Blueberry polyphenols increase lifespan and thermotolerance in Caenorhabditis elegans

Summary The beneficial effects of polyphenol compounds in fruits and vegetables are mainly extrapolated from in vitro studies or short-term dietary supplementation studies. Due to cost and duration, relatively little is known about whether dietary polyphenols are beneficial in whole animals, particularly with respect to aging. To address this question, we examined the effects of blueberry polyphenols on lifespan and aging of the nematode, Caenorhabditis elegans, a useful organism for such a study. We report that a complex mixture of blue-berry polyphenols increased lifespan and slowed aging-related declines in C. elegans. We also found that these benefits did not just reflect antioxidant activity in these compounds. For instance, blueberry treatment increased survival during acute heat stress, but was not protective against acute oxidative stress. The blueberry extract consists of three major fractions that all contain antioxidant activity. However, only one fraction, enriched in proanthocyanidin compounds, increased C. elegans lifespan and thermotolerance. To further determine how polyphenols prolonged C. elegans lifespan, we analyzed the genetic requirements for these effects. Prolonged lifespan from this treatment required the presence of a CaMKII pathway that mediates osmotic stress resistance, though not other pathways that affect stress resistance and longevity. In conclusion, polyphenolic compounds in blueberries had robust and reproducible benefits during aging that were separable from antioxidant effects.


Introduction
Plants synthesize an array of chemical compounds that are not involved in their primary metabolism. These 'secondary compounds' instead serve a variety of ecological functions, ultimately to enhance the plant's survival during stress (Winkel-Shirley, 2002). In addition, these compounds may be responsible for the beneficial effects of fruits and vegetables on an array of healthrelated measures (Liu, 2003). Previous research suggested that the combination of antioxidant/anti-inflammatory polyphenol compounds found in fruits and vegetables may show efficacy in reversing aging (Joseph et al ., 2005). In particular, 8-week dietary supplementation with spinach, strawberry or blueberry (BB) extracts was effective in reversing age-related deficits in neuronal function and behavior in aged (19 month) F344 rats (Joseph et al ., 1999). However, only the BB-supplemented group exhibited improved performance on tests of motor function. Specifically, the BB-supplemented group displayed improved performance on two motor tests that rely on balance and coordination, rod walking and the accelerating rotarod, while none of the other supplemented groups differed from controls on these tasks. A subsequent study using a BB-supplemented diet replicated these findings (Youdim et al ., 2000). Results from both studies indicated that the significant effects of BB on behavior were due to a multiplicity of actions, in addition to those involving antioxidant and anti-inflammatory activity.
To further investigate the effects of blueberries on parameters related to aging, we designed a study to determine whether BB polyphenols can delay aging and prolong lifespan in a whole organism. For these studies, we required an organism with relatively short lifespan that could be assayed reproducibly and robustly, and for which the genetic and environmental factors affecting lifespan were well defined. The experimental organism that could best accommodate these requirements was the nematode, Caenorhabditis elegans , which has become a popular model for studying aging and longevity, due to its short 2-to 3-week lifespan, rapid generation time and experimental flexibility (Guarente & Kenyon, 2000).
Studies have shown that specific genetic and environmental factors influence aging and lifespan of C. elegans (Gems & Riddle, 2000;Garigan et al ., 2002;Herndon et al ., 2002). In addition, aspects of aging are similar between nematodes and mammals, including humans. For instance, sarcopenia, the loss of muscle mass, is a major feature of aging in humans and C. elegans and contributes to aging-related behavioral declines in both (Herndon et al ., 2002;Glenn et al ., 2004). Second, calorie restriction, the only known intervention that successfully extends lifespan in mammals, can prolong C. elegans lifespan (Lakowski & Hekimi, 1998). Finally, oxidative stress appears to be a major factor limiting lifespan in both C. elegans and humans (Larsen, 1993;Finkel & Holbrook, 2000). These findings show that studies of aging in C. elegans provide useful stepping stones for identifying genes and compounds that can prolong lifespan in humans.
