T.R., R.M.P., D.C.C. and D.A.L. contributed equally to this work.
Fenton chemistry and oxidative stress mediate the toxicity of the β-amyloid peptide in a Drosophila model of Alzheimer’s disease
Article first published online: 23 MAR 2009
© The Authors (2009). Journal Compilation © Federation of European Neuroscience Societies and Blackwell Publishing Ltd
European Journal of Neuroscience
Volume 29, Issue 7, pages 1335–1347, April 2009
How to Cite
Rival, T., Page, R. M., Chandraratna, D. S., Sendall, T. J., Ryder, E., Liu, B., Lewis, H., Rosahl, T., Hider, R., Camargo, L. M., Shearman, M. S., Crowther, D. C. and Lomas, D. A. (2009), Fenton chemistry and oxidative stress mediate the toxicity of the β-amyloid peptide in a Drosophila model of Alzheimer’s disease. European Journal of Neuroscience, 29: 1335–1347. doi: 10.1111/j.1460-9568.2009.06701.x
Re-use of this article is permitted in accordance with the Terms and Conditions set out at http://www3.interscience.wiley.com/authorresources/onlineopen.html
- Issue published online: 30 MAR 2009
- Article first published online: 23 MAR 2009
- Received 6 October 2008, revised 2 February 2009, accepted 9 February 2009
- chelation therapy;
- hydroxy radical;
- Top of page
- Materials and methods
- Supporting Information
The mechanism by which aggregates of the β-amyloid peptide (Aβ) mediate their toxicity is uncertain. We show here that the expression of the 42-amino-acid isoform of Aβ (Aβ1–42) changes the expression of genes involved in oxidative stress in a Drosophila model of Alzheimer’s disease. A subsequent genetic screen confirmed the importance of oxidative stress and a molecular dissection of the steps in the cellular metabolism of reactive oxygen species revealed that the iron-binding protein ferritin and the H2O2 scavenger catalase are the most potent suppressors of the toxicity of wild-type and Arctic (E22G) Aβ1–42. Likewise, treatment with the iron-binding compound clioquinol increased the lifespan of flies expressing Arctic Aβ1–42. The effect of iron appears to be mediated by oxidative stress as ferritin heavy chain co-expression reduced carbonyl levels in Aβ1–42 flies by 65% and restored the survival and locomotion function to normal. This was achieved despite the presence of elevated levels of the Aβ1–42. Taken together, our data show that oxidative stress, probably mediated by the hydroxyl radical and generated by the Fenton reaction, is essential for Aβ1–42 toxicity in vivo and provide strong support for Alzheimer’s disease therapies based on metal chelation.
- Top of page
- Materials and methods
- Supporting Information
There is a growing consensus that smaller, soluble aggregates of the β-amyloid peptide (Aβ), rather than mature amyloid plaques, are the pathogenic species in Alzheimer’s disease (AD) (Lambert et al., 1998; Walsh et al., 2002). However, the mechanism by which these aggregates mediate their toxicity remains unclear. As Aβ is generated in the extracellular space, or more likely within the lumen of endocytic vesicles (Koo & Squazzo, 1994; Refolo et al., 1995), its toxic effects may be mediated by membrane damage or by interactions with membrane-bound proteins. There is evidence that Aβ aggregates can degrade the electrical resistance of membranes (Kayed et al., 2004; Demuro et al., 2005), possibly by forming pores (Lashuel et al., 2002), or alternatively they may interact with membrane receptors or even gain access to the cytoplasm. Membranes can also be damaged by the reactive oxygen species that are generated by Aβ aggregates in the presence of metals such as copper, zinc or iron (Bush, 2003). Subsequent pathological processes include mitochondrial damage (Abramov et al., 2004), tau phosphorylation with consequent axonal transport dysfunction and the initiation of cell death (Kienlen-Campard et al., 2002; Wei et al., 2002; Jo et al., 2004). However, until recently (Cao et al., 2008) it has been impossible to take a global view to ask which biological processes are essential for the development of the disease and which are downstream consequences of neurotoxicity. Knowing which biological processes are directly involved in initiating AD will allow us to focus on those upstream targets that have the greatest therapeutic potential.
We have developed a model of AD that is based on the expression of the human Aβ in fly neurons by coupling it to an N-terminal secretion signal peptide (Crowther et al., 2005). The Aβ1–42 but not the Aβ1–40 control accumulates in the brain and results in decreased lifespan and impaired locomotor performance. These phenotypes are more marked in flies expressing the E22G (Arctic) mutant of the Aβ1–42, which causes increased aggregation of Aβ and is responsible for early onset familial AD (Nilsberth et al., 2001). Here, we use our Drosophila model of AD to identify the pathways and intermediates that are critical for Aβ-mediated toxicity in vivo.
Materials and methods
- Top of page
- Materials and methods
- Supporting Information
The following stocks were generous gifts: UAS-CAT [catalase (CAT) upstream activating sequence (UAS) inducible transgene] (Anderson et al., 2005), UAS-SOD1 [CuZn-superoxide dismutase (SOD)1] (Anderson et al., 2005), UAS-SOD1-IR (RNAi line for SOD1) (Missirlis et al., 2003) and UAS-mitSOD2 (Mn-SOD2) (Anderson et al., 2005) from Professor John Phillips (Guelph, Canada), UAS-Sniffer (carbonyl reductase) (Botella et al., 2004) from Professor Stephan Schneuwly (Regensburg, Germany) and UAS-GST [glutathione-S-transferase (GST) S1] (Whitworth et al., 2005) from Dr Alex Whitworth (Sheffield, UK). We used previously characterized UAS-CAT, UAS-SOD1 and UAS-mitSOD2 stocks (Anderson et al., 2005) as they increase enzymatic activity by up to 200%. Stocks of mutant SOD1n108 (Phillips et al., 1995) and elavc155-GAL4 were from the Bloomington stock centre (Indiana, USA). Flies carrying Aβ transgenes [Aβ1–40 (Alz40.1), Aβ1–42 (Alz42.2 and Alz42.3) or Arctic Aβ1–42 (AlzArc1)] have been described previously (Crowther et al., 2005; Luheshi et al., 2007). The transgenes are each representative of six independent transgenic lines and each transgene drives similar levels of mRNA whether alone or in combination (such as Alz42.2 + Alz42.3).
