Neural progenitor-like cells have been isolated from bone marrow and the cells have the ability of tracking intracranial tumor. However, the capacity of the cells to deliver molecules for activating immune response against intracranial tumor and the identity of cellular and molecular factors that are involved in such immune responses have yet to be elucidated. Here, we isolated neural stem-like cells from the bone marrow of adult mice. The isolated cells were capable of producing progenies of three lineages, neurons, astrocytes, and oligodendrocytes, in vitro and tracking glioma in vivo. By genetically manipulating bone marrow–derived neural stem-like cells (BM-NSC) to express a recently discovered cytokine, interleukin (IL)-23, the cells showed protective effects in intracranial tumor-bearing C57BL/6 mice. Depletion of subpopulation lymphocytes showed that CD8+ T cells were critical for the antitumor immunity of IL-23–expressing BM-NSCs and that CD4+ T cells and natural killer (NK) cells participated in the activity. Furthermore, the IL-23–expressing BM-NSC-treated survivors were resistant to the same tumor rechallenge associated with enhanced IFN-γ, but not IL-17, expression in the brain tissue. Taken together, these data suggest that IL-23–expressing BM-NSCs can effectively induce antitumor immunity against intracranial gliomas. CD8+ T cells are critical for such antitumor activity; in addition, CD4+ T cells and NK cells are also involved. (Cancer Res 2006; 66(5): 2630-8)

Malignant gliomas represent a significant cause of tumor-related mortality and continue to be associated with poor prognosis despite extensive surgical excision and adjuvant radiotherapy and chemotherapy (1, 2). Their resistance to treatment is related to the tumor cell's exceptional migratory nature and ability to grow extensively into normal neural tissue away from the main tumor mass. These cells preclude complete surgical resection and eventually serve as a source for recurrent tumor growth. It is this characteristic that also makes it difficult for otherwise promising gene therapeutic strategies and other local interventions to efficiently access the tumor cells (3). Recent evidence indicated a new strategy for targeting tumor cells within the central nervous system by using neural stem cells to deliver therapeutic gene and/or their products efficiently to treat brain tumors (46).

Bone marrow represents an attractive source for generating neural stem cells, which have an advantage over fetal neural stem cells, or allogeneic cell lines as a cellular vehicle for brain tumor therapy because of their autologous characteristic. The bone marrow–derived neural stem cells obviate the accessibility and ethical problems associated with fetal neural stem cells as well as potential immunologic incompatibility due to the requirement for allogeneic application (713). In the present study, we sought to isolate neural stem cells from murine bone marrow as a cellular vehicle for glioma therapy. Using bone marrow–derived neural stem-like cells (BM-NSC) as a vehicle, we studied the antiglioma function of a newly discovered cytokine, interleukin (IL)-23, which consists of a heterodimer of IL-12 p40 subunit and a novel protein, p19. It has been shown that IL-23 acts on memory T cells and dendritic cells directly to promote IL-12 and IFN-γ production in vitro (14, 15). Recently, IL-23–expressing tumor cell showed antitumor and antimetastatic function (16). We showed that neural stem-like cells can be generated from adult bone marrow and the generated BM-NSCs can track migratory glioma cells and deliver IL-23 in situ. IL-23–expressing BM-NSCs when injected into brain are able to induce tumor-specific antitumor activity and protective immunity against intracranial glioma.

Cell Cultures and Cell Lines

Whole bone marrow was harvested from the femurs of adult C57BL/6 mice as described previously (11). Cells were plated on poly-d-lysine–coated 24-well plates (BD Biosciences, San Jose, CA) and cultured in serum-free DMEM/F-12 (Invitrogen, Gaithersburg, MD) supplemented with B27 (Invitrogen), 20 ng/mL basic fibroblast growth factor (Peprotech, Rocky Hill, NJ), and 20 ng/mL epidermal growth factor (Peprotech) along with antibiotics. The cells were plated at a density of 1 × 106 per well. Fetal neural stem cells were harvested from the frontoparietal regions of day 15 fetal C57BL/6 mice as described previously (17). The cells were grown in culture under the same conditions as mentioned above. Spheres, formed in both bone marrow cell and fetal neural stem cell culture, were processed for immunocytochemistry staining. Bone marrow–derived spheres were subjected to cell differentiation by plating onto laminin-coated 24-well plates in growth medium devoid of growth factors containing retinoic acid (1 μmol/L, Sigma, St. Louis, MO) and dibutyryl cyclic AMP (1 mmol/L, Sigma). Medium was replenished every 3 days. Cells were differentiated for 7 to 14 days and processed for immunocytochemistry staining.

NIH-3T3 cells (murine fibroblasts) and GL26 cells (murine glioma) were grown in DMEM and RPMI 1640, respectively, supplemented with 10% fetal bovine serum, 2 mmol/L l-glutamine, and antibiotics (all the reagents from Invitrogen). Green fluorescent protein (GFP)–expressing GL26 cells were derived from stable transfection of GL26 cells with pEGFP-N1 vector (BD Biosciences) by calcium phosphate precipitation method as described previously (18). The stable transfectants were selected with geneticin.

Adenoviral Vectors Construction and In vitro Infection

The viral vector bearing the gene for murine IL-"23 (AdIL-23) was constructed by subcloning the single-chain IL-23 cDNA (15) into a shuttle vector pMH5 (Microbix Biosystems, Inc., Toronto, Ontario, Canada) downstream of the mCMV promoter. The resultant shuttle vector was cotransfected with pBHGE3 (Microbix Biosystems) into 293 cells (Microbix Biosystems). Recombinants were subsequently subjected to three rounds of plaque purifications. The purified viral vector was clarified by Southern hybridization and immunocytochemistry to confirm the bearing of IL-23 expression cassette within the vector genome and the expression of IL-23 from the vector transduced cells, respectively.3

3

Yuan et al., unpublished data.

The viral vectors bearing the gene for LacZ (AdLacZ) or a 2.4-kb noncoding stuffer DNA (AdEmpty) were constructed the same way as AdIL-23. The construction of IL-12-bearing vector (AdIL-12) was described previously (6). For in vitro gene transduction, BM-NSCs, fetal neural stem cells, and NIH-3T3 cells were infected with 100 multiplicities of infection of AdIL-23, AdEmpty, AdLacZ, and AdIL-12 as indicated. Twenty-four hours after infection, the cells were washed thrice with PBS to ensure that final intracranial injections or in vitro analysis were devoid of free viral particles. NIH-3T3 cells were treated by either mitomycin C (25 mg/mL, Sigma) or irradiation (5,000 rads) to induce growth arrest.

Animal Studies In vivo

Intracranial tumor implantation. C57BL/6 wild-type mice, athymic nude mice, and CD4 T-cell knockout mice (6-8 weeks old, all from The Jackson Laboratory, Bar Harbor, ME) were anesthetized with i.p. ketamine and medetomidine and stereotactically implanted with 1 × 104 GL26 cells or GFP-expressing GL26 glioma cells in 2.5 μL of 1.2% methylcellulose/PBS in the right striatum as described previously (6).

