Current therapies for gliomas often fail to address their infiltrative nature. Conventional treatments leave behind small clusters of neoplastic cells, resulting in eventual tumor recurrence. In the present study, we have evaluated the antitumor activity of neural progenitor cells against gliomas when stereotactically injected into nucleus Caudatus of Fisher rats. We show that the rat neural progenitor cell lines HiB5 and ST14A, from embryonic hippocampus and striatum primordium, respectively, are able to prolong animal survival and, in 25% of the cases, completely inhibit the outgrowth of N29 glioma compared with control animals. Delayed tumor outgrowth was also seen when HiB5 cells were inoculated at the site of tumor growth 1 week after tumor inoculation or when a mixture of tumor cells and HiB5 cells were injected s.c. into Fisher rats. HiB5 cells were additionally coinoculated together with two alternative rat gliomas, N32 and N25. N32 was growth inhibited, but rats inoculated with N25 cells did not show a prolonged survival. To evaluate the possibility of the involvement of the immune system in the tumor outgrowth inhibition, we show that HiB5 cells do not evoke an immune response when injected into Fisher rats. Furthermore, the rat neural progenitor cells produce all transforming growth factor β isotypes, which could explain the observed immunosuppressive nature of these cells. Hence, some neural progenitor cells have the ability to inhibit tumor outgrowth when implanted into rats. These results indicate the usefulness of neural stem cells as therapeutically effective cells for the treatment of intracranial tumors.

Progenitor cells are regarded as undifferentiated cells capable of proliferation, self-renewal, production of large numbers of differentiated progeny, and with an ability to regenerate tissues (1, 2). Studies by a number of investigators have now confirmed that mammalian adult neural progenitors exist and are capable of extensive cell division and self-renewal (3, 4, 5, 6). Neural progenitors can migrate and home in on specific sites of damage or regeneration, e.g., to the olfactory bulb of rodents (7), the hippocampus of humans (8), and to sites of central nervous system (CNS) tumors such as gliomas (9). Recent studies emphasize their feasibility in acting as vehicles for delivering transgenes (10, 11, 12).

Astrocytomas originating from cells of the CNS are one of the most difficult cancers to treat because of their location and their resistance to conventional therapy, and they continue to present a challenge in the attempts to find the most suitable therapeutic strategy.

Foremost, among these problems is the tendency of these tumors to present as disseminated within the CNS at the time of diagnosis. Although this type of cancer seldom, if ever, develops a metastatic phenotype, they spread long distances within the CNS, rendering complete surgical removal possible only in a small number of cases. Moreover, most of these cancers are resistant to treatment with chemotherapy and irradiation. The dispersed nature of gliomas indicate that an effective treatment will require the development of new methods to eliminate cells that have migrated long distances from their site of origin. There are a number of theoretical means to achieve this goal. One possibility has been to devise a therapeutic delivery system that could itself migrate along the same route used by the glioma cells and thus distribute its activity throughout the CNS in a manner similar to the distribution of glioma cells themselves. That this type of approach would yield a beneficial outcome is in part expected because the various treatment modalities are already known to enhance survival (10, 11, 12, 13, 14, 15, 16, 17, 18, 19). A surprising finding, however, is that the neural progenitor cells themselves increase the efficacy of the treatment and, even more intriguing, by themselves increase the survival of the recipient animals.

In this study, we show that the rat embryonic neural progenitor cell lines ST14A and HiB5 can inhibit glioma outgrowth in vivo. As glioma models, we use ethyl-nitroso-urea-induced rat gliomas. These tumors grow well in vitro and can be injected stereotactically for growth intracerebrally in a highly reproducible manner. We additionally show that not all neural progenitor cell lines have this capability and that the outgrowth inhibition is not effective on all glioma cell lines. Rats coinoculated with HiB5 or ST14A cells together with the glioma N29 into the CNS exhibited a longer life span or even completely inhibited the tumor outgrowth. HiB5 cells were also coinoculated with the glioma cell lines N25 and N32 and prolonged the life span of the N32 glioma inoculated animals, but not the life span of N25 glioma inoculated rats. Tumor outgrowth inhibition was additionally observed when HiB5 was inoculated 1 week after the N29 tumor cells as compared with the controls injected with tumor cells only. Coinoculation of N29 glioma cells with different adult neural progenitor cell lines or normal Fisher rat fibroblasts had no effect on animal survival. These studies indicate that tumors coinoculated with progenitor cell lines HiB5 or ST14A are demarcated, with fewer colonies of expanding tumor isles distant from the main tumor mass, indicating a less aggressive tumor with reduced migratory potential. These findings support a new approach for brain tumor therapy based on the grafting of neural progenitor cells with an ability of not only eliminating the tumor cells but possibly of repairing damage in the CNS as well.

Cell Lines.

The temperature sensitive SV40 large T conditionally immortalized cell lines HiB5, RN33B, and ST14A were kindly provided by Cecilia Lundberg (Wallenberg Neurocenter at Lund University). HiB5 was originally prepared from Sprague Dawley embryonic (E16) hippocampus (20), ST14A was generated from E14 rat striatum (21) and RN33B originate from the raphe region of E12.5 rats (22). The cells were cultured as previously described in a humidified chamber at 33°C (20, 21, 22).

The neural progenitor cells ANSC-2 and ANSC-3 were isolated from the subventricular zone of adult Fisher rat brains as described previously (23). The prepared pool of cells were divided into two cell fractions and transferred to PRIMARIA tissue culture flasks and cultured in DMEM:Ham’s F-12 supplemented with N2 (Life Technologies, Inc., Paisley, United Kingdom) and 10 ng/ml fibroblast growth factor 2 (R&D Systems, Abingdon, United Kingdom). One cell fraction was immortalized with C-Myc-plxsn (kindly provided by Bengt Widegren, Department of Cell and Organism Biology, Lund, Sweden) by retroviral transfection (ANSC-3-C-Myc), whereas the other fraction denoted ANSC-2 was grown without immortalization. The N25/N29/N32 Fisher rat glioma cell lines were prepared and cultured in vitro as described previously (24, 25). H1D2 is a subclone from the BN7005 rat adenocarcinoma of the colon, originally induced by 2-dimethylhydrazine in a Brown Norway rat (26). H1D2 was cultured in vitro as described previously (26). Fisher rat fibroblasts were prepared from the peritoneal wall and cultured in RPMI 1640 supplemented with 20% FCS.

