Introduction
With the dramatic rise in the availability and abuse of illicit opioids, the mortality rate of overdose associated with opioid use disorder has reached epidemic proportions in the world [
1,
2], particularly the United States [
3,
4]. Most of the synthetic opioids are mu-opioid receptor (MOR) agonists with the potential of causing respiratory depression [
5]. Naloxone is a highly effective opioid antagonist that has been widely used to reverse the effect of opioids [
6,
7]. GSK1521498 is a MOR selective antagonist or conditional inverse agonist [
8,
9] that has been studied in alcohol use disorder and binge eating in obesity [
8,
10,
11]; it has also been suggested as an alternative to naloxone in the rescue of opioid overdose [
12]. Current studies of GSK1521498 in the treatment of binge-eating disorders and compulsive alcohol seeking have used either oral or intraperitoneal administration of the drug [
8‐
10,
13,
14] to evaluate the long-lasting pharmacokinetics for MORs. The acute effects of GSK1521498 by intramuscular (IM), intravenous (IV), or intranasal administration to determine the ability to reverse the effects of opioids for MORs has not yet been studied.
MORs are widely expressed in the central nervous system (CNS), peripheral organs, and immune system [
15,
16]. The mechanism of MORs in brain have been extensively studied in vivo by using positron emission tomography (PET) imaging with radioligands such as [
11C]carfentanil ([
11C]CFN) in human [
17‐
22], non-human primates (NHPs) [
23,
24], and rodents [
25]. However, only a small number of in vivo studies regarding the function and the density of MORs in the spinal cord and peripheral organs have been reported [
26].
In this study, we demonstrated the distribution of total-body MORs in both CNS and peripheral organs of rhesus macaques by using the PennPET Explorer, long axial field of view instrument, with the MOR radioligand [11C]CFN under baseline and opioids blockade conditions. The effects of opioid blockade by naloxone via IM administration prior to the [11C]CFN injection, and naloxone or GSK1521498 given by IV administration during the [11C]CFN scans, were conducted to evaluate the effect of the MOR antagonists on the pharmacokinetics of [11C]CFN in the CNS, spinal cord, and peripheral organs. The comparison of receptor occupancy (RO) and the rate of target engagement for GSK1521498 and naloxone were also evaluated.
Discussion
This study demonstrated the total-body imaging and test-retest variability of the MOR agonist, [
11C]CFN. Blocking studies were also conducted using the MOR antagonists, naloxone and GSK1521498, which were given via IM or IV administration in rhesus macaques. The %Var of the test-retest measures across brain regions for DVR of [
11C]CFN was approximately 8%, which is consistent with the previous test-retest study in humans [
18], demonstrating the high reproducibility of [
11C]CFN in imaging brain MOR. Taking advantage of the high sensitivity total-body scanner, PennPET Explorer, [
11C]CFN uptake in spinal cord and peripheral organs was also measured. A high reproducibility of [
11C]CFN DVR in cervical spinal cord was also observed (~ 6% test-retest variability). Due to the low perfusion and uptake of [
11C]CFN in the thoracic (SUV at peak = 1.4 ) and lumbar (SUV at peak = 1.2 ) spinal cord, the semi-quantitative parameter, SUVR
70 − 90 min, was used to evaluate the reproducibility of [
11C]CFN; this showed a somewhat higher test-retest variability than what was observed in brain. The %Var of test-retest was also higher in peripheral organs than in the CNS, which may be due to either the mixed signal of the parent compound and radiolabeled metabolites, and the variability of the biological clearance of [
11C]CFN and its radiolabeled metabolites among the NHPs.
Logan graphical analysis using the occipital cortex as the reference region has been widely used in quantitative [
11C]CFN brain imaging studies in human [
18‐
21], NHP [
23,
24] and porcine [
31]. In the current study, although the occipital/visual cortex showed low uptake of [
11C]CFN in the baseline scan, there was a notable decrease in [
11C]CFN uptake in the naloxone pretreatment study, and the naloxone or GSK1521498 displacement studies (Supplementary Fig.
