Chronic phenmetrazine treatment promotes D₂ dopaminergic and α2-adrenergic receptor desensitization and alters phosphorylation of signaling proteins and local cerebral glucose metabolism in the rat brain

Abstract

Phenmetrazine (PHEN) is a putative treatment for cocaine and psychostimulant recidivism; however, neurochemical changes underlying its activity have not been fully elucidated. We sought to characterize brain homeostatic adaptations to chronic PHEN, specifically on functional brain activity (local cerebral glucose utilization), G-Protein Coupled Receptor-stimulated G-protein activation, and phosphorylation of ERK1/2^{Thr202/Tyr204}, GSK3β^Tyr216, and DARPP-32^Thr34. Male Sprague-Dawley rats were implanted with sub-cutaneous minipumps delivering either saline (vehicle), acute (2-day) or chronic (14-day) low dose (25 mg/kg/day) or high dose (50 mg/kg/day) PHEN. Acute administration of high dose PHEN increased local cerebral glucose utilization measured by 2-[¹⁴C]-deoxyglucose uptake in basal ganglia and motor-related regions of the rat brain. However, chronically treated animals developed tolerance to these effects. To identify the neurochemical changes associated with PHEN’s activity, we performed [³⁵S]GTPγS binding assays on unfixed and immunohistochemistry on fixed coronal brain sections. Chronic PHEN treatment dose-dependently attenuated D₂ dopamine and α₂-adrenergic, but not 5-HT_1A, receptor-mediated G-protein activation. Two distinct patterns of effects on pERK1/2 and pDARPP-32 were observed: 1) chronic low dose PHEN decreased pERK1/2, and also significantly increased pDARPP-32 levels in some regions; 2) acute and chronic PHEN increased pERK1/2, but chronic high dose PHEN treatment tended to decrease pDARPP-32. Chronic low dose, but not high dose, PHEN significantly reduced pGSK3β levels in several regions. Our study provides definitive evidence that extended length PHEN dosage schedules elicit distinct modes of neuronal acclimatization in cellular signaling. These pharmacodynamic modifications should be considered in drug development for chronic use.

Graphical Abstract

1. Introduction

Psychostimulants are commonly abused drugs that evoke a wide range of behavioral responses by increasing or prolonging extracellular levels of dopamine (DA), norepinephrine (NE), serotonin (5-HT), and other neurotransmitters in the brain. The euphoric and reinforcing effects of these drugs are primarily attributed to their ability to either increase DA levels in mesocorticolimbic brain regions or prolong its residence time extracellularly in these brain regions, resulting in exaggerated activation of the brain’s natural pleasure and reward circuitry. In general, psychostimulant compounds produce this increase in extracellular DA in one of two ways: ‘DA transporter (DAT) inhibitors’ such as cocaine and methylphenidate block DA reuptake via direct inhibition of DAT. On the other hand, ‘monoamine releasers’ such as amphetamine (AMPH) and phenmetrazine (PHEN) have the capacity to be actively transported into presynaptic neurons by DAT and promote the release of DA from storage vesicles into the extracellular space including the synapse (Freyberg et al., 2016; Riddle et al., 2005).

Although the acute behavioral and clinical effects of DAT inhibitors and monoamine releasers are similar (Ciccarone, 2011), the difference between these two mechanisms can have profound implications for chronic psychostimulant use. Prolonged use of these compounds can cause lasting changes in receptor-mediated signal transduction (Calipari et al., 2014), downstream signaling cascades (Perrine et al., 2008), and other neuroadaptations thought to underlie the neurological changes, cognitive impairment, cravings, and psychotic features exhibited by long-term stimulant users.

PHEN was first discovered in Europe in 1952, and the monoamine releaser was approved for use in the clinic as an anorectic in 1954. Abuse of the psychostimulant spread in the U.S. during the 1960s and early 1970s, which ultimately led to its current classification as a Schedule III-controlled substance. However, evidence now suggests that PHEN may be useful in the treatment of cocaine addiction. Continuous 14-day PHEN treatment of rats decreased responding for cocaine (but not food), and PHEN was more effective than AMPH at preventing cocaine-induced increases in responding during reinstatement (Czoty et al., 2015). Furthermore, continuous 7-day treatment of rhesus monkeys with PHEN reduced cocaine (but not food) self-administration (Negus et al., 2009). Importantly, PHEN was shown to be significantly less potent than cocaine as a reinforcer under a progressive ratio schedule of administration in male rhesus monkeys (Minkiewciz et al., 2020)

In contemporary pharmacotherapies, the clinically-available anorectic drug phendimetrazine is used as a prodrug for the monoamine releaser PHEN. Phendimetrazine is slowly metabolized to PHEN via N-demethylation by cytochrome P450 liver enzymes, resulting in prolonged plasma levels of PHEN. Like PHEN, phendimetrazine reduces cocaine self-administration in rhesus monkeys (Banks et al., 2013a), but does so without affecting DA uptake, DA release, or extracellular DA levels (Rothman et al., 2002). Importantly, daily phendimetrazine administration decreased cocaine self-administration in rhesus monkeys by 30-90% for at least four weeks (Czoty et al., 2016). For these reasons, phendimetrazine is being considered as a treatment for cocaine addiction.