Immunohistology.

For histological analysis, brains were surgically removed and frozen in isopentane. Cryosections (5–7 μm) were fixed in acetone and washed in PBS or Saponin (0.1%), depending on the targeted antigen. Sections were blocked with 5% donkey serum and incubated with primary antibody from BD PharMingen, mouse antirat nestin monoclonal antibody (0.5 μg/ml), and biotin-labeled anti-rat monocyte antibody (5 μg/ml, 1C7) 1 h at room temperature. After washing in PBS, the appropriate samples were incubated with 2 μg/ml secondary biotinylated donkey antimouse IgG antibody (Fab2; Jackson Immunoresearch, Baltimore, PA) and 1 μg/ml peroxidase-labeled streptavidin (Jackson Immunoresearch) for 30 min each. The chromogen AEC+ (DAKO, Glostrup, Denmark) was added for ∼3 min before inactivation in H2O. Hematoxylin or eosin (Harry’s Histolab, Gothenburg, Sweden) staining was performed before mounting the slides with DAKO Faramount (DAKO) mounting media. Sections were visualized with a standard light microscope (BX-60; Olympus) and captured with a three-color charge-coupled device RGB Practica ColorScan. Pictures were analyzed in Image Pro. Carboxyfluorescein diacetate labeling of NSC cells was performed according to the protocol from Molecular Probes (Leiden, the Netherlands).

Animal Experiments.

All animals were maintained by continuous single line, brother sister mating. Intracerebral injections were done in the Fisher 344 rat strain. The animals were anesthetized with Isoflourane. After drilling a hole in the skull, intracerebral inoculations were performed into the right nucleus Caudatus of male Fisher rats, using a small stereotactic frame (Braintree Scientific, Inc., Braintree, MA). Coordinates for injection were 2.5 mm laterally and 1.7 mm rostrally measured from the bregma and 5 mm ventrally measured from the outer surface of the skull. Injection of the cells and withdrawal of the canula was made slowly to diminish backflow through the insertion canal. The hole in the skull was closed with bone wax. A maximum of 5 μl fluid was inoculated containing glioma cells (5 × 103 cells) alone or together with progenitor cells (25 × 103, 1 × 105, or 25 × 104 cells) or fibroblasts (1 × 105). After the first signs of motoric dysfunction, animals were sacrificed, and brains were taken for frozen tissue examination. Subcutaneous inoculations in the flank with the same types of cells were performed in Fisher rats, and SCID FOX CHASE mice and follow-up palpation’s were made at 1-week intervals. Tumors were taken for immunohistochemistry when reaching a size of ∼1000 mm3 in rat and in mouse. The tumor volume was calculated on the basis of measurements of the longest diameter (a) and the diameter perpendicular to this (b) according to this formula [volume = 0.4 × (a) × (b)2] and expressed as mm3. All cells were inoculated in RPMI medium supplemented with 1% Fisher rat serum. For retroperitoneal injections, F1 rats from Brown Norway crossed with Wistar/Furth were used. Retroperitoneal inoculations were made with the colon carcinoma cell line H1D2 (1 × 105) together with HiB5 (1 × 106) cells. All animals were treated according to the Swedish guidelines for humane treatment of laboratory animals, and the experiments were approved by the Lund/Malmö animal ethical committee.

Evaluation of HiB5 Immunogenicity.

Two Fisher rats were immunized s.c. at day 0 and day 14 with 106 viable HiB5 cells. After 2 weeks, lymphocytes were harvested from the spleen of immunized and nonimmunized animals and cocultured with irradiated HiB5 cells (11,000 rad, 37Cs source) for 5 days or stimulated with 0.01 ng/ml Staphylococcal Enterotoxin A (Toxtech, Henderson, NV) for 3 days. Irradiated HiB5 cells were also cocultured with spleen cells from a nonimmunized animal for 3 days together with Staphylococcal Enterotoxin A (0.01 ng/ml). The cells were pulsed with 0.5 μCi of tritiated [3H]thymidine (Amersham, Buckinghamshire, United Kingdom) and incubated for another 6 h. Incorporated radioactivity was determined in a Wallac 1450 Microbeta Liquid Scintillation Counter (Turku, Finland). HiB5 cells and YAC-1 cells were each plated together with spleen cells as effector cells and used in a 17 h 51Cr release cytotoxic assay as described previously (27).

In Vitro Tumor Proliferation Assays.

N29 glioma cells or H1D2 colon cancer cells were cocultured with irradiated HiB5 cells (11,000 rad; 37Cs source). The cells were plated together in a 96-well plate and incubated at 37°C for 3 days before adding 0.5 μCi of tritiated [3H]thymidine (Amersham) and incubated for another 6 h. The plates were thereafter harvested in a Tomtec cell harvester and the filters analyzed as described above.

PCR Analysis.

Before RNA extraction from HiB5 cells, ST14A cells, ANSC-2 cells, and lymphocytes, the cells were washed in PBS, and total RNA was extracted with Trizol (Invitrogen-Life Technologies, Inc., Carlsbad, CA) according to the manufacturers protocol. The quality and quantity of the total RNA was determined by gel electrophoresis. cDNA was synthesized from 2 μg of total RNA using 400 units of Superscript (Invitrogen-Life Technologies, Inc.) and 100 ng of random hexanucleotides. The following temperature profile was used: 10 min at 25°C and 50 min at 42°C. Parallel controls in the absence of room temperature were also performed. After cDNA synthesis, an amount of cDNA mixture equivalent to 20 ng of total RNA was added to a PCR reaction mixture containing 2.5 mm MgCl2 and hot-started with AmpliTaq (Perkin-Elmer Cetus). The following cycle profile was used: 30 s at 94°C; 30 s at the annealing temperature; and 30 s at 72°C. The primer sequence, optimal annealing temperature, and length of the PCR product for each mRNA is shown in Table 1. The PCR products were analyzed on a 1.5% agarose gel stained with ethidium bromide. The primer pairs were checked for amplification from genomic DNA, and control PCR was performed to confirm the absence of DNA contamination in the RNA preparations.

ELISA.

Conditioned medium from 1 million HiB5 cells saturated over a period of 4 days was analyzed for transforming growth factor β1 (TGF-β1), TGF-β2, bone morphogenic protein 4 (BMP4), and tumor necrosis factor α (TNF-α) expression using an ELISA kit (R&D Systems) according to manufacturers protocol.

