By John S. Yu, MD and Moneeb Ehtesham, MD

• Figure 1
• Figure 2
• Figure 3
• Figure 4
• Figure 5
• Figure 6

The possibility of identifying all tumor microsatellites to treat each neoplastic pocket individually remains unrealistic given the potentially staggering number of microsatellites that may be associated with a large intracerebral tumor, and the limitations of current imaging technology and the difficulty of detecting the often extremely small microsatellites. An ideal treatment should, therefore, be capable of independently targeting glioma microsatellites without the need for operator involvement. Therefore, the therapeutic agent must be capable of distinguishing neoplastic cells from normal neurons or glia on the basis of either cell surface or secretory characteristics.

The demonstrated ability of NSC to track to intracerebral glioma microsatellites has brought to light the potential of their use as a tumor tropic treatment vehicle. Initial treatment options utilizing NSC generally focused on their ability to serve as reconstructive agents following tissue destruction resulting from infarction or neurodegenerative disease. The identification of pluripotent NSC within the subventricular zones in the adult CNS, as well as the ability of these cells to migrate to areas of intracranial pathology has raised the interesting hypothesis that intracranial NSC populations persist well into adulthood and may play a key role in endogenous tissue repair and repopulation following injury or insult [6]. The fulfillment of such a role would necessitate that NSC be responsive to chemotactic signals that would emanate from zones of pathology within the brain, thereby enabling NSC to follow chemokine gradients as they migrate towards sites of injury. The first clear demonstration that NSC could exhibit migratory activity towards sites of intracranial tumor was forwarded by Aboody and colleagues [7]. Aboody et al demonstrated that the immortalized murine NSC line C17.2 [8] expressing the reporter gene product B-galactosidase, could be identified within areas of disseminated intracranial tumor following injection either intracerebrally or intravenously into glioma bearing rodents. The authors also demonstrated that their NSC could be engineered to carry the bioactive transgene for cytosine deaminase, which could convert the systemically administered non-toxic pro-drug 5-fluoro-cytosine to the cytotoxic 5-fluoro-uracil, resulting in significant shrinkage of treated tumors compared to controls. These results clearly pointed to the potential of NSC as tumor tropic vehicles for therapeutic gene delivery. A further attractive aspect of the methodology described by Aboody and colleagues revolves around the ease of expanding an immortalized cell line, thereby facilitating the availability of patho-tropic NSC populations as "off the shelf products" for rapid in vivo transplantation when required. In comparison to relying on autologous NSC for therapeutic purposes, a mass-produced and widely distributed allogeneic NSC product would provide a more uniform and accessible source of these cells for therapy. However, it is important to note that despite these advantages, the use of immortalized NSC is limited by certain factors that could substantially restrict its relevance in a clinical setting. The use of a transformed cell line such as C17.2 is inherently problematic, as their remain concerns of potential tumorigenicity given the proliferative potential of cells immortalized by means of viral oncogenes. Additionally, the use of an allogeneic transplant carries with it the significant likelihood of immune rejection, and would therefore necessitate iatrogenic immune suppression in patients treated in this manner. Given the already existent immune defects found in patients with high grade gliomas, and the potential role that these immune deficiencies may play in enhancing the tumorigenicity of these neoplasms, the use of additional immune suppressive treatments may be highly undesirable. Also, it is unclear whether an immortalized cell line that has been grown in laboratory culture for many passages will retain the same degree and spectrum of chemotactic, and therefore tropic, sensitivities that primary NSC may exhibit. In light of these limitations, we hypothesized that the use of a syngeneic, non-transformed NSC population may be more relevant as a tool for delivering tumor toxic therapy to disseminated tumor pockets within the brain. Benedetti and colleagues first described the use of cytokine expressing primary NSC in the treatment of experimental gliomas [9]. Primary fetal- and post-natal derived NSC were engineered to secrete interleukin 4 (IL-4) by means of retroviral mediated gene transfer, and then transplanted into established intracranial rodent gliomas. This therapy resulted in a significant treatment benefit as evidenced by decreased tumor burden and prolonged survival. However, in their study Benedetti et al. did not characterize the migratory potential of these cells and therefore the mechanisms underlying survival benefit associated with their therapy was not readily apparent.

