Influence of non-steroidal anti-inflammatory drugs on Drosophila melanogaster longevity.

Most age-related diseases and aging itself are associated with chronic inflammation. Thus pharmacological inhibition of inflammatory processes may be effective antiaging strategy. In this study we demonstrated that treatment of Drosophila melanogaster with 10 non-steroidal anti-inflammatory drugs (NSAIDs: CAY10404, aspirin, APHS, SC-560, NS-398, SC-58125, valeroyl salicylate, trans-resveratrol, valdecoxib, licofelone) leads to extension of lifespan, delays age-dependent decline of locomotor activity and increases stress resistance. The effect of the lifespan increase was associated with decrease of fecundity. Depending on the concentration, NSAIDs demonstrated both anti- and pro-oxidant properties in Drosophila tissues. However, we failed to identify clear correlation between antioxidant properties of NSAIDs and their pro-longevity effects. The lifespan extending effects of APHS, SC-58125, valeroyl salicylate, trans-resveratrol, valdecoxib, and licofelone were more pronounced in males, valdecoxib and aspirin - in females. We demonstrated that lifespan extension effect of NSAIDs was abolished in flies with defective genes involved in Pkh2-ypk1-lem3-tat2 pathway.


INTRODUCTION
Chronic inflammation is one of the main processes that causes disruption of normal functioning of tissues and contributes to age-associated diseases and aging [1,2]. There is a popular industry term «inflammaging», which is used to describe aging associated with the development of chronic inflammation [3]. Inflammation is accompanied by increase of the activity of pro-inflammatory pathways with age [4,5]. Inflammatory cause of different age diseases: metabolic syndrome, Alzheimer's disease and Huntington's disease [6][7][8]. Pharmacological and genetic inhibition of inflammatory processes is considered as effective and proven anti-aging strategy [9].
However, according to growing number of recent publications, NSAIDs have multiple molecular targets, and the range of their activity may be broader than COX inhibition [17,20].
Many studies have experimentally demonstrated that NSAIDs have antioxidant activity via anti-radical www.impactjournals.com/oncotarget activity and membrane-stabilizing action. Anti-radical activity of NSAIDs is mediated by free radical scavenging and antioxidant enzyme activation [21]. NSAIDs exhibit antioxidant activity on model membranes [22], in cells [23,24], as well as on the organismal level [25,26].
Anti-inflammatory effects of rapamycin and resveratrol may be associated with pharmacological stimulation of autophagic activity via mTOR inhibition and AMPK activation, respectively. Autophagy prevents the activation of inflammasomes and induction of inflammatory responses [27,28].
The purpose of this study is to explore the geroprotective properties of NSAIDs, their impact on lifespan, life quality (locomotor activity), fecundity, and stress resistance in Drosophila model, reveal toxicity, the antioxidant activity and membrane-protective activity of these compounds using in vitro and in vivo models. In addition it is planned to clarify the role of Pkh2-ypk1-lem3-tat2 signaling pathway in the formation of geroprotective effects.

Effects of NSAIDs on longevity
While exposing the flies to test substances, we observed a significant increase in median survival age and 90% mortality in males and females ( Figure 1).
When exposed to NSAIDs in females at concentration of 0.05 μM we observed increases in median lifespan by 5.9-12%. Also SC-560, trans-resveratrol and valdecoxib at this concentration increased the age of 90% mortality by 3.4-6.9%. Aspirin and NS-398 at concentration 0.5 μM extended median lifespan by 7.7 and 9.3% respectively, and the age of 90% mortality by 1.6 and 6.3% respectively. APHS at concentration 0.5 μM increase median lifespan by 9.6%. The most significant effects on the longevity of females we observed when exposed to NSAIDs at concentration of 1 μM. Aspirin at this concentration extended median lifespan by 32.7%. Trans-resveratrol, CAY10404, APHS, NS-398, valdecoxib and licofelone at concentration of 1 μM increased median lifespan by 9.6-15.4%. All test substances except valeryl salicylate increased the age of 90% mortality by 4.1-13.7% ( Figure 1C, 1D).
Some of the substances we tested reduced lifespan of flies. SC-560 at a concentration of 0.5 μM significantly reduced the median lifespan of males by 6% and decreased the age of 90% mortality in males by 8.3%, and in females by 3.2% Valeryl salicylate at a concentration of 1 μM reduced the age of 90% mortality of females by 4.1%. CAY10404 at a concentration of 0.5 μM reduced the age of 90% mortality of female by 4.8% ( Figure 1A, 1B, 1C, 1D).

