REVIEW etal muscle stem cells (MuSCs, or satellite cells)

play a minimal role in muscle maintenance but

vigorously engage in regeneration after injury.
Stem cells and healthy aging Hematopoietic stem cells (HSCs) and intestinal
stem cells (ISCs) do both, contributing to on-
going production of differentiated cells and also
Margaret A. Goodell1* and Thomas A. Rando2*
repairing tissue after injury.
During protected aging, the extent to which
Research into stem cells and aging aims to understand how stem cells maintain tissue
stem cells continue to maintain their cognate
health, what mechanisms ultimately lead to decline in stem cell function with age, and
tissues depends on their own health. Although
how the regenerative capacity of somatic stem cells can be enhanced to promote healthy
stem cells have characteristics (e.g., turnover rate,
aging. Here, we explore the effects of aging on stem cells in different tissues. Recent
a specialized niche) that may protect them from
research has focused on the ways that genetic mutations, epigenetic changes, and the
insults associated with aging, data also indicate
extrinsic environmental milieu influence stem cell functionality over time. We describe
that stem cells deteriorate with age (2). Further-
each of these three factors, the ways in which they interact, and how these interactions
more, any aberrations in stem cells may be carried
decrease stem cell health over time. We are optimistic that a better understanding of these
forward into their differentiated progeny, contrib-
changes will uncover potential strategies to enhance stem cell function and increase tissue
uting to tissue aging. As such, the question
resiliency into old age.
remains as to how effectively stem cell popula-

Downloaded from [Link] on December 3, 2015

T
tions maintain tissue health, what the limitations
he aging process remains one of the central mechanisms beyond it, likely account for the of that capacity are, and what the mechanisms
mysteries of biology, both from an evolu- progressive degradation of tissue function that are that ultimately lead to decline in stem cell
tionary perspective (why we age) and from characterizes protected aging. As such, what is function. Research in stem cells and aging is
a mechanistic perspective (how we age). generally called “aging” represents the inter- geared toward these questions, with one long-
Organismal aging is the failure of an in- section of a gradually failing system selected term goal being the maintenance or restoration
tegrated system that balances genetic programs for early growth and reproductive fitness with of youthful characteristics in aged somatic stem
for survival and reproduction. As reflected in the the cumulative effects of growth-suppressive mech- cells to promote healthy tissue aging. Here, we
“disposable soma” theory (1), resources that are anisms and acquired somatic insults. Modulation focus on three major areas of recent research in
available to an organism are allocated either to of any aspect of this network could potentially stem cell aging: genetic mutations, epigenetic
survival or to reproduction, both of which are accelerate or decelerate the process of aging. changes, and extrinsic factors. We also consider
essential for the propagation of the species as The ability of an organism to ensure healthy how these influences are interrelated, and how
evolved in the wild. However, additional factors function during adult life depends on homeo- in the future we might be able to enhance stem
come into play when species are protected from static mechanisms. In many organs of mature cell function and increase tissue resiliency into
extrinsic causes of mortality (such as predation, vertebrates, resident stem cells participate in tis- old age by modulating these factors.
starvation, and exposure), as is the case for modern sue maintenance and regeneration after injury,
Homo sapiens as well as animals on farms, in with variations in these roles across different Somatic mutations, stem cells, and age
zoos, and in the laboratory. In those cases, in- tissues. For example, neural stem cells (NSCs) are For decades, we have understood that environ-
dividuals within the species are more likely to important for ongoing generation of new neu- mental insults such as irradiation and xenobiotic
live far beyond the ages of their wild counter- rons in specific regions of the brain but play a exposure can lead to accumulation of somatic
parts, allowing the emergence of phenotypes of limited role in damage repair. In contrast, skel- mutations in a variety of tissues. Indeed, this
aging and age-related diseases that would rarely
if ever be manifest in the wild; we consider this
“protected aging.” Outside of evolutionary
One of the central features of protected aging Shaped by natural selection selective pressure
is prolonged survival beyond the ages of peak
reproductive fitness (Fig. 1). The fact that indi- Development and growth Adult reproductive years “Protected aging”
vidual members of any species rarely live beyond
this stage in the wild means that there would
Stem cell activity

