Biomaterials 30 (2009) 2985–2994
Contents lists available at ScienceDirect
Biomaterials journal homepage: www.elsevier.com/locate/biomaterials
The support of neural stem cells transplanted into stroke-induced brain cavities by PLGA particles Ellen Bible a, David Y.S. Chau b, Morgan R. Alexander c, Jack Price a, Kevin M. Shakesheff b, Michel Modo a, * a
Kings College London, Institute of Psychiatry, Department of Neuroscience, London SE5 9NU, UK Division of Advanced Drug Delivery, Centre for Biomolecular Sciences, School of Pharmacy, University of Nottingham, Nottingham NG7 2RD, UK c Laboratory of Biophysics and Surface Analysis, School of Pharmacy, University of Nottingham, Nottingham NG7 2RD, UK b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 December 2008 Accepted 8 February 2009 Available online 10 March 2009
Stroke causes extensive cellular loss that leads to a disintegration of the afﬂicted brain tissue. Although transplanted neural stem cells can recover some of the function lost after stroke, recovery is incomplete and restoration of lost tissue is minimal. The challenge therefore is to provide transplanted cells with matrix support in order to optimise their ability to engraft the damaged tissue. We here demonstrate that plasma polymerised allylamine (ppAAm)-treated poly(D,L-lactic acid-co-glycolic acid) (PLGA) scaffold particles can act as a structural support for neural stem cells injected directly through a needle into the lesion cavity using magnetic resonance imaging-derived co-ordinates. Upon implantation, the neuroscaffolds integrate efﬁciently within host tissue forming a primitive neural tissue. These neuro-scaffolds could therefore be a more advanced method to enhance brain repair. This study provides a substantial step in the technology development required for the translation of this approach. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Stroke Neural stem cells Cell transplantation PLGA Scaffold particle Tissue engineering
Stroke is the most common cause of adult disability in industrialised countries . Unfortunately, to date, no effective treatment is available to reverse the brain damage caused by stroke. Replacement of lost brain tissue by transplantation of neural stem cells (NSCs) is a potential therapeutic avenue that can offer some hope to patients . Experimental studies in animal models have shown that NSCs can bring about some recovery of function  although recovery is never complete. In our own studies for example, NSC transplantation brought about a 30% reduction of lesion volume by 1 year following engraftment, but a large cavity of approximately 80 mm3 remained . In general, transplanted neural stem cells do not create de novo tissue, but integrate into the existing tissue matrix from which cells have been lost . The lesion cavity, the void left by the lost ischaemic tissue, is void of all structural support and cells transplanted into this area typically migrate into surviving host tissue and show site-appropriate differentiation . It can therefore be hypothesised that NSCs might improve tissue repair in the stroke-damaged area if sufﬁcient structural support is provided within the lesion cavity. A more
* Corresponding author at: Centre for the Cellular Basis of Behaviour, The James Black Centre, Kings College London, 125 Coldharbour Lane, London SE5 9NU, United Kingdom. Tel.: þ44 207 848 5315; fax: þ44 207 848 0986. E-mail address: [email protected]
(M. Modo). 0142-9612/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2009.02.012
controlled delivery (i.e. deﬁned number of cells per scaffold, amount of scaffold, reduction of carrier ﬂuids or products), as well as a greater localised retainment of cells, may be achieved by supporting cells with a scaffold rather than introducing them as a simple suspension . For instance, Park et al.  demonstrated that polyglycolic acid (PGA)-based scaffolds containing NSCs can enhance reciprocal interactions of donor and transplanted cells 7 days after a neonatal-induced hypoxia-ischaemia lesion. To date, this is the only demonstration of a transplantation of a scaffold complex into brain damage. The challenges that emerged from this work are to design a biomaterial on which neural stem cells can attach and grow in vitro , but at the same time can be injected through a ﬁne needle into the appropriate locale inside the brain to potentially completely ﬁll the lesion cavity. Delivery of these neuro-scaffolds to the brain therefore needs to be image-guided to ensure that grafts are not injected into intact tissue, which may cause additional brain damage. Ideally, the engineered construct will also adapt to the diverse topology that is present within a lesion cavity. Creating individual particles loaded with neural stem cells on their surface here provides an advantage, as they can adopt the precise shape of the lesion cavity. At the same time, cells can maintain contact with other cells to create connections. This will also reduce the risk of a collagen-like response that could seal off the graft from host tissue. Interaction between host and grafted cells is further
E. Bible et al. / Biomaterials 30 (2009) 2985–2994
Fig. 1. Scanning electron micrographs of fabricated untreated PLGA microparticles at (A) 800 magniﬁcation. (B) Microparticle following deposition with allylamine via plasma polymerization (0.5 KÅ layer, coating: incident power, 20 W; reﬂected power, 75% of MHP36 cells (Fig. 2). Although there was a signiﬁcant difference in the percentage of cells expressing nestin on particles (P < 0.05), this reﬂects the difﬁculty in counting cytoplasmic stains in 3 dimensions (Suppl. Fig. 2), where portions of the cells appear out of focus and hence were not enumerated. Overall, MHP36 cells adapted well to the PLGA particles (Fig. 2E) with >90% of cells viable (Fig. 2F) whilst showing continual proliferation (Fig. 2G). Hence a sufﬁcient number of cells can be attached to the PLGA particles to almost fully occupy its surface area (Suppl. Fig. 2) without affecting their stem cells status, viability or proliferation.
