Introduction

Drug abuse continues to be a major public health problem worldwide, resulting in deleterious consequences that affect health, social, and economic welfare [1]. In the United States, the societal cost of drug abuse is over $820 billion, with approximately 20.3 million people over the age of 12 having met the DSM-V criteria for substance use disorder (SUD) in 2018 [2, 3]. SUD is a multifactorial disorder resulting from complex interactions between the pharmacological effects of a drug, biological characteristics of the individual, and environmental factors that render some individuals vulnerable [4]. As such, only a subset of individuals that use drugs of abuse develop SUDs. Similarly, individual differences in SUDs are observed in treatment efficacy and response to exogenous challenges (e.g., social stress, environmental enrichment) [5,6,7]. These variables are likely to contribute to the heterogeneity common in clinical studies, thus stressing the importance of a personalized treatment approach for SUDs, similar to other CNS disorders [8]. The goal of the present study was to further evaluate a biological marker associated with vulnerability to cocaine use disorders (CUDs).

For these studies, the model of enrichment and chronic stress involves social housing of female and male cynomolgus macaques (e.g., [9, 10]). Macaques form a dominance hierarchy that is based on the outcomes of fights, with winners being considered dominant to losers. Evidence indicates that social status profoundly influences physiology. For example, the subordinate monkeys have been shown to be more susceptible to reproductive dysfunction [11], upper respiratory infection [12] and atherosclerosis [13]. Furthermore, the finding that subordinate monkeys had heavier adrenal glands compared to dominant monkeys [14], support the hypothesis that social status can differentially affect the physiology of the HPA axis [15]. On the other end of the continuum, we have hypothesized that becoming dominant in the social group represents conditions of environmental enrichment [16, 17].

Initial positron emission tomography (PET) research on cocaine abuse focused primarily on the dopamine (DA) receptor system, and in particular, the DA D2/D3 receptors (D2/D3R). In humans, almost exclusively men, individuals with a long-history of cocaine use had lower measures of D2/D3R availability compared with age-matched control subjects [18]. This finding was extended to nonhuman primates (NHPs) using a within-subjects design showing that cocaine self-administration (SA) resulted in dose-dependent decreases in D2/D3R availability in all monkeys [19]. As it relates to measures of vulnerability, Volkow et al. [20] noted that non-substance abusing men with lower measures of D2/D3R availability reported rewarding subjective responses to methylphenidate, suggesting an inverse relationship between D2/D3R availability and stimulant reinforcement. This finding was replicated in socially housed male monkeys, showing that dominant males had higher measures of D2/D3R than subordinate males, and were less vulnerable to cocaine SA [21]. Importantly, similar results have been observed in humans [22, 23]. However, in socially housed female monkeys, subordinate animals had lower D2/D3R availability, but were less sensitive to cocaine reinforcement compared with dominant female monkeys [24]. Thus, there appears to be sex differences in the role of D2/D3R availability and vulnerability to cocaine reinforcement.

While DA is associated with euphoria, an opposing neurobiological system, the kappa opioid receptor (KOR) system, and its endogenous ligand dynorphin, is an integral part of the brain’s stress response system and is implicated in the neurobiological regulation of aversive states, including negative reinforcement associated with SUDs [25,26,27,28,29], sometimes referred to as the “dark side” of addiction, an allostatic dynamic of decreasing brain reward and recruitment of stress systems [26, 30,31,32,33,34]. This allostatic dynamic, negative affective state, can lead to the transition from social drug use to SUD [31,32,33]. Both chronic stress and cocaine exposure increase dynorphin/KOR system function, promoting dysregulation, negative affective states and stress reactivity [26, 27, 30, 35].

