Purpose: Lentiviral vectors pseudotyped with glycoproteins of the lymphocytic choriomeningitis virus (LCMV-GP) are promising candidates for gene therapy of malignant glioma, as they specifically and efficiently transduce glioma cells in vitro and in vivo. Here, we evaluated the therapeutic efficacy of LCMV-GP and vesicular stomatitis virus glycoprotein (VSV-G) pseudotyped vectors.
Experimental Design: Therapeutic efficacy was tested for unmodified (9L) and DsRed-modified (9LDsRed) gliomas using the suicide gene thymidine kinase of the herpes simplex virus type 1 (HSV-1-tk). Positron emission tomography (PET) and magnetic resonance imaging were done to analyze transduction of tumors and monitor therapeutic outcome.
Results: LCMV-GP pseudotypes mediated a successful eradication of 9LDsRed tumors with 100% of long-term survivors. Before initiation of ganciclovir treatment, a strong HSV-1-tk expression within the tumor was detected by noninvasive PET using the tracer 9-[4-[18F]fluoro-3-(hydroxymethyl)butyl]guanine. Therapeutic outcome was successfully monitored by magnetic resonance imaging and PET imaging and correlated with the histopathologic data. In the 9L model, LCMV-GP and VSV-G pseudotyped lentiviral vectors displayed similar therapeutic efficacy. Further studies revealed that normal brain cells transduced with VSV-G pseudotypes were not eliminated by ganciclovir treatment and contributed significantly to the bystander killing of tumor cells.
Conclusions: Suicide gene transfer using pseudotyped lentiviral vectors was very effective in the treatment of rat glioma and therefore is an attractive therapeutic strategy also in human glioblastoma especially in conjunction with an imaging-guided approach. In addition, high selectivity of gene transfer to tumor cells may not always be desirable for therapeutic genes that exert a clear bystander effect.
Malignant gliomas, the most frequent primary brain tumors, still have a dismal prognosis despite advances in neurosurgery, radiation, and chemotherapy. Gene therapy using viral vectors to transduce tumor cells with therapeutic genes is an attractive alternative to conventional therapies. Gammaretroviral vectors derived from the murine leukemia virus have thus far been the most frequently used retroviruses for gene therapy of brain tumors (1–4). Exchange of the retroviral envelope protein by the rhabdoviral G protein of the vesicular stomatitis virus (VSV-G) even improved vector stability and broadened the host range (5–7). Use of these pseudotyped gammaretroviral vectors in a rat glioma model showed a high transduction and therapeutic efficiency following gene transfer of the herpes simplex virus type 1 thymidine kinase (HSV-1-tk) and ganciclovir treatment (8). Despite these promising results in animal models, clinical trials using retroviral vector supernatants or retroviral packaging cells failed (9–11).
One major problem may be the restricted gammaretroviral transduction of dividing cells, as in human glioma the majority of tumor cells will not divide within a given treatment window. Therefore, a vector system is required for transduction of both dividing as well as nondividing tumor cells. In this regard, lentiviral vectors are attractive candidates. However, lentiviral vectors pseudotyped with VSV-G transduced neurons and astrocytes at high levels in the normal rat brain as well as in the infiltrating area of a rat glioma model (12, 13). Another disadvantage of the VSV-G protein is its cytotoxicity, thus preventing establishment of stable recombinant packaging cell lines (7, 14).
