Introduction

Neonatal encephalopathy is a clinical syndrome that occurs in the first days of life, characterized by the presence of neurological symptoms, such as a depressed level of consciousness, hypotonia, difficulty with initiating and maintaining respiration, and frequently by seizures, in full-term and late preterm infants.

Although the etiology of neonatal encephalopathy is varied, including endocrine, metabolic and genetic disorders, intrapartum hypoxia-ischemia is present in 30–60% of the cases. Accordingly, when there is evidence that an intrapartum hypoxic-ischemic (HI) insult is the cause of the encephalopathy, the syndrome is called neonatal hypoxic-ischemic encephalopathy (HIE), which occurs in 1.5 per 1000 live births (Kurinczuk et al., 2010).

HIE can be classified into mild, moderate or severe encephalopathy, according to the classification of Sarnat and Sarnat, based on clinical presentation and electroencephalographic signs. The percentage of adverse outcomes, including cerebral palsy, motor/cognitive impairment or death, is 0% for mild, 32% for moderate and almost 100% for severe HIE, in infants under 3 years of age. In addition, both mild and moderate HIE may affect daily life behavioural functioning at the age of 9–10 years and most of the children (81%) with moderate HIE have cognitive deficits when evaluated at 15–18 years of age.

Besides the clinical classification, the pattern of brain injury on magnetic resonance imaging (MRI) scans is one of the best predictors of neurodevelopmental outcome in infants with HIE (de Vries and Jongmans, 2010). Briefly, MRI may identify two main injury patterns after HIE:

  1. (1)

    The basal ganglia-thalamus pattern (BGT) affects bilaterally the deep gray nuclei and perirolandic cortex, occurring more often after an acute sentinel event, such as placental abruption, uterine rupture or umbilical cord prolapse. Hippocampus, brain stem and white matter may also be affected. BGT is associated with cerebral palsy in 70% of the cases, and with epilepsy in 30–40% of HIE survivors. Visual impairments and dysarthria are also common in children with HIE and BGT injury.

  2. (2)

    The watershed predominant pattern (WS) involves the white matter, particularly the vascular watershed zones (anterior-middle cerebral artery and posterior-middle cerebral artery), and also the cerebral cortex when severe. WS is associated with cognitive deficits and epilepsy, but usually is not the cause of severe motor impairment (Martinez-Biarge et al., 2010).

Recently, large multicentric clinical trials showed that therapeutic hypothermia modestly improves the neurologic outcome of infants with moderate HIE, when started within 6 h of birth. However, given the limited benefits of therapeutic hypothermia, new treatments that could reduce or prevent the long-term neurodevelopmental sequelae of children with HIE are urgently needed. In this regard, neural stem/progenitor cell (NSCPs) transplantation represents a promising treatment to regenerate or reduce brain damage in HIE.

Animal Model of HIE

The most used model of neonatal HIE was developed by Rice and Vannucci in 1981. This model uses the Levine preparation, consisting of unilateral common carotid artery ligation followed by systemic hypoxia (8% oxygen-balance nitrogen) in post-natal day 7 (P7) rats.

The damage is restricted to the hemisphere ipsilateral to the common carotid artery occlusion, being observed in the cerebral cortex, thalamus, striatum, hippocampus and subcortical white matter. Hypoxic-ischemic animals also have cognitive and motor deficits, with a correlation between the degree of brain damage and the alterations found in motor and behavioral tests.

Since then, this model has been widely used to study mechanisms of brain damage, brain plasticity and possible therapeutic interventions after HIE.

Neural Stem/Progenitor Cells

NSPC are cells with a self-renewing capacity, and have the potential to generate cells of both neuronal and glial lineages. During brain development, there are at least three main types of NSPC:

  1. (1)

    Neuroepithelial progenitors (NE) are the proliferative cells that form the pseudo-stratified epithelium of the ventricular zone (VZ). Initially, NE have radial processes and divide symmetrically, increasing the pool of progenitors. In humans, at around embryonic day 33 (3 days after neural tube closure) these cells give rise to the first neurons of the brain. Thereafter, NE will also give rise to the other two types of NSPC: radial glia cells (RGC) and intermediate progenitors (basal progenitors).

