Introduction

Stroke is one of the leading causes of long-term disability and death in the world. Currently, the only treatment for stroke is thrombolytic tissue plasminogen activator (t-PA), which has been shown to reduce ischemic damage and permanent disability if administered within 3–4.5 h after onset1,2. However, the short effective time window, as well as the potential for a variety of complications, means that only 3%–5% of stroke patients receive t-PA treatment. A long-term treatment is needed to repair infarct damage and improve functional recovery in stroke patients.

Stroke results in the necrosis of brain tissue, leading to the loss of the intricate arrangement of neurons, glial cells, and vasculature. Neural stem cells in the adult brain hold great promise for the development of neural repair strategies. Specifically, neural stem and progenitor cells — collectively known as neural precursor cells (NPCs) — possess the ability to generate the neural cell types present in the brain. The use of NPCs in cell-based therapy has focused on two distinct approaches: endogenous and exogenous cell-based strategies. The exogenous approach involves the propagation of NPCs in vitro and their subsequent transplantation to achieve regeneration and recovery. In the endogenous approach, NPCs resident in the brain are induced to proliferate, migrate to the infarct site, differentiate into various neural cells needed for repair, and functionally integrate into the tissue to promote recovery. This review aims to provide an understanding of NPCs, and to present current findings in the two approaches of cell-based therapy for ischemic stroke, with a focus on endogenous repair strategies.

Neural precursor cells

By definition, a stem cell has the capacity to self-renew indefinitely and the ability to give rise to cells that generate differentiated progeny. A neural stem cell generates multipotent progeny which can differentiate into neurons, astrocytes, and oligodendrocytes. A neural progenitor cell possesses limited self-renewal capacity and differentiation potential3. Neural stem and progenitor cells are collectively referred to as NPCs.

Historically, NPCs were thought to exist only in the developing brain, while the adult mammalian brain was presumed post-mitotic and devoid of regenerative capacity. However, seminal works by Reynolds and Weiss showed the presence of NPCs in the adult mammalian central nervous system (CNS)4,5. These studies used a colony forming assay which involved the dissection and plating of mouse CNS cells, including the periventricular region lining the forebrain lateral ventricles. Cells were cultured in the presence of epidermal growth factor (EGF) and after 7 d in vitro, clonally derived free floating colonies (termed neurospheres) were observed. Individual neurospheres could be dissociated, replated in EGF, and passaged indefinitely thereby revealing the stem cell property of self-renewal. The cells were also shown to differentiate into the three CNS cell-types, confirming their multipotentiality. The discovery of NPCs in the adult brain offered new hope for the possibility of regeneration in the adult brain.

Adult NPCs are found in two discrete regions: the subventricular zone (SVZ)6, and the dentate gyrus (DG) of the hippocampus7. The SVZ is a region lining the walls of the forebrain lateral ventricles, and consists of the ependymal and subependymal layers. The NPCs are found in the subependyma, adjacent to the single-cell ependymal layer lining the ventricle6,8,9. Neural stem cells in the SVZ are slowly proliferating cells6 that express glial fibrillary acidic protein (GFAP)10,11 and are referred to as type-B cells12. Type-B cells give rise to rapidly dividing progenitor cells called type-C cells, which subsequently give rise to neuroblasts (type-A cells)12. In rodents, type-A cells migrate through the rostral migratory stream (RMS) to the olfactory bulb, where they differentiate into interneurons and become functionally integrated into the neuronal network13,14. Neurogenesis in the rodent olfactory bulb has been suggested to play a role in olfactory learning and memory15. In humans, samples from the SVZ of adult brains have shown the presence of GFAP+ cells capable of forming self-renewing, multipotent neuropheres in vitro, confirming the presence of NPCs in the adult human brain16,17. Furthermore, neuroblasts migrating from the SVZ to the olfactory bulb, analogous to the rodent RMS, have been observed in humans18, albeit in much smaller numbers than what is observed in rodents19.

