This study investigated the functional role of cyclin-dependent kinase inhibitor 1a (Cdkn1a or p21) in cocaine-induced responses using a knockout mouse model. Acute locomotor activity after cocaine administration (15 mg/kg, i.p.) was decreased in p21−/− mice, whereas cocaine-induced place preference was enhanced. Interestingly, κ-opioid–induced place aversion was also significantly enhanced. Concentration-dependent analysis of locomotor activity in response to cocaine demonstrated a rightward shift in the p21−/− mice. Pretreatment with a 5-hydroxytryptamine receptor antagonist did not alter the enhancement of cocaine-induced conditioned place preference in p21−/− mice, indicating a lack of involvement of serotonergic signaling in this response. Cocaine exposure increased p21 expression exclusively in the ventral sector of the hippocampus of rodents after either contingent or noncontingent drug administration. Increased p21 expression was accompanied by increased histone acetylation of the p21 promoter region in rats. Finally, increased neurogenesis in the dorsal hippocampus of p21−/− mice was also observed. These results show that functional loss of p21 altered the acute locomotor response to cocaine and the conditioned responses to either rewarding or aversive stimuli. Collectively, these findings demonstrate a previously unreported involvement of p21 in modulating responses to cocaine and in motivated behaviors.
Cocaine is a powerfully addictive psychoactive substance, and abuse of this drug has been a persistent problem for over 100 years (National Institutes of Health National Institute on Drug Abuse, 2010). The most recent National Survey on Drug Use and Health estimates that there are more than 1.5 million current cocaine users age 12 years or older in the United States alone, with over 850,000 of these individuals meeting the Diagnostic and Statistical Manual of Mental Disorders (fourth edition) criteria for dependence or abuse of cocaine (Substance Abuse and Mental Health Services Administration, 2014). Although there have been numerous studies on the neurochemical and neurobiological effects of various drugs of abuse including cocaine, many gaps in knowledge remain regarding the underlying mechanisms of substance use disorder.
Research concerning cocaine has typically focused on the nucleus accumbens and alterations in gene expression induced therein (Meyer et al., 2009; Renthal et al., 2009; Larson et al., 2010; LaPlant and Nestler, 2011; Robison and Nestler, 2011; Nestler, 2012), cocaine exposure can also affect the hippocampus, a region involved in the formation of long-term memories (Scoville and Milner, 1957; Larson et al., 2010; Robison and Nestler, 2011; Nestler, 2012). The hippocampus can be structurally and functionally divided into at least two distinct compartments or sectors: the dorsal hippocampus and the ventral hippocampus. The dorsal hippocampus is primarily involved in performing cognitive functions and in spatial learning and memory (Moser et al., 1995). The ventral hippocampus has been related to stress responses and emotional behaviors (Henke, 1990) and has been implicated in cocaine seeking in reports assessing either context-primed reinstatement (Lasseter et al., 2010) or drug-primed reinstatement (Rogers and See, 2007) models of relapse. The dentate gyrus of the hippocampus is a primary site for adult neurogenesis, or the rise of new neurons (Abrous et al., 2005). Interestingly, levels of adult hippocampal neurogenesis are negatively correlated with drug taking and drug seeking, suggesting an important role for neurogenesis in drug addiction (Noonan et al., 2010).
Cell cycle–related proteins, such as cyclin-dependent kinases (CDKs) and their inhibitors, mediate numerous cellular processes, including proliferation, a process integral to neurogenesis. In fact, neural progenitor cells derived from mice lacking cyclin-dependent kinase inhibitor 1a (Cdkn1a, p21) show higher proliferation than those derived from wild-type mice (Pechnick et al., 2011; Zonis et al., 2013); similarly, suppression of p21 using short hairpin RNA increases neural cell proliferation (Pechnick et al., 2011), suggesting that p21 may regulate neurogenesis. Although studies investigating the role of CDKs and their inhibitors in response to drugs of abuse are limited, published data illustrate a strong potential link between these proteins and cocaine-induced effects. For example, cocaine exposure has been shown to increase the expression of p21 in human neural precursor cells, correlating with decreased cell proliferation (Hu et al., 2006). In addition, Zhou et al. (2011) recently reported alterations in CDKN1b (p27) and other CDK-associated proteins in the postmortem brain tissue of cocaine users. Furthermore, a human genome-wide association study reported correlations between certain cocaine-induced behavioral responses and a single nucleotide polymorphism in CDK1, a target of p21 (Gelernter et al., 2014). These human data suggest a relationship between cell cycle regulators and cocaine-induced responses. The purpose of the current study was to further characterize the link between cell cycle–related proteins and cocaine-induced behaviors through the use of a functional p21 knockout (p21−/−) mouse model. We also determined the effect of cocaine and p21 on adult hippocampal neurogenesis.