In previous studies, we have developed gammaretroviral and lentiviral vectors pseudotyped with the glycoproteins of the lymphocytic choriomeningitis virus (LCMV-GP; refs. 15, 16). These vectors have a broad host range and can be concentrated by ultracentrifugation. In addition, LCMV-GP is not cytotoxic, and stable recombinant packaging cell lines can be established (17). In a recent study, lentiviral LCMV-GP pseudotypes were found to efficiently and specifically transduce glioma cells in vivo, whereas neurons were not transduced (12). In the current approach, we analyzed the therapeutic efficacy of LCMV-GP and VSV-G pseudotyped lentiviral vectors for intracranial rat glioma using the suicide gene HSV-1-tk. LCMV-GP pseudotypes mediated a successful eradication of 9LDsRed glioma, which was monitored by positron emission tomography (PET) and magnetic resonance imaging (MRI). Strong HSV-1-tk expression in the transduced tumor cells was detected by PET imaging using the specific tracer 9-[4-[18F]fluoro-3-(hydroxymethyl)butyl]guanine ([18F]FHBG). Unexpectedly, the VSV-G lentiviral pseudotypes also showed a significant therapeutic efficacy similar to that of LCMV pseudotypes, with 50% of long-term survivors in the 9L model. Transduction of normal brain cells with VSV-G pseudotypes before tumor implantation and subsequent ganciclovir application was also therapeutically effective, which outlines the potent contribution of gene-modified normal brain cells to the bystander killing of tumor cells.
Materials and Methods
Cell lines. The human embryonic kidney cell line 293T and the 9L cell line were obtained from the American Type Culture Collection and maintained in DMEM supplemented with 10% FCS and 1% glutamine. All cell lines were grown at 37°C in a humidified atmosphere of 5% CO2.
Vector construction. The lentiviral vector plasmid pRRL.sinCMVeGFPpre was published by Naldini et al. (18). The HSV-1-tk39–containing plasmid was obtained from Prof. M. Grez (Georg-Speyer-Haus, Frankfurt, Germany). The cloning plasmid pEGFP-N1 was obtained from BD Clontech and the TopoZeroBlunt cloning kit was purchased from Invitrogen.
The construction of the lentiviral vector containing the HSV-1-tk39 gene fused to a gfp gene was done as follows: The HSV-1-tk39 gene was amplified by PCR using the primers TTAAAGATCTGCCGCCATGGCTTCGTACCCCTGCC (HSV-1-tk forward primer) and TTAAACCGGTGGTAAGTTAGCCTCCCCCATCTCC (HSV-1-tk reverse primer). The primers were designed to mutate the stop codon of the HSV-1-tk gene and to insert AgeI and BglII restriction sites flanking the HSV-1-tk gene. The PCR product was ligated into the TopoZeroBlunt cloning vector. The HSV-1-tk gene was then excised by AgeI and BglII digestion and ligated into AgeI and BglII sites of pEGFP-N1 in frame with the enhanced green fluorescent protein (egfp) gene. The resulting HSV-1-tk/egfp fusion gene encoded eGFP fused to the COOH terminus of HSV-1-tk. The HSV-1-tk/egfp fusion gene was then ligated into BamHI and HincII sites of pRRL.sinCMVeGFPpre to replace the egfp gene present in the lentiviral vector. The final lentiviral vector was termed pRRL.sinCMV-TK/eGFPpre.
Preparation of lentiviral vector supernatants. The 293T cell line was used for transient lentiviral vector production. The lentiviral vector plasmid pRRL.sinCMV-TK/eGFPpre (5 μg), the HIV gag-pol-REV expression plasmid pCMV-dR8.91 (12.5 μg; ref. 18), and 2 μg of the envelope expression plasmid pHCMV-LCMV-GP (15) or pCMV-G (6) were cotransfected into 293T cells and concentrated as described previously (12).
Titration of lentiviral vector supernatants. Lentiviral vector titers were measured by transduction of TE671 cells as described previously (12).
Transduction of 9L cells with DsRed. 9L cells were transduced as described previously (12).
Killing of 9L cells in vitro by lentiviral expression of HSV-1-tk. 9L cells were seeded into six-well plates at 2 × 105 per well and transduced with 2 to 4 mL of unconcentrated lentiviral vectors pseudotyped with LCMV-GP carrying pRRL.sinCMV-tk/eGFPpre. The amount of vector supernatant added was chosen to achieve at least 80% transduction efficacy. After adding lentiviral vectors, the plates were centrifuged for 1 h at 2,000 rpm, 31°C. The expression of pRRL.sinCMV-TK/eGFPpre in 9L-tk cells was quantified 2 days after transduction by flow cytometry. 9L-tk cells were then mixed with native 9L cells to obtain cultures with 80%, 50%, or 10% HSV-1-tk–positive cells. The mixtures were seeded into 96-well plates at 1 × 104 cells per well in DMEM/10% FCS. After 4-h incubation at 37°C, ganciclovir was added to the wells in increasing concentrations in duplicates. Cells were incubated for 3 days, after which the cell viability was measured using a WST cell proliferation kit (Roche). The WST test was done according to the manufacturer's instructions.