A recent PET study in humans investigated the relationship between KOR system and social status [36] utilizing a novel KOR agonist tracer [11C]EKAP [37, 38]. KOR availability inversely correlated with social status such that lower social status was associated with higher KOR availability in “anti-reward”/stress brain regions. Additionally, sex differences in KOR availability were observed, with females having higher KOR availability than males [36], consistent with clinical literature showing women have greater responses to KOR agonists than men [39,40,41]. A subsequent PET study in subjects with CUD (all males) reported a significant association between KOR availability and cocaine choice [42], implicating differences in KOR availability with cocaine choice and stress-induced relapse [42, 43]. These studies suggest KOR availability as a potential mechanism mediating sex differences in vulnerability to cocaine abuse. Thus, one goal of the present study was to extend the characterization of brain KOR availability, using [11C]EKAP, to socially housed female and male monkeys.

While PET imaging is an excellent in vivo measure of brain function, identifying potential peripheral markers of KOR function would provide greater clinical utility. Thus, a second goal was to identify and quantify protein expression levels of the KOR gene OPRK1 in neuron-derived extracellular vesicles (NDE), obtained from plasma of socially housed monkeys. Extracellular vesicles (EVs) are lipid-bound vesicles secreted by cells into the extracellular space and play a key role in intercellular communication and maintenance of cellular homeostasis. EVs have been reported in almost every bodily fluid and their cargo (e.g., proteins, RNAs, lipids, and metabolites, etc.) has been extensively characterized and correlated with normal or pathophysiologic states [44,45,46,47,48]. Moreover, in recent years NDE in plasma have shown promise as a potential biomarker for early diagnosis and therapeutic outcomes in various neurocognitive disorders [46, 47]. A recent study from our lab reported the usefulness of NDE in better understanding the pro-inflammatory and neurodegenerative consequence of oxycodone SA in male cynomolgus monkeys [45]. Moreover, the cargos (both proteins and miRNAs) of NDE are shown to be associated with adverse clinical neurodevelopmental outcomes [49]. The present study is the first to examine OPRK1 expression in NDE, which will serve as a peripheral measure of KOR function.

Methods and materials

Subjects

Experimentally naïve female and male cynomolgus monkeys (Macaca fascicularis), living in same-sex, stable (≥ 18 months) social groups of four, served as subjects (Table 1). Social hierarchy is linear, from the most dominant (#1-ranked) to the most subordinate (#4-ranked) monkey, and was determined according to the outcomes of agonistic encounters as described previously [50,51,52]. For PET studies, N = 8/sex (4 dominant and 4 subordinate/sex) and for OPRK1 expression studies, N = 13/sex (6 dominant and 7 subordinate/sex). Monkeys were naïve to drugs except for infrequent exposure to ketamine (IM), used for veterinary or imaging procedures. Monkeys lived in stainless steel cages (Allentown Caging, Allentown, NJ) divided into four equal quadrants (0.71 × 0.84 × 0.84 m) by removable wire mesh partitions. Partitions allowed visual, auditory and limited tactile interactions. When partitions were removed, four (same-sex) monkeys occupied the entire cage (0.71 × 1.73 × 1.83 m). Each monkey was fitted with an aluminum collar (Primate Products, Redwood City, CA) and trained to sit in a standard primate chair (Primate Products). Monkeys were weighed weekly and fed enough food daily (Purina LabDiet 5045, St Louis, MO and fresh fruit and vegetables) to maintain a healthy body weight and appearance as determined by daily inspection and periodic veterinary examinations. Water was available ad libitum in the homecage. For female monkeys, menstrual cycle was monitored daily by vaginal swabs [53] and was approximately 28 days. The first indication of bleeding was indicative of menses and counted as day 1 of the cycle. We considered days 2–10 the follicular phase and days 19–28 the luteal phase of the menstrual cycle. All PET studies in females occurred in the follicular phase. All procedures were performed in accordance with the 2011 National Research Council Guidelines for the Care and Use of Mammals in Neuroscience and Behavioral Research and were approved by the Wake Forest University Animal Care and Use Committee.

