Original Article | Published:

The anti-glioma effect of suicide gene therapy using BMSC expressing HSV/TK combined with overexpression of Cx43 in glioma cells

Cancer Gene Therapy volume 17, pages 192202 (2010) | Download Citation


The disseminated neoplastic foci of malignant gliomas are essentially responsible for the limited efficacy of current available therapeutic modalities. Bone marrow-derived stem cells (BMSCs) have the ability to migrate into these tumors and even track infiltrating tumor cells, making them to be promising cellular vehicles for delivering therapeutic agents to glioma cells. The herpes simplex virus thymidine kinase (HSV–TK)/ganciclovir (GCV) suicide gene therapy with a potent bystander effect has been considered as one of the most promising therapeutic strategies for malignant gliomas. In this study, we evaluate the anti-glioma effect of suicide gene therapy using BMSCs expressing HSV–TK combined with overexpression of connexin 43 (Cx43), which can restore the gap junction of intercellular communication and may enhance the bystander effect of suicide gene therapy. To assess the potential of BMSCs to track glioma cells, a spheroid co-culture system in matrigel was used to show that some BMSCs migrated to C6 glioma cell microspheres. Transwell assay showed the tumor tropic property of BMSCs. In addition, BrdU-labeled BMSCs injected directly into the cerebral hemisphere opposite to the established C6 rat gliomas were capable of migrating into the xenograft gliomas. C6 cell growth was more intensively inhibited by HSV–TK/GCV treatment mediated by BMSCs, and could be further enhanced by combination with Cx43 transfection into glioma cells. The same result was observed in vivo by the growth of C6 gliomas and the survival analysis of rats bearing C6 glioma. In conclusion, Cx43 combined with HSV–TK/GCV gene therapy using BMSCs as vehicles was highly effective in a rat glioma model and therefore hold great potential as a novel approach for the gene therapy of human malignant gliomas.


Gliomas are highly infiltrative neoplasms, with solitary tumor cells or clusters of neoplastic cells migrating extensively throughout the brain. Despite aggressive surgical intervention, radiation therapy and chemotherapy, it is still not possible to eliminate successfully all the tumor foci interspersed within normal brain parenchyma. The inability of current therapies to improve outcome for patients with high-grade glioma has spurred the search for alternative treatment strategies, including the delivery of novel tumor-toxic molecules/agents into the tumor bed, molecular targeted therapies and the use of immunostimulatory therapies to induce endogenous immune responses against the tumor. However, none of these approaches has directly focused on the key issue of targeting tumor satellites. It is the disseminated neoplastic foci that form the obstacle to current therapeutic approaches, because they serve as a nidus for tumor recurrence, resulting in the failure of predestine standard treatment protocols. Therefore, targeting these tumor satellites is critical for the successful therapeutic strategy of gliomas.

Owing to the chemotaxic characteristics of stem cells, neural stem cells (NSCs) and bone marrow-derived stem cells (BMSCs) have been used as vectors for the immunogene therapy of gliomas in recent years.1, 2, 3 BMSCs are well suited for clinical application because they are easily obtained from patients themselves and the autologous transplantation is possible to avoid immunologic incompatibilities. Of the various progenitor cells that exist in bone marrow, BMSCs are particularly attractive for clinical use because they can be easily isolated and expanded in culture, and genetically manipulated using currently available molecular techniques.4, 5, 6 BMSCs have been genetically modified to express interferon-β with potent tropism for disseminating glioma cells and show strong anti-tumor effects both in vitro and in vivo.3

Herpes simplex virus thymidine kinase (HSV–TK)/ganciclovir (GCV) suicide gene therapy has been considered as one of the most promising therapeutic strategies for malignant gliomas. Theorectically, it can be used for various malignant tumors no matter what changes of molecular pathology they have. HSV–TK/GCV viral-directed enzyme prodrug gene therapy causes potent, tumor-selective cytotoxicity. The passage of toxic molecules from HSV–TK+ to neighboring HSV–TK cells during GCV therapy is an important mechanism that may account for this ‘bystander’ cytotoxicity.7, 8 As the extracellular virus moves by passive diffusion or fluid convection, it is not possible for the virus to spread far from the site of injection. In contrast, glioma cells are known to migrate a long distance along the white matter beyond the therapeutic radius.

In this study, we showed that BMSCs have a tropism for human gliomas after local delivery and that this tropism can be exploited therapeutically by engineering BMSCs with HSV–TK gene to release a soluble anti-glioma factor after the treatment with GCV. Furthermore, the ‘bystander’ therapeutic effect of BMSC–TK for glioma cells can be enhanced by restoring the gap junction of intercellular communication through simultaneous overexpression of connexin 43 (Cx43) in glioma cells both in vitro and in vivo.

Materials and methods

Mesenchymal stem cell isolation, culture, identification and transfection with HSV–TK

Rat BMSCs were isolated from long bones of male Sprague–Dawley (SD) rats using the methods described by Peister et al.9 Briefly, cells from each long bone were plated in 40 ml of complete isolation medium (low sugar Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum). After 24 h of incubation, nonadherent cells were removed and adherent cells were washed with phosphate buffered saline (PBS), and fresh complete isolation medium was replaced every 3 to 4 days for 4 weeks. Cells were collected by trypsinization and replated in 30 ml of complete isolation medium in 175 cm2 flasks. After incubation for 1 to 2 weeks, cells were trypsinized and plated in expansion medium (DMEM containing 10% fetal bovine serum). As cells were subcultured to six passages, adherent bone marrow-derived cells were uniformly shown fibroblast-like appearance. Cells at sixth passage were either frozen or further expanded.

