BDNF-TrkB signaling in striatopallidal neurons controls inhibition of locomotor behavior

Abstract

The physiology of brain-derived neurotrophic factor signaling in enkephalinergic striatopallidal neurons is poorly understood. Changes in cortical Bdnf expression levels, and/or impairment in brain-derived neurotrophic factor anterograde transport induced by mutant huntingtin (mHdh) are believed to cause striatopallidal neuron vulnerability in early-stage Huntington's disease. Although several studies have confirmed a link between altered cortical brain-derived neurotrophic factor signaling and striatal vulnerability, it is not known whether the effects are mediated via the brain-derived neurotrophic factor receptor TrkB, and whether they are direct or indirect. Using a novel genetic mouse model, here, we show that selective removal of brain-derived neurotrophic factor-TrkB signaling from enkephalinergic striatal targets unexpectedly leads to spontaneous and drug-induced hyperlocomotion. This is associated with dopamine D2 receptor-dependent increased striatal protein kinase C and MAP kinase activation, resulting in altered intrinsic activation of striatal enkephalinergic neurons. Therefore, brain-derived neurotrophic factor/TrkB signaling in striatopallidal neurons controls inhibition of locomotor behavior by modulating neuronal activity in response to excitatory input through the protein kinase C/MAP kinase pathway.

Figure 1. Generation of Trkb-enkephalin knockout mice.

(a) Adult brain image of BAC–Penk–Cre^tg/+::Z/EG–egfp^tg/+ mouse stained with G(Y)FP Abs revealing Cre-mediated GFP expression in specific brain areas, the caudate-putamen (CPu), the nucleus accumbens (NAc), scattered cells in cortical layer II and V–VI (CX). (b) Confocal images showing Cre-mediated EYFP expression in ENK striatal cells, (c) and absence of EYFP expression in striatal interneurons (arrows) identified by specific markers. (d,e) Dlx5 in situ expression at E11.5 in the striato-pallidal neural ephitelium (stiatal-ne, pallidal-ne), and absence of Penk1 expression in the same regions at this stage (images are from the Allen Mouse Brain Atlas20). (f) E11 BAC–Penk–Cre^tg/+::Z/EG–egfp^tg/+ transversal brain section with no GFP staining in the striato-pallidal neural ephitelium. (g) PCR analysis of Trkb recombination in brains carrying the BAC–Penk–Cre transgene at E12.5. Primers were specific for un-recombined (Nr) and recombined (r) Trkb^lx allele; each embryo carried a Trkb^lx allele and Trkb^WT allele. Recombination occurred only in presence of the Cre transgene (embryos 1–3). (h) Protein lysates from different brain regions of adult Trkb^lx/lx mice either in presence (+) or in absence (−) of the Cre transgene were analyzed for TrkB protein levels. Blots were probed with Abs to p145-TrkB. Striatal TrkB levels were 50% reduced, but unchanged in the cortex. Blots were reprobed with Abs to TUJ.1 to control for loading and quantification of fold changes (values are mean±s.d.). ChAT, choline acetyltransferase. Scale bars: a, 500 μm; b,c, 50 μm; f, 100 μm. See also Supplementary Fig. S1.

Figure 2. Enhanced spontaneous and cocaine-induced locomotion in Trkb^PENK-KO mice.

(a) Spontaneous locomotor measures obtained for control and mutant mice tested as independent groups at different stages (Trkb^PENK-WT, n=9, 9, 7; Trkb^PENK-KO, n=7, 9, 7, for the three groups, respectively). Mice were individually placed into an OF arena for 30 min and the distance traveled was recorded. While no difference was observed up to 8 M of age, a sharp significant increase in total distance traveled was found in Trkb^PENK-KO mice compared with that in Trkb^PENK-WT littermates by 9 M of age until later in life (two-way ANOVA, F_(1,42)=8.964; P=0.0046, main effect of genotype; F_(2,42)=6.406; P=0.0037, interaction between genotype and age, Trkb^PENK-KO versus Trkb^PENK-WT, *P=0.0001); values are means±s.e.m. (b) Cocaine-induced hyperlocomotion. After 2 days of habituation to the apparatus, mice (Trkb^PENK-WT, n=11; Trkb^PENK-KO, n=10) were i.p.-injected with either vehicle on day 0, showing no difference between mutants and controls (one-way ANOVA, F_(1,19)=0.386; P=0.5417), or with 10 mg kg⁻¹ cocaine for 3 consecutive days, immediately before being placed into the arena where locomotor activity was recorded for 30 min (one-way ANOVA, F_(1,19)=4.797; *P=0.0412). Cocaine values are mean across the 3 days±s.e.m. (c) Trkb^PENK-KO and Trkb^PENK-WT mice were tested for motor coordination as independent groups at different stages (Trkb^PENK-WT, n=9, 8, 8, and Trkb^PENK-KO mice, n=7, 11, 7, for the three groups, respectively). On day 1, they were habituated to the rotarod apparatus and subjected to three sessions of 3 min at 4 r.p.m. constant speed, separated by 15-min rest intervals. The next day they were placed on the rod accelerating from 4 to 40 rpm for three consecutive sessions (separated by 5-min intervals). Latency of each animal to fall from the rod was scored. The performance was not significantly different between mutants and controls (two-way ANOVA, F_(1,44)=0.298; P=0.5878 main effect of genotype; F_(2,44)=0.874; P=0.4244, interaction between genotype and age; F_(2,44)=9.702; P=0.0003 effect of age). Values are means±s.e.m.