Our hypothesis was that prolonged exposure to PHEN would result in pharmacodynamics changes in the brain that are dependent upon duration of exposure (acute versus chronic), and dosage. To test this hypothesis, we examined changes in brain activity, G-protein coupled receptor (GPCR) function, and signaling protein phosphorylation resulting from PHEN treatment of rats with acute (s.c. minipump implanted for 2 days), or chronic (s.c. minipump implanted for 14 days) low dose (25 mg/kg/day) or high dose (50 mg/kg/day) PHEN. We found that administration of PHEN produced widespread AMPH-like increases in local cerebral glucose utilization, which was reduced to control levels upon chronic administration. Rats treated chronically with PHEN showed AMPH-like desensitization in D₂ dopamine and α₂-adrenergic receptor-mediated activation of G-proteins compared to saline controls. We further found that PHEN treatment altered key protein phosphorylation signals dopamine- and cAMP-regulated phosphoprotein-32 (DARPP-32), extracellular signal regulated kinase (ERK1/2), and glycogen synthase kinase 3β (GSK3β) in a number of mesocorticolimbic brain regions.

2. Results

In order to identify brain regions undergoing increased metabolic activity associated with behavioral responses to acute PHEN administration in the rat, we examined local cerebral glucose utilization using the 2-[¹⁴C]-deoxyglucose method (Table 1). There were significant main effects of treatment (F(_2,15) = 8.101, p <001) and brain region (F(_18,15) = 253.498, p <.001), as well as a significant region x treatment interaction (F(_18,36) = 1.857, p <.005). Multiple comparisons demonstrated that the rates of glucose utilization were significantly greater in rats treated acutely with PHEN compared with controls. However, those rats treated chronically with PHEN minipumps did not differ from controls. Within individual brain regions, significantly greater rates of cerebral glucose utilization were observed in the caudate (+19%), nucleus accumbens (+16%), motor cortex (+15%), hippocampus (+15%), globus pallidus (+25%), entopeduncular nucleus (+20%), medial thalamus (+14%), subthalamic nuclei (+26%), substantia nigra compacta (+17%), and substantia nigra reticulata (+28%) of rats treated acutely with PHEN compared with the values of control animals.

Table 1.

Effects of the acute and chronic administration of phenmetrazine (50 mg/kg/day) on rates of local cerebral glucose utilization in rats.

Brain Region	Control (N=7)	Acute (N=6)	Chronic (N=5)
Prefrontal cortex	84.± 1.3	88 ± 2.5	85 ± 2.2
Anterior cingulate cortex	114 ± 2.1	121 ± 3.1	120 ± 5.9
Motor cortex	113 ± 2.0	131 ± 3.9**	122 ± 3.3
Corpus callosum	38 ± 1.4	43 ± 2.2	40 ± 2.6
Caudate	107 ± 2.4	121 ± 3.1**	113 ± 3.6
Nucleus accumbens	78 ± 2.3	96 ± 3.1*	85 ± 1.3
Lateral septum	59 ± 1.2	66 ± 2.7	63 ± 3.2
Amygdala	88 ± 3.6	94 ± 1.3	89 ± 3.8
Hippocampus	70 ± 2.3	80 ± 1.3*	77 ± 3.2
Globus pallidus	58 ± 1.5	73 ± 2.5**	67 ± 3.1
Entopeduncular nucleus	57 ± 2.3	68 ± 1.6*	64 ± 3.7
Lateral thalamus	99 ± 4.5	111 ± 3.2	114 ± 6.6
Medial thalamus	109 ± 4.2	124 ± 3.6*	118 ± 5.2
Subthalamic nucleus	100 ± 4.1	126 ± 4.7**	116 ± 4.9
Substantia nigra compacta	71 ± 3.4	84 ± 3.0	81 ± 4.7
Substantia nigra reticulata	63 ± 3.5	81 ± 2.8**	76 ± 3.6
Ventral tegmental area	63 ± 3.6	68 ± 2.8	70 ± 4.1
Auditory cortex	138 ± 5.1	156 ± 5.5	135 ± 3.5
Medial geniculate nucleus	116 ± 4.7	128 ± 2.8	126 ± 4.4

^†Data represent rates of local cerebral glucose utilization (μmol/100g/min) expressed as means ± S.E.M.

^*P < 0.05

^**P < 0.01, Bonferroni t-tests for multiple comparisons following two-way ANOVA (treatment group X brain region, with brain region considered a repeated measure).

In another set of rats, the neurochemical effects of acute low dose, chronic low dose, and chronic high dose PHEN treatments were compared with vehicle-treated controls. [³⁵S]GTPγS binding assays were performed by in vitro autoradiography on sections obtained from each animal as previously described (Keegan et al., 2015) in order to quantitate differences in receptor/G-protein coupling between treatment groups (Figure 1). With potential effects of PHEN on brain dopamine, serotonin and norepinephrine levels, receptor activation of G-proteins were determined for D₂, 5-HT_1A, and α₂-adreneregic receptors in adjacent sections from the same animals (Figure 2). These Gi-coupled receptors gave a strong [³⁵S]GTPγS response that could be quantitated in the mesocorticolimbic regions of interest. Acute PHEN treatment had no significant effect on D₂, 5-HT_1A or α₂-adrenergic receptor-stimulated [³⁵S]GTPγS binding in any brain region. However, rats chronically administered either low dose (25 mg/kg/day) or high dose (50 mg/kg/day) PHEN displayed significantly decreased D₂-stimulated binding in the nucleus accumbens compared with vehicle-treated controls. Animals from the chronic high dose treatment group also displayed significant reduction in D₂-stimulated binding in the caudate. Animals administered chronic high dose PHEN exhibited significantly decreased α₂-adrenergic receptor-stimulated [³⁵S]GTPγS binding in the dentate gyrus, hippocampus, and amygdala compared with saline controls. 5-HT_1A receptor activity was not significantly affected by any of the PHEN treatments in any of the brain regions studied.

Figure 1. Dopamine D₂ receptor activation of G-proteins, as measured by [³⁵S]GTPγS autoradiography in rat brain sections, is reduced following chronic 14-day treatment of rats with phenmetrazine (50 mg/kg/day), but not by acute treatment.

