In vivo directed differentiation of pluripotent stem
cells for skeletal regeneration
Benjamin Levia,1, Jeong S. Hyuna,1, Daniel T. Montoroa,1, David D. Loa, Charles K. F. Chanb, Shijun Huc, Ning Sunc,
Min Leed, Monica Grovaa, Andrew J. Connollye, Joseph C. Wuc, Geoffrey C. Gurtnera, Irving L. Weissmanb,2,
Derrick C. Wana,2, and Michael T. Longakera,b,2
aHagey Laboratory for Pediatric Regenerative Medicine, Department of Surgery, Plastic and Reconstructive Surgery Division,cDepartment of Medicine and
Radiology, andeDepartment of Pathology, Stanford University School of Medicine, Stanford, CA 94305;bInstitute for Stem Cell Biology and Regenerative
Medicine, Stanford University, Stanford, CA 94305; anddDivision of Advanced Prosthodontics, Biomaterials, and Hospital Dentistry, School of Dentistry,
University of California, Los Angeles, CA 90095-1668
Contributed by Irving L. Weissman, October 17, 2012 (sent for review July 20, 2012)
Pluripotent cells represent a powerful tool for tissue regeneration,
but their clinical utility is limited by their propensity to form tera-
tomas. Little is known about their interaction with the surround-
ing niche following implantation and how this may be applied to
promote survival and functional engraftment. In this study, we
evaluated the ability of an osteogenic microniche consisting of
a hydroxyapatite-coated, bone morphogenetic protein-2–releasing
poly-L-lactic acid scaffold placed within the context of a macroen-
vironmental skeletal defect to guide in vivo differentiation of both
embryonic and induced pluripotent stem cells. In this setting, we
found de novo bone formation and participation by implanted
cells in skeletal regeneration without the formation of a teratoma.
This finding suggests that local cues from both the implanted scaf-
fold/cell micro- and surrounding macroniche may act in concert to
promote cellular survival and the in vivo acquisition of a terminal
cell fate, thereby allowing for functional engraftment of pluripo-
tent cells into regenerating tissue.
Both human induced pluripotent stem cells (h-iPSCs) and em-
bryonic stem cells (h-ESCs) are capable of differentiating into
a multitude of cell types from each of three germ layers, allowing
investigators to devise novel platforms for research and thera-
peutic drug screening (3, 4). This same property has also made
these cells a much more powerful tool compared with mesen-
chymal stromal cells for regenerative medicine. In addition, as
h-iPSCs can be reprogrammed from a patient’s own somatic
cells, they have the added potential of mitigating some of the
concerns over immunogenic sequelae that are raised with other
cell types, yet simultaneously enabling development of patient-
specific disease modeling (5–7).
Despite dramatic progress made over recent years, widespread
application of pluripotent cells in clinical medicine has been ham-
pered by several challenges, chief among which is the propensity for
both h-iPSCs and h-ESCs to form tumors in vivo (8). As recent
studies have shown development of teratomas to directly correlate
withthe number ofresidualundifferentiatedcells implanted,several
strategies have been proposed to eliminate these persistent plurip-
can be completely successful in the context of the number of cells
required for in vivo tissue regeneration. Furthermore, few reports
have also demonstrated engraftment and functional integration of
how transplanted cells truly interact with the endogenous niche
in stabilizing fully pluripotent cells and guiding acquisition of cell
fate, while also minimizing teratoma formation (11).
In this study, we evaluated how a skeletal defect macroniche
combined with a pro-osteogenic biomimetic scaffold microniche
could provide cues affecting survival and differentiation of
implanted cells lacking in a developmental program. In response
luripotent stem cells hold significant promise for the treat-
ment of tissue deficiencies and other human diseases (1, 2).
to such an environment, not only did we find a high degree of
survival, but the transplanted pluripotent cells were also shown
to acquire a fully differentiated osteogenic state, integrating into
the surrounding bone without the formation of a teratoma. Our
data thus suggest that the surrounding niche is capable of not only
supporting cellular viability, but can also guide differentiation of
pluripotent cells for functional engraftment into regenerating tissue.
In Vitro Differentiation of Pluripotent Cells. As bone morphogenetic
proteins (BMPs) have been shown to both powerfully promote
osteogenesis and regulate differentiation of pluripotent cells, the
capacity for h-iPSCs and h-ESCs to respond to BMP-2 was first
evaluated (12–14). At baseline, pSmad1/5 could not be detected
in either type of pluripotent cell (Fig. S1 A and D). However,
culturing h-iPSCs or h-ESCs with BMP-2 (200 ng/mL) resulted in
increased levels of pSmad1/5, as demonstrated by Western blot
analysis just 2 h following treatment. Importantly, this result was not
accompanied by any appreciable change in baseline Smad5 levels
(Fig. S1 B and E). Therefore, the canonical BMP signal trans-
duction pathway is functionally active in both h-iPSCs and h-ESCs.
The effect of BMP-2 on pluripotency was next evaluated by
culturing cells in growth medium, standard osteogenic differen-
tiation medium (ODM), or ODM supplemented with BMP-2.
