Abstinence from Escalation of Cocaine Intake Changes the microRNA Landscape in the Cortico-Accumbal Pathway

Cocaine administration alters the microRNA (miRNA) landscape in the cortico-accumbal pathway. These changes in miRNA can play a major role in the posttranscriptional regulation of gene expression during withdrawal. This study aimed to investigate the changes in microRNA expression in the cortico-accumbal pathway during acute withdrawal and protracted abstinence following escalated cocaine intake. Small RNA sequencing (sRNA-seq) was used to profile miRNA transcriptomic changes in the cortico-accumbal pathway [infralimbic- and prelimbic-prefrontal cortex (IL and PL) and nucleus accumbens (NAc)] of rats with extended access to cocaine self-administration followed by an 18-h withdrawal or a 4-week abstinence. An 18-h withdrawal led to differential expression (fold-change > 1.5 and p < 0.05) of 21 miRNAs in the IL, 18 miRNAs in the PL, and two miRNAs in the NAc. The mRNAs potentially targeted by these miRNAs were enriched in the following pathways: gap junctions, neurotrophin signaling, MAPK signaling, and cocaine addiction. Moreover, a 4-week abstinence led to differential expression (fold-change > 1.5 and p < 0.05) of 23 miRNAs in the IL, seven in the PL, and five miRNAs in the NAc. The mRNAs potentially targeted by these miRNAs were enriched in pathways including gap junctions, cocaine addiction, MAPK signaling, glutamatergic synapse, morphine addiction, and amphetamine addiction. Additionally, the expression levels of several miRNAs differentially expressed in either the IL or the NAc were significantly correlated with addiction behaviors. Our findings highlight the impact of acute and protracted abstinence from escalated cocaine intake on miRNA expression in the cortico-accumbal pathway, a key circuit in addiction, and suggest developing novel biomarkers and therapeutic approaches to prevent relapse by targeting abstinence-associated miRNAs and their regulated mRNAs.


Introduction
Recent epidemiological data suggest a resurgence in cocaine use and cocaine-related problems in the United States [1]. The continued use of cocaine induces lasting maladaptive plasticity that can lead to cocaine use disorder (CUD). Similar to other substance use disorders (SUDs), CUD is a chronic and relapsing neuropsychiatric disorder, causing a serious global health problem that exerts a heavy toll on both individuals and society [2,3].
contingent cocaine in rodents have demonstrated coding and non-coding RNA expression changes at different stages of cocaine withdrawal in the prefrontal cortex (PFC) and nucleus accumbens (NAc) [29][30][31]. However, abstinence-induced small noncoding RNA (e.g., miRNA) expression changes using a contingent, escalated cocaine intake paradigm merit further investigation.
The present study explored cocaine withdrawal/abstinence-associated miRNA transcriptomic changes in brain reward regions using the HS rat. Extended access to cocaine was followed by two different durations of forced withdrawal and abstinence (18 h and 4 weeks, respectively). miRNA transcriptome profiling was carried out in the corticoaccumbal pathway, including the nucleus accumbens (NAc), a key brain region mediating a variety of behaviors, including reinforcement learning, as well as the pre-and infra-limbic areas of the prefrontal cortex (PFC), which normally exerts executive control of behavior.

