Long-term monitoring of transplanted human neural
stem cells in developmental and pathological
contexts with MRI
Raphael Guzman*, Nobuko Uchida†, Tonya M. Bliss*, Dongping He†, Karen K. Christopherson†, David Stellwagen‡,
Alexandra Capela‡, Joan Greve§, Robert C. Malenka‡, Michael E. Moseley§, Theo D. Palmer*, and Gary K. Steinberg*¶
*Department of Neurosurgery, Stanford University School of Medicine, 300 Pasteur Drive, R200, Stanford, CA 94305-5327;†StemCells, Inc., 3155 Porter Drive,
Palo Alto, CA 94304-1213,‡Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, MSLS P104, Stanford, CA 94305-5485;
and§Department of Radiology, Lucas Magnetic Resonance Spectroscopy and Imaging Center, Stanford University School of Medicine, P286, Stanford, CA 94022
Edited by Irving L. Weissman, Stanford University School of Medicine, Stanford, CA, and approved April 24, 2007 (received for review October 6, 2006)
Noninvasive monitoring of stem cells, using high-resolution mo-
lecular imaging, will be instrumental to improve clinical neural
nervous system stem cells grown as neurospheres with magnetic
nanoparticles does not adversely affect survival, migration, and
differentiation or alter neuronal electrophysiological characteris-
tics. Using MRI, we show that human central nervous system stem
cells transplanted either to the neonatal, the adult, or the injured
rodent brain respond to cues characteristic for the ambient micro-
environment resulting in distinct migration patterns. Nanoparticle-
labeled human central nervous system stem cells survive long-term
and differentiate in a site-specific manner identical to that seen for
transplants of unlabeled cells. We also demonstrate the impact of
graft location on cell migration and describe magnetic resonance
edge of migration patterns and implementation of noninvasive
stem cell tracking might help to improve the design of future
clinical neural stem cell transplantation.
superparamagnetic iron oxide ? stem cell biology
as Parkinson’s (1–3) and Huntington’s (4), and stroke (5). The
success of these trials for neurological diseases will depend not only
on patient selection criteria and on choosing the right cell type, but
also on the timing and site of transplantation. Both variables can
influence the transplanted stem cells’ migration pattern and sub-
sequent differentiation (6, 7). Therefore, long-term monitoring of
the graft in relation to the evolving lesion will be crucial.
MRI, with its high spatial resolution, is the ideal modality for in
vivo cell tracking. Tagging cells with superparamagnetic iron oxide
(SPIO) nanocomposites has been shown to induce sufficient MR
cell contrast for in vivo imaging of neural cell migration (8, 9).
Previous studies have demonstrated its application to track stem
cells after stroke, but these were done with non-human stem cells
Before this method can be considered to label human neural
stem cells for clinical application, further analysis of the effects of
SPIO on the biology of human stem cells are needed. For example,
it has been described that Feridex, a SPIO reagent approved by the
United States Food and Drug Administration for human use,
inhibits mesenchymal stem cells from differentiating into chondro-
cytes (12), emphasizing the need for in-depth analysis of the
influence of magnetic labeling on stem cell biology.
We investigated the effects of SPIO labeling on human central
nervous system stem cells grown as neurospheres (hCNS-SCns) (6,
13–15) in vitro and in vivo. We show that SPIO-labeled hCNS-SCns
proliferate and differentiate normally in vitro and exhibit neuronal
electrophysiological characteristics. We then used MRI to analyze
migration patterns of SPIO-labeled hCNS-SCns transplanted into
dvances in neural transplantation have paved the way for
newborn and adult (injured and uninjured) rodent brain. We find
that SPIO-labeled hCNS-SCns exhibit a similar extent of survival,
migration, integration, and differentiation after transplantation
into the rodent brain as their unlabeled counterparts.
In vitro Characterization and MRI of SPIO-Labeled Human Neural Stem
is feasible and does not impair cell viability and biological charac-
teristics in vitro. HCNS-SCns were labeled with Feridex and pro-
tamine sulfate and tested for labeling efficiency, cell proliferation,
differentiation as well as electrophysiological behavior. SPIO was
detected in the cytoplasm (Fig. 1A). We achieved a 98% labeling
efficiency as determined by Prussian Blue staining [supporting
information (SI) Fig. 5]. Cell viability for SPIO-labeled and unla-
beled hCNS-SCns was 92 ? 3% vs. 96 ? 2%, respectively. There
was no difference in the proliferation rate between SPIO-labeled
and unlabeled hCNS-SCns (Fig. 1B).
