Self-renewal induced efficiently, safely, and effective
therapeutically with one regulatable gene in a human
somatic progenitor cell
Kwang S. Kima,b,1, Hong J. Leea,b,1, Han S. Jeongb,c, Jianxue Lid, Yang D. Tenge, Richard L. Sidmand,2, Evan Y. Snyderf,2,
and Seung U. Kima,b,2
aMedical Research Institute, Chung-Ang University College of Medicine, 440-746 Seoul, Korea;bDivision of Neurology, Department of Medicine, University
of British Columbia, Vancouver, BC, Canada V6T 2B5;cDepartment of Physiology, Chonnam National University Medical School, 500-757 Gwangju, Korea;
dDepartment of Neurology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215;eDepartment of Neurosurgery, Brigham
and Women’s Hospital, Harvard Medical School, Boston, MA 02115; andfSanford-Burnham Medical Research Institute, La Jolla, CA 92037
Contributed by Richard L. Sidman, January 3, 2011 (sent for review December 17, 2010)
In the field of induced potency and fate reprogramming, it remains
unclear what the best starting cell might be and to what extent
a cell need be transported back to a more primitive state for
translational purposes. Reprogramming a committed cell back to
pluripotence to then instruct it toward a particular specialized
cell type is demanding and may increase risks of neoplasia and
undesired cell types. Precursor/progenitor cells from the organ of
therapeutic concern typically lack only one critical attribute—the
capacity for sustained self-renewal. We speculated that this could
be induced in a regulatable manner such that cells proliferate only
in vitro and differentiate in vivo without the need for promoting
pluripotence or specifying lineage identity. As proof-of-concept,
we generated and tested the efficiency, safety, engraftability, and
therapeutic utility of “induced conditional self-renewing progenitor
(ICSP) cells” derived from the human central nervous system (CNS);
we conditionally induced self-renewal efficiently within neural pro-
genitors solely by introducing v-myc tightly regulated by a tetracy-
cline (Tet)-on gene expression system. Tet in the culture medium
sion of homogeneous, clonal, karyotypically normal human CNS pre-
cursors ex vivo; in vivo, where Tet was absent, myc was not
expressed, and self-renewal was entirely inactivated (as was tumor-
igenic potential). Cell proliferation ceased, and differentiation into
ensued upon transplantation into rats, both during development
and after adult injury—with functional improvement and without
neoplasia, overgrowth, deformation, emergence of non-neural cell
types, phenotypic or genomic instability, or need for immunosup-
pression. This strategy of inducing self-renewal might be applied
to progenitors from other organs and may prove to be a safe, ef-
fective, efficient, and practical method for optimizing insights
gained from the ability to reprogram cells.
programming” of differentiated somatic cells (e.g., skin
fibroblasts) to a state of pluripotence by forcing expression of
three to four genes (1–3), the question remains whether an end-
committed cell from an irrelevant tissue is the best initiator of
therapy in a particular organ. Reprogramming all of the way
back to pluripotence may involve more reverse steps than is
necessary; returning to or preserving potential fates more di-
rectly pertinent to a given organ system or germ layer (i.e.,
“oligopotence” or “multipotence”) may suffice for many thera-
peutic indications. Furthermore, challenges still remain with all
forms of pluripotent cells, whether they be human embryonic
stem cells (hESCs) or newer induced pluripotent stem (iPS)
cells: (i) efficiently generating a homogeneous cell population
while excluding cells of undesired lineages; and (ii) eliminating
tumorigenic potential. For example, treatment of a neurologic
disease might not require reprogramming back to the pluripotent
lthough intense interest has been generated by the “re-
state only to then attempt generation of pure neural lineages. It
might be more prudent and efficient to start with cells firmly
within the neural lineage, endow them with self-renewal to
permit extensive expansion ex vivo while deferring terminal dif-
ferentiation, and then reverse those actions upon engraftment to
Myc is likely to remain the only required gene if one does not
reprogram back to pluripotence. (Sox 2 is already expressed in
neural progenitors.) Although not entirely necessary (4, 5), myc
increases efficiency of reprogramming (4, 6, 7) and is the one
gene consistently highly expressed in all reprogrammed colonies
(7). The Myc family of transcription factors are well-established
regulators of G1 progression, driving cells into S phase by
mechanisms impacting cyclin-dependent kinase (Cdk) activity.
myc is its tumorigenic potential (v- and L-myc less so than c-myc),
offsetting its great utility in promoting self-renewal. However, if
might be the only gene required to be invoked in vitro to enhance
self-renewal of a neural progenitor, preserving its multipotence/
oligopotence without inducing pluripotence.
In this study, we explore whether a progenitor cell (rather than
a terminally committed cell) with only a single gene for the pur-
efficacy, and safety of the reprogramming process for specific
organ therapy. To test this concept, we conditionally induced self-
renewal of a human neural progenitor cell and applied it to
a prototypical neural injury (stroke). We envision that this ap-
proach may be applicable to progenitor cells from other organs.
