By Dwain Morrris-Irvin, PhD and John S. Yu, MD
Malignant brain tumors, particularly gliomas, can be devastating to patients and shorten their lives to just a few months after diagnosis [1]. These tumors have a remarkable ability of being highly invasive and resistant to current methods of treatment like surgical resection, irradiation and chemotherapy. In the recent past, several reports have demonstrated the ability of transplanted neural stem/progenitor cells to migrate towards a tumor mass in animal models of brain tumors, primarily gliomas [2-10]. More recently, there has been evidence demonstrating that endogenous precursors from the subventricular zone of the forebrain will respond and migrate towards experimentally induced tumors in the forebrain [3]. Here, we discuss the implications of these findings and the potential future of this novel approach to treating malignant brain tumors.
Neural stem/progenitor cells (NSC), by definition, have the ability to self-renew and generate neurons, astrocytes and oligodendrocytes during development in-vivo and in the adult [11-12]. The proliferation, differentiation, migration and survival of NSC is regulated, at least in part, through epidermal growth factor receptor (EGF-R) and fibroblast growth factor receptor (FGF-R) signaling [13-14]. In the adult mammalian brain, there are only two sites of neurogenesis from endogenous progenitors and they are the forebrain subventricular zone and subgranular layer of the dentate gyrus [12]. While neurogenesis in the adult is limited to these sites, the migration potential of these progenitors is still being elucidated.
Recently, a wave of reports have demonstrated that endogenous neural progenitors from forebrain germinal zones have the ability to migrate towards sites of various brain injuries, including animals models of stroke, ischemia and chemical induced lesions. A recent report has also demonstrated that endogenous progenitors from the subventricular zone respond to experimentally-induced tumors by infiltrating the primary tumor mass [3]. The specific factors that regulate the migratory potential of endogenous progenitors in response to injury or disease are relatively unknown, however, some factors, like transforming growth factor-alpha,
TGF-α or platelet-derived growth factor, PDGF, have been implicated.
The migratory potential of NSC subsequent to transplantation in the damaged brain has been described by several reports, reviewed by Emsley et al [15]. One of the first reports to demonstrate NSC ability to migrate towards tumors was performed in a rodent model of intracranial glioma, in which the authors showed that both murine and human neural cell lines, transplanted into the brain parenchyma, ventricles or systemic circulation, had the potential to migrate towards, infiltrate and surround the primary tumor inoculation site [2]. The report also noted that NSC were able to track infiltrative tumor cells that had migrated away from the primary tumor mass. Additionally, the authors showed that NSC injected into the systemic circulation were able to "home in" and infiltrate the primary tumor mass but were not identified within the surrounding normal brain parenchyma. In addition to NSC tracking abilities, the authors showed that NSC implanted directly into the primary tumor inoculation site could be found infiltrating throughout the tumor mass but not beyond the borders of the tumor, suggesting that the tumor itself is more permissive for NSC migration compared to the healthy neighboring tissue. These data suggests that there are signaling molecules generated within the tumor that are tropic for NSC and promote their migration.
The authors also took advantage of the efficient tumor targeting capabilities of NSC and in another report showed that transplanted NSC genetically engineered to express cytosine deaminase, an enzyme that converts 3,5-fluorocytosine (5-FC) to the oncolytic drug, 5-fluorouracil, significantly reduced the experimental-induced tumor in rodents [16]. The above findings suggested great potential and utility for treating highly invasive intracranial tumors utilizing NSC as vehicles for gene delivery. Benedetti et al, also demonstrated that primary NSC genetically modified to express IL-4, a cytokine used experimentally to treat glioblastomas, improved survival of experimentally induced glioblastoma in mice
[4].
Additionally, IL-12, has been shown to have long lasting efficacy against glioma [17]. Ehtesham et al, demonstrated that NSC genetically modified to over-express IL-12 and subsequently transplanted into the contralateral hemisphere, in a rodent model of glioma, were able to migrate and infiltrate the primary tumor mass, promote infiltration of CD4+ and CD8+ lymphocytes within the tumor mass and improve survival in these mice [6]. Yang et al, demonstrated the utility of human embryonic NSC derived from the hippocampus by genetically engineering these cells to over-express IL-12 and subsequently implanting these genetically modified NSC into C6 glioma induced rats [8]. The authors were also able to show that these genetically modified NSC were able to promote the infiltration of lymphocytes, predominantly CD8+ lymphocytes, diminish tumor size and improve survival times.
Tumor necrosis factor-inducing ligand (TRAIL), a member of the TNF ligand family, has also been shown to induce apoptosis in various tumor cell lines while limiting its toxicity to normal tissue. Ehtesham et al, showed that genetically modified human NSC that express TRAIL and subsequently implanted intratumorally into a rodent model of glioma, significantly reduced the tumor mass [5]. This study was supported by a subsequent study that further examined the anti-tumor effects of genetically modified human NSC that express TRAIL [9]. These genetically modified NSC were transplanted into the brains of rodents with experimentally induced glioma and the authors were able to utilize eukaryotic luciferases, the firefly (Photinus) and sea pansy (Renilla reniformis) luciferases, Fluc and Rluc, to mark the cell types respectively, and examine the migratory potential of NSC, as well as, potential changes in tumor size.
