Abstract
Dysregulation of the input from the prefrontal cortex (PFC) to the nucleus accumbens (NAc) contributes to cue-induced opioid seeking but the heterogeneity in, and regulation of, prelimbic (PL)-PFC to NAc (PL->NAc) neurons that are altered has not been comprehensively explored. Recently, baseline and opiate withdrawal-induced differences in intrinsic excitability of Drd1+ (D1+) versus Drd2+ (D2+) PFC neurons have been demonstrated. Thus, here we investigated physiological adaptations of PL->NAc D1+ versus D2+ neurons after heroin abstinence and cue-induced relapse. Drd1-Cre+ and Drd2-Cre+ transgenic male Long–Evans rats with virally labeled PL->NAc neurons were trained to self-administer heroin followed by 1 week of forced abstinence. Heroin abstinence significantly increased intrinsic excitability in D1+ and D2+ PL->NAc neurons and increased postsynaptic strength selectively in D1+ neurons. These changes were normalized by cue-induced relapse to heroin seeking. Based on protein kinase A (PKA)-dependent changes in the phosphorylation of plasticity-related proteins in the PL cortex during abstinence and cue-induced relapse to cocaine seeking, we assessed whether the electrophysiological changes in D1+ and D2+ PL->NAc neurons during heroin abstinence were regulated by PKA. In heroin-abstinent PL slices, application of the PKA antagonist (R)-adenosine, cyclic 3′,5′-(hydrogenphosphorothioate) triethylammonium (RP-cAMPs) reversed intrinsic excitability in both D1+ and D2+ neurons and postsynaptic strength in only D1+ neurons. Additionally, in vivo bilateral intra-PL infusion of RP-cAMPs after abstinence from heroin inhibited cue-induced relapse to heroin seeking. These data reveal that PKA activity in D1+ and D2+ PL->NAc neurons is not only required for abstinence-induced physiological adaptations but is also required for cue-induced relapse to heroin seeking.
SIGNIFICANCE STATEMENT Neuronal plasticity in the medial prefrontal cortex is thought to underlie relapse to drug seeking, yet the subpopulation of neurons that express this plasticity to functionally guide relapse is unclear. Here we show cell type-specific adaptations in Drd1-expressing versus Drd2-expressing prelimbic pyramidal neurons with efferent projections to nucleus accumbens. These adaptations are bidirectionally regulated during abstinence versus relapse and involve protein kinase A (PKA) activation. Furthermore, we show that disruption of the abstinence-associated adaptations via site-specific PKA inhibition abolishes relapse. These data reveal the promising therapeutic potential of PKA inhibition for preventing relapse to heroin seeking and suggest that cell type-specific pharmacologies that target subpopulations of prefrontal neurons would be ideal for future therapeutic developments.
Introduction
The prefrontal cortex (PFC) is a critical component of the circuitry underlying the transition to opioid addiction (Kalivas and Volkow, 2011). While dysregulation of nucleus accumbens (NAc) neurons that receive a glutamatergic projection from the prelimbic (PL) PFC has been implicated in cue-induced heroin seeking (Shen et al., 2011), changes in the physiological characteristics of medial PFC neurons following exposure to opioids are less well understood. Few studies in the literature have investigated opioid-induced changes in PFC activity. However, two studies have demonstrated that self-administration (SA) of opioids followed by different durations of abstinence altered the intrinsic excitability of PL and anterior cingulate pyramidal neurons (PNs), respectively, in complex ways (Anderson et al., 2021; McKendrick et al., 2022). However, because little is known about opioid-induced physiological adaptations in heterogeneous subpopulations of mPFC neurons, the extant literature is difficult to interpret. An exception derives from a recent demonstration that morphine withdrawal increased intrinsic excitability and decreased HCN channel-mediated currents of thick-tufted (putative Drd2+)-expressing, but not thin-tufted (Drd1+)-expressing, anterior cingulate PNs, indicating different functional responses within these subpopulations of PFC neurons during opiate withdrawal (McDevitt et al., 2021).
Physiologic characterization of phenotypically heterogeneous PFC projection neurons has been limited. Drd1 and Drd2 mRNAs are expressed in neurons distributed in different layers of PL cortex, and both phenotypes project to dorsal striatum (Gaspar et al., 1995), but innervation of ventral striatum (i.e., NAc) has not been reported previously. Furthermore, Drd1-expressing (D1+) and Drd2-expressing (D2+) PNs in the PFC (of mice) show differences in their electrophysiological properties along with differences in their anatomic structure and distribution (Gee et al., 2012; Anastasiades et al., 2019; McDevitt et al., 2021). However, there are no studies that compare responses of Drd1+ (D1+) and Drd1+ (D2+) PL-PFC to NAc (PL->Nac) PNs after heroin SA, abstinence, and cue-induced relapse to heroin seeking. The ability to probe the phenotypic heterogeneity of projection neurons has become possible using viral-mediated approaches and Cre-recombinase driver rodent lines (Gerfen et al., 2013; Garcia-Keller et al., 2020). Thus, we investigated PL->NAc PNs that express Drd1-Cre or Drd2-Cre, focusing on the changes in intrinsic excitability and synaptic plasticity during abstinence from heroin SA that underlie cue-induced relapse to heroin seeking.
Compensatory upregulation of cAMP-dependent protein kinase A (PKA) leading to downstream CREB hyperphosphorylation is one of the most well established molecular neuroadaptations to chronic drug exposure (Nestler and Aghajanian, 1997; Nestler, 2001). Upregulation of PKA-mediated neuroadaptations is considered to be a persistent cellular form of withdrawal and is linked to complex changes in ion channels at multiple synapses (Christie, 2008; Adhikary and Williams, 2022). However, studies implicating changes in cAMP/PKA signaling in the regulation of PL neurons after opioid exposure are lacking. We have previously shown that bidirectional changes in PKA activity occur in PL cortex during cocaine abstinence that regulate relapse. Specifically, by the end of the first week of abstinence from cocaine SA, PKA-dependent GluA1 and CREB hyperphosphorylation emerges (Whitfield et al., 2011; Sun et al., 2014) and intra-PL infusion of a PKA inhibitor, (R)-adenosine, cyclic 3′,5′-(hydrogenphosphorothioate) triethylammonium (RP-cAMPs), reverses the hyperphosphorylation and decreases relapse (Sun et al., 2014). Additionally, changes in PKA activity have been demonstrated to influence the functioning of several voltage-gated ion channels that govern intrinsic excitability (Maurice et al., 2001; Scheuer and Catterall, 2006; Kisilevsky et al., 2008; Iftinca and Zamponi, 2009; Lin et al., 2011; Pathak et al., 2016). Therefore, in this study, we hypothesized that heroin would differentially alter the intrinsic excitability and synaptic plasticity of D1+ versus D2+ PL->NAc core neurons during abstinence from heroin SA and that these changes would be reversed by cue-induced relapse to heroin seeking as well as PKA inhibition.
Materials and Methods
Animals.
Male Drd1-Cre+ (N = 30) and Drd2-Cre+ (N = 25) transgenic rats on a Long–Evans background (LE-TgDrd1a-iCre-3Ottc and LE-TgDrd2-iCre-1Ottc, initially generated as part of the National Institute on Drug Abuse Transgenic Rat Project; Garcia-Keller et al., 2020) were bred in-house at the Medical University of South Carolina and used for all experiments. All rats were housed in a humidity/temperature-controlled colony room with a reverse 12 h light/dark cycle (lights on at 8:00 P.M.). Rats were individually housed 1 week before the beginning of experiments and were food restricted to 20–40 g of rat chow throughout the drug SA procedure. Rats were provided with ad libitum food during home-cage abstinence and before a cue-induced relapse test. All experimental procedures were approved by the Institutional Care and Use Committee of the Medical University of South Carolina and were performed in accordance with the National Institutes of Health guidelines.
Surgeries and intracranial viral infusions.