Neural Progenitor Cells Inhibit Intracranial Outgrowth of Gliomas.

We injected HiB5 cells into nucleus Caudatus of Fisher rats together with N29 glioma cells. A summary of several experiments show that the control animals injected with N29 cells alone had a rapid tumor progression, where the first rats show symptoms due to tumor progression at day 27 after tumor inoculation, ranging to day 64 (Fig. 1). Whereas animals coinjected with N29 and HiB5 cells at a ratio of 1:20 showed a prolonged symptom-free period, the days of death ranged from 34 to 145. Two of eight rats survived these experiments and had no sign of intracerebral tumor growth when investigated 6 months postinoculation. In addition, one of five animals that received injections of N29 cells and HiB5 cells at a ratio of 1:5 had no sign of tumor development when an autopsy was performed 6 months after inoculation. Two nestin-positive adult neural progenitor cell lines (ANSC-2, ANSC-3-C-Myc), prepared from the rat subventricular zone (SVZ), as well as Fisher rat fibroblasts, were used as control cells. These cells had no effect in vivo on the tumor outgrowth (Fig. 1). We performed an experiment to investigate if HiB5 cells had a therapeutic effect on a pre-established tumor in nucleus Caudatus (Fig. 2). HiB5 cells were injected into nucleus Caudatus 7 days after tumor inoculation. At day 84 after N29 inoculation, the 11 control animals injected with N29 cells had been sacrificed due to motor dysfunction caused by the tumor outgrowth, whereas only 6 of 12 animals that received injections of HiB5 cells had been sacrificed due to symptoms (Fig. 2). The remaining six animals had a prolonged survival of up to 109 days after the last control animal was sacrificed. Other types of neural progenitor cell lines were screened for the ability to inhibit glioma outgrowth. The progenitor lines RN33B and ST14A were coinoculated with N29 in nucleus Caudatus (Fig. 3). RN33B showed no sign of inhibiting the tumor outgrowth, whereas ST14A coinoculations resulted in prolonged survival of the animals. At day 54 after tumor inoculation, the last control animal was sacrificed, whereas 5 of 8 animals receiving ST14A were still alive. Two animals continued to be symptom free during this experiment. Macroscopic dissections showed no sign of tumor growth 150 days after the last control animal was sacrificed. To investigate whether the effect of neural progenitor cells on glioma cells was exclusive for the N29 tumor, two different rat glioma cell lines were tested for in vivo outgrowth in nucleus Caudatus (N25, N32). HiB5 had the ability to inhibit the glioma N32 and prolong animal survival of up to 53 days (Fig. 4). However, N25 gliomas were not growth inhibited either by HiB5, RN33B, or ANSC-2 (Fig. 5).

Progenitor Cell Immunogenicity.

To rule out the possibility that an immune response to HiB5 cells was involved in the tumor inhibitory effect, Fisher rats were immunized with HiB5 at day 0 and day 14 s.c. Two weeks after the last immunization, the spleen cells were tested in a mixed lymphocyte progenitor cell culture. No response to HiB5 cells was seen (Fig. 6). Irradiated HiB5 cells were also cultured together with normal spleen cells from a Fisher rat for three days in the presence of the polyclonal activator Staphylococcal Enterotoxin A. HiB5 cells strongly inhibited the activation of the T lymphocytes (Fig. 7). Furthermore, sensitivity of HiB5 cells to natural killer cell cytolytic activity was measured by using a 51Cr release assay, and no HiB5 sensitivity to natural killer cell cytolytic activity was seen (Fig. 8).

To investigate the role of leukocyte infiltration as a potential mechanism for tumor inhibition, HiB5 cells were injected s.c. together with N29 cells into SCID FOX CHASE mice. No tumor delay was seen in these experiments (data not shown). To further elucidate a possible immunological mechanism involved in the tumor inhibition during the first 14 days postintracranial injection, we examined if any difference in leukocyte infiltration could account for the observed effects on the tumor outgrowth. Animals injected with N29 cells alone or in combination with HiB5 cells showed no difference in the infiltration of monocytes/microglia (Fig. 9, A and B), granulocytes or T cells (data not shown). T-Cell infiltration was sparse during the first week and increased slightly up to day 14 (data not shown).

To investigate the mechanism behind the inability of HiB5 to evoke an immune response in Fisher rats and to get an idea about the tumor outgrowth inhibiting effect of HiB5, we investigated the expression of different members of the TGF-β family. Reverse transcription-PCR analysis was performed on total RNA from HiB5 cells, ST14A cells, ANSC-2 cells, and lymphocytes (as a control). The cells were tested for mRNA expression of BMP2, BMP4, Inhibinβ1A, TGF-β1, TGF-β2, TGF-β3, and TNF-α. HiB5, ST14A, and ANSC-2 cells expressed the three different isoforms of TGF-β, as well as Inhibinβ1A, BMP2, and BMP4 (Table 2). ELISA experiments for secreted TNF-α, BMP4, TGF-β1, and TGF-β2 proteins were performed on conditioned medium from HiB5 and ANSC-2 cells. These experiments showed that HiB5 and ANSC-2 produce high levels of TGF-β1, 1.8 ± 0.6 and 2.1 ± 0.8 ng/ml, respectively, and somewhat lower levels of TGF-β2, 43 ± 25 and 931 ± 12 pg/ml, respectively. No BMP4 or TNF-α protein expression could be detected (data not shown). These results confirm that the TGF-β family members could indeed be involved in the immune-suppressive phenotype of HiB5 cells, but because ANSC-2 cells also express high levels, these molecules are probably not involved in the effects on the tumor outgrowth inhibition in vivo.

Morphological Differences between Coinoculated Tumors and Control Tumors.

