REPORTS

aged animals were myogenic (Fig. 2A). After

Increased Wnt Signaling During another 1½ days in culture, the percentage of
nonmyogenic cells from the young cultures
Aging Alters Muscle Stem Cell remained ~1%, but the percentage from the aged
fibers had increased to ~17%, and they exhibited

Fate and Increases Fibrosis morphological changes characteristic of a fibro-

blastic lineage (Fig. 2B). No loss of cells due to
detachment or apoptosis was detected, and the
Andrew S. Brack,1 Michael J. Conboy,1 Sudeep Roy,1 Mark Lee,2 Calvin J. Kuo,2 proliferation of nonmyogenic cells from aged
Charles Keller,3 Thomas A. Rando1,4* muscle was, if anything, lower than that from
young muscle (fig. S2). Therefore, the increase in
The regenerative potential of skeletal muscle declines with age, and this impairment is associated the percentage of fibrogenic cells in cultures from
with an increase in tissue fibrosis. We show that muscle stem cells (satellite cells) from aged mice aged mice most likely arose by conversion of
tend to convert from a myogenic to a fibrogenic lineage as they begin to proliferate and that previously myogenic cells into nonmyogenic
this conversion is mediated by factors in the systemic environment of the old animals. We also cells.
show that this lineage conversion is associated with an activation of the canonical Wnt signaling With age, there is a decline in satellite cell
pathway in aged myogenic progenitors and can be suppressed by Wnt inhibitors. Furthermore, functionality (5). That aged muscle regeneration
components of serum from aged mice that bind to the Frizzled family of proteins, which are can be enhanced by direct activation of the Notch
Wnt receptors, may account for the elevated Wnt signaling in aged cells. These results indicate pathway (5) or by exposure to a youthful sys-
that the Wnt signaling pathway may play a critical role in tissue-specific stem cell aging and temic environment (6) indicates that those func-
an increase in tissue fibrosis with age. tional changes are largely reversible. To test for
reversibility of the age-related myogenic-to-
ging of skeletal muscle is characterized in mediating the fibrotic responses of aged fibrogenic conversion, we examined satellite cell

A by an increase in fibrous connective

tissue (1) and an impairment of muscle
regenerative potential (2, 3), manifested by a
tissues.
We examined cells derived from muscles of
young (~6-month-old) and aged (~24-month-
progeny from young and aged mice in hetero-
chronic parabiotic pairings (6). Aged tissues
exposed to this heterochronic systemic environ-
replacement of muscle by fibrous connective old) mice for any characteristics that might ment exhibited a reduction in the myogenic-to-
tissue and adipose tissue (fig. S1). In the account for the age-related fibrotic response. In fibrogenic conversion (from 17 to 10%), whereas
muscular dystrophies, there is also progressive single-fiber cultures, we found 99% of the fiber- the younger tissues exhibited an increase (from
muscle fibrosis with age (4). We examined the associated mononucleated cells appeared to be ~1 to 9%) (Fig. 2C). We also exposed young and
cellular and molecular mechanism of this age- myogenic as defined by expression of combina- aged cells in vitro to serum from young or aged
dependent increase in skeletal muscle fibrosis. tions of myogenic markers (Pax7, MyoD, and animals (“young serum” and “aged serum,” re-
The regeneration of aged muscle can be Desmin) 2 days after fiber isolation from young spectively). Aged serum increased the myogenic-
enhanced by exposure to a youthful systemic animals, and 98% of the cells from fibers from to-fibrogenic conversion of young cells, whereas
environment, an effect mediated at least in part
by restoration of normal signaling of the Delta- Fig. 1. Prevention of age-
Notch pathway (5, 6). Therefore, we tested related increase in fibro-
whether exposure of old tissue to a youthful genesis during muscle
systemic environment, established by parabiotic regeneration by hetero-
pairings of old animals to young animals chronic parabiosis. (A) Iso-
(heterochronic pairings) (7), might also reduce chronic or heterochronic
the fibrotic response of old muscle. Indeed, there parabiotic pairs were es-
was a reduction of collagen deposition in tablished for 4 weeks, and
regenerating areas in muscles of aged mice in hind limb muscles were
heterochronic pairings compared with that of subjected to freeze inju-
aged mice in isochronic pairings (pairings of ries. Bromodeoxyuridine
mice of the same age), and this was accompanied (BrdU) was injected intra-
by enhanced proliferation of myogenic progeni- peritoneally 2 days after in-
tors (Fig. 1, A and B). Conversely, collagen jury. Three days later,
deposition was increased and progenitor prolif- muscles were sectioned
eration was reduced in young partners of and (left) stained with
heterochronic pairings compared with young Gomori trichrome (red, mus-
partners in isochronic pairings. Thus, systemic cle fibers; green, connec-
tive tissue) or (right)
influences that change with age are important
immunostained for BrdU
1
Department of Neurology and Neurological Sciences, (green). 4′,6′-Diamidino-
Stanford University School of Medicine, Stanford, CA 2-phenylindole (DAPI) (blue)
94305, USA. 2Department of Medicine, Division of Hema- labels all nuclei. (B) Quanti-
tology, Stanford University School of Medicine, Stanford, tative analyses of histologic
CA 94305, USA. 3Department of Cellular and Structural studies as in (A). (Left)
Biology, The University of Texas Health Science Center,
Fibrotic index (percentage
San Antonio, TX 78229, USA. 4Geriatric Research, Education,
and Clinical Center (GRECC) and Neurology Service, Veterans of the injury area occupied
Affairs (VA) Palo Alto Health Care System, Palo Alto, CA by connective tissue in
94304, USA. Gomori-stained sections). (Right) Proliferative response (percentage of DAPI+ mononucleated cells that were
*To whom correspondence should be addressed. E-mail: also BrdU+). “Young” and “aged” refer to young and aged mice, respectively; “iso” and “hetero” to isochronic
[email protected] and heterochronic pairings, respectively. (*P < 0.05; **P < 0.01.)

