Aging trumps everything as a risk factor for sporadic AD, but after decades of study, researchers are still unclear why. One theory blames accumulating oxidative damage, a metabolic consequence of getting older. Three recent articles lend new support to that theory. Two cell culture studies point to mechanisms by which oxidative stress wreaks havoc on components of the γ-secretase complex and hence amyloid processing. Another reports that reducing oxidative damage in mice from birth alleviates amyloidβ accumulation. Together, the studies reveal that age-related oxidative stress is not just bad for neurons, it can specifically exacerbate AD pathology.

Presenilin (PS), the catalytic component of γ-secretase, cleaves APP to yield Aβ40 or Aβ42 (the species more prone to aggregation). Some familial AD PS mutations cause an increase in the Aβ42/Aβ40 ratio, a change that is pathogenic. Could age-related changes affect γ-secretase in similar ways? Bart De Strooper, K.U. Leuven in Belgium, and colleagues set out to determine just that, and published their findings April 10 in EMBO Molecular Medicine. First author Francesc Guix grew rat hippocampal neurons in culture and watched them age for four weeks, during which time they accumulated reactive oxygen species and underwent additional aging processes similar to what happens in vivo over much longer time frames. Between three and four weeks of age, the cells increased Aβ production and also raised the Aβ42:40 ratio. The group wondered if peroxynitrite, an oxidant that accumulates during aging and irreversibly modifies protein tyrosine residues by nitrating them (a process called nitrotyrosination), might somehow alter γ-secretase to change Aβ processing. Guix found that between three and four weeks of age (the same at which Aβ processing went awry), protein nitrotyrosination tripled relative to two-week-old neurons. Nitration is widespread in AD brains (see Smith et al., 1997 and Castegna et al., 2003) and nitrotyrosination has been implicated in AD pathogenesis (see Tran et al., 2003).

To see if peroxynitrite might be to blame for the modified Aβ levels observed in these aged cells, the team treated younger, two-week-old neurons with a peroxynitrite generator, SIN-1. Sure enough, Aβ42 and the Aβ42:40 ratio rose dramatically, just as it did in aged neurons. The same was true for human embryonic kidney (HEK) cells treated with SIN-1. Treatment with hydrogen peroxide however, did not produce the same results, suggesting the effect was specific to nitrosative stress. To determine if nitrotyrosination specifically modified γ-secretase, the research group isolated microsomes containing the protease from SIN-1 treated and untreated HEK cells and combined them with APP. The Aβ42:40 ratio nearly doubled in microsomes from treated HEK cells relative to that of the untreated ones.

How was γ-secretase being changed? On adding higher amounts of SIN-1, the C-terminal fragments and N-terminal fragments of presenilin 1 appeared to associate more strongly, based on immunoprecipitation reactions. Interaction between these two fragments is known to raise the Aβ42:40 ratio in familial AD (FAD) cases (see Berezovska et al., 2005). "We think that the nitration of the presenilin is mimicking, to some extent, this aspect of FAD," said Guix. Fluorescence-lifetime imaging microscopy, which estimates how close two protein partners are, confirmed that nitrosative stress brought the two PS ends closer together in HEK cells. After treating the cells with SIN-1, the C-terminal fragment of PS1 also bound an anti-nitrotyrosine antibody. Also, in postmortem AD patients' brains, high levels of presenilin nitrotyrosination turned up compared to age-matched controls. All this evidence points to nitrotyrosination of PS1 inducing a conformational change that leads to changes in Aβ.

Finally, Guix and colleagues wanted to know what causes the increase in peroxynitrite with age. Peroxynitrite is formed when superoxide anion, a product of the mitochondrial electron transport chain, reacts with nitric oxide. Superoxide dismutase 2 (SOD2) sops up superoxide, but its activity had dropped threefold in four-week-old neurons compared to three-week-old ones. SOD2 knockout neurons generated a higher Aβ42:40 ratio compared to wild type, and mice that produced only half the normal amount of SOD2 protein had widespread nitrotyrosination in their brains. The C- and N-terminal fragments of presenilin 1 were closer together in these mice and Aβ42:40 ratios were higher. The results indicate that a drop in SOD2 activity unfetters peroxynitrite, shifting γ-secretase processing toward Aβ42 production.