For identification of BMSCs, cell surface antigen expression was analyzed by flow cytometry on a FACS Calibur flow cytometer (Becton–Dickinson, Franklin Lakes, NJ). 1 × 106 cells were suspended in 100 μl of 10% PBS, stained with fluorescein isothiocyanate-labeled specific antibodies for 30 min (1:100 dilution) against CD29, CD31, CD34, CD45, CD71 and CD90. CD29 and CD34 antibodies were purchased from Serotec, Oxford, OX5, 1GE, UK. CD31 and CD71 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). CD45 and CD90 antibodies were purchased from Biolegend (San Diego, CA). Cells were washed with PBS, resuspended in 600 μl of 1% paraformaldehyde and analyzed with FACSCalibur. For transfection with HSV–TK, BMSCs were incubated with Ad–CMV–TK (provided by Institute of Life Science, Nankai University, Tianjin, China) at multiplicity of infection 100 for 24 h. Cells were washed and prepared for use. TK expression was detected by reverse transcriptase-PCR (RT-PCR) analysis as described below.

Reverse transcriptase-PCR analysis

The total RNA of BMSC–TK was extracted using TRIzol reagents (Invitrogen, Carlsbad, CA). Isolated RNA was electrophoresed on a 1% agarose-formaldehyde gel to verify the quality of RNA. PCR amplification was performed using a Perkin–Elmer DNA thermal cycler (PTC-200, a type of PCR Perkin–Elmer, Waltham, MA). The sequence of PCR primers were: TK, 5′-CGATGACTTACTGGCAGGTG-3′ and 3′-TGGGAGTAGAAGCTGGCG-5′ β-actin, 5′-TCCCTGGAGAAGAGCTACGA-3′ and 3′-GATCCACACGGAGTACTTGC-5′. The following PCR conditions were used: initial denaturation at 94 °C for 2 min, then 94 °C for 30 s, 54 °C for 30 s and 72 °C for 30 s for 35 cycles; after the final cycle, the reaction was held in 72 °C for 5 min. β-actin was used as a loading control. PCR products were analyzed by electrophoresis on a 2% gel containing 0.1 mg ml–1 of ethidium bromide. Relative quantification was calculated as the density of TK band divided by the density of β-actin.

Cell lines

Two human glioblastoma cell lines, TJ899 and TJ905, were established in our laboratory and their biological characteristics have been identified earlier.10 Rat C6 glioma and murine fibroblast NIH 3T3 cell lines were kindly provided by Dr Hao Jiang (National Institutes of Health). Human glioblastoma U251 and U87 cell lines were obtained from Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. All these cell lines were cultured in DMEM supplemented with 10% heat-inactivated fetal calf serum, 4 mM glutamine, 50 units ml–1 penicillin and 50 μg ml–1 streptomycin and subcultured every 2–3 days.

Establishment of rat glioma model

Rat C6 glioma model was set up as described earlier.11 Briefly, male SD rats (250–300 g) were anesthetized by intraperitoneal injection of chloral hydrate (3.5 ml kg–1) and were placed in a stereotactic apparatus. Using a Hamilton syringe (Sigma-Aldrich Inc., Shanghai, China), C6 glioma cells (1 × 106 cells in 25 μl of serum free DMEM) were injected through a burr hole into the right caudate nucleus.

Evaluation of tumor tropic property of bone marrow-derived stem cells

The tumor tropism of BMSCs was detected using transwell and spheroids co-culture assays in vitro, and the tumor-homing property of BMSCs in vivo was observed by the migratory activity of BrdU-labeling BMSCs implanted in the opposite side of xenograft tumors moving to the tumor bed and its surrounding tumor satellites.

Transwell invasion assay in vitro

Transwell filters (Costar, a brand of transwell filter, Corning Incorporated-Life Sciences, Lowell, MA) in the upper chambers were coated with matrigel (3.9 μg μl–1, 60–80 μl) on the upper surface of polycarbonic membrane (diameter 6.5 mm and pore size 8 μm). After incubation at 37 °C for 30 min as matrigel solidified, BMSCs or control cells (1 × 105) suspended in 200 μl of serum-free medium were added to the upper chamber, conditional medium of TJ899, TJ905, U251, U87, C6, NIH 3T3 cells and DMEM as well as DMEM supplemented with 10% fetal calf serum were placed into the lower chamber as chemoattractant. After 24 h of incubation at 37 °C with 5% CO2, the medium was removed from the upper chamber. The non-invaded cells on the upper surface of the inserted filter were gently scraped off with a wet cotton swab. The cells that had invaded to the lower surface of the filter were fixed with 4% paraformaldehyde and stained with hematoxylin. The migrated cells were counted by light microscopy (200 × magnification) and the average number of cells of at least five fields from each well was calculated.

Spheroid co-culture assay

Each well of a six-well plate was precoated with 1 ml of 0.75% agar. After incubation at 37 °C for 30 min, C6 cells and BMSCs in the log phase of growth were seeded on the surface of agar (2 × 104 cells per well). Cell spheroids at 60–80 μm diameters were prepared for the experiment. Each well of a 24-well dish was precoated with 200 μl of 1:2 DMEM diluted phenol-red-free matrigel (BD Biosciences, San Jose, CA). After gelling, the BMSC spheroids were labeled with Hoechst33258 for 15 min and C6 spheroids with similar diameter were laid over the bottom layer, 100 μl of serum-free medium was then added and covered with another 200 μl of 1:2 DMEM diluted Matrigel. The co-culture spheroids were then incubated for additional 24 h. The interaction between two cell spheroids was observed at 12 and 24 h after incubation using an IX inverted phase-contrast microscope (Olympus, Tokyo, Japan).

