SEARCH

SEARCH BY CITATION

Keywords:

  • aging;
  • autofluorescence;
  • Caenorhabditis elegans;
  • caloric restriction;
  • healthspan

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

Autofluorescent lipofuscin and advanced glycation end-products (age pigments) accumulate with age across phyla, yet little is understood about their formation under physiological conditions and their specific contributions to the aging process. We used in vivo spectrofluorimetry to quantitate autofluorescence in wild-type Caenorhabditis elegans and longevity mutants disrupted for distinct aspects of the aging process. In wild-type animals, age pigments increase into adulthood, accumulating slowly during the reproductive phase and more rapidly during the post-reproductive period. As in humans, insulin signaling influences age pigment accumulation – mutations that lower efficacy of insulin signaling and extend lifespan [daf-2(e1370) insulin receptor and age-1(hx546) PI3-kinase] dramatically lower age pigment accumulation; conversely, elimination of the insulin-inhibited DAF-16/FOXO transcription factor causes a huge increase in age pigment accumulation, supporting that the short-lived daf-16 null mutant is truly progeric. By contrast, mutations that increase mitochondrial reactive oxygen species production do not affect age pigment accumulation, challenging assumptions about the role of oxidative stress in generating these species in vivo. Dietary restriction reduces age pigment levels significantly and is associated with a unique spectral shift that might serve as a rapidly scored reporter of the dietary restricted state. Unexpectedly, genetically identical siblings that age poorly (as judged by decrepit locomotory capacity) have dramatically higher levels of age pigments than their same-aged siblings that appear to have aged more gracefully and move youthfully. Thus, high age pigment levels indicate a physiologically aged state rather than simply marking chronological time, and age pigments are valid reporters of nematode healthspan.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

Multiple aspects of aging biology, including roles for oxidative damage, insulin signaling and dietary restriction, appear remarkably conserved across species (Martin et al., 1996; Heilbronn & Ravussin, 2003; Tatar et al., 2003). Even the manner in which specific organ and tissue systems age can exhibit common features. For example, sarcopenia – the age-associated loss in muscle mass and strength over time – occurs with mid-life onset in nematodes and humans and progresses with a similar cellular profile (Herndon et al., 2002; Fisher, 2004). Taken together, these observations validate the use of simple animal models for deciphering the basic biology of aging and for revealing novel strategies by which deleterious consequences of aging might be postponed or circumvented.

A central goal of current aging research is to identify strategies that maximize human healthspan – the period of vitality that is relatively free of age-related disease and physical impairment, which precedes senescent decline. Given conserved aspects of aging biology, accomplishing this goal should be facilitated by identifying easily scored indicators of healthspan in genetic model systems such as Caenorhabditis elegans in which many mutations affecting longevity have been identified. A few studies have identified biomarkers of aging that can be used to track how well or how fast an animal is aging (Garigan et al., 2002; Herndon et al., 2002; Huang et al., 2004), but most involve laborious evaluation and thus rapidly scored reporters of age-related decline remain highly sought reagents in the field.

One well-conserved feature of aging is the accumulation of endogenous fluorescent compounds over time. ‘Age pigments’ have been found to accumulate with increasing age from nematodes to humans. These fluorescent compounds include lipofuscin, a heterogeneous mix of oxidized and cross-linked molecules (proteins, lipids and carbohydrates) thought to accumulate in lysosomes (Yin, 1996). Another group of age-related fluorophores is advanced glycation end-products (AGEs), which are formed by non-enzymatic sugar addition to free amino groups of proteins followed by autocatalyzed cross-linking (Ulrich & Cerami, 2001). Whether such compounds are a cause or a consequence of age-related cellular decline has not been resolved. It has been suggested that highly cross-linked lipofuscin compounds accumulate in lysosomes where they fail to be efficiently degraded and thereby impair normal protein turnover to diminish cell function (Brunk & Terman, 2002).

Previous studies have examined age pigment accumulation in senescing C. elegans (Klass, 1977; Davis et al. 1982; Hosokawa et al. 1994; Braeckman et al., 2002a,b). These studies, however, mostly recorded data using narrow excitation/emission ranges, which provide only partial profiles of autofluorescent entities, and measured extracted pigments, that might be chemically altered during extraction or when in non-physiological pH. Moreover, normalization of protein levels in extracted fractions is plagued by technical challenges. We therefore revisited characterization of age pigment accumulation by recording broad-range excitation/emission spectra in intact animals, examining wild-type nematodes over time and studying representative lifespan mutants. This new look at endogenous fluorescence accumulation provided several unexpected insights into age pigment biology.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

Autofluorescent species can be measured in living C. elegans using fiber optic-coupled spectrofluorimetry