Reduction of D₂ agonist (N-propylapomorphine)-stimulated [³⁵S]GTPγS binding after chronic treatment is seen in nucleus accumbens (top sections) and caudate (bottom sections). Abbreviations: N Acc – nucleus accumbens; Caud – caudate.

Figure 2. Effect of acute and chronic phenmetrazine treatments on D₂-dopamine, 5-HT_1A-, and α₂-adrenergic receptor-stimulated [³⁵S]GTPγS binding in coronal rat brain sections.

Slide-mounted tissue sections were assayed for D₂ dopamine-, 5-HT_1A-, and α₂-adrenergic receptor-stimulated [³⁵S]GTPγS binding, and densities from the autoradiograms for each region of interest are presented as mean ± SEM (n = 6-8 per group); * p < 0.05 significantly different from saline (one-way ANOVA, Tukey’s test for multiple comparisons). Abbreviations: N Acc – nucleus accumbens; Caud – caudate; Sep – septum; CC – cingulate cortex; Hippo – hippocampus; DG – dentate gyrus; Cort – cortex; Amyg – amygdala.

Quantitative immunohistochemical determinations on coronal brain sections were performed to detect changes in phosphorylation of key kinases and signaling proteins (Figure 3). PHEN administration produced two distinct patterns of changes in ERK1/2 phosphorylation. In the first pattern, acute PHEN significantly increased levels of pERK1/2 in the ventral tegmental area, and acute PHEN tended (0.05<p<0.10) to increase pERK1/2 in the substantia nigra, thalamus, and hippocampus relative to saline control levels. In those regions, pERK1/2 levels remained significantly elevated after chronic administration of both PHEN doses, and were significantly higher in rats treated with chronic low dose PHEN compared with those treated with chronic high dose PHEN.

Figure 3. Effect of acute and chronic phenmetrazine treatments on ERK1/2 phosphorylation at Thr202/Tyr204 in select rat brain regions.

Fixed coronal sections were immunostained for detection of pERK1/2^{Thr202/Tyr204} and total ERK1/2 and visualized by Li-Cor infrared imaging system. The integrated intensity of pERK1/2 was normalized to total ERK1/2 in each region and data are reported as mean ± SEM (n = 6-8 per group); * p < 0.05 significantly different from saline, ^#p < 0.05 significantly different from chronic high dose (50 mg/kg) phenmetrazine treatment (one-way ANOVA, Tukey’s test for multiple comparisons). Abbreviations: Cort – cortex, VTA – ventral tegmental area; SN – substantia nigra; Amyg – amydala; Caud – caudate; Thal – thalamus; Hippo – hippocampus; DG – dentate gyrus.

In a second pattern, acute PHEN significantly decreased (amygdala) or had no measurable effect on pERK1/2 levels (cortex, caudate, and dentate gyrus), and chronic administration of low dose PHEN produced significant and robust reductions in ERK1/2 phosphorylation (e.g., amygdala was decreased to 65% of saline control). However, reductions were not observed in rats administered chronic high dose PHEN. Further analyses revealed that levels of pERK1/2 in all four of these brain regions were significantly lower in rats treated with chronic low dose PHEN compared to those treated with chronic high dose PHEN.

The effects of each PHEN treatment on DARPP-32^Thr34 phosphorylation are presented in Figure 4. Acute PHEN administration tended to elevate levels of pDARPP-32^Thr34 in all brain regions tested; however, none of these effects achieved statistical significance (0.05<p<0.10). Significantly higher levels of pDARPP-32^Thr34 were detected in the cortex, amygdala, caudate, and hippocampus of rats treated chronically with low dose PHEN compared with saline controls. In contrast, pDARPP-32^Thr34 levels in the caudate and thalamus of rats treated with chronic high dose PHEN were significantly lower than in rats treated with chronic low dose PHEN or control (thalamus).

Figure 4. Effect of acute and chronic phenmetrazine treatments on DARPP-32 phosphorylation at Thr34 in select rat brain regions.

Fixed coronal sections were immunostained for detection of pDARPP-32^Thr34 and total DARPP-32 and visualized by Li-Cor infrared imaging system. The integrated intensity of pDARPP-32 was normalized to total DARPP-32 in each region and data are reported as mean ± SEM (n = 6-8 per group); * p < 0.05 significantly different from saline, ^# p < 0.05 significantly different from chronic high dose (50 mg/kg) phenmetrazine treatment (one-way ANOVA, Tukey’s test for multiple comparisons). Abbreviations: Cort – cortex, VTA – ventral tegmental area; SN – substantia nigra; Amyg – amydala; Caud – caudate; Thal – thalamus; Hippo – hippocampus; DG – dentate gyrus.

In animal models, GSK3β inhibition attenuates the hyperlocomotion induced by either acute cocaine or AMPH administration and interferes with the development of locomotor sensitization after chronic administration of cocaine, AMPH, or methamphetamine (Enman and Unterwald, 2012; Miller et al., 2009; Xu et al., 2011). Based on these reports, we hypothesized that PHEN would decrease GSK3β activity through inhibition of GSK3β auto-phosphorylation at Tyr216 (Lu et al., 2012). Figure 5 shows that neither acute administration of low dose PHEN nor chronic administration of high dose PHEN treatment produced a significant change in GSK3β Tyr216 phosphorylation in any brain region. However, chronic low dose PHEN treatment significantly decreased levels of pGSK3β^Tyr216 in the cortex, ventral tegmental area, caudate, thalamus, hippocampus, dentate gyrus, and bed nucleus of the stria terminalis compared with saline controls. Similar to the pattern that we observed for pERK1/2, the levels of pGSK3β^Tyr216 in all of these brain regions except the BNST were significantly lower in rats treated chronically with low dose PHEN compared with those treated chronically with high dose PHEN.

Figure 5. Effect of acute and chronic phenmetrazine treatments on GSK3β auto-phosphorylation at Tyr216 in select rat brain regions.