Here, we report that treatment with total BB polyphenols, or a proanthocyanidin (PAC)-enriched fraction from BB, could prolong adult lifespan and delay aging in C. elegans . These treatments increased thermotolerance, but did not improve resistance to oxidative stress. BB treatment was correlated with reduced basal levels of hsp mRNA, indicating these compounds had direct or indirect effects on gene expression. Our genetic analysis indicated that BB polyphenols may act through a CaMKII signaling pathway to affect C. elegans lifespan. These findings show that natural compounds from blueberries can provide antiaging benefits in vivo in an intact organism.

Blueberry polyphenols extend C. elegans lifespan
Adult wild-type animals grown under our standard laboratory conditions at 25 ° C have a mean lifespan of 12.7 days and average maximum lifespan of 19.7 days. On media containing either crude BB extract ( Vaccinium angustifolium ) or a C18 column fraction containing their bulk polyphenols, mean lifespan of wild-type animals was lengthened by 28% (Fig. 1A, Table 1). Maximum lifespan was also increased by an average of 14% in all trials ( P = 0.007; 0 µ g mL − 1 , n = 17 trials; 200 µ g mL − 1 , n = 16 trials). Similar results were obtained using a crude extract of a different species of blueberry ( Vaccinium asheii ) (not shown). Lifespan in C. elegans is affected by temperature (Gems et al ., 1998). Animals grown at 25 ° C have shorter adult lifespan than animals grown at 15 ° C, although both temperatures are considered within the normal range for C. elegans . The effects of BB polyphenols on C. elegans lifespan were also temperature dependent. BB treatment prolonged lifespan in animals grown at 25 ° C or 20 ° C, but no significant benefits were observed upon lifespan at 15 ° C (Table 1).
By several measures, BB treatment slowed aging in C. elegans , rather than simply improving survival at old age. One measure of aging in C. elegans is the speed of pharynx contraction, or pumping (Bolanowski et al ., 1981;Huang et al ., 2004). Young adult animals pump 250 -300 times per minute and pumping declines gradually with increasing age. BB treatment was associated with higher pumping rates at adult days 8 and 10 (Fig. 1B).
BB treatment also delayed the accumulation of aging-related cellular damage. One marker for cellular damage during aging is the intracellular level of lipofuscin, autofluorescent material that accumulates in aging cells. (Brunk & Terman, 2002). Lipofuscin levels increase with aging in many organisms, including C. elegans (Hosokawa et al. , 1994). We examined lipofuscin levels in the intestines of control and BB-treated adult day 16 animals. Consistent with the expectation that BB polyphenols could slow aging, intestinal lipofuscin levels were reduced 20%  in BB-treated animals, compared with controls ( Fig. 2A,B). 4-Hydroxynonenol (4-HNE) is a lipid peroxidation product that accumulates with aging in many animals (Chiarpotto et al. , 1995). In C. elegans , a correlation between 4-HNE levels and aging has not yet been demonstrated, although levels of 4-HNE-modified protein have been correlated with lifespan (Ayyadevara et al. , 2005). BB treatment also reduced levels of 4-HNE in 14-day adults (Fig. 2C,D).

BB treatment delayed aging-related increase in heatshock protein mRNA levels
We next examined the effects of BB treatment on a transcriptional marker of aging. Several independent analyses of gene expression during aging in C. elegans have revealed that heatshock protein mRNA levels increase with age (Lund et al ., 2002;Golden & Melov, 2004). Using real-time PCR, we confirmed that expression of hsp-12.6, -16.1, -16.49 and -70 increased markedly in control animals between days 0 and 4 of adulthood under our laboratory conditions (Fig. 3). BB treatment blocked this increase and mRNA levels for these hsp s remained at basal levels through to adult day 4. From this experiment, we conclude that BB polyphenols delayed changes in hsp expression that are normally associated with aging in C. elegans , consistent with the observation that BB treatment delayed morphological features of aging.