NPCs that express GFAP are also found in the subgranular zone of the DG, between the hippocampal hilus and granular layer7,20,21,22. These NPCs proliferate, migrate into the granular layer, and extend processes that functionally integrate into the overlying molecular layer23,24. Unlike the SVZ, however, NPCs from the DG show limited self-renewal and multipotent capacities when propagated in vitro, and there is controversy as to whether the DG contains “true” stem cells25,26. In humans, post-mortem analysis of hippocampi has revealed the presence of neurons derived from proliferating cells in patients injected with the proliferation marker bromodeoxyuridine27. NPCs in the live adult hippocampus have also been identified using proton magnetic resonance spectroscopy28. Given the critical role of the hippocampus in learning and memory, the NPCs in this region are considered an important target for treating the stroke injured brain29.

Stroke models and outcome measures

Given the clinical prevalence of ischemic over hemorrhagic stroke, most research has focused on rat and mouse models of ischemic stroke. Early stroke studies used the common carotid artery occlusion model, which induced global ischemia30. From a clinical perspective, ischemic stroke is more focal, commonly affecting the striatum and internal capsule. Accordingly, a majority of recent studies use the middle cerebral artery occlusion (MCAO) model as it induces focal ischemic damage to the striatum and overlying cortex. Variations in the location (proximal, distal) and duration of MCAO (transient, permanent) can be used to induce different extents of damage in distinct regions throughout the brain (reviewed by Lindvall and Kokaia31,32).

Other models are used to address the clinical heterogeneity of ischemic stroke. For instance, lesions restricted to the cortex have been modelled using a variety of techniques such as photothrombic ischemia33,34, devascularisation35, and cortical aspiration35. Endothelin-1, a potent vasoconstrictor that results in transient ischemia followed by reperfusion of the tissue, has also been used to induce stroke36.

A variety of outcome measures are used to determine the efficacy of the treatments employed. In cell-based therapies a comprehensive analysis of the cell phenotypes is performed using specific cell markers. In addition, infarct volume, as well as neuroimaging techniques including MRI and CT, have been used to examine the effects of treatments on angiogenesis and tissue repair. Arguably, the most clinically relevant outcome measure is functional recovery. As such, the degree of sensorimotor recovery, as determined by a number of behavioural assays, is commonly used to determine efficacy of the treatment employed (summarized in Table 1). In this review we will consider some of the parameters important for developing and employing NPC based strategies with a focus on the endogenous approach for ischemic stroke repair.

Table 1 Summary of the effects of the activation factors on endogenous repair processes in stroke models.
Full size table

Exogenous approach: transplanting neural precursor cells following stroke

A variety of stem cell types have been used for decades in cell transplantation studies to treat stroke, and while some have shown promise, the success of this treatment strategy is yet to be realized. Initial pioneering work using fetal brain cells in ischemic rat models provided some evidence of synaptic integration and cognitive recovery, although the limited supply, difficulty ensuring the purity of the cell source and the ethical issues surrounding the cell type have largely impeded the progress of these studies37,38. Clinical trials using human NT2 teratocarcinoma cells and fetal porcine neural cells were met with risks of adverse effects and limited improvement39,40. More promising was the use of autologous bone marrow mononuclear cells (BMMCs) and mesenchymal stem cells (MSCs), which are now in Phase II clinical trials as the treatments were shown to be safe and further demonstrated some neurological improvements41,42. More recently, induced pluripotent stem cells (iPSCs) have gained attention due to their non-embryonic origin and their potential to be induced to a neural lineage. Despite the exciting prospects of these cells, considerable safety concerns remain regarding their potential to form teratomas. (For a more detailed review of different stem cells, see Burns and Steinberg43).

Transplantation of NPCs has advantages over other cell types for a number of reasons. For example, their neural restricted lineage limits the potential for producing undesired non-neural phenotypes while permitting the production of CNS cells. Further, the fact that NPCs reside in a very well defined region of the adult brain raises the possibility of autologous transplantation, although the extraction of endogenous NPCs remains an issue. Indeed, the difficulties associated with accessing the NPCs in the adult brain have driven pre-clinical studies to employ NPCs derived from other stem cell types. Fetal and embryonic stem cell-derived NPCs, after intracranial transplantation, migrate towards site of ischemia, differentiate into neurons, and survive for up 12 weeks44,45,46. Behavioural improvements have been shown in forelimb tasks in rats, and this has coincided with improved electrophysiological cell recordings in the appropriate anatomical sites in the brain46,47. These promising results have led to an ongoing Phase I clinical trial of intracranial transplantation of embryonic stem cell-derived neural stem cells in patients with subcortical/basal ganglia stroke48. More recently, iPSC-derived NPCs have also been shown to produce mature neurons, though no functional assessments were examined49. Studies that employed adult NPCs isolated from the SVZ and expanded in neurospheres have also shown NPC survival and differentiation after transplantation, although their survival is lower than NPCs derived from embryonic neural precursors50,51,52.

