mediated herpes simplex thymidine kinase gene - Nature

Gene Therapy (1998) 5, 113–120  1998 Stockton Press All rights reserved 0969-7128/98 $12.00

In vitro and in vivo antitumor effects of retrovirusmediated herpes simplex thymidine kinase genetransfer in human medulloblastoma A Rosolen1, E Frascella1, C di Francesco1, A Todesco1, M Petrone1, M Mehtali2, F Zacchello1, L Zanesco1 and M Scarpa1 1

Department of Pediatrics, University of Padua, Italy; and 2Transgene, Strasbourg, France

Herpes simplex virus thymidine kinase gene (HSVtk) transfer together with treatment with the prodrug ganciclovir (GCV) represents the most commonly used suicide gene approach for the gene therapy of human central nervous system malignancies. Despite encouraging results reported in clinical trials conducted in adults, very little is known about the feasibility of this approach for the treatment of CNS tumors of childhood. We studied the effects of the HSVtk/GCV system on human medulloblastoma cells in vitro and in vivo. The transfer of tk gene to medulloblastoma cells was capable of mediating cell suicide in vitro

and in vivo upon treatment with GCV, but the overall effect in vivo appeared to be suboptimal. The relatively low sensitivity of the medulloblastoma cells to viral infection and a limited bystander effect, coupled with a low expression of connexin-43 protein, might partially explain these results. Whether this is a peculiarity of the cell line studied or a general characteristic of medulloblastoma remains to be determined. These findings should be taken into account for the future planning of gene therapy trials for human medulloblastoma.

Keywords: thymidine kinase; gene transfer; bystander effect; medulloblastoma

Introduction Tumors of the central nervous system (CNS) are the second most frequent cancer in childhood.1 Despite recent improvements in diagnosis and treatment, approximately half of the children affected by a CNS tumor die of their disease. While supratentorial astroglial brain tumors represent the great majority of CNS tumors in adults, infratentorial tumors, such as ependimoma and medulloblastoma, are more common in children.1 Medulloblastoma, which accounts for 15–20% of all pediatric CNS tumors, is a predominantly midline cerebellar tumor that arises in the cerebellar vermis area and in the region adjacent to the fourth ventricle. A complete surgical resection is achievable in approximately half of the patients, but even in cases of radical resection further pharmacologic treatment and radiation therapy are needed to prevent local relapses and leptomeningeal dissemination. Given the frequency of an incomplete tumor resection and the young age at tumor onset, which prevents administration of an effective radiation dose, new therapeutic approaches should be explored. Among them, gene therapy is one of the most attractive. It has been demonstrated that transfer of the herpes simplex thymidine kinase (HSVtk) gene to tumor cells sensitizes them to the antiherpetic drug ganciclovir

` Correspondence: A Rosolen, Dipartimento di Pediatria, Universita di Padova, Via Giustiniani 3, 35128 Padova, Italy Received 23 May 1997; accepted 26 August 1997

(GCV).2–7 HSVtk can phosphorylate ganciclovir, which is then incorporated into cell DNA as a chain terminator, and ultimately causes selective death of the cells harboring and expressing the gene.8,9 Regardless of the vector used to transfer the HSVtk gene, it was suggested that the antitumor effects of the thymidine kinase/ganciclovir system are in part a direct consequence of the ‘suicide’ of the cell harboring the HSVtk gene, and are partly dependent on the so-called ‘bystander effect’.3,10–14 According to recent experimental findings, cells which are in the vicinity of tk-infected cells are exposed to toxic metabolites which can be transferred to adjacent cells through gap junctions,15,16 composed of a protein matrix containing abundant connexin-43.17–19 Recently, replication-defective retroviruses and adenoviruses were described which are capable of efficiently infecting human tumor cells and are able to mediate the transfer of therapeutically active genes.20 Retroviruses are suitable vectors for gene therapy of CNS tumors, since they infect only actively dividing tumor cells,21 but not normal nonproliferative cells in the brain.22 In order to achieve the highest efficiency of transduction, the in situ injection of packaging cells has been proposed.3 In this article we describe the use of a helper-free replicationdefective retroviral vector, RVTG5360-12, expressing the herpes simplex virus-1 thymidine kinase gene to infect human medulloblastoma cells in vitro. Furthermore, TG5360-12-producing cells were injected into human medulloblastoma xenografts in athymic mice and were capable of inducing necrosis in the treated tumors.

