PINK1-associated Parkinson's disease is caused by neuronal vulnerability to calcium-induced cell death

Abstract

Mutations in PINK1 cause autosomal recessive Parkinson's disease. PINK1 is a mitochondrial kinase of unknown function. We investigated calcium homeostasis and mitochondrial function in PINK1-deficient mammalian neurons. We demonstrate physiologically that PINK1 regulates calcium efflux from the mitochondria via the mitochondrial Na(+)/Ca(2+) exchanger. PINK1 deficiency causes mitochondrial accumulation of calcium, resulting in mitochondrial calcium overload. We show that calcium overload stimulates reactive oxygen species (ROS) production via NADPH oxidase. ROS production inhibits the glucose transporter, reducing substrate delivery and causing impaired respiration. We demonstrate that impaired respiration may be restored by provision of mitochondrial complex I and II substrates. Taken together, reduced mitochondrial calcium capacity and increased ROS lower the threshold of opening of the mitochondrial permeability transition pore (mPTP) such that physiological calcium stimuli become sufficient to induce mPTP opening in PINK1-deficient cells. Our findings propose a mechanism by which PINK1 dysfunction renders neurons vulnerable to cell death.

Figure 1

Characteristics of Mitochondrial Membrane Potential in PINK1 KD Neurons (A) Immunofluorescence of human dopaminergic neurons (Ai) (blue, Hoechst; red, β3 tubulin; green, TH) and primary mouse neurons (Aii) (blue, Hoechst; red, GFAP; green, MAP). Scale bar, 10 μM. (B) PINK1 KD neuroblastoma cells showed a 43% reduction (n > 120, p < 0.0001) in basal mitochondrial membrane potential (Δψm) compared to control cells (Bi). PINK1 KD human neurons exhibited 16% reduction (n > 60, p < 0.005) in basal Δψm compared to controls (Bii). (C) In WT mouse neurons (Ci), oligomycin did not affect Δψm; rotenone induced partial depolarization; FCCP induced complete depolarization. In PINK1 KO mouse neurons (Cii), oligomycin induced mitochondrial depolarization (78.6% ± 4.8% decrease in Δψm, n = 72). (D) Application of methyl succinate to PINK1 KO mouse neurons increased basal Δψm (19.5% ± 2.1%); application of pyruvate or malate to PINK1 KO neurons also increased the basal Δψm (18.6% ± 0.7%, n = 87). Substrate provision abolished oligomycin-induced mitochondrial depolarization in PINK1 KO. Error bars represent mean ± SEM.

Figure 2

Redox State and Rate of Oxygen Consumption in PINK1 KD/KO and Control Cells (A and B) Graphs demonstrate averaged trace of NADH and FAD²⁺ autofluorescence in neuroblastoma cells ([A], control; [B], PINK1 KD). The response to FCCP (1 μM) is larger in the control cells; the response to cyanide (1 mM) is larger in PINK1 KD cells. (C and D) Quantification of the percentage change in NADH or FAD²⁺ fluorescence: 0 is response to FCCP, and 100% is response to cyanide; this is reversed for FAD fluorescence. PINK1 KD/KO neurons have lower NADH fluorescence compared to control neurons. PINK1 KD neuroblastoma cells have lowered NADH fluorescence and increased FAD²⁺ fluorescence than controls. This can be reversed by re-expression of WT PINK1 in PINK1 KD cells. (E) Preincubation of PINK1 KO mouse neurons with 5 mM pyruvate restored redox level to control values. (F) Preincubation of WT mouse neurons with IAA to inhibit glycolysis reduced redox level to PINK1 KD values. (G) The basal rate of oxygen consumption is reduced in PINK1 KD neuroblastoma cells compared to control cells. PINK1 KD cells show no response to oligomycin. The maximal rate of respiration and oxygen consumption induced by FCCP was significantly lower in PINK1 KD cells than control cells. Pyruvate 5 mM and methyl succinate increased basal oxygen consumption, and increased FCCP induced maximal respiration. Error bars represent mean ± SEM.

Figure 3

Physiological Calcium Stimuli Induce Mitochondrial Depolarization in PINK1 KD Cells (A–C) Mouse (Ai and Aii) and human (Bi–Biii) neurons were loaded with fura-2 am and Rh123. KCl (50 mM) produced a rise in [Ca²⁺]_c in mouse wild-type neurons (Ai) and human control neurons (Bi). In PINK1 KO mouse neurons (Aii) and in PINK1 KD human neurons (Bii), the [Ca²⁺]_c was associated with an increase in Rh123 fluorescence and Δψm depolarization. Preincubation of human PINK1 KD neurons with 0.5 μM CsA prevented the Δψm depolarization of cells, but not the [Ca²⁺]_c signal (C). (D and E) Human neuroblastoma cells were loaded with fluo-4 (green) and TMRM (red). ATP (100 μM) induced a rise in [Ca²⁺]_c and thus an increase in fluo-4 fluorescence. In PINK1 KD cells the ATP-induced [Ca²⁺]_c signal was associated with a decrease in TMRM fluorescence and Δψm depolarization. Error bars represent mean ± SEM.