BM-NSC migration and transgene delivery. To determine whether BM-NSCs were able to track glioma cells, deliver transgenes to tumor targets, and generate progenies of different phenotype, AdLacZ and AdIL-23 infected BM-NSCs (2 × 105 cells) were peritumorally (1 mm lateral and 3 mm behind the tumor implantation site) and intratumorally (at the tumor implantation site) injected into the brain with a 7-day established glioma. Animals were euthanized on days 12, 24, 28, and 42 after the BM-NSC injection by intracordic perfusion-fixation with 4% paraformaldehyde. Animal brains were cut into 40-μm coral sections and processed for immunohistochemistry staining.

Animal survival and brain tissue evaluation. For C57BL/6 mice survival experiments, the intracranial glioma-bearing mice were randomly divided into four groups 3 days after tumor implantation and treated with intratumoral injections of saline (3 μL, n = 12), 2 × 105 BM-NSCs infected with either AdEmpty (BM-NSC-E, n = 18) or AdIL-23 (BM-NSC-IL-23, n = 17), or NIH-3T3 cells infected with AdIL-23 (NIH-3T3-IL-23, n = 20) in 3 μL serum-free medium. Animals used for histologic evaluation were treated similarly to the survival experiment and euthanized 4 weeks after the BM-NSC injections by intracordic perfusion-fixation as mentioned above.

Depletion of CD4+ and CD8+ T cells and natural killer cells. For depletion with monoclonal antibodies (mAb), each mouse was injected i.p. with 0.5 mg rat anti-mouse CD8 (53-6.7) and anti-CD4 (GK1.5, both from American Type Culture Collection, Manassas, VA) mAb or normal rat IgG as control antibody in 200 μL PBS 1 day before tumor implantation, once daily for the following 3 consecutive days and then twice weekly. Natural killer (NK) cells were depleted by i.p. injection of 20 μL rabbit anti–asialo GM1 antiserum (Wako Chemicals, Richmond, VA) or normal rabbit serum as control using the same schedule as above. Intracranial tumor-bearing mice with specific cell population depletion and CD4 T-cell knockout mice with intracranial tumor implantation were treated with either BM-NSC-IL-23 or BM-NSC-E and were followed for survival.

Animal rechallenge. C57BL/6 wild-type mice that survived intracranial tumor implantation due to BM-NSC-IL-23 treatment were rechallenged with GL26 cells or GL26 cells plus either BM-NSC-IL-23 or BM-NSC-E. Age-matched naive mice were challenged as control. After the rechallenge/challenge, animals were followed for survival. Some animals were euthanized on days 1, 3, 5, 7, 10, and 14 after the rechallenge/challenge for reverse transcription PCR (RT-PCR) analysis of the brain tissues. All of the animals used were done in strict accordance with Institutional Animal Care and Use Committee guidelines in force at Cedars-Sinai Medical Center.

ELISA and RT-PCR

Supernatants from BM-NSCs and fetal neural stem cells, which were infected with AdEmpty, AdIL-12, or AdIL-23 for 24 hours, were analyzed by a sandwich ELISA specific for p40 as described previously (15). The cell pellets were subjected to total RNA extraction with a RNeasy Mini kit (Qiagen, Valencia, CA). Brain tissues from the rechallenged/challenged animals as described above were subjected to total RNA extraction with a RNeasy Lipid Tissue Mini kit (Qiagen). The RNA was reverse transcripted by using a Bioscript kit (Bioline, Randolph, MA) and oligo(dT)12-18 primer (Invitrogen). The PCR was carried out in a 20 μL reaction mixture that contained 1 μL cDNA as template, specific oligonucleotide primer pairs (Table 1), and Accuzyme (Bioline). Each specific gene was concurrently amplified with internal control β-actin in the same reaction tube as described previously (19). The amplified products were identified by agarose gel electrophoresis and ethidium bromide staining.

Table 1.

Oligonucleotides for PCR

mRNA targetsOligonucleotides (5′→3′)Product size (bp)
p19 Forward: CAGCAGCTCTCTCGGAATCT 361 
 Reverse: TAGAACTCAGGCTGGGCATC  
IFN-γ Forward: GGCTGTTTCTGGCTGTTACTG 427 
 Reverse: GAATCAGCAGCGACTCCTTT  
IL-17 Forward: TCTCTGATGCTGTTGCTGCT 408 
 Reverse: CACACCCACCAGCATCTTCT  
β-actin Forward: GGACTCCTATGTGGGTGACG 293 
 Reverse: TACGACCAGAGGCATACAGG  
mRNA targetsOligonucleotides (5′→3′)Product size (bp)
p19 Forward: CAGCAGCTCTCTCGGAATCT 361 
 Reverse: TAGAACTCAGGCTGGGCATC  
IFN-γ Forward: GGCTGTTTCTGGCTGTTACTG 427 
 Reverse: GAATCAGCAGCGACTCCTTT  
IL-17 Forward: TCTCTGATGCTGTTGCTGCT 408 
 Reverse: CACACCCACCAGCATCTTCT  
β-actin Forward: GGACTCCTATGTGGGTGACG 293 
 Reverse: TACGACCAGAGGCATACAGG  

Relative Quantitative Real-Time PCR

Total RNA isolation and cDNA preparation were done as described above. Detection of IFN-γ mRNA level was done by a real-time PCR assay using the iCycler iQ system (Bio-Rad Laboratories, Hercules, CA). A 12.5 μL reaction mixture containing 1 μL cDNA and 200 nmol/L of each primer was mixed with 12.5 μL of 2× iQ SYBR Green Supermix (Bio-Rad Laboratories). The reaction conditions were designed as follows: 95°C for 3 minutes to activate the iTaq DNA polymerase followed by 40 cycles with 30 seconds at 95°C and 30 seconds at 60°C. PCR amplification of the endogenous β-actin was done for each sample to control for sample loading and to allow normalization between samples. The threshold cycle (Ct; the cycle number at which the amount of amplified gene of interest reached a fixed threshold) was determined subsequently. Each data point was examined for integrity by analysis of the amplification plot and disassociation curves. All amplifications were conducted in triplicates. The relative quantitation of IFN-γ mRNA expression was calculated by the comparative Ct method. The relative quantitation value of target, normalized to endogenous control β-actin and relative to a calibrator, is calculated as follow: fold increased = 2−[ΔCt (survived animal) − ΔCt (naive animal)], where ΔCt = Ct (IFN-γ) − Ct (β-actin).

Cytotoxicity Assays

Spleen cells were harvested from each mouse. The harvested cells (5 × 106/mL) were restimulated in vitro by coculture with mitomycin C–treated GL26 cells (5 × 105/mL) for 5 days and used as effector cells in a lactate dehydrogenase release assay. GL26 parental tumor cells or p815 cells (1 × 104/well) and serial dilutions of effector cells were incubated in a 96-well U-bottomed plate at 37°C for 5 hours. Supernatants then were analyzed with a cytotoxicity detection kit (Roche Applied Science, Indianapolis, IN) according to the manufacturer's instructions. Results were expressed as the percentage of specific lysis.

