Effects of Long-Term Culture
on Human Embryonic Stem Cell Aging
Xiaoyan Xie,1–4Asimina Hiona,1,2Andrew Stephen Lee,1,2Feng Cao,1,2Mei Huang,1Zongjin Li,1,2
Athena Cherry,5Xuetao Pei,4and Joseph C. Wu1–3
In recent years, human embryonic stem (hES) cells have become a promising cell source for regenerative
medicine. Although hES cells have the ability for unlimited self-renewal, potential adverse effects of long-term
cell culture upon hES cells must be investigated before therapeutic applications of hES cells can be realized. Here
we investigated changes in molecular profiles associated with young (<60 passages) and old (>120 passages)
cells of the H9 hES cell line as well as young (<85 passages) and old (>120 passages) cells of the PKU1 hES cell
line. Our results show that morphology, stem cell markers, and telomerase activity do not differ significantly
between young and old passage cells. Cells from both age groups were also shown to differentiate into deriv-
atives of all 3 germ layers upon spontaneous differentiation in vitro. Interestingly, mitochondrial dysfunction
was found to occur with prolonged culture. Old passage cells of both the H9 and PKU1 lines were characterized
by higher mitochondrial membrane potential, larger mitochondrial morphology, and higher reactive oxygen
species content than their younger counterparts. Teratomas derived from higher passage cells were also found to
have an uneven preference for differentiation compared with tumors derived from younger cells. These findings
suggest that prolonged culture of hES cells may negatively impact mitochondrial function and possibly affect
the capacity for unlimited replication . As a result, begin-
ning with their isolation in 1998 by Dr. James Thomson, these
source in cell replacement therapy. Numerous articles have
since demonstrated the potential therapeutic use of hES-
derived cells in the treatment of diseases affecting the heart
[2,3], brain[4,5], pancreas,liver,andbone marrow[8,9].
Current models of cell replacement therapy used in clinical
trials can require billions of cells to achieve optimal effect in
lines are limited, it is likely that repeated and prolonged
passaging of hES cells will be necessary for clinical applica-
tions of hES cells to be realized. Therefore, it is critical to
determine whether long-term in vitro cell culture can ad-
versely affect their capacity to participate effectively in cell
Senescence is a process that affects all somatic cells of
human body and has traditionally been characterized by
uman embryonic stem (hES) cells can differentiate into
every somatic cell type of the human body and possess
telomere shortening, accumulation of nuclear mutation,
epigenetic silencing, and mitochondrial dysfunction, the
overall effect of which produces the loss of function
[12,13]. Recently, adult stem cell senescence has also come
under scrutiny . hES cells are generally considered to
be resistant to replicative senescence. A number of studies
have demonstrated that ES cells not only continue to
replicate, but also maintain constant telomere length and
undergo lower rates of genomic mutation than their so-
matic counterparts even after prolonged in vitro replica-
tion extending into 1 year or longer [15–17]. Stem cells
grown in culture for such periods have also been shown to
retain normal karyotypes [17–19] and epigenetic stability
[20–22], but several recent articles have disputed this claim
In our experience, very late passage hES cells have been
observed to have a reduced ability to differentiate into de-
rivatives of all 3 germ layers, which may affect their thera-
peutic potential. To document this reduction in pluripotency
and determine whether these changes are associated with
replicative senescence, we investigated the proliferation and
differentiation of young and old passage hES cells, and
1Department of Medicine,
5Department of Pathology, Stanford University School of Medicine, Stanford, California.
4Stem Cells and Regenerative Medicine Lab, Beijing Institute of Transfusion Medicine, Beijing, China.
2Molecular Imaging Program at Stanford,
3Institute for Stem Cell Biology and Regenerative Medicine,
STEM CELLS AND DEVELOPMENT
Volume 20, Number 1, 2011
ª Mary Ann Liebert, Inc.
intracellular indices of aging such as mitochondrial function,
telomerase activity, and chromosomal stability.
Materials and Methods
Culture of hES cells
H9 hES cells (WiCell) and PKU1 hES cells (non-federal-
approved hES cells, a gift from Peking University)  were
cultured on a feeder layer of irradiated mouse embryonic
fibroblasts using hES cell culture medium consisting of 80%
Dulbecco’s modified Eagle’s medium (DMEM)=F-12 (In-
vitrogen), 20% knock-out serum replacement (Invitrogen),
1mM L-glutamine, 1% nonessential amino acids, 0.1mM
b-mercaptoethanol, and 8ng=mL basic fibroblast growth fac-
tor (Invitrogen). Cells were disassociated with Collagenase
IV (Invitrogen) every 4–6 days. Before analysis, cells were
moved to a Matrigel (hES cell-qualified Matrix; BD Bios-
ciences)-coated plate and cultured for 2 passages with
mTeSR feeder-free medium (StemCell Technology). H9 cells
having undergone <60 passages or >120 passages were
defined as young or old passage cells, respectively. PKU1
cells having undergone <85 passages or >120 passage were
defined as young or old passage cells, respectively.
hES cell colonies plated on chamber slides (Lab-Tek,
Nunc, Thermo Fisher Scientific) were fixed in 4% parafor-
maldehyde at room temperature for 30min. After washing
with phosphate-buffered saline (PBS), 5% goat serum was
added to the cells at room temperature for 1h. Cells were
subsequently incubated with primary antibodies at 48C
overnight. Antibodies used for embryonic stem cell marker
(SSEA-4) and Oct-4 (Santa Cruz). For Oct-4 staining, cells
were permeabilized by 0.1% Triton X-100 for 20min at room
temperature before antibody incubation. Primary signals
were detected using tetramethyl rhodamine ISO-thiocyanate
(TRITC)-conjugated goat secondary antibodies (Santa Cruz)
at room temperature for 1h in the dark. Finally, each well
was washed with PBS, nuclei were highlighted with Hoechst
33342 (Molecular Probe=Invitrogen), and immunofluores-
cence was detected by fluorescent microscopy.
hES cell proliferation
hES cells were plated on Matrigel-coated 96-well plates.