Rat Brain Samples
Three groups of snap-frozen HS rat brains were obtained from the Cocaine Biobank located in the Department of Psychiatry at the University of California, San Diego (UCSD) La Jolla, CA 92093, USA. Abstinence rats (4 male rats) had an escalated cocaine selfadministration (SA) followed by a 4-week protracted abstinence; withdrawal rats (4 male rats) had an escalated cocaine SA followed by an 18-h withdrawal; and naïve rats (4 male rats) were naïve animals. The rats in the three groups were age-matched at the start of self-administration [the mean age (days) of rats for the 4-week abstinence experiment: 61 ± 9 (mean ± SD); the mean age (days) of rats for the 18-h withdrawal experiment: 64 ± 8 (mean ± SD); and the mean age (days) of naïve rats for the control experiment (naïve: 63 ± 8 (mean ± SD)]. The t-tests did not show significant age differences between the three groups of rats at the start of cocaine administration (abstinence vs. withdrawal: t = 0.47, p = 0.656; abstinence vs. naïve: t = 0.252, p = 0.809; withdrawal vs. naïve: t = 0.22, p = 0.829). Detailed information on cocaine SA and withdrawal, as well as behavioral procedures, can be found in a published article [28] and the George Lab protocol (https://www.protocols.io/workspaces/george-lab) (last two visits-November 2022 and December 2022). All procedures were performed following the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health (NIH), and the protocol (IACUC S1906-accessed on 23 February 2023) was approved by the Institutional Animal Care and Use Committees (IACUC) of The Scripps Research Institute and UCSD. Our studies were carried out blinded to the treatment conditions of the rats.
Briefly, rats were anesthetized with vaporized isoflurane (1-5%) and then underwent surgical insertion of an intravenous catheter into the right jugular vein. After surgery and recovery, rats were trained on a fixed ratio 1 (FR1) schedule for self-administering (SA) cocaine (0.5 mg/kg) for 10 sessions, with each session lasting for 2 h. Next, rats were given longer access (LgA) to SA cocaine (0.5 mg/kg) for 14 sessions, with each session lasting for 6 h. Following the last SA session, a subset of rats underwent a protracted period (4 weeks) of forced abstinence and were then euthanized (abstinence: 4 male rats). Another subset of rats were euthanized 18 h following the last SA session (withdrawal: 4 male rats). Finally, a third group of aged-matched naïve rats (naïve: 4 male rats) was used in parallel. All brains were extracted and submerged in a slurry of 2-methylbutane with dry ice (at −30 • C) until fully frozen. Snap-frozen rat brains were then shipped on dry ice to the Boston University School of Medicine. In addition, abstinence and withdrawal rats were subjected to a battery of behavioral tests to compute z-scores for specific addiction-like behaviors, as described previously [28]. These z-scores included the escalation index (Z ESC), the progressive ratio index (motivation or Z PR), and the shock index (compulsivitylike behavior or Z Shock). The overall addiction index was then obtained by averaging relevant behavioral indices. Boston University's investigators were blinded to whether the rats were naïve or cocaine-experienced rats, and whether they underwent withdrawal or protracted abstinence from cocaine.

Rat Brain Dissection and Small RNA Sequencing (sRNA-Seq)
Thick sections from snap-frozen rat brains were sliced using a razor blade and a Thermo Scientific HM525 NX Cryostat (Thermo Fisher Scientific, Waltham, MA, USA). The nucleus accumbens (NAc) and infra-and pre-limbic prefrontal cortices (IL and PL) were dissected (Supplementary Figure S1). The microdissection was carried out rapidly from thick frozen sections placed on the surface of a clean, RNase-free glass platform cooled to −30 • C on dry ice. Anatomical features were used to identify the brain nuclei prior to dissection of specific limbic nuclei.
Total RNAs were isolated from 10 to 30 mg of selected brain nuclei of the limbic circuit using the miRNeasy Mini kit (Qiagen, Valencia, CA, USA). RNA integrity numbers (RIN) and concentrations of 36 total RNA samples [extracted from three brain regions (IL, PL, and NAc) of three groups of rats (abstinence: 4 male rats; withdrawal: 4 male rats; and naïve: 4 male rats)] were measured using the Agilent 2100 Bioanalyzer and Agilent RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, CA, USA) (Supplementary Table S1). Small noncoding miRNA transcriptomes in three brain regions (IL, PL, and NAc) of the three groups of rats were profiled using small RNA sequencing (sRNA-seq), as we previously described [21,32]. In short, 250 g of total RNA was used to generate small RNA-seq libraries with the use of the NEBNext Multiplex Small RNA Library Prep Set for Illumina (Set 1) (NEB, Ipswich, MA, USA). cDNA libraries were purified and pooled in an equimolar ratio (12 libraries per pool). They were then sequenced by single-ended 75-bp RNA sequencing on an Illumina HiSeq 2500 Sequencing System (Illumina, CA, USA).
The workflow for the Comprehensive Analysis Pipeline for miRNA Sequencing Data (CAP-miRseq) [33] was used for raw reads (in fastq files) pre-processing, alignment (Reference Sequences: the Rat Genome Assembly rn6), and mature/precursor/novel miRNA qualification and prediction. The read counts of 723 rat mature miRNAs were obtained. The average number of reads per sample was 20,559,895, and the average mapping rate (aligned reads/reads sent to Aligner) was 72.9%. The sRNA-seq fastq files and normalized read counts can be downloaded from the NCBI Gene Expression Omnibus (GEO) database (accession number: GSE212651).

