Enhanced Autophagy from Chronic Toxicity of Iron and Mutant A53T α-Synuclein

Parkinson disease (PD), a prevalent neurodegenerative motor disorder, is characterized by the rather selective loss of dopaminergic neurons and the presence of α-synuclein-enriched Lewy body inclusions in the substantia nigra of the midbrain. Although the etiology of PD remains incompletely understood, emerging evidence suggests that dysregulated iron homeostasis may be involved. Notably, nigral dopaminergic neurons are enriched in iron, the uptake of which is facilitated by the divalent metal ion transporter DMT1. To clarify the role of iron in PD, we generated SH-SY5Y cells stably expressing DMT1 either singly or in combination with wild type or mutant α-synuclein. We found that DMT1 overexpression dramatically enhances Fe2+ uptake, which concomitantly promotes cell death. This Fe2+-mediated toxicity is aggravated by the presence of mutant α-synuclein expression, resulting in increased oxidative stress and DNA damage. Curiously, Fe2+-mediated cell death does not appear to involve apoptosis. Instead, the phenomenon seems to occur as a result of excessive autophagic activity. Accordingly, pharmacological inhibition of autophagy reverses cell death mediated by Fe2+ overloading. Taken together, our results suggest a role for iron in PD pathogenesis and provide a mechanism underlying Fe2+-mediated cell death.

Parkinson disease (PD) 3 is the most common motor neurodegenerative disorder, affecting 1-2% of the population over the age of 65. Pathologically, it is characterized by selective dopaminergic neuron loss and the presence of Lewy bodies immunoreactive for ␣-synuclein in the substantia nigra pars compacta. To date, the leading causes for the sporadic form of the disease remain unclear, although there is accumulating evidence implicating oxidative stress (1), including the finding that PD brains have increased levels of oxidative damage to DNA, proteins, and lipids (2)(3)(4). One potential player contributing to increased oxidative stress is iron, which can convert hydrogen peroxide to highly reactive hydroxyl radicals via the Fenton reaction. Indeed, increased deposition of iron was found in microglia, astrocytes, oligodendrocytes, and dopaminergic neurons of the substantia nigra pars compacta of post-mortem PD brains (5,6). The total iron content was found to be significantly higher in the substantia nigra pars compacta of PD patients together with a corresponding increase in divalent metal transporter-1 (DMT1) transcripts in the same region (7). This suggests a close association among DMT1 expression, iron overload, and PD.
Mutations in a number of genes have also been implicated in the pathogenesis of PD (8) of which the first to be discovered was ␣-synuclein. Besides the A53T, A30P, and E46K missense mutations (9 -11), duplication (12,13) and triplication (14) of the ␣-synuclein gene have also been linked to familial forms of PD. It has been suggested that the tendency of ␣-synuclein to undergo misfolding and aggregation may underlie its involvement in Lewy body formation and hence PD (15). Given that iron increases the propensity of ␣-synuclein to aggregate (16,17), it is tempting to speculate on a synergistic effect between the two. In line with this, there is a growing consensus that the combinatorial effects of environmental and genetic factors may underlie neuronal vulnerability in PD (18).
We therefore used the DMT1 transporter (19) as a genetic tool to overload SH-SY5Y cells to test the role of iron and its interplay with PD-linked genes. We first showed that DMT1 overexpression resulted in increased cell death upon treatment with iron. Next, DMT1 expression coupled with chronic iron treatment and ␣-synuclein led to increased cell death especially in cells expressing the A53T mutant, and this correlated with increased DNA damage and oxidized proteins. Interestingly, cell death did not appear to be due to apoptosis because activation of caspases was not observed. Instead, there was an increase in autophagy in A53T cells as seen from the increased conversion of LC3-I to -II, greater monodansylcadaverine (MDC) staining for autophagolysosomes, and greater numbers of autophagosomes observed with electron microscopy. Finally, treatment with autophagy inhibitors ameliorated cell death resulting from chronic iron treatment. Taken together, our results suggest that autophagy is involved in cell death mediated by the synergistic effects of chronic iron treatment and ␣-synuclein.
