.98, p=0.03) (Figure 1). No statistically significant diagnostic group (F_1,59=1.67, p=0.20) or sex (F_1,59=0.98, p=0.33) differences were observed in baseline dynorphin levels (Figure 1).

Figure 1.

Impact of diagnostic group and sex on kappa opioid receptor (KOR) and dynorphin levels (n = 64). ^a Comparisons of KOR and dynorphin levels between individuals with major depressive disorder (MDD) and healthy volunteers (HVs) were controlled for order of infusion, age, and sex. ^b Comparisons of KOR and dynorphin levels between women and men were controlled for order of infusion, age, and diagnostic group.

Abbreviation: CI = confidence interval; mean (SE) = adjusted mean (standard error)

Baseline KOR and dynorphin levels as moderators of ketamine’s therapeutic effects and side effects in MDD

KOR levels:

Baseline KOR levels did not moderate ketamine’s therapeutic effects on depressive, anhedonia, anxiety, or trauma-related symptoms or ketamine’s dissociative side effects before or after controlling for multiple comparisons (all drug*biomarker and drug*biomarker*time p-values>0.05; Table 3). These results indicate that baseline KOR levels did not impact ketamine’s ability to improve psychiatric symptoms or dissociation scores.

Table 3.

Moderation analyses investigating the relationship between baseline kappa opioid receptor (KOR) and dynorphin levels and ketamine’s therapeutic effects and side effects in major depressive disorder (MDD) (n=39)

Psychiatric symptoms	KORs a				Dynorphins a
	drug*biomarker		drug*biomarker*time		drug*biomarker		drug*biomarker*time
	F	p	F	p	F	p	F	p
Therapeutic effects b
Depressive symptoms (MADRS total score)	0.33	0.57	0.36	0.70	2.85	0.10 c	1.19	0.31 c
Overall anhedonia (SHAPS total score)	0.04	0.85	0.02	0.98	0.23	0.64	0.48	0.62
Overall anhedonia (TEPS total score)	0.43	0.52	0.60	0.55	0.08	0.78	1.10	0.34
Anticipatory anhedonia (TEPS subscale)	0.39	0.54	0.23	0.79	0.04	0.84	1.28	0.29
Consummatory anhedonia (TEPS subscale)	0.30	0.59	1.27	0.29	0.05	0.83	1.08	0.35
Anxiety symptoms (HAM-A)	2.20	0.15	0.30	0.74	3.30	0.08 d	0.21	0.81 d
Trauma-related symptoms (PCL-C)	0.15	0.71	1.15	0.32	1.81	0.20	0.43	0.65
Side effects b
Dissociative symptoms (CADSS total score)	0.69	0.41	NA	NA	0.40	0.53	NA	NA

^aControlled for order of infusion, age, sex, and baseline biomarker (KOR or dynorphin) levels.

^bAnalyses of therapeutic effects included data from baseline up to three days post-infusion while side effects included data only from baseline and 40 minutes post-infusion because side effects were negligible after 40 minutes post-infusion.

^cOne high leverage data point was removed.

^dTwo high leverage data points were removed.

Abbreviations: CADSS = Clinician-Administered Dissociative States Scale; HAM-A = Hamilton Anxiety Rating Scale; MADRS = Montgomery-Åsberg Depression Rating Scale; NA = not applicable; PCL-C = PTSD Checklist Civilian Version; SHAPS = Snaith-Hamilton Pleasure Scale; TEPS = Temporal Experience of Pleasure Scale

Dynorphin levels:

Initial analyses (not controlling for multiple comparisons) including all MDD participants revealed that higher baseline dynorphin levels were associated with lower MADRS (drug*baseline dynorphin interaction F_1,38=7.40, p=0.01) and HAM-A (drug*baseline dynorphin interaction F_1,33=9.51, p=0.004) scores post-ketamine but not post-placebo. However, the statistical significance of the drug*baseline dynorphin interactions for the MADRS and HAM-A did not survive sensitivity analyses that entailed removing observations from participants who appeared to have notable statistical leverage (that is, datapoints that strongly influenced results), especially in the ketamine condition (Table 3). After controlling for multiple comparisons, baseline dynorphin levels were not found to moderate ketamine’s therapeutic or dissociative effects.

Ketamine’s effect on KOR and dynorphin levels in MDD participants

Compared to placebo, ketamine was not associated with post-infusion changes in KOR (drug main effect F_1,32=3.45, p=0.07; drug*time interaction F_2,67=0.97, p=0.38) or dynorphin levels (drug main effect F_1,55=0.78, p=0.38; drug*time interaction F_2,132=0.22, p=0.81) (Figure 2).

Figure 2.

Impact of ketamine compared to placebo on kappa opioid receptor (KOR) and dynorphin levels in individuals with major depressive disorder (MDD) (n = 39). ^a Controlled for order of infusion, age, and sex.

Discussion

This study examined the relationship between KOR and dynorphin plasma levels, diagnosis of MDD, sex, and response to ketamine. At baseline, individuals with MDD had significantly lower KOR levels than HVs, and women (analyses including both diagnostic categories) had significantly higher KOR levels than men. KOR and dynorphin levels did not consistently moderate ketamine’s therapeutic effects (i.e., improve depressive, anhedonia, anxiety, or trauma-related symptoms) or dissociative side effects. In other words, KOR or dynorphin baseline levels did not impact ketamine’s clinical benefits or dissociative effects. In addition, KOR and dynorphin levels did not change post-ketamine infusion. Taken together, the results support in humans the extensive pre-clinical research suggesting that the KOR/dynorphin pathway is involved in the physiopathology of MDD and is affected by sex differences.