Rats were injected intraperitoneally with a ketamine (66 mg/kg; VetOne)/xylazine (1.33 mg/kg; Thermo Fisher Scientific) cocktail for initial induction of anesthesia and maintained on 0.5–1% isoflurane (Piramal Pharma) throughout the surgical procedures. A SILASTIC catheter was implanted into the right jugular vein attached to a catheter access button (SAI Infusion Technologies). In a cohort of both Drd1-Cre+ and Drd2-Cre+ rats, the NAc (coordinates from bregma: anteroposterior, +1.6; mediolateral, ±1.6; dorsoventral, −6.5 mm from dura) was bilaterally microinjected with the retrogradely transported AAVrg-hSyn-DIO-EGFP (pAAV-hSyn-DIO-EGFP was a gift from Bryan Roth; viral prep #50 457-AAVrg, Addgene (http://n2t.net/addgene:50457); RRID:Addgene_50457) to visualize D1+ or D2+ PL->NAc neurons. Injections (0.75 µl/hemisphere) were performed over 5 min using a Nanoject II injector (Drummond Scientific) and were left in place for an additional 5 min to allow viral diffusion. In a separate cohort, rats (both Drd1-Cre+ and Drd2-Cre+) were implanted with a 26 gauge bilateral stainless steel guide cannula (Plastics One) aimed at 0.5 mm above the dorsomedial PFC infusion target (coordinates from bregma: anteroposterior, +2.8; mediolateral, ±0.6; dorsoventral, −3.3 mm from skull surface). Guide cannulae were secured to the skull with steel screws (Stoelting) and Jet dental acrylic resin (Lang Dental). A dummy cannula (Plastics One) was immediately placed into the guide cannula to prevent blockage throughout the experiment. Following the surgical procedures, all rats were injected with ketorolac (2 mg/kg, i.p.; Sigma-Aldrich) and cefazolin (200 mg/kg, s.c.; Sandoz) and allowed 4–7 d for recovery. To maintain patency, catheters were flushed daily with taurolidine citrate lock solution (Access Technologies).
Heroin self-administration, abstinence, and postabstinence testing.
The SA drug procedure took place 6 d/week between 9:00 A.M. and 12:00 P.M. Heroin HCl [National Institute on Drug Abuse (NIDA) Drug Supply Program, National Institutes of Health/NIDA) SA was conducted in standard operant chambers equipped with two retractable levers, a cue light, and a tone generator (Med Associates). Each SA session began with the illumination of the house light and extension of the two levers. Rats were trained for 14 d (3 h/d) to lever press for heroin on a fixed ratio 1 reinforcement schedule. Pressing the active lever resulted in delivery of the light and tone followed by an intravenous infusion of either 100 µg/50 µl (days 1–2), 50 µg/50 µl (days 3–4), or 25 µg/50 µl (days 5–14) heroin HCl dissolved in sterile saline (Baxter) followed by a 20 s time-out period. Pressing the active or inactive levers during the time-out period had no programmed consequences. Criteria for acquisition of heroin SA were as follows: >5 infusions/session for days 1–4 and ≥10 infusions/session for days 5–14. Active and inactive lever presses, and the number of infusions received were recorded (Heroin group). Yoked-saline rats were run on a prerecorded program to receive 10 intravenous infusions of sterile saline at random intervals over 3 h. Active and inactive lever presses were also recorded in these rats (Saline group). Following the last SA session, rats underwent home-cage abstinence for 7 d. On the seventh day, a subset of rats underwent a 30 min cue-induced relapse test. These rats were placed in the operant chamber for 30 min and cues accompanied active lever presses, but no drug was delivered. Active and inactive lever presses were recorded (Heroin-Relapse group). Rats that pressed the active lever more than or equal to the number of average active lever presses within the first 30 min of the last 3 d of heroin SA were considered to have relapsed to heroin seeking. The cohort of rats that was implanted with guide cannulae underwent heroin SA, abstinence, and cue-induced relapse test procedures as described above. However, 30 min before the start of a cue-induced relapse test, a subset of these animals was anesthetized with 3% isoflurane (Piramal Pharma) and received bilateral infusions of a competitive antagonist of cAMP-induced activation of PKA, RP-cAMPs [40 nm/0.5 μl/side; Tocris Bioscience; RP-cAMPs treatment (tx) group].
Brain slice preparation for electrophysiology.
On the last day of home-cage abstinence or immediately after a cue-induced relapse test, rats were anesthetized with isoflurane and immediately decapitated. The brains were harvested, and coronal brain sections containing the PFC (300 µm) were sliced with a vibratome (model VT 1200, Leica Biosystems) in an oxygenated (95% O2, 5% CO2; Airgas) ice-cold solution consisting of the following (in mm): 75 sucrose, 87 NaCl, 2.5 KCl, 1.25 Na2HPO4, 25 NaHCO3, 7 MgCl26H2O, 0.1 CaCl2 (Sigma-Aldrich), and 305 ± 5 mOsm. Slices were transferred to a holding chamber containing oxygenated normal artificial CSF (aCSF) containing the following (in mm): 126 NaCl, 2.5 KCl, 1.2 Na2HPO4, 1.2 MgCl26H2O, 2.6 NaHCO3, 0.24 CaCl2, 13.88 glucose (Sigma-Aldrich) and 305 ± 5 mOsm. Brain slices were incubated at 34°C for 30 min and were allowed to recover at room temperature for an additional 45 min. Subsequently slices were kept at room temperature until transferred to the recording chamber.
Whole-cell electrophysiology.
Brain slices were transferred to the recording chamber and continuously perfused with oxygenated aCSF containing picrotoxin (100 μm; Tocris Bioscience) at a flow rate of 2 ml/min. Layer V D1+ or D2+ PL->NAc PNs were visually identified by EGFP expression and their morphologic features using a fluorescent CCD camera (SciCam Pro, Scientifica) according to landmarks illustrated in a rat brain atlas (Paxinos and Watson, 2018). Recording pipettes were constructed from thin-walled borosilicate glass capillaries (inner diameter, 0.58; outer diameter, 1.00; 3–8 MΩ; World Precision Instruments) pulled with a horizontal pipette puller (model P-97, Sutter Instrument). For measuring intrinsic properties and spontaneous EPSCs (sEPSCs) of PL->NAc PNs, recording pipettes were filled with intracellular solution containing (in mm) 135 K-gluconate, 4 KCl, 10 HEPES (Sigma-Aldrich), 4 Mg-ATP, and 0.3 Na3GTP (Tocris Bioscience), at pH 7.35 and 280–285 mOsm. After a stable gigaohm seal was formed, light suction was applied to break through the cell membrane and achieve whole-cell access. Recorded events were acquired with a MultiClamp 700B (Molecular Devices), digitized at a sampling rate of 10 kHz (filtered at 5 kHz) with a Digidata 1550 analog-to-digital converter (Molecular Devices) controlled by pClamp10.7 software (Molecular Devices). The resting membrane potential (Vm) of all neurons was first recorded and adjusted to −70 mV for electrophysiological assessments of intrinsic firing properties under current-clamp mode. Cells with Vm more negative than −58 mV were included for analyses. First, the rheobase [minimum amount of current to elicit a single action potential (AP)] was obtained by applying a series of 50 ms sweeps at +10 pA steps starting at 0 pA until the recorded neuron fired a single AP. Next, AP firing was examined by applying 500 ms depolarization sweeps at +25 pA steps from 0 to 400 pA. These data were acquired at a 1 kHz sampling rate and analyzed offline using Clampfit 10.7 software (Molecular Devices). To examine synaptic changes following heroin abstinence, sEPSCs were recorded in voltage-clamp mode (holding potential, −70 mV) over a 2 min period. Recordings were conducted using a Multiclamp 700B amplifier (Molecular Devices). Currents were low-pass filtered at 5 kHz and digitized at 10 kHz with a Digidata 1550 acquisition board and pClamp10.7 software (Molecular Devices). sEPSC traces were analyzed offline using Clampfit 10.7 software (Molecular Devices). A subset of PL slices from rats abstaining from heroin or saline controls were incubated with 40 nm RP-cAMPs (Tocris Bioscience) for 1 h before neurophysiological recordings (Heroin RP-cAMPs and Saline RP-cAMPs groups).
Electrophysiology data analysis.
AP properties for D1+ or D2+ PL->NAc PNs were derived using the AP trace obtained from rheobase recordings. AP threshold was manually designated from the inflection point at which the AP began, and AP amplitude was measured relative to threshold (Kaczorowski et al., 2012). AP half-width was designated as the width of the AP at half-amplitude between the threshold potential and the peak of the AP (contributed by activity of voltage-gated Na+ channels), whereas AP width (approximation of activity of both voltage-gated Na+ and K+ channels) was defined as the time taken for the membrane potential to return to threshold after firing an AP (Karayannis et al., 2007). Depolarization width was defined as the time taken for the membrane potential to reach peak potential relative to threshold and repolarization width as the time taken to return to threshold from peak potential. The number of APs at each depolarizing current step was counted and plotted to assess excitability of layer V D1+ and D2+ PL->NAc PNs. Another factor affecting the excitability of neurons is afterhyperpolarization (AHP). The amplitude of fast AHP (fAHP) was sampled using the second AP evoked from the current trace that elicited ≥2–4 APs and analyzed by comparing the antipeak of the AP to threshold (Karayannis et al., 2007; Kaczorowski et al., 2012; Otis et al., 2013). Medium AHP (mAHP) and slow AHP (sAHP) events were derived from an average trace generated by averaging current traces with first AP to the last current trace. mAHP amplitude relative to baseline was measured as the maximum antipeak following current offset, and the sAHP amplitude was measured 1 s following current offset of the averaged trace (Kaczorowski et al., 2012).