SPIO-labeled hCNS-SCns were grown under proliferative condi-
tions for 9 days, cell samples were stained with an anti-dextran
antibody and the relative iron content per cell measured. On
average, the SPIO content per cell was halved every 3 days in
culture, which correlated with cell doubling time (Fig. 1C and SI
We compared the differentiation potential of SPIO-labeled and
unlabeled hCNS-SCns in vitro. After 10 days under differentiation
conditions, an average of 35.5% (? 4) of the cells were Nestin-
positive, 56.2% (? 7.7) were ?-tubulin-positive, and 19.4% (? 3.5)
were glial fibrillary acidic protein (GFAP)-positive (Fig. 1D) in the
SPIO-labeled hCNS-SCns. There was no statistically significant
difference between the two groups (Fig. 1D).
Whole-cell patch-clamp recordings were made from SPIO-
labeled and unlabeled hCNS-SCns 28 days after differentiation.
Cells displaying a typical neuronal morphology (two to five well
defined primary processes (16) were chosen for the recordings.
Author contributions: R.G., R.C.M., M.E.M., T.D.P., and G.K.S. designed research; R.G.,
T.M.B., K.K.C., D.S., and J.G. performed research; N.U., D.H., K.K.C., and A.C. contributed
new reagents/analytic tools; R.G., T.M.B., D.S., J.G., R.C.M., and M.E.M. analyzed data; and
R.G., T.D.P., and G.K.S. wrote the paper.
Conflict of interest statement: N.U., D.H., K.K.C., and A.C. are employees of StemCells, Inc.
neural stem cells and human-specific antibodies.
This article is a PNAS Direct Submission.
Abbreviations: GFAP, glial fibrillary acidic protein; hCNS-SCns, human central nervous
SPIO, superparamagnetic iron oxide; SVZ, subventricular zone.
¶To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
© 2007 by The National Academy of Sciences of the USA
June 12, 2007 ?
vol. 104 ?
no. 24 ?
action potentials in cells from both groups (Fig. 1E), and voltage-
gated Na?and K?currents were evoked from stem cell-derived
neurons in both groups (Fig. 1F). Expression of ?-III tubulin and
the neuronal somatodendritic protein MAP2 was used to confirm
the neuronal identity of the measured cell populations.
We also found that the graft size of 5 ? 104SPIO-labeled cells
measured by in vivo MRI was similar to the graft size measured on
change seen on T2-weighted imaging caused by 5 ? 104SPIO-
labeled cells in an agarose gel was comparable with the signal
change of the same number of labeled cells transplanted to the rat
brain (SI Fig. 6).
SPIO-Labeled Human Neural Stem Cells Undergo Widespread Migra-
tion Along Natural Tracts in the Neonatal NOD-SCID Brain and Differ-
entiate into Neurons and Glia. We and others have shown that the
immature neonatal brain represents a unique environment to test
the natural migration, integration, and differentiation of trans-
the progeny of endogenous stem cells that have proliferated in the
subventricular zone (SVZ) enter the rostral migratory stream
(RMS) and migrate to the olfactory bulb, where they integrate in
a site-specific manner (14).
We asked whether MRI could be used to follow the natural
migration of SPIO-labeled hCNS-SCns and whether SPIO labeling
impairs migration and differentiation of hCNS-SCns in vivo. Im-
mediately after transplantation, distribution of hCNS-SCns in the
ventricular system could be seen by MRI. Three weeks after
transplantation, hCNS-SCns were distributed in the lateral and
fourth ventricle (Fig. 2A). Cells migrating along the RMS were
detected as a bilateral hypointense area on T2-weighted spin echo
and 3D gradient-echo MRI (Fig. 2A; gradient-echo images are
shown in SI Fig. 7). On subsequent MRI at 9 and 12 weeks (data
not shown), we detected migration from the SVZ into the corpus
callosum, extending posteriorly into the hippocampus and cortex
and anteriorly into the RMS and olfactory glomerulus. This mi-
gration remained visible up to 18 weeks (Fig. 2B). Migration,
integration, and differentiation were subsequently studied on sag-
ittal sections stained for the human-specific marker SC121. In
agreement with MRI, SC121 staining showed human cells in the
lateral ventricle integrating into the SVZ and migrating along the
RMS (Fig. 2D). Over 18 weeks, we found progressively more
widespread integration of hCNS-SCns into the host brain with
site-specific integration into the SVZ, the hippocampus, the cortex
and the olfactory bulb (OB) (Fig. 2 E and F). Microscopic images
at higher magnification clearly demonstrate integration into the
hippocampus along CA1, CA3, and the dentate gyrus (Fig. 2G). In
the olfactory bulb, hCNS-SCns adopt laminar positions (Fig. 2H).
