Oxycodone self-administration activates the mitogen-activated protein kinase/ mitogen- and stress-activated protein kinase (MAPK-MSK) signaling pathway in the rat dorsal striatum

To identify signaling pathways activated by oxycodone self-administration (SA), Sprague–Dawley rats self-administered oxycodone for 20 days using short—(ShA, 3 h) and long-access (LgA, 9 h) paradigms. Animals were euthanized 2 h after SA cessation and dorsal striata were used in post-mortem molecular analyses. LgA rats escalated their oxycodone intake and separated into lower (LgA-L) or higher (LgA-H) oxycodone takers. LgA-H rats showed increased striatal protein phosphorylation of ERK1/2 and MSK1/2. Histone H3, phosphorylated at serine 10 and acetylated at lysine 14 (H3S10pK14Ac), a MSK1/2 target, showed increased abundance only in LgA-H rats. RT-qPCR analyses revealed increased AMPA receptor subunits, GluA2 and GluA3 mRNAs, in the LgA-H rats. GluA3, but not GluA2, mRNA expression correlated positively with changes in pMSK1/2 and H3S10pK14Ac. These findings suggest that escalated oxycodone SA results in MSK1/2-dependent histone phosphorylation and increases in striatal gene expression. These observations offer potential avenues for interventions against oxycodone addiction.

The opioid epidemic remains a public health crisis 1,2 . This is related, in part, to the over-prescription of the opioid agonist, oxycodone, for pain management [3][4][5][6] . Its illicit abuse has also contributed to the high number of overdose-related deaths 7,8 . Other complications of oxycodone use disorder include moderate to severe withdrawal symptoms 1 and repeated episodes of relapses during attempts to quit through psychological or pharmacological interventions 9 . Chronic use of opioid drugs is also accompanied by cognitive deficits 10 and post-mortem evidence of neuropathological abnormalities in the brain 11 . These biopsychosocial complications make the development of effective treatment of paramount importance.
Pharmacological approaches to treat opioid use disorders (OUDs) have mainly included the use of agents that interact with opioids receptors 3,12,13 . Upon activation, opioid receptors transmit signals to the nucleus via intracellular events that involve modulation of some kinase cascades [14][15][16] , with consequent changes in gene expression 17,18 . Because of several of these experiments had done using in vitro systems, it was important to address the potential neurobiological impact of repeated oxycodone self-administration in humans by using model systems that better mimic human conditions. Therefore, we decided to use the model of escalated oxycodone self-administration (SA) in rats 19,20 to identify potential biochemical and molecular pathways that might be perturbed by this drug.
Herein, we used that model to measure alterations in various proteins that may impact the flow of intracellular signals from the mu opioid receptor consequent to its repeated interactions with oxycodone during a drug SA experiment. The rat dorsal striatum was dissected and processed for biochemical and molecular analyses because this structure is thought to play essential roles in the manifestation of habitual drug taking behaviors [21][22][23] . Thus, we report that the mitogen-activated protein kinase (MAPK)/mitogen-and stress-activated protein kinase (MSK) signaling cascade is activated preferentially in rats that consume large quantities of oxycodone over a period of 20 days. This was manifested by increased phosphorylation of extracellular signal-regulated kinases Training phase. Rats (n = 42) were randomly assigned to either saline (Sal) (n = 8) or oxycodone (n = 33) conditions. Rats were trained to self-administer oxycodone-HCL (NIDA Pharmacy, Baltimore, MD) for one 3 h daily session for the short-access (ShA) condition (n = 10) or one to three 3 h sessions for long-access (LgA) condition (n = 23) (Fig. 1A). For the LgA group, the 3 h sessions were separated by 30 min intervals from day 6 to day 20 (Fig. 1A). Lever presses were reinforced using a fixed ratio-1 with a 20-s timeout accompanied by a 5-s compound tone-light cue. We used a scheduling pattern of 5 days of drug SA and 2 days off to control for weight loss, a common side effect of oxycodone intake in laboratory animals 25 . Rats self-administered oxycodone at a dose of 0.1 mg/kg per infusion over 3.5-s (0.1 ml per infusion). The house light was turned off, and the active lever retracted at the end of the 3 h session.
