PGC-1{alpha} and PGC-1{beta} regulate mitochondrial density in neurons

Abstract

Recent studies indicate that regulation of cellular oxidative capacity through enhancing mitochondrial biogenesis may be beneficial for neuronal recovery and survival in human neurodegenerative disorders. The peroxisome proliferator-activated receptor gamma coactivator-1alpha (PGC-1alpha) has been shown to be a master regulator of mitochondrial biogenesis and cellular energy metabolism in muscle and liver. The aim of our study was to establish whether PGC-1alpha and PGC-1beta control mitochondrial density also in neurons and if these coactivators could be up-regulated by deacetylation. The results demonstrate that PGC-1alpha and PGC-1beta control mitochondrial capacity in an additive and independent manner. This effect was observed in all studied subtypes of neurons, in cortical, midbrain, and cerebellar granule neurons. We also observed that endogenous neuronal PGC-1alpha but not PGC-1beta could be activated through its repressor domain by suppressing it. Results demonstrate also that overexpression of SIRT1 deacetylase or suppression of GCN5 acetyltransferase activates transcriptional activity of PGC-1alpha in neurons and increases mitochondrial density. These effects were mediated exclusively via PGC-1alpha, since overexpression of SIRT1 or suppression of GCN5 was ineffective where PGC-1alpha was suppressed by short hairpin RNA. Moreover, the results demonstrate that overexpression of PGC-1beta or PGC-1alpha or activation of the latter by SIRT1 protected neurons from mutant alpha-synuclein- or mutant huntingtin-induced mitochondrial loss. These evidences demonstrate that activation or overexpression of the PGC-1 family of coactivators could be used to compensate for neuronal mitochondrial loss and suggest that therapeutic agents activating PGC-1 would be valuable for treating neurodegenerative diseases in which mitochondrial dysfunction and oxidative damage play an important pathogenic role.

FIGURE 1.

Effects of PGC-1 α and PGC-1 β on mitochondrial density in primary cortical neurons. Cortical neurons were transfected with plasmids expressing wild type PGC-1 coactivators or their shRNAs, GFP, and mitochondria-targeted pDsRED2 at day 2 after plating. A (expressing wild type PGC-1α) and B (expressing shRNA suppressing PGC-1α) demonstrate substantial differences in axonal mitochondrial density 3 days later. Further morphometric analysis demonstrates that axonal mitochondrial density is controlled both by PGC-1α (C) and PGC-1β (D). To estimate mitochondrial density in the neuron's body, we performed three-dimensional scan with a confocal laser microscope, followed by three-dimensional reconstruction of mitochondrial network in neuronal body. E (expressing wild type PGC-1α) and F (expressing shRNA suppressing PGC-1α) demonstrate also a clear difference in mitochondrial density in the neuronal body. Stereological analysis of three-dimensional reconstructed mitochondrial networks demonstrates a decrease of density in response of PGC-1α (G) and PGC-1β (H) suppression. *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus control (one-way ANOVA followed by Newman-Keuls post hoc test).

FIGURE 2.

Effects of PGC-1 α and PGC-1 β on the COX8a and COX6c promoter activity. Cortical neurons were cotransfected with plasmids expressing COX8a or COX6c promoter-luciferase reporter, Renilla luciferase, and PGC-1α or PGC-1β constructs at day 1–2 after plating. Overexpression of PGC-1α (A) or PGC-1β (B) led to increased luciferase activity performed 24 h later, suggesting increased expression of these reporter genes. Results are normalized to Renilla luciferase signal (**, p < 0.01; ***, p < 0.001 versus control, t test).

FIGURE 3.

Effects of PGC-1 α and PGC-1 β on the ATP level. Cortical neurons were cotransfected with plasmids expressing firefly luciferase, Renilla luciferase and PGC-1α or PGC-1β at day 2 after plating. Firefly luciferase activity was measured 24 h later, using 1-(4,5-dimethoxy-2-nitrophenyl)diazoethane-caged luciferin as a substrate, after which the neurons were lysed and Renilla luciferase was measured. Overexpression of PGC-1α and PGC-1β led to an increased firefly to Renilla luciferase activity ratio. **, p < 0.01; ***, p < 0.001 versus control (one-way ANOVA followed by Newman-Keuls post hoc test).

