Mice with altered gene activity in reward system cells show reduced fentanyl seeking

All procedures were approved by the Institutional Animal Care and Use Committee at the Pennsylvania State University College of Medicine (PSUCOM) and conducted in accordance with NIH guidelines for the use of laboratory animals. Drd1-cre^120Mxu mice on a C57BL/6 J background were obtained from The Jackson Laboratory (B6;129-Tg(Drd1-cre)120Mxu/Mmjax; Jax strain #024860; RRID:MMRRC_037156-JAX) [55]. At PSUCOM, Drd1-cre^120Mxu hemizygotes were bred with C57BL/6 J wildtype mice obtained from The Jackson Laboratory. Mice were given food and water ad libitum and housed in the PSUCOM vivarium on a 12:12 h light:dark cycle with lights on at 7:00. All mice were housed in corncob bedding and provided with nestlets. Mice were group housed until intravenous surgery, after which they were pair-housed across a perforated acrylic divider throughout self-administration to prevent isolation stress. All experiments used mice of both sexes aged 8–10 weeks at the start of behavior.

All procedures are detailed in. Supplemental

Data were analyzed with Graphpad Prism 10 (La Jolla, CA) and JASP [71]. In the absence of significant sex effects or interactions, we collapsed data by sex. Any repeated measures data (e.g. self-administration, locomotion) were analyzed with repeated-measures ANOVA (RM-ANOVA) employing a Greenhouse-Geisser sphericity correction, with sex and genotype as between-subjects factors. Seeking data were analyzed with ANOVA using sex and genotype as between-subjects factors, and nose-poke as within-subject factor. Data without repeated measures (e.g. conditioned place preference, stereotypy) were analyzed by unpaired t-test due to no sex effects, or ANOVA when sex was significant. Gene expression data were analyzed with 2-way ANOVA using sex and genotype. For chemogenetic experiments, seeking was analyzed with ANOVA using sex, genotype, and virus as between factors, and nose-poke as within-subject factor. c-fos expression was analyzed with Welch’s ANOVA due to unequal variance. All post-hoc tests employed Sidak’s correction, except for the chemogenetic conditioned place preference and c-fos experiments which employed Dunnett’s comparing to the mCherry control.

We next asked if Drd1-cre^120Mxu mice exhibited fentanyl seeking under extinction conditions. As expected, after 24 h abstinence, wildtype mice exhibited fentanyl seeking behavior marked by increased responding at the active nose-poke. By contrast, Drd1-cre^120Mxu mice exhibited reduced fentanyl seeking behavior (Fig. 1E, 2-way ANOVA, Genotype: F_1,71 = 5.2, p = 0.02; Active nose-poke, wildtype vs. Drd1-cre, p = 0.006, Sidak’s). We tested a separate set of mice after 14 d abstinence. Like the 24 h timepoint, Drd1-cre^120Mxu mice exhibited reduced fentanyl seeking behavior relative to wildtype mice (Fig. 1F, 2-way ANOVA, Genotype: F_1,64 = 4.23, p = 0.04; Active nose-poke, wildtype vs. Drd1-cre^120Mxu, p = 0.03, Sidak’s). To see if this generalized to other forms of fentanyl reward, we used conditioned place preference (CPP) in a separate set of mice (Timeline in Fig. 1G). We found slightly reduced CPP scores in Drd1-cre^120Mxu mice relative to wildtype (Fig. 1H, t_1,126 = 2.17, p = 0.03). When we examined fentanyl-induced hyperlocomotion, another proxy for drug reward, we found Drd1-Cre^120Mxu mice exhibit greater movement counts during fentanyl conditioning (Fig. 1I, drug x genotype: F_1,109 = 13.22 p = 0.0004; wildtype vs. Drd1-cre^120Mxu fentanyl, p = 0.0047, Sidak’s). To determine if reduced drug reward was restricted to fentanyl, we asked if this generalized to other drugs with misuse potential, and compared cocaine self-administration and seeking in Drd1-cre^120Mxu mice relative to wildtype mice in our published dataset [77] (Supplemental Fig. 1A). Unlike fentanyl, Drd1-cre^120Mxu had reduced cocaine intake relative to wildtype mice (Supplemental Fig. 1B, C, Infusions: Genotype: F_1,24 = 10.8, p = 0.003; Active responses: Genotype F_1,24 = 17.3, p = 0.0003). Like fentanyl, Drd1-cre^120Mxu had fewer active responses during the non-reinforced seeking test at 24 h (Supplemental Fig. 1D, Genotype: F_1,48 = 22.6, p < 0.0001; Active responses, wildtype vs. Drd1-cre^120Mxu, p = 0.0001, Sidak’s). Finally, to determine if Drd1-cre^120Mxu mice had generally impaired reward seeking or if this behavior was specific to misused drugs, we compared sucrose self-administration and seeking in Drd1-cre^120Mxu mice to wildtype mice (Supplemental Fig. 2A). We found that Drd1-cre^120Mxu mice had no difference in number of sucrose pellets earned or consumed (Supplemental Fig. 2B, RMANOVA, Genotype F_1,22 = 0.2; 2 C F_1,22 = 0.8, p’s>0.3) and no differences in nose-poke discrimination (Supplemental Fig. 2D, E, Genotype F_1,22 = 0.2, p = 0.7) relative to wildtype littermates. D1-cre^120Mxu mice also exhibited similar sucrose seeking as wildtype littermates under both food-restriction and ad libitum conditions (Supplemental Fig. 2F, G, Genotype F’s< 0.21, p’s>0.6).