Effects of NSAIDs on fly feeding
According to the feeding assay flies ate less paste containing SC-560 (37.3% lower level of fluorescence), NS-398 (43.1% lower), SC-58125 (44% lower). On the contrary, we observed a higher level of fluorescence in the flies, when aspirin (58.6%) was added into the paste. In other experiment variants we observed no change in the amount of food consumed. Regression analysis did not reveal statistically significant correlation between the amount of food consumed and the changes in lifespan (Supplemental Figure 1).

Effects NSAIDs on locomotor activity and fertility
Analysis of locomotor activity showed an increase activity in the test for negative geotaxis of males in the last measuring point on the 45th day. Females showed no significant changes in the negative geotaxis test when exposed to NSAIDs.
We often observed a decrease in spontaneous activity of males and females in the first half of life in 5-25 days. However, in 35 and 45 days, spontaneous activity, in most embodiments a exposure was equal to control or higher than control values. Likofelon had the strongest www.impactjournals.com/oncotarget effect on the conservation and improvement of locomotor function with age ( Figure 2).
Test substances have different effects on female fertility. We observed an increase in fertility in the middle of the measurement time when exposed to almost all substances. However, we did not find the extension of the reproductive period females. It may be noted that NSAIDs at concentration of 0.05 μM and especially the 0.5 μM were more positive effects on female fertility. In contrast, NSAIDs at concentration of 1 μM had more negative effects on fertility. Some substances with most geroprotective properties often had a fairly strong negative effect on fertility. It was aspirin, APHS and NS-398 ( Figure 3).

Effects of NSAIDs on stress resistance and level of lipid peroxidation products
Preliminary experiments have shown that all tested compounds at a concentration of 0.5 µМ were not toxic to Data are represented as percentage changes in comparison with control group. A. Males median lifespan. B. Males age of 90% mortality. C. Females median lifespan. D. Females age of 90% mortality. *p < 0.001, **p < 0.01, ***p < 0.05, Gehan-Breslow-Wilcoxon and Mantel-Cox tests for median lifespan; Wang-Allison test for age of 90% mortality. red blood cells, the hemolysis level in the presence of the compounds relative to spontaneous hemolysis was 1.17.
The results of the comparative evaluation of the membrane-protective activity are shown in Figure  4A. Nine of the ten compounds studied were able to protect cells from death under stress of acute oxidative stress. Resveratrol proved to be the most active one in this respect. Only licofelone, a complex inhibitor of cyclooxygenase (COX) and lipoxygenase did not possess the membrane-protective activity; however, it statistically significantly increased the degree of the induced hemolysis. Relative values of the membrane-protective and antioxidant activity have been closely correlated (Rs = 0.782, p = 0.008, Figure 4B).
Biologically active compounds can have an impact on organism's lifespan, indirectly, through a complex of different mechanisms. Since a vast majority of the compounds investigated are not only COX inhibitors, they also have a powerful antioxidant activity, it is possible that this property could also affect the results obtained when evaluating the parameters of lifespan of model animals. In order to analyze the contribution of the antioxidant and membrane-protectiveactivity of NSAIDs on lifetime variation in Drosophila maintained on a medium supplemented with test compounds at 0.05, 0.5 and 1 µМ concentrations, an appropriate regression analysis was performed. However, none of the cases were statistically significant for an association between the antioxidant/ membrane-protective activity of the compounds and the parameters characterizing the Drosophila lifespan (median lifespan and age of 90% mortality). Thus, among all compounds studied resveratrol possessed the most significant antioxidant and membrane-protective activity; however, supplementing the nutrient medium with this compound did not lead to a dramatic increase in the lifespan of model animals.
An analysis of the secondary products of lipid peroxidation in Drosophila homogenates has demonstrated that at both concentrations (0.05 and 1μM) the test compounds have a more considerable effect on males. At the same time males are characterized by a higher level of lipid peroxidation. Furthermore, the effect of the compounds significantly depended on both the concentration and the sex of animals. Previously, we observed a similar phenomenon in relation to other biologically active substances (BAS) in experiments conducted on laboratory mice [34].
For example, maintenance of flies on the nutrient medium supplemented with NSAIDsat a concentration of 1 µM in most cases resulted in a statistically significant decrease in the LPO level in male organisms, while in females half of the compounds caused an increase in the secondary products of lipid peroxidation ( Figure 4C, 4D).
A use of lower concentrations of NSAIDs in the nutrient medium (0.05 µM) caused an increase in the intensity of lipid peroxidation in male tissues and did not significantly affect females ( Figure 4C, 4D). Regression analysis showed no association between the intensity of lipid peroxidation in Drosophila tissues and the parameters characterizing their lifespan.
Thus, despite the fact that the vast majority of the NSAIDs examined in this study have prominent antioxidant properties, other mechanisms of their pharmacological activity are implemented as well. This underlies an ambiguous effect on such an integral indicator, as lifespan.

