Gene Therapy (2009) 16, 734–745
& 2009 Macmillan Publishers Limited All rights reserved 0969-7128/09 $32.00
www.nature.com/gt
ORIGINAL ARTICLE
Discrepancies between the fate of myoblast xenograft
in mouse leg muscle and NMR label persistency after
loading with Gd-DTPA or SPIOs
C Baligand1,2,3,4,9, K Vauchez3,4,5,6,7,8,9, M Fiszman3,4,5,6,7, J-T Vilquin3,4,5,6,7 and PG Carlier1,2,3,4
1
Institute of Myology, NMR laboratory, Paris, France; 2CEA, I2BM, MIRCen, IdM NMR Laboratory, Paris, France; 3UPMC Univ Paris
06, Paris, France; 4IFR14, Pitie Salpetriere Hospital, Paris, France; 5INSERM UMR S974, Paris, France; 6CNRS UMR 7215, Paris,
France; 7Institute of Myology, Paris, France and 8Genzyme SAS, Paris, France
1
H-NMR (nuclear magnetic resonance) imaging is regularly
proposed to non-invasively monitor cell therapy protocols.
Prior to transplantation, cells must be loaded with an NMR
contrast agent (CA). Most studies performed so far make use
of superparamagnetic iron oxide particles (SPIOs), mainly for
favorable detection sensitivity. However, in the case of
labeled cell death, SPIO recapture by inflammatory cells
might introduce severe bias. We investigated whether NMR
signal changes induced by preloading with SPIOs or the low
molecular weight gadolinium (Gd)-DTPA accurately monitored the outcome of transplanted cells in a murine model of
acute immunologic rejection. CA-loaded human myoblasts
were grafted in the tibialis anterior of C57BL/6 mice. NMR
imaging was repeated regularly until 3 months post-trans-
plantation. Label outcome was evaluated by the size of the
labeled area and its relative contrast to surrounding tissue. In
parallel, immunohistochemistry assessed the presence of
human cells. Data analysis revealed that CA-induced signal
changes did not strictly reflect the graft status. Gd-DTPA label
disappeared rapidly yet with a 2-week delay compared with
immunohistochemical evaluation. More problematically, SPIO
label was still visible after 3 months, grossly overestimating
cell survival (o1 week). SPIOs should be used with extreme
caution to evaluate the presence of grafted cells in vivo and
could hardly be recommended for the long-term monitoring of
cell transplantation protocols.
Gene Therapy (2009) 16, 734–745; doi:10.1038/gt.2009.12;
published online 12 March 2009
Keywords: magnetic resonance imaging; cell transplantation; gadolinium; SPIO; muscle
Introduction
Cell therapy has progressively become an important
modality of regenerative medicine. Through the use of
now classical cell types, such as hematopoietic stem cells,
skin or cartilage progenitors, pancreatic islets, cell
therapy has entered the therapeutic armamentarium for
dedicated diseases, and should be completed in the
future by new emerging adult, embryonic or induced
pluripotent stem cells. In the specific clinical field of
muscular disorders, clinical and social expectations are
high. Autologous approaches are being developed when
localized repairs are mandated, whereas heterologous
approaches are used in the context of generalized
advanced degenerative diseases.1 Whatever be the
immunological situation, however, biological, conceptual
and technical set-up of strategies have been hampered by
several issues, and particularly in large organisms.2
Several major obstacles remain to be abated: the amount
of tissues to be treated may represent dozens of
kilograms; immediate cell death post-injection may
Correspondence: Dr PG Carlier, Laboratoire de RMN AIM-CEA,
Institut de Myologie––Bat. Babinski, CHU Pitié–Salpêtrière, 83, Bd
de l’Hôpital, 75651 Paris 13, France.
E-mail: [email protected]
9
These authors contributed equally to this work.
Received 14 August 2008; accepted 5 October 2008; published online
12 March 2009
exceed 80%; grafted cell migration and colonization of
target muscles are extremely limited. Nevertheless,
transplantation of myoblasts has been considered as a
possible therapeutic approach for several pathologies
such as Duchenne muscular dystrophy,3 heart failure4
and urinary incontinence.5 Moreover, myoblasts are not
the only candidates for muscular cell therapy approaches. Mesoangioblasts,6 pericytes7 or CD133-positive
cells8 have also been proposed. The efficacy of these
different approaches has to be improved and so requires
the development of follow-up and evaluation tools.
When new ideas and possible technical improvements
are to be tested, there is a great need for non-invasive
imaging techniques that would allow the mapping of
therapeutic cell distribution in the organs, the detection
of apoptotic and necrotic cells, the measurement of cell
migration and fusion with muscle fibers, the determination of cell proliferation and survival.2 All this information would be of invaluable help to select and optimize
procedures, without having to kill multiple groups of
animals or having to perform multiple biopsies on the
same individuals over a period of time to determine
event kinetics.
In this context, 1H-NMR (nuclear magnetic resonance)
imaging might have a role to play to evaluate the gross
response of muscles to cell therapy: is there a gain in
muscle mass? Are fibrosis and fatty degeneration volume
fractions decreased? With current technologies, the
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
amount of therapeutic cells transferred into skeletal
muscle represents a very small percentage of the
recipient cell population. This implies that imaging
techniques used to monitor organ response must be of
high resolution and high precision, which NMR imaging
can offer today. Although conventional histological and
physiological analyses have demonstrated myoblast
transfer efficiency,9–12 there have so far been no NMR
studies reporting changes in muscle mass or composition
after myoblast transplantation.
In most instances, one might prefer to directly visualize
the grafted cells in vivo. With NMR as with all other
available imaging techniques, there is no intrinsic contrast
between transplanted cells and host tissues when they are
of identical nature, as in this case, with myoblasts among
muscle fibers (J-TV, unpublished data). To distinguish the
transplanted cells in the target organ, they need to be
labeled prior to injection. In nuclear medicine, they are
typically labeled with 111In-oxime or 99Tcm.13–15
Contrast agents (CAs) are available for NMR imaging.
Two main classes of CAs are commonly used for clinical
diagnostic MRI: the lanthanide-based paramagnetic
complexes and the superparamagnetic iron oxide particles (SPIOs). They differ from each other not only by the
nature and mechanisms of the contrast generated but
also by their molecular size and biodistribution.16 Both
classes were developed for i.v. injection but, whereas
lanthanide chelates distribute exclusively in the extracellular space in normal conditions, the high molecular
weight SPIOs are rapidly taken up by reticulo-histiocytic
cells and have been used to track these specific cell types
in vivo, including macrophages, for many years.17 They
were subsequently successfully tested for in vitro cell
loading prior to in vivo injection.18
Although internalization pathways and long-term
intracellular behavior of these molecules are not yet
fully understood, it is well established that intracellular
loading with these CAs is possible without significant
impact on cell viability or function.19,20 Accurate evaluation and monitoring of cell therapy protocols require a
very sensitive technique, because of the small amount of
injected therapeutic cells. For these reasons, SPIOs,
which have very high T2 (and T2*) relaxivities, have
been privileged over the past 10 years. The high
susceptibility difference between CA and surrounding
tissue can be exploited to generate a negative contrast,
and several studies have indeed demonstrated that SPIO
labeling allows the detection of grafted cells in vivo with
high sensitivity.21,22
A serious issue is the particulate nature of SPIOs,
which make them a target for phagocytosis by activated
reticulo-histiocytic cells. SPIOs have sizes ranging from
120 to 180 nm. If they are released from damaged or
dying cells, there is a high probability that they will be
rapidly phagocytosed by macrophages. This would result
in false-positive images, indicating the persistence of
grafted cells, whereas it is really inflammatory cells that
are detected instead. This problem was recognized and
nicely documented by Cahill et al.22 But the caveat seems,
however, to have been ignored by most investigators so far.