After 3 d in ODM, 83.5% of cells were still found to be stage-
specific embryonic antigen 4+(SSEA-4+) (Fig. S1G, Center). In
contrast, only 61.9% of cells were SSEA-4+when cultured with
BMP-2 (Fig. S1G, Right). This result was associated with an in-
crease in runt-related protein-2 (RUNX-2) expression, as 10.0%
of cells in ODM were RUNX-2+but 31.4% of cells in ODM +
BMP-2 were RUNX-2+(Fig. S1H, Center and Right, respectively).
Therefore, treatment of pluripotent cells with BMP-2 accelerated
acquisition of a more differentiated state.
To evaluate the ability to guide in vitro differentiation of
pluripotent cells along an osteogenic lineage, h-iPSCs and h-
ESCs were subsequently cultured in ODM supplemented
with BMP-2. Immunostaining these cells for octamer-binding
transcription factor 4 (OCT-4), NANOG, and SRY-related
HMG-box (SOX)-2 after 7 d demonstrated reduction of ex-
pression for each of these pluripotent genes (Fig. 1 A and B) (15).
Author contributions: B.L., J.S.H., D.T.M., I.L.W., D.C.W., and M.T.L. designed research;
B.L., J.S.H., D.T.M., D.D.L., C.K.F.C., M.G., and D.C.W. performed research; S.H., N.S., and
M.L. contributed new reagents/analytic tools; B.L., J.S.H., D.T.M., D.D.L., C.K.F.C., A.J.C.,
J.C.W., G.C.G., I.L.W., D.C.W., and M.T.L. analyzed data; and B.L., J.S.H., D.T.M., I.L.W.,
D.C.W., and M.T.L. wrote the paper.
The authors declare no conflict of interest.
1B.L., J.S.H., and D.T.M. contributed equally to this work.
2To whom correspondence may be addressed. E-mail: firstname.lastname@example.org, dwan@stanford.
edu, or email@example.com.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| December 11, 2012
| vol. 109
| no. 50
Not surprisingly, concurrent up-regulation of osteogenic markers
RUNX-2 and osteocalcin (OCN) was also appreciated among
h-iPSCs and h-ESCs cultured in ODM with BMP-2 (Fig. 1 C
and D) and Alizarin red staining revealed significantly greater
extracellular matrix mineralization (*P < 0.05) (Fig. 1 E and F).
To confirm our histological staining, gene-expression analysis by
quantitative real-time PCR was performed for RUNX-2 and
OCN. At 7 d following culture in ODM with BMP-2, significantly
increased levels of transcripts were noted for both of these genes
(*P < 0.05) (Fig. 1 G and H).
Osteogenic Microniche Promotes in Vivo Bone Regeneration. To
demonstrate the baseline in vivo pluripotent capacity of our h-
iPSC and h-ESC lines, kidney capsule injection of one-million
pluripotent cells was performed in immunodeficient athymic mice
(Fig. S2) (16). h-iPSCs were transduced with a firefly luciferase/
red fluorescent protein/herpes simplex virus thymidine kinase
triple-fusion reporter construct and h-ESCs were labeled with
a firefly luciferase/green fluorescent protein construct to follow in
vivo cell viability (17). With both cell types, there was a significant
increase in luciferase imaging over time consistent with biologic
activity and teratoma formation (Fig. S2B). Furthermore, tissue
formation from all three embryonic germ layers could be readily
observed histologically 8 wk following injection (Fig. S2C).
The ability for an increasingly osteogenic environment to
guide in vivo differentiation of h-iPSCs and h-ESCs was then
evaluated by manipulating the surrounding microniche upon
which cells were delivered. Because pluripotent cells can respond
to BMPs, scaffolds were constructed from hydroxyapatite-coated
poly-L-lactic acid engineered to release BMP-2 (HA-PLGA +
BMP-2) (18). One-million h-iPSCs or h-ESCs were seeded onto
this microniche and placed into the context of a larger 4-mm
critical-sized calvarial defect macroniche. These critical-sized
defects do not spontaneously heal and any significant bone re-
generation observed is a direct result of treatment (19). We
hypothesized that by placing pluripotent cells in such an envi-
ronment, robust osteogenesis could be stimulated and selected.
Immunostaining of the regenerate was performed 7 d after im-
plantation. The most intense staining for pSmad1 was appreci-
ated in defects treated with h-iPSCs or h-ESCs seeded onto
HA-PLGA + BMP-2 (Fig. 2A, Top, fourth and sixth box). No-
ticeably less staining was observed when cells were seeded onto
HA-PLGA alone or when no cells were seeded at all (Fig. 2A,
top, third and fifth box, and second box, respectively). Similar
findings were appreciated for RUNX-2 and OCN, with the
greatest staining demonstrated from pluripotent cells seeded on
HA-PLGA + BMP-2 (Fig. 2A, Middle and Bottom, respectively).
These findings therefore suggest greater in vivo osteogenic dif-
ferentiation for h-iPSCs and h-ESCs in response to a BMP-2–
releasing microniche compared with the standard HA-
NANOG, and SOX-2 for h-iPSCs (Top) and h-ESCs (Bottom) before treatment with ODM and (B) after 7 d of culture in ODM. (C) Immunofluorescent staining for
RUNX-2 and OCN before treatment with ODM and (D) after 7 d of culture in ODM. (Scale bars, 50 μm.) (E) Alizarin red staining of h-iPSCs and h-ESCs before
and after 7 d in ODM with (F) spectrophotometric quantification at both time points. (Scale bars, 75 μm.) (G) quantitative real-time PCR analysis of gene
transcripts for RUNX-2 and OCN in h-iPSCs and (H) h-ESCs (*P < 0.05).