The migration assay of bone marrow-derived stem cells in vivo

The ability of BMSCs to migrate toward gliomas was assessed in vivo by implanting C6 glioma cells into the right caudate nucleus of six SD rats as described above; seven days later, BrdU-labeled BMSCs were injected into left cerebral hemisphere contralateral to the established tumors. Briefly, monolayer cultures of BMSCs were labeled with BrdU for 12–24 h, and then trypsinized and suspended in 10% PBS at a density of 1 × 106 BrdU-BMSCs/10 μl. In all, 10 μl of BrdU–BMSCs cell suspension were stereotactically injected into the contra-lateral brain of tumor-bearing SD rats. Injections were carried out manually over 3 to 5 min. After injection, every two rats were killed on day 2, 5 and 7, respectively. The brain tissues were removed and frozen sections were prepared. Migration of BMSCs was detected by immunofluorescent staining with fluorescein isothiocyanate -labeled secondary antibody, or immunohistochemical staining and hematoxylin and eosin staining. Migration toward the tumor was assessed by direct visualization using optical or fluorescent microscopy.

Scrape loading and dye transfer assay of C6 cells before and after transfection with Cx43

For transfection, 2 × 105 cells were plated into each well of six-well plates and grown overnight until they were 50–80% confluent. The plasmid with cytomegalovirus promoter (pCMV)–Cx43 complementary DNA plasmids (kindly provided by Dr David Kiang, University of Minnesota) (2 μg for each well) were transfected into C6 cells by lipofectamine. Cells were subcultured at a 1:5 dilution in G418-containing (1000 μg ml–1) medium. Stable transfectants (Cx43–C6) were selected and expanded for 4 weeks. Cx43 expression was detected by immunohistochemical staining.

C6 and Cx43–C6 cells were plated in the 35-mm dishes and grown to confluency. After rinsing with PBS, monolayer cells were immersed in 0.05% of Lucifer Yellow (MW 457.2, Sigma-Aldrich Inc., Shanghai, China) in PBS. Scrape loading was performed using a sharp knife to draw several clear straight lines on the culture dishes. Cells were incubated in dye solution for additional 3 min at room temperature, then washed with PBS and observed under inverted fluorescence microscope. Cells competent in gap junction intercellular communication (GJIC) showed transfer of Lucifer Yellow to neighboring cells from the border of scraped line, while cells incompetent in GJIC did not show dye transfer and Lucifer Yellow was retained in the original loaded cells at the injured border.

In vitro effect of treatment BMSC–TK/GCV combined with Cx43 on glioma cell growth

To determine the most efficient dose of GCV to BMSC–TK, BMSCs and BMSC–TK were plated in 24-well plates (1 × 105 cells per well). When they were 50–80% confluent, various concentrations of GCV (10−3–102 μg ml–1) were added in each well. After incubation for 96 h, viable and nonviable cells were counted using a hemocytometer after Trypan blue staining. All experiments were carried out in triplicate.

To determine the best mixed ratio between BMSC–TK and C6/Cx43–C6, the various percentage of BMSC–TK (0, 10 or 100%) was co-cultured with C6/Cx43–C6 in DMEM medium supplemented with 10% fetal calf serum containing 1 μg ml–1 GCV in 24-well plates for 96 h. Viable and nonviable cells were counted using a hemocytometer after Trypan blue staining. All experiments were carried out in triplicate.

Growth inhibition of BMSC–TK/GCV combined with Cx43 was evaluated using a 3-(4,5-dimethylthiazol- 2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma) assay. Briefly, C6, Cx43–C6 and mixtures of BMSCs/C6 (1:1), BMSCs/Cx43–C6 (1:1), BMSC–TK/C6/GCV (1:1, 1 μg ml–1), BMSC–TK/Cx43–C6/GCV (1:1, 1 μg ml–1) were seeded in a 96-well plate (4 × 103 per well). On each day of consecutive 7 days, 20 μl of MTT (5 mg l–1) was added into each well and cells were incubated for additional 4 h. After the medium containing MTT was removed, the formazan crystals were dissolved in 200 μl of dimethyl sulfoxide for 5 min. The quantification measurement (optical density) was recorded using a Teacan 96-well spectrophotometer at a wavelength of 570 nm with a wavelength of 630 nm as the reference. Eight samples of each cell group were tested and C6 cell group was served as control.

Therapeutic efficacy of BMSC–TK/GCV combined with overexpression Cx43 in rat glioma model

SD rats were divided into four groups (eight rats per group): (1) C6 group: parental C6 cells were implanted without any treatment; (2) Cx43–C6 group: Cx43–C6 cells were implanted; (3) BMSC–TK/C6 group: on day 5 after the implantation of C6 cells, BMSC–TK (5 × 106 cells/25 μl) were injected directly into the tumor; 3 days later GCV was administrated intraperitoneally at 15 mg kg–1 per 48 h for 14 days; (4) BMSC–TK/Cx43–C6 group: Cx43–C6 cells were used instead of C6 cells as was the case in BMSC–TK/C6 group.

The general behavior and the survival of the rats in each group, the findings on magnetic resonance imaging and apoptosis in the developing gliomas were observed. Tumor volume of three rats selected randomly in each group was measured at regular intervals by high-resolution magnetic resonance imaging performed as reported earlier.11 This procedure has the advantage of allowing a comparison of tumor size at different periods in the same animal. The histopathological changes were examined using paraffin-embedded sections stained with hematoxylin and eosin. Cell apoptosis was detected on frozen sections by TUNEL method using an in situ cell death kit (Roche, Basel, Switzerland).

Statistical methods

All data were analyzed with SPSS 10.0 statistics software (SPSS Inc., Chicago, Illinois). The results were presented as mean±s.e. Average results were evaluated by one-way analysis of variance and Bonferroni t-test. For efficacy experiments, differences in survival among groups were determined by a log-rank test.