There are a number of questions that remain outside of the cell source to be employed. For instance, the route of administration and the time to transplant post-stroke must also be examined. While the previously mentioned studies used intracranial administration, there has also been increasing interest in intravascular delivery of NPCs. Reports indicate that NPCs can “home” into sites of ischemic injury, reduce infarct volume and lead to behavioural improvements53,54,55. The NPCs were shown to have immunomodulatory and anti-apoptotic effects54. A recent meta-analysis examined the effect of intravenous delivery of NPCs in 60 preclinical studies for neurological diseases, and summarized the overall outcomes into morphological, behavioural, and molecular effects56. The study showed that NPCs have a greater effect on behavioural outcomes than MSCs, BMMCs, and other stem cell types56, which suggests that intravascular delivery of NPCs may be a clinically viable therapy.

Successful application of exogenous NPCs in the clinic requires the optimization of various parameters including the time of transplantation and dosage. Indeed, the survival of transplanted cells is dependent on changes in the brain environment after ischemia. There is no current consensus on the most effective time of delivery. The meta-analysis by Janowski et al found no significant difference in behavioural, molecular, or morphological outcomes in relation to timing of cell delivery56. The meta-analysis, however, found a significant positive correlation between cell dose (up to 107 cells) and improved molecular outcomes. Hence, the cell type, dosage, and time of delivery and administration route need to be considered when developing exogenous strategies for human use.

The endogenous approach: activating neural precursor cells to promote self-repair

NPCs in the adult brain demonstrate the fundamental properties in vivo that would be necessary for developing strategies to promote their contribution to neural repair: proliferation, migration, and differentiation into neural phenotypes. In the first demonstration of the recruitment of NPCs in stroke repair, transient ischemia was induced in rats by MCAO, and immunohistochemistry demonstrated SVZ-derived NPC proliferation, migration to the site of infarct, and the generation of neurons in the striatum as well as the parietal lobe57,58. Further studies suggested that this activation and recruitment process persists for several months after the ischemic attack59,60, and that newly formed neurons become synaptically integrated as determined by morphological and electrophysiological studies60,61. These phenomena have more recently been examined in humans, where post-mortem biopsies of stroke patients have shown the presence of proliferating and differentiating cells in the ischemic penumbra as well as the ipsilateral SVZ62,63. However, this endogenous activation is clearly not sufficient for functional recovery as demonstrated by the persistent and devastating functional impairments observed in patients following stroke.

Notably, NPCs in the DG have also been shown to proliferate and differentiate into mature neurons after ischemia30,64. However, no conclusive evidence to date has shown the migration of NPCs out of the DG into other brain regions65. As a result, the regenerative capacity of NPCs in the DG are restricted to the hippocampus, and much of the research has focused on the more migratory NPCs observed in the SVZ.

Developing endogenous activation strategies to promote tissue repair and functional recovery relies on the ability to stimulate endogenous precursors in the brain using a variety of exogenous “activation factors”, including growth factors and cytokines. The premise of the endogenous repair strategy is to take advantage of the inherent properties of NPCs to proliferate (and permit an expansion of the NPC pool), migrate to the site of the injury and differentiate into mature cells in the injured brain. Similar to the challenges faced with exogenous transplant therapies, there are many unanswered questions such as which types of mature cells — glial and/or neuronal — need to be replaced in the stroke injured tissue in order to facilitate recovery. Understanding the mechanisms that promote regeneration and recovery will ultimately underlie the choice of NPC activation factors. Activation factors may mediate their effects by: modifying cell proliferation kinetics, promoting cell survival, enhancing cell migration to infarct site and/or promoting neurogenesis, and angiogenesis. Table 1 provides a summary of recent research on various activation factors. Herein we discuss how the above processes are enhanced by activation factors, with functional improvement as the end-goal.