Recently, gadolinium (Gd) chelates have also been
investigated as CA for cell labeling.20,23,24 Grafted on
Matrigel structures and subcutaneously implanted,
endothelial progenitor cells labeled with Gd-HPDO3A
were identified as hyperintensities 14 days after
injection.20 Plugged in mouse kidney, they were visible
24 h after injection. The authors demonstrated the
possibility of using Gd chelates for cell monitoring, at
least in tissues with low intrinsic NMR signal such as
kidney and suggested that it might be tested in muscle.
Indeed, the figures of CA relaxivities are more in favor of
SPIOs than Gd chelates, with a potential negative impact
on detection sensitivity. However quantitation issues are
likely to be less problematic because Gd chelate contrast
is based mainly on T1 changes, which are much more
proportional to CA concentrations and hence easier to
treat analytically.
Very interestingly, after i.v. injection, low molecular
weight Gd chelates are rapidly eliminated through
glomerular filtration, with a blood half-life shorter than
30 min.25,26 Similarly, lanthanide-based CAs, which are
released in the interstitium in the case of labeled cell
death, also ought to be rapidly cleared by the kidney.
With this class of CA, label recapture by other cells can
be expected to be less of a problem than with SPIOs.
The aim of this study was to test this hypothesis in
skeletal muscle, and to investigate and compare the
ability of Gd-DTPA and SPIOs to accurately report the
fate of labeled human myoblasts in the case of high cell
death rate after transplantation. We carried out cell
labeling and transplant experiments with two molecules
in parallel, one representative of each class of CA. Wildtype mice were used as transplant recipients of human
cells to take advantage of the massive and rapid
immunorejection induced and the very high and rapid
death rates of transplanted cells.27 Immunohistochemical
analyses were confronted with the NMR measurements
and provided a reference frame for correct interpretation
of the in vivo imaging data.
735
Results
In vitro assessment of loading conditions
We determined the effective concentration ranges for
each of the CAs.
First, their potential cellular toxicity was evaluated in
vitro. Gd-DTPA and SPIOs were incubated for 24 h with
the human myoblasts in the concentration ranges of 10–
80 mM and of 10–1000 mg ml1, respectively. Nonspecific
endocytosis and pinocytosis mechanisms warrant penetration of these agents into the cells, which are
concentrated in endosomes and lysosomes, and stored
in the cytosol.19,28 Immediate cell death and induction of
apoptosis were quantified by flow cytometric evaluation
of propidium iodide uptake and annexin V surface
expression, respectively. As compared with unloaded
cells, the increase in death rate was statistically significant for the highest concentrations only (60 and
80 mM of Gd-DTPA; 700 and 1000 mg ml1 of SPIOs;
Po0.05, n ¼ 4). The apoptosis rate tended to increase
with concentrations without being significant below
80 mM of Gd-DTPA (Po0.05, n ¼ 4).
Second, the proliferation characteristics of myogenic
cells were not modified at the lowest concentrations
mentioned above, but the rate tended to decrease at
highest concentrations. In this experimental situation, the
calculation of cell proliferation might be partly biased by
the increased ratios of death and apoptosis.
Gene Therapy
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
736
Third, flow cytometry quantitation of cell population
distribution indicated that phenotypes were not altered by
the CAs. Indeed, the percentage of CD56+ cells remained
high (more than 78%) throughout cell expansions.
Finally, the differentiation capacity of loaded myogenic cells was assessed by the formation of myotubes in
differentiation conditions, and calculation of fusion
indexes. In comparison with unloaded cells, the fusion
indexes obtained for all the Gd-DTPA or SPIO loading
doses were not statistically different. Indeed, direct
loading of both CAs did not alter the differentiation
ability of myoblasts.
On the basis of these results, we chose 25 and 40 mM of
Gd-DTPA and 100–500 mg ml1 of SPIOs and performed a
preliminary feasibility assay in vivo. Its aim was to
determine the concentration of each CA that would
provide sufficient contrast for NMR detection while
labeling comparable areas at injection sites. Loaded human
cells were injected into mouse muscles and NMR images
were obtained and analyzed on the same day. We selected
the concentrations of 25 mM Gd-DTPA and 100 mg ml1
SPIOs for further in vivo experiments (see Table 1).
The ranges of CA concentrations, determined to be
safe in vitro, then in vivo in this study, have also been
proven non-toxic and effective in other studies using
SPIOs21,28,29 or Gd chelate.20
NMR imaging characterization of the injection rejection
paradigm in mouse muscle
Contrast in NMR images is governed by various
endogenous processes such as inflammation, on top of
which CA relaxivity properties are further added. We
first showed that the lesion caused by the surgical act
induced only slight and transient modifications of
muscle NMR signal. Mouse leg muscles were injected
with a phosphate-buffered saline (PBS) solution that was
visible only as hyperintensities on T2 maps, and had
totally disappeared at D6 (data not shown).
The monitoring of C57BL/6 mice legs injected with
unlabeled human myoblasts demonstrated no visible
changes on sectional T1-weighted images and T1 map.
T2 maps, however, revealed the edema caused by the
rejection process: we observed that T2 was increased by
19±5.5 ms at the injection site as compared with normal
tissue at D4 (Figure 1) and progressively decreased
within the next 3 weeks.
NMRI follow-up of grafted SPIO-labeled myoblasts and
Gd-DTPA-labeled myoblasts in the mouse leg muscle
Taking advantage of the rejection process of human
myoblasts grafted into the leg muscle of immunocompetent mice, we tested the ability of SPIO and Gd-DTPA cell
loading to monitor the graft status. The NMR scans of each
animal were carefully inspected at D0. As expected at the
selected concentrations, the legs that received Gd-DTPAloaded cells showed a hyperintense spot at the injection
site on T1-weighted images. All those treated with SPIO
label presented a hypointense area in the injected muscle.
A few of the observed signal voids, evidently much larger
than the injection site and probably resulting from an
inhomogeneous distribution of the graft during the
injection, were excluded from the study.
Table 1 Viability, apoptosis, phenotype, proliferation and myogenic differentiation of human myoblasts loaded with or without contrast
agents at the selected concentration rates of 25 mM of Gd-DTPA and 100 mg ml1 of SPIOs
SPIO (100 mg ml1)
Gd-DTPA (25 mM)
Viability assay (%, n ¼ 4)
Apoptosis assay (%, n ¼ 4)
CD56 phenotype (%, n ¼ 4)
Proliferation assay ( 103 cells, n ¼ 3)
D3
D6
Fusion Index (%, n ¼ 3)
Loaded cells
Control
Loaded cells
Control
94±4
5±2
97±2
98±0.5
4±1
97±2
84±8
9±4
83±10
89±5
9±5
87±7
51±19
186±82
70±27
263±57
62±17
167±43
66±15
204±58
45±8
47±5
54±16
55±8
Abbreviations: Gd, gadolinium; SPIO, superparamagnetic iron oxide particle.
a
b
c
Figure 1 Nuclear magnetic resonance (NMR) imaging of a C57BL/6 mouse leg 4 days after injection with unloaded human myoblasts.
T1-weighted imaging (a) or T1 mapping (c) did not show any change all along the experiment. T2 maps (b) indicated the inflammation site
(white arrow), which covered the widest area on D4 and tended to disappear in 3 weeks. This phenomenon had to be taken into account
when analyzing images of labeled human myoblasts injected into muscles.
Gene Therapy
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
Among the set of T1-weighted transversal images
obtained on D0 for each animal leg, we chose a
representative slice and monitored the time course of
the graft signal at each time point of the experiment: D1,
D4, D6, D8, D11, D14, D21 and 3 months posttransplantation (illustrated on Figures 2b and c). We
performed a quantitative analysis of the T1-weighted
a
Immunostaining
b
Gd-DTPA
images by monitoring the following three parameters:
the size N of the labeled area, the contrast C between
surrounding muscle and labeled area, and finally the
product C N, which accounts for the two phenomena.