H&E, Luxol Fast Blue, and Immunohistochemistry Staining of Brain Sections

The perfusion-fixed brains were cut into 10-μm coronal sections and stained with either H&E or luxol fast blue as per standard protocol. To characterize the brain tissue by immunohistochemistry, free-floating 40-μm sections were treated with 10% donkey serum (Sigma) for 30 minutes at room temperature and then stained with primary antibodies for anti-β-galactosidase protein (mouse mAb, 1:1,000, Promega, Madison, WI), anti-p40 subunit of murine IL-23 (mouse mAb, 1:50, BD Biosciences), anti–β-tubulin III (TuJ1, mouse mAb, 1:200, Chemicon, Temecula, CA), anti–glial fibrillary acidic protein (GFAP; rabbit polyclonal, 1:1,000, Chemicon), anti-myelin/oligodendrocyte (mouse mAb, 1:1,000, Chemicon), anti-F4/80 (rat mAb, 1:50, Serotec, Raleigh, NC), anti-CD4 (rat mAb, 1:50, BD Biosciences), anti-CD8 (rat mAb, 1:100, BD Biosciences), and isotype control antibodies. The primary antibodies were detected with either Texas red–conjugated donkey anti-mouse, anti-rat IgG (1:200, The Jackson Laboratory) before mounting the sections or Vector Elite ABC kit (Vector Laboratories, Burlingame, CA) and developed with diaminobenzidine (Sigma) and counterstained with hematoxylin before mounting the sections.

Neural stem-like cells can be generated from adult mouse bone marrow. We harvested adult mouse whole bone marrow cells and cultured the cells under the conditions described previously that are permissive for neural stem cells (11, 20). Free floating spheres were formed after 8 to 12 days of culture. These bone marrow–derived spheres were morphologically indistinguishable from the neurospheres of fetal brain neural stem cells and were stained positive for the neural stem cell marker, nestin (Fig. 1A). Upon being switched into differentiation conditions (11), the spheres attached and spread out. After 7 to 14 days of in vitro differentiation, we analyzed the culture for cells expressing neural and glial markers. TuJ1-positive (neuron-specific β-tubulin III), GFAP-positive (astrocyte marker), and myelin/oligodendrocyte-positive (oligodendrocyte marker) cells were clearly identified in the differentiated progenies of the bone marrow–derived spheres (Fig. 1B).

Figure 1.

Generation of neural stem-like cells from adult mouse bone marrow. A, neurospheres were formed within the culture of adult mouse bone marrow cells (a). The formed spheres were self-renewing in the free-floating pattern (b), morphologically indistinguishable from the neurospheres of fetal neural stem cells (c), and expressed a similar pattern of nestin (d) as that expressed by fetal neural stem cells (e). B, bone marrow–derived spheres were able to differentiate into β-tubulin III–expressing neurons (a), GFAP-expressing astrocytes (b), and myelin/oligodendrocyte marker–expressing oligodendrocytes (c). Bar, 50 μm.

Figure 1.

Generation of neural stem-like cells from adult mouse bone marrow. A, neurospheres were formed within the culture of adult mouse bone marrow cells (a). The formed spheres were self-renewing in the free-floating pattern (b), morphologically indistinguishable from the neurospheres of fetal neural stem cells (c), and expressed a similar pattern of nestin (d) as that expressed by fetal neural stem cells (e). B, bone marrow–derived spheres were able to differentiate into β-tubulin III–expressing neurons (a), GFAP-expressing astrocytes (b), and myelin/oligodendrocyte marker–expressing oligodendrocytes (c). Bar, 50 μm.

Close modal

BM-NSCs can track intracranial tumors and express transgene both in vitro and in vivo. We sought to determine whether bone marrow–derived sphere cells have migratory properties similar to that of fetal neural stem cells (46). To investigate their ability to migrate toward intracranial gliomas, the bone marrow–derived sphere cells were transduced with AdLacZ and injected intratumorally or peritumorally (at a site distant from the tumor bed) into mouse brain harboring an established glioma. The AdLacZ transduced cells migrated into the tumor mass (Fig. 2A,, a) or tumor “satellite,” which was away from the main tumor mass (Fig. 2A,, b). Some β-galactosidase-positive cells were seen in immediate proximity to and intermixed with invading tumor “islands” (Fig. 2A,, c, arrow). Based on the above characteristics, the sphere cells were named BM-NSCs. To test the ability to express transgenes, the BM-NSCs were transduced with adenoviral vectors carrying either stuffer DNA, IL-12, or single-chain IL-23. In parallel, fetal neural stem cells were also transduced as controls. After 24 hours, the transduced cells and conditioned medium were harvested for mRNA analysis and protein assay, respectively. One subunit of IL-23, p19, was detected as mRNA in both BM-NSC and fetal neural stem cell lysate by RT-PCR (Fig. 2B,, a). Another subunit of IL-23 and of IL-12, p40, was detected in the protein format by ELISA in the conditioned medium (Fig. 2B,, b). The mRNA of p19 and the protein of p40 were similarly expressed by both BM-NSCs and fetal neural stem cells after the adenoviral vectors transduction. To test their ability to delivery transgene in vivo, the BM-NSCs transduced with AdIL-23 were injected into tumor-bearing mice brain. The p40 expression was detected both in and around the tumor mass (Fig. 2C). To examine the phenotype of BM-NSCs after intracranial implantation, brain sections were double stained with anti-β-galactosidase or anti-p40 antibodies together with cell type–specific markers. The implanted cells were differentiated into positive staining of TuJ1, myelin/oligodendrocyte, or GFAP. There were no F4/80 (specific for macrophage)–positive cells noted (Fig. 3). Taken together, these data suggest that BM-NSCs have the ability to track glioma cells in vivo, to express transgenes both in vitro and in vivo after transduction with adenoviral vectors, and to differentiate into the neural cell types of the central nervous system.

Figure 2.

BM-NSCs displayed strong tropism for intracranial gliomas and were able to express transgenes after transduction with adenoviral vectors. A, BM-NSCs (red) distributed throughout the intracranial tumor mass (green) and intermingled among green tumor cells after peritumoral implantation into brain (a). BM-NSCs migrated away from the main tumor mass to a tumor satellite following intratumoral implantation (b), and BM-NSCs were also seen in the immediate proximity to and intermixed with an invading tumor island deep into normal neural tissue (c). The BM-NSCs were identified by immunohistochemistry as β-galactosidase-positive cells (red) due to AdLacZ vector transduction before the cells were implanted. Tumor cells were GFP stably transfected (green). Arrowheads, tumor edge; arrow, tumor island. Bar, 100 μm. B, BM-NSCs transduced by adenoviral vectors to express transgenes in vitro. BM-NSCs and fetal neural stem cells were transduced by AdEmpty (E), AdIL-23 (23), and AdIL-12 (12), respectively. BM-NSCs were capable of expressing transgene with a similar pattern as that of fetal neural stem cells, which were verified with either RT-PCR to amplify the p19 subunit of IL-23 (a) or ELISA to detect the p40 subunit of IL-23 and IL-12 (b). Columns, mean of triplicate determinations; bars, SD. ND, not detected. C, BM-NSC, as a tumor-targeting cellular vehicle, delivered transgene into intracranial tumor. After BM-NSCs were transduced with AdIL-23, when implanted into brain, the cells were able to deliver the transgene into glioma as identified by immunohistochemistry staining of p40 in both wild-type C57BL/6 mice (a and b) and athymic nude mice (c and d) intracranial tumor models. Isotype control antibody was stained negative for both wild-type mice (e and f) and athymic nude mice (g and h). Bar, 100 μm.

