Journal of Gerontology: BIOLOGICAL SCIENCES
2006, Vol. 61A, No. 6, 557–561
Copyright 2006 by The Gerontological Society of America
Guest Editorial
Is (Your Cellular Response to) Stress Killing You?
Felipe Sierra
Biology of Aging, National Institute on Aging, National Institutes of Health, Bethesda, Maryland.
Free radicals provide a generally accepted explanation for age-related decline in tissue function.
However, the free radical hypothesis does not provide a mechanistic course of action to explain
exactly how damage to macromolecules translates into the recognizable pathophysiology of aged
organisms. Recent advances in the fields of DNA damage and cellular senescence point towards
a substantial role for the DNA damage response, rather than DNA mutations per se, in the genesis
of cellular and/or tissue damage. Furthermore, several studies suggest that protein damage can be
at least as important as DNA damage in bringing about the aging phenotype. Here we propose
that a ‘‘protein damage response,’’ namely the ER/UPR (endoplasmic reticulum/unfolded
protein) stress response is likely to play an important role in the aging process.
I
N 1956, Harman proposed the free radical hypothesis of
aging (1). The hypothesis has ebbed and waned, but it has
survived fairly intact through almost half a century. Much
understanding of the aging process has come from experiments designed to test Harman’s tenets, but a central
question still remains: How exactly does random damage to
macromolecules translate into the reproducible and recognizable organismic phenotype we call ‘‘aging’’? This issue
was first raised in 1959 (2), but it still remains unsolved (3).
Most often, free radicals or other noxious agents will
damage macromolecules in a very stochastic, idiosyncratic
fashion. It is difficult to understand how such a random
process can lead to the significant and reproducible loss
in tissue function observed during aging. Therefore, the
premise of this writing is that it is not damage per se that
leads to the phenotype of aging, but rather, it is the cellular
response to that damage that is at the center of the action.
Thus, we propose that aging ensues only when damage is
extensive enough, and of a type capable of inducing a
cellular response, which often results in either cell senescence or apoptosis. We further propose that these cellular
responses play a pivotal role on the road to the tissue
deterioration observed during aging.
The cellular response to irreparable DNA damage is better
understood than is the response to protein damage. Point
mutations in DNA, either as a result of damaging insults
or errors in replication, do not often activate a cellular
response. However, more severe damage (for example,
genomic instability brought about by double-stranded
breaks or bulky adducts) leads to activation of a p53dependent response, which can result in apoptosis or
senescence (4). There are accumulating data in the literature
suggesting that premature aging is not a sine qua non result
of increased DNA damage, but rather, premature aging only
occurs in those models where repair is inadequate and the
cellular response to DNA damage has been activated (see
below). There are now also data that suggest that damage
to proteins might be relevant to aging (5), and it is therefore tempting to suggest that, in this case too, aging will
ensue only if a cellular response to that damage is activated.
Here we propose that the ER or UPR stress response
(Endoplasmic Reticulum or Unfolded Protein Response,
respectively) (6) could play this role, and thus be central to
the development of the aging phenotype. As in the case of
the DNA damage response, activation of the ER/UPR
stress response by damaged proteins can also lead to both
loss of cellular function and, if overwhelmed, apoptosis
(Figure 1).
DNA DAMAGE AND AGING
DNA is believed to suffer different types of damage, such
as breaks, abasic sites, and base lesions. It has been
estimated that at least 2,000–10,000 purine bases are turned
over per day per cell (7). This creates an enormous burden,
and consequently, a complex and interwoven system of
DNA repair enzymatic activities exists within each cell, to
reverse such damage. It has been estimated that hundreds
of genes in the human genome are directly devoted to DNA
repair mechanisms, with many more ancillary factors
suspected (8). The genome is kept intact by two major mechanisms: caretakers and gatekeepers (9). The functions of
caretakers are to ensure that the DNA is not damaged and,
if it is, to get it repaired. Caretakers include enzymes that
scavenge reactive oxygen species (ROS) and DNA repair
mechanisms. However, if the caretakers fail and DNA gets
irreparably damaged, then gatekeepers come into play,
and they ‘‘decide’’ the fate of the affected cell as a whole.
Classical gatekeepers include the tumor suppressor proteins
p53 and pRB. Gatekeepers are not activated by point
mutations; therefore, point mutations do not engage
a cellular response. However, when gatekeepers sense that
the damage is irreparable, they activate a cellular response
557
558
SIERRA
(either apoptosis or senescence), which effectively removes
the affected cell from the pool of proliferating cells (4).
