Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 May 15.
Published in final edited form as: Biol Psychiatry. 2021 Jan 27;89(10):959–969. doi: 10.1016/j.biopsych.2021.01.008

The PI3-kinase p110β isoform controls severity of cocaine-induced sequelae and alters the striatal transcriptome

Lauren P Shapiro 1,2,#, Elizabeth G Pitts 2,3,#, Dan C Li 2,3,#, Britton R Barbee 1,2, Elizabeth A Hinton 2,3, Gary J Bassell 3,5, Christina Gross 4, Shannon L Gourley 2,3,6
PMCID: PMC8202243  NIHMSID: NIHMS1687766  PMID: 33773752

Abstract

Background:

The PI3-kinase (PI3K) complex is a well-validated target for mitigating cocaine-elicited sequelae, but pan-PI3-kinase inhibitors are not viable long-term treatment options. The PI3K complex is composed of p110 catalytic and regulatory subunits, which can be individually manipulated for therapeutic purposes. This possibility has largely not, however, been explored in behavioral contexts.

Methods:

Here we inhibited PI3K p110β in the medial prefrontal cortex (mPFC) of cocaine-exposed mice. Behavioral models for studying relapse, sensitization, and decision-making biases were paired with protein quantification, RNA-sequencing, and cell-type-specific chemogenetic manipulation and RNA quantification to determine whether and how inhibiting PI3K p110β confers resilience to cocaine.

Results:

Viral-mediated PI3K p110β silencing reduced cue-induced reinstatement of cocaine seeking by half, blocked locomotor sensitization, and restored mPFC synaptic marker content following cocaine. Cocaine blocked the ability of mice to select actions based on their consequences, and p110β inhibition restored this ability. Silencing dopamine D2 receptor-expressing excitatory mPFC neurons mimicked cocaine, impairing goal-seeking behavior, and again, p110β inhibition restored goal-oriented action. We verified p110β presence in mPFC neurons projecting to the dorsal striatum and orbitofrontal cortex and found that inhibiting p110β in the mPFC altered the expression of functionally-defined gene clusters within the dorsal striatum and not orbitofrontal cortex.

Conclusions:

Subunit-selective PI3K silencing potently mitigates drug seeking, sensitization, and decision-making biases following cocaine. We suggest that inhibiting PI3K p110β provides neuroprotection against cocaine by triggering coordinated cortico-striatal adaptations.

Keywords: prelimbic, reward, addiction, incentive, action-outcome, goal, psychostimulant

Introduction

Drug misuse is a major public health concern. Despite decline in global cocaine manufacture in the early 2000s, production rose in recent years, reaching record highs in 2017, the last year data were available (1). Cocaine-related deaths more than doubled between 1999-2017 in the U.S., an increase only partially attributable to the opioid crisis (2). Nevertheless, FDA-approved pharmacotherapies for cocaine use disorder do not exist.

PI3-kinase (PI3K) is a signaling complex that phosphorylates phosphoinositides, regulating receptor trafficking and function at the membrane and translating cell surface signals into the phosphorylation of Protein kinase B (also termed Akt) and the serine-threonine protein kinase mammalian target of rapamycin (mTor). Cocaine stimulates PI3K, Akt, and mTor signaling in the prefrontal cortex (in humans, see ref.3; in rodents, summarized in ref.4). These phenomena are durable, detectable despite drug washout (4–9) and may thus be associated with long-term consequences of cocaine: craving, relapse, and altered decision making. In experimental models, PI3K/Akt/mTor inhibitors block the expression of cocaine-induced locomotor sensitization, conditioned place preference, and drug-seeking behaviors after they have formed (reviewed 4,10); however, these drugs are not viable treatment approaches in humans due to broad-spread effects that would make them intolerable for long-term use, and further, they would be expected to interfere with cell survival, differentiation, and division.

The PI3K complex is composed of p85 regulatory and p110 catalytic subunits determining the intracellular positioning and activation properties of PI3K (11–13). p110β is highly expressed in neurons, augmenting the capacity for PI3K activation (12,14–16). Meanwhile, inhibiting p110β interferes with PI3K signaling (14). PI3K subunits, including p110β, can be genetically and pharmacologically manipulated individually, a strategy long used in cancer research. Yet, behavioral neuroscience-focused investigations still tend to treat the PI3K complex as a monolithic protein target, rather than the sum of dissociable parts. Here we silenced the neuronally-expressed p110β subunit in the medial prefrontal cortex (mPFC) following cocaine, reducing cocaine-seeking behavior and virtually eliminating sensitization and drug-induced decision-making abnormalities. Further, prefrontal cortical p110β silencing altered the striatal transcriptome. Understanding the discrete functions of different PI3K subunits could reveal novel, targeted strategies for treating disease.

Methods and Materials

Subjects.

Mice were wildtype C57BL/6 mice from Jackson Labs or transgenic mice expressing Tg(Drd2-cre)ER44Gsat and their Cre-null littermates (C57BL/6 background). Here, Cre-recombinase is introduced upstream of the ATG start codon of Drd2 (17). Both sexes were used. Procedures were approved by the Emory University IACUC.

Viral vectors.

Lentiviruses expressing mCherry and an shRNA against Pik3cb or a scrambled construct under the CMV promoter were generated by the Emory University Viral Vector Core. They were validated in mouse mPFC, where they reduce p110β protein content by half and dampen phosphorylated mTor (18). Note that lentiviruses preferentially transfect excitatory neurons and largely spare inhibitory neurons; modest glial transfection would also be expected (19).

Cre-activated Designer Receptor Exclusively Activated by Designer Drugs (DIO-DREADDs) (AAV5-hSyn-DIO-hM4D(Gi)-mCherry; AddGene; ref.20) were infused into both Drd2-Cre+ and Cre− littermates. In this case, Cre− mice infused with DIO-DREADDs served as the control group because DREADDS would not be expressed in the absence of Cre.

For affinity purification experiments, a Cre-dependent fluorescently-tagged ribosomal subunit (AAV5-FLEX-EGFP-L10a) was infused in the mPFC of naïve mice, while a retrogradely transported virus encoding Cre-recombinase (AAVrg-hSyn-Cre-WPRE-hGH) was infused in either dorsomedial striatum (DMS) or orbitofrontal cortex (OFC) (AddGene). Thus, immunoprecipitation of the EGFP-L10a fusion protein from tissue collected from the mPFC allows for the isolation of translating mRNA transcripts from a projection-selective sub-population of mPFC neurons.

Procedures.

Details regarding immunoblotting from bulk mPFC tissue, surgery, cocaine and food self-administration, locomotor quantification, drug administration, virally-expressed translating ribosome affinity purification (vTRAP), and RNA-sequencing are provided in the Supplementary Methods.