Figure 2.

BM-NSCs displayed strong tropism for intracranial gliomas and were able to express transgenes after transduction with adenoviral vectors. A, BM-NSCs (red) distributed throughout the intracranial tumor mass (green) and intermingled among green tumor cells after peritumoral implantation into brain (a). BM-NSCs migrated away from the main tumor mass to a tumor satellite following intratumoral implantation (b), and BM-NSCs were also seen in the immediate proximity to and intermixed with an invading tumor island deep into normal neural tissue (c). The BM-NSCs were identified by immunohistochemistry as β-galactosidase-positive cells (red) due to AdLacZ vector transduction before the cells were implanted. Tumor cells were GFP stably transfected (green). Arrowheads, tumor edge; arrow, tumor island. Bar, 100 μm. B, BM-NSCs transduced by adenoviral vectors to express transgenes in vitro. BM-NSCs and fetal neural stem cells were transduced by AdEmpty (E), AdIL-23 (23), and AdIL-12 (12), respectively. BM-NSCs were capable of expressing transgene with a similar pattern as that of fetal neural stem cells, which were verified with either RT-PCR to amplify the p19 subunit of IL-23 (a) or ELISA to detect the p40 subunit of IL-23 and IL-12 (b). Columns, mean of triplicate determinations; bars, SD. ND, not detected. C, BM-NSC, as a tumor-targeting cellular vehicle, delivered transgene into intracranial tumor. After BM-NSCs were transduced with AdIL-23, when implanted into brain, the cells were able to deliver the transgene into glioma as identified by immunohistochemistry staining of p40 in both wild-type C57BL/6 mice (a and b) and athymic nude mice (c and d) intracranial tumor models. Isotype control antibody was stained negative for both wild-type mice (e and f) and athymic nude mice (g and h). Bar, 100 μm.

Close modal
Figure 3.

In vivo differentiation of intracranial implantation of BM-NSCs. BM-NSCs transduced by AdIL-23 in vitro were intracranially implanted into mice bearing experimental brain tumor. The transplanted BM-NSCs (green) were identified within the brain (A, E, I, and M). The cells (red) showed neural differentiation as stained positive for Tuj1 (B-D), oligodendroglial differentiation as stained by myelin/oligodendrocyte-specific antibody (F-H), and astrocytic differentiation as identified by GFAP antibody (J-L). Macrophage marker-positive cell was not found in the implanted BM-NSC (N-P). The sections were labeled also with 4′,6-diamidino-2-phenylindole (blue) to identify nuclei (D, H, L, and P). Bar, 100 μm.

Figure 3.

In vivo differentiation of intracranial implantation of BM-NSCs. BM-NSCs transduced by AdIL-23 in vitro were intracranially implanted into mice bearing experimental brain tumor. The transplanted BM-NSCs (green) were identified within the brain (A, E, I, and M). The cells (red) showed neural differentiation as stained positive for Tuj1 (B-D), oligodendroglial differentiation as stained by myelin/oligodendrocyte-specific antibody (F-H), and astrocytic differentiation as identified by GFAP antibody (J-L). Macrophage marker-positive cell was not found in the implanted BM-NSC (N-P). The sections were labeled also with 4′,6-diamidino-2-phenylindole (blue) to identify nuclei (D, H, L, and P). Bar, 100 μm.

Close modal

Intracranial tumor-bearing C57BL/6 mice show prolonged survival after IL-23–expressing BM-NSC treatment. To determine whether the IL-23–expressing BM-NSCs provided a therapeutic benefit for brain tumor, we delivered BM-NSCs transduced with AdIL-23 or AdEmpty or delivered NIH-3T3 cells transduced with AdIL-23 into established intracranial gliomas in C57BL/6 mice. BM-NSC-IL-23–treated mice showed significantly prolonged survival compared with mice treated by BM-NSC-E (Fig. 4A). Approximately 60% of the BM-NSC-IL-23–treated mice survived beyond day 120 with tumor-free condition, which was verified by histology staining of the brain sections. Only 20% of the NIH-3T3-IL-23–treated mice survived. BM-NSC-E treatment was not protective and had the same “effect” as the saline control treatment. Statistical analysis revealed that the effects of both BM-NSC-IL-23 and NIH-3T3-IL-23 were significantly different from that of the BM-NSC-E (P < 0.001, BM-NSC-IL-23 versus BM-NSC-E; P = 0.0018, NIH-3T3-IL-23 versus BM-NSC-E, log rank). There was also a significant difference in effect between BM-NSC-IL-23 and NIH-3T3-IL-23 (P = 0.0310, log rank).

Figure 4.

Effect of IL-23–expressing BM-NSCs on the experimental glioma in C57BL/6 mice. A, Kaplan-Meier survival curve of intracranial glioma-bearing C57BL/6 mice that were injected intracranially with BM-NSC-IL-23, NIH-3T3-IL-23, BM-NSC-E, or saline. Representative of three independent experiments. B, intratumoral CD8+ and CD4+ T-cell infiltration 4 weeks after the transduced BM-NSCs were injected. BM-NSC-IL-23 treated mice showed robust infiltration of CD8+ (a and b) and CD4+ (d and e) T cells within the tumor mass. BM-NSC-E–treated mice showed negligible CD8+ (c) and CD4+ (f) T-cell infiltration. C, histology staining of glioma-bearing mice brains that were treated with either BM-NSC-IL-23 (a and d) or BM-NSC-E (b and e). Age-matched naive mice were included as control (c and f). Both H&E (a-c) and luxol fast blue (myelin stain) showed no evidence of demyelination (d-f). Bar, 100 μm.

Figure 4.

Effect of IL-23–expressing BM-NSCs on the experimental glioma in C57BL/6 mice. A, Kaplan-Meier survival curve of intracranial glioma-bearing C57BL/6 mice that were injected intracranially with BM-NSC-IL-23, NIH-3T3-IL-23, BM-NSC-E, or saline. Representative of three independent experiments. B, intratumoral CD8+ and CD4+ T-cell infiltration 4 weeks after the transduced BM-NSCs were injected. BM-NSC-IL-23 treated mice showed robust infiltration of CD8+ (a and b) and CD4+ (d and e) T cells within the tumor mass. BM-NSC-E–treated mice showed negligible CD8+ (c) and CD4+ (f) T-cell infiltration. C, histology staining of glioma-bearing mice brains that were treated with either BM-NSC-IL-23 (a and d) or BM-NSC-E (b and e). Age-matched naive mice were included as control (c and f). Both H&E (a-c) and luxol fast blue (myelin stain) showed no evidence of demyelination (d-f). Bar, 100 μm.