Statistical Analysis
miRNA transcriptomic changes in each of the three brain regions (IL, PL, and NAc) due to a 4-week abstinence or an 18-h withdrawal after escalated cocaine use were analyzed [abstinence (a 4-week abstinence) vs. naïve, as well as withdrawal (an 18-h withdrawal) vs. naïve] following the method applied in our recent study [32]. We also compared miRNA transcriptomes between rats with protracted abstinence and acute withdrawal (i.e., a 4-week abstinence vs. an 18-h withdrawal).
First, the filterByExpr function in edgeR [34] was used to filter out lowly expressed miRNAs, which were not expressed at a biologically meaningful level. Second, miRNA expression levels were normalized using the weighted trimmed mean of M-value (TMM) method through the calcNormFactors function in edgeR [35], which computed scaling factors to convert observed library sizes into effective library sizes. Third, the expression counts were converted into normalized log 2 CPM counts using the function voom [36], by which the mean-variance relationship of the log-counts was analyzed and a precision weight for each miRNA was calculated. Fourth, the lmfit function in Limma [37] was used to fit a linear regression model using weighted least squares for miRNAs. Estimated coefficients and standard errors for a given set of contrasts between groups were computed using the contrasts.fit function, which is also part of the lima package [37]. Finally, a t-statistic test was used to identify significant miRNAs. The RNA integrity number (RIN) of RNA samples was considered a covariate in the differential expression analysis. The decideTests function [38] for multiple testing across genes and contrasts was used to classify miRNAs as significantly negative, not significant, or significantly positive. The differential expression results were visualized by volcano plotting [39]. In addition, the correlation of normalized expression levels of differentially expressed miRNAs [fold-change (FC) ≥ 2.0 and p < 0.05] with the addiction index, as well as z-scores (Z ESC: the z-score of escalation of cocaine use; Z PR: the z-score of progressive ratio or motivation to seek cocaine; and Z Shock: the z-score of contingent foot shock or resistance to punishment) was analyzed by a partial correlation analysis [40] with "Group" as a confounding variable.

Bioinformatics Analysis
The function of cocaine withdrawal-associated miRNAs [|fold-change (FC)| ≥ 1.5 and p < 0.05], identified by comparing miRNA transcriptomes among the three groups of rats (acute withdrawal vs. naïve, protracted abstinence vs. naïve, and acute withdrawal vs. protracted abstinence), was analyzed using web-server mirPath v.3 (Volos, Greece) [41], which utilizes the DIANA-microT-CDS algorithm to predict miRNA targets (or mRNAs). mirPath v.3 was also used to annotate KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways overrepresented by mRNAs predicted to be targeted by differentially expressed miRNAs.
In addition, the predicted miRNA-mRNA pairs and their associated canonical pathways were analyzed using the miRNA Target Filter function in the Ingenuity Pathway Analysis (IPA, Ingenuity Systems, http://www.ingenuity.com) URL accessed on 29 July 2022. miRNAs that were differentially expressed between groups (|FC| ≥ 1.5 and p < 0.05) were run through the miRNA Target pipeline in the IPA to identify their target mRNAs with high confidence or experimentally validated. Next, target mRNAs were filtered to retain those belonging to specific classes of molecules, such as "nuclear receptor" or "transcription factor". miRNAs were filtered to include those predicted as target genes (or mRNAs) participating in canonical pathways associated with nervous system processes or signaling pathways. Finally, miRNAs and their target mRNAs (after filtering) along with relevant canonical pathways enriched in miRNA target mRNA sets (as determined by the IPA database) were visualized.

Brain Region-Specific miRNA Expression Changes Due to Acute Withdrawal or Protracted
Abstinence following Cocaine Self-Administration (SA) After normalizing miRNA expression levels and removing low-expression miRNAs, 377 mature miRNAs remained for differential expression analysis. Differentially expressed miRNAs in three brain regions (IL, PL, and NAc) of rats with acute withdrawal (18 h) are visualized using volcano plots ( Figure 1). An 18-h acute withdrawal (compared to naïve rats) resulted in differential expression (|FC| ≥ 1.5 and p < 0.05) of 21 miRNAs in the IL, 18 miRNAs in the PL, and two miRNAs in the NAc (Table 1). rno-miR-338-3p showed differential expression in both PL and NAc; however, it was upregulated in the PL (log 2 FC = 0.80 or FC = 1.74 and p = 0.046) but downregulated in the NAc (log 2 FC = −0.84 or FC = −1.78 and p = 0.038). rno-miR-382-3p was downregulated in both IL (log 2 FC = −1.29 or FC = −2.45 and p = 0.022) and NAc (log 2 FC = −1.08 or FC = −2.11 and p = 0.049). Moreover, differentially expressed miRNAs in three brain regions (IL, PL, and NAc) of rats with a protracted (4-week) protracted abstinence are visualized with volcano plots (Figure 2). A 4-week protracted abstinence (compared to naïve rats) led to differential expression (|FC| ≥ 1.5 and p < 0.05) of 23 miRNAs in the IL, seven miRNAs in the PL, and five miRNAs in the NAc (Table 2). Two miRNAs were upregulated in both the IL (rno-miR-935: log 2 FC = 0.81 or FC = 1.75 and p = 0.002; rno-miR-770-5p: log 2 FC = 0.83 or FC = 1.77 and p = 0.017) and the NAc (rno-miR-935: log 2 FC = 0.70 or FC = 1.63 and p = 0.005; rno-miR-770-5p: log 2 FC = 0.66 or FC = 1.58 and p = 0.047). Additionally, differential expression (|FC| ≥ 1.5 and p < 0.05) of 16 miRNAs in the IL, two miRNAs in the PL, and one miRNA in the NAc were observed when comparing miRNA transcriptome profiles between rats with a 4-week abstinence and an 18-h withdrawal after cessation of cocaine self-administration ( Figure 3 and Table 3).
in the NAc were observed when comparing miRNA transcriptome profiles between rats with a 4-week abstinence and an 18-hour withdrawal after cessation of cocaine self-administration ( Figure 3 and Table 3).