Phenotypic characterization of rat bone marrow-derived stem cells

Rat BMSCs was isolated from the bone marrow of adult SD rats. Cells appeared a typical spindle shape. The expression of cell-surface markers was identified by flow cytometry. BMSCs did not express CD31, CD34 and CD45, whereas CD29, CD71 and CD90 were strongly expressed. On the basis of the available criteria, the cells used in our experiments had the properties of BMSCs as described earlier.12

Capability of migration and tumor-homing property of rat bone marrow-derived stem cells in vitro and in vivo

In vitro transwell assay were performed to determine the tropism of BMSCs toward gliomas. We compared the capability of various glioma cell lines to stimulate the migration of BMSCs. BMSCs exposure to cell-free medium or to conditioned medium from fibroblasts resulted in a few migrated BMSCs. However, exposure to conditioned medium from various glioma cells showed significant migration of BMSCs (P<0.05). In addition, the number of migrated BMSCs stimulated by conditioned medium from C6 and U251 cells was higher than that from TJ905 cells, but significantly lower than those from TJ899 and U87 cells (P<0.05). These data indicated that BMSCs might have different tropism for various gliomas (Figures 1a and b).

Figure 1
Figure 1

Tumor tropism of bone marrow-derived stem cell (BMSC) detected by in vitro transwell assay and spheroid co-culture assay. (a) Photomicrograph of transwell assay. Cells staining in black were cells crossed microporous membrane from the upper to low well. (b) The column figure of percentage of cell migration. (c) Fluorescent photomicrograph of three-dimensional cell growth on matrigel showing the migratory ability of BMSCs toward glioma at 12 and 24 h. Fewer BMSCs labeled by Hoechst (blue mixed green showed in green fluorescent passage) migrated to C6 glioma cell sphere at 12 h, and many more BMSCs migrated at 24 h.

Spheroid co-culture system in matrigel was also used to explore the tropism of BMSCs toward glioma. We observed under the fluorescence microscope that some BMSCs migrated to C6 glioma cell sphere. However, fewer C6 glioma cells migrated to BMSCs spheroid after co-culture of both cell spheroids for 24 h (Figure 1c).

To determine the glioma-homing ability of BMSCs in vivo, BMSCs pre-labeled with BrdU were injected directly into the contralateral hemispheres of established gliomas. At day 2, a small number of BMSCs migrated to the opposite side near the tumor border (Figure 2g). At day 5, more BMSCs moved to the tumor border, even tracked and sieged tumor cells infiltrating the normal tissue, and engrafted a fraction of the solid tumor mass (Figures 2a–f, h). At day 7, a plenty of BMSCs engrafted the tumor mass, but less BMSCs were observed in the far lateral side of injection site (Figure 2i). These results showed that BMSCs possess tumor-homing property and are capable of tracking the glioma as well as engrafting the cells disseminated in the surrounding normal brain parenchyma.

Figure 2
Figure 2

Immunohistochemical and immunofluoresence staining of whole brain section showed the tumor tropism of bone marrow-derived stem cells (BMSCs) in vivo. (a) BrdU-labeled BMSCs were injected into the left caudate nucleus just opposite to the xenograft glioma in the right cerebral hemisphere. (b) Immunohistochemical staining of BrdU-labeled BMSCs migrated and engrafted to established C6 glioma in right hemisphere on day 5. (c) and (d) High-magnification view of area within the rectangle indicated in (b). BMSCs moved to the tumor, and mainly distributed in the boundary between C6 glioma and normal brain tissue, even tracked and sieged the disseminated neoplastic foci. (e) Local magnification of (d) showed that BMSCs sieged the disseminated neoplastic foci. (f) BMSCs did not appear at the far- lateral side of tumor. (gi) Immunofluorescent staining of BrdU-labeled BMSCs (green color) migrated to established C6 glioma in opposite hemisphere on day 2 (g), day 5 (h) and day 7 (i). A small number of BMSCs migrated to the tumor border on day 2. More BMSCs migrated to the tumor border and engrafted the tumor mass on day 5 and 7.

Overexpression of Cx43 in glioma cells enhances the GJIC

Scrape loading and dye transfer method showed that control C6 cells were poorly dye coupled indicating the blockage of GJIC, whereas the GJIC function was significantly restored in glioma cells transfected with Cx43 complementary DNA. The Lucifer Yellow was transmitted to neighboring cells from the loading cells on the injured scraping border (Figure 3a).

Figure 3
Figure 3

Scrape loading and dye transfer (SLDT) assay and reverse transcriptase-PCR (RT-PCR) analysis of herpes simplex virus thymidine kinase (HSV–TK) gene expression in bone marrow-derived stem cells (BMSC). (a) Gap junction intercellular communication (GJIC) of C6 and connexin 43 (Cx43)–C6 glioma cells, Lucifer Yellow (LY) dye was restricted in dye loaded C6 cells at the border of scraped line. For Cx43–C6 cells, LY dye was transferred from loaded cells at the border of scraped line to the neighboring cells. (b) Reverse transcriptase-PCR (RT-PCR) analysis of HSV–TK gene expression in BMSCs showing that the highest expression occurred at 48–72 h after transfection, and higher expression were remained at twentieth day.

Inhibitory effect of BMSC–TK/GCV combined with Cx43 overexpression on glioma cell growth in vitro

Bone marrow-derived stem cells were transduced with an adenoviral vector expressing the HSV–TK suicide gene. Reverse transcriptase-PCR analysis showed that TK gene was expressed in a time-dependent manner. The highest level of TK expression occurred at 48–72 h after transfection and higher expression level was still maintained 20 days (Figure 3b). Cx43 expression in Cx43–C6 cells was detected by immunohistochemical staining (data not shown) and the restoring of gap junction in C6 cells was shown above.

