CREB activity in dopamine D1 receptor expressing neurons regulates cocaine-induced behavioral effects

Abstract

It is suggested that striatal cAMP responsive element binding protein (CREB) regulates sensitivity to psychostimulants. To test the cell-specificity of this hypothesis we examined the effects of a dominant-negative CREB protein variant expressed in dopamine receptor D1 (D1R) neurons on cocaine-induced behaviors. A transgenic mouse strain was generated by pronuclear injection of a BAC-derived transgene harboring the A-CREB sequence under the control of the D1R gene promoter. Compared to wild-type, drug-naïve mutants showed moderate alterations in gene expression, especially a reduction in basal levels of activity-regulated transcripts such as Arc and Egr2. The behavioral responses to cocaine were elevated in mutant mice. Locomotor activity after acute treatment, psychomotor sensitization after intermittent drug injections and the conditioned locomotion after saline treatment were increased compared to wild-type littermates. Transgenic mice had significantly higher cocaine conditioned place preference, displayed normal extinction of the conditioned preference, but showed an augmented cocaine-seeking response following priming-induced reinstatement. This enhanced cocaine-seeking response was associated with increased levels of activity-regulated transcripts and prodynorphin. The primary reinforcing effects of cocaine were not altered in the mutant mice as they did not differ from wild-type in cocaine self-administration under a fixed ratio schedule at the training dose. Collectively, our data indicate that expression of a dominant-negative CREB variant exclusively in neurons expressing D1R is sufficient to recapitulate the previously reported behavioral phenotypes associated with virally expressed dominant-negative CREB.

Keywords: CREB; activity-dependent gene expression; addiction; cocaine-related behavior; dominant negative CREB; dopamine receptor D1.

Figure 1

Cell-type specificity of the A-CREB transgene. (A) Design of the transgene D1R-A-CREB BAC construct. This construct was inserted after the translational start of the gene encoding the dopamine D1 receptor in a BAC. (B) Southern blot of 3 founder lines of D1R-A-CREB with 1, 2, and 4 copies of the transgene. (C) Expression of the transgene in D1R-A-CREB mice in coronal brain sections. Brain slides were incubated with anti-Flag antibodies and a peroxidase-conjugated secondary antibody and stained with 3,3′-diaminobenzidine (overview). The D1-A-CREB construct was expressed in layer VI of the cortex (upper), striatum (middle) and NAc (lower). (D) Specific expression of A-CREB in dynorphin-expressing neurons as shown by double-immunoflorescence using anti-FLAG (green) and anti-dynorphin (red) show perfect overlay. Scale bars: (C) 30 μm; (D) 15 μm. cc, corpus callosum; aca, anterior commissure.

Figure 2

Expression of D1-A-CREB transgene causes no significant increase in apoptosis or cell loss. (A) Immunohistochemical staining of coronal sections with antibodies against the dopamine receptor D1 (D1R) and the neuronal marker NeuN revealed no loss of cells or decrease in D1R abundance. (B) In contrast, D1-A-CREB mice carrying two transgenes (T/T) show an increase of GFAP (brown) and cleaved caspase-3 staining in comparison to wild-type (+/+) and single-transgene mice (+/T). (C) Quantification of caspase-3 positive cells in striata of coronal section in wild-type, single transgene and double transgene mice. (D) Homozygous mice show reduced weight gain in comparison to wild-type and single transgene. Data are presented as mean + s.e.m., p-value of t-test (^***P < 0.001). Scale bars (A) 50 μm; (B) 70 μm (left panel), 30 μm (right panel), 15 μm (insert).

Figure 3

Effect of A-CREB transgene on the induction of activity-dependent genes. (A) The heatmap summarizes results from gene expression profiling in the striatum of wild-type (n = 5) and D1-A-CREB (n = 5) mice using Illumina MouseWG-6 v2 BeadChip arrays. The 40 transcripts included in the heatmap are the top 20 most significantly increased and 20 most decreased in abundance according to the signal-to-noise metric from GSEA 2.0. Each column represents one array and one animal. The color corresponds to a ratio of fold-change to standard error according to the scale shown below the heatmap. The results were clustered (neighbor-joining), Euclidean shortest distance correlations between transcript profiles are represented by the dendrogram on the left. (B) Validation by quantitative PCR (qPCR). The bars represent transcript abundance normalized to the levels observed in control animals. Expression of Arc, Egr2, and Fosb in naïve conditions in wild-type and D1-A-CREB mice. Abundance of the “house-keeping” Hprt transcript was identical in wild-type and D1-A-CREB mice. Data are presented as mean + s.e.m., P-value of t-test (^*P < 0.05).