We showed that Drd1-cre^120Mxu mice have reduced self-administration of cocaine but not fentanyl. Our study is not the first to report on a transgenic line with abnormal drug self-administration behavior; for example, ChAT-Cre mice (B6.FVB(Cg)-Tg(Chat-cre)GM60Gsat/Mmucd) have reduced self-administration of intravenous nicotine [73]. However, our findings are unique in that despite normal acquisition of fentanyl self-administration, Drd1-cre^120Mxu mice exhibited attenuated nonreinforced fentanyl seeking. This effect is strikingly robust: the blunted fentanyl seeking is observed at both 24 h and 14 days after self-administration, is also reflected in a reduced preference for a fentanyl-paired context, persists in both paradigms despite bidirectional chemogenetic manipulation of D1-MSNs, and it is specific to drugs as there is no impairment of sucrose seeking in these mice.

In wildtype mice, we found chemogenetic stimulation of putative NAc core D1-MSNs (ventral mesencephalon-projecting) attenuated fentanyl seeking. While reminiscent of D1R agonists decreasing cocaine seeking [82], this was also surprising, since research by O’Neal et al. shows activating VTA-projecting NAc neurons exacerbates heroin seeking in rats [83]. This discrepancy may reflect species differences in D1-MSN collaterals to the ventral pallidum (VP). In mice, over 90% of VTA-projecting MSNs send collaterals to VP [84]; in rats, O’Neal et al. report VTA-projecting MSNs do not send collaterals to VP [83]. MSNs make inhibitory synapses on mostly GABAergic neurons in VP [85]. Activating MSN collaterals in VP would increase inhibitory drive on GABAergic VP neurons, and inhibiting VP GABA neurons attenuates remifentanil seeking [86]. Thus, we suspect reduced fentanyl seeking with NAc core D1-MSN activation is explained by inhibition in VP, although this requires further investigation. It is worth noting these mechanisms may not extend to the NAc shell. For example, systemic D1R agonism attenuates cocaine seeking [82], but when delivered directly to the NAc shell, D1R agonists instead promotes cocaine seeking [87]. Since DREADDs expression in our study was limited to NAc core, future work is necessary to elucidate the specific contributions of NAc subregions to fentanyl seeking.

