Transcriptome Sequencing Reveals Candidate NF-κB Target Genes Involved in Repeated Cocaine Administration

Transcriptome Sequencing Reveals Candidate NF-κB Target Genes Involved in Repeated Cocaine... Background: Drug-induced alterations in gene expression play an important role in the development of addictive behavior. Numerous transcription factors have been implicated in mediating the gene expression changes that occur in drug addiction. Nuclear factor kappa B is an inducible transcription factor complex that is rapidly activated by diverse stimuli. Methods: We performed next-generation high-throughput sequencing of the prefrontal cortex in a mouse model of repeated cocaine administration combined with pharmacological nuclear factor kappa B inhibition to identify nuclear factor kappa B target genes that participate in the cocaine addiction process. Results: We found that the nuclear factor kappa B antagonist sodium diethyldithiocarbamate trihydrate significantly reversed the cocaine-induced expression changes of the amphetamine addiction pathway. Genes that demonstrated differential expression in response to cocaine treatment that was also reversed by sodium diethyldithiocarbamate trihydrate were enriched for the axon guidance pathway. Furthermore, the nuclear factor kappa B homo-dimer motif could be mapped to 86 of these sodium diethyldithiocarbamate trihydrate-reversed genes, which were also enriched for axon guidance. Conclusions: We suggest that nuclear factor kappa B directly modifies the expression of axon guidance pathway members, leading to cocaine sensitization. Our findings reveal the role of prefrontal cortex nuclear factor kappa B activity in addiction and uncover the molecular mechanisms by which nuclear factor kappa B drives changes in the addicted brain. Keywords: addiction, cocaine, nuclear factor kappa B, prefrontal cortex, sodium diethyldithiocarbamate trihydrate Introduction Drug addiction is a global issue that generates extensive harm alter the structure and function of numerous and distinct brain to the health and socioeconomic status of affected individuals regions, at least one of the major constinuents of the mesolim - and inflicts further burdens on society. Chronic addictive behav - bic reward pathway—the ventral tegmental area, the nucleus ior develops as a consequence of progressive and multifaceted accumbens, and the prefrontal cortex (PFC)—is modified by alterations to neural circuits upon repeated administration of all addictive substances ( Levy, 2013). In the context of cocaine drugs of abuse (Chao and Nestler, 2004 ). Though various drugs addiction, it has been demonstrated that the PFC is involved in Received: January 14, 2018; Revised: February 28, 2018; Accepted: March 19, 2018 © The Author(s) 2018. Published by Oxford University Press on behalf of CINP. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http:// creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, 697 provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 698 | International Journal of Neuropsychopharmacology, 2018 Significance Statement Despite a growing number of treatments, drug addiction continues to be a source of tremendous physical, psychological, social, and economic harm. Understanding the molecular mechanisms by which the brain is altered during the onset of addiction can lead to the development of more effective therapies. It is widely accepted that chronic drug exposure results in changes in gene expression. Here, we investigate the role of the transcription factor NF- κB in mediating these gene expression changes in the prefrontal cortex. Our findings demonstrate the role of NF- κB activity in the prefrontal cortrex in the establishment of cocaine addiction and uncover target genes and biological pathways through which this process occurs. both the primary reinforcing effects of cocaine and long-term through which NF-κB may mediate cocaine-induced cellular and sensitization to cocaine. Previous studies have further demon - behavioral plasticities. To do so, we examined the effects of an strated that PFC-dependent processes, such as executive func - NF-κB inhibitor on the behavioral response to repeated cocaine tion, explicit learning, and memory, are damaged in animal administration and combined our model with high-throughput models of cocaine addiction and in human cocaine abusers transcriptome sequencing to identify target genes activated by (Chao et  al., 2014). However, the molecular mechanisms driv- NF-κB in the PFC upon cocaine treatment. We demonstrate the ing altered PFC biology in response to cocaine addiction remain efficacy of pharmacological NF-κB inhibition in counteracting poorly understood, and, as a result, there continues to be a lack the effects of cocaine and, by establishing the genes and path - of effective therapies for addicted patients. ways modulated by NF-κB in response to cocaine, we reveal Though drug-induced changes in the expression of numer - novel targets for further therapeutic intervention. ous neuronal genes contribute to behavioral abnormalities, the origin and causality of these genetic signatures are unclear Materials and Methods (Zhou et  al., 2014). Consistent changes in characteristic path - ways suggest the action of specific transcription factors, which Animals respond to environmental changes by binding to regulatory regions of target genes to alter their expression ( Baldwin, 1996). Male C57BL/6J mice (purchased from the Academy of Military Medical Science) weighing 25 to 30 g (8 weeks old) at the beginning In the past several decades, numerous transcription factors, such as Fos family protein, cyclic adenosine monophosphate of the experiments were used in this study. Five mice were housed per cage and all mice were given food and water ad libitum. The response element binding protein, nuclear factor kappa B (NF- κB), and myocyte enhancing factor-2, have been implicated in colony was maintained at 22  ±  2°C with a standard 12-h-light/- dark cycle (lights on at 7:00 am). The animals were handled daily the addiction process ( Nestler, 2012 ). NF-κB represents a small family of inducible transcription for 1 week before injection. Testing was performed during the light cycle. All animal protocols were approved by the Review Board of factors that are rapidly activated by diverse external stimuli and therefore serve as critical mediators of many environmen - the Institute of Psychology, Chinese Academy of Sciences, and were performed strictly in accordance with the Guideline for Care and tal inputs ( Nestler, 2012 ). The NF-κB complex is typically held in an inactive form in the cytosol by interaction with members of Use of Laboratory Animals of the Chinese Academy of Sciences. the inhibitor of NF-κB (IκB) family of inhibitory proteins. Upon stimulation by extracellular inducers,κ I Bs are phosphoryl - Drugs ated by an IκB kinase complex, which leads to ubiquitination and proteosome-mediated degradation of the IκBs. Then, active Cocaine hydrochloride (Qinghai Pharmaceutical Co., Ltd.) and NF-κB, through classical or noncanonical pathways, translo - sodium diethyldithiocarbamate trihydrate (DDTC; NF- κB antag - cates to the nucleus, where it binds consensus κB sequences in onist, Sigma-Aldrich) were dissolved in sterile 0.9% saline and the promoter and enhancer regions of responsive genes ( Widera administered i.p. at 20 and 200 mg/kg, respectively. et  al., 2006). NF-κB binding sites are found in numerous genes that are known to participate in memory formation and drug Locomotor Sensitization Test addiction: brain-derived growth factor, inducible nitric oxide synthase, and various opioid receptors ( Simeonidis et al., 1999; Mice were placed in the center of the field (30 x 30 cm), and total Malek et al., 2001 ). However, these examples represent a small distance travelled by each animal was recorded for 60 minutes number of NF-κB targets, and there is no evidence whether and analyzed by a computer-based system (Anilab). After each these and other genes are regulated by NF- κB in the PFC during trial, the apparatus was cleaned with a 30% ethanol solution. cocaine addiction or whether they play a causal role in addiction The order of testing was balanced by group. onset and persistence. It had been previously demonstrated that the NF- κB signal- Treatment and Behavioral Testing Procedures ing pathway, including several candidate downstream genes, is Mice were randomly assigned to 4 treatment groups: (1) activated in the nucleus accumbens of mice exposed to chronic cocaine and inhibition of this pathway by a virally expressed saline + saline (SS, n = 10), (2) DDTC + saline (DS, n = 10), (3) DDTC + cocaine (DC, n = 10), and (4) cocaine + saline (CS, n = 10). dominant-negative form of IκB kinase attenuates the addictive effects of cocaine (Russo et al., 2009 ). However, in the context of The treatment period was 8 days, as shown in Figure 1. On days 1 to 3, baseline locomoter activity of every treatment group cocaine addiction, the importance of NF- κB signaling in the PFC has not been investigated, and the great genome-wide diversity was assessed for 1 hour immediately following an i.p. injection of saline. On days 4 to 8, groups (2) and (3) received an of NF-κB target genes has not been explored. The major goals of this study were to determine whether injection of saline or cocaine, respectively, 30 minutes after an injection of DDTC, while groups (1) and (4) also received an injec - NF-κB signaling in the PFC is important in the development of cocaine addiction and to identify the potential