The size N of the labeled area was measured on T1weighted images and normalized to the value at D0
(Figure 3a). For Gd-DTPA label, we noted a linear
c
SPIO
d
737
Prussian Blue
D0
D6
D 14
D 95
Figure 2 Parallel follow-up of injected cells by nuclear magnetic resonance (NMR) imaging and histology. Immunochemistry (a),
T1-weighted NMR imaging (b, c) and Prussian blue staining (d) were performed at D0, D6, D14 and D95 after transplantation in tibialis
anterior of C57BL/6 mice. (a) Lamin A/C staining (red) identified the nuclear envelope of human cells at the injection site. All nuclei were
labeled by 4,6-diamidino-2-phenylindole (DAPI) (blue), whereas laminin (green) stained the muscle and vessel basal lamina (magnification:
16). (b) The T1 shortening effect of gadolinium (Gd)-DTPA provided hyperintense spot on T1-weighted transversal images. This contrast
tended to disappear over the succeeding weeks and was not visible anymore at D95. (c) Superparamagnetic iron oxide (SPIO) label remained
visible after 3 months as a hypointensity on T1-weighted images. (d) Prussian blue staining identified iron particles (blue) in muscle
transplanted with SPIO-loaded myoblasts. Fast Red stains cytoplasm in pink and nuclei in red (magnification: 20).
Gene Therapy
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
738
N normalized to D0
Follow-up of the size
2.0
1.5
∗ ∗∗ ∗ ∗ ∗
∗
∗
1.0
0.5
0.0
0
10
20
30
90
100
Follow-up of the contrast
C normalized to D0
2.0
1.5
∗ ∗∗ ∗ ∗ ∗
∗
∗
1.0
0.5
0.0
0
10
20
30
90
100
CxN normalized to D0
Follow-up of the sizexcontrast product
2.0
1.5
∗ ∗∗ ∗ ∗ ∗
∗
∗
1.0
0.5
0.0
0
10
20
30
Time (days)
90
100
Figure 3 In vivo quantitative follow-up of the graft by nuclear
magnetic resonance (NMR) imaging. (a) The size N of the
gadolinium (Gd)-DTPA- and superparamagnetic iron oxide
(SPIO)-labeled area was determined as the number of pixels with
an intensity outside the 95% confidence interval for normal muscle
signal. (&: Gd-DTPA; ~: SPIO). (b) The relative contrast C was
defined as the signal intensity variation rate between the labeled
area and the normal muscle tissue. Size and contrast of Gd-DTPA
labeled area decreased linearly (linear regression curves are
represented by the dotted lines). (c) The product N C identified
both Gd-DTPA and SPIO labels on T1-weighted images. Their time
courses were significantly different for each parameter reported, as
demonstrated by repeated measures analysis of variance (ANOVA),
and Bonferroni comparison of means (*Pp0.05).
decrease (N ¼ 0.07+0.95 day; R2 ¼ 0.98) between D0 and
D14 and a total disappearance on D21. The size of the
SPIO-labeled area showed an initial increase of 23%
between D0 and D1, it then slowly diminished to 50% of
its initial size at D14 and later remained stable over the
rest of the experiment. The time to half label size was
t1/2N ¼ 39.5±42 days for SPIO and t1/2N ¼ 5.7±3 days for
Gd-DTPA.
The contrast between SPIO label and muscle was very
stable over the 3 months of the experiment, whereas the
contrast between Gd-DTPA label and muscle decreased
and totally disappeared within 3 weeks (Figure 3b).
Gene Therapy
Right from D1, the contrast between Gd-DTPA label and
muscle was only +58% of its value at D0. It then linearly
decreased until D21 (C ¼ 0.05+0.83 day; R2 ¼ 0.89).
The time to 50% decrease of the initial contrast value
was t1/2C ¼ 85±19 days for SPIO and t1/2C ¼ 5±4 days
for Gd-DTPA.
Finally, we also assessed the possibility of monitoring
the graft by the product C N for Gd-DTPA and SPIO
labels (Figure 3c), and the time to half the initial value
was t1/2CN ¼ 11.2±25 days for SPIO and t1/2CN ¼ 1.6±1.2
days for Gd-DTPA.
Because of the strong T2 relaxivity of SPIOs, T2
mapping systematically showed a dark spot where SPIOlabeled cells were grafted, whereas the milder T2
shortening due to the presence of Gd-DTPA was visible
only on D1. The increase in T2 induced by the local
inflammatory response competed with Gd-DTPA-induced T2 relaxation in such a way that the latter was
masked, as confirmed by measurements on unlabeled
cell injection sites (Figure 1) and computer simulations.
On the other hand, Gd-DTPA T1 shortening effect was
visible on T1 maps in spite of the inflammation. We
measured the difference over time in longitudinal
relaxation rate r1 ¼ 1/T1 of the labeled area relatively to
the normal tissue, which decreased from Dr1 ¼ 0.47 s–1 on
D1 to Dr1 ¼ 0.21 s–1 on D21 (over an area reduced to 38%
of its initial size). Unlike T2 variations, T1 variations of
unloaded cells were below our detection limit and
consequently did not affect T1 measurements of GdDTPA-loaded cells.
Finally and most importantly, the time courses of all
the parameters that we studied were significantly
different between SPIO- and Gd-DTPA-labeled legs
(Pp0.05) and showed that Gd-DTPA label disappearance
had occurred within 3 weeks after the injection of human
myoblasts in host tissue, whereas SPIO label had reached
a plateau by then and was still visible after 3 months.
The fate of human myoblasts transplanted into
immunocompetent mouse muscle
After NMR imaging, batches of animals were immediately killed. Their muscles were processed to identify
total nuclei (4,6-diamidino-2-phenylindole (DAPI) staining), the laminin delineating the endomysial and the
muscle fiber compartments, and the human-specific
nuclear lamin A/C (Figure 2a). The latter was selected
to identify human nuclei on fusion with mouse muscle
fibers, and to witness the formation of hybrid host–donor
structures, at the time when human cellular elements are
diluted within the large mouse muscle cytoplasm. SPIOs,
if any, were detected in blue with Prussian blue staining
(Figure 2d).
On the injection day, the presence of human cells was
confirmed by the detection of numerous nuclei (DAPI)
colocalized with human lamin A/C staining. These
pockets of cells were located in the endomysium, as
delineated by laminin staining. On serial sections,
Prussian blue staining confirmed the presence of
numerous ferrite nanoparticles inside the injected cells.
At this early time, most nuclei were of human origin, and
the inflammation process was not extensive yet.
On days 1 and 2, inflammatory cells, identified by
DAPI staining and devoid of human-specific lamin A/C
staining, extensively invaded the injection sites. Already,
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
some rare figures of fusion between donor human
myoblasts and recipient mouse muscle fibers were
witnessed by the presence of lamin A/C-positive nuclei
located beneath the laminin-delineated basement membranes of muscle fibers (Figure 4a–f).
From days 2–6, human cells were progressively
eliminated and scavenged from the injection site by
inflammatory cells, as demonstrated by the decrease in
the number of lamin A/C-positive nuclei (Figure 2a).
Inside mouse muscle fibers, some lamin A/C-positive
a
b
c
d
e
f
g
h
739
Figure 4 Evidence of human nuclei incorporation into host mouse muscular fiber. Immunostainings were performed at D1 (a, b), D2 (c, d),
D4 (g) and D6 (e, f, h) after transplantation of human loaded myoblasts in tibialis anterior of C57BL/6 mice. (b, d, f) Inset magnifications of
(a, c, e). White arrowheads indicate the human lamin A/C-positive nuclei (red), colocalized with 4,6-diamidino-2-phenylindole (DAPI) (blue)
and positioned in muscle fibers, under the basal lamina (green). Prussian blue staining highlights SPIO (black arrows) included in muscle
fibers (delineated by black arrowheads). Magnification: (a, c, e), 63; (b, d, f), 100; (g, h), 150.