Effects of blueberry polyphenols on intrinsic stress resistance
One possible explanation for the beneficial effects of BB polyphenols on aging in C. elegans is that these compounds were able to increase cellular stress resistance. In several studies, increased longevity was closely associated with improved survival under conditions of heat or oxidative stress (Lithgow et al., 1995;Muñoz & Riddle, 2003). To test this possibility, we examined stress resistance of BB-treated animals. Thermotolerance was increased significantly by BB treatment. In wild-type animals, treatment with BB polyphenols was correlated with a 2.5-fold increase in 16-h survival at 35 °C (Fig. 5A). Lifespan and thermotolerance were linked, as thermotolerance was increased by treatment with the PAC-enriched fraction, but not the ATC-enriched fraction nor CA (Fig. 5D). Interestingly, BB treatment did not improve survival under mild to severe oxidative stress. Resistance to oxidative stress was examined by exposing animals to hydrogen peroxide or paraquat, an intracellular free-radical-generating compound. BB treatment was not associated with any increase in survival in the presence of hydrogen peroxide or paraquat (Fig. 5B,C). To further examine whether BB polyphenols could protect against acute oxidative stress, we examined mev-1(kn1) animals, which have a mutation in the cytochrome b large subunit of mitochondrial complex II (Ishii et al., 1998). This mutation results in overproduction of superoxide and increased oxidative stress, along with accelerated aging and reduced lifespan (Hosokawa et al., 1994;Senoo-Matsuda et al., 2001). Consistent with the finding that BB did not protect against extrinsic oxidative stress, BB treatment also did not significantly affect lifespan of mev-1(kn1) animals ( Table 1).

Effects of BB treatment on expression of stressinducible genes
We next considered the possibility that BB treatment acted as a mild stressor that induced expression of protective enzymes and this increased gene expression delayed aging (Lithgow et al., 1995). To investigate this possibility, we examined the effects of BB treatment on the expression of several stress-inducible genes. Although BB treatment improved thermotolerance, the mRNA levels for a number of stress-inducible genes were not altered by BB treatment in any consistent way (Supplementary  Table S1). In this analysis, we were careful to select stress-inducible genes regulated by several different pathways. In addition, the earlier experiment showed that mRNA levels of inducible hsps were not increased in BB-treated animals (Fig. 3), nor was BB treatment associated with consistently greater hsp mRNA induction following heat-shock compared with untreated controls (Supplementary Table S2). Together, these gene expression analyses show that BB treatment did not cause a general induction of stress-response genes under normal growth conditions, but appeared to increase overall health, which may have promoted Fig. 4 A PAC-enriched fraction of BB contained components sufficient to extend Caenorhabditis elegans lifespan. The total blueberry polyphenols were fractionated by C18 and Sephadex LH20 to produce fractions enriched in anthocyanins (ATC), proanthocyanidins (PAC) or chlorogenic acid (CA). Purchased, purified CA was used for CA experiments. Each fraction was assayed for effects on C. elegans longevity at a concentration of 67 µg mL −1 , similar to that in the complete extract. (A) Only the PAC-enriched fraction (orange) prolonged lifespan. No longevity benefits were observed for the ATC-enriched fraction (purple) or purified CA (blue); 0 µg mL −1 , n = 97 animals, 2 trials; ATC, n = 99, 2; CA, n = 91, 2; PAC, n = 88, 2. Complete statistics for PAC trials are presented in Table 2. (B) The reconstituted extract, produced by mixing equal mass ratios of fractions in (A) also conferred the same lifespan extension as the starting mix of polyphenols (remix, n = 40 animals, 1 trial; start, n = 98, 2).

