Ageing Increases Vulnerability to Aβ42 Toxicity in Drosophila

Age is the major risk factor for many neurodegenerative diseases, including Alzheimer's Disease (AD), for reasons that are not clear. The association could indicate that the duration or degree of exposure to toxic proteins is important for pathology, or that age itself increases susceptibility to protein toxicity. Using an inducible Drosophila model of AD, we investigated these possibilities by varying the expression of an Aβ42 transgene in neurons at different adult ages and measuring the effects on Aβ42 levels and associated pathological phenotypes. Acute induction of Arctic Aβ42 in young adult flies resulted in rapid expression and clearance of mRNA and soluble Arctic Aβ42 protein, but in irreversible expression of insoluble Arctic Aβ42 peptide. Arctic Aβ42 peptide levels accumulated with longer durations of induction, and this led to a dose-dependent reduction in negative geotaxis and lifespan. For a standardised level of mRNA expression, older flies had higher levels of Arctic Aβ42 peptide and associated toxicity, and this correlated with an age-dependent reduction in proteasome activity. Equalising Aβ42 protein at different ages shortened lifespan in correlation with the duration of exposure to the peptide, suggesting that Aβ42 expression accumulates damage over time. However, the relative reduction in lifespan compared to controls was greater in flies first exposed to the peptide at older ages, suggesting that ageing itself also increases susceptibility to Aβ42 toxicity. Indeed older flies were more vulnerable to chronic Aβ42 toxicity even with a much lower lifetime exposure to the peptide. Finally, the persistence of insoluble Aβ42 in both young and old induced flies suggests that aggregated forms of the peptide cause toxicity in later life. Our results suggest that reduced protein turnover, increased duration of exposure and increased vulnerability to protein toxicity at later ages in combination could explain the late age-of-onset of neurodegenerative phenotypes.


Introduction
Ageing is the major risk factor for common, chronic, killer conditions, including cancer, cardiovascular disease and neurodegeneration. The aetiology of many neurodegenerative diseases includes the formation of toxic protein aggregates in neurons. For instance Alzheimer's disease (AD), the most prevalent form of senile dementia, is characterised by the widespread presence of extracellular amyloid plaques, predominantly composed of amyloid beta (Ab) peptides, and intraneuronal neurofibrillary tangles formed from insoluble fibrillar aggregates of the microtubulebinding protein tau [1]. Most cases of AD (.99%) are sporadic [2], with age as the main risk factor [3,4].
A large body of evidence suggests that the accumulation and deposition of Ab peptides are the primary influence driving the disease [5,6]. Support for this 'amyloid cascade' hypothesis comes from mutations causing early-onset, familial AD (FAD), which affect the amyloid precursor protein (APP), from which Ab peptides are derived, and presenilins PS1 and PS2, which are involved in the cleavage of APP to yield Ab peptides. These mutations increase production of Ab, levels of Ab42 relative to Ab40 or the propensity of Ab42 to aggregate (for review see [7]). In addition, many mouse models of AD, typically based on the overexpression of FAD associated APP alone or in combination with mutations in PS1, develop age-dependent Ab plaques and behavioral and memory deficits [8]. Furthermore, studies in the fruit fly Drosophila melanogaster [9,10,11,12] and the nematode worm Caenorhabditis elegans [13,14] have demonstrated that direct expression of the toxic Ab42 peptide leads to an age-dependent accumulation of Ab42, neuronal dysfunction and shortened lifespan.
Other neurodegenerative conditions share both the aggregation of toxic protein and a late age of onset [15]. For instance, Parkinson's Disease is associated with aggregation of a-synuclein in Lewy bodies, with most cases sporadic and age the main risk factor [16]. However the mechanisms linking protein aggregation and the appearance of disease to age remain to be identified. Protein aggregation may accumulate to a toxic threshold, or the duration and extent of exposure may be important for inducing neuronal dysfunction. In addition, the ageing process itself could increase vulnerability to toxic proteins by impairing their clearance or increasing vulnerability to their toxic effects [15,17]. All of these factors could be important -they are not mutually exclusive.
