PGC-1α and PGC-1β Regulate Mitochondrial Density in Neurons^*

Primary Neuronal Culture

Primary cultures of rat cortical cells were prepared from neonatal Wistar rats. Briefly, cortices were dissected in ice-cold Krebs-Ringer solution (135 mm NaCl, 5 mm KCl, 1 mm MgSO₄, 0.4 mm K₂HPO₂, 15 mm glucose, 20 mm HEPES, pH 7.4) containing 0.3% bovine serum albumin and trypsinized in 0.8% trypsin for 10 min at 37 °C. This was followed by trituration in a 0.008% DNase solution containing 0.05% soybean trypsin inhibitor. Neurons were resuspended in basal medium Eagle with Earle's salts (BME) containing 10% heat-inactivated fetal bovine serum, 25 mm KCl, 2 mm glutamine, and 100 μg/ml gentamicin and plated onto 35-mm glass bottom dishes (glass surface diameter 14 mm; MatTek, Ashland, MA) precoated with poly-l-lysine at a density of 10⁶ cells/ml (2 ml of cell suspension/dish) or 170 μl/well into white microwell plates (96 F Nuclon Delta, Nunc; Thermo Fisher Scientific). 3 h later, medium was changed to Neurobasal^TM-A medium containing B-27 supplement, 2 mm GlutaMAX^TM-I, and 100 μg/ml gentamicin.

For preparation of primary cultures of cerebellar granule cells, the cerebelli from 8-day-old Wistar rats were dissociated by trypsinization in 0.25% trypsin at 35 °C for 15 min, followed by trituration in a 0.004% DNase solution containing 0.05% soybean trypsin inhibitor. Cells were resuspended in BME containing 10% fetal bovine serum, 25 mm KCl, 2 mm glutamine, and 100 μg/ml gentamicin. Neurons were plated onto 35-mm glass bottom dishes precoated with poly-l-lysine at a density of 1.3 × 10⁶ cells/ml. 10 μm cytosine arabinoside was added 24 h after plating to prevent the proliferation of glial cells.

Primary cultures of mesencephalic neurons were prepared from mesencephali of mouse embryos at embryonic day 15. Midbrain anlages were dissected, cleaned from meninges, triturated mechanically through a 1-ml pipette tip in BME containing 10% fetal bovine serum, and plated as droplets containing 0.2 × 10⁵ cells directly on poly-l-lysine-precoated glass bottom dishes. 1 h after plating, 2 ml of Neurobasal^TM-B medium containing B-27 supplement, 2 mm GlutaMAX^TM-I, and 100 μg/ml gentamicin was added. All of the cultures were grown in a humidified 5% CO₂, 95% air incubator at 37 °C.

Plasmids

Plasmids overexpressing PGC-1α (Addgene plasmid 10974), Gal4-PGC-1α (plasmid 8892), PGC-1β (plasmid 1031), p160^MBP (plasmid 41), p160C (plasmid 46), p67^MBP (plasmid 42), SIRT1 (plasmid 10962), SIRT1 G261A (plasmid 10963), or PGC-1α promoter-luciferase (plasmid 8887) were from ADDGENE (Cambridge, MA). GFP and mitochondrial pDsRED2 were from Clontech. Plasmids expressing short hairpin RNA-s against PGC-1α, PGC-1β, and GCN5 were from SABiosciences (Frederick, MD). Plasmids expressing dominant negative or constitutively active AMPK isoforms were generous gifts from Dr. D. Carling (14). Luciferase reporter constructs of murine COX8a and COX6c promoters (15) were gifts from Dr. M. T. Wong-Riley, and luciferase reporter constructs of mitofusion promoter were from Dr. A. Zorzano (16). 120Q Htt-overexpressing plasmid (17) was a gift from Dr. L. Hasholt, and A53T α-synuclein-expressing plasmid was from Dr. S. Kõks. Expression vectors were confirmed in HEK293 cells using Western blot. shRNA-encoding plasmids were validated using reverse transcription-PCR according to the manufacturer's protocols (supplemental Fig. S1).