Perfusion, imaging, and mapping.
For rats that received intracranial viral injections in the NAc, coronal sections containing the NAc (300 µm thick) were sliced immediately following PFC-containing brain slice preparation for electrophysiology and drop fixed in 4% paraformaldehyde (PFA; Sigma-Aldrich) buffered in PBS. NAc sections were mounted and imaged using a Thunder Microscope and Leica imaging software (both Leica Microsystems) for EGFP expression to verify intracranial viral injection sites. Those sections with maximal EGFP expression were mapped using Illustrator software (Adobe). Intracranial injection placements were anatomically verified using a standard rat atlas (Paxinos and Watson, 2018). Rats implanted with cannulae were anesthetized with 30% urethane (Thermo Fisher Scientific) immediately following a cue-induced relapse test. Thereafter, the animals were transcardially exsanguinated using PBS and perfused with 4% PFA in PBS. Then they were decapitated, and the whole brain was extracted and postfixed in 4% PFA in PBS for 24 h followed by incubation in 30% sucrose (Sigma-Aldrich) solution overnight. The 30% sucrose solution was replaced with fresh solution to remove any residual PFA and further incubated for 48–72 h. Coronal sections containing the PFC (50 µm thick) were sliced on a cryostat (model CM 1520, Leica Biosystems), mounted, and visualized using a Leica Thunder microscope and Leica imaging software (Leica Microsystems) to verify placement of the cannula. Sections with track marks from the cannulae were mapped using Illustrator software (Adobe). To qualitatively validate differential distribution of D1+ and D2+ PL->NAc PNs, we injected AAVrg-hSyn-DIO-mCherry (pAAV-hSyn-DIO-mCherry; RRID:Addgene_50459) bilaterally into the NAc of Drd1-Cre+ (N = 2) and Drd2-Cre+ (N = 2) male LE rats, waited 4 weeks, then perfused the rats with 4% buffered paraformaldehyde, and processed for mCherry immunoreactivity and confocal microscopy, as described previously (Siemsen et al., 2019).
Experimental design.
Only singly housed LE transgenic (Drd1-Cre+, N = 23; Drd2-Cre+, N = 18) male rats were used for all experiments because of greater availability than females in the breeding colony. All animals underwent jugular intravenous catheterization followed by bilateral injections of AAVrg-hSyn-DIO-EGFP into NAc to label D1+ or D2+ PL->NAc PNs. Following surgery, the rats were allowed to recover for 4–7 d. To assess the influence of abstinence from heroin on electrophysiological properties of D1+ or D2+ PL->NAc PNs, the rats underwent heroin SA for 14 d while their yoked-saline counterparts received saline infusions. SA data from all rats were collapsed across groups within each genotype and analyzed together. Following the 14 d SA protocol, all animals underwent 6 d of forced home-cage abstinence. On the seventh day of abstinence, the rats were decapitated under isoflurane-induced anesthesia (Heroin and Saline groups). A subset of heroin-abstinent rats (Drd1-Cre+, N = 6; Drd2-Cre+, N = 5) underwent a 30 min cue-induced relapse test to assess the influence of cue-induced relapse to heroin seeking on the neurophysiological properties of D1+ or D2+ PL->NAc PNs. Immediately following the cue test, rats were killed under anesthesia and decapitated (Heroin-Relapse group). Brains were rapidly extracted and 300-µm-thick coronal sections containing the PL cortex were sliced and incubated in aCSF or RP-cAMPs (for 1 h) before recording (Heroin RP-cAMPs and Saline RP-cAMPs groups). EGFP-expressing PNs were identified and patched to record the number of APs fired at increasing levels of injected current, rheobase, and sEPSCs.
To establish the involvement of PKA in cue-induced relapse to heroin seeking, a second experiment was conducted in a separate group of singly housed LE transgenic (Drd1-Cre+, N = 7; or Drd2-Cre+, N = 7) male rats. Following jugular IV catheterization and bilateral intracranial cannula implantation in the PL cortex, rats were allowed to recover for 4–7 d. To assess the influence of intra-PL inhibition of PKA on cue-induced relapse to heroin seeking after forced abstinence, the rats underwent heroin SA for 14 d. SA data from all rats were analyzed together. Following the 14 d SA protocol, all animals underwent 6 d of forced home-cage abstinence. On the seventh day of abstinence, the PL cortex of a subset of rats was infused with RP-cAMPs (under anesthesia) 30 min before the start of a 30 min cue-induced relapse test (RP-cAMPs tx group). Control animals did not receive any infusions and directly underwent a 30 min cue-induced relapse test (Sham group). Immediately following the cue test, rats were exsanguinated and perfused with 4% PFA in PBS under anesthesia. Postperfusion, rats were decapitated, and brains were extracted and collected for verification of cannula placements.
Statistical analyses.
All statistics were conducted using IBM SPSS. SA data of rats that received intracranial viral injections were analyzed using two separate mixed ANOVAs. A four-way mixed ANOVA (within-subjects factors: lever presses – active vs inactive × 14 SA days; between-subjects factors: genotype – Drd1-Cre+/Drd2-Cre+ × drug – saline/heroin) was conducted to determine whether there were significant differences between active versus inactive lever presses across the 14 d of SA between Drd1-Cre+ versus Drd2-Cre+, and heroin-treated versus saline-treated rats. Multiple comparisons were conducted using a Bonferroni correction. Differences in the number of heroin infusions received between Drd1-Cre+ and Drd2-Cre+ were assessed using a three-way mixed ANOVA (within-subjects factors: number of infusions taken across 14 d of SA; between-subjects factors: Genotype – Drd1-Cre+/D2-Cre+ × drug – saline/heroin). To assess cue-induced relapse to heroin seeking, a two-way repeated-measures ANOVA (within-subjects factor: average active lever presses during SA and cue-induced active lever presses × between-subjects factor: genotype – Drd1-Cre+/D2-Cre+) was conducted. To compare baseline differences in AP firing, rheobase, AP properties, and AHP events between D1+ and D2+ PL->NAc PNs, we separately analyzed data from saline-treated rats. A two-way mixed ANOVA (within-subjects factor: injected current, 16 25 pA step increments from 0 to 400 pA; between-subjects factors: genotype, Drd1-Cre+/D2-Cre+) was conducted to analyze differences in AP firing. Baseline differences in rheobase, AP properties, and AHP events were analyzed using t tests. When the test for equality of variances was violated, Welch's t tests were performed. To analyze differences in AP firing of D1+ or D2+ PL->NAc PNs across treatment groups, the following two separate mixed ANOVAs were conducted: (1) three-way mixed ANOVA (within-subjects factor: injected current, 16 25 pA step increments from 0 to 400 pA; between-subjects factors: genotype, Drd1-Cre+/D2-Cre+ × groups, saline/heroin/heroin-relapse); and (2) four-way mixed ANOVA (within-subjects factor: injected current, 16 25 pA step increments from 0 to 400 pA; between-subjects factors: genotype, Drd1-Cre+/D2-Cre+ × groups Saline/Heroin × treatment, aCSF/RP-cAMPs) followed by multiple comparisons using a Bonferroni correction. Differences in rheobase and sEPSC amplitude and frequency were analyzed using two separate factorial ANOVAs: (1) two-way factorial ANOVA (Genotype – Drd1-Cre+/D2-Cre+ × Groups: Saline/Heroin/Heroin-Relapse); and (2) three-way factorial ANOVA (Genotype – Drd1-Cre+/D2-Cre+ × Groups – Heroin/Saline × Treatment: aCSF/RP-cAMPs) followed by multiple comparisons using Bonferroni correction. To assess cue-induced relapse and ex vivo RP-cAMPs induced differences in AP properties and AHP events, we used separate two-way and three-way factorial ANOVAs respectively like those described for analyzing rheobase and sEPSC amplitude and frequency. A one-way repeated-measures ANOVA (within-subjects factors: lever presses – Active vs Inactive × 14 d of SA) was used to determine whether there were significant differences between active versus inactive lever presses across the 14 d of SA in rats that were implanted with cannulae. To assess the influence of RP-cAMPs treatment on cue-induced relapse to heroin seeking, a two-way repeated-measures ANOVA (within-subjects factor: average active lever presses during SA and cue-induced active lever presses × between-subjects factor: Treatment – RP-cAMPs/Sham) was conducted. Multiple comparisons were conducted using a Bonferroni correction. Whenever tests of sphericity were violated, Greenhouse–Geisser-corrected degrees of freedom are reported.
