Download Bone marrow contribution to skeletal muscle: A physiological

Developmental Biology 279 (2005) 336 – 344
www.elsevier.com/locate/ydbio
Bone marrow contribution to skeletal muscle: A physiological
response to stress
Adam T. Palermob,1, Mark A. LaBargeb,1, Regis Doyonnasa,
Jason Pomerantza, Helen M. Blaua,b,*
a
Baxter Laboratory in Genetic Pharmacology, Department of Microbiology and Immunology, Stanford University School of Medicine,
Stanford, CA 94305-5175, USA
b
Department of Molecular Pharmacology, Stanford University School of Medicine Stanford, CA 94305, USA
Received for publication 1 November 2004, revised 16 December 2004, accepted 20 December 2004
Available online 22 January 2005
Abstract
Adult bone marrow-derived stem cells (BMDC) have been shown to contribute to numerous tissues after transplantation into a new host.
However, whether the participation of these cells is part of the normal response to injury remains a matter of debate. Using parabiotically
joined pairs of genetically labeled and wildtype mice, we show here that irradiation-induced damage of the target tissue, injection of bone
marrow into the circulation, and immunological perturbation that are consequences of bone marrow transplantation are not necessary for bone
marrow contribution to myofibers. Moreover, severe toxin-induced damage is not a prerequisite, as BMDC contribution to muscle is
enhanced in response to increased muscle activity resulting from muscle overloading or forced exercise. Indeed, these two forms of muscle
stress result in much more rapid contribution (within 1 month) than voluntary running (6 months). These results indicate that BMDC
contribute to myofibers in response to physiologic stresses encountered by healthy organisms throughout life.
D 2005 Elsevier Inc. All rights reserved.
Keywords: Adult stem cells; Bone marrow-derived cells; Muscle; Transplantation; Irradiation; Regeneration; Damage; Exercise; Parabiosis
Introduction
Following transplantation with genetically marked bone
marrow, donor-derived cells can be found integrated into
diverse somatic tissues of the transplant recipient, such as
skeletal muscle, brain, liver, kidney, lung, and gastrointestinal tract (Alvarez-Dolado et al., 2003; Brazelton et
al., 2000, 2003; Camargo et al., 2004; Ferrari et al., 1998;
Abbreviations: BMDC, bone marrow-derived cells; NTX, Notexin;
BMT, bone marrow transplantation; HSC, hematopoietic stem cell; TA,
tibialis anterior; EDL, extensor digitorum longus.
* Corresponding author. Department of Molecular Pharmacology,
Baxter Laboratory in Genetic Pharmacology, Stanford University School
of Medicine, 269 W. Campus Drive CCSR 4215, Stanford, CA 943055175, USA. Fax: +1 650 736 0080.
E-mail address: [email protected] (H.M. Blau).
1
These authors contributed equally to this work.
0012-1606/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.ydbio.2004.12.024
Gussoni et al., 1999; Harris et al., 2004; Houghton et al.,
2004; Kale et al., 2003; Krause et al., 2001; LaBarge and
Blau, 2002; Lagasse et al., 2000; Mezey et al., 2000;
Theise et al., 2002; Weimann et al., 2003a,b; for reviews,
see Pomerantz and Blau, 2004; Wagers and Weissman,
2004). In one striking example, the contribution of
BMDC to hepatocytes by fusion rescued mice with a
lethal metabolic liver disease, indicating that BMDC can
activate previously silent genes and function as nonhematopoietic cells (Camargo et al., 2004; Lagasse et al.,
2000; Vassilopoulos et al., 2003; Wang et al., 2002). In
other cases, mononucleate BMDC have been shown to
assume different phenotypes in the absence of fusion and
become epithelial cells (Harris et al., 2004) or muscle
satellite cells (Dreyfus et al., 2004; Fukada et al., 2002;
LaBarge and Blau, 2002; Sherwood et al., 2004).
However, these experiments have all used bone marrow
transplantation, which is highly invasive, and associated
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
with lethal irradiation as well as massive disregulation of
cytokines. These confounding variables do not allow
conclusions to be made regarding the potential of BMDC
to contribute to tissues under normal physiological
conditions.
In the case of skeletal muscle, extensive research has
convincingly shown that bone marrow-derived nuclei can
become incorporated into mature myofibers of transplant
recipients (reviewed in Grounds et al., 2002). Bone
marrow contains a variety of cell types, such as
hematopoietic cells, stromal cells, and endothelial progenitors, any of which could theoretically contribute to muscle
in experiments in which whole bone marrow is transplanted. That at least some bone marrow-derived myonuclei are of hematopoietic origin has been clearly shown by
transplantation of single hematopoietic stem cells (HSC)
(Camargo et al., 2003; Corbel et al., 2003). Further
investigation has revealed that one type of HSC derivative,
the myelomonocytic cell, can contribute to both skeletal
muscle and liver (Camargo et al., 2003, 2004; Doyonnas et
al., 2004; Willenbring et al., 2004).
