Mutant Huntingtin Impairs BDNF Release from Astrocytes by Disrupting Conversion of Rab3a-GTP into Rab3a-GDP

Abstract

Brain-derived neurotrophic factor (BDNF) is essential for neuronal differentiation and survival. We know that BDNF levels decline in the brains of patients with Huntington's disease (HD), a neurodegenerative disease caused by the expression of mutant huntingtin protein (mHtt), and furthermore that administration of BDNF in HD mice is protective against HD neuropathology. BDNF is produced in neurons, but astrocytes are also an important source of BDNF in the brain. Nonetheless, whether mHtt affects astrocytic BDNF in the HD brain remains unknown. Here we investigated astrocytes from HD140Q knock-in mice and uncovered evidence that mHtt decreases BDNF secretion from astrocytes, which is mediated by exocytosis in astrocytes. Our results demonstrate that mHtt associates with Rab3a, a small GTPase localized on membranes of dense-core vesicles, and prevents GTP-Rab3a from binding to Rab3-GAP1, disrupting the conversion of GTP-Rab3a into GDP-Rab3a and thus impairing the docking of BDNF vesicles on plasma membranes of astrocytes. Importantly, overexpression of Rab3a rescues impaired BDNF vesicle docking and secretion from HD astrocytes. Moreover, ATP release and the number of ATP-containing dense-core vesicles docking are decreased in HD astrocytes, suggesting that the exocytosis of dense-core vesicles is impaired by mHtt in HD astrocytes. Further, Rab3a overexpression reduces reactive astrocytes in the striatum of HD140Q knock-in mice. Our results indicate that compromised exocytosis of BDNF in HD astrocytes contributes to the decreased BDNF levels in HD brains and underscores the importance of improving glial function in the treatment of HD.

Significance statement: Huntington's disease (HD) is an inherited neurodegenerative disorder that affects one in every 10,000 Americans. To date, there is no effective treatment for HD, in part because the pathogenic mechanism driving the disease is not fully understood. The dysfunction of astrocytes is known to contribute to the pathogenesis of HD. One important role of astrocytes is to synthesize and release brain-derived neurotrophic factor (BDNF), which is vital for neuronal survival, development, and function. We found that mutant huntingtin protein (mHtt) at the endogenous level decreases BDNF secretion from astrocytes by disrupting the conversion of GTP-Rab3a into GDP-Rab3a and that overexpressing Rab3a can rescue this deficient BDNF release and early neuropathology in HD knock-in mouse brain. Our study suggests that astrocytic Rab3a is a potential therapeutic target for HD treatment.

Keywords: BDNF; Huntington's disease; Rab3; exocytosis; glial cell.

Figure 1.

BDNF secretion from HD astrocytes is reduced. A, ELISA assay showed that cultured primary astrocytes from TG (Student's t test, n = 6 independent experiments, p = 1.6381E-05) and KI (Student's t test, n = 4 independent experiments, p = 0.0208) mice release less BDNF than astrocytes from WT mice. ELISA results also indicated that BDNF secretion is decreased from brain slices of TG mice compared with WT mice (Student's t test, n = 3 independent experiments, p = 0.0449). However, BDNF secretion is unchanged from brain slices of N171-82Q mice compared with WT mice (Student's t test, n = 3 independent experiments, p = 0.1489). B, WT neurons were treated with ACM containing released BDNF from WT or KI astrocytes, and then analyzed via Western blotting to examine p-TrkB levels. p-TrkB is decreased in KI ACM-treated neurons, confirming decreased BDNF in KI ACM. C, Glutamate measurement assay showing no significant reduction in glutamate release in cultured astrocytes from TG mice (Student's t test, n = 5 independent experiments, p = 0.1027) or KI mice (Student's t test, n = 6 independent experiments, p = 0.3647) compared with WT astrocytes. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant.

Figure 2.