Stress resistance analysis
It was demonstrated that in males NSAIDs at a concentration of 0.05 µM significantly improved survival ratio upon exposure to paraquat, and at a concentration of 1 µM they significantly improved it under heat shock. There was no significant increase in survival during starvation. Upon exposure to CAY10404, aspirin, APHS,

The role of Pkh2-ypk1-lem3-tat2 pathway
In order to understand the mechanisms of geroprotective properties of NSAIDs we studied the effects of aspirin, valdecoxib, and NS-398 on the lifespan of Drosophila defective in Pkh2-ypk1-lem3-tat2 signaling pathway.
We observed a decrease in lifespan in flies with activated RNA interference of genes ypk1/S6k, Pkh2/Pdk1 and lem3/CG8679 upon exposure to aspirin, valdecoxib, NS-398 at a concentration of 1 μM (Figures 6, 7, 8). However, inActGS>ypk1/S6kRNAi#2 males exposure to NS-398 induced a slight extension in lifespan. In ActGS>lem3/CG8679RNAi females we observed an increase in the survival ratio in the second half of life. The treatment with NSAIDs of tat-2/CG14741 knockout males were characterized by a decreased lifespan, while females, had an extended lifespan.

DISCUSSION
Thus we investigated the effects of 10 NSAIDs at concentrations of 0.05, 0.5 and 1 μM on Drosophila lifespan, life quality (locomotor activity), fecundity, and stress resistance (oxidative stress, heat shock and starvation).
According to the feeding assay analysis flies consumed less food containing some substances (SC-560, NS-398, SC-58125). However, we did not observe any correlation between lifespan and the level of food consumption (Supplemental Figure 1).
In addition, NSAIDs delayed age-dependent decline in locomotor activity. This effect may be associated with neuroprotective action of NSAIDs [14,15]. As it was shown previously, aspirin reduces age-associated functional declines in C. elegans also [36].
We observed a decrease in female fecundity. It may be related to the fact that NSAIDs inhibits the activity of Pxt, a COX-like facilitator of follicle maturationin Drosophila [40]. Therefore, the effect of the lifespan increase may be associated with decrease offecundity.
Attention is drawn to the similarity of life-extending effects of various anti-inflammatory drugs (Figure 1). This suggests that non-specific component (hormesis) may also be implicated in triggering these effects [41]. Hormesis is usually associated with the activation of various cellular stress-resistance mechanisms, such as heat shock proteins, antioxidant enzymes, DNA repair mechanisms, immune response, selection of unfit cells and may increase both longevity and stress resistance [42][43][44][45][46][47][48][49].
Therefore, we tested whether NSAIDs might also increase Drosophila resistance to oxidative stress, heat shock and starvation. We found that NSAIDs increased resistance to various stresses. Increased resistance to heat shock and starvation may be related to the ability of NSAIDs to inhibit the components of the insulin/IGF-1 signaling [11,50,51].
Depending on the concentration and the experimental model NSAIDs demonstrate both anti-and pro-oxidant properties. The antioxidant properties were revealed in the erythrocytes model and Drosophila tissues at a concentrations of 0.5 μM and 1 μM, accordingly. However, at a concentration of 0.05 μM we observed prooxidant properties of NSAIDs in Drosophila males. Thus we failed to find correlation between antioxidant properties of NSAIDs and their antiaging action in Drosophila.