Proliferation

Proliferation represents the first step of the adult NPCs' response to ischemia. Indeed, the injury alone is sufficient to enhance the NPC proliferation and thereby increase the size of the NPC pool66,67. Given the fact that NPCs make up a very small fraction of cells in the adult brain (less than 0.5% of the cells in the SVZ are stem cells6,68, and the progeny comprise approximately 15%69), increasing the size of the NPC pool is an important factor in endogenous repair strategies. Moreover, some factors that promote NPC expansion have also lead to increased olfactory bulb neurogenesis in non-injured animals, or in the stroke lesion site of injured animals. For example, intraventricular administration of EGF has been shown to promote NPC proliferation in the SVZ of uninjured and stroke injured animals70,71, and NPCs in injured brains subsequently migrated to the lesion site where they developed into mature neurons71,72. Adult NPCs express the EGF receptor (EGFR), and increase their sensitivity to EGF and other EGFR ligands in response to ischemia by upregulating EGFR expression73. Indeed, heparin-binding EGF-like growth factor (HB-EGF), an EGFR ligand, delivered intraventricularly one day after ischemia in rats increased proliferation of NPCs in the SVZ and DG, and improved neurological deficits compared to saline-administered post-ischemic rats74.

A plethora of other growth and trophic factors have been shown to increase NPC proliferation: vascular endothelial growth factor (VEGF)75,76,77, glial cell-line derived neurotrophic factor (GDNF)78, granulocyte colony stimulating factor (G-CSF)79,80, basic fibroblast growth factor (FGF-2)81, insulin-like growth factor (IGF-1)82,83, bone morphogenic protein 7 (BMP-7)84, transforming growth factor α (TGF-α)85, and brain derived neurotrophic factor (BDNF)33. While many of these factors have been shown to act directly through their associated receptors, it is also possible that they mediate their proliferative effects indirectly via immunomodulation, neuroprotection, and angiogenesis. A summary of these factors are found in Table 1.

From a clinical perspective, enhancement of NPC proliferation increases the risk of aberrant cell growth and tumour formation, especially since some growth factors are well-known mitogens. Using certain hormones to stimulate NPCs may be a safer approach, as their physiological levels fluctuate dramatically with minimal tumorigenic effects. One molecule of interest is erythropoietin (EPO) which acts directly upon EPO receptors present on NPCs to stimulate their proliferation86,87. Intraperitoneal administration of EPO in rats after unilateral carotid artery occlusion led to an improvement in sensorimotor outcomes 14 d after ischemia compared to saline-injected stroke-injured rats88. In agreement with a more recent study by Wang et al, the studies have shown a decreased number of apoptotic nuclei ipsilateral to the infarct, suggesting an additional role for EPO in extending cell survival88,89. Growth hormone (GH) has also been shown in vitro to increase proliferation of neurosphere-derived NPCs by acting directly through the GH receptor90,91. In uninjured mice, GH increased NPC proliferation in the SVZ and olfactory bulb when delivered intraventricularly. However, there is no evidence to date on the effects of GH on NPCs in an animal stoke model, which may be a novel therapeutic target. Some recent work has demonstrated that melatonin can also promote NPC proliferation and reduce neurological deficits that result following stroke92,93. Melatonin, delivered orally to mice for 29 d after MCAO, leads to improved motor coordination, as well as decreased hyperactivity and anxiety92. In a recent study, estradiol was delivered subcutaneously in male and ovarectomized female mice and was shown to enhance NPC proliferation post-ischemia in the SVZ and DG, and to increase neurogenesis in the DG (but not in the SVZ)94. While the study provided some evidence that this enhanced proliferation and neurogenesis can result in functional recovery, the deleterious role of estradiol in some human cancers is well-documented, and therefore the use of this hormone in a clinical setting demands further study.

