Hematopoietic cells and osteoblasts are derived
from a common marrow progenitor after
bone marrow transplantation
Massimo Dominici*, Colin Pritchard†, John E. Garlits*, Ted J. Hofmann*, Derek A. Persons*, and Edwin M. Horwitz*‡§
Divisions of *Experimental Hematology and‡Stem Cell Transplantation and†Department of Genetics, St. Jude Children’s Research Hospital, 332 North
Lauderdale Street, Memphis, TN 38105
Communicated by Darwin J. Prockop, Tulane University, New Orleans, LA, June 28, 2004 (received for review February 19, 2004)
Bone and bone marrow are closely aligned physiologic compart-
ments, suggesting that these tissues may represent a single func-
tional unit with a common bone marrow progenitor that gives rise
to both osteoblasts and hematopoietic cells. Although reports of
multilineage engraftment by a single marrow-derived stem cell
support this idea, more recent evidence has challenged claims of
stem cell transdifferentiation and therefore the existence of a
multipotent hematopoietic?osteogenic progenitor cell. Using a
repopulation assay in mice, we show here that gene-marked,
transplantable marrow cells from the plastic-nonadherent popu-
lation can generate both functional osteoblasts?osteocytes and
hematopoietic cells. Fluorescent in situ hybridization for the X and
Y chromosomes and karyotype analysis of cultured osteoblasts
confirmed the donor origin of these cells and excluded their
generation by a fusion process. Molecular analysis demonstrated a
common retroviral integration site in clonogenic hematopoietic
cells and osteoprogenitors from each of seven animals studied,
establishing a shared clonal origin for these ostensibly indepen-
dent cell types. Our findings indicate that the bone marrow
contains a primitive cell able to generate both the hematopoietic
and osteocytic lineages. Its isolation and characterization may
suggest novel treatments for genetic bone diseases and bone
This capacity may solely reflect the activities of multiple discrete
stem cells with restricted genetic programs, such as hematopoi-
etic stem cells able to differentiate to leukocytes, erythrocytes,
or megakaryocytes and mesenchymal stem cells that can differ-
entiate to bone, cartilage, or adipose tissue (3). Alternatively, the
bone marrow may contain rare ‘‘marrow stem cells’’ with the
potential to differentiate to a spectrum of tissues, possibly in
response to specific environmental cues. This paradigm is sup-
ported by the findings of Krause et al. (4), which indicate that a
single marrow stem cell can generate both hematopoietic and
nonhematopoietic lineages. However, that study lacked a unique
molecular marker for confirming the multipotentiality of single
transplanted stem cells; further, recent evidence shows that
many examples of so-called stem cell developmental plasticity
may represent fusions of hematopoietic stem cells with com-
mitted progenitors of nonhematopoietic tissues, generating
cell hybrids that possess both donor and host genetic informa-
Bone marrow and bone are anatomically contiguous tissues
that show parallel age-related changes and share several genetic
features (9–11), suggesting a close developmental relationship.
Thus, a reasonable hypothesis is that the hematopoietic marrow
harbors a stem cell with osteogenic potential. Consistent with
this idea is the observation, reported over a decade ago, that
nonadherent CD34 bone marrow cells can differentiate to
osteoblasts (12) and the quite recent report that murine bone
marrow side population (SP) cells can engraft in bone after
transplantation (13). Moreover, in our human cell therapy trials,
one marrow cells contribute to many diverse tissues after
systemic transplantation in both mice and humans (1, 2).
donor osteoblast engraftment was demonstrated after transplan-
tation of unmanipulated bone marrow (14), but the percentage
of such engraftment could not be improved by transplanting as
many as 5 ? 106isolated plastic-adherent marrow stromal cells
per kg of body weight, a cell number that greatly exceeds the
marrow stromal cell content of unmanipulated marrow (15).
One interpretation of these observations is that cells other than
those in the adherent population, where mesenchymal stem cells
are thought to reside (3), are potent transplantable progenitors
of osteoblasts, consistent with laboratory studies showing that
nonadherent cells can give rise to bone (12, 13, 16). We tested
this prediction in a murine transplantation model by using
gene-marked bone marrow cells and retroviral integration site-
specific PCR analysis.
flushed from the dissected femurs and tibias of FVB?N mice
(The Jackson Laboratory), and the isolated adherent marrow
cells were transduced with a GFP-expressing retroviral vector
(multiplicity of infection, ?5) as described (17). In separate
studies, nonadherent marrow cells, isolated from a 5-day ex vivo
culture in which adherent cells stuck to the plastic dish, were
transduced with the same GFP-expressing retroviral vector
(70–80% efficiency) and transplanted into 4- to 6-week-old
lethally irradiated (1,100 cGy) FVB?N mice (18). In this phase
of the study, 106nonadherent cells in 500 ?l of PBS were infused
into the tail veins of recipient mice at ?4 h postirradiation.