8). In contrast, the cerebellar cortex showed the lowest [
11C]CFN uptake in the baseline study and no detectable effects of naloxone or GSK1521498 in both the blocking and displacement studies. It has been reported that there is no significant binding of MOR in a [
35S]GTPγS autoradiography study in cerebellum of cynomolgus monkey brain [
32]. These results indicate that the cerebellar cortex may serve as a better reference region for quantitative [
11C]CFN brain imaging studies in NHPs.
A study was also done to compare the route of administration on %RO of naloxone at MOR using [
11C]CFN. There was no difference in the %RO of the same dose of naloxone in MOR abundant brain regions when naloxone was given either via IM or IV injection (Supplementary Table
4). For instance, the %RO in thalamus was 87.9% and 88.1% for IM and IV administration respectively in NHP-3. Saccone et al. [
23]. has reported that the RO in thalamus is slightly greater in the same dose of naloxone given by IV versus intranasal administration in rhesus monkeys, whereas no difference in RO in thalamus was observed via IM and IV administration in a follow-up study [
24].The later study is consistent with our results. In the current study, IM administration of naloxone (0.14 mg/kg), the average %RO in thalamus and striatum was 86% and 81% respectively. This is higher than in the previous report from Scott et al. [
24], who observed a %RO of approximately 65% and 74% in thalamus and basal ganglia for the same dose of naloxone given by the same injection route. The difference in %RO might be attributed to the choice of the reference region (i.e., cerebellar cortex versus occipital cortex) for DVR calculation. By using occipital/visual cortex as reference region to calculate DVR in the current study, the average %RO in thalamus and striatum was 80% and 70% respectively (Supplementary Fig.
9). It should also be noted that gender differences in [
11C]CFN binding to MOR in human brain have been reported [
17]. Hence, the differences of RO in NHP brain between the two studies may also due to the gender differences, since male rhesus monkeys were used in the current study and female rhesus monkeys were used in the previous study [
24].
The function of displacement studies with a PET radiotracer is to measure the ability of an antagonist to compete with the radiotracer for binding to a CNS receptor. This is measured by the increased rate of washout from a region of interest following administration of a displacer, which prevents the rebinding of the radiotracer to its target receptor. In the naloxone displacement studies, the SUVRs of [
11C]CFN in the CNS began to decrease at the 4–6 min time frame post naloxone IV administration. This indicates that it takes naloxone approximately 4–6 min to reach concentrations in the CNS that prevents the rebinding of [
11C]CFN to the MOR. In addition, this result of the naloxone displacement studies is similar to the clinical reports suggesting that an average of 6–8 min response time is required for naloxone to reverse opiate overdose when given by IM administration [
33‐
35]. It is worth noting that the time of naloxone to affect the binding of opioids for MORs may differ depending on the route of administration.
In the studies comparing the pharmacokinetics of naloxone and GSK1521498 for [11C]CFN displacement, there were no differences in the shape of TACs, %RO, displacement rate, and the response time for displacing [11C]CFN from MOR in the brain and cervical spinal cord. These results indicate that the behavior of the MOR selective antagonist/inverse agonist GSK1521498 is similar to naloxone for MOR target engagement in vivo, and suggests that GSK1521498 may be a potential alternative to naloxone for opioid overdose rescue. However, more in vivo studies are needed, such as dose-response relationships and route of administration of the drug, in order to evaluate the capability of GSK1521498 to reverse opiate overdose.