  2. (2)

    RGC appear after the onset of neurogenesis and have some astroglial characteristics, such as the expression of astroglial markers and the presence of glycogen granules. Besides being a neuronal and glial progenitor, RGC have radial processes that extend from the ventricular zone to the pial surface, serving as scaffolds for the migration of newborn neurons. From midgestation (around 20 gestational weeks) until birth, RGC differentiate into astrocytes. Later on, RGC will also give rise to ependymal cells and to NSPC that persist in the adult brain.

  3. (3)

    Basal progenitors (BP) are generated by asymmetrical division of NE and RGC, accumulating at the basal border of the ventricular zone, forming the subventricular zone (SVZ). BP express a distinct set of transcription factors than RGC and generate neurons to the upper cortical layers from the 20th to at least 25–27 gestational weeks (Bystron et al., 2008).

In the perinatal period, when HIE occurs, NSPC can be found in two neurogenic niches that will persist throughout adult life. In the SVZ, in the walls of the lateral ventricule, NSPC continuously give rise to new neurons that migrate to the olfactory bulb, where they replace local interneurons. In the hippocampus, NSPC are localized in the subgranular layer of the dentate gyrus, giving rise to neurons of the inner granule cell layer.

As a source for cell-based therapies, NSPC can be obtained from the neurogenic regions of fetal or adult brain. Although several cell surface markers, such as Lex/CD15 and CD133, can be used to improve identification and isolation of NSPC, none of them can be used as a true NSPC marker. Thus, NSPC usually are defined based on their functional properties, such as self-renewal capacity and multipotency.

NSPC can be expanded in vitro in the presence of mitogens, forming floating cell clusters called neurospheres, which contain a heterogeneous population of proliferating cells, including self-renewing neural stem cells, multipotent progenitors, and more restricted progenitors. After removal of mitogens or growth factors, these cells spontaneously differentiate into astrocytes, oligodendrocytes and neurons. Upon exposure to different combinations of growth factors and culture conditions, NSCP may be induced to generate increased numbers of a desired cell population. As an example, numerous protocols have been created to optimize the formation of dopaminergic neurons from fetal and adult NSPC.

NSPC pools in the developing brain are regionally and temporally distinct. NE and RGC from different VZ subregions express a different combination of transcription factors, producing different subtypes of neurons. Moreover, different neuronal phenotypes are produced in a precise temporal sequence during development. When cultured in the presence of mitogens in vitro, NSPC seem to lose the regional specification. After long-term expansion in the neurosphere system, there is a shift towards gliogenesis and most of the generated neurons have an inhibitory gamma-aminobutyric acid (GABA)ergic phenotype, which represents one important limitation for a cell replacement therapy.

Alternatively, NSPCs can also be obtained from pluripotent stem cells. Embryonic stem cells (ESC), derived from the inner cell mass of blastocysts, spontaneously undergo neural differentiation when inhibitory signals for neural differentiation, such as bone morphogenetic protein 4 (BMP4), are absent. ESC neural differentiation occurs through progressive lineage restriction, recapitulating neural development. NE, RGC and BP can be efficiently obtained during this process. Accordingly, many protocols to obtain NSPCs and different neuronal phenotypes from ES were developed in the last years.

In one interesting study, ES-derived cortical progenitors sequentially generated several types of cortical neurons in vitro, in a temporal pattern that resembled the appearance of each neuronal subtype in vivo. However, ES-derived neurons expressed typical markers of the occipital cortex and exhibited a very specific pattern of projection when transplanted into neonatal cortex, suggesting a visual/limbic identity. Since the cells were transplanted in the frontal cortex, indicating that host factors were not involved in the specification of the neuronal phenotype, it remains to be determined how the intrinsic regional specification of cortical progenitors before transplantation could limit a cell replacement therapy (Gaspard et al., 2008).

An important risk associated with the use of ES-derived NSPC is the formation of teratomas if undifferentiated ES persist in the transplant pool. The risk of neural overgrowth, even when undifferentiated ES are not present, is also a critical issue.

Induced pluripotent stem cells (iPSC), obtained after epigenetic reprogramming of adult cells by a combination of transcription factors, can also be used to generate NSPC. Human iPSC can produce NE cells, which respond to extracellular cues, differentiating into regional progenitors and producing functional neurons. Although ES and iPSC use the same transcriptional networks for neural differentiation, iPSC are less efficient and show an increased variability (Hu et al., 2010). On the other hand, iPS-derived NSPC could be obtained after reprogramming of somatic cells from the own patient, allowing an autologous transplantation.