We have subsequently demonstrated in a murine model of experimental glioblastoma, that syngeneic NSC exhibit robust tropism for disseminated tumor and are capable of extensive migration within the brain [10]. NSC were harvested from the frontoparietal regions of day 15 fetal C57Bl/6 mice and cultured in vitro in combination with basic fibroblast growth factor (bFGF) and epidermal growth factor (EGF) as previously described [11]. Isolated cells grew primarily as spherical aggregates (Figure 1A) and were comprised mainly of cells that stained strongly for nestin, a marker of neural progenitor cells (Figure 1B). We confirmed their pluripotency by inducing in vitro differentiation into all three neural lineages, with identifiable populations of neurons, astrocytes, and oligodendrocytes (Figure 1C - 1E, respectively). NSC were then infected in vitro with replication defective adenovirus encoding the gene for B-galactosidase to facilitate in vivo tracking. After in vitro confirmation of robust B-galactosidase expression by infected cells (Figure 1F), NSC were stereotactically inoculated into established intracranial GL26 gliomas. Brains from treated animals were harvested and stained with an X-gal substrate to detect the presence of B-galactosidase expressing NSC. NSC were readily identifiable dispersed within inoculated tumors, and were clearly visible tracking glioma cells as they migrated away from the main tumor mass. We were able to detect several different patterns of tumor spread and found NSC tracking migrating glioma cells in each case. These included thin outgrowths of tumor deep into adjacent normal brain (Figure 2A and 2B), direct extension of the tumor mass into adjacent tissue (Figure 2C), migration of glioma cells along established white matter tracts (Figure 2D), and dissemination of solitary tumor pockets at considerable distance from the primary tumor bed (Figure 2E and 2F). To establish that the observed tumor tracking capacity of NSC was non-random, we inoculated NSC into the corpus striatum of mice, contralateral to existing gliomas. We found that NSC did not randomly disperse into adjacent normal tissue (Figure 2G, left panel) nor could they be seen migrating to any distant nontumorous region of the brain. However, some NSC were visible tracking directly across the brain into the immediate vicinity of the tumor (Figure 2G, center panel) and into the tumor itself (Figure 2G, right panel). We also found that NSC inoculated into nontumor-bearing brains did not randomly dissipate into adjacent tissue or to the contralateral hemisphere (Figure 2H). These results provide strong support for the potential of NSC to act as tumor tracking delivery vehicles for therapeutic gene product delivery. It should be noted, however, that ours and others assessment of the tropic potential of NSC for in vivo disseminated tumor is based on the qualitative assessment of observing NSC co-localized with tumor outgrowths and microsatellites at fixed time intervals. Studies to visualize the migration of NSC in real time as well as strategies to accurately quantitate the proportion of cells exhibiting patho-tropic properties would further enhance the relevance of these findings.

Based on our observations that syngeneic, primary NSC migrated in conjunction with tumor outgrowths and microsatellites, we wished to investigate whether this tropism could be exploited for therapeutic benefit. We and others have previously reported the ability of the immunostimulatory cytokine interleukin 12 (IL-12) to elicit tumoricidal T-cell responses against experimental gliomas [12, 13], and we have specifically established the ability of in situ IL-12 gene transfer in promoting T-cell infiltration into intracranial tumor tissue [12]. We hypothesized that the delivery of IL-12 by NSC that have migrated into the proximity of disseminated pockets of tumor cells, may induce a T-cell response against these tumor islands and thus elicit immune mediated clearance all tumor foci within the brain. We utilized a replication defective adenoviral vector encoding the genes for IL-12 to confer IL-12 protein expression in NSC. These cells when transplanted into established intracranial murine gliomas continued to secrete significant amounts of cytokine. To establish whether the migratory capacity of NSC was of therapeutic relevance in this setting, we intratumorally inoculated glioma bearing control animals with non-migratory fibroblasts that secereted similar levels of IL-12 to NSC in vitro and in vivo. Of significance was our finding that IL-12 expressing NSC (NSC-IL-12) could be found within tumor outgrowths and microsatellites, whereas transplanted fibroblasts were limited to the main tumor mass. Despite the allogeneic nature of the 3T3 fibroblast transplant, overall T-cell infiltration within the main neoplasm was similar to that seen with syngeneic NSC-IL-12 therapy. Additionally, whereas fibroblast mediated IL-12 secretion could only induce T-cell infiltration within the main tumor mass, treatment with NSC-IL-12 resulted in robust T-cell aggregation along the tumor / normal tissue boundary as well as within tumor outgrowths and microsatellites (Figure 3). This was most likely a result of a chemotactic co-localization of T-cells with marginating and migrating NSC-IL-12. To determine whether the observed ability of NSC-IL-12 to induce T-cell infiltration within disseminating tumor was therapeutically relevant, we followed glioma bearing animals treated with intratumoral NSC inoculation for long term survival. We found that animals treated with tumor tropic NSC-IL-12 survived longer than controls (Figure 4). These finding underscore the potential relevance of using NSC to deliver therapeutic agents directly to disseminated tumor satellites.

