- Email: [email protected]
Mapping transplanted stem cell migration after a stroke: a serial, in vivo magnetic resonance imaging study
www.elsevier.com/locate/ynimg NeuroImage 21 (2004) 311 – 317
Mapping transplanted stem cell migration after a stroke: a serial, in vivo magnetic resonance imaging study Michel Modo, a,* Karen Mellodew, b Diana Cash, a Scott E. Fraser, c Thomas J. Meade, d Jack Price, b and Steven C.R. Williams a a
Neuroimaging Research Group—Neurology P042, Institute of Psychiatry, King’s College London, London SE5 8AF, UK Department of Neuroscience, Institute of Psychiatry, King’s College London, London SE5 8AF, UK c Beckman Institute, California Institute of Technology, Pasadena, CA 91125, USA d Departments of Chemistry, Biochemistry, Molecular and Cell Biology, Neurobiology, and Physiology, Northwestern University, Chicago, IL 60208-3500, USA b
Received 5 June 2003; revised 22 August 2003; accepted 22 August 2003
Preferential migration of stem cells toward the site of a lesion is a highly desirable property of stem cells that allows flexibility in the site of graft implantation in the damaged brain. In rats with unilateral stroke damage, neural stem cells transplanted into the contralateral hemisphere migrate across to the lesioned hemisphere and populate the area around the ischaemic infarct. To date, the migration of neural stem cells in the damaged brain has been mainly inferred from snapshot histological images. In this study, we demonstrate that by prelabelling neural stem cells with the bimodal contrast agent GadoliniumRhodamIne Dextran [GRID, detectable by both magnetic resonance imaging (MRI) and fluorescent microscopy], the transhemispheric migration of transplanted neural stem cells contralateral to a stroke lesion can be followed in vivo by serial MRI and corroborated by subsequent histological analyses. Our results indicate that neural stem cells migrated from the injection tract mainly along the corpus callosum within 7 days of transplantation and extensively re-populated the peri-lesion area by 14 days following implantation. In contrast, neural stem cells transplanted into sham controls did not show any substantial migration outside of the injection tract, suggesting that the transcallosal migration observed in the stroke-lesioned animals is due to neural stem cells being attracted by the lesion site. In vivo tracking of the migration of neural stem cells responding to damage will greatly enhance our understanding of optimal transplantation strategies as well as how neural stem cells promote functional and anatomical recovery in neurological disorders. D 2003 Elsevier Inc. All rights reserved. Keywords: MCAo; Neural stem cells; Migration; Contrast agent; Cell tracking; Neural transplants; MRI
* Corresponding author. Neuroimaging Research Group—Neurology P042, Institute of Psychiatry, King’s College London, De Crespigny Park, London SE5 8AF, UK. Fax: +44-207-848-0055. E-mail address: [email protected] (M. Modo). Available online on ScienceDirect (www.sciencedirect.com.) 1053-8119/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2003.08.030
Introduction The clinical application of stem cell therapy to remedy brain damage is dependent on the potential of grafts to promote sustained recovery after transplantation (Bjo¨rklund and Lindvall, 2000; Park et al., 2002). At present, investigating the relationship between anatomical change and behavioural recovery is hindered by the necessity to use post-mortem tissue to determine how neural stem cells contribute to behavioural recovery. The migratory properties and the seamless integration of neural stem cells in damaged brains are considered to be beneficial to behavioural recovery. However, to date there is no indication as to how the migration or integration of neural stem cells into the damaged host parenchyma relates to behavioural recovery. Several recent studies have demonstrated that incorporation of contrast agents into cells destined for cellular therapy before transplantation can afford their in vivo identification by magnetic resonance imaging (MRI) (Bulte et al., 2002; Hoehn et al., 2002; Zhang et al., 2003). Recently, we described the use of a bimodal contrast agent Gadolinium-RhodamIne Dextran (GRID), detectable both in vivo by MRI and subsequently at post-mortem by fluorescent microscopy, as a novel tool to track migrating neural stem cells (Modo et al., 2002a). This approach now allows us to investigate how neural stem cells migrate and integrate into various brain regions in the same individual over time. Neural stem cells transplanted into the contralateral hemisphere of rats with stroke damage, not only populate areas around the injection tract, but show transhemispheric migration to the damaged hemisphere to integrate into areas around the lesion site (Modo et al., 2002c; Veizovic et al., 2001). However, post-mortem histological analyses only provide a single ‘‘snapshot’’ of the presence of cells in the brain. Although transplanted cells were found in the corpus callosum in these published studies, it is unclear whether this tract provided a channel for migration. From ex vivo data, it is also debatable if areas of damage function as ‘attractants’ to neural stem cells or if there is a process of nondirected dispersion followed by selective retention. Histological data also cannot resolve the issue if migrating neural stem cells
312
M. Modo et al. / NeuroImage 21 (2004) 311–317
need to populate an area of damage before behavioural recovery can ensue. The use of in vivo neuroimaging to address the relationship between anatomical/neurochemical and behavioural changes has the potential to further develop and refine neural stem cell therapy. Here we report the serial in vivo tracking of transhemispheric neural stem cell migration in rats with chronic brain damage following middle cerebral artery occlusion (MCAo, a rat model of stroke). To distinguish transplanted cells from host cells, neural stem cells were pre-labelled in vitro with the bimodal contrast agent GRID allowing the identification by MRI for serial in vivo evaluation of migration and by post-mortem fluorescent microscopy to validate the in vivo observations. Sham control animals were also followed serially to determine if neural stem cell migration is a consequence of focal brain damage attracting transplanted cells or if transplanted cells merely diffuse from the injection tract.
Materials and methods Animals Sprague – Dawley rats (Charles River, UK) were acclimatised for at least a week before surgery. Animals weighed between 250 and 270 g upon arrival with maintenance on a 10-h light/14-h dark schedule. All procedures were in accordance with the UK Animals (Scientific) Procedures Act 1986 and the ethical review process of Queen Mary and Westfield College, University of London. Middle cerebral artery occlusion Animals between 280 and 330 g were either allocated for sham (n = 3) or 60 min of transient MCAo surgery (n = 3) in the right hemisphere (Longa et al., 1989). Briefly, animals were anaesthetized with halothane (4% induction, 2% maintenance) in a mixture of O2 and N2O (30:70). Ligatures were placed on the external and common carotid to stop the flow of blood to the internal carotid artery. An aneurysm clip was placed on the internal carotid before the common carotid was opened, to allow insertion of a thread up to the aneurysm clip. The tip of the thread was advanced 18 – 20 mm from the cervical carotid bifurcation or until reaching resistance from the ostium of the middle cerebral artery in the circle of Willis. For MCAo, the thread was left in place for 60 min, whereas for sham surgery, the thread was immediately removed. Following occlusion, animals were tested for spontaneous circling and forelimb flexion (Modo et al., 2000). Occluded animals were reanaesthetized and the thread was removed.
day 14 (Sinden et al., 1997). These cells proliferate at 33jC providing an ample source of cells for transplantation, but cease proliferation at 37jC. In the damaged CNS, MHP36 cells show site-appropriate phenotypic differentiation (Modo et al., 2002c) and are therefore ideal candidates for brain repair (Gray et al., 2000). Frozen MHP36 cells (passage 42) were thawed and cultured for at least one passage before transplantation according to a standard protocol (Sinden et al., 1997). During labelling, MHP36 cells were grown in proliferative conditions (33jC with bFGF and IFN-g) with GRID added to the media. Addition of GRID to the growth medium resulted in a final concentration of 45 AM of Gd3 + and 2.73 AM of tetramethylrhodamine (chelated in one molecule) per millilitre of media. After 6 h of incubation with GRID, the medium was discarded and cells were removed from the flask by adding Hank’s balanced salt solution (HBSS, Gibco) without Ca2 + and Mg2 +. The suspension was centrifuged (1500 rpm for 5 min) and cells were resuspended for transplantation (25,000 cells/Al) with 1 mM Nacetyl-L-cysteine (NAC, Sigma) in HBSS (Gibco). Cell viability for transplantation, as determined by trypan blue (Sigma) exclusion, was >90% after GRID-labelling. At the end of grafting, cell viability was f85% as determined from the same vial used for transplantation by trypan blue exclusion. Grafting Three months after MCAo, animals were transplanted in the left (contralateral) hemisphere to probe their migration toward the stable ischaemic lesion in the right (ipsilateral hemisphere). Anesthesia was induced by intraperitoneal (i.p.) injection of a mixture of 0.1 mg medetomidine hydrochloride (Dormitor, C-Vet Products), 5.5 mg ketamine hydrochloride (Ketaset, C-Vet Products), in 0.1 ml of 0.9% saline (Baxter) per 100 g body weight. Animals were placed in a stereotaxic frame and an incision was made exposing bregma. Burr holes were drilled contralateral (Site 1. AP = + 0.7, L = 2, V = 5.5/ 2; Site 2. AP = 0.3, L = 3, V = 5.5/ 2.5) to the ischaemic lesion. Each deposit consisted of 2 Al at a speed of 1 Al/min. The syringe was left in place for 2 min after injection to allow dispersion of the cells. A total of 2 105 cells with a total volume of 8 Al was injected per animal. No immunosuppression was given since MHP36 cells survive in immunocompetent rats for at least the first 2 weeks following transplantation (Modo et al., 2002b). After grafting, animals were injected with 0.1 mg atipamezole hydrochloride (Anti-Sedan, C-Vet Products) in 0.1 ml of 0.9 saline per 100 g body weight i.p. to counteract anaesthesia followed by 2.5 ml of glucosaline (subcutaneous) in two sites. After suturing, a local analgesic (lignocaine hydrochloride) was applied to minimize post-operative discomfort of the wound.