Transcription or translation of BDNF is not changed in the HD astrocytes. A, B, qRT-PCR results revealing no significant reduction in BDNF mRNA levels in cultured astrocytes from TG mice (A, Student's t test, n = 5 independent experiments; BDNF transcripts generated by different promoters are indicated: BDNF promoter I, p = 0.5844; BDNF promoter II, p = 0.3669; BDNF promoter III, p = 0.5765; BDNF promoter IV, p = 0.67; BDNF promoter pan, p = 0.2406) or KI mice (B, Student's t test, n ≥ 5 independent experiments; BDNF promoter I, p = 0.0688; BDNF promoter II, p = 0.2283; BDNF promoter III, p = 0.6892; BDNF promoter IV, p = 0.0809; BDNF promoter pan, p = 0.7733) compared with WT astrocytes. C, D, Western blotting revealing similar mature mBDNF and precursor of BDNF (pro-BDNF) levels in cultured astrocytes from TG (C), KI (D), and WT mice. Quantifying ratios of pro-BDNF or mBDNF to actin in TG astrocytes (C, Student's t test; pro-BDNF, n = 8 independent experiments, p = 0.6983; mBDNF, n = 8 independent experiments, p = 0.4087) and KI astrocytes (D, Student's t test; pro-BDNF, n = 8 independent experiments, p = 0.8801; mBDNF, n = 5 independent experiments, p = 0.6349). ns, Not significant.

Figure 3.

mHtt associates with Rab3a in astrocytes. A, Knocking down Rab3a via siRNA in WT astrocytes. B, ELISA results showing that downregulation of Rab3a inhibits the release of BDNF and ATP from WT astrocytes compared with scrambled siRNA transfected astrocytes (Student's t test: BDNF release, n = 3 independent experiments, p = 0.0353; ATP release, n = 5 independent experiments, p = 0.0142). C, D, Association of mHtt with Rab3a is detected in cultured KI astrocytes (C) but not in cultured KI neurons (D). Endogenous mHtt in KI astrocytes or KI neurons was immunoprecipitated by 1C2 antibody, and the immunoprecipitates were probed with antibody to Rab3a. Immunoprecipitation with IgG served as a control. E, Association of mHtt with Rab3a-V5 was detected in cultured KI astrocytes infected with Rab3a-V5 adenovirus. Rab3a was immunoprecipitated by anti-V5 antibody, and the immunoprecipitates were probed with antibodies to 1C2 to detect mHtt. Immunoprecipitation with IgG served as a control. F, In vitro binding assay showed that in vitro translated Htt (1–212 aa) with 150Q binds to purified GST-Rab3a. G, Binding of mHtt to GTP-Rab3a was found in cultured HD KI astrocytes. To immunoprecipitate endogenous mHtt in HD KI astrocytes, 1C2 antibody was used, and the immunoprecipitates were probed with the antibody specific to GTP-Rab3a. *p < 0.05.

Figure 4.

Reduced association between Rab3-GAP1 and Rab3a by mHtt results in increased GTP-Rab3a in HD astrocytes. A, GTP-Rab3a protein levels were increased in cultured KI astrocytes (Student's t test, n = 3 independent experiments, p = 0.0402). B, C, Association between GTP-Rab3a with Rab3-GAP1 is decreased in both cultured KI astrocytes (C, Student's t test, n = 4 independent experiments, p = 0.0041) and in the corpus callosum of KI mice (C, Student's t test, n = 4 independent experiments, p = 0.0106). D, Rab3-GAP1 levels did not change in KI astrocytes. E, GTPase activity of purified GST-Rab3a was not affected by binding to mHtt (Student's t test, n = 4 independent experiments; 30 min, p = 0.4807; 60 min, p = 0.9353; 90 min, p = 0.4071). *p < 0.05, **p < 0.01; ns, not significant.

Figure 5.

Overexpression of Rab3a rescues defective release of BDNF from HD astrocytes. A, B, Overexpression of Rab3a in KI astrocytes by adenovirus infection was confirmed by immunostaining (A) and Western blotting (B). C, Rab3a was overexpressed in the striatum of TG mice by injecting Rab3a-V5 adenovirus into the striatum. D, E, ELISA results indicated that BDNF secretion was increased from cultured KI astrocytes infected with Rab3a-V5 adenovirus (Student's t test, n = 3 independent experiments, p = 0.0056) or from Rab3a-V5 adenovirus-injected brain slices containing striatum of TG mice (Student's t test, n = 3 independent experiments, p = 0.0464). F, WT neurons were treated with ACM containing released BDNF from KI or Rab3a-overexpressed KI astrocytes. p-TrkB levels are increased in neurons treated with Rab3a-overexpressed KI ACM relative to those treated with KI ACM (Student's t test, n = 5 independent experiments, p = 0.0292). *p < 0.05, **p < 0.01; ns, not significant.

