The Kappa Opioid Receptor Agonist 16-Bromo Salvinorin A Has Anti-Cocaine Effects without Significant Effects on Locomotion, Food Reward, Learning and Memory, or Anxiety and Depressive-like Behaviors

Kappa opioid receptor (KOR) agonists have preclinical antipsychostimulant effects; however, adverse side effects have limited their therapeutic development. In this preclinical study, conducted in Sprague Dawley rats, B6-SJL mice, and non-human primates (NHPs), we evaluated the G-protein-biased analogue of salvinorin A (SalA), 16-bromo salvinorin A (16-BrSalA), for its anticocaine effects, side effects, and activation of cellular signaling pathways. 16-BrSalA dose-dependently decreased the cocaine-primed reinstatement of drug-seeking behavior in a KOR-dependent manner. It also decreased cocaine-induced hyperactivity, but had no effect on responding for cocaine on a progressive ratio schedule. Compared to SalA, 16-BrSalA had an improved side effect profile, with no significant effects in the elevated plus maze, light–dark test, forced swim test, sucrose self-administration, or novel object recognition; however, it did exhibit conditioned aversive effects. 16-BrSalA increased dopamine transporter (DAT) activity in HEK-293 cells coexpressing DAT and KOR, as well as in rat nucleus accumbens and dorsal striatal tissue. 16-BrSalA also increased the early phase activation of extracellular-signal-regulated kinases 1 and 2, as well as p38 in a KOR-dependent manner. In NHPs, 16-BrSalA caused dose-dependent increases in the neuroendocrine biomarker prolactin, similar to other KOR agonists, at doses without robust sedative effects. These findings highlight that G-protein-biased structural analogues of SalA can have improved pharmacokinetic profiles and fewer side effects while maintaining their anticocaine effects.


Introduction
Substance use disorders (SUDs) exact an enormous medical, financial, and emotional toll on society, and there is an unmet need for new prevention and treatment strategies. Although it is now well established that SUDs are disorders of the brain, maintained, in part, by persistent changes in brain function, efficacious treatments for many SUDs have remained elusive [1,2]. Psychostimulant use disorders, in particular, still have no Food and Drug Administration (FDA)-approved treatments, despite their high prevalence.
Preclinical studies have demonstrated that the kappa opioid receptor (KOR) is a promising target for the management of psychostimulant use disorders. For example, the administration of KOR agonists reduced cocaine self-administration in rats, mice, and

16-BrSalA Has a Longer Duration of Action Compared to SalA In Vivo
KOR agonists, including SalA and its analogues, have well-established antinociceptive effects [43,53]. As such, a warm water tail withdrawal assay was used to evaluate the in vivo duration of action of 16-BrSalA compared with SalA ( Figure 2). A two-way ANOVA on the time course data revealed a significant treatment × time interaction, F(18, 234) = 2.21, p = 0.0039. Follow-up comparisons showed that treatment with SalA had significant antinociceptive effects primarily within the first 15 min, whereas 16-BrSalA had significant effects for over 90 min (p < 0.05; Figure 2a). A one-way ANOVA on the area under the curve (AUC) data similarly revealed a significant effect of treatment, F(2, 26) = 6.36, p = 0.0056, with 16-BrSalA, but not SalA, having a significant overall antinociceptive effect during these 150 min compared to vehicle-treated controls (p < 0.05; Figure 2b). These results are consistent with our previous, independent observations that used 16-BrSalA as well as another C16-based analogue, 16-ethynyl SalA [53]. These findings demonstrate that C-16 modifications to SalA can improve pharmacokinetic profiles, which are one of the primary limitations of SalA and its analogues as therapeutic agents. We have previously shown that SalA analogues derived from modifications at the C2 position similarly have a longer duration of action in antinociceptive assays, with improved metabolic stability in vivo [65]. This has been attributed to the lack of a hydrolysable ester [69]. How C16 alterations of SalA impact pharmacokinetic profiles have yet to be thoroughly investigated, however.

16-BrSalA Has a Longer Duration of Action Compared to SalA In Vivo
KOR agonists, including SalA and its analogues, have well-established antinociceptive effects [43,53]. As such, a warm water tail withdrawal assay was used to evaluate the in vivo duration of action of 16-BrSalA compared with SalA ( Figure 2). A two-way ANOVA on the time course data revealed a significant treatment × time interaction, F(18, 234) = 2.21, p = 0.0039. Follow-up comparisons showed that treatment with SalA had significant antinociceptive effects primarily within the first 15 min, whereas 16-BrSalA had significant effects for over 90 min (p < 0.05; Figure 2a). A one-way ANOVA on the area under the curve (AUC) data similarly revealed a significant effect of treatment, F(2, 26) = 6.36, p = 0.0056, with 16-BrSalA, but not SalA, having a significant overall antinociceptive effect during these 150 min compared to vehicle-treated controls (p < 0.05; Figure 2b). These results are consistent with our previous, independent observations that used 16-BrSalA as well as another C16-based analogue, 16-ethynyl SalA [53]. These findings demonstrate that C-16 modifications to SalA can improve pharmacokinetic profiles, which are one of the primary limitations of SalA and its analogues as therapeutic agents. We have previously shown that SalA analogues derived from modifications at the C2 position similarly have a longer duration of action in antinociceptive assays, with improved metabolic stability in vivo [65]. This has been attributed to the lack of a hydrolysable ester [69]. How C16 alterations of SalA impact pharmacokinetic profiles have yet to be thoroughly investigated, however.

16-BrSalA Attenuates Cocaine-Primed Reinstatement and Hyperactivity, but Not Progressive Ratio Responding
The anticocaine effects of 16-BrSalA are shown in Figure 3. As has been previously shown [60], 16-BrSalA produced a dose-dependent attenuation of cocaine + cue-induced reinstatement of drug-seeking behavior (Figure 3a). A one-way repeated measures ANOVA revealed a significant effect of treatment, F(3, 15) = 8.35, p = 0.0017, with SalA and both doses of 16-BrSalA significantly attenuating total active lever responses compared to vehicletreated controls (p < 0.05). Of note, while a low dose of 16-BrSalA (0.3 mg/kg) produced a significant attenuation of drug-seeking behavior, a larger dose (1 mg/kg) was required to produce effects equivalent to SalA (0.3 mg/kg), despite both compounds having a similar affinity, potency, and efficacy for the KOR in vitro [60,67]. This difference might be due to the greater potency of SalA in its ability to modulate DAT function compared to 16-BrSalA (discussed in the next section).