There are considerable data indicating that, if a cellular
response to damage is not activated, then cells can withstand
a significant amount of damage to DNA without the
organism showing signs of premature aging (10). Indeed,
in most of the mouse models where caretaker functions have
been inactivated, an increased level of DNA mutations
has been observed as expected, but the mice fail to display
an accelerated aging phenotype. For example, Xpc/ mice
accumulate up to 30-fold higher levels of DNA mutations
than do their wild-type counterparts, yet there is no effect on
their life span (11). A similar, though less dramatic result
has been observed in the case of scavenging proteins. For
example, Van Remmen and colleagues (12) reported that the
SOD2þ/ mouse suffers a 3-fold to 4-fold increase in DNA
mutations, with no detectable effect in life span. Because
these animals indeed display a significant increase in DNA
damage, it cannot be argued that the effect of the gene
knockout is being counterbalanced by a partly redundant
mechanism: There is a considerable accumulation of unrepaired mutations that has no deleterious effect on life span.
There are exceptions, including the Ku86/ mouse (13), the
XPDTTC mouse (14), and more recently, a model in which
catalase was directed to the mitochondria (15). In these
cases, the expected changes in life span were observed
(decrease in the two former, increase in the latter). The
mechanisms that explain the differences between the different models still need to be unraveled. In contrast, mouse
models in which the activity of gatekeepers (including
telomerase, Wrn, Blm, ATM, or p53) has been manipulated
do generally display an accelerated aging phenotype
(16–20). In the cases of telomerase, Wrn, Blm, and others,
the gene knockout results in generalized genomic instability,
including not only double-stranded breaks, but also telomere
shortening and/or stalled replication or transcription complexes, both of which appear to be interpreted by the cell
as an unrepaired double-stranded break. As previously
mentioned, this type of damage leads to activation of a
response that culminates in cellular senescence or apoptosis
(4). From these data, we must conclude either that genomic
instability (but not mutations) plays a role in aging, or that
longevity is related to the cellular response of the cell to
such DNA damage.
CELLULAR RESPONSE TO PROTEIN DAMAGE
The integrity of the cellular proteome can also be viewed
as being dependent on caretakers and gatekeepers. Caretakers would perform all the activities that prevent protein
damage (ROS scavengers and chaperones, for example) or
take care of the damaged proteins (proteolysis and protein
repair mechanisms). In contrast, gatekeepers would function
to assess this damage and decide whether it can be repaired.
In analogy to genome gatekeepers, the ER/UPR response
can be viewed as a gatekeeper function for proteins, because
its activation can, under certain circumstances, lead to
programmed cell death (6).
The ER/UPR stress response is a cell-protective mechanism that can be triggered by accumulation of damaged or
Figure 1. Stress responses to irreversible damage. Ordinarily, most damage to
nuclear DNA (left) is efficiently repaired by a complex set of DNA repair
mechanisms. However, when the damage is of a nature or extent that makes its
repair impossible or overwhelming (irreversible damage, such as double-strand
breaks, for example), then the cell activates the DNA stress response, which
allows it further time and capabilities for repair. If repair is still not possible,
then this same pathway unfolds a cellular response that takes the cell out of the
dangerous path that could lead to cancer, by forbidding its further cell division.
This can be accomplished by entering either cell senescence or apoptosis (or any
other kind of programmed cell death). We propose that damage to proteins
(right) can follow a similar course: If the damage cannot be repaired or the
damaged proteins cannot be eliminated via degradation (irreversible damage),
then the ensuing accumulation of damaged proteins activates the endoplasmic
reticulum (ER) stress response, the function of which is to allow the cell further
time and capabilities for repair (by inducing chaperones, for example). If the
damage becomes overwhelming, then the cell enters into apoptosis, or possibly
cell senescence, although this last point has not been proven (dotted line).
misfolded proteins in any cell compartment, including the
nucleus (21). In young organisms at least, accumulation of
damaged proteins activates a series of sensors, including
PERK, IRE1, and ATF6 (reviewed in 6). Activation of these
pathways converges in the activation of ER chaperones,
including several glucose-regulated proteins (GRP), primarily GRP78 (22). The overall effects are: (i) to increase
the levels of available chaperones, (ii) to reduce the rate
of protein synthesis, and (iii) to activate several protein
degradation pathways. Altogether, these effects result in an
alleviation of the stress produced by the misfolded proteins
(23). However, the need for the cell to devote resources to
protect itself comes at the price of a partial disregard for
its more differentiated functions. This switch is mediated by
the overall decrease in the rate of new protein synthesis,
which occurs concomitantly with the upregulation of genes
encoding chaperones. This temporary change in priorities
GUEST EDITORIAL
results in saving the cell from deleterious environmental conditions, and therefore saving the tissue from
permanent loss of function (21). However, this shift also
means that the ER/UPR stress response is even better
equipped than is the DNA damage response to produce
tissue-wide dysfunction, because its activation results in loss
of differentiated tissue functions even if damage is not
overwhelming enough to activate the apoptosis or senescence pathways.