Starting Material: Primary Culture of Human Neural Progenitor/
Precursor Cells. We first established primary cultures of cells dis-
sociated from telencephalic tissue of human fetal cadavers (11–14
wk gestational age). When grown in UBC1 serum-free medium
(DMEM with high glucose containing human insulin, human
transferrin, sodium selenite, hydrocortisone, triiodothyronine,
and other nutrients and antioxidants) with supplementary basic
fibroblast growthfactor (bFGF), the cultures grew as a monolayer
of cells with flat fibroblast morphology intermixed with smaller
Author contributions: K.S.K., H.J.L., R.L.S., E.Y.S., and S.U.K. designed research; K.S.K.,
H.J.L., and H.S.J. performed research; S.U.K. contributed new reagents/analytic tools;
K.S.K., H.J.L., J.L., Y.D.T., R.L.S., E.Y.S., and S.U.K. analyzed data; and K.S.K., H.J.L.,
R.L.S., E.Y.S., and S.U.K. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
1K.S.K. and H.J.L. contributed equally to this work.
2To whom correspondence may be addressed. E-mail: firstname.lastname@example.org,
email@example.com, or firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| March 22, 2011
| vol. 108
| no. 12www.pnas.org/cgi/doi/10.1073/pnas.1019743108
bipolar cells; the former were mostly glial fibrillary acidic protein-
positive (GFAP+), nestin−(consistent with an astrocytic phe-
notype), whereas the latter were either nestin+or β-tubulin III+
(consistent with an early neural progenitor/precursor phenotype).
The cultures were composed of ∼70% nestin+cells, ∼20%
GFAP+cells, and ∼9–10% β-tubulin III+cells; <0.5% of the cells
were O4+, galactocerebroside+, or 2′,3′-cyclic nucleotide 3′-
phosphodiesterase-positive (CNPase+), suggestive of an oligo-
dendrocytic phenotype. To assess cell division, cultures were in-
cubated in bromodeoxyuridine (BrdU) and processed for dual
immunoreactivity against antibodies to BrdU/nestin or BrdU/
β-tubulin III; >60% of cells on days 7–14 in vitro were dual
BrdU+/nestin+, whereas no dual BrdU+/β-tubulin III+cells were
found. (In the absence of tonic exposure to supraphysiological
concentrations of mitogens, progenitors in these primary cultures
did not sustain self-renewal.)
Transduction of Regulatable myc into Human Neural Progenitor/
Precursor Cells. Replication-incompetent amphotropic infection
witha retroviral vector, basedona Maloneymurineleukemiaviral
(MMLV) backbone, was carried out on the primary dissociated
some nestin+proliferative cells, i.e., the neural progenitor cell
population in the dish, would integrate the provirus. The culture
was first infected with a tetracycline (Tet)-response regulator vec-
v-myc, an isoform of myc (9, 10) (SI Materials and Methods).
Infectants were then selected for resistance to hygromycin
(200 μg/mL). During the selection with antibiotics (i.e., dual
resistance to both G418 and hygromycin), much cell death oc-
curred. Colonies of phase-dark bipolar or polygonal cells with
short processes began to appear (Fig. 1 A, A).
Isolation of Clonal Lines of Regulatable myc-Expressing Induced Self-
Renewing Human Neural Progenitor Cells. Cell lines (collectively
designated as “HB2 cells”) were isolated by limiting dilution and
their clonal identity confirmed by presence on Southern blot of
a single retroviral insertion site in all members of the population.
A representative clonal cell line, “HB2.G2,” is presented here in
detail but reflects the characteristics of all clones so generated
(Fig.1).Withdoxycycline (Dox+)in thegrowthmedium,HB2.G2
cells expressed nestin (Fig. 1 A, G), suggesting that they were
undifferentiated. Similarly, all Dox+HB2.G2 cells were immu-
nopositive for the v-myc gene product, whereas in the absence of
Dox, v-myc expression was completely inactivated (Fig. 1 A, C
and Western blot showed the v-myc gene product in Dox-con-
taining medium; in the absence of Dox, both message and protein
were undetectable (Fig. 1 B, A). Coincident with v-myc inac-
tivation, the cells ceased dividing, became phase bright, and
began extending long processes. In the presence of Dox in vitro,
70 ± 15.6% of HB2.G2 cells had incorporated BrdU (after a 12-h
incubation), but without Dox, only 6 ± 5% did so (Fig. 1 A, E
compared with F). Doubling time for HB2.G2 cells, investigated
by cell counts at 12, 24, 48, and 72 h, indicated mitosis every 24.2 h.
(As a control against attributing these effects to nonspecific retro-
viral integration, neural progenitors transduced with a vector en-
coding only the lacZ reporter gene did not evince these behaviors.
As a control against attributing these properties to integration of
a provirus within a particular “hot spot” in the cell’s genome, all
the “induced conditional self-renewing progenitor (ICSP) cell”
clones—each differing in integration site because retroviral in-
sertion is essentially random—possessed these same properties.)
Cytogenetic analysis of HB2.G2 cells showed a normal human
46XX karyotype (Fig. S1).
