Department of Pathology, University of Alabama at Birmingham, Birmingham, Alabama, USA
Center for Free Radical Biology, University of Alabama at Birmingham, Birmingham, Alabama, USA
Department of Veterans Affairs, Birmingham VA Medical Center, Birmingham, Alabama, USA
Address correspondence and reprint requests to Jianhua Zhang, Department of Pathology, University of Alabama at Birmingham, BMRII-534, 901 19th Street S., Birmingham, AL 35294-0017, USA. E-mail: email@example.com
Parkinson's disease is the second most common neurodegenerative disorder with both mitochondrial dysfunction and insufficient autophagy playing a key role in its pathogenesis. Among the risk factors, exposure to the environmental neurotoxin rotenone increases the probability of developing Parkinson's disease. We previously reported that in differentiated SH-SY5Y cells, rotenone-induced cell death is directly related to inhibition of mitochondrial function. How rotenone at nM concentrations inhibits mitochondrial function, and whether it can engage the autophagy pathway necessary to remove damaged proteins and organelles, is unknown. We tested the hypothesis that autophagy plays a protective role against rotenone toxicity in primary neurons. We found that rotenone (10–100 nM) immediately inhibited cellular bioenergetics. Concentrations that decreased mitochondrial function at 2 h, caused cell death at 24 h with an LD50 of 10 nM. Overall, autophagic flux was decreased by 10 nM rotenone at both 2 and 24 h, but surprisingly mitophagy, or autophagy of the mitochondria, was increased at 24 h, suggesting that a mitochondrial-specific lysosomal degradation pathway may be activated. Up-regulation of autophagy by rapamycin protected against cell death while inhibition of autophagy by 3-methyladenine exacerbated cell death. Interestingly, while 3-methyladenine exacerbated the rotenone-dependent effects on bioenergetics, rapamycin did not prevent rotenone-induced mitochondrial dysfunction, but caused reprogramming of mitochondrial substrate usage associated with both complex I and complex II activities. Taken together, these data demonstrate that autophagy can play a protective role in primary neuron survival in response to rotenone; moreover, surviving neurons exhibit bioenergetic adaptations to this metabolic stressor.
Exposure to the neurotoxin rotenone is a risk factor for Parkinson's disease. We tested the hypothesis that autophagy is protective against rotenone toxicity in primary neurons. Exposure to nanomolar concentrations of rotenone caused immediate mitochondrial dysfunction, associated with a suppression of macroautophagy. However, mitophagy occurred that was independent of LC3II accumulation, and the surviving neurons exhibited adaptations to their cellular bioenergetic profiles. Cotreatment with the autophagy enhancer rapamycin was protective, whereas further inhibition of autophagy with 3-methyladenine (3-MA) exacerbated cell death, resulting in additional bioenergetic adaptations in the surviving neurons.
Parkinson's disease (PD) brains exhibit decreased complex I activity (Schapira and Gegg 2011), which is important since this enzyme is both the rate-limiting step for mitochondrial respiratory chain activity, and a key site for generation of reactive oxygen species (Murphy 2009). Mitochondrial defects are widespread in both the substantia nigra and cortex in PD (Keeney et al. 2006; Parker et al. 2008), but the underlying mechanisms that lead to mitochondrial dysfunction are unclear. Meta-analyses of several studies indicated that pesticide exposure potentially increases PD risk on average by ~ 1.5–3 fold (Priyadarshi et al. 2000, 2001; van et al. 2012; Pezzoli and Cereda 2013). Also, a recent study reported that exposure to paraquat and rotenone in farming communities increased the incidence of PD with an odds ratio of ~ 2–3 fold (Tanner et al. 2011). Rotenone is an herbicide, insecticide, and piscicide, and is the most potent member of the rotenoid family of neurotoxins found in tropical plants. Rotenone can pass through the blood brain barrier and plasma membrane, bind to complex I in striatal sections from rodent brains with a Kd of ~ 55 nM, decreasing its activity, and inducing oxidative stress (Higgins and Greenamyre 1996; Panov et al. 2005). In animal models, dopaminergic neurons appear to be particularly susceptible to rotenone-induced degeneration (Betarbet et al. 2000). Oral administration of rotenone at 30 mg/kg for 56 days in mice has been shown to cause α-synuclein accumulation and dopaminergic neurodegeneration (Inden et al. 2011). In vitro studies have shown that short term (1 week), low concentrations of rotenone induce accumulation of soluble and insoluble α-synuclein, and after 4 weeks, increased apoptosis in SK-N-MC neuroblastoma cells (Sherer et al. 2002).