Promoting survival

Proliferation of NPCs can occur by shortening the cell cycle, and supporting this process may increase the risk of tumorigenesis. An alternative approach to expanding the NPC pool post-ischemia is by promoting cell survival through inhibition of apoptosis. Within the adult NPC population, 60% of the newborn progeny undergo cell death under baseline conditions68, hence sparing the NPCs from cell death would effectively increase the size of the NPC pool. The PI3K-Akt pathway has been shown to be important in NPC survival in the developing brain through the upregulation of anti-apoptotic genes95,96. Administration of G-CSF to cultured neurons and neuroblastoma cells led to strong activation of the PI3K-Akt pathway resulting in decreased caspase 3/7 activity and decreased apoptosis79. G-CSF receptors were found to be expressed on cultured adult NPCs, and intravenous administration of G-CSF in stroke-injured rats promoted NPC survival in vivo, suggesting a possible anti-apoptotic role for G-CSF on NPCs following ischemia79. In addition, the anti-apoptotic effect of G-CSF was shown to be mediated by the STAT3 pathway in ischemic rats97. Other factors, such as IGF-1, activate the PI3K-Akt pathway and inhibit apoptosis by the phosphorylation of glycogen synthase kinase 3β, thereby promoting cell survival98.

Other potential targets to promote cell survival include the tumour suppressor p53, which is known to negatively regulate proliferation and increase apoptosis in NPCs99. Pifithrin-α, a p53 inhibitor, has been shown to enhance NPC proliferation and survival in the SVZ when administered intraventricularly for 6 d post MCAO ischemia, and further, functional improvements were observed using a standardized motor disability scale100,101. Inhibition of p53 promotes survival through downregulation of p53-upregulated modulator of apoptosis (PUMA)100,102. More recently it was found that cyclosporine A (CsA), a commonly used immunosuppressive drug, acts directly on NPCs to increase their survival without affecting cell cycle kinetics103. While the pro-survival mechanism has not been fully elucidated, the administration of CsA in vivo resulted in >2 fold increase in the size of the NPC pool in control mice. Moreover, mice that received CsA following stroke showed an expansion of the NPC pool, migration of the NPCs to the site of injury, new tissue formation at the site of cortical ischemia, as well as recovery of motor function104. Hence, a number of pro-survival pathways have demonstrated efficacy in models of stroke.

Other possible mechanisms to increase the size of the NPC pool include activation of signaling pathways that promote a change in the mode of division of NPCs. For example, a change in the mode of division of stem cells from asymmetric division (producing one stem cell and one progenitor cells) to symmetric (producing two stem cell progeny) could effectively double the size of the stem cell pool with each division. One factor recently identified as a regulator of symmetric divisions in adult neural stem cells is Wnt signalling105. Active Wnt signaling was found to be necessary for regeneration of the SVZ following ablation of the NPCs in vivo. Indeed, in the absence of Wnt signaling no repopulation of the neural stem cell pool occurs as the stem cells do not undergo symmetric division. Instead, the asymmetric cell divisions result in maintenance of the stem cell pool but not expansion. Piccin and Morshead also demonstrated that Wnt signaling is upregulated in NPCs in the SVZ following stroke, suggesting that this pathway plays a role in the normal expansion of NPCs observed following stroke alone.

Notch signalling has also been implicated in the maintenance and expansion of NPC pool in the adult brain (reviewed by Ables et al106). The transmembrane Notch receptor and its downstream targets are expressed on NPCs in the SVZ, while the Notch ligands — Jagged1 and Delta1 — are expressed by astrocytic GFAP+ cells107,108. The interaction between the two cell types has been suggested to play a role in regulating the proliferation of NPCs in the SVZ108. In a rat MCAO model, a transient increase in the expression of Notch signalling molecules are seen up to 24 h post-ischemia108. In vitro administration of Notch ligands and inhibitors modulate NPC proliferation and differentiation in post-ischemic NPCs107,109. Intraventricular injection of Notch activators in vivo lead to increased cell proliferation in the SVZ for up to 24 h108, and has been shown to improve motor deficits in mice over a 45 d period108. Thus various signalling pathways contribute post-ischemia to the expansion of the NPC pool through the inhibition of apoptosis, change in the cell cycle kinetics, mode of cell division, and regulation of proliferation.