The authors were able to demonstrate enhanced migration of NSC and a reduction in tumor size over time in the living animal. This dual enzyme substrate bioluminescence imaging allows for labeling of both the NSC and glioma cell line for sequential and differential imaging in the living organism [18]. In addition, the authors utilized a more soluble form of S-TRAIL, that was shown to be more efficient at inducing apoptosis in tumor cells [19]. This advanced imaging analysis provides a new standard for demonstrating the efficacy of utilizing genetically engineered NSC for the purpose of treating intracranial tumors. In fact, the use of this imaging technology to study many disease paradigms will prove to be essential in providing a better understanding of the dynamic and plastic responses of brain tissue during disease and repair.
While the above studies have investigated the potential of NSC to act as vehicles for anti-tumor gene delivery, they did not identify the potential signals responsible for the homing ability of NSC for tumors. One report has examined the potential role of stromal derived factor-1 (SDF-1) and Chemokine receptor-4 (CXCR4) signaling in the homing ability of NSC for tumors [20]. CXCR4, a chemokine receptor that governs cellular migration and homing in hematopoetic, neural and tumor cell lines, has been shown to be expressed by neural progenitors during development [21-26]. Stromal cell-derived factor-1 (SDF-1) a ligand for the CXCR-4 receptor has been shown to be secreted by highly invasive tumors and is correlated to tumor survival and invasiveness [27-29] Ehtesam et al, showed that human fetal NSC transplanted intratumorally in experimentally induced GL26 glioma mice, disseminated into the neighboring parenchyma, following "islets" of tumor cells that had migrated away from the primary tumor inoculation site.
In addition, they also showed that the migrating NSC cells were strongly immunopositive for CXCR-4, while the non-migrating NSC and primary tumor mass were weakly immunopositive [20]. The authors also suggested that the tumor-tropic NSC were immature astrocytes based on immunohistochemical analysis and positive expression of A2B5 and GFAP, while the non-migrating NSC that remained at the primary tumor inoculation site were more differentiated astrocytes, as they were immunopositive for excitatory amino acid transporter 1 and 2, EAAT1 and EAAT2, proteins suggested to be expressed by mature astrocytes [30]. These data suggest that SDF-1/CXCR-4 signaling, at least in part, plays some functional role in NSC's tumor tropism.
Another recent report by Schmidt et al has implicated vascular endothelial growth factor (VEGF) in promoting NSC tropism towards tumors [31]. The authors were able to demonstrate through in-vitro analysis of human brain tumor extracts that VEGF protein was up-regulated in the tumor samples compared to normal human brain controls. Additionally, the authors infused VEGF protein into the striatum of nude mice. Two days later, human and mouse NSC were transplanted into the contralateral striatum. Within one week NSC showed long-range attraction and migration towards the site of VEGF infusion in the contralateral hemisphere. These data suggests that VEGF plays a role in NSC tumor tropism. Taken together, secreted molecules shown to be up-regulated in brain tumors are beginning to be examined for their potential role in NSC migration and tumor tropism. Continued investigation in this area will be the next critical step to better understand the molecular mechanisms underlying NSC attraction towards sites of tumor genesis and angiogenesis.
To date, the transplantation of NSC into the parenchyma, ventricular system or systemic circulation of experimentally induced rodent glioma models, has clearly demonstrated that NSC have a tropism for the primary tumor mass and invasive tumor cells that have migrated away from the primary tumor inoculation site. In addition, intra-tumor delivery of NSC, in glioma rodent models, has demonstrated the migrational capabilities of NSC within the tumor mass and to track tumor satellites that have migrated away from the primary tumor mass. A more recent report has also demonstrated in a experimentally induced glioblastoma murine model that endogenous precursors generated from the subventricular zone may also migrate towards the primary tumor mass [3]. The authors used a transgenic mouse expressing enhanced green fluorescent protein under the control of the nestin promoter to identify endogenous neural precursors [32-33] and showed that two weeks after glioblastoma inoculation into the caudate-putamen, nestin-GFP positive cells could be identified surrounding and throughout the tumor inoculated hemisphere compared to the contralateral side. The authors also showed that nestin-GFP positive cells were associated with tumor cells that had migrated away from the primary tumor site and that endogenous neural precursor tropism for glioblastoma decreased with age.
Taken together, the authors suggest that the tropism for tumor cells by endogenous neural precursors may represent an intrinsic tissue response that is very specific for tumor cells and not the normal tissue. In addition, the fact that many of these cells migrated long distances from the subventricular zone to reach the primary tumor inoculation site in the caudate-putamen supports earlier studies demonstrating similar long distance effects with exogenous transplanted NSC and may suggest that some aspects of immune signaling may be involved. The idea of immune signaling playing a role in the tropism of endogenous progenitors towards a tumor mass was suggested by Glass et al and they made reference to a previous study demonstrating an age-related decrease in survival, in an experimental glioblastoma mouse model, correlated to a significant decrease in CD8+ T-cell infiltration [34].