Autophagy plays an important role in mitochondrial quality control (Higdon et al. 2012; Hill et al. 2012; Lee et al. 2012; Benavides et al. 2013; Dodson et al. 2013a,b; Liang et al. 2013; Mitchell et al. 2013a,b; Zhang 2013; Boyer-Guittaut et al. 2014; Giordano et al. 2014; Levonen et al. 2014; Redmann et al. 2014), and changes in autophagy have been observed in postmortem PD brains (Anglade et al. 1997). In addition, some of the familial PD genes such as pink1, parkin, and DJ-1 play important roles in the process of autophagy of mitochondria, known as mitophagy (Youle and Narendra 2011). In undifferentiated SH-SY5Y cells, it has been shown that rotenone inhibits autophagic flux prior to inducing cell death, and that restoration of autophagic activity is partially protective (Pan et al. 2009; Xiong et al. 2011). Importantly, this transformed cell line is metabolically distinct from neurons, and in the undifferentiated state exhibits aerobic glycolysis that likely changes their susceptibility to mitochondrial neurotoxins (Schneider et al. 2011). Studies performed in primary cortical neurons have shown that concentrations of rotenone far in excess of that resulting in complete inhibition of complex I activity (e.g. 250 nM–1 μM) induce externalization of the mitochondrial-specific cardiolipin, which serves as a tag for the mitochondria to undergo mitophagy (Chu et al. 2013). How low nM concentrations of rotenone, which are more likely to occur during environmental exposure, affect mitochondrial function in primary neurons, and whether mitophagy is engaged as a protective response, have not been critically examined, and are the focus of the present study.
Primary rat cortical neurons were isolated from embryonic day 18 rats purchased from Charles River Laboratories Inc., Wilmington, MA, USA. All animal handling has been approved by University of Alabama at Birmingham IACUC. Briefly, 80 000 cells per well were plated in the XF24 plates for cell viability, western blotting, and mitochondrial function assays. A total of 250 000 cells per well were plated for microscopy in four-well chamber slides. Neurons were grown in neurobasal media with B27 supplement, l-glutamine, and Penn/Strep, with half of the media changed every 3 days.
Measurement of cellular bioenergetic function
To measure cellular bioenergetics in primary cortical neurons, the Seahorse Bioscience XF24 Extracellular Flux Analyzer (XF24; Seahorse Bioscience, Billerica, MA, USA) was used (Dranka et al. 2011; Schneider et al. 2011; Giordano et al. 2012; Chacko et al. 2013; Chu et al. 2013). Both the oxygen consumption rate (OCR) in pmol/min and the extracellular acidification rate (ECAR) in mpH/min were measured and normalized to total protein amount per each individual well, determined by the DC protein assay (Bio-Rad Laboratories, Hercules, CA, USA).
Mitochondrial function in live neuronal cultures was assessed using the sequential injection of oligomycin, Carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), and antimycin A, with concentrations optimized to be 1 μg/mL, 1 and 10 μM, respectively. Parameters of mitochondrial function were analyzed as described in our prior publications (Dranka et al. 2011; Giordano et al. 2012; Chacko et al. 2013). Briefly, specific inhibitors are used in the ‘mitochondrial stress test’ to characterize different bioenergetics parameters. The first inhibitor injected, after a stable baseline is established, is oligomycin which inhibits ATP synthase (mitochondrial complex V) and thus the portion of OCR that is inhibitable by oligomycin is termed ATP-linked OCR (OCR before oligomycin injection minus OCR after oligomycin injection). The remaining OCR is largely because of proton transport back into the mitochondrial matrix via non-mitochondrial complex V-mediated mechanisms, and is termed Proton leak OCR (OCR after oligomycin injection minus after antimycin A injection). Maximal OCR after uncoupling by FCCP represents the unconstrained activity of the mitochondrial electron transport chain that can be achieved with the cellular substrate supply. The reserve capacity is calculated by subtracting the basal OCR from the FCCP stimulated maximal OCR, and represents the bioenergetic function available for cellular energetics to combat cellular stress (Dranka et al. 2011). Basal ECAR was determined at the time point before oligomycin addition, and maximal ECAR was determined at the time point after oligomycin addition. Rotenone was freshly prepared in dimethylsulfoxide and diluted into the medium. Cotreatment with rapamycin or 3-methyladenine (3-MA) for 24 h prior to the experiment was conducted prior to bioenergetics analysis in the XF24.