Close modal

Based on the known ability of IL-23 to induce proliferation of memory T cells, we sought to assess whether the BM-NSC-IL-23 treatment could induce enhanced intratumoral T-cell infiltration. Stained by immunohistochemistry, brain tumors from the mice treated with BM-NSC-IL-23 showed robust infiltration of CD8+ and CD4+ T cells (Fig. 4B,, a, b, d, and e). There was negligible CD8+ and CD4+ T-cell infiltration, however, within tumors from mice treated with BM-NSC-E (Fig. 4B,, c and f). The NIH-3T3-IL-23 treatment also resulted in intratumoral T-cell infiltration, but the intensity of that infiltration was significantly lower than that seen in the BM-NSC-IL-23–treated tumors (data not shown). Analyzing the brains that were treated with transduced BM-NSCs revealed that there was no morphologic difference between BM-NSC-IL-23– and BM-NSC-E–treated mice after either luxol fast blue staining, which was used to identify neural demyelination, or H&E staining. Both groups of treated mice showed a similar staining pattern of brain tissue compared with that of naive control mice (Fig. 4C). These data indicated that IL-23–expressing BM-NSCs have a strong protective effect against intracranial tumor and that there is no deleterious effect with morphology significance associated with IL-23–expressing BM-NSCs.

Involvement of tumor-specific CTL and CD8+ and CD4+ T cells in the antitumor activity of IL-23–expressing BM-NSCs. To examine whether a tumor-specific immunity was involved in the protective activity of IL-23–expressing BM-NSCs, mice that survived intracranial tumor implantation for 12 weeks after BM-NSC-IL-23 treatment and age-matched naive mice were subjected to CTL activity assay. CTL activity against parental tumor GL26 but not irrelevant tumor p815 was augmented in spleen cells, which were obtained from mice that survived intracranial tumor implantation after BM-NSC-IL-23 treatment and were restimulated in vitro with mitomycin C–treated GL26 for 5 days (Fig. 5A,, a). To determine whether IL-23 mainly contributed to this CTL activity, spleen cells from brain tumor-bearing mice treated with either BM-NSC-IL-23 or BM-NSC-E were analyzed 4 weeks after treatment. As shown in Fig. 5A, GL26-specific CTL activity was detected only in BM-NSC-IL-23–treated mice. To examine the involvement of particular lymphocytes in the BM-NSC-IL-23–induced antitumor activity, depleting anti-CD4 or anti-CD8 mAb were injected before and after BM-NSC-IL-23 treatment. Normal rat IgG at the same dose and schedule was included as control. Flow cytometry showed that the injection of mAb depleted the appropriate cell population by 95% (data not shown). Depletion of CD8+ T cells greatly impaired the protective effect of BM-NSC-IL-23 in intracranial tumor-bearing mice (P = 0.0010, CD8 depletion versus wild type, log rank). There was no long-term survival observed in the CD8+ T-cell–depleted mice (Fig. 5B). Similarly but less significantly, the impairment of the protective effect was also observed in the CD4+ T-cell–depleted mice (P = 0.0097, CD4 depletion versus wild type, log rank). When compared with BM-NSC-E, BM-NSC-IL-23 showed different protective activity on CD8+ versus CD4+ T-cell–depleted mice (Fig. 5B). In addition, on CD4+ T-cell knockout transgenic mice BM-NSC-IL-23 produced an effect comparable with that of produced on the CD4+ T-cell–depleted wild-type mice (data not shown). These data suggest that BM-NSC-mediated IL-23 delivery induces a potent tumor-specific protective immunity, that CD8+ T cells play critical role in the antitumor activity of BM-NSC-IL-23, and that CD4+ T cells are also involved in the process.

Figure 5.

Involvement of tumor-specific CTL activity and CD8+ and CD4+ T cells in the effect of BM-NSC-IL-23. A, tumor-specific CTL activity of spleen cells induced by BM-NSC-IL-23. Spleen cells obtained from mice that survived intracranial GL26 glioma implantation for 12 weeks after BM-NSC-IL-23 treatment (n = 3) or from age-matched naive mice (n = 3) were restimulated with mitomycin C–treated GL26 cells for 5 days, and the CTL activity of the resultant effector cells was measured against parental GL26 tumor target in a lactate dehydrogenase release assay. Irrelevant p815 was also used as a tumor target for the effector cells that were from BM-NSC-IL-23–treated mice (a). Spleen cells obtained from glioma-bearing mice that were intracranially injected with either BM-NSC-IL-23 (n = 3) or BM-NSC-E (n = 3) for 4 weeks were restimulated with mitomycin C–treated GL26 cells for 5 days, and the CTL activity of resultant effector cells was measured against parental GL26 tumor target in a lactate dehydrogenase release assay. Irrelevant p815 was also used as a tumor target for the effector cells that were from BM-NSC-IL-23–treated mice (b). Points, mean of three mice; bars, SD. Representative of three independent experiments. B, Kaplan-Meier survival curve of intracranial glioma-bearing C57BL/6 mice. Mice were randomly divided into four groups, in which one group was intracranially injected with BM-NSC-E plus i.p. injections of normal rat serum and the other three groups were intracranially injected with BM-NSC-IL-23 plus i.p. injections of either CD8+ T-cell depleting antibody, CD4+ T-cell depleting antibody, or normal rat serum, respectively. Representative of three independent experiments.

Figure 5.

Involvement of tumor-specific CTL activity and CD8+ and CD4+ T cells in the effect of BM-NSC-IL-23. A, tumor-specific CTL activity of spleen cells induced by BM-NSC-IL-23. Spleen cells obtained from mice that survived intracranial GL26 glioma implantation for 12 weeks after BM-NSC-IL-23 treatment (n = 3) or from age-matched naive mice (n = 3) were restimulated with mitomycin C–treated GL26 cells for 5 days, and the CTL activity of the resultant effector cells was measured against parental GL26 tumor target in a lactate dehydrogenase release assay. Irrelevant p815 was also used as a tumor target for the effector cells that were from BM-NSC-IL-23–treated mice (a). Spleen cells obtained from glioma-bearing mice that were intracranially injected with either BM-NSC-IL-23 (n = 3) or BM-NSC-E (n = 3) for 4 weeks were restimulated with mitomycin C–treated GL26 cells for 5 days, and the CTL activity of resultant effector cells was measured against parental GL26 tumor target in a lactate dehydrogenase release assay. Irrelevant p815 was also used as a tumor target for the effector cells that were from BM-NSC-IL-23–treated mice (b). Points, mean of three mice; bars, SD. Representative of three independent experiments. B, Kaplan-Meier survival curve of intracranial glioma-bearing C57BL/6 mice. Mice were randomly divided into four groups, in which one group was intracranially injected with BM-NSC-E plus i.p. injections of normal rat serum and the other three groups were intracranially injected with BM-NSC-IL-23 plus i.p. injections of either CD8+ T-cell depleting antibody, CD4+ T-cell depleting antibody, or normal rat serum, respectively. Representative of three independent experiments.