2.3. Lesion localization and targeted delivery Neuro-scaffolds need to be accurately targeted to the brain in order to deliver the injectate directly to the core of the lesion and to avoid additional tissue damage that would result from such a large volume. To achieve this, we used MRI prior to transplantation to derive stereotactic co-ordinates to guide engraftment in rats with 60 min of middle cerebral artery occlusion (MCAO) damage. MRI scans were also acquired post-transplantation to identify the neuro-scaffolds in the lesion site. The grafts were indeed visible in the post-transplantation scans as hypointense regions within the hyperintense lesion area (Fig. 3A). PLGA particles are very dense and hence have stronger attenuation of the T2 signal than brain tissue. This allowed us to detect the scaffold particles in the hyperintense lesion even at 1 day post-transplantation. After 7 days, this hypointense signal had intensiﬁed and clearly distinguished the scaffold graft from the hyperintense lesion and the isointense host brain tissue (Fig. 3B). The extent of coverage varied across animals due to a degree of variability in the quantity of particles delivered. Volume measurements within the affected region indicate that the hypointense area gradually increased in size, whereas the hyperintense area of the lesion (i.e. cavity) decreased with time (Fig. 3C). The relaxivity (i.e. the rate of magnetization loss) of the lesion environment therefore gradually resembles that of intact tissue in the contralateral hemisphere (Fig. 3D). MRI here allowed us to scrutinize the neuro-scaffold in vivo, demonstrating the clear effect these particles have on the lesion environment.
2.4. Conﬁrmation of PLGA particles in the lesion cavity To conﬁrm the correct identity of the hypointense MRI signal and the location of the scaffold particles within the lesion cavity, excised post-mortem brains were sectioned and assessed using ﬂuorescent and brightﬁeld microscopy (Fig. 4A). As particle material does not autoﬂuoresce, the hypointense regions seen on MR images appeared as black areas in tissue sections when viewed under ﬂuorescence microscopy (Fig. 4C, E) and as a reddish brown region under brightﬁeld microscopy (Fig. 4F). The absence of red blood cells in the graft suggested that this colour change was not due to haemorrhage or tissue damage, but was due to the presence of the phenol red present in the suspension media that had been taken up by the particles. A phase contrast image of these areas further supported the presence of cells and particles within this region (Fig. 4G). Higher magniﬁcation revealed PKH26-labelled cells condensed in a central area with a smaller proportion dispersed to the outer edges of the graft (Fig. 4E). These
Fig. 3. In vivo monitoring of the neuro-scaffold. (A) Stroke-lesioned rats were imaged prior to implantation of the neuro-scaffold to establish a baseline of the size and relaxivity of the lesion. The hyperintensity on the T2-weighted scan is reﬂective of a stroke lesion that develops into a tissue cavity (red arrows). The injection (blue arrows) and presence (orange arrows) of the neuro-scaffold were clearly visible 1 day post-implantation as a hypointensity, but also the lesion environment itself was transformed from a hyperintense region into patches of hyper-, hypo- and isointensities. By day 7, almost the whole region was transformed into a hypointense region. (B) This hypointensity covered the whole lesion area from anterior to posterior regions. (C) Quantiﬁcation of the hyper-, hypointense and lesion area, indicated that gradually the hypointense area started to dominate the lesion environment. (D) Relaxivity measurements of the lesion environment also reﬂected this change in MR signal potentially providing a measure of de novo tissue formation.
