Parkinson's disease-associated human ATP13A2 (PARK9) deficiency causes zinc dyshomeostasis and mitochondrial dysfunction

Abstract

Human ATP13A2 (PARK9), a lysosomal type 5 P-type ATPase, has been associated with autosomal recessive early-onset Parkinson's disease (PD). ATP13A2 encodes a protein that is highly expressed in neurons and is predicted to function as a cation pump, although the substrate specificity remains unclear. Accumulation of zinc and mitochondrial dysfunction are established aetiological factors that contribute to PD; however, their underlying molecular mechanisms are largely unknown. Using patient-derived human olfactory neurosphere cultures, which harbour loss-of-function mutations in both alleles of ATP13A2, we identified a low intracellular free zinc ion concentration ([Zn(2+)]i), altered expression of zinc transporters and impaired sequestration of Zn(2+) into autophagy-lysosomal pathway-associated vesicles, indicating that zinc dyshomeostasis occurs in the setting of ATP13A2 deficiency. Pharmacological treatments that increased [Zn(2+)]i also induced the production of reactive oxygen species and aggravation of mitochondrial abnormalities that gave rise to mitochondrial depolarization, fragmentation and cell death due to ATP depletion. The toxic effect of Zn(2+) was blocked by ATP13A2 overexpression, Zn(2+) chelation, antioxidant treatment and promotion of mitochondrial fusion. Taken together, these results indicate that human ATP13A2 deficiency results in zinc dyshomeostasis and mitochondrial dysfunction. Our data provide insights into the molecular mechanisms of zinc dyshomeostasis in PD and its contribution to mitochondrial dysfunction with ATP13A2 as a molecular link between the two distinctive aetiological factors of PD.

Figure 1.

Zn²⁺-induced ROS-mediated cell death in ATP13A2^−/− hONs cells. hONs cells were tested for zinc sensitivity using the Neutral red uptake assay. (A) Increasing doses of ZnCl₂ (0, 100, 112.5 and 125 µm) significantly reduced the cell viability of ATP13A2^−/− cells (grey bar) in a dose-dependent manner, while the same treatment induced a significant change only at the highest dose in the control cells (white bar). (B) The cytotoxic effect of ZnCl₂ (112.5 µm) on ATP13A2^−/− cells was completely reversed by co-treatment with NAC (1 mm), an antioxidant agent. (C) H₂O₂ (1 mm) reduced cell viability in both hONs cell lines, but to a greater extent in ATP13A2^−/− cells compared with the control. The H₂O₂-mediated toxicity was significantly reduced by TPEN (1 µm), a Zn²⁺ chelator. (D) Western blot analysis detected the expression of V5-tagged ATP13A2 (V5ATP13A2) in the hONs cells transduced with V5ATP13A2 expressing lentivirus, but not in the cells transduced with lentivirus carrying an empty vector. β-Actin was used as a loading control. (E) ATP13A2 overexpression protected Zn²⁺-mediated cytotoxicity induced by 100 µm ZnCl₂ in ATP13A2^−/− cells compared with the empty vector control. Values in the graphs are represented as mean ± SD. NAC; N-acetyl-cysteine, TPEN, N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine. #P < 0.05 and ##P < 0.01 by Mann–Whitney U test and **P < 0.01 by Kruskal–Wallis one-way ANOVA followed by post hoc Tukey's HSD multiple comparison test.

Figure 2.

Reduced [Zn²⁺]_i in ATP13A2^−/− hONs cells. [Zn²⁺]_i was determined by quantification of the FluoZin-3 fluorescence in hONs cells. Cell images used for analysis were acquired at three to four randomly selected locations on a coverslip (a total of 3-4 coverslips per group in two independent experiments). There was no difference in the number of cells per coverslip between the cell lines (186.9 ± 46.3 for the control and 201.0 ± 38.4 for ATP13A2^−/− cells, P = 0.15 in a two-tailed Student's t-test). (A) Representative confocal images are presented for control (upper panel) and ATP13A2^−/− (bottom panel) cells treated as indicated. Scale bar = 100 µm. (B) Quantification of fluorescence from the cells revealed a significantly reduced FluoZin-3 signal for ATP13A2^−/− cells (grey bar) compared with the control (white bar) under vehicle treatment. H₂O₂ treatment increased FluoZin-3 signals to a similar extent in both the cell lines, while TPEN treatment reversed the H₂O₂-induced increase in FluoZin-3 signals to basal levels. Values in the graphs are represented as mean ± SD. TPEN, N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine. ##; P < 0.01 by Mann–Whitney U test.

Figure 3.

