Persistent variations in neuronal DNA methylation following cocaine self-administration and protracted abstinence in mice

Abstract

Continued vulnerability to relapse during abstinence is characteristic of cocaine addiction and suggests that drug-induced neuroadaptations persist during abstinence. However, the precise cellular and molecular attributes of these adaptations remain equivocal. One possibility is that cocaine self-administration leads to enduring changes in DNA methylation. To address this possibility, we isolated neurons from medial prefrontal cortex and performed high throughput DNA sequencing to examine changes in DNA methylation following cocaine self-administration. Twenty-nine genomic regions became persistently differentially methylated during cocaine self-administration, and an additional 28 regions became selectively differentially methylated during abstinence. Altered DNA methylation was associated with isoform-specific changes in the expression of co-localizing genes. These results provide the first neuron-specific, genome-wide profile of changes in DNA methylation induced by cocaine self-administration and protracted abstinence. Moreover, our findings suggest that altered DNA methylation facilitates long-term behavioral adaptation in a manner that extends beyond the perpetuation of altered transcriptional states.

Keywords: DNA methylation; MBD; cocaine self-administration; genome-wide; neuron; next-generation sequencing; relapse.

Figure 1. Mice continue to seek cocaine during protracted abstinence

A two-way ANOVA revealed that mice press the previously cocaine-paired lever significantly more than the inactive lever after 1 and 21 days of forced abstinence (F_1,128 = 198.6, p<0.0001, IVSA (Day 10-12): n=34, IVSA 1: n=18 and IVSA 21: n=15). There was no difference in the number of cocaine-paired lever presses performed during the last 3 days of cocaine self-administration or after 1 or 21 days of forced abstinence (F_2,128 = 0.40, not significant). Data are displayed as mean lever presses ± SEM; IVSA 1: mice tested after 1 day of abstinence, IVSA 21: mice tested after 21 days of abstinence.

Figure 2. Representative heatmap of 5mC enrichment

All RMEs that were supported by enrichment (relative to background genomic coverage as identified by MACS) in at least 50% of biological replicates in naïve (n=5), IVSA 1 (n=7) or IVSA 21 (n=6) animals are plotted. White indicates that no biological replicates displayed enrichment for 5mC, while dark red indicates that all biological replicates had enrichment for 5mC at the given genomic locus. The ratio is indicative of the number of animals within a group that displayed an enrichment for 5mC at a given RME.

Figure 3. Validation of candidate DMRs and expression of co-localizing genes

In an independent cohort of animals, MBD qPCR was performed to assess the relative levels of 5mC enrichment at select DMRs and qPCR was used to examine expression of co-localizing genes. (a) In accordance with sequencing results, the locus within Golgb1 was demethylated following cocaine self-administration; F_3,17= 3.57, p<0.05. (b) The decreased DNA methylation was correlated with a significant reduction in the expression of Golgb1 (all isoforms) at all time points regardless of relapse testing; F_3,26 = 5.42, p<0.01. (c) The DMR located within Kctd16 displayed a near-significant trend towards persistent methylation in all treatment groups following cocaine self-administration; F_3,16 = 2.73, p=0.07. (d) Interestingly, a trend towards a significant decrease in Kctd16 gene expression was observed only after 21 days of abstinence and relapse testing; F_3,26 = 6.55, p<0.01, Holm-Sidak post hoc test IVSA 21 R vs. Naïve, p= 0.07. After 21 days of abstinence, Kctd16 expression was decreased in animals that underwent relapse testing (IVSA 21 R) compared to those that were simply sacrificed (IVSA 21 NR), p<0.05 (Holm-Sidak post hoc test), which suggests that the relapse test may influence the relationship between altered DNA methylation and gene expression. (e) Demethylation was also replicated at the intergenic locus located proximal to Snw1; F_3,17= 4.35, p<0.05. (f) Decreased DNA methylation was not associated with a significant change in the expression of Snw1; F_3,26 = 0.07. (g) Finally, demethylation of the Glra1-associated DMR was reproduced following 21 days of abstinence regardless of relapse testing; F_3,17 = 5.59, p<0.01. (h) Demethylation of the Glra1-associated DMR was associated with a trend towards a significant reduction in the expression of Glra1 (all isoforms) after 21 days of abstinence; Welch's F_3,13.36 = 7.52, p<0.01, Games-Howell post hoc tests. All data are displayed as mean ± SEM, and p-values are derived from Holm-Sidak post hoc tests relative to naïve animals, except where specifically indicated. * p<0.05, ** p<0.01.

Figure 4. Expression of genes co-localizing with DMRs that cannot be validated by MBD qPCR

(a) Cocaine self-administration yielded a persistent increase in DNA methylation at a locus within Mctp1 (b) Increased DNA methylation correlated with a concomitant decrease in the expression of Mctp1 (−001 isoform) across treatment groups; F_3,26 = 4.07, p<0.05. (c) Cocaine self-administration resulted in a persistent increase in DNA methylation within an intragenic region of Cpeb4. (d) Increased DNA methylation was associated with a long-lasting upregulation of Cpeb4 (001/201 isoform) expression; F_3,26 = 5.96, p<0.01. (e) A significant increase in DNA methylation within an intron of Nkain3 was observed after 21 days of abstinence (f) The increase in DNA methylation was associated with a decrease in the expression of Nkain3 (all isoforms) after 21 days of abstinence and relapse testing, but this decrease was not observed when the relapse test did not occur; F_3,26 =4.56, p<0.05, naïve vs. IVSA 21 R, p<0.05, IVSA 21 R vs. IVSA 21 NR, p<0.01. Data are displayed as mean ± SEM, all p-values are derived from Holm-Sidak post hoc tests relative to naïve animals, * p<0.05 ** p<0.01.

Figure 5. Regulation of alternative splicing and isoform-specific expression by intragenic DNA methylation

Intragenic modifications of DNA methylation arising during cocaine self-administration may contribute to the differential regulation of select splice variants. The isoforms selected include the major protein-coding isoform of the select gene and isoforms transcribed from the genomic region proximal to the DMR of interest. (a) There was a trend towards the overall decreased expression of Cdh13, F_3,26 =2.53, p=0.07. (b) However, this is likely due to the regulation of non-coding transcripts of Cdh13, as the expression of the protein-coding transcript was not significantly altered following cocaine self-administration; F_3,26 = 1.35, not significant. (c). Subsequent to self-administration, overall Cpeb4 expression increased in all treatment groups; F_3,26 = 5.96, p<0.01. (d) Nevertheless, the expression of the common isoform, Cpeb4-001, displayed a different pattern of expression relative to that in naïve animals, with the sole significant difference in expression being between animals subject to relapse testing at 21 days of abstinence compared to those that were not (F_3,26 = 4.32, p<0.01, IVSA 21 R vs. IVSA 21 NR, p<0.01, Holm-Sidak post hoc test). (e) When the collective expression of all protein-coding isoforms of Mctp1 was explored, the sole significant difference was again between animals subject to relapse testing at 21 days and those that were not (F_3,26 = 3.62, p<0.05, Holm-Sidak post hoc, IVSA 21 R vs. IVSA 21 NR, p<0.05). (f) However, when explored individually, the expression of one protein-coding isoform (Mctp1-001, Ensembl 37) was persistently decreased at all time points relative to naïve animals; F_3,26= 4.07, p<0.05. Data are displayed as mean ± SEM; all p-values are derived from Holm-Sidak post hoc tests relative to naïve animals except where indicated, * p<0.05, ** p<0.01.