Migration

In order for endogenous NPCs to contribute to recovery, it is necessary for NPCs to migrate from the SVZ through the brain parenchyma into the peri-infarct region and further, which must occur in sufficient numbers in order to generate sufficient numbers of cells. The increased migration to the peri-infarct region has been suggested to occur at the expense of migration to the olfactory bulb, the default path of NPC migration110. A number of studies have looked at the receptor-ligand signalling pathways that are involved in the modest NPC migration that occurs post-stroke. Some of the factors involved include: stromal cell-derived factor 1 (SDF-1; also called CXCL12) and CXC chemokine receptor 4 (CXCR4)58,110; angiopoietin-1 (Ang1) and Tie-2110; monocyte chemoattractant protein-1 (MCP-1; also called CCL2) and CC chemokine receptor 2 (CCR2)111; osteopontin and integrin β1112; and, meteorin113. With the goal of enhancing migration, controlled delivery of ligands has been examined in vivo. In vitro, SDF-1 has been shown to promote NPC migration by activating the ERK 1/2 pathway, and treatment of NPCs with an ERK 1/2 pathway inhibitor abolished the stimulatory effect of SDF-1 on migration114. Intraventricular administration of SDF-1 in ischemic mice increases the number of doublecortin-expressing neuroblasts that migrate to the peri-infarct region 7 d post-stroke110. CXCR4 antagonists suppress the migration of neuroblasts into the peri-infarct region58, causing neuroblasts to migrate non-directionally and disperse diffusely throughout the cortex following stroke110. A similar effect on cell migration is observed with Ang1/Tie-2 signalling modulation by administering Ang1 and Tie-2 antagonist intraventricularly110. Finally, both in vitro115 and transplantation111 studies using NPCs have shown that MCP-1/CCR2 signalling promotes the directed migration of NPCs towards the infarct region. Notably, MCP-1/CCR2116,117, and SDF1/CXCR4118 are also involved in the chemotaxis of inflammatory cells, hence these molecules may lead to a more pronounced immune response and may subsequently counteract neurogenic recovery.

Matrix metalloproteinases (MMPs) have also been implicated in neuroblast migration from the SVZ. MMPs are capable of degrading extracellular matrix proteins, involved in the cleavage of cell surface receptors, and are thought to play a role in many cell behaviours including proliferation, migration, differentiation and apoptosis119. In the adult brain, MMP-3 and -9 have been shown to be expressed by NPCs, and inhibition of MMPs in in vitro siRNA experiments, as well as MMP inhibitors in a mouse model, have revealed a reduction in NPC migration following MCAO ischemia120,121. The working hypothesis is that MMPs released by the NPCs digest the ECM to clear the path for neuroblast migration122. Considering that MMPs are also expressed by endothelial cells and astrocytes123,124 and in light of the increased gliosis that occurs following stroke, these cells may also be contributing to the enhanced migration of NPCs following stroke.

In thinking of ways to enhance and direct cell migration following injury, it is helpful to consider what is known about cell migration during development of the nervous system when all newborn cells are migrating to their final destination within the developing brain parenchyma. In recent studies, Babona-Pilipos et al, applied direct current electric fields to adult derived NPCs in vitro125, as it is well established that disruption of electric fields during development, or during wound healing, completely disrupts cell migration, leading to aberrant development and tissue repair126. When adult NPCs were exposed to a 250 mV/mm electric field, the NPCs undergo a rapid and directed migration towards the cathode125. The migration could be entirely accounted for the application of the electric field and was not mediated by any chemotactic gradient that was established in the assay. These findings suggest that the application of electric fields may be a novel and clinically relevant approach to enhancing and directing the migration of NPCs to the site of injury and is worthy of further investigation.

