Abstract
The mesolimbic dopamine system is a crucial component of reward and reinforcement processing, including the psychotropic effects of drugs of abuse such as cocaine. Drugs of abuse can activate intracellular signaling cascades that engender long-term molecular changes to brain reward circuitry, which can promote further drug use. However, gaps remain about how the activity of these signaling pathways, such as ERK1/2 signaling, can affect cocaine-induced neurochemical plasticity and cocaine-associated behaviors specifically within dopaminergic cells. To enable specific modulation of ERK1/2 signaling in dopaminergic neurons of the ventral tegmental area, we utilize a viral construct that Cre dependently expresses Map kinase phosphatase 3 (MKP3) to reduce the activity of ERK1/2, in combination with transgenic rats that express Cre in tyrosine hydroxylase (TH)-positive cells. Following viral transfection, we found an increase in the surface expression of the dopamine transporter (DAT), a protein associated with the regulation of dopamine signaling, dopamine transmission, and cocaine-associated behavior. We found that inactivation of ERK1/2 reduced post-translational phosphorylation of the DAT, attenuated the ability of cocaine to inhibit the DAT, and decreased motivation for cocaine without affecting associative learning as tested by conditioned place preference. Together, these results indicate that ERK1/2 signaling plays a critical role in shaping the dopamine response to cocaine and may provide additional insights into the function of dopaminergic neurons. Further, these findings lay important groundwork toward the assessment of how signaling pathways and their downstream effectors influence dopamine transmission and could ultimately provide therapeutic targets for treating cocaine use disorders.
Significance Statement
Dopamine signaling is critically involved in mediating cocaine-associated behaviors. Here we demonstrate a role for the ERK1/2 signaling pathway and its associated phosphatase, MKP3, specifically in dopamine neurons in regulating dopamine signaling in rats. Furthermore, we demonstrate that this modulation of the ERK1/2 signaling pathway affects cocaine-associated behaviors, including the motivation for cocaine. This work could help identify downstream targets of the ERK1/2 signaling pathway that could be involved in the development of cocaine use disorders.
Introduction
Mesolimbic dopamine (DA) transmission plays a critical role in reward and reinforcement processes, with the activity of ventral tegmental area (VTA) DA neurons linked to the pursuit and consumption of virtually all known drugs of abuse (Koob, 1992; Brodie et al., 1999). In addition, many drugs of abuse have been shown to alter the activity of DA neurons and/or DA transporter (DAT) function, ultimately resulting in disruptions in extracellular DA in terminal regions. Mesolimbic DA function is regulated by a plethora of extrinsic signals, including hypocretins, GABA, glutamate, and many others (Morales and Margolis, 2017; Gantz et al., 2018; Shaw et al., 2019). These diverse signals converge on several key intracellular signaling pathways, which regulate the excitability and firing rate of DA neurons. One such pathway is the extracellular signal-regulated kinase (ERK1/2) signaling pathway (Chen et al., 2001).
It is known that repeated cocaine injections increase ERK1/2 phosphorylation in the VTA (Berhow et al., 1996; Martinez-Rivera et al., 2023) and that ERK1/2 affects several proteins within DA neurons. For example, ERK1/2 phosphorylates the serine (ser) residue 31 (Ser31) of TH, thereby increasing its catalytic activity and resulting in increased DA synthesis (Haycock, 2002). ERK1/2 also increases voltage-dependent calcium channel (VDCC) expression and function (Catterall and Few, 2008; Mortensen, 2013) and regulates DA transporter (DAT) surface expression (Moron et al., 2003; Bolan et al., 2007; Mortensen et al., 2008; Kivell et al., 2014). Furthermore, studies have established that ERK1/2 signaling influences behaviors associated with cocaine, including locomotor sensitization (Marin et al., 2009; Kim et al., 2011), reward (Valjent et al., 2006a; Gangarossa et al., 2019), and consolidation of cue memories associated with drug use (Lu et al., 2005; Miller and Marshall, 2005; Li et al., 2008).
MAP kinase phosphatases (MKPs) are regulators of ERK1/2 activity, and deficits or dysfunction in MKPs have been shown to result in many forms of cellular disruption; including in cancer, development, and CNS function (Seternes et al., 2019). In the CNS, MKP's are involved in depression (Duric et al., 2010; Barthas et al., 2017; Labonté et al., 2017; Lee et al., 2019), ADHD (Demontis et al., 2019), reward processing (Pytka et al., 2020; Qiao et al., 2020), and drug memory formation (Abdul Rahman et al., 2016). MKP3 is one of the most specific for modulating the activity of the ERK1/2 pathway, and it has no other known substrates (Mortensen et al., 2017; Seternes et al., 2019). In addition to its phosphatase activity, MKP3 is also able to anchor ERK1/2 in the cytoplasm, preventing it from shuttling to the nucleus where it normally activates gene expression, further contributing to its strong regulatory influence (Karlsson et al., 2004).
We previously established that MKP3 expression and ERK1/2 signaling regulate several processes of DA transmission through the regulation of two critical proteins in DA neurons (Mortensen et al., 2008; Mortensen, 2013). Overexpression of MKP3 in vitro stabilizes surface expression of DAT (Mortensen et al., 2008) and downregulates L-type voltage-dependent calcium channel Cav1.2 expression (Mortensen, 2013). The decrease in Cav1.2 leads to diminished calcium flux and neurotransmitter release (Mortensen, 2013).