Osteogenic differentiation medium stimulates in vitro osteogenic differentiation of pluripotent cells. (A) Immunofluorescent staining for OCT-4,
| www.pnas.org/cgi/doi/10.1073/pnas.1218052109 Levi et al.
Pluripotent markers were also evaluated by immunofluores-
cent staining of in vivo regenerates at 5 and 14 d following cel-
lular implantation. Although OCT-4, NANOG, SOX-2, SSEA-4,
TRA-1-60, and TRA-1-81 could all be detected at 5 d in the
region of the regenerate, these markers could no longer be
detected at 14 d when h-iPSCs or h-ESCs were seeded on HA-
PLGA (Fig. S3) (15). In contrast, these same markers were much
more difficult to detect at either 5 or 14 d when pluripotent cells
were seeded onto HA-PLGA + BMP-2, suggesting incorporation
of BMP-2 into the microniche accelerates loss of pluripotency by
both h-iPSCs and h-ESCs in vivo (Fig. S3).
Given the observed enhanced expression of osteogenic markers
in concert with down-regulation of pluripotent genes in response
to HA-PLGA + BMP-2, we next determined the capacity for
h-iPSCs and h-ESCs to promote in vivo bone regeneration.
MicroCT scans performed on skeletal defects treated with HA-
PLGA alone demonstrated minimal healing (Fig. 2B, first row,
and Fig. 2C), and h-iPSCs seeded onto this same scaffold re-
sulted in 49% regeneration by 8 wk (Fig. 2B, third row, and Fig.
2C). Using a more potent osteogenic microenvironment, HA-
PLGA + BMP-2 scaffolds alone were capable of inducing 67%
healing at 8 wk (Fig. 2B, second row, and Fig. 2C). This result
was likely secondary to BMP-2 stimulation of endogenous cells
surrounding the skeletal macroniche. h-iPSCs seeded onto HA-
PLGA + BMP-2 resulted in the greatest amount of bone re-
generation, with robust healing seen as early as 2 wk and com-
plete (96%) healing of the critical-sized defect at 8 wk (Fig. 2B,
fourth row, and Fig. 2C; *P < 0.05 for each respective time
point). Similar results were observed with h-ESCs, as bone re-
generation from cells seeded onto HA-PLGA + BMP-2 (99%
healing) far outpaced that observed when HA-PLGA was used
(Fig. 2B, fifth and sixth rows, and Fig. 2C; *P < 0.05 for each
respective time point). Therefore, the HA-PLGA + BMP-2
microniche placed within the larger context of a skeletal defect
macroniche was highly effective at promoting in vivo pluripotent
cell bone formation and repair of a critical-sized defect. Finally,
treatment groups were followed out to 28 wk to confirm dura-
bility of our findings, with little to no change noted beyond 8 wk
by microCT (Fig. S4).
Bone Formation by Pluripotent Cells Without Teratoma Formation.
Histological analysis with aniline blue and pentachrome staining
was performed on sections to evaluate the quality of the regenerate.
Robust bone formation was best appreciated in defects treated
with pluripotent cells seeded onto HA-PLGA + BMP-2 (Fig. 2 D
and E). Importantly, significant bone overgrowth was not ap-
preciated and no teratoma formation was observed when
h-iPSCs were implanted onto HA-PLGA (0 of 15 animals) or
HA-PLGA + BMP-2 (0 of 12 animals) (Table S1). With h-ESCs,
only two teratomas were appreciated among sections from five
animals with pluripotent cells seeded on HA-PLGA and 10
animals with HA-PLGA + BMP-2 (Table S1). This was not
entirely unexpected given the potency and differentiability of the
one-million h-ESCs implanted and known differences at the
epigenomic level with h-iPSCs incurred during reprogramming
and prolonged passage in vitro (20–24). Nonetheless, the in-
cidence of teratoma formation in the context of an osteogenic
microniche from both h-iPSCs and h-ESCs was only 2 of
Pluripotent Cells Directly Contribute to in Vivo Osteogenesis. To
confirm contribution of implanted pluripotent cells to the bony
regenerate, immunohistochemistry for human nuclear antigen
was performed (Fig. 3 A and B). Among defects with h-iPSCs on
HA-PLGA + BMP-2, many positively stained cells were appre-
ciated in either the osteogenic stroma or in appositional osteo-
blastic arrangements along the surface of newly formed bone
(Fig. 3A). Interestingly, osteoclasts and osteocytes were found to
be of mouse origin, but more than half of the osteogenic cells
were human. This finding may reflect BMP-2 stimulation of host
mouse cells from the underlying dura mater. Similar findings
were also appreciated with h-ESCs implanted on HA-PLGA +
BMP-2, as human cells were readily appreciated within the re-
generate (Fig. 3B). Immunofluorescent staining of the re-
generate for human nuclear antigen and OCN demonstrated
colocalization of these markers in defects treated with either
h-iPSCs or h-ESCs seeded on HA-PLGA + BMP-2 (Fig. 3C,
third and fourth rows, respectively). No staining for either pro-
tein was appreciated with HA-PLGA alone. Interestingly, at the
late time point of 8 wk, no human cells could be detected in the
regenerate. This finding, however, may be consistent with bone
turnover by the above noted mouse-derived osteoclasts. Finally,
microdissection of the regenerate was performed for PCR analysis
using human specific osteogenic primers (Table S2). In defects
treated with h-iPSCs on HA-PLGA + BMP-2, both hOCN and
hRUNX-2 could be detected (Fig. 3D, Lower, third column);
Immunohistochemistry staining for pSmad1 (Top), RUNX-2
(Middle), and OCN (Bottom) 7 d after treatment of critical-
sized calvarial defects with: (i) no antibody control, (ii) HA-
PLGA + BMP-2 scaffold alone, (iii) HA-PLGA with h-iPSCs,
(iv) HA-PLGA + BMP-2 with h-iPSCs, (v) HA-PLGA with h-
ESCs, and (vi) HA-PLGA + BMP-2 with h-ESCs. (Scale bars,
100 μm.) (B) Representative microCT images of the re-
generate at 1, 2, 4, and 8 wk after injury and treatment.