Results
Heterogeneity in anatomic organization and electrophysiological characteristics of D1+ and D2+ PL->NAc neurons
We first assessed baseline physiological characteristics of PL->NAc D1+ and D2+ subpopulations in Drd1-Cre and Drd2-Cre transgenic rats. Observations of mCherry-immunoreactive neurons after AAVrg-hSyn-DIO-mCherry infusion into NAc (Fig. 1A) demonstrated that D1+ PL->NAc PNs were abundantly expressed in layers II-III and Vb, whereas D2+ PL->NAc PNs were predominantly expressed more superficially in layer Va and were scattered in layers II-III (Fig. 1B).
Anatomical organization and physiological differences in intrinsic excitability of Drd1+ and Drd2+ PL->NAc PNs. A, Schematic of AAVrg-hSyn-DIO-mCherry intracranial administration in NAc and its expression in PL cortex. B, Representative images of differential distribution of D1+ (left) and D2+ (right) PL->NAc PNs labeled for mCherry immunoreactivity. C, Schematic of AAVrg-hSyn-DIO-EGFP intracranial administration in NAc and sample image of AAVrg-hSyn-DIO-EGFP viral expression in NAc core (inset). D, Schematic mapping of NAc core coronal sections with viral injection sites (red dots). All viral injection sites were confirmed to be within the NAc core (anteroposterior coordinates from bregma, +1.0 to +2.0 mm). E, Number of APs fired at increasing steps of injected current (0–400 pA with 25 pA step increments) in D1+ versus D2+ PL->NAc PNs of saline controls. D1+ PL->NAc PNs showed significantly greater AP firing at higher steps of injected current than D2+ PL->NAc PNs (*p < 0.05). F, Rheobase (current required to fire a single action potential) did not significantly differ between D1+ versus D2+ PL->NAc PNs. Data expressed as mean +/–SEM.
Next, to assess functional heterogeneity in physiological properties, we compared the number of APs fired, rheobase, AP properties, AHP events, and sEPSC amplitude and frequency between D1+ and D2+ PL->NAc PNs of rats that were injected with AAVrg hSyn-DIO-EGFP in the NAc (between AP coordinates of +1.0 to +2.0 mm from bregma; Fig. 1C,D) and received yoked-saline during the SA procedure. The number of APs fired was slightly greater in D1+ compared with D2+ PL->NAc PNs only at the highest depolarizing current steps (interaction effect between injected current and number of APs fired; F(2.67,85.41) = 2.67, p = 0.059; Fig. 1E). Half-width was significantly greater in D2+ versus D1+ PL->NAc PNs (p < 0.05; Table 1). However, there were no significant differences in Vm, membrane resistance (Rm), AP properties, AHP events, or sEPSC amplitude and frequency between these neuronal subpopulations (Table 1). Despite the apparent difference in the number of APs fired by D1+ versus D2+ PL->NAc PNs in response to high injection currents, no significant difference was found in the rheobase (Fig. 1F). Thus, D1+ PL->NAc PNs show organizational and physiological consistency with thin-tufted PFC neurons that display greater baseline intrinsic excitability than thick-tufted mPFC neurons (Anastasiades et al., 2019; McDevitt et al., 2021), and the latter are consistent with the anatomic organization and excitability of D2+ PL-NAc PNs. Therefore, we corroborated previously identified anatomic heterogeneity within PL->NAc PNs and demonstrated functional differences in intrinsic properties of D1+ versus D2+ PL->NAc PNs.
Baseline differences in membrane properties, intrinsic excitability, and synaptic plasticity
Drd1- and Drd2-Cre+ rats relapsed to heroin seeking following 7 d forced abstinence
All Drd1-Cre+ and Drd2-Cre+ rats that received intracranial viral infusions acquired heroin SA or received yoked-saline followed by 7 d of abstinence. Lever pressing significantly changed across the 14d of heroin SA (significant main effect of lever presses across SA days; F(3.20,118.27) = 3.87, p = 0.010). Specifically, the number of active lever presses significantly increased for both heroin-treated Drd1-Cre+ (F(13,25) = 4.07, p = 0.001) and Drd2-Cre+ (F(13,25) = 3.22, p = 0.006) rats compared with inactive lever presses over time (Fig. 2A,B). There was no significant difference in lever pressing between Drd1-Cre+ or Drd2-Cre+ saline-treated animals. While Drd1-Cre+ rats had a significantly higher number of active lever presses during heroin SA compared with Drd2-Cre+ rats (F(3.20,118.27) = 1.45, p = 0.038), there were no significant differences in the number of heroin infusions received between Drd1-Cre+ or Drd2-Cre+ rats (no significant interaction between number of infusions received, drug and genotype; F(4.19,155.21) = 1.08, p = 0.372; Fig. 2C). A greater number of active lever presses in Drd1-Cre+ versus Drd2-Cre+ rats may be attributed to greater stereotypy exhibited by Drd1-Cre+ rats that led to increased pressing during time-out periods in heroin SA sessions.
Heroin self-administration and relapse in Drd1-Cre+ and Drd2-Cre+ transgenic LE rats. A, The average number of active lever presses in Drd1-Cre+ heroin SA rats (filled red circles, n = 16) was significantly higher than inactive lever presses (filled gray circles), and active lever presses in yoked-saline controls (empty red circles, n = 7; *p < 0.05). B, The average number of active lever presses in Drd2-Cre+ heroin SA rats (filled blue circles, n = 11) was significantly higher than inactive lever presses (filled gray circles), and active lever presses in yoked-saline controls (empty blue circles, n = 7; *p < 0.05). C, The average number of heroin infusions received did not differ significantly between Drd1-Cre+ and Drd2-Cre+ rats. D, Postabstinence cue-induced relapse test: Average number of active lever presses during the first 30 min of the last 3 SA sessions for Drd1-Cre+ (red triangles) and Drd2-Cre+ (blue triangles) rats (left). Average number of active lever presses during a 30 min cue-induced relapse test for the same rats (right). The number of active lever presses significantly increased for both Drd1-Cre+ (**p < 0.01) and Drd2-Cre+ (*p < 0.05) rats during the cue-induced relapse test compared with the average number of active lever presses during first 30 min of last 3 SA sessions. Data expressed as mean +/– SEM.
To assess the influence of cue-induced relapse to heroin seeking on heroin abstinence-induced physiological adaptations in PL->NAc PNs, a subset of heroin SA rats underwent a 30 min cue-induced relapse test after 6 d of home-cage abstinence. Both Drd1-Cre+ or Drd2-Cre+ rats had a significantly greater number of active lever presses during the 30 min cue-induced relapse test than the average active lever presses during the first 30 min of the last 3 d of SA when SA maintenance was well established (F(1,9) = 21.89, p = 0.001; Fig. 2D). Additionally, the number of active lever presses between Drd1-Cre+ and Drd2-Cre+ rats either on the last 3 d of SA or during the cue-induced relapse test did not differ significantly (no significant interaction between active lever presses across genotype: F(1,9) = 0.80, p = 0.395). Together, these data demonstrate that both Drd1-Cre+ and Drd2-Cre+ rats acquired and maintained heroin SA and relapsed to heroin seeking when presented with cues associated with heroin acquisition.