Fixed coronal sections were immunostained for detection of pGSK3β^Tyr216 and total GSK3β and visualized by Li-Cor infrared imaging system. The integrated intensity of pGSK3β was normalized to total GSK3β in each region and data are reported as mean ± SEM (n = 6-8 per group); * p < 0.05 significantly different from saline, ^# p < 0.05 significantly different from chronic high dose (50 mg/kg) phenmetrazine treatment (one-way ANOVA, Tukey’s test for multiple comparisons). Abbreviations: PFC – prefrontal cortex; Cort – cortex, VTA – ventral tegmental area; SN – substantia nigra; NAc Core – nucleus accumbens core; NAc Shell – nucleus accumbens shell; Amyg – amydala; Caud – caudate; VP – ventral pallidum; Thal – thalamus; Hippo – hippocampus; DG – dentate gyrus; BNST – bed nucleus of the stria terminalis; Sep – septum.

3. Discussion

Acute injections of phenmetrazine in rats increase motor activity and rearing in an open field (Lindquist and Götestam, 1977) and evoke a cocaine-like drug discrimination response (Bauer et al., 2016). Our objective was to identify key neurochemical changes underlying the behavioral effects of PHEN that depend upon duration of exposure (i.e. acute versus chronic treatment), dose, and/or mechanism of drug action. We identified elevations in brain glucose metabolism in discrete brain areas in response to acute PHEN, but tolerance developed to this response within 14 days of chronic PHEN exposure. This report documents pharmacodynamic changes in both GPCR-stimulated G-protein activation and phosphorylation of ERK1/2, GSK3β^Tyr216, and DARPP-32^Thr34 observed in specific rat brain regions following both acute and chronic PHEN treatments.

3.1. PHEN administration evoked significant regional increases in metabolic activity that are subject to development of tolerance.

Acute administration of PHEN stimulated glucose utilization in a range of brain regions associated with psychostimulant addiction. Interestingly, rats treated chronically with PHEN appeared to have developed tolerance to these effects, as no differences in local cerebral glucose utilization were detected in any brain region of chronically-treated animals compared to saline-treated animals. This finding is reminiscent of results of a study in which rhesus monkeys developed tolerance to phendimetrazine’s ability to reduce cocaine choice over the course of a 14 day treatment (Banks et al., 2013b). These findings may point to a potential limitation of PHEN’s utility as a pharmacotherapy for psychostimulant addiction, although more studies would need to be performed to evaluate the drug’s effectiveness over a more prolonged period of time. These limitations may be overcome by new, recently discovered routes of administration of the drug (Jiang et al., 2019) or newly discovered PHEN analogs, although caution is warranted due to the fact that some of these analogs have exhibited significantly greater abuse potential than the parent compound (Grumann et al., 2019; Mayer et al., 2017; McLaughlin et al., 2018).

3.2. Chronic PHEN treatment desensitizes D₂ dopamine and α₂-adrenergic, but not 5-HT_1A, receptor-mediated G-protein activation

It is common for cells to desensitize surface receptors in response to repeated GPCR stimulation in order to maintain cellular homeostasis. In the past, we have used [³⁵S]GTPγS binding experiments to detect GPCR desensitization following exposure to several different drugs of abuse (Keegan et al., 2015; Maher et al., 2001; Selley et al., 1996; Sim et al., 1995, 1996b). For this study, we performed [³⁵S]GTPγS binding experiments to test the hypothesis that chronic administration of the AMPH-like catecholamine releaser PHEN would produce widespread increases in extracellular monoamine levels and lead to marked desensitization of monoamine receptors. In vitro experiments using rat brain synaptosomes not only characterized PHEN as a more potent monoamine releaser than monoamine uptake inhibitor, but also revealed that PHEN acts as a much more potent releaser of NE (EC₅₀ = 50.4 nM) and DA (EC₅₀ = 131 nM) than serotonin (EC₅₀ = 7765 nM) (Rothman et al., 2002). Given this functional profile, it is not surprising that results reported herein reveal significant, region-specific, and dose-dependent decreases in D₂ dopamine and α₂-adrenergic, but not 5-HT_1A serotonergic receptor-stimulated G-protein activation following chronic administration of PHEN.

It is worth noting that desensitization of nucleus accumbens D₂ receptors occurred after chronic administration of both doses of PHEN but that desensitization of α₂-adrenergic receptors only occurred in rats treated chronically with the higher dose of PHEN. This may be evidence that desensitization of D₂ dopamine and/or α₂-adrenergic receptors requires threshold concentrations of DA and NE and/or threshold residence times of these monoamines in the extracellular space including the synapse be breached. Under this model, our interpretation is that the amount of DA released by the low dose (25 mg/kg/day) of PHEN was sufficient to trigger the mechanisms underlying D₂ receptor desensitization, and that chronic treatment with twice the dose of PHEN (50 mg/kg/day) elicited even more DA to be released, triggering an even greater extent of D₂ desensitization. On the other hand, it would appear that the low PHEN dose was not sufficient to prompt the release of NE in amounts above the threshold needed to induce α₂-adrenergic receptor desensitization. That chronic high dose PHEN did cause α₂-adrenergic receptor desensitization may be an indication that the threshold level of extracellular NE needed for α₂-adrenergic receptor desensitization is achieved by a dose of PHEN between 25 and 50 mg/kg/day. Further studies are needed to confirm this hypothesis, determine threshold concentrations of DA and NE required for desensitization, and find the doses of PHEN needed to elicit each threshold concentration of these monoamines.