target genes tion of saline or cocaine, respectively, but 30 minutes following Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup Wang et al. | 699 (Tripathi et al., 2015 ) with a cutoff of P < .05 and a minimum over - lap of 3. These raw sequencing data sets were deposited in the Gene Expression Omnibus of the NCBI ( https://www.ncbi.nlm. nih.gov/geo) under accession number GSE108836. Real-Time PCR Validation Validation of DEGs was performed by quantitative reverse-tr- an scription PCR (RT-qPCR) using the Maxima SYBR Green qPCR Master Mix kit (Fermentas), according to the manufacturer’s instructions, in an ABI Prism 7500 Sequence Detection System machine (Applied Biosystems Inc). All real-time RT-qPCR data were normalized to SS expression level (see additional data file for primer information). Figure 1. Effect of behavior of sodium diethyldithiocarbamate trihydrate (DDTC; 200  mg/kg) in response to cocaine-addicted mice (20  mg/kg, i.p.). CS, saline+cocaine; DC, DDTC+cocaine; ds, DDTC+saline; SS, saline+saline. All Motif Search results represent means ± SD of 10 independent determinations (n = 10/group) in the behavioral test experiments. **** P < .0001, ***P < .001. To discover DEGs with NF-κB binding sites, we applied the UniProbe Database ( http://thebrain.bwh.harvard.edu/nfkb/ ) to an injection of saline. All mice were put into the locomotor box promoters (-500 to +200 bp relative to the transcription start site) immediately following the second injection and monitored for of DEGs. The z-score cutoff was set to the default value of 4. 1 hour. Data Analysis RNA Sequencing and Bioinformatics Data are expressed as the mean ± SEM. Statistical analyses were Two hours after the final (day 8) locomotor activity test, animals performed with GraphPad Prism 6.0 software. Comparisons were killed and their PFCs were surgically excised. The tissue was of means of behavioral sensitization were analyzed by 2-way stored in liquid nitrogen immediately and then transferred to a repeated-measures ANOVA followed by Tukey’s posthoc tests. -80°C freezer. Total RNA was extracted from the frozen tissues using the TIANamp DNA/RNA Isolation Kit (TIANGEN), including Results additional treatment with RNase-free DNase I  (Ambion) for 30 minutes at 37°C to remove contaminating DNA. Effects of DDTC on Cocaine-Induced Behavior Then 1  µg of RNA from a pool of 5 animals (200  ng each) per experimental condition was used to generate an mRNA First, we analyzed the effects of the injection of (1) saline + saline sequencing library using the TruSeq RNA Sample Prep Kit (SS), (2) DDTC + saline (DS), (3) DDTC + cocaine (DC), and (4) (Illumina) following the manufacturer’s protocol. Briefly, polyA cocaine + saline (CS), repeated daily for 5  days, on mouse loco - mRNA was purified and fragmented. Then, first and second motor activity after 3  days of baseline testing. As shown in strand cDNA synthesis was performed, ends were repaired and Figure  1, 2-way repeated-measures ANOVA of the locomotion adenylated, adapters were ligated, and fragments were enriched test results demonstrated that there were significant differences with PCR amplification. High library qualities were verified with in the total distance traveled between the 4 treatment groups an Agilent Technologies 2100 Bioanalyzer (Agilent Technologies). (F = 44.34, P < .001). Posthoc analysis revealed that there was (3,27) Libraries were sequenced using the Illumina HiSEQ2K platform no significant difference between the locomotor activities of (Illumina). each group at baseline (days 1–3). However, total distance tra - ve FastQC was used to assess RNA-seq quality. Because of the led increased significantly following cocaine treatment (SS vs low quality of the 3’ end of raw reads, the last 20 base-pairs were CS P < .001) and DDTC administration significantly inhibited trimmed from each read. Reads were also trimmed sequen - cocaine-induced activity (CS vs DC P < .001). Additionally, there tially for adapter content. The paired-end reads were aligned was no significant difference between the locomotor activi - to the mouse reference genome (mm9, Jul. 2007)  with TopHat ties of DC and SS groups on Day 5 onwards, suggesting that 2.0.9 (Trapnell et  al., 2012 ) using default parameters. Aligned DDTC can completely inhibit the hyperlocomotion induced by bam files were processed by Cufflinks v.2.2.1 ( Trapnell et  al., repeated cocaine administration. 2010) to generate a transcriptome assembly and to estimate the expression level (fragments per kilobase of transcript per mil - GO and KEGG Pathways Induced by Cocaine lion mapped reads [FPKM]) of all detected isoforms. FPKM was calculated as the number of paired-end reads (a single fragment To understand the molecular mechanisms underlying the effects per end-paired reads) mapped to a gene divided by the number of cocaine and to delineate the functional consequences of DDTC of all fragments mapped to the genome (in millions) and the treatment in the context of cocaine addiction, we performed length of the transcript (in kilobases). The tagwise dispersion mRNA-sequencing of extracted PFCs of mice from each of the 4 was estimated and then used for log fold change calculation. treatment groups, immediately following the 8th and final day Differentially expressed genes (DEGs) were identified by apply- of treatment and testing. A total of approximately 60 million 100 ing the threshold false discovery rate of 0.2 to adjusted P values base-pair paired-end reads were obtained per sample. following Fisher’s exact test Chen ( et  al., 2011). DEGs induced Of 266 DEGs, 175 genes were upregulated and 91 genes were by cocaine sensitization (logfold_change [CS/SS]) were plotted downregulated in the CS group compared with the SS group against baseline gene expression levels of the SS group (lo F- g (Figure  2A). We then analyzed whether certain (KEGG) path - PKM) using ggplot2 in R.  KEGG pathway, and Gene Ontology ways were enriched for in the list of PFC DEGs and discovered (GO) analyses of DEGs were performed with Metascape software 13 significantly enriched pathways upon repeated cocaine Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 700 | International Journal of Neuropsychopharmacology, 2018 Figure  2. Cocaine and sodium diethyldithiocarbamate trihydrate (DDTC) cause broad changes to the transcriptome. mRNA from prefrontal cortex (PFC) of SS (saline+saline), CS (saline+cocaine), and DS (DDTC+saline) was subjected to RNA sequencing and analysis as described in the methods. (A) MA-plot [M (log ratios); A (mean average)] of differentially expressed genes (DEGs) of cocaine+saline (CS), respectively < (q 0.2). DEGs are represented by green dots. Log2 fold change values for CS vs SS are plotted against average log expression values (fragments per kilobase of transcript per million mapped reads [FPKM]). (B) Chart of enriched pathways of significantly expressed genes in CS ( P < .05). (C) MA-plot of DEGs of DDTC+saline (DS), respectively (q < 0.2). DEGs are represented by green dots. Log2 fold change values for DS vs SS are plotted against average log expression values (FPKM). (D) Chart of enriched pathways of significantly expressed genes in DS ( P < .05). administration. Among these, 3 pathways were directly related GO and KEGG Analyses Reveal the Effects of to central nervous system function: most notably, the amphet - DDTC on Pathways Altered by Repeated Cocaine amine addiction pathway, as well as the circadian rhythm Administration pathway and amyotrophic lateral sclerosis (ALS) ( Figure  2B). Next, we applied the same analyses to study the effects of com - Additional GO analysis revealed multiple terms relevant to bining cocaine with DDTC treatment. As a major goal of this neurological function ( supplimentary Figure 1 ), including regu - study was to identify the target genes by which NF- κB mediates lation of circadian rhythm (GO:0042752), synaptic vesicles the develoment of cocaine addiction, we sought to find genes (GO:0008021), and MAPK function (GO:0043408 and GO:0017017), whose expression was altered by repeated cocaine treatment which were reported to be involved in the control of synaptic with the condition that this differential expression is abolished plasticity ( Thomas and Huganir, 2004). upon NF-κB inhibition by DDTC treatment. In the process of identifying these genes, the next step was to perform a differ - GO and KEGG Pathways Induced by DDTC ential expression analysis between DC and CS groups to obtain a list of DEGs whose expression is affected by DDTC treatment We simultaneously analyzed how DDTC treatment modi - in the context of cocaine treatment. A  total 451 DEGs were fies the mouse transcriptome ( Figure  2C). A  total 143 DEGs upregulated and 410 DEGs were downregulated in DC-treated were upregulated while 58 DEGs were downregulated in DS mice compared with CS-treated mice (Figure 3A). The top 20 sig- compared with SS. KEGG pathway analysis of all DS DEGs nificantly enriched pathways showed that DDTC can regulate revealed a stastically significant enrichment for 11 pathways pathways known to be related to addiction, such as ampheta - (Figure 2D). Interestingly, the MAPK signaling pathway, the cir - mine, calcium signaling, ABC transporters, gap junctions, glu - cadian rhythm (entrainment) pathway, and the focal adhesion tathione metabolism, ALS, glycolysis, Alzheimer’s disease, and pathway, which were significantly enriched in cocaine treat - neuroactive ligand-receptor interaction ( Li et  al., 2008), as well ment (CS vs SS), were also affected by DDTC treatment, indi - as other pathways related to neuronal function such as cir - ca cating that cocaine and DDTC modulate several of the same dian entrainment, neuotrophin signaling, and axon guidance molecular pathways in the PFC. GO analysis again revealed (Figure  3B). GO analysis (supplementary Figure  3) showed that numerous enriched pathways related to neuronal function, the terms neuron projection development (GO:0031175), behav - including MAPKs (GO:0017017), synapse (GO:0045202), neuro - ior (GO:0007610), calcium channel activity (GO:0005262), synapse transmitter transport (GO:0006836), central nervous system (GO:0045202), and axon (GO:0030424) were enriched. development (GO:0007417), learning or memory (GO:0007611), We were then interested in isolating pathways from DEGs and neuron projection development (GO:0031175) ( supplemen - that met 3 requirements: (1) are altered by cocaine alone tary Figure 2 ). Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup Wang et al. | 701 Figure 3. Transcriptome-wide response of sodium diethyldithiocarbamate trihydrate (DDTC) to cocaine addiction. (A) MA-plot of differentially expressed genes (DEGs) of DDTC+cocaine (DC), respectively (q < 0.2). DEGs are represented by green dots. Log2 fold change values for DC vs saline+cocaine (CS) are plotted against average log expression values (fragments per kilobase of transcript per million mapped reads [FPKM]). (B) Chart of enriched pathways of significantly expressed genes in DS ( P < .05). (C) Venn diagram depicting the overlap of DEGs among DC, DDTC+saline (DS), and CS. (D) Heatmap of comparison of enriched pathways of CS, DS, and DC. (E) Scatter plot of all genes in amphetamine addiction pathway. (Figure 2A); (2) are modulated by DDTC in cocaine-treated (DC) Interestingly, scatter plots ( Figure  3E; supplementary mice ( Figure  3A); and (3) result from DDTC treatment alone Figure  5) showed that DDTC (DC) significantly reversed the (DS) (Figure  2C). These DEGs and pathways would represent direction of cocaine-induced (CS) gene expression changes of changes in cocaine-sentization that are specifically caused by the amphetamine addiction pathway ( P = .0498) but not of the DDTC in DC-treated mice rather than representing a combined ALS pathway ( P = .078), indicating that the amphetamine addic - effect of DDTC and cocaine coadministration. The 3 requisite tion pathway is an important pathway by which DDTC modu - DEG data sets are plotted in a Venn diagram ( Figure 3C; http:// lates gene expression changes in response to repeated cocaine bioinfogp.cnb.csic.es/tools/venny/ ), which interestingly depicts administration. a significant overlap between all 3 data sets, between DC and DS DEGs as well as DC and CS DEGs, but not between CS and DS DDTC Reverses Cocaine-Induced Changes in DEGs, which are not present in the DC group (chi-squared test). Expression of Genes in the Axon Guidance Pathway Next, we performed a 3-way statistical comparison of enriched pathways and GO terms from DEGs induced in CS, DC, and DS While the previous analyses identified common pathways treatments ( Figure  3D). The KEGG pathways common to all 3 by which DDTC can reverse changes seen upon repeated DEG groups were fluid shear stress and atherosclerosis, focal cocaine treatment, pathways consist of many genes whose adhesion, MAPK signaling, and circadian entrainment. products exert contradictory functions on pathway activity, However, common pathways significantly enriched in both obscuring conclusions as to whether DDTC clearly reversed cocaine treatment alone (CS) and DDTC+cocaine (DC) treatment cocaine-induced behavior of these pathways. Therefore, we DEGs included pathways in cancer, ALS, and amphetamine wished to determine if DDTC is able to antagonize cocaine- addiction, and several common GO iterms were related to neu - related expression changes of specific genes. We plotted the ral signal conduction supplementar ( y Figure 4 ,) such as calcium log fold change of DEGs from DC, CS, and DS conditions in channel activity (GO:0005262), neurotransmitter receptor acti - v a heatmap, with which it was immediately visible that many ity (GO:0030594), and presynapse (GO:0098793). more genes exhibited fold changes in opposite directions in Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 702 | International Journal of Neuropsychopharmacology, 2018 the DC mice compared with CS and DS mice, an observation Candidate NF-κB Target Genes Involved in Response that was confirmed by hierarchial clustering ( Figure  4A). In to Repeated Cocaine Administration fact, the log fold changes of DEGs in DC and CS displayed -15 2 NF-κB is a protein complex composed of homo- or heterodimers a significantly negative relationship ( P = 2.994e , R = 0.2647), of 5 different family members: c-Rel (Rel), RelA/p65 (Rela), RelB while those of DS and CS had a significantly positive rela - -16 2 (Relb), NF-κB /p50/p105 (Nfkb1), and NF-κB /p52/p100 (Nfkb2) tionship (P = 2.2e , R = 0.4723) (Figure  4B; supplementary 1 2 (Natoli, 2006). Studies of protein-DNA crystal structures and Figure  5). These results indicate that DDTC reverses specific DNA-binding studies have shown that NF- κB and NF-κB recog- gene expression changes that occur in response to cocaine 1 2 nize the 5 base-pair motif of 5’-GGGRN-3’, whereas c-Rel, RelA/ treatment. p65, and RelB recognize a 4 base-pair 5’-GGRR-3’ motif, and the We then performed an additional pathway analysis to entire κB family recognizes a concensus motif of 5’-GGGRN(Y) determine which pathways were enriched in the list of YYCC-3’ (Hoffmann et  al., 2006; Bracchi-Ricard et  al., 2007 ). genes whose cocaine-induced expression changes were spe- Using this information, we applied online tools (see Methods) cifically reversed by DDTC (i.e., fold change in the opposite to scan the promoter regions of DEGs for NF- κB family bind- direction). There were a total of 142 such genes, 58 of which ing sites. The NF-κB motif could be mapped to the promoter were upregulated and 84 of which were downregulated by regions 86 DEGs (supplementary Table 3 ). This gene list was also cocaine, while these changes were reversed by DDTC ( supple- enriched in the axon guidance term (mmu04360, P = .089), sup- mentary Table  2). KEGG analysis revealed that the most sig - porting the hypotheses that NF- κB directly regulates the axon nificantly enriched pathway was axon guidance (mmu04360, guidance pathway in cocaine addiction and that DDTC inter - P = .0088). Interestingly, when we analyzed the lo gfold feres with this activity. changes of all DEGs in the axon guidance pathway posthoc, that is, in a less stringent manner, not only were the DEGs with significant cocaine-induced expression changes signifi - Validation of Genes Reversed by DDTC cantly reversed by DDTC but there was a significantly neg-a -08 tive relationship between CS and DC groups ( P = 4.211e ), The top-ranked DEGs by fold-change expression were Epha8 (Ephrin Type-A Receptor 8), a member of the axon guidance thereby further demonstrating that DDTC can reverse the axon guidance behavior that is modified by cocaine treatment pathway, and Dpf3 (Double PHD Finger 3), which belongs to the neuron-specific chromatin remodeling complex involved in (Figure 4C). Figure  4. Sodium diethyldithiocarbamate trihydrate (DDTC) reverses genes regulated by cocaine addiction. (A) Heatmap of all significantly differentially expressed genes in each group (saline+cocaine [CS], DDTC+saline [DS], and DDTC+cocaine [DC]). Red, upregulated; blue, downregulated. (B) Scatter plot of all differentially expressed genes of log2 fold_change of CS (X axis) and log2 fold_change of DC (Y axis). (C) Scatter plot of all genes in axon guidance pathway. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup Wang et al. | 703 Figure 5. Expression of cocaine and sodium diethyldithiocarbamate trihydrate (DDTC) target genes as determined by qPCR. (A) Normalized mRNA expression quanti - fied by RT-qPCR for genes whose expression is altered by cocaine treatment and reversed by DDTC: Epha8 and Dpf3. Results represent mean ± SD of 3 determinations in each of the 2 types of experiments; **** P < .0001; ***P < .001; **P < .01; *P < .05. (B) Chart of fragments per kilobase of transcript per million mapped reads (FPKM) of Epha8 and Dpf3 from mRNA-seq. dendrite growth. An RT-qPCR assay confirmed that expression amphetamine addiction pathway, indicating that this pathway, of these 2 tested genes was significantly altered by cocaine and but not MAPK and circadian pathways, is an important mechan - reversed by DDTC, which coincided with the results obtained ism by which DDTC modifies the effects of cocaine. from RNA-sequencing ( Figure 5). Axon guidance is modulated during the development of drug addiction (Bahi and Dreyer, 2005). For example, cocaine expos- ure alters the expression of axon guidance molecules that re-gu Discussion late the formation of axon-target connections ( Bahi et al., 2005 ), The first major finding of this study was that the NF- κB inhibitor which contribute to the structural adaptations that underlie the DDTC, when coadministered with cocaine, is capable of produc - long-term effects of prolonged drug exposure ( Pasterkamp et al., ing stastistically significant partial (>50%) rescue of cocaine- 2009). We found that DDTC can significantly reverse changes in induced locomotor stimulation on the first day of treatment, gene expression of the axon guidance pathway that are induced while completely desensitizing this response to cocaine when by repeated cocaine administration, suggesting that it may