Neural tissue engineering using embryonic and induced pluripotent stem cells

With the recent start of the first clinical trial evaluating a human embryonic stem cell-derived therapy for the treatment of acute spinal cord injury, it is important to review the current literature examining the use of embryonic stem cells for neural tissue engineering applications with a focus on diseases and disorders that affect the central nervous system. Embryonic stem cells exhibit pluripotency and thus can differentiate into any cell type found in the body, including those found in the nervous system. A range of studies have investigated how to direct the differentiation of embryonic cells into specific neural phenotypes using a variety of cues to achieve the goal of replacing diseased or damaged neural tissue. Additionally, the recent development of induced pluripotent stem cells provides an intriguing alternative to the use of human embryonic stem cell lines for these applications. This review will discuss relevant studies that have used embryonic stem cells to replicate the tissue found in the central nervous system as well as evaluate the potential of induced pluripotent stem cells for the aforementioned applications.


Introduction
Th e human nervous system can be divided into two major components: the central nervous system (CNS) and the peripheral nervous system [1]. Th e CNS consists of the brain and the spinal cord with the blood-brain barrier restricting the types of biomolecules that can reach these organs [2]. Th e majority of neural tissue found in the CNS consists of two cell types: neuronal cells and glial cells. Neurons serve as the main information transmitting unit of the nervous system, which can be classifi ed as either sensory, allowing them to detect stimuli from the environment, or motor, responsible for the generation of movement through signaling with muscle tissue [3]. Glial cells encompass a number of diff erent types of support cells, including astrocytes and oligodendrocytes found in the CNS [4,5]. Th is review will discuss the use of embryonic stem (ES) cell therapy as a method of treating injuries and diseases that infl ict damage to the CNS. Th ese studies are particularly relevant now as the fi rst human ES-cell-derived therapy is currently being evaluated in clinical trials as a potential method for treating spinal cord injury (SCI) [6].
ES cells possess two hallmark characteristics: the ability to self-renew and pluripotency [7]. Th e pluripotent nature of ES cells allows them to generate the cells found in neural tissue, including neurons and glia. As a result, ES-cell-based regeneration strategies have been investigated for a number of diseases as well as for repairing mechanically damaged nerve tissue. While many other types of stem cells have been evaluated for their potential to promote neural repair, this review will focus specifically on the attempts made with ES cells as this work will be most applicable to developing therapies using induced pluripotent stem (iPS) cells. First generated in 2006, iPS cells are produced from adult somatic cells, such as skin cells, by inducing specifi c factors that restore pluripotency [8-10] (Figure 1). Th e recent generation of these cell lines serves as an exciting alternative to traditional ES cell lines and recent research using iPS cells will also be critically examined in terms of the future of stem-cellbased therapies for repairing neural tissue.
Many of the studies detailed in this review use mouse ES cells as a model system for studying cell behavior with the goal of translating this knowledge to human ES cells. While this approach does have merit, it is important to note some of the diff erences between mouse and human ES cell lines. One of the major diff erences is that mouse ES cells can be maintained in the presence of leukemia inhibitory factor (LIF) on gelatin substrates in a relatively cost eff ective manner while human ES cells are cultured either on a feeder layer of cells or on a Matrigel surface in the presence of defi ned media [11]. Mouse and human ES cells diff er in protein expression patterns, including the Abstract With the recent start of the fi rst clinical trial evaluating a human embryonic stem cell-derived therapy for the treatment of acute spinal cord injury, it is important to review the current literature examining the use of embryonic stem cells for neural tissue engineering applications with a focus on diseases and disorders that aff ect the central nervous system. Embryonic stem cells exhibit pluripotency and thus can diff erentiate into any cell type found in the body, including those found in the nervous system. A range of studies have investigated how to direct the diff erentiation of embryonic cells into specifi c neural phenotypes using a variety of cues to achieve the goal of replacing diseased or damaged neural tissue. Additionally, the recent development of induced pluripotent stem cells provides an intriguing alternative to the use of human embryonic stem cell lines for these applications. This review will discuss relevant studies that have used embryonic stem cells to replicate the tissue found in the central nervous system as well as evaluate the potential of induced pluripotent stem cells for the aforementioned applications.
signaling pathways that regulate diff erentiation and the markers that indicate pluripotency [12]. For example, undiff erentiated mouse ES cells express stage-specifi c embryonic antigen (SSEA)-1 while undiff erentiated human ES cells express the SSEA-3 and SSEA-4 markers [13]. Th us, the information gained in mouse ES cell studies does not always directly translate to human ES cell lines due to these intrinsic diff erences.
Both human ES and iPS cells exhibit high variability between diff erent cell lines as illustrated by a recent study published in Cell that mapped the genome-wide DNA methylation patterns and gene expression for 20 ES and 12 iPS cell lines [14]. Th ey used the information obtained from the ES cell lines as a reference to evaluate the expression patterns of the iPS cell lines to see if they fell within an acceptable range of 'stemness' . Another study used transcriptional profi ling analysis to show that the iPS cell lines have residual gene expression from the donor cells after reprogramming, with certain donor cells being reprogrammed more effi ciently [15]. Both of these studies illustrate the need for defi ned standards to use for evaluating newly derived ES and iPS cell lines to determine their suitability for clinical applications.