Preparation of neural stem cells Magnetic resonance imaging The GRID chelate, consisting of gadolinium (Gd-DTPA) and rhodamine particles linked by a dextran polymer, was conjugated according to published methods (Hu¨ber et al., 1998) to yield a single molecule (total molecular weight 16,600) which is acting as a bimodal contrast agent detectable by both MRI and fluorescent histology. GRID was diluted (1:1) with distilled H2O at least 24 h before cell labelling (as described in Modo et al., 2002a). Conditionally immortalized neural stem cells from the Maudsley Hippocampal Clone 36 (MHP36) cell line are derived from the neuroepithelium of the H-2Kb-tsA58 immortomouse at embryonic
To acquire in vivo MR images, animals were anaesthetized using isoflurane (4% induction, 2% maintenance) in a O2/N2O (30:70) mixture supplied via a face mask in a head holder to minimize movement artefacts. All MRI was performed on a small animal, horizontal bore, 4.7T NMR system (Oxford Systems) controlled by a UNITYInova-200 imaging console (Varian). A quadrature birdcage radio frequency coil with 63 mm internal diameter (Varian) was used for signal transmission and reception. The raw time domain data were Fourier transformed using locally
M. Modo et al. / NeuroImage 21 (2004) 311–317
developed software. The choice of imaging protocol was determined by the need for a high spatial resolution as well as the distinction between the host tissue and the gadolinium-bearing graft. Our imaging protocol consisting of coronal T1-weighted (TR = 500 ms, TE = 15 ms), T2-weighted (TR = 4000 ms, TE = 45 ms), and proton density (PD)-weighted (TR = 4000 ms, TE = 15 ms) spin echo scans with a 2562 matrix size, 4 averages per phase encoding step, 30 contiguous 500-Am-thick slices, and a field of view of 2.5 2.5 cm was adopted to afford an in plane spatial resolution of 248 Am. Histology After scanning, animals were overdosed with pentobarbitone sodium (Pentoject, Animal Care) for perfusion. The descending aorta was clamped to restrict blood flow to the upper body half. Rats were pre-washed transcardially with heparinized saline (0.9% NaCl in dH2O containing 5000 units of heparin, CP Pharmaceuticals, per litre) followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate-buffered saline (PBS, Sigma). Brains were removed and placed overnight in 4% PFA at room temperature before being cryoprotected in 30% sucrose for at least 1 week before cutting. Sections of 30 Am thickness were cut on a cryostat microtome (Leica) at 24jC and collected free-floating in sucrose. For immunohistochemical staining, sections were washed 3 5 min in 0.01 M PBS before the application of the primary antibody used for identification of neurons (monoclonal mouse anti-NeuN, 1:1000, Chemicon) or astrocytes (polyclonal goat anti-GFAP, 1:4000, Chemicon). Primary antibodies were applied overnight at room temperature. Sections were washed 3 5 min in PBS and then incubated with a cross-adsorbed secondary antibody (goat anti-mouse Alexa488 for NeuN, and rabbit anti-goat Alexa488 for GFAP, 1:500, Molecular Probes) for 45 min at room temperature. Sections were washed 3 5 min in PBS before coverslips were applied with Vectashield for immunofluorescence (Vector). Macroscopic analyses of graft migration and integration were conducted on a fluorescent microscope (Nikon), whereas phenotypic differentiation of transplanted cells required the use of a confocal microscope (Leica). A confocal analysis of the overlay of GRID and phenotype markers is paramount to ensure that both markers co-localise within the same cells and are not the consequence of juxtaposed markers leading to an erroneous identification of phenotypic differentiation.
Results Lack of migration in sham controls MRI of sham control brains provided evidence of how GRIDlabelled transplants affect the MR signal over time without a possible confound of a lesion environment. GRID-labelled MHP36 cells were detected 1 day post-grafting as an elongated line on T2-weighted images reflecting the injection tract (Fig. 1). The baseline image of sham control brains, 1 day pre-transplantation, provides evidence that the changes observed on the MR scan are a consequence of the injection of GRID-labelled cells. From the MR sequences used in this study, T2-weighted MR images provided the best anatomical discrimination and also allowed a satisfactory distinction between grafted cells and host
313
tissue (Fig. 2). Although PD-weighted scans also allowed differentiation of the grafted cells in the injection tract from the host striatum, they provided less anatomical detail and were therefore considered inferior to T2-weighted images to study the migration of transplanted cells. Only minimal anatomical information could be derived from the T1-weighted MR images and GRID-labelled cells were mostly indistinguishable from the host tissue. T2-weighted MR images were therefore used to study how transplants evolved over time. The assessment of T2-weighted images at 1 day, 7 days, and 14 days following transplantation indicated that MHP36 cells did not migrate (or become passively dispersed) out of the injection tract (Fig. 1). Corroboration of these in vivo observations were provided by post-mortem fluorescent histology (of the rhodamine signal) indicating that grafted cells were densely packed within the injection tract, but were not found in regions distant from the injection tract. A few cells labelled with GRID, however, could be found in the immediate vicinity of the injection tract. Stroke-induced transhemispheric neural stem cell migration Animals with stroke damage were imaged 1 day before transplantation (approximately 3 months following ischaemia) to visualise the chronic lesion providing a baseline to which graft-induced changes were compared (Fig. 3). At 1 day post-grafting, an injection tract was visible from the top of the cortex to the medial part of the striatum. In comparison to sham controls, the injection tracts were located slightly more lateral due to a shift of the brain to the right side caused by the cavitation of the ischaemic lesion. Injection tracts appeared, as in control animals, as dark elongated lines packed with GRID-labelled cells causing a hypointense signal differentiating the injection tract from the host brain. By 7 days post-transplantation, a substantial number of cells had migrated out from the injection tract and induced a change in signal of the contralateral striatum rendering it patchy around the injection tract. However, the injection tract could still be identified, albeit the differentiation from host tissue was less obvious than at 1 day post-grafting. Furthermore, along the corpus callosum, some pockets of cells loaded with GRID could be differentiated from the transhemispheric white matter tract. This indicated that GRIDlabelled transplanted cells used the corpus callosum for transhemispheric migration. At this time point, there was no clear change in signal in the peri-lesion area that would indicate that transplanted cells already arrived at the lesion site. At the final time point, 14 days following contralateral implantation, a significant number of cells had migrated to the lesion along the corpus callosum and populated the peri-lesion area. A clear change in intensity in the region of the lesion close to the corpus callosum was observed on MR images reflecting the accumulation of GRID-labelled MHP36 cells in the peri-lesion area. The contralateral striatum still exhibited some hypointensities reflecting the presence of dispersed transplanted cells. Most clearly, the injection tract could still be detected at 14 days post-grafting, suggesting that not all cells had migrated/diffused out of the implantation site. The temporal dynamics of neural stem cell migration from the contralateral hemisphere therefore suggest that by 1 day following transplantation, cells could be mainly found within the injection tract with no significant dispersion out of the tract, but by 7 days post-grafting, cells disperse in the surrounding striatum and along the corpus callosum. By 14 days post-grafting,
314
M. Modo et al. / NeuroImage 21 (2004) 311–317
Fig. 1. Serial in vivo visualisation of stem cell transplants in control animals. The neural stem cell transplant in a control animal showed no significant migration away from the implantation tract. Despite lack of migration from the implantation site, the stem cell graft (arrows) showed good graft survival over the 14 days of in vivo assessment by T2-weighted MR images and post-mortem by fluorescent histology.