Cell counting using unbiased stereology revealed no statistically
significant differences in the distribution and number of surviving
the histological correlate to the MRI signal changes, adjacent
sections were stained with Prussian blue and the human-specific
marker SC121. At 3 weeks, we found distinct punctuate Prussian
blue staining in the olfactory bulb, a site to which many SC121-
labeling, two adjacent paraffin sections were stained with Prussian
blue or the human-specific marker SC121. The sections were
with the panmonocytic marker Iba-1 and SC121 did not reveal
double-labeled cells in the migrating cell population in the RMS,
making it unlikely that the observed migration was mainly caused
by SPIO-laden macrophages (SI Fig. 8 G and H).
To quantify cell fate at 18 weeks, double immunofluorescence
stainings were done by using human-specific and neuronal-specific
antibodies (SI Fig. 9 A–C). A human-specific anti-GFAP antibody
was used to quantify astroglial differentiation (SI Fig. 9D). Overall,
there was no statistically significant difference between SPIO-
labeled and unlabeled hCNS-SCns (Fig. 2 J and K). In the SVZ,
9.3% (?1.3) of the cells were positive for both ?-III tubulin and
SC101. In the RMS, 9.4% (?1.6) cells were double-positive. Few
hCNS-SCns in the SVZ and RMS expressed GFAP (5.5% ? 1.3).
Migrating hCNS-SCns in the RMS also expressed doublecortin. At
18 weeks, 19.6% of the hCNS-SCns were in the OB. Of these, 17%
(?4.5) expressed ?-III tubulin and numerous cells were double-
labeled for both MAP-2 and SC121. There was no GFAP expres-
sion in the OB.
pronounced inflammatory response than unlabeled hCNS-SCns.
hCNS-SCns (open bars) analyzed 48 h after labeling. (C) The SPIO content is reduced by 50% with each division over 9 days. (D) Analysis of labeled and unlabeled
hCNS-SCns grown under differentiation conditions for 10 days revealed no difference in cell fate. Results are means ? SEM. (E–F) Electrophysiological cell
injection of 80 pA. (F) Voltage-activated sodium and potassium currents in a neural progenitor cells-derived neuron, steps of 35–65 mV.
In vitro analysis of SPIO-labeled hCNS-SCns. SPIO labeling was efficient and did not affect proliferation, differentiation and electrophysiology of
www.pnas.org?cgi?doi?10.1073?pnas.0608519104Guzman et al.
transplantation (data not shown). Based on cell morphology, most
of the Iba-1-positive cells were considered resting microglia.
Stroke Induces Targeted Migration of SPIO-Labeled Human Neural
Stem Cells in the Adult Rat Brain. Using a stroke and transplantation
paradigm (6), we show a targeted intraparenchymal migration of
hCNS-SCns in vivo, whereas no migration was observed in the
uninjured rodent brain. Seven days after a distal middle cerebral
artery occlusion, rats received three cortical grafts of 1 ? 105
hCNS-SCns each, medial to the cortical stroke. MRI 1 day before
transplantation ruled out signal changes that could have been
similar to the signal of transplanted hCNS-SCns. SPIO-labeled
hCNS-SCns were detected as strongly hypointense areas in the
cortex medial to the cortical stroke on T2-weighted images at 1
week (Fig. 3A). At 5 weeks, transparenchymal-targeted migration
was clearly detected on MRI extending from the lateral graft edge
toward the injured brain (Fig. 3C). This was confirmed histologi-
MRI was visualized by using 3D reconstructions with surface
rendering (Fig. 3 B, D, and E and SI Movies 1 and 2).
ischemic brain tissue on targeted stem cell migration by transplant-
ing hCNS-SCns in stroked rats adjacent to the lesion border or
without connection to corpus callosum. We found that in animals
where hCNS-SCns were transplanted far from the lesion there was
the bolus of stem cells was in continuity with the corpus callosum,
the stem cells underwent transcallosal migration despite greater
has been shown for stem cells transplanted in the hemisphere
contralateral to the stroke (10, 18). If transplantation was adjacent
to the lesion, we observed transparenchymal-targeted migration as
shown in Fig. 3. In animals displaying a targeted cell migration the
mean graft to lesion distance was 0.81 mm, whereas it was 3.48 mm
in animals without migration (SI Fig. 10G).
Total surviving hCNS-SCns in immunosuppressed stroked rats 5
weeks after transplantation was 51.3% (38.9–74.6%). This com-
pared favorably with the data in ref. 6, using the same cells and the
same lesion and same transplantation paradigm where we found
33.4% cell survival.