Implantation of 9L and vector supernatants. Adult female Fisher 344 rats (Harlan Winkelmann) were anesthetized by i.p. injection of ketamine (50 mg/kg) and xylazine (2 mg/kg). Intracranial 9LDsRed or 9L tumors were established by injection of 8 × 104 9LDsRed cells or 7 × 104 9L cells (in 2 μL PBS) into the right striatum, respectively, using a Hamilton syringe in a stereotactic apparatus (Stoelting). The coordinates used were 4 mm lateral to the bregma and 5 mm in depth to the dural surface. Four days after or 5 days before tumor implantation, rats were anesthetized, and lentiviral vector pseudotypes with titers of 1 × 107 transducing units/mL were injected using the same stereotactic coordinates and 1 mm apart (seven different sites). A total volume of 10 μL was injected into each tumor.
Treatment of rat gliomas. Fischer rats bearing 9LDsRed or 9L tumors were treated by daily i.p. injections of 30 mg/kg ganciclovir.
Analysis of rat brains. Animals were euthanized and perfused with 4% paraformaldehyde. Brains were removed, suspended in 30% sucrose for 3 days, and then snap frozen in isopentane chilled with dry ice. Coronal sections (12 μm) were prepared on a cryostat. For immunofluorescence analysis, sections were stained with rabbit anti-glial fibrillary acidic protein antibodies (Dako) for astrocytes, mouse anti-NeuN (Chemicon) antibodies for neurons, mouse anti-nestin antibodies (Chemicon) for progenitor cells, or rabbit anti-caspase-3 antibodies (Becton Dickinson) for apoptotic cells. Primary antibodies were incubated overnight at 4°C. Cy3-goat anti-mouse and Cy3-goat anti-rabbit (Dianova) were used as secondary antibodies for 2 h at room temperature. The sections were examined under a fluorescence microscope (Zeiss) and analyzed by confocal scanning laser microscopy (Leica).
Radiosynthesis of methyl-[11C]-l-methionine and [18F]FHBG. The radiotracer for tumor detection, methyl-[11C]-l-methionine ([11C]MET), and the specific marker substrate for TKGFP, [18F]FHBG, were produced as described previously (19–21).
Positron emission tomography. PET imaging was done using a microPET (63 image planes; 2.0 mm full width at half maximum; Concorde Microsystems, Inc.). Radiotracer was given i.v. (tail vein) into experimental animals with the following doses: no-carrier-added [11C]MET: 670 to 1,111 μCi/rat; no-carrier-added [18F]FHBG: 390 to 480 μCi/rat. Emission scans (duration, 30 min) were obtained starting 10 min ([11C]MET) and >120 min ([18F]FHBG) after tracer application. Images were reconstructed using Fourier rebinning and two-dimensional filtered backprojection (microPET). For quantification of images, a reference standard sample of radiotracer was placed within the field of view of the PET scanner. Image coregistration and data evaluation based on a region-of-interest analysis of PET images to determine maximal radioactivity concentrations within tumors were done by the use of VINCI, a newly developed fast graphical image analysis tool (VINCI; ref. 22). After background (contralateral hemisphere) subtraction, data were decay corrected and divided by the total injected dose to represent percentage injected dose per gram (%ID/g).
Localization of tumor before and during gene therapy was done by MRI and [11C]MET PET. The respective PET images were coregistered to MR images as described (19, 21).