Effects of forced abstinence from heroin and cue-induced relapse to heroin seeking on intrinsic excitability of D1- and D2-expressing PL->NAc PNs
Figure 3A illustrates the experimental design used to identify the effects of heroin abstinence with or without a cue-induced relapse test on intrinsic excitability. D1+ or D2+ PL->NAc PNs expressing EGFP (Fig. 3B,D, respectively) were patched in PL slices. In both D1+ and D2+ PL->NAc PNs, the number of APs fired across increasing depolarized current steps significantly differed between groups (interaction between number of APs fired across increasing levels of depolarized current steps and groups; F(2.38,214) = 105.84, p < 0.001; Fig. 3C,E). Although there was a trend toward a greater number of APs fired in D1+ versus D2+ PL->NAc PNs after heroin abstinence, this trend was not quite significant (interaction between number of APs fired across different amounts of injected current and genotype; F(2.38,214) = 2.74, p = 0.057). Analysis of simple effects revealed that heroin abstinence increased AP firing in both D1+ and D2+ PL->NAc PNs compared with saline-treated controls. Interestingly, AP firing was significantly normalized in D1+ and D2+ PL->NAc PNs after cue-induced relapse to heroin seeking versus abstinence (Fig. 3C,E). Additionally, we saw a decline in AP firing after cue-induced relapse to heroin seeking in both D1+ and D2+ PL->NAc PNs, and in D2+ PL->NAc PNs of saline controls at higher depolarization current steps. Unlike AP firing, the rheobase was differentially affected in D1+ versus D2+ PL->NAc PNs across different groups (significant interaction between rheobase, genotype and groups; F(2,89) = 7.35, p = 0.001). Specifically, the rheobase was significantly decreased in D1+ PL->NAc PNs after heroin abstinence versus yoked-saline controls, and this decrease was reversed by cue-induced relapse to heroin seeking (Fig. 3F, left). In contrast, there was no change in the rheobase of D2+ PL->NAc PNs after heroin abstinence or relapse (Fig. 3F, right).
Changes in intrinsic excitability during abstinence from heroin and cue-induced relapse to heroin seeking. A, Schematic of experimental design. After 7 d of home-cage abstinence, a subset of heroin SA rats (n = 8) and all yoked-saline controls (n = 7) were killed for in vitro whole-cell slice electrophysiology. Another subset of heroin SA rats (n = 6) underwent a 30 min cue-induced relapse test immediately followed by in vitro whole-cell slice electrophysiology. B, Representative image of patched D1+ PL->NAc PN in bright-field (top) and EGFP fluorescence (bottom). C, Left, The number of APs fired in response to increasing steps of injected current (0–400 pA with 25 pA step increments) in D1+ PL->NAc PNs (increased significantly compared with saline controls postabstinence (*p < 0.05) and post-cue-induced relapse test (^p < 0.05, ^^p < 0.01). Right, Representative AP traces at 225 pA injected current (Saline, black; Heroin, red; Heroin + Relapse, orange). D, Representative image of patched D2+ PL->NAc PN in bright-field (top) and EGFP fluorescence (bottom). E, Left, The number of APs fired at increasing steps of injected current (0–400 pA with 25 pA step increments) in D2+ PL->NAc PNs increased significantly compared with saline controls postabstinence (*p < 0.05, **p < 0.01) and post-cue-induced relapse test (^p < 0.05, ^^p < 0.01). Right, Representative AP traces at 225 pA injected current (saline, black; heroin, blue; heroin plus relapse, purple). F, Left, Rheobase of D1+ PL->NAc PNs is decreased significantly during heroin abstinence compared with saline controls (**p < 0.01) and is reversed by cue-induced relapse to heroin seeking. Right, Rheobase in D2+ PL->NAc PNs remained unchanged during heroin abstinence and post-cue-induced relapse test. Data expressed as mean +/– SEM.
Differences in the firing rate of neurons (measure of intrinsic excitability) are influenced by changes in the functioning of ion channels. AP properties (AP amplitude, threshold, half-width, width, depolarization width, and repolarization width) can be used as an approximation of underlying changes in ion channel function that contribute to the generation and duration of an AP. Figure 4A–C illustrates how AP properties and AHP events were measured. Heroin abstinence and cue-induced relapse to heroin seeking differentially affected AP properties in D1+ and D2+ PL->NAc PNs (Fig. 4D). Specifically, in D1+ PL->NAc PNs (Fig. 4E–J, left), AP threshold, amplitude half-width, width, depolarization width, and repolarization width were not affected in heroin-abstinent versus saline-treated rats. However, AP width and repolarization width were significantly increased in the cue-induced relapse versus the heroin-abstinent group (F(14,162) = 2.21, p = 0.009; Fig. 4H,J, left). In D2+ PL->NAc PNs (Fig. 4E–J, right), AP threshold and amplitude remained unaffected, whereas AP half-width, width, depolarization, and repolarization width were significantly reduced in heroin-abstinent versus saline-treated rats. In these neurons, cue-induced relapse reversed the abstinence-induced decrease in these AP properties (F(14,162) = 3.73, p < 0.001). AHP events (fast, medium, or slow) in D1+ or D2+ PL->NAc PNs after heroin abstinence or cue-induced relapse to heroin-seeking did not significantly differ from saline controls (Fig. 4K–M).
Abstinence from heroin and cue-induced relapse to heroin seeking changed AP waveform measurements. A, Schematic of rheobase AP trace demonstrating how AP properties were derived. B, Schematic of fAHP amplitude determination. C, Schematic of average trace used to compute mAHP and sAHP amplitude. D, Overlaid representative traces of rheobase measurements used to derive AP waveform measurements. Rheobase trace for Saline (black), Heroin (red), and Heroin-Relapse (orange) in D1+ PL->NAc PNs (top) and rheobase trace for Saline (black), Heroin (blue), and Heroin-Relapse (blue) in D2+ PL->NAc PNs (bottom). E, F, Changes in AP properties after heroin abstinence and cue-induced relapse. Abstinence from heroin and cue-induced relapse to heroin seeking did not alter AP threshold (E) and amplitude (AP amp) (F). G, AP half-width was not affected by heroin or cue-induced relapse in D1+ PL->NAc PNs (left), but it was significantly decreased by heroin (*p < 0.05) and reversed by cue-induced relapse (^p < 0.05) in D2+ PL->NAc PNs (right). H, In D1+ PL->NAc PNs (left), AP width was not affected significantly by heroin, but cue-induced relapse significantly increased (^p < 0.05) it compared with heroin. In D2+ PL->NAc PNs (right), AP width was significantly decreased by heroin (*p < 0.05) and cue-induced relapse normalized it (^^p < 0.01). I, AP depolarization width was unaffected by heroin or cue-induced relapse in D1+ PL->NAc PNs (left). AP depolarization width was not affected significantly by heroin in D2+ PL->NAc PNs (right) but cue-induced relapse significantly increased (^^p < 0.01). J, AP repolarization width was not affected significantly by heroin in D1+ PL->NAc PNs (left) but cue-induced relapse significantly increased it (^^p < 0.01). In D2+ PL->NAc PNs (right), AP repolarization width was significantly decreased by heroin (*p < 0.05) and normalized by cue-induced relapse (^^p < 0.01). K, Abstinence from heroin-induced and cue-induced relapse to heroin seeking did not alter fAHP in both D1+ and D2+ PL->NAc PNs. L, Cue-induced relapse to heroin seeking marginally decreased mAHP in D1+ PL->NAc PNs (left) compared with heroin (p = 0.052). In D2+ PL->NAc PNs (right), mAHP was marginally decreased after cue-induced relapse to heroin seeking compared with saline controls (p = 0.051), but significantly decreased compared with heroin (**p < 0.01). M, Abstinence from heroin marginally increased sAHP in D2+ PL->NAc PNs (right) compared with D1+ PL->NAc PNs (left; p = 0.054). However, abstinence from heroin and cue-induced relapse did not affect sAHP in either D1+ or D2+ PL->NAc PNs. Data expressed as mean +/– SEM.
These data demonstrate that heroin abstinence increased firing of both D1+ and D2+ PL->NAc PNs that is reversed by cue-induced relapse to heroin seeking, but the amount of current necessary to fire one AP was reduced after abstinence and restored by relapse only in D1+ PL->NAc PNs. Additionally, differential effects of heroin abstinence and cue-induced relapse on AP properties between D1+ versus D2+ PL->NAc PNs suggest potential differences in the molecular mechanisms underlying changes in the firing of these neuronal subtypes.