Neuronal Differentiation of HB2.G2 Conditionally Induced Self-
Renewing Human Neural Progenitor Cells. HB2.G2 cells were ex-
media containing 10 ng/mL of bFGF (Fig. 2). In both conditions,
whether proliferative or nonproliferative, HB2.G2 cells expressed
ABCG2 or nestin, undifferentiated state markers (Fig. 2A).
Musashi 1, an even more immature neural marker, however, was
expressed only under Dox+conditions and was absent under Dox–
conditions. Low level expression of NF-L [low molecular weight
renewing by conditional expression of a single “stemness” gene. (A) Dox-
regulated expression of myc (isoform v-myc) and cell proliferation. (A and B)
Phase contrast microscopy of cells grown in presence (+) or absence (–) of Dox
extend long processes. (C and D) Immunofluorescence staining for the v-myc
gene product. In Dox−(differentiation-inductive conditions), the cells do not
BrdU, i.e., are not mitotically active. [Nuclei are stained with DAPI (blue)].
(G and H) Immunostaining for nestin (green). In Dox+, virtually all of the cells
fewer cells express nestin, indicating that a majority of the cells are now dif-
ferentiating—primarily into neurons (see Fig. 2) and, to a lesser extent, glia.
[Nuclei arestained withDAPI(blue)]. (Scalebars,A–D,100μm;E–H,50μm.)(B)
conditions (Dox–). Expression of the "self-renewal" gene myc in HB2.G2 cell is
well regulated and does not adversely alter the karyotype (Fig. S1).
Clonal human neural progenitor cells induced to become self-
Kim et al. PNAS
| March 22, 2011
| vol. 108
| no. 12
(MW) neurofilament], a marker for early immature neurons, was
observed in both Dox+and Dox–media. Expression of β-tubulin
III, another marker for immature neurons, increased markedly in
Dox–media, whereas its expression was low in Dox+. NF-M
(medium MW), NF-H (high MW), and MAP2—markers formore
Dox–conditions (Fig. 2A). GFAP (a neural stem/progenitor cell
marker when it coexpresses nestin and an astrocyte marker when
expressed alone) was present under both conditions, but the
emergence of MBP (a maturing oligodendrocyte-specific marker)
was evident only in the absence of Dox (Fig. 2A).
Corroborating the RT-PCR studies in Fig. 2A, immunocyto-
chemical analysis showed a similar expression pattern of neural
cell-type markers by HB2.G2 cells (Fig. 2B), with 92.5 ± 1.5%
β-tubulin III+under Dox–conditions, and only 23.5 ± 2.5% in
Dox+(Fig. 2 B, A compared with B). In Dox+, no cells stained
for MAP2 or NF-H (Fig. 2 B, C–F); however, 7 d after removal
of Dox, 79.4 ± 2.8% of the cells were immunopositive for NF-H
and 81.7 ± 1.1% for MAP2. Expression of several neurotrophic
factors is documented in Fig. S2.
Immunocytochemically identifiable glial cells were less abun-
dant than neurons (Fig. 2 B, G–J). Despite a prominent RT-PCR
amplification product for GFAP, <5% of GFAP immunopositive
cells and <1% of O4 immunopositive cells (the latter suggestive
of oligodendrocytes) were detectable in the absence of Dox—
typical profiles for human neural stem cells (hNSCs) in the ab-
sence of specific glia-promoting agents (10). No glial cells of
either type were seen in Dox+media.
Thus, under these in vitro differentiation conditions, this “condi-
tional induced self-renewing human neural progenitor cell” line is
with the well-accepted notion that neuronal differentiation is a de-
fault path for neural progenitor cells, particularly in nonconfluent
cultures. Importantly, no cells of any non-neuroectodermal lineages
were evident under any conditions.
Electrophysiological Assessment of HB2.G2 Cells. Electrophysiologi-
cal properties of differentiated HB2.G2 cells were further stud-
ied by whole-cell voltage clamping after growth in proliferative
self-renewing (Dox+) and differentiation-inductive (Dox–) con-
ditions (Fig. 2C). Dox+HB2.G2 cells exhibited virtually no so-
dium currents (n = 9), but upon removal of Dox, substantial
numbers of sodium currents were observed (n = 11) and were
blocked by tetrodotoxin (1 μM). Concomitant with sodium cur-
rent induction, regenerative action potentials were observed in
HB2.G2 cells following growth in the absence of Dox.