Magnetic resonance imaging. Animals were imaged after administration of ∼1 mL/g body weight gadolinium-diethylenetriaminepentaacetic acid (500 mmol/L; Magnevist, Schering AG). Animals were anesthetized using 1% halothane in O2:N2O (35:65%). Temperature was maintained at 37°C and monitored throughout the MRI measurements. All MR images were acquired using a Bruker BioSpin 4.7 Tesla small animal MR scanner (horizontal bore, 30 cm; Bruker BioSpin) equipped with actively shielded gradients (200 mT m−1). Purpose-built radiofrequency coils were used (12-cm Helmholtz coils for excitation and a 3-cm surface coil for detection). Following two-dimensional localization scans, three-dimensional gradient-echo scans (FLASH) with T1 weighting were acquired. The acquisition variables were as follows: spatial resolution, 121 × 121 × 242 μm3; field of view, 31 × 31 × 15.5 mm3; repetition time, 70 ms; echo time, 5 ms; flip angle, 60°. Images were processed using ParaVision 3.0.2 (Bruker BioSpin), NIH ImageJ, and, for coregistration with PET data, VINCI (MPI Cologne).
Statistical analysis. Survival was analyzed by a log-rank test based on the Kaplan-Meier test using Statistical Package for the Social Sciences software. Differences between pairs of groups were determined by the Student's t test. P values of <0.05 were considered significant.
Bystander-mediated killing of 9L cells in vitro by the HSV-1-tk/ganciclovir paradigm. The suicide gene HSV-1-tk is used as therapeutic gene for lentiviral vector therapy throughout the study. To verify the activity of HSV-1-tk in 9L glioma cells in vitro, 9L were transduced with the HSV-1-tk suicide gene (9L-tk; see Materials and Methods). The bystander killing effect of 9L-tk on wild-type 9L glioma cells was examined by coculture experiments. After addition of ganciclovir to the medium, 9L-tk exerted a strong and dose-dependent bystander effect on 9L glioma cells (Fig. 1). Killing of 100% of glioma cells was achieved in cultures containing even 10% 9L-tk, which was statistically significant with ganciclovir concentrations from 10−1 to 102 μmol/L (Fig. 1).
Therapeutic efficiency of LCMV-GP pseudotyped lentiviral vectors monitored by PET and MRI. In a previous study, we showed that LCMV-GP pseudotyped lentiviral vectors specifically and efficiently transduce 9LDsRed glioma in vivo (12). To confirm a therapeutic effect of these vectors, we injected LCMV pseudotypes carrying the suicide gene HSV-1-tk fused to eGFP into established intracranial 9LDsRed glioma 5 days after tumor implantation and did a therapeutic follow-up by PET and MRI. Noninvasive detection of HSV-1-tk expression in tumors by PET imaging is an important issue in clinical application as the success of gene transfer can be quantified before treatment initiation (23).
9LDsRed tumors were identified by MRI and [11C]MET PET (Fig. 2A1-A3 and B1-B3) before treatment initiation. A strong HSV-1-tk expression was detected within the tumor tissue (Fig. 2A1, A2, B1, B2, C1, and C2), with a relative expression level of 0.193 ± 0.06% ID/g (Fig. 3A). Efficient transduction of 9LDsRed by the vectors was confirmed by fluorescence microscopy of brain sections from animals that were sacrificed after the [18F]FHBG PET scan (Fig. 3B). In contrast, animals that had received glioma cells only (control group) had a significantly lower [18F]FHBG uptake in the tumor (0.037 ± 0.015% ID/g; P < 0.05; Figs. 2C3 and 3A).