16-BrSalA Attenuates Cocaine-Primed Reinstatement and Hyperactivity, But Not Progressive Ratio Responding
The anticocaine effects of 16-BrSalA are shown in Figure 3. As has been previously show [60], 16-BrSalA produced a dose-dependent attenuation of cocaine + cue-induced reinstat ment of drug-seeking behavior (Figure 3a). A one-way repeated measures ANOVA reveale a significant effect of treatment, F(3, 15) = 8.35, p = 0.0017, with SalA and both doses of 16 BrSalA significantly attenuating total active lever responses compared to vehicle-treated con trols (p < 0.05). Of note, while a low dose of 16-BrSalA (0.3 mg/kg) produced a significant a tenuation of drug-seeking behavior, a larger dose (1 mg/kg) was required to produce e fects equivalent to SalA (0.3 mg/kg), despite both compounds having a similar affinity, po tency, and efficacy for the KOR in vitro [60,67]. This difference might be due to the greate potency of SalA in its ability to modulate DAT function compared to 16-BrSalA (discusse in the next section).  Results of the warm water tail withdrawal assay in vehicle-, salvinorin A (SalA; 1.0 mg/kg)-, or 16-bromo SalA (16-BrSalA; 1.0 mg/kg)-treated mice (n = 8-11/treatment). (a) Effect of treatment on % maximal possible antinociceptive effect as a function of time. (b) Area under the curve (AUC) data for each treatment. * p < 0.05, ** p < 0.01, and *** p < 0.001 compared to the vehicle (Dunnett's test).

16-BrSalA Attenuates Cocaine-Primed Reinstatement and Hyperactivity, But Not Progressive Ratio Responding
The anticocaine effects of 16-BrSalA are shown in Figure 3. As has been previously shown [60], 16-BrSalA produced a dose-dependent attenuation of cocaine + cue-induced reinstatement of drug-seeking behavior (Figure 3a). A one-way repeated measures ANOVA revealed a significant effect of treatment, F(3, 15) = 8.35, p = 0.0017, with SalA and both doses of 16-BrSalA significantly attenuating total active lever responses compared to vehicle-treated controls (p < 0.05). Of note, while a low dose of 16-BrSalA (0.3 mg/kg) produced a significant attenuation of drug-seeking behavior, a larger dose (1 mg/kg) was required to produce effects equivalent to SalA (0.3 mg/kg), despite both compounds having a similar affinity, potency, and efficacy for the KOR in vitro [60,67]. This difference might be due to the greater potency of SalA in its ability to modulate DAT function compared to 16-BrSalA (discussed in the next section).  To determine if these anticocaine effects of 16-BrSalA were KOR-dependent, the effect of cotreatment with the KOR antagonist norbinaltorphimine (nor-BNI) was also determined in a subset of animals ( Figure 3b). A two-way fully repeated measures ANOVA on the time course data revealed a significant treatment × time interaction, F(46, 138) = 3.92, p < 0.0001.
Follow-up comparisons showed that 16-BrSalA treatment produced a significant attenuation of responding from 115 to 120 min (p < 0.05), which was reversed by nor-BNI treatment (p < 0.05), indicating KOR dependency. A one-way repeated measures ANOVA on the summed response data revealed a significant effect of treatment, F(2, 4) = 20.71, p = 0.0078, with 16-BrSalA also significantly attenuating total responding in a KOR-dependent manner (p < 0.05).
The effect of 16-BrSalA on cocaine-produced locomotor hyperactivity is shown in Figure 3c. The cocaine prime was the same dose (20 mg/kg) as that used during the reinstatement tests. A two-way ANOVA on the time course data revealed a significant treatment × time interaction, F(34, 187) = 2.50, p < 0.0001, with significant decreases in ambulatory counts as a function of 16-BrSalA treatment at 15 and 20 min (p < 0.05), which was prevented by nor-BNI pretreatment. The total cocaine-produced ambulatory counts were also significantly decreased via 16-brSal treatment in a KOR-dependent manner (p < 0.05). These results are consistent with those produced by other KOR agonists, including SalA and analogues [55]. Of note, we have previously shown that 16-BrSalA did not decrease spontaneous locomotor activity [60]; therefore, these results are unlikely to be due to a generalized decrease in motor activity, but rather 16-BrSalA counteracting the locomotor-activating effects of cocaine.
These results suggest that, as with other KOR agonists, 16-BrSalA was effective at attenuating the cocaine-induced reinstatement of drug-seeking behavior as well as cocaineinduced locomotor activity. To further expand on these findings, we also determined for the first time if KOR activation via 16-BrSalA or other KOR agonists was similarly effective at attenuating the reinforcing effects of cocaine on a progressive ratio (PR) schedule, where the response requirement increases with each successive infusion (Figure 3d). A oneway repeated measures ANOVA revealed a significant effect of treatment, F(6, 78) = 4.88, p = 0.0003, which was driven by a significant attenuation of PR responding produced by U69,593 (p < 0.05). No significant effects from other treatments were found. This might be due to the longer session duration used during the PR tests (up to 5 h) and the relatively short duration of action of KOR agonists. SalA has been previously shown to significantly decrease breakpoints for sucrose self-administration, but this was during 2 h sessions [70]. In rhesus monkeys, SalA and nalfurafine both decreased responding for oxycodone on a PR schedule; however, in this study SalA/nalfurafine was mixed with oxycodone infusions, rather than administered separately at the start of the session [71].