Mitochondrial complex I and complex II activity in permeabilized neuronal cultures was measured by injection with 20 μg/mL saponin plus substrates (ADP, pyruvate, malate, and succinate). Then rotenone (2 μM) was added to determine respiration because of complex I, followed by antimycin A (10 μM) for non-mitochondrial OCR (Salabei et al. 2014).
Cell viability was measured by the trypan blue exclusion assay as described previously (Giordano et al. 2012). About 80 000 cells/well were plated on a XF24 plate. Cells were exposed to rotenone with or without rapamycin and 3-MA for 24 h, after which cells were trypsinized and mixed with trypan blue before counting. Cells that excluded trypan blue were considered viable.
About 250 000 cells per well were plated on four-well Lab-Tek chambered cover-glass slides previously coated with poly-l-lysine. On DIV 7, cells were treated with rotenone for 24 h. Then cells were washed with phenol free media and incubated with 25 nM MitoTracker Red and 100 nM LysoTracker Green (Sigma, St Louis, MO, USA). After 30 min, cells were imaged using the Zeiss LSM 710 Confocal Microscope (Zeiss, Thornwood, NY, USA). Red and green pixel intensity overlay was determined using quantification software on the Nikon eclipse Microscope. n > 20 cells per group.
Western blot analysis
About 80 000 cells per well were grown in XF24 plates and treated for 2 or 24 h with rotenone alone, rotenone + rapamycin, or rotenone + 3-MA. Medium was removed and 20 μL of lysis buffer were added to each well. Samples were triturated in the wells, collected, and 5 μL of sample buffer was added. Protein extracts were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and probed with anti-LC3 and anti-actin antibodies (Sigma). Relative levels of protein were quantified using NIH Image J software (NIH, Bethesda, MD, USA) and all samples were compared to a standard control on all gels.
Relative mtDNA copy number
DNA was extracted from primary neurons following treatment. Quantitative real-time PCR was performed by using a SYBR Green master mix (Life Tech Corp, Grand Island, NY, USA) in an ABI 7500. The primer sequences used for mtDNA were mtDNA-F (5′-CCAAGGAATTCCCCTACA CA-3′) and mtDNA-R (5′-GAAATTGCGAGAATGGTGGT-3′). The primer sequences for the nuclear DNA were 18S-F (5′-CG AAAGCATTTGCCAAGAAT-3′) and 18S-R (5′-AGTCGGCATC GTTTATGGTC-3′) and targeted the human nuclear 18S DNA. Cycling conditions were as follows: 94°C for 15 s, followed by 40 cycles at 94°C for 15 s, 60°C for 1 min, and 60°C for 1 min. The mtDNA copy number was normalized to the amplification of the 18S nuclear amplicon.