"It's a nice study with a provocative set of data that may end up suggesting some therapeutic approaches down the road," said Michael Wolfe, Brigham and Women's Hospital, Boston, Massachusetts. "The fact that a neuron undergoing this nitrosative stress has a change in its γ-secretase enzymatic properties so that you get more Aβ42 to Aβ40 is very novel and potentially important." However, he cautioned that cellular models do not necessarily provide an exact replica of what happens inside the brain, and that more work is needed to determine whether more Aβ is produced and a higher Aβ42:40 ratio exists in people with sporadic AD.

Additional work is needed to figure out why SOD2 activity drops with aging, said Gunnar Gouras, Lund University, Sweden, though overall, "it's a rigorous, logical paper that provides a mechanistic link between oxidative stress, aging and elevation of Aβ42—which is key to AD—via the nitration of presenilin," he said.

Antioxidants, which would counteract oxidative stress, have been tested before as AD therapeutics and most have shown little to no effect on people with the disease (for an overview, see the AlzRisk). But the current study supports early administration of such treatments—at mid-life or even before, Guix said. "I think it's important to treat the patient earlier so that the intervention is done before the damage occurs."

Nicastrin, another component of the γ-secretase complex, may also be modified by oxidative stress, suggests a study led by Mark Mattson, National Institute on Aging, Baltimore, Maryland and Dong-Gyu Jo, Sungkyunkwan University, Suwon, Korea and published online March 10 in Aging Cell. Co-first authors A-Ryeong Gwon, Jong-Sung Park and Thiruma Arumugam found that nicastrin, which acts as an APP receptor, had a higher binding affinity for the substrate after modification by 4-hydroxynonenal (HNE), a product of membrane lipid peroxidation. Higher HNE-nicastrin levels correlated with more γ-secretase activity as well as greater Aβ plaques in cultured neurons and in the brains of people with AD. Could blocking this nicastrin modification prove beneficial? Gwon and colleagues found that a histdine analog called AG/01, which scavenges HNE, diminished γ-secretase activity in cultured rat neurons and reduced Aβ42 production in human neuroblastoma-derived (SH-SY5Y) cells overexpressing the Swedish APP mutant. Treating triple transgenic mice (3xTg-AD) every other day for a month with AG/01 suppressed γ-secretase activity, Aβ42, HNE-modifed nicastrin and lowered the Aβ42:40 ratio in the brain compared to untreated mice, suggesting HNE-targeted treatments could be possible AD therapies, the authors wrote.

"The combination of both studies gives very strong support to the idea that oxidative stress links aging with γ-secretase and provides some mechanisms by which aging increases the risk for Alzheimer's disease," said Guix. Further, the two papers "are the first to identify specific oxidative stress-induced molecular modifications of proteins involved in APP processing that result in increased neurotoxic Aβ42," Mattson told Alzforum in an email. He and his co-authors also pointed out that Aβ reportedly enhances oxidative stress on cells, meaning a vicious, self-perpetuating cycle could be in play whereby lipid peroxidation leads to Aβ production, which in turn leads to further lipid peroxidation.

"There are almost too many smoking guns to decide that a single one is dominant," said Douglas Galasko, University of California, San Diego. "We need to understand biochemical mechanisms and pathways that predispose towards sporadic Alzheimer's disease to replace ' aging'—which is a black box—with a series of specific events, to study how they can effect pathways that are relevant to Alzheimer's disease." Galasko recently completed an unsuccessful trial of antioxidants aimed at treating people with mild to moderate AD (see ARF related news story). Studies such as the one from De Strooper and colleagues do make antioxidants attractive AD therapies, he said, but before undertaking any more large preventative studies, researchers should identify antioxidants most likely to enter the brain and protect from relevant damage, he added.