FIGURE 4.

Effects of PGC-1 α and PGC-1 β on the autophagy. Cortical neurons were transfected with plasmids expressing EGFP-LC3, mitochondria-targeted pDsRed2, and PGC-1α or PGC-1β at day 2 after plating. 3 days later, the LC3-positive dots were visualized. A, EGFP-LC3-expressing autophagosome (left), mitochondria-targeted pDsRed2 expressing mitochondria (middle), and superimposed channels to demonstrate colocalization of signals (right and below). The number of autophagosomes (B and D) as well as the number of mitochondria engulfed by autophagosomes per mm of axonal length (C and E) was unchanged in PGC-1α- or PGC-1β-overexpressing neurons. Rapamycin used as a positive control increased the number of LC3-positive mitochondria.

FIGURE 5.

Effects of SIRT1 and GCN5 shRNA. A–G, cortical neurons were transfected with plasmids expressing GFP and mitochondrial pDsRED2 and with different plasmids interfering with PGC-1 at day 2 after plating. A demonstrates that overexpression of SIRT1 but not its inactive mutant, SIRT G261A, increases mitochondrial density. Effect of SIRT1 disappeared in the presence of shRNA suppressing PGC-1α (B) (p < 0.001 for interaction, two-way ANOVA) but not PGC-1β (C) (p = 0.89). D demonstrates the positive effects of GCN5 suppressing shRNA that disappeared in the presence of PGC-1α shRNA (E) (p = 0.003) but not in the presence of PGC-1β shRNAs (F) (p = 0.009). G demonstrates that although combination of SIRT1 and GCN5 shRNA increased mitochondrial density more that each treatment separately; both treatments were still dependent (p = 0.008 for interaction). H and I, cortical neurons were cotransfected with Gal4-PGC-1α fusion protein, GAL4-UAS-luciferase reporter, and SIRT1 or with PGC-1α promoter reporter and SIRT1 plasmids. H demonstrates that overexpression of SIRT1 increases PGC-1α transcriptional activity whereas the activity of PGC-1α promoter reporter (I) remains unchanged (*, p < 0.05; **, p < 0.01; ***, p < 0.001 versus control; one-way ANOVA followed by Newman-Keuls post hoc test was used to compare the difference from control group, and two-way ANOVA was used to analyze the interaction between treatments).

FIGURE 6.

Effect of p160^MBP. A demonstrates that overexpression of p160^MBP and p67^MBP but not the deletion mutant of p160^MBP decreases mitochondrial density, whereas p160^MBP shRNA increases it. Experiments with Gal4-PGC-1α (B) and PGC-1 promoter reporter (C) confirm that p160^MBP and p67^MBP regulate only PGC-1α transcriptional activity and not the expression. Effects of p160^MBP overexpression/suppression disappeared in the presence of shRNA suppressing PGC-1α (D) (p < 0.011 for interaction, two-way ANOVA). Effects of p160^MBP overexpression/suppression and SIRT1 were additive (E) (p = 0.45 for interaction). (*, p < 0.05; **, p < 0.01; ***, p < 0.001 versus control; one-way ANOVA followed by Newman-Keuls post hoc test).

FIGURE 7.

PGC-1 α, PGC-1 β, and SIRT1 overexpression and GCN5 suppression protect against mitochondrial loss. Cortical neurons were transfected with plasmids expressing A53T-mutated α-synuclein or 120Q huntingtin, GFP, or mitochondrial pDsRED2 and with different plasmids modulating expression or activity of PGC-1 coactivators. A and B demonstrate that PGC-1α and PGC-1β protect against A53T-mutated α-synuclein or 120Q huntingtin-induced mitochondrial loss, respectively. C and D demonstrate that also SIRT1 and GCN5 shRNA protect against A53T-mutated α-synuclein or 120Q huntingtin-induced mitochondrial loss, respectively. E and F demonstrate that overexpression of PGC-1α and SIRT1 protect also against A53T α-synuclein or 120Q huntingtin-induced neuronal death, respectively (*, p < 0.05; **, p < 0.01; ***, p < 0.001 versus control; one-way ANOVA followed by Newman-Keuls post hoc test).