mtDNA damage assay
Mitochondrial DNA damage (mtDNA) was evaluated by modified quantitative PCR (QPCR) method as described previously (Schapira and Gegg 2011). Briefly, total DNA was extracted and used as PCR sample. The primer sequences used for mtDNA long segment (16 kb) were mtLongF (5′-GGACAAATATCATTCTGAGGAGCT-3′) and mtLongR (5′-GGATTAGTCAGCCGTAGTTTACGT-3′). The primer sequences for mtDNA short (80 bp) segment were mtShortF (5′-CCAAGGAATTCCCCTACACA-3′) and mtShortR (5′-GAAATTGCGAGAATGGTGGT-3′). The mtDNA long segment and the short segment were amplified using AccuPrime™ Taq DNA Polymerase High Fidelity kit (Life Tech Corp) and separated by agarose gel electrophoresis, respectively. mtDNA long PCR conditions were as follows: 94°C for 11 s, followed by 25 cycles of denaturation at 94°C for 15 s, annealing and extension at 67°C for 12 min, final extension at 72°C for 10 min. mtDNA Short PCR conditions were as follows: 94°C for 6 s, followed by 18 cycles of denaturation at 94°C for 20 s, annealing and extension at 65°C for 1 min, and final extension at 72°C for 10 min. The gels were stained by ethidium bromide and visualized with Alpha Imager, and densitometry analysis performed using Image J software. Lesion frequency per 16 kb of mtDNA was calculated using the following equation (Knight-Lozano et al. 2002).
Data are reported as mean ± SEM. Comparisons between two groups were performed with unpaired Student's t-tests. A p value of < 0.05 was considered statistically significant.
Effects of rotenone on cell viability and cellular bioenergetics function
Primary cortical neurons (DIV 7) were treated with increasing concentrations of rotenone for 2 and 24 h and cell viability assessed. Rotenone induced a concentration and time-dependent cell death, after exposure for 24 h (Fig. 1a). To test if early mitochondrial dysfunction precedes cell death, cellular bioenergetics were measured. After four baseline measurements of OCR, rotenone (0–100 nM) was injected into the wells, and OCR measured over the next 2 h (Fig. 1b) followed by a mitochondrial stress test (Dranka et al. 2011) (Fig. 1c–h).
Concentrations of rotenone as low as 10 nM caused an immediate decrease in the major parameters of mitochondrial function consistent with inhibition of functional electron transport at complex I (Fig. 1b and c). These include a decrease in ATP-linked respiration (Fig. 1d), as well as maximal OCR and reserve capacity (Fig. 1f and g). There was no change in non-mitochondrial OCR (Fig. 1h) following rotenone treatment at any concentration.
The ECAR, which represents changes in glycolytic flux, was also measured for 2 h after the injection of rotenone in the XF24. Immediately after the injection of 1, 10, or 100 nM rotenone, the glycolytic rate was increased compared to 0 nM and remained constant for 2 h (basal ECAR, Fig. 1i). Injection of mitochondrial complex V inhibitor oligomycin further increased ECAR in control cells as a result of inhibition of mitochondrial function and the expected compensatory unregulation of glycolysis to meet cellular energy demands (maximal ECAR, Fig. 1i). At the concentrations of 1–100 nM rotenone, oligomycin was unable to further stimulate glycolysis after ATP synthase inhibition (Fig. 1i). After 2 h rotenone exposure, there is the anticipated association between the decrease in basal OCR and an increase in basal ECAR with increasing concentrations of rotenone (Fig. 1j).
Effects of rotenone on autophagic flux and mitophagy
Autophagic flux was measured in response to 10 nM rotenone at both 2 and 24 h, by treating primary cortical neurons with rotenone in the presence or absence of chloroquine (CQ) or bafilomycin (Baf), both of which block autophagy completion (Higdon et al. 2012; Benavides et al. 2013; Dodson et al. 2013b; Liang et al. 2013; Mitchell et al. 2013b; Boyer-Guittaut et al. 2014). Western blot analysis for LC3 protein levels showed that 10 nM rotenone inhibited autophagy at both 2 (Fig. 2a) and 24 h (Fig. 2b).
To determine if dysfunctional mitochondria are being cleared by mitophagy, cells were treated with 10 nM rotenone for 24 h, and stained with MitoTracker and LysoTracker. We found that rotenone had a significant impact on MitoTracker fluorescence, but did not affect LysoTracker staining, which is consistent with partial suppression of the mitochondrial membrane potential (Fig. 3). The increased colocalization of mitochondria with lysosomes is consistent with increased mitophagy in response to rotenone (Fig. 3). Interestingly, the greatest extent of colocalization was seen in the soma and was absent in neuronal processes (Fig. 3).