Close modal

NK cells are involved in the antitumor immunity induced by IL-23–expressing BM-NSCs. To examine the involvement of NK cells in the antitumor activity induced by BM-NSC-IL-23, the first series of experiments were carried out with athymic nude mice, in which T lymphocytes are absent. As shown in Fig. 6, no mice survived the intracranial GL26 glioma after either BM-NSC-IL-23 or BM-NSC-E treatment. However, there was a significant difference in survival time between the two treatments (P = 0.0195, log rank). A longer survival rate was observed in mice treated with BM-NSC-IL-23 compared with those treated with BM-NSC-E (Fig. 6A). Furthermore, after depleting NK cells in the athymic mice by injecting an anti–asialo GM1 antibody, no significant difference (P = 0.8801, log rank) in the survival rate between BM-NSC-IL-23 and BM-NSC-E treatments was observed (Fig. 6B). In C57BL/6 wild-type mice, anti–asialo GM1 antibody injection, which depleted NK cells in the wild-type animals, greatly impaired the protective effect of BM-NSC-IL-23 against intracranial tumor. Only 20% of the mice survived intracranial glioma implantation, but there was an ∼60% survival rate after normal rabbit serum injection, which served as control (Fig. 6C). The survival rate between the two groups was significantly different (P = 0.0457, log rank). These data suggest that NK cells are involved in the antitumor activity produced by IL-23–expressing BM-NSCs.

Figure 6.

Involvement of NK cells in the antitumor immunity induced by BM-NSC-IL-23. A, Kaplan-Meier survival curve of intracranial glioma-bearing athymic nude mice that were intracranially injected with either BM-NSC-IL-23 or BM-NSC-E. B, Kaplan-Meier survival curve of intracranial glioma-bearing athymic nude mice that were depleted of NK cells with an anti–asialo GM1 antibody injection and intracranially injected with either BM-NSC-IL-23 or BM-NSC-E. C, Kaplan-Meier survival curve of intracranial glioma-bearing C57BL/6 mice that were intracranially injected with BM-NSC-IL-23 plus i.p. injection of either anti–asialo GM1 antibody for NK cell depletion or normal rat serum as control. Representative of three independent experiments.

Figure 6.

Involvement of NK cells in the antitumor immunity induced by BM-NSC-IL-23. A, Kaplan-Meier survival curve of intracranial glioma-bearing athymic nude mice that were intracranially injected with either BM-NSC-IL-23 or BM-NSC-E. B, Kaplan-Meier survival curve of intracranial glioma-bearing athymic nude mice that were depleted of NK cells with an anti–asialo GM1 antibody injection and intracranially injected with either BM-NSC-IL-23 or BM-NSC-E. C, Kaplan-Meier survival curve of intracranial glioma-bearing C57BL/6 mice that were intracranially injected with BM-NSC-IL-23 plus i.p. injection of either anti–asialo GM1 antibody for NK cell depletion or normal rat serum as control. Representative of three independent experiments.

Close modal

IL-23–expressing BM-NSCs induce long-term immune memory and enhanced IFN-γ expression in brain. To examine whether a memory protective immunity was established in those animals that survived intracranial GL26 glioma after IL-23–expressing BM-NSC treatment, the surviving C57BL/6 mice were subjected to rechallenge with intracranial injection of parental GL26 glioma cells. All of the rechallenged animals survived beyond day 90 after the rechallenge and were tumor-free upon sacrifice as verified by H&E staining of the brain sections. In contrast, age-matched naive mice were uniformly susceptible to GL26 glioma challenge and died of brain tumor within 35 days of the challenge.

To study the possible molecules that may involve in the activity of IL-23–expressing BM-NSC in brain, we focused on IL-17 and IFN-γ, because these molecules have been shown to play important roles for the dendritic cell and T-cell activation states, which were promoted by IL-23 (14, 15, 21). The BM-NSC-IL-23–treated survivors of the earlier experiment were rechallenged with parental GL26 cells alone or GL26 with either BM-NSC-IL-23 or BM-NSC-E. After the rechallenge, brain tissues were analyzed by RT-PCR. We were consistently unable to detect any IL-17 expression. However, IFN-γ was detected from day 1 and reached the maximum level 5 to 7 days after the rechallenge (Fig. 7A). Then, the expression declined to an undetectable level 2 weeks after the rechallenge. The IFN-γ expression was always higher in the GL26 plus BM-NSC-IL-23 rechallenged animals than in the GL26 plus BM-NSC-E or the GL26 alone rechallenged animals. Age-matched naive animals did not have detectable IFN-γ expression in their brains after GL26 plus BM-NSC-IL-23 or GL26 alone challenges (Fig. 7B). To further examine the IFN-γ expression level, we used relative quantitative real-time PCR to analyze the brain samples 7 days after tumor rechallenge. As shown in Table 2, IFN-γ mRNA was up-regulated in all rechallenged mice that survived the earlier experiment compared with the age-matched naive mice that were challenged with GL26 plus BM-NSC-IL-23 or GL26 alone. We did not find IFN-γ mRNA up-regulation in mice that survived the earlier experiment but without rechallenge. These data suggest that IL-23–expressing BM-NSCs used to treat intracranial gliomas can induce long-term antitumor memory that is associated with enhanced IFN-γ, but not IL-17, up-regulation in the brain.

Figure 7.

RT-PCR assay of cytokine expression within brain after the rechallenges/challenges. A, IFN-γ expression was detected in the brain tissues of mice that survived earlier experiment and rechallenged with GL26 alone (lane 4) or GL26 with either BM-NSC-IL-23 (lane 2) or BM-NSC-E (lane 3). IL-17 expression was not detected whether GL26 alone (lane 8) or GL26 with either BM-NSC-IL-23 (lane 6) or BM-NSC-E (lane 7) was used in the rechallenges. Neither IFN-γ nor IL-17 expression was detected in age-matched naive mice after the GL26 plus BM-NSC-IL-23 challenge (lanes 5 and 9). B, IFN-γ expression profile in brain tissue of different animals and different rechallenges/challenges. IFN-γ expression was detected in mice that survived earlier experiments and rechallenged with GL26 alone (lanes 4 and 7) or GL26 with either BM-NSC-IL-23 (lanes 2 and 5) or BM-NSC-E (lanes 3 and 6). IFN-γ expression was not detected in age-matched naive mice that were challenged with GL26 plus BM-NSC-IL-23 (lane 8) or GL26 alone (lane 9). Brain tissues harvested on day 7 after the rechallenge/challenge. Representative of three independent experiments. L, 1 kb plus DNA ladder.

Figure 7.