In vivo fluorescence spectroscopy has been previously used on humans and other mammals to measure endogenous autofluorescent species of skin and other tissues (Leffell et al. 1988; Kollias et al. 1998; Brancaleon et al. 1999; Gillies et al., 2000; Doukas et al., 2001; Na et al., 2001; Sandby-Moller et al., 2003). Fluorescent species that are common in many biological tissues include the aromatic amino acids (tryptophan, tyrosine and phenylalanine), cross-linked collagen and elastin, lipofuscin, NADH, and FAD. Each species fluoresces at characteristic excitation and emission wavelengths (for reviews see Richards-Kortum & Sevick-Muraca, 1996; Kollias et al., 2002; DaCosta et al., 2003). We used spectrofluorimetry to quantitate autofluorescent material in living C. elegans (thickness ∼80 µm; translucent to UV and visible light). We placed 150 wild-type (N2) 10-day-old nematodes on a fiber optic probe (Fig. 1A) and excited with electromagnetic radiation ranging from 240 nm to 580 nm in wavelength; we collected excitation and emission profiles over this range. The spectra generated consistently featured two distinct major emission maxima (contour map in Fig. 1B; two-dimensional representations are given in supplemental Fig. S1A,B). The excitation–emission pair, 290 nm/330 nm, corresponds to the well-documented fluorescence signal generated by aromatic amino acids (Udenfriend, 1962; Lakowitz, 1999) that can be used as a measure of overall protein content. Tryptophan is known to be the major source of fluorescence in this region and thus we refer to the 290 nm/330 nm excitation/emission maximum as the TRP signal. The TRP fluorescence was relatively constant during adulthood in wild-type worms (Fig. 1C and supplementary Fig. S2A) and in all tested mutants (supplementary Fig. S2B–D), except for young adult worms under food restriction (supplemental Fig. S2E,F).

image

Figure 1. In vivo spectrofluorimetric analysis of endogenous fluorescence in C. elegans. (A) Schematic diagram of the spectrofluorimeter used for data collection. A drop of 10 µL sodium azide solution including 150 wild-type (N2) nematodes is placed on a quartz spacer over the fiber optic probe for data collection in a light-tight chamber. (B) The excitation–emission matrix comprises data from 36 consecutive excitation scans acquired from 150 wild-type nematodes on day 10 after hatching. Each color on the contour plot corresponds to fluorescence intensity, increasing from blue to red as shown in the color bar. The TRP band maximum is at 290 nm excitation and 330 nm emission; age pigment band at 340 nm excitation and 430 nm emission. (C) Plot of fluorescence intensity emitted vs. excitation wavelength for TRP (left peak) and age pigment bands (right peak) for nematodes of increasing age. The TRP signal remains relatively constant over time, while the age pigment signal intensity increases dramatically late in life (a.u. = arbitrary units). (D) Representative photographs of fluorescent species in aging nematode intestines (excitation filter 360 nm, emission 420 nm).

Download figure to PowerPoint

The other major autofluorescence maximum (excitation/emission pair 340 nm/430 nm) corresponds to signals previously described to include age-associated biomarkers such as advanced glycation end-products and lipofuscin (Yin, 1996) (supplementary Table S1 gives the autofluorescence wavelengths and molecular identities of some of these components). In C. elegans this spectral region includes two additional smaller shoulder peaks likely to reflect distinct, although unknown, fluorescent species (excitation maximum at 340 nm, emission at 400 nm and 460 nm; supplementary Fig. S1B). As has been observed across animal species, we find that the fluorescence intensity/worm at these wavelengths increases in magnitude with age (Fig. 1C,D). We refer to the endogenous fluorescent species in the 400–460 nm emission region as age pigments.

In C. elegans lipofuscin is thought to accumulate in secondary lysosomes, which appear as granules in intestinal cells (Clokey & Jacobson, 1986). We found that autofluorescence was predominantly localized in gut granules (Fig. 1D) and we used confocal microscopy to show that the spectral properties of individual granules were generally in agreement with our spectrofluorimeter studies (supplementary Fig. S3).

Age pigment fluorescence signals increase during adulthood, with accelerated accumulation beginning in the post-reproductive phase

A key question we sought to address with in vivo spectrofluorimetry was how age pigment components change over lifespan and how mutations or environmental conditions that extend lifespan or healthspan influence their abundance. In marked contrast to the generally maintained protein levels indicated by quantitation of the TRP signal, the age pigment fluorescence increases significantly as populations of wild-type animals grow older (Figs 1C and 2A). Age pigment components accumulate at a low rate during adult days 5–10 (0.027 a.u. day−1, 20 °C; we refer to this period as reproductive midlife), but after day 10 (advanced post-reproductive age), the rate of age pigment accumulation increases significantly (day 10 to day 15 rate = 0.125 a.u. day−1). In this and all following experiments, we came to identical conclusions whether we normalized absolute fluorescence values to the number of animals or whether we scored the age pigment/TRP ratio (see supplementary Fig. S4). Our data suggest that a metabolic transition, reflected by increased age pigment accumulation, occurs in aging adults at approximately day 10, 20 °C. This transition point to accelerated age pigment accumulation occurs earlier for animals reared at 25 °C (∼7 days; data not shown).