To date, a major caveat of most reports of bone
marrow contribution to myofibers has been that they
utilized bone marrow transplantation (BMT), as this
method enables bone marrow genetically marked with
green fluorescent protein (GFP), beta-galactosidase, or sex
chromosome differences to be tracked and studied in a
background of unmarked tissues. The problem is that
BMT is a severe procedure that introduces a number of
complex variables. Specifically, transplantation involves
the delivery of high doses of whole body radiation to
ablate the endogenous bone marrow, and create a niche
that is essential for donor marrow engraftment and
reconstitution of the blood lineages. Such radiation doses
are well documented to damage numerous tissues that
contain proliferative cells in addition to the marrow. For
example, irradiation of skeletal muscle impairs the local
satellite cell population (muscle stem cells) and can lead
to muscle atrophy, inflammation, and vascular leakiness
(Baker and Krochak, 1989; Rosenblatt and Parry, 1992;
Van der Meeren et al., 2001). Furthermore, the mechanical extraction of bone marrow from its native environment and injection directly into the circulation may allow
cells that are normally confined to the bone marrow
compartment to access tissues where they are not found
under normal circumstances. Thus, published results to
date do not effectively address whether bone marrow
contribution to non-hematopoietic tissues occurs during the
normal life of an organism. Consequently, it is critically
important to determine whether the contribution of BMDC
to muscle can also occur independently of BMT-associated
variables.
BMDC contribution to muscle is a response to tissue
damage and was previously missed when studied in the
absence of overt muscle injury (Wagers et al., 2002). By
contrast, damage induced by delivery of toxins such as
337
notexin or cardiotoxin has resulted in an increased
contribution to skeletal muscle (Abedi et al., 2004;
Camargo et al., 2003; Corbel et al., 2003; Fukada et al.,
2002). Although these stimuli have resulted in significant
increases in the number of myofibers that integrate bone
marrow-derived nuclei, these are severe forms of muscle
damage in which widespread necrosis and inflammation
are typical (Lefaucheur and Sebille, 1995). The question
remains as to whether BMDC respond in this manner to
stresses typical of normal life such as increased muscle
use.
In this study, we determine whether bone marrowderived cells contribute to muscle in response to physiological stimuli encountered throughout life. To determine
whether damage due to muscle overuse could substitute for
toxins, we exposed mice to treadmill running or surgically
removed a muscle so that an otherwise underutilized
muscle was overloaded. In both cases, a significant
increase in the number of GFP+ muscle fibers was
observed after 1 month. To eliminate transplant-associated
variables, we used parabiotically joined mice to genetically
label and track blood cells. The data reported here provide
evidence that cells from the circulation, which derive from
the bone marrow, can incorporate into myofibers independently of transplantation and in response to damageinduced signals typically encountered in the course of
normal adult life.
Materials and methods
Mice and bone marrow transplantation
C57/B6 (wildtype) mice were purchased from Jackson
laboratory (Bar Harbor, ME) and C57/B6 GFP transgenic
(GFP) mice (gift from Dr. Okabe, Osaka, Suita, Japan)
were bred in-house and maintained in a pathogen-free
environment. Bone marrow of GFP donor mice was
flushed from the femurs and tibias to make single cell
suspensions. Recipient wt mice that had received total
body irradiation (9.6 Gy) were transplanted at 8 weeks of
age with 106 cells of total bone marrow. Peripheral blood
of transplanted recipients was analyzed at various time
points after transplantation and only mice with more than
90% GFP+ cells in their peripheral blood were used. All
protocols were approved by the Administrative Panel on
Laboratory Animal Care at Stanford University School of
Medicine.
Parabiosis
Parabiotic pairs were created by a modification of the
procedure of Bunster and Meyer (1933). Mice were
anesthetized with a cocktail of Ketamine (2.4 mg per
mouse, Fort Dodge Animal Health, Overland Park, KS) and
Xylazine (240 Ag per mouse, Phoenix Scientific, St. Joseph,
338
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
MO) delivered by intraperitoneal injection. Mice were
shaved on either their right side or left side. The shaved
skin of each mouse was sterilized and incisions were made
from the olecranon to the knee joint, and the subcutaneous
fascia was bluntly dissected to create free skin. A single 4-0
silk suture and tie was used to join the opposing olecranon
and knee joints, and opposing dorsal and ventral skins were
approximated by 9 mm wound clips (Becton-Dickinson,
Franklin Lakes, NJ). Mice were given supportive care
including buprenorphine, moistened food, and subcutaneous
saline as needed, and each pair was housed separately.
Cross-circulation and the degree of peripheral blood
chimerism was evaluated by retro-orbital bleeding followed
by analysis on a FACSCalibur flow cytometer (BectonDickinson, Franklin Lakes, NJ).
Notexin injury
Mice were anesthetized with inhaled methoxyflurane, the
skin over the tibialis anterior (TA) was shaved, and a single
10 Al dose of 10 Ag/ml NTX (Sigma, St. Louis, MO) in PBS
was injected.
Forced exercise
C57Bl/6 mice were transplanted with bone marrow from
GFP transgenic mice 3 months prior to exposure to forced
exercise. The mice ran on a motor-driven treadmill
(Exercise 4 treadmill, Columbus Instrument, Columbus,
OH) at a constant speed of 20 m/min on a 108 decline
(downhill) for 60 min, 3 times per week with 1 day of rest
between sessions. All mice were trained for 15 min before
the first session. Training involved 5 min at 10 m/min
followed by a progressive increase in speed from 10 m/min
to 20 m/min at a rate of 1 m/min. To encourage animals to
move forward and start running, a mild electric shock was
applied from a grid behind the treadmill. Mice in the 4 week
group were submitted to the protocol described above for 4
consecutive weeks. For both groups (1 or 4 weeks running),
the TA, extensor digitorum longus (EDL), and soleus
muscles were harvested for analysis 8 weeks after the onset
of the exercise regimen.