Table 1 Individual characteristics of cocaine-naïve subjects used in the NDE and PET imaging procedures.
Full size table

Experiment 1: Effects of sex and social rank on KOR availability

On separate days, socially housed female and male monkeys underwent a structural MRI and [11C]EKAP PET scans. All subjects received a 15 min structural MRI scan on a Siemens MAGNETOM Skyra 3-T scanner with TIM technology for image co-registration and assistance with the anatomic localization of regions of interest (ROI). Prior to the MRI scan, the monkey was anesthetized with ketamine (5.0 mg/kg, IM) and dexmedetomidine (0.04 mg/kg, IM) and then transported to the MRI Building. The reversal agent, atipamezole (0.4 mg/kg, IM) was administered following the scan. On the morning of the PET study, the monkey was anesthetized with ketamine (10 mg/kg, IM) and transported to the Wake Forest University School of Medicine PET Center. The monkey was intubated and anesthesia (isoflurane 1.5%) was maintained throughout the scan. A catheter was inserted into an external vein for tracer injection and fluid replacement throughout the study. Body temperature was maintained at 40°C and vital signs (e.g., heart rate, respiration rate, O2 saturation, etc.) were monitored throughout the scanning procedure. All PET studies were conducted with a GE 64 slice PET/CT Discover VCT scanner (GE Medical Systems, Milwaukee, WI). Before tracer administration, a 5 min transmission scan was acquired in three-dimensional mode. Next, [11C]EKAP (5 ± 1 mCi) was administered intravenously via manual injection, followed by a 120 min three-dimensional dynamic brain scan. At the end of the scan and following recovery from anesthesia/extubation, the monkey was returned to its homecage and remained singly housed for the day. The time between ketamine-induced anesthesia and the PET study was at least 60-min, thereby minimizing any pharmacological effects of ketamine on binding potential (BP). Two of the male monkeys were scanned twice, to obtain a measure of test-retest variability, which was ~3% (Table S1).

PET imaging analysis and kinetic modeling

PET data were collected in list mode for 120 min and reformatted into 33 successive frames of increasing durations (6 × 10 s, 3 × 1 min, 2 × 2 min, and 22 × 5 min). PET emission data were attenuation-corrected using the transmission scan, and dynamic images (33 frames and 256 slices over 120 min) were reconstructed using an automated QCHD8-based vue point HD rebinning algorithm. PET images were co-registered to individual MRIs using PMOD Biomedical Image Quantification Software (version 3.1; PMOD Technologies, Zurich, Switzerland). Spherical ROIs, all 2.5 mm radii unless noted below, were drawn on individual MR images for the following brain regions: dorsal prefrontal cortex (dPFC), medial prefrontal cortex (mPFC), orbitofrontal cortex (OFC), anterior cingulate cortex (ACC), caudate nucleus, putamen, ventral striatum (2.0 mm radii), amygdala, globus pallidus, ventral claustrum (1.0 mm radii), insula (2.0 mm radii), cingulate cortex, hippocampus, thalamus (3.0 mm radii), temporal cortex and cerebellum (4.0 mm radii), which served as the reference region. Previous studies with KOR radiotracers, including [11C]EKAP, indicated that cerebellum can be used as a reference region in nonhuman primates to calculate BP [37, 54]. The primary dependent variable is the ratio of the distribution of radioligand in the region of interest to the distribution in the reference region. The distribution volume ratio can be converted to BP, which is a unitless number representing the ratio of receptor density to affinity (Bmax/Kd). For these studies, the BP was calculated for each ROI using the Simplified Reference Tissue model [55], which allowed for the generation of binding parameters without obtaining arterial blood samples [37]. BPs for each region were not different between left and right sides and therefore data from each monkey were expressed as a mean of both sides.