Thus, our data support the new paradigm that ROS is not involved in aging under physiological, nonstressed  conditions [52]. Numerous experiments on animals also showed that manipulation of antioxidant gene expression often has little effect on life span, and that levels of mitochondrial ROS do not limit it and moreover an increased ROS production correlated with extended life span [53,54]. Clinical trials demonstrated that current antioxidants do not prolong life and reduction of ROS by antioxidants can even shorten life span [54]. Perhaps this is due to ROS important physiological roles, including regulatory role in cellular signaling pathways, and elimination of ROS may not be favorable for the organism.
Previous studies demonstrated that antioxidant and neuroprotective effects of NSAIDs are associated with inhibition of COX and reduction in the production of reactive oxygen species (ROS) in the metabolism of arachidonic acid, a decrease of the prostaglandin synthesis and a decline of β-amyloid accumulation [8,14,15]. Licofelone, due to its conformational similarity with arachidonic acid, is capable of binding to the active sites of both classes of enzymes COX-1 and COX-2, blocking their catalytic activity [55], and normalizing the parameters of lipid peroxidation in brain tissues with an experimental Huntington's syndrome [26].
At the same time, NSAIDs are known to have intrinsic antioxidant activity [22,24,56]. For example, a complex biological activity of resveratrol is associated not only with its ability to interact with multiple molecular targets, in particular, cyclooxygenase, but also with a high antioxidant activity due to both a direct elimination of ROS and inhibition of enzymes involved in their formation, as well as enhancement of activity of antioxidant protection enzymes [57]. It has been experimentally demonstrated that resveratrol possesses a high antioxidant activity on model membranes [22], in cells [24,58,59], as well as in model organisms [25,60]. Furthermore, many of the effects of resveratrol are due to an action of its metabolites, and even very low doses of resveratrol can affect an organism through an indirect action [57].
Thus, the antioxidant effect of NSAIDs may be less pronounced on the level of lipid peroxidation due to the absence of COX in Drosophila. However intrinsic antioxidant activity of NSAIDs may protect Drosophila from acute toxicity of paraquat.
In our study, RNA interference of gene Pkh2/Pdk1 led to an increase in longevity of Drosophila females. Our data is consistent with the effects of pharmacological inhibition of Pkh2/PDK-1 in nematode [11]. Simultaneous RNAi of Pkh2/Pdk 1and exposure to NSAIDs reduced the positive effect of RNA interference on female longevity. It should be noted that the effects of NSAIDs on flies with other components of Pkh2-ypk1-lem3-tat2 signaling pathway being affected also led to a decrease in longevity. Thus the lifespan extending effects of NSAIDs in Drosophila model is mediated by Pkh2-ypk1-lem3-tat2 signaling pathway.
Thus, we demonstrated the ability of 10 NSAIDs (CAY10404, aspirin, APHS, SC-560, NS-398, SC-58125, valeroyl salicylate, trans-resveratrol, valdecoxib, licofelone) to extend lifespan and increase resistance to stress in Drosophila accompanied by positive effects on locomotor activity. The lifespan extending effects of APHS, SC-58125, valeroyl salicylate, trans-resveratrol, valdecoxib, and licofelone was more pronounced in males, valdecoxib and aspirin-in females. The effect of the lifespan increase was associated with decrease of fecundity. No correlation was found between the antioxidant properties of NSAIDs and lifespan-extending effect. At the same time, increasing of the survival rate of NSAIDs-treated flies under the influence of paraquat can be associated with the antioxidant activity of substances. The lifespan extension effect of NSAIDs was abolished in flies that are defective in genes of Pkh2-ypk1-lem3-tat2 pathway.