The cDNA for Drosophila ferritin 1 heavy chain (Fer1HC) was isolated from EST clone GH24060 (Berkley Drosophila Genome Project). This cDNA lacked the iron response element resulting in the expression of a constitutively active form of Fer1HC. The cDNA for Drosophila ferritin 2 light chain (Fer2LC) naturally lacks an iron response element and was isolated from the EST clone AT16780 (Berkley Drosophila Genome Project). Both cDNAs were inserted downstream of GAL4 UASs (UAS-Fer1HC and UAS-Fer2LC) in the pUAST plasmid by directional cloning following EcoRI and XhoI digestion.
Affymetrix cDNA microarray
Individual samples were each hybridized to a Drosophila Genome GeneChip® Array following standard Affymetrix protocols. Affymetrix Microarray Suite 5 was used to generate signal values and detection calls. Probe level intensity data were adjusted for background, normalized and log transformed using the robust multichip average pre-processing method (Irizarry et al., 2003) using Rosetta Resolver® 7.1.
Ratio data for each individual probe were created as follows. Aβ1–40 baseline samples were created for each time-point by pooling all replicates (see Supplementary material, Fig. S1a). The individual replicates for each test condition (Arctic Aβ1–42 and Aβ1–42) were then compared with the age-matched control Aβ1–40 pool. All statistical analyses were performed on log10 ratio data.
Gene expression differences were determined by one-way, error-weighted anova on ratio data by comparing Aβ1–42 and Arctic mutants with age-matched Aβ1–40 controls (factor = genotype, n = 4 per group). Genes were only considered in the analysis of over-represented biological themes if the differential expression was highly significant (P < 0.01).
Characterization of over-represented biological themes
Gene Ontology (Ashburner et al., 2000) enrichment analysis was performed using the Gene Ontology tree machine (http://bioinfo.vanderbilt.edu/gotm/) (Zhang et al., 2004). In summary, a hypergeometric test was performed to determine whether a particular biological process or molecular function was disproportionately represented in the set of genes that were differentially expressed as compared with all of the genes in the GeneChip® array. Where n = number of genes that are differentially expressed between two experimental conditions (e.g. between Arctic Aβ1–42 and Aβ1–40 on day 0), N = total number of genes on the array, K = number of genes on the array that belong to the category of interest and k = number of genes that belong to the category that are differentially expressed, the significance (P) of enrichment for a given category is determined by
A ratio of enrichment, R, is calculated as R = k/ke where ke = (n/N)K is the expected value for a given category if n = number of genes were a random sample of uniformly selected genes from the reference set of all genes on the chip.
Gene Search element screen
A library of 3000 unique insertions of the Gene Search (GS) element was generated by mobilizing the GS element from the X chromosome (DGRC number 200079) to the autosomes by crossing with flies expressing the Delta2-3 transposase. The first eclosed fly with a stably-jumped GS element from each mobilization cross was used to establish a GS line. Stocks were maintained by monitoring eye colour. The effect of the Gal-4 activated GS element on the longevity of flies expressing the Arctic mutant of Aβ1–42 was determined by crossing male flies with elavc155-Gal4 (elav-Gal4) on the X chromosome and the Arctic Aβ1–42 transgene on the second chromosome with virgin females with a floating GS element. All of the female offspring had both elav-Gal4 and the Arctic Aβ1–42 transgene but only half had the GS element; the unmodified population (without the GS element) provided an internal control for each longevity assay. The null hypothesis was that the presence of the GS element made no difference to the longevity of the flies expressing the Arctic Aβ1–42 transgene.
Longevity assays for the primary screen were performed at 25°C, blind to the identity of the GS elements. Live flies were counted and the food changed on days 1, 3 and 5 of a 7 day cycle. A mean number of 17 flies was assessed per GS line (total number of flies assessed, 50 320; minimum number of flies assessed per GS line, 10). To detect suppression of the longevity phenotype we determined the time to 75% death, which is the median survival of the flies in which the GS element has prolonged their life. These survival times were normally distributed (n = 2893, mean 21.6 days, SD 3.8 days) and any lifespan that was more than two SDs greater than the mean was defined as significant.
In the secondary screen the chromosomal site of insertion was determined by classical genetics and homozygous GS flies were crossed so that all flies in the longevity assays expressed the activated GS element. This secondary screen used homozygous GS element stocks and assessed more than 30 flies, permitting robust comparison of the survival of GS-modified and unmodified Arctic Aβ1–42 flies. The site of the GS element insertion was identified in those lines with confirmed modifier activity by inverse polymerase chain reaction using the FlyChip facility (http://www.drosdel.org.uk/molecular_methods.php). The molecular function of genes with inserts within the coding sequence was determined from FlyBase. For inserts in non-coding DNA, the molecular function was determined for the genes on either side of the insert. The effect of the activated GS element on the survival of control flies and flies expressing the AlzArc1 transgene was also determined. GS lines in which the modification of the Aβ longevity phenotype was not confirmed or in which the GS element had a marked non-specific effect on control flies were discarded.
Longevity assays for the secondary screen and genetic modifier assessment
Longevity assays in the secondary screen and for each of the subsequent genetic modifiers followed at least 80 flies per genotype in groups of 10 flies per vial. Live flies were counted and their food changed on days 1, 3 and 5 of a 7 day cycle. For assessing the efficacy of metal chelation, clioquinol (Calbiochem) was dissolved in dimethylsulphoxide and the solution added to fly food to give a final concentration of 0.2% v/v. Survival curves were plotted using the Kaplan–Meier estimator. The statistical significance was calculated using the log rank test within the spss 11.0 statistical package. The null hypothesis in all of the longevity assays was that the presence of the GS element made no difference to the longevity of the flies expressing the Arctic Aβ1–42 transgene.