To investigate the mechanisms linking protein toxicity to age, we have induced expression of a toxic protein at different ages in neurons of adult Drosophila. If ageing plays a direct role then we would expect that standardised exposure of older neurons to a transgene encoding a toxic protein would lead either to greater levels of toxic protein, or greater toxicity of a standard dose of protein, in older flies. The few studies that have addressed this issue have suggested that ageing is indeed important. Brewer et al., (1998) isolated neurons from embryonic, young and old-aged rat hippocampus and exposed them to toxic Ab fragments (25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) finding that toxicity, as measured by cell death, was age-, doseand time-dependent. Guela et al., (1998) demonstrated that aged rhesus monkeys were more vulnerable than young to injected plaque-equivalent concentrations of fibrillar Ab, with the young monkeys developing no pathology at all. However, these studies used either extracellular application of Ab fragments or microinjection of fibrillar Ab to induce toxicity, which makes some assumptions about the nature and site of Ab toxicity. Evidence is increasing that intracellular [18], soluble oligomers of Ab, rather than extracellular, fibrillar forms of the peptide, are pathogenic [19]. Furthermore, ageing could increase vulnerability to Ab toxicity through the mechanisms by which the peptide is produced and broken down. Hence, we have taken advantage of powerful systems for conditional and tissue-specific gene over-expression in Drosophila [20] to examine the effects of age on Ab peptide toxicity under physiological conditions in vivo.
We have previously shown that induced expression of the FADassociated Arctic Ab42 (Arc Ab42) isoform in adult neurons of Drosophila results in age-dependent locomotor and electrophysiological deficits and shortened lifespan [21]. In the present study, we show that these pathological phenotypes are dependent on the concentration of Ab42. For equivalent Ab42 mRNA levels, higher levels of Arc Ab42 peptide were present in older flies, and this correlated with an age-dependent reduction in proteasome activity and resulted in shortened lifespan. Controlling for this agedependent reduction in protein turnover, expression of equivalent amounts of Ab42 protein at different ages shortened lifespan in correlation with the duration of exposure to the peptide. This suggests that the relationship between Ab42 toxicity and age reflects an accumulation of damage over time. Despite this, however, the relative reduction in lifespan compared to non-Arctic Ab42-expressing controls was greater in flies first exposed to the peptide at older ages, suggesting that ageing itself also increases susceptibility to Ab42 toxicity. Indeed older flies were more vulnerable to death from even much lower levels of exposure to Ab42 protein. Our results therefore suggest that increasing age is associated with greater protein toxicity through a combination of reduced clearance, increased exposure and increased vulnerability to pathogenic proteins.

Dynamics of Ab42 expression and toxicity using the elavGS-UAS system
We previously characterised an adult-onset, fly model of Alzheimer's disease [21], generated by expressing the Arctic Ab42 peptide (UAS-Arc Ab42; [11]) under the control of an RU486-inducible pan-neuronal driver (elav GeneSwitch (elavGS); [20,22]). Under chronic induction conditions the levels of Arc Ab42 mRNA and protein expression appeared to vary with age [21]. Ab42 RNA expression declined with age, probably reflecting a lower intake of the RU486 inducer as a consequence of the known age-dependent reduction in feeding behaviour [23], whereas protein levels increased with age, reflecting either a time-dependent accumulation of the peptide or an age-dependent reduction in protein turnover, or both. We now show further that, under acute induction conditions, Ab42 mRNA ( Figure 1A) and soluble protein ( Figure 1B) levels are rapidly cleared following removal of RU486, whereas aggregated forms of the peptide are stable for up to one week ( Figure 1B). Varying the duration of RU486 exposure for 2, 4, 7 or 14 days (see Table S1 for pulse conditions), did not alter the level of Arc Ab42 mRNA expressed at the end of the induction period ( Figure 1C), suggesting that the age-dependent reduction in transcript observed under chronic conditions occurs at ages over 21 days. The concentration of Arc Ab42 protein increased in correlation with the duration of RU486 exposure ( Figure 1D) and this led to progressively more severe reductions in negative geotaxis ( Figure 1E) and survival ( Figure 1F), presumably due to the persistence of insoluble forms of peptide after cessation of induction of gene expression.