Mechanistic differences underlying D₂ dopamine and α₂-adrenergic receptor desensitization in each brain region are unclear; they likely involve differences in expression and activation of GPCR receptor kinases (GRKs), β-arrestins, and other molecules involved in receptor desensitization and/or down-regulation. We must emphasize, however, that results from this study cannot differentiate between true receptor desensitization (i.e., uncoupling between GPCRs and G-proteins) and receptor down-regulation (i.e., decrease in the number of receptors), since a decrease in agonist-stimulated [³⁵S]GTPγS binding would be produced by either mechanism. To differentiate between these different mechanisms is beyond the scope of the current study. Nevertheless, it is important to point out that either desensitization or down-regulation could be responsible for the reductions in receptor-mediated signal transduction observed in this study.

3.3. PHEN induces two distinct patterns of ERK1/2^{Thr202/Tyr204} and DARPP-32^Thr34 phosphorylation

In addition to altering receptor functionality, chronic drug exposure can cause changes in phosphorylation-driven signaling events implicated in a number of long-term neuroplasticity changes associated with addiction (e.g., axonal growth cone dynamics, dendritic spine/actin cytoskeleton remodeling, long-term changes in transcription) (Thomas et al., 2008). During the process of GPCR desensitization, phosphorylation of the GPCR by GRKs allows for binding of β-arrestins that serve as scaffolds for G-protein-independent signaling cascades involving ERK1/2 and GSK3β (Beaulieu et al., 2008; Kohout and Lefkowitz, 2003).

ERK1/2 can become activated/phosphorylated by a wide variety of stimuli and be involved in the regulation of many DA-associated behaviors (Berhow et al., 1996), including several plasticity changes associated with addiction (Girault et al., 2006; Lu et al., 2006). pERK2 has been directly implicated in the induction of conditioned place preference and in the development of psychomotor sensitization associated with chronic cocaine administration (Girault et al., 2006; Lu et al., 2006; Zhai et al., 2007).

Repeated administration of psychostimulants results in increases in cellular phosphorylation of ERK1/2. Both acute and repeated injections of the monoamine releaser AMPH increased ERK1/2 phosphorylation in the rat dorsal striatum, reportedly through a mechanism involving D₁ dopamine receptors and ionotropic glutamate receptors (Choe et al., 2002; Choe and Wang, 2002; Shi and McGinty, 2010; Valjent et al., 2004), and chronic treatment with cocaine enhanced ERK1/2 activity in the rat ventral tegmental area (Berhow et al., 1996; Valjent et al., 2000). A central role for these kinases in mediating cellular responses to psychostimulant drugs is further evidenced by the observation that pharmacological inhibition of ERK1/2 altered several short- and long-term behavioral effects of these drugs (Pan et al., 2011; Valjent et al., 2004, 2000).

We report that PHEN administration induces two distinct patterns of change in ERK1/2 phosphorylation. Chronic low dose PHEN produced significant reductions in pERK1/2 in the cortex, amygdala, dentate gyrus and caudate, where α₂-adrenergic and/or D₂ coupling to G proteins was not desensitized. D₂-like receptors mediate the inhibition of ERK1/2 signaling via the D₂/Ras/MEK pathway (Beaulieu and Gainetdinov, 2011; Zhang et al., 2004), and perhaps the α₂-adrenergic receptor can evoke a similar signaling pathway. One can reasonably speculate that stimulation of undesensitized D₂ and/or α₂-adrenergic receptors was primarily responsible for the decreases in pERK1/2 observed.

DARPP-32, also known as protein phosphatase 1 regulatory subunit 1B (PPR1B), is a cellular inhibitor of protein phosphatase 1 (PP1) that can be activated by phosphorylation at Thr34 by cyclic AMP-activated PKA. PKA-activated pDARPP-32^Thr34 can in turn affect downstream regulation of GSK3β, CREB, and c-Fos proteins via inhibition of PP1. The extent of DARPP-32 phosphorylation is used as a common measure of psychostimulant activity, as this protein is abundantly expressed in mesolimbic dopaminergic cells, has been shown to mediate several effects of cocaine and AMPH (Svenningsson et al., 2005), and is implicated in addiction pathology (Novère et al., 2008).

PHEN appears to produce opposing effects on pERK1/2 and pDARPP-32. In regions where chronic low dose PHEN decreased pERK1/2 (cortex, amygdala, caudate, and dentate gyrus), chronic low dose PHEN significantly increased pDARPP-32 levels. However, in those regions where acute increases in pERK1/2 were maintained after chronic treatments (ventral tegmental area, substantia nigra and thalamus), chronic high dose PHEN treatment tended to decrease pDARPP-32. Region-specific elevations in pDARPP-32^Thr34 were observed in areas that receive large amounts of DA and NE input. Moreover, elevated pDARPP-32 levels were detected in animals treated with chronic low dose PHEN and that showed normal receptor function, but not in those that were treated with high dose PHEN and displayed desensitization. These results might therefore suggest that repeated stimulation of dopaminergic and noradrenergic receptors in these regions potentiates levels of cyclic AMP synthesis and cyclic AMP-stimulated PKA activity. Conversely, decreases in pDARPP-32 were observed in the ventral tegmental area and substantia nigra, two regions in which DA is produced in the brain. These decreases were only observed in animals treated with high dose PHEN displaying marked D₂ dopamine receptor desensitization. One might reasonably infer that while repeated D₂ receptor stimulation appears to have induced desensitization, the consequent adenylyl cyclase, unrestricted by inhibiting G_i proteins, would lead to heterologous desensitization of cyclic AMP production and ultimately attenuated PKA activity. Still, more investigation into the receptor-mediated mechanism of control over these signaling cascades is warranted. Several studies have shed light on a D₁ dopamine receptor/cAMP/PKA pathway that results in the phosphorylation and activation of DARPP-32 (Napolitano et al., 2010; Neve et al., 2004). Activation of a D₁/D₂ dopamine receptor heterodimer in the rat striatum can block the cocaine-mediated activation of a D1R/cAMP/PKA/pDARPP-32^Thr34/pERK pathway (Hasbi et al., 2018). The possible role of D1 dopamine receptors in producing the observed effects warrants further investigation.