Despite these and other findings, progress in understanding the molecular mechanisms that underlie the reinforcing effects of drugs of abuse in the intact, behaving animal is lacking, as the majority of prior studies have employed ex vivo approaches or pharmacologic interventions that are system-wide and not neuroanatomically or neurochemically specific. To overcome the limitations of prior in vitro studies, we developed a novel adeno-associated virus (AAV) that overexpresses MKP3 in a Cre-dependent manner and thus effectively inactivates ERK1/2 signaling in vivo. We took advantage of transgenic Long–Evans rats with Cre expression limited to tyrosine hydroxylase (TH)-positive cells to achieve selective inactivation of ERK1/2 in DA neurons of the VTA. Using this genetic approach, we examined the influence of MKP3 expression and resulting ERK1/2 inhibition on DAT protein expression and phosphorylation, DA transmission, and cocaine-associated behaviors.
Materials and Methods
Animals
Adult, male, hemizygous Long–Evans-Tg3.1Deis (TH:Cre+; n = 50) transgenic rats and Cre-negative (TH:Cre−; n = 46) wild-type littermates were bred in-house (Witten et al., 2011). Animals were maintained on a 12 h light/dark cycle (with experimental testing occurring during the dark phase) and given ad libitum access to food and water. Animals were pair- or triple-housed prior to receiving infusions of viruses and subsequently single-housed. All protocols and animal procedures were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals under the supervision of the Institutional Animal Care and Use Committee at Drexel University College of Medicine.
Preparation of viral vectors
The AAV9-FLEX-MKP3-R virus is designed to overexpress the MKP3 protein in a Cre-dependent manner (Fig. 1A). MKP3 C-terminally tagged with GFP was cloned in the reverse direction between doubly floxed incompatible loxP and lox2272 sites (FLEX). This results in re-orientation and expression of the MKP3 ORF in the presence of Cre recombinase. The construct is under control of the CAG promoter and was used to produce AAV9 particles. AAV9 virus is likely to transfect only the cells in the area of injection, as AAV9 is not believed to be retrogradely transported (Cearley and Wolfe, 2007). An AAV9 virus with GFP expression was used as control (AAV9-control). This control virus has previously been shown to have no effects on the fast scan cyclic voltammetry (FSCV) and self-administration assays used here (Bernstein et al., 2018). AAV9 serotype viruses were produced in accordance with standard procedure (Penn Vector Core).
Novel construct enables Cre-dependent ERK1/2 inactivation. A, Diagram of virus construct (AAV9-FLEX-MKP3-R) that enables Cre-dependent expression of MKP3. B, Representative multi-fluorescence photomicrograph showing specific expression of MKP3 (green) only in TH-positive cells (red) of the VTA. C, Number of TH-positive and GFP-positive neurons in the VTA. Data points are from separate animals (n = 12). D, Immunoblotting demonstrates successful Cre-dependent inhibition of ERK1/2 activation in HEK293 cells. ERK1/2 activation (phosphoERK1/2) is observed when cells are treated with epidermal growth factor (EGF). MKP3 expression is only observed in the presence of both cre recombinase (Cre) and MKP3 construct (MKP3-R), which results in the inhibition of EGF-induced activation of ERK1/2. E, Immunoblotting demonstrates decreased ERK1/2 phosphorylation/activation (pERK) in the VTA of rats injected with AAV9-FLEX-MKP3-R (MKP3) compared with AAV9-control (CTRL). F, Quantitative analysis of immunoblot band intensities using total ERK1/2 (ERK1/2) for normalization (pERK/ERK). Results are shown as mean ± SEM. Student's t test; *p ≤ 0.05.
Virus infusion
To deliver the AAV9-FLEX-MKP3-R or AAV9-control viruses, animals were anesthetized with isoflurane and placed into a stereotaxic frame, with the skull flat. To target VTA, we drilled a 1 mm hole in the skull (−5.25 mm A/P, +0.95 mm M/L relative to bregma) and lowered a glass injection pipette (Sigma Aldrich) into the VTA, reaching a final depth of 7.8 mm ventral to the brain surface. Glass pipettes were cut to an inner diameter of ∼30 µm. After waiting 10 min to allow tissue to settle, 0.5 µl of AAV9-FLEX-MKP3-R or AAV9-control was delivered into the VTA over 10 min using a picospritzer. Figure 1B shows an example infusion location in the general region of the VTA. This region of the VTA has previously been shown to express high levels of TH+ cells and therefore represents an ideal target for TH-specific manipulation of ERK1/2. Following surgery, rats were housed singly, administered ketoprofen (5 mg/kg, s.c.) at 12 h intervals for 36 h, and then designated for behavioral, neurochemical, or molecular experiments.
Immunohistochemistry
Viral expression was verified by visualization of endogenous GFP fluorescence for all but the Western blot experiments. Rats were anesthetized with isoflurane and transcardially perfused with 140 ml of saline, followed by 150 ml of 10% formalin in 0.01 M phosphate-buffered saline (PBS), pH 7.4. Brains were extracted and fixed in 10% formalin solution for a further 60 min, then placed in a solution of 30% sucrose in PBS overnight, at 4°C. Prior to cryosectioning, a tracking mark was made in the right lateral cortex for identification of hemispheres. Using a microtome, 40 µM coronal sections through the VTA were collected and then stored in a solution of 0.1% sodium azide in 0.01 M PBS, at 4°C. To identify the site of infusion, sections were incubated overnight with primary antibodies for EGFP (600-101-215, 1:10,000, Rockland Immunochemicals) and TH (#AB152, 1:2,000, EMD Millipore) at 4°C, followed by Alexa Fluor 488 donkey anti-goat (A11055, 1:1,000, Life Technologies) and Alexa Fluor 594 donkey anti-rabbit (A21207, 1:1,000, Life Technologies) secondary antibodies, respectively, for 90 min at 20°C. Sections were mounted and coverslipped with fluorescent mounting medium (Vector Laboratories). Cells positive for TH and GFP staining, respectively, were counted and totaled from three separate sections from each animal.