Alternatively, generation of induced neuronal cells directly from fibroblasts or other somatic cells, using a combination of neural-lineage-specific transcription factors, is also possible, although a better characterization of the neuronal subtypes obtained with this technique is still needed.

Finally, mesenchymal stem/progenitor cells (MSC) and human umbilical cord blood cells (HUCB) can be induced to differentiate into neural-like cells under specific culture conditions. However, in most of the studies, characterization of newly generated neural cells was based on morphological criteria and on the expression of neural markers. While a few studies showed the presence of functional properties, such as the firing of action potentials, by MSC- and HUCB-derived neuron-like cells, other studies failed to reproduce it.

When transplanted in the developing or adult brain, MSC and HUCB do not differentiate into neurons, even when injected after an ischemic brain injury. Therefore, while is still controversial if MSC and HUCB can be induced to differentiate into neural cells in vitro, it is clear that these cells do not spontaneously adopt a neural cell fate in the host brain.

Endogenous NSPC Response to HIE

NSPC biology is tightly controlled by the local environment. Throughout embryonic development and adult life, NSPC are localized in specific brain regions, the neural stem cell niches, where NSPC are exposed to diffusible factors, matrix glycoproteins and cell-to-cell contacts that control their proliferation, differentiation and migration capacity. After a HI brain injury, several components of the neural stem cell niches are altered, including changes in glial cells, blood vessels, cerebrospinal fluid, and oxygen tension. All these alterations result in increased NSPC proliferation and neurogenesis in the SVZ and in the hippocampus.

In the SVZ, an increased generation of new neurons persisted for at least 5 months after the injury. It was shown that some of these newborn neurons migrated from the SVZ to the cerebral cortex, occupying the cell-sparse columns produced by the hypoxic-ischemic injury, while most of them migrated to the damaged striatum. Nevertheless, two major limitations were pointed out by these studies. First, around 85% of the newborn neurons died before maturation (Yang et al., 2007). Second, most of the neuroblasts that migrated to the striatum differentiated into calretinin-positive interneurons. Since in the striatum, most of the neurons are GABAergic medium-sized spiny projection neurons and <5% are interneurons (divided into 4 populations, based on the expression of calretinin, parvalbumin, somatostatin, or choline acetyltransferase), the restricted fate of newborn striatal neurons is an obstacle for the regeneration of this brain region (Yang et al., 2008).

Recently, Yang and colleagues showed that almost all the newborn calretinin-positive neurons in the HI striatum are derived from NSPC expressing the transcription factor Emx1 in the dorsolateral SVZ. Moreover, they showed that Emx1-expressing NSPC from the postnatal SVZ give rise almost exclusively to calretinin-positive neurons, both in the intact and in the damaged striatum (Wei et al., 2011).

A relevant aspect of both embryonic and postnatal neurogenesis is the regional heterogeneity of progenitor zones. Different SVZ regions give rise to different neuronal phenotypes in the postnatal brain, indicating a restriction in the types of neurons that can be generated by NSPC from different SVZ rostrocaudal regions. The results of Wei et al. (2011) suggest that this restriction continues in the dorsolateral SVZ after HI, representing an important limitation of the endogenous regenerative response.

Microglial activation, endothelial activation, and the infiltration of blood-derived monocytes and lymphocytes may also control the NSPC response after an injury. Microglia activation occurs in the first hours after the HI insult, persisting for at least 1 week. Microglia regulates NSPC differentiation in vitro and this effect is critically dependent on microglial phenotype.

Moreover, immune cells produce chemokines that are involved in the recruitment of blood-derived cells to the damaged brain. Some of these chemokines, such as stromal-derived factor 1 (SDF-1) and monocyte chemotactic protein-1 (MCP-1), are involved in the migration of NSPC to sites of ischemic injury. While MPC-1 expresion is up-regulated in the HI brain, increasing in the first few hours after the injury, SDF-1 is constitutively expressed in several regions of the developing early post-natal brain. In the first 3 days after HI, SDF-1 expression increases in the striatum, hippocampus, SVZ and white matter. Importantly, SDF-1 is produce mainly by reactive astrocytes and endothelial cells after HI (Imitola et al., 2004; Miller et al., 2005).

Taking advantage of this self-repair mechanism, increasing the formation of new neurons, their migration, survival and differentiation into the appropriate phenotypes, is a promising new strategy to induce neuronal replacement. Although the manipulation of the endogenous NSPC pool is still far from clinical translation, there are a few candidate drugs that successfully increased neurogenesis in preclinical studies.