Figure 1.
Volcano plotting of differentially expressed rat brain miRNAs due to an 18-hour cocaine withdrawal (an 18-hour cocaine withdrawal vs. naïve rats). Scattered points represent miRNAs. The x-axis is the log2 fold-change (FC) of miRNA expression levels between the two groups of rats, and the y-axis is the p-value based on −log10. . Volcano plotting of differentially expressed rat brain miRNAs due to an 18-h cocaine withdrawal (an 18-h cocaine withdrawal vs. naïve rats). Scattered points represent miRNAs. The x-axis is the log 2 fold-change (FC) of miRNA expression levels between the two groups of rats, and the y-axis is the p-value based on −log10. Log 2 FC: the log2 fold-change (FC) between cases and controls; AveExpr: average miRNA expression level; t: the t-statistic used to assess differential expression; B: the empirical Bayes log odds of differential expression; p.Value: the p-value for differential expression (unadjusted for multiple testing); adj.p.Val: the p-value adjusted for multiple testing (the default is Benjamini−Hochberg). *, **: Differentially expressed miRNAs in multiple brain regions. Log2FC: the log2 fold-change (FC) between cases and controls; AveExpr: average miRNA expression level; t: the t-statistic used to assess differential expression; B: the empirical Bayes log odds of differential expression; p.Value: the p-value for differential expression (unadjusted for multiple testing); adj.p.Val: the p-value adjusted for multiple testing (the default is Benjamini−Hochberg). *, **: Differentially expressed miRNAs in multiple brain regions.

Figure 2.
Volcano plotting of differentially expressed rat brain miRNAs due to a 4-week cocaine abstinence (a 4-week cocaine abstinence vs. naïve rats). Scattered points represent miRNAs. The xaxis is the log2 fold-change (Fc) of miRNA expression levels between the two groups of rats, and the y-axis is the p-value based on −log10.

Figure 2.
Volcano plotting of differentially expressed rat brain miRNAs due to a 4-week cocaine abstinence (a 4-week cocaine abstinence vs. naïve rats). Scattered points represent miRNAs. The x-axis is the log 2 fold-change (Fc) of miRNA expression levels between the two groups of rats, and the y-axis is the p-value based on −log10.   Log2FC: the log2 fold-change (FC) between cases and controls; AveExpr: average miRNA expression level; t: the t-statistic used to assess differential expression; B: the empirical Bayes log odds of differential expression; p.Value: the p-value for differential expression (unadjusted for multiple testing); adj.p.Val: the p-value adjusted for multiple testing (the default is Benjamini−Hochberg).

Figure 3.
Volcano plotting of miRNAs with differential expression between rats with two stages of cocaine withdrawal (18-h cocaine withdrawal vs. 4-week cocaine abstinence). Scattered points represent miRNAs. The x-axis is the log 2 fold-change (Fc) of miRNA expression levels between the two groups of rats, and the y-axis is the p-value based on −log10. the t-statistic used to assess differential expression; B: the empirical Bayes log odds of differential expression; p.Value: the p-value for differential expression (unadjusted for multiple testing); adj.p.Val: the p-value adjusted for multiple testing (the default is Benjamini−Hochberg).