image

Figure 2. Age pigment accumulation is influenced by the DAF-2 insulin-like signaling pathway and significantly lowered by reduced food availability/consumption but is not increased by mutations that increase free radical damage. We obtained excitation and emission maxima by running excitation scans from 240 to 580 nm and we plotted emission fluorescence at its maximum value. We report the ex340/em430 wavelength pair for all but dietary restricted animals; for dietary restricted animals (F,G) we report the ex320–340/em430 pair. Note also that for A–G, plotting the ratio of age pigment/TRP scores reveals the same results as does normalization on a per animal basis that is shown here (supplementary Fig. S4). (A) Age pigment measurements for wild-type (N2) animals over time indicate accelerated age pigment accumulation late in life. In early adulthood, rate of accumulation is steady (rate for day 5 to day 10 = 0.027 a.u. day−1); after 10 days, however, accumulation rate rises substantially (rate from day 10 to day 15 = 0.125 a.u. day−1). (B) Fluorescence over most of adulthood is lower in insulin pathway mutants age-1(hx546) and daf-2(e1370) that diminish insulin signaling. (C) The mutants daf-16(m26) and daf-16(mgDf50) show increased rates of post-reproductive stage fluorescence accumulation as compared with wild-type. (D) The double mutant strain daf-2(e1370); daf-16(m26), which has a lifespan similar to wild-type, exhibits age pigment levels similar to wild-type on day 5 and day 12. (E) Age pigment fluorescence intensities of mev-1(kn1) and gas-1(fc21) compared with wild-type over time. Note that very low scores for gas-1 early in life may reflect slow growth early in this experiment. mev-1(kn1) and gas-1(fc21) mutations were confirmed by sequence analysis. (F) Comparison of age pigment measurements for wild-type animals reared on agar plates (black fill) or in liquid culture (white fill) (P-value < 0.05). (G) Comparison of age pigment signals for plate-reared wild-type N2 (black circles), plate-reared eat-2(ad465) (white circles), plate-reared unc-26(e205) (black triangles) and liquid culture-reared N2 (gray squares). We noted a 1-day developmental delay in egg-laying onset in eat-2 and unc-26 mutants. However, even an adjustment for relative age as compared with wild-type would not change the conclusion that age pigments are present at very low abundance in these backgrounds. Likewise, although the graphs in G correspond to excitation maximum of 340 nm for N2 and 320–340 nm for eat-2, comparison of both at 320 nm or both at 340 nm would not markedly change the absolute values plotted or the conclusions reached. (H) The excitation maximum for the age pigment fluorescence band shifts with age for eat-2(ad467), unc-26(e205) and wild-type animals reared in liquid culture, while it remains constant (at 340 nm) for wild-type animals reared on plates. The excitation maximum remains at 340 nm for all other longevity mutants scored in this study (data not shown).

Download figure to PowerPoint

Lowering DAF-2 insulin-like signaling lowers age pigment abundance

Reduction-of-function mutations affecting the DAF-2 insulin receptor protein and the AGE-1 PI3-kinase increase lifespan and extend locomotory healthspan (Kenyon et al. 1993; Adachi et al. 1998; Herndon et al., 2002). We found that levels of age pigment fluorescence are significantly lower in daf-2(e1370) and age-1(hx546) mutant backgrounds as compared with wild-type at all measured time points, with the exception of the age-1 mutant at day 5 (Fig. 2B). Although age pigment signals do progressively increase over time in insulin longevity mutant backgrounds, daf-2 and age-1 mutants accumulate age pigment fluorescence with seemingly different kinetics from wild-type and from each other. The age pigment fluorescence profile in the age-1(hx546) mutant roughly parallels that of wild-type but the time at which age-1 mutants enter the phase of accelerated age pigment accumulation appears delayed. In the daf-2(e1370) mutant background, age pigment products accumulate at a lower rate late in adulthood as compared with wild-type, with no clear transition to elevated accumulation rates. Despite these kinetic differences, it is clear that lowered insulin-like signaling is associated with low age pigment concentrations over much of C. elegans adulthood.

daf-16/FOXO transcription factor deficiency causes enhanced age pigment accumulation

The longevity effects of daf-2 and age-1 mutations are mediated by the function of the FOXO transcription factor DAF-16 (Kenyon et al. 1993; Larsen et al. 1995). daf-16 null mutants have modestly shortened mean and maximum lifespans as compared with wild-type nematodes (Larsen et al., 1995, 2002). We monitored age pigment signals in two daf-16 mutant backgrounds, the null allele daf-16(mgDf50) and daf-16(m26) (partial loss of function allele) (Fig. 2C). At all time points during the measured adult lifespan after day 5, daf-16 mutants harbor significantly elevated levels of age pigments, with the null allele in particular conferring extremely high accumulation later in life. Notably, the rate of increase of endogenous fluorescence appears significantly accelerated during the advanced post-reproductive phase of life (day 10 to day 12 rate: N2 = 0.200 a.u. day−1; daf-16(m26) = 0.436 a.u. day−1; daf-16(mgDf50) = 0.660 a.u. day−1). We conclude that age pigment products accumulate at accelerated rates when daf-16 activity is low or eliminated and infer that DAF-16 may normally help limit age pigment accumulation.

Because daf-16 mutations suppress daf-2 lifespan extension phenotypes, we wondered how daf-16 deficiency influences age pigment levels in the daf-2 mutant background. We found that in the daf-2(e1370); daf-16(m26) double mutant [which has a lifespan similar to wild-type (Larsen et al. 1995)], age pigment levels are restored to approximately wild-type values at day 12, although scores do not fall to values as low as in the daf-2(e1370) mutant alone (Fig. 2D). We conclude that daf-16(m26) partially suppresses the daf-2 age pigment reduction phenotype. In addition, our data suggest that DAF-2 activity is needed to promote the massive accumulation of age pigment products that occurs when daf-16 is mutant.