MRI Detects Clearance of Dead Human Neural Stem Cells. MRI
detection of graft rejection or death and clearance of the trans-
planted cells will be a critical aspect in the application of MRI for
future clinical studies. SPIO-labeled hCNS-SCns were killed by
repeated freeze–thaw cycles before transplantation into the left
striatum of immunosuppressed Sprague–Dawley rats. Viable cells
were transplanted on the contralateral side. Imaging 2 days after
transplantation showed bilateral hypointense areas, representing
the grafts (Fig. 4A). Whereas only minimal signal changes over 35
reduction in signal strength and graft volume on the side with dead
was 1.64 mm3and dropped by 52% over the 35 days of observation
(Fig. 4G). Fluorescence microscopy on sections stained for Iba-1
and the human-specific cytoplasmic marker SC121 demonstrated
significantly more activated Iba-1-positive cells in rats with dead
grafts as compared with rats with viable grafts (Fig. 4 D–F).
Tracking cells in vivo has traditionally been performed by using
extensive longitudinal studies where animals are killed at multiple
time points. This can be circumvented by using noninvasive meth-
ods for tracking cells in real time. Several papers have explored the
feasibility of tracking stem cells transplanted to the injured brain
and spinal cord (19, 20), including studies in stroke (10, 11, 21).
Although the ability to track various types of tagged stem cells by
the P0/P1 NOD-SCID mouse brain. (A and D) (A) Three weeks after transplantation, sagittal MRI shows hypointensities representing SPIO-labeled hCNS-SCns in
the human-specific cytoplasmic marker SC121. (Inset) Shows RMS in adjacent section. (B and E) (B) sagittal MRI 18 weeks after transplantation showing that
migration along the corpus callosum (arrowheads). (E) Corresponding histological section. (C and F) (C) sagittal MRI of a control animal transplanted with
unlabeled hCNS-SCns 18 weeks after transplantation shows no cell signal. (F) Corresponding histological section. (G–H) Higher magnification sagittal images
(areas boxed in Fig. 2E) show hCNS-SCns in the CA1, CA3, and dentate gyrus of the hippocampus (G) and in the OB (H) of the NOD-SCID mouse. (I) There was
a robust cell survival at 18 weeks after transplantation without statistically significant difference between SPIO-labeled (n ? 4 animals) and unlabeled cells (n ?
in anatomical subregions in NOD-SCID mice 18 weeks after intraventricular transplantation of hCNS-SCns (K). There was no statistically significant difference
between SPIO-labeled and -unlabeled cells in terms of cell fate. OB, olfacatory bulb; RMS, rostral migratory stream; SVZ, subventricular zone. Results are mean
Migration and integration of SPIO-labeled hCNS-SCns. MRI detects widespread migration of SPIO-labeled hCNS-SCns after intraventricular injection in
Guzman et al.
June 12, 2007 ?
vol. 104 ?
no. 24 ?
using MRI has been shown, there is a lack of information con-
cerning aspects of stem cell biology including cell proliferation,
differentiation, migration, and integration and electrophysiological
characteristics. Moreover, most of the studies have used rodent-
derived stem cells, and this significantly limits the conclusions that
can be made about human cells for use in clinical trials. Here, we
is not altered in vitro and in vivo, allowing us to monitor the fate of
these cells in vivo under a variety of transplant conditions over
extended periods of time (18 weeks). These findings may have
significant impact on the eventual translation of stem cell regen-
erative medicine to the clinic.
Multipotency is a hallmark of neural stem cells, and we have
shown that our hCNS-SCns can differentiate into neurons, astro-
cytes, and oligodendrocytes (15). Kostura et al. raised the concern
labeling of mesenchymal stem cells inhibits chondrogenesis but not
adipogenesis (12). Here, we found no statistically significant dif-
ference in the differentiation potential between SPIO-labeled and
unlabeled hCNS-SCns. In addition to cell fate analysis, we bring
evidence that magnetic labeling does not affect the electrophysio-
and classic sodium currents were not different between SPIO-
labeled and unlabeled hCNS-SCns. Depolarizing pulses elicited a
single action potential. We did not observe spontaneous spiking in
either group, which probably reflects the early maturation stage of
go through the same phases as neurons do during development in
of the cell membrane.
MRI of SPIO-Labeled Human Neural Stem Cells. A number of factors
affect the MRI detection threshold of SPIO-labeled cells, such as
the SPIO concentration per cell, the cell density once the cells
integrate into the host, and intrinsic MRI parameters such as field
strength, signal to noise ratio (SNR), pulse sequence, and acquisi-
tion parameters (23).
Long-term observation of SPIO-labeled stem cells might be
in vitro. Nevertheless, it has been shown that cells can be detected
for several weeks (10), and we were able to detect clusters of cells
by MRI up to 18 weeks. Heyn et al. (23) have shown that as little
as 1.4–3.0 pg of iron per cell is sufficient for detection with MRI.