Immunoblotting
Rats received unilateral infusion of either AAV9-FLEX-MKP3-R or AAV9-control into the VTA. After 3 weeks of incubation, rats were killed and the brain was extracted for Western blotting. Immediately following extraction, NAc and VTA tissue were flash frozen in dry ice. Sections were then digested with proteinase K in membrane prep buffer (250 mM sucrose, 1 mM EDTA, 10 mM Tris HCl) and homogenized with beads (Fisher, #15340153) for 40 s. Following 10 min of centrifuging at 4°C and 1,000 rpm, the supernatant was extracted and stabilized with 10% DTT. Following the Pierce BCA protein quantification assay (Thermo Fisher Scientific), samples were accordingly diluted to contain 100 µg protein in 100 µl. A total of 10 µg of protein was then loaded into each well of 4–12% Tris-glycine gels. Samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) for a period of 1 h at ∼200 V. All proteins were transferred to nitrocellulose membranes and then blocked for 1 h at room temperature with blocking medium (Fisher, #PI37515). The membranes were then probed overnight at 4°C with either rabbit DAT primary antibody (1:1,000; Padmanabhan et al., 2008), rabbit DAT-phosphoThr53 primary antibody (1:1,000; p435-53; PhosphoSolutions), with mouse monoclonal Actin primary antibody (1:2,000; #3700, Cell Signaling Technology) used as a control. All incubations with primary antibodies were followed by washing with PBS-Tween followed by incubation for 1 h at room temperature with secondary antibodies (1:10,000). Blots were assayed for fluorescence at the 700 and 800 nm wavelengths, using the LI-COR system. For standardization across blots, each blot contained all experimental groups.
For ERK1/2 immunoblotting, samples were isolated immediately following the post-testing in the conditioned place preference (CPP) assay as described below. Brain samples were processed as above and membranes were probed with either rabbit ERK1/2 (1:1,000; #4695, Cell Signaling Technology) or rabbit phospho-ERK1/2 antibodies (1:1,000; #4370, Cell Signaling Technology) overnight at 4°C.
Striatal slice biotinylation
Rats received unilateral infusion of either AAV9-FLEX-MKP3-R or AAV9-control into the VTA. After 3 weeks of incubation, rats were killed and the brains were removed and immediately chilled in sucrose-supplemented ACSF (SACSF; in mM, 2.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2, 2.4 CaCl2, 26 NaHCO3, 1 ascorbic acid, 11 glucose, and 250 sucrose) and oxygenated with 95%O2/5%CO2. Brains were mounted on Precisionary Instruments Compresstome (model VF-310-0Z), and 300 µm coronal slices were taken. Slices containing the striatum were recovered in ACSF (in mM, 125 NaCl, 2.5 KCl, 1.2 NaH2PO4, 1.2 MgCl2, 2.4 CaCl2, 26 NaHCO3, 1 ascorbic acid, and 11 glucose), 37°C, 40 min with oxygenation from 95%O2/5%CO2, and then were immediately used for trafficking studies. Surface proteins were covalently labeled with 1.0 mg/ml sulfo-NHS-SS-biotin (Thermo Fisher Scientific) in ice-cold ACSF, 45 min, 4°C, with gentle and constant bubbling 95%O2/5%CO2. Following biotinylation, the residual active biotin was quenched; slices were quenched twice with ice-cold ACSF supplemented with 100 mM glycine, 20 min, 4°C, with oxygenation from 95%O2/5%CO2. Then, slices were washed twice with ice-cold ACSF. Tissue was lysed by homogenizing in fresh RIPA buffer with a glass Dounce homogenizer and then transferred to microcentrifuge tubes to fully lyse for 30 min at 4°C. Samples were pelleted by centrifugation at 18,000× g, 20 min, 4°C. Protein concentrations were determined using the BCA protein assay. Biotinylated proteins were isolated by Neutravidin-Agarose Beads, and bound proteins were eluted in SDS-PAGE sample buffer. Total lysate samples were stored in SDS-PAGE sample buffer. Samples were run through SDS-PAGE, and proteins were detected by immunoblotting as above.
