Cell type- and region-specific modulation of cocaine seeking by micro-RNA-1 in striatal projection neurons

The persistent and experience-dependent nature of drug addiction may result in part from epigenetic alterations, including non-coding micro-RNAs (miRNAs), which are both critical for neuronal function and modulated by cocaine in the striatum. Two major striatal cell populations, the striato-nigral and striato-pallidal projection neurons, express respectively the D1 (D1-SPNs) and D2 (D2-SPNs) dopamine receptor, and display distinct but complementary functions in drug-evoked responses. However, a cell-type-specific role for miRNAs action has yet to be clarified. Here, we evaluated the expression of a subset of miRNAs proposed to modulate cocaine effects in the nucleus accumbens (NAc) and dorsal striatum (DS) upon sustained cocaine exposure in mice and showed that these selected miRNAs were preferentially up-regulated in the NAc. We then focused on miR-1 considering the important role of some of its predicted mRNA targets, such as fosb and npas4, in the effects of cocaine. We validated these targets in vitro and in vivo and further showed that overexpression of miR-1 in D1-SPNs of the DS reduced cocaine-induced CPP reinstatement, whereas it increased cue-induced reinstatement of cocaine-SA, without affecting other cocaine-mediated adaptive behavior. In addition, miR-1 overexpression in D2-SPNs of the DS reduced the motivation to self-administer cocaine but did not modify other measured behaviors. Together, our results highlight a precise cell-type- and region-specific control of relapse to cocaine-seeking behaviors by miR-1, and illustrate the importance of cell-specific investigations.


ABSTRACT
The persistent and experience-dependent nature of drug addiction may result in part from epigenetic alterations, including non-coding micro-RNAs (miRNAs), which are both critical for neuronal function and modulated by cocaine in the striatum. Two major striatal cell populations, the striato-nigral and striato-pallidal projection neurons, express respectively the D1 (D1-SPNs) and D2 (D2-SPNs) dopamine receptor, and display distinct but complementary functions in drug-evoked responses. However, a cell-type-specific role for miRNAs action has yet to be clarified. Here, we evaluated the expression of a subset of miRNAs proposed to modulate cocaine effects in the nucleus accumbens (NAc) and dorsal striatum (DS) upon sustained cocaine exposure in mice and showed that these selected miRNAs were preferentially up-regulated in the NAc. We then focused on miR-1 considering the important role of some of its predicted mRNA targets, such as fosb and npas4, in the effects of cocaine. We validated these targets in vitro and in vivo and further showed that overexpression of miR-1 in D1-SPNs of the DS reduced cocaine-induced CPP reinstatement, whereas it increased cue-induced reinstatement of cocaine-SA, without affecting other cocaine-mediated adaptive behavior. In addition, miR-1 overexpression in D2-SPNs of the DS reduced the motivation to self-administer cocaine but did not modify other measured behaviors. Together, our results highlight a precise cell-type-and region-specific control of relapse to cocaine-seeking behaviors by miR-1, and illustrate the importance of cell-specific investigations.
INTRODUCTION association between cocaine and contextual cues and drive cocaine seeking induced by those cues [19]. 282 While most of the studies investigating the role of D1-and D2-SPNs in the behavioral effects of 283 cocaine focused on the NAc, a recent work showed a very specific role of D1-SPNs of the dorsomedial 284 striatum in cocaine responses, where their inhibition had no impact on loss of control over cocaine 285 intake, motivation to obtain cocaine or compulsive cocaine self-administration, but specifically 286 reduced cue-induced cocaine-seeking in high-risk addiction groups of rats [20]. evidence that miRNA-mediated gene regulation plays an important role in cocaine-induced changes in 297 neurotransmission and behavior [26], but there is so far, no direct evidence regarding cell-type-298 specificity of cocaine-induced miRNA in the striatum, and more specifically in D1-SPNs versus D2-299 SPNs. One study began to broadly address this aspect by specific ablation in D2-SPNs of Ago2 [27], a 300 key protein in the processing and production of miRNAs [28][29][30]. These mice showed loss of 301 motivation to self-administer cocaine, along with a decrease of around one third of miRNAs induced 302 by acute cocaine, thus indicating that miRNA induction occurs in both D1-and D2-SPNs. However, 303 the question of whether and how miRNAs regulation in D1-SPNs versus D2-SPNs controls gene 304 expression and behaviors induced by cocaine, remains to be addressed. 305 Here we aimed to identify the regional pattern, i.e. DS versus NAc, of expression of a set of selected 306 miRNAs, proposed to play a key role in cocaine action or neuronal plasticity after sub-chronic cocaine 307 administrations [24,25,31]. We next focused on miR-1, based on its predicted mRNA targets, including 308 fosb and npas4 mRNAs. We used a validated viral strategy [32][33][34] in order to thoroughly evaluate 309 miR-1 cell-type-specific roles in D1-or D2-SPNs of the NAc or DS in multiple behaviors related to 310 cocaine addiction. We found a key role of miR-1 overexpression in D1-SPNs of the DS on cocaine-and 311 cue-induced relapse to cocaine seeking, with a reduction of cocaine-induced reinstatement of cocaine 312 seeking in conditioned place preference (CPP) paradigm, and an increase in cue-induced reinstatement 313 of cocaine self-administration (SA). We also observed a reduction of motivation for cocaine self-314 administration after miR-1 overexpression in D2-SPNs of the DS. Altogether our data unveil a 315 complex cell-type-and region-specific regulatory role for miR-1 in different relapse behaviors, and the 316 importance of cell-specific investigations. 317