E. Bible et al. / Biomaterials 30 (2009) 2985–2994
E. Bible et al. / Biomaterials 30 (2009) 2985–2994
observations demonstrate that scaffold particles ﬁlled out the lesion cavity to provide structural support for transplanted cells. 2.5. Cell–particle interaction within the cavity The structural support the particles provided within the lesion allowed cells to either remain attached or to migrate along other particles inside the lesion. At 1 day post-transplantation, cells were often found as a conglomerate either at the core or edge of the lesion (Fig. 4H). Some transplanted cells were found to have migrated off particles into surrounding tissue and were more dispersed within the graft itself after 7 days (Fig. 4E and G). The architecture of the neuro-scaffold typically differed from the core to the periphery of the graft. When present at the core, cells formed primitive de novo tissue as a tightly packed mass of cells. Over time, cells became more dispersed and, towards the periphery of the graft, were found as a ﬁbrous web of connective tissue. This was interspersed amongst a honeycomb-like structure formed from degrading particle material (Fig. 4E inset). Scaffold particles hence provide a structural support to neural stem cells, but the behaviour of transplanted cells depends on their location within the lesion cavity. 2.6. PLGA particle degeneration The degradation of particles is part of the attraction to provide a temporal support to cells while these differentiate and organise into de novo tissue. Therefore establishing if some degradation can be seen is an important step that could lead to a reconstitution of lost tissue. In cases, where particles degraded, some transplanted cells remained in situ to form a mass of cells at the lesion edge contributing to new tissue formation (Fig. 4H and I). In contrast, whole particles with transplanted cells attached were more frequently encountered towards the lesion edge integrating with host tissue (Fig. 4H–J). Particles with PKH26-labelled cells attached (Fig. 4J) indicated that cells remained adherent to particles through the transplantation process, but can migrate from these into surrounding tissue (Fig. 4K) or into the space ﬁlled by degrading particle material (Fig. 4L and M). It is clear that the particle scaffold allows for migration, but at the same time provides structural support enabling cells to integrate with host tissue at the lesion edge. 2.7. Neuro-scaffold and host tissue interaction The interaction between neuro-scaffold and host tissue is clearly deﬁned by the area of glial scarring that seals off the ischaemic lesion. It is evident that a close association between particles and host tissue is needed for grafted cells to interact and migrate off the particle. This is predominately in areas where there is little or no glial scaring. In the core of the graft and in the absence of further structural support and chemical attraction, MHP36 cells remained on the scaffold particles and did not migrate off to integrate with host tissue. Although glial scarring in some areas was minimal, an
interaction between grafted cells and host tissue is more difﬁcult in the presence of an extensive glial scar. Hence neuro-scaffolds can integrate with the host brain in terms of attachment of particles to the host tissue and migration of cells into the damage brain. Glial scarring, however, reduces this interaction.
2.8. Neuro-scaffolds and NSC differentiation To replace lost tissue, appropriate differentiation of cells into neurons and astrocytes is required. Immunostaining for neuronal, astrocytic and stem cell markers revealed a mixed population of NeuN-, GFAP- and SOX2-positive transplanted cells at both 1 and 7 days post-transplantation (Fig. 5A). There was no difference in NeuN/GFAP (Fig. 5B) or SOX2-positive (Fig. 5C) cells at either time point. Around the lesion edge, an ongoing inﬂammatory reaction was indicated by the presence of CD11b-positive microglia/ macrophages (Fig. 5E), but this was not extensive and did not cover the whole graft. The MRI hypointensities could therefore not have been caused by an extensive inﬁltration of microglia/macrophages into the neuro-scaffold. Nevertheless, to guarantee the long-term survival of transplanted cells, small blood vessels need to be present to supply the grafted area with nutrients. Although there was evidence of blood vessels and angiogenesis in the peri-infarct area (RECA-1 positive cells), the graft was mostly devoid of blood vessels (Fig. 5D). The long-term survival of the graft (>1 week) might therefore be dependent on enhancing angiogenesis within the neuro-scaffold.