Materials and Methods
Adult male Sprague-Dawley rats (Harlan, Indianapolis, IN), wild-type mice (B6;129SF2/J; Jackson Laboratory, Bar Harbor, ME), and p21−/− mice (B6;129S2-Cdkn1atm1Tyj/J; Jackson Laboratory) were used. All studies were approved by the University of Georgia Institutional Animal Care and Use Committee and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals. Rats were housed in pairs and mice were kept in the groups in which they arrived (3–5 mice). They were maintained on a 12-hour light/dark cycle (7:00 AM/7:00 PM) with food and water available ad libitum. Animals were allowed to acclimate to their home cages for at least 1 week and were habituated to handling (rats, 3 days; mice, 1 day) prior to testing. Sessions were conducted between 10:00 AM and 4:00 PM. All mice were genotyped for p21 genetic status based on the method suggested by the Jackson Laboratory.
Cocaine hydrochloride was obtained from the National Institutes of Health National Institute on Drug Abuse (RTI International, Research Triangle Park, NC), (±)-U50,488 [2-(3,4-dichlorophenyl)-N-methyl-N-[(1R,2R)-2-pyrrolidin-1-ylcyclohexyl]acetamide] hydrochloride was obtained from Tocris (Bristol, UK), and methiothepin maleate was obtained from Santa Cruz Biotechnology (Dallas, TX). All drugs were dissolved in 0.9% saline and filter sterilized prior to use.
Behavioral testing was carried out in 43.2-cm × 43.2-cm chambers with clear plastic walls and a smooth solid floor (Med Associates Inc., St. Albans, VT). Each chamber was located in a sound-attenuating box equipped with two house lights (20 lx) and a ventilation fan. Two banks of 16 infrared photo beams detected horizontal activity, and Activity Monitor software (Med Associates Inc.) was used to count beam breaks. Conditioned place preference (CPP) testing took place in a modified open field chamber using a two-compartment insert (Med Associates Inc.) specifically designed for use with mice. The insert divided the chamber into two compartments separated by a black partition with a guillotine door. When the door was removed, the mouse could access either compartment. When in place, the door confined the animal’s activity to one compartment for conditioning purposes. One compartment contained flooring of rod-like steel bars and black plastic walls, which allowed for the darkening of that side of the chamber insert. The other compartment contained a wire mesh grid floor and transparent walls. These were the only ways that the two compartments differed, and preliminary studies suggested that this arrangement yielded an equal preference between compartments.
The behavioral results for the initial rat locomotor sensitization (LMS) experiment have been published as part of another recent study, and these animals exhibited the expected increase in locomotor activity on the challenge day relative to the activity day (Cummings et al., 2015). For the assessment of LMS, the animal was placed in the center of an open field chamber for 30 minutes to determine baseline movement during the activity day (day 2). An intraperitoneal injection of either 0.9% saline, cocaine (15 mg/kg), or U50,488 (5 mg/kg) was given and the animal was placed back into the open field chamber for an additional 60 minutes. One week after the post-test, the open field challenge (day 14) test was conducted in the same manner as the activity test to assess sensitization. The timeline for this assay can be found in (Supplemental Fig. 1C).
CPP/Conditioned Place Aversion.
The experimental design for the mouse behavior assays was based on that described in previous studies (Seymour and Wagner, 2008). For the place preference assay, mice were first tested in a pretest session (day 1) in which each was placed in the lighter/grid floor compartment and the guillotine door was removed, allowing free access to both compartments for 15 minutes. The time spent in each compartment was measured to ensure that no animals had a strong bias (>65%) toward one side or another.
Conditioning sessions took place on experiment days 3–6 (Supplemental Fig. 1, A and B). Mice were divided into six groups: wild-type saline (n =14), wild-type cocaine (n =12), wild-type U50,488 (n = 6), p21−/− saline (n = 14), p21−/− cocaine (n = 12), and p21−/− U50,488 (n = 6). Animals were given an intraperitoneal injection of saline, cocaine (20 mg/kg), or U50,488 (5 mg/kg) and placed in one of the two insert compartments for 15 minutes and then returned to their home cage. Four hours later, animals were injected with either saline or drug and were confined to the opposite compartment for the second daily conditioning session. After the 4 days of conditioning, the mice underwent a drug-free place preference post-test on experiment day 7.
Dose Response and Serotonin Assays.
Mice underwent LMS assays as described above using concentrations of 15 mg/kg for wild-type mice, and 15, 20, or 25 mg/kg for p21−/− mice (n = 6 per group) to determine the locomotor response to an initial dose of cocaine at various concentrations. These animals also underwent the remainder of the LMS/CPP assay as described.
To determine the potential involvement of serotonin on the enhanced cocaine-induced CPP in p21−/− mice, on the four conditioning days (days 3–6), a subset of cocaine-paired mice (n = 6) were pretreated with 50 μg/kg (i.p.) of the serotonin receptor antagonist methiothepin maleate in the home cage 30 minutes prior to cocaine exposure (Hnasko et al., 2007). The mice were exposed to saline in the home cage 30 minutes prior to saline conditioning sessions.
Self-Administration Behavior Assays.