www.sciencemag.org SCIENCE VOL 317 10 AUGUST 2007 807

REPORTS
young serum had the opposite effect on aged When myogenic progenitor cultures from these muscle expressed more Axin2 than did such
cells (Fig. 2D). Nearly 100% of young cells tamoxifen-treated mice were incubated in young cells from young muscle (Fig. 3A). In addition,
previously maintained in young serum expressed serum, all b-gal+ cells had a myogenic pheno- analysis of TOPGAL mice [in which b-gal
myosin heavy chain under differentiation con- type (either MyoD+ or Pax7+, and ER-TR7–). expression is a read-out of Wnt signaling (11)]
ditions, but that declined to ~75% in cultures of However, in aged serum, 18% of b-gal+ cells revealed a progressive increase in Wnt signaling
young cells previously maintained in aged serum were Pax7– and MyoD–, and 10% were ER-TR7+ in myogenic cells during aging (Fig. 3B).
(fig. S3). Together, these results suggest that (Fig. 2, E and F), which confirmed that the aged We also analyzed other components of the
myogenic progenitors tend to deviate from their environment promotes a myogenic-to-fibrogenic canonical Wnt signaling cascade, glycogen syn-
myogenic lineage in the aged environment. conversion. thase kinase 3b (GSK3b) and its substrate b-catenin.
We used genetic lineage tracing to confirm Activation of the Wnt pathway can lead to In aged satellite cells, fluorescence-activated
that the aged systemic environment promotes a fibrogenic conversion of cells in other lineages cell sorting (FACS) analysis demonstrated that
myogenic-to-fibrogenic conversion. To identify (8, 9). Thus, we examined markers of the the amounts of active GSK3b decreased and
fibrogenic cells, we used an antibody to a steady-state activation of the Wnt pathway in active b-catenin increased (Fig. 3C), both changes
fibroblast-specific marker (ER-TR7) that was muscle and purified satellite cells (fig. S6) from being indicative of active Wnt signaling (12, 13).
highly specific (0% of Pax7+ or MyoD+ cells young and aged mice, as well as the effects of Furthermore, these changes were due to Wnt
were ER-TR7+) and highly sensitive [93% of modulating Wnt signaling on myogenic-to- signaling, because systemic adenoviral-mediated
nonmyogenic cells (Pax7– and MyoD–) were fibrogenic conversion and the fibrotic response. expression of Dickkopf-1 (DKK1) (14), a Wnt
ER-TR7+] (fig. S4). For myogenic lineage We analyzed a direct downstream target of antagonist, decreased the percentage of myo-
studies, we used Pax7.Cre-ER.ROSA26 mice, Wnt signaling, Axin2 (10), in uninjured muscle genic progenitors expressing active b-catenin
a strain in which tamoxifen administration leads from young and aged mice. Axin2 transcript (fig. S7B).
to permanent b-galactosidase (b-gal) expression levels were increased in aged muscle (fig. S7A). To analyze the Wnt signaling cascade in
only in myogenic cells in the adult (fig. S5). Furthermore, purified satellite cells from aged myogenic progenitors during muscle regenera-