coadministered over several consecutive days. While virus- be able to prevent structural changes associated with chronic mediated gene therapy that results in NF- κB inhibition has cocaine addiction. Though DDTC is known to have other func - shown success in animal models of cocaine sensitization, it is tions, such as zinc chelation and and superoxide dismutase faced with obvious barriers to clinical implementation. DDTC is inhibition, it is important to note that the NF- κB homo-dimer a pharmacological inhibitor that has been demonstrated to be motif can be mapped to 86 genes whose expression changes safe in clinical trials and is a small lipid-soluble molecule that were reversed by DDTC and that these genes are also enriched is capable of crossing the blood-brain barrier ( Hersh et al., 1991 ; in axon guidance. Thus, we suggest that NF- κB responds to Kodama et  al., 2005) and as such could be considered for the cocaine exposure by directly modifying the expression levels treatment of cocaine-addicted patients pending more thorough of members of the axon guidance pathway, leading to cocaine analyses of its effects on the development and persistence of sensitization. The amphetamine addiction pathway may then addictive behavior. The major goal of this study, however, was be modified indirectly by downstream NF- κB pathway effectors. not to evaluate the efficacy of DDTC as a treatment for cocaine In summary, our results shed light on NF- κB target genes at addiction but rather to use DDTC to reveal genes and pathways a genome-wide level in an effort to understand the mechanisms regulated by NF-κB during repeated cocaine administration. by which this critical transcription factor complex mediates the We found that the MAPK signaling cascade is a key molecu - functional effects of repeated cocaine administration. Our work lar pathway in the PFC that is altered in response to cocaine and provides new insight into the molecular mechanisms underl-y that inhibition of NF- κB by DDTC with or without cocaine treat - ing the role of NF-κB in addiction. ment also results in gene expression changes of MAPK pathway members. The MAPK signaling cascade involves the activation of Supplementary Material extracellular signal-regulated kinases-1 and -2 (ERK1 and ERK2) and is known to play an important role in the control of synaptic Supplementary data are available at International Journal of plasticity in the adult brain ( Thomas and Huganir, 2004). In fact, Neuropsychopharmacology online. MAPK/ERK signaling contributes to various behavioral effects of cocaine, such as psychomotor sensitization, conditioned place preference, and reconsolidation of memories for cocaine cues Funding (Lu et  al., 2004 , 2005, 2006) (Radwanska et  al., 2005 ). Our previ- ous study also demonstrated that cocaine treatment induced This work was financially supported by the National Natural long-lasting changes of the ERK signaling pathway in the PFC Science Foundation of China (nos. 91132728, 30870821, and (Li et  al., 2017). Consistent with this ( Li et  al., 2017) and other 61401459) and Key Laboratory of Mental Health, Institute of studies (Abarca et al., 2002 ; Lynch et al., 2008 ; Imbesi et al., 2009), Psychology, Chinese Academy of Sciences. our results also show that the circadian pathway can play a role in cocaine sensitization. However, while both the MAPK and cir - Acknowledgments cadian pathways were involved in cocaine response and DDTC also generally affected both pathways, they were not enriched We appreciate the support we received from the National Natural in the list of DEGs reversed by DDTC. On the other hand, DDTC Science Foundation of China and the Institute of Psychology, significantly reversed the cocaine-induced changes in the Chinese Academy of Sciences during our study. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 704 | International Journal of Neuropsychopharmacology, 2018 Lu L, Grimm JW, Hope BT, Shaham Y (2004) Incubation of Statement of Interest cocaine craving after withdrawal: a review of preclinical data. None. Neuropharmacology 47:214–226. Lu L, Hope BT, Dempsey J, Liu SY, Bossert JM, Shaham Y (2005) Central amygdala ERK signaling pathway is critical to incuba - References tion of cocaine craving. Nat Neurosci 8:212–219. Abarca C, Albrecht U, Spanagel R (2002) Cocaine sensitization Lu L, Koya E, Zhai H, Hope BT, Shaham Y (2006) Role of ERK in cocaine addiction. Trends Neurosci 3:3. and reward are under the influence of circadian genes and rhythm. Proc Natl Acad Sci U S A 99:9026–9030. Lynch WJ, Girgenti MJ, Breslin FJ, Newton SS, Taylor JR (2008) Gene profiling the response to repeated cocaine self-admin- Bahi A, Boyer F, Kolira M, Dreyer JL (2005) In vivo gene silenc - ing of CD81 by lentiviral expression of small interference istration in dorsal striatum: a focus on circadian genes. Brain Res 1213:166–177. rnas suppresses cocaine-induced behaviour. J Neurochem 92:1243–1255. Malek S, Chen Y, Huxford T, Ghosh G (2001) Ikappabbeta, but not ikappabalpha, functions as a classical cytoplasmic Bahi A, Dreyer JL (2005) Cocaine-induced expression changes of axon guidance molecules in the adult rat brain. Mol Cell inhibitor of NF-kappab dimers by masking both NF-kappab nuclear localization sequences in resting cells. J Biol Chem Neurosci 28:275–291. Baldwin AS Jr (1996) The NF-kappa B and I kappa B proteins: new 276:45225–45235. Natoli G (2006) Tuning up inflammation: how DNA sequence and discoveries and insights. Annu Rev Immunol 14:649–683. Bracchi-Ricard V, Brambilla R, Levenson J, Hu WH, Bramwell A, chromatin organization control the induction of inflamma - tory genes by NF-kappab. FEBS Lett 580:2843–2849. Sweatt JD, Green EJ, Bethea JR (2007) Astroglial nuclear factor- kappaB regulates learning and memory and synaptic plasti - Nestler EJ (2012) Transcriptional mechanisms of drug addiction. Clin Psychopharmacol Neurosci 10:136–143. city in female mice. J Neurochem 22:22. Chao J, Nestler EJ (2004) Molecular neurobiology of drug addic - Pasterkamp RJ, Smidt MP, Burbach JP (2009) Development and engineering of dopamine neurons. Preface. Adv Exp Med Biol tion. Annu Rev Med 55:113–132. Chao MR, Fragou D, Zanos P, Hu CW, Bailey A, Kouidou S, Kovatsi 651:v–vi. Radwanska K, Caboche J, Kaczmarek L (2005) Extracellular sig - L (2014) Epigenetically modified nucleotides in chronic heroin and cocaine treated mice. Toxicol Lett 229:451–457. nal-regulated kinases (erks) modulate cocaine-induced gene expression in the mouse amygdala. Eur J Neurosci 22:939–948. Chen Z, Liu J, Ng HK, Nadarajah S, Kaufman HL, Yang JY, Deng Y (2011) Statistical methods on detecting differentially expressed Russo SJ, et  al. (2009) Nuclear factor kappa B signaling re -gu lates neuronal morphology and cocaine reward. J Neurosci genes for RNA-seq data. BMC Syst Biol 5(Suppl 3):S1. Hersh EM, Brewton G, Abrams D, Bartlett J, Galpin J, Gill P, 29:3529–3537. Simeonidis S, Stauber D, Chen G, Hendrickson WA, Thanos Gorter R, Gottlieb M, Jonikas JJ, Landesman S (1991) Ditiocarb sodium (diethyldithiocarbamate) therapy in patients D (1999) Mechanisms by which ikappab proteins control NF-kappab activity. Proc Natl Acad Sci U S A 96:49–54. with symptomatic HIV infection and AIDS. A  randomized, double-blind, placebo-controlled, multicenter study. Jama Thomas GM, Huganir RL (2004) MAPK cascade signalling and synaptic plasticity. Nat Rev Neurosci 5:173–183. 265:1538–1544. Hoffmann A, Natoli G, Ghosh G (2006) Transcriptional regu - Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L (2010) Transcript lation via the NF-kappab signaling module. Oncogene 25:6706–6716. assembly and quantification by RNA-seq reveals unanno- tated transcripts and isoform switching during cell differ - en Imbesi M, Yildiz S, Dirim Arslan A, Sharma R, Manev H, Uz T (2009) Dopamine receptor-mediated regulation of neuronal tiation. Nat Biotechnol 28:511–515. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, Pimentel “clock” gene expression. Neuroscience 158:537–544. Kodama H, Sato E, Gu YH, Shiga K, Fujisawa C, Kozuma T (2005) H, Salzberg SL, Rinn JL, Pachter L (2012) Differential gene and transcript expression analysis of RNA-seq experiments with Effect of copper and diethyldithiocarbamate combination therapy on the macular mouse, an animal model of menkes tophat and cufflinks. Nat Protoc 7:562–578. Tripathi S, et al. (2015) Meta- and orthogonal integration of influ - disease. J Inherit Metab Dis 28:971–978. Levy N (2013) Addiction is not a brain disease (and it matters). enza “omics” data defines a role for UBR4 in virus budding. Cell Host Microbe 18:723–735. Front Psychiatry 4:24. Li CY, Mao X, Wei L (2008) Genes and (common) pathways under- Widera D, Mikenberg I, Kaltschmidt B, Kaltschmidt C (2006) Potential role of NF-kappab in adult neural stem lying drug addiction. Plos Comput Biol 4:e2. Li M, Xu P, Xu Y, Teng H, Tian W, Du Q, Zhao M (2017) Dynamic cells: the underrated steersman? Int J Dev Neurosci 24:91–102. expression changes in the transcriptome of the prefrontal cortex after repeated exposure to cocaine in mice. Front Zhou Z, Enoch MA, Goldman D (2014) Gene expression in the addicted brain. Int Rev Neurobiol 116:251–273. Pharmacol 8:142. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Neuropsychopharmacology Oxford University Press

Transcriptome Sequencing Reveals Candidate NF-κB Target Genes Involved in Repeated Cocaine Administration

Free
8 pages

Loading next page...
 
/lp/ou_press/transcriptome-sequencing-reveals-candidate-nf-b-target-genes-involved-Q0WsbNCbd2
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of CINP.