One candidate is erythropoietin, a drug used for the treatment of anemia in premature infants. Erythropoieting improves neurological outcome in animal models of HIE through several mechanisms, including neuroprotection and induction of revascularization and neurogenesis. In addition, erythropoietin treatment in newborns with HIE was shown to be safe and feasible in two recent clinical trials (Xiong et al., 2011). Other candidates are modulators of the endocannabinoid system, such as WIN55212-2, a synthetic cannabinoid receptor agonist, that increased both neurogenesis and the generation of mature oligodendrocytes after HI in rats (Fernández-López et al., 2011). Trophic factors could also be used to increase neurogenesis, as showed by the administration of basic fibroblast growth factor (bFGF) or a combination of brain-derived neurotrophic factor (BDNF)/epidermal growth factor (EGF) after brain ischemia or HI damage in early postnatal rats, respectively (Im et al., 2010; Jin-Qiao et al., 2009).

One important question that remains to be addressed is how HIE affects the function of RGC. At the time of HIE, RGC are differentiating into astrocytes. In a model of acute HI injury in the preterm-like brain it was shown that the radially oriented pattern of RGC processes is disrupted, correlating with the appearance of reactive astrocytes (Sizonenko et al., 2008). However, it is still unknown if HIE accelerates the transformation of RCG into astrocytes.

SVZ glial progenitors are also affected by HIE and there is an increased production of SVZ-derived astrocytes after the injury, contributing to reactive gliosis and to glial scar formation. Overproduction of EGF, leukemia inhibitory factor (LIF) and transforming growth factor beta 1 (TGF-β1) are involved in this process, increasing the differentiation of glial progenitors into astrocytes and reducing the formation of oligodendrocytes (Bain et al., 2010).

Finally, although new oligodendrocytes are generated after HIE, most of the new preoligodendrocytes fail to initiate myelination due to an arrest in maturation (Segovia et al., 2008). Since this arrest is probably caused by alterations in the microenvironment, such as the presence of hyaluronic acid in the glial scar, it could also affect the differentiation of transplanted NSPC or oligodendrocyte progenitor cells into functional oligodendrocytes.

Therefore, understanding the role of different glial and immune cells after HIE may be a necessary step to improve the efficacy of the endogenous regenerative response and of NSPC-based therapies. Besides regulating the regenerative potential of the endogenous NSPC pool, glial and immune cells may also influence migration, differentiation, survival and integration of transplanted NSPC and their progeny.

NSPC Transplantation in HIE

In an attempt to re-establish neural circuitry in the damaged newborn brain, two groups performed the first neural transplantation studies in hypoxic-ischemic rats. Elsayed et al. (1996) showed that fetal neocortical tissue blocks from embryonic day 13 rats could survive in up to 80% of the transplanted animals for at least 2–6 weeks, even in animals with severe brain infarcts. Then, Jansen et al. (1997) performed an intracerebral transplantation of cell suspensions from the sensorimotor cortical region of embryonic day 16 rats, 3 days after HI. The grafts were well integrated into the sensorimotor cortex in 62% of the animals, 10–12 weeks later. Importantly, the animals with well integrated grafts had a better motor performance than those in which the grafts were not integrated into the brain, as well as than the animals that were not treated.

One of the main contributions of these studies was to show the long term survival of neural grafts in the damaged brain. On the other hand, despite the functional benefit observed after treatment, ethical concerns related to the use of neural tissue from post-mortem embryos, as well as problems related to the availability of donors and to standardization of the grafts, would limit the clinical utilization of fetal tissue as a routine therapy.

In this regard, transplantation of NSPC offers a promising alternative for neural cell replacement therapies. Since these cells can be obtained from different sources and expanded in culture, large-scale generation of NSPC could overcome the problem of limited donor tissue availability.

Unlike Parkinson’s disease, in which the main damage is restricted to a well localized region, HIE affects multiple brain regions. The need to replace a large variety of neurons in several brain regions represents a major obstacle for a successful therapy using NSPC in HIE. The transplanted cells should migrate to the damaged areas, where they should differentiate into the appropriate neuronal subtypes, form new connections and extend long axons, restoring function. In this regard, several studies showed increased migration capacity and neuronal differentiation of NSPC in the HI brain.

