Beta-amyloid peptides induce mitochondrial dysfunction and oxidative stress in astrocytes and death of neurons through activation of NADPH oxidase

Abstract

Beta-amyloid (betaA) peptide is strongly implicated in the neurodegeneration underlying Alzheimer's disease, but the mechanisms of neurotoxicity remain controversial. This study establishes a central role for oxidative stress by the activation of NADPH oxidase in astrocytes as the cause of betaA-induced neuronal death. betaA causes a loss of mitochondrial potential in astrocytes but not in neurons. The mitochondrial response consists of Ca2+-dependent transient depolarizations superimposed on a slow collapse of potential. The slow response is both prevented by antioxidants and, remarkably, reversed by provision of glutamate and other mitochondrial substrates to complexes I and II. These findings suggest that the depolarization reflects oxidative damage to metabolic pathways upstream of mitochondrial respiration. Inhibition of NADPH oxidase by diphenylene iodonium or 4-hydroxy-3-methoxy-acetophenone blocks betaA-induced reactive oxygen species generation, prevents the mitochondrial depolarization, prevents betaA-induced glutathione depletion in both neurons and astrocytes, and protects neurons from cell death, placing the astrocyte NADPH oxidase as a primary target of betaA-induced neurodegeneration.

Figure 1.

β-Amyloid causes depolarization of mitochondria in astrocytes but not in neurons. Changes in Δψ_m were measured using Rh123 in dequench mode; the loss of potential is seen as an increase in fluorescence. A, Changes in Δψ_m from a neuron (black line) and an astrocyte (gray symbols) in a mixed culture from rat hippocampus (15 DIV) after exposure to 50 μm βA 25-35. In the example shown, βA 25-35 caused a slow progressive collapse of Δψ_m in the astrocyte, whereas no change at all was seen in a nearby neuron. Application of 300 μm glutamate caused a rapid collapse of Δψ_m in the neuron but promoted recovery of Δψ_m in the astrocyte. On subsequent removal of external Ca²⁺, neuronal Δψ_m recovered, showing that mitochondrial injury was still reversible. In this and all subsequent records, the protonophore FCCP was added at the end of the experiment to determine the extent of the Rh123 signal in response to complete mitochondrial depolarization. The Rh123 fluorescence signal in these traces is normalized between 0, representing the resting Rh123 fluorescence, and 100, representing the maximal increase in Rh123 fluorescence in response to complete mitochondrial depolarization by 1 μm FCCP. In traces in which the Rh123 signal was lost in some cells during exposure to βA, this normalization was not possible, and so the data are shown normalized only to a baseline set at 100%. Thus, the trace shown in A is actually very similar to some of the traces in B, in which FCCP responses were maintained. In B is shown an example of an experiment using astrocytes in culture in response to the full peptide βA 1-42 (5 μm). The cells responded with a slow loss of mitochondrial potential on which were superimposed abrupt depolarizing transitions. Some of these were reversible, but the larger depolarizations were followed by loss of the dye, suggesting cell death. Some examples of these traces are extracted from another data set to illustrate these three types of response in C. The series of images shown in D illustrate examples from an extended time sequence showing the transient increases in signal and the gradual increase in basal signal in response to βA 25-35. The image after FCCP saturates the display at this range because it is much brighter than the rest of the sequence and so is not shown. The time of each extracted image is indicated in minutes. In Eii and Eii are shown records from two astrocytes co-loaded with fura-2 and Rh123 to measure [Ca²⁺]_c (black triangles) and Δψ_m (gray lines and filled circles) simultaneously during exposure to 50 μm βA 25-35. As we have shown previously, βA caused fluctuations in [Ca²⁺]_c in the astrocytes. Although the abrupt mitochondrial depolarizations were clearly associated with [Ca²⁺]_c signals, they did not show a tight or fixed correlation in time or amplitude. Thus, large changes in Δψ_m followed large changes in [Ca²⁺]_c, and some small transient depolarizations of Δψ_m could be seen associated with [Ca²⁺]_c transients (asterisks in Eii).

Figure 2.

Transient mitochondrial depolarizations are dependent on βA-induced Ca²⁺ influx. Simultaneous measurements of [Ca²⁺]_c and Δψ_m were made from astrocytes in mixed hippocampal culture (14 DIV) co-loaded with fura-2 (Ai, Bi) and Rh123 (Aii, Bii). Traces are shown from two cells in each case. A, As shown previously (Abramov et al., 2003), the [Ca²⁺]_c signals were abolished in the absence of external Ca²⁺ (a saline with no added Ca²⁺ and with the addition of 500 μm EGTA). The transient mitochondrial depolarizations were also suppressed, but a significant slow depolarization was still evident as shown by the analysis of different components of the response shown in C. Note that the appearance of irreversible abrupt depolarizations was abolished and that of transient reversible depolarizations much reduced, whereas the slow increase in signal was unaffected. It may appear to increase simply because cells are not undergoing abrupt loss of signal. B, When βA was added in the absence of external Ca²⁺, as in A, no change in [Ca²⁺]_c and only a modest mitochondrial depolarization were seen. Washing with a Ca²⁺ containing saline then caused the abrupt appearance of [Ca²⁺]_c fluctuations and the appearance of large mitochondrial depolarizations. The data shown in D summarize the mean changes in the normalized Rh123 signal measured at 30 min exposure to 50 μm βA 25-35 in the presence (Control) or absence of calcium and in the presence of 1 mm Zn²⁺ and 2 μm clioquinol, which inhibit completely the [Ca²⁺]_c response to βA (Abramov et al., 2003). In all three cases, the mitochondrial depolarization was reduced significantly. ^*p < 0.01; ^**p < 0.001.