Figure 4

A Rise in [Ca²⁺]_c Induces [Ca²⁺]_m Overload and Δψm Depolarization in PINK1 KD/KO Cells (A) Arrows mark UV-induced flash photolysis of cells loaded with Ca-NPEGTA, fluo-4, and TMRM. In (A), mouse WT neurons demonstrated an increase in [Ca²⁺]_c in response to flash photolysis, with no change in Δψm. (B) In mouse PINK1 KO neurons, flash photolysis induced an increase in [Ca²⁺]_c with profound depolarization of the mitochondria. (C) In the same experiment performed on human PINK1 KD neurons, the photolysis-induced rise in [Ca²⁺]_c resulted in a dramatic increase in [Ca²⁺]_m, as demonstrated by the fluo-4 signal in the mitochondrial area (Figure 4Cg). This was rapidly followed by mitochondrial depolarization and subsequent release of fluo-4 from the mitochondrial area. (D) Application of 10 μm CGP-37157 to control neurons induced the same [Ca²⁺]_m overload and Δψm depolarization seen in PINK1 KD.

Figure 5

Mitochondrial Calcium Capacity Is Reduced in PINK1 KD/KO Neurons (A–D) Flash photolysis of permeabilized neurons loaded with Sodium Green and Rhod-5n demonstrated a flash-induced increase in [Ca²⁺]_m followed by Ca²⁺ efflux and Na⁺ influx in control cells (A). In PINK1 KD cells (B), there was no recovery of the [Ca²⁺]_m signal and reduced influx of Na⁺. Application of CGP-37157 (10 μM) to control neurons (C) inhibited the Na⁺/Ca²⁺ exchanger. Removal of Na⁺ from the medium in control neurons (D) also inhibited the Na⁺/Ca²⁺ exchanger. (E and F) Application of 50 mM NaCl in the presence of 7 μM ruthenium red stimulated an increase in Na⁺ and decrease in Ca²⁺ in control mitochondria (E). In contrast, there was minor activation of Na⁺/Ca²⁺ exchange in PINK1 KD mitochondria (F). (G–J) Increasing concentrations of Ca²⁺ were applied to permeabilized human neurons (G and H) or mouse neurons (I and J). Arrows indicate the final free calcium concentration to which mitochondria are exposed. Control human mitochondria (G) demonstrated a much higher Ca²⁺ capacity compared to PINK1 KD mitochondria (H). Control neurons are able to partially maintain the Δψm, until the collapse of the fluo-4 fluorescence. WT mouse neurons (I) also exhibited a higher Ca²⁺ capacity than PINK1 KO neurons (J). Note that human control neurons showed a significantly higher calcium capacity compared to mouse WT neurons.

Figure 6

Increase in Mitochondrial and Cytosolic ROS Production in PINK1 KD/KO Neurons (A and B) (A) WT cortical neurons (Ai) demonstrated an increase in HEt fluorescence, and thus cROS production, in response to 50 mM KCl, which was prevented by inhibition of NOX using DPI. PINK1 KO neurons (Aii) exhibited a higher basal level of ROS production than WT neurons. The basal ROS production and the response to KCl were blocked by inhibition of NOX in PINK1 KO neurons. (B) Histogram shows percentage values of rate of HEt fluorescence compared to 100% for WT neurons. (C and D) PINK1 KO neurons displayed a higher basal rate of increase in MitoSOX fluorescence, demonstrating a higher basal production of mROS compared to controls. Inhibition of complex 1 with rotenone induced a significant increase in ROS production in control neurons but only a small increase in ROS production in PINK1 KO neurons. (D) Histogram demonstrates percentage values of rate of MitoSOX fluorescence compared to 100% for WT neurons. (E) The rate of 2-NBDG fluorescence was lower in PINK1 KO neurons compared to WT neurons, reflecting lower glucose uptake. Incubation of PINK1 KO neurons with ROS scavenger MnTBAP increased the glucose uptake. Incubation of PINK1 KO neurons with the NOX inhibitor DPI restored glucose uptake to control levels. (F) The increase in the rate of HEt fluorescence in PINK1 KD cells was abolished by treatment of PINK1 KD cells with NOX-2 siRNA. (G) The reduced rate of 2-NBDG fluorescence in PINK1 KD cells was rescued by treatment of PINK1 KD cells with NOX-2 siRNA. Error bars represent mean ± SEM.

Figure 7

Schematic Diagram Illustrating the Effects of PINK1 within the Mitochondria and Cytosol