On day 4, immediately after the initial [18F]FHBG scan, one group of vector implanted rats and one control group implanted with 9LDsRed only received daily i.p. injections of 30 mg/kg ganciclovir for 10 days. A second control group of vector implanted rats was not treated with ganciclovir. Five days after initiation of ganciclovir treatment, no tumors were visible on [11C]MET PET scans of treated animals, whereas control animals had large tumors (Fig. 2D1-D3). On MRI, treated animals showed a cystic lesion/residual tumor, whereas control animals displayed large tumors with central necrotic areas (Fig. 2E1-E3). Tumor volumes were quantified based on the gadolinium-diethylenetriaminepentaacetic acid–enhanced three-dimensional MRI (Fig. 4). The tumor size in treated animals was only 15% of both controls on day 15 after tumor initiation. Three to five days after termination of ganciclovir application, treated animals exhibited no signal on [11C]MET PET scans (Fig. 2F). Although there was still a lesion visible on MRI on day 22, no hyperintensity after gadolinium-diethylenetriaminepentaacetic acid administration was observed, indicating a cystic/necrotic lesion (Fig. 2G). No tumor cells were detected by histology, whereas the initial tumor was replaced by scar tissue (Fig. 3C). In both control groups, animals became terminally ill during the treatment and were euthanized. Histology revealed large tumor masses (Fig. 3D and E). Thus, the observations in PET and MRI correlated with histopathology and are therefore a useful tool to noninvasively measure gene transfer and treatment response.
Long-term survival of LCMV-tk/ganciclovir–treated rats harboring 9LDsRed glioma. To investigate the long-term therapeutic effect of lentiviral LCMV-GP pseudotyped vectors, we did a survival study. Animals were divided into three groups and treated as described for the imaging study above. Kaplan-Meier survival analysis showed 100% of long-term survivors in the therapeutic group surviving 100 days (Fig. 5A). Thus, treated animals showed a prolonged survival compared with both control groups, which was statistically highly significant (P < 0.001). However, there were 16% long-term survivors in both control groups. We investigated the brains of long-term survivors in the control groups and detected pseudocystic cavities filled with collagenous fibers and infiltrated by immune cells, which was similar to the histologic findings in treated animals (data not shown). Thus, we suspected that the DsRed protein may be immunogenic as described for GFP previously (24, 25) and thereby facilitate tumor eradication. Consequently, we used the naive 9L model in the following survival studies.
Both lentiviral LCMV-GP and VSV-G pseudotypes mediate killing of glioma cells in vivo. VSV-G pseudotyped lentiviral vectors transduce normal brain cells at a higher efficiency than rat glioma cells, whereas LCMV-GP pseudotypes were found to transduce glioma cells with great specificity and efficacy in vivo (12). To compare the therapeutic efficacy of these two lentiviral pseudotypes, a survival study was done with these vectors carrying HSV-1-tk fused to eGFP as therapeutic gene. Kaplan-Meier survival analysis showed that both LCMV-GP and VSV-G vector-treated animals had a significant survival advantage (P < 0.001 and 0.005, respectively) compared with the control groups (Fig. 5B). The tumor-free long-term survival was 50% in the VSV group and 37.5% in the LCMV group. The difference was not significant.
Bystander killing of 9L cells by normal brain cells in vivo. As 50% of the VSV-G vector-treated animals were long-term survivors in the study above, we addressed the question whether this therapeutic outcome was in part mediated by a bystander effect from HSV-1-tk–transduced normal brain cells. For this purpose, we first injected lentiviral VSV-G pseudotyped vectors carrying HSV-1-tk fused to eGFP into brains of Fischer rats at seven different coordinates in 1-mm distance to the planned tumor injection site. On day 5 after vector injection, 9L cells were implanted central to the previous vector injections. Four days later, ganciclovir therapy was done for 30 days. Kaplan-Meier survival analysis showed a prolonged survival of animals in the therapeutic group compared with both control groups, which was statistically significant (P < 0.01; Fig. 6A). There was one long-term survivor (80 days) without detectable tumor cells in the therapeutic group. To exclude that residual vector particles transduced tumor cells and therefore contributed to the therapeutic effect, we investigated the brains of vector-treated animals before ganciclovir application. We used 9LDsRed cells to distinguish tumor cells from normal brain cells. Histologic sections were analyzed by fluorescence microscopy and revealed transduction of normal brain cells only (Fig. 7A).