3.4. Chronic PHEN reduces GSK3β^Tyr216 phosphorylation

The constitutively active Ser/Thr protein kinase GSK3β has been described as a ‘master regulator’ of cellular processes due to its wide range of substrates, which include metabolic proteins, structural proteins, and transcription factors (Silva et al., 2014). Neuronal GSK3β is regarded as a regulator of tau protein, β-catenin, NMDA glutamate receptors, and circadian clock proteins. Auto-phosphorylation at Tyr216 is associated with maximum activation of GSK3β and permits this kinase to function constitutively (Hughes et al., 1993). The importance of GSK3β auto-phosphorylation is evidenced by the fact that pGSK3β^Tyr216 plays a critical role in important processes such as memory formation through regulation of long-term potentiation (LTP) (Peineau et al., 2007), inhibition of cyclic AMP response element-binding protein (CREB) (Bullock and Habener, 1998; Hansen et al., 2004), and promotion of actin and tubulin assembly during memory formation (Koivisto et al., 2003).

Disruption of GSK3β signaling has been implicated in several DA-associated disorders, including addiction, bipolar disorder, schizophrenia, and attention deficit disorder (Li and Gao, 2010). GSK3β activity is significantly enhanced by cocaine (Miller et al., 2014; Perrine et al., 2008; Xu et al., 2009), AMPH (Enman and Unterwald, 2012), cues associated with these drugs (Shi et al., 2014), as well as DAT knockout and other means of elevating extracellular DA levels (Beaulieu et al., 2011, 2006, 2004). GSK3β also mediates the initiation and expression of sensitization to many behavioral effects of cocaine (Xu et al., 2009) and methamphetamine (Xu et al., 2011). Thus, it is apparent that GSK3β plays an integral role in the cellular neuroadaptations that contribute to addiction.

We report that chronic low dose PHEN significantly decreases pGSK3β^Tyr216 in the caudate, hippocampus or dentate gyrus where D₂ or α₂-adrenergic receptors were not desensitized. We could speculate that PI3K/Akt signaling predominates in rats chronically administered low dose PHEN.

In regions where chronic high dose PHEN promoted marked D₂ and α₂-adrenergic receptor desensitization, receptor coupling to β-arrestin and internalization could promote the activation of the β-arrestin-2/Akt pathway to disinhibit GSK3β and potentiate its activity. Repeated D₂ receptor stimulation promotes β-arrestin-2 recruitment to the plasma membrane for desensitization of the receptor (Li and Gao, 2010). The resulting D₂ receptor/β-arrestin-2 complex causes dephosphorylation of Akt Thr308, thereby inhibiting its kinase activity. Hindrance of Akt allows for disinhibition of GSK3β by removal of the phosphate from Ser9, a reaction catalyzed by the DA-activated protein phosphatase 2A (PP2A). By this mechanism, cocaine and other DA-promoting agents are believed to enhance GSK3β activity (Xu et al., 2009). Further investigations are needed to validate or refute these speculations.

Results from this study provided the strongest evidence that dose is a major determinant of the brain’s response to chronic PHEN. In animals treated acute and chronically with low dose (25 mg/kg/day) PHEN we found significant changes in levels of phosphorylated ERK1/2, GSK3β, and DARPP-32 in several brain regions where we were unable to detect differences in receptor functionality. This indicates that the principal mode of neuronal adaptation to this PHEN dose proceeds primarily through phosphorylation mechanisms. On the other hand, animals chronically treated with high dose (50 mg/kg/day) PHEN displayed profound desensitization of D₂- and α₂-stimulated [³⁵S]GTPγS binding activity but exhibited levels of pERK1/2, pGSK3β, and pDARPP-32 that were generally indistinguishable from vehicle-treated controls. This leads us to believe that neuronal adaptations to chronic administration of this high PHEN dose proceed primarily through mechanisms associated with D₂ and α₂-adrenergic receptor desensitization. It appears that high PHEN-induced desensitization of these receptors allowed for re-equilibration of ERK1/2, GSK3β, and DARPP-32 phosphorylation back to basal levels.

3.5. Conclusions

We and others have conducted a number of preclinical studies which have demonstrated the potential utility of PHEN and the prodrug phendimetrazine as pharmacotherapies for stimulant addiction; specifically, a slow onset, prolonged duration of action, and the ability to prevent cocaine self-administration (Banks et al., 2013a, 2013b, 2012; Czoty et al., 2016, 2015; Negus et al., 2009). Although further studies are needed to confirm and better understand the pharmacodynamic parameters observed, our study provides direct evidence that both dose and duration of PHEN treatment can elicit such different modes of neuronal acclimatization. Future studies should seek to clarify more of the molecular mechanisms underlying these different neuronal responses to PHEN in order to rationally design, dose, and optimize the therapeutic effects of medications to treat psychostimulant addictions.

4. Experimental Procedures

4.1. Animal Treatments

Animal procedure protocols were approved by the Wake Forest School of Medicine Institutional Animal Care and Use Committee, conformed to the principles set forth in the NIH Guide for the Care and Use of Laboratory Animals, and were conducted by the Wake Forest Center for the Neurobiology of Addiction Treatment Tissue Core. All animal experiments were performed in an effort to minimize the number of animals used and the degree of animal suffering. Male Sprague-Dawley rats (280-300 g) (Harlan Industries, Indianapolis, IN, USA) were housed in a temperature- and humidity-controlled vivarium on a 12-hour light/dark cycle (lights on at 7:00 am) and were given unrestricted access to food and water.