Statistics.

Response rates, locomotor counts, protein densities, and gene expression (qPCR) values were compared by ANOVA or 2-tailed t-test, as appropriate, using SPSS or SigmaPlot. Post-hoc comparisons were applied following interactions or main effects between >2 groups. p<0.05 was considered significant, and significant comparisons are reported graphically.

In one experiment, we also converted instrumental response patterns to preference scores (response rates in the non-degraded/degraded conditions). Group means were compared by 1-sample t-test to 1, reflecting no preference.

Analysis of the transcriptomic data is described in the Supplementary Methods.

Results

To begin, we quantified p110β in the mPFC and lateral PFC of intact, drug-naïve mice. PI3K p110β was detected at stable levels throughout postnatal development (no main effects or interactions, Fs<1)(fig.1a). The same pattern has been reported in humans for PIK3CB, which encodes p110β (21).

Fig. 1. p110β inhibition mitigates cocaine seeking.

Fig. 1.

Open in a new tab

(a) p110β is detectable in the mPFC and lateral PFC throughout postnatal development. Representative blots loaded in the same order are below with an HSP-70 loading control. (b) sh-Pik3cb-expressing viral vectors were infused into the mPFC and largely contained within the PL. Black and white traces on images from The Mouse Brain Library (63) convey representative large and small viral vector spread, respectively. (c) Experimental timeline. (d) Mice were first trained to self-administer cocaine, then designated to viral vector groups by matching response rates. (e) sh-Pik3cb reduced cocaine seeking in a test for cue-induced reinstatement. Means + SEMs, *p<0.05, n=7 control, 5 sh-Pik3cb. Behavioral experiments were conducted in 2 independent cohorts, and western blots were conducted at least twice.

Knockout of Pik3cb is lethal (22), but viral-mediated gene silencing allows for targeted reduction of PI3K p110β. We packaged an shRNA against Pik3cb into a lentivirus (18) and delivered it into the mPFC. Infusions were largely contained within the prelimbic PFC (PL; fig.1b), a key node in the brain’s “reward circuitry” overwhelmingly implicated in reinstatement behavior (23,24).

First, we implanted indwelling jugular catheters and trained naïve mice to self-administer i.v.-delivered cocaine (fig.1c). Mice were designated “to be control” vs. “to be sh-Pik3cb” by matching response rates during the acquisition (fig.1d) and extinction phases (fig.1e). Accordingly, we detected no main effects of group or group interactions (Fs<1). Main effects of session and nose poke and an interaction between the two indicated that mice responded more as training proceeded, and they responded more on the “active” nose poke, which triggered cocaine, than the “inactive” nose poke, which did not [main effect of session F(9,90)=8.22, p=0.02; main effect of nose poke F(1,90)=25.01, p=0.001; session × nose poke F(1,90)=31.16, p<0.001]. Responding was then extinguished, with the final extinction session in fig.1e.

Throughout, we reduced Pik3cb after mice had been exposed to cocaine, as would occur when treating problematic cocaine use in humans. Then, mice were re-exposed to discrete cocaine-associated cues daily for 3 days. sh-Pik3cb reduced response rates by roughly half [main effect of sh-Pik3cb F(1,10)=5.2, p=0.046; no effect of day F(2,20)=1.65, p=0.22; no interactions F<1](fig.1e), considerably mitigating the reinstatement of cocaine seeking.

We next assessed locomotor sensitization to cocaine, first using a within-subjects design. We treated new mice with cocaine (10 mg/kg daily, 5 days), then introduced the viral vector (fig.2a). Mice were later habituated to a locomotor monitoring chamber, then administered saline, and finally, cocaine (10 mg/kg). Mice locomoted more following cocaine than saline, but importantly, this locomotor response was greatly reduced by Pik3cb knockdown [cocaine x sh-Pik3cb F(17,221)=1.9, p=0.02; main effect of cocaine F(17,221)=21.1, p<0.001; no main effect of sh-Pik3cb F<1](fig.2b).

Fig. 2. p110β inhibition mitigates cocaine-induced behavioral sensitization.

Fig. 2.

Open in a new tab

(a) Experimental timeline: Mice were repeatedly injected with cocaine, then given a washout period, then viral vector infusion and ultimately, a cocaine “challenge.” (b) sh-Pik3cb mitigated the locomotor response to cocaine. n=9 control, 6 sh-Pik3cb. (c) In separate experiments, inclusion of mice that were drug-naïve upon cocaine challenge revealed that sh-Pik3cb fully blocked the expression of sensitization. n=12-15/group. (d) We next assessed phospho-mTor, downstream of PI3K p110β, in one cohort of these mice. We used a tissue punch, represented by the red circle, of frozen brains infused with viral vectors, represented by a histological trace on a coronal section from The Mouse Brain Library (63). sh-Pik3cb reduced mTOR phosphorylation in gross mPFC tissue punches. (e) The presynaptic marker synaptophysin was also measured: Synaptophysin was reduced with repeated cocaine exposure, and sh-Pik3cb restored levels. For (d) and (e), representative blots loaded in the same order are below with an HSP-70 loading control and molecular weights indicated. Means+SEMs, *p<0.05. This figure represents 4 independent cohorts of mice; western blots were conducted at least twice.

Next, we repeated our experiment using a 2x2 between-subjects design, this time including mice that received only saline prior to the cocaine “challenge.” Control mice receiving cocaine for the first time locomoted twice as much as when they received saline, while mice with a history of cocaine exposure locomoted nearly 4 times as much – a sensitized response (fig.2c). Pik3cb knockdown reduced locomotor activity in cocaine-exposed mice, such that these mice were indistinguishable from animals that had never had cocaine before; i.e., Pik3cb knockdown blocked the expression of sensitization [cocaine × sh-Pik3cb F(1,49)=10.98, p=0.002; no main effects Fs<1](fig.2c).

We next dissected tissue from the mPFC. sh-Pik3cb reduced phospho-mTor, as expected [main effect of sh-Pik3cb F(1,17)=9.3, p=0.007; no cocaine effects or interactions Fs≤1](fig.2d). Effects were modest, presumably because tissue punches contained both transfected and unaffected cells, and p110β is only one of multiple PI3K subunits. Nevertheless, in addition to the main effect of viral vector, planned comparisons confirmed that sh-Pik3cb reduced phospho-mTor even in cocaine-exposed mice [cocaine + scrambled control vs. cocaine + sh-Pik3cb t(9)=3.1, p=0.01](fig.2d). Thus, reducing Pik3cb blunts phopsho-mTor following cocaine.