The CyQuant cell proliferation assay (Molecular Probes) was
conducted using a microplate spectrofluorometer (Gemini
EM) at 24-, 48-, and 72-h time points. Eight samples were
assayed and averaged.
Spontaneous differentiation of hES cells
hES cells were detached enzymatically and washed as
described above. The cells were resuspended in an embryoid
body (EB) medium containing DMEM supplemented with
20% FBS (Hyclone), then plated on 100mm ultra-low at-
tachment tissue culture dishes (Corning). The medium was
changed every 2–3 days. EBs at day 8 were transferred to
gelatin-coated dishes for adhesive culture. At day 14, EBs
were removed, pelleted, and frozen at ?808C for further
To track teratoma formation in vivo, hES cells of the H9
cell line were stably transduced with a self-inactivating
lentiviral vector carrying a human ubiquitin promoter
driving firefly luciferase and enhanced green fluorescence
protein (Fluc-eGFP) as previously described . After
selection for stable populations, 5?105low (passage 35–45)
and 5?105high (passage 120–130) Fluc-eGFP hES cells were
suspended in 25mL PBS, mixed with an equal volume of
Matrigel for injection into the subcutaneous regions of the
backs of the animals (n¼4, 2 spots for each group per
mouse). Eight weeks after transplantation, teratomas
were harvested and weighed. Cell differentiation was
assayed by histological analysis and reverse transcriptase-
polymerase chain reaction (RT-PCR). For histology, tera-
tomas were fixed with 4% paraformaldehyde, and then
embedded with paraffin, sectioned, and stained with he-
matoxylin and eosin. Light microscopy was used to ob-
serve the cells.
Measurement of teratoma growth via
Cell signal was measured from day 2 after transplantation
on a weekly basis for 8 weeks using a Xenogen IVIS 200
system (www.caliperls.com) as previously described .
After intraperitoneal injections of reporter probe D-Luciferin
(375mg=kg body weight), animals were imaged for a dura-
tion of 1s to 1min. Imaging signals were quantified in units
of maximum photons per second per square centimeter per
Reverse transcriptase-polymerase chain reaction
Tissue samples were homogenized in Trizol (Invitrogen).
Total RNA was isolated from cells using the RNeasy kit from
Qiagen according to the manufacturer’s instructions. cDNAs
were obtained using 1mg RNA with an iScript cDNA syn-
thesis kit (Bio-Rad). PCRs were carried out with 2mL cDNA
template. The specific primers and reaction conditions are
listed in Table 1.
Measurement of intracellular reactive
DCFH-DA (Invitrogen) was used for reactive oxygen
species (ROS) detection. When oxidized by ROS intracel-
lularly, the nonfluorescent compound will become fluo-
rescent. Cells from replicate cultures were dissociated with
0.5mM ethylenediaminetetraacetic acid (EDTA) to make a
single cell suspension, and resuspended in DMEM=F12
medium containing 10mM DCFH-DA to a final density at
106cell=mL. Cells were incubated at 378C for 20min, wa-
shed once, resuspended in PBS, and kept on ice for an
immediate detection by FACSCalibur (Becton, Dickinson
Biosciences). Average measurements from 4 replicates
were quantified as mean fluorescence intensity (MFI)=105
128XIE ET AL.
Assessment of mitochondrial membrane potential
Changes in mitochondrial membrane potential were esti-
mated using the cationic fluorescent dye (JC-1; Molecular
Probes) according to the manufacturer’s instructions. Briefly,
5?105cells were dissociated with 0.5mM EDTA and re-
suspended in 1mL fresh complete medium as a single-cell
suspension. The cell suspension was incubated with JC-1
(2.5mM) for 30min at 378C in the dark, followed by washing
with PBS. Cells were analyzed using a flow cytometer
equipped with a 488nm argon laser (BD FACSCalibur). JC-1
is a dual-emission potential-sensitive probe that localizes to
different sides of the mitochondrial membrane after cellular
uptake. The ratio of red to green fluorescence from JC-1 can
be quantified using flow cytometry and used as a measure of
mitochondrial membrane potential. Cells treated with JC-1
and carbonyl cyanide 3-chlorophenylhydrazone, a potent
uncoupler of oxidative phosphorylation, served as controls
of dissipation of mitochondrial membrane potential.
Measurement of oxygen consumption
Cells from replicate cultures were dissociated with 0.5mM
EDTA to make a single-cell suspension, and resuspended in
mTeSR medium for immediate polarographic measurement
of oxygen consumption using a Clark-type oxygen electrode
(Hansatech) at 378C. mTeSR was used as the background
oxygen value. This number was subtracted from the final
oxygen consumption values obtained. Cells were maintained
during the measurements at 378C in a temperature-jacketed
chamber, and oxygen consumption was monitored for
10min. Measurements with potassium cyanide were also
performed as controls to ensure that the oxygen consump-
tion observed was related to mitochondrial oxygen con-
sumption. Average measurements from 4 replicates were
quanitified as nM oxygen consumed=min=106cells.
Fluorescent staining of mitochondria
Mitochondria were stained with MitoTracker Green FM
(Invitrogen), which preferentially accumulates in the mito-
chondrial matrix irrespective of changes in membrane po-
tential. Cells were plated into chamber slides 1 day before
staining. Adherent cells were exposed for 45min to a 200nM
MitoTracker Green FM solution at 378C together with
Hoechst 33342 nuclei staining. The resulting fluorescent
signal was imaged with a laser-scanning confocal micro-
scope (talamasca LSM510; Carl Zeiss). Mitochondrial volume
was quantified using Volocity software (www.improvision
.com). 3D mitochondrial imaging was reconstructed from a
z-stack of optical sections. 3D image analysis tool from the
software was used to quantify volume of mitochondria and
expressed in mm3.
hES cell karyotyping
hES cellsgrowinginlog phaseweretreatedwith0.1mg=mL
of colcemid for induction of mitotic arrest. Cell cultures were
Table 1.Primers and Reaction Conditions for Reverse Transcriptase-Polymerase Chain Reaction
temperature (8C)Cycles Product (bp)
AFP60 35 216
Ncam60 32 164
NeuroD 56 32578
RunX252 37 375
HNF4A50 37 270
Glut2 5737 350
LONG-TERM CULTURE ON HUMAN ES CELL AGING129
subsequently harvested by standard cytogenetic methods of
trypsin dispersal, hypotonic shock, and fixed with 3:1 metha-
nol:acetic acid . Mitotic cell slide preparations were ana-
lyzed by the G-banding method and interpreted by an
investigator blinded to study conditions .
