Sorting nexin 27 regulation of G protein-gated inwardly rectifying K⁺ channels attenuates in vivo cocaine response

Abstract

The subcellular pathways that regulate G protein-gated inwardly rectifying potassium (GIRK or Kir3) channels are important for controlling the excitability of neurons. Sorting nexin 27 (SNX27) is a PDZ-containing protein known to bind GIRK2c/GIRK3 channels, but its function in vivo is poorly understood. Here, we investigated the role of SNX27 in regulating GIRK currents in dopamine (DA) neurons of the ventral tegmental area (VTA). Mice lacking SNX27 in DA neurons exhibited reduced GABABR-activated GIRK currents but had normal Ih currents and DA D2R-activated GIRK currents. Expression of GIRK2a, an SNX27-insensitive splice variant, restored GABABR-activated GIRK currents in SNX27-deficient DA neurons. Remarkably, mice with significantly reduced GABABR-activated GIRK currents in only DA neurons were hypersensitive to cocaine and could be restored to a normal locomotor response with GIRK2a expression. These results identify a pathway for regulating excitability of VTA DA neurons, highlighting SNX27 as a promising target for treating addiction.

Figure 1. Viability of mice with targeted disruption of SNX27 expression in VTA DA neurons

(A) Schematic shows conditional deletion of exon 3 in Snx27 occurs in DAT-expressing cells. Floxed animals with SNX27 exon 3 flanked by loxP sites were crossed to DAT-Cre^+/− line. SNX27 exon 3 is excised by Cre in DAT-expressing dopamine neurons, referred to as SNX27_DA KO mice. (B,C) Dual immunofluorescence for SNX27 (red) and TH (green) in coronal sections from WT and SNX27_DA KO. (D) Immunofluorescence for TH in coronal sections from WT and SNX27_DA KO. No apparent difference in TH+ density (7.5± 0.99 for WT, n=2 and 6.0 ± 0.26 for KO, n=2 per 150µm². (E) Resting membrane potentials were indistinguishable in DA neurons from WT, DAT-Cre^+/− and SNX27_DA KO mice. (F) I_h currents were indistinguishable in DA neurons from WT, Dat-Cre^+/− and SNX27_DA KO mice.

Figure 2. Reduced GABA_BR-GIRK currents in VTA DA neurons lacking SNX27 protein

(A,C) Whole-cell recordings from VTA DA neurons show baclofen-activated GIRK currents (I_Baclofen; 300 µM) from WT and SNX27_DA KO mice. Extracellular Ba²⁺ (1 mM), a selective inhibitor of inwardly rectifying potassium channels (K_ir), decreases I_Baclofen. Outward current measured at −50 mV is plotted as a function of time. (B,D) Scatter plot of I_Baclofen for indicated DA neurons. **p<0.01, one-way ANOVA with Bonferroni post hoc test.

Figure 3. Loss of GABA_BR-dependent inhibition of firing in DA neurons of SNX27_DA KO mice

(A) Current clamp recording shows action potentials in a wild-type DA neuron (200pA injection) before (ACSF) and after application of 300µM baclofen (+Baclofen, red trace). (B) Input-output plot shows firing frequency increases with larger current injections (solid circles, ACSF; n=16) and suppression with baclofen (open circles) (p<0.01 at 100–300pA, 2-way ANOVA with Bonferroni post hoc test). (C) Current clamp recording from DA neuron (200pA injection) before (ACSF) and after application of 300µM baclofen (+Baclofen, red trace). (D) Firing frequency is plotted as a function of current injection before (solid triangles, ACSF; n=24) and after baclofen (open triangles). Input-output plot shows loss of baclofen-dependent inhibition of firing at all current injection levels (n.s. all steps, 2-way ANOVA).

Figure 4. Attenuation of direct G protein-activation of GIRK channels in VTA DA neurons of SNX27_DA KO mice

(A,B) Outward currents recorded with 100µM GTPγS in the recording pipet from WT or SNX27_DA KO DA neuron. A time-dependent increase in outward current occurs with GTPγS, which is inhibited by extracellular Ba²⁺, consistent with direct activation of GIRK channels (Logothetis et al., 1987). (C) Scatter plot shows Ba²⁺-sensitive activated currents. GTPγS-induced current is significantly reduced in SNX27_DA KO neurons (**p<0.01, one-way ANOVA with Bonferroni post hoc test).