Fig. 2. Multiparametric MRI of stem cell transplants. T1-weighted images were very poor in detecting the gadolinium-bearing grafted cells (arrows), whereas T2-weighted images provided a satisfactory contrast to distinguish grafted cells from host brain. PD-weighted images also allowed the identification of grafted cells, but provided less anatomical detail than T2-weighted images.
a substantial number of transplanted cells populated the ipsilateral peri-lesion area (contralateral to the implantation site), but some cells remained within the injection tract and along the corpus callosum. Histological analyses corroborated the in vivo observation that transplanted cells migrated from the contralateral implantation site to populate the peri-lesion area (Fig. 4). In the peri-lesion area, a glial scar surrounded the cavity resulting from the ischaemic insult and transplanted cells were found both within this glial scar and within the non-gliotic tissue surrounding the lesion. However, histologically GRID-labelled transplanted cells were also present at lower concentrations in regions of the brain in which no significant signal change was detected on any of the MR images. These small clusters of cells could be found in the contralateral striatum away from the injection tract. GRID-labelled cells also lined the injection tract that could still be detected in the MR images 14 days post-injection. At the bottom
of the injection tract, the deposit showed a marked lateral dispersion suggesting that most of the cells from the deposit had migrated away from their site of implantation. Within the injection tract, cells were still strongly labelled with GRID, whereas cells at more remote sites, such as the ipsilateral hemisphere, showed less GRID fluorescence, suggesting that these cells had undergone some cell division that had diluted the total GRID. The wider dispersion of MHP36 cells with low amounts of GRID in remote locations (such as the ipsilateral hemisphere) therefore affords a better identification of the graft (at least visually) compared to a high concentration of GRID inside the cells in the densely packed injection tract. Phenotypic differentiation of transplanted stem cells In sham control animals, cells within the injection tract did not, in the main, co-localise with either GFAP (as an astrocyte marker) or NeuN (as a neuronal marker), but a few cells in and especially
Fig. 3. Serial in vivo tracking of stroke-induced transhemispheric stem cell migration. One day after transplantation of MHP36 cells, an injection tract (arrow) packed with GRID-labelled cells can be detected in the medial striatum in comparison to pre-transplant T2-weighted MR images of the same animal. Over the following 14 days post-grafting in the same animal, packs of transplanted cells migrated along the corpus callosum to reach the site of damage (arrows). This indicates that migration of stem cells to the contralateral hemisphere mainly relies on transcallosal migration. However, some cells remained within the injection tract in the contralateral hemisphere (arrow).
M. Modo et al. / NeuroImage 21 (2004) 311–317
315
Fig. 4. Histological corroboration of in vivo observations. The advantage of GRID to label cells for transplantation is its bimodality allowing identification of MHP36 cells by two independent imaging modalities, MRI and fluorescent histology. On the T2-weigted MR image (A), a hypointense region in the peri-lesion area indicates an infiltration of GRID-labelled MHP36 cells (B). Fluorescent histology of this area indicated a glial scar (as determined by GFAP labelling) sealing off the cavity (B), with a higher magnification image ( 400) indicating that indeed transplanted cells infiltrated this peri-lesion area (C – E). At 14 days, the injection tract was also still visible on the MRI scans (F) and fluorescent histology indicated that the GRID-bearing grafted cells were still lining the injection tract (F – H).
around the injection tract did, however, co-label with GFAP, suggesting that some of these cells differentiated into astrocytes. In animals with stroke damage, transplanted cells co-labelled with GFAP and NeuN, indicating that in damaged brains, cells differentiated into both astrocytes and neurons (Fig. 5). There was
little difference in neuronal differentiation between the striatum surrounding the injection tract on the contralateral side and the remaining striatum around the ischaemic lesion. However, there were no NeuN-positive cells found in the direct vicinity of the lesion, that is, within the glial scar.
Fig. 5. Phenotypic differentiation of transplanted cells. Transplanted MHP36 cells showed co-labelling of GFAP (for astrocytes) and NeuN (for neurons) with GRID indicating that labelling of cells with a contrast agent did not impair their ability to differentiate into appropriate phenotypes.
316
M. Modo et al. / NeuroImage 21 (2004) 311–317
Within the glial scar, GRID-labelled cells co-localised with GFAP, with only few transplanted cells not showing an astrocytic differentiation. Since stroke-damaged animals were only transplanted 3 months following the lesion, it can be assumed that these transplanted cells infiltrated the glial scar after it had formed and differentiated into astrocytes. Also, in the peri-lesion area, some of the transplanted cells showed a morphology characteristic of reactive astrocytes indicating that transplanted cells participated in on-going, lesion-associated astrocytosis.
Discussion Until recently, investigating the migration of neural stem cells has been restricted to post-mortem histological studies inferring rather than capturing the dynamic nature of migrating cells. Preloading of cells with contrast agents allows both ex vivo and in vivo visualisation of transplanted cells in damaged brains (Bulte et al., 2002) and therefore can uncover the extensive migration of neural stem cells. The use of a bimodal contrast agent, such as GRID, additionally provides a direct corroborative visualisation of transplanted cells by post-mortem fluorescent histology. The serial MRI of transplanted brains therefore may provide an unique insight into the dynamic migration of neural stem cells. Damage-induced transhemispheric stem cell migration In rats with stroke damage, transplanted neural stem cells will migrate from the contralateral intact hemisphere to the site of damage in the ipsilateral hemisphere (Modo et al., 2002c; Veizovic et al., 2001). In the present study, serial in vivo MRI provided a dynamic assessment of neural stem cell migration within 2 weeks following transplantation. The results indicate that by 1 day post-grafting, the injection tract was still packed with contrast agent-loaded neural stem cells which could be differentiated from the surrounding striatum by MRI. During the first week post-grafting, the cells diffused/migrated out from the injection tract and by 7 days post-grafting, cells migrating along the corpus callosum to the damaged hemisphere were apparent on MR images. Cells in the striatum did not only remain in the injection tract, but also populated the surrounding striatum. The injection tract therefore could be differentiated from the rest of the striatum by MRI during the 2 weeks post-grafting. During the second week following grafting, more cells migrated along the corpus callosum causing a significant change in MR image contrast within this structure. The migration along the corpus callosum by 7 days appears to be the product of transplanted cells from both the cortical and striatal deposit with both showing dispersion away from the site of implantation. The distribution of transplanted cells by 14 days following transplantation was corroborated by fluorescent microscopy which confirmed the heterogeneous distribution of cells in the contralateral striatum. This suggests that cells do not merely disperse in a radial fashion from the injection tract, but move along particular routes/pathways. Although the movement out of the injection tract can differ from animal to animal, the movement of cells toward the corpus callosum on the contralateral hemisphere and the migration along the corpus callosum is very similar across all lesioned animals. At 14 days post-transplantation, a homogenous gradient