Gene Therapy
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
740
nuclei were still observed. Of note, Prussian blue staining
illustrated the presence of ferrite nanoparticles in the
endomysium, and in some rare mouse muscle fibers
(Figure 4g–h).
As expected,30 the complete rejection of human
myoblasts was observed 8 days after injection. The
injection sites were identified by the presence of
numerous nuclei not expressing human antigens. Human myoblasts were then considered as eliminated by
the inflammatory cells. We did not observe differences in
rejection kinetics between legs injected with SPIOloaded, Gd-DTPA-loaded or unloaded human cells. This
observation was confirmed on D14.
Prussian blue staining confirmed the presence of a
high concentration of iron along aponevrotic sheet
between fibers at D95, which was in good agreement
with the persistence of the negative NMR contrast over
time, and findings by Cahill et al.22
Discussion
In this study, we showed that sufficient contrast can be
generated by Gd chelate and iron oxide CAs preloading
of myoblasts, in media containing 25 mM Gd-DTPA and
100 mg ml1 ferumoxide, respectively, to detect approximately 106 cells immediately after transplantation into
skeletal muscle. No significant toxicity was associated
with the CA concentrations needed to obtain a measurable contrast on NMR images. High cell mortality, which
is the rule with direct intramuscular injection of
myoblasts, releases CA into the interstitium. Low
molecular weight soluble Gd chelates are more or less
rapidly washed out, with a time lag that remains to be
explained. A significant proportion of the released iron
oxide particles are captured by neighboring or inflammatory cells that reside locally for very long periods of
time. The long-term remanence of SPIO-induced hypointensities is no guarantee of transplanted cell survival. In
contrast, there is a very high risk of seeing inflammatory
cells, secondarily attracted on site. As a consequence,
SPIO loading of transplanted cells is valid and will yield
interpretable results only when imaging is performed
immediately after injection to map the initial distribution
volume. Long-term studies will be biased by unspecific
label transfer to macrophages, the severity of the
problem being directly proportional to grafted cell
mortality. It might be less severe when cells are injected
intravascularly. In this study, we addressed only the
question of confounding remanence of contrast and poor
specificity of cell labeling with SPIOs. One also needs to
investigate the long-term follow-up and sensitivity of cell
detection with Gd chelates, for example, in immunodeficient SCID mice. A slow progressive release of Gd
chelate, by the same pinocytosis processes that resulted
in CA intracellular accumulation, is likely to occur, with
kinetic characteristics that remain to be determined.
There is an important lesson to draw from this study.
Potentially therapeutic counts of transplanted cells can be
visualized by in vivo NMR imaging equally well after
loading with Gd chelates or SPIOs. These results were
obtained without the use of cell uptake facilitating factors
and for intracellular concentrations of CAs that had no
detectable deleterious effect on cell viability. In protocols
where cell detection sensitivity is not a critical issue, Gd
Gene Therapy
chelate loading is an alternative for in vivo cell imaging,
which has been somewhat overlooked so far, with only
seven published papers over the last 3 years20,31–36 versus
more than 45 using SPIOs and USPIOs.
There is much more than feasibility, safety and
convenience considerations to encourage the investigators to opt for Gd chelate loading in many conditions.
Although the extent of labeling was initially comparable
with the two CAs, a crucial observation was the very
divergent outcome of Gd chelate and SPIO labeling as
monitored by in vivo NMR imaging. The labeled area
shrank at a slower rate after SPIO loading than after Gd
chelate loading. Although Gd labeling totally disappeared, it remained an SPIO-labeled area throughout the
90 days of follow up. This area was stable with a surface
corresponding to approximately 40% of the initial zone.
In the caricatured model of acute immune rejection that
we had chosen, long-term remanence of SPIOs was a
proof of their release from myoblasts and re-uptake by
other cells. Very similar observations were made in both
earlier and very recent publications.22,37 However, with
our head-to-head comparison of labeling time courses
obtained with two different classes of CAs and supplemented by histological confrontations at many time
points, we clearly demonstrated the severity of the
positive bias generated by SPIOs for the long-term
monitoring of grafted cells. As SPIOs are avidly phagocytosed by reticulo-histiocytic cells, these compounds
were initially developed for intravenous injections to map
reticulo-histiocytic cell distribution. Hence, re-uptake by
inflammatory cells of SPIOs released by dead labeled cells
came as no surprise and was documented here by
persistent positive Prussian blue staining of infiltrating
cells in grafted muscles. Smaller size iron oxide complexes
are less avidly taken up by phagocytic cells and choosing
USPIOs instead would probably somewhat alleviate the
particle recapture problem, at the expense of a lesser
stability of the compound and its progressive catabolism
over a few weeks time.
From a quantitative point of view, very large magnetic
moments of SPIOs, which are responsible for their
remarkable detection sensitivity, induce magnetic susceptibility changes at a relatively long distance and water
spins dephase over a volume larger than the actual
volume of labeled cells. These susceptibility effects are
complex: they depend on the SPIO distribution in cells,
on labeled cell dispersion and orientation with regard to
the main magnetic field and their visualization also
depends on the orientation of the imaging slice. This
makes accurate localization and quantification of SPIO
distribution quite difficult to achieve. In some instances,
computer simulations and in vitro assessment of the
labeled cells prior to injection can help,38 but quantitation
remains generally an unresolved issue.
As a consequence, the key message of this study is the
following. Although a vast majority of studies of cell
therapy monitoring by NMR imaging have been performed with SPIOs, the use of this CA is inadequate and
will lead to erroneous conclusions in most instances, that
is, each time a significant amount of grafted cells die in
host tissue, which is the most frequent situation. Most
investigators give preference to SPIO labeling because of
the higher detection sensitivity with this CA. It remains
legitimate to take advantage of this property to map
therapeutic cell distribution but only immediately after
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
grafting or in the rare event that transplanted cell
population is not to be prone to massive death.
Rather unexpectedly, Gd chelate labeling remained
visible up to 2 weeks after transplanted cell rejection. Reuptake by other cell types of Gd chelates released by
dead transplanted cells is a priori an unlikely event
because of the low molecular weight and the very short
biological half-life of these molecules. Poor drainage of
the interstitium due to local inflammation might facilitate uptake by reticulo-histiocytic cells or granulocytes,
to an extent that remains to be determined. Another
possible event is the fusion of human myoblasts with
native muscle fibers and formation of hybrid fibers, the
rejection of which is delayed and in which CA remained
diluted but present for a longer period of time. However,
such hybrid fibers were rare, as assessed by the
immunohistochemical analysis. Also, some of the Gd
chelate released by the many transplanted cells dying
immediately after injection might diffuse into the
cytoplasm of host muscle fibers that were damaged by
needle trajectory but survived and subsequently resealed
their sarcolemma. As a consequence, a quantity of Gd
chelate would remain entrapped in host cells for an
undetermined period of time. On the basis of measurements of Gd chelate persistence in electroporated muscle
fibers,39 it might be several weeks. After direct injection
of Gd chelate into muscle (data not shown), local
response was biphasic, most of the initial hyperintensity
(70% of the initial product size contrast) disappeared
during the next 20 h but a slow component of 30%
persisted for weeks, which is in favor of the latter
hypothesis. More conclusive evidence would be obtained
by tracking the outcome of myoblasts loaded with Eu
chelate by fluorescence microscopy. This would indicate
whether any of this CA is present in inflammatory cells
or in host muscle fibers. Their detection seems, however,
to be considerably more complex on histological samples
than on cultured cells.