Sections from the central part of the N29 control tumor and the tumor coinoculated with N29 cells and HiB5 cells at day 12 clearly indicated that the control tumor (stained with nestin) was larger and had more extensive finger-like projections throughout the right hemisphere than the coinoculated tumor (Fig. 9, C and D). Fig. 9, E and F, show the periphery of the same tumors, additionally illustrating the extensive invasion throughout the control animal brain. Tumor sections from the central part of N29 coinoculated with ST14A at day 21 show a similar pattern of reduced tumor growth (Fig. 9, G and H). HiB5 cells became nestin negative within the first 2 days in vivo, whereas the tumor stained strongly for nestin during the 14-day period of examination, showing only patches of unstained regions (data not shown). Reactive glia-like cells also stained positive for nestin and were present at day 1 and continued to be numerous at day 14, contributing to the scar formation (data not shown). In both the coinoculated animal brains and the animal brains with N29 growing alone, no significant difference in angiogenesis was found up to day 23 (data not shown). The gliomas N29 and N25 were positive for nestin in vitro; however, N32 lacked nestin-staining (data not shown). The growth characteristics of N32 differed from N29. The expansion of the tumor was relatively even and lacked the extensive protrusions characteristic for N29 (Fig. 10). One other important difference is that the N32 tumor reached a final volume within 22 days, whereas the median of the N29 tumor outgrowth was 38 days and the median for N25 outgrowth was 43 days.

HiB5 Cells Inhibit Glioma and Colon Carcinoma Growth Outside the CNS.

We injected HiB5 cells together with N29 tumor cells s.c. into Fisher rats at various concentrations. We showed that tumor inhibition is correlated to the amount of progenitor cells present (Fig. 11). At a 1:10 ratio of N29 cells versus HiB5 cells, four of six animals continued to be tumor free 9 months after injections. When comparing tumors at day 18 after inoculation, the control animals had an average tumor size of 853 mm3 (±81 mm3), whereas animals injected with N29 cells and HiB5 cells at a ratio of 1:10 and 1:1 had no signs of tumor growth. At day 24 after inoculation, the 1:1 ratio had an average tumor size of 90 mm3 (±85 mm3), but the 1:10 group was still tumor free.

To test whether HiB5 cells could inhibit a tumor cell line originating from another site than the CNS, we tested HiB5 cells in a rat colon cancer model (H1D2; Ref. 26). H1D2 cells (1 × 105) were coinoculated retroperitoneally with HiB5 cells (1 × 106) in three F1 Brown Norway × Wistar/Furth rats. Tumor growth was measured at day 43 after injection. The control animal tumors had a mean volume of 1922 ± 179 mm3, whereas the animals injected with H1D2 and progenitor cells continued to be tumor free during this experiment.

HiB5 Cells Have An Inhibitory Effect on Tumor Proliferation in Vitro.

We cocultured HiB5 progenitor cells with tumor cells to investigate if the progenitor cells in vitro were able to inhibit cell proliferation. Irradiated HiB5 cells cultured together with rat tumor cells (N29/H1D2) for 3 days inhibited tumor proliferation as measured by [3H]thymidine incorporation (Fig. 12).

Successful treatment of CNS tumors is an exceptionally complex problem, in part, because of the devastating effects that injury to the brain or spinal cord can have on normal functions. Moreover, it is clear that therapeutic modalities used to treat CNS tumors may themselves cause injury. This fact indicates that a useful advance in brain tumor therapy would be to develop therapeutic approaches that would both eliminate tumor cells and repair injury to the damaged CNS. Over the past years, several studies have used genetically modified neural progenitor cells, which could promote tumor regression and prolong survival of recipient animals (10, 11, 12, 13, 14, 15, 16, 17, 18, 19). The possibility that transplanted neural progenitor cells might also repair damage associated with the occurrence and/or treatment of a brain tumor has not been assessed. Transplantation studies with HiB5 and ST14A cells have shown that these rat neural progenitor lines have the capacity to survive and integrate functionally into a developing or adult rodent brain (28). Previous in vivo studies have shown that upon HiB5 or ST14A striatal implantation, the cells extensively migrate from the implantation site into surrounding adult brain regions over a period of ∼2–5 days and undergo two to three cell divisions. The majority of the grafted cells convert to a glial morphology, and only a small percentage express neural markers (28, 29). Upon induction of a lesion in the striatum, the transplanted cells were able to respond to the environment, up-regulating glial markers, and become active in the process of gliosis, even when the lesion was made distant from the transplantation site (30). Other evidence for neural progenitor cell integration and function in the injured brain has been shown where progenitor cells can restore the dopaminergic neurotransmission in substancia nigra, which is lost in Parkinson’s disease (31). Both the local environment and the intrinsic properties of the progenitor cells are important for determining the differentiated fate of the engrafted cells (32). Whether similar behavior will be seen with neural progenitor cells transplanted into a tumor-bearing brain is not yet known.

In correlation with earlier studies (29) we observed that HiB5 cells, after engraftment into nucleus Caudatus together with N29 cells at day 1, were clustered at the injection site. At this time point, the cells expressed nestin, indicating an undifferentiated state. Nestin was down-regulated around day 1 or 2. We have observed a positional localization of progenitor cells correlating with tumor cells during the first 4–5 days after inoculation into nucleus Caudatus. After day 5, the staining with which the HiB5 cells were identified (carboxyfluorescein diacetate) faded, with only a weak green signal remaining around the tumor border up to day 14. However, we do not exclude the possibility that the HiB5 cells migrated or integrated into the brain environment at a distance from the primary tumor mass because these cells have previously been shown to have migratory and differentiation capabilities (28, 29, 30). HiB5 cells exhibit plasticity in response to the local environment, and the tumor microenvironment could modify their phenotypic differentiation. It is possible that the fate of differentiation of the neural progenitor cells is of importance for the tumor outgrowth inhibiting effect. HiB5 and ST14A cells that possess tumor outgrowth-inhibiting properties preferentially differentiate toward a glial phenotype (28, 29), and RN33B, which does not show any tumor outgrowth-inhibiting effects in our rat glioma model, shows a higher capacity for differentiation toward a neural phenotype in vivo(22). It will be of importance to evaluate the fate of differentiation of the HiB5 and ST14A cells when coinoculated with glioma cells into the CNS.

It has been shown that mouse and human progenitor cells are able to migrate and home in on the site of glioma growth and target single glioma cells surrounding the tumor mass when injected directly into the tumor bed or at a site distant from the tumor (9). A useful advance in brain tumor therapy with progenitor cells was a study showing that a genetically modified mouse progenitor cell expressing interleukin 4 could promote glioma tumor regression in mice and rats (13). In addition, it was found that the unmodified stem cells also had a potential for tumor outgrowth inhibition. One of the questions raised was whether this phenomenon was exclusive for this particular progenitor cell, or if there exists a mechanism common to more progenitor cells. We show that the two rat embryonic progenitor cell lines HiB5 and ST14A have similar qualities in vivo as cells described previously (9, 13).