16-BrSalA Increases DAT Activity in a KOR-and ERK-Dependant Manner
The activation of central dopaminergic mechanisms is crucial for the acute reinforcing as well as locomotor-activating effects of cocaine and other psychostimulants [72,73]. Microdialysis studies have shown that KOR agonists can attenuate cocaine-induced increases in synaptic DA concentrations in the NAcc, and, as such, this has been suggested to be a mechanism underlying the antipsychostimulant effects of KOR agonists [19][20][21][22][23]. We have previously shown that SalA and some analogues can increase DAT activity [24,62]. Here, we determined whether 16-BrSalA would have the same effect ( Figure 4).
To confirm the KOR-dependent effect of 16-BrSalA on DAT activity, the effect of pretrea ment with nor-BNI on the uptake of fixed concentrations of DA was also determined. Separa one-way ANOVAs revealed significant effects of treatment on both NAcc (F(2, 24) = 8.   To confirm the KOR-dependent effect of 16-BrSalA on DAT activity, the effect of pretreatment with nor-BNI on the uptake of fixed concentrations of DA was also determined. Separate one-way ANOVAs revealed significant effects of treatment on both NAcc in both assays (p < 0.05), which was blocked by nor-BNI cotreatment (p < 0.05), indicating KOR dependency.
In cells, we compared the effect of SalA and two doses of 16-BrSalA on the uptake of fixed concentrations of ASP + and determined the effect of pretreatment with nor-BNI or the ERK1/2 inhibitor, U0126 (Figure 4f). A one-way ANOVA revealed a significant effect of KOR agonist treatment, F(3, 224) = 12.83, p < 0.0001, with both SalA and the higher dose 16-BrSalA significantly increasing ASP + uptake (p < 0.05). Separate two-way ANOVAs revealed a significant interaction between 16-BrSalA treatment and U0126 treatment, F(1, 204) = 7.28, p = 0.0076, and a trending interaction between 16-BrSalA treatment and nor-BNI treatment, F(1, 195) = 3.15, p = 0.078. Follow-up comparisons indicated that 16-BrSalA-induced increases in ASP + uptake were both KOR-and ERK1/2-mediated (p < 0.05). ERK1/2 regulates the transport capacity and intracellular trafficking of the DAT in the striatum [74]. We have previously shown that SalA-induced increases in DAT activity were mediated by ERK1/2 [24,62]. The data we collected using 16-BrSalA are consistent with this and support the idea that the anticocaine effects of KOR agonists are, in part, mediated through ERK-dependent DAT modulation. These data also show that SalA is more potent than 16-BrSalA at modulating DAT activity, which might explain why a larger dose of 16-BrSalA was required to produce anticocaine effects comparable to SalA ( Figure 3a).

16-BrSalA Has an Improved Side Effect Profile Compared to SalA
In order to develop clinically viable KOR-targeted treatments, it is vital to identify novel compounds that have reduced side effects compared to U50,488 or SalA. 16-BrSalA shows a greater G-protein signaling bias compared to SalA [53,55] as well as a more stable pharmacokinetic profile (Figure 2), and was, therefore, predicted to produce fewer adverse side effects. This was tested in the current study ( Figure 5). Of note, the dose of 16-BrSalA evaluated in these side effect assays was the dose (1.0 mg/kg) that was required to produce significant anticocaine effects equivalent to SalA (0.3 mg/kg; Figure 3).
To evaluate the potential anxiogenic effects of 16-BrSalA, we used two preclinical models: the elevated plus maze and the light-dark test. A one-way ANOVA on the elevated plus maze data revealed a significant effect of treatment, F(3, 75) = 3.46, p = 0.0205, with both doses of SalA, but not 16-BrSalA, significantly decreasing the time spent on the open arms of the maze (p < 0.05; Figure 5a). Similar results were found in the light-dark test ( Figure 5b): a one-way ANOVA produced a significant effect of treatment, F(3, 55) = 5.96, p = 0.0014, but with only the higher dose of SalA (1.0 mg/kg) having a significant effect (p < 0.05). These results indicate that, in contrast to the parent compound, 16-BrSalA does not have significant anxiogenic effects at these doses, which is consistent with our previous observations in mice using the elevated zero maze and the marble burying task [53]. Of note, neither agonist impacted the total distance travelled during the light-dark test (p > 0.05). Moreover, we have previously shown that 16-BrSalA (1 mg/kg) or SalA (0.3 mg/kg) had no effect on spontaneous locomotor activity in rats [14,60], suggesting that 16-BrSalA does not induce sedation and that sedation was not a confound in the current experiments.
Depressive-like behavior, anhedonia, and reduced non-drug reward behavior are related side effects associated with KOR activation. These effects have been primarily attributed to KOR-mediated decreases in central DAergic activity [6,70,75]. Given that this has been the proposed mechanism through which KOR agonists attenuate the effects of cocaine and other psychostimulants, finding a therapeutic window where KOR agonists can produce antipsychostimulant effects without these side effects is critical. In the current study we examined the effect of 16-BrSalA on the forced swim test, a measure of learned helplessness and depressive-like behavior, and on reinforcement maintained by a sucrose reward. In contrast to what has been previously shown with SalA or other SalA analogues [14,55,64], there was no effect of 16-BrSalA treatment on mobility or immobility time in the forced swim test (p > 0.05; Figure 5c). Similarly, 16-BrSalA had no effect on sucrose reinforcement ( Figure 5d); a one-way repeated measures ANOVA revealed a significant effect of treatment, F(5, 30) = 60.89, p < 0.0001, but this was driven by a large attenuation of responding produced by U50,488 and a small increase produced by U69,593 (p < 0.05). This effect of U50,488 treatment is consistent with previous observations [14,76] and could be driven by the sedative effects of U50,488 [77] or differences in DA-independent mechanisms related to satiety and food consumption [78][79][80].  Depressive-like behavior, anhedonia, and reduced non-drug reward behavior are related side effects associated with KOR activation. These effects have been primarily attributed to KOR-mediated decreases in central DAergic activity [6,70,75]. Given that this has been the proposed mechanism through which KOR agonists attenuate the effects of cocaine and other psychostimulants, finding a therapeutic window where KOR agonists can produce antipsychostimulant effects without these side effects is critical. In the current study we examined the effect of 16-BrSalA on the forced swim test, a measure of learned helplessness and depressive-like behavior, and on reinforcement maintained by a sucrose reward.
In contrast to what has been previously shown with SalA or other SalA analogues [14,55,64], there was no effect of 16-BrSalA treatment on mobility or immobility time in the forced KOR activation has been known to disrupt processes related to learning and memory, with stress-induced impairments playing a particularly important role [81][82][83][84][85]. Here, we show that 16-BrSalA had no effect on learning and memory, as assessed by the novel object recognition task (p > 0.05; Figure 5e). This might be due to a lack of stress induced by 16-BrSalA, given that there was no effect of 16-BrSalA in our anxiety or forced swim tests; however, the further evaluation of both behavioral and biological markers of stress (i.e., corticotrophin-releasing factor) would be needed to confirm this. In our previous report, neither SalA nor mesyl salvinorin B affected novel object recognition [55], although significant effects of SalA on other measures of learning and memory, such as the eight-arm radial maze and passive avoidance, have been reported [86]. U50,488, on the other hand, has been previously shown to cause deficits in the novel object recognition task [84] and showed a trending effect in our study (p = 0.0860), which could be stress-induced [81].
Lastly, we determined the aversive properties of 16-BrSalA in a conditioned place aversion (CPA) paradigm (Figure 5f). As with other SalA analogues previously tested [65], 16-BrSalA treatment had significant conditioned aversive effects in our tests, as did SalA and U50,488 (p < 0.05). There is evidence to suggest that KOR-induced aversion requires β-arrestin and, more specifically, p38 signaling [87,88]. White and colleagues, however, found that conditioned place aversion to U50,488, U69,593, and SalA was still produced in β-arrestin knockout mice [52]. On the other hand, studies have shown that KORmediated aversion also involves interactions with both dopaminergic and serotoninergic mechanisms [89,90]. These studies suggest that the mechanisms underlying KOR-induced aversion are multifaceted; more research on the development of non-aversive KOR agonists is needed.