Cocaine reduces the intrinsic excitability of PL neurons (25), with a loss of dendritic spines and presynaptic boutons evident upon repeated exposure (26–28). Meanwhile, Pik3cb silencing reinstates dendritic spine homeostasis in certain disease models (18). We thus quantified the glycolipid protein synaptophysin, a commonly-used synaptic marker, in the same tissues. Cocaine decreased synaptophysin by approximately 20%, which was blocked by sh-Pik3cb [interaction F(1,19)=6.0, p=0.02; no main effects Fs≤1](fig.2e). This pattern suggests that reducing p110β activities or levels can restore synaptic presence after cocaine.

Correction of cocaine-induced decision-making biases

Next, we aimed to understand whether reducing Pik3cb could normalize some of the decision-making abnormalities that manifest with repeated cocaine exposure. We utilized a procedure that tests whether mice behave based on outcome expectations (goals) or instead, reflexively (habitually) (fig.3a). We exposed mice to cocaine (30 mg/kg daily, 14 days), then reduced Pik3cb, then trained mice to perform two distinct food-reinforced nose poke responses (fig.3a). All groups acquired the responses, with cocaine-exposed mice generating slightly higher response rates at the end of training [cocaine x session F(6,180)=4.2, p=0.001; main effect of session F(6,180)=121.81, p<0.001; no other main effects or interactions ps>0.5](fig.3b). Importantly, we found no response biases (“contingency to be intact” vs. “contingency to be violated”) that could affect our subsequent findings (no main effect or interactions Fs<1)(fig.3b).

Fig. 3. p110β inhibition blocks cocaine-induced habit-like behavior.

Fig. 3.

Open in a new tab

(a) Experimental timeline and schematic: Mice were exposed to cocaine, then infused with viral vectors, then trained to generate two food-reinforced operant responses. Next, the predictive relationship between one response and associated reinforcer was violated. The inhibition of that response reflects sensitivity to the contingency. (b) A history of cocaine exposure modestly elevated food-reinforced response rates during training. *p≤0.05 saline vs. cocaine. (c) When a familiar action failed to produce outcomes, control mice inhibited that behavior. Cocaine blocked sensitivity to changes in contingencies, while sh-Pik3cb restored sensitivity. *p≤0.05 contingent vs. noncontingent. (d) The same data are represented as response ratios, in which case, a preference ratio of 1 represents no preference (dashed line). Cocaine-exposed mice generated a score of 1, while sh-Pik3cb+cocaine mice preferred the response most likely to be reinforced (values >1, *p=0.02 vs. 1). Means+SEMs, n=9 control, 7 cocaine only, 11 sh-Pik3b only, 7 cocaine + sh-Pik3b. Experiments were conducted in 2 independent cohorts.

We next violated the predictive relationship between one response and the associated outcome by delivering food pellets noncontingently, and responding was not reinforced. Control mice then inhibited that response. Cocaine-exposed mice failed to adjust their response strategies, generating both responses equivalently. Meanwhile, Pik3cb knockdown reinstated outcome sensitivity, such that Pik3cb knockdown mice preferred the behavior that was most likely to be reinforced [cocaine x sh-Pik3cb x response choice F(1,30)=4.1, p=0.05; no main effect of nose poke F=1; all other main effects and interactions ps>0.1](fig.3c).

These same response rates can be distilled down to preference scores – the ratio of responses directed towards the intact vs. violated contingency. Scores of 1 reflect no change in behavior based on contingency, and scores >1 reveal outcome-oriented response preferences. Cocaine-exposed mice generated a score of 1 as expected [1-sample t6=0.39, p=0.7 vs. 1], while Pik3cb inhibition increased that score, improving goal sensitivity in cocaine-exposed mice [1-sample t6=3.0, p=0.02 vs. 1](fig.3d).

Cocaine diminishes availability of dopamine D2 receptors (D2R) (29,30), which are otherwise preferentially expressed on PL neurons necessary for goal-oriented action (31–33). We hypothesized that optimal responding in our task would require the activity of D2R+ excitatory PL neurons, and that reducing p110β content might normalize decision-making behavior if D2R+ PL neurons were compromised. We utilized D2R-Cre mice and simultaneously infused Cre-dependent Gi-DREADDs to inducibly reduce the excitability of D2R+ neurons, and sh-Pik3cb or a control viral vector into D2R-Cre+ and D2R-Cre− mice (fig.4a–b).

Fig. 4. D2R+ PL neurons are necessary for mice to select actions based on outcome expectancies; inhibiting p110β improves this ability.

Fig. 4.

Open in a new tab

(a) D2R-Cre+ mice or their Cre-littermates were infused into the PL with Cre-dependent Gi-coupled DREADDs (“DIO-DREADDs”). Mice were co-administered either a scrambled control viral vector or an sh-Pik3cb-expressing viral vector, ultimately creating four groups. (b) Experimental timeline. (c) Mice acquired the nose poke responses. (d) Chemogenetic inhibition of D2R+ excitatory neurons ablated the ability of mice to generate goal-directed response preferences; meanwhile, silencing Pik3cb restored response preferences. We next repeated the test procedure. In this case, all mice preferred the action likely to be reinforced, indicating that PL involvement is transient. Means+SEMs, *p<0.05, n=9 Cre− scrambled control, 10 Cre− sh-Pik3cb, 8 Cre+ scrambled control, 7 Cre+ sh-Pik3cb. Experiments were conducted in 3 independent cohorts.

Mice acquired the food-reinforced nose poke responses [main effect of session F(6,180)=64.3, p<0.001](fig.4c). We detected a main effect of Cre [F(1,30)=5.46, p=0.03] – consistent with unexpected consequences of the Cre transgene in bacterial artificial chromosome (BAC)-Cre mice. Importantly, we found no responses biases (“contingency to be intact” vs. “contingency to be violated”) that could affect our subsequent findings [no nose poke main effect or interactions Fs<1; no effect of sh-Pik3cb F(1,30)=1.31, p=0.26](fig.4c).

We then provided food pellets associated with one behavior noncontingently, as above, and injected all mice with the DREADDs ligand Clozapine N-oxide immediately following, when new memories are being formed. We tested response preference the next day when mice were drug-free. Chemogenetic inhibition of D2R+ excitatory neurons ablated the ability of mice to generate goal-oriented response preferences; meanwhile, silencing Pik3cb restored response preference [D2R-Cre+ x sh-Pik3cb x response interaction F(1,30)=4.18, p=0.05; main effect of nose poke F(1,30)=20.49, p<0.001; no other group or interaction effects, all ps>0.1](fig.4d). Thus, inhibiting Pik3cb in the PL improves outcome sensitivity following both cocaine and chemogenetic silencing of D2R+ neurons.