The homing of NSPC to the HI brain was first demonstrated by one study that showed the potential of NSCP to migrate long distances, crossing the corpus callosum, from the contralateral hemisphere to areas with increased SDF-1 expression in the HI hemisphere. In addition, the role of SDF-1 in NSPC migration was also observed in vitro, showing that this chemokine is a potent attractant for NSPC in the HI brain (Imitola et al., 2004).

The permissive environment for NSPC migration in the HI brain was further demonstrated by one study in which it was observed that SVZ-derived NSPC migrated from the injection site to the surrounding cortical areas with no directional preference, when transplanted in sham-operated or control postnatal day 7 animals. However, when injected in the HI brain, NSPC migrated preferentially towards the area of injury. While NSPC differentiated mainly into astrocytes in sham-operated and control animals, neuronal differentiation was observed in the HI brain, although only a small number of the grafted cells expressed a marker of mature neurons 14–21 days after transplantation. Regarding the best time for NSPC transplantation, it was shown that the migration pattern of NSPC in the HI brain was the same whether the cells were transplanted 24 h or 5 days after the injury. (Zheng et al., 2006).

Accordingly, Snyder and colleagues observed that a robust engraftment could be achieved if the cells were transplanted shortly after the HI injury, especially 3–7 days following HI, while no engraftment was observed if the cells were transplanted 5 weeks after the injury. They also demonstrated that while 5% of the donor NSPC differentiated into neurons in the HI brain, no neuronal differentiation was observed when the cells were transplanted in the intact neocortex. Likewise, NSPC gave rise to five times more oligodendrocytes in the HI brain than in the intact brain (Park et al., 2006a).

Taken together, these studies showed that the neonatal HI brain induces the migration of intracerebrally transplanted NSPC to damaged areas in the acute/subacute phase of the injury, through a SDF-1-dependent mechanism. Moreover, the HI brain induces an increase differentiation of donor cells into neurons and oligodendrocytes, although the percentage of newly formed neurons is still small.

Alternatively, NSPC could be seeded in biomaterial scaffolds before transplantation. These scaffolds serve as a temporary extracellular matrix, increasing survival and integration of the cells, when the host tissue microenvironment is hostile. This approach could be used in more severe cases of HIE, when a cavity (porencephalic cyst) that would not support the survival of new cells is present. Transplantation of NSPC seeded in a polyglycolic acid scaffold into the cyst of HI animals resulted in long-distance connections between donor neurons and their targets in the penumbra and in the contralateral hemisphere. Neovascularization of the graft by the host tissue and alteration of the trajectory of host neuronal processes toward the scaffold further demonstrated the complex interactions between the host tissue and the implant (Park et al., 2002).

The increased migration capacity of NSPC in the HI brain could also be explored to deliver genes to the brain. Several trophic factors, including nerve growth factor (NGF), BDNF and insulin-like growth factor 1 (IGF-1), were shown to reduce HI brain injury, improving cognitive and motor function through multiple mechanisms. It remains to be determined if genetically engineered NSPC could be used to deliver these trophic factors to the HI brain, combining cell replacement with gene therapy.

In addition, genetic manipulation could be used to modulate NSPC properties, such as proliferation and differentiation potential, in an autocrine/paracrine fashion. In this regard, one study showed that genetically modified NSPC overexpressing the neurotrophic factor neurotrophin-3 (NT-3) generated more neurons in the HI brain than control NSPC. Remarkably, more than 80% of the NT-3 overexpressing NSPC-derived cells differentiated into neurons in the penumbra, giving rise to GABAergic, cholinergic and glutamatergic neurons (Park et al., 2006b).

In another elegant study, Kiss and colleagues observed that fibroblast growth factor 2 (FGF-2) overexpression increased proliferation and migration of NSPC in brain slices. FGF-2 overexpression also increased the number of NSPC-derived immature neurons after transplantation into the HI brain. They observed the formation of neurons expressing the GABA synthesizing enzyme glutamic acid decarboxylase-67 (GAD67) and the calcium binding protein calretinin in the ischemic brain, but none of the cells differentiated into calbindin- or parvalbumin-expressing neurons (Dayer et al., 2007).

Potential Benefits of NSPC Transplantation in HIE

In the last years, the potential therapeutic effects of NSPC have been demonstrated in many models of central nervous system injury, including neurodegenerative diseases, acute ischemic, excitotoxic and traumatic insults, and demyelinating disorders. In all these models, NSPC transplantation provided functional benefits through multiple mechanisms, combining cell replacement and bystander mechanisms. NSCP can secrete many trophic factors and increase the production of these factors by host cells. NSPC also produce cytokines and chemokines, being able to promote local immunomodulation. As a result, NSPC transplantation have been shown to decrease neuronal death and to increase brain plasticity and endogenous regenerative mechanisms in the adult, as well as in the neonatal damaged brain.

