To evaluate our hypothesis that Mito-Met is more potent than its parent compound metformin in providing neuroprotection, we wanted to compare the effects of metformin and Mito-Met on the activation of pro-survival kinase PKD1, reported by our group to have neuroprotective effects against oxidative stress-induced cell death in DAergic neurons (Asaithambi et al., 2011, 2014). We first treated N27 DAergic neuronal cells with the widely accepted dose range of metformin (100 and 1,000 μM) and Mito-Met (100 and 300 nM) for 3 h and checked the phosphorylation of PKD1. We selected Mito-Met concentrations of 100 nM and 300 nM as optimal doses after performing a preliminary dose–response study ranging from 10 nM to 1,000 nM. Also, this dose range of mitochondria-targeted compounds has been shown to greatly increase the mitochondrial concentration (Zielonka et al., 2017; Kalyanaraman et al., 2018). As evident from Figures 1A,B, although both metformin and Mito-Met induced the phosphorylation of PKD1 activation loop (S744/748), Mito-Met was more than 300-fold more efficacious than metformin in activating PKD1. To determine the time course response, we next treated N27 DAergic cells with 100 and 300 nM Mito-Met for 3 h and 6 h and measured both the PKD1 activation loop phosphorylation and the C-terminal phosphorylation (S916). As shown in Figures 1C,D, Mito-Met significantly increased the phosphorylation of both S744/748 and S916 of PKD1 at both 3 and 6 h time points. However, expression of total PKD1 was not affected by Mito-Met treatment (Figure 1C). Following the observation that Mito-Met enhances PKD1 activation, we wanted to determine whether Mito-Met can also increase the activation of Akt because Akt has been previously shown to have neuroprotective effects in cell culture and animal models of PD (Ries et al., 2006; Gunjima et al., 2014). Immunoblot analysis revealed that treatment of N27 cells with Mito-Met (100 nM and 300 nM) for 3 and 6 h increased Akt phosphorylation (Figure 2A). To further explore the potential neuroprotective mechanism of Mito-Met, we examined whether Mito-Met can activate AMPK in DAergic neuronal cells since metformin, the parent compound of Mito-Met, is known to activate AMPK in different cell types, including neurons (Ouyang et al., 2011; Stephenne et al., 2011; Hou et al., 2015). AMPK has also been reported to exert a neuroprotective effect against DAergic neurodegeneration (Ng et al., 2012). To determine AMPK activation, we used an antibody directed against phosphorylated Thr172 within the catalytic subunit (alpha) of AMPK. As shown in Figure 2A, Mito-Met treatment increased AMPK Thr172 phosphorylation in N27 cells. Next, we reasoned that PKD1 might lie upstream of Akt and AMPK kinase during Mito-Met treatment. To test this hypothesis, N27 cells were pretreated with a pharmacological inhibitor of PKD1 (CID755673) for 1 h and then cotreated with CID755673 (50 μM) and Mito-Met (100 nM) for 3 h. Interestingly, treatment with CID755673 blocked the Mito-Met-induced activation of both Akt and AMPK kinase (Figure 2B). Collectively, these data suggest that Mito-Met can effectively activate PKD1 and its potential downstream neuroprotective effectors, Akt and AMPK kinase signaling, in DAergic cells.

After observing that Mito-Met can enhance the activation of PKD1, we investigated whether Mito-Met stimulates mitochondrial biogenesis. We first investigated the effect of Mito-Met on TFAM expression. TFAM is a transcription factor that is required for mtDNA replication and transcription to regulate mitochondrial biogenesis (Virbasius and Scarpulla, 1994; Scarpulla, 2008). N27 cells were treated with 100 and 300 nM Mito-Met for 3 and 6 h and TFAM levels were determined by Western blot analysis. As seen in Figures 3A,B, the expression of TFAM protein was significantly increased in N27 cells treated with Mito-Met. Next, we measured TFAM mRNA expression using real-time RT-PCR at 3 h time point. Exposure of N27 cells to Mito-Met (100–1,000 nM) significantly increased TFAM mRNA expression (Figure 3C). We also determined mtDNA copy number using mitochondrial ND1 and nuclear β-actin genes and found that Mito-Met treatment (1,000 nM) significantly increased mtDNA content in N27 cells (Figure 3D). Furthermore, we evaluated the effect of Mito-Met on cellular bioenergetics in DAergic neuronal cells. N27 cells were treated with 100 and 300 nM Mito-Met for 3 h and the oxygen consumption rates (OCR) were measured using a Seahorse XF96 analyzer. Interestingly, we observed significant increases in basal OCR and ATP-linked respiration in Mito-Met-treated N27 cells, while rotenone, a well-known mitochondrial complex I inhibitor, almost completely blocked mitochondrial respiration (Figures 4A–C). These results suggest that Mito-Met can stimulate mitochondrial biogenesis and oxidative phosphorylation in DAergic cells.