Because the ER stress response is a transient protective
mechanism that leads to the re-establishment of proper
homeostasis, under most circumstances, activation of the
ER stress response should not lead to the progressive
deterioration observed during aging. However, it is possible
that during aging, decreased efficiency of the response, or
other factors including persistent insults or a decreased
capacity to activate proteolytic pathways (24,25) might lead
to a sustained ER stress response. This sustained response
could be directly responsible for the loss of tissue function,
through the persistent diversion of cellular efforts away
from differentiated activities, as described above. Some
additional facts about aging and the ER stress response
become relevant at this point: (i) GRP levels increase with
age, and the increase is rapidly reversed by diet restriction
(26); (ii) Induction of heat shock protein (hsp) 70 family
members by a variety of stressors is diminished in old
organisms [reviewed in (27)]; (iii) Overexpression of
hsp70 family members extends life span in Caenorhabditis
elegans (28,29); (iv) One of the hallmarks of long-lived
organisms is their increased resistance to oxidative as well
as many other types of stress (30); and (v) Having said that,
cells under severe stress do die! We propose that it is this
cell death, in conjunction with the loss of functionality of
the surviving cells, that is at the root of the age-related
decline in tissue mass and function.
The main issue here is that stochastic DNA mutations or
random damage to proteins or other macromolecules will
not lead to a coordinated cellular response, and therefore,
will not lead to the phenotype of accelerated aging. In
contrast, activation of a programmed damage response will
lead to a decrease in cell and tissue function that does not
depend on which molecule was originally affected, and
therefore is no longer idiosyncratic. Instead, all the cells
subjected to the insult would have a concordant, preestablished response, which includes a decrease in differentiated
functions (Figure 2).
In summary, activation of the ER stress response will lead
to loss of tissue functionality both by diversion of energy
away from differentiated functions and, eventually, by a loss
in cell number. It logically follows that a loss in cell
numbers driven by a programmed cellular response should
proceed via programmed cell death, and not by necrosis.
Interestingly, it has been proposed (11) that the reason
Xpc/ mice fail to show a premature aging phenotype is
because the mice do not have an increased apoptotic
response, in contrast to the XpdTTC model, which is indeed
short-lived (14). Finally, it should be pointed out that, in
most tissues derived from young healthy organisms,
a decrease in the number of differentiated cells can be
compensated either by self-duplication (31) or by activation
559
of the stem cell pools, the asymmetric division of which
allows for the replenishment of the missing cells, while
still retaining the stem cell pool. There are several
indications in the literature that suggest that activation of
stem cells is suboptimal in aged individuals (32,33).
Although this is still a budding area in aging research, data
available suggest that aged organisms suffer from a variety
of deficits in their stem cells, both quantitative (decreasing
numbers of stem cells) and qualitative [decreased proliferation capacity, mobilization, and/or differentiation properties; reviewed in (34)]. It is possible that stem cells suffer
the same type of damage as do other cells, and activation of
a stress response in these cells also leads to their declining
functionality (or in some niches, increased death), just as in
other cellular compartments. Thus, stem cells represent
a niche where stress-induced loss of either cell function or
cell numbers could be of particular relevance to the aging
process. Because stem cells divide asymmetrically, it would
be interesting to establish whether they (and germline cells)
have more sophisticated ways to deal with damage, as has
been shown in asymmetrically dividing yeast cells (35). It
should be pointed out that, although this proposal invokes
a programmed response to cellular damage as the leading
cause of aging, it does not contradict evolutionary theory,
because the response did not evolve with the purpose of
producing aging, but rather, aging is the side effect of
a protective mechanism active in younger organisms. In that
sense, the ER/UPR stress response could represent a case
of antagonistic pleiotropy.
