cAMP-mediated upregulation of HCN channels in VTA dopamine neurons promotes cocaine reinforcement

Abstract

Chronic cocaine exposure induces enduring neuroadaptations that facilitate motivated drug taking. Hyperpolarization-activated cyclic nucleotide-gated (HCN) channels are known to modulate neuronal firing and pacemaker activity in ventral tegmental area (VTA) dopamine neurons. However, it remained unknown whether cocaine self-administration affects HCN channel function and whether HCN channel activity modulates motivated drug taking. We report that rat VTA dopamine neurons predominantly express Hcn3-4 mRNA, while VTA GABA neurons express Hcn1-4 mRNA. Both neuronal types display similar hyperpolarization-activated currents (I_h), which are facilitated by acute increases in cAMP. Acute cocaine application decreases voltage-dependent activation of I_h in VTA dopamine neurons, but not in GABA neurons. Unexpectedly, chronic cocaine self-administration results in enhanced I_h selectively in VTA dopamine neurons. This differential modulation of I_h currents is likely mediated by a D₂ autoreceptor-induced decrease in cAMP as D₂ (Drd2) mRNA is predominantly expressed in dopamine neurons, whereas D₁ (Drd1) mRNA is barely detectable in the VTA. Moreover, chronically decreased cAMP via Gi-DREADD stimulation leads to an increase in I_h in VTA dopamine neurons and enhanced binding of HCN3/HCN4 with tetratricopeptide repeat-containing Rab8b-interacting protein (TRIP8b), an auxiliary subunit that is known to facilitate HCN channel surface trafficking. Finally, we show that systemic injection and intra-VTA infusion of the HCN blocker ivabradine reduces cocaine self-administration under a progressive ratio schedule and produces a downward shift of the cocaine dose-response curve. Our results suggest that cocaine self-administration induces an upregulation of I_h in VTA dopamine neurons, while HCN inhibition reduces the motivation for cocaine intake.

Fig. 1. Expression of Hcn1–4 mRNA in rat VTA dopamine and GABA neurons.

Representative 63x images of VTA sections labeled with RNAscope in situ hybridization for Hcn1–4 (red), Gad1 (GABA neuron, yellow) and Slc6a3 (dopamine neuron, green) mRNA. A, B Hcn1 and Hcn2 mRNA were highly expressed in Gad1 neurons, but barely detectable in Slc6a3 neurons. Hcn2 mRNA was also expressed on a cell type that remains to be detected. C, D Hcn3 and Hcn4 mRNA were expressed in both Gad1 neurons and Slc6a3 neurons. E Left, percentage of VTA dopamine neurons co-expressing Hcn1–4 (Hcn1: 208 of 2245 neurons; Hcn2: 286 of 2350 neurons; Hcn3: 1591 of 1661 neurons; Hcn4: 1109 of 2392 neurons). Right, the number of Hcn1–4 mRNA puncta expressed in VTA dopamine neurons. F Left, percentage of VTA GABA neurons co-expressing Hcn1–4 (Hcn1: 428 of 619 neurons; Hcn2: 628 of 736 neurons; Hcn3: 509 of 545 neurons; Hcn4: 514 of 634 neurons. Right, the number of Hcn1–4 mRNA puncta expressed in VTA GABA neurons. (n = 8 to 14 imaged sections from 4 rats).

Fig. 2. I_h currents were not significantly different between VTA dopamine and GABA neurons and were sensitive to cAMP stimulation.

A Left: Voltage protocol for recording I_h current. Right: Representative I_h traces recorded from dopamine (DA) neurons and GABA neurons in the VTA. B There was no significant difference in I_h between dopamine and GABA neurons at all corresponding hyperpolarization potentials (two-way RM ANOVA, cell-type, F_1,29 = 0.02, p = 0.894; holding potential, F_7,203 = 37.7, p < 0.001; cell-type ⨯ holding potential interaction, F_7,203 = 0.6, p = 0.770; DA neuron, n = 18 cells; GABA neuron, 13 cells; n = 4 rats). C I_h amplitude was calculated by subtracting the instantaneous current from the steady-state current achieved during the voltage step at −130 mV, and no significant difference was detected (t-test, t₂₉ = 0.5, p = 0.640, n = 13–18 cells). D There was no significant difference in membrane capacitance (C_m) between dopamine and GABA neurons (t-test, t₂₉ = 0.4, p = 0.728, n = 13–18). E I_h current density (=I_h amplitude at −130 mV/C_m) was not significantly different between dopamine neurons and GABA neurons (t-test, t₂₉ = 0.5, p = 0.655, n = 13–18). F I_h activation curves in dopamine neurons and GABA neurons were generated by the tail current protocol. Tail current amplitudes were fitted with a Boltzmann function. G There was no significant difference in the midpoint activation voltage (V_1/2) between dopamine and GABA neurons (t-test, t₂₉ = 1.0, p = 0.322). H, I Increasing cAMP (cAMP ↑ ) via bath perfusion of forskolin (20 µM) and rolipram (1 µM) led to a significant rightward shift in the I_h activation curve of and a significant depolarizing shift in the V_1/2 of VTA dopamine neurons (paired t-test t₆ = 6.4, p < 0.001, n = 7 from 3 rats). J, K Increasing cAMP led to a significant depolarizing shift in the I_h activation curve and a depolarizing shift in the V_1/2 of VTA GABA neurons (paired t-test, t₇ = 3.9, p = 0.006, n = 8 from 3 rats). ns, not significant, p > 0.05, **p < 0.01, ***p < 0.001.