Tissue collection. Rats were euthanized during early withdrawal, which is defined as the 2 h time point after cessation of drug self-administration. Dorsal striata tissue was dissected as previously described 26 . In brief, we used stereotaxic coordinates (A/P + 2 to -2 mm bregma, M/L ± 2 to 5 mm, D/V − 3 to − 6 mm) according to the rat atlas 27 and we used the position of anatomical structures (corpus callosum and lateral ventricles) for further accuracy. In brief, the dorsal striata was removed from the skulls and snap frozen on dry ice. Tissue was later used for western blotting and quantitative RT-PCR experiments.
Western blotting. Western blotting was conducted as previously described 19 . Ten-twenty µg of lysate was prepared in solutions that contained 1 × NuPage LDS Sample Buffer (ThermoFisher Scientific, Waltham, MA), and 1% β-Mercaptoethanol. Protein samples were heated to 70 °C and loaded on 3-8% Tris-Acetate Protein Gels (ThermoFisher Scientific, Waltham, MA) or NuPAGE 4-12% Bis-Tris Protein Gels (ThermoFisher Scientific, Waltham, MA). Proteins were electrophoretically transferred on the Trans-Blot Turbo System (Bio-Rad, Hercules, CA). Membrane blocking, antibody incubations, and chemiluminescence reactions were performed according to the manufacturer's instructions. Primary and secondary antibodies are listed in Supplemental Table S1. Supplemental Table S1 also includes Research Resource Identifiers (RRIDs) where antibodies were previously validated. All antibodies ran at approximate predicted sizes according to manufacturer's instructions. Cyclophilin B or alpha-tubulin was used as loading controls. Following secondary antibody incubation, ECL clarity (Bio-Rad, Hercules, CA) was used to visualize gel bands on ChemiDoc Touch Imaging System (Bio-Rad, Hercules, CA), and intensities were quantified with Image Lab version 6.0 (Bio-Rad, Hercules, CA) software.
Quantitative PCR. Total RNA was collected as previously described 19 . PCR experiments were performed using the LightCycler 480 II (Roche Diagnostics, Indianapolis, IN) with iQ SYBR Green Supermix (Bio-Rad, Hercules, CA). Primers were purchased from Johns Hopkins University (Baltimore, MD) Synthesis and Sequence Facility. Primer sequences are listed in Supplementary Table S2. The data was normalized to Qaz1 or B2m reference genes. The standard curve method was used to analyze data and the results are reported as fold change relative to Sal.
Effects of early withdrawal and oxycodone SA on the activation of PKC. Rats were euthanized 2 h after cessation of oxycodone SA and their dorsal striata were used in Western Blot analyses of several phospho-proteins involved in the MAPK/MSK signaling pathway. The kinase, PKC, is known to be involved in the MAPK signaling cascade stimulated by opioid receptors [28][29][30] . Supplementary Figure S1 shows the effects of oxycodone SA on PKC and pPKC protein expression in the dorsal striata of rats euthanized at 2 h after cessation of drug SA. There were no significant changes in striatal PKC protein levels [F (3,18) = 0.73, p = 0.540; Supplementary Fig. S1]. However, changes in phosphorylated PKC (pPKC) abundance trended towards significance Withdrawal from oxycodone SA increases ERK phosphorylation in LgA rats. ERK1/2 are members of MAPK kinases that are regulated by opioid drugs 32 and are also activated by PKC 33 . We thus measured their expression in oxycodone-exposed rats and found no significant changes [F (3,20) = 2.18, p = 0.1224] in striatal ERK1/2 protein expression ( Fig. 2A,B). However, there were increases [F (3,20) = 6.41, p = 0.0032] in the abundance of pERK1/2 in the LgA-H group in comparison to Saline and ShA groups ( Fig. 2A,C). pERK/ERK ratios were also significantly increased [F (3,20) = 7.10, p = 0.0020] in the LgA-H group in comparison to the other 3 groups (Fig. 2D). Regression analysis showed oxycodone amount-dependent increases in pERK/ERK ratios (Fig. 2E).