Differentiation and neurogenesis

The importance of generating new neurons in cell-based therapies has not been completely established, and this is partly due to the fact that when NPCs do appear to differentiate into mature neurons, it is challenging to determine the extent to which these newborn neurons have synaptically and functionally integrated into the neural circuitry. Using markers for mature neurons, studies have shown several activations factors to facilitate differentiation: FGF-281, EGF71, HB-EGF127, VEGF128, IGF-183, G-CSF80, GDNF78, TGF-α85, estradiol94, and EPO89 (Table 1). Indeed, it is possible that neurogenesis is not required for functional recovery. At least two studies have shown improvements in functional outcomes despite the lack of neurogenesis at the infarct site after activation factor administration. In the case of Erlandsson et al, CsA administration led to newly formed tissue at the site of the infarct that contained only newborn glial cells104. In Chern et al, intraperitoneal administration of melatonin after MCAO stroke in mice resulted in a reduction in neurological deficit, even though the increase in mature neurons was not significant93. These studies support the possibility that NPCs contribute to recovery through other mechanisms, such as providing factors that prevent secondary stroke damage, or by providing trophic support to existing cells in the infarct region65,129.

Role of angiogenesis

Angiogenesis plays an integral role in post-ischemic recovery, and is closely tied to neurogenesis130. Vasculature not only provides blood supply to the infarct region, but also stimulates NPCs by the endothelial secretion of various growth and chemotactic factors. VEGF and IGF-1 are released by endothelial cells in response to ischemia, and promote NPC proliferation and differentiation82,83,131. The receptor ligand interaction is also thought to act reciprocally, that is, signalling from NPCs promotes neovascularization, leading to the “coupling” of angiogenesis and neurogenesis83,131.

Vasculature can also provide a scaffold on which NPCs can migrate, observed by the close association of migrating NPCs and neuroblasts with vascular endothelial cells110,132,133. Selective inhibition of VEGFR2 attenuates post-ischemic NPC migration, supporting the role of VEGF in coupled angiogenesis and neurogenesis in stroke recovery134. Similarly, IGF-1, SDF-1/CXCR4, and Ang1/Tie-2 signalling from endothelial cells also facilitate the migration of NPCs along blood vessels83,110,132. These studies provide strong support for the development of strategies that promote angiogenesis, which will ultimately lead to enhanced migration of NPCs and neurogenesis.

Challenges

One of the challenges to the application of NPC activation to the clinic is the delivery of activation factors. Some studies have used peripheral modes of administration (intravenous, subcutaneous, intraperitoneal), while others have attempted to target the brain directly. Peripheral administration minimizes trauma to the brain, but does not target NPCs directly and may lead to systemic effects that confound the activation factor's effect on NPCs. Alternatively, those that target the brain directly (intracerebral/intraventricular infusion of factors/viral vectors) may avoid systemic effects, but damage to the brain tissue and the increased risk of infection make them clinically impractical options. Recently, some bioengineering strategies have looked at a less invasive delivery of factors using an epicortically implanted hydrogel that releases factors in a spatially and temporally defined fashion89,135. The hydrogel technology delivers the factors of interest locally and resulted in greater brain tissue penetration than when the factors were delivered systemically or intraventricularly. This delivery strategy is a promising methodology for clinical application.

Another barrier to successful activation of endogenous NPCs is the difference in the neurogenic response between the young and aged brains. Many studies of activation factors use young adult rodents for stroke models. However, stroke occurs predominantly in the elderly in humans, and consequently studies need to take this into consideration in order for the results to be translational. While reports have confirmed that post-ischemic neurogenesis occurs in the aged brain, the response is significantly attenuated136,137. Furthermore, the response of NPCs to activation factors such as G-CSF and VEGF have been shown to be attenuated in aged mice80,138, or absent with FGF-2139. Gene expression analysis in the ischemic young and aged brains have shown that the aged brain upregulates the expression of genes related to DNA damage, cell cycle arrest, apoptosis, and inflammation, while genes related to axon and dentrite growth are downregulated140. These global changes in gene expression may explain the differential response of the aged brain to ischemia.

As Table 1 summarizes, there are a plethora of activation factors that affect NPCs, and many others that may work through indirect mechanism. Successful translation of the endogenous approach to the clinic requires a thorough evaluation of the magnitude of the effect of individual and combinations of factors. Potential therapies need rigorous studies in multiple rodent models with different baseline conditions that are relevant to clinical conditions (hypertensive, diabetic, age, sex). Safety and efficacy testing in nonhuman primate (NHP) models are also recommended by STEPS II (Stem Cell Therapies as an Emerging Paradigm in Stroke), given the anatomical differences, especially white matter, in gyrencephalic primate brains141,142. To date, little work has been done in primates, given the lack of an established NHP model for ischemic stroke, as well as the cost and limited availability of animals141,143. Ultimately, however, a transition to clinical trials is necessary to assess functional outcomes in humans, as animal models provide limited information about the recovery of functions such as speech, language, higher cognitive function, learning, and memory.