have been no evolutionary pressure to select for

genetic mechanisms to assure maintenance of
somatic tissues into old age. The homeostatic
mechanisms that are necessary to preserve func-
tion throughout life are thus predicted to lose
robustness over time. Furthermore, for most spe-
cies, growth-suppressive mechanisms take over
from growth-promoting mechanisms around the
time of reproductive maturity. This critical tran- Age
sition, and the perpetuation of growth-suppressive
Fig. 1. Model of stem cell use over the life span. During embryogenesis and organismal growth, stem
cells are highly active and contribute to tissue formation and growth. During the prime reproductive phase,
1
Stem Cells and Regenerative Medicine Center, Center for growth is suppressed. Stem cells maintain and repair tissues. Properties of stem cells during these first
Cell and Gene Therapy, and Department of Pediatrics, Baylor
two phases would be subject to forces of natural selection, because these phases of survival and re-
College of Medicine, Houston, TX 77030, USA. 2Glenn Center
for the Biology of Aging and Department of Neurology and production would be critical to the propagation of the species. Beyond the period of reproductive maturity
Neurological Sciences, Stanford University School of as fecundity declines, which is also the period of “protected aging,” cell and tissue functions are predicted
Medicine, Stanford, CA 94305, USA, and Center for to be under little or no evolutionary pressure, both because they are predicted to have negligible effect on
Regenerative Rehabilitation, Veterans Administration Palo
species survival and also because, in the wild, survival beyond this point markedly diminishes. It is during
Alto Health Care System, Palo Alto, CA 94304, USA.
*Corresponding author. E-mail: goodell@[Link] (M.A.G.); this phase that stem cell functionality (although not necessarily stem cell number) declines in most
rando@[Link] (T.A.R.) tissues, in some cases precipitously.

SCIENCE [Link] 4 DECEMBER 2015 • VOL 350 ISSUE 6265 1199

 A GI N G

concept underlies views of increased Blood

cancer incidence with age. How-
ever, it has remained unclear how
the accumulation of somatic mu-
tations affects either organismal
aging or aging of stem cells. Recent
work from the hematopoietic sys-
tem has begun to shed some light
on these issues.
If somatically acquired mutations
are rare, most cells in the peripheral
blood should be virtually identical ~1000 HSCs active Collapse of diversity: one dominant clone
in their genome sequence; hence,
blood can be used to assess the Skin
“germline status” of an individu-
al’s genetic complement. However,
deep genome sequencing studies
investigating mutations that con-
tributed to leukemia development
revealed that normal blood cells of-
ten harbor passenger mutations Skin stem cells have limited domain Marked expansion of some clones
(unrelated to the leukemia) much
more frequently than expected (3). Young Aged
These data indicate that blood pro- Time
genitors acquire random mutations
Fig. 2. Stem cell diversity and dynamics with age. Top: Peripheral blood from young individuals is generated
constantly, on the order of 10 muta-
from around 1000 active stem cells. By the age of 70, the clonal diversity collapses, resulting in dominance of one
tions per HSC per year, many of
HSC clone, such that about 20% of individuals have one clone that dominates 20 to 80% of blood cell production.
which then appear in small frac-
Bottom: Representation of the surface of skin. Young skin is continuously replenished from stem cells, each with a
tions of differentiated blood prog-
highly restricted domain (represented by circles). Random mutations generate small variations across the surface
eny (3). When this deep sequencing
in terms of stem cells and their progeny (colored circles). With time, some clones expand markedly, resulting in
approach was extended to blood
clonally derived patches with a common set of genetic variants (18).
samples from tens of thousands of
individuals across many ages, a
striking pattern arose: Specific so-
matically acquired genetic variants were often third recurrently mutated gene, TP53, is the most the strongest association (5). Furthermore, aging-
present in a large fraction of blood cells (Fig. 2), frequently mutated gene across all cancers (14) associated diseases such as type 2 diabetes have
up to 70% in some cases (4–7). Although blood and its product, p53, is considered the “guardian also been correlated with blood-based genomic
cells were thought to be generated from ~1000 of the genome” because of its central role in re- aberrations (17). The mechanisms behind these
active stem cells in young adults (8), the ex- gulating cellular responses to stress and DNA correlations are not understood, but mutations
planation for the high proportion of blood cells damage. In mice, Tp53 has been shown to regu- in stem cell regulators likely affect the function
with particular somatic mutations was that a late HSC proliferation (15) and its deficiency leads of their progeny; HSCs are continuously gen-
single stem cell clone was dominating the gen- to age-related stem cell expansion (16). With each erating an array of immune cells, platelets, and
eration of the peripheral blood in some indi- HSC acquiring around 10 mutations per year, red blood cells, all of which have an impact on
viduals. This collapse of highly polyclonal into any individual stem cell will harbor 700 to 800 disease resistance, inflammation, clotting, and
quasi-monoclonal hematopoiesis increases with mutations in later decades of life (3). Most of tissue oxygenation. Because the blood system
age; between 5 and 20% of 70-year-olds show these mutations will be neutral, although some supports all tissues in the organism, any age-
clear evidence of this state, and almost all in- will have a deleterious effect, potentially leading related impairment of stem cells manifest in their
dividuals above the age of 90 are estimated to to the arrest or elimination of those cells. Thus, progeny could conceivably affect distant tissues
have a single dominant stem cell clone generat- mutations seen in advanced age are those com- and therefore healthy aging.
ing a substantial proportion of their blood (9). patible with the cell’s long-term survival. Is there evidence for aging-associated clonal
The notion of clonal collapse is revolutionizing Together, data from humans and mice suggest expansion of stem cells in other tissues, and does
our view of HSC dynamics with age, and there is that acquired somatic mutations can confer a this affect tissue heath? This phenomenon, al-
intense interest in understanding the potential growth or survival advantage to the target stem though not yet examined on a large scale, is
mechanisms of this phenomenon. One clue has cell, enabling its expansion and leading to a known to occur in other tissues. For example,
come from examining the genes that are most preponderance of its progeny in the blood with stem cells are generally responsible for generat-
commonly mutated in these large clones. Muta- age. What are the long-term health implications? ing the differentiated epithelial cells in a very re-
tions in about 20 genes are recurrently asso- The vast majority of individuals who exhibit clo- stricted surface area of young skin. However, in
ciated with clonal hematopoiesis. The top two, nal hematopoiesis will not develop hematologic normal aged Sun-exposed skin, marked expansion
DNMT3A and TET2, are epigenetic regulators malignancy during their remaining life span. of clones associated with specific mutations—
that control DNA methylation status. Both are Nonetheless, they are at significantly higher risk including TP53 and the stem cell regulator
frequently mutated across many hematologic of developing age-associated blood diseases such NOTCH1—occurs well beyond those original bound-
malignancies, and their loss leads to increased as myelodysplastic syndrome, aplastic anemia, aries (18). Remarkably, almost 20% of normal
numbers of stem and progenitor cells (promot- and leukemias (5) associated with the acquisition skin cells have these potent expansion-promoting
ing self-renewal) while hindering their ability of additional mutations. Moreover, individuals NOTCH1 mutations. The relatively low incidence
to differentiate in mouse models (10–13), al- with clonal hematopoiesis are at higher risk for of cancer acquisition despite the high frequency
though the precise mechanisms through which earlier mortality when all causes are considered, of clones bearing cancer-associated mutations
these mutations act is not well understood. The with myocardial infarctions and strokes having is a testament to the mechanisms that restrain