Effect of 16-BrSalA on ERK1/2 and p38
KOR agonism activates many signal transduction cascades, namely MAPK pathways, which include ERK1/2 and p38. As shown previously, as well as in the current study, the KOR-induced modulation of the DAT is mediated by ERK1/2, although no differences in the KOR modulation of ERK1/2 have been identified. KOR-induced ERK1/2 activation has both β-arrestinand G-protein-dependent mechanisms, with increases in early phase ERK1/2 activation (5-15 min) being G-protein-dependent and late-phase ERK1/2 activation (~2 h) being β-arrestin-dependent [91,92]. Here, we quantified phosphorylated ERK1/2 levels in rat NAcc (Figure 6a), dSTR (Figure 6b), and prefrontal cortex (PFC; Figure 6c) tissue, as well as in HEK-293 cells coexpressing the DAT and KOR (Figure 6d), at different time points following 16-BrSalA treatment. The aim was to determine if, firstly, 16-BrSalA increased ERK1/2 activation, and, secondly, if the time course of activation reflects that of a G-protein-biased agonist. Peak increases in ERK1/2 phosphorylation were observed 10-15 min following 16-BrSalA treatment, though these increases were only significant in cells (p < 0.05). In order to determine if this effect was KOR-and ERK1/2-dependent, the effect of nor-BNI or U0126 cotreatment on 16-BrSalA-induced changes in phosphorylated ERK1/2 levels after a 10-minute incubation was also determined in cells (Figure 6e). Relative to vehicle treatment, 16-BrSalA significantly increased ERK1/2 phosphorylation, which was blocked by nor-BNI, while U0126 abolished all ERK1/2 activity (p < 0.05). The full Western blot membranes from Figure 6 are shown in Figure S1.
We also examined the effect of 16-BrSalA on p38 phosphorylation over time in rat NAcc (Figure 6f), dSTR (Figure 6g), and PFC tissue (Figure 6h). Because p38 phosphorylation is largely dependent on β-arrestin recruitment [87], G-protein-biased agonists, such as 16-BrSalA, would be expected to produce minimal effects. We found small increases in all brain regions, which peaked at 15 min, but this was only statistically significant in the NAcc, t(4) 4.23, p < 0.0134. The KOR-induced activation of p38 has been suggested to contribute to some of the undesirable side effects of KOR agonists [44,87,93,94]. The small but significant increase in p38 activity in the NAcc following 16-brSalA treatment might therefore contribute to the conditioned aversive response observed in the current study. Phosphorylated p38 (p-p38) was similarly measured in rat NAcc (f), dSTR (g), and PFC (h) tissue collected 0-120 min following treatment with 16-BrSalA (n = 5-7/timepoint/region). Representative Western blot scans are displayed above each graph. p-ERK1/2 and p-p38 expression were normalized to total ERK1/2 and p38, respectively, and expressed as fold change from the baseline (time 0/vehicle + vehicle treatment). * p < 0.05, ** p < 0.01, and **** p < 0.0001 compared to the baseline (one-sample t-test).

Effect of 16-BrSalA on Prolactin as a Neuroendocrine Biomarker in Non-Human Primates
Lastly, in order to provide translational data on 16-BrSalA, we evaluated the effect of 16-BrSalA on prolactin levels in blood samples collected from male rhesus monkeys (n = 3). Peripheral blood levels of prolactin can be used as a translationally valid, quantitative neuroendocrine biomarker of KOR-mediated effects, including the potency and apparent efficacy of ligands [95][96][97]. Figure 7a shows the cumulative dose-effect curves produced by 16-BrSalA, SalA, the partial KOR agonist nalorphine, and the vehicle on prolactin levels. Figure 7b shows the duration of effect of 16-BrSalA and Figure 7c shows the sedation score. Compared with Sal A, 16-BrSal A was approximately 0.5 log units less potent at inducing the release of prolactin, with similar maximal effects, larger than those of the KOR partial agonist nalorphine. Importantly, at doses that produced robust neuroendocrine effects, 16-BrSalA only produced slight sedation. The difference in potency between 16-BrSalA and SalA is comparable with the effects observed in drug-seeking models in rats (Figure 3a), anxiety models in rats (Figure 5a,b) as well as mice [53], motor incoordination tests in mice [53], and DAT modulation in vitro (Figure 4f). Together, these would suggest that the differences in the effectiveness of 16-BrSalA and SalA to produce several relevant effects may be similar across species.
4.23, p < 0.0134. The KOR-induced activation of p38 has been suggested to contribute to some of the undesirable side effects of KOR agonists [44,87,93,94]. The small but significant increase in p38 activity in the NAcc following 16-brSalA treatment might therefore contribute to the conditioned aversive response observed in the current study.