3. Discussion Tissue engineering approaches are increasingly being adapted to repair the damaged central nervous system. Although delivery of tissue-engineered constructs to the spine  and eye [5,13–16] has progressively developed over the last few years, application of these techniques to the damaged brain has been rare due to limitations of delivery [6,17]. For success, this approach requires systematic technology development in order to identify and optimise the parameters critical for stem cell delivery. In this study, we have undertaken this systematic approach. Speciﬁcally, we have 1) demonstrated that scaffold particles can be engineered to attach NSCs in vitro at high density; 2) optimised conditions for cell attachment in order to preserve highly viable cells with their stem cell properties; 3) determined optimal particle size in relation to density of cell delivery, consistency and viability of injection; 4) developed a targeted delivery by use of MRI-derived implantation co-ordinates; 5) utilised MRI to monitor the integration of scaffold particles into the stroke lesion cavity over time providing us with a valuable temporal understanding of how the graft evolves within the same animal. Finally, we demonstrate a primitive de novo tissue formation within 7 days after implantation into the lesion cavity. Thus, we have taken substantial steps towards developing translatable stem cell–scaffold matrices for injection into stroke-lesioned brain.
Fig. 4. Macroscopic validation of MRI detection of neuro-scaffolds. Post-mortem, the neuro-scaffolds (small red bulbar structures) were clearly visible in an extensive stroke lesion that encompassed sub-cortical and cortical regions on day 1 (A). Two small blood clots were also visible (B), but these cannot explain the large progressive hypointense areas visible on the MRI scans. Whole ﬂuorescent (C) and brightﬁeld (D) brain sections stained for DAPI and GFAP revealed a good correspondence of the grafted area (red ﬂuorescent PKH26labelled cells) to the MR images. Within the neuro-scaffold there is de novo tissue (GFAP in green) that has formed from the transplant, whereas within the surrounding area fewer cells agglomerate in a more ﬁbrous structure (inset image) (E). The black area under ﬂuorescence (E) contains particle material that is visible in brightﬁeld, scale bars 50 mm (F). A phase contrast image shows transplanted cells in close association with particle material, scale bar 50 mm (G). Transplanted cells (in red, GFAP in green) surrounding particles (black areas) migrate into the host tissue if these interface with host tissue (H and I, scale bar 50 mm). In some cases, transplanted cells after 1 day can be seen still attached to particles as these attach to host tissue, scale bar 20 mm (J). From these attached particles, MHP36 cells can also be seen migrating off into host tissue, scale bar 20 mm (K). When particles degrade (L), cells can remain attached, but in some instances invade/replace the degrading particle suggesting that the gradual increase in porosity will help tissue formation, scale bar 20 mm (M).
E. Bible et al. / Biomaterials 30 (2009) 2985–2994
Fig. 5. Post-mortem cellular characterization of neuro-scaffolds. (A) A core of transplanted cells (in red) is already present with a surrounding ﬁbrous web-like area. Black areas indicate particles that support the transplanted cells. At 7 days, the core area (*) of cells was generally less dense and an interconnecting area of cells at the periphery (**) was still present. On days 1 and 7 post-transplantation, cells had differentiated into astrocytes (GFAPþ), neurons (NeuNþ) (B), or remained undifferentiated (SOX2þ) (C). Insets show positive NeuN and GFAP staining in addition to the ﬁbrous appearance of the neuro-scaffold at the periphery versus the core regions. Reca-1 staining for blood vessels showed little positive staining within the core area (D) with more endothelial cells being present in the peri-infarct area (inset). Microglia/macrophages (CD11b) were found to have inﬁltrated part of the transplant, but was conﬁned to small portions of the grafts, mainly in its periphery at the interface with host tissue (E). Scale bars 50 mm, insets scale bars 20 mm.
The use of MRI is becoming invaluable to study tissue engineering in vivo [18–20]. Most importantly, it affords the visualisation of the damage prior to therapy, thus predicting lesion size and location. Histological techniques are inappropriate for these assessments and leave open questions as to the extent of damage and accuracy of delivery. Guidance of the injection to the core of the cavity is important here as delivery into intact tissue could lead to iatrogenic complications caused by the large volumes that are injected. This provides more experimental control compared with previously used standard implantation co-ordinates for suspension grafts [3,4]. This longitudinal design is especially important in studies, such as stroke,
where the lesion can be quite heterogeneous and variable. As PLGA particles were much denser than normal tissue, it was possible here to detect these as hypointensities on T2-weighted scans. It also allowed us to measure changes in relaxivity in the lesion area that reﬂects the degree to which the graft replaces the cavitation as the MR signal is decreasing due to the cyst being replaced. This potentially could provide a particle degradation measure for longer time points, where one would expect the hypointense signal to disappear as particles are replaced by de novo tissue. MRI is therefore an exquisite tool to monitor in vivo brain tissue formation that currently cannot be achieved using any other imaging technique.