Survival, Integration, and Differentiation of HB2.G2 Cells in Imma-
ture Normal Rat Brain. We next tested the responsiveness of the
HB2.G2 clone to normal developmental cues in vivo. The cells
were identified with a human mitochondria-specific antibody
(hMito). At 2 wk following transplantation into the lateral ven-
tricles of newborn rats, HB2.G2-derived cells had integrated into
the germinal subventricular zone (SVZ) and had migrated into
subcortical brain parenchyma (Fig. 3 A, A–C). hMito+cells were
found in the outer granular layer of the olfactory bulb, indicating
that some HB2.G2 cells had migrated along the rostral migratory
stream (RMS) to their appropriate developmental target, where
a small number had differentiated appropriately into granule cell
neurons (Fig. 3 A, D). Additional HB2.G2 cells had migrated
appropriately to the hippocampus (Fig. 3 B, E and F). The
neuronal phenotype of donor-derived cells was supported by
double immunostaining for hMito and neuron-specific markers
(Fig. 3B). hMito+HB2.G2 cells not only expressed the mature
neuronal markers MAP2 and NF-H (Fig. 3 B, A–F) but also the
neurotransmitter-related markers glutamate and GAD (the
synthetic enzyme for GABA), suggestive of at least partial dif-
ferentiation into glutamatergic and GABAergic interneurons
(Fig. 3 B, G–L). When implanted into the developing substantia
nigra, HB2.G2 cells expressed tyrosine hydroxylase (TH) 4 wk
after transplantation (n = 3 rats) (Fig. 3 B, M–O). Only neural
lineage markers were expressed; no non-neuroectodermal line-
Neuronal differentiation appears somewhat more prominent
than glial differentiation, consistent with the typical default
fate of most normal neural progenitors upon exiting the cell
cycle. (A) Analysis by RT-PCR of ICSP cells when self-renewal is
inactivated in the absence of Dox. Transcripts of immature
neural markers (ABCG2, nestin, and Musashi1) as well as of
cell type-specific markers—for neurons (β-tubulin III, NF-L,
NF-M, and NF-H), astrocytes (GFAP), and oligodendrocytes
(MBP)—were determined in cells grown in Dox+or Dox−
media for 3 d. GAPDH was an internal control. (B) Immuno-
cytochemical analysis of ICSP cells when self-renewal is inac-
tivated in the absence of Dox. Immunostaining is shown for
cardinal neuronal and glial markers (red) in cells grown in
Dox+or Dox–media for 7 d: β-tubulin III (A and B), MAP2 (C
and D), and NF-H (E and F) for neurons; GFAP (G and H) for
astrocytes; and O4 (I and J) for oligodendrocytes. Cell nuclei
were stained with DAPI (blue). (C) Whole cell patch-clamp
recordings of HB2.G2 ICSP cells in Dox+and Dox−media. (A) In
Dox+(self-renewing conditions), no sodium currents were
present, as recorded with depolarization pulses to 0 mV from
a holding potential of 80 mV. (B, a) Representative Na+cur-
rents from cells in Dox−differentiation-inductive conditions.
(b) Tetrodotoxin (TTX) at 1 μM blocked inward currents. (c)
Recovery of Na+currents after wash out of TTX.
Differentiation profile of HB2.G2 neural ICSP cells.
| www.pnas.org/cgi/doi/10.1073/pnas.1019743108Kim et al.
ages or tumors were generated for at least 20 wk, the longest
interval examined (Fig. S3).
Transplantation of HB2.G2 Cells Improves Function in Rats with
Hemorrhagic Stroke. A rodent model of intracerebral hemor-
rhagic (ICH) stroke provided proof of principle that HB2.G2
cells, in which self-renewal was stringently inactivated in vivo,
can be safely and effectively transplanted into the brains of adult
animals with neurological dysfunction and help mediate func-
tional recovery without tumor formation, deformation, or the
emergence of inappropriate cell types. The ICH rats receiving
HB2.G2 cells showed improved behavioral performance on the
rotarod compared with the PBS sham-transplanted control
group (Fig. 4 A, A) and marked improvement in the limb
placement test (Fig. 4 A, B), effects that persisted for the 2-mo
period until histological analysis of the brains. The mechanisms
underlying improvement are elaborated in Discussion.
Transplanted HB2.G2 Cells Migrate to ICH Lesions and Differentiate
into Neurons and Glia. Immunostaining with the hMito antibody
demonstrated preferential migration of grafted HB2.G2 cells
toward the perihematomal border by 1–2 wk posttransplantation.
Double labeling with hMito/NF-H antibodies revealed a large
number of donor-derived cells differentiating toward a neuronal
phenotype with extended processes at the lesion site (Fig. 4 B,
A–D), and progressively by 4 and 8 wk posttransplantation,
hMito+/NF-H+cells in the HB2.G2 transplanted group dis-
played a more mature neuronal morphology with long processes
(Figs. 4 B, E–L). Also, many grafted cells at the perihematomal
border area were hMito+/GFAP+, indicating that some cells of
the HB2.G2 clone had differentiated toward an astrocytic line-
age (Fig. 4 B, D–F). Finally, some hMito+/CNPase+cells (sug-
gestive of an oligodendrocytic lineage) were present within the
lesion sites (Fig. 4 C, A–C).
Importantly, at 4 and 8 wk posttransplantation, no v-myc+
cells, including dual hMito+/v-myc+cells, were detected by im-
munohistochemistry (Fig. 4 C, G–I), indicating that v-myc ex-
pression was inactivated in vivo in the Dox–environment. No
tumors, deformations, or donor-derived nonneural cells were
seen at 2, 4, 8, 12, and 20 wk (Fig. S3).