Detection of telomerase activity
Telomerase activity was assayed in triplicates using the
TRAPeze ELISA Detection Kit (Chemicon=Millipore) as per
the manufacturer’s instructions. Briefly, cells were lysed in
CHAPS lysis buffer, and the cell extracts were frozen on dry
ice. Telomerase was allowed to add telomeric repeats
(GGTTAG) onto the 30end of a biotinylated telomerase
substrate oligonucleotide at 308C for 30min. The extended
products were amplified by PCR using biotinylated telo-
merase substrate oligonucleotide and reverse primers and a
deoxy-cytidine triphosphate (dCTP). The labeled PCR
products were immobilized on streptavidin-coated micro-
titer plates and detected by an anti-dinitrophenyl antibody
conjugated to horseradish peroxidase (HRP). The amount of
product was determined by HRP activity using the HRP
substrate 3,30,5,50-tetramethylbenzidine and subsequent col-
or development. The absorbance of the samples was mea-
sured at 450 and 690nm with an automatic microplate reader
(Multiskan EX; Thermo Scientific). Telomerase activity was
determined using the following equation: absorbance at
450nm–690nm. Lysis buffer alone and heat-inactivated
lysed cells were used as negative controls. Heat inactivation
was conducted by incubating cells at 858C for 10min.
The data are represented as average?standard error of the
mean and analyzed for statistical significance (P<0.05) using
1-way analysis of variance with the Bonferonni correction.
Morphology and cell marker expression
over long-term culture
hES cells maintained on feeder layers or in feeder-free
culture grew as colonies of undifferentiated cells. No mor-
phological changes were observed during the culture period
of 82 passages (p38 to p120) for H9 or 62 passages (p61 to
p123) for PKU1 cells. Immunostaining revealed retainment
of stem cell surface markers (SSEA-4) and transcription fac-
tor expression (Oct4) from young to old passage cells (Fig. 1).
FACS analysis was used to further quantify SSEA-4 surface
marker expression on both young and old cells of the H9 and
PKU1 cell lines. The percentage of hES cells found to stain
positive for SSEA-4 did not significantly differ between
young and old cells of either the H9 (p48: 97.1%?1.4 vs.
p102: 98.0%?0.8) or PKU1 (p81: 69.2%?2.0 vs. p123:
57.5%?6.4) cell lines.
In vitro hES cell proliferation and differentiation
To determine whether long-term culture had any impact
on hES cell proliferation, we used a CyQuant cell prolifera-
tion assay to quantify cell division in low and high passage
H9 cells, PKU1 cells, and H9 cells stably transduced with a
double fusion (eGFP-Fluc) reporter gene (H9DF) . Al-
though proliferation rates were similar between young and
old H9 cells at 24 and 48h, older passage H9 cells began
proliferating at an elevated rate beginning 72h after the
initiation of the assay. For the H9DF cell line, a higher pro-
liferation rate for older passage cells was observed at 48h,
but led to reduced expansion at 72h due to overconfluency.
Increased proliferation was also detected in older passage
PKU1 cells as compared with younger passage cells (Fig. 2A–
C). To compare the capacity of young and old hES cells to
differentiate into derivatives of all 3 germ layers, EBs were
formed from low and high passage cells in vitro. EBs were
dissociated into pellets at day 14 and analyzed by semi-
quantitative RT-PCR to detect mRNA expression for plur-
ipotency markers (Oct-4), ectoderm (Ncam), mesoderm
(Flk-1), and endoderm (AFP). Low and high passage H9
derivatives cultured under the same differentiation condi-
tions expressed no significant differences at the mRNA level
after EB formation (Fig. 2D). However, older passage PKU1
cells appeared to yield EBs that differentiated less robustly as
detected by semiquantitative RT-PCR for the expression of
Ncam, Flk-1, and AFP (Fig. 2E).
In vivo teratoma formation by hES cells
Teratoma formation upon transplantation into immuno-
deficient animals is a hallmark of hES cells . To confirm
whether high passage cells would retain the capacity to form
derivatives of all 3 germ layers in vivo, we transplanted
5?105low (p49) and 5?105high (p126) passage H9DF cells
into the subcutaneous regions of the backs of SCID mice
(n¼4, two spots for each group per mouse). H9DF cells were
stably transduced with a reporter gene expressing enhanced
green fluorescent protein and firefly luciferase (eGFP-FLuc)
for noninvasive tracking of proliferation by bioluminescence
imaging . No significant differences between growth
rates of young and old passage cells were observed during
the 8-week period (Fig. 3A). All mice developed teratomas,
which were extracted 8 weeks after cell transplantation.
Teratomas from the old passage cells weighed slightly more
(0.16?0.08gm) than those arising from young passage cells
(0.15?0.08gm), but this difference was not statistically sig-
nificant (P¼0.83) (Fig. 3B). Hematoxylin and eosin staining
of tumor samples revealed fairly similar visual patterns of
differentiation between the 2 groups. Teratomas formed
from young and old passage cells contained derivatives of all
3 germ layers that were easily identifiable via light micro-
scopy (Fig. 3C). To quantify differentiation, semiquantitative
RT-PCR was performed on the RNA of explanted tumors for
pluripotency markers (Nanog, Oct4, and Rex1), ectodermal
markers (Ncam and NeuroD), mesodermal markers (Runx2,
HNF4a, and Nkx2.5), and endodermal markers (Sox17, Albu-
min, Glut2, and Insulin). Gene expression was normalized to
the expression of a house keeping gene, GAPDH (Fig. 3D).