Normal brain cells transduced with HSV-1-tk-GFP are not eliminated by ganciclovir therapy. We investigated apoptosis during ganciclovir treatment of VSV-tk-GFP–injected and 9L-injected animals (day 4 after start of ganciclovir treatment). Sections were stained with anti-caspase-3 antibody and analyzed by confocal microscopy. Tumor cells became apoptotic during ganciclovir treatment, whereas the majority of GFP-positive brain cells remained caspase negative (Fig. 7B). Furthermore, we were interested in identifying the different cell populations transduced by VSV-G pseudotypes. GFP-expressing normal brain cells were found on histologic sections next to the tumor by confocal microscopy and composed of mainly glial fibrillary acidic protein–expressing astrocytes (Fig. 7C) and nestin-positive progenitor cells (Fig. 7D). Only few transduced NeuN-positive neurons were found (Fig. 7E). However, the general density of neurons was reduced near to the tumor, most likely because neurons are more vulnerable to glioma invasion than glial cells. GFP-positive cells of all three normal cell populations were also detected in the brain of one long-term surviving and tumor-free animal (data not shown). In conclusion, normal brain cells are not eliminated by ganciclovir therapy and contribute to the bystander effect of suicide gene therapy.
The present study shows that both lentiviral LCMV-GP and VSV-G pseudotypes are effective in HSV-1-tk suicide gene therapy for 9L glioma in vivo. We first analyzed the therapeutic efficiency of lentiviral LCMV-GP pseudotypes as these vectors transduce glioma more specifically and efficiently than VSV-G pseudotypes (12). In an imaging-guided approach based on PET and MRI, we showed eradication of 9LDsRed glioma using HSV-1-tk as therapeutic gene fused to eGFP. The expression of HSV-1-tk, which selectively incorporates certain radiolabeled nucleoside analogues into DNA (26–28), can be noninvasively imaged in distinct regions of a transduced tissue (19, 21, 23, 26, 29–34). In particular, HSV-1-tk–based PET imaging has been used to evaluate the transduction efficiency of viral vector-mediated gene therapy in various applications in vivo (19, 21, 35, 36). We extended this approach using PET and MRI techniques to a complete therapeutic follow-up with terminal histology. Before application of the prodrug ganciclovir, a high level of HSV-1-tk expression was measured in tumor tissue (0.193 ± 0.06% ID/g) by PET imaging using the specific tracer [18F]FHBG compared with nontransduced controls (0.037 ± 0.015% ID/g). Eradication of tumors in the treated group versus control groups was shown by [18C]MET PET, MRI-based volumetry, and histopathology. Thus, combined PET and MRI are useful tools to visualize viral gene transfer before treatment initiation and to follow the treatment course in a suicide gene therapy approach. In a previous study, we used a progenitor cell–based suicide gene therapy approach for malignant glioma and showed a similar correlation of histopathology to PET and MRI (37).
To further prove the therapeutic potential of lentiviral LCMV-GP pseudotypes, we did a survival study and verified complete tumor eradication and long-term survival of all animals in the treatment group. However, 16% of animals in both control groups also were long-term survivors. As we used 9L stably expressing the marker gene DsRed for tumor cell implantation, we postulated that the DsRed protein was immunogenic and that an immune response eliminated the tumor in a subset of animals. Although immunogenicity of DsRed has not been shown previously, the marker gene GFP was found to be immunogenic in different animal tumor models (24, 25). Consequently, we repeated the survival study using wild-type 9L cells and additionally compared the therapeutic efficacy of lentiviral LCMV-GP and VSV-G pseudotyped vectors. In accordance with the assumption that an immune response to DsRed had contributed to tumor eradication in 9LDsRed gliomas, the suicide gene therapy for 9L glioma was less effective and no long-term survivors were observed in the control groups. Nevertheless, a significant therapeutic effect was observed for both pseudotyped lentiviral vectors.