The glioma models used in our experiments reproduce important characteristics of human malignant gliomas. The N29 glioma cell line shows a growth pattern similar to most clinical glioblastomas with a central area of necrosis, a marked rim of tumor cells, and a peripheral zone of infiltrating cells. The tumor infiltrates with both finger-like projections into the surrounding brain and with perivascular single cells spreading far from the core of the tumor. In the presence of HiB5 and ST14A, we show evidence of a different infiltrative pattern with an overall smaller tumor core, fewer finger-like projections stretching into the surrounding brain, and little evidence of perivascular spreading except within the vicinity of the tumor core. The mechanism behind these differences is not yet known and could be attributed either to inhibition of tumor proliferation or a partial impairment of the migratory machinery. One possible mechanism for tumor inhibition is that progenitor cell treatment stimulates an immune reaction because HiB5 and ST14A are not syngeneic to the Fisher strain. Indications supporting this theory come from the lack of tumor outgrowth inhibition in SCID/FOX CHASE mice (data not shown). However, this could also be explained by the possibility that molecules produced by HiB5, mediating the tumor growth-inhibiting effect in vivo, may be species selective and unable to mediate its effect in mice. Evidence against the involvement of the immune system includes the fact that coinoculated animals should attract cells of the immune system, but there is no obvious difference in the number of infiltrating leukocytes between the control animals and the animals that received injections of HiB5 cells and tumor cells. Furthermore, we show that HiB5 cells are resistant to lysis by natural killer cells and that HiB5 does not evoke a cell-mediated immune response when Fisher rats are immunized with viable HiB5 cells. We also show that the presence of irradiated HiB5 cells strongly inhibit polyclonal activation of spleen lymphocytes from Fisher rats. Other facts arguing against the involvement of the immune system are earlier studies using HiB5 in nonhistocompatible rats and mice (33, 34). To get an indication of a possible mechanism for the immunosuppressive state of HiB5 and maybe even to the tumor growth-suppressive effect, we investigated HiB5 cells for expression of TGF-β family members. Cytokines expressed by neural progenitor cells have not been extensively studied, but a recent article shows that human, rat, and mouse neural progenitor cells express TGF-β1, TGF-β2, and TNF-α (35). We show that rat neural progenitor cell lines HiB5, ST14A, and ANSC-2 are not only positive for TGF-β1 and TGF-β2 but also for TGF-β3. Contradictory to their results, we could not detect any TNF-α expression on the mRNA or on the protein level. The high expression of TGF-β molecules is probably of importance for the absence of immunogenicity of the HiB5 cells but not for the tumor growth inhibition in vivo because ANSC-2 cells do not show any effect on glioma outgrowth.

We additionally show that the tumor-inhibiting effect elicited by HiB5 cells is not area restricted or dependent on the origin of the tumor because transplantation of gliomas intracranially and s.c. and transplantation of colon carcinomas retroperitoneally show the same type of inhibition when coinjected with HiB5. We found that HiB5 also inhibited the outgrowth of N32 glioma cells, but animals coinoculated with N25 and HiB5 had a rapid tumor progression showing no difference in survival compared with the control animals inoculated with N25 alone. We have no explanation at present as to why the observed inhibition of tumor growth seems to be dependent on the glioma model used, but many gliomas could be of neural progenitor origin themselves (36). This indicates the possibility that neural progenitor cells situated near or in direct contact with glioma cells could influence the proliferative potential and differentiation status of the glioma cells by signals originating from the differentiating neural progenitor cell line HiB5 or ST14A. We investigated the three glioma cell lines for the expression of the neural progenitor marker nestin, astroglia marker GFAP, and the neural marker MAP2. All N25 cells were nestin positive and negative for GFAP and MAP2, whereas N29 cells were mixed nestin and GFAP positive, and N32 cells were nestin negative and <5% were GFAP positive (data not shown). This could be of importance for the negative effect on N25 cells because we have observed that nestin-positive neural progenitor cells are very resistant against apoptosis-inducing agents, i.e., irradiation and immunotoxins (data not shown). It will be of interest to investigate if expression of nestin is a negative prognostic factor for glioma treatment.

In conclusion, our results suggest that the neural progenitor cell lines HiB5 and ST14A can inhibit the growth of tumor cells in vivo and in vitro. This effect is neither specific for tumors of the neuroepithelial lineage nor for tumor growth localized in the CNS. It will be of importance to define the phenotype specific for cells with this effect and find a human counterpart of this cell type.

We also conclude that HiB5 cells can act on a pre-established tumor and delay its progression as compared with the untreated control. Hence, progenitor cells seem to provide a strong antitumor effect and might have a good potential in tumor therapy.

Grant support: Swedish Children’s Cancer Foundation Grant PROJ02/001.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Karin Staflin, Department of Cell and Molecular Biology, Section for Tumor Biology, University of Lund, BMC: I 12, S-221 84 Lund, Sweden. Phone: 46-46-2229264; Fax: 46-46-2224606; E-mail: karin.staflin@tumor.lu.SE

Fig. 1.

Kaplan-Meier survival curve of Fisher rats inoculated with nucleus Caudatus with either 5000 N29 glioma cells alone or coinoculated with neural progenitor cells or fibroblasts. The Kaplan-Meier curve illustrates the symptom-free period of 26 control animals, animals receiving a mixture of N29 glioma cells and HiB5 cells at a ratio of 1:20 (n = 8) or at a ratio of 1:5 (n = 5), and animals coinoculated with N29 and the adult progenitor line ANSC-2 (n = 8), ANSC-3 cells (n = 9), or fibroblasts (Fib.) (n = 8) prepared from Fisher rats.

Fig. 1.

Kaplan-Meier survival curve of Fisher rats inoculated with nucleus Caudatus with either 5000 N29 glioma cells alone or coinoculated with neural progenitor cells or fibroblasts. The Kaplan-Meier curve illustrates the symptom-free period of 26 control animals, animals receiving a mixture of N29 glioma cells and HiB5 cells at a ratio of 1:20 (n = 8) or at a ratio of 1:5 (n = 5), and animals coinoculated with N29 and the adult progenitor line ANSC-2 (n = 8), ANSC-3 cells (n = 9), or fibroblasts (Fib.) (n = 8) prepared from Fisher rats.

Close modal
Fig. 2.