Effect of 16-BrSalA on Prolactin as A Neuroendocrine Biomarker in Non-Human Primates
Lastly, in order to provide translational data on 16-BrSalA, we evaluated the effect of 16-BrSalA on prolactin levels in blood samples collected from male rhesus monkeys (n = 3). Peripheral blood levels of prolactin can be used as a translationally valid, quantitative neuroendocrine biomarker of KOR-mediated effects, including the potency and apparent efficacy of ligands [95][96][97]. Figure 7a shows the cumulative dose-effect curves produced by 16-BrSalA, SalA, the partial KOR agonist nalorphine, and the vehicle on prolactin levels. Figure 7b shows the duration of effect of 16-BrSalA and Figure 7c shows the sedation score. Compared with Sal A, 16-BrSal A was approximately 0.5 log units less potent at inducing the release of prolactin, with similar maximal effects, larger than those of the KOR partial agonist nalorphine. Importantly, at doses that produced robust neuroendocrine effects, 16-BrSalA only produced slight sedation. The difference in potency between 16-BrSalA and SalA is comparable with the effects observed in drug-seeking models in rats (Figure 3a), anxiety models in rats (Figure 5a,b) as well as mice [53], motor incoordination tests in mice [53], and DAT modulation in vitro (Figure 4f). Together, these would suggest that the differences in the effectiveness of 16-BrSalA and SalA to produce several relevant effects may be similar across species.

Subjects
Adult male B6-SJL mice (23-26 g) were used in the tail withdrawal assay. Mice were obtained from the Malaghan Institute of Medical Research (Wellington, New Zealand) and housed within the School of Biological Sciences animal facility, Victoria University of Wellington (Wellington, New Zealand). Adult male Sprague Dawley rats (300-400 g) were used in all other rodent experiments. Rats were bred and housed within the School of Biological Sciences or the School of Psychology animal facilities (Wellington, New Zealand). Food and water were available ad libitum for all of the rodents, except for the rats in the sucrose selfadministration experiment, which were maintained at 85% of their feeding weight via food restriction. Mice were housed in groups of 4-5, while rats were housed either individually

Subjects
Adult male B6-SJL mice (23-26 g) were used in the tail withdrawal assay. Mice were obtained from the Malaghan Institute of Medical Research (Wellington, New Zealand) and housed within the School of Biological Sciences animal facility, Victoria University of Wellington (Wellington, New Zealand). Adult male Sprague Dawley rats (300-400 g) were used in all other rodent experiments. Rats were bred and housed within the School of Biological Sciences or the School of Psychology animal facilities (Wellington, New Zealand). Food and water were available ad libitum for all of the rodents, except for the rats in the sucrose self-administration experiment, which were maintained at 85% of their feeding weight via food restriction. Mice were housed in groups of 4-5, while rats were housed either individually (cocaine self-administration) or 2-3/cage (all other experiments). The animal facilities were temperature (19-21 • C)-as well as humidity (55%)-controlled and on a 12 h light/dark cycle (lights on at 07:00), with testing conducted during the light cycle. All experimental protocols involving rodents were approved by the Victoria University of Wellington Animal Ethics Committee, New Zealand.
Adult, gonadally intact, and captive-bred male rhesus monkeys (Macaca mulatta, weight range: 9-12 kg) were used for the prolactin assay. They were singly housed in a stable colony room maintained at 20-22 • C with controlled humidity and lights on from 07:00 to 19:00. Monkeys had visual, auditory, and olfactory access to other conspecifics, and an environmental enrichment plan was in place. They were fed appropriate amounts of primate chow (PMI Feeds, Richmond, VA, USA), supplemented by treats. Water was available ad libitum. Consecutive experiments in the same subject were typically separated by at least 72 h, with all experiments carried out between 10:00 and 14:00 h. All experiments involving non-human primates were approved by the Rockefeller University Animal Care and Use Committee, in accordance with the Guide for the Care and Use of Animals (National Academy Press; Washington, DC, USA).

Drugs and Treatment
SalA was isolated and purified from Salvia divinorum leaves as previously described [68]. 16-BrSalA was synthesized as previously described [60]. U50,488, U69,593, U0126, and DA were purchased from Sigma Aldrich (St. Louis, MO, USA). ASP + was purchased from Tocris Biosciences (supplied by Pharmaco NZ Ltd., Auckland, New Zealand). KOR agonists were suspended in 75% DMSO for reinstatement, locomotor activity, and forced swim tests, as well as for ex vivo tissue collection. A vehicle with a 4:1:5 ratio of propylene glycol:dimethyl sulfoxide (DMSO):phosphate-buffered saline (PBS) was used for the tail withdrawal procedure, and a 2:1:7 ratio of DMSO:Tween-80:water was used for all of the other in vivo tests in rodents. A vehicle containing a 1:1:8 ratio of ethanol:Tween-80:water was used for the prolactin assay. For rodent tests, all of the agonists were administered at a volume of 1 mL/kg intraperitoneally (i.p.), except for U69,593, which was administered subcutaneously (s.c.). Agonists were administered either 5 min (SalA), 10 min (16-BrSalA, U50,488), or 15 min (U69,593) prior to testing based on previous work [14,56] and from the results of the duration of action experiment in the current study. For the prolactin assay, agonists were administered intravenously (i.v.) at a volume of 0.05-0.1 mL/kg. The KOR antagonist nor-BNI was dissolved in 0.9% NaCl and administered subcutaneously 24 h prior to testing at a volume of 1 mL/kg. Cocaine-HCl (BDG synthesis; Wellington, New Zealand) was dissolved in 0.9% NaCl containing sodium heparin (3 U/mL) for intravenous