Effects of rapamycin and 3-MA on autophagy and cell viability in response to rotenone exposure
Autophagic flux analysis was performed by measuring LC3II/LC3I ratio in the presence and absence of 40 μM chloroquine (CQ). A 2 h exposure of rotenone (10 nM), alone or in combination with rapamycin (1 μM), decreased autophagy (Fig. 4a). At the 24 h time point, rotenone alone still decreased autophagy. In contrast, rapamycin significantly increased autophagy above basal levels, and this was substantially inhibited by rotenone (Fig. 4b). The inhibitor of autophagy, 3-MA (10 mM), significantly decreased the LC3II/LC3I ratio compared to CQ treated cells, indicating a decrease in autophagy in the absence or presence of rotenone (Fig. 4b).
To assess the impact on viability, primary cortical neurons were cotreated with rotenone ± rapamycin, or rotenone ± 3-MA for 2 or 24 h. At the 2 h time point, neither of the modulators of autophagy showed a biologically significant impact on viability, suggesting that the effects on bioenergetics are an early response, and autophagy is engated later. In contrast, at the 24 h time point, rapamycin modestly enhanced cell viability at rotenone concentrations of 1–100 nM, while 3-MA exacerbated cell death at 0.1–100 nM rotenone (Fig. 5).
Effects of rapamycin and 3-MA on mitochondrial function
In primary cortical neuronal cultures, 10 nM rotenone exposure for 24 h resulted in decreased maximal and reserve capacity OCR in the remaining viable cells, but with no significant effect on basal mitochondrial respiration, suggesting some recovery of bioenergetic function (Fig. 6a). Interestingly, exposure to rapamycin alone for 24 h had similar effects on maximal and reserve capacity OCR compared to 10 nM rotenone, suggesting an off-target effect on mitochondrial function. Cotreatment of rapamycin with rotenone resulted in a further decrease in basal, ATP-linked, maximal, and reserve capacity OCR. The most striking effect was with 3-MA alone or in combination with rotenone, which additively decreased basal, ATP-linked, maximal, and reserve capacity OCR (Fig. 6a). There were only subtle differences among the six groups regarding ECAR. 3-MA exposure led to a slight decrease in basal ECAR, and 3-MA + rotenone decreased maximal ECAR (Fig. 6b and c). To further assess the impact on mitochondrial function, permeabilization of these cultured neurons after 24 h rotenone exposure and/or cotreatment with rapamycin or 3-MA was performed, followed by complex I and complex II activity assays. Interestingly, at 24 h, complex I inhibition was modest with rotenone alone even though the basal OCR after rotenone in intact cells is comparable to control. Surprisingly, rapamycin increased complex I activity but inhibited complex II suggesting a reprogramming of mitochondrial substrate usage associated with treatment (Fig. 6d). Cells exposed to a cotreatment of rapamycin and rotenone exhibited a similar complex I activity to that of rotenone alone with partial restoration of complex II activity. 3-MA decreased both complex I and complex II activities, with an additive effect on complex I with rotenone treatment (Fig. 6d). mtDNA copy number and damage are not significantly different among the six treatment groups (Figure S1).
In this study, we investigated changes to mitochondrial function in response to the environmental neurotoxin rotenone in primary neurons, and the impact of modulating autophagy on neuronal survival in response to this compound.
As expected, rotenone has a significant effect on mitochondrial function consistent with inhibition at complex I (Figs 1 and 6). Of particular note is the sensitivity of the key aspects of cellular mitochondrial function at only 10 nM rotenone. These data highlight a critical role of complex I in maintaining the bioenergetics of primary neurons, and an inability of other metabolic substrates to compensate. It is important to note that in a cellular setting, inhibition of complex I will also inhibit flux through the TCA cycle, increasing the impact of rotenone inhibition on mitochondrial electron transport. Autophagic flux was inhibited by rotenone (Fig. 2), suggesting that the cells could also exhibit deficits in removal of damaged cellular constituents. In contrast, despite the fact that autophagy is inhibited, rotenone only decreased MitoTracker fluorescence without affecting LysoTracker staining (Fig. 3). This might be as a result of partial inhibition of mitochondrial function rather than a decrease in mitochondrial mass, since there was no apparent change in mtDNA copy number (Figure S1).