Kaplan-Meier survival curve illustrating the inhibition of an established tumor in nucleus Caudatus by HiB5 cells. Fisher rats were inoculated intracranially with 5000 N29 (n = 23) cells at day 0. At day 7, 12 of the rats were inoculated with additional 250,000 HiB5 cells. The control animals were sham-inoculated with an empty canula.

Fig. 2.

Kaplan-Meier survival curve illustrating the inhibition of an established tumor in nucleus Caudatus by HiB5 cells. Fisher rats were inoculated intracranially with 5000 N29 (n = 23) cells at day 0. At day 7, 12 of the rats were inoculated with additional 250,000 HiB5 cells. The control animals were sham-inoculated with an empty canula.

Close modal
Fig. 3.

Kaplan-Meier survival curve illustrating the inhibition of glioma N29 growth by ST14A but not by RN33B cells after coinoculation into nucleus Caudatus. Fisher rats were injected with N29 (n = 11) alone or together with ST14A cells (n = 8) or coinjected with RN33B (n = 9) at a ratio of 1:20.

Fig. 3.

Kaplan-Meier survival curve illustrating the inhibition of glioma N29 growth by ST14A but not by RN33B cells after coinoculation into nucleus Caudatus. Fisher rats were injected with N29 (n = 11) alone or together with ST14A cells (n = 8) or coinjected with RN33B (n = 9) at a ratio of 1:20.

Close modal
Fig. 4.

Kaplan-Meier survival curve illustrating the inhibition of the N32 glioma cell outgrowth after coinoculation with HiB5 cells into nucleus Caudatus (n = 5 rats/group) at a ratio of 1:20.

Fig. 4.

Kaplan-Meier survival curve illustrating the inhibition of the N32 glioma cell outgrowth after coinoculation with HiB5 cells into nucleus Caudatus (n = 5 rats/group) at a ratio of 1:20.

Close modal
Fig. 5.

Kaplan-Meier survival curve illustrating the outgrowth of N25 glioma cells and the lack of growth inhibition by neural progenitor cells. N25 cells (control group n = 8) were coinoculated with HiB5 cells, RN33B cells (n = 5 rats/group), or ANSC-2 (n = 4) cells into nucleus Caudatus of Fisher rats.

Fig. 5.

Kaplan-Meier survival curve illustrating the outgrowth of N25 glioma cells and the lack of growth inhibition by neural progenitor cells. N25 cells (control group n = 8) were coinoculated with HiB5 cells, RN33B cells (n = 5 rats/group), or ANSC-2 (n = 4) cells into nucleus Caudatus of Fisher rats.

Close modal
Fig. 6.

Immunogenicity testing of the HiB5 cells in the Fisher model. Two animals were immunized with 1 × 106 HiB5 cells at day 0 and day 14. The spleen cells were harvested 14 days after last immunization and tested in a mixed lymphocyte assay together with irradiated HiB5 cells or stimulated with SEA (0.01 ng/ml) as control.

Fig. 6.

Immunogenicity testing of the HiB5 cells in the Fisher model. Two animals were immunized with 1 × 106 HiB5 cells at day 0 and day 14. The spleen cells were harvested 14 days after last immunization and tested in a mixed lymphocyte assay together with irradiated HiB5 cells or stimulated with SEA (0.01 ng/ml) as control.

Close modal
Fig. 7.

Spleen lymphocyte inhibition by irradiated HiB5 cells. Staphylococcal Enterotoxin A-stimulated lymphocytes (Lymph.), (0.01 ng/ml; 300,000 cells/well) were cocultured with 5,000 irradiated HiB5 cells for a period of 4 days.

Fig. 7.

Spleen lymphocyte inhibition by irradiated HiB5 cells. Staphylococcal Enterotoxin A-stimulated lymphocytes (Lymph.), (0.01 ng/ml; 300,000 cells/well) were cocultured with 5,000 irradiated HiB5 cells for a period of 4 days.

Close modal
Fig. 8.

A natural killer cell cytotoxicity test was done on Cr51-labeled HiB5 cells cultured with spleen cells from Fisher rats for 17 h. YAC-1 cells were used as natural killer-sensitive positive controls for the specific release.

Fig. 8.

A natural killer cell cytotoxicity test was done on Cr51-labeled HiB5 cells cultured with spleen cells from Fisher rats for 17 h. YAC-1 cells were used as natural killer-sensitive positive controls for the specific release.

Close modal
Fig. 9.

Histopathological features of tumor growth in nucleus Caudatus 12 days after inoculation of N29 glioma cells alone or coinoculated with HiB5 cells. An equal amount of monocyte infiltration was seen (×100 magnification) as shown by 1C7 antibody staining when comparing the control tumor (A), with the HiB5 + N29 coinoculated tumor (B). A central section (×40 magnification) of the N29 tumor stained with nestin (C), a central section of the HiB5-coinoculated tumor stained with nestin (D), peripheral part of the control tumor, illustrating finger-like projections stained with nestin (E), and the corresponding site in the coinoculated animal stained with nestin (F), illustrating a smaller tumor with fewer projections. Tumor growth in nucleus Caudatus 21 days after inoculation of N29 glioma cells alone or coinoculated with ST14A cells (×40 magnification). The central part of the N29 tumor stained with nestin (G), the central part of a ST14A-coinoculated tumor stained with nestin (H), illustrating reduced tumor size and fewer finger-like projections. All slides were counterstained with hematoxylin. The scale bar in B indicates 1 μm in A and B; the scale bar in H indicates 1 μm in C–H.

Fig. 9.

Histopathological features of tumor growth in nucleus Caudatus 12 days after inoculation of N29 glioma cells alone or coinoculated with HiB5 cells. An equal amount of monocyte infiltration was seen (×100 magnification) as shown by 1C7 antibody staining when comparing the control tumor (A), with the HiB5 + N29 coinoculated tumor (B). A central section (×40 magnification) of the N29 tumor stained with nestin (C), a central section of the HiB5-coinoculated tumor stained with nestin (D), peripheral part of the control tumor, illustrating finger-like projections stained with nestin (E), and the corresponding site in the coinoculated animal stained with nestin (F), illustrating a smaller tumor with fewer projections. Tumor growth in nucleus Caudatus 21 days after inoculation of N29 glioma cells alone or coinoculated with ST14A cells (×40 magnification). The central part of the N29 tumor stained with nestin (G), the central part of a ST14A-coinoculated tumor stained with nestin (H), illustrating reduced tumor size and fewer finger-like projections. All slides were counterstained with hematoxylin. The scale bar in B indicates 1 μm in A and B; the scale bar in H indicates 1 μm in C–H.