of cells can be observed along the corpus callosum in the vicinity of the lesion. Clusters of cells are mainly found in the initial contralateral and medial aspect of the corpus callosum, although many overlapping clusters can give the appearance of a homogenous distribution of cells along this pathway on MRI. These results are in accordance with previous histological analyses of neural stem cell migration from later time points (Modo et al., 2002c; Veizovic et al., 2001) suggesting that approximately one third of transplanted cells will migrate to the ipsilateral hemisphere after contralateral implantation. At later time points, transplanted cells were also found to extensively populate ipsi- and contralateral cortical areas. The present study indicates that most of these cells arrive within the first 2 weeks following transplantation. Still, in comparison to previous histological studies (Modo et al., 2002c; Veizovic et al., 2001), it can be assumed that some migration to populate cortical areas might follow the transhemispheric migration of cells. A more extensive study at later time points is needed to answer this question. These data echo recent findings by Hoehn et al. (2002). In that study, they observed the migration of embryonic stem cells along the corpus callosum by means of pre-labelling cells for transplantation with iron oxide particles (pre-treated with a transfection agent). However, the migrating embryonic stem cells appear to be mainly derived from the cortical deposits at the intersection of the corpus callosum and cerebral cortex, with little observed dispersion of embryonic cells away from the striatal deposit which could be observed in the present study after grafting of neural stem cells. Our present study also demonstrates the advantage of a bimodal gadolinium-based contrast agent to visualise the migration of a large number of cells along the corpus callosum both in vivo and ex vivo to corroborate the presence of the contrast agent within transplanted cells and to determine their differentiation by fluorescent immunohistochemistry. Moreover, the incorporation of GRID into cells did not necessitate the use of a transfection agent and therefore provides a more efficient approach to label cells for transplantation. Zhang et al. (2003) recently also reported the visualisation of transplanted subventricular zone cells derived from neurospheres in animals with stroke damage which migrated toward the ischaemic lesion after intracisternal implantation. It is noteworthy, that although there are substantial differences in the cellular properties of embryonic, subventricular zone cells and neural stem cells, these three different sources of grafted cells have all shown migration toward the ischaemic lesion. Therefore, in vivo MRI has the potential to further assess differences in migration and effectiveness between these different populations of cells to determine optimal implantation sites for stem cell therapy. Likewise, it will be important to determine why not all cells migrate/ disperse out of the implantation site during the same time course and its significance for behavioural recovery. From the present data, it can be concluded that the stroke damage functions as an attractant to the neural stem cells, since in intact brains, no significant migration (apart from a few scattered cells in the surrounding striatum) away from the injection tract was observed. The migration and differentiation of stem cells therefore appears to depend on the presence of ‘damaged’ tissue. Neural stem cell migration and behavioural recovery At present, it is still unclear if the migration of stem cells after transplantation has any relevance to the functional recovery ob-
M. Modo et al. / NeuroImage 21 (2004) 311–317
served in animals with stroke damage. We recently reported that neural stem cells transplanted contralateral to the lesioned hemisphere resulted in significant behavioural improvements between 4 and 6 weeks post-grafting relative to animals with stroke damage (Modo et al., 2002c). If neural stem cells arrive at the target within 2 weeks following transplantation, the 2 – 4 weeks discrepancy between arriving in the peri-lesion area and behavioural changes could be a reflection of delayed cellular differentiation and integration into host circuitry. However, differentiation per se, could not account for this temporal discrepancy as differentiated (GFAPand/or NeuN-positive) cells can be found by 2 weeks following transplantation. It can therefore be speculated that other anatomical/functional changes are required during this 2- to 4-week period which are necessary for facilitated behavioural recovery. Recruitment of transplanted cells might, however, be an ongoing process with neural stem cells still in the corpus callosum by 3.5 months after transplantation and a continued expression of proteins associated with plasticity or neuroprotection which in turn could support cellular migration and the process of behavioural recovery (Modo et al., 2002c, in press; Ourednik et al., 2002). Therefore, a dynamic, serial assessment of cellular migration and behaviour within subjects will be instrumental in disentangling the functional anatomy underlying behavioural recovery after neural stem cell transplantation. Combining the in vivo assessment of graft survival and the functional anatomical deficits associated with behavioural impairments as determined by fMRI and behavioural analyses (Dijkhuisen et al., 2003) will provide a better understanding of stem cell-mediated neuroplasticity. An individualization of neural stem cell therapy, by determining how implantation site, extent of damage, and other relevant in vivo factors affect the migration of cells in normal and/or damaged brains, will open a new vista in our study of transplant efficacy.
Acknowledgments The authors thank the UK Medical Research Council (MRCROPA grant G0000966) for their generous support. The MRI spectrometer was provided by the University of London Intercollegiate Research Service scheme and is located at Queen Mary College London managed by Dr. Alisdair Preston. The authors also thank ReNeuron for the use of their confocal microscope and the use of MHP36 cells.
References Bjo¨rklund, A., Lindvall, O., 2000. Cell replacement therapies for central nervous system disorders. Nat. Neurosci. 3, 537 – 544. Bulte, J.W., Duncan, I.D., Frank, J.A., 2002. In vivo magnetic resonance tracking of magnetically labelled cells after transplantation. J. Cereb. Blood Flow Metab. 22 (8), 899 – 907.
317
Dijkhuisen, R.M., Singhal, A.B., Mandeville, J.B., Wu, O., Halpern, E.F., Finkelstein, S.P., Rosen, B.R., Lo, E.H., 2003. Correlation between brain reorganization, ischemic damage, and neurologic status after transient focal cerebral ischemia in rats: a functional magnetic resonance imaging study. J. Neurosci. 23 (2), 510 – 517. Gray, J.A., Grigoryan, G., Virley, D., Patel, S., Sinden, J.D., Hodges, H., 2000. Conditionally immortalized, multipotential and multifunctional neural stem cell lines as an approach to clinical transplantation. Cell Transplant. 9, 153 – 168. Hoehn, M., Ku¨stermann, E., Blunk, J., Wiedermann, D., Trapp, T., Wecker, S., Fo¨cking, M., Arnold, H., Heschler, J., Fleischmann, B.K., Schwindt, C., Bu¨hrle, C., 2002. Monitoring of implanted stem cell migration in vivo: a highly resolved in vivo magnetic resonance imaging investigation of experimental stroke in rat. Proc. Natl. Acad. Sci. 99 (25), 16267 – 16272. Hu¨ber, M.M., Staubli, A.B., Kustedjo, K., Gray, M.H.B., Shih, J., Fraser, S.E., Jacobs, R.E., Meade, T.J., 1998. Fluorescently detectable magnetic resonance imaging agents. Bioconjug. Chem. 9, 242 – 249. Longa, E.Z., Weinstein, P.R., Carlson, S., Cummins, R., 1989. Reversible middle cerebral artery occlusion without craniectomy in rats. Stroke 20, 84 – 91. Modo, M., Stroemer, R.P., Tang, E., Veizovic, T., Sowinski, P., Hodges, H., 2000. Neurological sequelae and long-term behavioural assessment of rats with transient middle cerebral artery occlusion. J. Neurosci. Methods 104, 99 – 109. Modo, M., Cash, D., Mellodew, K., Williams, S.C.R., Fraser, S.E., Meade, T.J., Price, J., Hodges, H., 2002a. Tracking transplanted stem cell migration using bifunctional, contrast agent-enhanced, magnetic resonance imaging. NeuroImage 17 (2), 803 – 811. Modo, M., Rezaie, P., Heuschling, P., Patel, S., Male, D.K., Hodges, H., 2002b. Transplantation of stem cells in a rat model of stroke: assessment of short-term graft survival and acute host immunological response. Brain Res. 958, 70 – 82. Modo, M., Stroemer, R.P., Tang, E., Patel, S., Hodges, H., 2002c. Implantation site of stem cell grafts affects behavioural recovery from stroke damage. Stroke 33 (9), 2270 – 2278. Modo, M., Hopkins, K., Virley, D., Hodges, H., 2003. Apolipoprotein E expression after transplantation of neural stem cells in a rat model of stroke. Exp. Neurol. 183 (2), 320 – 329. Ourednik, J., Ourednik, V., Lynch, W.P., Schachner, M., Snyder, E.Y., 2002. Neural stem cells display an inherent mechanism for rescuing dysfunctional neurons. Nat. Biotechnol. 20, 1103 – 1110. Park, K.I., Ourendik, J., Ourendik, V., Taylor, R.M., Aboody, K.S., Auguste, K.I., Lachyankar, M.B., Redmond, D.E., Snyder, E.Y., 2002. Global gene and cell replacement strategies via stem cells. Gene Ther. 9, 613 – 624. Sinden, J.D., Rashid-Doubell, F., Kershaw, T.R., Nelson, A., Chadwick, A., Jat, P.S., Noble, M.D., Hodges, H., Gray, J.A., 1997. Recovery of spatial learning by grafts of a conditionally immortalized hippocampal neuroepithelial cell line into the ischaemia-lesioned hippocampus. Neuroscience 81, 599 – 608. Veizovic, T., Beech, J.S., Stroemer, R.P., Watson, W.P., Hodges, H., 2001. Resolution of stroke deficits following contralateral grafts of conditionally immortal neuroepithelial stem cells. Stroke 32, 1012 – 1019. Zhang, Z.G., Jiang, Q., Zhang, R., Zhang, L., Wang, L., Zhang, L., Arniego, P., Ho, K.-L., Chopp, M., 2003. Magnetic resonance imaging and neurosphere therapy of stroke in rat. Ann. Neurol. 53, 259 – 263.