Embryonic stem-cell-based therapies for neural tissue replacement
Many diff erent studies have used ES cells to generate replacement neural tissue for a variety of diseases and disorders ( Table 1). One of the major considerations when working with ES cells is how to induce them to diff erentiate into the specifi c neural phenotypes necessary for treating the particular application. Regenerationpromoting strategies can include directly deriving the desired cell type to be replaced or generating supporting glial cells that secrete factors to help restore lost functionality. In terms of diff erentiation protocols, the most desirable methods would produce a highly purifi ed population of specifi c cellular phenotypes for trans plantation as any undiff er en tiated ES cell clusters remaining can proliferate in an uncontrolled fashion, leading to teratoma formation [16]. To avoid teratoma formation, many diff erent methods have been investigated to  While specifi c protocols for directing stem cell diff erentiation into neural lineages will not be reviewed in depth here, several reviews on the subject describe these processes in more detail [21,22]. When developing EScell-based treatments for neural diseases and disorders, it is important to consider what specifi c cell populations could potentially restore lost function. For certain neurodegenerative disorders that aff ect specifi c neuronal popula tions, the goal is to transplant a highly diff erentiated mature population of neurons to replace the lost cells. For promoting recovery after traumatic CNS injury, a variety of ES-cell-based therapies have been explored as neural progenitors could potentially secrete regenerationpromoting factors while the transplantation of ES-cellderived neurons and oligodendrocytes to restore the lost mature cellular populations has also been studied. Th e method of transplanting the cells in the desired location in the CNS should also be carefully evaluated to ensure cell viability and prevent unwanted diff erentiation. Other relevant issues relating to ES-cell-based therapies include the potential of the transplanted cells to induce an immune response. Th ese issues will be discussed along with the relevant studies for each of the following diseases and disorders.

Traumatic brain injury
It is estimated that 3.2 million people in the United States currently suff er from reduced function after hospitalization as a result of traumatic brain injury (TBI) [23]. Th e impact and resulting lesion from TBI often results in cognitive, sensory, motor, and emotional impairments. Several studies have investigated the use of ES-cellderived therapies, specifi cally the transplantation of neural progenitors and undiff erentiated ES cells, as a way of treating TBI with the hope of alleviating the aforementioned symptoms as these cells could secrete factors that would induce regeneration. In one of the fi rst studies to investigate such a strategy, neural progenitors produced by treating mouse ES cells with retinoic acid were transplanted into the lesion site one week after administration of a cortical impact injury [24]. Th ey found that these cells were able to prevent the formation of necrotic cavities that would normally occur after injury and were able to improve the sensorimotor function. Another group showed transplanted mouse ES cells migrated to the injury sites consisting of lesions induced by fl uid injection in the mouse brain, suggesting that the damaged brain tissue secreted factors to induce ES cell migration [25]. In a similar study, the ability of undiff erentiated mouse ES cells to promote recovery after TBI was evaluated [26]. Th e animals that received the ES cells performed better on the Rotorod test, measuring the animal's ability to stay on a moving rod, and had better Neuroscores, refl ecting an improvement in their neuromotor function. However, they observed tumor formation in two of the ten animals receiving the cells, illustrating a limitation of using undiff erentiated ES cells as a therapy. In a follow-up study, they reported that the infl ammatory response associated with TBI impaired ES cell survival and integration after implantation into injured rat brain [27]. Th ese studies illustrate the potential of ES cell therapies for the treatment of TBI, but much work remains before such therapies will be suitable for evaluation in clinical trials.