If we assume an initial content of 20 pg Fe per cell (24) and
symmetrical cell division, detection would be possible up to six cell
divisions. We know from earlier studies that hCNS-SCns trans-
planted to the neonatal Nod-Scid brain do continue to proliferate
in neurogenic areas of the brain (14). One could hypothesize that
cells staying in the SVZ eventually loose their cytoplasmic SPIO.
SPIO over longer time period.
Reports and discussions on MRI cell detection have mainly
focused on the minimum number of SPIO-labeled cells detectable
with MRI. Although single-cell detection has been reported (25),
occlusion, with 3 ? 105cells in three deposits. (A and C) two consecutive coronal
stroke area in T2-weighted images 1 week (A) and 5 weeks (C) after transplan-
tation. There is some callosal (arrow) but no intraparenchymal migration at 1
week (A). At 5 weeks after transplantation, robust migration of hCNS-SCns
toward the lesion is visible as a hypointense area on the edge of the bolus
extending laterally (arrows) (C). Three-dimensional reconstruction and surface
rendering of the rat brain based on high resolution T2-MRI as illustrated in (E).
the graft (pink) and the stroked area (green). Note the broad migration of
hCNS-SCns along the medial border of the stroke in the anterior posterior and
craniocaudal direction, resulting in a significant increase in graft volume. (E)
Three-dimensional reconstruction of the rat brain illustrating the segmentation
(F) Histological section corresponding to the MRI (C) at 5 weeks stained with the
side (asterisk) and robust migration (arrows) toward the infarct.
planted with viable hCNS-SCns into the right striatum and dead hCNS-SCns
graft volume indicates graft/cell death. (A–C) MRI showing a left sided intra-
striatal graft of killed cells at days 2 (A), 7 (B), and 35 (C) after transplantation.
(D), showing number and morphology of Iba-1-positive cells (green) in close
relation to hCNS-SCns (red) in a viable graft. In the dead graft (E), morpho-
logically activated Iba-1-positive cells (green) are more abundant and have
phagocytosed hCNS-SCns (red) at 35 days after transplantation. (F) Iba-1
expression in healthy brain tissue (insert at higher magnification). Nuclei
counterstained with DAPI (blue). (G) Volumes of dead grafts (filled bar) and
viable grafts (open bar) based on MR images at 2, 7, and 35 days after
transplantation clearly demonstrates the early volume loss (*, P ? 0.05). (H)
viable grafts (VG), and in normal parenchyma (NP) at 35 days after transplan-
tation. Results are means ? SEM.
Graft undergoing necrosis. Immunosuppressed rats were trans-
www.pnas.org?cgi?doi?10.1073?pnas.0608519104Guzman et al.
most in vivo reports have found higher detection limits. We found
a detection limit of 1,000 cells in vitro, where the cells are highly
concentrated (SI Fig. 6E). However, once the cells start to migrate
and integrate in site-specific manner as observed in the OB and the
hippocampus, the density of cells is reduced considerably, leading
to a gradual loss of MRI cell signal. Dilution of SPIO during cell
division and reduction of cell density during migration and inte-
gration are the most important cellular factors determining the
detection thresholds of MRI. In contrast to the slow reduction in
cells as demonstrated in the last experiment (Fig. 4) results in a
signal loss in the first 35 days after transplantation. Because dead
SPIO-labeled cells result in early loss of MRI signal (Fig. 4) and
Iba-1/SC121 double-labeled cells are absent from the main migra-
that the observed MRI signal changes are only caused by SPIO
Early after transplantation, the intracellular SPIO causes a
susceptibility artifact most prominent when using gradient echo
imaging. Signal changes observed on gradient echo images at this
time point significantly overestimated the size of the transplants.
especially if volumetric analyses are required. Indeed, quantitative
analysis of graft size on T2-weighted images was comparable with
measurements on histological sections (SI Fig. 6). At later time
points however, when cell density decreases as cells integrate,
gradient echo imaging proved to be more reliable for cell detection
than T2-weighted spin echo imaging.
The Migration Pattern of hCNS-SCns Depends on the Microenviron-
ment. Whereas hCNS-SCns migrated along the RMS and the
corpus callosum and integrated into the OB and the hippocampus
MRI, allowed this process to be followed longitudinally.
In a previous study (14), we demonstrated widespread migration
of human cells when transplanted into the neonatal brain. The
extent and pattern of hCNS-SCns migration transplanted to the
(26), which migrate from the SVZ to the OB where they differen-
tiate into local interneurons (27). MRI showed that hCNS-SCns
migrate along the RMS leading them into the core of the OB at a
very early time point after transplantation. This may be due to
directed currents of CSF generated by ciliary beating of ependymal
cells leading to rapid distribution of transplanted hCNS-SCns (32).