We next repeated the procedure in the same mice, again violating a familiar contingency, while leaving another intact. In analyzing response preferences during the second choice test, only a main effect of response was detected [F(1,27)=53.42, p<0.001](fig.4d). While a trend for a Cre x nose poke interaction was noted [F(1,27)=3.73, p=0.06], the key 3-factor interaction was ablated [F(1,27)=0, p=0.99]. This pattern is consistent with evidence that multiple brain regions are involved in detecting novel action-outcome contingencies, and that PL involvement is transient (reviewed ref.34); this experiment is the first, to our knowledge, to specifically implicate D2R+ neurons.

Silencing mPFC Pik3cb alters the transcriptomic profile of the DMS

We next wanted to identify circuit-level changes resulting from Pik3cb silencing in the mPFC. We focused on the DMS, since the DMS is a major PL projection target (35). Further, prefrontal-to-striatal connections are overwhelmingly implicated in cocaine-driven reward-seeking behavior and disruptions in adaptive, outcome-oriented decision making (36). First, we confirmed that Pik3cb was present in these projection neurons. Using naïve, wildtype mice, we deployed vTRAP (37) to isolate ribosome-bound mRNA transcripts from projection-defined PL neurons (fig.5a–b). We extracted PL-to-DMS neurons and for comparison, PL-to-OFC neurons, given that they are also involved in expectancy updating (38). We then quantified Pik3cb, and also Pik3cd (encoding the p110ō catalytic PI3K subunit), Pik3r1 (encoding the p85 regulatory PI3K subunit), and Bdnf (encoding Brain-derived Neurotrophic Factor: BDNF) as a positive control, with qPCR. Transcripts were highly enriched (~100x order of magnitude) in both DMS-and OFC-projecting PL projection neurons relative to bulk PL tissue (fig.5c), indicating that key projection neurons express Pik3cb, as well as the obligate Pik3r1 regulatory subunit.

Fig. 5. p110β inhibition in the PL changes the transcriptional profile of the DMS.

Fig. 5.

Open in a new tab

(a) Retrogradely transported Cre-recombinase constructs infused into either OFC or DMS drive Cre-dependent expression of the EGFP-L10a “ribotag” selectively within OFC- or DMS-projecting PL neurons, respectively. (b) EGFP-L10a expression in the mPFC. (“IL” refers to the infralimbic cortex.) Scale bar=250 μm. Tissue punches from the PL (red circle) were dissected and processed by immunoprecipitation to isolate ribosome-bound mRNA transcripts in projection-defined PL neurons. (c) Quantification of mRNA expression for Pik3cb, Pik3cd, Pik3r1, and Bdnf in PL➜OFC and PL➜DMS neurons. Enrichment value for each transcript calculated as the fold change of mRNA quantity from samples following ribotag immunopurification versus from “input” controls prior to immunopurification, i.e., bulk tissue expression. Tissue from 3 animals were pooled for each group (PL➜OFC vs. PL➜DMS). Samples were prepared and amplified in duplicate, and this entire protocol was performed twice. (d) A control viral vector or sh-Pik3cb was infused into the PL, and tissue was collected from the DMS or OFC. (e) DEGs were defined using a nominal significance criteria of p<0.05 and FC >20%. Comparing the number of DEGs per region revealed a more robust pattern of gene expression change in the DMS vs. OFC. (f) Volcano plot showing pattern of transcriptomic changes in DMS and OFC across all genes analyzed. (g) When considering genes whose expression differences satisfied p<0.05 only, overlap remained sparse. (h) Comparison of DEGs across both regions (264 total genes; see supplementary tables for full gene list) revealed a striking discordance in transcriptomic changes in the DMS vs. OFC. (i-j) Among the 9 DEGs in both DMS and OFC, the direction of expression change was not uniform between regions (DMS is red; OFC is blue). n=6/group.

We next used bulk tissue RNA-sequencing to generate a transcriptomic profile of the DMS and OFC (fig.5d–g) from mice with sh-Pik3cb in the PL. Differentially expressed genes (DEGs) following knockdown vs. control were defined using a nominal significance criteria of p<0.05 and fold change (FC) >20%. Comparing the number of DEGs per region (DMS: 220 genes; OFC: 44 genes) revealed a more robust pattern of gene expression change in the DMS compared to OFC (fig.5e–g; Supplementary Table 1). Notably, although the number of comparisons in the OFC that satisfied p<0.05 (429 genes) was greater than in the DMS (259 genes), only a small proportion of those OFC genes (~10%) satisfied our relatively modest FC criteria (cf.,39–41), compared to the DMS (~85%).

Interestingly, we found no shared DEGs between DMS and OFC following Pik3cb silencing in the PL. Even when considering genes whose expression differences satisfied p<0.05 only, overlap was sparse (9 genes; fig.5g; Supplementary Table 2); and while these genes could conceivably represent some shared response to Pik3cb knockdown, the direction of expression change was not uniform (fig.5h–i; Supplementary Table 2). Further, comparison of the global patterns of transcriptomic changes in DEGs across both regions (again, DMS: 220 genes; OFC: 44 genes) revealed a striking discordance between the DMS and OFC (fig.5h). Together, our analyses reveal a divergent response to PL-specific Pik3cb knockdown in the DMS vs. OFC, and these effects are distinct in both the magnitude and patterning of gene expression changes.

We next determined whether the observed transcriptomic changes were organized into functional categories. We employed a gene set enrichment analysis (GSEA; ref. 42), which identified transcriptomic signatures across our entire gene expression dataset using experimentally-derived gene sets (GSs). We tested 8,166 individual GSs (accessed via MSigDB; ref.43) for both regions. In the OFC, we did not find any enriched GSs (significance defined as p<0.05 and FDR<0.25; fig.6a), suggesting that while some individual genes in the OFC changed as a function of Pik3cb silencing in the nearby PL, these changes are not obviously organized around known biological processes. In contrast, 142 GSs were enriched in the DMS (fig.6a–b; Supplementary fig. 1). Among these, the most common group of significantly enriched GSs was related to brain/neuronal processes (29 gene sets; fig.6b–c), supporting the ability of our analysis to detect biologically relevant patterns.

Fig. 6. p110β inhibition in the PL changes gene set profiles in the DMS.

Fig. 6.

Open in a new tab

(a) We employed a GSEA, which identified transcriptomic signatures across our entire gene expression dataset using experimentally derived GSs. In the OFC, we did not find any significantly enriched GSs. In contrast, 142 GSs were found to be significantly enriched in the DMS. (b) Among these, the most common group of significantly enriched GSs was related to brain/neuronal processes (29 GSs). The next most frequently enriched GSs were related to cytoskeleton projection and vesicle trafficking (22 GSs each). (c) Normalized enrichment scores (ES) for significantly enriched GSs related to brain/neuronal processes, with more negative values indicating greater enrichment following Pik3cb knockdown vs. control. n=6/group.