Southern Blot Analysis. Genomic DNA was isolated (PureGene
kit, Gentra Systems) from bone fragments that had been rigor-
ously flushed to remove marrow cells, repeatedly minced and
washed, and treated with collagenase to remove any blood cell
remnants, by using the same procedure applied in the prepara-
tion of osteoblasts (14). DNA was also isolated from murine
hematopoietic colony-forming unit-spleen (CFU-S) colonies
and from culture-expanded osteoblasts and stromal cells. The
resultant Southern blots were hybridized with a GFP-specific
probe and visualized with the Storm 860 PhosphorImager
Proviral Integration Analysis. Inverse PCR was performed as
described in detail elsewhere (19), except that primers B and C
were modified to complete complementarity to the murine stem
with the CpG methylation-insensitive TaqI restriction endonu-
clease. Primer sequences were as follows: VirA, 5?-TCCATGC-
Abbreviations: SP, side population; CFU-S, hematopoietic colony-forming unit-spleen;
FACS, fluorescence-activated cell sorting.
§To whom correspondence should be addressed at: Divisions of Stem Cell Transplantation
and Experimental Hematology, St. Jude Children’s Research Hospital, Mail Stop 321, 332
North Lauderdale Street, Memphis, TN 38105. E-mail: firstname.lastname@example.org.
© 2004 by The National Academy of Sciences of the USA
August 10, 2004 ?
vol. 101 ?
no. 32 ?
CTTGCAAAATGGC-3?; VirB-MSCV, 5?-AGGACCTGAA-
AATGACCCTGTGCCTTATT-3?; VirC-MSCV, 5?-TTACT-
TAAGCTAGCTTGCCAAACCTACAGGT-3?; and VirD, 5?-
CAACCCCTCACTCGGCGCGCCAGTC-3?. For integration
site analysis, 100 ng of DNA was amplified with site-specific
primers (forward, VirB-MSCV as above; reverse, 5?-AAAG-
CAAAAACAAAAATGGTTCCCTTTC-3?) for 25 cycles at
94°C for 1 min, 55°C for 1 min, and 72°C for 1 min. PCR products
containing32P-labeled nucleotides, were analyzed by polyacryl-
amide gel electrophoresis and visualized with the Storm 860
RT-PCR. RT-PCR was performed with the Titan One Tube
RT-PCR System (Roche Applied Sciences) according to the
manufacturer’s directions. Primer sequences were as follows:
?-actin, 5?-CATTGTGATGGACTCCGGAGACGG-3? and 5?-
CATCTCCTGCTCGAAGTCTAGAGC-3?; CD45, 5?-CTTC-
GACGGAGAGTTAATGC-3? and 5?-GTCGCCTTAGCTT-
GACAACA-3?; collagen I, 5?-GCAATCGGGATCAGTAC-
tin, 5?-TCACCATTCGGATGAGTCTG-3? and 5?-ACTTGTG-
GCTCTGATGTTCC-3?; and osteocalcin 5?-CTCTGTCT-
CTCTGACCTCACAG-3? and 5?-GGAGCTGCTGTGACA-
TCCATAC-3?. Conditions for amplification were 27 cycles at
94°C for 1 min, 55°C for 1 min, and 72°C for 1 min. Products
containing32P-labeled nucleotides were analyzed and visualized
Immunohistologic Staining. Immunohistochemical staining of cells
was performed according to standard protocols. Briefly, GFP
expression in sections of bone and cartilage was identified by
incubating formalin-fixed, decalcified, paraffin-embedded sec-
tions overnight at 4°C with a rabbit anti-GFP antibody (1:300)
(Molecular Probes). The primary antibody was visualized with a
room temperature, dilution of 1:200) and peroxidase-conjugated
avidin (ABC kit, Vector Laboratories) by using NovaRED
(Vector Laboratories) as a substrate; counterstaining was with
Harris hematoxylin (Surgipath Medical Industries, Richmond,
IL). Immunohistochemical staining of collagen I and collagen II
was performed with rabbit anticollagen primary antibodies
(Chemicon) and visualized by use of a biotinylated goat anti-
rabbit secondary antibody followed by peroxidase-conjugated
avidin with NovaRED as a substrate. Osteocalcin staining was
performed with a goat anti-mouse osteocalcin primary antibody
(Biomedical Technologies, Stoughton, MA) and visualized with
a biotinylated donkey anti-goat secondary antibody, as above.