Inhibiting cAMP-dependent PKA mimicked the effects of cue-induced relapse to heroin seeking on intrinsic excitability of D1+ and D2+ PL->NAc PNs
To assess the involvement of PKA signaling on the changes in intrinsic excitability of D1+ and D2+ PL->NAc PNs during abstinence, we incubated a set of PL slices obtained from rats after 7 d of heroin abstinence and yoked-saline controls with 40 nM of competitive antagonist of cAMP-dependent PKA activity, RP-cAMPs (Fig. 5A). In both D1+ and D2+ PL->NAc PNs, incubation with RP-cAMPs significantly changed the number of APs fired in heroin-treated or saline-treated rats (significant interaction between the number of APs fired by D1+ and D2+ PL->NAc PNs of heroin-treated or saline-treated rats; aCSF vs RP-cAMPs treatment: F(2.40,223.17) = 7.83, p < 0.001). Specifically, RP-cAMPs treatment significantly decreased the abstinence-induced increase in the number of APs in heroin-abstinent D1+ and D2+ PL->NAc PNs. Additionally, there was no significant effect of RP-cAMPs on the number of APs fired by either D1+ or D2+ PL->NAc PNs of saline-treated control rats, demonstrating that RP-cAMPs treatment alone does not influence neuronal firing (Fig. 5B,C, respectively). RP-cAMPs treatment did not significantly alter the rheobase of D1+ or D2+ PL->NAc PNs in heroin-treated versus saline-treated rats (interaction among genotype, drug, and RP-cAMPs: F(1,98) = 3.01, p = 0.086; Fig. 5D). It is important to note that RP-cAMPs treatment did not affect the rheobase of either D1+ or D2+ PL->NAc PNs of saline-treated rats, reiterating that RP-cAMPs alone does not affect the intrinsic properties of these neuronal subpopulations (Fig. 5D). However, while AP threshold and amplitude remained unaffected (Fig. 6A,B), AP half-width, AP width, and, AP depolarization and repolarization widths were significantly increased in both D1+ and D2+ PL->NAc PNs after RP-cAMPs treatment (significant interaction between drug and RP-cAMPs treatments; Heroin vs Heroin-RP: AP half-width: F(1,85) = 10.23, p = 0.002; AP width: F(1,85) = 10.67, p = 0.002; AP depolarization width: F(1,85) = 8.45, p = 0.005; AP repolarization width: F(1,85) = 9.15, p = 0.003; Fig. 6C–F), associating increased PKA activity during heroin abstinence with enhanced functioning of ion channels. Interestingly, AP half-width, AP width, and AP depolarization and repolarization widths in D1+, but not D2+, PL->NAc PNs after RP-cAMPs treatment was significantly greater in heroin-abstinent rats than in saline control rats (Fig. 6C–F), suggesting that heroin enhanced the effects of PKA inhibition on D1+ PL->NAc PNs. This effect was not seen in D2+ PL->NAc PNs (F(7,79) = 0.78, p = 0.609). RP-cAMPs did not alter AHP events of D1+ and D2+ PL->NAc PNs (Fig. 6G–I).
Effects of in vitro PKA inhibition on intrinsic excitability of D1+ and D2+ PL->NAc PNs after heroin abstinence. A, Schematic of experimental design. After 6 d of home-cage abstinence, both heroin and yoked-saline SA rats underwent in vitro whole-cell slice electrophysiology with or without in vitro application of Rp-cAMPs. B, Heroin abstinence induced increased firing of D1+ PL->NAc PNs (*p < 0.05; left) was reduced to baseline/control levels by in vitro application of RP-cAMPs (^p < 0.05, ^^p < 0.01). The number of APs fired was unaffected by RP-cAMPs application in saline-treated rats. Example traces of AP firing at 225 pA of D1+ PL->NAc PNs (Saline, black; Heroin, red; Saline-RP-cAMPs, gray; Heroin-RP-cAMPs, pink; right). C, Heroin abstinence-induced increased firing of D2+ PL->NAc PNs (*p < 0.05, **p < 0.01; left) was reduced to baseline/control levels by in vitro application of RP-cAMPs (^p < 0.05, ^^p < 0.01). The number of APs fired was unaffected by RP-cAMPs application in saline-treated rats. Example traces of AP firing at 225 pA of D2+ PL->NAc PNs (Saline, black; Heroin, blue; Saline-RP-cAMPs, gray; Heroin-RP-cAMPs, green; right). D, Rheobase of D1+ PL->NAc PNs (left) is decreased significantly during heroin abstinence compared with controls (*p < 0.05) but RP-cAMPs treatment did not significantly reverse this effect. Rheobase in D2+ PL->NAc PNs (right) remained unchanged during abstinence and was unaffected by in vitro RP-cAMPs application. Data expressed as mean +/– SEM.
Effects of in vitro PKA inhibition on AP waveform measurements of D1+ and D2+ PL->NAc PNs after heroin abstinence. A, B, Abstinence from heroin or in vitro RP-cAMPs treatment did not affect AP threshold (A) or amplitude (AP amp) (B) of D1+ and D2+ PL->NAc PNs. C, In D1+ PL->NAc PNs (left), heroin abstinence did not affect AP half-width, but RP-cAMPs significantly increased it in heroin-treated rats (^^p < 0.01), while in D2+ PL->NAc PNs (right), heroin abstinence significantly decreased AP half-width (*p < 0.05) and RP-cAMPs restored it to baseline levels in heroin-treated rats (^^p < 0.01). D, In D1+ PL->NAc PNs (left), heroin abstinence did not significantly decrease AP width, but RP-cAMPs restored it to baseline levels in heroin-treated rats (^p < 0.05). In D2+ PL->NAc PNs (right), heroin decreased AP width (**p < 0.01) and RP-cAMPs restored it to baseline levels (^p < 0.05). E, In D1+ PL->NAc PNs (left), heroin abstinence did not significantly decrease AP depolarization width, but RP-cAMPs restored it to baseline levels in heroin-treated rats (^^p < 0.01). In D2+ PL->NAc PNs (right), heroin significantly decreased AP depolarization width (**p < 0.01) and RP-cAMPs restored it to baseline levels in heroin-treated rats (^^p < 0.01). F, In D1+ PL->NAc PNs (left), heroin abstinence did not significantly decrease AP repolarization width, but RP-cAMPs restored it to baseline levels in heroin-treated rats (^^p < 0.01) while in D2+ PL->NAc PNs (right), heroin significantly decreased AP depolarization width (**p < 0.01) and RP-cAMPs restored it to baseline levels in heroin-treated rats (^p < 0.05). G, fAHP remained unaffected in D1+ PL->NAc PNs (left). In D2+ PL->NAc PNs (right), fAHP was significantly reduced in heroin-RP-cAMPs group compared with Saline-RP-cAMPs (*p < 0.05). H, RP-cAMPs treatment significantly reduced mAHP in D1+ PL->NAc PNs (left) of heroin-treated rats (*p < 0.05) and saline controls treated with RP-cAMPs (*p < 0.05) but did not affect mAHP in D2+ PL->NAc PNs (right). I, sAHP remained unaffected by heroin abstinence and RP-cAMPs treatment in both D1+ and D2+ PL->NAc PNs. Data expressed as mean +/– SEM.
These data demonstrate that the activation of PKA is required for heroin abstinence-induced increase in intrinsic properties of PL->NAc PNs and that PKA differentially modulates physiological responses of D1+ versus D2+ PL->NAc PNs. Additionally, the overlap in the physiological effects of PKA inhibition and those induced by cue-induced relapse on these neurons suggests that PKA inhibition underlies cue-induced relapse to heroin seeking.
Changes in synaptic plasticity of D1- and D2-expressing PL->NAc PNs after cue-induced relapse and in vitro RP-cAMPs treatment
Changes in amplitude and frequency of sEPSCs were used as indicators of altered synaptic plasticity after heroin abstinence and relapse. Abstinence from heroin-induced differential effects on sEPSC amplitude and frequency in D1+ and D2+ PL->NAc PNs. sEPSC amplitude significantly differed between D1+ versus D2+ PL->NAc PNs across different groups (significant interaction between genotype and groups for sEPSC amplitude; F(2,90) = 18.42, p < 0.001). Specifically, abstinence significantly increased sEPSC amplitude in D1+, but not in D2+, PL->NAc PNs compared with saline controls (Fig. 7A–D). Interestingly, this increase in D1+ PNs was reversed by cue-induced relapse to heroin seeking (Fig. 7A,B). In contrast, sEPSC frequency did not differ by genotype or group (interaction between genotype and groups for sEPSC frequency; F(2, 90) = 1.46, p = 0.537; Fig. 7A–D).