The experimental design for the cocaine self-administration assay using male Sprague-Dawley rats was described in Kelamangalath and Wagner (2010). Briefly, animals were allowed to self-administer cocaine in an operant environment (Med Associates Inc.) containing both active and inactive levers. Upon an active lever press, an infusion of cocaine was delivered at a volume of 0.125 ml/kg body weight (0.5 mg/kg per infusion) and a 30-second light/tone (conditioning stimulus) was presented. Inactive lever pressing had no programmed consequences. MED-PC software was used to record the number of active and inactive lever presses and the number of infusions. The self-administration protocol can be divided into three stages: stage 1, short access (90 minutes) under a fixed ratio (FR) schedule 1 (FR-1; 10 days); stage 2, short access/FR-3 (5 days); and stage 3, long access (6 hours)/FR-3 (7 days). The maximum number of infusions per session was 200, at which point that session was terminated and the animal was placed back in its home cage. Saline self-administration was carried out in the same manner as cocaine self-administration, but the animals received saline instead of cocaine infusions. Otherwise, these saline rats were treated in the same manner as cocaine self-administration rats in all respects to provide a rigorous control group.
The sucrose self-administration assays were performed as described previously (Kelamangalath et al., 2009). A separate cohort of rats was allowed to self-administer sucrose pellets under the same conditions described previously for cocaine, in which active lever presses delivered a 45-mg sucrose pellet.
To identify persistent changes beyond acute drug effects, tissue was extracted 7 days after the final exposure to cocaine. Animals were anesthetized with halothane prior to decapitation in compliance with protocols approved by the University of Georgia Animal Care and Use guidelines. The brain was removed from the cranium and regions of interest were extracted and flash frozen in liquid nitrogen. For immunohistochemistry and immunofluorescence studies (n ≥ 3 mice per group), the left hemisphere was placed midline down on a coverslip, coated in O.C.T. Embedding Compound (VWR, Radnor, PA), placed in −90°C isobutanol for 30 seconds, and kept at 80°C until cryosectioning.
Quantitative Polymerase Chain Reaction.
RNA was isolated region specifically from flash-frozen tissue using TRIzol (Invitrogen, Carlsbad, CA) based on the manufacturer’s protocol. Total RNA (100 ng) was used for cDNA synthesis using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). cDNA (100 ng) with a 260:280 > 1.8 was used for quantitative polymerase chain reaction (qPCR) using iTaq Universal SYBR Green Supermix (Bio-Rad) with primers specific for p21 (mouse sense: 5′-CTCTCCCAGTCTCCAAACTTAAA-3′, anti-sense: 5′-TCATCCTAGCTGGCCTTAGA-3′; rat sense: 5′-ACGTGGCCTTGTCGCTGTCTT-3′, anti-sense: 5′-TAAGGCAGAAGATGGGGAAGAG-3′) and glyceraldehyde phosphate dehydrogenase (mouse sense: 5′-TCAACAGCAACTCCCACTCTTCCA-3′, anti-sense: 5′-ACCCTGTTGCTGTAGCCGTATTCA-3′; rat sense: 5′- CACGGCAAGTTCAACGGCACAGTCA-3′, anti-sense: 5′- GTGAAGACGCCAGTAGACTCCAGG AC-3′). PCR conditions were 95°C for 30 seconds, followed by 45 cycles of 95°C for 5 seconds and 60°C for 40 seconds, then 1 minute at 95°C and 1 minute at 60°C. qPCR results were analyzed via 2−ΔΔCt to determine fold change between saline- and cocaine-treated animals normalized to glyceraldehyde phosphate dehydrogenase.
Chromatin immunoprecipitation for acetylated histone 3 was performed using the EpiQuick Tissue Acetyl-Histone H3 ChIP Kit based on the manufacturer’s protocol (Epigentek, Farmingdale, NY). Approximately 40 mg flash-frozen tissue (dorsal and ventral hippocampus) was used to obtain input and immunoprecipitated DNA. qPCR was then performed on both using primers for the p21 promoter region obtained from Yuan et al. (2013). PCR conditions were 95°C for 5 minutes, followed by 50 cycles of 95°C for 30 seconds, 59°C for 45 seconds, and 72°C for 45 seconds, then 72°C for 10 minutes. qPCR results were analyzed via 2−ΔΔCt to determine fold change between saline- and cocaine-treated animals normalized to the input DNA.