To add further credence to the use of NSC therapy for high grade glioma, we tested the efficiency of NSC as a vehicle for delivery of a non-immune tumoricidal agent [14]. Using replication defective adenoviral mediated gene transfer, we engineered NSC to secrete the tumor toxic chemokine, tumor necrosis factor alpha related apoptosis inducing ligand (TRAIL). TRAIL has been shown to selectively induce apoptosis in transformed cells with negligible toxicity to normal tissue, and has been utilized as an experimental agent in a variety of cancer models including glioma [15, 16]. We demonstrated that in athymic T-cell incompetent mice bearing intracranial human U343 glioblastoma xenografts, the inoculation of TRAIL secreting NSC (NSC-TRAIL) significantly inhibited the growth of intracerebral gliomas (Figure 4). Of significance was our additional finding that NSC-TRAIL could migrate into disconnected tumor pockets at significant distance from the main tumor mass and induce overwhelming apoptotic activity in these glioma satellites (Figure 5). These findings point to the potential superiority (in an experimental setting) of NSC mediated protein delivery over more conventional methods such as direct gene transfer using viral vectors or direct infusion of recombinant protein into the tumor or post-resection tumor cavity. The ability of NSC to seek out pockets of disseminated tumor and specifically deliver therapeutic agents into the vicinity of these neoplastic foci, without distributing these potentially toxic chemokines throughout the brain, underscores their utility and potential.

In their report, Benedetti and colleagues forwarded evidence suggesting that NSC may independently be capable of inhibiting glioblastoma cell proliferation [9]. In our extensive experiments, we have not found evidence to support this earlier demonstration, as we could only demonstrate a therapeutic benefit when NSC were engineered to deliver tumor toxic chemokines [10, 14]. The exact mechanism underlying Benedetti et al.'s observations remains unclear, although they have reported that NSC may elaborate a secretory agent that could inhibit tumor cell growth [9]. The ability of NSC to counter the tumorigenic potential of glioma cells independent of the delivery of therapeutic proteins remains an interesting, albeit questionable, proposition. If confirmed, this phenomenon will add further credence to the utility of NSC as therapeutic agents for malignant glioma.

It is however important to note that despite these encouraging pre-clinical findings, a significant gap still remains between the experimentally demonstrated efficacy of NSC therapy for glioma and the clinical translation of this treatment strategy. The exact phenotypic characteristics of the NSC sub-populations that exhibit tumor tracking activity and the mechanisms that underly this migratory activity currently remain unclear. It is highly likely, given the reparative role that NSC populations are speculated to have within the adult CNS, that migratory NSC are tracking a chemotactic gradient established by the secretion of inflammatory mediator(s) by disseminating tumor cell pockets or by normal brain parenchyma secondary to injury by infiltrating neoplastic cells. Additionally, given the observation that only a portion of transplanted NSC exhibit patho-tropic characteristics, it is likely that specific phenotypic characteristics (most likely involving expression of receptor(s) for tumor elaborated chemokine(s)) present in this sub-population govern their ability to migrate. The elucidation of the exact mediators and specific signals governing NSC tropism for migrating tumor will allow further refinement in the use of these cells as, prior to transplantation, NSC populations could be purified in vitro in a phenotype specific manner, possibly based on the expression of cell surface receptor(s) for the chemokine(s) involved in tumor tracking. Conceivably, the identification of cell surface proteins associated with NSC migratory activity could allow for the enrichment of this expression by in vitro manipulation, thereby increasing the tropic potential of transplanted NSC.