To further examine the potential neuroprotective influence of Mito-Met on mitochondrial function, we examined whether Mito-Met can reduce MPP⁺-induced mitochondrial damage. N27 DAergic cells were pretreated with 300 nM Mito-Met for 1 h and then cotreated with Mito-Met and MPP⁺ (300 μM) for 16 h and stained with MitoTracker to visualize mitochondrial morphology. Expectedly, MPP⁺ treatment strongly induced mitochondrial fragmentation (Figure 5A); however, Mito-Met pre- and co-treatment significantly ameliorated mitochondrial morphology in MPP⁺-treated cells as evidenced by improved mitochondrial length and circularity (Figures 5B,C). Next, we tested the protective effect of Mito-Met in primary DAergic neurons cultured from E14 mouse embryos. The primary nigral neurons were treated with 10 μM MPP⁺ in the presence or absence of Mito-Met (100 nM) for 24 h and immunostained with a TH antibody. To evaluate whether Mito-Met treatment protects against MPP⁺-induced DAergic neurodegeneration, we measured the length of DAergic neurites. The immunocytochemical analysis revealed that Mito-Met treatment significantly protected against MPP⁺-induced neurotoxicity, as evidenced by a significant increase in the length of neuronal processes in primary nigral cultures cotreated with MPP⁺ and Mito-Met compared with cultures treated with MPP⁺ alone (Figures 6A,B). Together, these data clearly demonstrate that Mito-Met can effectively protect DAergic neurons from MPP⁺-induced toxicity.

The results of our in vitro experiments demonstrated that Mito-Met affords neuroprotection against DAergic neurotoxicity. Here, we wanted to test the neuroprotective efficacy of Mito-Met in the MitoPark mouse model of PD. As seen in Figure 7A, MitoPark and control mice were treated with either saline or Mito-Met (10 mg/kg) three times per week by oral gavage for 8 weeks. One day prior to sacrifice, mice were analyzed for their locomotor activities. MitoPark mice develop progressive motor deficits from 12 weeks of age onwards (Ekstrand et al., 2007). As expected, 20-week-old MitoPark mice had a dramatic reduction in total movement (Figure 7B). Notably, improved locomotion was observed in the MitoPark mice-treated with Mito-Met. Mito-Met treatment reversed the reduction in horizontal activity (Figure 7C), total distance traveled (Figure 7D), number of movements (Figure 7E), and stereotypy counts (Figure 7F) of MitoPark mice. Moreover, we also determined whether Mito-Met treatment attenuates striatal DA depletion in MitoPark mice. Mice were sacrificed 1 day after the last Mito-Met treatment and striatal tissues were processed for HPLC analysis of DA. Neurochemical analysis revealed that the DA and DOPAC levels were significantly higher in the MitoPark mice-treated with Mito-Met than the MitoPark mice-treated with saline (Figures 8A,B). Collectively, these results suggest that Mito-Met treatment can improve motor deficits and at least partially prevent striatal DA depletion in MitoPark mice.