MATERIALS AND METHODS 319
More details on the materials and methods used are given in Supplementary Methods. 320

Animals and drug treatments 321
All animal treatments and procedures were conducted in accordance with local and European 322 directives for the care and use of laboratory animals (86/609/EEC) and approved by the institutional 323

Stereotaxic injections 331
Mice were anesthetized using ketamine and xylazine (5:1 in amount; 0.1 mL/10g), placed in a 332 stereotaxic frame (David Kopf, Tujunga, CA), and 1 µl in the NAc or 2 µl in the DS was injected at a 333 constant rate of 0.20 µl/min by using a microinfusion pump (Harvard Apparatus, Holliston, MA). We 334 used the following coordinates to perform bilateral injections according to the mouse brain atlas of 335

Quantitative PCR analysis (qPCR) of miRNA and mRNA levels 338
Punches of NAc and DS were performed from frozen tissues and total RNA was isolated using 339 miRNeasy MiniKit (Qiagen) according to the manufacturer's instructions. For mRNA levels analysis, 340 total RNA (100 ng) isolated from each brain sample was reverse-transcribed with oligodT primers 341 using RevertAid First Strand cDNA Synthesis Kit (ThermoFisher). For miRNA levels analysis, total 342 RNA (100 ng) was reverse-transcribed using the same kit with a combination of specific RT-primers 343 for each miRNA of interest (0.5 pmol of each RT-primer), as previously described [35], allowing high 344 specificity, sensitivity and homogeneity. qPCR was performed with a LightCycler 480 device (Roche) 345 using LightCycler SYBR Green I Master Mix (Roche) and specific couples of primers; samples were 346 run in triplicates. The primers sequences for miRNAs and mRNAs are presented in supplementary 347  Luciferase assay to assess the interaction between miR-1 and the 3'UTR of Bdnf, FosB and 349 7 The pmirGLO-3'UTR_Bdnf, pmirGLO-3'UTR_FosB and pmirGLO-3'UTR_Npas4 plasmids were 351 purchased from Creative Biogene and the mimick-miR-1a-3p and mimick-miR-negative control from 352 Qiagen (miScript miRNA Mimics). The protocol used is described in DOI: 10.21769/BioProtoc.420. 353 The mean ratio of luminescence from the experimental reporter (firefly) to luminescence from the 354 control reporter (Renilla) was calculated for each triplicate and normalized to the ratio of control wells 355 (miRNA negative control which does not target the 356 UTR). 357