Age-dependent reduction in protein turnover correlates with Ab42 peptide accumulation and toxicity We next aimed to use our inducible model to investigate the intrinsic effects of age on vulnerability to Ab42 toxicity, by inducing comparable levels of Arc Ab42 at different ages and measuring the time to develop, and extent of, the subsequent reduction in survival. This output variable allowed us to induce Ab42 expression at a wide range of ages prior to the onset of ageing-related deaths, unlike negative geotaxis, which starts to decline already in the first week of life. Notably, chronic administration of RU486 at the highest (200 mM) dose used in our study had no effect on lifespan of non-Ab42-expressing flies ( Figure S1), demonstrating that our observed effects of Ab42 induction on survival are not attributable to the RU486 treatment conditions.
Our analysis of the dynamics of Ab42 expression, as described above, suggested that to standardise Ab42 levels at different ages both the concentration of RU486 and the length of exposure to the inducer must be finely tuned. We first equalised expression levels of Ab42 mRNA in flies at day 5 or day 20 post-eclosion, by varying the concentration of RU486, inducing for 1 week and measuring transcript levels at the end of the induction period ( Figure S2). Induction at all RU486 concentrations produced significant (P,0.05) Ab42 over-expression, and levels of Ab42 mRNA were similar in flies induced at day 5 with 50 mM (RU 50 [5-12d]) and at day 20 with 200 mM (RU 200 [20-27d]) RU486, respectively (Figures 2A and S2). However, these standardised levels of mRNA expression led to different levels of Ab42 protein, measured at the end of the 1-week pulse (0d) and 1, 2 and 3 weeks following switch-off ( Figure 2B). Total Arctic Ab42 peptide levels were significantly higher (P,0.0001, two-way ANOVA) in RU 200 [20-27d] than in RU 50 [5-12d] induced flies, and no significant reduction occurred for up to 3 weeks following switchoff in either age group, further confirming that Arctic Ab42 peptide is highly resistant to degradation. Presumably associated with the higher levels of Arctic Ab42 peptide present in the day-20-induced flies (RU 200 [20-27d]), only they displayed a shortened lifespan in response to Ab42 expression ( Figure 2C).
For a given level of mRNA, older flies thus had higher levels of Ab42 peptide. Protein turnover declines with age in several organisms [24,25], suggesting that age-dependent effects on protein synthesis or degradation could explain the accumulation of Ab42 peptide in older flies. As a measure of the rate of Ab42 fractions (two-way ANOVA). Soluble Ab42 was reduced to baseline levels following switch-off (see inset; *P,0.05 comparing +RU 7d to 2 or 7d following transfer to 2RU food and no significant difference comparing 2 or 7d following switch to 2RU to non-RU-treated controls; Tukey's HSD), whereas insoluble Ab42 was highly stable for 1 week following cessation of transgene expression (no significant difference between +RU 7d and 2 or 7d following switch to 2RU, Tukey's HSD). (C-F) Arctic Ab42 protein levels correlate with increasing duration of RU486 exposure and associate with impairments in function. (C) Arctic Ab42 mRNA levels were quantified at the end of each RU486 pulse, as indicated. Ab42 transcript was significantly increased under all RU treatment conditions compared to untreated controls (***P,0.001, n = 5, Tukey's HSD), with no significant difference (n.s.) between the varying RU486 pulse conditions. (D) Total (soluble and insoluble) Ab42 peptide was extracted using Guanidine HCl and levels were measured by ELISA at the end of each RU486 pulse (see methods). Total Arctic Ab42 protein levels increased with the length of RU486 exposure (P,0.0001, one-way ANOVA, n = 3). *P,0.05 comparing 2, 4 and 7 day pulses, but no significant difference between 7 and 14 day pulses. P,0.05 comparing 2RU to all + RU treatments conditions (Tukey's HSD). (E) Climbing ability was assessed at the indicated time-points. Arrows indicate the period of RU486 treatment for each pulse. Data were analysed by two-way ANOVA and Tukey's HSD (n = 3, number of flies per group = 39-45).