Function of Withdrawal-Associated Brain miRNAs Predicted by DIANA-mirPath
The function of differentially expressed (|FC| ≥ 1.5 and p <0.05) miRNAs associated with cocaine withdrawal was annotated by the miRNA pathway analysis web server DIANA-mirPath and presented using bubble plots (Figures 4-6). mRNAs predicted to be targeted by 39 miRNAs (Table 1)

Function of Withdrawal-Associated Brain miRNAs Predicted by DIANA-mirPath
The function of differentially expressed (|FC| ≥ 1.5 and p <0.05) miRNAs associat with cocaine withdrawal was annotated by the miRNA pathway analysis web server D ANA-mirPath and presented using bubble plots (Figures 4-6). mRNAs predicted to targeted by 39 miRNAs (Table 1) (Table  associated with an 18-hour cocaine withdrawal. All 39 miRNAs were associated with an 18-ho cocaine withdrawal in either of the three brain regions (IL, PL, and NAc). The mirPath v.3 onl tool (https://dianalab.e-ce.uth.gr/html/mirpathv3/index.php?r=mirpath; accessed on 27 Janua 2023) was used to annotate KEGG pathways overrepresented by mRNAs predicted to be target by these miRNAs.
Moreover, mRNAs predicted to be targeted by 33 differentially expressed (|FC| ≥ 1 and p < 0.05) miRNAs (Table 2) (Table 1) associated with an 18-h cocaine withdrawal. All 39 miRNAs were associated with an 18-h cocaine withdrawal in either of the three brain regions (IL, PL, and NAc). The mirPath v.3 online tool (https://dianalab.e-ce.uth.gr/html/mirpathv3/index.php?r=mirpath; accessed on 27 January 2023) was used to annotate KEGG pathways overrepresented by mRNAs predicted to be targeted by these miRNAs.
Moreover, mRNAs predicted to be targeted by 33 differentially expressed (|FC| ≥ 1.5 and p < 0.05) miRNAs (Table 2) (Table  associated with a 4-week cocaine abstinence. All 33 miRNAs were associated with a 4-week cocai abstinence in either of the three brain regions (IL, PL, and NAc). The mirPath v.3 online tool w used to annotate KEGG pathways overrepresented by mRNAs predicted to be targeted by the miRNAs.

Cocaine Withdrawal or Abstinence-Associated miRNA−mRNA-Pathway Network Predicted by IPA
IPA was used to gain further insights into the biological significance of differentially expressed miRNAs and their putative mRNA targets. The numbers of differentially ex-  (Table 3) with differential expression between two stages (18-h vs. 4-week) of cocaine withdrawal rats. Here, 19 miR-NAs showed differential expression between two stages (18-h vs. 4-week) of cocaine withdrawal for rats in either of the three brain regions (IL, PL, and NAc). The mirPath v.3 online tool was used to annotate KEGG pathways overrepresented by mRNAs predicted to be targeted by these miRNAs.

Cocaine Withdrawal or Abstinence-Associated miRNA−mRNA-Pathway Network Predicted by IPA
IPA was used to gain further insights into the biological significance of differentially expressed miRNAs and their putative mRNA targets. The numbers of differentially expressed (|FC| ≥ 1.5 and p < 0.05) miRNAs identified in each of the three limbic nuclei (IL, PL, and NAc) through the three comparisons (an 18-h acute withdrawal vs. naïve; a 4-week protracted abstinence vs. naïve; and acute withdrawal vs. protracted abstinence) were visualized using Venn diagrams (Figure 7). These miRNAs were associated with acute withdrawal, protracted abstinence, or both.  Figure S2).

Figure 7.
Venn diagram plotting of the numbers of differentially expressed miRNAs in three rat brain regions due to an 18-hour cocaine withdrawal, a 4-week cocaine abstinence, or both.
IL-PFC: number of differential IL-PFC miRNA expression profiles by three comparisons; PL-PFC: number of differential PL-PFC miRNA expression profiles by three comparisons; and NAc: number of differential NAc miRNA expression profiles by three comparisons. These miRNAs were associated with acute withdrawal, protracted abstinence, or both.