Cocaine-induced conditioned place preference (CPP), extinction and reinstatement 360
The CPP was evaluated in a Plexiglas Y-shaped apparatus (Immetronic) with one arm blocked, located 361 in a soundproof testing room with low luminosity using an unbiased procedure. Two of the three 362 chambers were used for the conditioning and distinguished by different floor textures and wall 363 patterns. The device was connected to an electronic interface for data collection. 364 Entries and time spent in each chamber were measured, as well as the locomotor activity of the mice 365 during the experiments. 366

Operant self-administration of cocaine 367
The self-administration experiments were conducted in mouse operant chambers (Model ENV307A-368 CT; Med Associates Inc., Georgia, VT, USA) as previously described [37]. Nose pokes were the 369 operant action that led to delivery of a cocaine infusion. 370

Statistical analysis 376
All statistical analyses were carried out using GraphPad Prism and Statview softwares. For all 377 experiments, ANOVAs were performed followed, when appropriate, by post hoc test Bonferroni for 378 multiple comparisons, except when data did not pass the normality test and/or when the number of 379 individuals was too low. In those cases, non-parametric tests, such as Mann-Whitney (for 2 groups) or 380 Kruskall-Wallis followed by post-hoc Dunn's test (for more than two groups) were used. The null 381 hypothesis was rejected at p < 0.05. All the statistical results for principal figures are reported in Table  382 1. In order to predict cocaine-related target genes of the selected miRNAs, we used a combination of 402 databases, including miRBase, Miranda, TargetScan and Tarbase, crossed with literature data on 403 cocaine addiction. We identified 9 potential interesting targets: bdnf, fos-B, npas4, gria2, grin2b, 404 gria3, pi3K, drd1a and mef2c (Figure 1d) and studied expression levels of these mRNAs in the DS 405 and NAc after our protocol of sustained cocaine administration. Among these mRNAs, we only 406 observed a significant increase in the expression levels of fosB, npas4 and pi3k mRNAs in the DS, but 407 not in the NAc (Figure 1e).

Validation of fosB and npas4 mRNAs as targets of miR-1 410
In most cases, miRNAs interact with the 3′ UTR of target mRNAs to induce mRNA degradation or 411 translational repression [38]. We hypothesized that the absence of detectable up-regulation of fosB and 412 npas4 mRNAs by cocaine in the NAc might be due to cocaine-induced up-regulation of miRNA(s) 413 targeting these mRNAs in this region. Among the selected miRNAs, we focused on miR-1 because the 414 "in silico" prediction indicated that it targeted both fosB and npas4 mRNAs (Figure 1d). which are 415 key players in cocaine-induced long-term adaptations [41,42]. We experimentally tested that these 416 potential targets were directly bound by miR-1 ( We further aimed to assess whether miR-1 could repress bdnf, fosB and npas4 mRNA levels in vivo. 423 To this end, we used an AAV-mediated cell-type specific strategy allowing to overexpress miR-1 in 424 either D1R-SPNs or D2R-SPNs. We used a combination of two AAVs, with one expressing miR-1 in a 425  We then measured the effects of overexpression of miR-1 in cocaine-induced conditioned place 452 preference (CPP), which can be used as a proxy for the rewarding properties of cocaine. 453 Overexpression of miR-1 in D1-SPNs of the DS did not alter cocaine-induced CPP (Figure 3a, left  454 panel). We then extinguished the CPP by freely exposing the mice to cocaine-paired and unpaired 455 environments after saline injection and observed that miR-1 overexpression in D1-SPNs of the DS did 456 not modify CPP extinction (Figure 3a, middle panel). In contrast, the reinstatement of cocaine CPP induced by an acute injection of cocaine after the extinction period (15 mg/kg) was prevented by miR-458 1 overexpression in DS D1-SPNs (Figure 3a, right panel). 459 We measured the levels of miR-1 and its target genes in the DS of these mice 1h after the start of the 460 reinstatement test. As expected we found, high levels of miR-1 in both saline and cocaine miR-1-461 injected groups, while miR-1 levels were not altered in control groups (Figure 3b, left panel). miR-1 462 overexpression in D1-SPNs blunted fosB and npas4 mRNA up-regulation observed in the control 463 scramble group after cocaine-induced reinstatement of CPP (Figure 3b, right panel). miR-1 464 overexpression in D1-SPNs of the NAc did not alter either cocaine-induced CPP, its extinction or 465 cocaine-induced reinstatement (Figure 3c). As shown in Figure 4a  We then investigated whether miR-1 overexpression in D1-SPNs of the DS altered cocaine self-473 administration. MiR-1 overexpression in DS D1-SPNs did not modify the acquisition of self-474 administration, the motivation for cocaine under a progressive ratio of reinforcement (Figure 5a-b), 475 nor the extinction of this behavior (Figure 5c). Furthermore, the number of days required for 476 acquisition of cocaine self-administration, or for achievement of extinction was similar between 477 groups. However, miR-1 overexpression in D1-SPNs of the DS induced a significant increase in cue-478 induced reinstatement of cocaine seeking (Figure 5d), a result that contrasted with the blockade of 479 cocaine-induced reinstatement of CPP (see Figure 3a). Interestingly, when analyzing FosB expression 480 after cue-induced reinstatement of cocaine seeking, we found a slight decrease in the DS and an increase in the NAc core after miR-1 overexpression in the D1-SPNs of the DS (Figure 5e), a result 482 that suggests the existence of crosstalk between the DS and NAc. 483 MiR-1 overexpression in D2-SPNs of the DS did not alter the acquisition of cocaine-SA (Figure 5f) 484 nor its extinction (Figure 5h) or cue-induced reinstatement of cocaine seeking (Figure 5i). 485 Interestingly, it significantly reduced the breaking point for cocaine -SA under a progressive ration 486 schedule of reinforcement (Figure 5g), suggesting a reduced motivation for cocaine-SA. 487