LC3 Overexpression-The tf-LC3 plasmid, which has been described previously (20), was used to study the different stages of autophagy.
Pharmacological Agents-For some assays, the phosphatidylinositol 3-kinase inhibitors 3-methyladenine (3MA) (5 mM; Sigma) and wortmannin (WM) (200 nM; Sigma) were added to investigate their effect on cell viability. Together with chloroquine (40 M; Sigma), these agents are known to inhibit autophagy. Cathepsin inhibitors, such as E64D (10 g/ml; Sigma) and pepstatin A (10 g/ml; Sigma), were also used. As a positive control for autophagy, cells were starved in amino acidfree Earle's balanced salt solution (EBSS; Sigma) for up to 12 h.
Iron Uptake Assay-To assess the iron uptake capability of DMT1, 55 Fe uptake assays were performed as described previously (21). Briefly, cells were incubated with serum-free DMEM for 1 h to deplete cells of transferrin followed by a first wash with PBS and a second wash in washing buffer (25 mM Tris-HCl, pH 5.5, 25 mM MES, 140 mM NaCl, 5.4 mM KCl, 5 mM D-glucose, 1.8 mM CaCl 2 , 800 M MgSO 4 ). Subsequently, the cells were maintained in incubation buffer for 30 min at either 37 or 0°C (nonspecific uptake). The incubation buffer contained washing buffer with 1 M 55 FeCl 3 (PerkinElmer Life Sciences) that had been conjugated with nitriloacetic acid in a 1:20 molar ratio and 50 M ascorbic acid. After incubation, cells were washed with ice-cold washing buffer and then with PBS. Cells were trypsinized and collected by centrifugation before solubilizing them with the scintillation mixture (PerkinElmer Life Sciences). Iron uptake was determined using a Beckman LS6000 scintillation counter. An aliquot of the resuspended cells was kept for determination of total protein using Bradford reagent (Pierce) for quantification of 55 Fe incorporation (cpm/g of protein). Normalized cellular iron uptake was obtained by subtracting the nonspecific iron uptake at 0°C from the iron uptake at 37°C.
Iron Quenching Assay-The uptake of iron in the double stable cell lines was determined using a cell-permeable calcein-AM dye (Sigma) as described previously (22). Briefly, cells were seeded at a concentration of 80,000 cells/well into 48-well plates and grown to 80% confluence. The cells were treated with 100 M FeCl 2 in DMEM (pH 5.5) for 30 min at 37°C. Cells were washed with DMEM and incubated with 0.25 M calcein-AM for 30 min at 37°C. The cells were then washed thrice in chilled Hanks' balanced salt solution and lysed in radioimmune precipitation assay buffer containing 1% Triton X-100. 150 l of the lysates were transferred to a 96-well plate, and fluorescence intensity was measured at 488/518 nm. The fluorescence of the lysates was normalized to the protein concentration.
Iron Treatments and Trypan Blue Exclusion Analysis-For a chronic cellular model of PD, a longitudinal iron treatment paradigm that exposed the stable cell lines to elevated iron for up to 6 days was carried out. Cells were cultured with daily changes of DMEM containing 100 M FeCl 2 for a period lasting up to 6 days. At each time point, total cells (live and dead) were harvested, and a trypan blue exclusion assay was used to determine cell viability (23).
Preparation of Cell Lysates-Lysis was carried out by disrupting harvested cells in lysis buffer containing 1% Triton X-100, 10 g/ml aprotinin, and 1 mM PMSF in PBS. Ultracentrifugation was carried out at 125,000 ϫ g for 60 min at 4°C (Hitachi Koki, CS150GXL micro-ultracentrifuge). The supernatants and pellet were then collected, and the protein content was quantified by Bradford assay (Pierce).
Ubiquitin-Proteasome System Assay-Proteasomal activity was determined using the ubiquitin-proteasome system kit (Chemicon, Billerica, MA). About 15 g of harvested cell lysates from day 4 of Fe 2ϩ treatment were incubated with the succinyl-LLVY-AMC substrate for 2 h at 37°C. Subsequently, the amount of AMC released was measured using a fluorometer with a 380/460-nm filter set (Tecan, Männedorf, Switzerland).