The first evidence that NSPC could have a therapeutic effect in HI animals, came from an study that showed that ESC-derived NSPC improved the spatial memory deficits, when transplanted 2–3 days after the insult. (Ma et al., 2007). However, the mechanisms involved in this improvement were not evaluated.

Recently, one study showed that human ESC-derived NSPC persisted in the brain of HI rats for at least 30 days after intracerebral transplantation, improving motor function of the treated animals. Grafted cells differentiated into neurons and astrocytes, with around 47% of the cells having a GABAergic phenotype, indicating a trend towards the formation of inhibitory neurons. It was also shown that NSPC enhanced axonal sprouting, possibly through activity-mediated and/or neurotrophic effects. Indeed, NSPC transplantation resulted in increased levels of host endogenous genes involved in neurogenesis and neurotrophic support, such as IGF-1, FGF-2, BDNF and neurturin. (Daadi et al., 2010).

Despite some evidences that NSPC transplantation can decrease neuronal death in stroke, a significant reduction of the infarct volume was not observed by Daadi et al. (2010).

It is possible that a combination of therapies may be required to enhance NSPC-mediated neuroprotection in HIE. NSPC obtained from the telencephalon of embryonic day 14 rat fetuses reduced HI brain injury only when injected in combination with chondroitinase ABC, which degrades glycosaminoglycans side chains of chondroitin sulphate proteoglicans, but no when injected alone (Sato et al., 2008). The presence of chondroitin sulphate proteoglicans in the glial scar is an obstacle for long-term survival and integration of NSPC in the chronically injured spinal cord, where the cells preferentially differentiate in oligodendrocytes (Karimi-Abdolrezaee et al., 2010). On the other hand, chondroitinase ABC reduces NSPC proliferation, as well as reduces neurogenesis in the embryonic VZ (Sirko et al., 2007). Therefore, it is possible that chondroitinase ABC could increase NSPC migration, survival and neuroprotective effect, at the expense of neuronal differentiation.

Translational Aspects of NSPC Transplantation in HIE

NSPC transplantation is still far from becoming a real treatment for patients with HIE. Most of the studies transplanting NSCP in HI animals assessed the migration and differentiation potential of these cells in the damaged brain, while only two studies evaluated the long-term effect of the treatment in cognitive and motor functions. Thus, the long-term benefit of a NSPC-based therapy in HIE still needs to be shown by multiple research groups.

Another important question that remains to be determined is the related to the best route of administration of NSPC in newborns with HIE. Is an intracerebral injection really necessary?

Up to now, all studies transplanting NSPC in animal models of neonatal HIE have injected the cells directly in the brain. Since NSPC have a surprisingly capacity to home to the adult damaged brain after an intravascular transplantation, it is still necessary to evaluate the migration capacity and the potential benefits of an intravascular NSPC transplantation in the newborn brain, comparing the effects of intravenous, intra-arterial and intracerebral transplantation after HIE. In this regard, it was demonstrated that more cells reached the brain after intra-arterial than after intra-venous transplantation in an adult model of cerebral hypoxia-ischemia (Pendharkar et al., 2010). The difference is probably related to the retention of the cells in the lungs after the pulmonary passage and/or to the clearance of the cells by the innate immune system after an intra-venous transplantation.

When transplanted systemically, NSPC could also exert a therapeutic effect through an immunomodulatory effect in peripheral organs. When injected intravenously 2 h after a hemorrhagic stroke in adult rats, NSPC migrated to the spleen, resulting in a reduction of the neurologic deficits through an anti-inflammatory effect that was abolished if the animals had their spleen removed before the injury (Lee et al., 2008).

If the same effect could be observed in HIE, would a combination of systemic and intracerebral injections be necessary to enhance the therapeutic potential of NSPC transplantation?

It will also be important to address if, despite the permissive environment for NSPC migration in the HI brain, multiple injections sites will be necessary to permit widespread dispersal of the cells in all the affected regions, such as hippocampus, thalamus, basal ganglia and cerebral cortex.

Accordingly, neuroimaging methods are valuable tools to answer some of these questions in both preclinical and clinical studies.