Mapping transplanted stem cell migration after a stroke: a serial, in vivo magnetic resonance imaging study Michel Modo, a,* Karen Mellodew, b Diana Cash, a Scott E. Fraser, c Thomas J. Meade, d Jack Price, b and Steven C.R. Williams a a
Neuroimaging Research Group—Neurology P042, Institute of Psychiatry, King’s College London, London SE5 8AF, UK Department of Neuroscience, Institute of Psychiatry, King’s College London, London SE5 8AF, UK c Beckman Institute, California Institute of Technology, Pasadena, CA 91125, USA d Departments of Chemistry, Biochemistry, Molecular and Cell Biology, Neurobiology, and Physiology, Northwestern University, Chicago, IL 60208-3500, USA b
Received 5 June 2003; revised 22 August 2003; accepted 22 August 2003
Preferential migration of stem cells toward the site of a lesion is a highly desirable property of stem cells that allows flexibility in the site of graft implantation in the damaged brain. In rats with unilateral stroke damage, neural stem cells transplanted into the contralateral hemisphere migrate across to the lesioned hemisphere and populate the area around the ischaemic infarct. To date, the migration of neural stem cells in the damaged brain has been mainly inferred from snapshot histological images. In this study, we demonstrate that by prelabelling neural stem cells with the bimodal contrast agent GadoliniumRhodamIne Dextran [GRID, detectable by both magnetic resonance imaging (MRI) and fluorescent microscopy], the transhemispheric migration of transplanted neural stem cells contralateral to a stroke lesion can be followed in vivo by serial MRI and corroborated by subsequent histological analyses. Our results indicate that neural stem cells migrated from the injection tract mainly along the corpus callosum within 7 days of transplantation and extensively re-populated the peri-lesion area by 14 days following implantation. In contrast, neural stem cells transplanted into sham controls did not show any substantial migration outside of the injection tract, suggesting that the transcallosal migration observed in the stroke-lesioned animals is due to neural stem cells being attracted by the lesion site. In vivo tracking of the migration of neural stem cells responding to damage will greatly enhance our understanding of optimal transplantation strategies as well as how neural stem cells promote functional and anatomical recovery in neurological disorders. D 2003 Elsevier Inc. All rights reserved. Keywords: MCAo; Neural stem cells; Migration; Contrast agent; Cell tracking; Neural transplants; MRI
* Corresponding author. Neuroimaging Research Group—Neurology P042, Institute of Psychiatry, King’s College London, De Crespigny Park, London SE5 8AF, UK. Fax: +44-207-848-0055. E-mail address: [email protected] (M. Modo). Available online on ScienceDirect (www.sciencedirect.com.) 1053-8119/$ - see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/j.neuroimage.2003.08.030
Introduction The clinical application of stem cell therapy to remedy brain damage is dependent on the potential of grafts to promote sustained recovery after transplantation (Bjo¨rklund and Lindvall, 2000; Park et al., 2002). At present, investigating the relationship between anatomical change and behavioural recovery is hindered by the necessity to use post-mortem tissue to determine how neural stem cells contribute to behavioural recovery. The migratory properties and the seamless integration of neural stem cells in damaged brains are considered to be beneficial to behavioural recovery. However, to date there is no indication as to how the migration or integration of neural stem cells into the damaged host parenchyma relates to behavioural recovery. Several recent studies have demonstrated that incorporation of contrast agents into cells destined for cellular therapy before transplantation can afford their in vivo identification by magnetic resonance imaging (MRI) (Bulte et al., 2002; Hoehn et al., 2002; Zhang et al., 2003). Recently, we described the use of a bimodal contrast agent Gadolinium-RhodamIne Dextran (GRID), detectable both in vivo by MRI and subsequently at post-mortem by fluorescent microscopy, as a novel tool to track migrating neural stem cells (Modo et al., 2002a). This approach now allows us to investigate how neural stem cells migrate and integrate into various brain regions in the same individual over time. Neural stem cells transplanted into the contralateral hemisphere of rats with stroke damage, not only populate areas around the injection tract, but show transhemispheric migration to the damaged hemisphere to integrate into areas around the lesion site (Modo et al., 2002c; Veizovic et al., 2001). However, post-mortem histological analyses only provide a single ‘‘snapshot’’ of the presence of cells in the brain. Although transplanted cells were found in the corpus callosum in these published studies, it is unclear whether this tract provided a channel for migration. From ex vivo data, it is also debatable if areas of damage function as ‘attractants’ to neural stem cells or if there is a process of nondirected dispersion followed by selective retention. Histological data also cannot resolve the issue if migrating neural stem cells
312
M. Modo et al. / NeuroImage 21 (2004) 311–317
need to populate an area of damage before behavioural recovery can ensue. The use of in vivo neuroimaging to address the relationship between anatomical/neurochemical and behavioural changes has the potential to further develop and refine neural stem cell therapy. Here we report the serial in vivo tracking of transhemispheric neural stem cell migration in rats with chronic brain damage following middle cerebral artery occlusion (MCAo, a rat model of stroke). To distinguish transplanted cells from host cells, neural stem cells were pre-labelled in vitro with the bimodal contrast agent GRID allowing the identification by MRI for serial in vivo evaluation of migration and by post-mortem fluorescent microscopy to validate the in vivo observations. Sham control animals were also followed serially to determine if neural stem cell migration is a consequence of focal brain damage attracting transplanted cells or if transplanted cells merely diffuse from the injection tract.
Materials and methods Animals Sprague – Dawley rats (Charles River, UK) were acclimatised for at least a week before surgery. Animals weighed between 250 and 270 g upon arrival with maintenance on a 10-h light/14-h dark schedule. All procedures were in accordance with the UK Animals (Scientific) Procedures Act 1986 and the ethical review process of Queen Mary and Westfield College, University of London. Middle cerebral artery occlusion Animals between 280 and 330 g were either allocated for sham (n = 3) or 60 min of transient MCAo surgery (n = 3) in the right hemisphere (Longa et al., 1989). Briefly, animals were anaesthetized with halothane (4% induction, 2% maintenance) in a mixture of O2 and N2O (30:70). Ligatures were placed on the external and common carotid to stop the flow of blood to the internal carotid artery. An aneurysm clip was placed on the internal carotid before the common carotid was opened, to allow insertion of a thread up to the aneurysm clip. The tip of the thread was advanced 18 – 20 mm from the cervical carotid bifurcation or until reaching resistance from the ostium of the middle cerebral artery in the circle of Willis. For MCAo, the thread was left in place for 60 min, whereas for sham surgery, the thread was immediately removed. Following occlusion, animals were tested for spontaneous circling and forelimb flexion (Modo et al., 2000). Occluded animals were reanaesthetized and the thread was removed.
day 14 (Sinden et al., 1997). These cells proliferate at 33jC providing an ample source of cells for transplantation, but cease proliferation at 37jC. In the damaged CNS, MHP36 cells show site-appropriate phenotypic differentiation (Modo et al., 2002c) and are therefore ideal candidates for brain repair (Gray et al., 2000). Frozen MHP36 cells (passage 42) were thawed and cultured for at least one passage before transplantation according to a standard protocol (Sinden et al., 1997). During labelling, MHP36 cells were grown in proliferative conditions (33jC with bFGF and IFN-g) with GRID added to the media. Addition of GRID to the growth medium resulted in a final concentration of 45 AM of Gd3 + and 2.73 AM of tetramethylrhodamine (chelated in one molecule) per millilitre of media. After 6 h of incubation with GRID, the medium was discarded and cells were removed from the flask by adding Hank’s balanced salt solution (HBSS, Gibco) without Ca2 + and Mg2 +. The suspension was centrifuged (1500 rpm for 5 min) and cells were resuspended for transplantation (25,000 cells/Al) with 1 mM Nacetyl-L-cysteine (NAC, Sigma) in HBSS (Gibco). Cell viability for transplantation, as determined by trypan blue (Sigma) exclusion, was >90% after GRID-labelling. At the end of grafting, cell viability was f85% as determined from the same vial used for transplantation by trypan blue exclusion. Grafting Three months after MCAo, animals were transplanted in the left (contralateral) hemisphere to probe their migration toward the stable ischaemic lesion in the right (ipsilateral hemisphere). Anesthesia was induced by intraperitoneal (i.p.) injection of a mixture of 0.1 mg medetomidine hydrochloride (Dormitor, C-Vet Products), 5.5 mg ketamine hydrochloride (Ketaset, C-Vet Products), in 0.1 ml of 0.9% saline (Baxter) per 100 g body weight. Animals were placed in a stereotaxic frame and an incision was made exposing bregma. Burr holes were drilled contralateral (Site 1. AP = + 0.7, L = 2, V = 5.5/ 2; Site 2. AP = 0.3, L = 3, V = 5.5/ 2.5) to the ischaemic lesion. Each deposit consisted of 2 Al at a speed of 1 Al/min. The syringe was left in place for 2 min after injection to allow dispersion of the cells. A total of 2 105 cells with a total volume of 8 Al was injected per animal. No immunosuppression was given since MHP36 cells survive in immunocompetent rats for at least the first 2 weeks following transplantation (Modo et al., 2002b). After grafting, animals were injected with 0.1 mg atipamezole hydrochloride (Anti-Sedan, C-Vet Products) in 0.1 ml of 0.9 saline per 100 g body weight i.p. to counteract anaesthesia followed by 2.5 ml of glucosaline (subcutaneous) in two sites. After suturing, a local analgesic (lignocaine hydrochloride) was applied to minimize post-operative discomfort of the wound.