1200 4 DECEMBER 2015 • VOL 350 ISSUE 6265 [Link] SCIENCE

malignancy development. The effect of these gene expression, contributing to the decline of sively examined, current data support the concept
expanded mutant clones on age-related changes stem cell and tissue function with age. Genetic that epigenetic regulation erodes in aging stem
of tissue function (including the critical barrier evidence in model organisms has supported the cells. In concordance with the C. elegans findings,
function of the skin), and on organismal health notion that aberrant epigenetic regulation affects H3K4me3 tends to increase in aging HSCs, par-
generally, is unknown but warrants investigation, organismal aging. For example, in Caenorhabditis ticularly on the genes involved in maintaining
given the striking associations of clonal collapse elegans, a genetic screen revealed that loss of HSC identity (23). In both HSCs and MuSCs, the
in the hematopoietic system. function of Wdr5—a gene encoding a histone repressive H3K27me3 mark increased with age
Together, these and other studies demonstrate methyltransferase that leads to trimethylation (23, 24). In HSCs, this increase was associated
that somatic mutations that arise in stem cells of lysine 4 on histone 3 (designated as H3K4me3 with repression of some genes that direct specific
confer an advantage that leads to their expansion and generally considered to be a mark of ex- differentiation programs such as the lymphoid
within a tissue over many years. Many of the so- pressed genes)—led to extended life span (20). fate, known to diminish with age. In MuSCs,
matic mutations repeatedly observed are asso- Although H3K4me3 generally marks promoters, the increase was associated with repression of
ciated with cancer, and this may underlie the the histone mark covers the entire coding unit at genes encoding histone genes themselves (24);
correlation between tissue-specific cancer inci- a subset of genes controlling cellular identity; this this finding is of interest because of the relation-
dence and stem cell proliferation (19). More im- pattern is associated with very high gene expres- ship between histone gene expression and yeast
portant, we expect that clonal dominance in the sion and transcriptional elongation (21). However, replicative life span (25). Interestingly, mesenchy-
blood, and possibly in other tissues, can have broad research is needed to understand why reduction mal stem cells (MSCs) from aged individuals
effects on healthy aging of the cognate tissues. of H3K4me3 should be correlated with longer life show a decline in H3K9me3, a mark associated
span. Similar efforts in yeast have shown that with proper maintenance of heterochromatin.
Epigenetic erosion with age lower levels of another histone mark, H3K36me3, Loss of this mark is also found in MSCs har-
Epigenetic regulation refers to the mechanisms, reduced replicative life span, whereas ablating boring the Werner’s progeria syndrome muta-
mainly DNA methylation and histone modifica- genes that diminish the mark increased yeast tion (26), again linking epigenetic erosion with
tions, that license regions of the genome for ex- life span (22). Reduction of the H3K36me3 mark aging.
pression while shutting down others. There has was associated with transcriptional infidelity and Finally, age-related changes in DNA methyla-
been great interest in understanding the extent cryptic transcripts. tion have been examined in HSCs. DNA methyl-
to which erosion of these genome-scale regula- How do these and other findings relate to aging ation was decreased at genes associated with the
tory mechanisms leads to dysregulated control of in mammalian stem cells? Although not exten- promotion of self-renewal and was increased near
genes associated with differentiation (23, 27). Poly-
comb proteins are factors generally involved in
gene silencing. Regions with histones bearing
Transplantation Polycomb-associated marks (H3K27me3) tended
to become hypermethylated, a phenomenon noted
previously in other tissues in mouse and human
(28). Changes in methylation may play a role in
inhibiting the expression of tumor suppressor
genes, thereby increasing the possibility of malig-
nant transformation.
Overall, these epigenetic changes in aging stem
cells are consistent with the functional deficits
that have been repeatedly observed: With aging,
HSCs appear to increase in numbers but simul-
taneously lose differentiation capacity (23, 27).
Young mouse Nonetheless, the extent to which these changes
are correlative versus causal is not yet clear and
Old mouse merits further exploration. In HSCs and MuSCs,
the precise epigenetic regulation observed at young
ages appears to drift. This drift is aligned with
the general aging process: The identity and gen-
eral function of the stem cells remain the same,
Parabiosis
but they cannot regenerate their cognate tissues
quite as well as they did before.