translation, 35 S-methionine incorporation was assessed in day-2 (RU 50 [2-9d]) and day-20 (RU 200 [20-27d]) UAS-Arc Ab42; elavGS flies pulsed with RU486 for 1 week. 35 S-methionine incorporation was significantly lower in RU 200 [20-27d] flies, indicating a reduced rate of translation ( Figure 2D). An increase in protein synthesis is, therefore, unlikely to explain the accumulation of Arc Ab42 peptide in older flies. However, proteasome activity was also significantly reduced in RU 200 [20-27d] compared to RU 50 [2-9d] Arc Ab42 over-expressing flies, as well as in older non-RU-treated flies ( Figure 2E) and elavGS driver flies alone under the same treatment conditions ( Figure S3), suggesting a general age-dependent reduction in proteasomal degradation which may account for the higher levels of the Ab42 peptide following induction in older flies and, at least in part, for the consequently increased toxicity.

Duration of exposure and ageing per se increase vulnerability to insoluble Arctic Ab42 peptide
To measure the vulnerability of older flies to protein toxicity per se, we next aimed to control for changes in the rate of protein turnover with age by producing standardised levels of Ab42 peptide at different ages. UAS-Arc Ab42; elavGS flies were treated with a range of RU486 doses (50-200 mM RU) for varying exposure times (1 or 2 weeks) from 5, 20 and 30 days of age. At the end of the induction period, total Ab42 peptide concentration was measured ( Figure S4). A variety of treatments resulted in equivalent levels of Ab42 peptide at different ages ( Figure S4; nonsignificant difference in Ab42 peptide levels); (1) a 2 week pulse of 100 mM RU in 5-day old flies (RU 100 [5-19d] Arctic Ab42 toxicity was assessed by measuring survival under the RU486 treatment conditions described in (1) above (Figures 3 and S5A). Equivalent levels of Ab42 peptide expression were again confirmed at the end of the 2 week induction period in day-5, day-20 and day-30-induced flies (Figur 3A). All RU-treated flies had reduced median lifespan compared to non-RU-treated controls: 27% for RU 100 [5-19d] Figure S5A). Measuring survival from 44 days only, when Ab42 peptide (which once present is not cleared) and age are equivalent in all treatment groups, young-induced flies had a shorter life expectancy compared to old-induced flies (Figure S5B). This suggests that susceptibility to Ab42 toxicity correlates with the duration of exposure to the peptide, most probably reflecting an accumulation of damage over time. To isolate the effect of age specifically on vulnerability to Arc Ab42 toxicity, further data analysis controlled for both the duration of exposure to insoluble Arc Ab42 protein and the direct effect of age on survival ( Figure 3B), with older flies expected to die sooner even in the absence of any Arc Ab42. To control for the longer exposure to Ab42 in younger flies, we re-analysed survival as function of time from the age of induction. To control for the direct effect of age on survival, we expressed survival as a percentage of that of non-RU-treated controls at each time-point from the age of induction. Relative survival following RU486 treatment was progressively lower in the two groups of older flies ( Figure 3B; P,0.05 comparing RU 100 [5-19d] [30-44d] flies, Wilcoxon rank sign test), suggesting that ageing itself also increases vulnerability to toxicity from a standard dose of Ab42 peptide. Finally, measuring relative survival of RU 100 [5-19d] flies from progressively later time-points ( Figure S5C) revealed that flies over-expressing Arctic Ab42 early in life were more severely disadvantaged in their survival compared to controls at older ages. This reduced survival with age may be attributable to an accumulation of damage with time, as indicated by the increase in toxicity with longer exposures to Ab42 peptide ( Figure S5B), or an ageing-dependent increase in susceptibility to Ab42 toxicity ( Figure 3B), or both.