Protein analysis was performed region specifically using 10–15 mg flash-frozen tissue. Tissue was briefly sonicated in 300 µl standard radioimmunoprecipitation assay buffer. An additional 600 µl radioimmunoprecipitation assay buffer was then added and the samples were maintained at constant agitation for 2 hours at 4°C. The samples were then centrifuged for 20 minutes at 16,000 × g at 4°C, the supernatant was transferred to a new tube, and the pellet was discarded. Protein concentrations were determined using the BCA assay and samples (40 µg) were separated on 4%/9% stacked SDS-PAGE gels and transferred to nitrocellulose membranes. The membranes were blocked in 5% (w/v) milk powder in Tris-buffered saline/Tween 20 for 2 hours and exposed to antibodies overnight, including p21 F-5 (1:200; Santa Cruz Biotechnology Inc., Santa Cruz, CA), doublecortin (DCX) (1:1000; Abcam, Cambridge, MA), and neuronal nuclei (NeuN) (1:500; EMD Millipore, Billerica, MA). Membranes were then incubated with secondary antibody (1:2500 dilution; Promega, Madison, WI) for 4 hours and bands were visualized using SuperSignal Chemiluminescent Substrate (Thermo Scientific, Waltham, MA). Intensities were quantified using an Alpha Innotech FluorChem HD2 system (ProteinSimple, Santa Clara, CA) and normalized to β-actin.
Immunohistochemistry was performed on 10-µm slices using the Vectastain Universal (anti-mouse IgG/rabbit IgG) Elite ABC Kit (Vector Laboratories, Burlingame, CA) based on the manufacturer’s protocol for staining frozen sections. Briefly, sections were removed from −80°C and allowed to air dry at room temperature. They were then fixed for 10 minutes in prechilled acetone at −20°C before rinsing in phosphate-buffered saline (PBS). The sections were then placed in 0.3% H2O2 in methanol for 30 minutes to quench endogenous peroxidase activity. After washing for 5 minutes in PBS, the sections were incubated in diluted blocking serum for 20 minutes. The sections were again washed in PBS and incubated overnight with primary antibody against p21 (C-19, 1:100; Santa Cruz Biotechnology Inc.) diluted in 0.1% bovine serum albumin (BSA) in PBS. The sections were washed in PBS and incubated for 30 minutes each with diluted biotinylated secondary antibody solution and the Vectastain Elite ABC Reagent before staining with 3,3′-diaminobenzidine reagent (Vector Laboratories), counterstained using Gill’s hematoxylin, and mounted with Fluoromount (Sigma-Aldrich, St. Louis, MO). Visualization of staining was performed using a Nikon AZ100 microscope (Tokyo, Japan).
Ten-micrometer sections were removed from −80°C and allowed to air dry at room temperature for 20 minutes. They were then fixed in 1× paraformaldehyde fixing solution for 30 minutes, followed by three 10-minute washes with PBS. A permeabilization step was performed with 0.2% Triton X-100 for 2 hours. The tissue was washed three times for 10 minutes with PBS containing 0.05% Triton X-100 and then blocked with 5% BSA for 2 hours. The sections were exposed to a primary antibody against NeuN (1:100; Millipore) or DCX (1:500; Abcam) overnight and then washed three times in PBS-Triton wash buffer with 0.5% BSA for 30 minutes. After washing, slides were exposed to a fluorescent-tagged secondary antibody for 2 hours. The tissues were again washed with PBS-Triton with 0.5% BSA three times for 30 minutes, followed by three 5-minute washes with PBS. All steps were performed at 4°C and on an orbital shaker at low speed. Fluoromount was used to mount the tissue and visualization was performed using an Olympus IX71 microscope (Center Valley, PA).
Tissue isolated from a single animal or a single animal’s behavior data represent one experiment (n = 1). Behavior assays were performed using multiple experimental cohorts of four animals each to obtain the final n values presented. For data sets with confirmed Gaussian distribution after normality testing, the unpaired two-tailed t test was used to compare data across two groups. For those without Gaussian distribution, the nonparametric Mann–Whitney test was used. To compare across multiple groups, one-way analysis of variance followed by the Bonferroni post hoc test was performed. In all cases, GraphPad Prism software (GraphPad Software Inc., La Jolla, CA) was used and P < 0.05 was considered indicative of a statistically significant difference between mean values.
Locomotor Activity of Wild-Type Versus p21−/− Mice.
We first verified that wild-type mice respond to cocaine similarly to rats in a behavioral protocol that combines locomotor activity and CPP assessments in the same animal (Seymour and Wagner, 2008). Cocaine-treated wild-type mice exhibited increased locomotor activity postinjection (15 mg/kg i.p.) compared with saline-treated mice during the activity day (Supplemental Fig. 1A, protocol day 2). In addition, during the challenge day (protocol day 14), the cocaine-treated wild-type mice demonstrated significantly greater locomotor activity postexposure than that seen during the activity day (i.e., sensitization) (F = 12.60, t = 4.596; F = 17.57, t = 4.233; and F = 12.87, t = 2.980 for times 35, 40, and 45 minutes, respectively; df = 44; Fig. 1A).