Extrinsic factors affect aging stem cells

The influence of the local and systemic environ-
ment on stem cell function during protected
aging has been demonstrated by exposing young
stem cells to an aged environment, and vice
versa. These studies have used strategies such
Fig. 3. Intrinsic and extrinsic factors influence age-related changes in stem cell function. as heterochronic transplantation, in which cells
Distinguishing cell-intrinsic changes from cell-extrinsic changes (e.g., arising from the cellular environment) derived from a donor of one age are transplanted
in cell function has been aided by heterochronic studies. In heterochronic transplantation, stem cells into a recipient of a different age, or hetero-
isolated from either young or old donors are transplanted into young or old hosts, and cellular function chronic parabiosis, in which two mice of dif-
is then analyzed in these four-way comparisons (young into young, young into old, old into young, old ferent ages are adjoined to create a shared
into old). In parabiosis, animals are joined to promote the development of a single, shared circulatory circulatory system, thus exposing cells in one
system, thus exposing cells in one animal to the systemic environment of the other animal. In this case, animal to the systemic environment of the other
the cellular functions in young or old partners in the heterochronic pairs (young-to-old) are compared (29) (Fig. 3). When young stem cells were sub-
to those in isochronic pairs (young-to-young and old-to-old). jected to an aged systemic milieu by heterochronic

SCIENCE [Link] 4 DECEMBER 2015 • VOL 350 ISSUE 6265 1201

 A GI N G

V V V

C C Time

Fig. 4. Neutral drift of stem cell populations. Left: The small intestine comprises villi (v) containing differentiated cells that are replenished (arrows)
from the progeny of stem cells residing in the crypts (c). Right: When crypts are viewed from above, en face (as if cut on the dashed line), several stem
cell domains (each represented by a different color) are present. With age, the crypt stem cells continuously compete with each other, such that over
time, all the stem cells in an individual crypt originate from one stem cell. These presumptions are based on mouse data (44). Certain genetic alterations
and environmental conditions can accelerate this process.