Assay of sensorimotor performance
The sensorimotor performance of the flies was determined using a previously described negative geotaxis assay (Rival et al., 2004). Fifteen flies were placed in a sterile plastic column (25 cm tall × 1.5 cm internal diameter) and tapped to the bottom. After 45 s the flies at the top of the column (Ntop) and the flies remaining at the bottom (Nbot) were counted. Three trials were performed at 1 min intervals. The performance index was defined as (15 + Ntop–Nbot)/30. Statistical analysis was performed using the two-tailed Student’s t-test.
Quantitation of β-amyloid peptides
Drosophila were cultured at 25°C for 5 days after eclosion and then five heads were homogenized in 50 μL of 5 m guanidinium HCl, 5 mm EDTA and 50 mm hepes, pH 7.3. Following centrifugation for 5 min at 12 000 g, 20 μL of the clear supernatant was removed and mixed with 180 μL of 25 mm HEPES, pH 7.3, 1 mm EDTA and 0.1% w/v bovine serum albumin with protease inhibitors (Complete™, Roche). Triplicate 25 μL aliquots were mixed with an equal volume of phosphate-buffered saline containing 2% w/v bovine serum albumin, 0.2% v/v Tween-20 and protease inhibitors (Complete™, Roche) in wells on a MESO microtitre plate (Standard bind, MA6000, no. P11AA-1; Meso Scale Discovery, MD, USA). The reaction was started by adding 25 μL of 4 μg/mL solutions of biotinylated 6E10 or 4G8 (Signet Laboratories, MA, USA) monoclonal antibodies. After mixing a further 25 μL aliquot of 1 μg/mL Ruthenium-labelled G2-10 or G2-11 (The Genetics Company, Switzerland), monoclonal antibody solution was added to each well. Following an overnight incubation at 25°C, the plates were washed twice with phosphate-buffered saline, 150 μL of S Read Buffer (R92SC-1, Meso Scale Discovery, MD, USA) was added and the measurement was taken in a Sector PR instrument (Meso Scale Discovery). Statistical analysis was performed using the two-tailed Student’s t-test.
Hydrogen peroxide sensitivity assay
Two-day-old flies were cultured in a vial containing only filter paper soaked in an aqueous solution of 2% w/v sucrose and 10% v/v H2O2 or a control solution of 2% w/v sucrose alone. Ten vials of 15 flies were cultured for each genotype at 25°C with the soaked paper being replaced twice per day. Statistical analysis was performed using the two-tailed Student’s t-test.
Five adult flies (28 days old, grown at 25°C) were decapitated and the heads were immediately homogenized in water, sonicated and centrifuged at 12 000 g for 10 min and the clear supernatant assayed for protein content using the Bradford method. The protein concentration was adjusted to 5 μg/mL by the addition of phosphate-buffered saline. Protein carbonyl groups were assayed using an enzyme-linked immunosorbent assay-based protocol described by Alamdari et al. (2005). Protein carbonyl groups were reacted with dinitrophenol hydrazine (Sigma-Aldrich) and the resulting dinitrophenol adducts were detected using an anti-DNP rabbit polyclonal antibody (Sigma-Aldrich). Statistical analysis was performed using the two-tailed Student’s t-test.
Iron and zinc determination in fly head extracts by mass spectrometry
Flies were cultured at 29°C and decapitated at 10 days old. Triplicates of 20 fly heads per condition were weighed by difference into a new 15 mL screw-top polypropylene centrifuge tube (part no. 2086-500, Elkay, UK). Nitric acid (500 μL) (‘Trace Select’, part no. 84385, Fluka, UK) was added and the tubes were sealed and incubated overnight at 65°C such that there was no visible solid matter remaining. The sample was prepared by resuspending the extracts in 5 mL of double-distilled water before analysis using an Elan 6100 DRC ICP/MS (Perkin-Elmer). The reaction cell was used for all measurements with the following operating conditions: 1.04 L/min nebulizer flow, 1100 W radio frequency power and a dynamic reaction cell gas flow of 0.7 mL/min of ammonia. Iron (Mr 55.9349), zinc (Mr 65.9260) and calcium (Mr 43.9555), but not copper (Mr 62.9298), were reliably detected. As calcium is not chelated by clioquinol it was used to control for the quantity of tissue in each sample.
- Top of page
- Materials and methods
- Supporting Information
Microarray analysis of gene expression in flies expressing β-amyloid peptide supports the importance of oxidative stress in β-amyloid peptide toxicity
Affymetrix microarrays were used to identify gene expression signatures (one-way anova, P < 0.001) for flies expressing wild-type Aβ1–42 at day 0, 3 and 8 of adult life and Arctic Aβ1–42 at day 0 and 3 by comparison to age-matched Aβ1–40 controls. The derived gene expression signatures for flies expressing wild-type and Arctic Aβ1–42 were then used to determine the biological processes, molecular functions and cellular components that were over-represented in each set (in supplementary Fig. S1b). The role of redox genes appeared most apparent in the early stages of Aβ1–42 expression (Fig. 1); this is most clearly seen for flies expressing wild-type Aβ1–42 at day 3 (top part, shaded) where these genes comprised six of the 10 families of enriched genes. Genes that were classified as having ‘oxidoreductase activity’ were significantly enriched at both day 3 and 8 in flies expressing wild-type Aβ1–42 (P < 0.01). Similarly, carbonyl reductase (NADPH) activity was enriched in the day 3 signature of flies expressing Arctic Aβ1–42 (P < 0.01). Taken together, these observations provide strong support for the role of oxidative stress-related genes in the response of the fly brain to the expression of the Aβ.