Survival of Engrafted HB2.G2 Cells in ICH Rat Brain. The total num-
ber of hMito+HB2.G2 cells in brain sections from ICH animals
was determined by unbiased stereological analysis at 2 and 8 wk
posttransplantation (Fig. 4 C, J). At 2 wk, 83,100 ± 1,140 cells
(41.6 ± 1.7% of the initial implant of 200,000 cells) and at 8 wk,
40,340 ± 2,850 cells (20.2 ± 2.5% of the initial implant) had
survived. (Because the HB2.G2 cells grow as an adherent
monolayer, it is necessary to remove them from the culture dish
by trypsin digestion for transplantation; this procedure evidently
produces some cell death, so that although 200,000 HB2.G2 cells
were prepared for transplantation, the actual number of live cells
implanted, as determined by trypan blue exclusion immediately
before transplantation, was 20% lower, i.e., 160,000–170,000 cells.
Taking this finding into account, the survival of grafted HB2.G2
cells in the ICH animals would be 51.9% at 2 wk posttrans-
plantation and 25.2% at 8 wk.) HB2.G2 cells survived to at least
20 wk posttransplantation (Fig. S4). It bears noting that this sur-
vival, which was sufficient to impart a behavioral improvement in
the ICH rats (mechanism discussed below), was performed with-
out immunosuppression. Although addition of immunosup-
pressive drugs might have increased grafted cell survival, the
achievement of therapeutic efficacy with safety while avoiding
such pharmacological complications was appealing. There was
no evidence of an inflammatory immunorejection reaction by
the host to the presence of human cells.
In the rapidly evolving field of cell reprogramming, choice of the
best starting cell in any particular experimental or therapeutic
back to a more primitive state remain unsettled. iPS cells offer
the advantage of a genetic match to the patient from whom the
skin cells were obtained, minimizing prospects of immunological
rejection. However, for diseases of a particular organ system
where lineage-restrictedprogenitorsare available,reprogramming
of an end-committed cell from an irrelevant tissue (e.g., a dermal
fibroblast) all of the way back to pluripotence and then efficiently,
specifically, and uniformly to an alternate organ identity may un-
necessarily add time, effort, genes, and manipulations, increasing
the risk of neoplasia, undesired cell types, and other sources of
of therapeutic interest may be preferable. We tested this hypoth-
esis in the CNS where progenitor/precursor cells are well known.
capacity for self-renewal. They matched the predilection of other
neural progenitor cell preparations to become neuronal (10) and
to be effective in clinical disease models (10–16). Although, in
these experiments, the neural progenitors were obtained from the
brain, they could likely be obtained from such more accessible
sources as olfactory neuroepithelium or skin. Also, given the ab-
sence of immunorejection, one ICSP line could serve many
To summarize, we have established stable clonal ICSP cells—
in other words, lines in which the starting cell is a non–self-
renewing progenitor from the target organ of ultimate thera-
peutic interest (rather than a terminally differentiated unrelated
cell) in which expression of a gene stimulating self-renewal is
under the conditional control of a tightly regulatable promoter.
of ICSP cells following transplantation. In vivo, Dox is not present and hence,
self-renewal is not active. (A) Survival and migration of HB2.G2 ICSP cells
following intraventricular transplantation into newborn rat brain. The
engrafted cells were detected by immunostaining with antibodies specific
either for human mitochondria (hMito) or β-galactosidase (β-gal). A sche-
matic sagittal section view of newborn rat forebrain 7 d after trans-
plantation of HB2.G2 ICSP cells (A), showing HB2.G2-derived cells in the
germinal SVZ adjacent to the rostral part of the lateral ventricle (B), and in
the RMS (D). B2and D2are enlargements of the boxed areas in B1and D1. An
HB2.G2 cell with a typical mature granule neuron morphology (hMito+) is
found in the olfactory bulb 2 wk posttransplantation (C). A schematic cor-
onal view of a rat brain (E) shows the distribution of hMito+HB2.G2-derived
cells at 2 wk posttransplantation. The hMito+HB2.G2 cells transplanted in-
traventricularly earlier are seen to have migrated to the hippocampus
(within E, note red cells in box F). F1is an enlargement of the boxed area F in
E. F2and F3are magnified views of the boxed areas in F1. (Scale bars in B1, D1,
and F1, 200 μm; B2and D2, 100 μm; and C, F2, and F3, 50 μm.) (B) Immuno-
histological analysis of differentiation fate in vivo of HB.G2 ICSP cells in rat
brain following neonatal intraventricular transplantation 2 wk earlier. Dif-
ferentiation of the cells toward a neuronal phenotype is suggested by dual
staining for hMito+(green) and the following neuronal markers: MAP2+(red)
(A–C); NF-H+(red) (D–F); glutamate (red) (consistent with a glutamatergic
interneuronal phenotype) (G–I); and GAD+(red) (consistent with a GABAer-
gic interneuronal phenotype) (J–L). Four weeks following transplantation of
ICSP cells into adult substantia nigra, occasional TH+(red) donor-derived cells
were, nevertheless, surprisingly detected (M–O). (Scale bar, 50 μm.)