Fast scan cyclic voltammetry
Rats received unilateral infusion of either AAV9-FLEX-MKP3-R or AAV9-control into the VTA. After 3 weeks of incubation, rats were anesthetized with isoflurane, implanted with an acute jugular catheter, and placed into a stereotaxic frame. Rats were implanted with a bipolar stimulating electrode (Plastics One) in the VTA (+5.2 mm A/P, +1.1 mm M/L, −7.5 to −8.0 mm D/V), a carbon fiber microelectrode in the core of the NAc (−1.3 mm A/P, +1.3 mm M/L, −6.5 to −7.0 mm D/V), and a reference electrode in the contralateral cortex (−2.5 mm A/P, −2.5 mm M/L, −2.0 mm D/V). Monophasic electrical pulse trains (60 Hz, 4 ms, 600 mA, 60 pulses) in the VTA were used to evoke DA release in the NAc, in accordance with established methods (España et al., 2010, 2011; Prince et al., 2015; Bernstein et al., 2018). During baseline recordings, electrically evoked DA release was elicited every 5 min for a minimum of 30 min or until a stable baseline recording was obtained (DA release within 10% across three consecutive stimulations). After obtaining 15 min of stable baseline recordings, rats received a 2 s, ∼200 µl intravenous cocaine injection (1.5 mg/kg). The 1.5 mg/kg dose of cocaine was chosen to compare with previous FSCV studies using the same dose and paradigm (España et al., 2010, 2011; Prince et al., 2015; Bernstein et al., 2018). Electrically evoked DA release was elicited 5 min post-cocaine delivery and every 5 min thereafter for 60 min. Resulting changes in DA release and uptake were monitored. The electrode potential was linearly scanned (−0.4 to 1.2 V and back to −0.4 V vs Ag/AgCl), and cyclic voltammograms were recorded at the carbon fiber electrode every 100 ms with a scan rate of 400 V/s using a voltammeter/amperometer (Chem-Clamp; Dagan Corporation). Quantification of peak evoked DA release and DA area under the curve was achieved by comparing current at the peak oxidation potential for DA in consecutive voltammograms with calibration factors obtained from electrodes exposed to 3 μM DA. DA overflow curves were fitted to a Michaelis–Menten-based kinetic model using Demon Voltammetry and Analysis software (Yorgason et al., 2011). DA uptake rates prior to cocaine delivery were modeled by setting the affinity of DA for the DAT between 0.16 and 0.20 μM and then fitting the overflow curve to establish a baseline, maximal uptake rate (Vmax) for each subject. Baseline release, area under the curve, and uptake were expressed as the average of three collections that occurred prior to the injection of cocaine. Following cocaine injection, Vmax was held constant for the remainder of the experiment, and changes in DA uptake rate due to cocaine-induced uptake inhibition were calculated as a change in the apparent affinity for the DAT and defined as (Km).
Cocaine conditioned place preference
Rats received bilateral infusion of either AAV9-FLEX-MKP3-R or AAV9-control into the VTA. After 3 weeks of incubation, the rats were taken through the CPP assay. A three-chamber apparatus was used (Med Associates, #MED-CPP-013AT). The apparatus consists of manual guillotine style doors on either side of a neutral gray center compartment with PVC flooring (4.75″ L × 8.25″ W × 8″ H) to allow for access to the remaining two choice compartments (11.75″ L × 8.25″ W × 8″ H). One choice chamber had white walls with wire mesh flooring, and the opposing choice chamber had black walls with grid rod flooring. The apparatus is outfitted with IR photobeam detectors to track animal movement and time spent in each chamber. Rats were habituated in the behavioral room in their home cages for 30 min prior to the start of all sessions. Baseline compartment preference was determined for each rat (n = 19) during a “pre-test,” in which rats had access to all three compartments for 20 min. An unbiased apparatus was used (i.e., compartments were equally preferred by the test subjects during the pre-test). Preference was defined as >20% time spent in one compartment over the other, and animals were paired with the CS+ (cocaine, 10 mg/kg) or CS− (saline, 1 ml/kg) via a biased stimulus-assignment procedure (rats were paired with the CS+ in the compartment they spent less time in on pre-test day). While as a group the rats did not display significant preference for one compartment over the other, we chose this stimulus-assignment procedure in case any specific rat did display subtle preference for one compartment and thus could interfere with the development of place conditioning. Rats were paired with the CS+ or CS− for a total of four pairings each, with sessions taking place twice a day. The CS− was administered during the morning session, and the CS+ was administered 6 h later, during the afternoon session. Pre- and post-testing sessions were performed in the middle of the day. The post-testing session was performed the day following the last pairing session. CPP scores were calculated as the difference in time spent between the CS+ paired chamber on post-test day versus pre-test day. This post–pre difference score allows for the analysis of the within-subject change in preference produced by exposure to the conditioning procedure.
Cocaine self-administration
Rats received bilateral infusions of either the AAV9-FLEX-MKP3-R or AAV9-control viruses into the VTA. After 2 weeks of incubation, rats were anesthetized with a ketamine/xylazine (80/10 mg/kg, i.p.) mixture and implanted with jugular catheters as previously described (España et al., 2010; Brodnik et al., 2015; Prince et al., 2015; Bernstein et al., 2018). Immediately following surgery, rats were placed in operant behavioral chambers, where they were housed for the duration of the experiments. On the third day after surgery, rats were placed on a fixed ratio 1 (FR1) schedule of reinforcement, whereby a single lever press resulted in delivery of 0.75 mg/kg intravenous (i.v.) cocaine (in saline; National Institute on Drug Abuse). Rats were given 6 h daily access to a lever and allowed a maximum of 20 injections per session. Following acquisition of self-administration behavior (defined as 10 injections per session for 3 consecutive days), rats were allowed to self-administer on the FR1 schedule until they reached stable responding (defined as 3 d of 20 injections per session). The latency to obtain the first cocaine injection and rate of intake were monitored during these sessions. Rats were then switched to the progressive ratio (PR) schedule of reinforcement, with 6 h access to a lever, and single cocaine (0.75 mg/kg) injections now contingent upon an increasing number of responses (Hodos, 1961; Richardson and Roberts, 1996). The 0.75 mg/kg dose was selected for self-administration experiments because this dose is on the ascending limb of the “inverted-U” dose–response curve under the PR schedule (Richardson and Roberts, 1996) and to compare with previous observations (España et al., 2010; Brodnik et al., 2015; Prince et al., 2015; Bernstein et al., 2018). Average breakpoint and total lever presses were recorded for 14 d. The timing of catheterization surgeries and behavioral training were selected to ensure that all animals were self-administering at ∼3 weeks post-viral infusion.