For double immunohistochemical staining, GFP was visualized
with nickel DAB (Vector Laboratories) as a substrate. Negative
control specimens were serial sections of bone and cartilage
incubated with isotypic rabbit IgG primary antibody (Vector
Laboratories) and visualized in the same manner as the exper-
imental specimens. As an additional negative control, bone
sections from animals transplanted with freshly harvested (non-
transduced) nonadherent bone marrow cells were studied with
the rabbit anti-GFP primary antibody and visualized as above.
Background staining was not apparent.
Osteoblast Isolation. Osteoblasts were isolated from mouse bone
as described for human osteoblasts (14).
Stromal Cell Isolation. Stromal cells were isolated and expanded in
culture as described for human stromal cells (14). Fibroblastic
adherent cells that were CD45??CD11b?by flow cytometry
Adherent Marrow Cell Transplantation.To establish a suitable assay
for transplantable osteoprogenitor cells and to evaluate the
osteoprogenitor capacity of mesenchymal stem cells, we isolated
plastic-adherent marrow stromal cells from FVB?N mice and
transduced them (82–93% efficiency) with a murine stem cell
virus (MSCV) retroviral vector encoding GFP (17, 18). The cells
(106per mouse) were then infused into lethally irradiated (1,100
cGy, n ? 3) or sublethally irradiated (400 cGy, n ? 3) mice. Flow
cytometric and PCR analyses of peripheral blood demonstrated
the absence of GFP-transduced cells (data not shown), indicat-
ing that these transplanted adherent cells did not contribute to
hematopoietic reconstitution. The fraction of GFP-positive os-
teoblasts and osteocytes identified by immunohistochemical
staining at 3 months posttransplantation ranged from 0% to 2%
(median, 1.5%; no difference between groups; data not shown),
indicating that the systemically infused adherent marrow stromal
cells lack robust repopulating activity in this murine system.
Transplanted Nonadherent Marrow Cells Engraft in Bone. To test the
hypothesis that transplantable stem cells capable of generating
committed osteoprogenitor cells after systemic infusion reside
within the nonadherent marrow cell population, we isolated
nonadherent marrow cells and transduced them with a GFP-
encoding retroviral vector (20). Transduced GFP-expressing
cells were isolated by fluorescence-activated cell sorting (FACS;
?98% purity) and transplanted into lethally irradiated recipient
mice. Two to 6 months after transplantation, flow cytometric
analysis for GFP expression showed that bone marrow mono-
nuclear cells (mean percentage with the fluorescent marker,
?99%), blood leukocytes (mean, 96%), erythrocytes (mean,
92%), and platelets (mean, 95%) from seven mice were all
predominantly derived from transduced marrow cells.
Microscopic evaluation of bone sections from these mice
demonstrated GFP-expressing osteocytes and chondrocytes in
the metaphysis?epiphysis and articular cartilage, respectively
(Fig. 1 A–F). Double staining for GFP and collagen I, osteocal-
cin, or collagen II (Fig. 1 G–I) indicated osteogenic and chon-
drogenic differentiation and function, as well as engraftment, by
the transplanted marrow cells. Osteoclasts, large multinucleated
cells derived from the hematopoietic stem cell, were not ob-
served in the specimens shown but could be identified by
examining numerous sections [?1% of all bone cells were
osteoclasts, consistent with published findings (21)]. Scanning of
several histologic sections revealed a median of 18% gene-
marked cells per 20? field (range, 0–50%; n ? 60 fields). Quite
strikingly, the gene-marked cells were observed in clusters
accounting for up to 50% of the total cellular content of some
regions of bone and cartilage.
A Unique Bone Marrow Cell Gives Rise to Both Blood and Bone. To
address whether the same gene-marked, nonadherent marrow
cells that reconstituted hematopoiesis were also progenitors for
CFU-S clones derived from the bone marrow of a primary
transplant recipient and killed at 6 months posttransplantation.
Southern blot analysis of DNA isolated from these clones after
digestion with restriction enzymes revealed three unique pat-
terns designated type 1 (3 of 47 clones; Fig. 2A), type 2 (43 of 47
clones; Fig. 2A), and type 3 (1 of 47 clones; data not shown).