Treatment groups included: (i) HA-PLGA scaffold alone,
(ii) HA-PLGA + BMP-2 scaffold alone, (iii) HA-PLGA with h-
iPSCs, (iv) HA-PLGA + BMP-2 with h-iPSCs, (v) HA-PLGA
with h-ESCs, and (vi) HA-PLGA + BMP-2 with h-ESCs. (C)
Quantification of percent healing in the region of the
defect at 1, 2, 4, and 8 wk following injury (*P < 0.05 for
each respective time point). (D) Aniline blue (Left) and
pentachrome staining (Right) of regenerates at 8 wk. For
orientation, the lower panel highlights region of interest
in the box. PB, parietal bone. (Scale bars, 1 mm.) (E) Quan-
tification of Aniline blue staining demonstrating total
amount of osteoid in the defect.
In vivo bone regeneration by pluripotent cells. (A)
Levi et al.PNAS
| December 11, 2012
| vol. 109
| no. 50
however, neither were observed when defects were treated with
HA-PLGA + BMP-2 scaffolds alone (Fig. 3D, Lower, fifth col-
umn). Evaluation of the contralateral uninjured mouse parietal
bone demonstrated absence of hOCN and hRUNX-2, confirm-
ing specificity of these primers for human transcripts (Fig. 3D,
Lower, fourth column). Collectively, these data therefore strongly
suggest that pluripotent cells placed onto a HA-PLGA + BMP-2
microniche within the larger context of a skeletal defect macro-
niche are capable of contributing to new bone formation with a
greatly reduced propensity to form teratoma in vivo.
Extraskeletal Bone Induction by Osteogenic Microniche. To evaluate
the capacity for HA-PLGA + BMP-2 scaffolds to guide in vivo
osteogenic differentiation of pluripotent cells irrespective of mac-
roenvironmental cues, extraskeletal subcutaneous implantation of
h-iPSC and h-ESC on HA-PLGA + BMP-2 was performed. His-
tological analysis 8 wk later revealed de novo bone formation by
bothpluripotent cell types (threeofseven animals, h-iPSC;threeof
four animals, h-ESC) (Fig. 4A and Table S1). Furthermore, im-
munofluorescent staining for both human nuclear antigen and
OCN demonstrated colocalization and human nuclear antigen
could be detected in the region of new bone by immunohisto-
chemistry (Fig.4 B andC). In contrast, nobonewasobserved when
the inability for the extraskeletal macroniche to promote bone
formation by implanted cells (Table S1). Thus, using multiple mo-
dalities, we have demonstrated that a strongly osteogenic micro-
niche (HA-PLGA + BMP-2) is sufficient to guide bone formation
by pluripotent cells in an in vivo extraskeletal macroenvironment.
These findings reveal that acquisition of terminal cell fate by
directly implanted pluripotent cells may be guided by local in
vivo cues provided by the surrounding niche. In the setting of our
strong osteogenic microniche placed within the context of
a larger skeletal macroniche, h-iPSCs and h-ESCs were observed
to undergo full differentiation and functional integration into
newly forming bone regenerate. Surprisingly, this effect was so
dominant that placement into such an environment of one-mil-
lion pluripotent cells resulted in a very low frequency of tera-
toma formation. To our knowledge, this finding represents
a unique demonstration of a direct effect from the surrounding,
albeit artificial niche on implanted pluripotent cells, guiding in
vivo engraftment and formation of bone in a meaningful manner.