P-element screen for modifiers of β-amyloid peptide toxicity implicate redox regulation as an important target of modifier activity
A 3000-line library of unique GS (Toba et al., 1999) element inserts was screened for lines that modified the longevity of Drosophila expressing Aβ1–42 in their central nervous system. The GS elements can either disrupt gene function when they insert within essential coding or non-coding DNA, or else they can bidirectionally upregulate neighbouring genes. In our screen 1.5% of the GS inserts resulted in an increase in median survival that was more than two SDs away from the mean, whereas 0.5% of the inserts significantly reduced survival. A secondary longevity assay was performed to confirm the initial findings but also to determine the effect of the GS elements on the survival of control flies that did not express Aβ. The insertions that specifically modified the longevity of Aβ-expressing flies, and not controls, were classified into 18 suppressor and three enhancer groups according to the identity of the neighbouring genes. Seven of these 21 classes were adjacent to genes with a predictable role in oxidative stress (Table 1 and Fig. 2). The selected suppressor GS elements increased median survival by 25–96% in flies expressing Aβ but had no, or little, effect in control flies (Fig. 2, filled and empty bars, respectively).
|GS elements||Gene name||Gene identifier||Molecular function||GS insertion sites (range of sequence locations)|
|148, 174, 274, 1360, 2225, 2823, 3063||Lethal (2) 01289||CG9432||Thioredoxin-like domain||2252526–2252617|
|2703, 2521, 3102||Fer1HC and Fer2LC||CG2216 and CG1469||Iron chelator and ferroxidase activity||26215006–26215015|
|532, 1287, 1472.2, 1652, 2484||Fer1HC and Fer2LC||CG2216 and CG1469||Iron chelator and ferroxidase activity||26212564–26212784|
|811||Cyp6a20 Cytochrome P450-6a9||CG10245 and CG10246||Oxidoreductase activity in steroid metabolism||10396945|
|1264, 2155||Carbonyl reductase||CG11200||Reduction of aldehydes||15823515–15823522|
|1357||Orthologues of human retinol dehydrogenase||CG2064 and CG2065||Short chain dehydrogenases||20961357|
|1774||Orthologue of human aldehyde dehydrogenase 4A1 and Hsc70Cb||CG6661 and CG6603||Oxidation of aldehydes and chaperone||14029850–14036659|
Flies expressing β-amyloid peptide1–42 are more sensitive to oxidative stress and have higher levels of oxidative damage
The genetic screen and microarray data implicated oxidative stress as playing a central role in the toxicity of the Aβ1–42. This was assessed by testing the response of flies to an oxidative insult in the form of H2O2 in their food. In these experiments we observed that flies expressing wild-type Aβ1–42 were more likely to die if their food was supplemented with 10% v/v H2O2 than either control flies (Fig. 3a, elav-Gal4 w1118 vs. elav-Gal4 UAS-Aβ1–42, P < 0.05) or flies expressing Aβ1–40 (Fig. 3a, elav-Gal4 UAS-Aβ1–40 vs. elav-Gal4 UAS-Aβ1–42, P < 0.05). When we quantified oxidative damage by measuring the levels of carbonyl groups in protein extracts of fly heads we found that flies expressing Aβ1–40 had carbonyl levels that were similar to those of control flies (Fig. 3b, elav-Gal4 w1118 vs. elav-Gal4 UAS-Aβ1–40), whereas flies expressing a single Aβ1–42 transgene had almost double the carbonyl levels of flies expressing Aβ1–40 (Fig. 3b, elav-Gal4 UAS-Aβ1–42 vs. elav-Gal4 UAS-Aβ1–40, P < 0.05).
Transgenic over-expression of single antioxidative stress genes rescues the β-amyloid peptide-induced longevity phenotype
The ability of antioxidative stress genes identified in the screen, and genes from the canonical oxidative stress pathway, to modify the toxicity of Aβ was confirmed in flies by specifically over-expressing each transgene in combination with wild-type or Arctic Aβ1–42. In this way we assessed candidate genes from the GS element screen, i.e. the heavy and light chains of Drosophila ferritin (Fer1HC and Fer2LC) and carbonyl reductase by over-expression of the Drosophila enzyme Sniffer (SNI). We also tested candidate genes including cytoplasmic CuZn-SOD1, the mitochondrial Mn-SOD2 (mitSOD2) and CAT, and GST (Fig. 4). This approach allowed comparison of the efficacy of the upstream enzymes that modulate the generation of free radicals (SOD1, mitSOD2, CAT, Fer1HC and Fer2LC) with the downstream enzymes that repair oxidative damage (GST and carbonyl reductase/SNI).
Over-expression of ferritin heavy chain resulted in a 105% increase in median survival of Arctic Aβ1–42 flies (P < 0.0001), whereas over-expression of ferritin light chain gave a 49% increase in median survival (P < 0.0001) (Fig. 4a and b). The most powerful canonical oxidative stress-related protein was CAT, which increased the median survival of Arctic Aβ1–42 flies by 56% (P < 0.0001, Fig. 4a and c). MitSOD2 prolonged median survival by 18% (P < 0.0001, Fig. 4a and d). In contrast, and contrary to our expectations, the over-expression of cytoplasmic SOD1 enhanced the toxicity of Arctic Aβ1–42. Flies expressing mutants of SOD1 that are dominant-negative for activity were then assessed to see if they had a similar modifying activity. Co-expression of both the SOD1n108 mutant and RNAi for SOD1 resulted in a 43% and 32% increase in median survival, respectively (Fig. 4a and c), indicating that it is the catalytic activity of SOD1 that potentiates the toxic effect of the Aβ. The protection afforded by CAT combined with the toxicity of SOD1 suggests that the uncompensated production of H2O2 is a vital step in the oxidative stress caused by the Aβ. An additional stress resulting from increased SOD1 activity may be the co-production of O2 by the dismutation of superoxide radicals. This oxygen may go on to generate further H2O2 by reacting with iron or copper ions that are complexed with Aβ (Huang et al., 1999a,b).
Much smaller protective activities were observed when enzymes, specifically Drosophila carbonyl reductase (SNI) and GST, that are involved in steps further downstream in the pathway were upregulated (Fig. 4a). Control experiments, in which the antioxidative stress genes were over-expressed in flies that were identical except that they lacked the Aβ transgene, did not reveal non-specific prolongation of life (Fig. 4d). In flies expressing wild-type Aβ1–42 we were able to confirm the prolongation of lifespan by the heavy and light chains of ferritin and the toxic effect of SOD1 (data not shown).