Compared with low passage cells, teratomas arising from old
passage cells were found to have depressed levels of ex-
pression for markers of endodermal lineage (Sox17 and
Glut2), and elevated levels of mRNA specific to ectoder-
mal lineage (neuroD) (P<0.05). Older passage cells were
also characterized by slightly elevated expression levels
for markers of undifferentiation (Nanog and Rex1), and
130 XIE ET AL.
depressed levels of mRNA specific to mesodermal lineages
(Runx2), although these differences were not statistically
Mitochondrial function of young versus old hES cells
Although a causal relationship between abnormalities in
mitochondria and premature aging has only been established
in recent years , it is commonly accepted within the sci-
entific community that mitochondrial dysfunction is a mar-
ker of senescence . To investigate whether mitochondrial
function is impaired by prolonged passage, we compared
several parameters of mitochondrial function in young and
old hES cells. Functional analysis was determined by the
measurement of intracellular ROS, oxygen consumption, and
mitochondrial membrane potential. Mitochondrial volume
within the cell was also assessed using confocal microscopy.
Late passage H9 cells were found to have higher levels of
intracellular ROS (1,071?23MFI=105cells) than younger
passage cells (847?147MFI=105cells, P<0.05, Fig. 4A). Late
passage PKU1 cells were also observed to have elevated
levels of ROS (2,117?150MFI=105cells) compared with
young passage cells (1,758?151MFI=105cells) (Fig. 4B). Older
passage H9 cells were also observed to have a slightly de-
creasedrateof oxygen consumption
O2=min=106cells) compared with young cells (4.04?0.47
nmol O2=min=106cells) (Fig. 4C).
Mitochondrial membrane potential (DCm) is a measure
of the transmembrane electrical gradient of mitochondria.
We measured DCm in low and high passage hES cells using
a JC-1 assay kit. DCm was numerically calculated as the ratio
between the intramitochondrial aggregate (red) signal to
cytoplasmic monomeric (green) signal of the dyes . Late
passage H9 cells were found to have significantly elevated
DCm (34.38?8.70) when compared with young passage
cells (12.97?2.92) (P<0.001, Fig. 4D). Similar findings were
observed in old passage PKU1 cells (2.33?0.05) compared
with younger passage cells (1.87?0.14) (P<0.01, Fig. 4E).
To observe the mitochondria and evaluate mitochondrial
mass using confocal fluorescence microscopy, we used a
MitoTracker?mitochondrion-selective probe assay. The total
mitochondrial volume in low and high passage hES cells was
documented (Fig. 4F). In young passage H9 cells, the total
volume of mitochondria was on average, 220?106mm3=cell.
In contrast, older cells displayed a much higher mean vol-
ume of 528?81mm3=cell (P<0.01, Fig. 4G). Mitochondria
volume of young passage PKU1 cells was 390?143mm3=cell,
and increased to 990?209mm3=cell for old passage cells
(P<0.005, Fig. 4H).
Karyotypic analysis of hES cells
Karyotyping is a traditional measure of cell line stability
and safety . To confirm that hES cells maintained
tained in an undifferentiated state
over prolonged in vitro culture.
hES cells (H9 line) were cultured
for 90 passages on mouse embry-
onic fibroblast feeder layers (from
passage 38 to 128). Expression of
the pluripotency transcription fac-
tor Oct4 and cell surface marker
SSEA-4 was examined by immu-
nofluorescent staining (red). Nu-
clei were costained with DAPI
(blue). Late passage cells were not
significantly different from youn-
ger passage cells in cell morphol-
ogy or expression of pluripotency
markers. DAPI, 40,6-diamidino-2-
phenylindole; hES cells, human
embryonic stem cells; SSEA, stage-
specific embryonic antigen.
hES cells can be main-
LONG-TERM CULTURE ON HUMAN ES CELL AGING131
chromosomal integrity over the period of this experiment,
we karyotyped H9 and PKU1 hES cells by standard G-
banding techniques. For H9 cells, young passage cells (p48)
presented with a normal female karyotype as 46, XX. Cells
passaged to p120 were found to have a karyotype change
event consisting of a translocation between the long arms of
chromosomes 1 and 9, resulting in the deletion of one of the
long arms of chromosome 9 and the duplication of one of the
long arms of chromosome 1. The new karyotype presented
itself as 46, XX, der(9)t(1;9)(q31;q22). PKU1 cells were ob-
served to have a stable karyotype for the culture period of 1
year. Both early and late passage PKU1 cells had normal
female karyotype as 46, XX (Fig. 5).
Telomerase activity of hES cells
High telomerase activity or expression of telomerase re-
verse transcriptase (TERT), the catalytic protein subunit of
telomerase, is regarded as a marker of stem cells . Telo-
mere shortening, which occurs as TERT activity declines, is
one of the fundamental molecular mechanisms underlying
cell aging [36–38]. A telomere repeat amplification protocol
assay was performed to compare telomerase activity in low
(p39) and high passage (p110) H9 cells. Both young and old
H9 cells showed equally elevated levels of telomerase ac-
tivity (5,014?294 vs. 5,030?377 unit=mg, P¼NS, Fig. 6). To
compare these levels of enzyme activity with cells known to
have upregulated levels of TERT, we also measured telo-
merase activity in 4 human cancer cell lines: (1) a human
pancreatic cancer cell line BXPC3; (2) a human glioblastoma-
astrocytoma, epithelial-like cell line U87MG; (3) a human
leukemic monocyte lymphoma cell line U937; and (4) a hu-
man colon adenocarcinoma cell line HT29. On average, tel-
omerase activity in both young and old passage H9 hES cells
was 2-fold higher than in tumor cell lines (2,562?1,909
unit=mg). Further, the level of telomerase activity in hES cells
remained stable from passage 39 to passage 110.
hES cells are defined by their capacity for unlimited self-
renewal and pluripotency. However, the issue of whether
hES cells can be maintained stably under prolonged in vitro
conditions is still unresolved. In this study, the long-term
stability of hES cells was confirmed by the preservation of
stem cell surface marker expression, telomerase activity, and
PKU1 hES cells was quantified at 24, 48, and 72h by a CyQuant cell proliferation assay. Late passage hES cells were generally
found to have a higher proliferation rate compared with younger passage cells. Specifically, older passage H9 cells exhibited
elevated proliferation a 72h, H9DF cells at 48h, and PKU1 cells at 48 and 72h compared with younger counterparts. (D, E)
Young and old passage cells were allowed to spontaneously differentiate into embroid bodies in vitro, after which cell
derivatives of all 3 germ layers were examined by semiquantitative reverse transcriptase-polymerase chain reaction. H9 cells
exhibited no significant differences in the expression patterns of endoderm marker (AFP), mesoderm marker (Flk-1), and
ectoderm marker (Ncam) between young and old passage cells. Older passage PKU1 cells exhibited reduced capacity to form
cell derivatives of the 3 germ layers however.