To determine the most effective dose of GCV for BMSC–TK, BMSCs and BMSC–TK were incubated in the medium containing various concentrations of GCV (from 10−3 to 102 μg ml–1) for 96 h. More than 90% of BMSC–TK was killed at 1 μg ml–1 of GCV. However, no more than 1% of BMSCs was killed at the same concentration. Therefore, 1 μg ml–1 of GCV was used in the subsequent study (Figure 4a).

Figure 4
Figure 4

Therapeutic potential of bone marrow-derived stem cells– thymidine kinase (BMSC–TK) combined with connexin 43 (Cx43) overexpression in vitro. (a) Dose-response curves of ganciclovir (GCV) at 96 h showed that more than 90% of BMSC–TK was killed at 1 μg ml–1 GCV compared with no more than 1% of BMSCs killed. (b) Cell mortality curves of various mixed ratio of BMSC–TK to C6/Cx43–C6 exposed to GCV (1 μg ml–1) at 96 h showed that the most significant bystander effect was observed at the mixed ratio of 1:1. (c) MTT assay measuring cell survival rate.

To determine the best mixed ratio of BMSC–TK to C6/Cx43–C6, the various percentage of BMSC–TK was co-cultured with C6/Cx43–C6 in DMEM medium containing 1 μg ml–1 GCV for 96 h. When the mixed ratio of BMSC–TK to C6/Cx43–C6 was 1:1, the favorable bystander effect was observed with 82 and 100% cell mortality rate, respectively. Meanwhile, the bystander effect was significantly enhanced by Cx43 using the same ratio (Figure 4b).

To evaluate the effect of combination treatment of BMSC–TK/GCV and Cx43 on cell growth inhibition, six cell groups were tested, including C6, Cx43–C6 and mixtures of BMSCs/C6 (1:1), BMSCs/Cx43–C6 (1:1), BMSC–TK/C6/GCV (1:1, 1 μg ml–1) and BMSC–TK/Cx43–C6/GCV (1:1, 1 μg ml–1). MTT assay showed that there was no significant difference in cell survival rate between C6 and BMSCs/C6, or Cx43–C6 and BMSCs/Cx43–C6 at any time point tested (P>0.05). However, the cell survival rate of Cx43–C6 and BMSCs/Cx43–C6 was significantly lower than that of C6 and BMSCs/C6 after day 4 (P<0.05) (Figure 4c). These data suggest that Cx43 can inhibit the growth of C6 cells, but BMSCs has no inhibitory effect on C6 cells and Cx43–C6 cells.

The cell survival rate of BMSC–TK/C6/GCV and BMSC–TK/Cx43–C6/GCV was markedly lowered as compared with other four groups after day 2 (P<0.05). Furthermore, the lowered survival was more significant in BMSC–TK/Cx43–C6/GCV than that in BMSC–TK/C6/GCV at day 2–4 (P<0.05) and nearly all the cells in BMSC–TK/Cx43–C6/GCV group were killed at day 7 (Figure 4c). Taken together, the results show that TK/GCV mediated by BMSCs has more intensive inhibition of C6 cells than Cx43, and Cx43 can further enhance the C6 cell growth inhibition by TK/GCV mediated by BMSCs in vitro.

Therapeutic potential of BMSC–TK/ GCV combined with Cx43 in vivo

To analyze the therapeutic effect of BMSC–TK/GCV combined with Cx43 in vivo, we set up two control groups (n=8), C6 and Cx43–C6 (1 × 106) cells were injected into the right caudate nucleus of SD rats, respectively, and two treated groups (n=8), C6 and Cx43–C6 cells were implanted as was the case with control group rats, on day 5 after implantation, BMSC–TK (5 × 106 cells/25 μl) were injected to the tumor site by using the same coordinates with stereotactic guidance, From day 3 after BMSC–TK injection, BMSC–TK/C6 and BMSC–TK/Cx43–C6 group rats received intraperitoneal injections of GCV at 15 mg kg–1 per 48 h for 14 days, whereas two control group rats were not treated with GCV. The observation period for the glioma models was ended on day 120 after implantation. The mean survival time of two control groups (C6 and Cx43–C6) was 15.88±2.36 days and 41.63±11.78 days, respectively. Cx43–C6 group rats had a longer survival time than rats of C6 group (P<0.01). In BMSC–TK/C6 group, six out of eight rats died at day 45, 84, 96, 98, 107 and 112 after tumor implantation, respectively, and two out of eight rats survived more than 120 days. In BMSC–TK/Cx43–C6 group, four out of eight rats died at day 88, 99, 100 and 113 after tumor implantation, respectively, and four out of eight rats survived more than 120 days. Thus, rats in two treated groups showed a prolonged survival compared with both control groups that was statistically highly significant (P<0.001), but also, rats in BMSC–TK/Cx43–C6 group showed a significantly prolonged survival compared with BMSC–TK/C6 group (P<0.05) (Figure 5A).