Transgenic over-expression of single antioxidative stress genes rescues the β-amyloid peptide-induced locomotor phenotype
The modifiers were then assessed for their ability to protect against the Aβ1–42-mediated decline in locomotor function. Flies co-expressing either ferritin subunit with Arctic Aβ1–42 were significantly more mobile than flies expressing Arctic Aβ1–42 alone (Fig. 5a, elav-Gal4 UAS-Arctic Aβ1–42) from day 10 onwards, and both the heavy and light chain flies performed like control flies (Fig. 5a, elav-Gal4) up to day 25. The effect of the heavy chain was more potent, maintaining wild-type locomotor performance in Arctic Aβ1–42 flies to day 35. The behavioural effects of over-expressing the canonical antioxidative genes were in accord with their longevity data; CAT (Fig. 5b, elav-Gal4 UAS-Arctic Aβ1–42UAS-CAT) and the mitSOD2 (Fig. 5b, elav-Gal4 UAS-Arctic Aβ1–42UAS-mitSOD2) were suppressors of the Arctic Aβ1–42 locomotor phenotype, whereas the cytoplasmic SOD1 (Fig. 5b, elav-Gal4 UAS-Arctic Aβ1–42UAS-SOD1) accelerated the decline in locomotor function. Over-expression of the heavy chain of ferritin (Fig. 5c, elav-Gal4 UAS-Aβ1–42UAS-FerHC) and CAT (Fig. 5c, elav-Gal4 UAS-Aβ1–42UAS-CAT) had similar beneficial effects in flies expressing wild-type Aβ1–42 (Fig. 5c, elav-Gal4 UAS-Aβ1–42). As expected, flies co-expressing SOD1 (Fig. 5c, elav-Gal4 UAS-Aβ1–42UAS-SOD1) with wild-type Aβ1–42 demonstrated an accelerated decline in locomotor function. The co-expression of GST (Fig. 5d, elav-Gal4 UAS-Arctic Aβ1–42UAS-GST) gave a weak but significant improvement in locomotor function from day 10 onwards. Carbonyl reductase (Fig. 5d, elav-Gal4 UAS-Arctic Aβ1–42UAS-SNI) did not significantly improve locomotor function.
The metal chelator clioquinol reduces β-amyloid peptide-mediated neuronal toxicity and specifically reduces iron levels in the brain
That the protection against Aβ toxicity afforded by ferritin is mediated in part by the chelation of Fe2+ and Fe3+ ions was investigated by treating flies expressing Aβ1–40, wild-type Aβ1–42 and Arctic Aβ1–42 with the metal chelator clioquinol (Kaur et al., 2003) (Fig. 6). Although clioquinol had no effect on the longevity of flies expressing the non-toxic Aβ1–40 (Fig. 6a, circles) there was, in contrast, a clear dose-related increase in longevity for Arctic Aβ1–42 (Fig. 6a, diamonds) flies. Expression of wild-type Aβ1–42 (Fig. 6a, triangles) gave an intermediate response to clioquinol treatment with an optimum dose of approximately of 20 μm. The degree of functional rescue in flies expressing Arctic Aβ1–42 closely reflected the concentration of iron in the extracts of fly heads (Fig. 6b). We found that the presence of Arctic Aβ1–42 (Fig. 6b, triangles) increased the uptake of iron as compared with control flies (Fig. 6b, circles), a striking effect that is only completely reversed by treatment with 200 μm clioquinol. The levels of zinc (1.5–1.7 mm) were the same in Arctic Aβ1–42-expressing and control flies and remained unchanged following clioquinol treatment. Copper could not be detected in the head extracts.
Oxidative damage in flies expressing β-amyloid peptide is reduced by ferritin and increased by superoxide dismutase 1
The quantity of carbonyl groups was assessed in flies expressing Arctic Aβ1–42 in the presence or absence of the heavy and light chains of ferritin. The expression of Arctic Aβ1–42 significantly increased the concentration of carbonyl groups (Fig. 7a, elav-Gal4 w1118 vs. elav-Gal4 UAS-Arctic Aβ1–42, P < 0.001) but this was reduced by 30% following the co-expression of ferritin light chain (Fig. 7a, elav-Gal4 UAS-Arctic Aβ1–42UAS-FerLC and elav-Gal4 UAS-Arctic Aβ1–42, P < 0.01). Remarkably, ferritin heavy chain (Fig. 7a, elav-Gal4 UAS-Arctic Aβ1–42UAS-FerHC) reduced the carbonyl level (P < 0.001) almost to that of control flies (Fig. 7a, control elav-Gal4 UAS-Arctic Aβ1–42). These findings are consistent with ferritin exerting its suppression of the Aβ phenotype by an antioxidant effect. Conversely, the enhancer effect of SOD1 was accompanied by a significant increase in carbonyl levels in Arctic Aβ1–42 flies (Fig. 7a, elav-Gal4 UAS-Arctic Aβ1–42 vs. elav-Gal4 UAS-Arctic Aβ1–42UAS-SOD1, P < 0.01). These data suggest that the rapid production of H2O2 by cytoplasmic SOD1 can overwhelm endogenous CAT resulting in the Fe2+-mediated generation of the hydroxyl radical (Fenton reaction) that damages proteins as demonstrated by the increased carbonyl load.
Ferritin subunits suppress the toxicity of β-amyloid peptide1–42 despite increased levels of β-amyloid peptide1–42 in the brain
The effect of over-expressing antioxidant transgenes on the level of Aβ1–42 in the brains of flies was assessed by an enzyme-linked immunosorbent assay that measured 5 m guanidinium hydrochloride-soluble Aβ1–42. Six independent lines of flies expressing Arctic Aβ1–42 and carrying GS elements that cause over-expression of ferritin subunits were shown to accumulate significantly higher levels of Aβ1–42 in their brains (65 pm in control flies vs. 100 pm for GS elements co-expressing the light chain gene, P < 0.01, and 141 pm over-expressing the heavy chain gene, P < 0.01). This elevation of Aβ1–42 when over-expressing ferritin subunits was confirmed in flies co-expressing Aβ1–42 with UAS-linked transgenes for either the heavy or light chains (Fig. 7b). In contrast, the levels of Aβ1–42 were unaffected by the co-expression of CAT or the over-expression, or the knockdown, of SOD-1. Thus, the rescue mediated by over-expression of oxidative stress genes was not mediated by suppression of the Aβ1–42.