Preparation of neural stem cells Magnetic resonance imaging The GRID chelate, consisting of gadolinium (Gd-DTPA) and rhodamine particles linked by a dextran polymer, was conjugated according to published methods (Hu¨ber et al., 1998) to yield a single molecule (total molecular weight 16,600) which is acting as a bimodal contrast agent detectable by both MRI and fluorescent histology. GRID was diluted (1:1) with distilled H2O at least 24 h before cell labelling (as described in Modo et al., 2002a). Conditionally immortalized neural stem cells from the Maudsley Hippocampal Clone 36 (MHP36) cell line are derived from the neuroepithelium of the H-2Kb-tsA58 immortomouse at embryonic
To acquire in vivo MR images, animals were anaesthetized using isoflurane (4% induction, 2% maintenance) in a O2/N2O (30:70) mixture supplied via a face mask in a head holder to minimize movement artefacts. All MRI was performed on a small animal, horizontal bore, 4.7T NMR system (Oxford Systems) controlled by a UNITYInova-200 imaging console (Varian). A quadrature birdcage radio frequency coil with 63 mm internal diameter (Varian) was used for signal transmission and reception. The raw time domain data were Fourier transformed using locally
M. Modo et al. / NeuroImage 21 (2004) 311–317
developed software. The choice of imaging protocol was determined by the need for a high spatial resolution as well as the distinction between the host tissue and the gadolinium-bearing graft. Our imaging protocol consisting of coronal T1-weighted (TR = 500 ms, TE = 15 ms), T2-weighted (TR = 4000 ms, TE = 45 ms), and proton density (PD)-weighted (TR = 4000 ms, TE = 15 ms) spin echo scans with a 2562 matrix size, 4 averages per phase encoding step, 30 contiguous 500-Am-thick slices, and a field of view of 2.5 2.5 cm was adopted to afford an in plane spatial resolution of 248 Am. Histology After scanning, animals were overdosed with pentobarbitone sodium (Pentoject, Animal Care) for perfusion. The descending aorta was clamped to restrict blood flow to the upper body half. Rats were pre-washed transcardially with heparinized saline (0.9% NaCl in dH2O containing 5000 units of heparin, CP Pharmaceuticals, per litre) followed by 4% paraformaldehyde (PFA) in 0.1 M phosphate-buffered saline (PBS, Sigma). Brains were removed and placed overnight in 4% PFA at room temperature before being cryoprotected in 30% sucrose for at least 1 week before cutting. Sections of 30 Am thickness were cut on a cryostat microtome (Leica) at 24jC and collected free-floating in sucrose. For immunohistochemical staining, sections were washed 3 5 min in 0.01 M PBS before the application of the primary antibody used for identification of neurons (monoclonal mouse anti-NeuN, 1:1000, Chemicon) or astrocytes (polyclonal goat anti-GFAP, 1:4000, Chemicon). Primary antibodies were applied overnight at room temperature. Sections were washed 3 5 min in PBS and then incubated with a cross-adsorbed secondary antibody (goat anti-mouse Alexa488 for NeuN, and rabbit anti-goat Alexa488 for GFAP, 1:500, Molecular Probes) for 45 min at room temperature. Sections were washed 3 5 min in PBS before coverslips were applied with Vectashield for immunofluorescence (Vector). Macroscopic analyses of graft migration and integration were conducted on a fluorescent microscope (Nikon), whereas phenotypic differentiation of transplanted cells required the use of a confocal microscope (Leica). A confocal analysis of the overlay of GRID and phenotype markers is paramount to ensure that both markers co-localise within the same cells and are not the consequence of juxtaposed markers leading to an erroneous identification of phenotypic differentiation.
Results Lack of migration in sham controls MRI of sham control brains provided evidence of how GRIDlabelled transplants affect the MR signal over time without a possible confound of a lesion environment. GRID-labelled MHP36 cells were detected 1 day post-grafting as an elongated line on T2-weighted images reflecting the injection tract (Fig. 1). The baseline image of sham control brains, 1 day pre-transplantation, provides evidence that the changes observed on the MR scan are a consequence of the injection of GRID-labelled cells. From the MR sequences used in this study, T2-weighted MR images provided the best anatomical discrimination and also allowed a satisfactory distinction between grafted cells and host
313
tissue (Fig. 2). Although PD-weighted scans also allowed differentiation of the grafted cells in the injection tract from the host striatum, they provided less anatomical detail and were therefore considered inferior to T2-weighted images to study the migration of transplanted cells. Only minimal anatomical information could be derived from the T1-weighted MR images and GRID-labelled cells were mostly indistinguishable from the host tissue. T2-weighted MR images were therefore used to study how transplants evolved over time. The assessment of T2-weighted images at 1 day, 7 days, and 14 days following transplantation indicated that MHP36 cells did not migrate (or become passively dispersed) out of the injection tract (Fig. 1). Corroboration of these in vivo observations were provided by post-mortem fluorescent histology (of the rhodamine signal) indicating that grafted cells were densely packed within the injection tract, but were not found in regions distant from the injection tract. A few cells labelled with GRID, however, could be found in the immediate vicinity of the injection tract. Stroke-induced transhemispheric neural stem cell migration Animals with stroke damage were imaged 1 day before transplantation (approximately 3 months following ischaemia) to visualise the chronic lesion providing a baseline to which graft-induced changes were compared (Fig. 3). At 1 day post-grafting, an injection tract was visible from the top of the cortex to the medial part of the striatum. In comparison to sham controls, the injection tracts were located slightly more lateral due to a shift of the brain to the right side caused by the cavitation of the ischaemic lesion. Injection tracts appeared, as in control animals, as dark elongated lines packed with GRID-labelled cells causing a hypointense signal differentiating the injection tract from the host brain. By 7 days post-transplantation, a substantial number of cells had migrated out from the injection tract and induced a change in signal of the contralateral striatum rendering it patchy around the injection tract. However, the injection tract could still be identified, albeit the differentiation from host tissue was less obvious than at 1 day post-grafting. Furthermore, along the corpus callosum, some pockets of cells loaded with GRID could be differentiated from the transhemispheric white matter tract. This indicated that GRIDlabelled transplanted cells used the corpus callosum for transhemispheric migration. At this time point, there was no clear change in signal in the peri-lesion area that would indicate that transplanted cells already arrived at the lesion site. At the final time point, 14 days following contralateral implantation, a significant number of cells had migrated to the lesion along the corpus callosum and populated the peri-lesion area. A clear change in intensity in the region of the lesion close to the corpus callosum was observed on MR images reflecting the accumulation of GRID-labelled MHP36 cells in the peri-lesion area. The contralateral striatum still exhibited some hypointensities reflecting the presence of dispersed transplanted cells. Most clearly, the injection tract could still be detected at 14 days post-grafting, suggesting that not all cells had migrated/diffused out of the implantation site. The temporal dynamics of neural stem cell migration from the contralateral hemisphere therefore suggest that by 1 day following transplantation, cells could be mainly found within the injection tract with no significant dispersion out of the tract, but by 7 days post-grafting, cells disperse in the surrounding striatum and along the corpus callosum. By 14 days post-grafting,
314
M. Modo et al. / NeuroImage 21 (2004) 311–317
Fig. 1. Serial in vivo visualisation of stem cell transplants in control animals. The neural stem cell transplant in a control animal showed no significant migration away from the implantation tract. Despite lack of migration from the implantation site, the stem cell graft (arrows) showed good graft survival over the 14 days of in vivo assessment by T2-weighted MR images and post-mortem by fluorescent histology.