Blueberry polyphenols do not appear to prolong lifespan through antimicrobial effects
One contributor to late-age mortality in C. elegans is the detrimental effect of the bacterial food source (Gems & Riddle, 2000;Garigan et al., 2002). Accordingly, C. elegans lifespan could be increased approximately 30% when bacterial growth was arrested by ampicillin (Table 1). We observed that lifespan of BB-treated wild-type animals was not further extended by the presence of ampicillin to arrest bacterial growth (Table 1), indicating that BB may have extended lifespan by relieving microbial stress. To investigate this possibility, we first tested whether the BB polyphenols that prolong lifespan inhibited bacterial growth. We monitored growth of the bacterial lawn under conditions identical to C. elegans aging assays, 11 days on NGM agar at 25 °C, in the presence or absence of BB. During this period, untreated bacterial lawns completed approximately one population doubling, while bacterial growth was arrested after ampicillin treatment. In contrast, BB polyphenols had no effect on bacterial population growth at the doses tested in lifespan assays (Fig. 6A). Furthermore, C. elegans nematodes grown on ampicillin-treated bacteria displayed no increase in thermotolerance, nor did the combination of ampicillin and BB together enhance thermotolerance (Fig. 6B). Together these findings show that ampicillin and BB polyphenols had distinct effects during aging, supporting the conclusion that BB polyphenols do not extend C. elegans longevity through simple antimicrobial effects.  Fig. 4, or untreated control animals, as described in Experimental procedures; n ≥ 25 animals / trial; *P ≤ 0.05, t-test, treated vs. untreated controls.

Genetic requirements for increased survival from BB treatment
Pathways for induction of stress-response genes that affect lifespan have been identified in C. elegans. Because BB treatment might improve survival by acting through these genes, we examined whether mutations in the four major stress response and longevity pathways impaired the ability of BB to prolong lifespan. The premise of these experiments was that BB treatment would not extend lifespan in mutant animals missing a gene required for BB's beneficial effects.
Treatment with the polyphenol, resveratrol, or related compounds, can increase C. elegans lifespan through sir-2.1, which encodes a histone deacetylase-like protein that integrates metabolic status with lifespan (Tissenbaum & Guarente, 2001;Wood et al., 2004). BB treatment extended lifespan of sir-2.1(ok434) animals which lacked sir-2.1 gene activity, showing that BB polyphenols and resveratrol did not act through the same mechanism to increase C. elegans lifespan (Table 1).
In C. elegans, two transcription factors, DAF-16 and SKN-1, promote expression of antioxidant or detoxification enzymes. The DAF-16/FOXO transcription factor promotes expression of genes that confer extended longevity and enhanced stress resistance (Lee et al., 2003;Murphy et al., 2003). SKN-1, which is related to vertebrate Nrf proteins, promotes expression of detoxification enzymes in response to oxidative stress (An & Blackwell, 2003). BB treatment prolonged lifespan of both daf-16(mgDf50) and skn-1(zu67) animals, showing that BB may act independently of these genes (Table 1, Supplementary Fig.  S1A). BB treatment also significantly improved thermotolerance in daf-16(mgDf50) animals ( Supplementary Fig. S1B). Interestingly, BB and ampicillin had additive effects on daf-16(mgDf50) lifespan, further supporting the conclusion that BB polyphenols do not act through simple antimicrobial effects (Fig. 6C).