In vitro hES cell proliferation and differentiation. (A–C) Proliferation of young and old passage H9, H9DF, and
132 XIE ET AL.
relative expression (/gapdh)
man ES cells. A half million H9 cells from early passage (P49)
and late passage (P126) were injected subcutaneously on the
backs of nude mice. Tetatomas were harvested 8 weeks after
injection, weighed, and processed for hematoxylin and eosin
histology. (A) Bioluminescence imaging of subcutaneous
teratoma formation revealed no significant differences in
growth between young and old passage hES cells implanted
subcutaneously. (B) No significant difference in weight was
observed between teratomas arising from young and late
passage H9DF cells. (C) Postmortem histological analysis of
transplanted hES cells showed (a) immature brain-like neural
cell formation; (b) squamous cell differentiation with keratin
pearls; (c, d) osteochondroid formation; (e, f) respiratory
epithelium with ciliated columnar and mucin producing
goblet cells. Both early and late passage hES cells can form all
3 germ layer morphologies. (D) Reverse transcriptase-poly-
merase chain reaction analysis of explanted teratomas at
week 8. Gene expression was normalized to a house keeping
gene, GAPDH. Both young and old passage hES cells formed
derivatives of all 3 germ layers. Old passage hES cells were
characterized by elevated levels of expression for ectodermal
markers (NeuroD) and decreased levels of expression for
endodermal markers (Sox17 and Glut2) as compared with
younger passage counterparts.
Teratoma formation by early and late passage hu-
LONG-TERM CULTURE ON HUMAN ES CELL AGING 133
capacity to differentiate into derivatives of all 3 germ layers
upon EB formation in vitro and teratoma formation in vivo.
Karyotyping and analysis of mitochondrial function, how-
ever, revealed differences between older and younger pas-
sage cells. The karyotype abnormalities [46,XX,der(9)t(1;9)
(q31;q22)] observed in H9 cells suggest prolonged culture of
hES cells or their derivatives may result in heightened risk of
cancerous transformation or functional abnormalities. While
late passage H9 cells were observed to have heightened cell
proliferation rates as compared with younger passage cells,
such increases in cell division were also found in the PKU1
cell line which possessed a normal karyotype. This suggests
that increases in cell proliferation for older passage hES cells
are not necessarily the result of a karyotypic abnormalities. A
recent study by Park et al. shows that correlations may exist
between increases in in vitro cellular division and tumorigene-
sis . However, in this study in vivo bioluminescence imaging
of teratoma formation did not reveal any significant differences
between tumor formation of young and old hES cells.
Interestingly, our examination of teratomas by semi-
quantitative PCR revealed that the age of hES cells may af-
fect their proclivity for differentiation into specific germ
lineages. Younger passage H9 cells differentiated evenly into
derivatives of ectodermal, mesodermal, and endodermal
germ layers. In contrast, higher passage hES cells primarily
formed derivatives of ectodermal lineage and were charac-
terized by lower levels of mRNA for derivatives of endo-
dermal origin. Preferential differentiation into ectodermal
derivatives was most likely influenced by the presence of
basic fibroblast growth factor in the culture media for the
duration of the experiment. Basic fibroblast growth factor
has been shown to be a strong neurotrophic factor and can be
used to induce differentiation of hES cells into neuroecto-
derm . These results indicate that both the culture me-
dium and cell age may potentially influence the preferential
differentiation of hES cells during prolonged in vitro culture.
Further studies will be required to examine the extent to
which prolonged cell culture affects the use of high passage
hES cells for derivation of terminally differentiated cells in
models of cell replacement therapy.
Mitochondria were first implicated in the aging process
for over 50 years ago by Denham Harman in his free radical
have mitochondrial dysfunction. (A, B) Intracellular ROS were quantified in young and old passage hES cells by DCFH-DA
fluorescent staining. Older passage cells were found to have higher ROS production than younger passage cells in both the
H9 and PKU1 cell lines, suggesting that hES cells are susceptible to oxidative stress with extended culture. (A) H9 cells; (B)
PKU1 cells. (C) Measurement of oxygen consumption revealed a decrease in O2 consumption in old passage H9 cells, but this
result was not significant. (D, E) FACS analysis of MitoProbe JC-1 staining showed higher mitochondrial membrane potential
in older passage H9 and PKU1 cells as compared with younger counterparts. (D) H9 cells; (E) PKU1 cells. (F) Confocal
microscopy was used to assess mitochondrial volume in fluorescent images of living H9 and PKU1 cells stained with
MitoTracker dye. (Scale bar: 7mm) (G, H) Total mitochondrial volume was quantified using confocal microcopy. Mito-
chondrial volume was found to be highly elevated in older passage hES cells as compared with younger passage cells. (G) H9
cells; (H) PKU1 cells. ROS, reactive oxygen species.
Mitochondrial function of young versus old hES cells. hES cells undergoing prolonged cell culture were observed to
134XIE ET AL.
theory of aging . In his original article, Harman postu-
lated that ROS made as a byproduct of oxidative phos-
phorylation drive the aging process via free radical damage
to mitochondrial DNA (mtDNA), which in turn leads to
decreased mitochondrial function and the release of pro-
apoptotic factors such as cytochrome C. A large body of
evidence has since validated this theory , and recently
several articles have shown a causal relationship between
mitochondrial defects and premature aging [31,42]. Declines
in mitochondrial function are commonly observed with ag-
ing in somatic cells . Despite these findings, little research
has been conducted on changes that occur in the mitochon-
dria of ES cells during prolonged culture.