Conclusion
If aging is controlled not by how much damage an
organism sustains, but rather, by its ability to respond to
such damage, then it is reasonable to conclude that aging
and longevity should be controlled genetically, at least in
part. This prediction has been proven to be true in the last
decade or so. A few more direct, testable hypotheses can be
drawn from this proposal. For example, life span should
be affected by genetic manipulation of GRP or heat shock
proteins (28,29), as well as by manipulations that affect the
equilibrium that dictates whether a challenged cell goes
into apoptosis or senescence. Precise tuning of either the
threshold or the overall activity of cellular responses such
as the ER response should also correlate with life span in
organisms of different species, or among members of
a cohort. Another testable corollary is that apoptosis and/or
senescence pathways should be activated in those mouse
models which display accelerated senescence, but not in
models where damage is increased, but longevity is not
affected. Activation of these pathways has been shown in
a subset of the models, but more research will be necessary
before a general conclusion can be drawn.
ACKNOWLEDGMENTS
I thank Drs. Judith Campisi, Huber Warner, and Ueli Schibler for critical
reading of the manuscript.
Address correspondence to Felipe Sierra, PhD, Biology of Aging,
National Institute on Aging, National Institutes of Health, 7201 Wisconsin
Ave., Suite 2C231, Bethesda, MD 20892. E-mail: [email protected]
560
SIERRA
Figure 2. Outcomes of damage to macromolecules. Macromolecules can be damaged by a variety of different agents and mechanisms. Some of the agents most
studied by gerontologists are depicted (top). In general, the cell deals with this damage either by activating repair mechanisms or by degrading the affected molecules
(except in the case of nuclear DNA, which can be repaired, but not degraded). In most instances, persistence of unrepaired damage to individual molecules, including
DNA, will lead to an ‘‘idiosyncratic functional loss,’’ meaning that in any given cell, a small fraction of any kind of molecule will be damaged and nonfunctional.
However, because this damage is random (idiosyncratic), the tissue as a whole will not be significantly affected because each cell within the tissue will suffer damage to
a different macromolecule. By and large, this should not affect tissue functionality (happy face). An exception is that, in a small number of cases, persistent damage to
a small handful of genes could lead to a growth advantage for the affected cell, which if unchecked, could lead to cancer (sad face). As depicted in Figure 1, the inability
to repair or eliminate the damage activates a stress response (in the case of lipid damage, this is only proposed, as indicated by the question mark). The stress response
allows one more chance for repair (happy face), but the persistence of damage can lead to a redistribution of cellular resources or activation of the cell senescence
pathway, both of which will result in a generalized functional loss, and aging (sad face). If, in contrast, the response results in cell death, then tissue function can still be
restored by activation of stem cell proliferation and differentiation. If this last resort fails for any reason, then the functional loss will still lead to aging (sad face).
REFERENCES
1. Harman D. Aging: a theory based on free radical and radiation
chemistry. J Gerontol. 1956;11:298–300.
2. Maynard Smith J. A theory of ageing. Nature. 1959;184:956–957.
3. Vijg J. Impact of genome instability on transcription regulation of aging
and senescence. Mech Ageing Dev. 2004;125:747–753.
4. Campisi J. Senescent cells, tumor suppression, and organismal aging:
good citizens, bad neighbors. Cell. 2005;120:513–522.
5. Moskovitz J, Bar-Noy S, Williams WM, Requena J, Berlett BS,
Stadtman ER. Methionine sulfoxide reductase (MsrA) is a regulator of
antioxidant defense and lifespan in mammals. Proc Natl Acad Sci U S
A. 2001;98:12920–12925.
6. Rao RV, Bredesen D. Misfolded proteins, endoplasmic reticulum stress
and neurodegeneration. Curr Opin Cell Biol. 2004;16:653–662.
7. Lindahl T. Instability and decay of the primary structure of DNA.
Nature. 1993;362:709–715.
8. Wood RD, Mitchell M, Lindahl T. Human DNA repair genes. Mutat
Res. 2005;557:275–283.
9. Kinzler KW, Vogelstein B. Cancer susceptibility genes: gatekeepers
and caretakers. Nature. 1997;386:761–763.
10. Ono T, Uehara Y, Saito Y, Ikehata H. Mutation theory of aging, assessed in transgenic mice and knockout mice. Mech Ageing Dev. 2002;
123:1543–1552.
11. Wijnhoven SWP, Kool HJM, Mullenders LHF, et al. Age-dependent
spontaneous mutagenesis in Xpc mice defective in nucleotide excision
repair. Oncogene. 2000;19:5034–5037.