EDL overload
Mice were anesthetized with Avertin (30 mg/kg, SigmaAldrich, St. Louis, MO) delivered by intraperitoneal
injection. Their left and right legs were sterilized and
incisions were made from below the ankle joint to above
the knee, along the front of the leg. An incision was made
in the epimysium, and the TA was bluntly dissected away
from the rest of the leg. In sham-operated control mice, the
skin was closed at this point, while in the contralateral
experimental legs, the TA was then removed by first
severing it below the myotendonous junction and gently
separating it from the underlying structures before cutting
just below the knee. The skin was then closed using a
single 4-0 silk suture. Mice were given supportive care
including buprenorphine, moistened food, and subcutaneous saline as above.
Tissue collection and immunohistochemistry
TA, EDL, or soleus muscles were dissected and
immersed in PBS/0.5% EM grade paraformaldehyde for 2
h at room temperature followed by overnight immersion in
PBS/20% sucrose at 48C. Fixed tissue was imbedded in
OCT, snap-frozen in isobutane and liquid nitrogen, and then
sectioned at 10 Am thickness. Sections were blocked in
PBS/20% normal goat serum/0.3% Triton X-100 for 2 h,
then incubated with primary antibodies against GFP (rabbit,
1:1000, Molecular Probes, Eugene, OR), and laminin-a2
(rat anti-mouse, 1:400, Upstate Biotech Associates, Waltham, MA), in blocking solution for 1 h. Following PBS/
0.1% Triton X-100 washes, sections were incubated in
secondary antibodies (goat anti-rabbit Alexa488, goat antirat Alexa546, Molecular Probes, Eugene, OR) in blocking
solution for 1 h. Following short washes in PBS/0.1% Triton
X-100, the sections were washed overnight in PBS, before
mounting them with Fluoromount G (Southern Biotechnology Associates, Birmingham, AL). Some sections were
counterstained with 5 Ag/ml Hoechst 33342 (Sigma, St.
Louis, MO).
Quantification of GFP+ muscle fibers
To quantitate muscle fibers, at least three sections from
each mouse that were z100 Am apart from each other
were imaged on a Zeiss Axioplan epifluorescent microscope (Carl Zeiss Inc., Thornwood, NY) with a low
magnification objective (5). Nuclear staining (Hoechst
33342) was visualized using a 345–385 nm BP excitation
filter and a 435–485 nm BP emission filter. Laminin
immunofluorescence images (Alexa546) were acquired
using a 530–560 nm BP excitation filter and a 573–648
nm BP emission filter. GFP immunofluorescence images
(Alexa488) were acquired using a 460–500 nm BP
excitation filter and a 510–560 nm BP emission filter.
All putative GFP+ fibers were visually confirmed using a
dual-band red-green filter set exciting at 485–505 nm/560–
585 nm with emissions at 515–550 nm/590–660 nm. This
filter set is useful for readily discriminating between
Alexa488 fluorescence and autofluorescence. GFP immunofluorescence images were then overlaid on laminin
immunofluorescence images and GFP+ muscle fibers
were only counted if they were completely surrounded
by an intact basal laminal membrane. Counting only
fibers with intact basal laminal membranes ensures that
dying fibers or large conglomerations of blood cells are
excluded from the analysis. The confocal images (1 Am
optical sections) were acquired on a Zeiss laser-scanning
microscope (LSM 510).
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
339
Results
Tissue irradiation is not a prerequisite for contribution to
muscle from the circulation
To determine whether irradiation of the muscle and
associated variables of bone marrow transplantation are
necessary for incorporation of circulating cells into myofibers, we generated surgically joined parabiotic pairs of mice.
In such pairs, one partner was wildtype (C57/B6) and the
other partner was a syngenic mouse in which the green
fluorescent protein was under the control of the beta actin
promoter and therefore expressed in all tissues (Fig. 1A). In
parabiotic animals, anastamosis of the circulatory systems
occurs and the peripheral blood can freely exchange between
the two animals as early as 2 weeks after joining (Bunster and
Meyer, 1933). Parabiosis thus enables GFP-labeled circulatory cells to gain access to the tissues of wildtype mice.
Three weeks after parabiotic joining, the peripheral blood
was evaluated by flow cytometry to confirm cross-circulation
between the animals and the extent of blood chimerism.
Mononuclear cells from peripheral blood were doublestained with either Mac1/Gr1 or B220/CD3 to identify their
myeloid or lymphoid origin, respectively. FACS analysis
indicated that peripheral blood of the wildtype partners
contained 37–51% GFP+ cells (Figs. 1B, E), and normal
proportions of the lymphoid and myeloid subsets.
The wildtype partner was then injured by injection of
NTX, a myotoxin commonly used to damage muscle, into
the TA contralateral to the surgical site. Mice were sacrificed
2 or 4 weeks later and both injured and non-injured TA
muscles of the wildtype mice were harvested. The TA
adjacent to the surgical site served as a negative control for
NTX damage. Circulating cell contribution to myofibers in
the TAs was evaluated by immunofluorescence staining
with antibodies against GFP and laminin, and Hoechst dye
was used to visualize the nuclei.
To control for the possibility that NTX injection had not
been administered to the muscle in question, central
nucleation of myofibers was used to assess whether the TA
had been damaged and regeneration had been induced. This
approach is based on the well-established finding that the
myonuclei in healthy muscle are on the perimeter of
myofibers, whereas in muscle regenerating either because
of severe dystrophy (Coulton et al., 1988) or because of
myotoxin injection (Martin and Ontell, 1988), myonuclei are
in the center of fibers. Only mice in which the contralateral
TAs contained N80% centrally nucleated fibers, indicative of
muscle regeneration, are reported here.