Transfection

Cultures were transiently transfected on the second day (on the first day in the case of cerebellar granule cells) in vitro with Lipofectamine^TM 2000. Briefly, conditioned medium was collected, and 100 μl of Opti-MEM® I medium containing 2% Lipofectamine^TM 2000 and 1–2 μg of total DNA was added directly to the cells on the glass bottom surface of dishes and incubated for 3–4 h in a humidified 5% CO₂, 95% air incubator. In multiwell plates, 50 μl of Opti-MEM® I medium with 2% Lipofectamine^TM 2000 and 0.6–0.7 μg of DNA were added to each well. For primary cerebellar granule cells, conditioned medium was added at the completion of incubation, and for cortical and mesencephalic cultures conditioned medium was mixed 1:1 with fresh Neurobasal^TM medium. The transfected DNAs were allowed to express and accumulate respective proteins, and experiments were performed on the fifth day in vitro unless specified otherwise. Controls were transfected as treated groups with an equal amount of non-interfering DNA. All of the culture media, supplements, and transfection reagents were obtained from Invitrogen.

Image Acquisition and Analysis

For mitochondrial density measurements in axons, the neuronal cultures were transfected with GFP, mitochondrial pDsRed2, and plasmids of interest. 3 days later, 10 fluorescence images were randomly captured from each dish using an Olympus IX70 inverted microscope equipped with WLSM PlanApo ×40/0.90 water immersion objective and Olympus DP70 CCD camera. Morphometric analysis was performed using MicroImage software (Media Cybernetics, Bethesda, MD). For mitochondrial density (mitochondrial length/axonal length), at least 40 axons per group were analyzed (1 axon/image).

For mitochondrial density measurements in cell body, the neuronal cultures were transfected with GFP, mitochondrial pDsRed2, and plasmids of interest. 3 days later, 512 × 512-pixel sections were acquired by an LSM 510 META Zeiss confocal microscope equipped with a Plan-Apo ×100/1.4 oil immersion objective. Voxels were collected at 43–74-nm lateral and 100-nm axial intervals. Raw images were three-dimensionally deconvolved and reconstructed by the AutoDeblur and Autovisualize X software package (Media Cybernetics). Reconstructed images were then subjected to stereologic analysis using the Computer-assisted Stereology Toolbox software (Olympus, Ballerup, Denmark). The mitochondrial volume as well as the volume of cell body was estimated using the Cavalieri principle. A grid of points was superimposed on the images, after which the points that overlaid the fluorescent signal were counted. Mitochondrial density in cell body was calculated by dividing the mitochondrial volume with the volume of cell body. Seven or eight cell bodies were analyzed in each group.

Estimation of Autophagy

For autophagy measurements, the cortical cultures were transfected with EGFP-LC3, mitochondrial pDsRed2, and plasmids of interest. 3 days later, fluorescence images were randomly captured using LSM 510 META Zeiss using CI Plan-Neofluar ×63/1.33 water immersion objective. The temperature was maintained at 37 °C using a climate chamber. Distribution and colocalization of GFP-LC3 punctates (for autophagosome visualization) and pDsRed2 (for mitochondrial visualization) were analyzed using MicroImage software. Altogether, 42 axons/group were analyzed (1 axon/image).

Luciferase Reporter Gene Assays

For luciferase assays, cortical neurons were transiently transfected on the second day after plating, as described above, using 0.6–0.7 μg of total DNAs in the desired combination of luciferase reporter plasmids or Gal4 fusions as well as SIRT1 or p160^MBP and p67^MBP plasmids. All of these transfections included 0.1 μg of Renilla luciferase (pRL-CMV) plasmid. Luciferase assays were performed 24 h after transfection using Dual-Glo Luciferase Assay reagent (Promega, Madison, WI) according to the manufacturer's instructions. The promoter activity of COX subunits (COX6c and COX8a) and PGC-1α as well as posttranscriptional activation of PGC-1α was determined as relative firefly luciferase luminescence normalized to Renilla luciferase signal measured by a MicroBeta® TriLux luminescence counter (PerkinElmer Life Sciences). In all luminescence experiments, eight independent samples were analyzed.

ATP Levels

Cortical neurons were transfected with plasmids expressing firefly luciferase, Renilla luciferase, and plasmids of interest. For measurement of firefly luciferase activity, we incubated cells for 10 min with 25 μm 1-(4,5-dimethoxy-2-nitrophenyl)diazoethane-caged luciferin at 37 °C and measured luminescence by MicroBeta® TriLux. For measurement of Renilla luciferase activity, the neurons were lysed, and luminescence was measured using Dual-Glo luciferase assay reagent. Firefly luciferase activity from living cells normalized per Renilla luciferase activity from lysed cells was used as an estimate of ATP level in transfected cells. In all experiments, 16 independent samples were analyzed.

Statistics

One-way ANOVA followed by Newman-Keuls post hoc test or t test were used to compare the difference from the control group, and two-way ANOVA was used to analyze the interaction between treatments.