One potential treatment, suggests a Human Molecular Genetics paper published April 5 by Hemachandra Reddy, Oregon Health and Science University, Beaverton, and colleagues, aims to enhance mitochondrial catalase (MCAT), one of the body's own antioxidants. This enzyme quenches hydrogen peroxide, which is readily converted into damaging radicals that cause lipid peroxidation, mitochondrial dysfunction and neuron damage. First author Peizhong Mao crossed mice that overexpress human MCAT (see ARF related news story) with those that overproduce Aβ and show cognitive deficits (Tg2576) and found that a lifelong boost in MCAT expression lessens Aβ pathology. Compared to control Tg2576 mice, the double mutants had reduced evidence of oxidative damage, less BACE1, and fewer Aβ monomers, oligomers and plaques. They also processed APP to a greater extent through the non-amyloidogenic alpha-secretase pathway. The double mutants also enjoyed a longer lifespan. Not only do these results implicate oxidative stress in AD pathology, but they also suggest a potential way to prevent the disease, by enhancing the cell's own mitochondrial anti-oxidants, wrote the authors.

"If we can somehow enhance brain mitochondrial catalase early on in life, we can possibly delay or prevent the disease process," said Reddy.

"It's a nice approach to blocking the oxidative stress at its main source very early on, upstream of altered APP processing and altered APP production," said Mattson. Even without drugs, it may be possible to give the cell's own antioxidants a boost with exercise and dietary restriction, which mildly stresses cells and enhances their ability to cope with more severe stress, he said.—Gwyneth Dickey Zakaib


  1. The interesting paper by Bart De Strooper and colleagues (1) points to nitrosative modification of the γ-secretase complex in aged neuronal cultures with consequent elevated Aβ(1-42) formation. The authors opine that aging, the major risk factor for Alzheimer's disease (AD), with consequent mitochondrial dysfunction (especially that of manganese superoxide dismutase), couples oxidative and nitrosative stress and AD via this nitrosative modification of the γ-secretase complex. The research is well done and the conclusions certainly seem to be supported by the data.

    However, a few comments appear to be in order regarding this interesting paper.

    1. Protein nitration is a formal oxidation, consistent with the well-known protein oxidation in brains of subjects with AD and amnestic mild cognitive impairment (MCI) (2,3).

    2. This work involves neuronal cultures. In AD brain, of course, many other cell types are involved and may significantly contribute to the nitrosative modification of γ-secretase in vivo. For example, microglia, when activated, secrete inducible nitric oxide synthase, which catalyzes formation of nitric oxide (NO), and, in the presence of superoxide radical (much of which emanates from mitochondria), peroxynitrite is formed by radical-radical recombination. In the presence of carbon dioxide, peroxynitrite undergoes a series of reactions that result in nitration of protein-resident tyrosine residues, especially in an oxidative environment such as the AD brain (2,3). Tyrosine nitration is highly detrimental to neurons, for example, blocking receptor tyrosine kinase phosphorylation and subsequent cellular signaling events, due to the steric hindrance associated with the NO2 functionality in the 3-position of tyrosine.

    3. Our laboratory showed that, in the brains of subjects with both AD and amnestic mild cognitive impairment (obtained with very short postmortem intervals), 3-nitrotyrosine was elevated (3,4), and we were the first to use redox proteomics to identify excessively nitrated brain proteins in both disorders (5-7). Accordingly, the interesting results shown in the De Strooper paper are a nice follow-up of already well-established findings in both disorders. Interestingly, that amnestic MCI brain demonstrates elevated 3-nitrotyrosine is consistent with the notion that this oxidative modification occurs early in the progression of AD (2-7).

    4. Mitochondria are known to be altered in AD and MCI (8), and MnSOD is an oxidatively modified protein in this dementing disorder (9) and nitrated in the APP/PS-1 human double mutant knock-in mouse model of AD (10).

    In summary, De Strooper and colleagues make a nice contribution to the literature, enhancing already well-established concepts.


    . Modification of γ-secretase by nitrosative stress links neuronal ageing to sporadic Alzheimer's disease. EMBO Mol Med. 2012 Jul;4(7):660-73. PubMed.

    . Evidence of oxidative damage in Alzheimer's disease brain: central role for amyloid beta-peptide. Trends Mol Med. 2001 Dec;7(12):548-54. PubMed.