Surgery and Cocaine Self-Administration
Indwelling intravenous catheters were surgically implanted in the external jugular vein as previously described [14]. Rats were trained to self-administer in standard operant chambers equipped with two levers (Med Associates, Fairfax, VT, USA; model ENV-001) and a mechanical syringe pump (Med Associates, USA; model PHM100A). The depression of the active lever resulted in a 0.1 mL i.v. infusion of cocaine (0.5 mg/kg/infusion) over 12 s and the illumination of a light located above the lever. The depression of the inactive lever resulted in no planned consequence. Drug delivery and recording the number of active/inactive lever responses made were controlled by Med PC software (v4.2, Med Associates, Fairfax, VT, USA).
All self-administration sessions were conducted during 2 h sessions, 6 days per week. Rats were initially trained on a fixed-ratio (FR) 1 schedule of reinforcement until responding had stabilized at ≥20 infusions per session with an active:inactive ratio of ≥2:1 for 3 consecutive sessions. Thereafter, the schedule of reinforcement was increased to FR2 until this criterion had been reached once more before the schedule of reinforcement was increased again to FR5. Rats were run for 10 days at FR5 to establish stable baseline responding before beginning the reinstatement or PR experiments.
As has been previously reported [60], the reinstatement experiment was run in three repeating phases. In phase 1 of reinstatement (baseline), rats (n = 6) self-administered cocaine on an FR5 schedule for at least 2 days, until their responses were within 20% of their prereinstatement baseline. In phase 2 (extinction) the light cue was removed and cocaine was substituted for the vehicle until responses had dropped to <20 within a single session (3-4 days). In phase 3 (reinstatement), rats received either the vehicle, SalA (0.3 mg/kg), or 16-BrSalA (0.3 or 1.0 mg/kg) prior to a priming injection of cocaine (20 mg/kg). The light stimulus that was previously paired with cocaine infusions was also reintroduced, and the number of responses made was recorded. Treatment was administered in a within-subjects Latin square design. The effect of nor-BNI (10 mg/kg) treatment combined with 16-BrSalA (1.0 mg/kg) was also determined in three of these animals, but this was carried out as the final test due to the long-lasting effects of nor-BNI. No significant differences in baseline or extinction responding were observed in-between reinstatement treatments (p > 0.05). The mean inactive lever responding ranged from 1.5 to 2.3 across treatments, and similarly did not significantly differ (p > 0.05).

Cocaine-Induced Locomotor Activity
Locomotor activity was measured in clear plexiglass chambers (42 × 42 × 30 cm; Med Associates, USA; model ENV-515) set in sound-attenuating boxes. Each chamber contained a lattice of 32 infrared beams, 1.7 cm above the floor of the chamber. The sequential interruption of 3 beams was recorded as one ambulatory count. Counts were recorded in 5 min bins via activity-monitoring software (Med Associates, USA). All locomotor experiments were conducted in the dark and in the presence of white noise. The rats were placed into locomotor activity chambers for 30 min prior to receiving either the vehicle (n = 6) or 16-BrSalA (1.0 mg/kg, n = 6). Another group of rats were pretreated with nor-BNI (10 mg/kg) before receiving 16-BrSalA (1.0 mg/kg; n = 2). Thereafter, all of the rats received an injection of cocaine (20 mg/kg) and were returned to the locomotor activity chambers for an additional 60 min. Ambulatory counts were recorded during the entire 90 min.

Light-Dark Test
A large white chamber (30 × 30 × 34 cm) connected to a smaller black chamber (15 × 30 × 34 cm) via a small grey corridor (8 × 10 × 34 cm) was used to carry out the light-dark tests. Testing was conducted in a dark room with three LED lamps: one directed at the light box, one at the corridor, and one at the ceiling. The intensity of the light in each chamber was as follows: 100 lux in the light chamber, 10 lux in the dark chamber, and 70 lux in the corridor. The rats were treated with either the vehicle (n = 24), SalA (0.3 mg/kg, n = 13; 1.0 mg/kg, n = 13), or 16-BrSalA (1.0 mg/kg, n = 9), and placed in the black chamber. The time spent in the light box as well as the total distance travelled were then determined during 15 min of free access. Activity was measured via the use of SMART 3.0 software (Panlab, Harvard Apparatus, Holliston, MA, USA).

Elevated plus Maze
The EPM maze consisted of four arms (50 cm long × 10 cm wide) elevated 55 cm above the ground. Two of the arms had a small parapet measuring 2.5 cm in height (open arms), while the other two arms were enclosed by 40 cm-high black walls (closed arms). The rats were habituated to the conditions in the testing room for 60 min and then treated with either the vehicle (n = 29), SalA (0.3 mg/kg, n = 16; 1.0 mg/kg, n = 19), or 16-BrSalA (1.0 mg/kg, n = 15), before being placed in the center of the apparatus facing an open arm. All subsequent activity was recorded for 5 min via the use of a Sony HDR-SR5E digital camera recorder, and the time spent in each arm was calculated by an observer blinded to the experimental treatment. Open arm time was calculated only when the rats had all four paws in the open arm.

Forced Swim Test
The forced swim test was conducted according to the methods described in [99], using a cylindrical chamber measuring 44 cm in height and 20 cm in diameter, which was filled with water (25 ± 1 • C) to a height of 35 cm. Testing was conducted over 2 days. On both days, rats were habituated to the testing room for 60 min. On day 1, drug-naïve rats were habituated to forced swimming conditions for 15 min. On day 2, rats received either vehicle (n = 8) or 16-BrSalA (1.0 mg/kg, n = 9) treatment before a 5 min test session. Sessions were recorded via the use of a SONY HDR-SR5E digital camera recorder and analyzed for mobility as well as immobility time, which was measured in 5 sec intervals by an observer blinded to experimental treatments.

Sucrose Self-Administration
The rats maintained at 85% of their free-feeding weight were trained to self-administer sucrose pellets (Dustless Precision Pellet, 45 mg sucrose; Able Scientific, Perth, Australia) in standard operant chambers (Med Associates, USA, ENV-011) with two levers: one lever was connected to a sucrose pellet dispenser and a light (active lever), while the second lever had no programmed function (inactive lever). Sucrose self-administration sessions were conducted during 45 min sessions, 6 days per week. Rats (n = 7) were initially trained on an FR1 schedule of reinforcement until they had self-administered ≥20 pellets during a single session with an active:inactive ratio of ≥2:1. Thereafter, rats progressed to an FR5 schedule for at least 5 days to establish stable baseline response rates. The effect of the vehicle, U50,488 (10 mg/kg), U69,593 (0.3 mg/kg), SalA (0.3 mg/kg), or 16-BrSalA (1.0 mg/kg) on sucrose self-administration was then determined in a within-subjects Latin square design. Sucrose delivery and the recording of the number of lever responses made were controlled via Med PC software (Med Associates, USA).