Fig. 3. Bath application of cocaine induced a hyperpolarizing shift of the V_1/2 of I_h currents in VTA dopamine neurons through activation of the D₂-autoreceptor.

A, B Bath perfusion of cocaine (10 μM) induced a significant leftward shift in the I_h activation curve and a hyperpolarizing shift of the V_1/2 of VTA dopamine neurons (paired t-test, t₆ = 6.3, p < 0.001, n = 7 from 3 rats). C, D The shifts of cocaine on the I_h activation curve and the V_1/2 were blocked by D₂ receptor antagonist sulpiride (1 μM) (paired t-test, t₅ = 0.5, p = 0.641, n = 6 from 3 rats). E, F Drd2 mRNA was abundantly expressed and completely colocalized with Slc6a3 mRNA but was expressed only in ~10% of GABA neurons (ns, not significant, p > 0.05, ***p < 0.001; n = 2 rats). G Drd1 mRNA was barely detected in both dopamine and GABA neurons in the VTA (n = 2 rats).

Fig. 4. Cocaine self-administration led to upregulation of I_h in VTA dopamine neurons.

A Timeline of jugular vein catheterization, cocaine self-administration and electrophysiology. B Representative I_h currents recorded from VTA dopamine neurons after 10 days of yoked saline (YS), cocaine self-administration (SA) and yoked cocaine infusions (YC). C Compared with yoked saline infusions, cocaine self-administration and yoked cocaine infusions led to significant increases in I_h currents at corresponding hyperpolarization potentials (two-way RM ANOVA, Cocaine, F_2,40 = 8.6, p < 0.001; holding potential, F_7,280 = 248.9, p < 0.001; Cocaine ⨯ holding potential interaction, F_14,280 = 5.5, p < 0.001; n = 14–15 from 5 rats; yoked saline vs cocaine SA, blue * and ^#, yoked saline vs yoked cocaine, red * and ^#; *p < 0.01, ^#p < 0.001). D–F Cocaine self-administration and yoked cocaine infusions led to significant increases in I_h amplitude (D; one-way ANOVA, F_2,42 = 6.5, p = 0.004, n = 14-15) and I_h density (F; one-way ANOVA, F_2,42 = 5.1, p = 0.010, n = 14–15 cells from 5 rats), but did not significant change the membrane capacitance (C_m) of VTA dopamine neurons (E; one-way ANOVA, F_2,42 = 0.5, p = 0.587, n = 14–15 cells). G, H Cocaine self-administration and yoked cocaine infusions led to depolarizing shifts in the I_h activation curve the V_1/2 of VTA dopamine neurons (one-way ANOVA, F_2,42 = 4.9, p = 0.012, n = 14–15 cells). I Left: A schematic diagram shows the location of the stimulating electrode (S) and recording electrode (R). Representative EPSPs (50 Hz ⨯ 5) showing temporal summation before and after application of ZD7288. Right, Comparison of changes in temporal summation of EPSPs (EPSP5/EPSP1 ratio) in cocaine self-administration, yoked cocaine and yoked saline groups in the absence and presence of ZD7288 (two-way RM ANOVA, Cocaine, F_2,24 = 0.2, p = 0.856; ZD7288, F_1,24 = 172.0, p < 0.001; Cocaine ⨯ ZD7288 interaction, F_1,24 = 18.1, p < 0.001; n = 9 cells from 4 rats). J–M Cocaine self-administration did not alter the I_h density (J, K; t-test, t₂₅ = 0.4, p = 0.716) and V_1/2 (L, M; t-test, t₁₈ = 0.5, p = 0.597) of GABA neurons in the VTA. For (D–I), *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 5. Chronically decreasing cAMP with hM4D(Gi) induced an up-regulation of I_h in VTA dopamine neurons.