Experiment 2: Peripheral measures of KOR using NDEs

All blood samples were obtained from the femoral vein of awake monkeys, while seated in a primate restraint chair. First, total EVs (TE) were isolated from the plasma of monkeys using the ExoQuick (System Biosciences, Palo Alto California, USA) precipitation method as described by us previously [56, 57]. Next, NDE were pulled out from TE using two surface biomarkers (L1CAM/CD171 and synaptophysin) as previously described [45]. Briefly, we first used L1CAM positive NDE from TE using L1CAM/CD171-Biotin antibody (Cat. No. 13-1719-82, ThermoFisher Scientific) and streptavidin-tagged agarose beads (ThermoFisher Scientific, Waltham, MA, USA). To further purify NDE using the second marker, synaptophysin-biotin antibody and streptavidin-tagged agarose beads were used and double positive (both L1CAM and synaptophysin) NDE were collected. TE and/or NDE were characterized by nanoparticle tracking analyses (NTA), immunogold labeling and transmission electron microscopy, Exo-check antibody array and ELISA assays (methodology described in Supplementary Methods) following our published methods [45, 56].

Data analyses

Two-way ANOVA was used to compare (1) PET measures of [11C]EKAP BP for each ROI (Experiment 1), and (2) NDE measure of OPRK1 expression by sex and social rank (Experiment 2). For these analyses, social rank was binary, with #1- and #2-ranked monkeys being “dominant” and #3- and #4-ranked monkeys being “subordinate”. An interaction was tested using a t-test. A post-hoc F-test with Bonferroni-Holm adjustment was used to determine social rank differences by sex. Normality assumptions were examined using the Shapiro-Wilk normality test. All data are presented as means and 95% confidence intervals. Spearman’s rank correlations were calculated between [11C]EKAP BP and social rank for each ROI and stratified by sex. Social rank was defined as an ordinal variable with values 1–4. Significance for all models was set at p < 0.05. For PET and NDE assessments, investigators were blind to the sex and social rank of the monkeys.

Results

Experiment 1: Effects of sex and social rank on KOR availability

The regional binding potentials (mean ± SD) of dominant and subordinate males and females for each of 15 ROIs, and the cerebellum, are shown in Table 2. A representative tissue-time activity curve is shown in Fig. S1. The normal distribution of the data was confirmed with the Shapiro-Wilk test for all ROIs, with the exception of the dPFC and the hippocampus. Log-transformations of these ROI did not impact findings; therefore, results are presented on the original scale. Two-way ANOVA indicated significant interactions between sex and social rank in the dPFC, ACC, caudate nucleus, putamen, ventral striatum, amygdala, globus pallidus, insula, claustrum, cingulate cortex, hippocampus and the temporal cortex (Table 2a). Three ROIs did not show an interaction between sex and social rank, though differential main effects were significant in these regions. The main effect of social rank on [11C]EKAP BP was significant in the mPFC [t(1,12) = 2.53, p < 0.05] and the OFC [t(1,12) = 2.55, p < 0.05], such that subordinates had higher BPs than their dominant counterparts. There was a significant main effect of sex on KOR BP in the thalamus [t(1,12) = 2.88, p < 0.05], such that males had higher BP than females, but no social rank differences. There was no significant main effect of sex in the mPFC or OFC.

Table 2 [11C]EKAP BP across all ROIs in cocaine-naïve socially housed female and male monkeys (N = 8/sex).
Full size table

Post-hoc tests (Table 2b) indicated no significant differences in KOR BP between dominant and subordinate males. However, subordinate females had significantly higher KOR BP than dominant females in the dPFC, ACC, caudate nucleus, putamen, ventral striatum, amygdala, globus pallidus, insula, claustrum, cingulate cortex, hippocampus, and the temporal cortex. Representative images of dominant (#1) and subordinate (#4), male and female monkeys are shown in Fig. 1.

Fig. 1: Representative PET/CT brain images with the KOR selective agonist [11C]EKAP in dominant (left) and subordinate (right), female (top) and male (bottom) cynomolgus monkeys.
figure 1

Subordinate females and dominant males have the higher KOR availability compared with dominant females and subordinate males (see also Table 2). Illustrated images are of slice 134, frame 17. The lower right image includes co-registered regions of interest labeled a–f. a Temporal cortex. b Hippocampus. c Insula and claustrum. d Globus pallidus. e Thalamus. f Cingulate cortex.

