Proteomics: a new approach to investigate oxidative stress in Alzheimer's disease brain

Abstract

In Alzheimer's disease (AD) brain oxidative stress is observed indexed by several markers, among which are protein carbonyls and 3-nitrotyrosine, markers for protein oxidation. We hypothesized that identity of these oxidatively modified proteins would lead to greater understanding of some of the potential molecular mechanisms involved in neurodegeneration in this dementing disorder. Proteomics is an emerging method for identification of proteins, and its application to neurodegenerative disorders, especially AD, is just beginning. Posttranslational modification of brain proteins, particularly that due of oxidation of proteins, provides an effective means of screening a subset of proteins within the brain proteome that likely reflects the extensive oxidative stress under which the AD brain exists, and this new methodology provides insights into mechanisms of neurodegeneration in and new therapeutic targets for AD. In this review, the use of proteomics to identify specifically oxidized proteins in AD brain is presented, from which new insights into mechanisms of neurodegeneration and synapse loss in this dementing disorder that is associated with oxidative stress have emerged.

Introduction

Alzheimer's disease (AD), a progressive, age-associated neurodegenerative disorder characterized by loss of memory and cognition [31], affects nearly 5 million persons in the United States currently. Estimates of 22 million persons worldwide with AD in the near future exist [30].

Oxidative stress, indexed by protein carbonyls (protein oxidation), 4-hydroxy-2-nonenal (HNE) and isoprostanes (lipid oxidation) and 8-hydroxy-2-deoxyguanine (DNA oxidation) is extensive in AD brain [9], [11], [12], [35], [39]. Amyloid β-peptide (1–42) [Aβ(1–42)] is thought to be central to the pathogenesis of AD [48].

We combined these two notions—the centrality of Aβ(1–42) to the pathogenesis of AD and the oxidative stress under which the AD brain exists—into a comprehensive, Aβ(1–42)-centered model for neurodegeneration in AD brain [9], [11], [12], [56]. In support of this model, we and others demonstrated that Aβ(1–42) induces in neurons protein oxidation, lipid peroxidation, and reactive oxygen species formation and that free radial scavengers could inhibit Aβ(1–42)-induced oxidative stress (reviewed in [Refs. [9], [11], [12], [56]). In vivo, oxidative stress was found in animals that express human Aβ(1–42) [22], [58] or express a human mutated gene that leads to familial AD [33], [52].

The mechanisms by which Aβ(1–42)-associated oxidative stress occur in neurons and putatively in AD brain are under active investigation. We showed that the single methionine residue of Aβ(1–42) at residue 35 is critical for the oxidative stress and neurotoxic properties of this peptide, both in vitro and in vivo [4], [8], [27], [28], [58]. Others invoke the involvement of peptide-bound redox metal ions in these properties [25]. Aggregated, not fibrillar, Aβ(1–42) is likely the toxic species of this peptide [2], [22], [34], [45], [57], consistent with the notion of small aggregates of the peptide inserting into the membrane bilayer to induce oxidative stress. Such membrane insertions are less likely with large fibrillar structures.

As noted, protein oxidation is evident in AD brain that correlates with markers of AD histopathology [23], i.e., protein oxidation occurs in AD brain where Aβ(1–42) is but not in Aβ(1–42)-poor cerebellum. But which proteins are oxidized? And could their identity provide insight into potential mechanisms of neurodegeneration in AD brain?

Our initial attempts at addressing these questions involved immunochemical methods. Brain slices or tissue were double immunoprecipitated, one antibody for protein carbonyls and a second for the protein of interest, and densitometric analysis of bands was compared between control and AD samples [3]. Creatine kinase and β-actin were shown to be selectively oxidized in AD brain by this approach [3]. However, this means of protein identification, one at a time, is impractical for the brain proteome. Moreover, one has to know beforehand the identity of the protein of interest in order to employ a specific antibody. A different approach to identification of large numbers of potentially oxidized proteins was needed. Hence, we used proteomics for the first time to identify specifically oxidized proteins in AD brain [4], [5], [6], [7], [13], [14], [15], [16], and the results are presented in this review. Insights into potential neurodegenerative mechanisms that stem from oxidative stress-induced protein oxidation and that are consistent with the biochemical and pathological alterations in AD brain have emerged.