Figure 3. Reduced striatal enkephalin expression in Trkb^PENK-KO mice in the absence of striatal loss.

(a) Striatal images from 8-M-old mice immunostained for DARPP-32. DARPP-32 cell counts at 8 M (Trkb^PENK-KO n=3, versus Trkb^PENK-WT n=3, P=0.741, unpaired Student’s t-test). (b) Striatal images of 8-M-old Trkb^PENK-WT and Trkb^PENK-KO mice showing enkephalin in situ hybridization. Counts of ENK⁺MSNs at 8 M (Trkb^PENK-KO n=3, versus Trkb^PENK-WT n=3, *P=0.036, unpaired Student’s t-test). (c) Confocal stacks of Golgi-impregnated MSNs from 12-M-old mice analyzed for spine density. (d) Sholl analysis of striatal MSNs dendritic complexity. For both panels c and d (n=30, 10 randomly selected neurons per mouse/over three mice). One-way ANOVA, F_(1,58)=0.341; P=0.5615. See also Supplementary Figs S6 and S7. (e) Representative western blot of 12-M-old striatal tissue lysates of Trkb^PENK-WT (n=3) and Trkb^PENK-KO (n=3) mice. Membranes were blotted with the indicated Abs and densitometry analysis was used for quantification of the proteins relative abundance (Trkb^PENK-KO versus Trkb^PENK-WT for Syn, P=0.46; PSD-95, P=0.15; Neuroligin-2, P=0.09; GAD67, P=0.40; DARPP-32, P=0.21; D2R, P=0.52; D1R, P=0.33. P-values were generated by unpaired Student’s t-test). GAPDH was used to control for protein loading. Values are means±s.e.m. See also Supplementary Fig. S8a. CPu, caudate-putamen. Scale bars: a–b, 50 μm. Syn, synaptophysin; D1R and D2R, dopamine receptor type 1 and 2.

Figure 4. Functional integrity of the corticostriatal pathway in Trkb^PENK-KO mice.

(a) Averaged fEPSP traces (upper panel) evoked by a stimulation of 12, 16, 22, 28, 32 and 36 V. Input–output best-fit curves (lower panel) revealed no significant difference between Trkb^PENK-WT (n=7 slices/4 mice) and mutant mice (n=10 slices/4 mice) (P>0.05, unpaired Student’s t-test). Values are means±s.e.m. of fEPSP amplitude evoked by a given stimulation. (b) Histograms of group data showing similar paired-pulse ratio (PPR) of responses at 40 and 60 ms of interpulse intervals stimulation between Trkb^PENK-WT (n=8 slices/4 mice) and Trkb^PENK-KO mice (n=9 slices/4 mice), and similar ratio values (P>0.05, unpaired Student’s t-test). Values are means±s.e.m. of changes in the respective cell populations. (c) Graphs showing mean frequency and amplitude of the spontaneous postsynaptic currents (sEPSC) recorded from A2A-positive MSNs (Trkb^PENK-KO, frequency: 2.67±0.56 (n=9/5 mice) versus Trkb^PENK-WT: 2.94±0.32 (n=14/9 mice), P=0.64, unpaired Student’s t-test; Trkb^PENK-KO, amplitude: 13.3±0.87 (n=9/5 mice) versus Trkb^PENK-WT: 13.8±1.2 (n=14/9 mice), P=0.76, unpaired Student’s t-test. Values are means±s.e.m.

Figure 5. Reduced neuronal activation in Trkb-depleted striatopallidal cells in response to D2R-blockade.