Of note, the ability for in vivo induction of hESCs has been
demonstrated with urogenital sinus and seminal vesicle mesen-
chyme, as such a heterospecific tissue recombination approach
has been shown to be capable of directing small pluripotent cell
aggregates to prostate-like tissue (25). However, Taylor et al.
used far fewer cells (1 × 103) per construct, demonstrating ex-
pression of prostate specific antigen and other glandular epi-
thelial markers only on a microscopic scale (25). Furthermore,
use of such few cells would not be expected to typically yield
teratoma formation (9). Alternatively, LeBleu et al. have shown
the functional incorporation of h-ESCs into the kidney of a
Col4A3 knockout mouse model of Alport syndrome following
intravascular injection of one-million pluripotent cells (26). Al-
though histological improvement in glomerular basement mem-
brane and ultimate kidney function was observed, no evaluation of
teratoma formation was provided (26). Importantly, such an ap-
proach for cellular delivery has been shown to result in prolonged
localization of ESCs to the spleen and lung and observation of
teratoma development (27).
Recent studies have also evaluated the utility of pluripotent
cells in bone regenerative strategies (28–30). In all of these
reports, however, an extended period of ex vivo culture was re-
quired before implantation of constructs (28–30). Although
some have reported a reduction in teratoma formation, each of
these investigations have required in vitro predifferentiation of
h-ESCs into mesenchymal cells before seeding onto osteo-
conductive scaffolds for in vivo bone formation (30). These
studies therefore serve to underscore the significance of our
findings as we were able to demonstrate formation and func-
tional integration of de novo bone from direct placement of
generate. (A) Immunohistochemical stain for human nu-
clear antigen (HuNu) 4 wk after h-iPSC implantation. Note
that human cell (h) in apposition to adjacent mouse oste-
oblast (m). Osteocytes (ocyte) and osteoclasts (ocl) did not
stain for HuNu. (Scale bars, 35 μm for Upper photographs
and 15 μm for Lower photographs.) (B) Immunohistochemical
stain for HuNu three weeks after h-ESC implantation. Arrows
show positively stained human-derived osteogenic cells.
(Scale bars, 50 μm.) (C) Confocal microscopy of regenerate
site 4 wk following h-iPSC implantation and 2 wk following
h-ESC implantation. Immunofluorescent stain indicates cells
positive for HuNu (green) in the nucleus and OCN (red) in
cytosol. Rows: first row, no primary control; second row
staining for HA-PLGA scaffold alone; third and fourth rows,
h-iPSCs and h-ESCs with HA-PLGA + BMP-2 scaffold, re-
spectively. (Scale bars, 25 μm.) (D) (Upper) Specificity of
human primers for hOCN and hRUNX-2 using human and
mouse cDNA samples. (Lower) PCR of microdissection sam-
ples from defect site at 4 wk. Dissected regions shown be-
low in H&E stain, outlined in black boxes. Positive control
provided by in vitro osteogenic differentiated h-iPSCs.
hOCN and hRUNX-2 could be detected in regenerate from
defects treated with h-iPSCs on HA-PLGA + BMP-2 scaf-
folds. Neither transcript was observed from contralateral
uninjured mouse parietal bone or defects with HA-PLGA +
BMP-2 scaffolds alone. (Scale bars, 2 mm.)
Contribution of pluripotent cells to the bony re-
| www.pnas.org/cgi/doi/10.1073/pnas.1218052109 Levi et al.
pluripotent cells into skeletal defects. Furthermore, the ability to
minimize teratoma formation emphasizes the critical role of the
local environmental niche in guiding differentiation of implanted
stem cells. Despite delivery of one-million pluripotent cells, our
highly osteogenic microniche, in the context of a larger skeletal
defect macroniche, was capable of directing acquisition of cell
fate while simultaneously mitigating teratoma risk.
Using our model, we noted direct contribution of implanted
pluripotent cells to the early bony regenerate in skeletal defects.
However, the presence of human-derived cells at later time
points could not be demonstrated. Despite this inability, stable
bone formation was observed out to 28 wk, and the loss of
implanted human cells was likely secondary to bone turnover.
Bone formation and bone resorption are tightly coupled pro-
cesses, and circulating mouse-derived osteoclasts may have been
stimulated by either pluripotent cells themselves or by the BMP-
2 releasing microniche (31, 32). It can also be argued that either
the implanted cells or the HA-PLGA + BMP-2 scaffold may
have also promoted long-term viability of host osteoblasts through
enhanced vascularity. Multiple studies have shown h-iPSCs and
BMP-2 to promote revascularization of ischemic wounds and up-
regulation of various angiogenic genes (33–35).
Nevertheless, the unique model used in this present work for
directed differentiation (or selective amplification of subsets) of
pluripotent cells represents significant progress in our under-
standing of the interplay between implanted cells and their sur-
rounding niche. The acquisition of terminal cell fate by pluripotent
cells in the in vivo setting may thus be coaxed from complex
spatiotemporal signals and physical interaction with a specific
endogenous environment that help to mediate functional tissue
regeneration. Additional work still needs to be done, though, to
precisely characterize what each of these signals and interactions
are. Furthermore, although we have demonstrated this finding in
the context of a bone-promoting niche, other tissue regeneration
paradigms would benefit from a similar characterization of niche–
stem cell interactions. Finally, from a therapeutic standpoint, the
potential exists to fabricate alternative biomimetic scaffolds ca-
pable of directing differentiation of h-iPSCs or h-ESCs not just
to bone, but also to cartilage, fat, nerve, muscle, glandular tissue,
and so forth, yet simultaneously avoid teratoma formation (36,
37). Promising results have already been obtained using various
nanomaterial scaffolds to craft specific niches supportive of both
short-term cellular adhesion and proliferation, as well as longer-
term viability, lineage differentiation, and functionalization
(38, 39). Application of these unique biomimetic scaffolds to
regenerative strategies using pluripotent cells may allow for the
future development of innovative therapies to treat a wide range
Materials and Methods
Cell Harvest and Derivation. Human adipose-derived stromal cells were iso-
lated from the lipoaspirate of patients without medical comorbidities.