Many brain/neuronal-enriched GSs related to the structure and function of synapses, including synaptic plasticity, dendritic spine dynamics, and synapse organization, among others (fig.6c). Additionally, the next most frequently enriched GSs were related to cytoskeleton processes (22 GSs), vesicle trafficking (22 GSs), the endolysosome (13 GSs), and the trans-golgi network (12 GSs), all of which represent cellular processes fundamental to the regulation of synaptic structure and function. Thus, inhibiting PI3K p110β in the PL triggers robust and diverse striatal adaptations that may contribute to the ability of p110β silencing to comprehensively combat cocaine-elicited behavioral sequelae.

Discussion

PI3K is a signaling complex that phosphorylates second messengers to regulate neuronal plasticity and gene transcription. Several independent groups have revealed that cocaine elevates PI3K activity in the mPFC (see Introduction). Further, hyper-activation is long-lasting, existing beyond the period of drug exposure, and may thus be associated with the long-term consequences of cocaine such as craving and altered decision making. The complex is composed of p85 regulatory and p110 catalytic subunits, which determine the intracellular positioning and activation properties of PI3K. The subunits can be experimentally manipulated individually – a strategy long used in cancer research but largely not in behavioral contexts. We report that the neuronally enriched p110β subunit is expressed in the postnatal PFC in mice (fig.1), as in humans (21). Reducing Pik3cb, which encodes p110β, mitigates the cue-induced reinstatement of cocaine seeking (a model for studying relapse), cocaine-induced locomotor sensitization (a measure of cocaine-induced plasticity), and cocaine-induced habitual behavior (a measure of cocaine-induced decision-making biases). Pik3cb knockdown also has circuit-wide consequences, altering the transcriptomic profile of the striatum, which may provide neuroprotection against cocaine by triggering coordinated cortico-striatal adaptations.

Reducing p110β combats cocaine-elicited neurobehavioral sequelae

The mPFC is a central hub in the brain’s “reward circuitry” (23), and glutamatergic PL-to-striatal connections control the reinstatement of cue-induced drug seeking in multiple diverse rodent models for studying relapse (44). We first discovered that reducing Pik3cb in the PL interferes with the cue-elicited reinstatement of cocaine seeking in mice, reducing responding by roughly half. Then, in assessing locomotor sensitization, we found that silencing Pik3cb blocked its expression. Non-selective PI3K and mTor inhibitors also block the expression of sensitization (45–47), as well as cocaine- and methamphetamine-conditioned place preference (46,48) – thus inhibiting drug-induced behaviors after they have formed. Given that reducing p110β levels or activity is sufficient to decrease phosphorylated (active) mTor (18; fig.2) and Akt (49), we imagine that silencing Pik3cb imposes a break between cocaine and intracellular sequelae that trigger drug seeking and sensitized locomotion. What is remarkable, though, is that silencing the p110β isoform, while leaving others intact, is sufficient to impose this break.

We next tested whether reducing Pik3cb could mitigate cocaine-induced habit-based behavior, thought to contribute to drug misuse in humans (36,50). We trained mice to generate two food-reinforced instrumental behaviors, then reduced the likelihood that one response would be reinforced and, instead, provided food noncontingently. Rodents that are sensitive to the predictive relationship between actions and their consequences will inhibit that response. By contrast, equivalent engagement of both responses reflects a reliance on familiar behaviors that are insensitive to outcomes – habitual behavior. Cocaine weakened the ability of mice to select actions based on causal knowledge, causing them to defer to habit-like behaviors, as has been reported by several groups (reviewed ref.24). Meanwhile, reducing Pik3cb improved this ability, blocking cocaine-induced habitual behavior.

Repeated cocaine reduces the intrinsic excitability of PL neurons (25), concurrent with a loss of synapses, dendritic spines, dendrite complexity, presynaptic markers and boutons, and mediators of vesicular glutamate release (26–28,51). Pik3cb silencing, meanwhile, reinstates typical, homeostatic dendritic spine densities in certain disease models (18), raising the possibility that it could reinstate synaptic homeostasis here. To begin to address this issue, we quantified the synaptic marker synaptophysin. Cocaine reduced synaptophysin levels as expected, and Pik3cb silencing corrected levels, suggesting that it restored typical synaptic presence.

Pik3cb silencing had no effects on locomotion or operant responses in cocaine-naïve mice. In other investigations, PFC-selective Pik3cb knockdown had no effects on appetitive extinction conditioning and nest building in otherwise naïve mice, but normalized these behaviors in Fmr1−/− mice, which lack Fragile X Mental Retardation Protein and suffer disinhibited PI3K-Akt-mTor signaling (18). Together, these patterns suggest that silencing Pik3cb largely impacts contexts in which PI3K-Akt-mTor signaling has been disrupted, presumably because other PI3K subunits are sufficient to support adaptive PI3K function under typical circumstances.

The notion that silencing Pik3cb has homeostatic-like consequences led us to next determine cell types within the PL that control reward-related action selection and identify the consequences of silencing Pik3cb when they were compromised. In humans, cocaine reduces cortical dopamine D2/3 receptor availability, with strongest effects in the mPFC (30), and it reduces Drd2 mRNA in the rodent mPFC (29). We hypothesized that silencing D2R+ neurons would, like cocaine, impair the ability of mice to select actions based on their outcomes, which was indeed the case. Simultaneously silencing Pik3cb, meanwhile, corrected response strategies. We envision that silencing Pik3cb increases the availability of the p110α PI3K isoform to form a complex with regulatory p85 subunits and AMPARs (52), thereby supporting excitatory plasticity of D2R+ neurons. Dopamine D1- and D2-type receptors express on excitatory PFC neurons in a largely segregated fashion (53–55), and our viral vector would be expected to transfect both cell types. Thus, another possibility, which is not mutually exclusive, is that reducing Pik3cb buffers the intracellular consequences of D1R binding, which is instead associated with risk-taking behavior (56,57).

Notably, prior experiments using a similar task structure, coupled with inducible inactivation techniques, indicated that the PL is essential for consolidating memories about actions and outcomes (reviewed ref.34). The present findings refine our understanding of that process because they reveal that D2R+ PL neurons, in particular, are involved in instrumental memory consolidation, countering reliance on familiar habits. Coincident stimulation of D2Rs and excitation of D2R-expressing projection neurons generates voltage fluctuations and spiking for hundreds of milliseconds following stimulation (31), which might transmit critical signals from the PFC to other structures to coordinate memory formation.