Altered expression of ZnTs/ZIPs in ATP13A2^−/− cells. cDNA from hONs cells was analysed to determine the mRNA expression level of ZnTs/ZIPs using a quantitative real-time RT–PCR. White bars in the graphs represent the control and grey bars represent ATP13A2^−/− cells. Among the genes investigated, 19 genes (8 ZnTs and 11 ZIPs) were expressed in the hONs cells, while the expression of ZnT2, ZIP5, ZIP8 and ZIP12 mRNA was not detected with the PCR conditions used (see Materials and Methods). There was no difference detected in the expression of β-actin (ACTB) encoding between the two cell lines (inset in the left of the top corner).The expression levels for six ZnTs (ZnT1, ZnT3, ZnT4, ZnT7, ZnT8 and ZnT9) and seven ZIPs (ZIP1, ZIP2, ZIP3, ZIP4, ZIP7, ZIP9 and ZIP10) were significantly upregulated in ATP13A2^−/− cells, while ZnT8 mRNA expression was significantly downregulated. ZnT, zinc transporters encoded by solute carrier family 30 genes (SLC30A1 to A9); ZIP, ZRT/IRT-related proteins encoded by solute carrier family 39 genes (SLC39A1 to A14). All reactions were repeated twice in triplicate. Values in the graphs are represented as mean ± SD. #P < 0.05 and ##P < 0.01 by Mann–Whitney U test.

Figure 4.

Reduction of Zn²⁺ levels in the ALP vesicles in ATP13A2^−/− cells. hONs cells expressing mRFP-LC3 were stained with FluoZin-3 after induction of accumulation of the mRFP-LC3-positive vesicles and release of Zn²⁺ from zinc-bound proteins (see Materials and Methods). (A) Representative confocal images are presented for the control (upper panels) and ATP13A2^−/− cells (bottom panels). Merged images (right panels) of mRFP-LC3 (red, left panel) and FluoZin-3 (green, middle panel) show yellow puncta, indicating co-localization of ALP vesicles with increased Zn²⁺. Scale bar = 20 µm. (B) Pearson's coefficient for co-localization between mRFP-LC3 and FluoZin-3 signals was significantly decreased in ATP13A2^−/− cells (grey bar) compared with the control (white bar) (n = 47, 11–13 cells per coverslip from four coverslips in two independent experiments). The area fraction occupied by mRFP-LC3-positive vesicles per cell (C), the number of FluoZin-3-positive vesicles per cell (D) and the FluoZin-3 intensity per vesicle (E) were not significantly different between the control and ATP13A2^−/− cells. Values in the graphs are represented as mean ± SEM. NS, not significant; #P < 0.05 by Mann–Whitney U test.

Figure 5.

Mitochondrial dysfunction and Zn²⁺-mediated ROS production in ATP13A2^−/− cells. Mitochondrial function and ROS production were assessed in hONs cells exposed to ZnCl_2. (A) ATP production rate was significantly lower in ATP13A2^−/− cells (grey bars) compared with the control (white bars) under basal conditions. Upon exposure to 100 µm ZnCl₂, ATP production rate was significantly reduced in the ATP13A2^−/− cells transduced with lentivirus carrying an empty vector, which was completely reversed by overexpression of V5-tagged wild-type ATP13A2 (V5ATP13A2). (B) TMRM labelling was significantly reduced in ATP13A2^−/− cells compared with the control. Treatment of the cells with the mitochondrial uncoupler CCCP decreased TMRM labelling to a similar extent in both the cell lines. (C) CM-H₂DCFDA was used to detect H₂O₂ in hONs cells. In the vehicle-treated groups, ATP13A2^−/− cells displayed a significantly lower CM-H₂DCFDA signal compared with the control. When treated with increasing concentrations of ZnCl₂ (0, 100, 500, 1000 µm) for 30 min, ATP13A2^−/− cells displayed a dose-dependent increase in CM-H₂DCFDA fluorescence signals with a significant increase at concentrations >500 µm. (D) Quantitative real-time RT–PCR detected a significant upregulation in the expression level of genes encoding the cellular antioxidant enzymes; superoxide dismutase 1 (SOD1), catalase (CAT) and glutathione peroxidase 1 (GPX1) and a significant down-regulation in superoxide dismutase 2 (SOD2), while β-actin mRNA (ACTB) was expressed at similar levels. All reactions were repeated twice in triplicate. Values in the graphs are represented as mean ± SD. CCCP, carbonyl cyanide 3-chlorophenylhydrazone. #P < 0.05 and ##P < 0.01 by Mann–Whitney U test and *P < 0.05 and **P < 0.01 by Kruskal–Wallis one-way ANOVA followed by post hoc Tukey's HSD multiple comparison test.

Figure 6.