Reprogramming to h-iPSCs was performed using a lentiviral vector (OCT-4,
SOX-2, KLF-4, andc-MYC) andpluripotent characterization was performed by
immunostaining for common h-ESC markers, pluripotency gene-expression
analysis, and evaluation of promoter methylation status, as previously de-
scribed (15). h-ESCs were derived from the H9 cell line (National Institutes of
Health ID WA09).
Western Blot. Pluripotent cells were incubated in standard RIPA buffer with
0.5% phosphatase inhibitor and 0.5% protease inhibitor mixture mix (Sigma-
Aldrich). Separation was performed on a NuPAGE Novex 4–12% Bis•Tris gel
(Invitrogen) and membranes were probed with either monoclonal rabbit
antiphospho-Smad1/5, anti-Smad5, or anti–α-tubulin antibodies in 1:1,000
dilution (Cell Signaling Technology). Detection was performed by enhanced
chemiluminescence using the appropriate horse-radish peroxidase-linked
secondary antibody (Jackson ImmunoResearch Laboratories).
Assessment of in Vitro Osteogenic Differentiation. Equal numbers of plurip-
otent cells were cultured on Matrigel in six-well plates. Cells were treated
with mTeSR-1 or ODM (consisting of DMEM, 10% FBS, 100 μg/mL ascorbic
acid, and 10 mM β-glycerophosphate) with or without rhBMP-2 (200 ng/mL).
To evaluate marker expression, cells were trypsinized, resuspended in FACS
buffer (2% FBS in PBS) and blocked before staining with fluorochrome-
conjugated antibodies against SSEA-4 or RUNX-2 for flow cytometry. To
evaluate osteogenesis, immunofluorescent staining was performed for OCT-
4, NANOG, and SOX-2 at day 0 and day 7 of culture. Alizarin red staining
was performed following 7 d of osteogenic differentiation, as previously
Kidney Capsule Injection. Evaluation of pluripotency through teratoma for-
mation was assessed by injecting one-million h-iPSCs or h-ESCs beneath the
kidney capsule of adult (60-d-old) Crl:NU(NCr)-Foxn1nuCD-1 nude mice
(Charles River) (16). All animals were cared for in accordance with approved
protocols by the Institutional Animal Care and Use Committee at Stanford
(A) Aniline blue, pentachrome, and H&E stains 8 wk after
subcutaneous implantation of h-iPSCs and h-ESCs on HA-
PLGA + BMP-2. (Scale bars, 1 mm.) (B) Confocal microscopy
showing colocalization of human nuclear antigen (HuNu) in
green with OCN (red) for h-iPSCs on HA-PLGA + BMP-2 at 4
wk (third row) and h-ESCs on HA-PLGA + BMP-2 at 2 wk
(fourth row). (Scale bars, 25 μm.) (C) Immunohistochemical
human nuclear antigen stain following subcutaneous im-
plantation of h-iPSCs at 4 wk (Upper) and h-ESCs (Lower) on
HA-PLGA + BMP-2 scaffolds at 3 wk. Arrows demonstrate
positively stained human-derived cells adjacent to de novo
bone. (Scale bars, 25 μm for Upper h-iPSC photographs and
35 μm for Lower h-ESC photographs.)
Extraskeletal bone formation by pluripotent cells.
Levi et al. PNAS
| December 11, 2012
| vol. 109
| no. 50
University. After 8 wk, tumors were dissected and fixed. Parrafin embedded Download full-text
sections were stained with H&E.
Preparation and Seeding of Scaffolds. Hydroxyapatite-coated PLGA scaffolds
were fabricated from 85 of 15 poly(lactic-coglycolic acid) by solvent casting
and a particulate leaching process as previously described (41). For BMP-2–
loaded scaffolds, 1.25 μg of recombinant human BMP-2 (rhBMP-2; Med-
tronic) was adsorbed onto fabricated scaffolds at a final concentration of
200 μg/mL and scaffolds were further lyophilized on a freeze drier overnight
(Labconco). Scaffolds were seeded with one-million h-iPSCs or h-ESCs 8 h
before implantation (42).
Creation of Skeletal Defects. Critical sized 4-mm skull defects were created in
the right parietal skull bones of adult (60-d-old) male Crl:NU(NCr)-Foxn1nu
CD-1 nude mice. Scaffolds were seeded with one million h-iPSCs or h-ESCs 8 h
before implantation (42). Animals were divided into six groups (n = mini-
mum of five mice per group): (i) HA-PLGA scaffold without cells; (ii) HA-
PLGA + BMP-2 scaffold without cells; (iii) HA-PLGA scaffold with h-iPSCs; (iv)
HA-PLGA + BMP-2 scaffold with h-iPSCs; (v) HA-PLGA scaffold with h-ESCs;
and (vi) HA-PLGA + BMP-2 scaffold with h-ESCs.