Another significant issue with respect to the clinical viability of NSC therapy for glioma involves the question as to whether the demonstrated ability of these cells to track to sites of disseminated tumor in rodent models of glioma will be relevant in the setting human tumors, given the significantly larger distances that these cells would have to translocate across in order to be of therapeutic relevance. Additionally, the significant heterogeneity found in primary human gliomas, as opposed to the relatively homogenous populations found in cultured glioma cell lines, might result in only a proportion of disseminating tumor satellites elaborating the chemokine signals necessary for NSC taxis. Even though ours and others results indicate that NSC mediated therapeutic protein delivery is efficacious in the setting of rodent gliomas, factors including elaboration of adequate quantities of therapeutic gene product(s) as well as duration of secretion from engineered NSC could also limit the effectiveness in human patients. Further investigation utilizing primary human GBM xenografts in pre-clinical models, preferably in larger mammalian species, would help to further define the relevance of NSC therapy in the setting of human glioma therapy.

It is evident that the use of NSC as therapeutic vehicles for glioma holds significant potential. However, in addition to the factors described above, the clinical relevance of this approach is also compounded by significant ethical and legal issues that surround the use of fetal/embryonic tissue for cell harvesting. Additionally, the use of allogeneic fetal derived NSC in a clinical setting would be complicated by concerns of tissue rejection, and would therefore necessitate a degree of iatrogenic immunosuppression, which for reasons discussed earlier is undesirable in patients with malignant gliomas. An appropriate cellular therapy for use in clinical practice should, therefore, ideally comprise of autologous cells that can be harvested without difficulty, processed efficiently in vitro, and then re-inoculated into the same patient. With the aim of developing such a system, we have demonstrated that neural progenitors can be derived from adult bone marrow and that these cells have characteristics similar to fetal NSC, and may therefore represent a promising and clinically relevant source of tumor tropic cells for glioma therapy.

The use of fetal derived NSC, although highly effective in experimental rodent models of glioblastoma, is precluded in a clinical setting given significant ethical, legal, and tissue rejection related impediments. The translation of this therapy into a viable modality for patients with high grade malignant glioma will, therefore, require an alternate source of tumor tropic cells that can replicate the migratory and therapeutic delivery characteristics of fetal derived NSC. Additionally, to minimize concerns of tissue rejection and tumorigenecity, an ideal clinical therapy in this setting should consist of autologous cells derived from a readily accessible tissue source as opposed to mass-produced immortalized cell lines. The identification of NSC populations within the adult brain raised the possibility of utilizing adult neural tissue as an autologous cell source. However the yields of NSC from adult brain have been highly variable, and generally necessitate protracted in vitro culture periods. This and the need for invasive intracranial tissue sampling have limited the applicability of the adult CNS as a viable source for NSC. More recently, several reports have described the presence of precursor cells within adult bone marrow that can yield cells with neuronal and glial phenotypes. Initial descriptions involved the identification of bone marrow stromal cell populations that could be induced to assume expression of neural and glial protein markers [17, 18]. Subsequently, Jiang et al. [19] and Zhao and colleagues [20] have described a distinct cell population within adult bone marrow that can be identified following 25 to 30 passages in in vitro culture and is capable of self-renewal and differentiation into neurons, astrocytes, and oligodendroglia. More recently, our laboratory has described a rapid culture process whereby multipotent neural precursors, phenotypically and morphologically distinct from either bone marrow stromal cells or the multipotent precursors described by Jiang et al. and Zhao et al., can be generated from unfractionated rodent adult bone marrow within 1 week of culture initiation [21]. These bone marrow derived neural progenitors were morphologically and phenotypically indistinguishable from fetal derived NSC and could differentiate into neurons, astrocytes, and oligodendroglia. Furthermore, these cells demonstrated tumor tropic behaviour in vivo, and could engraft into existing histological architecture. These findings indicate that adult bone marrow may serve as an effective source for the isolation of therapeutically relevant neural precursors that could serve either as agents for cell replacement or therapeutic protein delivery in a variety of neuropathological conditions, including disseminated intracranial glioma.