Associated with the decrease in mitochondrial membrane potential, MitoTracker and LysoTracker colocalization studies (Fig. 3) demonstrated that 10 nM rotenone for 24 h increased the delivery of mitochondria to the lysosomes, indicating an increase in mitophagy. The observation that rotenone induces mitophagy while general autophagic flux is inhibited suggests that a mitochondrial-specific lysosomal degradation pathway may be activated, as has recently been suggested, via a formation of mitochondrial spheroids (Ding et al. 2012; Ni et al. 2013; Yin and Ding 2013). Potential signaling mechanisms include, but are not limited to, preferential inhibitory modification of general macroautophagy but not mitophagy proteins, or preferential fission/fusion machinery activities and PARKIN/PINK1 function in response to loss of mitochondrial membrane potential. Increasing mitophagy without changing mtDNA copy number may be because of the following possibilities: (i) at 24 h after rotenone, although the mitophagic events are increased, mtDNA has not been digested at this time point, (ii) there is concurrent mitochondrial biogenesis that maintains the mitochondrial population, or (iii) the population of mitochondria that colocalizes with lysosomes are mainly mitochondrial fragments that are devoid of mtDNA.
The clearance of damaged mitochondria may improve the quality of the mitochondrial population, and so protect the primary cortical neurons from rotenone toxicity. These data are supported by the finding that inhibition of autophagy exacerbated rotenone toxicity, whereas activation of autophagy was partially protective (Fig. 5). Since we observed that in the presence of rotenone, rapamycin no longer activates autophagic flux, the partially protective effect of rapamycin may be independent of general macroautophagy, but may be dependent on mitophagy. In addition, rapamycin inhibits protein translation by inhibition of mammalian target of rapamycin, thereby decreasing the levels of short-lived proteins without significantly changing long-lived proteins; this may lead to changes in cellular signaling pathways. In our study, rapamycin enhanced complex I activity and decreased complex II activity on its own. In the presence of 10 nM rotenone, rapamycin did not change complex I activity, but further decreased complex II activity. Whether the decreased complex II activity contributed to rapamycin's ability to partially protect rotenone toxicity is unclear. To conclusively determine this possibility would require specific means to selectively up-regulate complex II activity to restore them to control levels in the presence of rapamycin. Removal of damaged mitochondria may also contribute to the protective function of rapamycin. However, since the mtDNA damage assay did not reveal any significant differences between rotenone exposed cells and rotenone + rapamycin exposed cells, more extensive studies using mitochondrial proteomic and lipidomic analysis are needed.
The interpretation of the bioenergetic profiles in response to 3-MA and rapamycin are complicated, since both these compounds had an effect on cellular mitochondrial function and complex I and II activities in the absence of rotenone. Cellular mitochondrial parameters were most severely affected with the combination of both rotenone and 3-MA, which also incidentally exhibited the greatest toxicity (Fig. 5). This finding is consistent with the primary hypothesis that the prevention of autophagy is detrimental to cellular bioenergetics and viability in the presence of rotenone. Interestingly, the approximately 90% inhibition of basal respiration in the presence of both 3-MA and rotenone is greater than the additive effect of rotenone or 3-MA alone, thus suggesting that inhibition of autophagy can accelerate bioenergetic defects. The rapamycin data are more difficult to explain, since the increase in viability (Fig. 5) was not reflected by an obvious improvement in cellular bioenergetics (Fig. 6). It is likely that the bioenergetic function in the presence of rapamycin and rotenone is adequate to maintain cell viability. Taken together, these observations indicate that general autophagy and mitophagy are differentially regulated in neurons, and that enhancement of general autophagy by rapamycin may result in metabolic modulation of the surviving cells.
Acknowledgments and conflict of interest disclosure
We thank members of the Zhang and the Darley-Usmar laboratories for discussions and technical help. This work was supported by NIH (NIHR01-NS064090) and a VA merit award (to JZ).
All experiments were conducted in compliance with the ARRIVE guidelines. VDU is a member of the Seahorse Biosciences Scientific Advisory Board.