RT-PCR assay of cytokine expression within brain after the rechallenges/challenges. A, IFN-γ expression was detected in the brain tissues of mice that survived earlier experiment and rechallenged with GL26 alone (lane 4) or GL26 with either BM-NSC-IL-23 (lane 2) or BM-NSC-E (lane 3). IL-17 expression was not detected whether GL26 alone (lane 8) or GL26 with either BM-NSC-IL-23 (lane 6) or BM-NSC-E (lane 7) was used in the rechallenges. Neither IFN-γ nor IL-17 expression was detected in age-matched naive mice after the GL26 plus BM-NSC-IL-23 challenge (lanes 5 and 9). B, IFN-γ expression profile in brain tissue of different animals and different rechallenges/challenges. IFN-γ expression was detected in mice that survived earlier experiments and rechallenged with GL26 alone (lanes 4 and 7) or GL26 with either BM-NSC-IL-23 (lanes 2 and 5) or BM-NSC-E (lanes 3 and 6). IFN-γ expression was not detected in age-matched naive mice that were challenged with GL26 plus BM-NSC-IL-23 (lane 8) or GL26 alone (lane 9). Brain tissues harvested on day 7 after the rechallenge/challenge. Representative of three independent experiments. L, 1 kb plus DNA ladder.

Close modal
Table 2.

Relative IFN-γ mRNA up-regulation in brain tissue

Survived mice
GL26 rechallengeGL26/BM-NSC-E rechallengeGL26/BM-NSC-IL-23 rechallengeNo rechallenge
Compared with naive mice + GL26 challenge 10.607 ± 1.177 10.733 ± 0.802 12.406 ± 1.049 1.132 ± 0.106 
Compared with naive mice + GL26/BM-NSC-IL-23 challenge 10.049 ± 1.231 10.168 ± 1.008 11.753 ± 1.061 1.079 ± 0.055 
Survived mice
GL26 rechallengeGL26/BM-NSC-E rechallengeGL26/BM-NSC-IL-23 rechallengeNo rechallenge
Compared with naive mice + GL26 challenge 10.607 ± 1.177 10.733 ± 0.802 12.406 ± 1.049 1.132 ± 0.106 
Compared with naive mice + GL26/BM-NSC-IL-23 challenge 10.049 ± 1.231 10.168 ± 1.008 11.753 ± 1.061 1.079 ± 0.055 

NOTE: Mean ± SD fold of increase. When 2−[ΔCt (survived animal) − ΔCt (naive animal)] = 1, there is no IFN-γ up-regulation.

In this study, we have presented evidence that neural stem-like cells can be isolated from whole bone marrow of adult mice. The mice BM-NSCs were able to be propagated and differentiated into neural and glial cells in vitro and were able to track intracranial glioma cells when implanted into glioma-bearing brain. More importantly, these cells displayed the ability to track invading tumor islands that had infiltrated deep into normal neural tissue well away from the main tumor mass. By incorporating the tumor tropic properties of NSCs with a novel cytokine, IL-23, which acts on memory T cells and dendritic cells (14, 15), we found that IL-23–expressing BM-NSCs when injected into brain showed antitumor activity against intracranial glioma. Mice that survived brain tumor implantation after BM-NSC-IL-23 treatment acquired a tumor-specific protective immunity. CD8+ T cells were crucial for this potent antitumor activity; in addition, CD4+ T cells and NK cells were also involved in the induction of this antitumor activity. Furthermore, enhanced IFN-γ production in the brain tissue was important for the BM-NSC-IL-23–elicited immune memory response against subsequent parental tumor rechallenge.

Effective eradication of primary or metastatic brain tumors and generation of a long-lasting immune response with an effective gene delivery system are important goals for cancer gene immunotherapy. Most gene delivery strategies employ viral vectors to deliver genes directly to tumor cells in vivo; however, the limitation of transgene distribution to extensive areas, especially the invading tumor foci, has limited the efficacy of such approaches. The present study sought to take advantage of the migratory ability of neural stem cells to track brain tumors and to deliver therapeutic molecules into expansive tumor regions. We generated neural stem-like cells from adult mice bone marrow. The BM-NSCs displayed the ability to track brain tumor in a mouse intracranial glioma model. After intratumoral or peritumoral implantation, the BM-NSCs extensively migrated throughout the established tumor mass. In addition, the BM-NSCs tracked into the invading tumor islands that had migrated away from the main tumor mass. This tracking characteristic has not been shown previously in bone marrow–derived neural progenitor-like cells (12).

By transducing the BM-NSCs with an adenoviral vector encoding a newly discovered cytokine, IL-23, we were able to deliver the molecule into intracranial tumor sites. It has just been shown that IL-23–expressing tumor cells produced antitumor activity (16). Delivering IL-23 by using BM-NSCs in the current study resulted in a protective effect in intracranial glioma-bearing mice with a significant prolongation of survival. These results strongly suggest that IL-23–expressing BM-NSCs produce antiglioma activity. To verify the potential role of IL-23 in this antiglioma activity, we compared the effect of IL-23–transduced BM-NSCs with that of an empty control vector transduced BM-NSCs. All of the intracranial glioma-bearing mice died within 40 days after treatment with the empty control vector transduced BM-NSCs. To further verify the potential benefit of the migration of BM-NSC in producing antiglioma activity, we compared the survival benefit offered by BM-NSC-IL-23 to that conferred by IL-23 secretion by nonmigratory NIH-3T3 cells, which produce similar levels of IL-23 to BM-NSC-IL23 in vitro and in vivo. NIH-3T3-IL-23 treatment produced significantly less protective effect compared with BM-NSC-IL-23. Taken together, the antiglioma activity of IL-23–expressing BM-NSCs may be a direct combined consequence of the ability of BM-NSCs to target migrating tumor cells and the ability of IL-23 to activate immune responses against tumor cells.

It has been shown that IL-23 can act directly on dendritic cells to promote tumor peptide presentation to T cells (15). In addition, T-cell responses may be amplified by the potent effect of IL-23 on memory-activated T cells (14). Our experiments in immunocompromised hosts and in animals selectively depleted of various lymphocyte populations suggest that T-cell immune responses are important for antitumor function of IL-23–expressing BM-NSCs and that CD8+ T cells play a crucial role in BM-NSC-IL-23–mediated antitumor activity, because the immune protective effect was greatly impaired in T-cell–deficient athymic nude mice and in wild-type mice depleted of CD8+ T cells. In CD4 knockout mice and in wild-type mice depleted of CD4+ T cells, we found that CD4+ T cells were also involved in the antitumor activity of BM-NSC-IL-23 albeit to a lesser extent than CD8+ T cells. In addition, in both athymic nude mice and wild-type mice, we observed the role of NK cells in the immune protective effect of BM-NSC-IL-23. NK cell depletion abolished and reduced the protective effect in athymic nude and wild-type mice, respectively. Lo et al. described the role of CD8+ T cells in the antitumor activity of IL-23–expressing tumor cells in a colon adenocarcinoma model (16). Our results confirmed that CD8+ T cells are crucial to the IL-23–mediated antitumor activity. However, the finding that CD4+ T cells and NK cells were involved in the IL-23 activity in our current study represents a significant divergence from the results forwarded by Lo et al. (16), in which IL-23–expressing tumor cells induced antitumor activity but did not require CD4+ T cells or NK cells. This may possibly be explained by their use of a different tumor cell line (CT26) in their colon adenocarcinoma s.c. tumor model, which is distinct from the GL26 glioma we used in our study as an intracranial brain tumor model. Alternatively, this difference may be attributable to the unique environment of the central nervous system with regard to antigen presentation by microglia or recruited dendritic cells. Interestingly, when we used B16-F10 cells as intracranial tumor model, we also observed the involvement of NK cells in the antitumor activity of IL-23 (data not shown). The variance in the cellular subsets crucial to an antitumor effect needs further delineation.