We next examined whether the increased vulnerability of older flies to Ab42 toxicity was due to differences in protein aggregation with age. To standardise the dose of total Ab42 peptide at different ages, UAS-Arc Ab42; elavGS flies were induced for 1 week as described in (2) above. Soluble and insoluble proteins were then separated and Ab42 measured in each fraction at 1 and 7 days after induction, then at 7, 21 and, in the case of RU 200 [5-12d]induced flies, 35 days following switch-off ( Figure 3C). Equivalent levels of Ab42 were detected in soluble and insoluble pools immediately following RU treatment (no significant difference comparing insoluble Ab42 at 1 day versus soluble Ab42 at 1 and 7 days following RU486 treatment). Ab42 aggregated during the induction period with a significant increase in insoluble Ab42 between 1 and 7 days following RU treatment (P = 0.016 comparing flies on RU for 1 day versus 7 days). Soluble Ab42 was cleared to baseline levels within 1 week following switch-off, but insoluble protein remained stable with no significant reduction in peptide levels for several weeks following transfer to RU486-free food (P = 0.886 comparing flies switched to -RU486 food to those at the end of the 7 day induction period). No significant difference in the rate of Ab42 aggregation or clearance was observed between young and old-induced flies (P = 0.428 comparing soluble, and P = 0.258 comparing insoluble, Ab42 levels between Older flies are more vulnerable to chronic Ab42, despite a lower lifetime exposure to the peptide Finally, we examined the effects of age on Ab42 toxicity when induced chronically, to allow protein to accumulate continuously under conditions that may more closely re-capitulate the disease situation. UAS-Arctic Ab42; elavGS flies were induced from 2 days and 20 days post-eclosion and total Arctic Ab42 protein levels assayed every 5 days until death ( Figure 4A & B). Arctic Ab42 accumulated following induction in both RU 200 [2d chronic] and RU 200 [20d chronic]-induced flies, but at a slower rate and to a much lower maximum concentration in older flies ( Figure 4A). This most probably reflects reduced induction of gene Negative geotaxis decreased with increasing RU486 exposure length. Pulse lengths $4 days were significantly different from 2RU controls, whereas pulse lengths #4 days were significantly different from chronically treated (+RU) controls (P,0.05, Tukey's HSD). A 2-day pulse of RU486 did not significantly induce climbing defects compared to untreated flies, whereas a 7-day pulse did not significantly improve the reduced climbing ability compared to chronically treated +RU controls. (F) Survival decreased with increasing RU486 exposure time. Median lifespans were: 62 days for -RU (chronic), 55 days for +2d RU, 52 days for +4d RU, 45 days for +7d RU and 27 days for + RU (chronic). All survival curves were significantly different from each other (P,0.001, log rank test   Fig S1. (B) Arctic Ab42 peptide accumulates in older flies at equivalent levels of Ab42 mRNA. Total (soluble and insoluble) Ab42 peptide was extracted using GnHCl (see methods), and levels were measured by ELISA at the end of the RU486 pulse (0d) then 1, 2 and 3 weeks following the switch to RU486-free medium. Data were analysed by two-way ANOVA followed by Tukey's HSD post-hoc (n = 4   Figure S5B, 45% versus 25%, P,0.05, log rank test), possibly due to either the longer duration of exposure or increased concentration of the peptide in young-induced flies. However, as described above, we controlled for the longer exposure to Ab42 in young-induced flies, by re-calculating survival as function of the time from the age of induction, and for the effects of age on survival by expressing survival as a percentage of non-RU-treated controls at each timepoint ( Figure 4C). Remarkably, RU 200 [20d chronic]-induced flies exhibited a significantly greater reduction in relative survival than did RU 200 [2d chronic]-induced flies (P,0.05, Wilcoxon matched pairs sign rank test). Since old-induced flies were exposed to much lower maximum and cumulative amounts of Ab42 peptide, these results further support the conclusion that older flies are more vulnerable to Arctic Ab42 protein toxicity.
RU486 induction period (P = 0.016 comparing flies on RU486 for 1 day versus 7 days, P,0.05 comparing insoluble Ab42 levels in flies on RU for 7 days versus soluble Ab42 levels at days 1 and 7 following induction). Insoluble Ab42 was not significantly reduced following switch-off (P = 0.886 comparing flies at the end of the 7-day induction period with flies switched to 2RU food) and remained stable for several weeks following transfer to RU486-free food (P = 0.744 comparing RU 200 [5-12d]

Discussion
Many neurodegenerative diseases share in common the accumulation of toxic proteins and a late age of onset [15]. Our study aimed to address the mechanisms linking protein toxicity to age, by inducing expression of a transgene encoding the toxic Arctic Ab42 peptide at different ages and measuring the effects on levels of the toxic peptide and the extent of the resulting pathology.