Unlike wildtype mice, we found that chemogenetic manipulation of D1-MSNs did not affect seeking in Drd1-cre^120Mxu mice. It is likely D1-MSN stimulation had no effect in Drd1-cre^120Mxu mice because their seeking responses are so low at baseline, they cannot be further attenuated (a “floor effect”); additionally, it is perhaps unsurprising that chemogenetic inhibition of D1-MSNs did not increase seeking in Drd1-cre^120Mxu mice since this did not increase seeking in wildtype mice. Importantly, the lack of response in Drd1-cre^120Mxu mice does not reflect an inability for DREADDs to stimulate or inhibit D1-MSNs, as we identified expected increases and decreases in cfos expression after DCZ treatment. However, the DREADDs manipulations in wildtype and Drd1-cre^120Mxu are not identical. In Drd1-cre^120Mxu mice, we expressed DREADDs in all NAc core D1-MSNs; in wildtype mice, we targeted only ventral mesencephalon-projecting MSNs in NAc core. While these neurons almost exclusively express D1 receptor, there may be minor differences in the populations targeted by both approaches, given the collateralization discussed above.

It is worth noting that neither Drd1-cre^120Mxu nor wildtype mice demonstrated increased fentanyl seeking after a longer period of abstinence, a phenomenon termed “incubation of craving” that is well-documented for other misused drugs [88, 89]. A potential explanation is the time course of our seeking experiments. Indeed, incubation of heroin seeking follows an inverted U-shaped time course, with higher responding at 6 and 12 days compared to 1 and 25 [90]. The specific time course for incubation of fentanyl seeking in mice is unknown, although we suspect it arises later than the 14 day timepoint tested here, since rats demonstrate incubation of fentanyl seeking after 30 days [91], and mice exhibit high rates of fentanyl seeking after 4 weeks [92]. Additional work is necessary to fully understand the nuance of time-dependent changes to fentanyl seeking behavior, and whether this differs between genotypes.

We showed that Drd1-cre^120Mxu mice exhibit several locomotor phenotypes, including increased locomotor response to fentanyl, reduced locomotion in a novel environment, and increased sensitivity to a D1R agonist. There are several reports on transgenic mouse lines with abnormal locomotor responses to misused drugs. For example, Drd2-EGFP mice [93] have increased striatal Drd2 expression and reduced locomotor response to cocaine [72]. Similarly, DAT-Ires-Cre mice (B6.SJL-Slc6a3^{tm1.1(cre)Bkmn}/J, Jackson Laboratory No: 006660) have reduced dopamine transporter (DAT) expression [94] and function [95], and exhibit attenuated amphetamine-induced locomotion [74] that generalizes to a different DAT-Cre line (Slc6a3^tm1(cre)Xz, Jackson Laboratory No: 020080) [75]. Here we show Drd1-cre^120Mxu mice have elevated expression of NAc Drd1 and a greater locomotor response to fentanyl. D1-MSNs form inhibitory synapses on GABAergic interneurons in VTA [96, 97]. Increased D1R activation may potentiate dopamine neuron disinhibition, increasing the locomotor response to fentanyl. In support of increased D1R sensitivity, Drd1-cre^120Mxu mice exhibited a relatively stronger locomotor stimulatory effect with D1R agonist SKF-38393. Interestingly, Drd1-cre^120Mxu mice exhibit baseline hypolocomotion in a novel environment. Altered baseline locomotion is similar to observations in the Drd2-EGFP and the DAT-IRES-Cre line, which exhibit baseline hyperactivity in a novel environment [72, 74] and dysregulated dopamine regulation mechanisms [72, 95]. These baseline locomotor differences are to be expected given the importance of D1- and D2-MSN signaling for normal locomotion [98]. Importantly, the increase in Drd1 does not reflect an increase in the number of D1-MSNs relative to D2-MSNs in NAc, as our viral labeling indicates there are an equivalent number of D1 positive and negative neurons in Drd1-cre^120Mxu mice.