Our analysis indicated that the undamaged TAs contained
very few GFP+ myofibers (Figs. 1C, E), whereas the
regenerating contralateral TAs contained substantially more
GFP+ myofibers (Figs. 1D, E). In this study, the average
number of myofibers in a single transverse section of the TA
approximated 3000. Thus, the average undamaged TA had
less than 0.05% GFP+ myofibers per cross-section as
Fig. 1. Contribution of circulating cells to skeletal muscle in parabiotic mice
in the absence of irradiation. (A) Wildtype C57/Bl6 mice were parabiotically conjoined with GFP transgenic mice. Three weeks after joining, the wt
partner was injured by notexin (NTX) injection into the tibialis anterior
muscle (TA) contralateral to the surgical site. TA muscles were harvested 2
or 4 weeks later. (B) Peripheral blood mononuclear cells were evaluated by
flow cytometry (FACS) for expression of GFP as well as staining with
either a cocktail of lymphoid markers (B220/CD3) or a cocktail of myeloid
markers (Mac1/GR1). FACS plots are shown for a GFP transgenic mouse
(GFP+ ctrl), a C57/Bl6 control mouse (wt ctrl), and for either partner of a
parabiotic pair. (C) Representative transverse section from a TA of a
wildtype parabiotic partner not treated with NTX, showing intact, nonGFP+ muscle fibers. (D) GFP+ muscle fibers (green) surrounded by intact
basal laminal membranes (red) in a representative TA transverse section
from a NTX-injured parabiotic mouse (nuclei in blue, Hoechst 33342). (E)
Number of GFP+ fibers observed in mice at either 2 or 4 weeks post injury,
along with the percentage of GFP+ cells (chimerism) in the peripheral
blood (% chim), and whether central nuclei were observed. Number of
fibers is reported as mean F SD for 3 sections separated by at least 100 Am.
340
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
compared to approximately 0.4% GFP+ myofibers in NTXinjured TAs, an increase of almost one order of magnitude.
Thus, using parabiosis, we could demonstrate that cells
present in the circulation become incorporated into regenerating muscle even when the muscle has not been damaged
by irradiation and the local muscle-specific stem cells
known as satellite cells are intact. Note that by contrast,
the level of irradiation required for a bone marrow transplant
ablates two thirds of the satellite cell population (LaBarge
and Blau, 2002).
While this experiment demonstrates the existence of
circulating cells that can contribute to muscle, it does not
identify the source of these cells, as all tissues of the GFP
transgenic partner expressed GFP. Thus, GFP+ cells from
tissues other than bone marrow could have been present in
the circulation and become incorporated into myofibers of
the wildtype partner.
Circulating cells contributing to muscle regeneration are of
bone marrow origin
To test the hypothesis that the circulating cells with the
ability to incorporate into myofibers can arise from the bone
marrow, wildtype mice were joined to wildtype mice that
had received bone marrow transplants from GFP transgenic
mice 6 months earlier (Fig. 2A). In these experiments, bone
marrow-derived cells, not cells from other tissues, are
labeled with GFP in the parabiotic animals. Three weeks
after joining, the peripheral blood of the non-transplanted
partner was found to be 22–42% GFP+ by flow cytometry
(Fig. 2D). Contralateral TAs of the non-transplanted mice
were injured with NTX and evaluated 2 or 4 weeks later. In
this experiment, as above, localized NTX-induced injury
caused a substantial increase in the number of GFP+
myofibers compared to non-injured TA muscles (Figs. 2B,
C, D). The fraction of GFP+ myofibers increased from less
than 0.04% in undamaged muscle to approximately 0.4% in
NTX-injured muscle, a magnitude similar to that observed
when GFP transgenic rather than transplanted partners were
used. We therefore conclude that the population of circulating cells able to incorporate into muscle in the previous
experiment can be functionally reconstituted by transplanted
bone marrow alone and does not require other cells.
Taken together, these data demonstrate that BMDC
incorporation into myofibers following transplantation is
similar to that observed in GFP–wildtype parabiotic pairs,
i.e., independent of irradiation or other bone marrow
transplantation-associated variables.
Forced downhill running elicits BMDC contribution to
muscle fibers
Most previous studies of BMDC contribution to muscle
have employed injury models unrelated to the physiological
stresses experienced by muscle in the course of normal life, or
have required half a year to see effects. BMDC incorporation
Fig. 2. Contribution of circulating bone marrow-derived cells to skeletal
muscle in parabiotic mice. (A) Wildtype C57/Bl6 mice were parabiotically
conjoined with wildtype mice that had received bone marrow transplants
from GFP transgenic mice 6 months earlier. Three weeks after joining, the
wt partner was injured by notexin (NTX) injection into the tibialis anterior
muscle (TA) contralateral to the surgical site at which the mice were
conjoined. TA muscles were harvested 2 or 4 weeks later. (B) Representative transverse section from non-notexin-injured TA of wildtype parabiotic
partner, showing a single rare, GFP+ muscle fiber. (C) GFP+ muscle fibers
(green) surrounded by intact basal laminal membranes (red) in a
representative TA transverse section from a NTX-injured parabiotic mouse
(nuclei in blue, Hoechst 33342). (D) Number of GFP+ fibers observed in
mice at either 2 or 4 weeks post injury, along with the percentage of GFP+
cells (chimerism) in the peripheral blood (% chim), and whether central
nuclei were observed. Number of GFP+ fibers is reported as mean F SD for
3 sections separated by at least 100 Am.
has been shown to be elevated after acute myotoxin injury
which results in chronic inflammation, necrosis, and regeneration (Abedi et al., 2004; Bossolasco et al., 2004; Camargo
et al., 2003; Corbel et al., 2003; Fukada et al., 2002). In the
course of normal life, however, the majority of ongoing
myogenesis is the result of exercise-induced damage.