ISSN
1461-1457
eISSN
1469-5111
D.O.I.
10.1093/ijnp/pyy031
Publisher site
See Article on Publisher Site

Abstract

Background: Drug-induced alterations in gene expression play an important role in the development of addictive behavior. Numerous transcription factors have been implicated in mediating the gene expression changes that occur in drug addiction. Nuclear factor kappa B is an inducible transcription factor complex that is rapidly activated by diverse stimuli. Methods: We performed next-generation high-throughput sequencing of the prefrontal cortex in a mouse model of repeated cocaine administration combined with pharmacological nuclear factor kappa B inhibition to identify nuclear factor kappa B target genes that participate in the cocaine addiction process. Results: We found that the nuclear factor kappa B antagonist sodium diethyldithiocarbamate trihydrate significantly reversed the cocaine-induced expression changes of the amphetamine addiction pathway. Genes that demonstrated differential expression in response to cocaine treatment that was also reversed by sodium diethyldithiocarbamate trihydrate were enriched for the axon guidance pathway. Furthermore, the nuclear factor kappa B homo-dimer motif could be mapped to 86 of these sodium diethyldithiocarbamate trihydrate-reversed genes, which were also enriched for axon guidance. Conclusions: We suggest that nuclear factor kappa B directly modifies the expression of axon guidance pathway members, leading to cocaine sensitization. Our findings reveal the role of prefrontal cortex nuclear factor kappa B activity in addiction and uncover the molecular mechanisms by which nuclear factor kappa B drives changes in the addicted brain. Keywords: addiction, cocaine, nuclear factor kappa B, prefrontal cortex, sodium diethyldithiocarbamate trihydrate Introduction Drug addiction is a global issue that generates extensive harm alter the structure and function of numerous and distinct brain to the health and socioeconomic status of affected individuals regions, at least one of the major constinuents of the mesolim - and inflicts further burdens on society. Chronic addictive behav - bic reward pathway—the ventral tegmental area, the nucleus ior develops as a consequence of progressive and multifaceted accumbens, and the prefrontal cortex (PFC)—is modified by alterations to neural circuits upon repeated administration of all addictive substances ( Levy, 2013). In the context of cocaine drugs of abuse (Chao and Nestler, 2004 ). Though various drugs addiction, it has been demonstrated that the PFC is involved in Received: January 14, 2018; Revised: February 28, 2018; Accepted: March 19, 2018 © The Author(s) 2018. Published by Oxford University Press on behalf of CINP. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http:// creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, 697 provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 698 | International Journal of Neuropsychopharmacology, 2018 Significance Statement Despite a growing number of treatments, drug addiction continues to be a source of tremendous physical, psychological, social, and economic harm. Understanding the molecular mechanisms by which the brain is altered during the onset of addiction can lead to the development of more effective therapies. It is widely accepted that chronic drug exposure results in changes in gene expression. Here, we investigate the role of the transcription factor NF- κB in mediating these gene expression changes in the prefrontal cortex. Our findings demonstrate the role of NF- κB activity in the prefrontal cortrex in the establishment of cocaine addiction and uncover target genes and biological pathways through which this process occurs. both the primary reinforcing effects of cocaine and long-term through which NF-κB may mediate cocaine-induced cellular and sensitization to cocaine. Previous studies have further demon - behavioral plasticities. To do so, we examined the effects of an strated that PFC-dependent processes, such as executive func - NF-κB inhibitor on the behavioral response to repeated cocaine tion, explicit learning, and memory, are damaged in animal administration and combined our model with high-throughput models of cocaine addiction and in human cocaine abusers transcriptome sequencing to identify target genes activated by (Chao et  al., 2014). However, the molecular mechanisms driv- NF-κB in the PFC upon cocaine treatment. We demonstrate the ing altered PFC biology in response to cocaine addiction remain efficacy of pharmacological NF-κB inhibition in counteracting poorly understood, and, as a result, there continues to be a lack the effects of cocaine and, by establishing the genes and path - of effective therapies for addicted patients. ways modulated by NF-κB in response to cocaine, we reveal Though drug-induced changes in the expression of numer - novel targets for further therapeutic intervention. ous neuronal genes contribute to behavioral abnormalities, the origin and causality of these genetic signatures are unclear Materials and Methods (Zhou et  al., 2014). Consistent changes in characteristic path - ways suggest the action of specific transcription factors, which Animals respond to environmental changes by binding to regulatory regions of target genes to alter their expression ( Baldwin, 1996). Male C57BL/6J mice (purchased from the Academy of Military Medical Science) weighing 25 to 30 g (8 weeks old) at the beginning In the past several decades, numerous transcription factors, such as Fos family protein, cyclic adenosine monophosphate of the experiments were used in this study. Five mice were housed per cage and all mice were given food and water ad libitum. The response element binding protein, nuclear factor kappa B (NF- κB), and myocyte enhancing factor-2, have been implicated in colony was maintained at 22  ±  2°C with a standard 12-h-light/- dark cycle (lights on at 7:00 am). The animals were handled daily the addiction process ( Nestler, 2012 ). NF-κB represents a small family of inducible transcription for 1 week before injection. Testing was performed during the light cycle. All animal protocols were approved by the Review Board of factors that are rapidly activated by diverse external stimuli and therefore serve as critical mediators of many environmen - the Institute of Psychology, Chinese Academy of Sciences, and were performed strictly in accordance with the Guideline for Care and tal inputs ( Nestler, 2012 ). The NF-κB complex is typically held in an inactive form in the cytosol by interaction with members of Use of Laboratory Animals of the Chinese Academy of Sciences. the inhibitor of NF-κB (IκB) family of inhibitory proteins. Upon stimulation by extracellular inducers,κ I Bs are phosphoryl - Drugs ated by an IκB kinase complex, which leads to ubiquitination and proteosome-mediated degradation of the IκBs. Then, active Cocaine hydrochloride (Qinghai Pharmaceutical Co., Ltd.) and NF-κB, through classical or noncanonical pathways, translo - sodium diethyldithiocarbamate trihydrate (DDTC; NF- κB antag - cates to the nucleus, where it binds consensus κB sequences in onist, Sigma-Aldrich) were dissolved in sterile 0.9% saline and the promoter and enhancer regions of responsive genes ( Widera administered i.p. at 20 and 200 mg/kg, respectively. et  al., 2006). NF-κB binding sites are found in numerous genes that are known to participate in memory formation and drug Locomotor Sensitization Test addiction: brain-derived growth factor, inducible nitric oxide synthase, and various opioid receptors ( Simeonidis et al., 1999; Mice were placed in the center of the field (30 x 30 cm), and total Malek et al., 2001 ). However, these examples represent a small distance travelled by each animal was recorded for 60 minutes number of NF-κB targets, and there is no evidence whether and analyzed by a computer-based system (Anilab). After each these and other genes are regulated by NF- κB in the PFC during trial, the apparatus was cleaned with a 30% ethanol solution. cocaine addiction or whether they play a causal role in addiction The order of testing was balanced by group. onset and persistence. It had been previously demonstrated that the NF- κB signal- Treatment and Behavioral Testing Procedures ing pathway, including several candidate downstream genes, is Mice were randomly assigned to 4 treatment groups: (1) activated in the nucleus accumbens of mice exposed to chronic cocaine and inhibition of this pathway by a virally expressed saline + saline (SS, n = 10), (2) DDTC + saline (DS, n = 10), (3) DDTC + cocaine (DC, n = 10), and (4) cocaine + saline (CS, n = 10). dominant-negative form of IκB kinase attenuates the addictive effects of cocaine (Russo et al., 2009 ). However, in the context of The treatment period was 8 days, as shown in Figure 1. On days 1 to 3, baseline locomoter activity of every treatment group cocaine addiction, the importance of NF- κB signaling in the PFC has not been investigated, and the great genome-wide diversity was assessed for 1 hour immediately following an i.p. injection of saline. On days 4 to 8, groups (2) and (3) received an of NF-κB target genes has not been explored. The major goals of this study were to determine whether injection of saline or cocaine, respectively, 30 minutes after an injection of DDTC, while groups (1) and (4) also received an injec - NF-κB signaling in the PFC is important in the development of cocaine addiction and to identify the potential target genes tion of saline or cocaine, respectively, but 