parabiosis, they exhibited functional decline that decrease in factors that promote tissue repair, Studies of the murine small intestine have borne
resembled accelerated aging (30, 31). On the likely contributes to aging phenotypes in many out these ideas. Differentiated cells of the intes-
other hand, the converse was also true: Aged tissues. tinal villi are continuously replenished by stem
cells placed in a young environment or exposed cells that reside at their bases in specialized crypts.
to a youthful systemic milieu exhibited more Somatic natural selection of stem cells Each crypt contains several stem cells that com-
youthful characteristics, suggesting that it may From the studies discussed above, we can ap- pete to populate the adjacent villi. Even in the
be possible to ameliorate certain aging features. preciate at least three key influences on somatic absence of selection, individual stem cells overtake
These findings have led to the search for “age- stem cells within an aging organism. Stem cells an entire crypt in a completely random fashion, as
promoting” factors in old blood (see below) and acquire many somatic mutations over time, they predicted by neutral drift theory (Fig. 4) (44, 45).
for “youth-promoting” factors in young blood experience epigenetic drift, and they are bathed Mutations that even marginally increase prolif-
(32–34) [but see also (35)]. Together, these studies in a broader milieu that can negatively influence eration, such as in the K-ras oncogene, accelerate
demonstrate that stem cells are profoundly in- function. Among the cells in any given stem cell crypt clonality (46). In specific environmental con-
fluenced by their environment and imply that population, each cell brings its unique character- texts, some mutations appear particularly adaptive.
blood-borne factors may be responsible for at istics (e.g., mutations) and experiences (e.g., ex- Stem cells with a Tp53 mutation have no advan-
least some of the age-associated declines in stem posure to local cytokines) to the evolving adaptive tage in a normal crypt, but in an inflammatory
cell functionality. landscape. We propose that these forces interact environment that mimics colitis, the Tp53-mutant
Efforts to identify specific aging-associated cir- over time to result in selective pressure on indi- stem cells rapidly take over and their progeny
culating factors have repeatedly highlighted pro- vidual stem cells: Stem cells that have acquired, dominate production of the entire crypt (47).
inflammatory molecules, such as cytokines in the through somatic mutation or epigenetic drift, the Although not established experimentally, we
blood, as key drivers of cell and tissue aging (36). characteristics best adapted to the aged environ- speculate that similar forces lead to the emer-
One of the first “aging factors” identified by hete- mental milieu will become enriched in the popu- gence of clonal dominance as a feature of the
rochronic parabiotic studies was the cytokine lation, as a result of Darwinian-like natural aging hematopoietic system (Fig. 2). This hypoth-
CCL11 (31). The levels of this protein increase selection that occurs in vivo during aging (“so- esis would predict that few of the acquired mu-
with age in the blood, and administration of this matic natural selection”). These stem cells have tations in HSCs would confer an advantage in
cytokine into the circulation of young animals characteristics that confer optimal survival in the the young environment, allowing hematopoie-
led to a decline in NSC activity, as occurs during protected aging environment, regardless of other sis to remain highly polyclonal. By contrast, the
normal aging. Likewise, the related inflammatory functional capabilities. Any detrimental charac- same mutations, compounded by epigenetic drift,
cytokine Rantes was found to be elevated in the teristics they have assumed will be conveyed to could confer a distinct survival and/or prolifer-
HSC niche with age and to contribute to the age- the tissue via their progeny, in proportion to their ative advantage in the changing milieu of age,
related myeloid skewing in the hematopoietic relative abundance. allowing for the expansion of specific HSCs and
lineage (37). This view of clonal dynamics has both theo- clonal dominance. Indeed, in young mice, Tp53
In addition to cytokines, other immune system– retical and experimental foundations. Mathemat- knockout HSCs have no particular advantage,
associated molecules have been shown to change ical modeling of stem cell populations predicts but in old hosts they expand relative to their nor-
with age and promote aging phenotypes. Plasma that constant competition in a closed environ- mal counterparts (16). We speculate that DNMT3A
levels of the complement protein C1q were shown ment will lead to domination by one stem cell, and TET2 mutations may similarly confer an
to increase with age and to promote age-related even in the absence of any selection (so-called advantage in the aging environment. This view
MuSC decline by activation of the Wnt signaling “neutral drift”) (41), analogous to population drift is supported by the observation that some mu-
pathway (38, 39). Similarly, b2-microglobulin, a of species leading to fixation of traits even without tations associated with clonal hematopoiesis
component of the major histocompatibility com- selection, particularly in small populations (42). (e.g., in splicing factors) only become prevalent
plex, was found to be elevated in the blood of aged Moreover, certain mutations or environmental after the seventh decade (5, 7, 9), which suggests
mice and to contribute to an age-related decline pressures should accelerate this phenomenon, that the aging environment is particularly impor-
in NSC function (40). The increase in proinflam- selecting for stem cells with particular adaptions tant for the emergence of HSC clones with these
matory factors, along with the concomitant [akin to a changed fitness landscape (43)]. mutations. This model of continuous Darwinian