PGC-1α and PGC-1β Control Mitochondrial Density

Overexpression of PGC-1α in cortical neurons led to a 13% increase in mitochondrial density in axons, and PGC-1α suppression by shRNAs decreased mitochondrial density up to 17% (Fig. 1, A–C). Similarly, overexpression of PGC-1β increased mitochondrial density in axons 19%, and PGC-1β suppression by shRNAs decreased mitochondrial density up to 25% (Fig. 1D). Similar trends were observed also in midbrain and cerebellar granule neurons (supplemental Fig. S2). Fig. 1, E–H, demonstrates that PGC-1α and PGC-1β control mitochondrial density also in the neuronal body; suppression of either PGC-1α or PGC-1β decreased significantly mitochondrial mass in the cortical neuron body, whereas overexpression of PGC-1α and PGC-1β did not increase mitochondrial density over control.

Observed changes were accompanied by changes in expressional level; overexpression of PGC-1α and PGC-1β activated transcription of downstream reporter genes containing promoter areas from nuclear mitochondrial genes known to be activated by nuclear respiratory factor-1. Luciferase reporter constructs of murine COX8a or COX6c promoters (15) were cotransfected with PGC-1α and PGC-1β expression plasmids. Luciferase activity performed 24 h later demonstrated (Fig. 2, A and B) that overexpression of PGC-1α and PGC-1β increase relative luminescence, suggesting increased expression of COX8a and COX6c fusion proteins. PGC-1α overexpression led also to an 85 ± 8% expression increase of luciferase reporter construct of the mitofusin-2 promoter known to be directly regulated by PGC1α (16).

Next, we estimated the relative ATP levels in PGC-1α- and PGC-1β-overexpressing neurons. Neurons were cotransfected with firefly and Renilla luciferase-expressing plasmids, and firefly luciferase activity was measured 24 h later in living cells using 1-(4,5-dimethoxy-2-nitrophenyl)diazoethane-caged luciferin as a substrate. Results normalized to Renilla luciferase activity measured after the lysis of cells demonstrate a 31 and 39% increase in PGC-1α- or PGC-1β-overexpressing neurons, respectively (Fig. 3).

We also tested the possibility that overexpression of PGC-1α and PGC-1β increases mitochondrial density via decreasing mitochondrial degradation by autophagy. PGC-1α or PGC-1β was cotransfected with EGFP-LC3 plus mitochondrial pDsRED2, and the number of autophagosomes per axonal segment was estimated. However, there were no major changes in macroautophagy or in mitochondrial removal by macroautophagy in PGC-1-overexpressing neurons (Fig. 4).

PGC-1α and PGC-1β control mitochondrial density also in other neuron types. Overexpression of PGC-1α and PGC-1β in cerebellar neurons increased mitochondrial density 30% (p < 0.001) and 20% (p < 0.001), respectively, whereas suppression of PGC-1α and PGC-1β led to a 20% (p < 0.01) or 34% (p < 0.001) decrease, respectively. Overexpression of PGC-1α and PGC-1β in midbrain neurons increased mitochondrial density 19% (p < 0.001) and 17% (p < 0.001), respectively, whereas suppression of PGC-1α and PGC-1β led to a 17% (p < 0.001) or 14% (p < 0.01) decrease, respectively.

It should also be noted that the effects of PGC-1α and PGC-1β were additive and independent. In these experiments, PGC-1α increased mitochondrial density in cortical neurons 16% and PGC-1β 19%, but when overexpressed together, the mitochondrial density increased 30% (p = 0.53 for interaction). Similarly, suppression of PGC-1α and PGC-1β by the respective shRNAs was additive and independent; suppression of PGC-1α decreased mitochondrial density 29%, suppression of PGC-1β 23%, and both together 51% (p = 0.77 for interaction) (supplemental Fig. S3).

Overexpression of SIRT1 and Suppression of GCN5 Increase Mitochondrial Density

Fig. 5A demonstrates that overexpression of wild type SIRT1 increased mitochondrial density in cortical neurons up to 30%, whereas overexpression of inactive G261A-mutated SIRT1 had no effect. Fig. 5, B and C, demonstrates that this effect was mediated exclusively by PGC-1α, since the positive effect of SIRT1 on mitochondrial density disappeared in the absence of PGC-1α (p = 0.0002 for interaction) but not in the absence of PGC-1β (p = 0.89 for interaction). A similar effect was observed when suppressing GCN5 acetyltransferase. Fig. 5D demonstrates that GCN5 suppression by specific short hairpin RNA-s leads to significant increase in mitochondrial density. Similarly to SIRT1, this effect disappeared when PGC-1α was suppressed (p = 0.0033 for interaction; Fig. 5E) but sustained when PGC-1β was suppressed (p = 0.86 for interaction; Fig. 5F). It should be noted that the effects of SIRT1 overexpression and GCN5 suppression were dependent (p = 0.0084 for interaction; Fig. 5G).