Fig. 2. Influences of age and systemic environment on muscle stem cell

fate. (A) Myogenic progenitors were obtained from single muscle fibers
from young or old mice. A panel of antibodies to myogenic markers
(Desmin, Pax7, and MyoD) was used to assess myogenic phenotype at an
early stage of satellite cell activation (2 days after isolation). (B) The fate of fiber-
associated cell progeny at a later stage of satellite cell activation (3 days after
isolation). An antibody to fibronectin (“FN”) highlights the morphology of the
fibroblastic cells. (C) Relative abundance of nonmyogenic cells (Pax7–/MyoD– or
Desmin–/MyoD–) in “late” cultures [as in (B)] derived from myofibers obtained
from mice in parabiotic pairings. Axis labels as in Fig. 1B. (*P < 0.05; **P <
0.01.) (D) Analysis of cell fate in vitro after exposure to young or aged serum.
Pure myogenic progenitor cultures from young or aged mice were incubated in
plating media for 1½ days, exposed to young or aged serum for 1½ days, and analyzed for the percentage of nonmyogenic cells. (*P < 0.05; **P <
0.01.) (E) Effect of young or aged serum on myogenic stem cell fate. Myogenic progenitors isolated from Pax7.Cre-ER.ROSA26 mice were incubated in
young or aged serum for 2 days. Cells were stained for b-gal with 5-bromo-4-chloro-3-indolyl b-D-galactopyranoside (X-gal, blue) and with antibodies
to MyoD (green) and Pax7 (red). Arrow indicates a cell that has undergone myogenic-to-fibrogenic conversion. (F) Myogenic progenitors isolated from
Pax7.Cre-ER.ROSA26 mice were incubated in young or aged serum for 2 days. Cells were stained with an antibody to b-gal (red) or to a fibroblast-
specific marker, ER-TR7 (green). White arrows in the bottom panel indicate previously myogenic cells showing early fibrogenic phenotype.