1-Methyl-4-phenyl pyridinium iodide (MPP⁺ iodide) and anti-β-actin antibody were purchased from Sigma-Aldrich (St. Lois, MO). The Bradford protein assay kit was purchased from Bio-Rad. Antibodies against PKD1, α-tubulin, and mtTFA were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-phospho PKD1 (p-S744/748 and p-S916), phospho-Akt (p-S473), total Akt, phospho-AMPKα (p-Thr172) antibodies were obtained from Cell Signaling Technology (Danvers, MA). Anti-TH antibody was purchased from Millipore (Billerica, MA). The PKD1 inhibitor CID755673 was purchased from Tocris Bioscience (Ellisville, MO). Alexa 680-conjugated anti-mouse secondary antibody, MitoTracker Red FM, and all cell culture reagents were obtained from Invitrogen (Carlsbad, CA). IRDye 800-conjugated anti-rabbit secondary was purchased from Rockland Labs (Gilbertsville, PA).

The rat DAergic N27 cell line was a kind gift from Dr. Kedar N. Prasad (University of Colorado Health Sciences Center, Denver, CO) and cultured as described previously (Kaul et al., 2003). Briefly, N27 cells were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 50 units penicillin, and 50 μg/mL streptomycin and maintained at 37°C in a 5% CO2 atmosphere. We also used the primary mesencephalic neuronal culture to assess the protective effect of Mito-Met against MPP⁺-induced TH-positive neurite loss. The primary mesencephalic culture was prepared from the ventral mesencephalon of gestational 14-d-old mouse embryos as described previously (Zhang et al., 2007). Briefly, the mesencephalic tissues from E14 mouse embryos were dissected and maintained in ice cold Ca²⁺-free Hank’s balanced salt solution, and then dissociated in Hank’s balanced salt solution containing trypsin-0.25% EDTA for 30 min at 37°C. Cultures were maintained in Neurobasal medium supplemented with B27. Half of the culture medium was replaced every 2 d. Approximately 6- to 7-d-old cultures were used for experiments. Primary mesencephalic neuronal cells were treated with 10 μM MPP⁺ in the presence or absence of Mito-Met (100–300 nM) for 24 h.

Total RNA was isolated from cells using the Absolutely RNA Miniprep Kit (Stratagene, La Jolla, CA), and 1–2 μg RNA was used to generate cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Real-time PCR was performed on an Mx300P QPCR system (Stratagene) using the RT² SYBR Green qPCR Master Mix Kit (Qiagen) and QuantiTect Primer Assay Kit (Qiagen). All RT-qPCR reactions were performed in triplicate and normalized to the β-actin housekeeping gene. The primer sequences and PCR conditions are available upon request. Dissociation curves were run to verify the singularity of the PCR product. The data were analyzed using the comparative threshold cycle (Ct) method.

N27 cells were plated in 6-well plates at a density of 2 × 10⁵ cells/well. Twenty-four hours later, cells were treated with 0.1–1 μM Mito-Met for 24 h. Genomic DNA was isolated using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol, and mtDNA content was determined by quantitative PCR with SYBR green. 10 ng nuclear DNA and 1 ng mitochondrial DNA were amplified using the primers for mitochondrial ND1 and nuclear β-actin genes. Reactions were performed using an Mx300P QPCR system (Stratagene).

After treatments, cells or brain tissues were collected and resuspended in a modified RIPA buffer containing protease and phosphatase inhibitors. Briefly, lysates containing equal amounts of protein were loaded in each lane and separated on a 10–15% SDS-PAGE gel and transferred onto a nitrocellulose membrane (Bio-Rad). The membranes were blocked in LI-COR blocking buffer for 1 h at room temperature (RT) and then incubated overnight at 4°C with the corresponding primary antibodies. Western blot was performed using IRDye 800 anti-rabbit, Alexa Fluor 680 goat anti-mouse, and Alexa Fluor 680 donkey anti-goat secondary antibodies. Western blot images were captured and analyzed with an Odyssey Infrared Imaging System (LI-COR, Lincoln, NE).