- Top of page
- Materials and methods
- Supporting Information
Two genome-wide genetic screens were used in our Drosophila model of AD to identify genes that play an important role in the toxicity of the Aβ. The first screen, using Drosophila cDNA microarrays, quantified the changes in gene transcription that occurred in response to the expression of Aβ1–42 as compared with Aβ1–40. The second complementary screen, using a library of flies with unique GS element insertions, identified genes, which when transcriptionally up- or downregulated modified the lifespan of flies expressing Aβ1–42. Taken as a whole the GS screen was remarkable for the number of genes adjacent to modifying P-elements that had redox or antioxidant activities. Similarly, the cDNA microarray studies showed that redox-associated transcripts were the most robustly represented functional group of genes that were differentially regulated by Aβ1–42 expression. This combination of observations is powerful because, whereas the microarray data provide genetic evidence of a response by the brain to oxidative stress, the GS screen, in contrast, points to a direct pathogenic role for oxidative stress in the generation of Aβ-related phenotypes. The role of oxidation in causing disease was further emphasized by the remarkably close correlation between the severity of the phenotypes observed for Aβ1–40, Aβ1–42 and Arctic Aβ1–42 and the oxidative modification of brain proteins as determined by carbonyl levels. These findings are concordant with the raised levels of oxidized proteins and lipids in post-mortem AD brains where it is known that oxidative damage is present from the earliest clinical stages of the disease (Nunomura et al., 2001; Markesbery et al., 2005). Our work therefore represents an advance on current clinical work and mouse models that have not clarified whether oxidative stress plays a direct role in the pathogenesis of AD or whether it is a consequence of the disease process. Indeed, the view that oxidative stress is a bystander effect is supported by clinical trial data that show, despite early encouraging data from observational studies (Morris et al., 2002a,b; Zandi et al., 2004) and one prospective trial (Sano et al., 1997), that the antioxidant vitamin E is not able to prevent the onset of dementia or prevent progression of established disease (Petersen et al., 2005).
In this work we have undertaken a molecular dissection of the oxidative stress pathway (Fig. 8) in order to gain a more detailed understanding of which oxidative species were of particular importance in Aβ-mediated neurodegeneration. This work has emphasized three main sources of oxidative stress: firstly, the generation of hydroxyl radicals via the Fenton reaction (Smith et al., 1997; Dikalov et al., 2004), secondly, mitochondrial superoxide levels and thirdly, reactive aldehyde species produced by lipid oxidation. Although each of these sources contributes significantly to oxidative stress, the conversion of H2O2 to the hydroxyl radical is the most important toxic process. The most immediate support for this comes from the profound protection afforded to flies when CAT is co-expressed with Aβ1–42 because hydroxyl radical generation will predictably be abolished by converting H2O2 to water.
Only co-expression of ferritin subunits offered a more effective rescue of the longevity phenotype than CAT. Ferritin proteins are highly conserved in evolution and have been extensively characterized (Harrison & Arosio, 1996). Functional ferritin complexes may be composed of various proportions of heavy and light subunits. For instance, in the human, ferritin from the brain is predominantly composed of heavy chain, whereas in the liver the ferritin has a high proportion of light chain. In our model, the over-expression of Fer1HC or Fer2LC occurred in the context of endogenous Fer1HC and Fer2LC expression. Thus, in our transgenic animals we should produce ferritin complexes composed of both subunits but with one or the other representing the major component. Remarkably, ferritin heavy chain was able to restore the longevity and locomotor phenotype of flies expressing the highly toxic Arctic Aβ1–42 to that of control flies. That this rescue was mediated by an antioxidant effect is supported by the reductions in carbonyl levels in the brains of flies co-expressing either the light or, again more potently, heavy chain of ferritin. It is likely that the sequestration of iron by both ferritin subunits, and the conversion of Fe2+ to Fe3+ by the heavy chain, slows hydroxyl radical production. Our data show for the first time that it is likely to be the removal of Fe2+, by the ferroxidase activity of the heavy chain, that is specifically beneficial. Previous in-vitro data have shown that synthetic Aβ can directly generate H2O2 in the presence of metal ions (Huang et al., 1999a; Tabner et al., 2005) and in particular iron (Khan et al., 2006). Using molecular oxygen as a substrate, the production of H2O2 by synthetic Aβ depends on Fe2+ ions generated via a redox cycling of iron (Huang et al., 1999a; Khan et al., 2006). Thus, it is possible that ferritin can not only prevent the Fenton reaction but additionally the Fe2+-scavenging activity of ferritin heavy chain can also protect neurones against the intrinsic redox properties of Aβ (Fig. 8).
The powerful antioxidant properties of ferritin can also rescue the locomotor deficits associated with Aβ expression despite an accompanying twofold increase in Aβ load. These data make the modulation of brain iron metabolism an attractive therapeutic target not least because the concentration of Fe2+ in amyloid plaques is 1 mm, almost three times the normal level (Bush, 2003). Furthermore, clioquinol, a metal chelator, reduces plaque deposition in mouse models of AD (Cherny et al., 2001), is safe in clinical trials (Ritchie et al., 2003) and, as we show here, specifically reduces iron levels in the brain and prolongs the survival of flies expressing Arctic Aβ1–42.
Oxidative stress has been linked with mitochondrial dysfunction in several neurodegenerative disorders including AD and also Parkinson’s and Huntington’s diseases (Bowling & Beal, 1995; Maier & Chan, 2002). In AD there is evidence that soluble oligomeric aggregates of Aβ may damage the insulating properties of the neuronal plasma membrane (Kayed et al., 2004) resulting in calcium influx (Demuro et al., 2005). Mitochondria from patients with AD are impaired in their ability to buffer calcium influxes and also show defects in their respiratory complexes, particularly complex IV, resulting in increased generation of reactive oxygen species (Sheehan et al., 1997; Abramov et al., 2004). Our data are consistent with these hypotheses showing that over-expression of the mitochondrial superoxide scavenging enzyme mitSOD2 offered modest but significant protection against Aβ toxicity.
