Cytochrome c oxidase deficiency in neurons decreases both oxidative stress and amyloid formation in a mouse model of Alzheimer's disease

Abstract

Defects in the mitochondrial cytochrome c oxidase (COX) have been associated with Alzheimer's Disease, in which the age-dependent accumulation of beta-amyloid plays an important role in synaptic dysfunction and neurodegeneration. To test the possibility that age-dependent decline in the mitochondrial respiratory function, especially COX activity, may participate in the formation and accumulation of beta-amyloid, we generated mice expressing mutant amyloid precursor protein and mutant presenilin 1 in a neuron-specific COX-deficient background. A neuron-specific COX-deficient mouse was generated by the Cre-loxP system, in which the COX10 gene was deleted by a CamKIIalpha promoter-driven Cre-recombinase. COX10 is a farnesyltransferase involved in the biosynthesis of heme a, required for COX assembly and function. These KO mice showed an age-dependent COX deficiency in the cerebral cortex and hippocampus. Surprisingly, COX10 KO mice exhibited significantly fewer amyloid plaques in their brains compared with the COX-competent transgenic mice. This reduction in amyloid plaques in the KO mouse was accompanied by a reduction in Abeta42 level, beta-secretase activity, and oxidative damage. Likewise, production of reactive oxygen species from cells with partial COX activity was not elevated. Collectively, our results suggest that, contrary to previous models, a defect in neuronal COX does not increase oxidative damage nor predispose for the formation of amyloidgenic amyloid precursor protein fragments.

Fig. 1.

Characterization of the CNS-specific COX10 KO mice. (A and B) Homogenates from cortex at different ages were analyzed for the activity of COX (complex IV) and complexes II+III and expressed as a ratio to citrate synthase activity. (C) In-gel complex I (C1) activity of cortical samples of different aged mice was performed on a BN-PAGE gel. (D) Western blots of cortical samples at different ages. COX I, COX subunit; ISP, complex III subunit Rieske iron sulfur protein; TUJ1, a neuronal-specific cytoskeleton protein. C, control; K, COX10 KO.

Fig. 2.

Western blot analyses of the alterations in the levels of mitochondrial and neuronal proteins in 4-month-old AD and COXd/AD brain homogenates. (A) Cortical homogenates of 1-, 2-, 4-, and 6-month-old female mice were blotted against APP, COX I, SDH, and VDAC1. Note that COX I and APP levels start to decline at ≈2 months of age and between 4 and 6 months of age, respectively. Other mitochondrial proteins (SDH and VDAC1) did not show marked alteration with aging in COXd/AD cortices. (B) Cortical (left 5 lanes) and hippocampal (right 4 lanes) homogenates from 4-month-old KO, AD, and COXd/AD mice were examined for the levels of APP (derived from a transgene), COX I, VDAC1, SDH, SOD2, and TUJ1. The gender of each sample was indicated below the figure: M, male; F, female. (C) The level of each protein indicated was quantified in the 4-month-old female cerebral cortex from four independent Western blot signals per each genotype (n = 4 per group), and percentage changes in the level of a protein in the COXd/AD samples relative to that in AD samples was expressed in the graphs. Western blots for TUJ1 and Core1 are not shown in B.

Fig. 3.

Quantification of Aβ accumulation in the 4-month-old AD and COXd/AD brains. (A) The distributions of dense and diffuse plaques optically detected by scanning 75.2% of total area examined were mapped on eight consecutive sections from representative 4-month-old AD and COXd/AD female mice. (B) Total plaque number was stereologically quantified in the cortex and hippocampus of AD and COXd/AD, male (M) and female (F) brains (n = 3 per group). See Materials and Methods for details. (C) Total levels of human Aβ42 were quantified in the 4-month-old AD and COXd/AD female forebrains by ELISA (n = 6 per group). Aβ42 was not detected in control 4-month-old KO male forebrains that do not express the AD transgene (data not shown).

Fig. 4.

Quantification of oxidative damage. (A) Quantification of oxidized proteins. The total amount (arbitrary unit) of DNP-derivatized protein was quantified and normalized by the amount of β-tubulin, and the average of protein carbonyls (DNP signal)/β-tubulin ratios was compared between AD and COXd/AD groups (n = 4 per group). P = 0.006. (B) Determination of oxidative damage to nucleic acids. Brain sections from 4-month-old female AD (Upper) and 4-month-old female COXd/AD (Lower) were stained with anti-8OHG antibody. For a negative control, AD sections were pretreated with DNase/RNase (Upper Right). For a positive control, COXd/AD sections were pretreated with 30% H₂O₂ (Lower Right). (Scale bar, 25 μm.)