Full size image

In addition, Spearman’s rank correlations were estimated within each sex, to determine the relationship between social rank on an ordinal scale (1,2,3,4) and KOR BPs. There was a strong, negative correlation between social rank and BPs in females (rs(6) = −0.85, p = 0.01), confirming that dominant females had lower BPs compared with subordinate females. In contrast, there was a significant positive correlation in males (rs(6) = 0.74, p = 0.046), confirming the opposite was true, with dominant males having higher BPs compared with subordinate males.

Experiment 2: Characterization of TE, NDE, and identification of OPRK1

Characterization of TE using NTA from both dominant and subordinate male and female monkeys did not show any significant difference in concentration, average size, protein concentration and distribution of size (Supplementary Results; Fig. S2A–B). Still, the size of the TE was <150 nm, which corresponds to a size range of small extracellular vesicles (sEVs). Typical sEVs/ exosomes markers were confirmed on isolated sEVs from plasma using antibody arrays (Fig. S2C). Moreover, characterization of NDE with immunogold labeling and transmission electron microscopy showed the presence of L1CAM and synaptophysin (markers used for NDE isolation) on NDE surface (Fig. S2D). Exo-check antibodies array for neuron- and exosome-specific markers confirmed the specificity and purity of the isolated NDE (Fig. S2E). Importantly, OPRK1 was quantified in NDE isolated from plasma in all animals (Fig. 2). Two-way ANOVA revealed there was no interaction [F(1,21) = 2.03, p = 0.16] of sex and social rank on NDE OPRK1 concentrations. Furthermore, there were no differences in NDE OPRK1 concentrations between dominant and subordinate females [F(1,12) = 0.06, p > 0.05], however, the trend is for OPRK1 concentrations to be higher for dominant males compared to subordinate males [F(1,11) = 3.66, p = 0.084] (Fig. 2).

Fig. 2: Relationship between social rank and NDE OPRK1 concentration in dominant (Dom) and subordinate (Sub) female (left) and male (right) cynomolgus monkeys (N = 13/sex).
figure 2

OPRK1 expression in NDE was anlayzed by ELISA and presented as ng/ml per 100 ug of NDE. Each bar represents mean and 95% confidence intervals for 26 monkeys (N = 13/sex).

Full size image

Discussion

The goals of the present study were to examine the relationship between social rank and brain KOR availability in drug-naïve female and male monkeys and to assess the role of NDE OPRK1 expression as a peripheral biomarker of KOR availability. The main findings of this study were: (1) significant interaction between sex and social rank in [11C]EKAP BP in 12/15 brain regions; (2) [11C]EKAP BP correlated positively with social rank in males and negatively in females. Furthermore, the overall lowest receptor availability across ROIs was observed in dominant females and subordinate males, the two most vulnerable phenotypes to cocaine reinforcement [21, 24]; and (3) peripheral measures of KORs were identified and quantified by OPRK1 expression in NDE, with concentrations trending higher in dominant males compared with their subordinate counterpart, while no rank differences were observed in females.

Previous PET imaging studies from our laboratory focused on the role of DA D2/D3R availability in socially housed male and female monkeys. Consistent with other studies in male subjects, we noted that dominant male monkeys had higher levels of D2/D3R availability compared with subordinates and the latter were more vulnerable to cocaine reinforcement [21]. However, while D2/D3R availability was also higher in dominant females compared with subordinate females, it was the dominant females that were more vulnerable to cocaine reinforcement [24]. Thus, the present study aimed to extend the characterization of social status and sex to include another neurotransmitter system, the KOR, which appears to operate in opposition to the DA system [26, 27]. The current findings in males are analogous to D2/D3R availability, where [11C]EKAP BP is higher in dominant males compared with subordinate males. However, in females, the current findings are opposite to D2/D3R availability, where [11C]EKAP BP is significantly higher in subordinate compared with dominant females. Of note, the overall lowest BPs across ROI were observed in dominant females and subordinate males; these are the two most vulnerable phenotypes to cocaine reinforcement.