Statistical analysis
For analysis of Western blotting, fluorescence intensities were analyzed using independent Student's t tests to compare between virus treatments. Baseline stimulated peak DA release, DA uptake, and DA area under the curve were also analyzed using independent Student's t tests to compare between virus treatments. Stimulated peak DA release, DA area under the curve, and cocaine-induced DA uptake inhibition (Km) across the course of the experiment were assessed using mixed design, repeated-measures ANOVAs, with virus treatment as the between-subjects variable and time as the within-subjects variable. For analysis of CPP, acquisition of cocaine self-administration, and rate of cocaine intake, data were analyzed using independent Student's t tests to compare between virus treatments. For PR data, breakpoints and lever presses were analyzed using a mixed design repeated-measures ANOVA, with virus treatment as the between-subjects variable, and days as the within-subjects variable. Average breakpoints and lever presses over the 7 d PR experiment were assessed using independent Student's t tests. Data were analyzed using SPSS (SPSS) or Prism (GraphPad Software).
Results
Cre-dependent expression of MKP3 results in ERK1/2 inactivation
To limit expression of MKP3 and resulting ERK1/2 inactivation to DA neurons of the VTA, we developed an AAV-based viral vector that overexpresses GFP-tagged MKP3 in a Cre-dependent manner (Fig. 1A). Injection of the virus into the VTA resulted in specific MKP3 expression in TH-positive neurons (Fig. 1B). We found that 73 ± 3% of TH-positive cells also expressed MKP3 (Fig. 1C), and there were no cases where MKP3 expression was observed in TH-negative cells, confirming the high efficiency and specificity of this model system. We also did not observe any expression in substantia nigra. To confirm that the construct resulted in Cre-dependent control of ERK1/2, we transiently transfected HEK293 cells with the individual construct alone or in combination with Cre. The construct showed no observable expression of MKP3 in the absence of Cre but when co-expressed with Cre, the gene was excised and reoriented, resulting in a forward orientation and expression of the protein. Importantly, we also observed that when ERK1/2 signaling is activated by epidermal growth factor (EGF) treatment, MKP3 overexpression attenuates this activation. p38 Map kinase activation is not affected by MKP3 expression (data not shown) demonstrating specificity of MKP3 toward the ERK1/2 Map kinase (Fig. 1D). Finally, we demonstrated that the construct similarly results in inhibition of ERK1/2 activation in vivo in rats. It has previously been demonstrated that ERK1/2 is activated following reward seeking in the CPP assay (Martinez-Rivera et al., 2023). Employing the same CPP assay, we found that MKP3 overexpression significantly decreased (t(10) = 3.04; p ≤ 0.0124) activation of ERK1/2 (Fig. 1E,F).
MKP3 overexpression affects total dopamine transporter and surface dopamine transporter protein levels and phosphorylation
To investigate the effects of MKP3 overexpression and resulting ERK1/2 inactivation on DAT expression and phosphorylation, we injected AAV9-FLEX-MKP3 into one hemisphere and AAV9-control virus into the other hemisphere of the VTA in TH-cre rats (n = 8) resulting in ERK1/2 inactivation in only one hemisphere. Following 3 weeks of incubation, tissue was collected from the NAc, and immunoblotting was performed against DAT, DAT phosphorylation at threonine 53 (pDAT), and actin as control. We observed a significant increase in DAT expression when MKP3 was overexpressed (t(14) = 3.61; p ≤ 0.003; Fig. 2). Overall, pDAT was not increased (t(14) = 1.71; p ≤ 0.109) but when taking into account the increased expression of total DAT, pDAT per DAT taken as a measure of phosphorylation level per transporter was significantly decreased by MKP3 overexpression (t(14) = 4.42; p ≤ 0.0006; Fig. 2A–D).
ERK1/2 inactivation affects DAT expression and phosphorylation. A, Immunoblots from nucleus accumbens tissue obtained from rats injected with AAV9-FLEX-MKP3-R (MKP3) or AAV9-control (CTRL). Quantitative analysis of immunoblot band intensities for (B) total dopamine transporter (tDAT), (C) phosphorylated dopamine transporter at threonine 53 (pDAT), and (D) pDAT/tDAT. E, Surface, total DAT, and actin immunoblots of samples prepared employing striatal slice surface biotinylation from MKP3 or CTRL rats. F, Quantitative analysis of immunoblot band intensities for surface DAT. Arrows indicate bands of interest. All intensities depicted as arbitrary units (A.U.). Results are shown as mean ± SEM. Student's t test; **p ≤ 0.005, ****p ≤ 0.0001 versus CTRL.
Furthermore, we investigated if the increased DAT expression translated into an increase in surface expression of DAT. To examine this, we performed surface biotinylation assays on rat striatal slices. We injected AAV9-FLEX-MKP3 into one hemisphere and AAV9-control virus into the other hemisphere of the VTA in TH-cre rats (n = 8) resulting in ERK1/2 inactivation in only one hemisphere. Following 3 weeks of incubation, striatal slices were prepared and surface biotinylation assays were performed. Similar to what we observed with overall DAT expression, we found a significant increase (t(14) = 4.68; p ≤ 0.0004) in DAT surface expression as a result of MKP3 overexpression (Fig. 2E,F).
MKP3 overexpression affects dopamine transmission
To examine the impact of MKP3 overexpression and resulting ERK1/2 inactivation in DA neurons of the VTA on DA transmission in the NAc, TH-cre rats received unilateral injections of AAV9-FLEX-MKP3 (n = 6) or AAV9-control (n = 7). Following a 3-week incubation, rats were anesthetized with 1.5 g/kg urethane and placed into a stereotaxic apparatus. Recording and stimulating electrodes were lowered into the NAc core and VTA, respectively, until an electrical stimulation produced stable DA release and uptake. Under these baseline conditions, ERK1/2 inactivation did not affect peak amplitude of stimulated DA release (t(11) = 1.15; p ≤ 0.2739) or DA uptake rate (t(11) = 0.721; p ≤ 0.4860). However, ERK1/2 inactivation significantly reduced DA area under the curve (t(11) = 2.708; p ≤ 0.0203), which incorporates both peak release and DA uptake dynamics into one measure (Fig. 3A–D).