808 10 AUGUST 2007 VOL 317 SCIENCE www.sciencemag.org

 REPORTS
tion in vivo, we isolated progenitors 2 days after from aged TOPGAL mice (Fig. 3D and fig. S7, We tested whether the observed hetero-
injury of young and aged muscle and analyzed D and E). When we incubated myogenic pro- chronic parabiotic effects (Fig. 1) were associ-
them by FACS. The percentage of progenitors with genitors from TOPGAL mice in young or aged ated with corresponding changes in the Wnt
detectable levels of active b-catenin was increased serum (Fig. 3E), Wnt signaling was increased in signaling pathway. Indeed, in heterochronic
in aged myogenic progenitors (fig. S7C). Increased progenitors exposed to aged serum, and this in- pairings, cells from the aged partners exhibited
b-gal activity was observed in injured areas of crease was prevented by a soluble Wnt inhibitor, decreased Wnt signaling, and cells from the
muscle and also in isolated muscle progenitors Frizzled-related protein 3 (sFRP3). young partners showed increased Wnt signaling
compared with that seen in cells in isochronic
controls (Fig. 3F). Thus, circulating factors in
aged animals appear to convey signals that result
in enhanced Wnt signaling.
To test for the presence of components of
serum capable of activating Wnt signaling by
binding to Frizzled receptors, we used a chi-
meric Wnt receptor fusion protein, Frizzled-Fc,
to deplete the serum of such activity (fig. S8).
Serum was incubated with conjugated Frizzled-
Fc or immunoglobulin IgG-Fc control and sub-
sequently tested for its effects on a cell line
(LSL) in which luciferase expression is de-
pendent on activation of the Wnt pathway (15).
When incubated with Frizzled-Fc, the activity in
aged serum that promoted Wnt signaling de-
creased; there was no change in the activity of
young serum subjected to the same treatment
(Fig. 3G).
We altered Wnt signaling experimentally to
test directly for its effects on cell fate and muscle
regeneration. Addition of Wnt3A protein to young
serum resulted in an increased myogenic-to-
fibrogenic conversion of young progenitors in
vitro. Conversely, the myogenic-to-fibrogenic
conversion by aged serum was abrogated by
Wnt inhibitors (Fig. 4A).
In vivo, the injection of Wnt3A into young
regenerating muscle 1 day after injury resulted in
increased connective tissue deposition (Fig. 4B),
phenotypically similar to regenerating aged mus-
cle (fig. S1). Exogenous Wnt also reduced cel-
lular proliferation in young regenerating muscles
(fig. S9A). We therefore tested whether inhibiting
Wnt signaling in aged muscle would reduce fi-
brosis and enhance muscle regeneration. Indeed,
there was reduced fibrosis in aged regenerating
muscle injected with DKK1, whereas no change
was observed in young muscle similarly treated
Fig. 3. Enhanced Wnt signaling in aged muscle and in myogenic progenitors exposed to aged serum. (A) (Fig. 4C). Injection of sFRP3 also enhanced myo-
Axin2 transcript levels assessed by real-time reverse transcription polymerase chain reaction from satellite genic progenitor proliferation in aged muscle
cells obtained by FACS-sorted Syndecan4+ cells (fig. S6) from uninjured muscles of young and aged mice. (fig. S9, B and C).
(B) b-Gal activity (normalized to DNA content) in fiber-associated cells isolated from uninjured muscle of These findings demonstrate that, with
TOPGAL mice of different ages. b-Gal activity of aged-matched wild-type controls was subtracted to age, the systemic environment is less effec-
normalize for any endogenous activity. (*P < 0.05.) (C) FACS analysis of active GSK3b and b-catenin tive in maintaining the myogenic fate of mus-
(GSK3b* and b-catenin*, respectively) in myogenic progenitors isolated from myofibers cultures from young cle stem cells and, instead, facilitates conversion
and aged animals. Graphs show the percentage of all myogenic progenitors (Syn-4+) that were also positive to a fibrogenic fate. In vivo, this is associated
for GSK3b* or b-catenin*. (*P < 0.05.) (D) Muscles of young and aged TOPGAL mice were injured, and the
with impaired muscle regeneration and an en-
muscles were analyzed 2 days later. Cryosections were incubated with a b-gal substrate (X-gal) that is blue
hanced fibrotic response. These effects are as-
after enzymatic conversion (top) and DAPI (bottom). X-gal staining in aged, wild-type littermate showed
negligible endogenous b-gal activity. (E) Analysis of Wnt signaling activity in myogenic progenitors derived sociated with increased Wnt signaling in the
from muscle fibers of TOPGAL mice. Cells were incubated in young or aged serum for 17 hours in the myogenic progenitors, possibly resulting from
presence or absence of a soluble Wnt inhibitor, sFRP3. The b-gal activity was quantified and normalized to increased amounts of Wnt or Wnt-like mole-
DNA content. The data are presented relative to the levels of b-gal activity in cells exposed to young serum. cules in the serum of aged animals. This gen-
(*P < 0.05.) (F) Analysis of GSK3b* in myogenic progenitors isolated from parabiotic pairs. Labeling as eralized role of Wnt signaling in promoting an
in Fig. 1B. (*P < 0.05; **P < 0.01.) (G) Wnt reporter gene expression in LSL cells exposed to serum aging phenotype is consistent with the find-
obtained from young and aged mice and incubated with agarose beads conjugated with chimeric Frizzled ings of Liu et al. from studies of the role of
receptors (Frz1-Fc, Frz7-Fc) or IgG control. Cells were incubated for 24 hours, and luciferase activity was Klotho in tissue aging (16). It is clear that
quantified and normalized to b-gal activity. (*P < 0.05.) Wnts have multiple actions both developmen-

www.sciencemag.org SCIENCE VOL 317 10 AUGUST 2007 809

REPORTS
Fig. 4. The effects of the aged environment on
myogenic progenitor cell fate and muscle regeneration
are mediated by the Wnt signaling pathway. (A) (Top)
The fate of myogenic progenitors from young mice
incubated in young serum, with or without exogenous
Wnt3A, or incubated in aged serum, with or without the
Wnt inhibitor sFRP3, for 1½ days. (Bottom) Fate of
myogenic progenitors from aged mice incubated in
young serum or in aged serum, with or without sFRP3 or
DKK1, for 1½ days. The percentages of cells that
acquired a nonmyogenic cell fate were analyzed
morphologically and immunohistochemically as
described in Fig. 2. (*P < 0.05; **P < 0.01.) (B) Effects
of exogenous Wnt on muscle regeneration. Muscles of
young mice were injured, and either Wnt3A (200 ng/10
ml) or control solution [10 ml of 0.1% bovine serum
albumin (BSA)] was injected into regenerating tissues
1 day after injury. Cryosections were stained with Gomori
stain. (C) Effects of Wnt inhibition on fibrosis in
regenerating muscle. Muscles of young and aged mice
were injured, and either DKK1 (500 ng/10 ml) or control
solution (10 ml of 0.1% BSA) was injected into
regenerating tissues 1 day after injury. Muscles were
analyzed 5 days later. Cryosections were stained with
antibodies against collagen VI (green) and embryonic
myosin heavy chain (red). The histogram represents the
fibrotic index (as in Fig. 1B). (**P < 0.01; *P < 0.05.)