Effects of heroin abstinence, cue-induced relapse, and in vitro RP-cAMPs treatment on sEPSC amplitude and frequency (freq) of D1+ and D2+ PL->NAc PNs. A, Abstinence from heroin significantly increased the AP amp of sEPSCs in D1+ PL->NAc PNs (left) compared with saline controls (**p < 0.01) and was normalized by cue-induced relapse (^^p < 0.01); sEPSC frequency of D1+ PL->NAc PNs (right) was unaffected by heroin abstinence or cue-induced relapse. B, Example traces of sEPSCs of D1+ PL->NAc PNs. C, sEPSC amplitude (left) and frequency (right) of D2+ PL->NAc PNs remained unaffected by heroin abstinence or cue-induced relapse. D, Example traces of sEPSCs of D2+ PL-> PNs. E, Abstinence from heroin significantly increased the amplitude of sEPSCs in D1+ PL->NAc PNs (left) compared with saline controls (**p < 0.01) and was normalized by in vitro RP-cAMPs treatment (^^p < 0.01), whereas sEPSC frequency of D1+ PL->NAc PNs (right) was unaffected by heroin abstinence and in vitro RP-cAMPs treatment. F, Example traces of sEPSCs of D1+ PL->NAc PNs. G, In D2+ PL->NAc PNs, both sEPSC amplitude (right) and frequency (right) remained unaffected by heroin abstinence and in vitro RP-cAMPs treatment. H, Example traces of sEPSCs of D2+ PL-> PNs. Data expressed as mean +/– SEM.
Changes in amplitude and frequency of sEPSCs were used as indicators of altered synaptic plasticity after heroin abstinence and PKA inhibition. Based on the data described above and our previous observations of intrinsic excitability measures, we predicted that RP-cAMPs-induced inhibition of PKA activity would reverse the changes in amplitude of sEPSCs seen in D1+ PL->NAc PNs during heroin abstinence. Indeed, RP-cAMPs treatment significantly affected sEPSC amplitude (interaction between RP-cAMPs treatment on sEPSCs of D1+ or D2+ PL->NAc PNs across heroin-treated or saline-treated rats; F(2,98) = 4.26, p = 0.017). Specifically, RP-cAMPs significantly decreased the abstinence-induced increase in sEPSC amplitude of D1+ PL->NAc PNs (Heroin vs Heroin-RP; F(2,98) = 18.73, p < 0.001; Fig. 7E,F). sEPSC amplitude remained unaffected by RP-cAMPs treatment in D1+ PL->NAc PNs of saline-treated controls (D1+ PL-NAc PNs: Saline vs Saline-RP; F(2,98) = 1.340, p = 0.267; Fig. 7E,F). RP-cAMPs had no effect on sEPSC amplitude in D2+ PL-NAc PNs of heroin-abstinent animals (Fig. 7G,H). In addition, RP-cAMPs had no effect on the frequency of sEPSCs in either D1+ or D2+ PL->NAc neurons (Fig. 7E–H).
These data demonstrates that PKA regulates synaptic plasticity of D1+ PL->NAc neurons. They also indicate differential influence of PKA on D1+ versus D2+ PL->NAc PNs during heroin abstinence, such that PKA influences both intrinsic and synaptic properties of D1+ PL->NAc PNs but only modulates intrinsic properties of D2+ PL->NAc PNs during heroin abstinence, suggesting that multiple protein regulators of these properties are PKA dependent. Additionally, they further corroborate PKA inhibition as an underlying mechanism of the effects that cue-induced relapse to heroin seeking has on synaptic properties of D1+ and D2+ PL->NAc PNs.
An intra-PL microinfusion of RP-cAMPs inhibits cue-induced relapse to heroin seeking
To determine the functional relevance of PKA inhibition on cue-induced relapse to heroin seeking, a cohort of Drd1-Cre+ and Drd2-Cre+ rats underwent heroin SA and 7 d abstinence followed by an intra-PL infusion of RP-cAMPs 30 min before a 30 min cue-induced relapse test. The experimental design is illustrated in Figure 8A. Of the 14 rats included in this group, cannula placements in three rats were inaccurate and excluded from further analyses. Intracranial cannula placements in the remaining rats were anatomically verified to lie between AP coordinates of +3.2 and +4.2 mm from bregma (Fig. 8B) using a standard rat atlas (Paxinos and Watson, 2018). All remaining rats acquired heroin SA (Fig. 8C). Lever pressing significantly changed across the 14 d of heroin SA (significant main effect of lever presses across SA days; F(1,10) = 55.80, p < 0.001). The number of active lever presses significantly increased compared with inactive lever presses over time (significant interaction between lever presses and SA days; F(3.71,37.08) = 3.36, p = 0.022). Since there were no significant differences in lever presses across the 14 d of SA between Drd1-Cre+ and Drd2-Cre+ rats (as shown earlier), SA data for this cohort were analyzed without splitting animals based on their genotype.
Cue-induced relapse to heroin seeking is blocked by intra-PL administration of RP-cAMPs. A, Schematic of the experimental design. B, Schematic mapping of PL cortex coronal sections with cannula traces (black vertical lines). All viral injection sites were confirmed to be within the prelimbic cortex (anteroposterior coordinates from bregma, +3.0 to +4.2 mm). C, The average number of active lever presses and heroin infusions in heroin SA rats. The average number of active lever presses was significantly higher than the number of inactive lever presses (*p < 0.05). D, Postabstinence cue-induced relapse test: average number of active lever presses during the first 30 min of the last 3 SA sessions for sham and RP-cAMPs-pretreated rats (left). Average number of active lever presses during a 30 min cue induced the relapse test for the same rats (right). Sham rats significantly increased their active lever presses during cue-induced relapse test compared with their average active lever presses in the first 30 min of the last 3 SA sessions (*p < 0.05), while the RP-cAMPs-treated rats did not. Average active lever presses in sham rats were significantly greater than RP-cAMPs-treated rats (^p < 0.05) during the cue-induced relapse test. Data expressed as mean +/– SEM.
A subset of Drd1-Cre+ and Drd2-Cre+ rats (N = 6) received a bilateral intracranial infusion of RP-cAMPs 30 min before the start of the cue-induced relapse test. Overall, the average number of active lever presses was significantly different during the first 30 min of the last 3 d of SA and the cue-induced relapse test (significant main effect of average active lever presses; F(1,9) = 7.72, p = 0.021; Fig. 8D). The average number of active lever presses during heroin SA also significantly differed between sham-treated and RP-cAMPs-treated groups (significant interaction effect of average active lever presses during heroin SA and cue-induced relapse test between groups; F(1,9) = 6.37, p = 0.033). In sham control rats, but not in Rp-cAMPs-infused rats, the average number of active lever presses was significantly greater during the cue-induced relapse test than during the first 30 min of the last 3 d of SA (p = 0.006; Fig. 8D). The average number of active lever presses during the cue-induced relapse tests of rats that received RP-cAMPs was significantly lower than in rats that did not receive the RP-cAMPs infusion (p = 0.020; Fig. 8D). Together, these data demonstrate that intra-PL administration of RP-cAMPs inhibited cue-induced relapse to heroin seeking after abstinence. Thus, PKA signaling in PL was not only required for heroin abstinence-induced physiological adaptations in PL->NAc neuronal subpopulations but was also required for cue-induced relapse to heroin seeking.
Discussion
This is the first study of the effects of heroin abstinence on D1+ and D2+ PNs in the PFC that project to the NAc and the involvement of PKA signaling in physiological adaptions of these neurons during heroin abstinence and cue-induced relapse to heroin seeking. These data demonstrate that abstinence from heroin increased the intrinsic excitability and postsynaptic strength of D1+ PL->NAc PNs, but only increased the firing of D2+ PL->NAc PNs. Further, reversal of these physiological adaptations by cue-induced relapse to heroin seeking reinforces the involvement of such adaptations as an underlying feature of heroin abstinence and cue-induced relapse. Additionally, these data demonstrate the importance of PL PKA signaling in heroin abstinence-induced physiological adaptations and solidified its role in cue-induced relapse to heroin seeking. Finally, our study extends and adds to important evidence that heterogeneous subpopulations of PL neurons differentially regulate specific addiction-related behaviors.
Differential D1+ and D2+ PL->NAc intrinsic excitability and AP waveform dynamics
Previously, subtypes of layer V PNs have been characterized separately by their apical dendritic morphology (thin tufted or thick tufted), projection targets, and electrophysiological signatures (Morishima and Kawaguchi, 2006; Hattox and Nelson, 2007). More recently, thick-tufted PNs in the rodent PFC have been shown to express Drd2 and thin-tufted PNs express Drd1 (Gee et al., 2012; McDevitt et al., 2021). Gee et al. (2012) demonstrated that thick-tufted D2+ PFC neurons in mice have a prominent h-current and project to subcortical targets. Specifically, they had a novel quinpirole-induced afterdepolarization characterized by a larger voltage sag and rebound afterdepolarization, whereas the PNs that lacked the afterdepolarization were thin-tufted and expressed Drd1 (Gee et al., 2012). We identified differences in intrinsic excitability of D1+ and D2+ PL->NAc PNs under control conditions. Specifically, there were an increased number of APs fired at the highest depolarizing currents and decreased AP half-width, but no differences in other AP properties, in D1+ versus D2+ PL->NAc PNs (Table 1). Decreased half-width in D1+ PL->NAc PNs may lead to faster membrane depolarization and therefore activation of voltage-gated sodium channels that remain resistant to depolarization-induced inhibition of AP firing and in turn contribute to sustained firing of D1+ PL->NAc PNs at high depolarization currents.