Fig. 2. Multiparametric MRI of stem cell transplants. T1-weighted images were very poor in detecting the gadolinium-bearing grafted cells (arrows), whereas T2-weighted images provided a satisfactory contrast to distinguish grafted cells from host brain. PD-weighted images also allowed the identification of grafted cells, but provided less anatomical detail than T2-weighted images.
a substantial number of transplanted cells populated the ipsilateral peri-lesion area (contralateral to the implantation site), but some cells remained within the injection tract and along the corpus callosum. Histological analyses corroborated the in vivo observation that transplanted cells migrated from the contralateral implantation site to populate the peri-lesion area (Fig. 4). In the peri-lesion area, a glial scar surrounded the cavity resulting from the ischaemic insult and transplanted cells were found both within this glial scar and within the non-gliotic tissue surrounding the lesion. However, histologically GRID-labelled transplanted cells were also present at lower concentrations in regions of the brain in which no significant signal change was detected on any of the MR images. These small clusters of cells could be found in the contralateral striatum away from the injection tract. GRID-labelled cells also lined the injection tract that could still be detected in the MR images 14 days post-injection. At the bottom
of the injection tract, the deposit showed a marked lateral dispersion suggesting that most of the cells from the deposit had migrated away from their site of implantation. Within the injection tract, cells were still strongly labelled with GRID, whereas cells at more remote sites, such as the ipsilateral hemisphere, showed less GRID fluorescence, suggesting that these cells had undergone some cell division that had diluted the total GRID. The wider dispersion of MHP36 cells with low amounts of GRID in remote locations (such as the ipsilateral hemisphere) therefore affords a better identification of the graft (at least visually) compared to a high concentration of GRID inside the cells in the densely packed injection tract. Phenotypic differentiation of transplanted stem cells In sham control animals, cells within the injection tract did not, in the main, co-localise with either GFAP (as an astrocyte marker) or NeuN (as a neuronal marker), but a few cells in and especially
Fig. 3. Serial in vivo tracking of stroke-induced transhemispheric stem cell migration. One day after transplantation of MHP36 cells, an injection tract (arrow) packed with GRID-labelled cells can be detected in the medial striatum in comparison to pre-transplant T2-weighted MR images of the same animal. Over the following 14 days post-grafting in the same animal, packs of transplanted cells migrated along the corpus callosum to reach the site of damage (arrows). This indicates that migration of stem cells to the contralateral hemisphere mainly relies on transcallosal migration. However, some cells remained within the injection tract in the contralateral hemisphere (arrow).
M. Modo et al. / NeuroImage 21 (2004) 311–317
315
Fig. 4. Histological corroboration of in vivo observations. The advantage of GRID to label cells for transplantation is its bimodality allowing identification of MHP36 cells by two independent imaging modalities, MRI and fluorescent histology. On the T2-weigted MR image (A), a hypointense region in the peri-lesion area indicates an infiltration of GRID-labelled MHP36 cells (B). Fluorescent histology of this area indicated a glial scar (as determined by GFAP labelling) sealing off the cavity (B), with a higher magnification image ( 400) indicating that indeed transplanted cells infiltrated this peri-lesion area (C – E). At 14 days, the injection tract was also still visible on the MRI scans (F) and fluorescent histology indicated that the GRID-bearing grafted cells were still lining the injection tract (F – H).
around the injection tract did, however, co-label with GFAP, suggesting that some of these cells differentiated into astrocytes. In animals with stroke damage, transplanted cells co-labelled with GFAP and NeuN, indicating that in damaged brains, cells differentiated into both astrocytes and neurons (Fig. 5). There was
little difference in neuronal differentiation between the striatum surrounding the injection tract on the contralateral side and the remaining striatum around the ischaemic lesion. However, there were no NeuN-positive cells found in the direct vicinity of the lesion, that is, within the glial scar.
Fig. 5. Phenotypic differentiation of transplanted cells. Transplanted MHP36 cells showed co-labelling of GFAP (for astrocytes) and NeuN (for neurons) with GRID indicating that labelling of cells with a contrast agent did not impair their ability to differentiate into appropriate phenotypes.
316
M. Modo et al. / NeuroImage 21 (2004) 311–317
Within the glial scar, GRID-labelled cells co-localised with GFAP, with only few transplanted cells not showing an astrocytic differentiation. Since stroke-damaged animals were only transplanted 3 months following the lesion, it can be assumed that these transplanted cells infiltrated the glial scar after it had formed and differentiated into astrocytes. Also, in the peri-lesion area, some of the transplanted cells showed a morphology characteristic of reactive astrocytes indicating that transplanted cells participated in on-going, lesion-associated astrocytosis.
Discussion Until recently, investigating the migration of neural stem cells has been restricted to post-mortem histological studies inferring rather than capturing the dynamic nature of migrating cells. Preloading of cells with contrast agents allows both ex vivo and in vivo visualisation of transplanted cells in damaged brains (Bulte et al., 2002) and therefore can uncover the extensive migration of neural stem cells. The use of a bimodal contrast agent, such as GRID, additionally provides a direct corroborative visualisation of transplanted cells by post-mortem fluorescent histology. The serial MRI of transplanted brains therefore may provide an unique insight into the dynamic migration of neural stem cells. Damage-induced transhemispheric stem cell migration In rats with stroke damage, transplanted neural stem cells will migrate from the contralateral intact hemisphere to the site of damage in the ipsilateral hemisphere (Modo et al., 2002c; Veizovic et al., 2001). In the present study, serial in vivo MRI provided a dynamic assessment of neural stem cell migration within 2 weeks following transplantation. The results indicate that by 1 day post-grafting, the injection tract was still packed with contrast agent-loaded neural stem cells which could be differentiated from the surrounding striatum by MRI. During the first week post-grafting, the cells diffused/migrated out from the injection tract and by 7 days post-grafting, cells migrating along the corpus callosum to the damaged hemisphere were apparent on MR images. Cells in the striatum did not only remain in the injection tract, but also populated the surrounding striatum. The injection tract therefore could be differentiated from the rest of the striatum by MRI during the 2 weeks post-grafting. During the second week following grafting, more cells migrated along the corpus callosum causing a significant change in MR image contrast within this structure. The migration along the corpus callosum by 7 days appears to be the product of transplanted cells from both the cortical and striatal deposit with both showing dispersion away from the site of implantation. The distribution of transplanted cells by 14 days following transplantation was corroborated by fluorescent microscopy which confirmed the heterogeneous distribution of cells in the contralateral striatum. This suggests that cells do not merely disperse in a radial fashion from the injection tract, but move along particular routes/pathways. Although the movement out of the injection tract can differ from animal to animal, the movement of cells toward the corpus callosum on the contralateral hemisphere and the migration along the corpus callosum is very similar across all lesioned animals. At 14 days post-transplantation, a homogenous gradient