Primary mesencephalic DAergic neurons were fixed with 4% paraformaldehyde for 30 min at RT. After washing, cells were permeabilized with blocking agent (2% bovine serum albumin, 0.5% Triton X-100, and 0.05% Tween-20 in PBS) for 1 h. Cells were then incubated with an anti-TH antibody (1:1,200, Millipore) at 4°C overnight. Fluorescently conjugated secondary antibody (Alexa Fluor 488-conjugated anti-mouse antibody, 1:2,000) was used followed by incubation with 10 μg/mL Hoechst 33342 for 5 min at RT to stain the nucleus. Cover slips were mounted on glass slides and cells were visualized by Z-Stack imaging with the Leica confocal microscope. 3D image reconstruction was performed using IMARIS software. The length of TH-positive neuronal processes in primary DAergic neurons from each coverslip was measured using MetaMorph software (Molecular Devices) as described previously (Ghosh et al., 2013; Charli et al., 2015). For quantification of neuronal processes, pictures were taken at 60X magnification. TH-positive neuronal processes were counted in six individual cultures for each treatment.

N27 cells were stained with MitoTracker Red FM (200 nM) for 30 min at 37°C, and then fixed with 4% paraformaldehyde for 30 min at RT. After washing, cells were permeabilized with 0.5% Triton X-100 in PBS for 5 min at RT, as described previously (Charli et al., 2015). Cover slips were mounted on glass slides and imaged through a Nikon TE2000 microscope with a SPOT color digital camera (Diagnostic Instruments, Sterling Heights, MI). Quantification of mitochondrial parameters such as mitochondrial length and degree of circularity was performed using a macro text file plug-in for ImageJ (Dagda et al., 2009).

The Seahorse XF96 Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA) was used to measure oxygen consumption rates (OCR) in N27 cells. Briefly, N27 cells were seeded at 10,000 cells/well in a Seahorse Bioscience polystyrene microplate. Twenty-four hours later, cells were treated with 100 and 300 nM Mito-Met for 3 h. After the treatment, cells were changed to assay media and the N27 cell-containing culture plates were loaded into the Seahorse instrument to collect baseline OCR measurements. The experimental agents (1 μg/mL oligomycin, 3 μM FCCP, and 10 μM antimycin A) were injected using the included ports on the XF96 cartridges and OCR measurements were collected. Basal OCR, ATP-linked OCR, maximal OCR, and spare respiratory capacity were calculated, as described previously (Dranka et al., 2010).

MitoPark and C57BL/6 mice (12 weeks old) were housed under standard conditions: constant temperature (22 ± 1°C), humidity (relative, 30%), and a 12 h light/dark cycle. Mice were given free access to food and water. Animal care and protocol procedures were approved and supervised by the Institutional Animal Care and Use Committee (IACUC) at Iowa State University (Ames, IA). Mice were treated with either saline or Mito-Met (10 mg/kg) via oral gavage three times weekly for 8 weeks. After the treatments, mice were subjected to behavioral, neurochemical, and biochemical measurements.

The striatal DA, dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA) levels were quantified using HPLC with electrochemical detection. Samples were prepared and quantified, as described previously (Ghosh et al., 2013). Briefly, neurotransmitters from striatal tissues were extracted in 0.1 M perchloric acid containing 0.05% Na₂EDTA and 0.1% Na₂S₂O₅ and isoproterenol (internal standard). DA, DOPAC, and HVA were separated isocratically using a C-18 reversed-phase column with a flow rate of 0.6 mL/min. An HPLC system (ESA, Bedford, MA) with an automatic sampler equipped with a refrigerated temperature control was used for these experiments.

The open field test was performed using an infra-red (IR) activity-tracking system (VersaMax monitor, model RXYZCM-16, and analyzer, model VMAUSB, AccuScan, Columbus, OH). We recorded spontaneous exploratory activity for a 10-min test period after a 2-min acclimatization period. The system quantifies activity by counting and mapping IR beam breaks within a 20 × 20-cm arena. The measures for horizontal activity, number of movements, and stereotypy count (beam breaks) was obtained and plotted. Before test day, animals were acclimatized to the instrument for 10 min for two consecutive days. Data were collected and analyzed using VersaPlot and VersaDat software.

Data analysis was performed using Prism 4.0 software (GraphPad Software, San Diego, CA). Data were first analyzed by using one-way ANOVA and then the Tukey multiple comparison test was used for statistical comparisons. Differences with p ≤ 0.05 were considered significant.