lifespan as well as improved thermotolerance. While this effect was modest as compared with mutations in longevity genes, it was significant, reliable and robust . Finally, we showed that the beneficial effects of BB treatment appear to require the activity of the osr-1 pathway that also governs osmotic stress resistance. By using C. elegans for these experiments, a genetic analysis of the beneficial effects of BB was possible. We showed that BB could not protect animals against the elevated oxidative stress imposed by the mev-1(kn1) mutation, consistent with our finding that BB also did not protect wild-type animals against oxidative stress from paraquat or hydrogen peroxide treatments. In addition, we found that three stress-response pathways were dispensable for the beneficial effects of BB polyphenols upon lifespan. These pathways were represented by: (i) sir-2.1, which promotes longevity during calorie restriction, (ii) daf-16, which promotes longevity and stress resistance, and (iii) skn-1, which promotes oxidative stress resistance. Specifically, BB treatment was able to extend lifespan of animals lacking any of these three genes. In contrast, BB treatment did not prolong lifespan of animals with defects in the osr-1/unc-43/sek-1 pathway that promotes resistance to osmotic stress. One interpretation of this latter finding is that the effects of BB treatment are mediated, at least in part, through these genes. Alternatively, mutations in the osr-1 pathway could place animals under some stress that is not affected by BB treatment. To date, osr-1 pathway mutants have not been thought to suffer from different causes of death than wild-type animals, suggesting that the latter explanation is less likely. Therefore, we propose that BB exerts beneficial effects through interactions with the osr-1 pathway.
Fractionation of the bulk polyphenols showed that BB's benefits on longevity and thermotolerance cofractionated with the PAC-enriched fraction, and not with ATC or CA, two other major constituents of blueberries that have antioxidant activity in vitro. These findings may either show that (i) the antioxidant effects of these compounds in vitro are not relevant to their in vivo effects in this model, or (ii) differential bioavailability of these compounds is a major determinant of their effects in whole animals. Due to the structural diversity in these fractions, we did not attempt to directly measure the levels of polyphenols inside the animals. However, we did observe the accumulation of BB pigments (primarily ATC) inside the intestines of treated animals, leading us to conclude that components of BB extracts could enter the body of these animals. Our analysis showed that BB did not protect animals against oxidative stress from paraquat, hydrogen peroxide or genetic mutation [mev-1(kn1)], which increases superoxide production in vivo (Senoo-Matsuda et al., 2001). These findings are consistent with the hypothesis that in vitro antioxidant activities of these compounds are not necessarily strong predictors of their benefits in vivo in whole animals.
As a part of this study, we examined possible mechanisms for the beneficial effects of BB treatment. As mentioned, BB treatment was not correlated with increased oxidative stress resistance, nor did BB appear to act solely by relieving microbial stress. One reasonable hypothesis is that BB's antioxidant activity was only sufficient to cope with low-level oxidative stress, such as induced by thermal stress or during aging. Alternatively, BB PACs might have other activities that protected cells specifically during thermal stress.
Given this evidence, we propose the following model for BB's effects on C. elegans lifespan. BB appears to protect animals against some aging-associated stress that overlaps with the stress imposed by high temperature. One possibility is that BB polyphenols protect cells from low levels of free radicals. Alternatively, BB polyphenols might alter the activity of signaling pathways required for response to thermal stress. Given the beneficial effect of BB on lifespan and thermotolerance, it was  Table 1. somewhat unexpected that the osr-1/unc-43/sek-1 pathway for osmotic stress resistance was required for these effects of BB polyphenols, especially since osr-1 has not been found to affect thermotolerance (Solomon et al., 2004). One possibility is that BB and osr-1 differentially affect outputs that can increase resistance to several types of stress. Further investigation of the cellular effects of BB treatment may reveal the nature of this genetic interaction.
Regardless of the specific mechanism involved, it is clear from these experiments that natural compounds available in blueberries can prolong lifespan of a whole organism, under certain conditions. This is a significant finding that lends support to previous experiments on cultured cells or short-term rodent studies showing beneficial effects in aging-related declines and stress resistance. Other studies have also shown that some other naturally available compounds can also prolong C. elegans lifespan under laboratory conditions (Harrington & Harley, 1988;Adachi & Ishii, 2000;Wu et al., 2002;Wood et al., 2004). With the exception of resveratrol, which acts through sir-2.1, the genetic requirements for these effects have not been examined (Wood et al., 2004). This work thoroughly examined in vivo the beneficial effects of compounds found in blueberries, and thus represents a significant advance in the study of the biological effects of natural compounds.