Although the inducible GeneSwitch (GS) system has been widely used to study both ageing [26] and neurodegeneration [22], most studies have examined the effects of chronic induction. Characterisation of the dynamics of Ab42 expression using our GS-inducible AD model [21] was, therefore, required. The responses of Ab42 mRNA and peptide to cessation of induction differed, with a rapid return to baseline levels of mRNA and soluble Ab42 expression, but insoluble Ab42 peptide levels remained stable for several weeks. Importantly, switching-off transgene expression in young flies did not reverse Ab42 toxicity, because both subsequent climbing ability and survival were reduced in a dose-dependent manner, probably attributable to the persistence of the highly stable, insoluble Ab42 peptide. Studies using inducible APP mouse models of AD are congruent with these observations, because APP expression and Ab production are suppressed upon removal of the inducer doxycyline, but the mice retain a considerable Ab load for the following six months [27,28]. The high stability of insoluble Ab42 in our fly model may be related to the ability of the Arctic mutation to promote rapid aggregation [29]. The Z Ab3 Affibody, a conformation-specific Ab binding protein that prevents Ab42 aggregation, clears Ab42 and rescues toxicity when co-expressed with Arctic Ab42 in Drosophila neurons [30], suggesting that aggregated Ab42 can indeed induce neuropathological phenotypes in the fly.
We found that Ab42 toxicity is related to the concentration of the peptide, with a dose-response in negative geotaxis and lifespan, in agreement with previous studies in cells showing that toxicity is dependent upon the dose of Ab42 [31]. Importantly, when Ab42 mRNA levels were equalised at different ages of induction, Ab42 peptide accumulated to higher levels in older-induced flies, and subsequently shortened lifespan only when induced at later timepoints. This difference correlated with an alteration in protein turnover with age [24]. Measurement of 35 S-Methionine incorporation showed that general protein translation was reduced in older flies, consistent with published studies showing that the rate of protein synthesis decreases with age in a wide variety of cells, tissues, organs and organisms, including humans (reviewed in [25]). As there is no pronounced reduction in protein mass with age [32], this reduction in protein synthesis must be accompanied by a reduction in protein degradation, and indeed we found that proteasome activity declined with age. Reduced autophagy and activity of Ab42-degrading enzymes such as insulin degrading enzyme [33,34] and neprilysin [35] could also potentially contribute to a reduced rate of degradation of Ab42 peptide at later ages.
We then examined the intrinsic effect of age on vulnerability to protein toxicity by inducing a fixed level of Ab42 peptide in flies of different ages. Due to the persistence of Arctic Ab42 peptide, at the end of the oldest induction period all treatment groups expressed the same level of Ab42 at the same age. Measuring lifespan from this time-point only (44 days of age) revealed that those flies exposed to Ab42 peptide for longer had a reduced survival, suggesting that the connection between age and Ab42 toxicity, in part, reflects an accumulation of damage over time. However, in order to isolate the effect of age specifically on Ab42 toxicity we controlled for the different durations of exposure by measuring survival from the time of transgene induction, and for the direct effect of ageing on lifespan by measuring the proportional reduction in survival compared to un-induced flies of the same age. This may explain why our conclusions differ from those of an earlier study, using a similar experimental design [36]. At standardised levels of mRNA expression, Ling et al., observed a shorter maximum lifespan in young-induced flies compared to oldinduced flies, as did we, but concluded that younger flies were therefore more vulnerable to Ab42 toxicity. However, the effects of age on protein turnover and survival were not considered, and our study highlights the importance of controlling for effects of age on protein turnover, duration of exposure and the direct effect of ageing on the output variable when investigating the relationship between ageing and protein toxicity. Our analysis showed that the relative survival of older flies was more reduced than that of young flies by an equivalent dose of Ab42 peptide, indicating that ageing itself increased their vulnerability. Moreover this age-dependent increase in susceptibility appeared to be independent of the rate of Arc Ab42 aggregation, suggesting that the relationship between age and the onset of proteotoxicity is not only due to the time required for an otherwise asymptomatic protein to become pathogenic.