Nontumorigenic engraftment, migration, and neural differentiation
Kim et al. PNAS
| March 22, 2011
| vol. 108
| no. 12
The gene is myc, well established as pivotal in endowing self-re-
newal competence to somatic cells in transition to a pluripotent
state. When myc is omitted because of its potential oncogenic risk,
efficiency of reprogramming is significantly impaired. In the
present study, we established conditional self-renewing clonal
neural progenitor cell lines (representative clone HB2.G2 was
presented here) in which expression of the isoform v-myc is reg-
established that v-myc was the most effective and safest isoform of
the presence of the Tet derivative Dox, resulting in enhanced cell
proliferation, but not malignant growth. (The cells still respected
normal growth controls, i.e., were contact inhibited, did not grow
readily with a doubling time of 24.2 h in the presence of Dox, thus
providing an accessible, renewable, homogenous, reliable, abun-
dant, safe, effective, and practical source of human CNS pre-
cursors. In the absence of Dox, myc was no longer expressed, cell
proliferation ceased, and differentiation into CNS cell types rap-
rats in vivo both under normal developmental conditions (where
they engrafted, migrated, and differentiated appropriately) and
following stroke in the adult, a model providing proof of principle
that these ICSP cells, with proliferation stringently regulated, can
be safely and effectively transplanted into injured brains without
tumor formation or deformation. In short, these ICSP cells, which
now behave transiently like multipotent somatic stem cells, em-
body the benefits of rapid scale-up offered by hESCs or iPS cells,
but without their risks (teratoma formation or emergence of
a premalignant state; emergence of nonneural cell types in-
appropriate to the brain or other types of uncontrolled heteroge-
neity; and phenotypic or genomic instability). Furthermore,
a degree of plasticity (often dangerous in pluripotent cells and
unattainable in end-differentiated somatic cells) is preserved,
which facilitates engraftment, integration, and multi-neural cell-
type reconstitution. The potential clinical advantages of such
The induction of conditional self-renewal in progenitor cells
with a single gene is far more efficient than has been achieved
with iPS cells, although improvements in the iPSC methodology
are being sought intensively (1–3). ICSP cells can be generated
with an efficiency of 33% or better, with colonies forming within
1 wk and sufficient cells for multiple transplants within 2 wk.
Role and Control of myc. Theroleofmycininducingpluripotencein
end-differentiated cells, and its conditional deletion in the neural
stem cells of transgenic mice, causing a failure of normal brain
(17), have led to recognition of myc as a “stemness” gene, both
promoting self-renewal and holding terminal differentiation in
abeyance (17–21). Its precise mechanism of action has been at-
tributed to one or a combination of the following: (i) Inducing
a cell-cycle gene expression program necessary specifically for self-
renewal and inhibition of differentiation by repression of cdk
21); it also may play a nontranscriptional role in initiation of DNA
replication (17–21). (ii) Maintaining key components of tran-
scriptionally active chromatin required for self-renewal, pluri-
potency, and suspension of differentiation (including histone H3
and H4 acetylation as well as H3 methylation at lysine 4) (19–21).
Furthermore, previous work on v-myc–enhanced transplanted
cells has documented that extra copies of myc are spontaneously
and constitutively inhibited by normal developmental mecha-
nisms of the transplant recipient, just as endogenous c-myc is
a regulatable promoter adds an extra level of safety and control.
Controlled expressionofmyc in ICSPcells witha Tet-onsystem
obviates the risk of tumorigenesis in vivo following trans-
proliferation in vitro (for scale-up), however (ii) cease dividing
once transplanted into CNS, while retaining the ability to differ-
entiate into neurons and glia in response to local microenviron-
mental cues. By comparison, previous studies (22) have indicated
that a Tet-off–v-myc gene expression system was “leaky,” so that
v-myc mRNA and protein were found in newly generated cell
lines. Also, cell lines generated via the Tet-off system grew slowly
with a doubling time of >80 h (22). More concerning, Tet-off cells
might continue to proliferate after transplantation into the brain
where continuous Dox exposure is not feasible. In contrast, the
HB2.G2 ICSP cells bear a Tet-on system, with Dox-promoting
expression of v-myc solely in vitro; the absence of Dox in vivo
renders v-myc mRNA and protein undetectable (Fig. 2).
Neurobiological Fidelity of Human Neural ICSP Cells. In vitro, HB2.
G2 cells under Dox−conditions show a predilection for neuro-
nal, more than glial, differentiation, consistent with the typical
tion of ICSP cells into an adult rat model of ICH stroke. (A) Functional tests of
adult rats subjected to stroke followed by intracerebral transplantation of
HB2.G2 ICSP cells. (A) In the rotarod test, ICH-lesioned rats receiving HB2.G2
cells showed improved performance from 1 to 9 wk compared with control
animals receiving PBS. *P < 0.05. (B) In the limb placement test, ICH-lesioned
rats receiving HB2.G2 cells showed improved performance from 3 to 9 wk
posttransplantation. *P < 0.05. (B) Immunohistochemistry suggests the
presence of donor-derived neurons. Three each of ICH-lesioned adult rats
were killed at 2, 4, and 8 wk posttransplantation (Left, Center, and Right
columns, respectively). Coronal sections of their brains were double labeled
with anti-hMito antibody (green) (A, E, and I) and anti–NF-H antibody (red)
(B, F, and J); merged images are shown in C, G, and K; the boxed area in each
merged image is shown at higher magnification in D, H, and L, respectively.