Novel Object Recognition
The novel object recognition test was carried out based on previously published methods [85]. On days 1-3 (habituation), rats (n = 21) were habituated to the testing chamber (45 × 45 × 35 cm, open field) for 30 min/day. On day 4 (familiarization), rats were familiarized with two identical objects, which were introduced on either side of the chamber (3 times for 6 min, with an inter-trial interval of 10 min). On day 5 (test), one of the familiar objects was replaced with a novel object, and the time spent interacting with both objects was recording via the use of SMART 3.0 software (Panlab, Harvard Apparatus, Holliston, MA, USA). The recognition index was calculated as follows: Molecules 2023, 28, 4848 15 of 22 where N is the time (in seconds) spent with the novel object and F is the time spent with the familiar object. Rats that only interacted with a single object were excluded. The effect of treatment with either the vehicle, U50,488 (10 mg/kg), SalA (0.3, 1.0 mg/kg), or 16-BrSalA (1.0 mg/kg) was determined via the use of a within-subjects Latin square design.

Conditioned Place Aversion
CPA was conducted via the use of a biased procedure based on previously published methods [100]. A 3-chamber apparatus was used (PanLab, Harvard Apparatus, USA), which had two large chambers (30 × 30 × 34 cm) connected by a small corridor (8 × 10 × 34 cm) with removable sliding doors. One of the large chambers had a textured black floor with black dotted patterns on its walls (black chamber), and the other had a smooth white floor with black striped patterns on its walls (white chamber). The corridor was a neutral zone with grey walls as well as floor, and illuminated at an intensity of 70 lux. The average light intensity in both conditioning chambers was 20 lux. Experiments were conducted in the presence of white noise.
The CPA procedure was conducted over 9 days. On day 0, rats were habituated to the CPA apparatus for 15 min. On day 1 (preconditioning), rats were given free access to all of the chambers for 15 min. Animals that showed >80% preference for a particular chamber or >40% preference for the corridor were excluded. On days 2-7 (conditioning), rats were treated with either the vehicle (n = 9), U50,488 (10 mg/kg, n = 11), SalA (0.3 mg/kg, n = 9), or 16-BrSalA (1.0 mg/kg, n = 9) and confined in their preferred chamber for 45 min. On alternating days, in a counterbalanced order, rats received the vehicle and were confined to their less preferred chamber for 45 min. On day 8 (postconditioning), rats were placed in the corridor and given free access to the apparatus for 15 min. The time spent in each chamber was recorded during the pre-and postconditioning sessions via the use of SMART 3.0 software (Panlab, Harvard Apparatus, Holliston, MA, USA).

Prolactin Assay
Following extensive habituation and chair training, an indwelling catheter (27 gauge, Surflo; Terumo, Tokyo, Japan) was placed percutaneously in a saphenous vein and attached to a Luer multisample injection plug (the catheter was removed at the end of each experiment). The catheter and plug were flushed with 0.3 mL of heparinized sterile saline (50 U/mL) prior to use, as well as after each injection or sampling. Approximately 15 min following catheter placement, two baseline blood samples were obtained (approximately 2 mL each). These baseline samples were collected 5 min apart from each other (approximately −10 and −5 min, relative to the onset of dosing). These blood samples were placed in a plain vacutainer and kept at room temperature until the time of spinning (3000 rpm at 4 • C) and serum separation. Experiments were carried out with a cumulative dosing procedure, where doses of SalA (0.001-0.032 mg/kg, i.v.) or 16-BrSalA (0.0032-0.1 mg/kg, i.v.) were administered in increasing 0.5 log unit steps every 30 min, and a blood sample was taken 15 min after each dose. The KOR partial agonist nalorphine (0.1-3.2 mg/kg) was examined in an identical manner [97]. A repeated vehicle condition was studied under identical timing and sampling conditions. Prior studies have shown that such cumulative dose-effect curve procedures can be efficiently used in this assay to examine the potency and apparent efficacy of opioid ligands [97,101]. Serum samples were kept at −40 • C until the time of analysis, typically within 2 weeks of collection. Samples were analyzed in duplicate with a standard human prolactin immunoradiometric kit (MP Biomedicals; Solon, OH, USA), following the manufacturer's instructions. Prolactin data were expressed as the change from the mean baseline (∆ng/mL), via subtracting the mean value obtained at time −10 and −5 min, on each test day.
In an exploratory study, the sedation rating scores of 16-BrSalA and SalA were compared during the neuroendocrine experiments (just prior to blood sampling) via the use of a previously validated observational rating scale [59].

Rotating Disk Electrode Voltammetry
DA uptake in rat brain tissue suspensions was measured via the use of RDEV, as has been previously described [24,62]. Briefly, NAcc and dSTR tissues from drug-naïve rats were dissected, weighed, minced in an ice-cold KREBS buffer, and transferred to microcentrifuge vials. The tissues were then washed 8 times in a carbogen-aerated KREBS buffer at 37 • C, before being resuspended in 296 µL of KREBS and transferred to the RDEV chamber. A rotating glassy carbon electrode was lowered into the chamber and rotated at 2000 rpm with an MSR rotator (Pine instruments, AFMDO3GC, Durham, NC, USA). A +450 mV potential versus an Ag/AgCl reference electrode was applied, and the resulting current was measured with an eDAQ recorder (Denistone, NSW, Australia). Tissue suspensions were left to reach a stable baseline ( ∼ =10 min) before each test.
A low to infinite trans model was used to determine the uptake kinetics of the DAT, following sequential additions of increasing DA concentrations (0.5-4 µM, 4 µL) in tissue suspensions pretreated with 16-BrSalA (0 or 500 nM, 4 min pretreatment; n = 9/treatment/region). A zero trans model was used to determine DAT uptake following a single addition of DA (2 µM, 4 µL) in tissue suspensions pretreated with 16-BrSalA (0 or 500 nM, 4 min pretreatment) and nor-BNI (1 µM, 30 min pretreatment; n = 8-9/treatment/region). Uptake data were collected for 10 s, beginning 1 s after the addition of DA. A linear regression was calculated and normalized to a standard concentration curve in order to determine DA uptake, which was expressed as the pmol/s/g of tissue. For the low to infinite trans model, a Michaelis-Menten curve was fitted using GraphPad Prism (San Diego, CA, USA), and the V max as well as K m were determined for each suspension.