The highest KOR available were observed in subordinate females, with the four highest regions being the claustrum > insula > ACC > putamen. In dominant male monkeys, the claustrum and insula were also the regions with the highest KOR availability, followed by the putamen and caudate nucleus. The claustrum is believed to play a key role in decision-making [58], so higher KOR availability, perhaps indicative of lower dynorphin concentrations, in less vulnerable phenotypes is consistent with proposed relationships between vulnerability and decision making [59, 60]. Similarly, the higher KOR availability measures in brain regions associated with reinforcement, the caudate nucleus and putamen, are consistent with lower stress response and less vulnerable phenotypes. It remains to be determined whether these sex x social rank differences reflect stress-related effects. Previous studies from socially housed female and male monkeys have not reported significant differences in cortisol concentrations [51, 53, 61]. As described below, we believe the reasons for these sex and social rank differences are due to a combination of rank-related changes in KOR densities and the influence of circulating dynorphin.

While human PET studies showed no differences between people with CUD vs. controls, high-dose cocaine self-administration displaced the PET KOR tracer in humans, suggesting cocaine-induced elevations in dynorphin [42]. Taken together, the PET and receptor autoradiography data point to a critical role for dynorphin (in autoradiography, it is washed away and for PET, it is competing with the radiotracer which will influence the binding potentials). If dynorphin is important in the PET signal, then studying cocaine-naïve subjects and using BP as a predictor of vulnerability may be the most appropriate phase in the addiction cycle to study KORs. Future studies will be required to better address this. In addition, it should be noted that KOR system has been implicated in several psychiatric diseases, including depression, bipolar disorder, post-traumatic stress disorder, and schizophrenia [27, 62,63,64], therefore the study of [11C]EKAP in socially housed female and male monkeys may have implications for other diseases.

The highest [11C]EKAP BP in the claustrum is consistent with earlier autoradiography studies measuring KOR receptor densities in human and NHP brains [65, 66]. When stratified for sex and social rank, [11C]EKAP BP was: Sub Female > Dom Male > Sub Male > Dom Female. The claustrum has extensive connectivity with the neocortex, reciprocally connected with almost all cortical areas (most densely with medial regions of frontal cortex) and receives input from subcortical regions [67,68,69,70]. The broad and unique connections of the claustrum suggest it might serve as a central network hub, and may play a fundamental role in a range of functions, associating sensory and limbic information to prompt attention by means of the frontal cortex and executive function control systems [68, 71, 72]. The proposed functional implications of claustrum connectivity with limbic–sensory-motor information could allow one to recognize a stimulus’ contextual importance and coordinate across the cortex to be able to direct focus to the most relevant stimuli [68, 72].

The second highest [11C]EKAP BP was observed in the insula, which is a structure involved in feelings of anxiety, pain, cognition, mood, threat recognition and conscious desires [73,74,75,76,77]. The insula is heavily involved in “interoception”, the integration of internal and external (environmental) stimuli, conveying it to the ACC, ventral striatum, mPFC to initiate adaptive responses, to guide behavior for the purposes of maintaining homeostasis [78,79,80]. Accordingly, the insula plays a major role in all aspects of decision making [77, 81]. The relevance of the insula to addiction first emerged with a study that showed that smokers with damage to the insula were able to quit smoking immediately and without cravings or relapse [82]. Other lesion studies have emerged, indicating that loss of activity in the insula best predicts smoking cessation with five-fold greater probability of quitting smoking [83, 84], experiencing less urges to smoke, along with less frequent and severe withdrawal symptoms [85, 86]. Imaging studies have further supported the relevance of the insula in addiction, showing differential insular activation during craving across multiple drug classes [87,88,89]. Of clinical significance, insular activity has been proposed to represent a potential biomarker for relapse risk and, as such, a target for transcranial magnetic stimulation (TMS) as a potential treatment for SUD [77, 80, 90,91,92,93].