In principle, LCMV-GP pseudotyped vectors have major advantages about gene therapy of malignant glioma. In contrast to VSV-G, LCMV-GP is not cytotoxic and can be used to establish stable packaging cell lines. In previous studies, we produced stable packaging cell lines for gammaretroviral LCMV-GP pseudotyped vectors in epithelial cells (17) and in bone marrow–derived progenitor cells, which specifically migrate to and into malignant glioma (38). We showed transduction of glioma cells on implantation of these stable packaging progenitor cells in vivo (38). Such migratory packaging cell lines for lentiviral vectors that specifically infiltrate malignant gliomas may be the only future option to efficiently deliver therapeutic genes to large and invasively growing tumors.
LCMV-GP pseudotyped vectors transduce glioma cells in vivo with great efficiency and specificity. In contrast, VSV-G pseudotyped lentiviral vectors are more efficient in transducing normal brain cells than glioma cells (12). Thus, the good therapeutic efficacy of VSV-G pseudotypes against 9L gliomas was surprising. We postulated that the unexpected long-term survival of a fraction of VSV-G–treated animals was in part mediated by bystander killing of tumor cells through HSV-1-tk–transduced normal brain cells. It has been shown that cytotoxic ganciclovir phosphate is transported via gap junctions to neighboring cells and thereby exerts a bystander effect on nongene-modified cells (39). The role of gap junctions in this process has been verified by coculture experiments of connexin-expressing and connexin-deficient cells (40). The expression of gap junctions on 9L glioma cells used in this study has been shown (41). In addition, communication of astrocytes with glioma cells via connexins has been demonstrated in vivo (41, 42).
To show the bystander effect of normal brain cells, we first injected VSV-GP pseudotyped lentiviral vectors carrying HSV-1-tk-GFP into the brain of Fischer rats and injected 9L cells 5 days later. Before ganciclovir treatment, we investigated the brains for green fluorescence and found GFP protein exclusively in normal brain cells. On ganciclovir treatment, animals survived significantly longer than controls. In brain sections analyzed during therapy, apoptotic tumor cells surrounded by GFP-positive normal brain cells were seen. During and after ganciclovir treatment, many GFP-expressing cells were detected, showing that the transduced normal brain cells survived ganciclovir treatment to mediate the therapeutic effect. The GFP-positive cells consisted mostly of glial fibrillary acidic protein–positive reactive astrocytes and nestin-positive progenitor cells. Transduced normal brain cells are in a postmitotic state and thus can resist ganciclovir phosphate toxicity exerted by chain termination during mitotic DNA synthesis.
Cancer gene therapy has focused on specific transduction of cancer cells, avoiding the transduction of normal tissue. However, normal tissue in the immediate vicinity of the tumor undergoes reactive changes and is in direct contact to the tumor. Particularly, in the brain, it has been shown that reactive astrocytes communicate with tumor cells (41, 42). Furthermore, malignant glioma cells migrate into normal brain tissue and escape current therapies. In conclusion, normal brain cells in these infiltrating tumor areas might support therapy by long-term expression of therapeutic genes, and high specificity of gene transfer to cancer cells may not always be required.
Grant support: Köln Fortune Program at the University of Cologne grant 108/2003 (H. Miletic) and Deutsche Forschungsgemeinschaft grant DFG-Ja98/1-2, Center for Molecular Medicine Cologne grant CMMC-TV46, and 6th Framework EU grants European Molecular Imaging Laboratories grant LSHC-CT-2004-503569 and Diagnostic Molecular Imaging grant LSHB-CT-2005-512146 (M.A. Rueger, A. Winkeler, H. Li, and A.H. Jacobs). This publication was generated in the context of the CellPROM project, funded by the European Community as contract no. NMP4-CT-2004-500039 under the 6th Framework Program for Research and Technological Development in the thematic area of “Nanotechnologies and nano-sciences, knowledge-based multifunctional materials and new production processes and devices”.
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.
Note: H. Miletic and Y.H. Fischer contributed equally to this work.
This publication reflects only the authors' views; CellPROM is not liable for any use that may be made of the information contained therein.
We thank Mariana Carstov, Roswitha Seyd, and Ayla Derici for expert technical assistance and M. Wodak for photographic help.