HSVtk gene transfer in medulloblastoma A Rosolen et al

114

Results D341Med cell proliferation studies D341Med human medulloblastoma cells were transduced in vitro with the HSVtk gene by superinfection with replication-defective RVTG5360 vector particles and then selected in puromycin-containing medium. Before proceeding to further studies, cell growth characteristics of control D341Med and D341Med-TK cells were evaluated. DNA synthesis, determined by standard 3H-thymidine incorporation assay, and cell growth, studied by MTT assay, were significantly higher in D341Med-tk cells (Table 1). Conversely, the ability to form colonies in semisolid medium (soft agar cloning) was decreased in the transduced D341Med cells compared to control D341Med cells. In order to ascertain whether the difference in cell growth and clonogenic capacity in vitro had any impact on the tumorigenicity in vivo, five athymic mice were injected subcutaneously in both inguinal regions with 5 × 106 D341Med and D341Med-TK cells, respectively. Within 15–18 days, all animals developed bilateral tumors which were similar in size and in growth rate (Table 1). In vitro evaluation of ganciclovir treatment of D341Med cells Various human cell lines including the medulloblastoma cell lines D341Med, D283Med and DAOY, the fibrosarcoma cell line HT1080, the melanoma cell line A2058, an ependimoma and an anaplastic astrocytoma short-term cell culture and the NIH3T3 cell line were used to determine the viral titer of our preparation. Viral titer varied from 1 × 106 c.f.u./ml in A2058 cells to 1 × 103 c.f.u./ml in primary ependimoma and astrocytoma cells, in the medulloblastoma cell lines and in the fibrosarcoma cell line (data not shown). D341Med-TK cells were then selected in puromycin and tested for their sensitivity to ganciclovir (GCV) in vitro. Control and infected cells were plated at a concentration of 2.5 × 104/ml in 96-well plates and 24 h later different concentrations of GCV were added. Cell viability was determined by MTT assay and by trypan blue exclusion test after 6 days of treatment, a time period similar to the duration of in vivo treatment used for CNS tumors in animals and in humans. MTT assay showed that no D341Med-TK survived at GCV concentrations

higher than 5 mg/ml (Figure 1) and this result was confirmed by trypan blue exclusion test. In contrast, no significant toxicity was observed when D341Med cells were treated with the same concentration of GCV (Figure 1). Only at concentrations 20 times higher than the 100% lethal dose for infected cells did we observe a 40% decrease in cell number (data not shown).

In vivo gene transfer and antitumor effect on D341Med xenografts Since D341Med-TK cells were killed by GCV treatment in vitro at concentrations usually reached in a clinical setting,23 we assessed whether D341Med-TK cells were sensitive to in vivo treatment, as well. Five mice were injected with 5 × 106 D341Med or D341Med-TK cells on each inguinal region, respectively, and when tumor masses reached a volume in the range of 500–1000 mm3, 150 mg/kg GCV was administered intraperitoneally twice a day for 6 days. While D341Med tumors continued to increase in size, D341Med-TK gradually decreased in volume and no mass was visible after 3 weeks (Figure 2). In order to mimic a clinical situation, in situ HSVtk gene transfer in vivo was attempted. When D341Med tumors grown bilaterally in the inguinal regions of athymic mice reached a diameter of 4–5 mm, one of the tumors was injected with 2 × 106 packaging cells, while the controlateral mass served as a control. Five days later GCV treatment was initiated at 150 mg/kg twice daily i.p. and continued for 6 days. In all five animals the treated tumors showed a decrease in growth during the following 15 days, although the difference between treated and control tumors was not statistically significant (data not shown). Histologic analysis of D341Med tumors injected with packaging cells showed some areas of necrosis which were absent in control tumors (Figure 3). Any attempt to establish a CNS D341Med xenograft in the cerebral parenchyma was unsuccessful. Bystander effect It was reported that the antitumor effect observed in vivo with a suicide gene therapy approach is in part mediated through the ‘bystander-effect’.3,10–14 Since in our experimental settings the in vivo antitumor effect was limited, we studied the role of the bystander effect in the GCVmediated death of D341Med cells. Five thousand D341Med cells were co-cultured with various numbers of D341Med-TK cells, ranging from 0 to 104, in medium