Figure 6.

Overexpression of Rab3a rescues the deficient release of ATP from HD astrocytes. A, Quinacrine staining was performed by incubating living cultured WT, KI, and Rab3a-overexpressed KI astrocytes for 15 min after treatment of PMA. B, The quantitative results showed that ATP levels are higher in KI astrocytes than in WT astrocytes after stimulation, and overexpression of Rab3a reduces ATP levels in astrocytes compared with KI astrocytes (Student's t test; WT vs KI: WT, n = 12 cells; KI, n = 14 cells, p = 0.0204; KI vs KI+Rab3a-V5: KI, n = 14 cells; KI+Rab3a-V5, n = 15 cells, p = 0.0068). C, Bioluminescence assay showed that ATP secretion was rescued from cultured KI astrocytes infected with Rab3a-V5 adenovirus (Student's t test, n = 5 independent experiments, WT vs KI, p = 0.00006; KI vs KI+Rab3a-V5, p = 0.0125). *p < 0.05, **p < 0.01, ***p < 0.001. Scale bar, 10 μm.

Figure 7.

Defective docking of BDNF-containing and ATP-containing vesicles are rescued by Rab3a overexpression in HD astrocytes. A–D, TIRFM revealing that the docking of BDNF-containing vesicles (A, C; Student's t test; WT vs KI: WT, n = 11 cells, KI, n = 16 cells, p = 0.0043; KI vs KI+Rab3a-V5: KI, n = 16 cells; KI+Rab3a-V5, n = 8 cells, p = 0.0275) and ATP-containing vesicles (B, D; Student's t test; WT vs KI: WT, n = 16 cells; KI, n = 13 cells, p = 0.0435; KI vs KI+Rab3a-V5: KI, n = 13 cells; KI+Rab3a-V5, n = 15 cells, p = 0.0226) are compromised in KI astrocytes, which can be improved by overexpressing Rab3a. *p < 0.05, **p < 0.01. Scale bar, 10 μm.

Figure 8.

Overexpression of Rab3a reduces reactive astrocytes in the striatum of HD140Q KI mice A, Five-month-old WT mice were injected with Rab3a-V5 adenovirus into the left side of the striatum. After 21 d, immunofluorescent staining showed that Rab3a-V5 preferentially infected astrocytes in the striatum. B, Stereotaxic injection of 9-month-old HD140Q KI mice. C, Low-magnification (10×) micrographs showing NeuN and GFAP staining in adenoviral GFP and Rab3a-V5-injected striatum. High-magnification (63×) micrographs showing GFAP (red) or mutant Htt aggregates (arrows, red) staining in the merged images in which the nuclei are stained by Hoechst (blue). Scale bars, 10 μm. D, Quantitative analysis of the GFAP immunofluorescent density showing GFAP staining is significantly decreased in the Rab3a-V5-injection side compared with GFP-injection side (Student's t test, randomly selected 7–10 images per section, n = 8 sections per group, p = 7.51767E-13). The percentage of NeuN-positive cells and the number of mutant Htt aggregates per image are unchanged (Student's t test, n = 600 cells per group, NeuN-positive cells, p = 0.1016; aggregates number, p = 0.1511). ***p < 0.001; ns, not significant.

Figure 9.

Proposed model for the decreased dense-core vesicles released from HD astrocytes. GTP/GDP-Rab3 exchange is essential for docking of dense-core vesicles in astrocytes. In HD, the binding of mHtt to GTP-Rab3a keeps GTP-Rab3a from associating with Rab3-GAP1, disrupts GTP/GDP-Rab3 exchange, inhibits docking of dense-core vesicles, and results in the decreased release of secreted molecules, such as BDNF.