Addiction-like Behaviors and Their Correlation with Expression Levels of Differentially Expressed miRNAs
The three z-scores (Z ESC, Z PR, and Z Shock) and the overall addiction index (based on the three z-scores) of four rats in abstinence (4-week abstinence) and four rats in withdrawal (18-hour withdrawal) are summarized in Supplementary Table S2. There was no significant difference in the addiction index between abstinence and withdrawal rats (t = 0.25, p = 0.808).
We further analyzed the correlation of the addiction index with expression levels of the top differentially expressed miRNAs (|FC| ≥ 2.0 and p < 0.05) that were identified by comparing brain transcriptome profiles of the three groups of rats (abstinence vs. naïve, withdrawal vs. naїve, and abstinence vs. withdrawal). Four miRNAs (rno-miR-582-3p, rno-miR-292-5p, rno-miR-551b-3p, and rno-miR-382-3p) in the IL, three miRNAs (rno-miR-381-3p, rno-miR-1188-3p, and rno-miR-872-5p) in the PL, and one miRNA (rno-miR- IPA pathway analysis predicted canonical pathways for all three groups of DE miRNA (withdrawal only, withdrawal and protracted abstinence, and uniquely differentially expressed during protracted abstinence, depicted in the Venn diagram). Additionally, IPA analysis provided predicted targets. We focus on ligand-gated receptors, transcription factors, and signaling pathways. Only targets and pathways that were classified as "experimentally validated and with high confidence" in the IPA database were used to obtain insight into the possible functional role of miRNAs.
All differentially expressed miRNAs in the three groups (WD, abstinence, and both in WD and abstinence) have ligand-gated nuclear receptors and transcription factors as predicted targets. These miRNAs are also predicted to be involved in highly relevant canonical signaling pathways. Canonical signaling pathways identified by IPA analysis include Glutamate Receptor signaling, Long Term Potentiation (LTP) and Long Term Depression (LTD), and GABA Receptor signaling; CREB signaling in Neurons; Endocannabinoid Developing Neuron Pathway; Endocannabioid Neuronal Synapse pathway; Protein Kinase A signaling and Glucocorticoid Receptor signaling; Dopamine-DARP32 feedback in cAMP signaling; and Nervous system pathways-Endocannabinoid, Serotonin, and Dopamine.
IL-PFC: number of differential IL-PFC miRNA expression profiles by three comparisons; PL-PFC: number of differential PL-PFC miRNA expression profiles by three comparisons; and NAc: number of differential NAc miRNA expression profiles by three comparisons. These miRNAs were associated with acute withdrawal, protracted abstinence, or both.

Addiction-like Behaviors and Their Correlation with Expression Levels of Differentially Expressed miRNAs
The three z-scores (Z ESC, Z PR, and Z Shock) and the overall addiction index (based on the three z-scores) of four rats in abstinence (4-week abstinence) and four rats in withdrawal (18-h withdrawal) are summarized in Supplementary Table S2. There was no significant difference in the addiction index between abstinence and withdrawal rats (t = 0.25, p = 0.808).