As supported by our observations, we can at least
assume that Gd chelate re-uptake is a problem of much
lesser gravity than it is with SPIOs. The mechanisms
responsible for the progressive and rapid decrease in Gd
chelate label size and intensity cannot be fully elucidated
at the moment. Of course, it is directly related to the loss
of transplanted cells. For instance, the dramatic decrease
in the number of positive pixels and in the relative
contrast between D0 and D1 reflects the fact that at least
50% of grafted cells die within the first hours after
injection. The time course of Gd labeling disappearance
roughly follows the timing of an acute graft rejection,
corroborated by the histological and immunohistological
data collected in this study. A progressive release of CA
from grafted cells by exocytosis can potentially contribute to the label fading, as suggested by our wash-out
experiments in vitro (data not shown). Recently, Effectene-mediated transfection has proven to increase
intracellular labeling longevity as compared with direct
pinocytosis.35 However, the long-term remanence in vivo
of Gd chelate in electroporated muscles suggests a minor
contribution from this mechanism. The dependence of
CA release on loading procedures has to be investigated
further. For instance, CA diffuses in cytoplasm after
electroporation, and in perinuclear vesicles after pinocytosis.24 Also, as recently pointed out, differential in vivo
stability of commercial Gd chelates is another variable to
be take into consideration for longer term studies,40 but
should have no role to play in the relative acute process
we had been monitoring.
There are other issues to be considered when using Gd
chelate loading to track grafted cells. Hyperintensities
generated by Gd chelate might be difficult to distinguish
from areas of fatty degeneration in diseased muscle.
Ambiguities can be removed by adding chemically
selective fat saturation modules to the imaging sequences. An angiographic effect by inflow of fast moving
spins is frequent in T1-weighted sequences used to map
Gd chelate-loaded cells. Discrimination is usually very
simple based on the well-known topography of vasculature. If needed, as might be the case when therapeutic
cells are administrated intravascularly, spatially selective
saturation modules positioned upstream could be used
to annihilate the blood signal.
Quantitative analysis of labeling size and intensity on
NMR images was performed in this study. Yet, by no
means does this imply that quantitative determination of
grafted cell number or even of CA’s concentration can be
easily derived. But again, in most, if not all instances,
quantitation issues would appear more problematic with
SPIOs than with Gd chelates. This in our view represents
a strong argument in favor of the latter. There are reports
that claim quantitative estimates of ferrite-labeled cells.41
Without any doubt, the size of the hypointensities
generated by SPIOs is a function of the number of
particles in the voxel. They also depend on local
distribution of particles and on their orientation in the
static magnetic field. Ferrite-induced hypointensities
extend into adjacent voxels in a complex manner, also
depending on the angulation with the magnetic field. All
this makes the in vivo determination of SPIO concentration more than hypothetical as most of the controlling
parameters are unknown and unquantifiable. It is
unclear why the size of the SPIO spot increased between
D0 and D1. The strong initial decrease in T2 relaxation
associated with fluid injection and early inflammatory
response certainly intervened. Labeled cell dispersion
over a somewhat larger volume than the injection
probably played a role as well. Their relative contributions would be difficult to determine precisely, knowing
that, in addition, a large proportion of cells die during
that specific period of time, and an undetermined
fraction of released SPIOs entered lymphatic and venous
circulations. In contrast, signal intensity changes induced
by Gd chelate depend primarily on the CA concentration
and relaxivity and hence would be more easily amenable
to quantitation. However, intracellular relaxivities of Gd
CAs cannot be determined in vivo and have to be
extrapolated from in vitro measurements on biological
samples. Relaxivity quenching at increasing concentrations of CAs has been identified after passive loading by
unfacilitated pinocytosis. No such problem was detected
when electroporation pulses were used to load cells.20
Despite possible acute toxic effects associated with this
technique, it should probably be given preference in
protocols aiming at quantitative estimations. Labeling
dilution by cell division or by fusion with unlabeled
muscle fibers is often put forward as a drawback of cell
detection based on prior loading with a CA. However,
with Gd chelates, it should not significantly alter contrast
as long as cells remain within the initial voxel after
division or fusion.
741
Gene Therapy
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
742
Analysis of the negative contrast induced by SPIOs
can also present some specific difficulties. Possible
confusion with air bubble artifacts has been described,42
but this seems easy to avoid if injection is carefully
performed. In our numerous in vivo experiments, this
problem was encountered only once. More frequently,
iron susceptibility artifacts can prevent accurate localization of the graft and proper appreciation of response of
surrounding tissues to treatment.
Hemorrhage, with the complex time course of signal
abnormalities caused by red blood cell lysis and
hemoglobin catabolism, is another potential confounding
factor.43 Some degree of bleeding might indeed occur as a
consequence of the damages directly caused by injection.
Such lesions would be similarly created during PBS
injection as well but no abnormal signal was detected on
NMR images taken in this experimental condition, which
precludes a frequent contribution of hemorrhage to the
observed NMR signal changes after intramuscular cell
injection.
Inflammatory reaction caused by surgery and injection
proved to be transient. After an injection of PBS, T2 maps
showed only a discrete and focal increase in muscle at
the injection site and T2 was back to normal values
within less than 6 days. The inflammation caused by the
immune reaction did not affect the detection of the graft
on T1-weighted images. However, the Gd shortens T1
and T2, whereas inflammation increases T1, T2 and
proton density. T1-weighted imaging is also sensitive to
changes in T2 and proton density. The effects are
opposite and may reduce or hide label contrast.
As just discussed above, all these sources of potential
interferences had only a minor impact on our experimental results, the main significance of which remains
unquestioned. Nevertheless, novel approaches might be
appealing, with higher specificity and with a contrast
depending exclusively on cell presence in situ and clearly
separable from other coinciding events. Activatable CAs
would be of great interest for the assessment of tissues
heavily damaged by pathological changes. Extensive
skeletal muscle remodeling develops in dystrophic
muscles and it might be difficult to recognize dark or
white spots generated by CA-labeled cells in a tissue that
appears as a patchwork of hypo-, iso- and hyperintensities, corresponding to fibrosis, regenerating fibers and
fatty replacement, respectively. These activatable NMR
CAs exist and are known as CEST or PARACEST
agents.44 Contrast is generated by magnetization transfer
between exchangeable protons in specific molecules and
water. Although the principle of mechanism is very
elegant, the preliminary results are encouraging, but
these CAs are at an early stage of development and are
not ready for use in preclinical studies. Besides, toxicity
issues have yet to be addressed.
Another approach would be to use a reporter gene
expression to identify cells transfected prior to transplantation. Plane scintigraphy and more recently single
photon emission computed tomography (SPECT) and
mSPECT tomography have been proposed to detect the
sodium/iodide symporter (NIS) activity in cells transfected with the NIS gene for several months in rats.45
Novel NMR reporter gene solutions might prove
competitive, in particular the ones coding for bacterial
membrane iron transporter.46 When controlled by an
inducible promoter, these transporters accumulate iron
Gene Therapy
in eukaryotic cells and generate a contrast very similar to
SPIOs, whereas the confounding risk of CA recapture
and recycling by macrophages is largely prevented. It
has also been proposed to introduce a reporter gene
coding for polypeptides with CEST properties, such as
polylysine. Proof of concept was recently reported.47
However, all these novel developments, although very
promising in terms of contrast improvement, raise a lot
of questions about toxic and immunologic side effects.
For these reasons, the standard, safe CAs studied here
and clinically approved for i.v. injections long back, are
likely to be the only ones to be considered for cell
labeling in human protocols in the foreseeable future.
Their use for these specific applications deserves to be
further optimized.