Reduced lifespan mutations that increase mitochondrial ROS damage do not elevate age pigment signals

In vitro mammalian studies have suggested that age pigment formation is potentiated by oxidative stress (Yin et al., 1995; Nilsson & Yin, 1997). We therefore probed the potential relationship of oxidative load and age pigment accumulation using a genetic approach. The majority of damaging cellular reactive oxygen species (ROS) are thought to be generated by the mitochondria during electron transport. mev-1(kn1) encodes a missense mutation in succinate dehydrogenase cytochrome b of mitochondrial complex II that confers enhanced sensitivity to paraquat (Ishii et al., 1990, 1998) and is associated with elevated mitochondrial superoxide production, elevated carbonyl accumulation (Adachi et al., 1998; Senoo-Matsuda et al., 2001; Yanase et al., 2002) and short lifespan (Hartman et al., 2001). We found that levels of age pigment products are lower in mev-1 than those of age-matched wild-type animals, with rates of age pigment accumulation appearing especially diminished in advanced post-reproductive age (Fig. 2E). Even when much of the mev-1 population has died, age pigments are clearly not higher relative to wild-type.

We also examined age pigment accumulation in the gas-1(fc21) mutant background, which encodes a subunit of the mitochondrial NADH:ubiquinone-oxidoreductase of complex I of the electron transport chain (Kayser et al. 1999). The gas-1(fc21) mutant has a modest reduction in lifespan in atmospheric oxygen, is hypersensitive to ROS damage (Kayser et al., 2001) and accumulates oxidative damage to mitochondrial proteins (Kayser et al., 2004), and thus has been proposed to experience enhanced endogenous oxidative stress. We found age pigment levels in the gas-1(fc21) mutant significantly lower than wild-type over time (Fig. 2E). In summary, unlike short-lived daf-16, the short-lived mev-1 and gas-1 mutants do not produce abnormally elevated age pigment levels. We conclude that elevated age pigment accumulation is not an obligate component of shortened lifespan. In addition, our data suggest that genetically induced increases in mitochondrial oxidative stress do not markedly influence age pigment accumulation in an in vivo setting.

Limited food availability/consumption lowers age pigment signals

We also investigated how food availability influences age pigment accumulation. Growth of C. elegans in liquid culture in the presence of bacteria alters metabolism and extends lifespan relative to plate-reared animals, possibly due to a dietary restriction effect (Vanfleteren & Braeckman, 1999; Houthoofd et al., 2002a,b, 2003). We found that age pigment fluorescence in liquid-reared animals was significantly lower than that of age-matched counterparts grown on plates, comparing these at day 5 and day 10 (Fig. 2F,G).

We also examined temporal age pigment profiles in the putatively dietary-restricted eat-2(ad465) and unc-26(e205) mutants, which have reduced pharyngeal pumping and long lifespan (Lakowski & Hekimi, 1998; Harris et al., 2000; Fig. 2G). We found that at all adult ages tested, age pigment levels are significantly lower in feeding-defective mutants than in age-matched wild-type animals. Moreover, the rate of age pigment accumulation is significantly reduced over the first 2 weeks of lifespan as compared with wild-type in feeding-defective mutants, with no clear transition to a higher accumulation rate apparent (day 5 to day 10: N2 = 0.027 a.u. day−1; eat-2(ad465) = 0.017 a.u. day−1; unc-26(e205) = 0.004 a.u. day−1; day 10 to day 15: N2 = 0.125 a.u. day−1; eat-2(ad465) = 0.009 a.u. day−1; unc-26(e205), ND). Taken together, our data support that low food consumption slows accumulation of fluorescent age pigments and establishes that feeding limitation is associated with low age pigment signals late into adult lifespan.

A novel fluorimetric property of dietary-restricted animals

Importantly, our excitation/emission scans for the eat-2, unc-26 and liquid-reared animals revealed an unexpected change in age pigment fluorescence properties over time. For all three food-restricted cultures, we found that the age pigment fluorescence excitation maximum is shifted to shorter wavelengths during early adulthood as compared with wild-type grown on agar plates (on day 5 P < 0.02 for eat-2; P < 0.001 for unc-26 and N2 in liquid culture). With increasing age, the fluorescence excitation maximum, which begins at ∼320 nm, gradually shifts towards 340 nm (for eat-2 by day 18; Fig. 2H). This shift in fluorescence excitation maximum likely reflects changes in the proportions or identities of different fluorescent molecules that accumulate under food limitation. We never observed a similar shift in wild-type or mutant strains assayed (15 different strains). Taken together, our data constitute a strong line of evidence that liquid-reared animals and the eat-2 and unc-26 feeding-impaired mutants share at least some metabolic properties during adulthood. Moreover, these data suggest that the excitation maximum shift might serve as a rapidly scored indicator of dietary restriction (see Discussion).