Network-wide consequences of Pik3cb silencing

The PL is situated within a complex and interconnected network of brain regions that together govern reward-related behaviors. We thus hypothesized that Pik3cb silencing in the PL might impact PL projection targets. The PL forms action-outcome associations via monosynaptic connections with the DMS (58), and PL-to-OFC interactions are thought to be involved in updating reward-related information (38). We found that both DMS- and OFC-projecting PL neurons express Pik3cb. Further, Pik3cb knockdown in the PL changed the transcriptomic landscape of the DMS, altering the levels of several genes (interestingly, to a greater degree than in the OFC). PL-to-DMS neurons arborize considerably, with collaterals terminating in the brainstem, spinal cord, claustrum, and striatum itself (59). Thus, the behavioral effects of Pik3cb knockdown reported here may be attributable to network-wide adaptations that support the reversal of, or recovery from, cocaine-induced defects.

Because cortico-striatal connections are unidirectional, the transcriptomic changes we observe in DMS tissue likely reflect adaptations by DMS neurons to Pik3cb knockdown-mediated changes to PL inputs. Indeed, among the significantly enriched GSs, a large proportion related to the structure or function of synapses, dendrites and dendritic spines, or processes related to (post-)synaptic plasticity (15 of 29 GSs in total; see fig.6). Additionally, among the significantly enriched GSs that were not explicitly linked to brain/neuronal functions, the most common categories were related to cytoskeletal projection and vesicular trafficking, which are basic cellular processes required for synaptic modulation and plasticity. Medium spiny neurons (MSN) comprise 90-95% of all striatal neurons, so it is likely that the transcriptomic changes we observed in the DMS predominantly reflect alterations in MSNs (60). Since MSNs are projection-type neurons, it is unsurprising that many of the significantly enriched GSs in the DMS were related to axonal projection structure and function (see fig.6). Overall, these patterns suggest that Pik3cb knockdown in the PL has “knock-on” effects, triggering multi-synaptic adaptations in brain regions downstream of the PL.

Our findings raise the possibility that distal PL connections contribute to the therapeutic-like effect of Pik3cb knockdown in the PL. Recent single-cell transcriptomic analyses revealed that in the PFC, deep-layer excitatory neurons, the principle source of long-range afferent and efferent projections (61), are preferentially impacted by cocaine self-administration relative to other cell types, and effects on their gene expression patterns are persistent and especially pronounced following prolonged withdrawal (62). Given that Pik3cb silencing combatted long-term effects of cocaine here, future experiments could focus in particular on p110β control of deep-layer neuron structure and function. Further, our transcriptomics findings establish clear, testable predictions: that many striatal DEGs and GSs identified here might be modified by cocaine; that their levels would be corrected by Pik3cb silencing within the PL; and that these changes would be causally related to behavior. We believe that understanding these relationships is important because inhibiting p110β activity could have more selective biological consequences relative to pan-PI3K or mTor inhibitors, making it an appealing drug target. New knowledge regarding the pathophysiology of drug misuse may also be revealed.

Supplementary Material

Supplement 1
Table S1

KEY RESOURCES TABLE

Resource Type Specific Reagent or Resource Source or Reference Identifiers
Antibody anti-mTOR, Rabbit polyclonal Cell Signaling lot 6, Cat.# 2972S
anti-phospho-mTOR (S2448), Rabbit polyclonal Cell Signaling lot 18, Cat.# 2971S
anti-synaptophysin, Rabbit monoclonal Abcam lot Gr196393-2, Cat.# 32127
anti-HSP-70, Mouse monoclonal Santa Cruz lot F0413, Cat.# 7298
horseradish peroxidase-conjugated goat anti-rabbit Vector Cat.# PI-1000
antibodies Htz-GFP-19C8 and Htz-GFP-19F7 Memorial-Sloan Kettering Monoclonal Antibody Facility Htz-GFP-19C8 and Htz-GFP-19F7
Drug Cocaine hydrochloride Sigma-Aldrich Cat.# C5776
Clozapine-N-oxide RTI International Cat.# C-929
RNA-sequencing conducted according to established SOPs Yerkes Nonhuman Primate Genomics Core http://www.yerkes.emory.edu/nhp_genomics_core/services/index.html
Organisms Mouse: C57BL/6, both sexes The Jackson Labs number 000664
Mouse: Tg(Drd2-cre)ER44Gsat, both sexes MMRC, GENSET collection RRID:MMRRC_032108-UCD
Software; Algorithm Cufflinks see Trapnell et al., 2012 see Trapnell et al., 2012
Broad Institute’s Molecular Signatures Database see Liberzon et al., 2015 see Liberzon et al., 2015
SPSS IBM SPSS v.27
SigmaPlot Systat Software, Inc. SigmaPlot v.13
Viral Vectors LV-CMV-sh-Pik3cb and scrambled control Emory University Viral Vector Core n/a
AAV5-hSyn-DIO-hM4D(Gi)-mCherry AddGene 44362-AAV5
AAV5-FLEX-EGFP-L10a AddGene 98747-AAV5
AAVrg-hSyn-Cre-WPRE-hGH AddGene 105553-AAVrg
Open in a new tab

Acknowledgements:

The authors thank members of the Emory Yerkes Genomics Core, and Dr. Marina Wheeler, Ms. Yong Yang, Dr. Aurelie Menigoz, Dr. Rachel Davies, and Dr. Kenneth McCullough for valuable advice and assistance. We thank Dr. Kerry Ressler for generous resource sharing. This work was supported by PHS MH103748, GM008602, and DA044297; the Georgia Research Alliance; and the Marcus Foundation and Children’s Healthcare of Atlanta. The Yerkes National Primate Research Center is supported by NIH OD011132. The Emory Viral Vector Core is supported by an NINDS Core Facilities grant, P30 NS055077.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Disclosures: The authors report no biomedical financial interests or potential conflicts of interest.