For the G-protein activation and cellular signaling studies, rats were randomly assigned to one of four treatment groups: control (n=9), acute low dose PHEN (n=8), chronic low dose PHEN (n=8), or chronic high dose PHEN. Subcutaneous (s.c.) osmotic minipumps (ALZET Model 2001; Durect, Cupertino, CA) were implanted in each animal as previously described (Czoty et al., 2015). Animals in the acute groups were administered either 150 mM NaCl saline (control) or (+)-(+)-phenmetrazine hemifumarate (PHEN) (PAL-56; RTI International, Raleigh, NC) (25 mg/kg/day) s.c. via osmotic minipump for 2 days. It should be noted that the approach to steady state levels over the first 24 hours includes the initiation of release from the minipump, drug uptake and distribution to the tissues and establishment of a pharmacokinetic equilibrium. The chronic treatment groups were infused with saline (control) or PHEN at 25 mg/kg/day for the low-dose group or 50 mg/kg/day for the high-dose group via osmotic minipump for 14 days. The doses stated here are those that were placed in the minipump at the time of implantation. These doses of PHEN had previously been shown to decrease self-administration of cocaine, but not food pellets, in rodents and attenuate increases in cocaine-induced responding during reinstatement (Czoty et al., 2015). Animals were sacrificed via sodium pentobarbital (100 mg/kg, i.v.). Brains were removed, quick-frozen in isopentane at −45 °C, and preserved at −80 °C.

4.2. Local Cerebral Glucose Utilization

Four additional groups of rats (acute and chronic saline, and acute and chronic PHEN) were generated for analysis of local cerebral glucose utilization using the 2-[¹⁴C]-deoxyglucose procedure. Animals were implanted with minipumps to administer either 150 mM NaCl saline or PHEN (50 mg/kg/day) via osmotic minipump for 2 days (acute group) or 14 days (chronic group) as described above.

To prepare for the 2-[¹⁴C]-deoxyglucose procedure, cannulations for the acute PHEN treatment group were carried out at the same time as implantation of the minipump to avoid multiple surgeries on consecutive days. The chronic PHEN treatment group was cannulated on day 13 of the 14-day drug administration period (24 h before the 2-[¹⁴C]-deoxyglucose procedure). Animals were also implanted unilaterally with indwelling catheters in the femoral artery and vein, as previously described (Martin et al., 2019). Briefly, animals were administered a prophylactic dose of ketoprofen (3 mg/kg, s.c.) and anesthetized with isoflurane (2-3%), and catheters, filled with heparinized saline, were inserted into the vessels. Catheters extended subcutaneously from the inguinal area to an exit point between the scapulae where they were coiled and secured with a rodent harness (Instech, Plymouth Meeting, PA) until the 2-[¹⁴C]-deoxyglucose procedure.

On day 2 (acute group) or day 14 (chronic group) following minipump implantation, the 2-[¹⁴C]-deoxyglucose procedure was undertaken according to the method of Sokoloff et al. (Sokoloff et al., 1977), as adapted for use in freely moving animals (Crane and Porrino, 1989). Briefly, animals were administered a pulse of 2-[¹⁴C]-deoxyglucose (75 μCi/kg; specific activity 55 mCi/mmol; Perkin Elmer, Waltham, MA, USA), and timed arterial blood samples were collected over a period of 45 min, after which rats were euthanized with pentobarbital sodium (100 mg/kg). Brains were rapidly removed, flash-frozen in isopentane at −30°C, and stored at −80°C until they were processed for ¹⁴C autoradiography.

For autoradiography, coronal sections (20 μm) were prepared on a cryostat microtome and maintained at −20°C. Sections were flash-dried on a hot plate at 60°C, and apposed to Carestream Min-R 2000 film for 12 days. Autoradiograms were analyzed by quantitative densitometry with a computerized image analysis system (MCID, Imaging Research, Liinton, UK). Tissue ¹⁴C concentrations were determined from densitometric analysis of autoradiograms of the calibrated standards. Rates of glucose utilization were calculated using the optical densities and a calibration curve obtained from local ¹⁴C tissue concentrations, time-courses of the plasma glucose and ¹⁴C concentrations, and the constants according to the operational equation of the 2DG method (Sokoloff et al., 1977). Glucose utilization measurements were determined for 19 discrete brain regions according to the rat brain atlas of Paxinos & Watson (2014). Each brain region was analyzed bilaterally in a minimum of five sections per animal.

4.3. [³⁵S]GTPγS Autoradiography

Frozen brains were sliced as coronal sections (20 μm) using a cryostat microtome maintained at −22 °C and the sections were thawed onto glass slides for [³⁵S]GTPγS autoradiography (Sim et al., 1995). Brain sections were washed with TME (50 mM Tris-HCl, pH 7.4; 3 mM MgCl₂; 0.2 mM EGTA; 100 mM NaCl) for 10 min at 25 °C prior to incubation with TME assay buffer containing 2 mM GDP for 15 min at 25 °C. Sections were then incubated for 2 h at 25 °C in TME assay buffer containing 2 mM GDP and 0.04 nM [³⁵S]GTPγS in the presence or absence of agonists (D₂ dopamine agonist: 10 μM propyl-norapomorphine; 5-HT_1A agonist: 3 μM 8-OH-DPAT; α₂-adrenergic agonist: 10 μM ST-91) (Sim et al., 1996a). The sections were washed twice with 50 mM Tris-HCl, pH 7.4 at 4 °C, rinsed once with deionized water, and were exposed to phosphor-imaging screens overnight. Screen images were captured with a Sony XC-77 video camera, and quantitative densitometric analysis was performed on regions of interest using NIH ImageJ software (National Institute of Health, Bethesda, MD, USA). Regions of interest were defined by user-defined settings in NIH Image software that selected areas of highest optical density. Optical densities were quantitated by comparison with [¹⁴C] brain paste standards and values corrected to nCi/g [³⁵S]. [³⁵S]GTPγS binding data were expressed as percent of net agonist-stimulated binding in sections from saline-treated control rats.