Parkinson's disease
Th e loss of dopaminergic neurons located in the substantia nigra is one of the major hallmarks of Parkinson's disease (PD) [28]. Th ese neurons secrete dopamine, which regulates cortical and thalamic activity. Loss of these dopamine neurons in the substantia nigra results in motor dysfunction, including tremor, rigidity, and bradykinesia, as well as non-motor symptoms, including anxiety and depression. Th e current treatment for PD consists of the drug levodopa, often referred to as L-dopa, which is a dopamine precursor that can cross the bloodbrain barrier, where it is then metabolized into dopamine [29]. However, long-term use of levodopa leads to more motor function dysregulation [30]. Additionally, the transplantation of fetal tissue as a method of replacing the lost dopamine neurons has also been investigated, but the most recent clinical trials did not show any benefi t to receiving this treatment [31][32][33]. Th e goal of ES-cell-based therapies for PD is to generate a highly defi ned, dopaminergic neuron population suitable for transplantation into the substantia nigra.
Many groups have investigated the use of ES-cellderived dopaminergic neurons for their potential to treat PD as an alternative to the current standard of care. One group implanted undiff erentiated mouse ES cells directly into the mid-brain, where the cells diff erentiated into functional dopaminergic neurons and promoted functional recovery in a rat model of PD induced by injections of 6-hydroxydopamine [34]. Another study developed an extensive fi ve step protocol for generating dopaminergic neurons from mouse ES cells and showed that transplantation of these cells improved function in the same PD rat model [35]. Similar studies were also performed using human ES-cell-derived dopaminergic neurons and showed that transplantation of these cells into a rat model of PD produced similar improvements in function [36][37][38][39]. Other groups have shown that co-cultures of ES cells with astrocytes and stromal cells induces diff erentiation into dopaminergic neurons [37,40,41]. Finally, a recent study developed an effi cient protocol for large scale production of dopaminergic neurons from human ES cells that reduces the potential for tumor formation, bringing this therapy closer to standards required for clinical testing [39].

Huntington's disease
Huntington's disease (HD) is a rare, inherited neurodegenerative disorder characterized by a loss of medium spiny projection neurons in the striatum [42]. Symptoms include loss of muscular coordination along with progressive cognitive decline. Many clinical studies have investigated the transplantation of fetal-derived tissue into the brain as a potential treatment for HD [43][44][45][46] and ES cells provide an alternative to the use of such tissues by providing an alternative way to replace the lost neuronal population.
One study induced human ES cells to diff erentiate into neural progenitors and then transplanted these cells into a rat model of HD induced using quinolinic acid [47]. Th e animals receiving the treatment performed better on an apomorphine-induced rotation test compared to sham treated animals and no tumor formation was observed. While they did observe neuronal diff erentiation of the transplanted progenitors, they did not investigate the mechanism behind the observed recovery to determine if it was due to the secretion of factors that preserved the existing cells or if the transplanted cells were contributing to the functional recovery. A second study utilized a three step protocol to induce human ES cells to diff erentiate specifi cally into striatal spiny neurons for transplantation [48]. While the initial results were promising, after 13 weeks, the grafts overgrew the implantation site, leading to deleterious side eff ects. However, these grafts did not contain undiff erentiated ES cells, but the regions of overgrowth did contain nestin-positive neural progenitor cells. A diff erent study showed that treatment of neural progenitors derived from human ES cells with the protein Noggin enhanced neuronal diff erentiation post-implantation in a rat model of HD [49]. Th ese studies demonstrate the viability of using ES-cell-derived therapies as a way to replace the lost neuronal populations due to HD. However, the issue of graft overgrowth as well as determining how these transplanted cells contribute to function recovery will have to be addressed and characterized before these therapies can be translated to the clinic.

Alzheimer's disease
Similar to HD, Alzheimer's disease (AD) is another neuro degenerative disease that tends to aff ect people over the age of 65 years. Th e clinical symptoms of AD include progressive cognition deterioration due to the loss of cholinergic neurons [50]. In terms of histology, AD is associated with the appearance of amyeliod plaques and neurofi brillary tangles in the brain. Many groups have investigated the use of ES-cell-derived therapies as a means of treating AD in pre-clinical models by diff erentiating ES cells into cholinergic neurons as this cell population is one the most aff ected by AD.
One group derived neurospheres from mouse ES cells and transplanted the resulting cells into a mouse model of AD [51]. Th ese cells diff erentiated into choline acteyltransferase-positive neurons and reduced memory deteri oration compared to control mice that received undiff erentiated ES cells. Th e control mice receiving the undiff erentiated ES cell transplants also showed tumor formation and signifi cant memory deterioration. Another study took a similar approach to diff erentiating mouse ES cells and also examined the eff ect of adding the growth factor sonic hedgehog (Shh) during the neurosphere formation step. Th ey observed that the diff erentiation of neural progenitors into cholinergic neurons was enhanced by priming the neurospheres with Shh and the treated animals that received these cells showed a significant improvement in memory as indicated by performance in a water maze test [52]. Th ese studies provide preliminary evidence for an ES-cell-based therapy for AD, but such results would need to be repeated with human ES cells in pre-clinical testing to determine the viability of such a strategy.