Close modal
Fig. 10.

Histopathological features of N32 growth in Fisher rats. A central section of N32 tumor 16 days after inoculation, stained with hematoxylin (×40 magnification). The scale bar indicates 1 mm.

Fig. 10.

Histopathological features of N32 growth in Fisher rats. A central section of N32 tumor 16 days after inoculation, stained with hematoxylin (×40 magnification). The scale bar indicates 1 mm.

Close modal
Fig. 11.

HiB5 cells inhibit s.c. N29 tumor growth. A total of 200,000 N29 cells was injected into the flank of Fisher rats (n = 14) alone or injected together with HiB5 at a ratio of 1:10 (n = 6) and at a ratio of 1:1 (n = 8). Tumor size was measured at 1-week intervals after injection. Four rats of the 1:10 ratio of tumor versus HiB5 cells continued to be tumor free over time. The animals were sacrificed at a tumor size of ∼1000 mm3.

Fig. 11.

HiB5 cells inhibit s.c. N29 tumor growth. A total of 200,000 N29 cells was injected into the flank of Fisher rats (n = 14) alone or injected together with HiB5 at a ratio of 1:10 (n = 6) and at a ratio of 1:1 (n = 8). Tumor size was measured at 1-week intervals after injection. Four rats of the 1:10 ratio of tumor versus HiB5 cells continued to be tumor free over time. The animals were sacrificed at a tumor size of ∼1000 mm3.

Close modal
Fig. 12.

Inhibition of tumor cell proliferation in vitro by irradiated HiB5 cells. Cell growth was measured after coculture of 5000 irradiated HiB5 cells together with N29 glioma (500 cells/well) cells or H1D2 colon carcinoma cells (300 cells/well) for 4 days.

Fig. 12.

Inhibition of tumor cell proliferation in vitro by irradiated HiB5 cells. Cell growth was measured after coculture of 5000 irradiated HiB5 cells together with N29 glioma (500 cells/well) cells or H1D2 colon carcinoma cells (300 cells/well) for 4 days.

Close modal
Table 1

Primer sequence, amplification conditions, and product length of the analyzed genes

Primer sequenceProduct approximatesAnnealing temperature
Rat BMP2 TGGTGTCCAATAGTCTGGTC 240 bp 60°C 
 CGAAGAAGCCATCGAGGAAC   
Rat BMP4 GTCACATTGTGACGGACTAG 260 bp 58°C 
 ACGAAGAACATCTGGAGAAC   
Rat TGF-β1a ATTCCGTCTCCTTGGTTCAG 230 bp 60°C 
 ACGTCAGACATTCGGGAAGC   
Rat TGF-β2 ACGGTATGAAGGTACAGCAG 260 bp 60°C 
 GCAGAGTTCAGGGTCTTTCG   
Rat TGF-β3 TTCCAGTATGTCTCCATTGG 220 bp 60°C 
 AGAGAATCGAGCTCTTCCAG   
Rat TNF-α CAGAGGAGGAGCTGGAGTG 280 bp 58°C 
 TCATACCAGGGCTTGACCTC   
Rat inhibin β1A TGAGGATGGTCTTCAGACTG 190 bp 58°C 
 TGCGGATTGCTTGTGAACAG   
Primer sequenceProduct approximatesAnnealing temperature
Rat BMP2 TGGTGTCCAATAGTCTGGTC 240 bp 60°C 
 CGAAGAAGCCATCGAGGAAC   
Rat BMP4 GTCACATTGTGACGGACTAG 260 bp 58°C 
 ACGAAGAACATCTGGAGAAC   
Rat TGF-β1a ATTCCGTCTCCTTGGTTCAG 230 bp 60°C 
 ACGTCAGACATTCGGGAAGC   
Rat TGF-β2 ACGGTATGAAGGTACAGCAG 260 bp 60°C 
 GCAGAGTTCAGGGTCTTTCG   
Rat TGF-β3 TTCCAGTATGTCTCCATTGG 220 bp 60°C 
 AGAGAATCGAGCTCTTCCAG   
Rat TNF-α CAGAGGAGGAGCTGGAGTG 280 bp 58°C 
 TCATACCAGGGCTTGACCTC   
Rat inhibin β1A TGAGGATGGTCTTCAGACTG 190 bp 58°C 
 TGCGGATTGCTTGTGAACAG   
a

TGF, transforming growth factor; TNF, tumor necrosis factor.

Table 2

Gene expression analysis of TGFβ family members and TNFα

Cells/PCRLymphocyteANSC2ST14AHiB5
TNF-αa − − − 
TGF-β1 
TGF-β2 − 
TGF-β3 − 
BMP2 
BMP4 − 
Inhibin β1A 
Cells/PCRLymphocyteANSC2ST14AHiB5
TNF-αa − − − 
TGF-β1 
TGF-β2 − 
TGF-β3 − 
BMP2 
BMP4 − 
Inhibin β1A 
a

TNF, tumor necrosis factor; TGF, transforming growth factor.

We thank Anna Darabi for supplying histological artwork for Fig. 10.