Age pigment levels correlate with physiological, rather than chronological, aging

We previously documented the fact that not all individual animals in a synchronized C. elegans culture undergo age-related decline at the same rate, despite their identical genotypes and identical growth conditions (Herndon et al., 2002). A mid-to-late age adult culture will include animals that are similar to young adults in locomotion, swimming and touch sensory behavior (A class), animals that are moderately impaired in these behaviors (B class), animals that are virtually paralyzed and non-responsive (C class), and animals that are dead. We also previously found that behavioral class type is a more accurate predictor of survival probability than is chronological age – for example, at day 12, B class animals live an average of 9.0 more days, whereas C class animals live on average 4.7 more days (Herndon et al., 2002). Thus, despite identical chronological age, identical genetic backgrounds and a common environment, nematodes can age gracefully (exhibiting a long healthspan and a relatively long life expectancy) or can age poorly (exhibiting a short healthspan and a relatively short life expectancy). The factors underlying the substantial stochastic influence(s) on aging have not been defined.

We wondered whether physiologically ‘old’ animals, impaired in locomotory capabilities or sensory responses, might have age pigment profiles distinct from their age-matched ‘youthful’ counterparts. We separated 11-day-old wild-type nematodes into A, B and C classes and measured age pigment fluorescence levels in these three groups (Fig. 3). Quite strikingly, we found the C class 11-day-old nematodes contain more than four times the age pigment levels of A class 11-day-old nematodes; ∼2.5-fold more than B class animals. The TRP signals from A, B and C animals, however, are similar. Thus, the dramatically impaired C class nematodes are not simply globally dysregulated for macromolecular accumulation. We conclude that age pigment levels closely parallel ‘quality of life’ phenotypes in old nematodes. Age pigment levels do not rise coincidentally with chronological age – rather it is the physiological condition of the animals that appears to be reflected in the age pigment scores.

image

Figure 3. Age pigment fluorescence reflects the physiological age of C. elegans rather then the chronological age. We sorted 11-day-old nematodes into A (youthful), B (partially impaired) and C (severely impaired) classes based on locomotion and response to gentle touch (as described in Herndon et al., 2002) and measured autofluorescence intensities. Bars represent three repeats of scans of 50 nematodes per trial; gray bars, TRP levels; black bars, age pigment levels. Note that all animals are of the same chronological age – A, B, C classes reflect differences in behavioral vigor. Age pigment fluorescence levels of class C animals are significantly different from class A and B (P-value < 0.05). No statistical significance for TRP levels was detected.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

We report here the first quantitative in vivo characterization of fluorescent signals in aging C. elegans. Our work is distinct from previous measurements of age pigments in nematodes (Epstein & Gershon, 1972; Epstein et al. 1972; Klass, 1977; Davis et al., 1982; Hosokawa et al., 1994; Braeckman et al., 2002a,b) in that we measured products in intact animals, enabling us to normalize signals on a per animal basis (or, alternatively, to endogenous TRP scores) and avoiding complications of differential extraction, chemical changes that occur during fractionation and pH alterations that affect signals (Yin, 1996; Lakowitz, 1999). Additionally, we recorded broad-range excitation/emission spectra, which revealed novel properties of age pigments in senescing nematodes under specific environmental and genetic conditions.

Age-pigment accumulation rate reveals a post-reproductive mid-life transition

Early in adulthood and into midlife, age pigments accumulate at a low rate in C. elegans; later age pigments accumulate ∼5-fold faster (around day 10 after hatching, 20 °C, which is just after reproduction by self-fertilization normally ceases), suggesting a midlife post-reproductive shift to a distinct physiological state. A significant portion of this midlife transition may reflect the increasing numbers of animals in the population that appear physiologically ‘aged’– class C animals that have much higher levels of age pigments than their youthful same-age counterparts (Fig. 3) become more abundant after day 9 (Herndon et al., 2002). Although the molecular basis of this accumulation rate shift is not known, it is clear that the accelerated rate of age pigment accumulation is correlated with the period of post-reproductive decline. Interestingly, discontinuities in rates of change in locomotory rate and defecation frequency have also been noted around the end of the reproductive phase in other studies (Bolanowski et al., 1981; Huang et al., 2004), indicating that the midlife transition is apparent using multiple aging biomarkers.

Insulin-like signaling has major impacts on age pigment accumulation

Mutations that diminish insulin-like signaling through the DAF-2 insulin receptor pathway significantly lower age pigment levels in senescing animals provided the DAF-16/FOXO transcription factor is functional. It may be noteworthy that mutations in different components of the DAF-2 insulin-like signaling pathway can affect the bi-phasic curve of age pigment accumulation in different ways. daf-2(e1370) mutants continue to accumulate age pigments at relatively constant low rates over time, such that a shift to accelerated rate is never apparent, whereas age-1(hx546) mutants exhibit a delay in the transition to accelerated accumulation rate. These differences may reflect differences in signaling levels through DAF-2 insulin-like pathways, which likely involve branches and cross-talk with other signaling pathways to exert their physiological effects. Alternatively, differences could reflect distinct actions of daf-2 and age-1 on processes that regulate accumulation.