4.4. Immunohistochemistry of Phosphorylated Proteins

Frozen coronal sections (30 μm) were alternate sections from the same brains used above for [³⁵S]GTPγS binding, and were placed flat onto frozen phosphate-buffered formalin (1.5 mM KH₂PO₄, 2.7 mM KCl, 8 mM Na₂HPO₄, 150 mM NaCl; 30% sucrose (w/v); 3% paraformaldehyde (v/v), pH 7.4) in a well of a 24-well plate. Thawed sections were stored at 4°C in fixative. Protein phosphorylation was measured using an In-Cell Western assay as previously described (Blume et al., 2013; Kearn, 2004; Keegan et al., 2015) by which rat brain sections were rinsed six times with Tris-buffered saline (TBS, 20 mM Tris-HCl, pH 7.4; 137 mM NaCl), blocked overnight at 4°C in Blocking Buffer (TBS containing 0.1% IGEPAL and 50% Odyssey® Blocking Buffer (LI-COR Biosciences®, Lincoln, NE, USA)), and incubated with a pair of affinity-purified primary antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA; dilutions 1:250 or 1:500) in Blocking Buffer overnight at 4° C. The primary antibodies used were anti-ERK1/2 (H-72) plus anti-pERK1/2 (Thr202/Tyr204), anti-GSK3β (11B9) plus anti-pGSK3β (Tyr216), or anti-DARPP-32 (N-19) plus anti-pDARPP-32 (Thr34). Tissue sections were then washed four times with TBS containing 0.1% Tween-20 and incubated with secondary antibodies conjugated to one of two IRDye® fluorophores (700 nm and 800 nm) (LI-COR Biosciences®; dilution 1:1500) in Blocking Buffer for 2 h at 25 °C. Sections were washed four times with TBS-0.1% Tween-20, allowed to dry overnight at 4 °C, and visualized using LI-COR Odyssey Infrared Imaging System software (LI-COR Biosciences®). Signal intensity values were quantified using NIH ImageJ software, and phosphoprotein fluorescence intensity was normalized to the total protein fluorescence intensity in each of the indicated brain regions and expressed as relative fluorescence units (RFU).

4.5. Data & Statistical Analysis

Brain regions of interest were selected for analyses from each section according to the stereotaxic atlas of the rat brain (Paxinos and Watson, 2014). Densitometric data from [³⁵S]GTPγS binding were expressed as net agonist-stimulated binding (nCi/g) in sections. Agonist-stimulated G-protein activation is expressed as mean ± standard error of the mean (SEM) of triplicate sections from 6 to 8 animals per group for comparison between treatment groups. The mean ± SEM of normalized RFU values from immunohistochemical phosphorylated proteins (n=8) were expressed as percent of net intensity ratio (RFU value) of saline-treated rat brain regions. Statistically significant differences in agonist-stimulated [³⁵S]GTPγS binding and protein phosphorylation between groups were determined by one-way ANOVA (between-subjects factor: treatment condition) followed by Tukey’s multiple comparisons tests with single pooled variance, and the family-wise significance and confidence threshold (α) was 0.05.

Standard statistics software (SPSS, Chicago, IL) was used for statistical analysis of the 2DG experiments. Rates of cerebral metabolism from the two control groups (for acute and chronic) were combined as no significant statistical differences between the two groups were observed in any brain region. Statistical analysis was accomplished using a two-way ANOVA, brain region group X treatment (control vs. acute vs. chronic) with brain region as a repeated measure. These were followed by planned Bonferroni’s tests for multiple comparisons. Statistical significance was considered as p < 0.05.

Highlights:

• Phenmetrazine (PHEN) is a putative treatment for cocaine addiction

• PHEN acutely stimulated glucose utilization in basal ganglia and motor brain regions

• Chronic PHEN decreased D₂ and α₂-adrenergic receptor-mediated G-protein activation

• PHEN altered the phosphorylation states of ERK1/2, GSK3β, and DARPP-32

• PHEN promoted desensitization and changes in cell signaling and brain metabolism

Abbreviations:

5-HT

serotonin

[³⁵S]GTPγS

guanosine 5’-O-[gamma-thio]triphosphate

Akt

protein kinase B

AMPH

amphetamine

Amyg

amygdala

ANOVA

analysis of variance

BNST

bed nucleus of the stria terminalis

cAMP

cyclic adenosine 3’,5’-monophosphate

Caud

caudate

Cdk5

cyclin-dependent kinase 5

Cere

cerebellum

Cing

cingulate cortex

Cort

cortex (cerebral cortex)

CREB

cAMP response element-binding-protein

DA

dopamine

DARPP-32

dopamine-and cAMP-regulated phosphoprotein-32

DAT

dopamine transporter

Dent Gyrus / DG

dentate gyrus

ERK1/2

extracellular signal regulated kinases 1 and 2

G-protein

guanine nucleotide binding protein

Glob Pal

globus pallidus

GPCR

G-protein-coupled receptor

GSK3β

glycogen synthase kinase 3-β

Hippo

hippocampus

mPFC

medial prefrontal cortex

NAc

nucleus accumbens

NE

norepinephrine

PAG

periaqueductal gray

PBN

parabrachial nucleus

PFC

prefrontal cortex

PHEN

phenmetrazine

PPR1B

protein phosphatase 1 regulatory subunit 1B

Sept

lateral septum

SN

substantia nigra

TBS

Tris-buffered saline

Thal

thalamus

Tris

tris(hydroxymethyl)aminomethane

VP

ventral pallidum

VTA

ventral tegmental area