1202 4 DECEMBER 2015 • VOL 350 ISSUE 6265 [Link] SCIENCE

selection acting on variants across a population over, internal competition and selective forces gradual loss of cell and tissue health. Instead, a
in a changing landscape also applies to cancer may be less important. Even in the gut, stem cell variety of genetic, epigenetic, and environmental
development, an inherently aging-associated competition occurs almost exclusively within, and factors allow drift until the aging environment
phenomenon (48). not between, individual crypts. Thus, gaining a acts strongly enough to select for particular (usu-
Considered more broadly, highly dynamic broader understanding of the role that somatic ally detrimental) characteristics (Fig. 5). The
tissues such as the gut and bone marrow have stem cell competition plays across many tissues mutant stem cells that accumulate with age are
stem cells that, in effect, continuously compete will be important in the future. not causes of aging per se; they simply exploit the
with each other. Thus, any cell-intrinsic change, Finally, how do the environment and somatic aged environment to become dominant. In turn,
genetic or otherwise, that confers a growth or mutations interact together with epigenetic mod- the functional deficiencies they confer on their
survival advantage may lead to predominance of ulation? Epigenetic drift likely occurs even in the progeny contribute to the phenotypes associated
particular stem cells (Fig. 5). This is important absence of particular somatic mutations, driven with aging.
because the selective environment (adaptive land- by, and promoting adaption to, the changing en- With these views, what strategies or interven-
scape) changes with age. In aged organisms, vironmental milieu. The frequent selection for tions can be envisioned to extend healthspan by
factors such as systemic inflammation may offer mutations in epigenetic regulators (e.g., DNMT3A targeting stem cells? Two approaches emerge
an advantage to stem cells with particular char- and TET2) may suggest that these afford a degree naturally from the studies discussed. First, it
acteristics. The stem cells that respond best in an of epigenetic plasticity that hastens adaption to the seems as if the phenotypes of aging stem cells
aged or injured environment may not be the most aged environment. may be at least partially reversible. As noted
effective at regenerating healthy tissue if their above, heterochronic parabiosis studies suggest
progeny also bear somatic mutations or exhibit Conclusions that factors in young blood might partially ame-
epigenetic drift. These general principles are Taken together, recent studies on somatic muta- liorate the functional deficits of aged stem cells
likely in effect throughout many tissues, albeit tions, epigenetic drift, and the environmental (30, 31). Furthermore, the injection of plasma
manifesting differently depending on factors influences on stem cells usher in a new view of from young mice into the circulation of aged
such as tissue turnover rate, local interactions aging and the challenges to preserving healthy mice has recently been shown to induce a more
among stem cells, physical constraints on cell tissue function over time. After organismal growth youthful state of cells in the brain of the old
interchange, and the magnitude of alterations ceases, stem cells effectively maintain tissues animal (32). These findings indicate that at least
in the adaptive landscape. For example, in tissues through the peak reproductive years. Subse- some aspects of cellular aging may be reversible,
such as skeletal muscle, in which there is thought quently, there are no effective mechanisms that perhaps through reprogramming of the epige-
to be lower stem cell interchange and tissue turn- have been evolutionarily selected to preclude the nome (49). Indeed, interventions that clearly ex-
tend organismal life span even when applied late
in life, such as rapamycin treatment (50), may
enhance stem cell function in aged animals. As
such, it may be that treatments that directly
enhance the function of aged stem cells do so
by acting on the epigenome to adopt a more
Adaptive landscape youthful state.
Typical aging; Is it possible to reduce the acquisition of
clonal dominance somatic mutations? Most are probably the in-
evitable consequence of cell division and de-
Possible amination events that result in C→T transitions
interventions (51, 52).
Assuming we cannot eliminate mutations
altogether, another approach would be to alter
the adaptive landscape so as to select for more
functional cells. By the time that the protected
aging phase begins, stem cells will already have
Diversity maintenance acquired a burden of somatic mutations (a largely
inevitable consequence of cell division) and have
drifted epigenetically. We suggest that monitoring
Clonal switch changes and modulating the environment early—
such as reducing inflammatory mediators and
Young Aged otherwise slowing the transition of the systemic
Time environment that occurs with age—may limit
Fig. 5. Model of age-related selection for stem cells with new characteristics and potential out- the development of clonal dominance, allowing
comes. In young individuals, a polyclonal population of HSCs gives rise to a heterogeneous population the polyclonal state to be sustained longer (Fig.
of blood cells. With age, stem cells acquire somatic mutations and experience epigenetic drift. 5). Alternatively, providing a new adaptive
Concurrently, attrition (“collapse”) of some clones occurs and the adaptive landscape gradually shifts. landscape in which different stem cell variants
This may be caused in part by changes in humoral factors (such as circulating cytokines or inflammatory are better adapted to thrive could likewise con-
factors) or changes in the cellular environment that regulates the stem cells, the so-called niche (change tribute to the maintenance of tissue health. Along
could occur in niche composition or behavior). The result of a changing landscape and clonal attrition these lines, it is interesting that genetic manipu-
may be a population bottleneck that provides an opportunity for clones with a selective advantage to lations early in life that preserve proliferative
expand. Ultimately, this leads to quasi-monoclonality and to the dominance of particular clones. In prin- homeostasis of gut stem cells in Drosophila lead
ciple, it could be possible to intervene to alter the forces that drive toward deleterious quasi-monoclonality to life-span extension (53). It may be that early
(dashed arrows). Any intervention that suppresses the change in the adaptive landscape would tend to life treatments that extend life span, such as
preserve the healthy polyclonality (diversity maintenance) of youth. Likewise, understanding any growth caloric restriction, do so in part by altering the
or survival advantages conferred upon HSCs with “nondeleterious” somatic mutations or epigenetic adaptive landscape to prevent detrimental clo-
changes could allow for a rational modulation (“clonal switch”) of the adaptive landscape to select for nal dominance and preserve tissue function.
those clones rather than deleterious clones. We are optimistic that better understanding of