Mutations affecting DAF-16/FOXO, a key transcription factor targeted for down-regulation by insulin signaling, markedly elevate endogenous age pigment levels, accelerating accumulation rate especially dramatically in late adulthood. Thus, in addition to promoting antioxidant and antimicrobial defenses (Honda & Honda, 1999; Lee et al., 2003; McElwee et al., 2003; Murphy et al., 2003; Moy et al., 2004), DAF-16/FOXO normally functions in some capacity to limit age pigment accumulation. The fact that age pigments accumulate in young daf-16 mutants to levels characteristic of wild-type of older age supports that daf-16 mutants are truly progeric.

Human diabetics, in whom insulin signaling is abnormal, accumulate high levels of advanced glycation end-products (Monnier & Cerami, 1982; Monnier et al., 1986; Hudson et al., 2003) and rats injected with streptozotocin to induce diabetes experimentally exhibit high levels of fluorescent age pigments in trigeminal neurons (Sugaya et al., 2004). Coupled with our observations in C. elegans, these data suggest that a conserved outcome of insulin signaling perturbation is age pigment accumulation.

Elevated mitochondrial oxidative stress is not correlated with exacerbated age pigment accumulation in vivo

Our work confirms basic observations regarding age pigments in aging C. elegans, supporting that specific autofluorescent species accumulate with chronological age. However, in contrast to some previous work (Hosokawa et al., 1994; Braeckman et al., 2002b), we found no evidence for elevated accumulation of age pigments in the mev-1 and gas-1 mutants, despite the fact that such strains have been reported to have increased oxidative damage [elevated superoxide generation in mev-1 (Senoo-Matsuda et al., 2001) and increased carbonyls in both (Adachi et al., 1998; Kayser et al., 2004)]. Discrepancies in our conclusions might be attributed to different age pigment measurement protocols and different normalization standards; we confirmed the presence of mutations in both mutant strains.

Studies with mammalian age pigments in vitro under conditions of extreme oxidative stress have suggested that free radicals promote lipofuscin and advanced glycation end-product formation (Yin et al., 1995; Yin, 1996; Nilsson & Yin, 1997; Brunk & Terman, 2002). As both mev-1 and gas-1 exhibit exacerbated phenotypes under hyperoxic conditions that elevate oxidative stress (Hosokawa et al., 1994; Adachi et al., 1998), it remains possible that age pigment accumulation might be observed under increased O2, a condition that we did not address. Nonetheless, under conditions of atmospheric oxygen in which mev-1 and gas-1 lifespans are shortened, we find there is no major acceleration of age pigment accumulation above wild-type levels, even later in life when a significant number of mev-1 or gas-1 mutants have died. Thus, early death is not necessarily associated with age pigment accumulation in these backgrounds; elevated mitochondrial ROS production does not necessarily cause an increase in accumulation of autofluorescent compounds in vivo. In addition to challenging some assumptions about age pigment formation in vivo, our observations also raise the question as to whether mev-1 and gas-1 mutations are truly progeric; it is possible that these mutations are toxic because of defects in mitochondrial function that confer sickness rather than specific acceleration of the aging process. ROS production and short lifespan are not always correlated (Miwa et al., 2004).

Age pigment levels identify impaired animals that have aged poorly relative to their same-aged counterparts

We found that same-aged animals that senesce gracefully (i.e. move and respond to stimuli relatively youthfully for their age) have lower levels of age pigments than their decrepit same-aged siblings reared under identical conditions. We thus correlate high in vivo autofluorescence values with impaired behavioral phenotypes and suggest that the ‘physiological’ aging state might be reported by the age pigment level. That age pigments might reflect physiological aging state has been suggested based on other work that correlated metabolic rate with lipofuscin levels (Sohal, 1981; Sohal & Brunk, 1989). Our observations suggest that screening for genetic or pharmacological interventions that block or delay age pigment accumulation might identify genetic manipulations that extend middle age healthspan, a goal of considerable importance. Evaluating healthspan by age pigment scores is far more rapid than other indicators that have been carefully quantitated, such as locomotory or cellular decline (Herndon et al., 2002; Huang et al., 2004).

Do age pigments promote age-related decline?

In higher organisms, advanced glycation end-products have been proposed to contribute to diabetic complications, including peripheral neuropathy, atherosclerosis, retinopathy, cataract formation and renal dysfunction (Monnier & Cerami, 1982; Monnier et al., 1986; Ulrich & Cerami, 2001; Hudson et al., 2003; Yamagishi et al., 2005), as well as to Alzheimer's disease pathology (Smith et al., 1994; Yan et al., 1994; Sasaki et al., 1998). The dramatic increase in age pigment accumulation during the C. elegans post-reproductive phase and the striking relative elevation of age pigments in decrepit class C animals that have low life expectancy are consistent with the hypothesis that age pigments could actually promote the aging process. Also consistent with such a possibility is that all the long-lived mutants we tested exhibited low levels of age pigments relative to wild-type animals (summarized in supplementary Fig. S5) and that, conversely, the highest degree of age pigment accumulation in progeric daf-16 mutants was correlated with the greatest shortening of lifespan (supplementary Table S2).