Previous studies have shown that the MOR is expressed in the dorsal horn of lumber spinal cord in rats [
16,
36], and a moderate level of MOR expression has been reported in human spinal cord [
15]. Our results revealed a low level of specific binding of [
11C]CFN in baseline/retest studies in the cervical spinal cord (average DVR = 1.1), and demonstrated a significant blocking effect of MOR antagonists, naloxone and GSK1521498. In the thoracic and lumber spinal cord, there is a hint of naloxone blockade in the SUVR
70 − 90 min in the pretreatment versus baseline/retest studies. However, there was no displacement of [
11C]CFN following IV administration of naloxone or GSK1521498. These data suggest that [
11C]CFN may be feasible for imaging MOR density in the cervical spinal cord, but not sufficient for imaging the thoracic and lumber spinal cord due to the low radiotracer uptake. This is the first study demonstrating the ability to image MORs in spinal cord in vivo with PET.
Mixed results have been reported on the expression of MOR in human heart. For example, Peng et al. [
15]. reported no MOR expression in heart tissue by using absolute quantitative real-time reverse transcriptase PCR to quantitate MOR mRNA. However, in a later study, Sobanski et al. [
37] used immunohistochemical techniques to demonstrate MOR expression in myocardial cells present in the heart wall of the left ventricle. An earlier PET imaging study demonstrated a 25% reduction of [
11C]CFN binding potential in human heart when naloxone was given in a dose of 0.2 mg/kg via IV administration 5 min prior to injection of the radiotracer [
26]. These results suggest the potential of [
11C]CFN to image cardiac MORs in vivo. In our study, the results of naloxone blockade in the pretreatment study were mixed. In comparison of the [
11C]CFN SUVR
70 − 90 min to the two control scans (baseline and retest) (Supplementary Fig.
10a), an average of 17% reduction of [
11C]CFN SUVR
70 − 90 min was observed in 2 NHPs, a 4% increase of [
11C]CFN SUVR
70 − 90 min was observed in one NHP, and mixed results were observed in another NHP. In the displacement study, there was gradual change of the TACs after naloxone or GSK1521498 administration (Supplementary Fig.
10b). These results indicate that the dose of 0.14 mg/kg naloxone used in the current study may not be high enough to block the binding of [
11C]CFN to cardiac MORs. Hence, more studies are needed to evaluate the feasibility of imaging cardiac MORs by PET in vivo.
It has been reported that there is low density of MORs expressed in small intestines [
15,
16]. However, there was no effect of naloxone or GSK1521498 in the blocking or displacement studies on the uptake of [
11C]CFN in the current study. This may be due to the mixed signal of [
11C]CFN and radiolabeled metabolites in the small intestines, since [
11C]CFN is metabolized in the liver and excreted via the small intestines.
In conclusion, specific uptake of [11C]CFN was observed in the MOR-abundant regions in both CNS and peripheral organs under baseline conditions. The blocking effect of both naloxone and GSK1521498 under pretreatment or displacement conditions were also observed in MOR-abundant regions including brain and cervical spinal cord. It took approximately 4–6 min for naloxone or GSK1521498 to distribute to CNS and displace the binding of [11C]CFN. The pharmacokinetic behavior of GSK1521498 and naloxone displacement were similar, suggesting that GSK1521498 could be a potential alternative to naloxone for reversing opioid overdose. In peripheral organs, only heart wall showed a hint of naloxone or GSK1521498 blockade.
By using the high sensitivity total-body imager, PennPET Explorer, in this study, we were able to image MORs in the entire CNS (i.e., brain and spinal cord). In addition to imaging MORs in the CNS, our study demonstrated the capability to image MORs in the cardiovascular system and peripheral organs (e.g., spleen) in vivo. The use of total body PET in combination with the blocking and displacement studies revealed the potential to study the dynamic interactions between a MOR agonist (i.e., [11C]CFN) and MOR antagonists (i.e., naloxone and GSK1521498) across the brain, spinal cord and peripheral organs in vivo. These data indicate that [11C]CFN total-body PET scans could provide a novel approach for studying mechanism of action of MOR drugs used in the treatment of acute and chronic opioid use disorder, and the effect of chronic opioid use disorder on the expression on MOR in the brain and spinal cord.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.