ERK1/2 inactivation disrupts DA signaling under baseline conditions and in response to cocaine. A, Example traces of stimulated DA release prior to (BL) and following 1.5 mg/kg intravenous cocaine (Coc) in rats injected with AAV9-FLEX-MKP3-R (MKP3) or AAV9-control (CTRL). B, Peak DA release, (C) DA uptake rate, and (D) DA area under the curve in the NAc under baseline conditions. E, Peak DA release prior to and following 1.5 mg/kg cocaine. F, Average DA peak height after cocaine delivery. G, Inhibition of DA uptake prior to and following 1.5 mg/kg cocaine. H, Average inhibition of DA uptake after cocaine delivery. I, Area under the curve prior to and following 1.5 mg/kg cocaine. J, Average area under the curve after cocaine delivery. E, G, I, Dashed red lines indicate the first collection after cocaine delivery. Results shown as mean ± SEM. D, Student's t tests or (E) Holm–Bonferroni; **p ≤ 0.01, ***p ≤ 0.001 versus CTRL.
Following the 30 min of baseline recording, an intravenous injection of cocaine (1.5 mg/kg) was delivered and resulting alterations in DA dynamics were monitored. Similar to what we observed for the baseline condition, in the presence of cocaine, ERK1/2 inactivation did not significantly affect peak DA release (Fig. 3E,F). However, ERK1/2 inactivation significantly reduced the ability of cocaine to inhibit DA uptake and increase DA area under the curve (Fig. 3G–J). A two-way ANOVA with virus group as a between-subjects variable and cocaine as the repeated-measures variable indicated an effect of virus (F(1,11) = 4.98; p ≤ 0.047), a significant effect of cocaine (F(11,121) = 25.99; p ≤ 0.0001) and a cocaine × virus interaction (F(11,121) = 9.36, p ≤ 0.0001) on DA uptake inhibition (Fig. 3G,H). Similarly, a two-way ANOVA with virus group as a between-subjects variable and cocaine as the repeated-measures variable showed an effect of virus (F(1,11) ≤ 5.81; p ≤ 0.035), a significant effect of cocaine (F(1.8,20.2) = 21.37; p ≤ 0.0001) but no cocaine × virus interaction (F(11,121) = 0.73; p ≤ 0.74) on DA area under the curve (Fig. 3I,J).
MKP3 overexpression in dopamine neurons does not affect conditioned place preference for cocaine
To examine the effect of MKP3 overexpression and resulting ERK1/2 inactivation specifically in VTA DA neurons on CPP for cocaine, rats received AAV9-FLEX-MKP3 (n = 9) or AAV9-control (n = 9) bilaterally into the VTA. We did not observe differences in time spent in the cocaine-associated chamber between MKP3 overexpressing rats and control animals (Fig. 4A). Both MKP3-overexpressing rats and control rats demonstrated place preference for cocaine (10 mg/kg) after eight total conditioning sessions (four sessions/context), indicating that DA neuron-specific ERK1/2 inactivation in the VTA did not influence the acquisition or development of cocaine CPP at this dose (10 mg/kg; t(16) = 0.912; p ≤ 0.928).
ERK1/2 inhibition reduces motivation for cocaine without affecting associative learning. A, CPP preference score, (B) days to acquire FR1 self-administration, (C) FR1 intake rate (i.e., injections/h), (D) breakpoints per PR session, (E) average breakpoints across the 7 d PR experiment, (F) lever presses per PR session, and (G) average lever presses across the 7 d PR experiment for rats injected with AAV9-FLEX-MKP3-R (MKP3) or AAV9-control (CTRL). Results shown as mean ± SEM. D, Main effect of virus or (E) Student's t tests; *p ≤ 0.05 versus CTRL.
MKP3 overexpression in dopamine neurons reduces motivation for cocaine
To examine the effect of MKP3 overexpression and resulting ERK1/2 inactivation specifically in VTA DA neurons on cocaine self-administration, rats received AAV9-FLEX-MKP3 (n = 8) or AAV9-control (n = 7) bilaterally into the VTA. After 2 weeks of incubation, rats were implanted with jugular catheters and placed on an FR1 schedule to acquire cocaine self-administration. Rats with ERK1/2 inactivation did not differ from control rats in the acquisition of cocaine self-administration (t(13) = 1.11; p ≤ 0.289) or the rate of cocaine intake (t(13) = 1.778; p ≤ 0.099; Fig. 4B,C). Rats were then switched to a PR schedule of reinforcement in which single cocaine (0.75 mg/kg) injections were contingent on a progressively increasing number of lever responses. ERK1/2 inactivation reduced motivation for cocaine as indicated by a reduction in breakpoints and lever presses. A two-way ANOVA showed a significant effect of virus (F(1, 13) = 5.434; p ≤ 0.0365), but no effect of session (F(3.34, 43.41) = 0.4081; >p ≤ 0.7684) or virus × session interaction (F(6,78) = 1.119; p ≤ 0.3593) on breakpoints (Fig. 4D). ERK1/2 inactivation also reduced average breakpoints across the 7 d experiment (t(13) = 2.330, p ≤ 0.0366) (Fig. 4E). A two-way ANOVA showed a nonsignificant trend of virus (F(1,13) = 3.735, p ≤ 0.0754), but no effect of session (F(2.57,33.39) = 0.8011; p ≤ 0.4851) or virus × session interaction (F(6,78) = 1.175; p ≤ 0.3281) on lever presses (Fig. 4F). Likewise, ERK1/2 inactivation produced a nonsignificant trend for reduced lever pressing across the 7 d experiment (t(13) = 1.933; p ≤ 0.0754; Fig. 4G).