DISCUSSION 488
Our data show that miR-1 is up-regulated in the NAc, but not the DS, by a 10-day regimen of 489 cocaine and that at least three mRNAs encoding proteins associated with cocaine-induced 490 neuronal plasticity, fosB, npas4 and bdnf are targeted by miR-1. We also demonstrate that housed rats increases operant responding for sucrose when motivated by hunger, but decreases 523 responding in satiated animals, in cocaine self-administration and cocaine-induced relapse. In our 524 study, the mice were housed five per cage, and we did not observe a significant increase of fosB 525 mRNAs in the NAc upon sustained cocaine administration, similarly to previous studies in rats 526 [53]. The absence of effect of miR-1 overexpression in D1-SPNs of the NAc on cocaine-induced 527 locomotor sensitization and CPP might thus be due to our environmental conditions. 528 Alternatively, we cannot exclude that ∆FosB overexpression previously observed in D1-SPNs of 529 the NAc [47-49] results from an accumulation of the stable protein rather than from de novo 530 gene expression, or that a partial decrease induced by miR-1 overexpression may not be 531 sufficient to prevent the behavior. 532 Another validated target of miR-1 was bdnf mRNA encoding brain-derived neurotrophic factor 533 Altogether, these observations can explain the lack of effect of miR-1 overexpression in bdnf 540 mRNA levels in vivo, despite a clear demonstration that miR-1 targets its 3'UTR region in vitro. 541 Further experiments, which are out of the scope of our study, could include overexpression of 542 miR-1 in prefrontal cortical neurons. 543 Finally, we demonstrate that npas4, is a target of miR-1 both in vitro and in vivo. npas4 encodes 544 an activity-dependent transcription factor that plays a key role in neuronal plasticity and is 545 In conclusion, our data show a specific and distinct role of miR-1 in cocaine-and cue-induced 585 relapses of two different cocaine-dependent behaviors after an extinction period and on 586 motivation for cocaine-SA. This contributes to the dissection of the mechanisms underlying addictive behaviors. Our results show the cell type-specific and region-specific role of miR-1 and 588 underlines differences between cocaine-and cue-induced reinstatement in non-operant and 589 operant paradigms, and thus highlights the complexity of the control of addictive behaviors and 590 emphasizes the importance of regional cell-type specific investigations. Schultz W. Multiple reward signals in the brain. Nature reviews Neuroscience. The 2way ANOVA indicated a significant effect of session but not of AAV and no interaction.