Unfractionated bone marrow was harvested from adult Fisher F344 rats and plated in culture in combination with bFGF and EGF, using protocols similar to those utilized for isolation and propagation of fetal NSC. After 4 to 7 days of culture, viable cellular aggregates in the shape of spheres started forming, which upon disassociation could undergo further propagation. These spheres were morphologically indistinguishable from fetal derived neurospheres were also strongly positive for nestin and neurogenin 1, indicating their neural characteristics [21]. We then investigated the differentiation potential of these cells and subjected them to an in vitro culture environment used to differentiate fetal NSC. The adult bone marrow derived cells attached to culture surfaces, and cells at the outer margins of each sphere began to migrate away from the primary site of attachment. Migrating cells developed processes and displayed varied morphology including large flat cells and multipolar and bipolar neuron-like cells (Figure 6A). Upon subsequent phenotypic analysis, we identified populations of cells expressing the glial markers GFAP and EAT1 (Figure 6B and 6C), the oligodendroglial antigen CNPase (Figure 6D), and the neuronal markers neurofilament 200 (Figure 6E), NSE (Figure 6F and 6G) and MAP2 (Figure 6H). As would be expected, given the completely different nature of the parent tissue, the efficiency of differentiation with bone marrow derived cells was significantly lower than that usually experienced with fetal NSC. Despite the lower yields of differentiated neural and glial progeny, our results conclusively confirmed the pluripotency of our adult derived neural progenitors.

As has previously been demonstrated for fetal derived NSC, we then wished to test the potential of bone-marrow derived neural progenitor for neural differentiation in vivo. We inoculated B-galactosidase expressing cells into the hippocampi of syngeneic animals. The hippocampus is a neurogenically active region of the adult mammalian brain rich in trophic factors, and presents a favorable environment for differentiation [22]. Some of the transplanted bone marrow cells visibly integrated into the structure of the hippocampus and were positive for the neuronal marker NeuN (Figure 6I). We have since demonstrated that bone marrow derived neural progenitors, when implanted into established intracranial gliomas, can migrate into disseminating pockets of tumor cells (Kabos et al, unpublished results). These results demonstrate that adult bone marrow can be utilized as source for cellular spheres that are similar in morphology, phenotype, and behavior to fetal NSC. These cells with their clinical accessibility, potential for in vitro expansion and demonstrable migratory activity and tumor tropism represent an important step forward that may prove critical in the clinical implementation of NSC therapy for malignant glioma.

Despite these promising findings, the efficiency of bone marrow derived neural precursor generation currently remains low when compared to NSC yields from fetal or embryonic tissue. It is likely that the frequency within bone marrow of mutlipotent precursors capable of yielding neural progenitors is extremely low, and little is known about the trophic signals and transcriptional cues that govern the development of neural precursors from this primitive progenitor population. Given the importance of developing efficient alternative sources for neural precursor isolation, the refinement and optimization of neural precursor harvest from adult bone marrow, among other potential tissue sources, will be an important pre-requisite to the clinical use of NSC based therapies. Currently employed harvest strategies rely on the use of defined growth supplements and culture conditions with little understanding of the mechanisms underlying the trans-differentiation of marrow precursors into neural cells. The elucidation of the mechanisms governing expansion of bone marrow derived mutlipotent precursor cells and the neurogenic trans-differentiation of this tissue will allow for more focused and efficient culture processes capable of yielding the large number of neural precursors that will be necessary for clinical applications. Additionally, with the aim of bringing the clinical translation of this methodology to fruition, it will be necessary to establish and refine a similar culture process for human tissue.