This early migration pattern was followed by a radial migration
where hCNS-SCns adopted a laminar organization in the granular
and periglomerular layers of the OB. It was striking that in the
olfactory bulb and the hippocampus the human cells displayed
remarkable regional specificity and also adopted cell body diame-
ters of the local host cells as described in ref. 17.
In contrast to the immature brain, transplanted hCNS-SCns did
not show spontaneous migration in the uninjured adult brain.
However, if an injury occurs, transplanted hCNS-SCns undergo
targeted migration toward stroke (as described in ref. 6). Stem cells
have been shown to cross from the contralateral hemisphere via a
sequential in vivo MR imaging of ipsilateral transcortical-targeted
migration of human stem cells with correlative histology. In con-
20), we found that the extent of transcortical migration was
determined by the distance between the graft site and the lesion. If
the distance between the lesion and the graft site in the rat brain
exceeded 1 mm, we did not find targeted migration. This contrasts
brain was observed (10, 11). These other studies, however, were
done by using rodent-derived stem cells. There are fundamental
differences between cells lines used in our and others’ studies, such
as migratory behavior, the degree of maturation, and the species
from which the cells are derived. Hence, comparison between
to control for post transplantation graft location in relation to an
Analysis of migration and cell fate revealed equivalent biological
properties of SPIO-labeled and unlabeled hCNS-SCns. It has been
speculated that release of iron oxide particles from dying cells into
the host parenchyma could lead to a prolonged inflammatory
expression between animals receiving SPIO-labeled hCNS-SCns as
compared with nonlabeled cells.
cell biology after magnetic labeling and transplant to the rodent
brain. We have shown that human neural stem cells labeled with
SPIO survive in large numbers, are able to differentiate into
neuronal and glial lineages, have neuronal-like electrophysiological
characteristics, and appropriately respond to microenvironmental
cues after transplantation to the rodent brain. Using MRI and
histology, we show that upon transplantation into the immature
rodent brain, SPIO-labeled hCNS-SCns migrate and integrate in a
manner appropriate for their location. Conversely, in the mature
brain, targeted migration is observed only when an injury occurs.
Using a cortical stroke model, we demonstrate that MRI is able to
finally, we demonstrate that death of transplanted cells can be
monitored by using in vivo MRI.
Neurotransplantation holds great promise for the treatment of
acute and chronic central nervous system disease. In vivo cellular
imaging will eventually serve multiple purposes, including postop-
erative visualization of graft location, tracking cell migration and
integration, and monitoring graft survival.
Human CNS Stem Cells Derived Neurosphere Cells. hCNS-SCns
(StemCells, Inc., Palo Alto, CA) were isolated by flow cytometry
from fetal brain tissue (16–20 wk) as described in refs. 13 and 14.
hCNS-SCns cells were plated at 105cells per ml in human neuro-
sphere culture media (X-VIVO 15 medium; BioWhittaker, Walk-
ersville, MD), N-2 supplement (GIBCO, Carlsbad, CA), and 0.2
ng/ml), EGF (20 ng/ml), and leukemia inhibitory factor (10 ng/ml).
On the day of transplantation, cells were prepared at a density of
1 ? 105cells per ?l.
Magnetic Labeling of Neural Stem Cells. SPIO particles (Feridex IV;
Pharmaceuticals Partner, Schaumburg, IL) (5 ?g/2.5 ?g/ml) were
incubated for 30min at 37° in culture medium before being added
and stained with an anti-dextran antibody (DX-1, StemCell Tech-
by using a particle analyzer function (ImageJ software). For cell
differentiation assay, SPIO-labeled and unlabeled cells were grown
in ex vivo base without mitogens and supplemented with BDNF,
GDNF, and laminin for 10 days. Immunofluorescence staining for
Nestin, ?-tubulin, and GFAP was done, and the percent of immu-
noreactive cells was counted.
Whole-Cell Patch Clamp Experiments. Whole-cell recordings were
made essentially as described in ref. 28. Morphologically, neuronal
appearing cells (phase-bright somata with two to five primary
processes) present in each condition (SPIO-labeled and unlabeled
For evoking sodium currents, cells were held at ?70 mV in voltage
Guzman et al.
June 12, 2007 ?
vol. 104 ?
no. 24 ?
clamp and subjected to depolarizing step functions. For evoking, Download full-text
action potentials cells were maintained at ?70mV in current clamp
and stimulated with a depolarizing current injection.