Determination of Oxidative Stress-Cell lysates were prepared as described above after 4 days of Fe 2ϩ treatment. Cellular oxidative stress levels were determined using the OxyBlot Protein Oxidation Detection kit (Chemicon). Briefly, 50 g of protein were diluted with 6% SDS and incubated with 2,4-dinitrophenylhydrazine for 15 min. The samples were then neutralized before resolving on a 10% polyacrylamide gel and transferring to a nitrocellulose membrane. After blocking in 5% milk, membranes were incubated with a rabbit anti-2,4-dinitrophenylhydrazine antibody overnight at 4°C followed by incubation with HRP-conjugated secondary antibody for another hour. Finally, the immunocomplexes were visualized by chemiluminescence using the ECL method.
Comet Assay-The extent of single and double strand breaks after 4 days of iron treatment was measured using the comet assay procedure described by Low et al. (24). Harvested cells were trypsinized, resuspended in Hanks' balanced salt solution (Sigma), mixed with low melting point agarose (Trevigen, Gaithersburg, MD), and spread on Comet slides (Trevigen). Cell lysis was done at 4°C for 1 h using the buffer provided (Trevigen). To survey both single and double strand breaks, alkaline buffer electrophoresis was carried out for 20 min before neutralization with 0.5 M Tris-HCl, pH 7.5 for 15 min and dehydration in 70% ethanol. Slides were then dried at 37°C to bring all the cells to a single plane before staining with SYBR Green (Trevigen). 100 comets were randomly captured and analyzed for each slide. A minimum of triplicate sets for each treatment and cell line were carried out. Images were visualized using the 20ϫ objective of a Zeiss Axioplan 2 microscope (Zeiss, Oberkochen, Germany) and scored with Comet Imager 2.2.100 software (Metasystems GmbH, Altlussheim, Germany).
MDC Assay-Cells grown on coverslips were incubated with 0.05 mM MDC in PBS at 37°C for 10 min. The dye is a marker for autophagolysosomes (25). Following incubation, cells were washed four times with PBS and immediately analyzed by fluorescence microscopy using an inverted microscope (Olympus BX51, Tokyo, Japan) equipped with a filter system (excitation filter V-2A, 380 -420 nm; barrier filter, 450 nm). LSM Zeiss Imaging analysis software was used to measure the fluorescence intensity of MDC-stained cells.
Electron Microscopy-Cells were seeded at a density of 500,000 cells/ml and fixed in 2.5% glutaraldehyde in PBS for 2 h. They were subsequently postfixed in osmium tetroxide (OsO 4 ) in 0.2 M buffer for 1 h after which they were dehydrated under an increasing ethanol gradient. The samples were embedded in epoxy resin mixture and allowed to polymerize at 60°C for 48 h. The blocks were ultrasectioned using a diamond knife with the ultramicrotome set to cut at approximately 100 nm using heat advance. The sections were subsequently mounted onto thin bar copper grids. For quantification purposes, a standard stereological approach using the "selector" probe (26) was used to account for the number of autophagosomes and autophagolysosomes in each field.
Statistics-All statistics were carried out using SPSS (Version 11) or GraphPad Prism (Version 4) software. One-way analysis of variance was used to compare the values between two or more groups supported by multiple comparisons. This was followed by the Bonferroni's post hoc test. The level of statistical significance was set at p Ͻ 0.05 (*).