Increased pCREB in LgA-H rats. Because CREB phosphorylation can be mediated by several upstream kinases that include PKC, ERK1/2 and MSKs 37 , we examined the possibility that activation of these kinases might have led to increased pCREB after oxycodone SA. There were no significant changes [F( 3, 18) = 2.97, p = 0.0592] in CREB protein expression (Fig. 4A,B). However, the abundance of pCREB was significantly increased [F( 3, 20) = 36.0, p < 0.001] in the LgA-H groups in comparison to the Saline group (Fig. 4A,C). Furthermore, pCREB/CREB ratios were also substantially increased [F (3,18) = 24.8, p < 0.001] in the LgA-H group compared to other groups (Fig. 4D), with there being a significant positive correlation between these ratios and amount of oxycodone taken (Fig. 4E).
Oxycodone SA induces increased phosphoacetylation of histone H3 in LgA rats. In addition to CREB phosphorylation, activated MSK1 and MSK2 participate in the phosphorylation of histone H3 at serine residue S10 and can cause increases in histone H3 phosphoacetylation of H3S10pK14Ac 34,37,38 . We therefore sought to determine the effects of oxycodone SA on the abundance of this histone marker. Figure 5 shows the results for histone H3 and H3S10pK14Ac. Unexpectedly, we found significant decreases [F (3,20) = 5.02, p = 0.0094] in histone H3 protein levels in the ShA group in comparison with Sal and the LgA-H groups (Fig. 5A,B). In contrast, H3S10pK14Ac abundance was significantly increased [F (3,20) = 20.0, p < 0.0001] in the LgA-L and LgA-H groups compared with Saline rats. Moreover, H3S10pK14Ac in the LgA-H group was substantially increased compared to the other 3 groups (Fig. 5A,C). H3S10pK14Ac/H3 ratios were also increased [F (3,20) = 20.0, p < 0.0001] in the LgA-L and LgA-H groups compared to the Saline and ShA groups (Fig. 5D). Regression analyses revealed that both H3S10pK14Ac abundance and H3S10pK14Ac/H3 ratios positively correlated with the amount of total oxycodone taken during the experiment (Fig. 5E, F).
Differential protein expression of CBP and H3K27Ac in oxycodone exposed rats. Phosphorylated CREB recruits CBP, under certain circumstances, to promote changes in gene expression 39,40 . We thus tested the possibility that oxycodone SA might have influenced striatal CBP protein expression and found that CBP protein expression was significantly increased [F (3,17) = 6.82, p = 0.0032] in the LgA-L and LgA-H rats compared with the Saline and ShA groups (Fig. 6A,B). Increased CBP expression correlated with the amount of oxycodone consumed by the rats (Fig. 6C).
In addition to CBP's transcriptional co-activity, it functions as a histone acetyltransferase 41,42 that mediates acetylation of H3K27 43-45 , a marker of active enhancers 46-48 that is involved in regulating neuronal gene expression. We therefore measured the abundance of H3K27Ac, which had been previously shown to be impacted in the brains of heroin addicts 49 . We found significant increases [F (3,20) = 4.53, p = 0.0140] in H3K27Ac abundance in all oxycodone groups, including ShA rats that did not escalate their intake (Fig. 6D,E), suggesting that oxycodone exposure is enough to increase striatal H3K27Ac abundance. We found that H3K27Ac abundance positively correlated with the amount of oxycodone self-administered by rats (Fig. 6F). These increases confirm the data in heroin-using individuals 49 . H3K27Ac/H3 ratios were also significantly increased [F (3,20) = 3.82, p = 0.0259] in the 3 oxycodone groups (Fig. 6G) and correlated with the amount of oxycodone taken (Fig. 6H).
Because CBP expression was only increased in the two LgA groups while H3K27Ac was increased in the 3 oxycodone groups, we sought to determine if the expression of another histone H3 acetyltransferase, Tip60, with putative activity towards the lysine 27 residue 50 was affected in the 3 oxycodone groups. Indeed, Tip60 protein expression was significantly increased [F (3,20) = 3.63, p = 0.0307] in all three oxycodone groups ( Supplementary  Fig. S3).