30 minutes following Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup Wang et al. | 699 (Tripathi et al., 2015 ) with a cutoff of P < .05 and a minimum over - lap of 3. These raw sequencing data sets were deposited in the Gene Expression Omnibus of the NCBI ( https://www.ncbi.nlm. nih.gov/geo) under accession number GSE108836. Real-Time PCR Validation Validation of DEGs was performed by quantitative reverse-tr- an scription PCR (RT-qPCR) using the Maxima SYBR Green qPCR Master Mix kit (Fermentas), according to the manufacturer’s instructions, in an ABI Prism 7500 Sequence Detection System machine (Applied Biosystems Inc). All real-time RT-qPCR data were normalized to SS expression level (see additional data file for primer information). Figure 1. Effect of behavior of sodium diethyldithiocarbamate trihydrate (DDTC; 200  mg/kg) in response to cocaine-addicted mice (20  mg/kg, i.p.). CS, saline+cocaine; DC, DDTC+cocaine; ds, DDTC+saline; SS, saline+saline. All Motif Search results represent means ± SD of 10 independent determinations (n = 10/group) in the behavioral test experiments. **** P < .0001, ***P < .001. To discover DEGs with NF-κB binding sites, we applied the UniProbe Database ( http://thebrain.bwh.harvard.edu/nfkb/ ) to an injection of saline. All mice were put into the locomotor box promoters (-500 to +200 bp relative to the transcription start site) immediately following the second injection and monitored for of DEGs. The z-score cutoff was set to the default value of 4. 1 hour. Data Analysis RNA Sequencing and Bioinformatics Data are expressed as the mean ± SEM. Statistical analyses were Two hours after the final (day 8) locomotor activity test, animals performed with GraphPad Prism 6.0 software. Comparisons were killed and their PFCs were surgically excised. The tissue was of means of behavioral sensitization were analyzed by 2-way stored in liquid nitrogen immediately and then transferred to a repeated-measures ANOVA followed by Tukey’s posthoc tests. -80°C freezer. Total RNA was extracted from the frozen tissues using the TIANamp DNA/RNA Isolation Kit (TIANGEN), including Results additional treatment with RNase-free DNase I  (Ambion) for 30 minutes at 37°C to remove contaminating DNA. Effects of DDTC on Cocaine-Induced Behavior Then 1  µg of RNA from a pool of 5 animals (200  ng each) per experimental condition was used to generate an mRNA First, we analyzed the effects of the injection of (1) saline + saline sequencing library using the TruSeq RNA Sample Prep Kit (SS), (2) DDTC + saline (DS), (3) DDTC + cocaine (DC), and (4) (Illumina) following the manufacturer’s protocol. Briefly, polyA cocaine + saline (CS), repeated daily for 5  days, on mouse loco - mRNA was purified and fragmented. Then, first and second motor activity after 3  days of baseline testing. As shown in strand cDNA synthesis was performed, ends were repaired and Figure  1, 2-way repeated-measures ANOVA of the locomotion adenylated, adapters were ligated, and fragments were enriched test results demonstrated that there were significant differences with PCR amplification. High library qualities were verified with in the total distance traveled between the 4 treatment groups an Agilent Technologies 2100 Bioanalyzer (Agilent Technologies). (F = 44.34, P < .001). Posthoc analysis revealed that there was (3,27) Libraries were sequenced using the Illumina HiSEQ2K platform no significant difference between the locomotor activities of (Illumina). each group at baseline (days 1–3). However, total distance tra - ve FastQC was used to assess RNA-seq quality. Because of the led increased significantly following cocaine treatment (SS vs low quality of the 3’ end of raw reads, the last 20 base-pairs were CS P < .001) and DDTC administration significantly inhibited trimmed from each read. Reads were also trimmed sequen - cocaine-induced activity (CS vs DC P < .001). Additionally, there tially for adapter content. The paired-end reads were aligned was no significant difference between the locomotor activi - to the mouse reference genome (mm9, Jul. 2007)  with TopHat ties of DC and SS groups on Day 5 onwards, suggesting that 2.0.9 (Trapnell et  al., 2012 ) using default parameters. Aligned DDTC can completely inhibit the hyperlocomotion induced by bam files were processed by Cufflinks v.2.2.1 ( Trapnell et  al., repeated cocaine administration. 2010) to generate a transcriptome assembly and to estimate the expression level (fragments per kilobase of transcript per mil - GO and KEGG Pathways Induced by Cocaine lion mapped reads [FPKM]) of all detected isoforms. FPKM was calculated as the number of paired-end reads (a single fragment To understand the molecular mechanisms underlying the effects per end-paired reads) mapped to a gene divided by the number of cocaine and to delineate the functional consequences of DDTC of all fragments mapped to the genome (in millions) and the treatment in the context of cocaine addiction, we performed length of the transcript (in kilobases). The tagwise dispersion mRNA-sequencing of extracted PFCs of mice from each of the 4 was estimated and then used for log fold change calculation. treatment groups, immediately following the 8th and final day Differentially expressed genes (DEGs) were identified by apply- of treatment and testing. A total of approximately 60 million 100 ing the threshold false discovery rate of 0.2 to adjusted P values base-pair paired-end reads were obtained per sample. following Fisher’s exact test Chen ( et  al., 2011). DEGs induced Of 266 DEGs, 175 genes were upregulated and 91 genes were by cocaine sensitization (logfold_change [CS/SS]) were plotted downregulated in the CS group compared with the SS group against baseline gene expression levels of the SS group (lo F- g (Figure  2A). We then analyzed whether certain (KEGG) path - PKM) using ggplot2 in R.  KEGG pathway, and Gene Ontology ways were enriched for in the list of PFC DEGs and discovered (GO) analyses of DEGs were performed with Metascape software 13 significantly enriched pathways upon repeated cocaine Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 700 | International Journal of Neuropsychopharmacology, 2018 Figure  2. Cocaine and sodium diethyldithiocarbamate trihydrate (DDTC) cause broad changes to the transcriptome. mRNA from prefrontal cortex (PFC) of SS (saline+saline), CS (saline+cocaine), and DS (DDTC+saline) was subjected to RNA sequencing and analysis as described in the methods. (A) MA-plot [M (log ratios); A (mean average)] of differentially expressed genes (DEGs) of cocaine+saline (CS), respectively < (q 0.2). DEGs are represented by green dots. Log2 fold change values for CS vs SS are plotted against average log expression values (fragments per kilobase of transcript per million mapped reads [FPKM]). (B) Chart of enriched pathways of significantly expressed genes in CS ( P < .05). (C) MA-plot of DEGs of DDTC+saline (DS), respectively (q < 0.2). DEGs are represented by green dots. Log2 fold change values for DS vs SS are plotted against average log expression values (FPKM). (D) Chart of enriched pathways of significantly expressed genes in DS ( P < .05). administration. Among these, 3 pathways were directly related GO and KEGG Analyses Reveal the Effects of to central nervous system function: most notably, the amphet - DDTC on Pathways Altered by Repeated Cocaine amine addiction pathway, as well as the circadian rhythm Administration pathway and amyotrophic lateral sclerosis (ALS) ( Figure  2B). Next, we applied the same analyses to study the effects of com - Additional GO analysis revealed multiple terms relevant to bining cocaine with DDTC treatment. As a major goal of this neurological function ( supplimentary Figure 1 ), including regu - study was to identify the target genes by which NF- κB mediates lation of circadian rhythm (GO:0042752), synaptic vesicles the develoment of cocaine addiction, we sought to find genes (GO:0008021), and MAPK function (GO:0043408 and GO:0017017), whose expression was altered by repeated cocaine treatment which were reported to be involved in the control of synaptic with the condition that this differential expression is abolished plasticity ( Thomas and Huganir, 2004). upon NF-κB inhibition by DDTC treatment. In the process of identifying these genes, the next step was to perform a differ - GO and KEGG Pathways Induced by DDTC ential expression analysis between DC and CS groups to obtain a list of DEGs whose expression is affected by DDTC treatment We simultaneously analyzed how DDTC treatment modi - in the context of cocaine treatment. A  total 451 DEGs were fies the mouse transcriptome ( Figure  2C). A  total 143 DEGs upregulated and 410 DEGs were downregulated in DC-treated were upregulated while 58 DEGs were downregulated in DS mice compared with CS-treated mice (Figure 3A). The top 20 sig- compared with SS. KEGG pathway analysis of all DS DEGs nificantly enriched pathways showed that DDTC can regulate revealed a stastically significant enrichment for 11 pathways pathways known to be related to addiction, such as ampheta - (Figure 2D). Interestingly, the MAPK signaling pathway, the cir - mine, calcium signaling, ABC transporters, gap junctions, glu - cadian rhythm (entrainment) pathway, and the focal adhesion tathione metabolism, ALS, glycolysis, Alzheimer’s disease, and pathway, which were significantly enriched in cocaine treat - neuroactive ligand-receptor interaction ( Li et  al., 2008), as well ment (CS vs SS), were also affected by DDTC treatment, indi - as other pathways related to neuronal function such as cir - ca cating that cocaine and DDTC modulate several of the same dian entrainment, neuotrophin signaling, and axon guidance molecular pathways in the PFC. GO analysis again revealed (Figure  3B). GO analysis (supplementary Figure  3) showed that numerous enriched pathways related to neuronal function, the terms neuron projection development (GO:0031175), behav - including MAPKs (GO:0017017), synapse (GO:0045202), neuro - ior (GO:0007610), calcium channel activity (GO:0005262), synapse transmitter transport (GO:0006836), central nervous system (GO:0045202), and axon (GO:0030424) were enriched. development (GO:0007417), learning or memory (GO:0007611), We were then interested in isolating pathways from DEGs and neuron projection development (GO:0031175) ( supplemen - that met 3 requirements: (1) are altered by cocaine alone tary Figure 2 ). Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup Wang et al. | 701 Figure 3. Transcriptome-wide response of sodium diethyldithiocarbamate trihydrate (DDTC) to cocaine addiction. (A) MA-plot of differentially expressed genes (DEGs) of DDTC+cocaine (DC), respectively (q < 0.2). DEGs are represented by green dots. Log2 fold change values for DC vs saline+cocaine (CS) are plotted against average log expression values (fragments per kilobase of transcript per million mapped reads [FPKM]). (B) Chart of enriched pathways of significantly expressed genes in DS ( P < .05). (C) Venn diagram depicting the overlap of DEGs among DC, DDTC+saline (DS), and CS. (D) Heatmap of comparison of enriched pathways of CS, DS, and DC. (E) Scatter plot of all genes in amphetamine addiction pathway. (Figure 2A); (2) are modulated by DDTC in cocaine-treated (DC) Interestingly, scatter plots ( Figure  3E; supplementary mice ( Figure  3A); and (3) result from DDTC treatment alone Figure  5) showed that DDTC (DC) significantly reversed the (DS) (Figure  2C). These DEGs and pathways would represent direction of cocaine-induced (CS) gene expression changes of changes in cocaine-sentization that are specifically caused by the amphetamine addiction pathway ( P = .0498) but not of the DDTC in DC-treated mice rather than representing a combined ALS pathway ( P = .078), indicating that the amphetamine addic - effect of DDTC and cocaine coadministration. The 3 requisite tion pathway is an important pathway by which DDTC modu - DEG data sets are plotted in a Venn diagram ( Figure 3C; http:// lates gene expression changes in response to repeated cocaine bioinfogp.cnb.csic.es/tools/venny/ ), which interestingly depicts administration. a significant overlap between all 3 data sets, between DC and DS DEGs as well as DC and CS DEGs, but not between CS and DS DDTC Reverses Cocaine-Induced Changes in DEGs, which are not present in the DC group (chi-squared test). Expression of Genes in the Axon Guidance Pathway Next, we performed a 3-way statistical comparison of enriched pathways and GO terms from DEGs induced in CS, DC, and DS While the previous analyses identified common pathways treatments ( Figure  3D). The KEGG pathways common to all 3 by which DDTC can reverse changes seen upon repeated DEG groups were fluid shear stress and atherosclerosis, focal cocaine treatment, pathways consist of many genes whose adhesion, MAPK signaling, and circadian entrainment. products exert contradictory functions on pathway activity, However, common pathways significantly enriched in both obscuring conclusions as to whether DDTC clearly reversed cocaine treatment alone (CS) and DDTC+cocaine (DC) treatment cocaine-induced behavior of these pathways. Therefore, we DEGs included pathways in cancer, ALS, and amphetamine wished to determine if DDTC is able to antagonize cocaine- addiction, and several common GO iterms were related to neu - related expression changes of specific genes. We plotted the ral signal conduction supplementar ( y Figure 4 ,) such as calcium log fold change of DEGs from DC, CS, and DS conditions in channel activity (GO:0005262), neurotransmitter receptor acti - v a heatmap, with which it was immediately visible that many ity (GO:0030594), and presynapse (GO:0098793). more genes exhibited fold changes in opposite directions in Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 702 | International Journal of Neuropsychopharmacology, 2018 the DC mice compared with CS and DS mice, an observation Candidate NF-κB Target Genes Involved in Response that was confirmed by hierarchial clustering ( Figure  4A). In to Repeated Cocaine Administration fact, the log fold changes of DEGs in DC and CS displayed -15 2 NF-κB is a protein complex composed of homo- or heterodimers a significantly negative relationship ( P = 2.994e , R = 0.2647), of 5 different family members: c-Rel (Rel), RelA/p65 (Rela), RelB while those of DS and CS had a significantly positive rela - -16 2 (Relb), NF-κB /p50/p105 (Nfkb1), and NF-κB /p52/p100 (Nfkb2) tionship (P = 2.2e , R = 0.4723) (Figure  4B; supplementary 1 2 (Natoli, 2006). Studies of protein-DNA crystal structures and Figure  5). These results indicate that DDTC reverses specific DNA-binding studies have shown that NF- κB and NF-κB recog- gene expression changes that occur in response to cocaine 1 2 nize the 5 base-pair motif of 5’-GGGRN-3’, whereas c-Rel, RelA/ treatment. p65, and RelB recognize a 4 base-pair 5’-GGRR-3’ motif, and the We then performed an additional pathway analysis to entire κB family recognizes a concensus motif of 5’-GGGRN(Y) determine which pathways were enriched in the list of YYCC-3’ (Hoffmann et  al., 2006; Bracchi-Ricard et  al., 2007 ). genes whose cocaine-induced expression changes were spe- Using this information, we applied online tools (see Methods) cifically reversed by DDTC (i.e., fold change in the opposite to scan the promoter regions of DEGs for NF- κB family bind- direction). There were a total of 142 such genes, 58 of which ing sites. The NF-κB motif could be mapped to the promoter were upregulated and 84 of which were downregulated by regions 86 DEGs (supplementary Table 3 ). This gene list was also cocaine, while these changes were reversed by DDTC ( supple- enriched in the axon guidance term (mmu04360, P = .089), sup- mentary Table  2). KEGG analysis revealed that the most sig - porting the hypotheses that NF- κB directly regulates the axon nificantly enriched pathway was axon guidance (mmu04360, guidance pathway in cocaine addiction and that DDTC inter - P = .0088). Interestingly, when we analyzed the lo gfold feres with this activity. changes of all DEGs in the axon guidance pathway posthoc, that is, in a less stringent manner, not only were the DEGs with significant cocaine-induced expression changes signifi - Validation of Genes Reversed by DDTC cantly reversed by DDTC but there was a significantly neg-a -08 tive relationship between CS and DC groups ( P = 4.211e ), The top-ranked DEGs by fold-change expression were Epha8 (Ephrin Type-A Receptor 8), a member of the axon guidance thereby further demonstrating that DDTC can reverse the axon guidance behavior that is modified by cocaine treatment pathway, and Dpf3 (Double PHD Finger 3), which belongs to the neuron-specific chromatin remodeling complex involved in (Figure 4C). Figure  4. Sodium diethyldithiocarbamate trihydrate (DDTC) reverses genes regulated by cocaine addiction. (A) Heatmap of all significantly differentially expressed genes in each group (saline+cocaine [CS], DDTC+saline [DS], and DDTC+cocaine [DC]). Red, upregulated; blue, downregulated. (B) Scatter plot of all differentially expressed genes of log2 fold_change of CS (X axis) and log2 fold_change of DC (Y axis). (C) Scatter plot of all genes in axon guidance pathway. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup Wang et al. | 703 Figure 5. Expression of cocaine and sodium diethyldithiocarbamate trihydrate (DDTC) target genes as determined by qPCR. (A) Normalized mRNA expression quanti - fied by RT-qPCR for genes whose expression is altered by cocaine treatment and reversed by DDTC: Epha8 and Dpf3. Results represent mean ± SD of 3 determinations in each of the 2 types of experiments; **** P < .0001; ***P < .001; **P < .01; *P < .05. (B) Chart of fragments per kilobase of transcript per million mapped reads (FPKM) of Epha8 and Dpf3 from mRNA-seq. dendrite growth. An RT-qPCR assay confirmed that expression amphetamine addiction pathway, indicating that this pathway, of these 2 tested genes was significantly altered by cocaine and but not MAPK and circadian pathways, is an important mechan - reversed by DDTC, which coincided with the results obtained ism by which DDTC modifies the effects of cocaine. from RNA-sequencing ( Figure 5). Axon guidance is modulated during the development of drug addiction (Bahi and Dreyer, 2005). For example, cocaine expos- ure alters the expression of axon guidance molecules that re-gu Discussion late the formation of axon-target connections ( Bahi et al., 2005 ), The first major finding of this study was that the NF- κB inhibitor which contribute to the structural adaptations that underlie the DDTC, when coadministered with cocaine, is capable of produc - long-term effects of prolonged drug exposure ( Pasterkamp et al., ing stastistically significant partial (>50%) rescue of cocaine- 2009). We found that DDTC can significantly reverse changes in induced locomotor stimulation on the first day of treatment, gene expression of the axon guidance pathway that are induced while completely desensitizing this response to cocaine when by repeated cocaine administration, suggesting that it