Our finding that individuals with MDD had lower KOR levels than HVs raises the question of what KOR plasma levels reflect and what their functional relevance is. KORs are abundantly expressed in the nervous system^12,49–51 and, in the brain, are especially found in areas related to reward (e.g., the mesolimbic pathway) and mood/stress (e.g., the amygdala and prefrontal cortex)^12,49. KORs are also observed in the spinal cord, dorsal root ganglion, and peripheral sensory neurons, where they modulate pain and neurogenic inflammation^51,52. KORs are also constitutively expressed in peripheral white blood cells, where receptor activity may modulate inflammatory processes^53–55. Although KORs are G- protein-coupled receptors primarily found in the plasma membrane, like hundreds of other membrane proteins they can also be detected in plasma^56,57. Membrane proteins and other cell peptides/proteins may be released in the blood after tissue remodeling or cell death^56,57. Therefore, KOR plasma levels may, to some degree, correlate with levels in nervous tissue (the primary site of KOR expression) and with white blood cells (a cell type with high turnover). However, the degree of correlation is unknown.

The present finding that individuals with MDD had lower KOR levels than HVs supports the role of the opioid system in MDD^8–11. Because the available evidence suggests that the KOR/dynorphin pathway is hyperactive in depressed individuals^10,12,25, the lower levels of KORs found in our study may reflect a compensatory downregulation of the KOR secondary to hyperactivation of the KOR/dynorphin pathway^10,54. However, other factors may also contribute to the lower KOR levels observed in MDD participants, including genetic variations and other processes involved in response to stress, such as activation of the hypothalamic-pituitary-adrenal (HPA) axis^16,22. Interestingly, our findings disagree with those of the only other study to examine circulating KOR levels in MDD and HVs, which observed higher KOR levels in depressed individuals¹⁹. That sample, however, consisted exclusively of males, and KOR levels were examined in the serum, not plasma as in our study. Future studies are needed to clarify the correlations between KOR levels in plasma, serum, white blood cells, CNS, and the peripheral nervous system.

Another key finding was that women had significantly higher KOR levels than men. This is consistent with pre-clinical evidence indicating the importance of sex differences in the KOR/dynorphin pathway^16,21,22. Gonadal hormones, sex chromosomes, and sex-linked autosomal polymorphisms all affect KOR expression^16,22. Interestingly, evidence in rodents suggests that, compared to females, males are more sensitive to both KOR agonism²⁰ and antagonism²¹. Future studies are needed to examine if sex differences in the KOR/dynorphin pathway may help explain the higher rates of MDD in women, and whether there are sex differences in response to interventions that modulate the KOR/dynorphin pathway in humans.

In this study, KOR and dynorphin levels neither consistently moderated ketamine’s therapeutic effects (i.e., improved depressive, anhedonia, anxiety, or trauma-related symptoms) nor altered post-ketamine infusion. These results suggest that KOR and dynorphin levels in human plasma may not be a major underlying mechanism of ketamine’s therapeutic effects or its dissociative side effects. However, there was a trend for ketamine to increase post-infusion KOR levels (p=.07). Notably, multiple ketamine infusions—as opposed to a single infusion—are associated with higher response rates and more prolonged effects ²⁷, indicating a potentially greater impact on neurobiological targets. Therefore, it remains possible that multiple ketamine infusions may significantly impact KOR levels. Preliminary evidence, mostly from pre-clinical studies, suggests that ketamine may at least partly act through other opioid receptors (i.e., MORs and DORs); thus, it is also possible that KORs are not involved in ketamine’s therapeutic or side effects^11,29–32. Given the likely underpowered nature of the current study, future studies are needed to further investigate ketamine’s impact on the opioid system—not only on the KOR/dynorphin pathway but also on the MOR/β-endorphin and DOR/enkephalins pathways.

Despite these intriguing preliminary results, the study is associated with several limitations. First, this was an unplanned post-hoc study. Given that most biological findings in psychiatry have modest effect sizes^58,59, our study had a relatively small sample size that may possibly be underpowered. Second, the study examined plasma levels (not brain levels) of KORs and dynorphins; while circulating KOR and dynorphin levels have been hypothesized to correlate with brain levels, only limited and indirect evidence supports that assumption in humans⁵⁴. Third, this study had a crossover design. Although this study design has significant advantages (e.g., minimizing between-participant confounding factors), no KOR or dynorphin levels were available after Day 3 and, therefore, the possibility of carryover effects cannot be fully excluded. Finally, the sample comprised individuals with recurrent treatment-resistant MDD, and caution is thus needed when generalizing our results to other samples.

Conclusion

In humans, diagnosis of MDD and biological sex are involved with changes in components of the KOR/dynorphin pathway. Neither KOR nor dynorphin levels consistently moderated ketamine’s therapeutic effects or side effects, nor were levels altered post-ketamine.

Data Availability

The data that support the findings of this study are available from the corresponding author upon request. The data are not publicly available due to privacy or ethical restrictions.