Similar analysis of DNA isolated from bone in the same mouse
revealed a restriction pattern that was a composite of the
patterns of two CFU-S colonies (Fig. 2A), providing evidence
for a common progenitor of the osteoblastic and hematopoietic
lineages. The complex restriction pattern in stromal cell DNA
suggests that many transduced, nonadherent progenitors had
contributed to the reconstitution of stroma over the first 6
www.pnas.org?cgi?doi?10.1073?pnas.0404626101 Dominici et al.
months posttransplantation. Although none of the cells seem
to have been generated by the specific CFU-S colonies we
identified, the possibility of their origin from a progenitor cell
with both hematopoietic and osteogenic potential cannot be
To confirm the Southern blot analysis indicating a common
progenitor for hematopoietic cells and bone cells (Fig. 2A), we
used inverse PCR to isolate a proviral integration site on
chromosome 3 in clone 16, a member of the predominant group
of CFU-S clones (type 2). Using site-specific primers (Fig. 2B),
we detected the predicted 616-bp PCR product from this clone
and then verified its presence in another CFU-S clone of the type
2 pattern (clone 17). By PCR analysis, the clonal integration site
in the hematopoietic CFU-S colonies was also present in bone.
These findings were extended by isolating an integration site
from the predominant CFU-S clones, identified by Southern blot
analysis, in two additional mice (Fig. 2C, animals 2 and 3). In
animal 2, site-specific PCR amplified the expected 361-bp
to osteocytes and chondrocytes. Immunohistochemical staining with horse-
radish peroxidase demonstrates GFP expression in bone, cartilage, and mar-
6 months after transplantation (A) and a negative control specimen (B).
Original magnification, ?100. (C and D) Enlarged view to demonstrate the
GFP-expressing cells within cartilage and bone (C), compared with the nega-
tive control (D). BM, bone marrow. Original magnification, ?400. (E and F)
High-power views demonstrating individual GFP-expressing osteocytes (E)
and chondroctyes (F). Arrows indicate positive cells. Original magnification,
?1,000. (G–I) Double immunohistochemical staining to demonstrate coex-
pression of GFP (black) and collagen I (orange) in osteocytes (G), osteocalcin
(orange) in metaphyseal osteoblasts (H), and collagen II (orange) in chondro-
cytes (I). Original magnification, ?1,000.
Engraftment and differentiation of transplanted nonadherent cells
1 and 2 were compared with those from bone and stroma. Isolated DNA (10
?g) was digested with the indicated restriction enzymes, and the resulting
blots were hybridized with a GFP-specific probe. (B) PCR analysis using inte-
gration site-specific primers of DNA isolated from two colonies of CFU-S clone
controls. PCR amplification of the globin ?-major sequences was used as a
control for the quality and quantity of DNA. (Upper) The integration site-
specific PCR target sequence is depicted schematically. (C) Integration site-
the CFU-S clone from which the integration site was isolated; another CFU-S
clone of the same origin (by Southern blot analysis) is also shown. Analysis of
DNA from bone marrow, peripheral blood, and splenocytes provides a com-
parison with other hematopoietic cells.
Retroviral integration site analysis of CFU-S cells and nonhematopoi-
Dominici et al.PNAS ?
August 10, 2004 ?
vol. 101 ?
no. 32 ?
product from chromosome 9 in the DNA of bone, splenocytes,
bone marrow, and peripheral blood. In animal 3, a 578-bp
product was amplified from chromosome 1 in splenocytes, bone,
and peripheral blood. Each of four additional transplanted
animals also had a common hematopoietic?osteoblastic integra-
tion site, as demonstrated by Southern blot analysis (Figs. 3 and
4 and data not shown)
Lack of Hematopoietic Contamination. To further exclude hemato-
poietic contamination as a major source of bias in these assays,
we obtained cultured mitotically active osteoblasts from bone
of a mouse killed at 2 months posttransplantation and sorted
them by FACS to eliminate possible contamination by
CD45??CD11b?(hematopoietic) cells. Subsequent flow cyto-
metric analysis demonstrated the lack of CD45??CD11b?(Fig.
3A) and CD34?cells (Fig. 3B). The cell population secreted an
extracellular matrix with mineralizing capacity as shown by
alizarin red staining (Fig. 3C) and expressed collagen I as shown
by immunocytochemical staining (Fig. 3D). One to 2% of the
sorted cells expressed GFP (donor marker; Fig. 3E), consistent
with our findings in previous clinical studies (14, 22). RT-PCR
analysis showed that the cells expressed collagen I, osteopontin,
and osteocalcin but not CD45 (Fig. 3F), indicative of the
osteoblast phenotype without blood cell contamination. Site-
specific PCR analysis demonstrated a common retroviral inte-
gration site in the hematopoietic and osteoblast DNA of this
mouse (Fig. 3G).
To enrich for donor cells, we next sorted the osteoblasts by
FACS for GFP expression, obtaining a population that was
?99% GFP-positive (Fig. 4A). Further analysis of the cells
confirmed the lack of CD45 and CD11b expression (Fig. 4B).