Detrimental effect of elevated [Zn²⁺]_i on ΔΨ_m in ATP13A2^−/− hONs cells. hONs cells were treated with H₂O₂ to increase [Zn²⁺]_i and stained with 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1) to examine the effect of elevated Zn²⁺ levels on mitochondrial membrane potential (ΔΨ_m). (A) Representative confocal images are presented for the control (upper panels) and ATP13A2^−/− cells (bottom panels) treated as indicated. Red fluorescence indicates mitochondria with normal ΔΨ_m and green fluorescence mitochondria with low ΔΨ_m (e.g. damaged mitochondria). The increase of green fluorescence detected in the H₂O₂-treated groups is due to cytoplasmic diffusion of JC-1 monomers. Scale bar = 20 µm. (B) The area occupied by red fluorescing mitochondria in the vehicle-treated groups was significantly lower in ATP13A2^−/− cells (grey bar) compared with the control (white bar). H₂O₂ (0.95 mm) treatment significantly decreased the red signal in ATP13A2^−/− cells compared with the vehicle treatment, while co-treatment of H₂O₂ and TPEN (1 µm) reversed the toxic effect of H₂O₂ alone on ΔΨ_m (n = 24–28, 11–16 cells per coverslip were analysed in two independent experiments). Values in the graphs are represented as mean ± SEM. TPEN, N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine. ##P < 0.01 by Mann–Whitney U test and **P < 0.01 by Kruskal–Wallis one-way ANOVA followed by post hoc Tukey's HSD multiple comparison test.

Figure 7.

Zn²⁺-mediated mitochondrial fragmentation in ATP13A2^−/− hONs cells. hONs cells were treated with either ZnCl₂ alone or ZnCl₂ with IBMX and assessed for mitochondrial interconnectivity, ATP production rate and cell viability. (A) Cells were immunologically stained for Grp75 (green), a mitochondrial matrix protein and the nuclei were stained with 4',6-diamidino-2-phenylindole (blue). Mitochondrial form factor was calculated to determine the degree of mitochondrial interconnectivity (see Materials and Methods). Representative confocal images are presented for the control (upper panels) and ATP13A2^−/− cells (bottom panels) that were treated as indicated. Scale bar = 20 µm. (B) The mitochondrial form factor was found to be comparable between the control (white bar) and ATP13A2^−/− (grey bar) cells in the vehicle control groups, while CCCP treatment reduced the mitochondrial form factor significantly in both cell lines, indicating mitochondrial fragmentation (n = 65, 15–18 cells per coverslip from four coverslips in two independent experiments). Conversely, ZnCl₂ treatment decreased the mitochondrial form factor in ATP13A2^−/− cells, while there was only mild reduction detected in the control. Promotion of mitochondrial fusion using IBMX treatment, prevented ZnCl₂-mediated mitochondrial fragmentation in ATP13A2^−/− cells and further increased mitochondrial interconnectivity in the control. (C) ZnCl₂ (100 µm) treatment caused a significant reduction in ATP production rate in both the cell lines, although to a greater extent in ATP13A2^−/− cells (grey bars) compared with the control (white bars). IBMX co-treatment significantly blocked the Zn²⁺-mediated reduction in the ATP production rate in ATP13A2^−/− cells. (D) The viability of ATP13A2^−/− cells was significantly reduced upon exposure to ZnCl₂ (112.5 µm), while no difference was observed in the control. Further to this, co-treatment with IBMX (100 µm) blocked Zn²⁺-mediated cytotoxicity in ATP13A2^−/− cells. Values in the graphs are represented as mean ± SD. CCCP, carbonyl cyanide 3-chlorophenylhydrazone; IBMX, 3-isobutyl-1-methylxanthine. ##P < 0.01 by Mann–Whitney U test and *P < 0.05 and **P < 0.01 by Kruskal–Wallis one-way ANOVA followed by post hoc Tukey's HSD multiple comparison test.

Figure 8.

Schematic model of zinc dyshomeostasis and abnormal energy metabolism in ATP13A2 deficiency. Loss of ATP13A2 (green ellipse) results in a limited cellular buffering capacity for cytosolic Zn²⁺ due to the impairment in sequestration of Zn²⁺ (black circle) into LC3 (red circle) positive vesicles (single and double membraned organelles) associated with the ALP. The ensuing zinc dyshomeostasis results in mitochondrial dysfunction (lower ΔΨ_m and ATP levels). When the [Zn²⁺]_i is elevated, cytosolic Zn²⁺ levels also increase due to inefficient sequestration by LC3-positive vesicles in the setting of ATP13A2 deficiency and instead induce the accumulation of Zn²⁺ in mitochondria, which increases ROS production. An elevated level of ROS in turn causes mitochondrial damage, worsening mitochondrial dysfunction that subsequently leads to reduced energy production, fragmentation of the mitochondrial network and cellular degeneration due to ATP depletion.