To investigate whether functional changes occurred in
the mitochondria of hES cells over time, we compared ROS
content, mitochondrial membrane potential, and oxygen
consumption in young and old passage hES cells. The re-
sults from our mitochondrial experiments support the idea
that hES cells maintained in culture are not totally immune
to the effects of senescence. A number of studies have
tracked changes in the mitochondria of hES cells that occur
with aging and development. These studies have shown
that not only does the intracellular localization of mito-
chondria change with blastocyst development, but also the
differentiation of hES cells is associated with increases in
mitochondrial mass, ATP production, and ROS production
. Schieke et al. recently found that when mouse ES
cells are separated by mitochondrial membrane potential,
However, passage 120 H9 cells were found to have an unbalanced chromosomal translocation involving the long arms of
chromosomes nos. 1 and 9, resulting in the karyotypic change 46,XX,der(9)t(1;9)(q31;q22). Both passage 82 and 124 PKU1
cells retained normal female karyotype as 46, XX.
Karyotyping of early and late passage hES cells. Passage 48 H9 cells showed a normal female karyotype as 46, XX.
old tumor cells
amplification protocol assay. Both early and late passage H9
hES cells maintained high levels of telomerase activity as
compared with tumor cells (averaged values for BXPC3,
U87MG, U937, and HT29). There was no significant differ-
ence found between young and old passage hES cells
5030?377 unit=mg for old passage cells, P¼NS).
Detection of telomerase activity by telomere repeat
unit=mg foryoung passagecells versus
LONG-TERM CULTURE ON HUMAN ES CELL AGING135
different groups showed distinct germ layer differentiation
tendency and teratoma formation rate . Our results are
highly similar to these reports, in that we found older pas-
sage hES cells to have an increased mitochondrial mass,
a higher ROS content, and an elevated mitochondrial
membrane potential compared with their younger passage
Senescent mitochondria can affect clinical use of hES cells
in several ways. In somatic cells, defects in mitochondria that
occur with aging have been shown to impair ATP produc-
tion and cause early cellular senescence [46,47]. Acceleration
in mitochondrial proliferation and development has also
been correlated to a loss of pluripotency. Two recent reports
have shown that in the developing blastocyst, elevation in
mtDNA copy number and increased drive for aerobic res-
piration correspond to an ensuing loss of pluripotency [48,49].
Similarly, repression of mitochondrial development by plac-
ing hES cells in hypoxic conditions has been shown to slow
differentiation . In this report, the elevated volume of
mitochondria observed in older passage hES cells by confo-
cal microscopy suggests elevated proliferation and expan-
sion of mitochondrial number. This may indicate impaired
removal of damaged and nonfunctional mitochondria in
older cells, and=or fusion of existing mitochondria as a com-
pensatory mechanism against the observed mitochondrial
dysfunction. This may potentially result in the accumulation
of ROS, which would explain the increased ROS content
observed in higher passage cells.
Finally, mitochondrial membrane potential is a pivotal
controller of respiratory rate, ATP synthesis, and the gener-
ation of ROS, and is itself controlled by electron transport
and proton leaks. Although it is unclear what causes the
increase of mitochondrial membrane potential in late pas-
sage hES cells, a high mitochondrial membrane potential in
our results may be the cause for the observed increase in ROS
production, as it has been clearly shown that at high DCms
even a small increase in membrane potential can give rise to
a large stimulation of H2O2production . If older passage
hES cells are susceptible to the generation of increased
amounts of ROS due to changes in DCm, the accumulation
of oxidative damage over the course of prolonged culture
may have serious implications for the use of late passage
cells in therapeutic medicine.
In summary, we found that although hES cells are resis-
tant to most aspects of senescence such as loss of plur-
ipotency, telomerase activity, and stem cell marker expression,
they are not completely immune to the aging process as their
mitochondrial function declines, making hES cell mitochon-
dria vulnerable to insults. Changes in mitochondrial function
observed with prolonged culture are significant because of
their effect upon cellular metabolism and their potential to
adversely affect ROS generation and integrity of mtDNA.
The nuclear genome may also be vulnerable to these insults
as evidenced by a chromosomal translocation after 80 pas-
sages in H9 cells. Karyotypic abnormalities did not result in
significant changes in teratoma formation upon transplan-
tation into immunodeficient animals, but have the potential
to place the cell at an increased risk for malignant transfor-
mation later. PKU1 cells cultured over 120 passages were not
observed to have a karyotypic change, but exhibited similar
differences in ROS production, mitochondrial membrane
potential, and volume, suggesting that these changes result
from prolonged cell culture and not chromosomal instability.
Preferential differentiation of older passage hES cells in vivo
into derivatives of ectodermal lineage is another issue that
arises from long-term culture, but is most likely a result of
the culture medium used to grow the cells. Taken together,
this is the first study to conduct an in-depth functional
evaluation of hES cells after long-term culture. This study
adds valuable new knowledge regarding the little-under-
stood senescence process of hES cells, as well as insights into
how long-term passaged hES cells can be used for future
We are grateful to Andrew J. Connolly for assistance with
histological analysis. We thank funding support from NIH
DP2OD004437, AI085575, HL091453, HL089027 (J.C.W.),
HHMI Research Fellowship (A.S.L.), and the National 973
project of China 2011CB964800 (T.C.).
Author Disclosure Statement
No competing financial interests exist.
1. Thomson JA, J Itskovitz-Eldor, SS Shapiro, MA Waknitz, JJ
Swiergiel, VS Marshall and JM Jones. (1998). Embryonic
stem cell lines derived from human blastocysts. Science
2. Laflamme MA, KY Chen, AV Naumova, V Muskheli, JA
Fugate, SK Dupras, H Reinecke, C Xu, M Hassanipour, S
Police, O’Sullivan C, L Collins, Y Chen, E Minami, EA Gill, S
Ueno, C Yuan, J Gold and CE Murry. (2007). Cardiomyo-
cytes derived from human embryonic stem cells in pro-
survival factors enhance function of infarcted rat hearts. Nat
3. Cao F, RA Wagner, KD Wilson, X Xie, Fu J-D, M Drukker, A
Lee, RA Li, SS Gambhir, IL Weissman, RC Robbins and JC
Wu. (2008). Transcriptional and functional profiling of hu-
man embryonic stem cell-derived cardiomyocytes. PLoS
4. Ben-Hur T, M Idelson, H Khaner, M Pera, E Reinhartz, A
Itzik and BE Reubinoff. (2004). Transplantation of human
embryonic stem cell-derived neural progenitors improves
behavioral deficit in Parkinsonian rats. Stem Cells 22:1246–
5. Kim JH, JM Auerbach, JA Rodriguez-Gomez, I Velasco, D
Gavin, N Lumelsky, SH Lee, J Nguyen, R Sanchez-Pernaute,
K Bankiewicz and R McKay. (2002). Dopamine neurons
derived from embryonic stem cells function in an animal
model of Parkinson’s disease. Nature 418:50–56.