Chronic over-expression of Arctic Ab42 was also more toxic to older flies, despite a significantly lower duration of exposure, maximum concentration and lifetime Arctic Ab42 burden. Thus, in all instances, either the equivalent or a lower Arctic Ab42 load resulted in relatively higher levels of toxicity per time of exposure in older flies, demonstrating that ageing does increase vulnerability to protein-mediated toxicity. These results agree with other studies demonstrating that older neurons are more vulnerable to extracellularly applied Ab42 fibrils both in cell culture [31] and in the brain of rhesus monkeys [37]. Our study builds upon these findings by providing the first evidence that this is the case in vivo under physiological conditions of Ab42 expression and aggregation.
Our study suggests that, in addition to an accumulation of damage over time, ageing results in lowered degradation of a toxic protein and in increased vulnerability to its toxic effects. Hence, both the duration of exposure to toxic protein and the ageing process itself may contribute to making age the main risk factor for protein toxicity and neurodegeneration.

Lifespan analyses
Flies were raised at a standard density on SY medium in 200 mL bottles. Two days after eclosion once-mated females were transferred to experimental vials containing SY medium with or without RU486 (at indicated concentrations) at a density of 10 flies per vial. Deaths were scored and flies were transferred to fresh food 3 times a week. Data are presented as cumulative survival curves, and survival rates were compared using log-rank tests. To control for the effects of duration of exposure to Arc Ab42 peptide and the direct effects of ageing on survival, survival from the age of induction was expressed as a percentage of non-RU-treated controls of the same age. Relative survival was compared between groups using the Wilcoxon matched pairs sign rank test.

Negative Geotaxis Assays
To characterise effects of Arctic Ab42 peptide on neuronal function, climbing assays were performed according to previously published methods [38]. Climbing ability was analysed every 2-3 days post-RU486 treatment. Briefly, 15 adult flies were placed in a vertical column (25 cm long, 1.5 cm diameter) with a conic bottom end, tapped to the bottom of the column, and their subsequent climb to the top of the column was observed. Flies reaching the top and flies remaining at the bottom of the column after a 45 sec period were counted separately, and three trials were performed at 1 min intervals for each experiment. Scores recorded were the mean number of flies at the top (n top ), the mean number of flies at the bottom (n bottom ) and the total number of flies assessed (n tot ). A performance index (PI) defined as K(n tot + n top -n bottom)/ n tot ) was calculated.

Quantification of total or soluble and insoluble Ab42
Total Ab42 was extracted based on previously published methods [30]. Five fly heads were homogenised in 50 ml GnHCl extraction buffer (5 M Guanidinium HCl, 50 mM Hepes pH 7.3, protease inhibitor cocktail (Sigma, P8340) and 5 mM EDTA), centrifuged at 21,000 g for 5 min at 4uC, and cleared supernatant retained as the total fly Ab42 sample. Alternatively, soluble and insoluble pools of Ab42 were extracted using a protocol adapted from previously published methods [39]: 50 fly heads were homogenised in 50 ml tissue homogenisation buffer (250 mM sucrose, 20 mM Tris base, 1 mM EDTA, 1 mM EGTA, protease inhibitor cocktail (Sigma)) then mixed further with 50 ml diethyl acetate (DEA) buffer (0.4% DEA, 100 mM NaCl and protease inhibitor cocktail). Samples were centrifuged at 135,000 g for one hour at 4uC (Beckman Optima TM Max centrifuge, TLA120.1 rotor), and supernatant retained as the cytosolic, soluble Ab42 fraction. Pellets were further resuspended in 200 mls ice-cold formic acid (FA; 70%), and sonicated four times for 15 sec on ice. Samples were re-centrifuged at 135,000 g for one hour at 4uC, then 100 ml of supernatant diluted with 1 ml FA neutralisation buffer (1 M Tris base, 0.5M Na 2 HPO 4 , 0.05% NaN 3 ) and retained as the insoluble, formic acid-extractable Ab42 fraction.