Beyond survival and migration, therapeutic cell functionality is the key issue, and is very difficult to
demonstrate in vivo. High-resolution functional NMR
imaging has the potential to quantify fiber groups’
contractility with velocity-encoding or -tagging techniques,
to map local perfusion and blood oxygenation with arterial
spin labeling and BOLD sequences.48 It could be adapted
to measure contractility, perfusion and oxygenation in
areas with high therapeutic cell density detected by
contrast enhancement on high-resolution anatomical
images. Similarly, high-resolution chemical shift NMR
spectroscopy can quantify the creatine pool,49 an index of
muscle metabolic activity, which could be used to evaluate
energy metabolism in grafted areas as well. If successful,
the combination of therapeutic cell topographical mapping
and functional evaluation by NMR techniques would give
NMR a definitive advantage over competing techniques
for cell transplantation monitoring in vivo.
Materials and methods
Cell culture
Human muscle biopsies were obtained in agreement
with the French Regulatory Health Authorities and our
Ethics Committee from the Tissue Bank for Research of
the French Association against Myopathies. The muscle
fragments were minced and digested using liberase (1 h
at 37 1C; Roche, Meylan, France), then trypsin-EDTA
(0.05%, 20 min at 37 1C; Hyclone-Perbio, Brebieres,
France). The cell suspension was filtered through 100mm and then 40-mm cell strainers (Becton-Dickinson
(BD), Le Pont de Claix, France) and pelleted. The cells
were seeded in the proliferation medium containing 80%
of modified MCDB medium (custom made by HyclonePerbio), 20% defined fetal bovine serum (HyclonePerbio), gentamicin (50 mg ml1; Invitrogen, CergyPontoise, France), human recombinant bFGF (10 ng ml1;
R&D Systems, Lille, France) and dexamethasone (1 mM;
Merck, Clermont-Ferrand, France). The cells were settled
for 5 days, and then harvested by trypsinization and
expanded until 80% confluence. Cells were amplified
in cell factory (Nunc, Rochester, NY, USA). At each
passage, one fraction was collected for phenotypical
analysis (see below).
Cell loading with CAs
Here, 5 million cells were seeded in 75-cm2 culture flasks.
The following day, the required amounts of Gd-DTPA
(Magnevist; Schering, Berlin, Germany) or SPIO (Endor-
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
em; Guerbet, Villepinte, France) were added to the
culture medium to have final concentrations ranging
from 10 to 80 mM and from 100 to 1000 mg ml1,
respectively. After 24 h, loaded cells were harvested by
trypsinization and suspended in PBS for in vitro
measurements, or in PBS+0.5% of bovine serum albumin
(Sigma, St Quentin Fallavier, France) for in vivo studies.
Cell analysis by cytofluorimetry: phenotyping, viability
and apoptosis
CD56 evaluation. The cells were incubated with
phycoerythrin-coupled anti-CD56 antibody (clone
NCAM 16.2, 1/5, 15 min at 4 1C; BD). Nonspecific
labeling was evaluated using control isotype antibodies
(IgG2b; BD). The CD56 phenotype was determined by
fluorescence-activated cell sorting (FACS Calibur; BD)
and recorded data (at least 104 events) were analyzed
using the Cell Quest Software.
Cell viability. Cell viability of loaded cells was evaluated by propidium iodide exclusion (0.1 mg ml1, 2 min
at room temperature (RT); Sigma) and FACS analyses.
Cells apoptosis. Apoptosis of loaded cells was evaluated by annexin expression study. Briefly, cells were
suspended in 100 ml annexin buffer (BD) containing
annexin V-FITC antibody (BD) and incubated for 15 min
at RT in the dark. Then, 100 ml of annexin buffer was
added followed by data acquisition, which was performed immediately on a flow cytometer.
Proliferation assay. From the previously described cell
cultures, 2 104 myoblasts, loaded with or without CAs,
were harvested and seeded into 12-well plates containing
the proliferation medium. Cells from two wells were
harvested at settled time points (D1, D2, D3, D4, D6 and
D7) and counting was performed with a hematocytometer. Mean proliferation curves were established for
each condition (n ¼ 3 for each time point).
Differentiation assay. In total, 50 000 loaded cells were
seeded into 12-well plates. Differentiation was induced
after 24 h by replacement of the proliferation medium by
the differentiation medium containing DMEM medium
(Invitrogen) and 10% horse serum (AbCys, Paris,
France). At 5 days after the induction of differentiation,
cells were permeabilized and fixed by cold methanol.
The expression of fast MyHC in differentiated myotubes
was assessed using monoclonal anti-fast MyHC antibody
(clone MY32, 1/400, 1 h; Sigma) followed by a goat antimouse IgG (H+L) antibody (Alexa Fluor 568, 1/1000, 1 h;
Invitrogen). The nuclei were labeled with DAPI (Vectashield+DAPI; AbCys). The presence or absence of
labeled structures was then evaluated qualitatively using
an inverted Olympus microscope equipped for fluorescence analysis. The fusion index was calculated on three
microscopic fields at magnification 100 (at least 1000
nuclei counted).
In vivo experiments
The experiments complied with the Principles of
Laboratory Animal Care. Transplantation experiments
were performed in immunocompetent C57BL/6 mice (9
weeks old) (Charles River Laboratories, L’Arbresle,
France). Animals were anesthetized with ketamine/
xylazine (30/15 mg kg1). The skin was incised to expose
the tibialis anterior muscle and 2 106 myoblasts
suspended in 10 ml PBS-BSA were slowly injected along
the length of the tibialis anterior with a microsyringe
(Hamilton, Bonaduz, Switzerland). This was repeated on
both legs. In the first group A, 14 animals were injected
in both legs with unloaded cells or with cells pre-loaded
either with Gd-DTPA or SPIO. Nine were monitored by
NMR during 3 months, and five were killed on D6, D8,
D11 or D16 to compare with histological results. The
other group B was set up with 20 animals, primarily to
histologically assess cell survival following the graft.
Animals received unloaded myoblasts in one of their legs
and either Gd-DTPA- (n ¼ 14) or SPIO (n ¼ 6)-loaded
myoblasts in the other leg. They were picked out on D0,
D1, D2, D4, D6, D8, D11 or D14 for histology,
immediately after NMR imaging.
743
NMR imaging
Mice underwent a series of NMR examinations immediately after cell injection (D0) (groups A and B) and
repeatedly on D1, D4, D6, D8, D11, D14, D21 and 3
months post-transplantation (group A) or just before
their killing (group B).
All acquisitions were performed on a 4T Bruker
Biospec NMR spectrometer, equipped with a 20 cm
diameter 200 mT m1 gradient insert (Bruker BioSpin
MRI GmbH, Ettlingen, Germany). Mice were anesthetized with 2% isoflurane (Forene; Abbot, Rungis, France)
mixed with medical air delivered at 1.5 l min1 and
placed supine in a 28 mm inner diameter volume coil
with quadrature polarization (Rapid Biomedical GmbH,
Würzburg, Germany). A heating pad ensured a constant
temperature during the experiment.
Axial T1-weighted spin echo images of the lower
limbs were acquired. Imaging parameters were
TR ¼ 150 ms, TE ¼ 5.7 ms, NEX ¼ 20, in-plane resolution ¼ 120 mm 120 mm, slice thickness ¼ 1.35 mm and
inter-slice gap ¼ 2 mm. T1 was measured by saturation
recovery RARE imaging (effective TE ¼ 8 ms; RARE
factor ¼ 4; TR ¼ 71 ms; matrix size: 256 128; in-plane
resolution: 133 mm 137 mm; slice thickness ¼ 1.35 mm).
T2 measurements were also systematically obtained by
using a multiple spin-echo sequence (40 echoes, equally
spaced between 7.4 and 296 ms; TR ¼ 2000 ms; matrix
size: 256 256; in-plane resolution: 120 mm 120 mm;
slice thickness ¼ 1.35 mm).