How might age pigment accumulation shorten lifespan? The mitochondrial–lysosomal axis theory of aging (Brunk & Terman, 2002) suggests that age pigments accumulate in secondary lysosomes, which become metabolically inert due to the dedication of lysosomal enzyme action to undegradable lipofuscin. As a consequence, normal protein turnover and degradation of damaged mitochondria does not transpire – misfunctioning mitochondria and diminished protein turnover under conditions in which damaged proteins accumulate in the cell contribute to decline. If such a mechanism is operative, experimental elimination of age pigments should extend both healthspan and lifespan. In one report, aminoguandine was used to limit formation of advanced glycation end-products in Drosophila without consequent lifespan extension, but the compound was clearly toxic, limiting interpretation of this result (Oudes et al. 1998). We tried to use AGE inhibitor aminoguanidine to manipulate endogenous age pigments, but were unable to detect any effects, possibly because the drug did not easily enter the worm, a common problem in nematode pharmacology (Rand & Johnson, 1995). Thus, although it is tempting to speculate that high levels of age pigments contribute to age-related decline by impairing the protein and mitochondrial turnover function of lysosomes, we cannot at this time distinguish whether high age pigment levels contribute to the cause of, or reflect the consequence of, poor aging.

Discovery of an easily measured signature of the dietary restricted state?

Among all strains tested in our study, we found the lowest aging pigment fluorescence levels and the lowest accumulation rates in animals previously suggested to experience dietary restriction and documented to have extended lifespan. Moreover, we noted a striking distinction in spectral properties of autofluorescent entities that we never found in any other mutants or under any other environmental conditions assayed. Specifically, we found that the age pigment excitation wavelength maximum, which is usually invariant, changes over time in food-limited animals and is notably shifted to low values in young adult animals. In addition, we found TRP signals are low in young, food-restricted animals. We propose that these features (extremely low age pigment levels, excitation shift for the age pigment fluorescence and low early TRP) may constitute a fluorimetric ‘signature’ for the dietary restricted state. This biomarker might constitute the first rapidly measured phenotype of dietary restriction in a genetic model, opening up the opportunity to apply classic and reverse genetic strategies to define genetic influences on dietary restriction.

New insights into age pigment biology

Given that the fluorescent species we measured are in their native context and cannot be differentially lost or chemically modified during measurement, our data advance understanding of age pigment biology by providing a quantitative, high-precision in vivo evaluation of age-related changes in autofluorescent species and the genetic influences upon this accumulation. In general, long-lived mutants we tested have relatively low levels of age pigments (summarized in supplemental Fig. S5). Age pigment accumulation is highly sensitive to insulin signaling and food limitation and appears to include unique fluorescent properties during dietary restriction. Age pigment accumulation occurs at a markedly increased rate in post-reproductive animals and its preponderance is associated with a decrepit state. Our data should rekindle hypotheses regarding the deleterious consequences of age pigments for the aging process.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

C. elegans cultures

Strains were obtained from the Caenorhabditis Genetics Center (CGC), and grown on NGM plates seeded with E. coli OP50 (Brenner, 1974). For aging studies gravid adults were allowed to lay eggs on seeded plates for a few hours; day 1 was the day of egg deposition, animals reach adulthood at day 3 and lay eggs until days 8 or 9. Plates were incubated at 20 °C until screening, except for the temperature-sensitive daf-2(e1370) strain, which was reared at the permissive temperature (15 °C) and shifted to 20 °C after the larvae bypassed the L3 stage (which slowed development by ∼1 day). Before spectrofluorimetry, animals were washed in a drop of M9 on an unseeded NGM plate. For liquid culture, 50 µL of E. coli OP50 (OD600 approx. 1) was inoculated into 25 mL LB, grown for 6 h at 37 °C and resuspended in 12 mL complete S medium (Stiernagle, 1999). About 25 eggs were added to wells containing 600 µL of the bacteria-supplemented S medium. The medium was changed every other day.

Segregation of animals into classes A, B and C was performed as described in Herndon et al. (2002). Briefly, we scored the locomotion phenotype of age-synchronized worms by gently prodding them with a platinum wire. Worms that moved quickly away from the touch stimulus in a sinusoidal pattern were categorized as class A. Class B comprised animals that did respond to touch, but with somewhat uncoordinated, slow movement. Worms that were unable to move forward or backward, but were clearly alive (as seen by slight movement of the head in response to touch), were classified as class C. Although this scoring is somewhat subjective, different researchers reproducibly sort according to class with similar results.

Fluorescence spectroscopy

Fluorescence spectroscopy was performed using an in vivo spectrofluorimeter (SkinSkan, JY Horiba, Edison, NJ, USA) (Stamatas et al., 2002). A 1-mm quartz spacer (SPI, West Chester, PA, USA) was used between the probe and the sample to provide the necessary distance to maximize the signal detection for this probe geometry. Before each set of measurements the instrument was spectrally calibrated for excitation and emission following the manufacturer's instructions. For each experiment exactly 150 worms were placed in a 10 µL drop of 10 mm NaN3 on a Teflon sheet (YSI Inc., Yellow Springs, OH, USA). The Teflon sheet was placed on the quartz spacer and the drop was centered over the active probe area. The probe–spacer–sample assembly was shielded from ambient light.