Our next step was to find out whether or not the effect of SIRT1 was related with increased transcriptional activity of PGC-1α or its expression. Cortical neurons were cotransfected with a fusion protein connecting the yeast Gal4 DNA binding domain with full-length PGC-1α, along with an appropriate Gal4-UAS-luciferase reporter, and SIRT1. Luciferase assays were conducted 24 h later and revealed that SIRT1 increased Gal4-PGC1α transcriptional activity more than 2-fold (Fig. 5H). In another set of experiments cortical neurons were cotransfected with reporter gene plasmid containing mouse PGC-1α promoter area and SIRT1. Luciferase activity performed 24 h later, showed no difference between control and SIRT1-overexpressing cortical neurons (Fig. 5I).

p160 Myb-binding Protein Represses PGC-1α Activity but Independently of SIRT1

We next tested whether the effect of SIRT1 could be related to the PGC-1α repressor, p160 Myb-binding protein (p160^MBP), which interacts with the negative regulatory domain of PGC-1α and reduces its transcriptional activity. Fig. 6A demonstrates that suppression of p160^MBP leads to a considerable, 20% increase in mitochondrial density. On the other hand, overexpression of p160^MBP or its N-terminal fragment of 67-kDa p67^MBP suppressed mitochondrial density about 20%. In experiments performed with an N-terminal deletion mutant of p160^MBP (which expresses amino acids 580–1330), no effect on mitochondrial density was observed. Experiments also demonstrate that p160^MBP suppresses PGC-1α transcriptional activity but not expression. Cortical neurons were cotransfected with Gal4-PGC1α, Gal4-UAS-luciferase reporter, and p160^MBP or p67^MBP. Luciferase assays revealed that p160^MBP or p67^MBP suppressed Gal4-PGC1α transcriptional activity around 25% (Fig. 6B). At the same time, no effect of p160^MBP or p67^MBP on reporter gene plasmid containing mouse PGC-1α promoter area was observed (Fig. 6C). The effects of p160^MBP were mediated only by PGC-1α; effects of p160^MBP suppression/overexpression disappeared when PGC-1α was suppressed by shRNA against PGC-1α (Fig. 6D). The experiments depicted in Fig. 6E demonstrate also that effects of p160^MBP and SIRT1 were additive and independent. Another set of experiments demonstrated that the effects of SIRT1 were also independent from AMPK, since the effect of SIRT was present also in neurons expressing negative dominant forms of AMPK1 and -2 (p = 0.54 for interaction).

PGC-1α, PGC-1β, and SIRT1 Overexpression and GCN5 Suppression Protect A53T α-Synuclein or 120Q Huntingtin-overexpressing Cortical Neurons from Mitochondrial Loss

Our next aim was to see whether PGC-1α and PGC-1β could reverse mitochondrial loss observed in models of Huntington and Parkinson disease. Both overexpression of mutant A53T α-synuclein and 120Q huntingtin in cortical neurons induced considerable, 15–25% mitochondrial loss. However, when A53T α-synuclein or 120Q huntingtin was cotransfected with PGC-1α or PGC-1β (Fig. 7, A and B), no mitochondrial loss was observed. Similarly, overexpression of SIRT1 or suppression of GCN5 protected A53T α-synuclein- and 120Q huntingtin-overexpressing cortical neurons from mitochondrial loss (Fig. 7, C and D). The protective effects of SIRT1 and GCN5 suppression were mediated by PGC-1α, since no protective effects were observed in an experiment where PGC-1α was suppressed by respective shRNA (data not shown).

PGC-1α or SIRT1 Overexpression Protect A53T α-Synuclein or 120Q Huntingtin-overexpressing Cortical Neurons from Degeneration

In final experiments, we tested whether overexpression of PGC-1α or SIRT1 could protect the neurons against A53T α-synuclein- or 120Q huntingtin-induced degeneration. Fig. 7, E and F, demonstrates that both PGC1α and SIRT1 increased the number of surviving neurons in A53T α- synuclein- or 120Q huntingtin-transfected cultures.