DMT1 Facilitated Fe 2ϩ Uptake and Enhanced Cytotoxicity in
Combination with A53T ␣-Synuclein-Using DMT1 as a means of overloading Fe 2ϩ into our cellular model, we first demonstrated via the 55 Fe uptake assay that S-DMT1 cells stably expressing DMT1 mediated a dramatic 300-fold increase in iron uptake compared with control SH-SY5Y cells (Fig. 1A). This demonstrated that DMT1 was indeed a viable system for loading iron into the cell. Additionally, treatment with iron resulted in greater cell death in the S-DMT1 stable cell lines compared with the S-V vector cell line after 24 h especially for 10 and 15 mM FeCl 2 (Fig. 1, B and C). Next, we examined how chronic iron overload could affect the cell in combination with a genetic component of PD, ␣-synuclein. With additional ␣-synuclein expression in combination with DMT1, the double stable SH-SY5Y cell line SD-A53T displayed synergistic toxicity when treated with iron over several days even when using a lower concentration of 100 M FeCl 2 . There was an approximately 2-fold decrease in SD-A53T cell viability with Fe 2ϩ treatment compared with SD-V especially from day 4 onward (Fig. 2A). In addition, it can be seen that the cell density was highest for SD-V and SD-WT and lowest for A53T (Fig. 2B). This was in contrast to the ␣-synuclein cell lines without stable DMT1 expression (SH-V, SH-WT, SH-A53T, and SH-A30P), which showed no significant difference in cell viability or density upon treatment with iron for the same period of time (supplemental Fig. S3, A and B). Additionally, the iron quenching assay using calcein-AM showed that upon treatment with iron there was no difference in iron uptake among the ␣-synucleinexpressing double stable cell lines when compared with the control SD-V cell line (supplemental Fig. S3C). This shows that WT or mutant ␣-synuclein does not affect iron uptake and suggests that cell death in the SD-A53T double stable cell line is due to the synergistic effect of iron and A53T rather than an A53T-mediated increase in iron uptake.
Enhanced Oxidative Stress-Oxidative stress can damage the cell in a number of ways, including the formation of protein carbonyls due to peptide bond cleavage or amino acid oxidation, formation of protein aggregates due to oxidant-induced cysteine cross-linking, lipid peroxidation, and oxidative DNA damage (27). We evaluated the extent to which iron in combination with WT or mutant ␣-synuclein could affect the level of oxidative stress in cells and found that there was indeed oxidative damage at both protein and DNA levels. After 4 days of treatment with 100 M FeCl 2 , the level of oxidized proteins (protein carbonyls) as shown by the oxyblot assay was about 4 times higher for SD-A53T compared with SD-WT and SD-A30P cell lines. In contrast, untreated cells showed very little oxidative damage (Fig. 3A). The increase in oxidized proteins for iron-treated SD-A53T cells suggested that iron treatment coupled with the A53T mutation could raise levels of reactive oxygen species. However, although oxyblot is a sensitive assay for protein carbonyls, it is unable to measure other forms of protein modification by reactive oxygen species, such as the formation of sulfoxide and disulfide bridges from cysteine and methionine and nitrotyrosine from tyrosine residues (27). In addition, it is unable to measure lipid peroxidation or DNA damage. Thus, to obtain a better idea of the level of oxidative stress, we examined oxidative stress-induced DNA damage using the comet assay. After 4 days of iron treatment, it can be seen that there was substantial DNA damage in SD-A53T cells, whereas the SD-V controls displayed minimal damage (Fig. 3B). Quantification of tail moments of the four cell lines (shown as population data) indicated significant differences between treated and untreated cells as well as between WT and mutant ␣-synuclein (Fig. 3B, right panel).
Cell Death Did Not Involve Ubiquitin-Proteasome System or Apoptosis-Cell death for SD-A53T and SD-A30P cell lines peaked at day 4 and day 5, respectively, after exposure to 100 M FeCl 2 . However, we did not detect any changes in ubiquitinproteasome system activity between iron-treated and untreated cells on the 4th day (supplemental Fig. S4). This suggested that the proteasome was still functional and capable of clearing damaged proteins. Thus, the cell death seen in SD-A53T and SD-A30P cells was unlikely to be due to proteasomal impairment.