Fluorescence measurements of the samples consisted of a series of 32 excitation scans (240–580 nm in steps of 10 nm). Each excitation scan was acquired after setting the emission monochromator position to a selected emission wavelength. Starting at 260 nm the emission monochromator was advanced 10 nm for every consecutive scan until it reached 600 nm. Combining all these measurements, a matrix of fluorescence values can be synthesized with points every 10 nm in excitation and in emission and the fluorescence intensities can be represented as contour maps (excitation–emission matrix, Fig. 1B). The excitation–emission matrix allows for easy localization of the fluorescence maxima and comparison of their relative intensities.

The excitation/emission = 290 nm/330 nm band is characteristic for aromatic amino acids tryptophan, tyrosine and phenylalanine with tryptophan being the primary contributor due to its high quantum yield (Udenfriend, 1962; Lakowitz, 1999). Another broad fluorescence band is evident at 340 nm excitation, which is typical of age-sensitive pigments. All fluorescence signals increased linearly with the number of worms in the range of 50–200 animals of the same age (data not shown). The NaN3 solution alone and small debris of agar did not fluoresce; bacteria gave a negligible fluorescence.

For the experiments on A, B and C class animals we used a spectrofluorimeter with a higher sensitivity photomultiplier (Model R928B, Hammamatsu, Japan). The higher detector sensitivity allowed us reliably to use fewer than 150 worms per scan. Owing to differences in the detector sensitivity, lamp output, small variations in the optical path, etc., the readings between the two instruments were not identical. Therefore, whenever we used the second instrument the fluorescence readings were scaled to those of the first instrument using a scaling factor, which was calculated as a ratio of the values of a synchronous scan of a reflectance standard (Minolta, Japan) measured with each instrument. The scaling was confirmed by comparing the tryptophan fluorescence value of wild-type worms at day 5 measured with the two instruments.

Because repeated scans of the same sample gave the same fluorescence intensity values, photobleaching or change in pigment properties due to specific excitation does not appear to occur. Likewise, we did not note any differences in TRP or age pigment scores in an experiment that lacked NaN3, suggesting that worm anesthesia does not alter age pigment properties. In multiple experiments, we found that scoring age pigment values per animal and scoring the ratio of age pigment/TRP values both indicated identical conclusions, supporting that both are legitimate normalization strategies.

Microscopy

A Zeiss Axioplan 2 Microscope with a UV cube (excitation bondpass filter centered at 360 nm and 420 nm emission longpass filter) was used to image worms and to verify fluorescence intensity with the results from the spectrofluorimeter. For confocal microscopy we used a Zeiss LSM510 Meta Confocal Laser Scanning Microscope with an Ar laser at 364 nm. Emission was collected at 398–483 nm. Spectral analysis was performed with LSM510 Meta Version 3-2.

Data analysis

All experiments were repeated at least in triplicate. The average values and standard deviations are reported in the figures. Statistical significance was accepted at the level of P < 0.05 using the standard t-test.

Acknowledgments

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

We thank G. Patterson, J. Lenard and B. Grant for critically reading the manuscript. Nematode strains used in this work were provided by the Caenorhabditis Genetics Center (CGC), which is supported by the NIH National Center for Research Resources. This work was supported by grant R01 AG024882-01 from the National Institute on Aging. M.D. is an Ellison Medical Foundation Senior Scholar In Aging (AG-SS-1307-04). B.G. was supported in part by a Louis Bevier Fellowship and a Fulbright Scholarship.

Supplementary material

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

Table S1 Selected studies on fluorescent material in biological tissues.

Table S2 Lifespan of insulin signaling mutants and age pigment fluorescence.

Fig. S1 Details on the excitation and emission scans of the TRP and the age pigment band.

Fig. S2 TRP fluorescence levels remain constant throughout adult life except under conditions of food limitation.

Fig. S3 Confocal images of an N2 and an eat-2 worm, at 11 days of age.

Fig. S4 Ratio of age pigment fluorescence to tryptophan fluorescence.

Fig. S5 The relationship of age pigment accumulation to lifespan.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgments
  8. Supplementary material
  9. References
  10. Supporting Information

Table S1. Selected studies on fluorescent material in biological tissues. Table S1. Lifespan of insulin signaling mutants and age pigment fluorescence. Fig. S1. Details on the excitation and emission scans of the TRP and the age pigment band. Fig. S2. TRP fluorescence levels remain constant throughout adult life except under conditions of food limitation. Fig. S3. Confocal images of an N2 and an eat2 worm, at 11 days of age. Fig. S4. Ratio of age pigment fluorescence to tryptophan fluorescence. Fig. S5. The relationship of age pigment accumulation to lifespan. Appendix S1. References relating to the supplementary material.

FilenameFormatSizeDescription
ACEL_153_sm_appendixA1.doc33KSupporting info item
ACEL_153_sm_figureS1.doc228KSupporting info item
ACEL_153_sm_figureS2.doc206KSupporting info item
ACEL_153_sm_figureS3.doc583KSupporting info item
ACEL_153_sm_figureS4.doc73KSupporting info item
ACEL_153_sm_figureS5.doc277KSupporting info item
ACEL_153_sm_tableS1.doc59KSupporting info item
ACEL_153_sm_tableS2.doc28KSupporting 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.