Engrafted cells that differentiated toward a neuronal phenotype show yel-
low fluorescence. (C) Double immunofluorescence staining of engrafted HB2.
G2-derived cells in ICH-lesioned adult rat brain 4 wk posttransplantation
suggests the presence of donor-derived glia and down-regulation of myc.
(A–F) In brain sections from an ICH-lesioned adult rat, HB2.G2 cells double
labeled with anti-hMito (green) (A and D) and anti-CNPase (red) (B) or anti-
GFAP (red) (E) had differentiated toward oligodendrocytic (C) or astrocytic
(F) lineages, respectively. (G–I) Negative v-myc immunoreactivity (red) (H) in
hMito+engrafted (G) HB2.G2-derived cells in the absence of Dox—the typical
condition in vivo in the brain. Accordingly, the neural progenitors do not
undergo mitosis, thereby precluding tumor formation or brain deformation.
(J) Survival of engrafted HB2.G2-derived cells 2 and 8 wk posttransplantation
in ICH-lesioned adult rat brains. The percentage of surviving cells per mm2of
engrafted brain was 41.6% at 2 wk and 20.2% at 8 wk; functional improve-
ment persisted until at least 20 wk (study termination).
Behavioral improvement following safe engraftment and integra-
| www.pnas.org/cgi/doi/10.1073/pnas.1019743108 Kim et al.
default behavior of most normal neural progenitors upon exiting Download full-text
the cell cycle. Their properties include expression of neuronal
cytoskeletal proteins (NF-L, M, and H, β-tubulin III and MAP2),
ion channels (for Na+, K+, and Ca2+), neurotransmitter syn-
thesizing enzymes (ChAT, GAD, and TH) and neurotransmitters
(GABA and glutamate). In addition, such differentiated HB2.G2
action potentials, critical hallmarks of functional neurons.
In vivo, HB2.G2 cells migrated along at least two different
pathways after implantation into the cerebral ventricles of neo-
natal rats: (i) along the RMS to olfactory bulb and (ii) into hip-
pocampus, as reported for other neural progenitors (10–12, 14),
hence suggesting that fundamental progenitor properties are
unaffected by the transgene and its Tet regulatory system. Al-
though the precise mechanism is unclear by which HB2.G2
hNSCs migrate extensively yet selectively to pathological lesions,
signals produced at CNS injury sites, such as stromal cell-derived
factor 1α (SDF1α) (12), vascular endothelial cell growth factor
(VEGF) (11, 13), or stem cell factor (SCF) (11, 13). Other cyto-
kines with important functions in CNS development, including
bFGF, EGF, and TGFα, also contribute to ischemia-induced pro-
liferation and migration of neural progenitor cells. HB2.G2 cells
express CXCR4 (the receptor for SDF1α), VEGFR1, VEGFR2,
and c-kit, again suggesting that conditionally augmenting self-re-
It bears emphasis that we do not claim functional improvement
15) the HB2.G2 cells providing trophic and neuroprotective sup-
(beyond the scope of this report), ICSP cells invoke it with a safety
and efficacy equal to that of other progenitor cell populations.
an immunosuppressive agent, consistent with previously reported
immunotolerance for NSCs in transplantation studies (9, 14).
Whereas immunosuppression might increase survival of donor-
derived cells, we, by intent, wished to determine how well the cells
would persist, differentiate, and impact function without the in-
evitable side effects, health risks, andconfounders imposed by the
human patients. Our results suggest that a sufficient impact may
be attainable under this more desirable treatment regime.
Therefore, induced conditional self-renewing progenitors will
likely be valuable for future studies in developmental biology,
drug discovery, and derivation of stem cell-based therapies.
Human neural ICSP cells—self-renewing in vitro and differentiat-
ing in vivo within a given organ’s range of lineages—offer an at-
tractive alternative to pluripotent cells (which require additional
isolated from a given organ (which are heterogeneous, of dubious
clonality, and difficult to transplant, engineer, synchronize, scale-
up, and render uniform and invariant from experiment to experi-
screening of small molecules (23). Usually, the number of neural
progenitors obtained from a human brain is small, their expansion
slow, and the risk of cell line senescence sufficiently prevalent that
routine clinical transplantation is difficult. Therefore, for the CNS,
ICSP cells—rapidly expandable and standardizable in vitro as
monolayers—provide an efficient, expeditiously generated source
of abundant, uniform, well-characterized, safe, therapeutically
effective neural cells for transplantation—possibly preferable to
hESCs and iPS cells in this regard (24).
In conclusion, we believe this strategy of inducing conditional
self-renewal within lineage-bounded progenitor cells (for in vitro
expansion after which self-renewal is permanently inactivated in
vivo) may serve broadly as a safe, effective, and practical method
for optimizing the revolutionary insights gained of late from the
the advantage of requiring fewer genes and fewer intervening steps
in the derivation of therapeutically useful cells for a given organ.