of cells can be observed along the corpus callosum in the vicinity of the lesion. Clusters of cells are mainly found in the initial contralateral and medial aspect of the corpus callosum, although many overlapping clusters can give the appearance of a homogenous distribution of cells along this pathway on MRI. These results are in accordance with previous histological analyses of neural stem cell migration from later time points (Modo et al., 2002c; Veizovic et al., 2001) suggesting that approximately one third of transplanted cells will migrate to the ipsilateral hemisphere after contralateral implantation. At later time points, transplanted cells were also found to extensively populate ipsi- and contralateral cortical areas. The present study indicates that most of these cells arrive within the first 2 weeks following transplantation. Still, in comparison to previous histological studies (Modo et al., 2002c; Veizovic et al., 2001), it can be assumed that some migration to populate cortical areas might follow the transhemispheric migration of cells. A more extensive study at later time points is needed to answer this question. These data echo recent findings by Hoehn et al. (2002). In that study, they observed the migration of embryonic stem cells along the corpus callosum by means of pre-labelling cells for transplantation with iron oxide particles (pre-treated with a transfection agent). However, the migrating embryonic stem cells appear to be mainly derived from the cortical deposits at the intersection of the corpus callosum and cerebral cortex, with little observed dispersion of embryonic cells away from the striatal deposit which could be observed in the present study after grafting of neural stem cells. Our present study also demonstrates the advantage of a bimodal gadolinium-based contrast agent to visualise the migration of a large number of cells along the corpus callosum both in vivo and ex vivo to corroborate the presence of the contrast agent within transplanted cells and to determine their differentiation by fluorescent immunohistochemistry. Moreover, the incorporation of GRID into cells did not necessitate the use of a transfection agent and therefore provides a more efficient approach to label cells for transplantation. Zhang et al. (2003) recently also reported the visualisation of transplanted subventricular zone cells derived from neurospheres in animals with stroke damage which migrated toward the ischaemic lesion after intracisternal implantation. It is noteworthy, that although there are substantial differences in the cellular properties of embryonic, subventricular zone cells and neural stem cells, these three different sources of grafted cells have all shown migration toward the ischaemic lesion. Therefore, in vivo MRI has the potential to further assess differences in migration and effectiveness between these different populations of cells to determine optimal implantation sites for stem cell therapy. Likewise, it will be important to determine why not all cells migrate/ disperse out of the implantation site during the same time course and its significance for behavioural recovery. From the present data, it can be concluded that the stroke damage functions as an attractant to the neural stem cells, since in intact brains, no significant migration (apart from a few scattered cells in the surrounding striatum) away from the injection tract was observed. The migration and differentiation of stem cells therefore appears to depend on the presence of ‘damaged’ tissue. Neural stem cell migration and behavioural recovery At present, it is still unclear if the migration of stem cells after transplantation has any relevance to the functional recovery ob-
M. Modo et al. / NeuroImage 21 (2004) 311–317
served in animals with stroke damage. We recently reported that neural stem cells transplanted contralateral to the lesioned hemisphere resulted in significant behavioural improvements between 4 and 6 weeks post-grafting relative to animals with stroke damage (Modo et al., 2002c). If neural stem cells arrive at the target within 2 weeks following transplantation, the 2 – 4 weeks discrepancy between arriving in the peri-lesion area and behavioural changes could be a reflection of delayed cellular differentiation and integration into host circuitry. However, differentiation per se, could not account for this temporal discrepancy as differentiated (GFAPand/or NeuN-positive) cells can be found by 2 weeks following transplantation. It can therefore be speculated that other anatomical/functional changes are required during this 2- to 4-week period which are necessary for facilitated behavioural recovery. Recruitment of transplanted cells might, however, be an ongoing process with neural stem cells still in the corpus callosum by 3.5 months after transplantation and a continued expression of proteins associated with plasticity or neuroprotection which in turn could support cellular migration and the process of behavioural recovery (Modo et al., 2002c, in press; Ourednik et al., 2002). Therefore, a dynamic, serial assessment of cellular migration and behaviour within subjects will be instrumental in disentangling the functional anatomy underlying behavioural recovery after neural stem cell transplantation. Combining the in vivo assessment of graft survival and the functional anatomical deficits associated with behavioural impairments as determined by fMRI and behavioural analyses (Dijkhuisen et al., 2003) will provide a better understanding of stem cell-mediated neuroplasticity. An individualization of neural stem cell therapy, by determining how implantation site, extent of damage, and other relevant in vivo factors affect the migration of cells in normal and/or damaged brains, will open a new vista in our study of transplant efficacy.
Acknowledgments The authors thank the UK Medical Research Council (MRCROPA grant G0000966) for their generous support. The MRI spectrometer was provided by the University of London Intercollegiate Research Service scheme and is located at Queen Mary College London managed by Dr. Alisdair Preston. The authors also thank ReNeuron for the use of their confocal microscope and the use of MHP36 cells.
References Bjo¨rklund, A., Lindvall, O., 2000. Cell replacement therapies for central nervous system disorders. Nat. Neurosci. 3, 537 – 544. Bulte, J.W., Duncan, I.D., Frank, J.A., 2002. In vivo magnetic resonance tracking of magnetically labelled cells after transplantation. J. Cereb. Blood Flow Metab. 22 (8), 899 – 907.
317
Dijkhuisen, R.M., Singhal, A.B., Mandeville, J.B., Wu, O., Halpern, E.F., Finkelstein, S.P., Rosen, B.R., Lo, E.H., 2003. Correlation between brain reorganization, ischemic damage, and neurologic status after transient focal cerebral ischemia in rats: a functional magnetic resonance imaging study. J. Neurosci. 23 (2), 510 – 517. Gray, J.A., Grigoryan, G., Virley, D., Patel, S., Sinden, J.D., Hodges, H., 2000. Conditionally immortalized, multipotential and multifunctional neural stem cell lines as an approach to clinical transplantation. Cell Transplant. 9, 153 – 168. Hoehn, M., Ku¨stermann, E., Blunk, J., Wiedermann, D., Trapp, T., Wecker, S., Fo¨cking, M., Arnold, H., Heschler, J., Fleischmann, B.K., Schwindt, C., Bu¨hrle, C., 2002. Monitoring of implanted stem cell migration in vivo: a highly resolved in vivo magnetic resonance imaging investigation of experimental stroke in rat. Proc. Natl. Acad. Sci. 99 (25), 16267 – 16272. Hu¨ber, M.M., Staubli, A.B., Kustedjo, K., Gray, M.H.B., Shih, J., Fraser, S.E., Jacobs, R.E., Meade, T.J., 1998. Fluorescently detectable magnetic resonance imaging agents. Bioconjug. Chem. 9, 242 – 249. Longa, E.Z., Weinstein, P.R., Carlson, S., Cummins, R., 1989. Reversible middle cerebral artery occlusion without craniectomy in rats. Stroke 20, 84 – 91. Modo, M., Stroemer, R.P., Tang, E., Veizovic, T., Sowinski, P., Hodges, H., 2000. Neurological sequelae and long-term behavioural assessment of rats with transient middle cerebral artery occlusion. J. Neurosci. Methods 104, 99 – 109. Modo, M., Cash, D., Mellodew, K., Williams, S.C.R., Fraser, S.E., Meade, T.J., Price, J., Hodges, H., 2002a. Tracking transplanted stem cell migration using bifunctional, contrast agent-enhanced, magnetic resonance imaging. NeuroImage 17 (2), 803 – 811. Modo, M., Rezaie, P., Heuschling, P., Patel, S., Male, D.K., Hodges, H., 2002b. Transplantation of stem cells in a rat model of stroke: assessment of short-term graft survival and acute host immunological response. Brain Res. 958, 70 – 82. Modo, M., Stroemer, R.P., Tang, E., Patel, S., Hodges, H., 2002c. Implantation site of stem cell grafts affects behavioural recovery from stroke damage. Stroke 33 (9), 2270 – 2278. Modo, M., Hopkins, K., Virley, D., Hodges, H., 2003. Apolipoprotein E expression after transplantation of neural stem cells in a rat model of stroke. Exp. Neurol. 183 (2), 320 – 329. Ourednik, J., Ourednik, V., Lynch, W.P., Schachner, M., Snyder, E.Y., 2002. Neural stem cells display an inherent mechanism for rescuing dysfunctional neurons. Nat. Biotechnol. 20, 1103 – 1110. Park, K.I., Ourendik, J., Ourendik, V., Taylor, R.M., Aboody, K.S., Auguste, K.I., Lachyankar, M.B., Redmond, D.E., Snyder, E.Y., 2002. Global gene and cell replacement strategies via stem cells. Gene Ther. 9, 613 – 624. Sinden, J.D., Rashid-Doubell, F., Kershaw, T.R., Nelson, A., Chadwick, A., Jat, P.S., Noble, M.D., Hodges, H., Gray, J.A., 1997. Recovery of spatial learning by grafts of a conditionally immortalized hippocampal neuroepithelial cell line into the ischaemia-lesioned hippocampus. Neuroscience 81, 599 – 608. Veizovic, T., Beech, J.S., Stroemer, R.P., Watson, W.P., Hodges, H., 2001. Resolution of stroke deficits following contralateral grafts of conditionally immortal neuroepithelial stem cells. Stroke 32, 1012 – 1019. Zhang, Z.G., Jiang, Q., Zhang, R., Zhang, L., Wang, L., Zhang, L., Arniego, P., Ho, K.-L., Chopp, M., 2003. Magnetic resonance imaging and neurosphere therapy of stroke in rat. Ann. Neurol. 53, 259 – 263.