Figure 5
Figure 5

Therapeutic potential of bone marrow-derived stem cells– thymidine kinase (BMSC–TK) combined with connexin 43 (Cx43) overexpression in vivo. (A) Kaplan–Meier survival plots of tumor bearing rats showed that in total 50% of rats survived over 120 days in BMSC–TK/Cx43–C6 treated group and 25% of rats survived over 120 days in BMSC–TK/C6 treated group. The survival of rats in BMSC–TK/Cx43–C6 group was significantly longer than that of rats in BMSC–TK/C6 group (P<0.05; log-rank test). The difference in survival between the rats in two treated groups and control groups was also statistically significant (P<0.001; log-rank test). (B) The tumor volume of C6 group was significantly larger at 2 weeks compared with other three groups (P<0.01). At 6 weeks, the tumor volume of BMSC–TK/C6 group was smaller than that of Cx43–C6 group and larger than that of BMSC–TK/Cx43–C6 group significantly (P<0.05). (C) Enhanced magnetic resonance imaging (MRI) coronal scanning of tumors. (D) Cell apoptosis detected by TUNEL method in vivo. It was shown that there were nearly no apoptotic cells (green) in C6 and Cx43–C6 groups. However, an increase of apoptotic cells was observed in BMSC–TK/C6 and BMSC–TK/Cx43–C6 groups, especially in BMSC–TK/Cx43–C6 group. (a) C6 group; (b) Cx43–C6 group; (c) BMSC–TK/C6 group; (d) BMSC–TK/Cx43–C6 group.

C6 gliomas grew large enough to be detected by enhanced magnetic resonance imaging at 1 week after cell implantation. The tumor volume (three out of nine rats) was calculated by enhanced magnetic resonance imaging at 1, 2 and 6 weeks after cell implantation. The rats in C6 group developed a significantly larger tumor at 2 weeks compared with other three groups (P<0.01). At 6 weeks, the tumor volume of BMSC–TK/C6 group was smaller than that of Cx43–C6 group and significantly larger than that of BMSC–TK/Cx43–C6 group (P<0.05). The tumor volume of Cx43–C6 group at 6 weeks was even smaller than that of C6 group at 2 weeks. Unexpectedly, the tumor of a rat in BMSC–TK/Cx43–C6 group disappeared at 6 weeks and the rat survived more than 120 days (Figures 5B and C). TUNEL staining showed that there were nearly no apoptotic cells found in tumors of C6 and Cx43–C6 group rats. On the contrary, a great deal of apoptotic cells was observed in tumors of BMSC–TK/C6 and BMSC–TK/Cx43–C6 group rats, especially in BMSC–TK/Cx43–C6 group (Figure 5D). These data are similar with the in vitro observations showing that Cx43 further enhances the inhibition of C6 tumor growth by TK/GCV mediated by BMSCs.


The tumor-homing property of BMSCs renders them as attractive cell vehicles for the delivery of therapeutic molecules to tumor cells. In this study, BMSCs from long bones of SD rats are successfully isolated, cultured and identified. The significant tumor tropism of BMSCs to glioma cells has been observed both in vitro and in vivo. Transwell assay showed that BMSCs displayed different tumor tropism and migration ability toward four different human glioblastoma cell lines and rat C6 glioma cells. Similarly, Heese et al.13 found that conditioned media from 10 different human glioblastoma cell lines significantly stimulated the migration of C17.2 NSCs. However, the degree of stimulation was varied by cell lines, ranging from 2.5-fold to more than 14-fold.

Our study in vivo showed that BrdU-labeled BMSCs from the injection site in the left cerebral hemisphere can migrate to the border of the tumor in the right cerebral hemisphere and engraft solid tumor mass, indicating the strong ability of BMSCs to track glioma cells. The migratory activity of NSCs toward the site of intracranial tumor has also been shown by several studies.13, 14, 15 However, NSCs have a big obstacle for clinical application, because NSCs can only be isolated from fetal or adult brain or from embryonic stem cells, resulting in the limited use of autologous NSCs because of significant ethical issues. BMSCs are attractive as a cell vehicle, because they can be harvested without difficulty, processed efficiently in vitro, and then reinoculated into the same patient. Several studies have shown that mesenchymal stem cells have ability to track and envelop infiltrating glioma cells like NSCs, as described in this study.3, 4, 5, 6 Therefore, BMSCs might be a rational substitute for NSCs as promising therapeutic gene delivery vehicles for tumor because of a good tropism for malignant glioma.

The HSV–TK/GCV suicide gene therapy has been considered to be highly effective for cancer therapy. A bystander effect, meaning HSV–TK non-expressing cells near the HSV–TK expressing cells can also be killed, is mainly responsible for a dramatic tumor ablation or regression by HSV–TK/GCV treatment in animal studies.7, 8, 16, 17, 18 However, clinical trials for glioma using this treatment have failed.19, 20, 21, 22 One of the major reasons may be due to the delivery manner of HSV–TK into glioma cells located only in the direct vicinity of the injection site, but fail to transduce infiltrating tumor cells or distant tumor areas that eventually serve as the reservoirs for tumor recurrence. In this study, we used BMSCs as HSV–TK gene vector that show an innate tumor-homing capability. When BMSC–TK and C6 were co-cultured in the presence of 1 μg ml–1 GCV for 96 h, the favorable bystander effect was observed. An 82% of cell mortality rate was achieved, and nearly 90% of cells were killed at day 7. In the in vivo experiment, rats in BMSC–TK/C6 group had a significantly longer survival time than rats in C6 group, and two out of eight rats survived more than 120 days. Similar results were reported by several groups showing that a bystander effect is the main mechanism of HSV–TK/GCV therapy mediated by stem cells in glioma. Li et al.22 showed a potent bystander effect between NSC–TK and C6 cells, and complete tumor eradication was obtained by NSC–TK therapy when the same number of NSC–TK and the tumor cells were intracranially implanted. Uhl et al.23 reported that the admixture of HSV–TK-transduced NSC to U87 MG and LN-18 human malignant glioma cell lines at ratios of 1:10 or 1:1 eliminated more than 50 or 90% of glioma cells in the presence of GCV (25 μM). Meanwhile, they showed that the presence of this bystander effect is correlated with the expression of Cx43 in untransduced glioma target cells. Furthermore, Miletic et al.24 confirmed the gap junction communication between bone marrow (BM) and tumor-infiltrating cells (TICs) (a highly proliferative subpopulation of bone marrow-derived mesenchymal stem cells) and 9L cells. Their cell culture experiments showed that BM–TIC–TK cells exert a strong bystander effect on 9L glioma cells on GCV treatment. In vivo, more than 60% of BM–TIC–TK plus GCV-treated animals were long-term survivors. The anti-tumor effect of HSV–TK/GCV mediated by BMSCs in our study is similar to that in the above study. However, the GCV doses we used for in vivo study is lower than that reported earlier (15 mg kg–1 per 48 h compared with 30 mg kg–1 per 24 h in the above study), and that may have some affect on the antiglioma effect. Benedetti and Staflin et al. described the ability of nonengineered rat neural progenitors to inhibit glioma growth in vivo.25, 26 The exact mechanism underlying the observations remains unclear, although they have reported that NSCs may elaborate a secretory agent that could inhibit tumor cell growth. For verifying this demonstration, we co-cultured BMSCs and C6 at the ratio of 1:1. The result shows that there was no significant difference in cell survival rate between C6 and BMSCs/C6, so we conclude that BMSCs has no inhibitory effect on C6 cells at least in vitro.