    . Roles of amyloid beta-peptide-associated oxidative stress and brain protein modifications in the pathogenesis of Alzheimer's disease and mild cognitive impairment. Free Radic Biol Med. 2007 Sep 1;43(5):658-77. PubMed.

    . Elevated levels of 3-nitrotyrosine in brain from subjects with amnestic mild cognitive impairment: implications for the role of nitration in the progression of Alzheimer's disease. Brain Res. 2007 May 7;1148:243-8. PubMed.

    . Proteomic identification of nitrated proteins in Alzheimer's disease brain. J Neurochem. 2003 Jun;85(6):1394-401. PubMed.

    . Identification of nitrated proteins in Alzheimer's disease brain using a redox proteomics approach. Neurobiol Dis. 2006 Apr;22(1):76-87. PubMed.

    . Proteomic identification of nitrated brain proteins in amnestic mild cognitive impairment: a regional study. J Cell Mol Med. 2007 Jul-Aug;11(4):839-51. PubMed.

    . Oxidatively modified, mitochondria-relevant brain proteins in subjects with Alzheimer disease and mild cognitive impairment. J Bioenerg Biomembr. 2009 Oct;41(5):441-6. PubMed.

    . Redox Proteomics in Selected Neurodegenerative Disorders: From Its Infancy to Future Applications. Antioxid Redox Signal. 2012 Jan 18; PubMed.

    . Beta-amyloid mediated nitration of manganese superoxide dismutase: implication for oxidative stress in a APPNLH/NLH X PS-1P264L/P264L double knock-in mouse model of Alzheimer's disease. Am J Pathol. 2006 May;168(5):1608-18. PubMed.

    View all comments by Allan Butterfield
  2. There appear to be several links of nitrosative stress (related to nitric oxide) and AD, and one exciting pathway is described here in this new and elegant work from Bart De Strooper's laboratory concerning nitration of γ-secretase. In this case, peroxynitrite (formed by reaction of nitric oxide [NO] and superoxide anion) nitrates a tyrosine residue on PS1.

    Other types of nitrosative stress in AD lead to S-nitrosylation (reaction of NO with a critical cysteine thiol on a protein), as our group reported for Drp1 in Science in 2009 (see Cho et al., 2009). Additional such reactions are emerging in AD and related diseases, and will be published soon.

    Please note that the chemistry of the events occurring here is important, but often confused in the literature: The new findings of De Strooper and colleagues report nitration of a tyrosine residue on PS1. In contrast, S-nitrosation (or what Jonathan Stamler and I have termed S-nitrosylation because of its regulation of protein function being akin to phosphorylation) involves reaction of NO with a cysteine residue (probably involving the thiolate anion of the cysteine residue).


    . S-nitrosylation of Drp1 mediates beta-amyloid-related mitochondrial fission and neuronal injury. Science. 2009 Apr 3;324(5923):102-5. PubMed.

    View all comments by Stuart A. Lipton
  3. This is a very interesting paper, carefully studying levels of secreted Aβ peptides with time in culture in wild-type neurons and noting a marked elevation of Aβ42 secretion between 21 and 28 days in vitro (DIV). The authors go on to define a mechanistic reason for such an increase in Aβ secretion and elevated 42/40 ratio by showing higher levels of the γ-secretase complex and nitrosylation of presenilin fragments concomitant with elevated nitrosative stress between 21 and 28 DIV. They demonstrate experimentally that PS1 nitrotyrosination elevates the 42/40 ratio, and show that levels of nitrosylated PS1-N-terminal fragments are elevated in AD brain. Here, it would have been interesting to also compare young versus old healthy control brains. Remarkably, using a reporter of PS1 conformation (GFP-PS1-RFP construct), they show that nitrosylated PS1 attains a conformation that resembles the conformational change seen with PS1 FAD mutations. They also link falling superoxide dismutase 2 (SOD2) levels to the increases in nitrosylation in aging cultured neurons and provide additional supportive data using SOD2 knockout neurons and SOD2+/- mice.