We then investigated the relationship between p21 and cocaine-induced effects utilizing a p21 functional knockout (p21−/−) mouse model. The locomotor response to an initial cocaine exposure, as measured on the activity day, was significantly diminished in the p21−/− mice compared with the wild-type mice, with the activity of p21−/− mice being not significantly greater with cocaine than with saline administration (wild-type cocaine versus saline: t = 4.729, df = 22; knockout cocaine versus saline: U = 44; wild-type cocaine versus knockout cocaine: U = 27; Fig. 1, B and C). There was no difference in locomotor activity between saline-treated p21−/− and wild-type mice, and there was no difference in activity observed preinjection (Fig. 1, A and B). After conditioning with cocaine, the locomotor response to the drug was enhanced in p21−/− mice, as assessed on the challenge day, similar to that seen in the wild-type mice (F = 13.27, t = 5.606; and F = 6.605, t = 3.634 for times 35 and 40 minutes, respectively; df = 44; Fig. 1B). Furthermore, the magnitude of this cocaine-induced LMS did not significantly differ between mouse models (t = 0.4091, df = 22; Fig. 1D).
CPP of Wild-Type versus p21−/− Mice.
Wild-type mice again acted similarly in response to cocaine as rats during the CPP assay (Seymour and Wagner, 2008). Both cocaine-treated wild-type and p21−/− mice spent significantly more time in the designated drug-paired compartment during the CPP post-test than during the pretest (wild-type: t = 2.758, df = 22; knockout: t = 5.991, df = 22; Fig. 2A). Interestingly, this preference for the drug-paired compartment during the post-test was significantly greater for cocaine-treated p21−/− mice than wild-type mice (t = 2.065, df = 22). Saline-treated mice spent the same amount of time in the drug-paired compartment during the pretest and post-test.
κ-Opioid U50,488-Induced Conditioned Place Aversion in Wild-Type versus p21−/− Mice.
Other groups of wild-type and p21−/− mice were tested in a conditioned place aversion (CPA) assay using U50,488, a selective κ-opioid receptor agonist that causes dysphoria and stress-like effects in rodents (Muschamp et al., 2012; Smith et al., 2012; Valdez and Harshberger, 2012). U50,488 exposure induced a decrease of approximately 150 seconds in the time spent in the drug-paired compartment in wild-type mice during the post-test compared with the pretest. This value, however, was not significantly different from that seen with saline treatment (t = 1.513, df = 16; Fig. 2B). Unlike the wild-type mice, the U50,488-treated p21−/− mice spent significantly less time (approximately 280 seconds) in the drug-paired compartment during the post-test compared with the pretest (t = 3.173, df = 18; Fig. 2B). To maintain consistency, LMS was also investigated as part of this protocol. Similar to that observed by Ukai and Kameyama (1985), exposure to 5 mg/kg U50,488 did not alter locomotor activity compared with saline (Supplemental Fig. 2, A and B).
Locomotor Activity Dose Response in p21−/− Mice.
The diminished locomotor activity after an initial 15-mg/kg dose of cocaine observed in the p21−/− mice could be attributable to a rightward shift in dose response in these animals. To assess this, we determined the concentration-dependent effect of locomotor response to cocaine in p21−/− mice. Similar to data shown in Fig. 1, p21−/− mice displayed significantly less locomotor movement in response to exposure to 15 mg/kg cocaine compared with wild-type controls (U = 27; Fig. 3A). Furthermore, exposure of p21−/− mice to 20 and 25 mg/kg cocaine significantly increased locomotor activity compared with saline (20 mg/kg versus saline: t = 4.370, df = 16; 25 mg/kg versus saline: t = 2.861, df = 16), at which point locomotor movement in p21−/− mice was similar to that seen in wild-type mice exposed to 15 mg/kg. These data suggest that genetic ablation of p21 results in a rightward shift in the dose curve after an initial exposure to cocaine.
A possible contributor to the diminished locomotor response in p21−/− mice could be a deficient dopamine pathway (Giros et al., 1996; Sora et al., 1998, 2001; Uhl et al., 2002). In dopamine-deficient mice, CPP is mediated by serotonin (Sora et al., 1998, 2001; Hnasko et al., 2007). We pretreated p21−/− mice with the serotonin receptor antagonist methiothepin maleate, which has been shown to inhibit cocaine-induced CPP in mice with dopamine deficiencies (Hnasko et al., 2007). p21−/− mice developed comparable CPP after cocaine conditioning both with and without pretreatment with the antagonist (cocaine post-test versus pretest: t = 5.991, df = 22; cocaine plus antagonist post-test versus pretest: t = 3.476, df = 16; cocaine post-test versus cocaine plus antagonist post-test: t = 0.5045, df = 18; Fig. 3B), suggesting that serotonergic signaling is not contributing to CPP in these animals.
Effect of Cocaine Administration on p21 Expression and H3 Acetylation in Rats.
The above results support the hypothesis that the absence of functional p21 alters cocaine-induced behavior in mice; however, the data do not indicate the effect of cocaine on p21, or indicate how p21 may mediate cocaine-induced changes in behavior. Thus, we examined the effect of cocaine exposure on p21 expression. Analysis of striatal and hippocampal tissue samples demonstrated hippocampal-specific increases in p21 mRNA expression after noncontingent cocaine administration (t = 2.566, df = 6), with no changes being observed in either the dorsal or ventral striatum (dorsal: t = 1.391, df = 6; ventral: t = 1.2972, df = 6; Fig. 4A). On the basis of these findings, we confined our subsequent analyses to the hippocampus, further dividing this brain region into dorsal and ventral sectors.