Oxycodone SA increases GluA2 and GluA3 glutamate receptor mRNA levels. Changes in gene expression in response to exogenous stimuli include many target genes that are expressed with different time courses of induction. Because respective changes in histone phosphorylation and acetylation generated by MSKs and CBP are known regulators of gene expression 51-54 , we tested the idea that some of their target genes might be affected in the striata of oxycodone-exposed rats. We also measured the mRNA levels of some glutamatergic genes whose expression was altered in the brains of heroin users based on a previous microarray study 49 . Figure 7 shows the effects of oxycodone SA on the mRNA expression of GluA1, GluA2, GluA3, and GluA4 subunits of AMPA receptors 55 . We found no substantial changes in GluA1 [F (3,29)  www.nature.com/scientificreports/ oxycodone consumed during the experiment (Fig. 7B,H). In contrast, striatal GluA2 [F (3,27) = 3.49, p = 0.0291; Fig. 7C] and GluA3 [F (3,31) = 15.7, p < 0.0001; Fig. 7E] were increased in the LgA-H group compared to other groups. Regression analyses revealed significant oxycodone amount-dependent changes in their mRNA levels (Fig. 7D,F). Moreover, the changes in GluA2 and GluA3 mRNAs correlated with changes in pMSK1 (Fig. 8A,B). However, only changes in GluA3, but not in GluA2, mRNA levels correlated with changes in pMSK2, H3S10p-K14Ac, and CBP protein expression (Fig. 8C-H).  Fig. S4). However, we observed no significant correlations between Ncoa1, Ncoa2, and Ncoa3 mRNA levels and CBP suggesting that histone phosphorylation might play a more important role in regulating their expression than histone acetylation ( Supplementary Fig. S4).

Discussion
In the present study, we show that long-access to oxycodone SA over a period of 20 days leads to activation of several kinases involved in the MAPK/MSK signaling pathway with consequent CREB and histone H3 phosphorylation in the rat dorsal striatum. These results are consistent, in part, with previous evidence of the involvement of the MAPK in the biochemical effects of morphine 56,57 . We also found significant increases in CBP and H3K27 acetylation in oxycodone-exposed rats. These findings are consistent with observations that epigenetic mechanisms are involved in models of opioid abuse 58 . The changes in signaling pathways are accompanied by oxycodone-induced increased gene expression of GluA2 and GluA3 subunits of AMPA receptors and of the acetyltransferases, Ncoa1-3. Our results provide novel insights into the role of H3S10pK14Ac in oxycodoneinduced gene expression and hint to a model whereby this histone marker is involved in the regulation of genes that might be responsible for some long-term molecular adaptations that drive compulsive oxycodone intake. The dorsal striatum is a brain region that is integral to various behavioral changes consequent to drug taking behaviors including habit forming and drug seeking during periods of opioid withdrawal 19,22,23,59 . Similar to other observations with cocaine, methamphetamine, and other drugs 58,60-66 , we found increased phosphorylation of PKC, ERK1/2, MSK1/2, and pCREB in the rat dorsal striatum after repeated exposure to oxycodone SA. Increased phosphorylation of H3S10pK14Ac, a marker that is downstream of these kinases 34,37,67 is of interest because these findings suggest that repeated exposure to long-access oxycodone self-administration might engender a permissive molecular state characterized by increased histone H3 phosphoacetylation and a more open chromatin structure. The hypothesized permissive state might also facilitate pCREB binding at the cAMP-Response Element (CRE) on the promoters of genes that have been implicated in the regulation of synaptic plasticity 51,68,69 . CREB activation is also known to enhance the recruitment of co-activators 39,40 such as CBP, an acetyltransferase that acetylates H3K27 [43][44][45]70 to increase the transcription of downstream genes in diverse cell populations 39,71,72 . This suggestion is further supported by observations of increased expression of co-activators for the steroid hormone receptor family, Ncoa1 73 , Ncoa2 74 , and Ncoa3 75 , that can enhance transcription, in part, via histone acetylation 76,77 and recruitment of CBP 75,78,79 , whose expression is also increased in rats exposed to relatively large quantities of oxycodone. Our proposal of an oxycodone-induced permissive state in the dorsal striatum is consistent with observations that histone H3 phosphorylation and acetylation can work in concert to regulate gene expression 80 . This discussion is supported by the observations that histone H3 phosphoacetylation also participates in heroin-induced conditioned place preference 81 , thus indicating a role of phosphoacetylation in the effects of opioids in general. His discussion notwithstanding, given the diversity of neurons within the dorsal striata 82 , follow-up studies are needed to identify which specific neuronal subtypes might exhibit the oxycodone-associated changes in that structure.