Spinal cord injury
Th e complex arrangement of neurons, oligodendrocytes, and astrocytes allow the spinal cord to coordinate movement and sensation between the brain and the limbs and disruption of this structure due to mechanical injury can result in paralysis [53]. Th e extent of paralysis depends on the degree of trauma and the location of the injury in the spinal cord. Th e initial mechanical injury triggers a secondary cascade of events and damage from both of these processes must be addressed when designing a suitable therapeutic repair strategy. ES-cell-based therapies for the treatment of SCI have focused on generating both neurons and oligodendrocytes to promote recovery after injury.
Th e McDonald group [54] transplanted retinoic-acidtreated ES cells into a rat model of SCI and these cells were able to survive and diff erentiate into neurons, oligodendrocytes, and astrocytes while promoting recovery as indicated by regained hind limb function. Other groups using similar approaches observed improve ments in motor and sensory function after transplantation of pre-diff erentiated mouse ES cells [55][56][57]. Another group showed that transplantation of bone marrow stromal cells along with retinoic-acid-treated mouse ES cells prevented tumor formation in a rat model of SCI [58]. Other studies have used genetically modifi ed ES cell lines to improve cell survival and diff erentiation after transplantation, but the use of virus-mediated transfection methods limit the feasibility of such approaches [59][60][61][62].
More recent work has focused on the production of specifi c cellular populations from ES cells for treatment of SCI. Many groups have chosen to diff erentiate ES cells into motor neurons as a therapeutic strategy. One approach involved inducing mouse ES cells to diff erentiate into motor neurons using a combination of retinoic acid and Shh and transplanted these cells into a virusmediated rat model of SCI, with approximately 25% of these cells surviving one month after transplantation [63]. Another group transplanted ES-cell-derived motor neurons along with olfactory ensheathing cells into a rat model of SCI and observed partial functional recovery [64]. Th e Keirstead group has shown that motor neurons derived from human ES cells can also promote functional recovery after SCI in a rat model [65]. Th is strategy also holds potential for the treatment of amyotrophic lateral sclerosis, a neurodegenerative disease characterized by the loss of motor neurons [66].
Th e other major therapeutic strategy has involved diff er entiating ES cells into oligodendrocytes for treating SCI [67][68][69][70][71][72][73][74]. One of the most promising approaches to treatment of SCI involves diff erentiating human ES cells into oligodendrocyte precursors using an extensive 42-day diff erentiation protocol [71] and then transplanting these cells into the site of SCI [70,[72][73][74]. Th is therapy showed promising results in pre-clinical trials as these oligodendrocyte precursors diff erentiated into mature oligodendrocytes, promoting functional recovery in two distinct models of SCI [72,74]. Th is therapy has become the fi rst human ES-cell-derived therapy to be evaluated in clinical trials [6], with the fi rst patient having already been enrolled [75]. Th e Keirstead group [65] has also begun researching the eff ects of transplanting motor neuron progenitors into the injured rat spinal cord with some promising results.
Several groups have used tissue engineering approaches that combine biomaterial scaff olds with ES cells for the treatment of SCI. Th e Sakiyama-Elbert group [76][77][78][79][80][81] has used a fi brin-based scaff old to deliver growth factors that promote ES cells to diff erentiate into neurons and oligodendrocytes and this approach has been shown to enhance functional recovery in a rat model of SCI. Th e Baharvand group [82] has investigated the transplantation of neural progenitors derived from human ES cells inside of collagen scaff olds for the treatment of SCI and this approach also resulted in an increase in locomotor function post-transplantation. Another group developed electrospun poly (ε)-caprolactone scaff olds, which support mouse ES cell culture and promoted diff erentiation into nestin-positive neural progenitors in an in vitro setting [83]. Many other biomaterials have been investigated for the treatment of SCI, but not in combination with stem cells [84,85].