Prior to assessing the effect of cocaine exposure on hippocampal p21 expression, we analyzed basal p21 expression in the dorsal and ventral sectors of control animals. We observed higher basal p21 mRNA expression in the dorsal hippocampus than in the ventral hippocampus of both mice and rats (rat: t = 4.173, df = 8; mouse: t = 3.199, df = 22; (Supplemental Fig. 3A). We next assessed the effect of contingent cocaine self-administration on the expression of p21 in both dorsal and ventral sectors of the hippocampus. Cocaine did not alter the expression of either p21 mRNA or protein in the dorsal hippocampus (mRNA: t = 0.3392, df = 14; protein: t = 1.409, df = 4; Fig. 4, B and C; (Supplemental Fig. 3A). By contrast, cocaine exposure significantly increased both p21 mRNA and protein in the ventral hippocampus of rats allowed to self-administer cocaine compared with saline self-administration rats (mRNA: t = 2.963, df = 14; protein: t = 3.1824, df = 6; Fig. 4, B and C; (Supplemental Fig. 3A).
We tested the hypothesis that cocaine-induced increases in p21 expression were mediated by epigenetic modification using chromatin immunoprecipitation for acetylated histone 3 followed by qPCR analysis for the p21 promoter region. Histone 3 acetylation of the p21 promoter region in untreated animals was higher in the dorsal sector than in the ventral sector of control rats (t = 3.740, df = 4; (Supplemental Fig. 3B), correlating with basal p21 mRNA expression (rat: t = 4.173, df = 8; mouse: t = 3.199, df = 22; (Supplemental Fig. 3A). Similar to above changes in p21 expression, cocaine exposure did not alter p21 promoter region histone 3 acetylation in the dorsal hippocampus compared with rats that had administered saline (Fig. 4D). By contrast, cocaine self-administration induced an approximate 5-fold increase in histone 3 acetylation of the p21 promoter region in the ventral hippocampus compared with saline self-administration (t = 2.666, df = 4; Fig. 4D).
Effect of Cocaine Treatment on p21 Expression in Wild-Type Mice.
Studies assessing alterations in p21 expression in response to cocaine were performed in rats, whereas much of our behavioral data were performed in mice. We addressed this limitation by assessing the effect of cocaine on p21 expression in mice. We also performed immunohistochemistry to identify changes in the specific subregions of the hippocampus. As observed in rats, repeated cocaine exposure increased p21 mRNA and protein expression selectively in the ventral sector of the hippocampus of mice (t = 2.875, df = 22; Fig. 5A) and there was no significant difference in staining between saline- and cocaine-treated animals in the dorsal hippocampus. Similar to data reported by Pechnick et al. (2011), positive p21 staining was present in the subgranular zone of the dentate gyrus (Fig. 5B, black arrows). In addition, positive staining was also observed in the pyramidal cell bodies of the cornu ammonis regions of the dorsal and ventral hippocampus of both saline- and cocaine-treated mice. Increased staining was particularly prominent in the cornu ammonis 1/ventral subiculum region (Fig. 5B, white arrows).
Effect of p21 on Hippocampal NeuN and DCX Expression.
The results obtained to this point show that the absence of functional p21 alters cocaine-induced behavior and that cocaine exposure increases p21 expression in the ventral hippocampus. However, these data do not discern the mechanism by which p21 affects behavior. We began to address this question by examining the effect of p21 on neurogenesis. We observed increased expression of DCX, a common marker for neurogenesis, in the dorsal hippocampus of p21−/− mice compared with wild-type mice (t = 5.087, df = 4; Fig. 6). We also assessed NeuN, a marker of adult neurons, which increases in expression as DCX expression decreases (Brown et al., 2003), to investigate more persistent effects of functional absence of p21 on neurogenesis. Similar to DCX, NeuN expression was also increased in the dorsal hippocampus of p21−/− mice compared with wild-type mice (t = 4.402, df = 4). These findings were confirmed using both immunofluorescence (Fig. 6A) and immunoblot analysis (Fig. 6, B and C). By contrast, significant differences in DCX or NeuN expression were not observed in the ventral hippocampus (data not shown).
Effect of Cocaine and U50,488 on Hippocampal NeuN Expression.
To further understand the role of neurogenesis in mediating drug-induced alterations in behavior, we investigated the effect of cocaine and U50,488 on neurogenesis in wild-type mice. NeuN expression was increased in the dorsal hippocampus of cocaine-treated wild-type mice compared with saline-treated mice (t = 3.098, df = 4; Fig. 7, A and C), whereas there was no difference in DCX expression (data not shown). Mice exposed to U50,488 also displayed increased NeuN protein expression in the dorsal hippocampus. These mice also had significantly decreased NeuN expression in the ventral hippocampus compared with those exposed to saline (t = 3.224, df = 4; Fig. 7, B and C). NeuN expression was not altered in either the dorsal or ventral sectors of the hippocampus of cocaine-treated p21−/− mice (data not shown).