In order to translate these pre-clinical findings into clinical therapies we must begin to examine the potential obstacles related to the source of NSC for the purpose of transplantation. Using NSC lines as a source of cells for transplantation has many advantages. For example, these cells may be uniform, well characterized and readily available, however, there is some evidence that NSC lines may have tumorigenic potential after transplantation. Additionally, if NSC lines, embryonic or fetal derived NSC are to be utilized for transplantation to treat brain tumors, patient would have to be immunosuppressed in order to minimize tissue rejection. Ideally, autologous cells would be the preferred choice and presumably avoid the need for immunosupression. In the recent past, many reports have demonstrated that bone marrow cells have the potential to generate NSC. The cell fate potential of BM-NSC is still under investigation, as well as, their ability to engraft into the brain and generate functional neurons. If it is determined that these cells truly have the potential to generate functional neurons, astrocytes and oligodendrocytes after transplantation in-vivo, then they would be an ideal autologous cell source for neural replacement therapy, as well as, delivery vehicles for treating brain tumors.
Other potential cell sources from bone marrow are Sca1+ cells [35]. Additionally, a recent report has demonstrated that bone-marrow stromal cells transfected to over-express EGF-R and subsequently transplanted intracranially in an experimentally-induced GL261 glioma mouse model showed enhanced migratory response towards tumor compared to non-transfected controls 36. [36] Sato et al also demonstrated that bone marrow stromal cells adenovirally transduced to secrete interferon-a upon intra-tumor transplantation resulted in significantly prolonged survival compared to controls. These reports further demonstrate and support the idea of utilizing bone marrow derived cells as sources for tumor targeting and therapy. The potential benefit of isolating bone marrow cells from a patient is that these cells can be readily isolated from a brain tumor patient, manipulated ex-vivo and then transplanted back into the patient for brain tumor therapy. Presumably minimizing the need for immunosupression or tissue rejection from an autologous transplant.
Using NSC as vehicles for gene delivery for the treatment of gliomas has great therapeutic potential. Mounting pre-clinical evidence of this potential demonstrates that NSC upon transplantation into a tumorigenic brain have both a direct apoptotic effect on tumor cells, as well as, an ability to "track down" infiltrating tumor cells. This kind of therapy could prove very useful when considering the limitation with current therapies, including surgical resection, irradiation and chemotherapy. However, many questions still need to be answered before we can examine the true potential of the cells to treat human patients with brain tumors. For example, the molecules primarily responsible for NSC's tumor tropism have not been fully elucidated. Many more molecules will probably be implicated and their specific role in NSC attraction towards tumor cells determined. While there is some demonstration that CXCR-4/SDF-1 or VEGF signaling plays a role in NSC tropism for tumors, these molecules also effect migration and communication between many other cell types. In addition, the cell fates of NSC transplanted into the brains of experimentally induced tumors has not been fully elucidated.
The reports we have discussed here, have provided some evidence that the NSC remain relatively undifferentiated at the primary tumor inoculation site or become more differentiated glia upon tracking infiltrating tumor cells that migrate away from the primary tumor mass and a relatively small percentage of these NSC may represent immature neuronal precursors [3,20]. Determining the fate and functional integration process of transplanted NSC in response to experimentally induced intracranial tumors is critical, if we are to better understand and prepare for the potential side effects of using NSC for brain tumor therapy.
The endogenous precursor response is very interesting and brings more questions than answers when trying to understand the pathology of brain tumors. Clearly, endogenous progenitors do not provide enough protection for patients with malignant gliomas, but what, if any, protection do they provide? Do they respond differently to different tumors, and if so, why? Can they also be manipulated to act as vehicles for gene delivery, and if so, is their tropism strong enough to utilize them as a potential cell source for brain tumor therapy? We look forward to future reports regarding what stimulates their tropism towards tumors.
(Figure 1) Migration of NSC towards experimentally induced tumor and tumor satellite (A-C). A. Illustration of intratumor injection of exogenous NSC infiltrating experimentally induced striatal tumor and NSC tracking tumor satellite. B. Illustration showing exogenous NSC transplanted into the contralateral striatum of experimentally induced tumor. C. Illustration showing endogenous progenitor migration from subventricular zone towards and infiltrating experimentally induced tumor.
John S. Yu, MD is Co-Director of the Comprehensive Brain Tumor Program at the Maxine Dunitz Neurosurgical Institute at Cedars-Sinai Medical Center in Los Angeles.
Dwain Morris-Irvin, PhD, is a researcher at the Maxine Dunitz Neurosurgical Institute at Cedars-Sinai Medical Center in Los Angeles.