Figure 3.

The slow mitochondrial depolarization induced by βA is reversed by mitochondrial substrates. Changes in Rh123 signal were measured in a culture of cortical astrocytes in response to βA 25-35 (50 μm) in the presence of 1 mm glutamate (A), 10 mm methyl-succinate (B), and 1 mm glutamate in a Ca²⁺-free saline with the addition of 500 μm EGTA (C). In the presence of either glutamate or me-succinate, fast reversible transient mitochondrial depolarizations were still seen, but the slow progressive mitochondrial depolarization was abolished almost completely; compare with the day-matched control trace shown in A as a dashed line. In the presence of glutamate and the absence of external Ca²⁺, the entire response was abolished. In D are shown a series of images extracted from a time sequence showing changes in Δψ_m in response to βA in the presence of methyl succinate. Note that the Rh123 signal in some cells becomes transiently very bright but then is restored and shows a further increase with FCCP in the final image. The final image with FCCP shows the first image acquired immediately after application of FCCP, because the signal continued to get even brighter with a saturation of the display at this range. The time of each image in the sequence is indicated (minutes). An analysis of the effect of glutamate on different components of the response is shown in E, and the measurements of the mean increase in Rh123 fluorescence at 30 min in response to βA in the presence of glutamate, methyl succinate, and TMPD/ascorbate are shown in F. ^*p < 0.01; ^**p < 0.001.

Figure 4.

The mitochondrial response to βA is suppressed by antioxidants. A, Cortical astrocytes co-loaded with Rh123 and fura-2 were preincubated for 30 min with 500 μm TEMPO plus catalase (250 U/ml) before exposure to 50 μm βA 25-35. Only the Rh123 trace is shown; the fura-2 record was indistinguishable from controls. The antioxidants remained in the chamber during the exposure to βA. All aspects of the response were much reduced, as indicated in the analysis in B, and the mean mitochondrial depolarization measured after 30 min in the presence of βA was reduced significantly, as shown in C ^*p < 0.01; ^**p < 0.001.

Figure 5.

Role of the mitochondrial permeability transition pore in the mitochondrial depolarization induced by βA. A, Hippocampal astrocytes in coculture were loaded with Rh123 and preincubated with 0.5 μm cyclosporin A before exposure to 50 μm βA 25-35. Large irreversible mitochondrial depolarizations were suppressed completely, and the incidence of smaller transient and reversible depolarizations was also reduced, whereas the slow mitochondrial depolarization was not affected, as indicated by the analysis of the components of the response shown in D. The traces in B show that combined treatment of cells with CsA and the antioxidants TEMPO/catalase almost completely suppressed any response to βA, as did a combination of TEMPO and catalase combined with the removal of extracellular Ca²⁺, as illustrated in C. E summarizes measurements of the normalized Rh123 signal at 30 min of exposure to βA with each of these manipulations. These data strongly suggest that the mitochondrial response to βA involves the synergistic action of Ca²⁺ and oxidative stress leading together to the activation of the MPTP. ^*p < 0.01; ^**p < 0.001.

Figure 6.

The mitochondrial response to βA but not the change in [Ca²⁺]_c is mediated by the NADPH oxidase. Astrocytes were co-loaded with fura-2 and Rh123 for simultaneous measurement of [Ca²⁺]_c and Δψ_m. Cells were preincubated with 0.5 μm DPI for 30 min. [Ca²⁺]_c transients were seen as usual after application of 50 mm βA, whereas the mitochondrial response was suppressed completely.

Figure 7.

βA-increased ROS generation in astrocytes is dependent on Ca²⁺ and activation of the NADPH oxidase. Hippocampal cultures were loaded with DCF. A, Addition of 5 μm βA 1-42 caused a clear increase in the rate of appearance of the fluorescent product, suggesting increased ROS generation. The line shown was fitted by a linear regression with a correlation coefficient of 0.98. The rate of increase of ROS production was reduced significantly in the absence of external Ca²⁺ (B) and was blocked almost completely by 0.5 μm DPI (C). The data are summarized in D, which shows the mean rate of rise of DCF signal under these conditions and also after exposure to 1 mm apocynin, another inhibitor of the NADPH oxidase. ^*p < 0.01.

Figure 8.

GSH depletion and cell death induced by βA are dependent on free radical generation by NADPH oxidase. βA 25-35 or 1-42 caused GSH depletion in neurons (A) and astrocytes (B) measured in hippocampal cocultures loaded with monochlorobimane. This was prevented almost completely by 0.5 μm DPI and by 1 mm apocynin and was reversed by provision of 1 mm γ-glutamyl-cysteine. βA 25-35 or 1-42 also increased cell death of neurons (C) to ∼40% and astrocytes (D) to ∼15% above background. In both cell types, cell death was reduced significantly by all manipulations that also suppressed the mitochondrial depolarization in astrocytes: in the absence of external Ca²⁺, with the addition of antioxidants, and in the presence of 0.5 μm DPI or 1 mm apocynin. The most effective protection was afforded by the combination of antioxidants and CsA. The addition of 1 mm glutamate suppressed βA-induced cell death in astrocytes and made no significant difference to cell death in the neurons. Both neurons and astrocytes were also substantially protected from death induced by βA by previous incubation with 1 mm γ-GluCys (see Results). ^*p < 0.01; ^**p < 0.001.