Drosophila melanogaster strains
Strains obtained from the Bloomington Drosophila Stock Center were used in this study.
Standard laboratory wild-type strain Canton-S (stock #1) was used to test geroprotective properties of NSAIDs.

Treatment with substances
We greased fly medium by paste of hydrolyzed yeast containing one of the substances. Control untreated animals were fed by yeast past without substances. To make the hydrolyz at yeast were boiled in water bath for 30 minutes. To prepare the 100 ml of paste 50 g of dry yeast per 60 mL of water were used.
Flies were treated throughout their whole lives with the substances as follows: aspirin (

Feeding assay
To evaluate the effects of NSAIDs on the food intake we used the fluorescein dye (Sigma-Aldrich, USA) as the food tracer. To perform the test, three-to-four day-old females were placed on the paste (with or without NSAIDs in concentration 1 μM) containing 50 μM fluorescein. After 1 h of feeding, the flies were anesthetised and placed in liquid nitrogen for 20 sec. The frozen flies were then briefly vortexed in tube to separate fly heads from the bodies. The fly bodies (5 flies per 1 mL) were then homogenized in distilled water, and the homogenates were centrifuged 2 min at 15,000g. Then 0.8 mL of supernatant was transferred to a new tube, the volume was adjusted to 1.5 mL, and centrifugation was repeated. Before the measurements, the supernatant was diluted with distilled water 1:1. The fluorescence level in resulting solution was then measured on spectrofluorimeter Fluorat-02-Panorama (Lumex, Russia) with a fluorescence excitation of 480 nm and an emission registration of 521 nm. Nine independent replicates were performed for each experiment. The level of food intake was estimated on the base of average values of fluorescence intensity. Regression analysis was performed to assess the correlation between the amount of food consumed and the lifespan. To assess the statistical significance of differences fluorescence level was carried out using the t-test.

Lifespan analysis
To analyze the lifespan, the imagoes of the same age of both sexes were selected. Control and experimental flies were maintained at 25±0.5°C under a 12h light/12h dark cycle and at densities of 30 same sex and age flies per vial containing sugar-yeast medium covered with a yeast paste. Flies were transferred to fresh medium 2 times per week. The number of dead flies was counted daily. The longevity of males and females was analyzed separately. Analysis was carried out in three replications, with 80-120 flies in each.

Locomotor activity analysis
In order to measure locomotor activity a hardware and software complex «Drosophila population monitor» (TriKinetics Inc., USA) was used. Spontaneous locomotor activity was measured (for 3 min), and a negative geotaxis test was performed. The flies were shaken off to the bottom of a tube, and movement was measured for 20 sec in three replications. When evaluating the spontaneous activity, an integral activity of flies for 3 min was registered. For the negative geotaxis test, an arithmetical mean for three replications was calculated to offset an impact of random factors. The measurements in each variant were carried out, as long as there was the sufficient number of living flies to perform the analysis, which was 30 pieces.

Female fecundity analysis
Fecundity was evaluated by the mean number of eggs and pupae per female. For this purpose, the groups consisting of 10 males and 10 females with the same age were placed into vials with fresh medium. After one day, the flies were transferred to a new medium, and the number of eggs laid was counted. After 10 days pupae formed were counted. A mean fecundity was calculated as a ratio between the number of eggs or pupae and the number of females in a group.