Subcutaneous Scaffold Placement. A 1-cm incision was made just over the
inguinal fat pad of nude mice and a subcutaneous pocket was dissected.
h-iPSCs and h-ESCs seeded HA-PLGA + BMP-2 scaffolds were placed under
the skin just over this fat pad. Specimens were removed at four weeks for
h-iPSCs and 3 wk for h-ESCs to evaluate bone formation.
Histological Staining. Human cells were positively stained with human specific
anti-nuclear antigen antibodies (clone 235–1 mAB 1281; Millipore) using
a mouse on mouse staining kit (Vector Laboratories) on 8-μm formalin-fixed
paraffin-embedded tissue, according to the manufacturer’s instructions.
Immunohistochemistry and immunofluorescent staining were performed
using Vectastain Elite kits (Vector Laboratories) on 8-μm formalin-fixed par-
affin-embedded tissue according to the manufacturer’s instructions. Immu-
nohistochemistry was exposed using a biotinylated secondary antibody and
DAB kit, and immunofluorescence was visualized with fluorescent secondary
antibodies (Alexa Fluor; Invitrogen) on a confocal microscope. Aniline blue
stain was carried out to stain collagen within bone tissue according to
Masson’s trichrome method. De novo bone was imaged with light microscopy
and quantified with Adobe Photoshop CS5 (Adobe Systems) by pixel density.
Pentachrome stain was carried out according to Movat’s method.
In Vivo Imaging. Micro-CT was performed on live animals postoperatively
(through 8-wk healing) using a high-resolution MicroCAT II (ImTek) imaging
system (43). An in vivo imaging system was performed using the IVIS 200B
imaging system. Luciferin (150 mg/kg in 200 μL) was injected into the peri-
toneal cavity and, after 10 min, the animals were placed into the imaging
device. Images were acquired over 3 min (43).
ACKNOWLEDGMENTS. We thank Shane Morrison for his assistance with
bioluminescent imaging; Divya Nag for her assistance with kidney capsule
injection; Emily R. Nelson for herassistance with histologicalsectioning; Li Wang
for his assistance with PCR experiments; and Andrew Lee for his assistance with
pluripotent cell lines. This work was funded in part by National Institutes of
Health (NIH) Grant 1F32AR057302-02 (to B.L.); NIH Grants 5U01HL099776,
DP2OD004437, AG036142, and AI085575 (to J.C.W.); NIH Grants 1 R21
DE019274-01, RC2 DE020771-0, RC1 HL100490, 5U01HL099776, the Oak
Foundation, California Institute for Regenerative Medicine Grant RL1-
00662, and Hagey Laboratory for Pediatric Regenerative Medicine (to
M.T.L.); and NIH Grants 5RC2DE20771-2, R01-HL058770, and the Stinehart/
Reed awards (to I.L.W.).
1. Tsuji O, et al. (2010) Therapeutic potential of appropriately evaluated safe-induced
2. Furth ME, Atala A (2009) Stem cell sources to treat diabetes. J Cell Biochem 106(4):
3. Thomson JA, et al. (1998) Embryonic stem cell lines derived from human blastocysts.
4. Tiscornia G, Vivas EL, Izpisúa Belmonte JC (2011) Diseases in a dish: Modeling human
genetic disorders using induced pluripotent cells. Nat Med 17(12):1570–1576.
5. Wernig M, et al. (2007) In vitro reprogramming of fibroblasts into a pluripotent ES-
cell-like state. Nature 448(7151):318–324.
6. Zhao T, Zhang ZN, Rong Z, Xu Y (2011) Immunogenicity of induced pluripotent stem
cells. Nature 474(7350):212–215.
7. Apostolou E, Hochedlinger K (2011) Stem cells: iPS cells under attack. Nature 474
8. Tang C, et al. (2011) An antibody against SSEA-5 glycan on human pluripotent stem
cells enables removal of teratoma-forming cells. Nat Biotechnol 29(9):829–834.
9. Miura K, et al. (2009) Variation in the safety of induced pluripotent stem cell lines. Nat
10. Cao F, et al. (2007) Molecular imaging of embryonic stem cell misbehavior and suicide
gene ablation. Cloning Stem Cells 9(1):107–117.
11. Moore KA, Lemischka IR (2006) Stem cells and their niches. Science 311(5769):
12. Yu P, Pan G, Yu J, Thomson JA (2011) FGF2 sustains NANOG and switches the
outcome of BMP4-induced human embryonic stem cell differentiation. Cell Stem Cell
13. Xu RH, et al. (2002) BMP4 initiates human embryonic stem cell differentiation to
trophoblast. Nat Biotechnol 20(12):1261–1264.
14. Xu RH, et al. (2005) Basic FGF and suppression of BMP signaling sustain un-
differentiated proliferation of human ES cells. Nat Methods 2(3):185–190.
15. Sun N, et al. (2009) Feeder-free derivation of induced pluripotent stem cells from
adult human adipose stem cells. Proc Natl Acad Sci USA 106(37):15720–15725.