When the CD45??CD11b??CD34?GFP-positive cells were
expanded in culture and analyzed by RT-PCR for osteocalcin
highly enriched cells (Fig. 4C). Site-specific PCR analysis again
demonstrated a common retroviral integration site in DNA from
both hematopoietic and osteoblastic cells (Fig. 4D). Finally,
Southern blot analysis after digestion with three different re-
striction enzymes showed a common proviral integration pattern
(four integrants per cell) for CFU-S clone 18 and the GFP-
positive osteoblasts (Fig. 4E), confirming a common clonal
origin and excluding low-level contamination.
Lack of Cellular Fusion. Finally, to exclude the possibility that
fusion of hematopoietic stem cells with osteoprogenitors might
have been responsible for these observations, we performed
fluorescence in situ hybridization for the X and Y chromosomes
on cultured osteoblasts from three mice killed at 2–6 months
posttransplantation (Fig. 5). All osteoblasts (n ? 200 per
mouse), including all donor cells (denoted by the Y chromo-
some), contained 40 chromosomes (normal mouse), indicating
that cell fusion was not the source of the clonal bilineage
differentiation seen after bone marrow transplantation.
Our results provide compelling evidence that a unique progen-
itor cell with both hematopoietic and osteocytic differentiation
potential resides in the nonadherent subset of bone marrow cells.
This conclusion is reinforced by previous studies showing that
both hematopoietic and osteogenic cells express the AML?
Cbfa?Runx family of transcription factors (11), by in vitro data
suggesting the existence of a transforming growth factor ?1
(TGF-?1)-responsive premesenchymal?prehematopoietic mar-
row progenitor (23), and by evidence that endoglin (CD105), a
TGF-? receptor, may serve as a marker of both mesenchymal
stem cells (24) and long-term repopulating hematopoietic stem
cells (25, 26). Furthermore, sca-1-null mice have recently been
reported to have age-related osteoporosis as well as hematopoi-
etic deficiencies (9, 10).
The major strength of our analysis is the use of gene marking
of the normal cellular constituents of bone marrow. Retroviral
integration site analysis unequivocally identifies a single trans-
duced cell and all subsequent progeny, allowing the fate of
individual cells to be tracked within a population of transplanted
cells (27–29). This strategy allows in vivo differentiation to be
critically evaluated against results obtained in vitro, which often
do not represent normal cellular function. Moreover, the use of
isolated bone marrow cells in our experiments permitted insights
into normal marrow cell transplantation biology that would not
Graphs of cells sorted by FACS demonstrating the absence of hematopoietic
(CD45??CD11b?) cells (A) and marrow progenitor (CD34?) cells (B). Straight
Alizarin red staining revealing mineral deposition within the extracellular
matrix secreted by the cultured osteoblasts. (D) Immunofluorescence staining
of the cultured osteoblasts for collagen I. (E) Graph of CD45??CD11b??CD34?
the cultured osteoblasts. (G) PCR analysis of the cultured osteoblasts with
Clonal analysis of cultured osteoblasts by site-specific PCR. (A and B)
www.pnas.org?cgi?doi?10.1073?pnas.0404626101Dominici et al.
have been possible had the cells been subjected to extensive
processing or ex vivo cultivation. The theoretical possibility that
integration of the proviral sequence into the genome may have
altered the normal biology of the progenitor cells is highly
unlikely because animals with ?99% transduced marrow mono-
nuclear cells maintained functional engraftment in both marrow
and bone with normal blood and marrow cellularity. Addition-
ally, the animals were free of opportunistic infections and
hematologic or nonhematologic malignancies. In short, our
findings clearly implicate a distinct, physiologic bone marrow
progenitor cell as the source of both hematopoietic and osteo-
cytic lineages after transplantation of whole bone marrow.
Our immunohistochemical data showed uniform engraftment
of transplanted adherent cells throughout the histologic fields,
representing only 1.5% of osteocytes and osteoblasts. Trans-
plantation of nonadherent marrow cells, in contrast, yielded
clusters of donor cells that accounted for 18% of such bone cells.
These data indicate that nonadherent marrow cells have more
robust (1 log greater) bone-repopulating activity than do adher-
are two, presumably distinct, populations of marrow cells with
the capacity to generate osteoprogenitors. One source may be
more important for bone homeostasis, whereas the other may
contribute to the osteogenic compartment in response to phys-
iologic stress or to demands for the repair of injured bone.