References

  • (1).World Drug Report (2019) United Nations publication, Sales No. E.19.XI.8. [Google Scholar]
  • (2).Centers for Disease Control and Prevention, National Center for Health Statistics (2020) Multiple Cause of Death 1999-2017 on CDC WONDER Online Database. [Google Scholar]
  • (3).Alvaro-Bartolome M, Garcia-Sevilla JA (2013) Dysregulation of cannabinoid CB1 receptor and associated signaling networks in brains of cocaine addicts and cocaine-treated rodents. Neuroscience 247:294–308. [DOI] [PubMed] [Google Scholar]
  • (4).Szumlinski KK et al. (2019) PI3K activation within the ventromedial prefrontal cortex regulates the expression of drug-seeking in two rodent species. Addict Biol 24:1216–1226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (5).Zhang X et al. (2006) PI3 kinase is involved in cocaine behavioral sensitization and its reversal with brain area specificity. Biochem Biophys Res Comm 340:1144–1150. [DOI] [PubMed] [Google Scholar]
  • (6).Zhang Y, Zhu X, Huang C, Zhang X (2015) Molecular changes in the medial prefrontal cortex and nucleus accumbens are associated with blocking the behavioral sensitization to cocaine. Sci Rep 5:16172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (7).Park HJ, Cui FJ, Hwang JY, Kang UG (2010) Effects of clozapine on behavioral sensitization induced by cocaine. Psychiatry Res 175:165–170. [DOI] [PubMed] [Google Scholar]
  • (8).Giannotti G et al. (2014) Prolonged abstinence from developmental cocaine exposure dysregulates BDNF and its signaling network in the medial prefrontal cortex of adult rats. Int J Neuropsychopharmacol 17:625–634. [DOI] [PubMed] [Google Scholar]
  • (9).Caffino L, Giannotti G, Messa G, Mottarlini F, Fumagalli F (2019) Repeated cocaine exposure dysregulates BDNF expression and signaling in the mesocorticolimbic pathway of the adolescent rat. World J Biol Psychiatry 20:531–544. [DOI] [PubMed] [Google Scholar]
  • (10).Garcia-Pardo MP, Roger-Sanchez C, Rodriguez-Arias M, Minarro J, Aguilar MA (2016) Pharmacological modulation of protein kinases as a new approach to treat addiction to cocaine and opiates. Eur J Pharmacol 781:10–24. [DOI] [PubMed] [Google Scholar]
  • (11).Guillermet-Guibert J et al. (2008) The p110beta isoform of phosphoinositide 3-kinase signals downstream of G protein-coupled receptors and is functionally redundant with p110gamma. Proc Natl Acad Sci USA 105:8292–8297. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (12).Gross C et al. (2010) Excess phosphoinositide 3-kinase subunit synthesis and activity as a novel therapeutic target in fragile X syndrome. J Neurosci 30:10624–10638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (13).Vanhaesebroeck B, Guillermet-Guibert J, Graupera M, Bilanges B (2010) The emerging mechanisms of isoform-specific PI3K signaling. Nat Rev Mol Cell Biol 11:329–341. [DOI] [PubMed] [Google Scholar]
  • (14).Gross C, Bassell GJ (2014) Neuron-specific regulation of class I PI3K catalytic subunits and their dysfunction in brain disorders. Front Mol Neurosci 7:12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (15).Rong R et al. (2003) PI3 kinase enhancer-Homer complex couples mGluR1 to PI3 kinase, preventing neuronal apoptosis. Nat Neurosci 6:1153–1161. [DOI] [PubMed] [Google Scholar]
  • (16).Hawkins PT, Anderson KE, Davidson K, Stephens LR (2006) Signalling through class I PI3Ks in mammalian cells. Biochem Soc Trans 34:647–662. [DOI] [PubMed] [Google Scholar]
  • (17).Gong S et al. (2007) Targeting Cre recombinase to specific neuron populations with bacterial artificial chromosome constructs. J Neurosci 27:9817–9823. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (18).Gross C et al. (2015) Selective role of the catalytic PI3K subunit p110β in impaired higher order cognition in fragile X syndrome. Cell Rep 11:681–688. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (19).Ehrengruber MU, Hennou S, Büeler H, Naim HY, Déglon N, Lundstrom K (2001) Gene transfer into neurons from hippocampal slices: comparison of recombinant Semliki Forest Virus, adenovirus, adeno-associated virus, lentivirus, and measles virus. Mol Cell Neurosci 17:855–871. [DOI] [PubMed] [Google Scholar]
  • (20).Urban DJ, Roth BL (2015) DREADDs (Designer receptors exclusively activated by designer drugs): Chemogenetic tools with therapeutic utility. Ann Rev Pharmacol Toxicol 55:399–417. [DOI] [PubMed] [Google Scholar]
  • (21).Rapoport SI, Primiani CT, Chen CT, Ahn K, Ryan VH (2015) Coordinated expression of phosphoinositide metabolic genes during development and aging of human dorsolateral prefrontal cortex. PLoS One 10:e0132675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (22).Bi L, Okabe I, Bernard DJ, Nussbaum RL (2002) Early embryonic lethality in mice deficient in the p110beta catalytic subunit of PI3-kinase. Mamm Genome 13:169–172. [DOI] [PubMed] [Google Scholar]
  • (23).Moorman DE, James MH, McGlinchey EM, Aston-Jones G (2015) Differential roles of medial prefrontal subregions in the regulation of drug seeking. Brain Res 1628:130–146. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (24).Gourley SL, Taylor JR (2016) Going and stopping: Dichotomies in behavioral control by the prefrontal cortex. Nat Neurosci 19:656–664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (25).Chen BT, Yau HJ, Hatch C, Kusumoto-Yoshida I, Cho SL, Hopf FW, Bonci A (2013) Rescuing cocaine-induced prefrontal cortex hypoactivity prevents compulsive cocaine seeking. Nature 496:359–362. [DOI] [PubMed] [Google Scholar]
  • (26).Caffino L, Giannotti G, Malpighi C, Racagni G, Fumagalli F (2015) Short-term withdrawal from developmental exposure to cocaine activates the glucocorticoid receptor and alters spine dynamics. Eur Neuropsychopharmacol 25:1832–1841. [DOI] [PubMed] [Google Scholar]
  • (27).Rasakham K et al. (2014) Synapse density and dendritic complexity are reduced in the prefrontal cortex following seven days of forced abstinence from cocaine self-administration. PLoS One 9:e102524. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (28).Zhu W, Ge X, Gao P, Li M, Guan Y, Guan X (2018) Adolescent cocaine exposure induces prolonged synaptic modifications in medial prefrontal cortex of adult rats. Brain Struct Funct 223:1829–1838. [DOI] [PubMed] [Google Scholar]