Total, soluble or insoluble Ab42 content was measured in Arctic Ab42 flies and controls using the High Sensitivity Human Amyloid Ab42 ELISA kit (Millipore). Total Ab42 samples were diluted 1:100, and soluble versus insoluble Ab42 samples diluted 1:10 in sample/standard dilution buffer, then ELISA performed according to the manufacturers' instructions. Protein extracts were quantified using the Bradford protein assay (Bio-Rad laboratories Ltd, UK) and the amount of Ab42 in each sample expressed as a ratio of the total protein content (pmoles/g total protein).

Quantitative RT-PCR
Total RNA was extracted from 20-25 fly heads using Trizol (GIBCO) according to the manufacturers' instructions. The concentration of total RNA purified for each sample was measured using an eppendorf biophotometer. 1 mg of total RNA was then subjected to DNA digestion using DNAse I (Ambion), immediately followed by reverse transcription using the Superscript II system (Invitrogen) with oligo(dT) primers. Quantitative PCR was performed using the PRISM 7000 sequence-detection system (Applied Biosystems), SYBR Green (Molecular Probes), ROX Reference Dye (Invitrogen), and Hot Star Taq (Qiagen, Valencia, CA) by following manufacturers' instructions. Each sample was analysed in triplicate with both target gene (Arctic Ab42) and reference gene (RP49, eIF-1A and aTubulin84) primers in parallel. The primers for the Ab transgenes were directed to the 59 end and 39 end of the Ab coding sequence: forward GATCCTTCTCCTGCTAACC, reverse CACCAT-CAAGCCAATAATCG. The reference gene primers were as follows: RP49 forward ATGACCATCCGCCCAGCATCAGG and reverse ATCTCGCCGCAGTAAACG; eIF-1A forward ATCAGCTCCGAGGATGACGC and reverse GCCGAGACA-GACGTTCCAGA; aTubulin84 forward TGGGCCCGTC-TGGACCACAA and reverse TCGCCGTCACCGGAGTCCAT [40].

Proteasome activity
Fly heads were homogenized in 25 mM Tris, pH 7.5 and protein content determined by Bradford assay. Chymotrypsin-like peptidase activity of the proteasome was assayed using the fluorogenic peptide substrate Succinyl-Leu-Leu-Val-Tyr-amidomethylcoumarin (LLVY-AMC), based on a previously published protocol [41]. 20 mg of crude fly head homogenate total protein was incubated at 37uC with 25 mM LLVY-AMC in a final volume of 200 mLs. Enzymatic kinetics were conducted in a temperaturecontrolled microplate fluorimeter (Tecan Infinite M200), at excitation/emission wavelengths of 360/460 nm, measuring fluorescence every two minutes for 30 min. Proteasome activity was determined as the slope of AMC accumulation over time per mg of total protein (pmoles/min/mg).

S-methionine incorporation
Protein translation was measured in fly heads using a method adapted from [42]. Standard SY medium was first supplemented with 100mCi 35 S methionine/ml of food (American Radiolabeled Chemicals 1 mCi/37MBq ARS0104A). 15 flies were transferred to each vial containing 1 ml radioactive SY medium. After threehours of feeding, flies were transferred to non-radioactive SY for 30 min in order to purge any undigested radioactive food from the intestines. Flies that were in contact with the radioactive food for 1 minute were used as a background control. Flies were then decapitated using liquid nitrogen and the heads homogenized in 1% SDS and heated for 5 minutes at 95uC. Samples were then centrifuged twice for 5 minutes at 16,000 g. Proteins were precipitated by the addition of the same volume of 20% cold TCA (10% TCA final concentration) and incubated for 15 minutes on ice. Samples were then centrifuged at 16,000 g for 15 minutes, the pellet washed twice with acetone and then resuspended in 200 ml of 4 M guanidine-HCl.
Samples (100 mls) were mixed with 3 mls of scintillant (Fluoran-Safe 2, BDH) and radioactivity counted in a liquid scintillation analyzer (TriCarb 2800TR, Perkin Elmer), with appropriate quench corrections. SDS-homogenates, prior to TCA precipitation, were also sampled and analysed as a measure of the total radioactivity (incorporated and un-incorporated) present. Total protein for each sample was determined by Bradford assay and a translation index was calculated as follows: (TCA protein cpm/ total cpm)/mg protein per sample.