Our previous work suggested that exercise-induced
damage could elicit BMDC contribution to myofibers, but
this occurred slowly over the course of 6 months (LaBarge
and Blau, 2002). To determine whether a more acute stress
could elicit the same effect, we placed wildtype mice 3
months post-GFP+ BMT on inclined treadmills and forced
the mice to run for 1 h, 3 times per week for 1 or 4 weeks
(Fig. 3A) (Lerman et al., 2002). Forced downhill running
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
Fig. 3. Bone marrow contribution to myofibers increases in frequency in
response to physiological cues. (A) Wildtype C57/Bl6 mice that had
received bone marrow transplants from GFP transgenic mice 3 months
earlier were placed on a treadmill on a 108 decline at a speed of 20 m/min
(downhill) for 60 min, 3 times per week with 1 day rest between sessions.
TA, EDL, and soleus muscles were harvested after 1 or 4 weeks of exercise
followed by 7 or 4 weeks of recovery. (B) Wildtype C57/Bl6 mice that had
received bone marrow transplants from GFP transgenic mice 4 months
earlier underwent surgery to remove their TA muscles. This procedure
results in compensatory hypertrophy of the underlying EDL muscle. EDL
muscles were harvested after 4 weeks. (C) Number of mice per group and
number of GFP+ fibers per muscle for mice after forced exercise or EDL
overload. GFP+ fibers are reported as mean F SD per section analyzed in
each experimental group.
involves acute eccentric contraction of the TA, EDL, and
soleus muscles (Smith et al., 1997), which is distinct from
voluntary running on a wheel (LaBarge and Blau, 2002).
Eccentric muscle contraction can lead to muscle damage and
delayed-onset muscle soreness in humans and mice
(Kuipers, 1994). On the eighth week after the initiation of
exercise, lower leg muscles of the exercised mice were
evaluated and compared to control legs from unexercised,
transplanted animals by immunostaining for GFP and
laminin. Significant increases in GFP+ myofibers were
observed in these animals after both 1 and 4 weeks of forced
exercise (Fig. 3C), a result never seen with voluntary
exercise. The fraction of GFP+ fibers increased from 0.02%
in unexercised mice to 0.2% and 0.1% in mice exercised for
1 and 4 weeks, respectively.
These results indicate that sustained exercise, in which
fibers are thought to acquire membrane damage due to shear
stress, is sufficient to elicit BMDC incorporation into
muscle. Furthermore, these experiments demonstrate that
BMDC can respond extremely rapidly to muscle damage,
incorporating into myofibers in a single week.
Overloading causes GFP+ myofibers to appear in the EDL
BMDC incorporation into myofibers was also tested in
another model of acute physiological muscle stress, in
341
which hypertrophy of the EDL muscle is induced by
surgically resecting the TA (Fig. 3B). The TA and the
EDL normally function synergistically during ankle flexion,
and resecting the TA increases demand on the EDL, which
undergoes a compensatory hypertrophy. This treatment has
been shown to induce satellite cell contribution to myofibers
as the fibers compensate, thus enabling the EDL to function
alone (Rosenblatt and Parry, 1992).
To test the hypothesis that EDL hypertrophy also results
in increased BMDC contribution to myofibers, we resected
the TA muscles of wildtype mice that had received GFP+
BMT 4 months earlier. In the contralateral leg, a sham
operation was performed in which the TA was bluntly
dissected away from the underlying EDL, but left intact.
Mice were sacrificed 4 weeks after surgery. The EDL
muscles of experimental legs were grossly hypertrophic as
compared to the sham-operated legs. EDL sections from
both legs were immunostained for GFP and laminin
expression. Very low numbers of GFP+ myofibers were
observed in the control muscles, whereas the hypertrophic
EDLs all exhibited significant GFP incorporation (Fig. 3C).
On average, 0.05% of fibers were GFP+ in non-overloaded
muscle, whereas 1.6% of fibers were GFP+ in overloaded
muscle, a 30-fold increase.
From these two experiments, in which muscle usage was
forcibly and acutely increased, we conclude that BMDCs can
respond relatively rapidly to physiological stresses as well as
to more extreme damage by incorporating into myofibers.
Discussion
This report demonstrates that previous findings by us and
others that bone marrow-derived cells can contribute to
muscle fibers are not merely a result of the harsh
experimental paradigms used, but also occur under physiological conditions. To date, most studies, including our own,
have used bone marrow transplantation to follow the fate of
genetically marked cells and have monitored the contribution of bone marrow derivatives to tissues extensively
damaged due to destructive toxins. Here, using skeletal
muscle as a model, we provide evidence that cells from the
bone marrow can contribute to other tissues in response to
types of acute stress that occur during the normal life of a
animal. First, in parabiotically joined animals, blood
chimerism allowed tracking of circulating BMDC originating in the bone marrow of one mouse and detection of those
cells in the damaged muscle of its parabiotic partner in the
absence of BMT and associated variables. Second, during
hypertrophy, a physiologic response to two forms of forced
increased muscle use, BMDC incorporated into muscle as
efficiently and rapidly as after severe toxin-induced damage.