E. Bible et al. / Biomaterials 30 (2009) 2985–2994
However, MRI by itself, at present, is insufﬁcient to characterise the graft. Histological analyses provided further evidence that the MRI hypointensity indeed corresponded to the grafted particles, rather than to macrophages or a haemorrhage. Although a haemorrhage was unlikely due to its quite distinct MRI signature that is a qualitatively different hypointensity, histology further conﬁrmed that the lesion cavity was not ﬁlled with blood. Although there was some minor microglia/macrophage inﬁltration, this did not extend throughout the graft in the cavity, but was conﬁned to the periinfarct region and small incursions into the graft. There was also a general lack of vascular support from the host within the neuroscaffold. Although the peri-infarct area was densely vascularised, these cells only rarely extended beyond the glial scar. Further scaffold particle engineering might overcome this by releasing speciﬁc growth factors, such VEGF or rhGH, from particles [21,22] or by adding endothelial cells to the neuro-scaffold transplant . To ensure long-term survival of cells and efﬁcient tissue formation, an adequate blood supply to the graft will need to be established. Using MRI, it should be possible to establish if there is a restoration of blood ﬂow within the cavity . Although PLGA co-polymers have a long track record as biomaterial and drug delivery systems, careful consideration must be given to design the particles’ degradation. It has been suggested that most of the degradation products of PLGA particles are absorbed in the target tissue with the remaining portions being cleared through the blood, liver and kidney . It will hence be important to examine if the degradation of PLGA particles and their prolonged presence inside the brain might exert any deleterious effects on cellular functions  or behaviour. There was no evidence here that this is the case. However, longer studies (>1 month) including behavioural assessment will be needed to establish the true therapeutic potential of this system.
4. Conclusion Engineered scaffolds will undoubtedly become a major research effort to ensure that grafted cells will be guided to provide the right kind of repair mechanism to improve brain damage. We here demonstrate that an injectable scaffold can be transplanted into the brain using conventional tools and that these can be further adapted to develop de novo tissue in situ. The combination of neural stem cells, tissue engineering and image-guidance will bring new hope to patients suffering from stroke and other debilitating neurological conditions.
Acknowledgements This work was supported by a BBSRC project grant (BB/ D014808/1) and the generous support by the Charles Wolfson Charitable Trust Foundation. The authors thank Dr Natalia Gorenkova for assisting with the transplantations and Dr Mieke Heyde for generating early versions of the PLGA particles. We thank the British Heart Foundation for supporting the 7 T MRI scanner (Preclinical Imaging Unit, King’s College London) used in this study. Authors contribution: Ellen Bible (conception of study, biological data acquisition, data interpretation, writing of manuscript, ﬁnal approval); David Chau (conception of study, engineering of particles, data acquisition, data interpretation, writing of manuscript, ﬁnal approval); Morgan Alexander (Plasma polymerization, revising of manuscript, ﬁnal approval); Jack Price (grant support, conception of study, data interpretation, writing of manuscript, ﬁnal approval); Kevin Shakesheff (grant support, conception of study, data interpretation, writing of manuscript, ﬁnal approval);
Michel Modo (grant support, conception of study, data interpretation, writing of manuscript, ﬁnal approval).
Appendix. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.biomaterials.2009.02.012.