Discussion
ERK1/2 signaling has previously been demonstrated to play a role in dopamine function and associated behaviors, but most previous studies have lacked anatomical resolution or have been performed in vitro or ex vivo. Here we advance previous studies by providing the first in vivo evidence that ERK1/2 signaling specifically in DA neurons affects both DA transmission and cocaine-associated behaviors in the living animal. We achieve this by overexpressing the negative ERK1/2 regulator and phosphatase MKP3 in dopaminergic cells of the VTA in vivo to manipulate ERK1/2 signaling. By utilizing TH-Cre animals and Cre-dependent MKP3 overexpression and resulting inactivation of ERK1/2 signaling, we identify a critical role for ERK1/2 in regulating DA function, resulting in significant neurochemical and behavioral effects.
Previous studies demonstrated that cocaine induces transient ERK1/2 activation in a specific subset of brain regions, such as the VTA (Berhow et al., 1996; Martinez-Rivera et al., 2023), NAc (Valjent et al., 2004), central amygdala (Lu et al., 2005), and the prefrontal cortex (Valjent et al., 2000). Other studies have established that ERK1/2 influences behaviors associated with cocaine use including locomotor sensitization (Pierce et al., 1999; Valjent et al., 2006b), reward (Valjent et al., 2000), and the consolidation of cue memories associated with drug use (Miller and Marshall, 2005; Valjent et al., 2006a). Of direct relevance to the studies proposed here, reports indicate that ERK1/2 signaling in the VTA plays a role in some of these behaviors (Pan et al., 2011). For example, pretreatment with an upstream MAP kinase inhibitor, PD98059, prevented the development of cocaine-induced psychomotor sensitization when bilaterally injected into the VTA (Pierce et al., 1999), suggesting a role of VTA ERK1/2 signaling in the long-term adaptations associated with sensitization. In addition, systemic pretreatment with a high dose of a MEK inhibitor, SL327 (30 mg/kg), before repeated administration of cocaine prevented the development of locomotor sensitization to cocaine (Valjent et al., 2006b).
Given the importance of DA transmission in the reinforcing effects of cocaine, we used Western blotting and FSCV to examine to what extent MKP3 overexpression and resulting ERK1/2 inactivation impacts the expression and phosphorylation of the DAT protein and DA release and uptake dynamics under baseline conditions and following a cocaine challenge. Previous in vitro studies have demonstrated that ERK1/2 plays a role in the regulation of DAT (Moron et al., 2003; Bolan et al., 2007; Mortensen et al., 2008; Kivell et al., 2014). Although ERK1/2 inactivation caused only nonsignificant decreases in peak DA release and nonsignificant increases in DA uptake rate, these effects resulted in a significant reduction in DA area under the curve following ERK1/2 inactivation. As area under the curve incorporates both peak DA release and rate of uptake, a decrease in this measure suggests an overall reduction in amount and time that DA is available extracellularly, once it is released. The lack of a stronger change in DA uptake following ERK1/2 inactivation was surprising given the robust increase we observed in surface DAT expression in the current studies and a finding that is in line with our prior work indicating that overexpression of MKP3 in vitro stabilizes surface expression of DAT (Mortensen et al., 2008). However, when considered together, our DA transmission and biochemistry results all suggest increased DA uptake.
We next ran a cocaine challenge experiment using FSCV and observed that ERK1/2 inactivation significantly reduced inhibition of DA uptake and DA area under the curve following intravenous cocaine. These effects are consistent with our observation of decreased pDAT and increased surface DAT expression following ERK1/2 inactivation, as recent studies have demonstrated that the phosphorylation state of this particular residue affects the interaction of DAT with cocaine and also affects cocaine-induced locomotor activity (Brodnik et al., 2020; Ragu Varman et al., 2021). Thus, we propose that a reduction in DAT phosphorylation and increased expression of DAT on the membrane surface could be among the molecular mechanisms underlying the observed reduction in inhibition of DA uptake by cocaine.
Disruptions in DA transmission have been repeatedly shown to influence DA-dependent behaviors, and thus the DA changes observed herein are posited to impact behavioral responses to cocaine (Duvauchelle et al., 2000; Quiñones-Jenab et al., 2001; Prus et al., 2009; España et al., 2010; Calipari et al., 2013; Prince et al., 2015; Shaw et al., 2016; Brodnik et al., 2017, 2020; Levy et al., 2017; Siciliano and Jones, 2017; Black et al., 2023). To test this possibility, we examined the effects of ERK1/2 inactivation on CPP for cocaine and cocaine self-administration. We observed that ERK1/2 inactivation did not affect CPP for cocaine, suggesting that functional ERK1/2 signaling in DA neurons is not critical for the learning of contextual cue and/or cocaine cue associations. These findings are in apparent contradiction with previous work suggesting that disruptions in DA transmission impair acquisition and/or expression of cocaine CPP. For example, treatment with a D2 receptor antagonist during cocaine conditioning blocked CPP for cocaine but had no effect on CPP expression when the antagonist was given right before the post-test (Banasikowski et al., 2010). Conversely, blockade of D3 receptors did not affect acquisition of cocaine CPP but did prevent the expression of cocaine CPP (Grakalic et al., 2007; Banasikowski et al., 2010). Although there is evidence that blockade of D1 receptors impairs both acquisition and expression of cocaine CPP (Kramar et al., 2014), other evidence suggests that D1 receptor blockade preferentially blocks acquisition of cocaine CPP but does not affect its expression (Cervo and Samanin, 1995).