Although the neurochemical mechanisms underlying the effects of cocaine are well studied, the potential role for p21 in these effects has received minimal attention. Recent data in humans describe cocaine-induced alterations in p21 expression, as well as in the expression of CDKs, targets of p21, suggesting a relationship between these proteins and drug-induced response (Hu et al., 2006; Zhou et al., 2011; Gelernter et al., 2014). To our knowledge, no studies exist, however, that directly examine the effect of p21 on cocaine-induced behaviors. This study addressed this gap in knowledge and further investigated the link between p21 and cocaine-induced effects using a p21−/− mouse model. Major findings obtained from the behavior assays performed in this study were 2-fold. First, altering p21 expression significantly diminished the response to an initial cocaine exposure, demonstrating a previously unreported relationship between p21 and drug-induced locomotor behavior. Second, the absence of functional p21 significantly enhanced either CPP in response to cocaine or CPA in response to U50,488.
Many studies concerning the actions of addictive drugs focus on the nucleus accumbens (Meyer et al., 2009; Renthal et al., 2009; Larson et al., 2010; LaPlant and Nestler, 2011; Robison and Nestler, 2011; Nestler, 2012). However, the regulatory involvement of cell cycle–related proteins such as p21 in proliferation and subsequently neurogenesis, in combination with the data presented here, suggest the importance of studies targeting the hippocampus. Not only did we observe that p21 expression was increased in response to cocaine in the ventral hippocampus of rodents, but we also saw that epigenetic alterations may be responsible. Several studies show that acute and chronic cocaine exposure alter gene expression in the limbic system via epigenetic mechanisms (Renthal et al., 2009; LaPlant and Nestler, 2011; Robison and Nestler, 2011; Wong et al., 2011; Zhou et al., 2011; Nielsen et al., 2012; Schmidt et al., 2013). Supplemental material in a study by Renthal et al. (2009) described a cocaine-induced increase in histone acetylation of p21 in the nucleus accumbens. Our data indicate that this extends to the ventral hippocampus as well. Interestingly, we found that under basal conditions, p21 expression was significantly lower in the ventral hippocampus than in the dorsal hippocampus, correlating with histone acetylation. In response to cocaine, p21 promoter histone acetylation within the ventral hippocampus rose to levels comparable to that of the dorsal hippocampus. Therefore, the increase in p21 expression observed in response to cocaine may be confined to the ventral hippocampus because of histone acetylation saturation in the dorsal sector.
A commonality among the cocaine treatments found to induce ventral hippocampal p21 expression was the occurrence of sensitization. In addition to the animals from noncontingent studies demonstrating LMS to cocaine, self-administration has also been found to augment behavioral and neurochemical responses to a challenge dose, resembling sensitization (Hooks et al., 1994). Thus, these behaviors could be related to the changes in p21 expression observed in the ventral hippocampus. In support of this, the aversive drug U50,488 induced neither LMS nor alterations in hippocampal p21 expression (Supplemental Fig. 2). However, sensitization also occurred in p21−/− animals, demonstrating that p21 is not required for this drug-induced behavioral response. These data suggest that cocaine-induced increases in p21 are not specific to species or route of administration, potentially indicating a more comprehensive effect. In support of this, we also observed increased p21 mRNA and protein expression in the ventral hippocampus of rats that had undergone a sucrose self-administration assay, demonstrating that a natural rewarding stimulus could mimic the effects of cocaine (Supplemental Figs. 4 and 5). Although the full implications of these findings remain to be determined, these results collectively indicate that not only is this effect on p21 expression in the ventral hippocampus not protocol specific, but it is also not cocaine specific.
Although the enhanced cocaine-induced CPP in the p21−/− mice could be attributable to increased salience of the drug reward, the performance differences observed with U50,488-induced CPA under the same protocol support the hypothesis that p21−/− mice have increased learning capacity, which may be mediated by increased neurogenesis. Multiple studies suggest a correlation between adult neurogenesis and learning capability (Shors et al., 2001, 2002; Raber et al., 2004; Winocur et al., 2006; Wojtowicz et al., 2008; Snyder et al., 2009a,b; Deng et al., 2010; Cameron and Glover, 2015), such as that necessary for successful CPP and CPA. Data presented here support the finding by Pechnick et al. (2011) that p21−/− mice express higher levels of both DCX and NeuN and further demonstrate that these increases are specific to the dorsal hippocampus. In wild-type mice, we note that NeuN expression is higher in the ventral hippocampus relative to the dorsal hippocampus, consistent with lower basal expression of p21 in the ventral sector. The dorsal hippocampus is primarily involved in performing cognitive functions and spatial learning and memory (Moser et al., 1995). Thus, our data confirm that heightened ongoing dorsal hippocampal neurogenesis occurs in the p21−/− mice, and our findings suggest that this may contribute to the enhanced place preference after conditioning with motivational stimuli. The differential effect of cocaine on NeuN expression in the dorsal hippocampus in the absence of increases in p21 expression itself is puzzling. A similar phenomenon was seen with U50,488. The sector specificity of the relationship between p21 and hippocampal neurogenesis is a novel finding and many questions remain. It is possible that the relatively high basal expression of NeuN in the ventral hippocampus, compared with the dorsal hippocampus, occluded any further increases. It is also possible that regulation of NeuN by p21 expression is dissociated between these two sectors. Further characterization of these relationships is the goal of future research.