A Timeline of viral injection, yoked DCZ infusions and electrophysiology. B AAV8-hSyn-DIO-hM4D(Gi)-mCherry [hM4D(Gi)] or AAV8-hSyn-DIO-mCherry (mCherry) was bilaterally microinjected into the VTA of TH-Cre rats. Immunohistochemistry showed that hM4D(Gi) was expressed in the majority of TH⁺ dopamine neurons (green) but was not expressed in TH^- neurons in the VTA of TH-Cre rats (n = 5 rats). C Representative I_h currents recorded from rats that received yoked DCZ infusions expressing mCherry or hM4D(Gi) in VTA dopamine neurons. D I_h amplitude was significantly increased in rats with expressing hM4D(Gi) compared to mCherry at the corresponding hyperpolarization potentials (two-way RM ANOVA, hM4D(Gi), F_1,21 = 2.7, p = 0.117; holding potential, F_7,147 = 118.2, p < 0.001; hM4D(Gi) ⨯ holding potential interaction, F_7,147 = 3.1, p = 0.005; n = 11–12 cells from 4 rats). E I_h current density was significantly increased in hM4D(Gi)-expressing rats compared with mCherry-expressing rats (t-test, t₂₁ = 2.1, p = 0.049, n = 11–12 cells). F, G Yoked DCZ infusions led to a significant depolarizing shift of the V_1/2 in hM4D(Gi)-expressing rats compared to mCherry-expressing rats (t-test, t₂₁ = 2.4, p = 0.023, n = 11–12 cells). H Representative temporal summation of evoked EPSPs (50 Hz ⨯ 5) recorded from mCherry and hM4D(Gi)-expressing VTA dopamine neurons in TH-Cre rats that received yoked DCZ infusions. I Compared with mCherry-expressing VTA dopamine neurons, hM4D(Gi)-expressing dopamine neurons exhibited a smaller increase in temporal summation of EPSPs. ZD7288 induced greater increase in temporal summation of EPSPs in hM4D(Gi)-expressing rats (two-way RM ANOVA, hM4D(Gi), F_1,11 = 0.004, p = 0.856; ZD7288, F_1,11 = 56.1, p < 0.001; hM4D(Gi) ⨯ ZD7288 interaction, F_1,11 = 16.4, p = 0.002; n = 6–7 cells from 3 rats). J, K Representative Co-IP reaction was performed using anti-HCN3 (J) or HCN4 (K) antibodies in VTA lysate from rats that expressed mCherry and hM4D(Gi) in VTA dopamine neurons and received yoked DCZ administration. IP, immunoprecipitation; IB, immunoblotting. (HCN3: t-test, t₁₈ = 2.2, p = 0.044, n = 9 and 9. HCN4: t-test, t₁₈ = 2.3, p = 0.034, n = 9 and 9). For the entire figure, *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 6. Ivabradine dose-dependently decreased cocaine intake under FR/PR schedules and produced a downward shift of the cocaine dose-response curve.

A Timeline of catheterization, cocaine self-administration and ivabradine treatment. B Intra-VTA infusion of Ivabradine produced a significant downward shift in the number of cocaine infusions on the cocaine dose-response curve (two-way RM ANOVA: Ivabradine, F_2,21 = 13.0, p < 0.001; cocaine dose, F_6,126 = 22.0, p < 0.001; Ivabradine ⨯ cocaine dose interaction, F_6,126 = 2.5, p = 0.005; 25 ng vs 0 ng, red *; 50 ng vs 0 ng, blue *; 25 ng vs 50 ng, red ^#; n = 8 rats for each group). C Intra-VTA infusion of significantly attenuated the number of cocaine infusions under PR reinforcement conditions (one-way ANOVA, F_2,21 = 9.1, p = 0.001, n = 8 rats in each group). D Intra-VTA infusion of significantly attenuated the PR breakpoint (one-way ANOVA, F_2,21 = 9.7, p = 0.001, n = 8 rats in each group). E, F Systemic ivabradine administration (0, 3, 10 mg/kg, i.p.) following iv elacridar (5 mg/kg), dose-dependently decreased the mean number of active lever presses (E, one-way ANOVA: F_2,23 = 27.3, p < 0.001) and cocaine infusions under an FR2 reinforcement schedule (F, one-way ANOVA: F_2,23 = 35.6, p < 0.001) (n = 8 rats in each group). G, H Ivabradine pretreatment produced a significant downward shift in the dose-response curve for cocaine infusions (G, two-way RM ANOVA: ivabradine pretreatment, F_2,21 = 16.9, p < 0.001; cocaine dose, F_6,126 = 28.0, p < 0.001; ivabradine pretreatment × cocaine dose interaction, F_12,126 = 6.3, p < 0.001; 3 vs. 0 mg/kg ivabradine, red *; 10 vs. 0 mg/kg ivabradine, green *; 3 vs. 10 mg/kg ivabradine black ^#) and decreased total cocaine intake (H, two-way RM ANOVA: ivabradine pretreatment, F_2,21 = 5.8, p = 0.010; cocaine dose, F_6,126 = 68.7, p < 0.001; ivabradine × cocaine dose interaction, F_12,126 = 1.6, p = 0.090; 10 vs. 0 mg/kg ivabradine green *; 3 vs. 10 mg/kg ivabradine red *) on the dose-response curve (n = 8 rats in each group). I, J Ivabradine dose-dependently decreased the number of cocaine infusions (I, one-way ANOVA: F_2,23 = 34.2, p < 0.001; n = 8 rats in each group) and breakpoint (J, Brown-Forsythe equal variance: p < 0.05; Kruskal-Wallis one-way ANOVA on ranks: ivabradine, H = 17.6, p < 0.001; n = 8 rats in each group) under a PR reinforcement schedule. *p < 0.05, **p < 0.01, ***p < 0.001; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001.