Stress resistance analysis
To study an effect of heat shock, the flies were placed into an incubator at 35° C.
To study an effect of the oxidative stress, a solution of 20 mM of the pro-oxidant paraquat dissolved in 5% sucrose was used. This oxidative stress medium was applied to a filter paper and placed into vials instead of the nutrient medium.
To study the effect of starvation on the flies they were placed into vials with 2% agar.

Toxicity, antioxidant activity and membraneprotective activity study
To study the toxicity, antioxidant activity and membrane-protective activity (in vitro) of the compounds, an erythrocyte suspension of laboratory mice blood in phosphate buffered saline (PBS, pH 7.4) was used. The toxicity of the compounds was evaluated by their ability to induce the death of erythrocytes (hemolysis). The solution of compounds dissolved in ethanol were added to the erythrocyte suspension at a final concentration of 0.5 μM and incubated at 37° C for 5 h in a thermostatically controlled shaker Biosan ES-20 (Latvia), control samples also contained ethanol. The membrane-protective and antioxidant activities were determined by a degree of inhibition of induced hemolysis and inhibition of accumulation of secondary products of lipid peroxidation in erythrocytes, respectively. For this purpose, after adding of the solutions of the different compounds (final concentration of 0.5 μM) to the erythrocyte suspension the hemolysis was induced by the addition of H 2 O 2 solution (1.8 mM). Then this reaction mixture was shaken gently while being incubated for 5 h, at 37°C. Every hour an aliquot of the incubation medium was taken and centrifuged for 5 min. The hemolysis degree was determined based on the hemoglobin content in the supernatant using a spectrophotometer Genesys 20 (Thermo Scientific, USA) at λ 524 nm, the percentage of hemolysis was calculated in relation to the total hemolysis of a sample [61][62][63]. Each experiment was performed in five parallels and two replications.

Analysis of the level of secondary products of lipid peroxidation (LPO)
The content of the secondary products of lipid peroxidation from a reaction with 2-thiobarbituric acid (TBA-RS) was determined using the spectrophotometer at λ = 532 nm [64] in a modification [65,66]. To prevent the oxidation of lipids in the process of analysis a butylated hydroxyltoluene (BHT) solution in ethanol (0.01%) was added to the samples. Samples without addition of the TBA solution were used as an external control in order to avoid artifacts of measuring LPO due to the Drosophila eye color. A concentration of TBA-RS products was calculated taking into account the extinction coefficient (1.56 × 10 5 M -1 cm -1 ) and expressed in nmol/mg protein or nmol/mg of fly mass.

Statistical analysis
Lifespan analyses were carried out using nonparametric methods were utilized. Survival was assessed using Kaplan-Meier survival curves [67]. A median lifespan and age of 90% mortality were calculated. When comparing the survival functions, a modified Kolmogorov-Smirnov's test was applied [68]. To assess the significance of differences for the median lifespan a Gehan-Breslow-Wilcoxon test [69] and Mantel-Cox test [70] were used. To assess a statistical significance of differences in the age of 90% mortality a Wang-Allison test was used [71]. The Kaplan-Meier curves were plotted using a program Statistica, version 6.1 (StatSoft, Inc., USA), the calculation of the lifespan parameters and their statistical analysis were performed in a statistical programming environment R [72].
To assess the statistical significance of differences in resistance to stress-factors, the Fisher's exact test was used [73,74]. Statistical analyses of the data were performed using OASIS: Online Application for the Survival Analysis of Lifespan Assays [74].
Statistical significance of differences in the fecundity and locomotor activity was evaluated using the χ2 criterion [75].
To process the locomotor activity and fecundity data Microsoft Excel 2010 (Microsoft, USA) was used.
Statistical significance of differences in toxicity, antioxidant and membrane-protecting activity of NSAIDs (in vitro erythrocytes model) and content of LPO secondary products in flies was assessed by a nonparametric Mann-Whitney test [76]. Significance was set at p < 0.05. Regression analyses were also performed to compare the changes of various parameters tested for using model objects. Spearman's rank correlation coefficient (Rs) was calculated. Analyses were performed by applying software packages Microsoft Office Excel 2007 and Statistica 6.0.