16. Wesselschmidt RL (2011) The teratoma assay: An in vivo assessment of pluripotency.
Methods Mol Biol 767:231–241.
17. Su W, et al. (2011) Bioluminescence reporter gene imaging characterize human em-
bryonic stem cell-derived teratoma formation. J Cell Biochem 112(3):840–848.
18. Kim S, et al. (2008) In vivo bone formation from human embryonic stem cell-derived
osteogenic cells in poly(d,l-lactic-co-glycolic acid)/hydroxyapatite composite scaffolds.
19. Cowan CM, et al. (2004) Adipose-derived adult stromal cells heal critical-size mouse
calvarial defects. Nat Biotechnol 22(5):560–567.
20. Phanstiel DH, et al. (2011) Proteomic and phosphoproteomic comparison of human ES
and iPS cells. Nat Methods 8(10):821–827.
21. Takahashi K, et al. (2007) Induction of pluripotent stem cells from adult human
fibroblasts by defined factors. Cell 131(5):861–872.
22. Takahashi K, Yamanaka S (2006) Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell 126(4):663–676.
23. Nishino K, et al. (2011) DNA methylation dynamics in human induced pluripotent
stem cells over time. PLoS Genet 7(5):e1002085.
24. Bilic J, Izpisua Belmonte JC (2012) Concise review: Induced pluripotent stem cells versus
embryonic stem cells: Close enough or yet too far apart? Stem Cells 30(1):33–41.
25. Taylor RA, et al. (2006) Formation of human prostate tissue from embryonic stem
cells. Nat Methods 3(3):179–181.
26. LeBleu V, et al. (2009) Stem cell therapies benefit Alport syndrome. J Am Soc Nephrol
27. Huang NF, et al. (2010) Embryonic stem cell-derived endothelial cells engraft into the
ischemic hindlimb and restore perfusion. Arterioscler Thromb Vasc Biol 30(5):984–991.
28. Kuznetsov SA, Cherman N, Robey PG (2011) In vivo bone formation by progeny of
human embryonic stem cells. Stem Cells Dev 20(2):269–287.
29. Arpornmaeklong P, Brown SE, Wang Z, Krebsbach PH (2009) Phenotypic character-
ization, osteoblastic differentiation, and bone regeneration capacity of human em-
bryonic stem cell-derived mesenchymal stem cells. Stem Cells Dev 18(7):955–968.
30. Marolt D, et al. (2012) Engineering bone tissue from human embryonic stem cells.
Proc Natl Acad Sci USA 109(22):8705–8709.
monocytes/macrophages depends on oncostatin M signaling. Stem Cells 30(4):762–772.
32. Zheng Y, et al. (2012) BMP2/7 heterodimer can modulate all cellular events of the in
vitro RANKL-mediated osteoclastogenesis, respectively, in different dose patterns.
Tissue Eng Part A 18(5-6):621–630.
33. Levi B, et al. (2012) Enhancement of human adipose-derived stromal cell angiogenesis
through knockdown of a BMP-2 inhibitor. Plast Reconstr Surg 129(1):53–66.
34. Suzuki H, et al. (2010) Therapeutic angiogenesis by transplantation of induced plu-
ripotent stem cell-derived Flk-1 positive cells. BMC Cell Biol 11:72.
35. Margariti A, et al. (2012) Direct reprogramming of fibroblasts into endothelial cells
capable of angiogenesis and reendothelialization in tissue-engineered vessels. Proc
Natl Acad Sci USA 109(34):13793–13798.
36. Nelson TJ, Martinez-Fernandez A, Terzic A (2010) Induced pluripotent stem cells:
Developmental biology to regenerative medicine. Nat Rev Cardiol 7(12):700–710.
37. Pozzobon M, Ghionzoli M, De Coppi P (2010) ES, iPS, MSC, and AFS cells. Stem cells
exploitation for Pediatric Surgery: Current research and perspective. Pediatr Surg Int
38. Zhao C, Tan A, Pastorin G, Ho HK (2012) Nanomaterial scaffolds for stem cell pro-
liferation and differentiation in tissue engineering. Biotechnol Adv (Aug):9.
39. Ku SH, Lee SH, Park CB (2012) Synergic effects of nanofiber alignment and electro-
activity on myoblast differentiation. Biomaterials 33(26):6098–6104.
40. Wan DC, et al. (2006) Osteogenic differentiation of mouse adipose-derived adult
stromal cells requires retinoic acid and bone morphogenetic protein receptor type IB
signaling. Proc Natl Acad Sci USA 103(33):12335–12340.
41. Chou YF, Huang W, Dunn JC, Miller TA, Wu BM (2005) The effect of biomimetic
apatite structure on osteoblast viability, proliferation, and gene expression. Bio-
42. Levi B, et al. (2010) Human adipose derived stromal cells heal critical size mouse
calvarial defects. PLoS ONE 5(6):e11177.
43. Levi B, et al. (2011) Studies in adipose-derived stromal cells: Migration and partici-
pation in repair of cranial injury after systemic injection. Plast Reconstr Surg 127(3):
| www.pnas.org/cgi/doi/10.1073/pnas.1218052109 Levi et al.