Figure 5. GABA_BR-GIRK currents and inhibition are restored with expression of SNX27b in VTA DA neurons of SNX27_DA KO mice

(A) Schematic shows stereotaxic injection of AAV DIO-Snx27b-ires-GFP into VTA of SNX27_DA KO mice. Fluorescence and DIC images of eYFP/GFP+ cell selected for recording from ex vivo midbrain section of SNX27_DA KO with either AAV DIO-eYFP or AAV DIO-Snx27b-ires-GFP (scale bar: 20 µm). (B) Top, whole-cell recordings from VTA DA neurons show baclofen-activated GIRK currents (I_Baclofen; 300 µM) from WT and SNX27_DA KO mice injected with AAV DIO-eYFP or AAV DIO-Snx27b-ires-GFP, and inhibition with Ba²⁺ (1 mM). Scale bar; 100 s and 50 pA. Bottom, baclofen-induced currents of Snx27b-ires-GFP positive cells are significantly larger compared to +eYFP (**p<0.01, Student’s t test). Scatter plot of I_Baclofen for indicated DA neurons with average current indicated by solid black bar. (C,D) Input-output plots show firing frequency increase with larger current injections in the absence (solid circles, ACSF) and presence (filled circles) of baclofen for DA neurons infected with eYFP (C) (n=6) or SNX27-ires-GFP (D) (n=6) (p<0.01 absence versus presence baclofen, 160–300pA, 2-way ANOVA with Bonferroni post hoc test).

Figure 6. Expression of SNX27-insensitive GIRK2a in VTA DA neurons of SNX27_DA KO mice restores GABA_BR-activated GIRK currents and inhibition

(A) Schematic shows stereotaxic injection of AAV DIO-GIRK2a-eYFP into VTA of SNX27_DA KO mice. Fluorescence and DIC images of eYFP positive cell selected for recording from ex vivo midbrain section of SNX27_DA KO mice injected with AAV DIO-Girk2a-eYFP. (scale bar: 20 µm) (B) Top, whole-cell recordings from VTA DA neurons show baclofen-activated GIRK currents (I_Baclofen; 300 µM) from WT and SNX27_DA KO mice injected with AAV DIO-eYFP or AAV DIO-Girk2a-eYFP, and response to Ba²⁺ (1 mM). Scale bar is 100 s and 100 pA. Bottom, baclofen-induced currents in GIRK2a-eYFP positive cells are significantly increased from eYFP+ cells (**p<0.05, Student’s t test). Scatter plot of I_Baclofen for indicated DA neurons with average current indicated by solid black bar. (C,D) Input-output plots show firing frequency increases as a function of larger current injections in the absence (solid circles, ACSF) and presence of baclofen for DA neurons cells infected with eYFP (n=6) (C) or GIRK2a-eYFP (n=8) (D). Baclofen inhibits firing activity in GIRK2a-eYFP expressing DA neurons (D) but not in eYFP expressing neurons (C) of SNX27_DA KO mice (p<0.01 absence versus presence baclofen at 140–300pA, 2-way ANOVA with Bonferroni post hoc test).

Figure 7. Enhanced cocaine locomotor response in SNX27_DA KO mice

(A) Total distance traveled in novel open field environment by WT (n=18), DAT-Cre^+/− (n=17), and SNX27_DA KO (n=17) mice over 1 hour (**p<0.01 vs WT, p<0.05 vs DAT-Cre^+/−, one-way ANOVA with Bonferroni post hoc test). (B) No significant change in thigmotaxis for SNX27_DA KO mice (n=12) vs. WT (n=11) or vs. DAT-Cre^+/− (n=11). Time spent in center of chamber (central 10” squared of 16” squared chamber, n.s. one-way ANOVA). (C) Heightened sensitivity to cocaine in SNX27_DA KO mice. Locomotor activity measured in WT (n=6), DAT-Cre^+/− (n=16) or SNX27_DA KO (n=14) mice following a single injection (i.p.) of saline (0.9%) or cocaine (20mg/kg). (**p<0.05 cocaine vs. saline, two-way ANOVA with Bonferroni post hoc test). (D) Fold increase in locomotor activity with cocaine measured over 30 minute period. (**p<0.05, One-way ANOVA with Bonferroni post hoc test).

Figure 8. SNX27-insensitive GIRK2a expression in DA neurons of SNX27_DA KO mice restores normal response to cocaine

(A) Schematic shows stereotaxic injection of AAV DIO GIRK2a-eYFP or AAV DIO eYFP into VTA of SNX27_DA KO mice. (B) Whole-cell recordings from VTA DA neurons show rescue of baclofen-activated GIRK currents (I_Baclofen; 300 µM) in SNX27_DA KO mice injected with AAV DIO-GIRK2a-eYFP. Scale bar is 100 s and 100 pA. (C) Locomotor activity plots show change in activity following a single cocaine injection (20 mg/kg) for a SNX27_DA KO mouse expressing eYFP (left) or GIRK2a-eYFP (right). Starting and ending points indicated by blue and red circles, respectively. Note reduced activity in mouse expressing GIRK2a. (D) Bar graph shows fold-increase in locomotor activity with cocaine (20 mg/kg) for uninfected WT mice (n=5) and SNX27_DA KO mice injected with either AAV DIO eYFP (n=3) or AAV DIO GIRK2a-eYFP (n=3) (**p<0.05 eYFP vs GIRK2a-eYFP, unpaired t-test).