Immunohistochemical analyses
Immediately after the NMR imaging session, mice were
killed by overdose of anesthetics and muscles were taken
and then snap frozen in nitrogen-cooled isopentane and
serially sectioned (7 mm) using a cryostat.
Immunofluorescence labeling. The presence of human
cells was detected by the species-specific labeling of
human lamin A/C protein. Labeling of laminin protein
defined the basement membrane localization. Sections
were first incubated with the polyclonal anti-laminin (1/
200, 1 h, RT; Abcam, Cambridge, UK) and monoclonal
IgG2b anti-human lamin A/C (1/100, 1 h, RT; AbCys),
and then with a goat anti-mouse IgG2b antibody (Alexa
Fluor 568, 1/1000, 1 h, RT; Invitrogen) and goat antirabbit IgG (H+L) antibody (Alexa Fluor 488, 1/1000, 1 h,
Gene Therapy
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
744
RT; Invitrogen). Slides were mounted under Vectashield
hard set+DAPI (AbCys) and observed using a Zeiss
fluorescence microscope.
Prussian blue staining. Serial sections were incubated
at first with a 0.12 M K4Fe(CN)6 and 1 M HCl solution
(20 min, RT) to reveal the presence of iron particles and
with nuclear fast red solution (15 min, RT; Sigma) to
reveal the cytoplasm and nucleus. Finally, they were
washed, dehydrated, mounted with Eukitt (EMS, Fort
Washington, PA, USA) and examined by light microscopy (Olympus IX microscope).
Data analysis
Registration between examinations was manually performed, based on the shape and the area of the tibia bone
section.
T1-weighted. images were analyzed using the Image
Sequences Analysis Tool of ParaVision (Bruker). To
discriminate between pixels from labeled cells and pixels
from muscle, we considered a confidence interval of 95%
for muscle intensity. Thus, we defined the detection
threshold as (S̄muscle+1.96 s.d.) for Gd-DTPA label and
(S̄muscle1.96 s.d.) for iron oxide label. S̄muscle is the mean
intensity of muscle tissue surrounding the injection site
and s.d. is the associated standard deviation. Number of
pixels and signal intensity of the detected area were
noted. The evolution of the graft was described by the
following parameters:
The size N, the number of pixels detected as label,
normalized to N at D0.
The relative contrast in absolute value,
C¼j
N
X
ðSlabeled pixels
pixels¼1
Sreference muscle Þ=ðNSreference muscle Þj;
normalized to C at D0.
The product of the size by the relative contrast: N C.
Half-life detection times of these parameters were
extracted.
T1 and T2 maps were calculated using Matlab 7.1 (The
Mathworks Inc., Natick, MA, USA) and Dr1 and Dr2, the
changes in the longitudinal and transversal relaxation
rates of water proton, respectively, were evaluated. For
calculation of T2 maps, the decay was assumed to be a
single exponential. Noise was evaluated on the first
magnitude image outside the object, and used to exclude
pixels with low signal-to-noise ratio from the data set to
fit according to SX7.6 Ms.d., where Ms.d. is the background noise standard deviation, which is a stricter
criterion than advised by Tofts.50
Statistics
All results are presented as mean±s.d. Statistical
analysis of NMR quantitative data were performed by
using repeated measures analysis of variance (ANOVA),
followed by Bonferroni comparison of means when
appropriate. Toxicity assays on cell cultures were
analyzed using ANOVA, then Tukey’s test when appropriate. Pp0.05 was considered significant (NCSS 2001;
NCSS, Kaysville, UT, USA).
Gene Therapy
Acknowledgements
This study was supported by the Association Française
contre les Myopathies (AFM). We thank Dr J Larghero,
Mrs B Ternaux, M-N Lacassagne and I Robert for
cytometric analyses (Cell Therapy Laboratory, St Louis
Hospital, Paris, France). We thank the Guerbet group for
kindly providing the SPIOs (Endorem) and Myosix.
References
1 Vilquin JT. Myoblast transplantation: clinical trials and perspectives. Mini-review. Acta Myol 2005; 24: 119–127.
2 Partridge T. Myoblast transplantation. Neuromuscul Disord 2002;
12 (Suppl 1): S3–S6.
3 Skuk D, Goulet M, Roy B, Chapdelaine P, Bouchard JP, Roy R
et al. Dystrophin expression in muscles of Duchenne muscular
dystrophy patients after high-density injections of normal
myogenic cells. J Neuropathol Exp Neurol 2006; 65: 371–386.
4 Menasche P, Alfieri O, Janssens S, McKenna W, Reichenspurner
H, Trinquart L et al. The Myoblast Autologous Grafting in
Ischemic Cardiomyopathy (MAGIC) trial: first randomized
placebo-controlled study of myoblast transplantation. Circulation
2008; 117: 1189–1200.
5 Strasser H, Marksteiner R, Margreiter E, Pinggera GM, Mitterberger M, Frauscher F et al. Autologous myoblasts and
fibroblasts versus collagen for treatment of stress urinary
incontinence in women: a randomised controlled trial. Lancet
2007; 369: 2179–2186.
6 Sampaolesi M, Blot S, D’Antona G, Granger N, Tonlorenzi R,
Innocenzi A et al. Mesoangioblast stem cells ameliorate muscle
function in dystrophic dogs. Nature 2006; 444: 574–579.
7 Zheng B, Cao B, Crisan M, Sun B, Li G, Logar A et al. Prospective
identification of myogenic endothelial cells in human skeletal
muscle. Nat Biotechnol 2007; 25: 1025–1034.
8 Torrente Y, Belicchi M, Marchesi C, Dantona G, Cogiamanian F,
Pisati F et al. Autologous transplantation of muscle-derived
CD133+ stem cells in Duchenne muscle patients. Cell Transplant
2007; 16: 563–577.
9 Satoh A, Labrecque C, Tremblay JP. Myotubes can be formed
within implanted adipose tissue. Transplant Proc 1992; 24:
3017–3019.
10 Arcila ME, Ameredes BT, DeRosimo JF, Washabaugh CH, Yang J,
Johnson PC et al. Mass and functional capacity of regenerating
muscle is enhanced by myoblast transfer. J Neurobiol 1997; 33:
185–198.
11 DeRosimo JF, Washabaugh CH, Ontell MP, Daood MJ, Watchko
JF, Watkins SC et al. Enhancement of adult muscle regeneration
by primary myoblast transplantation. Cell Transplant 2000; 9:
369–377.
12 Irintchev A, Langer M, Zweyer M, Theisen R, Wernig A.
Functional improvement of damaged adult mouse muscle by
implantation of primary myoblasts. J Physiol 1997; 500 (Part 3):
775–785.
13 Frier M. Leucocyte radiolabelling techniques: practical aspects.
Scand J Gastroenterol Suppl 1994; 203: 32–35.
14 McAfee JG, Subramanian G, Gagne G, Schneider RF, Zapf-Longo
C. 99mTc-HM-PAO for leukocyte labeling––experimental
comparison with 111In oxine in dogs. Eur J Nucl Med 1987; 13:
353–357.
15 Peters AM, Danpure HJ, Osman S, Hawker RJ, Henderson BL,
Hodgson HJ et al. Clinical experience with 99mTc-hexamethylpropylene-amineoxime for labelling leucocytes and imaging
inflammation. Lancet 1986; 2: 946–949.
16 Bulte JW, Duncan ID, Frank JA. In vivo magnetic resonance
tracking of magnetically labeled cells after transplantation.
J Cereb Blood Flow Metab 2002; 22: 899–907.
In vivo MRI of Gd-DTPA and SPIO-loaded myoblasts
C Baligand et al
745
17 Weissleder R, Elizondo G, Stark DD, Hahn PF, Marfil J, Gonzalez
JF et al. The diagnosis of splenic lymphoma by MR imaging:
value of superparamagnetic iron oxide. AJR Am J Roentgenol
1989; 152: 175–180.