Together, these experiments rule out a number of confounding variables, and suggest that bone marrow contribution to
tissues occurs relatively rapidly in response to injury
associated with strenuous physical activity.
342
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
Bone marrow transplantation (BMT) involves wholebody irradiation, mechanical extraction, and injection of
bone marrow cells into the bloodstream. In the course of
BMT, marked systemic alterations occur, such as changes in
the balance of cytokines (Holler et al., 1997). If these
transplant-associated variables were prerequisites for the
contribution of BMDC to muscle fibers, BMDC nuclei
would not be expected to be found in skeletal muscle fibers
in the course of normal life, a finding that would
significantly limit the therapeutic potential of BMDC. The
parabiosis studies presented here show that when GFP
transgenic mice were surgically joined to wildtype mice,
GFP+ muscle fibers were detected in the wildtype partner,
and their frequency was significantly increased by NTXinduced muscle damage. These experiments show that a cell
with the ability to transit through the circulation can
contribute to non-irradiated regenerating muscle in the
absence of bone marrow transplantation. Although
enhanced by NTX-mediated muscle damage, this participation of BMDC also occurred in low numbers in the absence
of NTX, in the leg adjacent to the conjoining surgical site.
When GFP+ bone marrow transplant recipients were used as
the parabiotic donors, similar results were found. This
finding suggests that the population of cells that contributed
to muscle was not derived from cells which entered the
bloodstream from other tissues, but rather from the bone
marrow via the circulation, in a manner similar to that
previously described using bone marrow transplantation.
Although parabiotic joining avoids most variables
associated with bone marrow transplantation and irradiation,
transient inflammation is associated with the surgical
procedure, similar to that which occurs after wounding.
We found evidence of systemic inflammation in a separate
cohort of parabiotic mice, but it occurred only in a subset of
animals and the extent of inflammation was not outside the
range of the response observed during wound repair, which
is a common physiologic occurrence (data not shown).
Additionally, serum creatine phosphokinase (CPK) levels,
indicative of muscle damage, were measured in parabiotic
animals. The results were within the normal range of the
assay (158 F 87 IU/L and 244 F 125 IU/L at 2 and 14 days
after surgery) and were similar to CPK levels measured in
uninjured mice. In contrast, work by others has shown that
in the dystrophic mdx mouse, CPK levels approximate 5000
IU/L (Hoyte et al., 2004). Therefore, by this measure,
parabiotic joining does not induce significant muscle
damage. Other groups have also used parabiotic joining to
emulate physiologic conditions, and have observed that
parabiosis does not alter the mobilization of bone marrow
progenitor cells in comparison with non-parabiotic mice
(Wright et al., 2001). Moreover, it should be noted that the
same group, using parabiotic animals in which no muscle
damage was induced, observed very few GFP+ muscle
fibers (Wagers et al., 2002), in good agreement with our
findings. Thus, the inflammation associated with the surgery
required for joining mice does not by itself account for bone
marrow contribution to myofibers. Nonetheless, the possibility that some systemic inflammation is necessary for
BMDC to contribute to muscle cannot be ruled out by this
or other published reports to date. The nature of such
inflammatory signals, their identity, and their potential role
in the recruitment and activation of BMDC is a subject of
major interest.
In a second set of experiments, we explored the nature of
regenerative stimuli that are able to elicit BMDC contribution to muscle and whether physiologic stimuli encountered
in daily life could suffice. Most muscle regeneration occurs
in response to use-related damage. BMDC contribution to
muscle was found to be inducible in two different models:
forced downhill running on a treadmill and compensatory
hypertrophy of the EDL after resection of the TA. Both of
these stresses resulted in significant increases in GFP+
myofibers within 4 weeks after forced exercise or surgery.
From these studies, we conclude that toxin-induced muscle
damage that is typically associated with widespread necrosis
and inflammation is not required for BMDC contribution to
muscle. Indeed, we found that contribution of BMDC to
muscle during the acquisition of new healthy muscle tissue
is on a par with that seen after toxin damage.
The number of bone marrow-derived nuclei in regenerating muscle is difficult to assess. In this, and other studies
(Brazelton et al., 2003; LaBarge and Blau, 2002), we have
observed muscles containing a range of 0.5% to 5% GFP+
myofibers. However, because a myofiber is a multinucleate
syncytium, it is unclear what proportion of total nuclei in
myofibers are derived from the bone marrow. Since GFP is
not restricted to the vicinity of the nucleus that encodes it,
but is a diffusible cytoplasmic protein, the fraction of
myonuclei derived from bone marrow is likely to be lower
than the fraction of GFP+ fibers. In addition, it is not known
whether the amount of GFP expressed by a single nucleus,
when distributed into the large volume of a myofiber, is
detectable by our assay. It is also possible that BMDC nuclei
fuse preferentially with certain fibers and not others. These
complicating factors make it difficult to estimate from this
type of study what fraction of myonuclei is of bone marrow
origin. However, the possibility that this fraction may be
very small does not affect the conclusion that it does occur,
and responds to physiological cues.