Sex differences in KOR availability have been reported using PET imaging in humans. Using the KOR antagonist radiotracer [11C]LY2795050, males had higher KOR availability than females across most brain regions [94]. In contrast, using the KOR agonist radiotracer [11C]EKAP, Matuskey et al. [36], reported that females had significantly greater KOR availability than males. These differences may be attributed to the use of a KOR agonist vs. KOR antagonist PET tracer. Agonists bind with high affinity and interact with only the active state of the receptor, whereas antagonists bind with equal affinity to both the active and inactive states [37]. This discrepancy suggests females may have an augmented proportion of receptors in the active state in comparison to males. Interestingly, there is clinical literature to support this discrepancy, with women showing stronger clinical responses to KOR agonists than men [39,40,41].

PET imaging represents a highly translational research tool to investigate brain function in disease states. However, with that said, the ability to identify peripheral correlates associated with KOR function, would provide a rapid and inexpensive biomarker that clinicians may use to determine treatment strategies. As a result, we evaluated the use of NDE as peripheral biomarkers for KOR availability. Our two-round approach to isolate NDE from TE is highly novel - first we used a generic neuronal marker (L1CAM/CD171) and then a CNS neuron-specific marker (synaptophysin), yielding highly pure double positive NDE population. The present study was also the first to identify KOR signals in NDE, as measured by OPRK1 expression. Recently, we reported that expression of neurodegenerative biomarkers (NFL and α-synuclein) correlated with differences in brain lobe volumes, using MRI analyses, between drug-naïve monkeys and those with an extended history of oxycodone SA [45]. In the present study, NDE OPRK1 expression trended towards significance in dominant male monkeys compared with subordinate males (see Fig. 2); the non-significant effect may be a result of a small sample size (N = 6–7/rank). However, the trend showing higher expression in dominant males is consistent with the higher KOR BPs in dominant males, suggesting that OPRK1 expression is indicative of CNS KOR availability. The fact that there were no differences in OPRK1 expression and social rank in females, while there were significant differences in KOR BP with [11C]EKAP, could indicate that the latter measures were, in fact, primarily influenced by circulating dynorphin, which would be consistent with greater vulnerability in dominant females. In the future, possible isolation of various brain-region specific NDE promises less invasive and more accurate measures of KORs in plasma.

There were some limitations to this study. For the PET imaging studies using a KOR agonist radiotracer, additional studies are needed to understand the mechanism(s) accounting for differences in BP between sexes and social groups. The human PET studies recently reported by Martinez et al. [42] utilized a KOR selective agonist [11C]GR103545 to investigate changes in the KOR/dynorphin system in subjects with CUD, before and after a cocaine binge. Following a 3-day cocaine binge, [11C]GR103545 binding showed a decrease in [11C]GR103545 VT of 14.4% across all brain regions. Martinez et al. [42] interpreted these decreases as representing increases in endogenous dynorphin concentrations. While multiple lines of evidence (rodent, NHP, human post-mortem) support binge cocaine greatly increasing dynorphin levels, the mechanism behind the decreases in [11C]GR103545 BP has not been identified and may involve receptor shift to inactive state, internalization or downregulation of the KOR, or a combination of these mechanisms [95]. A second limitation is conceptual: receptor binding studies of cocaine overdose victims indicate higher KOR densities [95] and elevated dynorphin concentrations [96] compared with controls, but we are speculating that lower KOR densities and higher dynorphin concentrations account for greater vulnerability. A particular advantage of using NHPs is that we can conduct longitudinal PET studies of how cocaine self-administration affects KOR availability, as well as study KOR agonists and antagonists on [11C]EKAP BP and behavior, to better address these potential discrepancies [25, 97,98,99,100,101]. Finally, there is the possibility that anesthetizing the monkeys with ketamine affected dynorphin concentrations, which could impact BP. We believe this is unlikely because of the 10–15 min half-life of IM ketamine [102] and the PET scan was not initiated for at least 60 min after ketamine-induced anesthesia.