Our observations of increased histone H3 phosphorylation and acetylation led us to test the possibility that some genes downstream of these molecular events might show differential expression in the brains of  Figure S5. Key to statistics: *, **, *** = p < 0.05, 0.01, 0.001, respectively, in comparison to Sal rats; #, ##, ### = p < 0.05, 0.01, 0.001, respectively, in comparison to SHA rats; $$ = p < 0.01 in comparison to LgA-L rats. Statistical analyses are as described in Fig. 2 www.nature.com/scientificreports/ oxycodone-exposed rats. Indeed, we found increased expression of several genes in the LgA-H rats that showed increased abundance of striatal pMSK1, pMSK2, and histone H3S10pK14Ac. Of interest among those are the changes in AMPA receptor subunits, GluA2 and GluA3, in LgA-H rats in an oxycodone amount-and pMSK1dependent fashion. MSK1 and MSK2 are known to play substantial roles in a number of biological events including synaptic plasticity 83 . The altered expression of GluA2 and GluA3 is of singular interest because Egervari et al. (2017) had reported that their microarray analyses, using tissues from the ventral striatum of heroin users, had detected changes in the expression of several genes, including GluA3, which are involved in glutamate neurotransmission. Increases in the expression of GluA2 and GluA3 receptor subunits in our study are consistent with the proposed roles of glutamate receptors in models of substance use disorders including cocaine 84 , methamphetamine 85 , and opioids 86,87 . For example, chronic cocaine increases GluA2 expression in the nucleus accumbens and increased expression of GluA2 via viral injections enhanced the sensitivity of mice to the behavioral effects of cocaine 84 , thus suggesting that increased GluA2 expression in the present study might have served to facilitate escalation of oxycodone intake in the LgA-H rats. A similar argument could be made for our novel findings of increased GluA3 expression after oxycodone SA. GluA3-containing AMPA receptors are located in various brain regions 88,89 . Because GluA3 exists in GluA2A3 combinations or GluA3 monomers or dimers 90 , it is possible that increased expression of both GluA2 and GluA3 might potentiate AMPAR-mediated changes in synaptic plasticity during repeated oxycodone exposure. Alternatively, GluA3 alone may regulate oxycodone intake because GluA3 knockout mice show decreased alcohol intake 91 . Thus, elucidation of the specific roles that GluA3 alone or in combination with GluA2 play in oxycodone SA will await future genetic and pharmacological studies.
In conclusion, we have demonstrated that rats that self-administer large quantities of oxycodone showed increased histone and CREB phosphorylation via activation of the MAPK/MSK phosphorylation signaling pathway in the rat dorsal striatum. Rats exposed to large quantities of oxycodone also showed increased striatal CBP and histone acetylation in oxycodone-exposed rats. Changes in histone modifications are proposed to lead to  Figure S5. Key to statistics: *, ** = p < 0.05, 0.01, respectively, in comparison to Sal rats; #, ## = p < 0.05, 0.01, respectively, in comparison to SHA rats; $$ = p < 0.01 in comparison to LgA-L rats. Statistical analyses are as described in Fig. 2 www.nature.com/scientificreports/ more permissive chromatin states that promoted changes in the expression in a diversity of classes of genes as exemplified by increased mRNA levels of acetyltransferases, Ncoa1-3, and of AMPA receptor subunits, GluA2 and GluA3, in an oxycodone amount-dependent fashion. Importantly, changes in the expression of both GluA2 and GluA3 mRNA levels correlated with altered abundance of pMSK1, a kinase that is involved in the regulation of synaptic plasticity 92 . These observations are illustrated schematically in Fig. 9. Furthermore, additional experiments that include genetic manipulations and pharmacological interventions that impact these pathways are necessary to identify the specific role that these transcriptional and post-transcriptional changes might play in promoting oxycodone self-administration or cessation of drug intake. Finally, confirmation of the involvement of MAPK/MSK/histone phosphorylation in oxycodone SA may serve as a stimulus to develop potential pharmacological agents against oxycodone use disorder. . Increased MSK phosphorylation is accompanied by increased histone H3 phosphoacetylation and CREB phosphorylation. In addition, there were increases in the protein expression of two acetyltransferases, CBP and Tip60 that acetylate H3K27. Recruitment of CBP by CREB and histone modifications serve to create a permissible transcriptional environment that led to increased mRNA levels of GluA2 and GluA3 in a MSK1dependent fashion. This kinase-histone modification cascade may serve as targets for therapeutic interventions against oxycodone use disorder. www.nature.com/scientificreports/