    This excellent and comprehensive paper brings together leading experts in AD and γ-secretase with colleagues that have pioneered cultured neurons as a model for aging. How aging acts as the most important risk factor for the development of AD remains a critical, unanswered question. Studies by numerous groups have supported the role of oxidative stress (OS) in the age-related risk of AD, and a few studies, including one of ours, have shown that OS modulates Aβ. Cell culture models have major advantages when trying to elucidate underlying biological mechanisms, and are much needed as complementary models to work on living animals and humans. When we began using neurons with time in culture as a model for AD (Takahashi et al., 2004) we were worried about using embryonic neurons to model an age-related disease. Over the years, we have been amazed by how much selective AD-like synapse and β amyloid alterations are reflected in "aging" AD transgenic neurons in culture. One can keep in mind that AD transgenic mice can develop behavioral dysfunction and Aβ pathology in a matter of months, which is also a very different time frame from the decades for the human disease.

    In summary, elucidating how aging can induce AD is of major importance, and that oxidative stress and aging can directly impact γ-secretase and the 42/40 ratio is therefore highly interesting. Not atypical for the field, intraneuronal Aβ receives no mention. The question is not whether, but when, this will change. The relationship between the extra- and intracellular pools of Aβ is turning out to be quite complex. I hope the authors consider comparing their results in aging wild-type neurons to FAD mutant neurons. Surprisingly, aging AD-prone neurons eventually cannot cope with or fail to secrete their elevated Aβ (Tampellini et al., 2011).


    . Oligomerization of Alzheimer's beta-amyloid within processes and synapses of cultured neurons and brain. J Neurosci. 2004 Apr 7;24(14):3592-9. PubMed.

    . Impaired β-amyloid secretion in Alzheimer's disease pathogenesis. J Neurosci. 2011 Oct 26;31(43):15384-90. PubMed.

    View all comments by Gunnar Gouras
  4. This paper places relationships between oxidative stress, BACE, and amyloidosis within a very interesting context. In doing so, it provides insight into AD etiology.

    View all comments by Russell Swerdlow
  5. Mitochondrial dysfunction is an early pathological event in the Alzheimer’s disease (AD) brain. Defects in brain energy metabolism and key respiratory enzyme activity, increased mitochondrial oxidative stress, and alterations in mitochondrial structure, including in the mitochondrial permeability transition pore, occur in the AD-affected regions. However, the mechanisms underlying mitochondrial damage and its association with AD pathology are poorly understood. The questions of whether and how targeting mitochondria serve as a therapeutic strategy for AD are worth addressing.

    This recent study led by Hemachandra Reddy at Oregon Health and Science University supports the concept that mitochondrial oxidative stress plays a primary role in amyloid pathology and cognitive decline in AD. The researchers have extensively analyzed the protective effects of the mitochondria-targeted antioxidant catalase (MCAT) and lifespan extension in mice expressing Aβ (Tg2576 mice) from birth to death. They provide substantial evidence that increased MCAT blunts not only oxidative stress and oxidative DNA damage, but also brain Aβ accumulation. Interestingly, MCAT also interferes with APP processing and Aβ production/accumulation by a reduction in the levels of full-length APP, CTF99, BACE1, and Aβ levels, and increased levels of soluble APPα, CTF83, and the Aβ-degrading enzymes neprilysin and insulin-degrading enzyme. Consistent with published studies (Schriner et al., 2005), lifespan extends four to five months in MCAT mice and double MCAT/APP mice compared to the non-Tg and single APP mice, respectively.

    These data significantly enhance our understanding of the contribution of mitochondrial reactive oxygen species to aging and AD-etiopathology. Mitochondrial oxidative stress could be an upstream modulator of amyloid pathology and APP processing, leading to impaired learning and memory in AD sufferers. Thus, mitochondria-targeted molecules, such as oxidative stressors or antioxidants, may be effective approaches for halting and preventing AD progression. Increasing antioxidants, such as catalase, at the early stage of AD might be one of the therapeutic approaches for prevention and treatment of the disease.