To determine whether the cell death was due to apoptosis, we examined cytochrome c release into the cytosol. After 4 days of treatment with 100 M FeCl 2 , no cytochrome c was observed in the cytosolic fraction (Fig. 4A), suggesting that apoptosis did not play a role in iron-mediated SD-A53T cell death. We also studied the possible involvement of caspases and found that cleaved caspase 9 was present but was not up-regulated in the presence of iron (supplemental Fig. S5A). More importantly, iron-treated SD-A53T cells did not show any cleaved caspase 3, which is the final executioner for apoptosis. We also studied another pathway for apoptosis via apoptosis-inducing factor, which plays a role in caspase-independent apoptosis. Apopto- sis-inducing factor is released from the mitochondria and migrates into the nucleus, binding to and triggering the destruction of DNA and subsequent cell death. We examined the potential translocation of apoptosis-inducing factor into the nuclear fraction but did not find any evidence of it doing so in the various untreated and treated cell lines (supplemental Fig. S5B).
Iron-mediated Stress Is Associated with Enhanced Autophagic Activity-After 4 days of FeCl 2 treatment, results from immunoblotting with LC3B antibody suggested that the SD-A53T cells had increased autophagic activity (Fig. 4B).
LC3-I is a cytoplasmic protein that when converted to LC3-II becomes incorporated into the double membrane of autophagosomes. Thus, the conversion of LC3-I to -II can be used as a marker for autophagy. NH 4 Cl was used to inhibit the degradation of LC3-II by the lysosome so that the conversion of LC3-I to -II could be visualized. Thus, it can be seen that for SD-A53T with NH 4 Cl (Fig. 4B, lane 10) there is a basal level of autophagy. On treatment with iron and NH 4 Cl (lane 11), the level of LC3-II increased, suggesting an increase in autophagic activity. Treatment with iron but not with NH 4 Cl showed a drastic reduction of LC3-II, demonstrating that it can indeed be degraded by the lysosome and that the increase in lane 11 was not due to upregulation of LC3. We next evaluated the induction of autophagy by using an autofluorescent dye, MDC, that is a marker for autophagolysosomes (25). After 4 days of exposure to iron, there was an increase in MDC fluorescence intensity (Fig. 5, A and B). Quantification of fluorescence intensity of at least 100 cells showed that after 4 days of iron treatment there was a 2-fold increase in fluorescence intensity compared with untreated cells (Fig. 5B). This suggests an increase in autophagolysosomes, which is the final step of autophagy in which the autophagosomes fuse to lysosomes to deliver their cargo for degradation.
We also used electron microscopy (EM) to examine the ultrastructures of SD-A53T stable cells that were exposed to iron and observed double membranous autophagosomes). Notably, there were more autophagosomes in SD-A53T cells compared with SD-V cells (Fig. 6, A-D). To quantify these vesicles, we used a stereological method and found that iron treatment increased the number of autophagosomes and autophagolysosomes in SD-A53T cells compared with treated SD-V cells as well as the untreated control (Fig. 6, E  and F).
Recently, a study by Shimizu et al. (28) demonstrated that both autophagy and JNK activation are required for autophagic cell death. Because we already demonstrated an increase in autophagy in iron-treated SD-A53T cells, we next examined the activation of JNK. Using an antibody specific for phosphorylated JNK, we observed an increase in the lower 46-kDa phosphorylated band in SD-A53T cells treated for 4 days with iron (Fig. 6G). Although there was also an increase in the phosphorylated band for iron-treated SD-A30P cells, these cells did not show an increase in autophagy, which could account for the lack of cell death when this line was treated with iron.