Experimental data indicating the differentiation of a bone
marrow progenitor cell to hematopoietic and nonhematopoietic
cells must be interpreted cautiously. Fusion of marrow-derived
cells with mature nonhematopoietic tissue cells could have lent
the appearance of stem?progenitor cell differentiation, much in
the manner that KTLS cells, highly purified murine hematopoi-
etic stem cells (30), were shown to fuse with hepatocytes (7, 8,
31), rather than differentiate into them, as first reported (32).
Similarly, transplanted marrow cells can fuse with neural ele-
ments to generate Purkinje cells (31), in contrast to transdiffer-
entiation (33). In our analyses, we excluded fusion of hemato-
poietic and osteocytic progenitors by demonstrating the
presence of donor (Y chromosome) osteoblasts with a diploid
genome (Fig. 5), substantiating our contention that a single bone
marrow progenitor cell gave rise to mature cells in blood and
Although retroviral integration site analysis affords an ex-
tremely specific means of assessing clonality (27–29), low-level
contamination by extraneous cells could confound the assay.
Several lines of evidence indicate that such interference was not
a factor in the present study. First of all, in animal 3 (Fig. 2C),
DNA isolated from the stroma and hematopoietic marrow, both
potential sources of contaminating cells, did not contain the
retroviral sequences by PCR analysis and therefore could not
have contributed to the bone DNA signals. Second, two exper-
imental approaches demonstrated that the FACS-purified pop-
ulation of GFP-expressing osteoblasts (Fig. 4 A–C) contained
cells derived from a progenitor that also produced hematopoi-
etic cells. PCR analysis demonstrated a very strong site-specific
signal indicating that the target sequence was represented in the
vast majority of osteoblasts (Fig. 4D), whereas Southern blot
analysis revealed a single GFP-positive clone, the most prevalent
in the presumably polyclonal GFP-expressing osteoblast popu-
lation, that was derived from the same marrow progenitor as the
hematopoietic CFU-S clone 18 (Fig. 4E).
Olmsted-Davis et al. (13) recently described mouse long bone
engraftment by a population of transplanted marrow-derived SP
cells, which contains hematopoietic repopulating cells (34).
However, primitive progenitor cells with an SP-like phenotype
(ability to efflux fluorescent DNA-binding dye) have been
identified in a variety of tissues (35), and the marrow population
of SP cells is quite heterogeneous with respect to dye-exclusion
fluorescence after FACS to enrich the population for GFP expression, demon-
strating the homogeneity of GFP expression. (B) Graph of cells sorted by FACS
demonstrating the lack of CD45??CD11b?(hematopoietic) cells. (C) RT-PCR
analysis for osteocalcin expression by the pure GFP-positive cultured osteo-
blasts. (D) PCR analysis with site-specific primers of DNA from the same sorted
osteoblasts in C, demonstrating a common, unique integration site. (E) Inte-
gration site analysis of CFU-S and osteoblasts by Southern blotting after
digestion with the indicated restriction enzymes and hybridization with a
GFP-specific probe. DNA (5 ?g) was isolated from CFU-S clone 18, the colony
21, a different colony from the same animal; and osteoblasts sorted by FACS.
confirms that these cells originated from a single hematopoietic?osteoblastic
Clonal analysis of GFP-enriched osteoblasts by site-specific PCR and
chromosomes (normal mouse) were stained blue with 4?,6-diamidino-2-
phenylindole (DAPI). The X (red) and Y (green) chromosomes were identified
by fluorescence in situ hybridization, using chromosome-specific probes.
Representative cytogenetic analysis of cultured osteoblasts. The 40
Dominici et al. PNAS ?
August 10, 2004 ?
vol. 101 ?
no. 32 ?
capacity, progenitor activity, and surface antigen expression (34,
36). This raises the possibility that multiple, disparate SP cells,
which may include the previously described CD34-nonadherent
marrow cell (12), contributed to the bone engraftment reported
by Olmsted-Davis et al. Our data, by contrast, demonstrate that
a single gene-marked marrow cell can engraft and differentiate
to both blood and bone. Although the marrow progenitor cell we
identified could well reside in the SP cell population, it may also
represent other populations of hematopoietic progenitor cells
Although providing proof of principle that a common pro-
genitor for the hematopoietic and osteocytic lineages resides in
bone marrow, we were unable to ascertain the frequency or the
potential clinical utility of this shared progenitor cell. However,
the robust regeneration of both functional hematopoietic and
osteocytic target tissues (up to 50% of bone cells in some
sections) suggests a physiologically important progenitor cell
response rather than a rare stochastic event that typifies most
instances of ‘‘stem cell plasticity’’ (37). This striking engraftment
of marked progenitor cells might be explained by observations
that murine osteoblasts can be recruited from progenitor cells in
only a few days in response to changes in stressful stimuli (38),
such as chemotherapy or irradiation. Indeed, the robust osteo-
poietic engraftment we observed may well depend on the
marrow-ablative effects of radiation (39), although this rela-
tionship will need to be assessed in carefully controlled com-
petitive repopulation experiments. We would emphasize, how-
ever, that marrow cell engraftment in patients with genetic
disorders of bone may be adequate without the use of a
preparative regimen (15).