Although the precise mechanisms underlying reduced fentanyl seeking in Drd1-cre^120Mxu mice are unknown, looking at the transcriptional differences provides some insight. At baseline, we found Drd1-cre^120Mxu mice had greater Drd1 expression relative to wildtype mice. Increased Drd1 was maintained following fentanyl self-administration in Drd1-cre^120Mxu mice, resulting in an exaggerated genotype difference, since fentanyl-experienced wildtype mice downregulated Drd1 expression. Tac1, another D1-MSN marker, demonstrated baseline sex differences that are consistent with findings in rat striatum [99], but did not initially differ between genotypes. Following fentanyl, wildtype mice upregulated Tac1, similar to findings in morphine-exposed rats [51]. In stark contrast, Drd1-cre^120Mxu mice downregulated Tac1 after fentanyl. Tac1 encodes the precursor for Substance P, a neuropeptide that is induced by morphine [100] and particularly relevant to opioid-dependent behaviors [101]. Substance P is released by local MSN collaterals [102] onto cholinergic interneurons expressing Substance P receptors [103, 104], where it participates in an excitatory microcircuit that promotes coordinated activity between D1- and D2-MSNs [105]. The downregulation of Tac1 in the NAc of Drd1-cre^120Mxu mice likely impairs the function of this local circuit, and as such, may contribute to the blunted fentanyl seeking in these mice.

We also looked at NAc expression of D2-MSN markers, including Drd2, Gpr6, Adora2a, and Penk. While expression of these genes did not differ between Drd1-cre^120Mxu and wildtype mice at baseline, fentanyl-experienced Drd1-cre^120Mxu mice had increased Gpr6 expression in both sexes, and increased Adora2a expression in males. Gpr6, which encodes G protein-coupled receptor 6, is enriched in NAc indirect pathway MSNs, where it plays a role in regulating instrumental conditioning [10], but a link to drug misuse has not yet been identified. Adora2a encodes adenosine A_2A receptors, which are colocalized on D2-MSNs [106, 107] and antagonize activity at the D2 receptor [108]. NAc A_2A activity has been shown to regulate cocaine seeking [109], but its role in opioid use is less studied. Two of the most promising molecular targets are Penk (encoding the enkephalin precursor) and Oprd1 (encoding delta opioid receptor, which binds enkephalin). The enkephalinergic system has close interactions with reward pathways: enkephalins released from D2-MSNs modulate neurotransmission in the VP, and enkephalins released locally in NAc modulate NAc signaling at various sites [110]. Prior studies in rats implicated a direct role for NAc Penk in opioid seeking, where Penk overexpression in NAc potentiates heroin seeking [111], consistent with upregulated Penk in our fentanyl-seeking wildtype mice. Since Drd1-cre^120Mxu mice downregulate Penk, and upregulate Oprd1 only after fentanyl, these adaptations could drive decreased fentanyl seeking and warrant further investigation.

In addition to the MSN markers, we also investigated other genes that, in the NAc, have known roles in opioid misuse. We found that Ntrk2, which encodes BDNF receptor TrkB, was downregulated following fentanyl abstinence in wildtype mice but was unchanged in Drd1-cre^120Mxu mice. This finding is consistent with previous work showing morphine exposure downregulates TrkB in D1-MSNs, resulting in impaired BDNF signaling that promotes morphine reward [17]. It is tempting to speculate that, given unchanged Ntrk2 expression, Drd1-cre^120Mxu mice preserve normal BDNF signaling in D1-MSNs after fentanyl, thereby reducing fentanyl seeking. We also found Drd1-cre^120Mxu mice have blunted downregulation of Grin2a and Grin2b following fentanyl abstinence. These genes encode the GluN2 subunits of the glutamatergic NMDA receptor. NAc glutamatergic signaling is essential for opioid seeking [112], and previous work shows chronic opioid use alters NMDA-mediated signaling in NAc [81]. The observed downregulation of NMDA receptor components in wildtype mice aligns with findings of an increased NAc AMPA/NMDA ratio that results from chronic morphine use and persists throughout protracted withdrawal [32, 35]. This adaptation is believed to contribute to an increased excitatory drive onto D1-MSNs that promotes drug seeking [113]. A similar adaptation results from fentanyl abstinence [52], and this process is likely disrupted in Drd1-cre^120Mxu mice, thereby impairing fentanyl seeking.