6. Kroon E, LA Martinson, K Kadoya, AG Bang, OG Kelly, S
Eliazer, H Young, M Richardson, NG Smart, J Cunningham,
AD Agulnick, D’Amour KA, MK Carpenter and EE Baetge.
(2008). Pancreatic endoderm derived from human embry-
onic stem cells generates glucose-responsive insulin-secreting
cells in vivo. Nat Biotechnol 26:443–452.
7. Soto-Gutierrez A, N Kobayashi, JD Rivas-Carrillo, N Na-
varro-Alvarez, D Zhao, T Okitsu, H Noguchi, H Basma, Y
Tabata, Y Chen, K Tanaka, M Narushima, A Miki, T Ueda,
HS Jun, JW Yoon, J Lebkowski, N Tanaka and IJ Fox. (2006).
Reversal of mouse hepatic failure using an implanted liver-
assist device containing ES cell-derived hepatocytes. Nat
136 XIE ET AL.
8. Narayan AD, JL Chase, RL Lewis, X Tian, DS Kaufman, JA
Thomson and ED Zanjani. (2006). Human embryonic stem
cell-derived hematopoietic cells are capable of engrafting
primary as well as secondary fetal sheep recipients. Blood
9. Tian X, PS Woll, JK Morris, JL Linehan and DS Kaufman.
(2006). Hematopoietic engraftment of human embryonic
stem cell-derived cells is regulated by recipient innate im-
munity. Stem Cells 24:1370–1380.
10. Chen SL, WW Fang, F Ye, YH Liu, J Qian, SJ Shan, JJ Zhang,
RZ Chunhua, LM Liao, S Lin and JP Sun. (2004). Effect on
left ventricular function of intracoronary transplantation of
autologous bone marrow mesenchymal stem cell in patients
with acute myocardial infarction. Am J Cardiol 94:92–95.
11. Wollert KC, GP Meyer, J Lotz, S Ringes-Lichtenberg, P
Lippolt, C Breidenbach, S Fichtner, T Korte, B Hornig, D
Messinger, L Arseniev, B Hertenstein, A Ganser and H
Drexler. (2004). Intracoronary autologous bone-marrow cell
transfer after myocardial infarction: the BOOST randomised
controlled clinical trial. Lancet. 364:141–148.
12. Goldstein S. (1990). Replicative senescence: the human fi-
broblast comes of age. Science 249:1129–1133.
13. Finkel T and NJ Holbrook. (2000). Oxidants, oxidative stress
and the biology of ageing. Nature 408:239–247.
14. Kamminga LM and GD Haan. (2006). Cellular memory and
hematopoietic stem cell aging. Stem Cells 24:1143–1149.
15. Cervantes RB, JR Stringer, C Shao, JA Tischfield and PJ
Stambrook. (2002). Embryonic stem cells and somatic cells
differ in mutation frequency and type. Proc Natl Acad Sci
16. Amit M, MK Carpenter, MS Inokuma, C-P Chiu, CP Harris,
MA Waknitz, J Itskovitz-Eldor and JA Thomson. (2000).
Clonally derived human embryonic stem cell lines maintain
pluripotency and proliferative potential for prolonged peri-
ods of culture. Dev Biol 227:271–278.
17. Rosler ES, GJ Fisk, X Ares, J Irving, T Miura, MS Rao and
MK Carpenter. (2004). Long-term culture of human em-
bryonic stem cells in feeder-free conditions. Dev Dyn 229:
18. Brimble SN, X Zeng, DA Weiler, Y Luo, Y Liu, IG Lyons, WJ
Freed, AJ Robins, MS Rao and TC Schulz. (2004). Karyotypic
stability, genotyping, differentiation, feeder-free mainte-
nance, and gene expression sampling in three human em-
bryonic stem cell lines derived prior to August 9, 2001. Stem
Cells Dev 13:585–597.
19. Buzzard JJ, NM Gough, JM Crook and A Colman. (2004).
Karyotype of human ES cells during extended culture. Nat
20. Rugg-Gunn PJ, AC Ferguson-Smith and RA Pedersen.
(2005). Epigenetic status of human embryonic stem cells. Nat
21. Mitalipov SM. (2006). Genomic imprinting in primate em-
bryos and embryonic stem cells. Reprod Fertil Dev 18:
22. Sun BW, AC Yang, Y Feng, YJ Sun, Y Zhu, Y Zhang, H Jiang,
CL Li, FR Gao, ZH Zhang, WC Wang, XY Kong, G Jin, SJ Fu
and Jin Y. (2006). Temporal and parental-specific expression
of imprinted genes in a newly derived Chinese human em-
bryonic stem cell line and embryoid bodies. Hum Mol Genet
L Meisner, TP Zwaka, JA Thomson and PW Andrews. (2004).
Recurrent gain of chromosomes 17q and 12 in cultured human
embryonic stem cells. Nat Biotechnol 22:53–54.
24. Maitra A, DE Arking, N Shivapurkar, M Ikeda, V Stastny, K
Kassauei, G Sui, DJ Cutler, Y Liu, SN Brimble, K Noaksson,
J Hyllner, TC Schulz, X Zeng, WJ Freed, J Crook, S Abra-
ham, A Colman, P Sartipy, S Matsui, M Carpenter, AF
Gazdar, M Rao and Chakravarti A. (2005). Genomic alter-
ations in cultured human embryonic stem cells. Nat Genet
25. Pannetier M and Feil R. (2007). Epigenetic stability of em-
bryonic stem cells and developmental potential. Trends
26. Peng H-m and G-a Chen. (2006). Serum-free medium culti-
vation to improve efficacy in establishment of human em-
bryonic stem cell lines. Hum Reprod 21:217–222.