SCIENCE [Link] 4 DECEMBER 2015 • VOL 350 ISSUE 6265 1203

 A GI N G

the mechanisms of stem cell dysregulation and REVIEW

selection with age will enable new rational inter-
ventions based on these principles.
Mitochondrial dysfunction and
RE FE RENCES AND N OT ES
1. T. B. Kirkwood, Cell 120, 437–447 (2005).
2. L. Liu, T. A. Rando, J. Cell Biol. 193, 257–266
longevity in animals:
(2011).
3. J. S. Welch et al., Cell 150, 264–278 (2012).
4. M. Xie et al., Nat. Med. 20, 1472–1478 (2014).
Untangling the knot
5. S. Jaiswal et al., N. Engl. J. Med. 371, 2488–2498
(2014).
Ying Wang and Siegfried Hekimi*
6. G. Genovese et al., N. Engl. J. Med. 371, 2477–2487
(2014). Mitochondria generate adenosine 5´-triphosphate (ATP) and are a source of potentially
7. T. McKerrell et al., Cell Rep. 10, 1239–1245 (2015). toxic reactive oxygen species (ROS). It has been suggested that the gradual mitochondrial
8. S. N. Catlin, L. Busque, R. E. Gale, P. Guttorp, J. L. Abkowitz,
Blood 117, 4460–4466 (2011).
dysfunction that is observed to accompany aging could in fact be causal to the aging
9. T. McKerrell, G. S. Vassiliou, Sci. Transl. Med. 7, 306fs38 process. Here we review findings that suggest that age-dependent mitochondrial
(2015). dysfunction is not sufficient to limit life span. Furthermore, mitochondrial ROS are not
10. G. A. Challen et al., Nat. Genet. 44, 23–31 (2012). always deleterious and can even stimulate pro-longevity pathways. Thus, mitochondrial
11. C. Quivoron et al., Cancer Cell 20, 25–38 (2011).
12. K. Moran-Crusio et al., Cancer Cell 20, 11–24 (2011).
dysfunction plays a complex role in regulating longevity.
13. M. Ko et al., Proc. Natl. Acad. Sci. U.S.A. 108, 14566–14571