Preliminary evidence has established that the use of NSC as tumor tropic delivery vehicles represents a promising new treatment modality for malignant glioma. It is important, however, to note that significant impediments remain before the true potential of this novel therapy can be fully realized in a clinical setting. The use of fetal derived NSC, although potentially of immense therapeutic relevance given their robust tropism for intracranial tumor, remains highly controversial given the ethical and legal reluctance to use embryonic tissue. Additionally, the necessity of immunosuppressing patients transplanted with allogeneic cells, whether those comprise of fetal NSC or an "off the shelf" immortalized stem cell pool as described by Aboody and colleagues [7] is undesirable given the known role of endogenous tumor induced immune defects in supporting the growth of malignant gliomas. The potential tumorigenicity of transplanted NSC, particularly oncogene mediated immortalized cells, also remains of concern. Even though no evidence has arisen in the extensive studies conducted by our group [10, 14] and Benedetti et al [9] indicating that primary NSC may be tumorigenic, further investigation is necessary to ensure the safety of this therapy. These limitations limit the clinical impact that currently validated experimental approaches as described by our group and others [7, 9, 10, 14] may have in the immediate future. The ability to translate the promising potential visualized with the use of fetal NSC into an alternate, clinically viable source of cells with similar tumor tropic characteristics will, therefore, prove critical to the eventual clinical translation of this therapeutic approach.

In light of these concerns, we have developed a novel technology utilizing unfractionated adult bone marrow as a source of neural progenitors that are morphologically and phenotypically similar to fetal derived NSC, and exhibit similar in vivo migratory and tumor tropic activity. The implications of this novel process are tremendous, as it would allow the generation of a pool of autologous cells, harvested from tissue that is routinely accessible in a clinical setting, and would therefore obviate the ethical and tissue rejection related concerns that preclude the use of fetal NSC in patients. However, before this strategy can be realistically implemented, it will be critical to refine the process further. Initially, the reproduction of this phenomenon in human tissues will prove critical to the future viability of this technology. However, given that other groups have already demonstrated that isolated phenotype specific populations of cells from human bone marrow can generate neural cells in vitro and in vivo [19, 20] it is a realistic expectation that our described rapid culture process using unfractionated whole bone marrow will be reproducible in human tissues.

The next five years should see rapid progress towards the development of alternate sources of stem cells that can serve as clinically viable avenues for transplantation. It is likely that tissues such as bone marrow and, potentially, whole blood may eventually serve as routine sources for cell harvest, with the establishment of dedicated core laboratories geared towards the rapid in vitro expansion of autologous neural progenitors. Once such methodology is developed, it is possible that we will see the initiation of clinical trials utilizing NSC therapy in patients with malignant glioma. Additionally, further investigation of the biological signals that govern stem cell migration towards sites of disseminating tumor will provide greater insight into how NSC track infiltrating tumor cells and may allow further refinement of NSC therapy, as specific populations could be purified or engineered on the basis of enhanced tropic characteristics, thereby improving the efficiency of NSC therapy for malignant glioma.

In conclusion, although the NSC technologies are currently in the initial stages of development and pre-clinical testing, there is tremendous promise in their potential to be of significant therapeutic benefit, particularly in patients with high grade gliomas. The field of neurooncology has been unable to make any satisfactory progress in combating this lethal disease process, largely as a result of the inability of currently employed clinical and experimental therapies in targeting and eliminating residual intracranial neoplastic foci. The use of NSC represents the first therapeutic attempt that is specifically geared towards addressing this problem. The encouraging results that have been observed to date underscore the potential of this therapy, and despite limitations that may make immediate clinical implementation impractical, further work in this field should be pursued aggressively.