It was surprising that the co-expression of SOD1 enhanced the toxicity of Aβ1–42. This result underlines the low toxicity of the superoxide anion and again emphasizes the importance of H2O2 in the oxidative stress pathway. By showing that dominant negative mutants of SOD1 prolong the life of the AD flies, our findings demonstrate that SOD1 catalytic activity mediates its toxicity. Furthermore, RNAi knockdown of SOD1 protein expression did not provide any additional rescue, indicating that our data are not confounded by toxicity caused by SOD1 misfolding and aggregation as is proposed for SOD1-linked familial amyotrophic lateral sclerosis (Lynch et al., 2004). The finding of elevated carbonyl levels in the brains of flies co-expressing SOD1 and Arctic Aβ1–42 suggests that rapid dismutation of superoxide can overwhelm the capacity of endogenous CAT to remove the consequent H2O2. In addition, the dismutation of superoxide radicals releases molecular oxygen that may become a substrate for the Aβ-dependent production of H2O2 (Fig. 8). Nevertheless, the enhancer activity of SOD1 in Drosophila was unexpected because previous studies in a mouse model of AD showed that SOD1 knockdown caused increased activation of apoptotic pathways (Chen et al., 2005). Further work is required to determine whether SOD1 upregulation is toxic in a mammalian model of AD.
Finally, we have evaluated the effect of two enzymes that are involved in the detoxification of reactive aldehydes (GST and carbonyl reductase). Reactive aldehydes, such as 4-oxonon-2-enal, result from hydroxyl radical-mediated lipid oxidation and accumulate in the brains of patients with AD (Lovell et al., 1997; Markesbery & Lovell, 1998). Despite being less reactive than the hydroxyl radical, the greater stability of the aldehydes permits damage to proteins and DNA over longer times and distances within the cell. We have shown that over-expression of GST or carbonyl reductase gave only a modest rescue of Aβ toxicity, suggesting that lipid damage is indeed downstream of the most important toxic events.
In summary, we have used genetic screens and a candidate gene approach to dissect the contribution of reactive oxygen species to the toxicity of Aβ in our Drosophila model of AD. We have found that the primary oxidative stressors are likely to be H2O2 and the consequent hydroxyl radical. Preventing oxidative stress, specifically by manipulating iron metabolism, provides a powerful strategy for reducing Aβ toxicity in AD.
- Top of page
- Materials and methods
- Supporting Information
This work was supported by the Wellcome Trust, Medical Research Council (UK), Papworth NHS Trust and an industrial collaboration grant between Merck Sharp & Dohme Ltd and the University of Cambridge. R.M.P. was funded by a studentship as part of an educational agreement between Merck Sharp & Dohme Ltd and the University of Cambridge. We are grateful to Professor M. Ashburner, Dr S. Russell and the Department of Genetics, University of Cambridge for the use of the fly laboratory facilities and thank Dr S. Imarisio (Cambridge Institute for Medical Research) for her comments on the manuscript.
ferritin 1 heavy chain
ferritin 2 light chain
mitochondrial Mn-superoxide dismutase 2
upstream activating sequence
- Top of page
- Materials and methods
- Supporting Information
- 2004) Beta-amyloid peptides induce mitochondrial dysfunction and oxidative stress in astrocytes and death of neurons through activation of NADPH oxidase. J. Neurosci., 24, 565–575. , & (
- 2005) High sensitivity enzyme-linked immunosorbent assay (ELISA) method for measuring protein carbonyl in samples with low amounts of protein. Free Radic. Biol. Med., 39, 1362–1367. , , , , , & (
- 2005) RNAi-mediated suppression of the mitochondrial iron chaperone, frataxin, in Drosophila. Hum. Mol. Genet., 14, 3397–3405. , , & (
- 2000) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet., 25, 25–29. , , , , , , , , , , , , , , , , , , & (
- 2004) The Drosophila carbonyl reductase sniffer prevents oxidative stress-induced neurodegeneration. Curr. Biol., 14, 782–786. , , , , , & (
- 1995) Bioenergetic and oxidative stress in neurodegenerative diseases. Life Sci., 56, 1151–1171. & (
- 2003) The metallobiology of Alzheimer’s disease. Trends Neurosci., 26, 207–214. (
- 2008) Identification of novel genes that modify phenotypes induced by Alzheimer’s β-amyloid overexpression in Drosophila. Genetics, 178, 1457–1471. , , , , , & (
- 2005) SOD-1 inhibits FAS expression in cortex of APP transgenic mice. Apoptosis, 10, 499–502. , , , , & (
- 2001) Treatment with a copper-zinc chelator markedly and rapidly inhibits beta-amyloid accumulation in Alzheimer’s disease transgenic mice. Neuron, 30, 665–676. , , , , , , , , , , , , , , , , , & (
- 2005) Intraneuronal Aβ, non-amyloid aggregates and neurodegeneration in a Drosophila model of Alzheimer’s disease. Neuroscience, 132, 123–135. , , , , , , & (
- 2005) Calcium dysregulation and membrane disruption as a ubiquitous neurotoxic mechanism of soluble amyloid oligomers. J. Biol. Chem., 280, 17294–17300. , , , , & (
- 2004) Cupric-amyloid beta peptide complex stimulates oxidation of ascorbate and generation of hydroxyl radical. Free Radic. Biol. Med., 36, 340–347. , & (
- 1996) The ferritins: molecular properties, iron storage function and cellular regulation. Biochim. Biophys. Acta, 1275, 161–203. & (
- 1999a) The A beta peptide of Alzheimer’s disease directly produces hydrogen peroxide through metal ion reduction. Biochemistry, 38, 7609–7616. , , , , , , , , , , & (
- 1999b) Cu(II) potentiation of alzheimer abeta neurotoxicity. Correlation with cell-free hydrogen peroxide production and metal reduction. J. Biol. Chem., 274, 37111–37116. , , , , , , , , , , , , , , , , , , & (
- 2003) Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res., 31, e15. , , , , & (