18 Yeh TC, Zhang W, Ildstad ST, Ho C. In vivo dynamic MRI
tracking of rat T-cells labeled with superparamagnetic ironoxide particles. Magn Reson Med 1995; 33: 200–208.
19 Bulte JW, Zhang S, van Gelderen P, Herynek V, Jordan EK,
Duncan ID et al. Neurotransplantation of magnetically labeled
oligodendrocyte progenitors: magnetic resonance tracking of cell
migration and myelination. Proc Natl Acad Sci USA 1999; 96:
15256–15261.
20 Crich SG, Biancone L, Cantaluppi V, Duo D, Esposito G, Russo S
et al. Improved route for the visualization of stem cells labeled
with a Gd-/Eu-chelate as dual (MRI and fluorescence) agent.
Magn Reson Med 2004; 51: 938–944.
21 Bulte JW, Kraitchman DL. Iron oxide MR contrast agents for
molecular and cellular imaging. NMR Biomed 2004; 17: 484–499.
22 Cahill KS, Gaidosh G, Huard J, Silver X, Byrne BJ, Walter GA.
Noninvasive monitoring and tracking of muscle stem cell
transplants. Transplantation 2004; 78: 1626–1633.
23 Aime S, Cabella C, Colombatto S, Geninatti Crich S, Gianolio E,
Maggioni F. Insights into the use of paramagnetic Gd(III)
complexes in MR-molecular imaging investigations. J Magn
Reson Imaging 2002; 16: 394–406.
24 Terreno E, Geninatti Crich S, Belfiore S, Biancone L, Cabella C,
Esposito G et al. Effect of the intracellular localization of a Gdbased imaging probe on the relaxation enhancement of water
protons. Magn Reson Med 2006; 55: 491–497.
25 Lin SP, Brown JJ. MR contrast agents: physical and pharmacologic basics. J Magn Reson Imaging 2007; 25: 884–899.
26 Oksendal AN, Hals PA. Biodistribution and toxicity of MR
imaging contrast media. J Magn Reson Imaging 1993; 3: 157–165.
27 Praud C, Vauchez K, Lombes A, Fiszman M, Vilquin JT.
Myoblast xenotransplantation as a tool to evaluate the appropriateness of nanoparticular versus cellular trackers. Cell
Transplant 2008; 17: 1035–1043.
28 Kaufman CL, Williams M, Ryle LM, Smith TL, Tanner M, Ho C.
Superparamagnetic iron oxide particles transactivator proteinfluorescein isothiocyanate particle labeling for in vivo magnetic
resonance imaging detection of cell migration: uptake and
durability. Transplantation 2003; 76: 1043–1046.
29 Walter GA, Cahill KS, Huard J, Feng H, Douglas T, Sweeney HL
et al. Noninvasive monitoring of stem cell transfer for muscle
disorders. Magn Reson Med 2004; 51: 273–277.
30 Skuk D, Caron N, Goulet M, Roy B, Espinosa F, Tremblay JP.
Dynamics of the early immune cellular reactions after myogenic
cell transplantation. Cell Transplant 2002; 11: 671–681.
31 Anderson SA, Lee KK, Frank JA. Gadolinium-fullerenol as a
paramagnetic contrast agent for cellular imaging. Invest Radiol
2006; 41: 332–338.
32 Biancone L, Crich SG, Cantaluppi V, Romanazzi GM, Russo S,
Scalabrino E et al. Magnetic resonance imaging of gadoliniumlabeled pancreatic islets for experimental transplantation. NMR
Biomed 2007; 20: 40–48.
33 Giesel FL, Stroick M, Griebe M, Troster H, von der Lieth CW,
Requardt M et al. Gadofluorine m uptake in stem cells as a new
magnetic resonance imaging tracking method: an in vitro and
in vivo study. Invest Radiol 2006; 41: 868–873.
34 Granot D, Kunz-Schughart LA, Neeman M. Labeling fibroblasts
with biotin-BSA-GdDTPA-FAM for tracking of tumor-associated
stroma by fluorescence and MR imaging. Magn Reson Med 2005;
54: 789–797.
35 Shyu WC, Chen CP, Lin SZ, Lee YJ, Li H. Efficient tracking of
non-iron-labeled mesenchymal stem cells with serial MRI in
chronic stroke rats. Stroke 2007; 38: 367–374.
36 Vuu K, Xie J, McDonald MA, Bernardo M, Hunter F, Zhang Y
et al. Gadolinium-rhodamine nanoparticles for cell labeling and
tracking via magnetic resonance and optical imaging. Bioconjug
Chem 2005; 16: 995–999.
37 Terrovitis J, Stuber M, Youssef A, Preece S, Leppo M, Kizana E
et al. Magnetic resonance imaging overestimates ferumoxidelabeled stem cell survival after transplantation in the heart.
Circulation 2008; 117: 1555–1562.
38 Pintaske J, Muller-Bierl B, Schick F. Geometry and extension
of signal voids in MR images induced by aggregations
of magnetically labelled cells. Phys Med Biol 2006; 51:
4707–4718.
39 Paturneau-Jouas M, Parzy E, Vidal G, Carlier PG, Wary C,
Vilquin JT et al. Electrotransfer at MR imaging: tool for
optimization of gene transfer protocols-feasibility study in mice.
Radiology 2003; 228: 768–775.
40 Kuo PH, Kanal E, Abu-Alfa AK, Cowper SE. Gadolinium-based
MR contrast agents and nephrogenic systemic fibrosis. Radiology
2007; 242: 647–649.
41 Mowat P, Franconi F, Chapon C, Lemaire L, Dorat J, Hindre F
et al. Evaluating SPIO-labelled cell MR efficiency by
three-dimensional quantitative T2* MRI. NMR Biomed 2007; 20:
21–27.
42 Bernsen MR, van Tiel ST, Wielopolski PA, Krestin GP. Susceptibility artifacts resulting from paramagnetically labeled cells:
understanding what you see. European Society for Magnetic
Resonance in Medicine and Biology 23rd Annual Meeing, Magma:
Warsaw, Poland, 2006, pp. 275–276.
43 Stark DD, Bradley Jr WG. Hemorrhage. In: Magnetic
Resonance Imaging, third edn, vol 3. Mosby: St Louis, 1999, pp
1329–1345.
44 Zhang S, Merritt M, Woessner DE, Lenkinski RE, Sherry AD.
PARACEST agents: modulating MRI contrast via water proton
exchange. Acc Chem Res 2003; 36: 783–790.
45 Bouchentouf M, Benabdallah BF, Dumont M, Rousseau J, Jobin
L, Tremblay JP. Real-time imaging of myoblast transplantation
using the human sodium iodide symporter. Biotechniques 2005;
38: 937–942.
46 Zurkiya O, Chan AW, Hu X. MagA is sufficient for producing
magnetic nanoparticles in mammalian cells, making it an MRI
reporter. Magn Reson Med 2008; 59: 1225–1231.
47 Gilad AA, McMahon MT, Walczak P, Winnard Jr PT, Raman V,
van Laarhoven HW et al. Artificial reporter gene providing MRI
contrast based on proton exchange. Nat Biotechnol 2007; 25:
217–219.
48 Carlier PG, Bertoldi D, Baligand C, Wary C, Fromes Y. Muscle
blood flow and oxygenation measured by NMR imaging and
spectroscopy. NMR Biomed 2006; 19: 954–967.
49 Boesch C. Musculoskeletal spectroscopy. J Magn Reson Imaging
2007; 25: 321–338.
50 Tofts P. Quantitative MRI of the Brain. Measuring Changes
Caused by Disease. John Wiley & Sons 2003, 650pp.
Gene Therapy