1. Liau LM, Black KL, Prins RM, et al. Treatment of intracranial gliomas with bone marrow-derived dendritic cells pulsed with tumor antigens. J Neurosurg. 90(6), 1115-24 (1999).
2. Ehtesham M, Kabos P, Gutierrez MA, et al. Intratumoral dendritic cell vaccination elicits potent tumoricidal immunity against malignant glioma in rats. J Immunother. 26(2), 107-16 (2003).
3. Yu JS, Wheeler CJ, Zeltzer PM, et al. Vaccination of malignant glioma patients with peptide-pulsed dendritic cells elicits systemic cytotoxicity and intracranial T-cell infiltration. Cancer Res. 61(3), 842-7 (2001).
4. Roszman TL, Brooks WH. Immunobiology of primary intracranial tumours. III. Demonstration of a qualitative lymphocyte abnormality in patients with primary brain tumours. Clin Exp Immunol. 39(2), 395-402 (1980).
5. Zou JP, Morford LA, Chougnet C, et al. Human glioma-induced immunosuppression involves soluble factor(s) that alters monocyte cytokine profile and surface markers. J Immunol. 162(8), 4882-92 (1999).
6. Alvarez-Buylla A, Herrera DG, Wichterle H. The subventricular zone: source of neuronal precursors for brain repair. Prog Brain Res. 127(1-11 (2000).
7. Aboody KS, Brown A, Rainov NG, et al. Neural stem cells display extensive tropism for pathology in adult brain: evidence from intracranial gliomas. Proc Natl Acad Sci U S A. 97(23), 12846-51 (2000).
8. Snyder EY, Deitcher DL, Walsh C, et al. Multipotent neural cell lines can engraft and participate in development of mouse cerebellum. Cell. 68(1), 33-51 (1992).
9. Benedetti S, Pirola B, Pollo B, et al. Gene therapy of experimental brain tumors using neural progenitor cells. Nat Med. 6(4), 447-50 (2000).
10. Ehtesham M, Kabos P, Kabosova A, et al. The use of interleukin 12-secreting neural stem cells for the treatment of intracranial glioma. Cancer Res. 62(20), 5657-63 (2002).
11. Martens DJ, Tropepe V, van Der Kooy D. Separate proliferation kinetics of fibroblast growth factor-responsive and epidermal growth factor-responsive neural stem cells within the embryonic forebrain germinal zone. J Neurosci. 20(3), 1085-95 (2000).
12. Liu Y, Ehtesham M, Samoto K, et al. In situ adenoviral interleukin 12 gene transfer confers potent and long-lasting cytotoxic immunity in glioma. Cancer Gene Ther. 9(1), 9-15 (2002).
13. Jean WC, Spellman SR, Wallenfriedman MA, Hall WA, Low WC. Interleukin-12-based immunotherapy against rat 9L glioma. Neurosurgery. 42(4), 850-6; discussion 856-7 (1998).
14. Ehtesham M, Kabos P, Gutierrez MA, et al. Induction of glioblastoma apoptosis using neural stem cell-mediated delivery of tumor necrosis factor-related apoptosis-inducing ligand. Cancer Res. 62(24), 7170-4 (2002).
15. Wu M, Das A, Tan Y, et al. Induction of apoptosis in glioma cell lines by TRAIL/Apo-2l. J Neurosci Res. 61(4), 464-70 (2000).
16. Hao C, Beguinot F, Condorelli G, et al. Induction and intracellular regulation of tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) mediated apotosis in human malignant glioma cells. Cancer Res. 61(3), 1162-70 (2001).
17. Woodbury D, Schwarz EJ, Prockop DJ, Black IB. Adult rat and human bone marrow stromal cells differentiate into neurons. J Neurosci Res. 61(4), 364-70 (2000).
18. Sanchez-Ramos J, Song S, Cardozo-Pelaez F, et al. Adult bone marrow stromal cells differentiate into neural cells in vitro. Exp Neurol. 164(2), 247-56 (2000).
19. Jiang Y, Jahagirdar BN, Reinhardt RL, et al. Pluripotency of mesenchymal stem cells derived from adult marrow. Nature. 418(6893), 41-9 (2002).
20. Zhao LR, Duan WM, Reyes M, et al. Human bone marrow stem cells exhibit neural phenotypes and ameliorate neurological deficits after grafting into the ischemic brain of rats. Exp Neurol. 174(1), 11-20 (2002).
21. Kabos P, Ehtesham M, Kabosova A, Black KL, Yu JS. Generation of neural progenitor cells from whole adult bone marrow. Exp Neurol. 178(2), 288-93 (2002).
22. Fricker RA, Carpenter MK, Winkler C, et al. Site-specific migration and neuronal differentiation of human neural progenitor cells after transplantation in the adult rat brain. J Neurosci. 19(14), 5990-6005 (1999).

* You will need the Adobe® Acrobat® Reader to view and print some items in this article. If you do not have this software, you can download it FREE from Adobe's website.