27. Li Z, Y Suzuki, M Huang, F Cao, X Xie, AJ Connolly, PC
Yang and JC Wu. (2008). Comparison of reporter gene and
iron particle labeling for tracking fate of human embryonic
stem cells and differentiated endothelial cells in living sub-
jects. Stem Cells 26:864–873.
28. Sun N, A Lee and JC Wu. (2009). Long term non-invasive
imaging of embryonic stem cells using reporter genes. Nat-
ure protocols 4:1192–1201.
29. Barch MJ, T Knutsen and JL Spurbeck. (1997). The AGT
Cytogenetics Laboratory Manual. Chapter 4 (3rd Edition).
Lippincott-Raven, New York.
30. Seabright M. (1971). A rapid banding technique for human
chromosomes. Lancet 2:971–972.
31. Vermulst M, J Wanagat, GC Kujoth, JH Bielas, PS Rabino-
vitch, TA Prolla and LA Loeb. (2008). DNA deletions and
clonal mutations drive premature aging in mitochondrial
mutator mice. Nat Genet 40:392–394.
32. Balaban RS, S Nemoto and T Finkel. (2005). Mitochondria,
oxidants, and aging. Cell 120:483–495.
33. Smiley ST, M Reers, Mottola-Hartshorn C, M Lin, A Chen,
TW Smith, GD Steele, Jr., and LB Chen. (1991). Intracellular
heterogeneity in mitochondrial membrane potentials re-
vealed by a J-aggregate-forming lipophilic cation JC-1. Proc
Natl Acad Sci USA 88:3671–3675.
34. Mamaeva SE. (1998). Karyotyptic evolution of cells in cul-
ture: a new concept. Int Rev Cytol 178:1–40.
35. Hiyama E and Hiyama K. (2007). Telomere and telomerase
in stem cells. Br J Cancer 96:1020–1024.
36. Collins K and JR Mitchell. (2002). Telomerase in the human
organism. Oncogene 21:564–579.
37. Blasco MA. (2005). Telomeres and human disease: ageing,
cancer and beyond. Nat Rev Genet 6:611–622.
38. Wright WE and JW Shay. (2005). Telomere biology in aging
and cancer. J Am Geriatr Soc 53(9 Suppl):S292–S294.
39. Park YB, YY Kim, SK Oh, SG Chung, SY Ku, SH Kim,
YM Choi and SY Moon. (2008). Alterations of prolifera-
tive and differentiation potentials of human embryonic
stem cells during long-term culture. Exp Mol Med 40:98–
40. Park S, KS Lee, YJ Lee, HA Shin, HY Cho, KC Wang, YS
Kim, HT Lee, KS Chung, EY Kim and J Lim. (2004). Gen-
eration of dopaminergic neurons in vitro from human em-
bryonic stem cells treated with neurotrophic factors. Neurosci
41. Harman D. (1956). Aging: a theory based on free radical and
radiation chemistry. J Gerontol 11:298–300.
42. Trifunovic A, A Wredenberg, M Falkenberg, JN Spelbrink,
AT Rovio, CE Bruder, YM Bohlooly, S Gidlof, A Oldfors, R
Wibom, J Tornell, HT Jacobs and NG Larsson. (2004). Pre-
mature ageing in mice expressing defective mitochondrial
DNA polymerase. Nature 429:417–423.
LONG-TERM CULTURE ON HUMAN ES CELL AGING137
43. Linnane AW, S Marzuki, T Ozawa, et al. (1989). Mitochon-
drial DNA mutations as an important contributor to ageing
and degenerative diseases. Lancet 1:642–645.
44. Brenner CA, HM Kubisch and KE Pierce. (2004). Role of the
mitochondrial genome in assisted reproductive technologies
and embryonic stem cell-based therapeutic cloning. Reprod
Fertil Dev 16:743–751.
45. Schieke SM, M Ma, L Cao, McCoy JP, C Liu, NF Hensel, AJ
Barrett, M Boehm and T Finkel. (2008). Mitochondrial me-
tabolism modulates differentiation and teratoma formation
capacity in mouse embryonic stem cells. J Biol Chem 283:
46. Nesti C, L Pasquali, F Vaglini, G Siciliano and L Murri.
(2007). The role of mitochondria in stem cell biology.
Bioscience reports. 27:165–171.
47. St. John JC, Ramalho-Santos J, HL Gray, P Petrosko, VY
Rawe, CS Navara, CR Simerly and GP Schatten. (2005). The
expression of mitochondrial DNA transcription factors
during early cardiomyocyte in vitro differentiation from
human embryonic stem cells. Cloning Stem Cells 7:141–153.
48. St. John J and R Lovell-Badge. (2007). Human-animal cyto-
plasmic hybrid embryos, mitochondria, and an energetic
debate. Nat Cell Biol 9:988–992.
49. Rivolta MN and MC Holley. (2002). Asymmetric segregation
of mitochondria and mortalin correlates with the multi-lineage
potential of inner ear sensory cell progenitors in vitro. Brain Res
Dev Brain Res 133:49–56.
50. Ezashi T, P Das and RM Roberts. (2005). Low O2 tensions
and the prevention of differentiation of hES cells. Proc Natl
Acad Sci USA 102:4783–4788.
51. Sergey SK, PS Vladimir and AS Anatoly. (1997). High pro-
tonic potential actuates a mechanism of production of re-
active oxygen species in mitochondria. FEBS Lett 416:15–18.
Address correspondence to:
Dr. Joseph C. Wu
Department of Medicine
Stanford University School of Medicine
300 Pasteur Drive, Grant S140
Stanford, CA 94305-5111
Dr. Xuetao Pei
Stem Cell and Regenerative Medicine Lab
Beijing Institute of Transfusion Medicine
27 Taiping Road
Received for publication November 22, 2009
Accepted after revision July 12, 2010
Prepublished on Liebert Instant Online July 14, 2010
138 XIE ET AL.