D1+ and D2+ PL->NAc intrinsic excitability and AP waveform dynamics were differentially altered by heroin abstinence and reversed by relapse to heroin seeking
Changes in intrinsic excitability in NAc MSNs have emerged as important neuroadaptations underlying the transition to addiction (Ishikawa et al., 2009; Kourrich et al., 2015). However, less is known about the contributions of PFC PN physiological adaptations. Recently, McDevitt et al. (2021) demonstrated that spontaneous opiate withdrawal induced subpopulation-specific alterations in layer V PNs in the ACC of mice. Specifically, morphine withdrawal increased intrinsic excitability, decreased HCN-channel-mediated currents of D2+ ACC PNs but did not affect D1+ ACC PNs (McDevitt et al., 2021). Similarly, PL PNs in male mice showed reduced rheobase and a trend toward increased firing frequency following 10–16 d of remifentanil self-administration and 14–21 d of abstinence (Anderson et al., 2021). We show that heroin abstinence enhanced the AP firing of both D1+ and D2+ PL->NAc PNs and decreased rheobase in D1+, but not in D2+, PL->NAc PNs, indicating that less current was necessary to fire APs in heroin-abstinent D1+ PL->NAc PNs. Further, using several different types of waveform measurements (Karayannis et al., 2007; Kaczorowski et al., 2012; Kourrich et al., 2015), we found that most waveform measurements of D2+ PNs were significantly altered by heroin abstinence. In contrast, only AP width and AP repolarization of D1+ PNs were altered by heroin. Together, these findings suggest that intrinsic excitability is a key molecular neuroadaptation of PL->NAc PNs induced by opiate exposure that differentially affects D1+ versus D2+ PL->NAc PNs. Further, increased intrinsic excitability and altered AP waveform dynamics were reversed by cue-induced relapse to heroin seeking in both D1+ and D2+ PL->NAc PNs, indicating that relapse reverses the abstinence-induced neuroadaptations in D1+ and D2+ PL->NAc PNs that contribute to drug seeking. A potential limitation of this finding is that significant differences in AP firing were only observed during 200–400 pA of depolarization current steps when AP firing in both D1+ and D2+ PL->NAc PNs of saline controls or heroin relapse rats had plateaued or even declined. The reduced number of APs in these groups may be because of depolarization-induced inactivation of sodium channels. Future studies will determine whether adaptations in Na+ channel dynamics in these neurons at high depolarization current steps is overcome by heroin abstinence.
PKA inhibition reverses the effects of abstinence on D1+ and D2+ PL->NAc PNs
The finding that inhibition of PKA activity reversed the enhancement of intrinsic excitability in both D1+ and D2+ PL->NAc PNs of heroin-abstinent rats indicates elevated PL PKA activity during abstinence. PKA and PKC are the predominant kinases that mediate GPCR regulation of voltage-gated Na+, Ca2+, and K+ ion channels (Schiffmann et al., 1995; Maurice et al., 2001; Chen et al., 2006; Kim et al., 2006; Scheuer and Catterall, 2006; Iftinca and Zamponi, 2009; Lin et al., 2011; Pathak et al., 2016). What PKA-regulated channels underlie the heroin-induced increase in intrinsic excitability in D1+ versus D2+ PL->NAc PNs are major questions to be addressed in future experiments. In fact, because RP-cAMPs was introduced into the slice bath, several GPCRs and ion channels in the local prefrontal network of heroin-abstinent rats may have been normalized by dephosphorylation. Enhancement of depolarization and repolarization occurred in D2+ PL->NAc PNs, whereas only repolarization enhancement predominated in D1+ PL->NAc PNs. Thus, in D2+ PL->NAc PNs, Na+ and K+ ion channels may be hyperphosphorylated by PKA during abstinence whereas phosphorylation of only K+ channels is implicated in D1+ PL->NAc PNs. A prominent PKA-regulated mediator of AP repolarization in PL->NAc PNs is the Kv4 channel subtype (Schrader et al., 2002; Pathak et al., 2016). Thus, one possibility is that PKA activity is increased after 1 week of abstinence, causing endocytosis of Kv4 channels and consequent increase in intrinsic excitability during heroin abstinence that is reversed by PKA inhibition. Further, the effects of RP-cAMPs mimicked those seen after cue-induced relapse to heroin seeking, suggesting that relapse reduced PKA activity causing dephosphorylation and exocytosis of Kv4 channels to the PN cell surface and a normalization of AP firing in both D1+ and D2+ PL->NAc PNs. Future studies will determine whether the changes in AP firing of D1+ and D2+ PL->NAc PNs after relapse are affected by Kv4 inhibitors (Pathak et al., 2016).
A unique feature of the D1+ PNs was that the AP waveform properties responded to RP-cAMPs with significantly greater magnitude after heroin abstinence than that for RP-cAMPs-treated D1+ PL->NAc PNs in saline controls, indicating that PKA phosphorylation of ion channels that regulate waveform dynamics in D1+ PL->NAc PNs has a more profound effect than in D2+ PL->NAc PNs in rats with a heroin history.
Differential effects of heroin abstinence on D1+ and D2+ PL->NAc synaptic plasticity were reversed by relapse or PKA inhibition
Previous research has demonstrated that synaptic plasticity, mediated by changes in trafficking and membrane surface expression of AMPA receptors in key brain areas associated with substance use disorders, is an important mechanism underlying the transition to addiction (Shen et al., 2011; Gipson et al., 2014; Hearing et al., 2018). In this study, it is likely that heroin abstinence and cue-induced relapse selectively and reciprocally altered D1+ PL->NAc PN sEPSC amplitude by altering AMPA receptor phosphorylation and surface expression in these neurons. Re-exposure of rats to audiovisual cues previously associated with heroin SA, whether after home-cage abstinence or extinction training, has been demonstrated to cause downregulation of the AMPA receptor subunit, GluR2, in synaptic membranes of PFC neurons and a decrease in the AMPA/NMDA ratio in PFC slices recorded immediately after a 30 min cue test (Van den Oever et al., 2008). In that study, blockade of GluR2 endocytosis by an in vivo injection of a GluR2-Tat protein attenuated the rectification index and cue-induced relapse to heroin seeking. However, these experiments were conducted in animals that underwent extinction training and a potential abstinence-induced or cell type-specific difference in excitability was not investigated. In contrast, we observed a selective increase in sEPSC amplitude in D1+ PL->NAc PNs during abstinence that was reversed by relapse or PKA inhibition, suggesting that phosphorylation-induced surface expression of AMPA receptors (Esteban et al., 2003; Sun et al., 2005; Oh et al., 2006; Man et al., 2007) may be involved. In future studies, we will investigate NMDAR and AMPAR expression in D1+ PL->NAc PNs and consequent NMDAR-mediated and/or AMPAR-mediated synaptic plasticity using AMPA/NMDA ratio measurements.
In conclusion, increased intrinsic excitability in both D1+ and D2+ PL->NAc PNs and increased sEPSC amplitude in D1+ PL->NAc PNs, suggests that abstinence increased glutamatergic tone in these neurons. Additionally, we demonstrated that cue-induced relapse to heroin seeking or PKA inhibition reverses these changes. Last, intra-PL PKA inhibition after abstinence inhibited cue-induced relapse to heroin seeking, demonstrating that PKA-dependent electrophysiological adaptations during abstinence in D1+ and D2+ PL->NAc PNs is an underlying molecular mechanism of cue-induced relapse to heroin seeking that has potential therapeutic implications.
Footnotes
This work was supported by National Institutes of Health Grants P50-DA-15369 and T32-DA-07288. We thank Dr. Elizabeth Doncheck and Lisa Green for technical training and assistance. We thank Dr. Matthew Hearing for reviewing preliminary electrophysiology data and providing valuable suggestions. In addition, we thank Ayteria Crow and Caroline Limbaker (Center for Opioid and Cocaine Addiction) for technical assistance.
The authors declare no competing financial interests.
- Correspondence should be addressed to Jacqueline F. McGinty at mcginty{at}musc.edu