12. Van Remmen H, Qi W, Sabia M, et al. Multiple deficiencies in antioxidant enzymes in mice result in a compound increase in sensitivity
to oxidative stress. Free Radic Biol Med. 2004;36:1625–1634.
13. Vogel H, Lim DS, Karsenty G, Finegold M, Hasty P. Deletion of Ku86
causes early onset of senescence in mice. Proc Natl Acad Sci U S A.
1999;96:10770–10775.
GUEST EDITORIAL
14. De Boer J, Andressoo JO, de Wit J, et al. Premature aging in mice
deficient in DNA repair and transcription. Science. 2002;296:1276–
1279.
15. Schriner SE, Linford NJ, Martin GM, et al. Extension of murine life
span by overexpression of catalase targeted to mitochondria. Science.
2005;308:1909–1911.
16. Blasco MA, Lee HW, Hande MP, et al. Telomere shortening and tumor
formation by mouse cells lacking telomerase RNA. Cell. 1997;91:
25–34.
17. Du X, Shen J, Kugan N, et al. Telomere shortening exposes functions
for the mouse Werner and Bloom syndrome genes. Mol Cell Biol.
2004;24:8437–8446.
18. Chang S. A mouse model of Werner syndrome: what can it tell us about
aging and cancer? Int J Biochem Cell Biol. 2005;37:991–999.
19. Wong KK, Maser RS, Bachoo RM, et al. Telomere dysfunction and
Atm deficiency compromises organ homeostasis and accelerates
ageing. Nature. 2003;421:643–648.
20. Tyner SD, Venkatachalam S, Choi J, et al. p53 mutant mice that display
early ageing-associated phenotypes. Nature. 2002;415:45–53.
21. Kopito RR. Aggresomes, inclusion bodies and protein aggregation.
Trends Cell Biol. 2000;10:524–530.
22. Lee AS. Mammalian stress response: induction of the glucose-regulated
protein family. Curr Opin Cell Biol. 1992;4:267–273.
23. Ron D. Translational control in the endoplasmic reticulum stress response. J Clin Invest. 2002;110:1383–1388.
24. Cuervo AM, Dice JF. When lysosomes get old. Exp Gerontol. 2000;
35:119–131.
25. Keller JN, Dimayuga E, Chen Q, Thorpe J, Gee J, Ding Q. Autophagy,
proteasomes, lipofuscin, and oxidative stress in the aging brain. Int J
Biochem Cell Biol. 2004;36:2376–2391.
561
26. Spindler SR, Crew MD, Mote PL, Grizzle JM, Walford RL. Dietary
energy restriction in mice reduces hepatic expression of glucoseregulated protein 78 (BiP) and 94 mRNA. J Nutr. 1990;120:1412–1417.
27. Macario AJL, Conway de Macario E. Sick chaperones, cellular stress,
and disease. N Engl J Med. 2005;353:1489–1501.
28. Yokoyama K, Fukumoto K, Murakami T, et al. Extended longevity of
Caenorhabditis elegans by knocking in extra copies of hsp70F, a homolog of mot-2 (mortalin)/mthsp70/Grp75. FEBS Lett. 2002;516:53–57.
29. Walker GA, Lithgow GJ. Lifespan extension in C. elegans by a
molecular chaperone dependent upon insulin-like signals. Aging Cell.
2003;2:131–139.
30. Guarente L, Kenyon C. Genetic pathways that regulate ageing in model
organisms. Nature. 2000;408:255–262.
31. Dor Y, Brown J, Martinez OI, Melton DA. Adult pancreatic beta-cells
are formed by self-duplication rather than stem-cell differentiation.
Nature. 2004;429:41–46.
32. Bell DR, Van Zant G. Stem cells, aging, and cancer: inevitabilities and
outcomes. Oncogene. 2004;23:7290–7296.
33. Nishimura EK, Granter SR, Fisher DE. Mechanisms of hair graying:
incomplete melanocyte stem cell maintenance in the niche. Science.
2005;307:720–724.
34. Carrington JL. Aging bone and cartilage: cross-cutting issues. Biochem
Biophys Res Commun. 2005;328:700–708.
35. Aguilaniu H, Gustafsson L, Rigoulet M, Nystrom T. Asymmetric
inheritance of oxidatively damaged proteins during cytokinesis.
Science. 2003;299:1752–1753.
Received October 14, 2005
Accepted November 22, 2005
Decision Editor: James R. Smith, PhD