Superparamagnetic iron oxide nanoparticles can be used to label NSCP, allowing non-invasive monitoring of migration and proliferation of the transplanted cells by MRI. In a recent study, iron-labeled murine NSPC were transplanted in the brain of hypoxic-ischemic rats, 3 days after the injury, or in age-matched controls. Whether injected in the striatum or into the ventricle, the cells remained in the original site of injection in controls. In the HI group, the cells migrated around 100–125 μm/day, reaching the border of the injury within 10–12 days, where an impressive increase in NSPC volume was observed over the first 4 weeks, indicating the high proliferative capacity of these cells in the damaged brain. Long-term follow-up of the animals revealed that, although NSPC volume decreased over time, many cells persisted for up to 58 weeks in the injured hemisphere (Obenaus et al., 2011).

In a case report, autologous umbilical cord blood-derived neural progenitor-like cells were labelled with iron oxide nanoparticles and transplanted in a 16-month-old child, 7 months after a global HI brain injury. The cells persisted in the vicinity of the lateral ventricles wall for up to 4 months after the intracerebroventricular transplantation and adverse effects were not observed (Jozwiak et al., 2010). Regardless of the controversy about the successful derivation of NSPC from blood cells, this study gives a preliminary support regarding the safety of transplanting iron-labelled cells in children.

The long-term survival of human and murine NSPC in post-natal day 7 rat brains contrasts with the demonstration that allogeneic and xenogeneic grafts might trigger an immune response leading to graft rejection in the adult central nervous system.

Although systemic immune suppression could be used to avoid graft rejection, immunosuppressive drugs are associated with an increased risk of opportunistic infections and with an increased susceptibility to malignancies.

Despite these risks, this is the approach that is being used in the first phase one clinical trial using ESC-derived cells. In this trial, adult patients with thoracic spinal cord injury will be immune-supressed from the time of human ESC-derived oligodendrocyte progenitor cell transplantation for 60 days (http://clinicaltrials.gov, identifier NCT01217008).

Immunosuppressive drugs will also be used in two ongoing clinical trials that are testing safety and preliminary efficacy of intracerebral human NSPC transplantation in children aged 6 months to 5–6 years with connatal Pelizaeus-Merzbacher disease (PMD) or infantile/late-infantile forms of neuronal ceroid lipofuscinosis (NCL) (http://clinicaltrials.gov, identifiers NCT00337636 and NCT01005004). In mouse models of infantile NCL, caused by a deficiency in the lysossomal enzyme palmitol protein thioesterase-1 (PPT1), transplanted NSPC produced PPT1 and reduced neurodegeneration. In PMD, a hereditary demyelinating disease generally caused by mutations of the gene that codes for the myelin protein proteolipid protein 1, it is expected that NSPC will produce oligodendrocytes to efficiently myelinate the brain.

The poor prognosis of infants with NCL and the lack of treatments for this disease justify the early translation from preclinical studies to clinical trials of a NSPC-based therapy. Given the great variation in neurodevelopmental outcome of chidren with HIE and the complex neuropathological alterations in the HI brain, preclinical studies are still needed before a clinical translation. These studies should address important questions regarding the time of transplantation, the safety and efficacy of different sources of NSPC, the delivery route, the need of multiple injections and the immunosuppressive strategy in cases of allogeneic transplantation.

In the future, the creation of public banks of human leukocyte antigens (HLA)-typed ESC lines and their differentiated cells, including NSPC, could provide an alternative to reduce the immunogenicity of ESC-derived NSPC transplantation. It was estimated that a bank with 150 ESC lines would provide a HLA-DR match for most of the potential recipients and only one HLA-A or HLA-B mismatch for 25–30% of the recipients (Taylor et al., 2005).

In conclusion, NSPC transplantation is a promising therapy for HIE that still has to be extensively tested in preclinical studies. When transplanted in the HI brain, these cells migrate long distances and differentiate into several subtypes of neurons, which survive and apparently integrate with the host cells. When injected in the acute phase of the HI injury, NSPC reduced motor and cognitive deficits, although the mechanisms involved in the neurological improvement are not completely clear. Therefore, preclinical studies addressing basic questions regarding the safety, efficacy and mechanisms of action of NSPC in the HI brain are urgently needed. Finally, the results of the first clinical trials using NSPC in infants with NCL and PMD will also be an important milestone for the clinical translation of a possible NSCP therapy in children.