To further demonstrate the effect of iron in up-regulating autophagy, we used the tandem fluorescence-tagged LC3 construct tf-LC3 (GFP-mRFP-LC3) (20) and transfected it into SD-A53T cells. Any localization of LC3 into the acidic autophagolysosome will result in a loss of green GFP-LC3 signal, whereas the red mRFP-LC3 fluorescence will remain intact. After 3 days of iron treatment, we observed the loss of GFP-LC3 fluorescence, whereas mRFP-LC3 fluorescence was visible as an increased number of red puncta compared with untreated cells (Fig. 7A, top and second panels). This result was very similar to induction of autophagy by amino acid starvation using EBSS, which resulted in a large number of red puncta (Fig. 7A,  fourth panel from top). Thus, it is likely that both EBSS and iron treatments led to autophagy induction with prolonged autophagic activation eventually resulting in cell death. We also added autophagy inhibitor 3MA to iron-treated cells and observed a reduction in red puncta (Fig. 7A, third panel from  top). This suggests that the inhibitor was able to reverse the up-regulation of autophagy caused by iron treatment. Finally, the addition of chloroquine neutralized the acidic pH of autophagolysosomes and prevented the quenching of GFP-LC3 fluorescence. This resulted in the colocalization of red and green puncta, showing that the puncta are a marker for autophagolysosomes. In addition, EBSS and iron treatments up-regulated cathepsin D (Fig. 7B and supplemental Fig. S6), a lysosomal enzyme, and this effect was reversed with autophagy (3MA and WM) and cathepsin inhibitors (pepstatin A and  E64D), further suggesting that iron treatment increased autophagy in SD-A53T cells.
Taken together, the above data suggest that chronic exposure of SD-A53T cells to iron increased autophagic activity. Conceivably, prolonged high levels of autophagy might promote cell death in iron-treated cells. Supporting this, we observed that cell death resulting from chronic 100 M FeCl 2 treatment could be reduced through the use of autophagy inhibitors 3MA and WM (Fig. 7C). Additionally, cell death due to acute treatment with 400 M FeCl 2 over 24 h could also be rescued by autophagy inhibition (supplemental Fig. S7). This strongly suggests that cell death in SD-A53T cells is likely to be autophagy-driven because pharmacological inhibition of autophagy was able to mitigate cell death.

DISCUSSION
Many cells in the substantia nigra pars compacta of PD postmortem brains have been shown to contain depositions of iron (29), suggesting that iron plays a role in the disease. In addition, it has also been shown that DMT1 transcripts in human PD brains are increased with a corresponding increase in iron de-position (7), suggesting that DMT1 might play an important role in iron accumulation. We thus used DMT1 as a tool to deliver iron into SH-SY5Y cells and demonstrated the synergistic interaction between chronic iron treatment and ␣-synuclein toxicity in mediating enhanced cell death. We observed that overexpression of DMT1 facilitated iron uptake and contributed substantially to cell death over 6 days in cells also expressing mutant A53T ␣-synuclein. Indeed, oxidative stress due to iron overload likely played a role in the toxicity of A53T ␣-synuclein and also to a lesser extent of A30P. Interestingly, the cell death did not appear to proceed via apoptosis because we did not observe cytochrome c release or caspase 3 activation. Instead, cell death appeared to be due to excessive induction of autophagy.
Although the role of autophagy in neurodegeneration is still being debated, there is evidence that it can contribute to both survival and cell death under different circumstances (30 -32). Studies of knock-out mouse models that are ablated of essential autophagic genes showed that the absence of autophagy in these animals can lead to neurodegeneration (33,34). Addition- FIGURE 7. Cell death in iron-treated SD-A53T is reversed by autophagy inhibition. A, SD-A53T cells were transfected with tf-LC3 (GFP-mRFP-LC3) for 24 h and treated with 100 M FeCl 2 for 3 days. In iron-treated cells, there was an almost total reduction of LC3-GFP puncta (because the GFP signal is quenched when it enters the lysosome), whereas the number of mRFP-LC3 puncta increased. This is similar to 2-h treatment with EBSS, which results in starvation-induced autophagy. With the addition of 5 mM 3MA (an autophagy inhibitor) for 16 h to iron-treated cells, there was a reduction in red puncta. The addition of 40 M chloroquine (CQ) for 1 h, which neutralizes the acidic pH of the lysosomes, prevents the quenching of the GFP signal and allows both red and green signals to be colocalized in puncta. Quantification of the average number of red puncta per cell was carried out and plotted onto a graph (*, p Ͻ 0.05). The data shown are representative of at least four independent experiments. Un, untreated. Error bars represent standard error of the mean (S.E.). B, representative immunoblot showing that both EBSS and iron treatment resulted in up-regulation of cathepsin D (25-kDa mature form). This was reversed with 5 mM 3MA or 200 nM WM (autophagy inhibitors) as well as with 10 g/ml pepstatin A or E64D (cathepsin inhibitors). C, a trypan blue assay of SD-A53T stable showed that autophagy inhibitors 3MA and WM were able to improve cell viability after 6 days of 100 M FeCl 2 treatment. Cells treated with both iron and wortmannin survived better compared with cells treated only with iron (*, p Ͻ 0.05). Error bars represent standard error of the mean (S.E.).