References  Rosamond W, Flegal K, Furie K, Go A, Greenlund K, Haase N, et al. Heart disease and stroke statisticsd2008 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 2008;117(4):e25–146.  Modo M. Brain repair how stem cells are changing neurology. Bulletin de la Societe des sciences medicales du Grand-Duche de Luxembourg 2008;(2):217–57.  Modo M, Stroemer RP, Tang E, Patel S, Hodges H. Effects of implantation site of stem cell grafts on behavioral recovery from stroke damage. Stroke 2002;33(9):2270–8.  Modo M, Beech JS, Meade TJ, Williams SC, Price J. A chronic 1 year assessment of MRI contrast agent-labelled neural stem cell transplants in stroke. NeuroImage. doi:10.1016/j.neuroimage.2008.06.017.  Lavik EB, Klassen H, Warfvinge K, Langer R, Young MJ. Fabrication of degradable polymer scaffolds to direct the integration and differentiation of retinal progenitors. Biomaterials 2005;26(16):3187–96.  Park KI, Teng YD, Snyder EY. The injured brain interacts reciprocally with neural stem cells supported by scaffolds to reconstitute lost tissue. Nature Biotechnology 2002;20(11):1111–7.  Bhang SH, Lim JS, Choi CY, Kwon YK, Kim BS. The behavior of neural stem cells on biodegradable synthetic polymers. Journal of Biomaterials Science 2007;18(2):223–39.  Sinden JD, Rashid-Doubell F, Kershaw TR, Nelson A, Chadwick A, Jat PS, et al. Recovery of spatial learning by grafts of a conditionally immortalized hippocampal neuroepithelial cell line into the ischaemia-lesioned hippocampus. Neuroscience 1997;81(3):599–608.  Modo M, Stroemer RP, Tang E, Veizovic T, Sowniski P, Hodges H. Neurological sequelae and long-term behavioural assessment of rats with transient middle cerebral artery occlusion. Journal of Neuroscience Methods 2000;104(1): 99–109.  Paxinos G, Watson C. The rat brain in stereotaxic coordinates. Amsterdam, NL: Academic Press; 2007.  Modo M, Rezaie P, Heuschling P, Patel S, Male DK, Hodges H. Transplantation of neural stem cells in a rat model of stroke: assessment of short-term graft survival and acute host immunological response. Brain Research 2002;958(1):70–82.  Teng YD, Lavik EB, Qu X, Park KI, Ourednik J, Zurakowski D, et al. Functional recovery following traumatic spinal cord injury mediated by a unique polymer scaffold seeded with neural stem cells. Proceedings of the National Academy of Sciences of the United States of America 2002;99(5):3024–9.  Ford MC, Bertram JP, Hynes SR, Michaud M, Li Q, Young M, et al. A macroporous hydrogel for the coculture of neural progenitor and endothelial cells to form functional vascular networks in vivo. Proceedings of the National Academy of Sciences of the United States of America 2006;103(8):2512–7.  Neeley WL, Redenti S, Klassen H, Tao S, Desai T, Young MJ, et al. A microfabricated scaffold for retinal progenitor cell grafting. Biomaterials 2008;29(4):418–26.  Ng TF, Lavik E, Keino H, Taylor AW, Langer RS, Young MJ. Creating an immuneprivileged site using retinal progenitor cells and biodegradable polymers. Stem Cells (Dayton, Ohio) 2007;25(6):1552–9.  Tomita M, Lavik E, Klassen H, Zahir T, Langer R, Young MJ. Biodegradable polymer composite grafts promote the survival and differentiation of retinal progenitor cells. Stem Cells (Dayton, Ohio) 2005;23(10):1579–88.  Menei P, Montero-Menei C, Venier MC, Benoit JP. Drug delivery into the brain using poly(lactide-co-glycolide) microspheres. Expert Opinion on Drug Delivery 2005;2(2):363–76.  Cheng HL, Wallis C, Shou Z, Farhat WA. Quantifying angiogenesis in VEGFenhanced tissue-engineered bladder constructs by dynamic contrastenhanced MRI using contrast agents of different molecular weights. Journal of Magnetic Resonance Imaging 2007;25(1):137–45.  Ko IK, Song HT, Cho EJ, Lee ES, Huh YM, Suh JS. In vivo MR imaging of tissueengineered human mesenchymal stem cells transplanted to mouse: a preliminary study. Annals of Biomedical Engineering 2007;35(1):101–8.  Terrovitis JV, Bulte JW, Sarvananthan S, Crowe LA, Sarathchandra P, Batten P, et al. Magnetic resonance imaging of ferumoxide-labeled mesenchymal stem cells seeded on collagen scaffolds-relevance to tissue engineering. Tissue Engineering 2006;12(10):2765–75.  Li Q, Ford MC, Lavik EB, Madri JA. Modeling the neurovascular niche: VEGFand BDNF-mediated cross-talk between neural stem cells and endothelial cells: an in vitro study. Journal of Neuroscience Research 2006;84(8): 1656–68.
E. Bible et al. / Biomaterials 30 (2009) 2985–2994
 King TW, Patrick Jr CW. Development and in vitro characterization of vascular endothelial growth factor (VEGF)-loaded poly(DL-lactic-co-glycolic acid)/ poly(ethylene glycol) microspheres using a solid encapsulation/single emulsion/solvent extraction technique. Journal of Biomedical Materials Research 2000;51(3):383–90.
 Hua N, Sun J. Body distribution of poly(D,L-lactide-co-glycolide) copolymer degradation products in rats. Journal of Materials Science 2008;10: 3243–8.  Shive MS, Anderson JM. Biodegradation and biocompatibility of PLA and PLGA microspheres. Advanced Drug Delivery Reviews 1997;28(1):5–24.