  • (29).Briand LA et al. (2008) Persistent alterations in cognitive function and prefrontal dopamine D2 receptors following extended, but not limited, access to self-administered cocaine. Neuropsychopharmacology 33:2969–2980. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (30).Narendran R, Mason NS, Himes ML, Frankle WG (2020) Imaging cortical dopamine transmission in cocaine dependent: A [11C]FLB 457-amphetamine positron emission tomography study. Biol Psychiatry, ahead of print. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (31).Gee S, Ellwood I, Patel T, Luongo F, Deisseroth K, Sohal VS (2012) Synaptic activity unmasks dopamine D2 receptor modulation of a specific class of layer V pyramidal neurons in prefrontal cortex. J Neurosci 32:4959–4971. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (32).Bradfield LA, Hart G, Balleine BW (2013) The Role of the Anterior, Mediodorsal, and Parafascicular Thalamus in Instrumental Conditioning. Front Sys Neurosci 7:51. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (33).Alcaraz F, Fresno V, Marchand AR, Kremer EJ, Coutureau E, Wolff M (2018) Thalamocortical and corticothalamic pathways differentially contribute to goal-directed behaviors in the rat. Elife 332517. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (34).Woon EP, Sequeira MK, Barbee BR, Gourley SL (2020) Involvement of the rodent prelimbic and medial orbitofrontal cortices in goal-directed action: A brief review. J Neurosci Res 98:1020–1030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (35).Mailly P, Aliane V, Groenewegen HJ, Haber SN, Deniau J-M (2013) The rat prefrontostriatal system analyzed in 3D: Evidence for multiple interacting functional units. J Neurosci 33:5718–5727. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (36).Everitt BJ, Robbins TW (2016) Drug addiction: Updating actions to habits to compulsions ten years on. Annu Rev Psychol 67:23–50. [DOI] [PubMed] [Google Scholar]
  • (37).Nectow AR et al. (2017) Rapid molecular profiling of defined cell types using viral TRAP. Cell Rep 19:655–667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (38).DePoy LM, Shapiro LP, Kietzman HW, Roman KM, Gourley SL (2019) β1-integrins in the developing orbitofrontal cortex are necessary for expectancy updating. J Neurosci 39:6644–6655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (39).Hodes GE et al. (2015) Sex differences in nucleus accumbens transcriptome profiles associated with susceptibility versus resilience to subchronic variable stress. J Neurosci 35:16362–16376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (40).Bagot RC et al. (2016) Circuit-wide transcriptional profiling reveals brain region-specific gene networks regulating depression susceptibility. Neuron 90:969–983. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (41).Bagot RC et al. (2017) Ketamine and imipramine reverse transcriptional signatures of susceptibility and induce resilience-specific gene expression profiles. Biol Psychiatry 81:285–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (42).Subramanian A et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 102:15545–15550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (43).Liberzon A, Birger C, Thorvaldsdottir H, Ghandi M, Mesirov JP, Tamayo P (2015) The molecular signatures database (MSigDB) hallmark gene set collection. Cell Syst 1:417–425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (44).Bobadilla A-C, Heinsbroek JA, Gipson CD, Griffin WC, Fowler CD, Kenny PJ, Kalivas PW (2017) Corticostriatal plasticity, neuronal ensembles, and regulation of drug-seeking behavior. Prog Brain Res 235:93–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (45).Izzo E, Martin-Fardon R, Koob GF, Weiss F, Sanna PP (2002) Neural plasticity and addiction: PI3-kinase and cocaine behavioral sensitization. Nat Neurosci 5:1263–1264. [DOI] [PubMed] [Google Scholar]
  • (46).Bailey J, Ma D, Szumlinski KK (2012) Rapamycin attenuates the expression of cocaine-induced place preference and behavioral sensitization. Addict Biol 17:248–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (47).Wu J, McCallum SE, Glick SD, Huang Y (2011) Inhibition of the mammalian target of rapamycin pathway by rapamycin blocks cocaine-induced locomotor sensitization. Neuroscience 172:104–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (48).Narita M et al. (2005) Involvement of mitogen-stimulated p70-S6 kinase in the development of sensitization to the methamphetamine-induced rewarding effects in rats. Neuroscience 132:553–560. [DOI] [PubMed] [Google Scholar]
  • (49).Gross C et al. (2019) Isoform-selective phosphoinositide 3-kinase inhibition ameliorates a broad range of fragile X syndrome-associated deficits in a mouse model. Neuropsychopharmacology 44:324–333. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (50).Ersche KD et al. (2016) Carrots and sticks fail to change behavior in cocaine addiction. Science 352:1468–1471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (51).Caffino L, Calabrese F, Giannotti G, Barbon A, Verheij MM, Racagni G, Fumagalli F (2015) Stress rapidly dysregulates the glutamatergic synapse in the prefrontal cortex of cocaine-withdrawn adolescent rats. Addic Biol 20:158–169. [DOI] [PubMed] [Google Scholar]
  • (52).Man H-Y, Wang Q, Lu W-Y, Ju W, Ahmadian G, Liu L, et al. (2003) Activation of PI3-kinase is required for AMPA receptor insertion during LTP of mEPSCs in cultured hippocampal neurons. Neuron 38:611–624. [DOI] [PubMed] [Google Scholar]
  • (53).Vincent SL, Khan Y, Benes FM (1993) Cellular distribution of dopamine D1 and D2 receptors in rat medial prefrontal cortex. J Neurosci 13:2551–2564. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (54).Gaspar P, Bloch B, Le Moine C (1995) D1 and D2 receptor gene expression in the rat frontal cortex: cellular localization in deifferent classes of efferent neurons. Eur J Neurosci 7:1050–1063. [DOI] [PubMed] [Google Scholar]
  • (55).Santana N, Menfod G, Artigas F (2009) Quantitative analysis of the expression of dopamine D1 and D2 receptors in pyramidal and GABAergic neurons of the rat prefrontal cortex. Cereb Cortex 19:849–860. [DOI] [PubMed] [Google Scholar]
  • (56).Onge JR St., Abhari H, Floresco SB (2011) Dissociable contributions by prefrontal D1 and D2 receptors to risk-based decision making. J Neurosci 31:8625–8633. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (57).Jenni NL, Larkin JD, Floresco SB (2017) Prefrontal dopamine D1 and D2 receptors regulate dissociable aspects of decision making via distinct ventral striatal and amygdalar circuits. J Neurosci 37:6200–6213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (58).Hart G, Bradfield LA, Fok SY, Chieng B, Balleine BW (2018) The bilateral prefronto-striatal pathway is necessary for learning new goal-directed actions. Curr Biol 28:2218–2229. [DOI] [PubMed] [Google Scholar]
  • (59).Levesque M, Parent A (1998) Axonal arborization of corticostriatal and corticothalamic fibers arising from prelimbic cortex in the rat. Cereb Cortex 8:602–613. [DOI] [PubMed] [Google Scholar]
  • (60).Castro DC, Bruchas MR (2019) A Motivational and Neuropeptidergic Hub: Anatomical and Functional Diversity within the Nucleus Accumbens Shell. Neuron 102:529–552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (61).Shipp S (2007) Structure and function of the cerebral cortex. Curr Biol 17:R443–449. [DOI] [PubMed] [Google Scholar]
  • (62).Bhattacharjee A, Djekidel MN, Chen R, Tuesta LM, Zhang Y (2019) Cell type-specific transcriptional programs in mouse prefrontal cortex during adolescence and addiction. Nat Commun 10:4169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (63).Rosen GD, et al. (2000) The mouse brain library at www.Mbl.Org. International Mouse Genome Conference 14:166. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1
NIHMS1687766-supplement-Supplement_1.pdf (680.6KB, pdf)
Table S1
NIHMS1687766-supplement-Table_S1.xlsx (98.7KB, xlsx)

RESOURCES