ally, overexpression of Beclin 1, another autophagy gene, was able to reduce neurodegeneration in a ␣-synuclein mouse model of PD (35). In contrast to the prosurvival role of autophagy in the above studies, overexpression of Alzheimer disease-associated peptide A␤  in Drosophila neurons results in accumulation of autophagic vesicles, neurodegeneration, and reduced lifespan. These detrimental effects could be partially rescued by inhibition of autophagy, demonstrating the prodeath role of autophagy in this case (36).
It has been suggested that autophagy under certain conditions can act as one of several cell death types known as type II programmed cell death or autophagic cell death (37,38). This mode of cell death is distinct from apoptotic cell death (type I programmed cell death) in that organelles are degraded early and cytoskeletal elements are preserved, whereas the exact opposite occurs in apoptosis. Experiments have shown that cell death via autophagy can occur, such as in the chemically induced cell death of Bax-and Bak-deficient mouse embryonic fibroblasts that does not occur when Beclin 1 and Atg5 are silenced (39). Additionally, autophagy inhibition can prevent this form of cell death, such as 3MA reduction of the death of human mammary carcinoma cells exposed to antiestrogen (40).
This correlates with our observations that iron overload and mutant ␣-synuclein could trigger autophagic cell death. Our data showed that there was increased conversion of LC3-I to LC3-II in SD-A53T cells following iron treatment, suggesting that autophagy is taking place. Although it is thought that upregulation of autophagy could help to clear ␣-synuclein aggregates and thus protect the cell, it is not the case here. This could be due to the combined effects of both iron and A53T ␣-synuclein. Iron has been found to accelerate the fibrillation of ␣-synuclein (16,41), especially the A53T mutant (17). Immunohistochemistry also showed an increased number of aggregate-positive cells for the iron-treated A53T mutant compared with wild type or A30P (42). With iron accelerating the aggregation of the A53T mutant, this could result in excessive autophagy induction in an attempt to clear the aggregates. Additionally, A53T ␣-synuclein can inhibit chaperone-mediated autophagy (43), leading to a further compensatory up-regulation of macroautophagy (44). Conceivably, under the twin insult of iron and A53T ␣-synuclein, the amount of autophagy could increase beyond the threshold of cell viability. Excessive up-regulation of autophagy could result in self-digestion and the degradation of important cellular components, resulting in cell death (32).
In addition, a recent study showed that autophagic cell death requires both autophagy and activated JNK (28). Examining the level of phosphorylated JNK, we found that iron treatment activated JNK in SD-A53T cells to a greater extent compared with SD-WT and SD-A30P. With a greater increase of both autophagy and activated JNK for A53T compared with WT and A30P, it is unsurprising that chronic iron treatment resulted in cell death for SD-A53T compared with the other cell lines. Correlating with this, expression of A53T ␣-synuclein in PC12 cells has been found to induce autophagic cell death (45). Finally, we found that cell death in SD-A53T cells upon chronic treatment with iron was attenuated when autophagy inhibitors WM and 3MA were added, suggesting that induction of autophagy contributed to cell loss.
In conclusion, we have demonstrated the use of DMT1 as a system for overloading iron into the cell. When used in conjunction with ␣-synuclein, this DMT1-mediated iron uptake leads to increased oxidative stress and DNA damage, resulting in cell death, particularly for cells expressing the A53T ␣-synuclein mutant. Interestingly, cell death did not proceed via apoptosis but was instead the result of excessive autophagy and JNK activation, and autophagy inhibitors were able to improve cell survival. Taken together, our results suggest a synergistic effect between environmental and genetic factors in PD pathogenesis and provide a mechanism underlying iron-mediated cell death.