The rightward shift in the dose response of cocaine-induced locomotor activity in p21−/− mice could indicate a deficiency in dopaminergic signaling. In dopamine-deficient tyrosine hydroxylase knockout mice, the serotonin receptor antagonist methiothepin maleate was shown by Hnasko et al. (2007) to inhibit CPP, suggesting that 5-hydroxytryptamine can support cocaine-induced CPP behavior in the absence of dopamine. However, the similarity in cocaine-induced CPP with and without methiothepin maleate pretreatment suggests that p21−/− mice are not compensating via a serotonergic mechanism. Instead, the blunted locomotor activity during the LMS assay, and the shifted dose response, may be attributed to altered neurogenesis. Noonan et al. (2010) showed that suppression of neurogenesis prior to drug taking resulted in a leftward shift in dose response to cocaine; therefore, the rightward shift in dose response in the p21−/− mice may be an effect of the enhanced neurogenesis. Furthermore, this rightward shift in the dose-response curve to cocaine (without a downward shift) supports the conclusion that the dopamine pathway is largely intact in p21−/− mice.
The effect of cocaine on neurogenesis remains controversial (Yamaguchi et al., 2004; Andersen et al., 2007; García-Fuster et al., 2010; Sudai et al., 2011). In the current study, cocaine-exposed wild-type mice had significantly increased NeuN expression exclusively in the dorsal hippocampus. If neurogenesis was measured in earlier reports using the hippocampus in its entirety, this sector specificity could contribute to the inconsistencies present in earlier literature. Furthermore, as with cocaine, we also observed increased NeuN expression in the dorsal hippocampus in response to conditioning with U50,488. Given the distinct pharmacological mechanisms and motivational properties of cocaine and U50,488, it is unlikely that the similar effects on NeuN are due to direct actions of the drugs themselves. Therefore, we hypothesize that the increase in NeuN in the dorsal hippocampus of wild-type mice is attributable instead to the associative learning that occurs during the behavior assays, because learning has been shown to enhance adult hippocampal neurogenesis (Gould et al., 1999). The concomitant effect of such learning is important to consider when analyzing the effect of a stimulus on neurogenesis.
The data presented here show the novel finding that loss of functional p21 alters behavioral response to cocaine in p21−/− mice. In addition, cocaine (either noncontingent or contingent exposure) was found to alter the expression of p21 in the ventral sector of the rodent hippocampus. These findings can be aligned with human data suggesting a potential relationship between CDK-related proteins and cocaine-induced behaviors (Zhou et al., 2011; Gelernter et al., 2014), further advocating the potential importance of cell cycle–related proteins in responses to drugs of abuse. Furthermore, hippocampal sector-specific cocaine-induced increases in p21 expression correlated with alterations in histone acetylation of the promoter region. Suggestive of the functional effect of this altered p21 expression, we also observed evidence of increased neurogenesis in p21−/− mice, consistent with their enhanced cognitive ability and altered dose response to cocaine. Taken together, these data further corroborate the existence of a relationship between p21 and drug-induced behaviors.
Participated in research design: Scholpa, Wagner, Cummings.
Conducted experiments: Scholpa.
Contributed new reagents or analytic tools: Wagner, Cummings.
Performed data analysis: Scholpa, Briggs.
Wrote or contributed to the writing of the manuscript: Scholpa, Wagner, Cummings.
- Received November 20, 2015.
- Accepted January 19, 2016.
This research was supported by the National Institutes of Health National Institute of Biomedical Imaging and Bioengineering [Grants EB08153 and EB0116100 (to B.S.C.)]; the National Institutes of Health National Institute on Drug Abuse [Grant DA016302 (to J.J.W.)]; the Achievement Rewards for College Scientists Foundation; the Alfred P. Sloan Foundation [Sloan Graduate Minority Fellowship]; and the Interdisciplinary Toxicology Program [Graduate Stipend (to N.E.S.)].
- bovine serum albumin
- cyclin-dependent kinase
- cyclin-dependent kinase inhibitor 1a
- conditioned place aversion
- conditioned place preference
- fixed ratio
- locomotor sensitization
- neuronal nuclei
- phosphate-buffered saline
- quantitative polymerase chain reaction
- Copyright © 2016 by The American Society for Pharmacology and Experimental Therapeutics