Imaging of ASP + Uptake
The DAT uptake kinetics was also determined in HEK-293 cells coexpressing yellow fluorescent protein human DAT (YFP-hDAT) and rat myc-tagged-KOR (myc-rKOR) via measuring the uptake of the monoamine transporter substrate ASP + , as has been previously described [24,102,103]. Briefly, transfected cells were aspirated of a medium and washed twice in a KREBS buffer, before being placed in a stage-mounted incubator (37 • C) mounted within an Olympus Fluoview FV1000 confocal microscope (Sydney, NSW, Australia). Cells were preincubated with 16-BrSalA (0 or 10 µM) in 1 mL of a KREBS buffer for 2 min. The microscope was then focused on a cell monolayer, and the KREBS buffer was removed so that background autofluorescence could be determined via the capturing of a referencing image. ASP + (1 mL, 1-16 µM) in a KREBS buffer containing 16-BrSalA (0 or 10 µM) was then added to different dishes. Cells were imaged every 5 s for 10 min to capture YFP (485 nm excitation, 545 nm emission) and ASP + fluorescence (570 nm excitation, 670 nm emission). The linear slope of ASP + accumulation was determined over a 60 sec period at the time of maximal effect, following correction for background fluorescence and normalization to YFP-hDAT expression (n = 17-51 cells/concentration/treatment). Data were entered into GraphPad Prism and Michaelis-Menten curves were fitted to determine the V max as well as K m .
To determine the KOR-and ERK1/2-dependant manner of ASP + uptake, separate dishes were preincubated with either nor-BNI (0 or 1 µM) or U0126 (0 or 20 µM) 30 min prior to the addition of a single concentration of ASP + (1 mL, 10 µM) and, 5 min later, 16-BrSalA (0 or 10 µM). The linear slope of ASP + accumulation was determined over a 60 sec period prior to the addition of 16-BrSalA and at the time of maximal effect. Cells without a linear uptake were discarded. The percentage change in the rate of ASP + uptake was then calculated by comparing the change in the slope of ASP + accumulation before and 16-BrSalA treatment (n = 38-65 cells/treatment).
The quantification of ERK1/2 and p38 was carried out as previously described [62]. Fifty micrograms of tissue or cell lysates was run on 10% SDS-PAGE gels. Following electrophoresis, separated proteins were transferred onto an Immobilon PVDF membrane, blocked with 5% bovine serum albumin (BSA) in tris-buffered saline (TBS) containing 0.1% Tween-20 (T-TBS) for 1 h at room temperature, and then probed with the primary antibody diluted in a blocking buffer overnight (1:500 mouse monoclonal p-ERK1/2, 1:1000 rabbit monoclonal ERK1/2, 1:1000 rabbit monoclonal p-p38, and 1:1000 rabbit monoclonal p38). The membrane was then washed 3 times with T-TBS before being probed with the secondary antibody (1:5000 goat anti-mouse Cy5, 1:5000 goat anti-rabbit Cy5) diluted in TBS at room temperature for 1 h. The membrane was then washed 3 times with TBS and scanned via the use of a FUJIFILM FLA-5000 scanner (Fujifilm, Tokyo, Japan). Membranes were first probed and scanned for the phosphorylated protein, before being stripped, probed, and scanned for the total protein (stripping buffer: 0.05 M Tris-HCl, 2% SDS, and 3.75% β-mercaptoethanol, pH 6.8, incubated at 40 min at room temperature followed by five-seven washes in T-TBS).
All Western blots were analyzed via the use of ImageJ (NIH). All band densities were background-corrected and corrected for protein loading differences through normalizing the density of phosphorylated protein bands to the corresponding total protein band. The data were normalized to vehicle-treated controls to enable comparisons between membranes. The size of visual protein bands was quantified via the use of their linear migration properties to enable the identification of proteins. Phosphorylated ERK1/2 and p38 expression was normalized to total ERK1/2 and p38, respectively, and expressed as fold change from the baseline (time 0/vehicle + vehicle treatment).

Statistical Analysis
Statistical analyses were carried out via the use of GraphPad Prism (v9.1.0, La Jolla, CA, USA) with the recommended parameters. Experiments with time course data were analyzed via the use of a two-way (treatment × time) analysis of variance (ANOVA), with time as a repeated measure. Sphericity was not assumed, and thus Greenhouse-Geisser corrections to degrees of freedom were applied. Experiments without time course data were analyzed via the use of one-way ANOVAs (repeated measures where appropriate) or t-tests. Significant interactions or effects of treatment were followed-up with multiple-comparison testing using the recommended tests., i.e., Šídák's test for comparing treatments vs. controls, Dunnett's test for comparing multiple treatment groups vs. controls, or Tukey tests for comparing all of the groups in experiments with antagonists/inhibitors. Normalized Western blot data were analyzed via the use of one-sample t-tests. The results were considered significant when p < 0.05.

Conclusions
The KOR has been identified as a promising target for the management of psychostimulant use disorders and for the development of non-addictive pain medications. Unfortunately, KOR agonists are typically associated with various adverse side effects that limit their clinical viability. In the current study, we showed that the SalA analogue 16-BrSalA had significant anticocaine effects in drug-seeking and locomotor hyperactivity models.
We showed that 16-BrSalA produced an ERK-dependent increase in DAT activity, which likely underlies these anticocaine effects. Importantly, 16-BrSalA produced minimal side effects, with no significant effects on all tests other than conditioned aversion. Furthermore, we showed that 16-BrSalA increased early phase ERK1/2 phosphorylation and had a small effect on p38 phosphorylation. These results provide support for the idea that KOR agonists with differential signaling can be developed to dissociate desirable therapeutic effects from side effects. More research into the development of non-aversive KOR agonists is still needed, however.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules28124848/s1, The full Western blot membranes from Figure 6 are shown in Figure S1.  Institutional Review Board Statement: All experimental protocols involving rodents were approved by the Victoria University of Wellington Animal Ethics Committee, New Zealand (AEC#22334). All experiments involving non-human primates were approved by the Rockefeller University Animal Care and Use Committee, in accordance with the Guide for the Care and Use of Animals (National Academy Press; Washington, DC, USA).

Data Availability Statement:
Data are presented within the manuscript. Additional raw data are available on request from the corresponding author.