(a–b) Representative images of CPu (a) and LGP (b) showing c-Fos expression 2 h after systemic i.p. injection of 1 mg kg⁻¹ haloperidol in 3-month-old Trkb^PENK-WT and Trkb^PENK-KO mice (n=5, each genotype). In (b) tissues are co-labeled with an Ab against parvalbumin. These cells are the target of the striatopallidal MSNs and do not express D2R/ENK. Therefore, these cells are insensitive to haloperidol in contrast to the other cell type within the LGP that is D2R/ENK+ (PV−). (c,d) Quantification of density of haloperidol-induced c-Fos-ir cells in the CPu (c) (Trkb^PENK-WT, 100±8.81 n=5, Trkb^PENK-KO, 53.03±9.40 n=5, P=0.006, unpaired Student’s t-test), and LGP (d) (Trkb^PENK-WT, 100±15.86 n=5, Trkb^PENK-KO, 46.52±13.68 n=5, P=0.034, unpaired Student’s t-test); values are means±s.d. (e,f) Schematic representation of TrkB signaling requirement in striatopallidal MSNs, and relative neuronal populations in the CPu and the LGP. Red and green cells in the dorsal striatum represent respectively striatopallidal (indirect pathway) and striatonigral (direct pathway) MSNs. Brown and orange cells in the LGP represent respectively PV+ neurons (indirect pathway) that projects to the subthalamic nucleus (STN), and PV−, D2R/ENK+ pallidostriatal neurons that project to the striatum. (e) Glutamatergic input from the cortex to the striatopallidal MSNs (D2R/ENK+) is integrated in these neurons by functional dopamine/D2R and BDNF–TrkB activation. This leads to normal transmission of inhibitory signal (GABA) to the target in the LGP, PV+ neurons. (f) Depletion of TrkB signaling alters indirect pathway MSN activation causing diminished inhibitory signaling to the target cells in the LGP (PV+), and PV−, D2R/ENK+ pallidostriatal neurons causing impaired feedback to the striatum. CC, corpus callosum; CPu, caudate-putamen; LGP, lateral globus pallidum; PV, parvalbumin. SNT, subthalamic nucleus. Arrowheads and arrows indicate c-Fos positive and PV-positive cells, respectively. Scale bars: a, 50 μm; b, 20 μm. See also Supplementary Methods.

Figure 6. Unchanged synaptic density in striatopallidal targets in the absence of presynaptic BDNF–TrkB signaling.

(a,b) Single-cell level confocal images of LGP neurons co-labeled with parvalbumin (red) and gephyrin (green) (a), or parvalbumin (red) and GAD67 (green) (b), in 3-M-old haloperidol-treated Trkb^PENK-WT and Trkb^PENK-KO mice highlighting somatodendritic synapses. (c,d) Density of GABAergic synapses in pallidostriatal neurons quantified by gephyrin (Trkb^PENK-WT, 100±14.59 n=20 neurons from 3 mice, Trkb^PENK-KO, 107.66±7.04 n=19 neurons from three mice, P=0.46, unpaired Student’s t-test) (c); and GAD67 (Trkb^PENK-WT, 100±7.78, n=27 neurons from three mice, Trkb^PENK-KO 86.38±35.79, n=24 neurons from three mice, P=0.55, unpaired Student’s t-test) (d). Values are means±s.d. Scale bars: a–b, 5 μm.

Figure 7. ENK-specific Trkb deletion affects striatal D2R-dependent PKC and ERK1/2 phosphorylation.

(a,b) Representative immunoblots of proteins extracted from dorsal striatum of 13-M- and 5-M-old Trkb^PENK-WT and Trkb^PENK-KO mice. Animals were either treated with vehicle or with a D2R-specific antagonist L-741,626 (20 mg/kg), and killed 30 min later. Phosphorylation states of ERK1/2, PKC, AKT and DARPP-32 were determined using phospho-specific Abs. See also Supplementary Fig. S8b–c. (c) Graphs represent the phosphorylation state of each protein normalized to its total amount. Fold changes were determined against vehicle-treated Trkb^PENK-WT (set as 1), (vehicle: Trkb^PENK-KO versus Trkb^PENK-WT, pERK1/2, P=0.020; p-PKC, P=0.028; p-DARPP32, P=0.065; pAKT, P=0.17), (L-741,626: Trkb^PENK-KO versus Trkb^PENK-WT, pERK1/2, P=0.07; p-PKC, P=0.19), (L-741,626 treated Trkb^PENK-KO versus vehicle treated Trkb^PENK-KO, pERK1/2, P=0.016; p-PKC, P=0.026). P-values were generated by unpaired Student’s t-test. Values are means±s.e.m., n=6 per genotype (three per condition) analyzed. (d) Graphs represent the phosphorylation state of each protein normalized to its total amount at 5 M of age. Total DARPP-32 was normalized to ERK1. Fold changes were determined against vehicle-treated Trkb^PENK-WT (set as 1), (vehicle: Trkb^PENK-KO versus Trkb^PENK-WT, pERK1/2, P=0.06; p-PKC, P=0.90; DARPP-32, P=0.33). P-values were generated by unpaired Student’s t-test. Values are means±s.e.m., n=5 per group analyzed.