MRI. High-resolution magnetic resonance images were acquired on
a 4.7T/40 cm animal scanner system (Varian, Palo Alto, CA). The
imaging protocol consisted of scout imaging in two planes followed
by a spin-echo sequence (TR ? 2500 msec, TE ? 45 msec, NEX ?
4, matrix ? 256 ? 256, FOV ? 40 mm, slice thickness ? 1 mm,
NEX ? 4, matrix 128 ? 128 ? 64, FOV 3 ? 3 ? 3 cm resulting in
a voxel size of 234 ? 234 ? 469 ?m). For in vitro studies cells were
injected into agarose gel phantoms. During imaging animals were
under isoflurane anesthesia and monitored for respiration and
heart rate, the temperature was kept at 37°C, using a feedback
Cell Transplantation in Neonatal NOD-SCID Mice. All animal proce-
dures were approved by the Stanford University Administrative
Panel on Laboratory Animal Care. Neonatal mice (postnatal day
5–10 min followed by stereotactic injection of either SPIO-labeled
(2 ?l of cells at 1 ? 105cells per ?l). MRI was performed at 3, 9,
was tested histologically for each time point.
Distal Middle Cerebral Artery Occlusion (dMCAo). Adult male
Sprague–Dawley rats (n ? 10) were anesthetized with isoflurane
and the left common carotid artery temporarily occluded for 1 h.
The distal middle cerebral artery was exposed, cauterized, and cut
just above the rhinal fissure.
Cell transplantation in rat after dMCAO. Seven days after dMCAO,
rats were given 1.0-?l deposits of suspended cells (1 ? 105cells per
?l), using a micropump in three cortical regions in the anterior-
posterior axis (SI Fig. 12). Rats were immunosuppressed by daily
i.p. injection of cyclosporine A (10 mg/kg). Five weeks posttrans-
plantation, rats were transcardially perfused with 4% phosphate-
buffered paraformaldehyde and processed for histology.
Histology and Immunohistochemistry. For Prussian blue staining,
mounted sections were washed and incubated for 20 min in 10%
potassium ferrocyanide and 20% hydrochloric acid. Triple-label
immunofluorescence staining was carried out on free-floating
sections. Sections were incubated overnight at 4°C with either the
human-specific nuclear (SC101, 1:1,000) or cytoplasmic marker
(SC121 1:3,000, both provided by StemCells, Inc.) and one of the
Research Products, Berkeley, CA), guinea-pig anti-GFAP (1:500;
Advanced ImmunoChemicals, Long Beach, CA), goat anti-
doublecortin (Dcx, 1:100; Santa Cruz Biotechnology, Santa Cruz,
CA), human-specific mouse anti-GFAP (1:3,000; provided by
StemCells Inc.), mouse anti-MAP2 (1:3,000, Abcam, Cambridge,
MA), rat anti-Iba-1 (ionized calcium binding adapter molecule 1,
used were Alexa Fluor 488 and 546, (1:1,000; Molecular Probes,
Eugene, OR) or Cy3 (1:2,000; Jackson ImmunoResearch
Louis, MO) was used to label nuclei. For total counting of SC101-
positive cells, a biotinylated anti-mouse secondary antibody (1:250;
Vector Laboratories, Burlingame, CA), followed by streptavidin-
HRP and DAB enhancement were used.
Microscopical Analysis. Total numbers of transplanted human cells
stained with SC101 were counted by using unbiased computational
stereology (Stereoinvestigator software, MicroBrightfield, Brattle-
boro, VT). For cell fate analysis, the proportion of SC101-labeled
cells that also stained with lineage-specific phenotype markers was
determined by confocal microscopy. Split-panel and z axis analyses
were used for all counting. Fifty or more SC101-positive cells were
scored for ?-tubulin and GFAP in each anatomical area (SVZ,
RMS, and OB) per animal.
Statistics. The statistic software package PRISM (GraphPad, San
Diego, CA) was used for data analysis. All means are presented ?
SEM. Statistical differences were assessed either by a nonparamet-
ric Mann–Whitney U test or by a one-way Kruskal–Wallis test. The
level of significance was set at P ? 0.05.
We thank Dr. Bruce Schaar for critical reading of the manuscript and Beth
National Institutes of Health National Institute of Neurological Disorders
Siegelman, the William Randolph Hearst Foundation, Bernard and Ronni
Lacroute (to G.K.S.), and Swiss National Science Foundation Grants
PBBEB-104450 and SSMBS-1194/PASMA-108940/1 (R.G.).
1. Mendez I, Sanchez-Pernaute R, Cooper O, Vin ˜uela A, Ferrari D, Bjo ¨rklund L,
Dagher A, Isacson O (2005) Brain 128:1498–1510.