tally and postnatally (17). In contrast to the in- 14. F. Kuhnert et al., Proc. Natl. Acad. Sci. U.S.A. 101, 266 NIH (AG23806), the Department of Veterans Affairs
hibition of myogenesis reported here, Wnt sig- (2004). (Merit Review), the Ellison Medical Foundation, and an
15. J. T. Blitzer, R. Nusse, BMC Cell Biol. 7, 28 NIH Director’s Pioneer Award to T.A.R.
naling may promote myogenic lineage progression (2006).
during development (18). Such pleiotropic ef- 16. H. Liu et al., Science 317, 803 (2007). Supporting Online Material
fects may relate to differences in the timing of 17. R. Nusse, Cell Res. 15, 28 (2005). www.sciencemag.org/cgi/content/full/317/5839/807/DC1
Wnt signaling with regard to the state of cellular 18. T. Holowacz, L. Zeng, A. B. Lassar, Dev. Dyn. 235, 633 Materials and Methods
(2006). Figs. S1 to S9
differentiation or to changes in other interact- 19. We thank E. Fuchs for TOPGAL mice, R. Nusse for References
ing signaling pathways during development and recombinant Wnt3A and LSL cells, and B. Olwin for
aging. Our results may provide a strategy to im- the Syndecan-4 antibody. Supported by a grant from 23 April 2007; accepted 9 July 2007
prove tissue repair, particularly under condi- the NIH (DK069989) to C.J.K. and by grants from the 10.1126/science.1144090
tions in which regeneration is impaired and
fibrosis is favored, such as in aging and mus-
cular dystrophies. International Conservation Policy
References and Notes
1. G. Goldspink, K. Fernandes, P. E. Williams, D. J. Wells, Delivers Benefits for Birds in Europe
Neuromuscul. Disord. 4, 183 (1994).
2. M. D. Grounds, Ann. N.Y. Acad. Sci. 854, 78 Paul F. Donald,1* Fiona J. Sanderson,1 Ian J. Burfield,2 Stijn M. Bierman,3
(1998). Richard D. Gregory,1 Zoltan Waliczky1
3. I. M. Conboy, T. A. Rando, Cell Cycle 4, 407 (2005).
4. C. Pastoret, A. Sebille, J. Neurol. Sci. 129, 97 Conservation of the planet’s biodiversity will depend on international policy intervention, yet
(1995).
5. I. M. Conboy, M. J. Conboy, G. M. Smythe, T. A. Rando,
evidence-based assessment of the success of such intervention is lacking. Poor understanding of
Science 302, 1575 (2003). the effectiveness of international policy instruments exposes them to criticism or abandonment
6. I. M. Conboy et al., Nature 433, 760 (2005). and reduces opportunities to improve them. Comparative analyses of population trends provide
7. Materials and methods are available as supporting strong evidence for a positive impact of one such instrument, the European Union’s Birds Directive,
material on Science Online.
and we identify positive associations between the rate of provision of certain conservation
8. M. Chilosi et al., Am. J. Pathol. 162, 1495 (2003).
9. F. Jiang, C. J. Parsons, B. Stefanovic, J. Hepatol. 45, 401 measures through the directive and the response of bird populations. The results suggest that
(2006). supranational conservation policy can bring measurable conservation benefits, although future
10. E. H. Jho et al., Mol. Cell. Biol. 22, 1172 (2002). assessments will require the setting of quantitative objectives and an increase in the availability
11. R. DasGupta, E. Fuchs, Development 126, 4557
of data from monitoring schemes.
(1999).
12. T. Hagen, E. Di Daniel, A. A. Culbert, A. D. Reith, J. Biol.
ecause global threats to biodiversity are (1), their solutions will depend largely on inter-
Chem. 277, 23330 (2002).
13. K. Willert, R. Nusse, Curr. Opin. Genet. Dev. 8, 95
(1998). B largely anthropogenic, already consider-
able in scale, and accelerating rapidly
national policy intervention. This was recognized
in the formulation of international agreements

810 10 AUGUST 2007 VOL 317 SCIENCE www.sciencemag.org