How durable is the regenerative contribution of this hemato-
poietic?osteocytic progenitor cell? In the present study, a sig-
nificant fraction of osteoblasts from mice killed at 2–6 months
posttransplantation were of donor origin, as were osteoblast
samples collected at 3 months posttransplantation from patients
(14, 22). However, in long-term follow-up studies of transplant
patients, samples of osteoblasts were exclusively of host origin
(40). Thus, bone repair and regeneration during the early
posttransplantation period seems to be driven by transplanted,
donor-derived marrow progenitor cells that either engraft di-
rectly in bone or are recruited from the marrow to bone. For
routine homeostasis, the integrity of bone seems to be main-
tained by repopulating cells normally present in the bone
microenvironment. This distinction between bone repair?
regeneration and maintenance will be important in developing
widely applicable cellular therapies for injured or genetically
We thank Dr. Richard Ashmun for flow cytometric analyses, Virginia
Valentine for cytogenetic analyses, John Gilbert for editorial review,
and Ms. Angie Williams for assistance in preparation of this manuscript.
This work was supported by Doris Duke Charitable Foundation Clinical
Scientist Development Award T99102B; National Heart, Lung, and
Blood Institute Clinical Scientist Award K08-HL0420; National Cancer
Institute Cancer Center Support CORE Grant P30-CA-21765; and the
American Lebanese Syrian Associated Charities.
1. Herzog, E. L., Chai, L. & Krause, D. S. (2003) Blood 102, 3483–3493.
2. Korbling, M., Katz, R. L., Khanna, A., Ruifrok, A. C., Rondon, G., Albitar, M.,
Champlin, R. E. & Estrov, Z. (2002) N. Engl. J. Med. 346, 738–746.
J. D., Moorman, M. A., Simonetti, D. W., Craig, S. & Marshak, D. R. (1999)
Science 284, 143–147.
4. 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.
5. Ying, Q. L., Nichols, J., Evans, E. P. & Smith, A. G. (2002) Nature 416, 545–548.
6. Terada, N., Hamazaki, T., Oka, M., Hoki, M., Mastalerz, D. M., Nakano, Y.,
Meyer, E. M., Morel, L., Petersen, B. E. & Scott, E. W. (2002) Nature 416,
7. Wang, X., Willenbring, H., Akkari, Y., Torimaru, Y., Foster, M., Al Dhalimy,
M., Lagasse, E., Finegold, M., Olson, S. & Grompe, M. (2003) Nature 422,
8. Vassilopoulos, G., Wang, P. R. & Russell, D. W. (2003) Nature 422, 901–904.
9. Ito, C. Y., Li, C. Y., Bernstein, A., Dick, J. E. & Stanford, W. L. (2003) Blood
10. Bonyadi, M., Waldman, S. D., Liu, D., Aubin, J. E., Grynpas, M. D. & Stanford,
W. L. (2003) Proc. Natl. Acad. Sci. USA 100, 5840–5845.
11. Lian, J. B., Balint, E., Javed, A., Drissi, H., Vitti, R., Quinlan, E. J., Zhang, L.,
van Wijnen, A. J., Stein, J. L., Speck, N., et al. (2003) J. Cell Physiol. 196,
13. Olmsted-Davis, E. A., Gugala, Z., Camargo, F., Gannon, F. H., Jackson, K.,
Kienstra, K. A., Shine, H. D., Lindsey, R. W., Hirschi, K. K., Goodell, M. A.,
et al. (2003) Proc. Natl. Acad. Sci. USA 100, 15877–15882.
14. Horwitz, E. M., Prockop, D. J., Fitzpatrick, L. A., Koo, W. W., Gordon, P. L.,
Neel, M., Sussman, M., Orchard, P., Marx, J. C., Pyeritz, R. E., et al. (1999) Nat.
Med. 5, 309–313.