We hypothesize that BMDC contribution to muscle is
mediated by a specific set of molecular cues associated with
physiologic muscle injury, inflammation, and the resulting
mobilization of BMDC. The identity of these cues is an
important matter for future investigation. Factors that
activate myelomonocytic progenitors, shown to be involved
in bone marrow contribution to muscle (Camargo et al.,
2003; Doyonnas et al., 2004), are a reasonable starting
point. For example, G-CSF is known to mobilize myelomonocytic progenitors from the bone marrow of mice
(Lapidot and Petit, 2002). Moreover, recent work in humans
has shown that endurance exercise results not only in
acutely higher blood levels of G-CSF and other cytokines,
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
but also in a chronic elevation of the number of circulating
bone marrow-derived progenitor cells (Bonsignore et al.,
2002). Another class of factors worthy of investigation
includes molecules involved in the proliferation and differentiation of muscle stem cells (satellite cells), such as IGF-1
(Florini et al., 1996). It has already been shown that
transgenic overexpression throughout life of the muscle
isoform of IGF-1 can enhance BMDC recruitment to
regenerating muscle in vivo (Musaro et al., 2004). Local
in vivo delivery of such factors in wildtype mice would
indicate whether these or other factors are involved in
signaling BMDC to contribute to myofibers.
In conclusion, the results reported here eliminate a
number of variables inherent in previous reports and
demonstrate that BMDC contribution to muscle occurs
under circumstances compatible with daily activity and is a
characteristic of normal physiology. These findings provide
impetus for understanding the underlying mechanism and
the role of inflammation and injury-related signals. From a
therapeutic perspective, the frequency of contribution may
currently be inadequate for improving muscle function.
However, the prospect of recruiting circulating cells to
regenerating muscle would circumvent the problems of cell
distribution associated with intramuscular injection of
myoblasts for therapy. We have shown here that this is a
natural phenomenon that is a response to damage. Elucidation of the relevant damage-associated factors could
potentially lead to a robust response sufficient for therapy.
Acknowledgments
We thank Frances L. Johnson for her teaching and
assistance with the parabiosis protocols, Andrew Patterson
for assistance with the treadmill exercise protocols, Peggy
Kraft and Kassie Koleckar for their excellent technical
support, and James Weimann and Alessandra Sacco for
useful comments. This work was supported by predoctoral
fellowships to MAL (TG AG00259) and to ATP (National
Science Foundation), and an NSRA to JHP (1F32AR05167801) and by NIH grants AG09521, HL65572, HD18179,
AG20961, a Senior Ellison Medical Foundation award AG33-0817, and support from the McKnight Endowment Fund
and the Baxter foundation to HMB.
References
Abedi, M., Greer, D.A., Colvin, G.A., Demers, D.A., Dooner, M.S., Harpel,
J.A., Weier, H.U., Lambert, J.F., Quesenberry, P.J., 2004. Exp. Hematol.
32, 426 – 434.
Alvarez-Dolado, M., Pardal, R., Garcia-Verdugo, J.M., Fike, J.R., Lee,
H.O., Pfeffer, K., Lois, C., Morrison, S.J., Alvarez-Buylla, A., 2003.
Nature 425, 968 – 973.
Baker, D.G., Krochak, R.J., 1989. Cancer Invest. 7, 287 – 294.
Bonsignore, M.R., Morici, G., Santoro, A., Pagano, M., Cascio, L.,
Bonanno, A., Abate, P., Mirabella, F., Profita, M., Insalaco, G., Gioia,
343
M., Vignola, A.M., Majolino, I., Testa, U., Hogg, J.C., 2002. J. Appl.
Physiol. 93, 1691 – 1697.
Bossolasco, P., Corti, S., Strazzer, S., Borsotti, C., Del Bo, R., Fortunato, F.,
Salani, S., Quirici, N., Bertolini, F., Gobbi, A., Deliliers, G.L., Pietro
Comi, G., Soligo, D., 2004. Exp. Cell Res. 295, 66 – 78.
Brazelton, T.R., Rossi, F.M., Keshet, G.I., Blau, H.M., 2000. Science 290,
1775 – 1779.
Brazelton, T.R., Nystrom, M., Blau, H.M., 2003. Dev. Biol. 262, 64 – 74.
Bunster, E., Meyer, R.K., 1933. Anat. Rec. 57, 339.
Camargo, F.D., Green, R., Capetenaki, Y., Jackson, K.A., Goodell, M.A.,
2003. Nat. Med. 9, 1520 – 1527.
Camargo, F.D., Finegold, M., Goodell, M.A., 2004. J. Clin. Invest. 113,
1266 – 1270.
Corbel, S.Y., Lee, A., Yi, L., Duenas, J., Brazelton, T.R., Blau, H.M., Rossi,
F.M., 2003. Nat. Med. 9, 1528 – 1532.
Coulton, G.R., Morgan, J.E., Partridge, T.A., Sloper, J.C., 1988. Neuropathol. Appl. Neurobiol. 14, 53 – 70.
Doyonnas, R., LaBarge, M.A., Sacco, A., Charlton, C., Blau, H.M., 2004.
Proc. Natl. Acad. Sci. 101, 13507 – 13512.
Dreyfus, P.A., Chretien, F., Chazaud, B., Kirova, Y., Caramelle, P.,
Garcia, L., Butler-Browne, G., Gherardi, R.K., 2004. Am. J. Pathol.
164, 773 – 779.
Ferrari, G., Cusella-De Angelis, G., Coletta, M., Paolucci, E., Stornaiuolo,
A., Cossu, G., Mavilio, F., 1998. Science 279, 1528 – 1530.
Florini, J.R., Ewton, D.Z., Coolican, S.A., 1996. Endocr. Rev. 17,
481 – 517.
Fukada, S., Miyagoe-Suzuki, Y., Tsukihara, H., Yuasa, K., Higuchi, S.,
Ono, S., Tsujikawa, K., Takeda, S., Yamamoto, H., 2002. J. Cell Sci.