Discussion
Substance use disorders (SUDs) are characterized by an overpowering loss of control over drug seeking and a compulsive intake of drugs despite adverse consequences [42]. The transition from casual or recreational drug use to compulsive use happens only in a small fraction of human subjects and animal models, but the mechanisms underpinning this transition are unclear [43][44][45][46]. These observations highlight the variability among individuals in the degree of vulnerability for compulsive drug use [47,48] and underscore the importance of using a genetically diverse rat strain (similar to the heterogeneity of human populations) to conduct research. Here we report on miRNA transcriptome alterations in HS rats during withdrawal and prolonged abstinence and compared with drug-naïve rats as a means of identifying changes relevant to the increased propensity for drug use and/or resilience.
Much is known regarding chronic drug use-induced neuroplasticity at the cellular and molecular levels in the case of cocaine in a region-specific manner in the brain [49,50]. In contrast, circuit-wide transcriptomic changes induced by chronic drug use followed by prolonged abstinence are less well understood. Mounting evidence shows that repeated misuse of drugs reprograms the genome throughout the cortico-accumbal circuit after prolonged abstinence [31,51] and may be key in causing vulnerability or resilience to chronic drug use. We have examined cocaine withdrawal-and abstinence-associated miRNA expression changes in the infralimbic (IL) and prelimbic (PL) subdivisions of the medial prefrontal cortex (mPFC), as well as the nucleus accumbens (NAc). Alterations in the expression of miRNAs in these limbic nuclei are of interest since the two subregions of the mPFC drive opposing drug use-related behaviors. The PL drives drug seeking, while the IL inhibits these behaviors [52,53]. Vulnerability to drug seeking increases in correlation with a long duration of abstinence. This phenomenon of incubation of cocaine craving [54][55][56], which is an increased responsivity to cocaine and withdrawal-associated cues, is seen not only in preclinical models but also in humans [57]. In human subjects and animal models, negative emotional states increase during withdrawal from drug use, and resumption of drug taking may alleviate these aversive states and increase relapse vulnerability [58][59][60][61].
Not surprisingly, miRNAs, by virtue of their ability for posttranscriptional regulation of mRNAs and hence protein synthesis, can directly affect cellular functions. Several studies have demonstrated a direct effect of miRNAs on modulating behavioral patterns of drug abuse-related behavior. Overexpression or knockdown of specific miRNAs in vivo in specific limbic nuclei influences addiction behaviors. For instance, motivation for cocaine [62] and vulnerability to cocaine use disorder [12,63] can be regulated by miRNAs. Viral vector-mediated overexpression of miR-9 in the NAc increases the self-administration of oxycodone in rats [64]. The direct effect of miRNAs on behavioral phenotypes indicates the importance of the miRNA transcriptome in determining an individual's vulnerability to drug addiction and the need for studying miRNA alterations in SUDs.
As an initial step in unraveling the critical functional role of miRNAs expressed in the limbic regions (IL, PL, and NAc) for regulating behaviors relevant to cocaine use disorder (CUD), we have investigated miRNA transcriptomic changes associated with acute withdrawal and protracted abstinence following escalated cocaine. The number of miRNAs that were selectively differentially expressed in specific time intervals, namely withdrawal or abstinence, or those that were differentially expressed persistently in both abstinence and withdrawal in the three brain regions were visualized by a Venn diagram (Figure 7). Importantly, the use of these three groups of rats enabled the identification of miRNAs that are differentially expressed and are unique to specific periods of withdrawal or abstinence (Figure 7) and has permitted insight into the possible functional roles of specific miRNAs in promoting craving or resilience in each of these abstinence periods.
To gain an understanding of mechanisms underlying drug use-related behaviors, bioinformatics platforms and tools were leveraged to predict canonical pathways regulated by miRNAs. Pathway identification provides insight into possible signaling pathways and mechanisms underlying drug use-relevant behaviors. The Diana-mirPath/KEGG Pathway analysis reveals that multiple addiction-relevant signaling pathways (including cocaine, morphine, and amphetamine addictions) are over-represented by genes targeted by miRNAs identified in the three limbic regions (Figures 4-6). KEGG pathways overrepresented by genes potentially targeted by miRNAs associated with both withdrawal and abstinence include cocaine addiction, neurotrophin signaling, MAPK signaling, and gap junctions (Figures 4 and 5). Ample evidence suggests that neurotrophins and the MAPK signal transduction cascade play important roles in behavioral sensitization to cocaine [65]. Studies have also shown that cocaine exposure impacts gap junctional communication between neurons [66]. The increase in the vasopressin pathway during withdrawal is significant since it promotes stress-related substance use behavior [67], whereas oxytocin, observed to increase during abstinence, is an anti-stress factor [68,69]. The FOXO family of transcription factors elicits the enhancement of SUDS-related behaviors [70].
Thus far, very few transcription factors other than members of the Fos family have been identified in drug abuse studies [71]. However, recent studies [31,51,72] support a prominent role of transcription factors and nuclear receptor families in the addictive process. Consistent with these recent studies, the IPA pathway analysis of our miRNA seq data using only experimentally validated and high confidence targets reveals that predicted targets of DE miRNAs in IL and PL during WD, abstinence alone, and both in WD and abstinence are enriched in CREB and other transcription factors (Supplementary Figure S3). This is of functional significance since transcription factors are critical upstream modulators of cocaine-induced plasticity [73].
It is known that dysregulation of endogenous plasticity mechanisms in the limbic circuit is likely to promote SUDs. For instance, in the NAc, cascades of homeostatic plasticity from bi-directional synaptic-membrane crosstalk occur and are critically involved in cocaine-induced plasticity and consequent cocaine craving [74,75], and there is a persistent strengthening of the PFC−NAc pathway [76]. Preventing dysregulation of homeostatic plasticity and restoring normal plasticity is likely to provide resilience. The IPA bioinformatic analysis using only experimentally validated and high confidence targets predict that DE miRNAs identified in the present study target many relevant plasticity pathways. These include the glutamate and GABA receptor pathways, LTP and LTD, the glucocorticoid signaling pathway, and the endocannabinoid signaling pathway. Predicted targets are interesting in that they include, for example, NMDA receptor subunits, AMPA receptor auxillary proteins, and glutamate transporters-proteins that are crucial for eliciting neuronal plasticity ( Supplementary Figures S4 and S5).
Additionally, even further restriction of identified miRNA candidates to only those with target mRNAs involved in nervous system processes or neurotransmitter signaling yielded three miRNAs (rno-miR101a-3p, rno-miR192-3p, and rno-miR137-3p). Pathways predicted to be regulated by these three miRNAs include the Endocannabinioid Neuronal Synapse Pathway, Neuroinflammation signaling Pathway, PI3/Akt Signaling, and Apoptosis Signaling for rno-miR101-3p; DNA Methylation and Transcriptional Repression Signaling Pathway and Myelination Signaling Pathway for rno-miR192-3p; and Aryl Hydrocarbon Receptor Signaling and Senescence Pathway for miR137-3p (Supplementary Figure S2). Once again, these pathways underscore the functional significance of the DE miRNA identified in our study and highlight the importance of plasticity, neuroinflammation, and epigenetic processes such as DNA methylation in CUD.
Recently, Domingo-Rodriguez and colleagues, using cocaine self-administration in mice, have shown that miRNAs regulate vulnerability vs. resilience, thus contributing to interindividual variability [63]. Consistent with this study, we find that specific miRNAs were significantly correlated with addiction-relevant behaviors, such as escalated cocaine intake (i.e., correlation with the z-score for escalation or Z ESC) and motivation to take cocaine (i.e., correlation with the z-score for progressive ratio or Z PR) (Supplementary  Materials Tables S3b,c). Of interest, the expression levels of rno-miR-101a-3p and rno-miR-137-3p in the NAc were significantly correlated with Z PR (Supplementary Table S3c). This finding implies that the altered expression of these two miRNAs in NAc could influence cocaine use behavior by changing the rats' motivation to seek cocaine. High motivation for drugs in rats, as measured by the maximal effort exerted for drug infusions, may mimic attributes seen in human subjects with SUDs.
Limitations of the study include a small sample size. Another is the lack of a larger number of differentially expressed miRNAs identified in the NAc. This may be due to extracting RNA from the entire NAc and not separating the shell from the core subdivisions. The shell and core subdivisions of the NAc have different roles in addiction [77]. Only male rats have been used in this study, but female rats will be tested in our future studies to elucidate sex differences since these are known to occur in impulsive and compulsive behaviors such as addiction [78].