Blueberry extracts and fractionation
Commercially prepared single strength wild blueberry juice (Vaccinium angustifolium) was applied to a preconditioned C18 column (Waters Canada Ltd, Mississauga, ON, Canada). The C18 column was washed with water to remove fructose, glucose and organic acids, which are abundant in blueberries, then with 100% methanol to obtain the total polyphenolic fraction. Methanol was removed under vacuum using a rotary evaporator (Buchi, Essen, Germany) at 30 °C. To obtain the proanthocyanidin fraction, the total polyphenol fraction was dissolved in 50% ethanol and applied to a column of Sephadex LH20 (Sigma-Aldrich, St. Louis, MO, USA), preconditioned with 50% ethanol. The column was washed with 60% ethanol until the eluant was colorless, and then with 70% aqueous acetone to elute the blueberry proanthocyanidins. To obtain the anthocyanin fraction, the total polyphenol fraction was dissolved in acidified water and washed four times with three volumes of ethyl acetate. Anthocyanins were partitioned into the water fraction, which was subsequently freeze dried. Pure chlorogenic acid (Sigma Chemical Company) was used instead of blueberry chlorogenic acid since the chlorogenic acid fraction obtained from blueberry fractionation is contaminated with minor flavonol, and flavonol glycoside components. These components are all contained in the ethyl acetate fraction obtained during anthocyanin isolation.

Phenotypic assays
For aging assays, synchronous populations were obtained by allowing 5 -10 hermaphrodites to lay eggs for 4 h to overnight. For ease of analysis, we used the adult sterile strain, fem-1(hc17), as the wild-type strain to avoid progeny overgrowth in lifespan assays. Therefore, eggs were shifted to 25 °C, the nonpermissive temperature for fertility of fem-1(hc17). Lifespan scoring was initiated after hermaphrodites completed the final larval molt, on the first day of adulthood. For aging assays with BB extracts, treatments were added to NGM agar plates on the first day of the lifespan assay. For lifespan assays with fertile strains, hermaphrodites were transferred daily for the first 4 days of adulthood to avoid progeny overgrowth. In these cases, all treatment plates were prepared on day 0 of adulthood. Statistical analyses and survival plots of lifespan data were performed with JMP analysis software (SAS Institute Inc, Cary, NC, USA). Pharynx pumping rates were scored on adults at room temperature (24 °C) under a Nikon SMZ1500 stereomicroscope (Nikon, Melville, NY, USA).
Thermotolerance assays were performed with hermaphrodites on adult day 5, after the majority of egg-laying had ceased. Animals were transferred onto 3-cm NGM agar plates supplemented as indicated and then incubated at 35 °C for 16 h. Survival was scored as the number of animals responsive to gentle touch as a fraction of the original number of animals on the plate. Animals that had died from dessication on the sides of the plate were censored. Paraquat-induced oxidative stress assays were performed with fem-1(hc17) hermaphrodites at 25 °C as for aging assays, except paraquat was added to NGM medium to 10 mM final concentration (ChemService, West Chester, PA, USA). For hydrogen peroxide, we scored 5-h survival of adult day 5 fem-1(hc17) animals in S-basal medium with indicated concentrations of hydrogen peroxide.
To determine lipofuscin levels, adult hermaphrodites were anesthetized in 0.2% sodium azide and mounted on 2% agarose pads for visualization of intestinal fluorescence on a Nikon E800 microscope using an Endow GFP filter with a mercury UV source (Nikon). Images were captured using a constant exposure time with a Hamamatsu ORCA digital CCD camera (Hamamatsu, Bridgewater, NJ, USA) using OpenLab software (Improvision, Lexingon, MA, USA). Lipofuscin levels were measured using ImageJ software (NIH Image) by determining average pixel intensity in each animal's intestine.