Materials and Methods
A replication-incompetent retroviral vector was used to transduce v-myc
transcribed from the RevTet-on system into primary human fetal telence-
phalic ventricular zone cells in culture, from which several clones were iso-
lated. Clone HB2.G2 was then transfected with an adenovirus encoding lacZ.
Please see SI Materials and Methods for details of clone genesis and sub-
sequent cytogenetic, RT-PCR, Western blot, immunohistochemical, electro-
physiological, and transplantation procedures. Most reagents and methods
were previously described (16).
ACKNOWLEDGMENTS. We thank I. Singec and D. Wakeman for their in-
sights. This work was supported by grants from the National RD Program for
Cancer Control, Korean Ministry of Health and Welfare (08203103-22650),
and the Canadian Myelin Research Initiative (to S.U.K.); from NS059770,
Sanford Children’s Research Center, and California Institute for Regenerative
Medicine; and from the A-T Children’s Project and the Nancy Lurie Marks
Family Foundation (to R.L.S.).
1. Yamanaka S, Blau HM (2010) Nuclear reprogramming to a pluripotent state by three
approaches. Nature 465:704–712.
2. Hanna JH, Saha K, Jaenisch R (2010) Pluripotency and cellular reprogramming: Facts,
hypotheses, unresolved issues. Cell 143:508–525.
3. Zaehres H, Kim JB, Schöler HR (2010) Induced pluripotent stem cells. Methods
4. Nakagawa M, et al. (2008) Generation of induced pluripotent stem cells without Myc
from mouse and human fibroblasts. Nat Biotechnol 26:101–106.
5. Wernig M, Meissner A, Cassady JP, Jaenisch R (2008) c-Myc is dispensable for direct
reprogramming of mouse fibroblasts. Cell Stem Cell 2:10–12.
6. Meissner A, Wernig M, Jaenisch R (2007) Direct reprogramming of genetically
unmodified fibroblasts into pluripotent stem cells. Nat Biotechnol 25:1177–1181.
7. Park IH, et al. (2008) Reprogramming of human somatic cells to pluripotency with
defined factors. Nature 451:141–146.
8. Singh AM, Dalton S (2009) The cell cycle and Myc intersect with mechanisms that
regulate pluripotency and reprogramming. Cell Stem Cell 5:141–149.
9. Snyder EY, et al. (1992) Multipotent neural cell lines can engraft and participate in
development of mouse cerebellum. Cell 68:33–51.
10. Flax JD, et al. (1998) Engraftable human neural stem cells respond to developmental
cues, replace neurons, and express foreign genes. Nat Biotechnol 16:1033–1039.
11. Jeong S-W, et al. (2003) Human neural stem cell transplantation promotes functional
recovery in rats with experimental intracerebral hemorrhage. Stroke 34:2258–2263.
12. Imitola J, et al. (2004) Inflammation’s other face: Directed migration of human and
mouse neural stem cells to site of CNS injury by the SDF1a/CXCR4-dependent path-
way. Proc Natl Acad Sci USA 101:18117–18122.
13. Kim SK, et al. (2006) Human neural stem cells target experimental intracranial
medulloblastoma and deliver a therapeutic gene leading to tumor regression. Clin
Cancer Res 12:5550–5556.
14. Lee J-P, et al. (2007) Stem cells act through multiple mechanisms to benefit mice with
neurodegenerative metabolic disease. Nat Med 13:439–447.
15. Redmond DE, Jr., et al. (2007) Behavioral improvement in a primate Parkinson’s
model is associated with multiple homeostatic effects of human neural stem cells.
Proc Natl Acad Sci USA 104:12175–12180.
16. Lee HJ, Lim IJ, Lee MC, Kim SU (2010) Human neural stem cells genetically modified
to overexpress brain-derived neurotrophic factor promote functional recovery and
neuroprotection in a mouse stroke model. J Neurosci Res 88:3282–3294.
17. Knoepfler PS, Cheng PF, Eisenman RN (2002) N-myc is essential during neurogenesis
for the rapid expansion of progenitor cell populations and the inhibition of neuronal
differentiation. Genes Dev 16:2699–2712.
18. Cartwright P, et al. (2005) LIF/STAT3 controls ES cell self-renewal and pluripotency by
a Myc-dependent mechanism. Development 132:885–896.
19. Knoepfler PS (2008) Why myc? An unexpected ingredient in the stem cell cocktail. Cell
Stem Cell 2:18–21.
21. Knoepfler PS, et al. (2006) Myc influences global chromatin structure. EMBO J 25:
22. Oberst C, Hartl M, Weiskirchen R, Bister K (1999) Conditional cell transformation by
doxycycline-controlled expression of the MC29 v-myc allele. Virology 253:193–207.
23. Singec I, et al. (2006) Defining the actual sensitivity and specificity of the neurosphere
assay in stem cell biology. Nat Methods 3:801–806.
24. Roy NS, et al. (2006) Functional engraftment of human ES cell-derived dopaminergic
neurons enriched by coculture with telomerase-immortalized midbrain astrocytes.
Nat Med 12:1259–1268.
Kim et al. PNAS
| March 22, 2011
| vol. 108
| no. 12