T
(2011).
14. C. Kandoth et al., Nature 502, 333–339 (2013). he primary and most essential function of the implications of these observations for the
15. Y. Liu et al., Cell Stem Cell 4, 37–48 (2009). mitochondria is to produce energy for aging process, a key question is whether the
16. M. Dumble et al., Blood 109, 1736–1742 (2007). the cell. The oldest explanation for aging, observed damage and dysfunction are severe
17. A. Bonnefond et al., Nat. Genet. 45, 1040–1043
(2013).
the rate-of-living theory, postulates that enough to cause the other degenerative pheno-
18. I. Martincorena et al., Science 348, 880–886 aging and life span are regulated by the types of aging.
(2015). rate of energy metabolism, with lower rates There is no doubt that mitochondrial dysfunction
19. C. Tomasetti, B. Vogelstein, Science 347, 78–81 leading to longer life spans. However, the ap- can severely damage the organism. Human patients
(2015).
20. E. L. Greer et al., Nature 466, 383–387 (2010).
peal of the rate-of-living theory has been weak- with mutations in mitochondrial DNA (mtDNA) or
21. K. Chen et al., Nat. Genet. 47, 1149–1157 (2015). ened by its failure to accurately predict the in nuclear genes coding for proteins that function in
22. P. Sen et al., Genes Dev. 29, 1362–1376 (2015). observed relationships between energy ex- the mitochondrial ETC are generally severely affect-
23. D. Sun et al., Cell Stem Cell 14, 673–688 (2014). penditure and life span. This is not to say that ed. They often show multisystem disorders that
24. L. Liu et al., Cell Rep. 4, 189–204 (2013).
25. J. Feser et al., Mol. Cell 39, 724–735 (2010).
mitochondrial and energy metabolism don’t include myopathy, encephalopathy, stroke, and
26. W. Zhang et al., Science 348, 1160–1163 (2015). play a crucial role in aging, but their relation- hearing loss (2). Most mitochondrial disorders
27. I. Beerman et al., Cell Stem Cell 12, 413–425 ship to aging might not be simple. Mitochondria present with neurological and muscular symp-
(2013). do much more than produce energy. Particularly toms. It is thus generally postulated that cells
28. S. Maegawa et al., Genome Res. 20, 332–340
(2010).
relevant to aging, the mitochondrial electron with high energy demands, such as those in the
29. I. M. Conboy, T. A. Rando, Cell Cycle 11, 2260–2267 transport chain (ETC) leaks electrons and gen- central nervous system and muscles, are more
(2012). erates reactive oxygen species (ROS) during susceptible to the reduced energy output of de-
30. I. M. Conboy et al., Nature 433, 760–764 (2005). normal respiration. Thus, a potentially harm- fective mitochondria and are consequently more
31. S. A. Villeda et al., Nature 477, 90–94 (2011).
32. S. A. Villeda et al., Nat. Med. 20, 659–663 (2014).
ful elevation of ROS production occurs when strongly affected by mitochondrial impairment.
33. C. Elabd et al., Nat. Commun. 5, 4082 (2014). the ETC function is perturbed. The mitochon- There is, however, considerable clinical variabil-
34. M. Sinha et al., Science 344, 649–652 (2014). drial free-radical theory of aging posits that ity among mitochondrial disease patients, and
35. M. A. Egerman et al., Cell Metab. 22, 164–174 biological aging results from the production of some mutations only affect particular tissues, re-
(2015).
36. C. Franceschi et al., Mech. Ageing Dev. 128, 92–105
ROS and the ensuing damage. However, direct flecting a diversity of distinct disease mechanisms
(2007). manipulation of cellular ROS levels within the that are still poorly understood. To understand
37. A. V. Ergen, N. C. Boles, M. A. Goodell, Blood 119, 2500–2509 biologically meaningful range does not accel- these conditions, a variety of mouse knockout
(2012). erate aging or decrease life span. Here we review (KO) models have been developed for nuclear-
38. A. S. Brack et al., Science 317, 807–810 (2007).
39. A. T. Naito et al., Cell 149, 1298–1313 (2012).
the relationships between normal mitochondrial encoded mitochondrial proteins (3). These include
40. L. K. Smith et al., Nat. Med. 21, 932–937 (2015). function, mitochondrial dysfunction, ROS genera- mutants carrying KO mutations in genes that
41. A. M. Klein, B. D. Simons, Development 138, 3103–3111 tion, and life span. are required for the assembly and function of
(2011). ETC complexes, mutants with defects in the
42. S. Wright, Am. Nat. 63, 556–561 (1929). Deleterious mitochondrial dysfunction
43. S. Wright, Proc. Sixth Int. Congr. Genet. 1, 356–366
production of mobile electron carriers [cyto-
(1932). Numerous studies have described damage to chrome c and ubiquinone (UQ)], and mutants
44. H. J. Snippert et al., Cell 143, 134–144 (2010). mitochondria in aged cells and organisms, in- lacking necessary factors for the maintenance
45. C. Lopez-Garcia, A. M. Klein, B. D. Simons, D. J. Winton, cluding in human samples. This damage in- of mitochondrial dynamics or the integrity of
Science 330, 822–825 (2010).
46. H. J. Snippert, A. G. Schepers, J. H. van Es, B. D. Simons,
cludes a gradual decline in respiratory chain mtDNA. In virtually every case, complete germline
H. Clevers, EMBO Rep. 15, 62–69 (2014). capacity, decreased activities of individual ETC KO causes embryonic to perinatal lethality.
47. L. Vermeulen et al., Science 342, 995–998 (2013). complexes, elevated oxidative damage, decreased Tissue-specific conditional KOs, mostly targeted
48. A. I. Rozhok, J. DeGregori, Proc. Natl. Acad. Sci. U.S.A. 112, mitochondrial content, morphological abnor- to neurons or muscles, result in abnormal
8914–8921 (2015).
49. T. A. Rando, H. Y. Chang, Cell 148, 46–57 (2012).
malities in mitochondrial structure, and in- mitochondria with severe deficits in respira-
50. D. E. Harrison et al., Nature 460, 392–395 (2009). creased fragility of aged mitochondria during tory chain function, giving rise to a variety of
51. L. B. Alexandrov et al., Nature 500, 415–421 experimental isolation (Fig. 1) (1). In exploring disease phenotypes. Most show severe progres-
(2013). sive loss of tissue function, such as progressive
52. S. Behjati et al., Nature 513, 422–425 (2014).
53. B. Biteau et al., PLOS Genet. 6, e1001159 (2010). Department of Biology, McGill University, Montreal, Quebec
skeletal muscle weakening, movement im-
H3A 1B1, Canada. pairment, and neurobehavioral abnormalities.
10.1126/science.aab3388 *Corresponding author. E-mail: [Link]@[Link] All result in death within the first year of life,

1204 4 DECEMBER 2015 • VOL 350 ISSUE 6265 [Link] SCIENCE

 Stem cells and healthy aging
Margaret A. Goodell and Thomas A. Rando
Science 350, 1199 (2015);
DOI: 10.1126/science.aab3388

This copy is for your personal, non-commercial use only.

If you wish to distribute this article to others, you can order high-quality copies for your
colleagues, clients, or customers by clicking here.

Downloaded from [Link] on December 3, 2015

Permission to republish or repurpose articles or portions of articles can be obtained by
following the guidelines here.

The following resources related to this article are available online at

[Link] (this information is current as of December 3, 2015 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
[Link]
A list of selected additional articles on the Science Web sites related to this article can be
found at:
[Link]
This article cites 52 articles, 18 of which can be accessed free:
[Link]
This article has been cited by 1 articles hosted by HighWire Press; see:
[Link]
This article appears in the following subject collections:
Cell Biology
[Link]

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
2015 by the American Association for the Advancement of Science; all rights reserved. The title Science is a
registered trademark of AAAS.