Reprogramming somatic cells and the potential for engineering neural tissue
As mentioned in the Introduction, the recent development of iPS cells provides an exciting alternative to the use of ES cells. Th ese cells are generated from somatic cells, such as fi broblasts, by upregulating the expression of specifi c genes (Oct3/4, Sox2, c-Myc, and Klf4) that restore pluripotency [9,10]. Unlike traditional ES cell lines, the use of iPS cells allows for generation of pluripotent cell lines without the use of embryos as well as for the production of patient-specifi c iPS cell lines, which should reduce the risk of rejection after transplantation.
Several studies have investigated iPS cells and their potential for diff erentiating into neural phenotypes. A recent study demonstrated that neural diff erentiation in human iPS cells uses the same transcription networks as traditional human ES cell lines [86]. Th ey also observed lower diff erentiation effi ciency and increased variability compared to ES cells, suggesting that more effi cient diff er entiation protocols may need to be developed to fully utilize the potential of iPS cells (Figure 2). In work also done by the Yamanaka lab [87], 36 mouse iPS cell lines were evaluated to determine their potential for generat ing neural phenotypes after secondary neurosphere formation as well as their potential safety for transplantation as indicated by teratoma formation in an in vivo setting. After an induction period using retinoic acid, these cell lines demonstrated the ability to diff er entiate into the three cell types found in the CNS (neurons, oligodendrocytes, and astrocytes) and certain iPS cell lines did not form teratomas after implantation, leading them to be classifi ed as a 'safe' cell line. In a follow-up study, neural progenitor cells derived from 'safe' iPS cell lines were implanted into both uninjured spinal cord tissue and a pre-clinical model of SCI in mice [88]. In the uninjured mice, the cells diff erentiated into neurons and glia, while in the injured animals, these cells diff erentiated into mature oligodendrocytes and promoted functional recovery in the hind limbs of mice. Other work has derived neurons from iPS cells, which were shown to promote functional recovery in a rat model of PD [89].
Th ese studies indicate the ability of iPS cells to diff erentiate into neural phenotypes, illustrating their potential as an alternative to the use of traditional ES cell lines. Th e generation of 'safe' , non-teratoma-forming cell lines serves as an added potential benefi t of using iPS cells.
An alternative approach to diff erentiating pluripotent stem cells is to directly convert one mature cell type into the desired cell type by manipulating cell signaling pathways. Th e Wernig group [90] recently demonstrated that mouse embryonic fi broblasts can be directly converted to neurons if the appropriate factors were expressed, off ering a potential method of engineering neural tissue without the use of pluripotent stem cells. Much work remains to be done to determine the feasibility of such an approach for clinical applications.

Conclusion
Overall, a large body of work exists showing the potential of using pluripotent stem cells to produce replacement Both iPS cells and ES cells were positive for OCT4 at day 0, for PAX6 but not SOX1 at days 8 to 10, and for both PAX6 and SOX1 at day 15. (c) Fluorescence activated cell sorting analysis indicates that diff erentiating cells from H9 human embryonic stem cells (hESCs), iPS(IMR90)-1 and -4, iPS-M4-10, iPS-DF6-9-12, and iPS109 began to generate PAX6-expressing cells at days 6 to 8, and reached a plateau at day 14 but with diff erent effi ciency. (d,e) By 12 weeks in culture, many MAP2+ neurons also expressed synapsin (SYN); (e) higher magnifi cation indicates a punctuate staining pattern on the cell bodies and neurites. (f) Glial fi brillary acidic protein (GFAP)-positive astrocytes were present in diff erentiated cultures at 12 weeks. (g) O4-positive oligodendrocytes were observed in cultures after 16 weeks. O4, oligodendrocyte marker referring to oligodendrocyte clone number 4. Scale bars: 50 μm. Reprinted from [86] with permission from the National Academy of Sciences.
tissue for the CNS as discussed in this review. Th e current evaluation of oligodendrocytes derived from human ES cells shows the promise of this technology for the treatment of SCI and other neurological disorders and diseases while the generation of iPS cell lines now allows for generation of patient-specifi c neural tissue derived from pluripotent stem cells. Additionally, the recent work from the Wernig group showing the direct conversion of somatic cells into neurons provides an intriguing alternative to diff erentiating pluripotent stem cells as a means of replacing lost neural tissue.