Our study showed that C6 cells that were deficient in endogenous Cx43 expression and transfected with Cx43 complementary DNA were proliferated at lower level, and the gap junction communication was restored in these transfected cells.27, 28 In view of the importance of the bystander effect on the treatment of HSV–TK/GCV mediated by stem cells (NSCs or BMSCs) for glioma, we evaluate the effect of introducing Cx43 into HSV–TK/GCV gene therapy mediated by BMSCs. This study indicates that BMSCs mediated TK/GCV therapy has more intensive inhibition of C6 cell growth than transfection with Cx43 alone. However, transfection of Cx43 into C6 cells further enhanced the inhibitory effect of TK/GCV mediated by BMSCs in vitro. Our in vivo study also showed that Cx43 overexpression combined with HSV–TK/GCV mediated by BMSCs can significantly prolong the survival time of rats bearing C6 glioma and 50% of rats bearing tumor survive over 120 days, as compared with the survival of rats bearing tumor solely treated with BMSC–TK/GCV. At 6 weeks, the tumor volume of BMSC–TK/C6 group was smaller than that of Cx43–C6 group and larger than that of BMSC–TK/Cx43–C6 group significantly (P<0.05).

It has been well known that the bystander effect has two major mechanisms: one involves the transfer of phosphorylated GCV through intercellular gap junction; and the other involves uptake of phosphorylated GCV by non-TK expressing cells through apoptotic vesicles produced by HSV–TK expressing cells.14, 15, 16, 17, 18 Owing to the loss of gap junction among C6 glioma cells, it is speculated that the second mechanism may have a major role in which, to a great extent, death of the neighboring C6 cells is induced by their uptake of cytotoxic GCV triphosphate released from BMSC–TK cells after GCV treatment. As transfection of Cx43 gene into C6 cells can restore gap junction,27, 28 cytotoxic phosphorylated GCV can be transported between C6 cells through gap junction, in addition to the uptake of phosphorylated GCV through apoptotic vesicles. As a result, the bystander effect of HSV–TK/GCV mediated by BMSCs is enhanced by the transfection of Cx43. Furthermore, it has been shown that Cx43 can suppress the tumor cell growth by itself.29, 30, 31 Therefore, co-treatment with BMSC–TK/GCV and Cx43 may have a synergistic effect on the suppression of glioma cell growth, especially for the surrounding infiltrated glioma cells.

In conjunction with earlier work showing that Cx43 expression is significantly downregulated in high-grade astrocytic tumors,32, 33, 34, 35 the clinical accessibility, potential for in vitro expansion and genetic modification, migratory activity and tumor tropism of BMSCs, it is conceivable that combination of Cx43 overexpression and HSV–TK/GCV mediated by BMSCs may be a potential and promising novel approach for the treatment of malignant gliomas.

Conflict of interest

The authors declare no conflict of interest.


  1. 1.

    , , , , , et al. Induction of glioblastoma apoptosis using neural stem cell-mediated delivery of tumor necrosis factor-related apoptosis-inducing ligand. Cancer Res 2002; 62: 7170–7174.

  2. 2.

    , , , , , . The use of interleukin 12-secreting neural stem cells for the treatment of intracranial glioma. Cancer Res 2002; 62: 5657–5663.

  3. 3.

    , , , , , , et al. Human bone marrowderived mesenchymal stem cells in the treatment of gliomas. Cancer Res 2005; 65: 3307–3318.

  4. 4.

    , , , . Bone marrow stromal stem cells: nature, biology, and potential applications. Stem Cells 2001; 19: 180–192.

  5. 5.

    , , , . Rapid expansion of recycling stem cells in cultures of plastic-adherent cells from human bone marrow. Proc Natl Acad Sci USA 2000; 97: 3213–3218.

  6. 6.

    , . Mesenchymal stem cells: heading into the clinic. Bone Marrow Transplant 2001; 27: 235–239.

  7. 7.

    , , , , , et al. Herpes simplex virus thymidine kinase gene therapy for rat malignant brain tumors. Hum Gene Ther 1996; 7: 197–205.

  8. 8.

    , , , , , et al. In vitro and in vivo antitumor effect of retrovirus-mediated herpes simplex thymidine kinase gene-transfer in human medulloblastoma. Gene Ther 1998; 5: 113–120.

  9. 9.

    , , , , , . Adult stem cells from bone marrow (MSCs) isolated from different strains of in bred mice vary in surface epitopes, rates of proliferation, and differential potential. Blood 2004; 103: 1662–1668.

  10. 10.