15. Horwitz, E. M., Gordon, P. L., Koo, W. K., Marx, J. C., Neel, M. D., McNall,
R. Y., Muul, L. & Hofmann, T. (2002) Proc. Natl. Acad. Sci. USA 99,
16. Long, M. W. (2001) Blood Cells Mol. Dis. 27, 677–690.
17. Marx, J. C., Allay, J. A., Persons, D. A., Nooner, S. A., Hargrove, P. W., Kelly,
P. F., Vanin, E. F. & Horwitz, E. M. (1999) Hum. Gene Ther. 10, 1163–1173.
18. Persons, D. A., Allay, J. A., Allay, E. R., Smeyne, R. J., Ashmun, R. A.,
Sorrentino, B. P. & Nienhuis, A. W. (1997) Blood 90, 1777–1786.
19. van Lohuizen, M., Verbeek, S., Scheijen, B., Wientjens, E., van der Gulden, H.
& Berns, A. (1991) Cell 65, 737–752.
20. Persons, D. A., Allay, J. A., Allay, E. R., Ashmun, R. A., Orlic, D., Jane, S. M.,
Cunningham, J. M. & Nienhuis, A. W. (1999) Blood 93, 488–499.
21. Roodman, G. D. (1996) Endocr. Rev. 17, 308–332.
22. Horwitz, E. M., Prockop, D. J., Gordon, P. L., Koo, W. W., Fitzpatrick, L. A.,
Neel, M. D., McCarville, M. E., Orchard, P. J., Pyeritz, R. E. & Brenner, M. K.
(2001) Blood 97, 1227–1231.
23. Hall, F. L., Han, B., Kundu, R. K., Yee, A., Nimni, M. E. & Gordon, E. M.
(2001) J. Hematother. Stem Cell Res. 10, 261–271.
24. Barry, F. P., Boynton, R. E., Haynesworth, S., Murphy, J. M. & Zaia, J. (1999)
Biochem. Biophys. Res. Commun. 265, 134–139.
25. Chen, C. Z., Li, M., De Graaf, D., Monti, S., Gottgens, B., Sanchez, M. J.,
Lander, E. S., Golub, T. R., Green, A. R. & Lodish, H. F. (2002) Proc. Natl.
Acad. Sci. USA 99, 15468–15473.
26. Chen, C. Z., Li, L., Li, M. & Lodish, H. F. (2003) Immunity 19, 525–533.
27. Kim, H. J., Tisdale, J. F., Wu, T., Takatoku, M., Sellers, S. E., Zickler, P.,
Metzger, M. E., Agricola, B. A., Malley, J. D., Kato, I., et al. (2000) Blood 96,
28. Nolta, J. A., Dao, M. A., Wells, S., Smogorzewska, E. M. & Kohn, D. B. (1996)
Proc. Natl. Acad. Sci. USA 93, 2414–2419.
29. Schmidt, M., Carbonaro, D. A., Speckmann, C., Wissler, M., Bohnsack, J.,
Elder, M., Aronow, B. J., Nolta, J. A., Kohn, D. B. & von Kalle, C. (2003) Nat.
Med. 9, 463–468.
30. Spangrude, G. J., Heimfeld, S. & Weissman, I. L. (1988) Science 241, 58–62.
31. Alvarez-Dolado, M., Pardal, R., Garcia-Verdugo, J. M., Fike, J. R., Lee, H. O.,
Pfeffer, K., Lois, C., Morrison, S. J. & Alvarez-Buylia, A. (2003) Nature 425,
32. 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.
33. Brazelton, T. R., Rossi, F. M., Keshet, G. I. & Blau, H. M. (2000) Science 290,
34. Goodell, M. A., Rosenzweig, M., Kim, H., Marks, D. F., DeMaria, M., Paradis,
35. Abbott, B. L. (2003) Hematol. Oncol. 21, 115–130.
36. Storms, R. W., Goodell, M. A., Fisher, A., Mulligan, R. C. & Smith, C. (2000)
Blood 96, 2125–2133.
37. Orkin, S. H. & Zon, L. I. (2002) Nat. Immunol. 3, 323–328.
38. Turner, C. H., Owan, I., Alvey, T., Hulman, J. & Hock, J. M. (1998) Bone 22,
39. Pateder, D. B., Sheu, T. J., O’Keefe, R. J., Puzas, J. E., Schwarz, E. M.,
Constine, L. S., Okunieff, P. & Rosier, R. N. (2002) Radiat. Res. 157, 62–68.
40. Koc, O. N., Peters, C., Aubourg, P., Raghavan, S., Dyhouse, S., DeGasperi, R.,
Kolodny, E. H., Yoseph, Y. B., Gerson, S. L., Lazarus, H. M., et al. (1999) Exp.
Hematol. 27, 1675–1681.
www.pnas.org?cgi?doi?10.1073?pnas.0404626101Dominici et al.