    Previously, we demonstrated that blockade of the cyclophilin D-dependent mitochondrial permeability transition pore significantly improves mitochondrial and cognitive function through increases in mitochondrial calcium buffer capacity, respiratory function, and ATP levels, and attenuation of mitochondrial oxidative stress in Alzheimer’s disease neurons and transgenic AD mice (Du et al., 2008; Du et al., 2011). Furthermore, in AD mice overexpressing amyloid precursor protein and Aβ (J-20 line), blocking the interaction of mitochondrial Aβ with amyloid binding alcohol dehydrogenase (ABAD) suppresses the production/accumulation of reactive free radicals (ROS) in mitochondria, and, as a result, reverses abnormal mitochondrial function and improves learning and memory (Lustbader et al., 2004; Yao et al., 2011).

    Taken together, these studies indicate mitochondria as a potential therapeutic target for AD, especially at the early stage of the disease before profound neuronal injury.


    . Extension of murine life span by overexpression of catalase targeted to mitochondria. Science. 2005 Jun 24;308(5730):1909-11. PubMed.

    . Cyclophilin D deficiency attenuates mitochondrial and neuronal perturbation and ameliorates learning and memory in Alzheimer's disease. Nat Med. 2008 Oct;14(10):1097-105. PubMed.

    . Cyclophilin D deficiency improves mitochondrial function and learning/memory in aging Alzheimer disease mouse model. Neurobiol Aging. 2011 Mar;32(3):398-406. PubMed.

    . ABAD directly links Abeta to mitochondrial toxicity in Alzheimer's disease. Science. 2004 Apr 16;304(5669):448-52. PubMed.

    . Inhibition of amyloid-beta (Abeta) peptide-binding alcohol dehydrogenase-Abeta interaction reduces Abeta accumulation and improves mitochondrial function in a mouse model of Alzheimer's disease. J Neurosci. 2011 Feb 9;31(6):2313-20. PubMed.

    View all comments by Shirley ShiDu Yan

Make a Comment

To make a comment you must login or register.


News Citations

  1. Antioxidants No Help for Alzheimer’s, Biomarker Trial Says
  2. Coaxing Longevity from Catalase

Paper Citations

  1. . Widespread peroxynitrite-mediated damage in Alzheimer's disease. J Neurosci. 1997 Apr 15;17(8):2653-7. PubMed.
  2. . Proteomic identification of nitrated proteins in Alzheimer's disease brain. J Neurochem. 2003 Jun;85(6):1394-401. PubMed.
  3. . Tyrosine nitration of a synaptic protein synaptophysin contributes to amyloid beta-peptide-induced cholinergic dysfunction. Mol Psychiatry. 2003 Apr;8(4):407-12. PubMed.
  4. . Familial Alzheimer's disease presenilin 1 mutations cause alterations in the conformation of presenilin and interactions with amyloid precursor protein. J Neurosci. 2005 Mar 16;25(11):3009-17. PubMed.

Other Citations

  1. 3xTg-AD

External Citations

  1. AlzRisk

Further Reading


  1. . Early memory deficits precede plaque deposition in APPswe/PS1dE9 mice: involvement of oxidative stress and cholinergic dysfunction. Free Radic Biol Med. 2012 Apr 15;52(8):1443-52. PubMed.
  2. . RNA modifications by oxidation: a novel disease mechanism?. Free Radic Biol Med. 2012 Apr 15;52(8):1353-61. PubMed.
  3. . Antioxidant and bioenergetic coupling between neurons and astrocytes. Biochem J. 2012 Apr 1;443(1):3-11. PubMed.

Primary Papers

  1. . Oxidative lipid modification of nicastrin enhances amyloidogenic γ-secretase activity in Alzheimer's disease. Aging Cell. 2012 Mar 10; PubMed.
  2. . Modification of γ-secretase by nitrosative stress links neuronal ageing to sporadic Alzheimer's disease. EMBO Mol Med. 2012 Jul;4(7):660-73. PubMed.
  3. . Mitochondria-targeted catalase reduces abnormal APP processing, amyloid β production and BACE1 in a mouse model of Alzheimer's disease: implications for neuroprotection and lifespan extension. Hum Mol Genet. 2012 Jul 1;21(13):2973-90. PubMed.