Figure 4

Locomotor, anxiety- and depressive-like responses in D1R-A-CREB mice. (A) Spontaneous home cage locomotor activity measured by the e-motion system is indistinguishable between wild-type (n = 11) and D1-A-CREB mice (n = 8) genotype effect [F_{(1, 102)} = 0, p = 1]. Two-Way ANOVA indicates a phase effect [F_{(1, 102)} = 141.8, P < 0.0001] and all day points are significantly different from all night points Phase × time point interaction effect [F_{(2, 102)} = 18.5, P < 0.0001]. (B) Habituation to novelty in activity chambers in wild-type (n = 14) and D1-A-CREB (n = 13) mice. During the first 30 min exposure both genotypes displayed a progressive decrease in locomotor activation indicating habituation to novelty. Habituation effect [F_{(1, 125)} = 21.2, P < 0.001] (C) Preference for sucrose in wild-type (n = 6–7) and D1-A-CREB (n = 6–8) mice. Both genotypes displayed similar preferences for 1 or 5% sucrose solutions. ANOVA analysis did not indicate any significance for Genotype [F_{(1, 23)} = 0.1, P = 0.7], sucrose solutions [F_{(1, 23)} = 2.7, P = 0.1] or Genotype × solutions interaction [F_(1,23) = 3.8, P = 0.1] effects. (D,E) During the second exposure to the FST, both the latency to the first episode of immobility [t₍₂₆₎ = 0.6, P = 0.5] and the immobility time during the last 4 min test session [t₍₂₆₎ = 1.7, P = 0.1] was similar in all mice. (F) Anxiety-related behavior is not different between both genotypes during the Elevated plus-maze test. The time spent in the open arms of the maze is almost identical in both genotypes (wild-type n = 14, D1-A-CREB n = 11), and Two-Way ANOVA indicated no genotype [F_{(1, 157)} = 0.09, P = 0.7 but an arm effect: F_{(1, 157)} = 3.5, P < 0.001]. (G) Similarly, in the light-dark box test, the time spent in the light area is not different between wild-type (n = 13) and D1-A-CREB (n = 13) mice [t₍₂₆₎ = −1.5; P = 0.1]. (H,I) The number of entries into the open arm of the elevated plus maze and the lit part of the light dark box did not differ between genotypes [E: t₍₂₃₎ = 0.8; P = 0.4 and F: t₍₂₄₎ = −0.1; P = 0.9]. All data represent mean ± s.e.m. (^**) indicates P < 0.05 and 0.01 vs. 5 m.

Figure 5

Cocaine-induced behavioral effects in D1-A-CREB mice. (A) Cocaine induced a higher increase in locomotor activity in the activity box at doses of 10 and 20 mg/kg cocaine in D1-A-CREB (n = 9) mice compared with wild-type (n = 13) mice. Two-Way ANOVA indicated a Genotype effect F_{(1, 19)} = 27.5, P < 0.001, a Dose effect F_{(3, 57)} = 132.3, P < 0.001, and a Genotype × Dose interaction F_{(3, 57)} = 4.3, P < 0.01. (B,C) Cocaine-induced development and expression of behavioral sensitization. At the dose of 5 mg/kg (B), wild-type and D1-A-CREB mice showed development of cocaine sensitization [Sensitization effect F_{(5, 70)} = 29.9, P < 0.001], but D1-A-CREB mice expressed a significantly higher response to repeated cocaine injections and after a drug free interval (day 11) than the wild-type mice. Two-Way ANOVA indicated genotype [F_{(1, 14)} = 21.9, P < 0.0005] and Genotype × Sensitization [F_{(3, 70)} = 2.3, P < 0.05] effect. Similarly, at the dose of 10 mg/kg (C) both development and the expression of behavioral sensitization were stronger in D1-A-CREB than in wild-type mice. Two-Way ANOVA indicated Genotype [F_{(1, 16)} = 10.3, P < 0.005], Sensitization [F_{(5, 80)} = 69.7, P < 0.0001] and Genotype × Sensitization [F_{(5, 80)} = 2.8, P < 0.05] effects. (D,E) All mice exhibited a conditioned locomotion in the CPP boxes after the cocaine treatment [Two-Way ANOVA, Conditioning effect F_{(1, 14)} = 29.6; P < 0.001 and F_{(1, 14)} = 75.2; P < 0.001 for (D,E) respectively]. This conditioned response was higher in D1-A-CREB mutants compared with wild-type mice [Two-Way ANOVA indicated a Genotype effect F_{(1, 14)} = 9.9; P < 0.01 for (D) and a Genotype × Conditioning effect: F_{(1, 14)} = 21.9; P < 0.0005 for (E)]. (F,G) Cocaine-induced CPP, extinction and reinstatement. Although both genotypes exhibited a significant cocaine-induced CPP, the D1-A-CREB mice displayed an increased—though non-significant- preference for the cocaine-paired compartment compared to wild-type mice at the dose of 5 mg/kg, which became significant when the mutants were conditioned with the dose of 10 mg/kg [F_{(1, 16)} = 3.3, P < 0.05]. There were no genotype differences during the extinction test, as indicated by the similar reduction of the time spent in the cocaine-paired floor after extinction training. However, after extinction, a challenge injection of cocaine [3 and 7.5 mg/kg, i.p. in (D,E), respectively] induced an increased reinstatement of the CPP in the D1-A-CREB mice, as indicated by the significantly stronger CPP score Genotype × Score [F_{(2, 32)} = 6.4, P < 0.005] interaction effect. (H,I) Cocaine self-administration. (H) Both genotypes learned to self-administer cocaine as indicated by the number of reinforcements across 12 consecutive sessions with a 0.5 mg/kg per infusion training dose (wild-type n = 8 and D1-A-CREB n = 10 mice). Two-Way ANOVA indicated a time [F_{(11, 176)} = 2.809; P < 0.01] but not a Genotype [F_{(1, 16)} = 0.5, P > 0.05] or Genotype × time interaction [F_{(11, 176)} = 1.1, P > 0.05] effects. (I) The inactive lever pressing was not different between genotypes. Two-Way ANOVA indicated a Time effect [F_{(11, 176)} = 6, P < 0.01], but not a Genotype [F_{(1, 16)} = 1.1, P > 0.05] or Genotype × time interaction [F_{(11, 176)} = 0.7, P > 0.05] effects. Data represent mean ± s.e.m. ^*P < 0.05 compared with 0. ^#P < 0.05 compared with wild-type mice.

Figure 6

Effect of A-CREB transgene on cocaine-dependent induction of activity-dependent genes. The graphs show mean abundances (8 mutants and 13 controls) of activity-dependent transcripts in mice 1 h after 7.5 mg/kg i.p. cocaine injection at the start of the reinstatement test. Values are normalized to control levels. Significant difference P < 0.05 (t-test) between wild-type and D1-A-CREB mice is labeled with a “^*.”