2. Lindvall O, Brundin P, Widner H, Rehncrona S, Gustavii B, Frackowiak R,
Leenders KL, Sawle G, Rothwell JC, Marsden CD (1990) Science 247:574–577.
3. Olanow CW, Goetz CG, Kordower JH, Stoessl AJ, Sossi V, Brin MF, Shannon
KM, Nauert GM, Perl DP, Godbold J, Freeman TB (2003) Ann Neurol
4. Freeman TB, Cicchetti F, Hauser RA, Deacon TW, Li XJ, Hersch SM, Nauert
GM, Sanberg PR, Kordower JH, Saporta S, Isacson O (2000) Proc Natl Acad
Sci USA 97:13877–13882.
5. Kondziolka D, Steinberg GK, Wechsler L, Meltzer CC, Elder E, Gebel J,
Decesare S, Jovin T, Zafonte R, Lebowitz J et al. (2005) J Neurosurg 103:38–45.
6. Kelly S, Bliss TM, Shah AK, Sun GH, Ma M, Foo WC, Masel J, Yenari MA,
Weissman IL, Uchida N et al. (2004) Proc Natl Acad Sci USA 101:11839–11844.
8. Norman AB, Thomas SR, Pratt RG, Lu SY, Norgren RB (1992) Brain Res
9. Hawrylak N, Ghosh P, Broadus J, Schlueter C, Greenough WT, Lauterbur PC
(1993) Exp Neurol 121:181–192.
10. Hoehn M, Kustermann E, Blunk J, Wiedermann D, Trapp T, Wecker S,
Focking M, Arnold H, Hescheler J, Fleischmann BK, et al. (2002) PNAS
11. Modo M, Mellodew K, Cash D, Fraser SE, Meade TJ, Price J, Williams SCR
(2004) NeuroImage 21:311–317.
12. Kostura L, Kraitchman DL, Mackay AM, Pittenger MF, Bulte JW (2004) NMR
13. Tamaki S, Eckert K, He D, Sutton R, Doshe M, Jain G, Tushinski R, Reitsma
M, Harris B, Tsukamoto A, et al. (2002) J Neurosci Res 69:976–986.
14. Uchida N, Buck DW, He D, Reitsma MJ, Masek M, Phan TV, Tsukamoto AS,
Gage FH, Weissman IL (2000) Proc Natl Acad Sci USA 97:14720–14725.
15. Cummings BJ, Uchida N, Tamaki SJ, Salazar DL, Hooshmand M, Summers R,
Gage FH, Anderson AJ (2005) Proc Natl Acad Sci USA 102:14069–14074.
16. Claiborne BJ, Amaral DG, Cowan WM (1990) J Comp Neurol 302:206–219.
17. Muotri AR, Nakashima K, Toni N, Sandler VM, Gage FH (2005) Proc Natl
Acad Sci USA 102:18644–18648.
18. Modo M, Hoehn M, Bulte JW (2005) Mol Imaging 4:143–164.
19. Bulte JW, Zhang S, van Gelderen P, Herynek V, Jordan EK, Duncan ID, Frank
JA (1999) Proc Natl Acad Sci USA 96:15256–15261.
20. Lee IH, Bulte JW, Schweinhardt P, Douglas T, Trifunovski A, Hofstetter C,
Olson L, Spenger C (2004) Exp Neurol 187:509–516.
21. Modo M, Cash D, Mellodew K, Williams SCR, Fraser SE, Meade TJ, Price J,
Hodges H (2002) NeuroImage 17:803–811.
22. Westerlund U, Moe MC, Varghese M, Berg-Johnsen J, Ohlsson M, Langmoen
IA, Svensson M (2003) Exp Cell Res 289:378–383.
23. Heyn C, Bowen CV, Rutt BK, Foster PJ (2005) Magn Reson Med 53:312–320.
24. Frank JA, Miller BR, Arbab AS, Zywicke HA, Jordan EK, Lewis BK, Bryant
LH, Jr, Bulte JW (2003) Radiology 228:480–487.
25. Shapiro EM, Skrtic S, Sharer K, Hill JM, Dunbar CE, Koretsky AP (2004) Proc
Natl Acad Sci USA 101:10901–10906.
26. Suhonen JO, Peterson DA, Ray J, Gage FH (1996) Nature 383:624–627.
27. Lois C, Alvarez-Buylla A (1994) Science 264(5162):1145–1148.
28. Stellwagen D, Beattie EC, Seo JY, Malenka RC (2005) J Neurosci 25:3219–
www.pnas.org?cgi?doi?10.1073?pnas.0608519104Guzman et al.