    , , , , , . The establishment and characterization of human glioblastoma cell lines TJ899 and TJ905. Tianjin Med J 1996; 24: 416–418. (in Chinese).

  11. 11.

    , , , , . Inhibitory effect of antisense epidermal growth factor receptor RNA on the proliferation of rat C6 glioma cells in vitro and in vivo. J. Neurosurg 2000; 92: 132–139.

  12. 12.

    , . Phenotypical and functional properties of human bone marrow mesenchymal progenitor cells. J Cell Physiol 1999; 181: 67–73.

  13. 13.

    , , , , . Migration of human neural stem cells toward an intracranial glioma. Neuro Oncol 2005; 7: 476–484.

  14. 14.

    , , , , . Migration of human neural stem cells toward an intracranial glioma. Exp Mol Med 2008; 40: 84–91.

  15. 15.

    , , , , , et al. Neural stem cells display extensive tropism for pathology in adult brain: evidence from intracranial gliomas. Proc Natl Acad Sci USA 2000; 97: 12846–12851.

  16. 16.

    , , , . The bystander effect in the HSVtk/ganciclovir system and its relation to gap junctional communication. Gene Ther 1998; 5: 1705–1711.

  17. 17.

    , , , . Enhancement of the herpes simplex virus thymidine kinase/ ganciclovir bystander effect and its antitumor efficacy in vivo by pharmacologic manipulation of gap junctions. Hum Gene Ther 1998; 9: 2385–2391.

  18. 18.

    , , , . In vitro evidence that metabolic cooperation is responsible for the bystander effect observed with HSV-TK retroviral gene therapy. Hum Gene Ther 1993; 4: 725–731.

  19. 19.

    , , , , , et al. AdvHSV-tk gene therapy with intravenous ganciclovir improves survival in human malignant glioma: a randomised, controlled study. Mol Ther 2004; 10: 967–972.

  20. 20.

    . Phase III clinical evaluation of herpes simplex virus type 1 thymidine kinase and ganciclovir gene therapy as an adjuvant to surgical resection and radiation in adults with previously untreated glioblastoma multiforme. Hum Gene Ther 2000; 11: 2389–2401.

  21. 21.

    , , , , , et al. Phase I study of adenoviral delivery of the HSV-tk gene and ganciclovir administration in patients with current malignant brain tumors. Mol Ther 2000; 1: 195–203.

  22. 22.

    , , , , , . Bystander effect-mediated gene therapy of gliomas using genetically engineered neural stem cells. Cancer Gene Ther 2005; 12: 600–607.

  23. 23.

    , , , , , . Migratory neural stem cells for improved thymidine kinase-based gene therapy of malignant gliomas. Biochem Biophys Res Commun 2005; 328: 125–129.

  24. 24.

    , , , , , et al. Bystander killing of malignant glioma by bone marrow-derived tumor-infiltrating progenitor cells expressing a suicide gene. Mol Ther 2007; 15: 1373–1381.

  25. 25.

    , , , , , et al. Gene therapy of experimental brain tumors using neural progenitor cells. Nat Med 2000; 6: 447–450.

  26. 26.

    , , , , , . Neural progenitor cell lines inhibit rat tumor growth in vivo. Cancer Res 2004; 64: 5347–5354.

  27. 27.

    , , , , . The effect of transfected Cx43 gene on the GJIC and proliferation of glioma cells. Chin J Cancer Res 2002; 14: 100–104. (in English).

  28. 28.

    , , , , , . Connexin 43 gene for the treatment of rat C6 cerebral gliomas in vivo. Chin J Oncol 2002; 24: 212–214. (in Chinese).

  29. 29.

    , , , , , . Reversion of the neoplastic phenotype of human glioblastoma cells by Connexin43 (Cx43). Cancer Res 1998; 58: 5089–5096.

  30. 30.

    , , , , . In vivo growth of C6 glioma cells transfected with connexin43 cDNA. Cancer Res 1992; 52: 4208–4213.

  31. 31.

    , , , . Connexin 43 reverses the phenotype of transformed cells and alters their expression of cyclin/cyclin-dependent kinases. Cell Growth Differ 1995; 6: 681–690.

  32. 32.

    , , , . Altered expression of Cx43 in astrocytic tumors. Clin Neurol Neurosurg 2004; 107: 49–54.

  33. 33.

    , , , , , et al. Protein and messenger RNA expression of connexin 43 in astrocytomas: implications in brain tumor gene therapy. J Neurosurg 1996; 84: 839–846.

  34. 34.

    , , . Reduced connexin 43 expression in high grade human brain glioma cells. J Surg Oncol 1999; 70: 21–24.

  35. 35.

    , , . Reduced expression of connexin-43 and functional gap junction coupling in human gliomas. Glia 2001; 33: 107–117.

Download references


This work was supported by grants from Tianjin Education Committee (Grant No. 20050227) and Tianjin Science and Technology Committee (Grant No. 09JCZDJC20500).

Author information

Author notes

    • Q Huang
    •  & X-Z Liu

    These authors contributed equally to this work.


  1. Department of Neurosurgery, Tianjin Medical University General Hospital, Tianjin, People's Republic of China

    • Q Huang
    • , C-S Kang
    • , Y Zhong
    •  & P-Y Pu
  2. Laboratory of Neuro-Oncology, Tianjin Neurological Institute, Tianjin, People's Republic of China

    • X-Z Liu
    • , C-S Kang
    •  & G-X Wang


  1. Search for Q Huang in:

  2. Search for X-Z Liu in:

  3. Search for C-S Kang in:

  4. Search for G-X Wang in:

  5. Search for Y Zhong in:

  6. Search for P-Y Pu in:

Corresponding authors

Correspondence to Y Zhong or P-Y Pu.

About this article

Publication history







Further reading