NPC proliferation and migration takes place after hemorrhagic stroke in rats144, and post-mortem analysis of human brains has suggested that increased neurogenesis occurs in patients with hemorrhagic stroke145. However, research has focused primarily on ischemic stroke models, and further research is necessary to determine whether the activation factors can enhance endogenous recovery in hemorrhagic models of stroke.

Combinatorial aproaches to enhance stroke recovery

An understanding of the mechanisms that underlie endogenous NPC activation will undoubtedly benefit transplantation strategies designed to promote stroke recovery. Indeed it has been hypothesized that NPCs in peri-infarct regions provide trophic support to existing cells to promote plasticity and/or prevent secondary damage104. If the latter mechanism has a greater contribution to stroke recovery, then combining endogenous and exogenous approaches may optimize the development of clinical therapies. Along these lines, transplanted NPC's may serve as vehicles for trophic factor delivery. This has been demonstrated in studies where transplanted NPCs expressing the BDNF gene lead to improved functional outcomes relative to control (non-expressing) NPC recipients146,147. Transplantation of GDNF-transfected NPCs not only improves functional outcomes, but also increases BDNF expression, ultimately leading to improved cell survival148. Hypoxia-induction factor (HIF)-transfected cells were shown to improve functional outcomes, and increase the proliferation of endothelial cells, suggesting increased angiogenesis149. These results prompt further study of HIF, especially given HIF's role in inducing the expression of various previously discussed factors, including VEGF, EPO, and HB-EGF. Transplanted NPCs have also been shown to promote endogenous NPC proliferation in the SVZ150,151, angiogenesis151, and to release various trophic factors to promote axonal and dendritic plasticity152. The latter study found that the predominantly released factor was VEGF, which may explain the proliferative and angiogenic effects.

Conclusion

There is much excitement surrounding the potential for NPCs to promote tissue repair and functional recovery following stroke. Both exogenous and endogenous approaches to cell-based therapy using NPCs have shown promising results in preclinical studies. However, it may be that neither alone will be sufficient for promoting full recovery post-stroke, especially when considering the need to apply these strategies in old age where the numbers of NPCs are fewer. Indeed, transition to the clinic will ultimately require rigorous evaluation of not only the activation factors being employed, but also how the exogenous and endogenous approaches can be used together to complement each other, with the end-goal of achieving functional and cognitive recovery.

Abbreviations

AAV, adeno-associated virus; Ang 1, angiopoietin 1; BDNF, brain derived neurotrophic factor; BMMC, bone marrow mononuclear cells; BMP-7, bone morphogenic protein 7; BrdU, bromodeoxyuridine; CNS, central nervous system; CsA, cyclosporine A; DG, dentate gyrus; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; EPO, erythropoietin; FGF-2, basic fibroblast growth factor; G-CSF, granulocyte colony-stimulating factor; GDNF, glial cell line-derived neurotrophic factor; GFAP, glial fibrillary acidic protein; GH, growth hormone; HB-EGF, heparin-binding EGF-like growth factor; HIF, hypoxia induction factor; IGF-1, insulin-like growth factor 1; iPSC, induced pluripotent stem cells; MCAO, middle cerebral artery occlusion; MCP-1, monocyte chemoattractant protein 1; MMP, matrix metalloproteinase; MSC, mesenchymal stem cells; NPC, neural precursor cell; PI3K, phosphoinositide-3-kinase; PUMA, p53-upregulated modulator of apoptosis; RMS, rostral migratory stream; SDF-1, stromal cell-derived factor 1; STEPS, Stem Cell Therapies as an Emerging Paradigm in Stroke; SVZ, subventricular zone; TGF-α, transforming growth factor alpha; t-PA, thrombolytic tissue plasminogen activator; VEGF, vascular endothelial growth factor.