may coadministered over several consecutive days. While virus- be able to prevent structural changes associated with chronic mediated gene therapy that results in NF- κB inhibition has cocaine addiction. Though DDTC is known to have other func - shown success in animal models of cocaine sensitization, it is tions, such as zinc chelation and and superoxide dismutase faced with obvious barriers to clinical implementation. DDTC is inhibition, it is important to note that the NF- κB homo-dimer a pharmacological inhibitor that has been demonstrated to be motif can be mapped to 86 genes whose expression changes safe in clinical trials and is a small lipid-soluble molecule that were reversed by DDTC and that these genes are also enriched is capable of crossing the blood-brain barrier ( Hersh et al., 1991 ; in axon guidance. Thus, we suggest that NF- κB responds to Kodama et  al., 2005) and as such could be considered for the cocaine exposure by directly modifying the expression levels treatment of cocaine-addicted patients pending more thorough of members of the axon guidance pathway, leading to cocaine analyses of its effects on the development and persistence of sensitization. The amphetamine addiction pathway may then addictive behavior. The major goal of this study, however, was be modified indirectly by downstream NF- κB pathway effectors. not to evaluate the efficacy of DDTC as a treatment for cocaine In summary, our results shed light on NF- κB target genes at addiction but rather to use DDTC to reveal genes and pathways a genome-wide level in an effort to understand the mechanisms regulated by NF-κB during repeated cocaine administration. by which this critical transcription factor complex mediates the We found that the MAPK signaling cascade is a key molecu - functional effects of repeated cocaine administration. Our work lar pathway in the PFC that is altered in response to cocaine and provides new insight into the molecular mechanisms underl-y that inhibition of NF- κB by DDTC with or without cocaine treat - ing the role of NF-κB in addiction. ment also results in gene expression changes of MAPK pathway members. The MAPK signaling cascade involves the activation of Supplementary Material extracellular signal-regulated kinases-1 and -2 (ERK1 and ERK2) and is known to play an important role in the control of synaptic Supplementary data are available at International Journal of plasticity in the adult brain ( Thomas and Huganir, 2004). In fact, Neuropsychopharmacology online. MAPK/ERK signaling contributes to various behavioral effects of cocaine, such as psychomotor sensitization, conditioned place preference, and reconsolidation of memories for cocaine cues Funding (Lu et  al., 2004 , 2005, 2006) (Radwanska et  al., 2005 ). Our previ- ous study also demonstrated that cocaine treatment induced This work was financially supported by the National Natural long-lasting changes of the ERK signaling pathway in the PFC Science Foundation of China (nos. 91132728, 30870821, and (Li et  al., 2017). Consistent with this ( Li et  al., 2017) and other 61401459) and Key Laboratory of Mental Health, Institute of studies (Abarca et al., 2002 ; Lynch et al., 2008 ; Imbesi et al., 2009), Psychology, Chinese Academy of Sciences. our results also show that the circadian pathway can play a role in cocaine sensitization. However, while both the MAPK and cir - Acknowledgments cadian pathways were involved in cocaine response and DDTC also generally affected both pathways, they were not enriched We appreciate the support we received from the National Natural in the list of DEGs reversed by DDTC. On the other hand, DDTC Science Foundation of China and the Institute of Psychology, significantly reversed the cocaine-induced changes in the Chinese Academy of Sciences during our study. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018 Copyedited by: oup 704 | International Journal of Neuropsychopharmacology, 2018 Lu L, Grimm JW, Hope BT, Shaham Y (2004) Incubation of Statement of Interest cocaine craving after withdrawal: a review of preclinical data. None. Neuropharmacology 47:214–226. Lu L, Hope BT, Dempsey J, Liu SY, Bossert JM, Shaham Y (2005) Central amygdala ERK signaling pathway is critical to incuba - References tion of cocaine craving. Nat Neurosci 8:212–219. Abarca C, Albrecht U, Spanagel R (2002) Cocaine sensitization Lu L, Koya E, Zhai H, Hope BT, Shaham Y (2006) Role of ERK in cocaine addiction. Trends Neurosci 3:3. and reward are under the influence of circadian genes and rhythm. Proc Natl Acad Sci U S A 99:9026–9030. Lynch WJ, Girgenti MJ, Breslin FJ, Newton SS, Taylor JR (2008) Gene profiling the response to repeated cocaine self-admin- Bahi A, Boyer F, Kolira M, Dreyer JL (2005) In vivo gene silenc - ing of CD81 by lentiviral expression of small interference istration in dorsal striatum: a focus on circadian genes. Brain Res 1213:166–177. rnas suppresses cocaine-induced behaviour. J Neurochem 92:1243–1255. Malek S, Chen Y, Huxford T, Ghosh G (2001) Ikappabbeta, but not ikappabalpha, functions as a classical cytoplasmic Bahi A, Dreyer JL (2005) Cocaine-induced expression changes of axon guidance molecules in the adult rat brain. Mol Cell inhibitor of NF-kappab dimers by masking both NF-kappab nuclear localization sequences in resting cells. J Biol Chem Neurosci 28:275–291. Baldwin AS Jr (1996) The NF-kappa B and I kappa B proteins: new 276:45225–45235. Natoli G (2006) Tuning up inflammation: how DNA sequence and discoveries and insights. Annu Rev Immunol 14:649–683. Bracchi-Ricard V, Brambilla R, Levenson J, Hu WH, Bramwell A, chromatin organization control the induction of inflamma - tory genes by NF-kappab. FEBS Lett 580:2843–2849. Sweatt JD, Green EJ, Bethea JR (2007) Astroglial nuclear factor- kappaB regulates learning and memory and synaptic plasti - Nestler EJ (2012) Transcriptional mechanisms of drug addiction. Clin Psychopharmacol Neurosci 10:136–143. city in female mice. J Neurochem 22:22. Chao J, Nestler EJ (2004) Molecular neurobiology of drug addic - Pasterkamp RJ, Smidt MP, Burbach JP (2009) Development and engineering of dopamine neurons. Preface. Adv Exp Med Biol tion. Annu Rev Med 55:113–132. Chao MR, Fragou D, Zanos P, Hu CW, Bailey A, Kouidou S, Kovatsi 651:v–vi. Radwanska K, Caboche J, Kaczmarek L (2005) Extracellular sig - L (2014) Epigenetically modified nucleotides in chronic heroin and cocaine treated mice. Toxicol Lett 229:451–457. nal-regulated kinases (erks) modulate cocaine-induced gene expression in the mouse amygdala. Eur J Neurosci 22:939–948. Chen Z, Liu J, Ng HK, Nadarajah S, Kaufman HL, Yang JY, Deng Y (2011) Statistical methods on detecting differentially expressed Russo SJ, et  al. (2009) Nuclear factor kappa B signaling re -gu lates neuronal morphology and cocaine reward. J Neurosci genes for RNA-seq data. BMC Syst Biol 5(Suppl 3):S1. Hersh EM, Brewton G, Abrams D, Bartlett J, Galpin J, Gill P, 29:3529–3537. Simeonidis S, Stauber D, Chen G, Hendrickson WA, Thanos Gorter R, Gottlieb M, Jonikas JJ, Landesman S (1991) Ditiocarb sodium (diethyldithiocarbamate) therapy in patients D (1999) Mechanisms by which ikappab proteins control NF-kappab activity. Proc Natl Acad Sci U S A 96:49–54. with symptomatic HIV infection and AIDS. A  randomized, double-blind, placebo-controlled, multicenter study. Jama Thomas GM, Huganir RL (2004) MAPK cascade signalling and synaptic plasticity. Nat Rev Neurosci 5:173–183. 265:1538–1544. Hoffmann A, Natoli G, Ghosh G (2006) Transcriptional regu - Trapnell C, Williams BA, Pertea G, Mortazavi A, Kwan G, van Baren MJ, Salzberg SL, Wold BJ, Pachter L (2010) Transcript lation via the NF-kappab signaling module. Oncogene 25:6706–6716. assembly and quantification by RNA-seq reveals unanno- tated transcripts and isoform switching during cell differ - en Imbesi M, Yildiz S, Dirim Arslan A, Sharma R, Manev H, Uz T (2009) Dopamine receptor-mediated regulation of neuronal tiation. Nat Biotechnol 28:511–515. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, Pimentel “clock” gene expression. Neuroscience 158:537–544. Kodama H, Sato E, Gu YH, Shiga K, Fujisawa C, Kozuma T (2005) H, Salzberg SL, Rinn JL, Pachter L (2012) Differential gene and transcript expression analysis of RNA-seq experiments with Effect of copper and diethyldithiocarbamate combination therapy on the macular mouse, an animal model of menkes tophat and cufflinks. Nat Protoc 7:562–578. Tripathi S, et al. (2015) Meta- and orthogonal integration of influ - disease. J Inherit Metab Dis 28:971–978. Levy N (2013) Addiction is not a brain disease (and it matters). enza “omics” data defines a role for UBR4 in virus budding. Cell Host Microbe 18:723–735. Front Psychiatry 4:24. Li CY, Mao X, Wei L (2008) Genes and (common) pathways under- Widera D, Mikenberg I, Kaltschmidt B, Kaltschmidt C (2006) Potential role of NF-kappab in adult neural stem lying drug addiction. Plos Comput Biol 4:e2. Li M, Xu P, Xu Y, Teng H, Tian W, Du Q, Zhao M (2017) Dynamic cells: the underrated steersman? Int J Dev Neurosci 24:91–102. expression changes in the transcriptome of the prefrontal cortex after repeated exposure to cocaine in mice. Front Zhou Z, Enoch MA, Goldman D (2014) Gene expression in the addicted brain. Int Rev Neurobiol 116:251–273. Pharmacol 8:142. Downloaded from https://academic.oup.com/ijnp/article-abstract/21/7/697/4948030 by Ed 'DeepDyve' Gillespie user on 03 July 2018

Journal

International Journal of NeuropsychopharmacologyOxford University Press

Published: Mar 21, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off