1
Watt FM, Hogan BLM. Out of Eden: stem cells and their niches.
Science (Wash. DC)
,
287
:
1427
-30,  
2000
.
2
Frisén J. Stem cell plasticity?.
Neuron
,
35
:
415
-8,  
2002
.
3
Gage FH. Mammalian neural stem cells.
Science (Wash. DC)
,
287
:
1433
-8,  
2000
.
4
Palmer TD, Takahashi J, Gage FH. The adult rat hippocampus contains primordial neural stem cells.
Mol Cell Neurosci
,
8
:
389
-404,  
1997
.
5
Reynolds BA, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system.
Science (Wash. DC)
,
255
:
1707
-10,  
1992
.
6
Van der Kooy D, Weiss S. Why stem cells?.
Science (Wash. DC)
,
287
:
1439
-41,  
2000
.
7
Goldman SA, Luskin MB. Strategies utilized by migrating neurons of the postnatal vertebrate forebrain.
Trends Neurosci
,
21
:
107
-14,  
1998
.
8
Eriksson PS, Perfilieva E, Bjork-Eriksson T, et al Neurogenesis in the adult human hippocampus.
Nat Med
,
4
:
1313
-7,  
1998
.
9
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 USA
,
97
:
12846
-51,  
2000
.
10
Ehtesham M, Kabos P, Gutierrez MAR, et al Induction of glioblastoma apoptosis using neural stem cell-mediated delivery of tumor necrosis factor-related apoptosis-inducing ligand.
Cancer Res
,
62
:
7170
-4,  
2002
.
11
Ehtesham M, Kabos P, Kabosova A, Neuman T, Black KL, Yu JS. The use of interleukin 12-secreting neural stem cells for the treatment of intracranial gliomas.
Cancer Res
,
62
:
5657
-63,  
2002
.
12
Barresi V, Belluardo N, Sipione S, Mudo G, Cattaneo E, Filippo Condorelli D. Transplantation of prodrug-converting neural progenitor cells for brain tumor therapy.
Cancer Gene Ther
,
10
:
396
-402,  
2003
.
13
Benedetti S, Pirola B, Pollo B, et al Gene therapy of experimental brain tumors using neural progenitor cells.
Nat Med
,
6
:
447
-50,  
2000
.
14
Herrlinger U, Woiciechowski C, Sena-Esteves M, et al Neural precursor cells for delivery of replication-conditional HSV-1 to intracerebral gliomas.
Mol Ther
,
1
:
347
-57,  
2000
.
15
Hendriks WT, Ruitenberg MJ, Blits B, Boer GJ, Verhaagen J. Viral vector-mediated transfer of neurotrophins to promote regeneration of the injured spinal cord.
Prog Brain Res
,
146
:
451
-76,  
2004
.
16
Corti O, Sabate P, Colin P, et al A single adenovirus vector mediates doxycycline-controlled expression of tyrosine hydroxylase in brain grafts of human neural progenitors.
Nat Biotechnol
,
17
:
349
-54,  
1999
.
17
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
:
9
-15,  
2002
.
18
Allessandri G, Emanueli C, Madeddu P. Genetically engineered stem cell therapy for tissue regeneration.
Ann N Y Sci
,
1015
:
271
-84,  
2004
.
19
Fulda S, Wick W, Weller M, Debatin KM. Smac agonists sensitize for Apo2L/TRAIL- or anticancer drug-induced apoptosis and induce regression of malignant glioma in vivo.
Nat Med
,
8
:
808
-15,  
2002
.
20
Renfranz PJ, Cunningham MG, McKay RD. Region-specific differentiation of the hippocampal stem cell line HiB5 upon implantation into the developing mammalian brain.
Cell
,
66
:
713
-29,  
1991
.
21
Cattaneo E, Magrassi L, Butti G, Santi L, Giavazzi A, Pezzotta S. A short term analysis of the behaviour of conditionally immortalized neuronal progenitors and primary neuroepithelial cells implanted into the fetal brain.
Dev Brain Res
,
83
:
197
-208,  
1994
.
22
Whittemore SR, White LA. Target regulation of neural differentiation in a temperature-sensitive cell line derived from medullary raphe.
Brain Res
,
615
:
27
-40,  
1993
.
23
Palmer TD, Markakis EA, Willhoite AR, Safar F, Gage FH. Fibroblast growth factor-2 activates a latent neurogenic program in neural progenitor cells from diverse regions of the adult CNS.
J Neurosci
,
19
:
8487
-97,  
1999
.
24
Koestner A, Swenberg JA, Wechsler W. Transplacental production with ethylnitrosourea of neoplasms of the nervous system in Sprague Dawley rats.
Am J Pathol
,
63
:
37
-56,  
1971
.
25
Siesjö P, Visse E, Lindvall M, Salford L, Sjögren HO. Immunization with mutagen-treated (tum) cells causes rejection of nonimmunogenic rat glioma isografts.
Cancer Immunol Immunother
,
37
:
67
-74,  
1993
.
26
Hegardt P, Widegren P, Li L, Sjögren B, Kjellman C, Sjögren HO. Nitric oxide synthase inhibitor and IL-18 enhance the anti-tumor immune response of rats carrying an intrahepatic colon carcinoma.
Cancer Immunol Immunother
,
50
:
491
-501,  
2001
.
27
Lindvall M, Eriksson H, Hedlund G, Sjögren HO. Selective cytotoxicity of two rodent T-cell lymphomas to rat yolc sac tumours involves a retroviral envelope protein expressed by the lymphoma.
Scand J Immunol
,
46
:
479
-87,  
1997
.
28
Lundberg C, Martinez-Serrano A, Cattaneo E, McKay RDG, Björklund A. Survival, integration, and differentiation of neural stem cell lines after transplantation to the adult rat striatum.
Exp Neurol
,
145
:
342
-60,  
1997
.
29
Martinez-Serrano A, Lundberg C, Björklund A. Use of conditionally immortalized neural progenitors for transplantation and gene transfer to the CNS Gage F Christen Y eds. .
Isolation, characterization and utilization of CNS stem cells
,
p. 151
-68, Springer-Verlag Berlin Heidelberg  
1997
.
30
Lundberg C, Björklund A. Host regulation of glial markers in intrastriatal grafts of conditionally immortalized neural stem cell lines.
NeuroReport
,
7
:
847
-52,  
1996
.
31
Björklund A, Lindvall O. Cell replacement therapies for central nervous system disorders.
Nat Am
,
3
:
537
-44,  
2000
.
32
Cao Q, Benton RL, Whittemore SR. Stem cell repair of central nervous system injury.
J Neurosci Res
,
68
:
501
-10,  
2002
.
33
Sheen VL, Arnold MW, Wang Y, Macklis JD. Neural precursor differentiation following transplantation into neocortex is dependent on intrinsic developmental state and receptor competence.
Exp Neurol
,
158
:
47
-62,  
1999
.
34
Philips MF, Mattiasson G, Wieloch T, et al Neuroprotective and behavioral efficacy of nerve growth factor transfected hippocampal progenitor cell transplants after experimental traumatic brain injury.
J Neurosurg
,
94
:
765
-74,  
2001
.
35
Klassen HJ, Imfeld KL, Tai L, Gage FH, Young MJ, Berman MA. Expression of cytokines by multipotent progenitor cells.
Cytokine
,
22
:
101
-6,  
2003
.
36
Recht L, Jang T, Savarese T, Litofsky NS. Neural stem cells and neuro-oncology: quo vadis?.
J Cell Biochem
,
88
:
11
-9,  
2003
.