Although there are comparatively fewer studies examining the effects of DAT manipulations on CPP for cocaine, findings do not offer a clear relationship between DAT function and expression of CPP. Indeed, engineered mice displaying DAT deficiency show relatively normal CPP for cocaine (Tilley et al., 2007) while mice with DAT overexpression—which may best match our finding of increased DAT surface expression—show enhanced CPP for cocaine (Donovan et al., 1999). Further, in a recent set of experiments, we found that DA uptake rate in the striatum did not correlate with the expression of CPP, though it did predict motivation for cocaine (Shaw et al., 2021).
In addition to a potential disconnect between our DA findings and lack of CPP effects, previous pharmacological findings also suggest a possible discrepancy between our CPP findings and ERK1/2 signaling. One study demonstrated that inactivation of ERK1/2 signaling via upstream MEK inhibitors directly into the NAc blocked acquisition of CPP for cocaine if treatments were given prior to the conditioning phase (Gerdjikov et al., 2004). Another study demonstrated that bilateral injections of an upstream ERK1/2 inhibitor into the VTA decreased the acquisition, but not the expression, of CPP for cocaine (Pan et al., 2011). Lastly, bilateral pretreatment with the MEK inhibitor U0126 directly into the NAc prior to post-testing blocked the expression of cocaine CPP, suggesting that this learned association can be abolished even after associative learning took place (Miller and Marshall, 2005). Together, these CPP observations further highlight the limitations of using pharmacological tools that impact ERK1/2 signaling in heterogeneous populations, and the importance of cell type- and region-specific manipulations to establish molecular mechanisms driving observed in vivo behaviors.
In addition to the CPP studies, we also used self-administration to examine the effects of ERK1/2 inactivation on the reinforcing effects of cocaine. First, using an FR1 schedule of reinforcement, we observed that inactivation of ERK1/2 in DA neurons of the VTA did not affect acquisition of cocaine self-administration and did not affect the rate of cocaine intake. These observations suggest the possibility that ERK1/2 signaling in DA neurons of the VTA is not critical for learning cocaine associations, nor for the reinforcing effects of cocaine under low effort conditions. This observation is generally in agreement with our CPP observation indicating no effect of ERK1/2 inactivation on associative learning. However, ERK1/2 inactivation significantly reduced the motivation for cocaine as indicated by decreased breakpoints and lever presses. Therefore, we conclude that ERK1/2 inactivation specifically influences volitional, effortful responding for cocaine.
The disruptions in motivation for cocaine following ERK1/2 inactivation are consistent with our observed disruptions in DA transmission. Although the exact mechanisms underlying how a reduction in DA responses to cocaine can lead to reduced motivation for cocaine are not entirely clear, there is considerable evidence showing that manipulations that result in disrupted inhibition of DA uptake in response to cocaine typically lead to reductions in motivation for cocaine (España et al., 2010; Calipari et al., 2013; Prince et al., 2015; Shaw et al., 2016, 2021; Brodnik et al., 2017, 2020; Levy et al., 2017; Siciliano and Jones, 2017; Black et al., 2023). Our finding of reduced levels of DAT phosphorylation could suggest this as a central mechanism for the behavioral effects we observe. It should be noted, however, that DAT is not the only protein that regulates DA signaling. For example, the reported involvement of ERK1/2 signaling in cocaine-induced long-term depression of DA neurons ex vivo (Pan et al., 2011) could be another process involved in mediating the behavioral and FSCV results observed herein. Therefore, future studies using a variety of methodologies and focusing on varying proteins of interest will be needed to further disentangle the effects of ERK1/2 signaling on DA release and uptake dynamics and the extent to which changes in DA responses to cocaine influence ongoing behavior.
Finally, it is interesting to note that MKP3 itself is an endogenous physiological regulator of ERK1/2 signaling, and previous studies investigating the role of MKP3 have found that the psychostimulant methamphetamine increases MKP3 mRNA by approximately 50% in the cortex, hippocampus, and striatum just 1 h after acute treatment. Because MKP3 is activated by its interaction with MAP kinases, this increase in MKP3 mRNA after methamphetamine is indicative that psychostimulant exposure activates both MAP kinases and MKP3 (Takaki et al., 2001; Ujike et al., 2002). Based on those earlier studies and our findings that demonstrate that MKP3 expression attenuates behavioral and neurochemical effects of cocaine, it is intriguing to speculate that MKP3 is involved in a compensatory homeostatic mechanism on a behavioral and neurochemical level to overcome the effects of increased DA signaling that result from psychostimulant exposure.
In conclusion the present findings lay important groundwork toward the assessment of how the ERK1/2 signaling pathways and their downstream effectors influence DA neurochemistry and behaviors. As such, the current findings could ultimately provide therapeutic targets for treating cocaine use disorders.
Footnotes
This work was funded by National Institutes of Health grants MH106912 and MH121453 (O.V.M.), DA043787 and DA031900 (R.A.E.), and DA057054 (O.V.M. and R.A.E.) and PA Department of Health CURE grants. We thank Bethan O’Connor for expert technical assistance.
The authors declare no competing financial interests.
- Correspondence should be addressed to Rodrigo A. España at rae39{at}drexel.edu or Ole V. Mortensen at ovm23{at}drexel.edu.