IL-23 can activate macrophages to produce proinflammatory cytokines that may contribute to autoimmune inflammation of the brain (2124). In the present study, we found that BM-NSC-IL-23–treated long-term survivors showed enhanced IFN-γ expression in brain tissue upon tumor rechallenge. We were consistently unable, however, to detect IL-17, a cytokine involved in the IL-23 signaling pathway and which induces inflammation (21, 23, 24), even when we delivered IL-23 with the tumor rechallenge. These findings were consistent with the results that elucidated that mice brains maintained normal morphologic characteristics after BM-NSC-IL-23 treatment as shown by H&E and demyelination-specific staining. These data suggest that BM-NSC-IL-23 treatment used in the current study does not induce detectable inflammation in the brain either because the IL-23 level delivered by BM-NSCs is not high enough to do so or because the in vivo expression of exogenous IL-23 predominantly enhances long-lasting memory T-cell immunity, such as antigen-specific CTL and IFN-γ–producing Th1 immune responses (25). To understand this, it will be required to closely investigate the signal pathways that are involved in inflammation and in protective immune responses induced by exogenous IL-23 expression.

In summary, in this study, we showed that neural stem-like cells generated from adult bone marrow can be used as a targeting vehicle to track migratory and invasive tumor cells within the central nervous system. In combination with the unique action of IL-23 on tumoricidal potency, autologous neural stem-like cell–mediated tumor targeting immune therapy represents an attractive new treatment modality for malignant brain tumors.

Note: X. Yuan and J. Hu contributed equally to this work.

Grant support: NIH grants 1K23 NS02232, 1R01 NS048959, and 1R21 NS048879 (J.S. Yu).

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.

We thank Drs. Sebastian Wachsmann-Hogiu and Daniel H. Farkas for their kind assistance in obtaining images of stained brain sections, Dr. Scot Macdonald for the critical review of the article and helpful comments, and Dr. Hong Zhou for discussions.

1
Maher EA, Furnari FB, Bachoo RM, et al. Malignant glioma: genetics and biology of a grave matter.
Genes Dev
2001
;
15
:
1311
–33.
2
Peter McL, Loeffler JS. Cancer of the nervous system. Cambridge (MA): Blackwell Publishers; 1997.
3
Kramm CM, Sena-Esteves M, Barnett FH, et al. Gene therapy for brain tumors.
Brain Pathol
1995
;
5
:
345
–81.
4
Benedetti S, Pirola B, Pollo B, et al. Gene therapy of experimental brain tumors using neural progenitor cells.
Nat Med
2000
;
6
:
447
–50.
5
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
2000
;
97
:
12846
–51.
6
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 glioma.
Cancer Res
2002
;
62
:
5657
–63.
7
Mezey E, Chandross KJ, Harta G, Maki RA, McKercher SR. Turning blood into brain: cells bearing neuronal antigens generated in vivo from bone marrow.
Science
2000
;
290
:
1779
–82.
8
Brazelton TR, Rossi FM, Keshet GI, Blau HM. From marrow to brain: expression of neuronal phenotypes in adult mice.
Science
2000
;
290
:
1775
–9.
9
Woodbury D, Schwarz EJ, Prockop DJ, Black IB. Adult rat and human bone marrow stromal cells differentiate into neurons.
J Neurosci Res
2000
;
61
:
364
–70.
10
Sanchez-Ramos J, Song S, Cardozo-Pelaez F, et al. Adult bone marrow stromal cells differentiate into neural cells in vitro.
Exp Neurol
2000
;
164
:
247
–56.
11
Kabos P, Ehtesham M, Kabosova A, Black KL, Yu JS. Generation of neural progenitor cells from whole adult bone marrow.
Exp Neurol
2002
;
178
:
288
–93.
12
Lee J, Elkahloun AG, Messina SA, et al. Cellular and genetic characterization of human adult bone marrow-derived neural stem-like cells: a potential antiglioma cellular vector.
Cancer Res
2003
;
63
:
8877
–89.
13
Nakamizo A, Marini F, Amano T, et al. Human bone marrow-derived mesenchymal stem cells in the treatment of gliomas.
Cancer Res
2005
;
65
:
3307
–18.
14
Oppmann B, Lesley R, Blom B, et al. Novel p19 protein engages IL-12p40 to form a cytokine, IL-23, with biological activities similar as well as distinct from IL-12.
Immunity
2000
;
13
:
715
–25.
15
Belladonna ML, Renauld JC, Bianchi R, et al. IL-23 and IL-12 have overlapping, but distinct, effects on murine dendritic cells.
J Immunol
2002
;
168
:
5448
–54.
16
Lo CH, Lee SC, Wu PY, et al. Antitumor and antimetastatic activity of IL-23.
J Immunol
2003
;
171
:
600
–7.
17
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
2000
;
20
:
1085
–95.
18
Graham FL, van der Eb AJ. A new technique for the assay of infectivity of human adenovirus 5 DNA.
Virology
1973
;
52
:
456
–67.
19
Yuan X, Curtin J, Xiong Y, et al. Isolation of cancer stem cells from adult glioblastoma multiforme.
Oncogene
2004
;
23
:
9392
–400.
20
Reynolds BA, Weiss S. Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system.
Science
1992
;
255
:
1707
–10.
21
Aggarwal S, Ghilardi N, Xie MH, de Sauvage FJ, Gurney AL. Interleukin-23 promotes a distinct CD4 T cell activation state characterized by the production of interleukin-17.
J Biol Chem
2003
;
278
:
1910
–4.
22
Cua DJ, Sherlock J, Chen Y, et al. Interleukin-23 rather than interleukin-12 is the critical cytokine for autoimmune inflammation of the brain.
Nature
2003
;
421
:
744
–8.
23
Langrish CL, McKenzie BS, Wilson NJ, de Waal Malefyt R, Kastelein RA, Cua DJ. IL-12 and IL-23: master regulators of innate and adaptive immunity.
Immunol Rev
2004
;
202
:
96
–105.
24
Langrish CL, Chen Y, Blumenschein WM, et al. IL-23 drives a pathogenic T cell population that induces autoimmune inflammation.
J Exp Med
2005
;
201
:
233
–40.
25
Ha SJ, Kim DJ, Baek KH, Yun YD, Sung YC. IL-23 induces stronger sustained CTL and Th1 immune responses than IL-12 in hepatitis C virus envelope protein 2 DNA immunization.
J Immunol
2004
;
172
:
525
–31.