115, 1285 – 1293.
Grounds, M.D., White, J.D., Rosenthal, N., Bogoyevitch, M.A., 2002. J.
Histochem. Cytochem. 50, 589 – 610.
Gussoni, E., Soneoka, Y., Strickland, C.D., Buzney, E.A., Khan, M.K.,
Flint, A.F., Kunkel, L.M., Mulligan, R.C., 1999. Nature 401, 390 – 394.
Harris, R.G., Herzog, E.L., Bruscia, E.M., Grove, J.E., Van Arnam, J.S.,
Krause, D.S., 2004. Science 305, 90 – 93.
Holler, E., Ertl, B., Hintermeier-Knabe, R., Roncarolo, M.G., Eissner, G.,
Mayer, F., Fraunberger, P., Behrends, U., Pfannes, W., Kolb, H.J., et al.,
1997. Leuk. Lymphoma 25, 217 – 224.
Houghton, J., Stoicov, C., Nomura, S., Rogers, A.B., Carlson, J., Li, H.,
Cai, X., Fox, J.G., Goldenring, J.R., Wang, T.C., 2004. Science 306,
1568 – 1571.
Hoyte, K., Jayasinha, V., Xia, B., Martin, P.T., 2004. Am. J. Pathol. 164,
711 – 718.
Kale, S., Karihaloo, A., Clark, P.R., Kashgarian, M., Krause, D.S., Cantley,
L.G., 2003. J. Clin. Invest. 112, 42 – 49.
Krause, D.S., Theise, N.D., Collector, M.I., Henegariu, O., Hwang, S.,
Gardner, R., Neutzel, S., Sharkis, S.J., 2001. Cell 105, 369 – 377.
Kuipers, H., 1994. Int. J. Sports Med. 15, 132 – 135.
LaBarge, M.A., Blau, H.M., 2002. Cell 111, 589 – 601.
Lagasse, E., Connors, H., Al-Dhalimy, M., Reitsma, M., Dohse, M.,
Osborne, L., Wang, X., Finegold, M., Weissman, I.L., Grompe, M.,
2000. Nat. Med. 6, 1229 – 1234.
Lapidot, T., Petit, I., 2002. Exp. Hematol. 30, 973 – 981.
Lefaucheur, J.P., Sebille, A., 1995. Neuromuscular Disord. 5, 501 – 509.
Lerman, I., Harrison, B.C., Freeman, K., Hewett, T.E., Allen, D.L.,
Robbins, J., Leinwand, L.A., 2002. J. Appl. Physiol. 92, 2245 – 2255.
Martin, H., Ontell, M., 1988. Muscle Nerve 11, 588 – 596.
Mezey, E., Chandross, K.J., Harta, G., Maki, R.A., McKercher, S.R., 2000.
Science 290, 1779 – 1782.
Musaro, A., Giacinti, C., Borsellino, G., Dobrowolny, G., Pelosi, L.,
Cairns, L., Ottolenghi, S., Cossu, G., Bernardi, G., Battistini, L.,
Molinaro, M., Rosenthal, N., 2004. Proc. Natl. Acad. Sci. 101,
1206 – 1210.
Pomerantz, J., Blau, H.M., 2004. Nat. Cell Biol. 6, 810 – 816.
Rosenblatt, J.D., Parry, D.J., 1992. J. Appl. Physiol. 73, 2538 – 2543.
Sherwood, R.I., Christensen, J.L., Conboy, I.M., Conboy, M.J., Rando,
T.A., Weissman, I.L., Wagers, A.J., 2004. Cell 119, 543 – 554.
344
A.T. Palermo et al. / Developmental Biology 279 (2005) 336–344
Smith, H.K., Plyley, M.J., Rodgers, C.D., McKee, N.H., 1997. Int. J. Sports
Med. 18, 94 – 100.
Theise, N.D., Henegariu, O., Grove, J., Jagirdar, J., Kao, P.N., Crawford,
J.M., Badve, S., Saxena, R., Krause, D.S., 2002. Exp. Hematol. 30,
1333 – 1338.
Van der Meeren, A., Monti, P., Lebaron-Jacobs, L., Marquette, C.,
Gourmelon, P., 2001. Radiat. Res. 155, 858 – 865.
Vassilopoulos, G., Wang, P.R., Russell, D.W., 2003. Nature 422, 901 – 904.
Wagers, A.J., Weissman, I.L., 2004. Cell 116, 639 – 648.
Wagers, A.J., Sherwood, R.I., Christensen, J.L., Weissman, I.L., 2002.
Science 297, 2256 – 2259.
Wang, X., Montini, E., Al-Dhalimy, M., Lagasse, E., Finegold, M.,
Grompe, M., 2002. Am. J. Pathol. 161, 565 – 574.
Weimann, J.M., Charlton, C.A., Brazelton, T.R., Hackman, R.C., Blau,
H.M., 2003a. Proc. Natl. Acad. Sci. 100, 2088 – 2093.
Weimann, J.M., Johansson, C.B., Trejo, A., Blau, H.M., 2003b. Nat. Cell
Biol. 5, 959 – 966.
Willenbring, H., Bailey, A.S., Foster, M., Akkari, Y., Dorrell, C., Olson,
S., Finegold, M., Fleming, W.H., Grompe, M., 2004. Nat. Med. 10,
744 – 748.
Wright, D.E., Wagers, A.J., Gulati, A.P., Johnson, F.L., Weissman, I.L.,
2001. Science 294, 1933 – 1936.