Conclusions
This work is an important, initial step in identifying miRNA-mediated regulation of plasticity induced by protracted abstinence from escalated cocaine intake. We have identified rat brain limbic region miRNAs that are likely to modulate behaviors that are relevant to cocaine abuse. By identifying escalated cocaine self-administration and withdrawal/abstinence-induced miRNAs, we have uncovered an additional layer of epigenetic modulation of CUD. Increased understanding of epigenome remodeling, brain plasticity, and vulnerability/resilience to drug abuse will provide insights for developing novel, effective therapeutic strategies for SUDs, including CUD.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/biomedicines11051368/s1. Figure S1: Dissection of rat brain nucleus accumbens and prefrontal cortex; Figure S2: Cocaine withdrawal/abstinence-associated miRNA-mRNA-Pathway network predicted by IPA; Figure S3: IPA miRNA Target Filter prediction of transcription factors (TFs) & nuclear receptors (NRs) potentially targeted by differentially expressed miRNAs in the IL-PFC of rats with protracted abstinence; Figure S4: IPA miRNA Target Filter prediction of miRNA-mRNA-pathway regulatory networks containing miRNAs differentially expressed in the IL-PFC of rats with protracted abstinence. Figure S5: IPA miRNA Target Filter prediction of miRNA-mRNA-pathway regulatory networks containing miRNAs differentially expressed in the PL-PFC of rats with either withdrawal or protracted abstinence. Table S1: The concentration and RNA integrity number (RIN)of 36 rat brain RNA samples; Table S2: Addiction behavior Z scores and addiction index of Group A and Group B rats; Table S3a: Correlation of addiction index with expression levels of top differentially expressed miRNAs (|FC| ≥ 2.0 & p < 0.05) in three rat brain regions by partial correlation analysis; Table S3b: Correlation of Z ESC with expression levels of top differentially expressed miRNAs (|FC| ≥ 2.0 & p < 0.05) in three rat brain regions by partial correlation analysis; Table S3c: Correlation of Z PR with expression levels of top differentially expressed miRNAs (|FC| ≥ 2.0 & p < 0.05) in three rat brain regions by partial correlation analysis; Table S3d: Correlation of Z Shock with expression levels of top differentially expressed miRNAs (|FC| ≥ 2.0 & p < 0.05) in three rat brain regions by partial correlation analysis.

Informed Consent Statement: Not Applicable.
Data Availability Statement: The raw data supporting the conclusions of this article can be downloaded from the NCBI Gene Expression Omnibus (GEO) database (Accession Number: GSE212651).