- 2004) Induction of pro-apoptotic calsenilin/DREAM/KChIP3 in Alzheimer’s disease and cultured neurons after amyloid-beta exposure. J. Neurochem., 88, 1570. , , , , , & (
- 2003) Genetic or pharmacological iron chelation prevents MPTP-induced neurotoxicity in vivo: a novel therapy for Parkinson’s disease. Neuron, 37, 899–909. , , , , , , , , , , , , , , & (
- 2004) Permeabilization of lipid bilayers is a common conformation-dependent activity of soluble amyloid oligomers in protein misfolding diseases. J. Biol. Chem., 279, 46363–46366. , , , , , & (
- 2006) Redox cycling of iron by Abeta42. Free Radic. Biol. Med., 40, 557–569. , & (
- 2002) Intracellular amyloid-beta 1–42, but not extracellular soluble amyloid-beta peptides, induces neuronal apoptosis. J. Biol. Chem., 277, 15666–15670. , , & (
- 1994) Evidence that production and release of amyloid beta-protein involves the endocytic pathway. J. Biol. Chem., 269, 17386–17389. & (
- 1998) Diffusible, nonfibrillar ligands derived from Abeta1–42 are potent central nervous system neurotoxins. Proc. Natl Acad. Sci. USA, 95, 6448–6453. , , , , , , , , , , , , , & (
- 2002) Neurodegenerative disease: amyloid pores from pathogenic mutations. Nature, 418, 291. , , , & (
- 1997) Elevated 4-hydroxynonenal in ventricular fluid in Alzheimer’s disease. Neurobiol. Aging, 18, 457–461. , , & (
- 2007) Systematic in vivo analysis of the intrinsic determinants of amyloid beta pathogenicity. PLoS Biol., 5, e290. , , , , , , , , & (
- 2004) Kinetic stability of Cu/Zn superoxide dismutase is dependent on its metal ligands: implications for ALS. Biochemistry, 43, 16525–16531. , & (
- 2002) Role of superoxide dismutases in oxidative damage and neurodegenerative disorders. Neuroscientist, 8, 323–334. & (
- 1998) Four-hydroxynonenal, a product of lipid peroxidation, is increased in the brain in Alzheimer’s disease. Neurobiol. Aging, 19, 33–36. & (
- 2005) Lipid peroxidation is an early event in the brain in amnestic mild cognitive impairment. Ann. Neurol., 58, 730–735. , , & (
- 2003) Compartment-specific protection of iron-sulfur proteins by superoxide dismutase. J. Biol. Chem., 278, 47365–47369. , , , , & (
- 2002a) Dietary intake of antioxidant nutrients and the risk of incident Alzheimer disease in a biracial community study. JAMA, 287, 3230–3237. , , , , , , & (
- 2002b) Vitamin E and cognitive decline in older persons. Arch. Neurol., 59, 1125–1132. , , , & (
- 2001) The ‘Arctic’ APP mutation (E693G) causes Alzheimer’s disease by enhanced Abeta protofibril formation. Nat. Neurosci., 4, 887–893. , , , , , , , , , , & (
- 2001) Oxidative damage is the earliest event in Alzheimer disease. J. Neuropathol. Exp. Neurol., 60, 759–767. , , , , , , , , , , , , & (
- 2005) Vitamin E and donepezil for the treatment of mild cognitive impairment. N. Engl. J. Med., 352, 2379–2388. , , , , , , , , , , , , & (
- 1995) Subunit-destabilizing mutations in Drosophila copper/zinc superoxide dismutase: neuropathology and a model of dimer dysequilibrium. Proc. Natl Acad. Sci. USA, 92, 8574–8578. , , , , & (
- 1995) Evidence that secretase cleavage of cell surface Alzheimer amyloid precursor occurs after normal endocytic internalization. J. Neurosci. Res., 40, 694–706. , , , & (
- 2003) Metal-protein attenuation with iodochlorhydroxyquin (clioquinol) targeting Abeta amyloid deposition and toxicity in Alzheimer disease: a pilot phase 2 clinical trial. Arch. Neurol., 60, 1685–1691. , , , , , , , , , , , , , , , , , & (
- 2004) Decreasing glutamate buffering capacity triggers oxidative stress and neuropil degeneration in the Drosophila brain. Curr. Biol., 14, 599–605. , , , , & (
- 1997) A controlled trial of selegiline, alpha-tocopherol, or both as treatment for Alzheimer’s disease. The Alzheimer’s Disease Cooperative Study. N. Engl. J. Med., 336, 1216–1222. , , , , , , , , , , & (
- 1997) Calcium homeostasis and reactive oxygen species production in cells transformed by mitochondria from individuals with sporadic Alzheimer’s disease. J. Neurosci., 17, 4612–4622. , , , , , & (
- 1997) Iron accumulation in Alzheimer disease is a source of redox-generated free radicals. Proc. Natl Acad. Sci. USA, 94, 9866–9868. , , & (
- 2005) Hydrogen peroxide is generated during the very early stages of aggregation of the amyloid peptides implicated in Alzheimer Disease and familial British dementia. J. Biol. Chem., 280, 35789–35792. , , , , , , , & (
- 1999) The gene search system. A method for efficient detection and rapid molecular identification of genes in Drosophila melanogaster. Genetics, 151, 725–737. , , , , & (
- 2002) Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term potentiation in vivo. Nature, 416, 535–539. , , , , , , & (
- 2002) Abeta 17-42 in Alzheimer’s disease activates JNK and caspase-8 leading to neuronal apoptosis. Brain, 125, 2036–2043. , , & (
- 2005) Increased glutathione S-transferase activity rescues dopaminergic neuron loss in a Drosophila model of Parkinson’s disease. Proc. Natl Acad. Sci. USA, 102, 8024–8029. , , , , & (
- 2004) Reduced risk of Alzheimer disease in users of antioxidant vitamin supplements: the Cache County Study. Arch. Neurol., 61, 82–88. , , , , , , , & (
- 2004) GOTree Machine (GOTM): a web-based platform for interpreting sets of interesting genes using Gene Ontology hierarchies. BMC Bioinformatics, 5, 16. , , & (
- Top of page
- Materials and methods
- Supporting Information
Fig. S1. (A) Experimental design for the Affymetrix chip analysis of differential gene transcription.
Fig. S2. In control experiments, the expression of the key modifiers of A toxicity in the absence of A did not have a beneficial effect on locomotor function at any time point.
Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
|EJN_6701_sm_Fig. S1.doc||217K||Supporting info item|
|EJN_6701_sm_Fig. S2.doc||186K||Supporting info item|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.