Table 1 Growth characteristics of D341Med and D341Med-TK cells

D341Med D341Med-TK

DNA synthesis c.p.m.a

Cell growth MTT-assay b

Clonogenicity % colonies c

Tumorigenicity tumors/mice d

12946 ± 1931 28294 ± 2404 P , 0.001

953 ± 88 1152 ± 73 P , 0.001

5.17 ± 0.33 3.9 ± 0.63 P , 0.05

5/5 5/5

DNA synthesis was determined by a standard 4-h 3H-thymidine incorporation assay in 7.5 × 103 cells plated in 96-well plates. Data are expressed as mean c.p.m. ± s.d. b Cell growth was studied by MTT assay, 24 h after plating. c Clonogenicity is expressed as percentage of colonies originated in soft agar. Data are expressed as mean percentage of colonies ± s.d. d Tumorigenicity in vivo is expressed as the number of mice which developed a tumor after a single subcutaneous injection of 5 × 106 cells. Statistical analysis was carried out with Student’s t test. a


115

Figure 1 Sensitivity of D341Med medulloblastoma cells measured by MTT assay. D341Med cells transduced or nontransduced with the HSVtk gene were cultured in complete medium in the presence of different concentrations of GCV for 6 days. On day 7, the viable cell number was determined by MTT assay. Data are expressed as mean absorbance at 540 nm ± standard deviation of triplicate samples.

containing 5 mg/ml GCV for 5 or 9 days. The inhibition of cell growth of D341Med cells, determined by MTTassay, increased dose-dependently as the D341-MedTK/D341Med ratio increased, suggesting that D341Med cell death was mediated by the presence of tk-positive cells (Figure 4). Bystander killing was evident at D341Med-TK/D341Med cell ratio as low as 2−2:1, and there was no significant difference in the 5-day versus 9day treatment. These findings were confirmed by trypan blue exclusion assay (data not shown).

Expression of connexin-43 in D341Med cells It has been suggested that the bystander effect is caused by passive transport of lethal metabolites from tk-positive cells to tk-negative cells through gap junctions.15,16 It has also been suggested that expression of connexin-43, which is a major structural component of the gap junction, is quantitatively related to the bystander effect.14,16,24 Furthermore, most tumor cell lines studied so far, showed significant bystander effect in vitro when tk+/tk− cells were plated at a ratio as low as 2−5:1.13,24 Thus, since the bystander effect in D341Med cells, although present, was lower than that reported for other tumor cell lines, we determined the expression of connexin-43 by Western blot analysis. Connexin-43 was expressed at very low levels in D341Med cells as compared with other human CNS tumor cell lines and normal rat brain and heart lysates, used as positive controls (Figure 5). These results were confirmed both in subconfluent and confluent cell cultures (data not shown).

Discussion The localized nature of human medulloblastoma and the impossibility of obtaining a complete surgical resection in all children affected by this disease render this tumor an ideal candidate for local therapy,1 including gene therapy. In situ retrovirus-mediated transfer of the HSVtk gene into CNS tumors was shown to cause preferential killing of malignant cells when followed by GCV treatment.3–7 Although there are various ongoing gene therapy clinical trials for the treatment of adult CNS malignancies,25,26 little data are available on the toxicity and efficacy of such an approach in pediatric CNS tumors,27,28 including medulloblastoma. Based on previous in vitro and in vivo studies, we determined the effects of HSVtk gene transfer to D341Med medulloblastoma cells and xenografts in order to evaluate the feasibility of retroviral HSVtk gene transfer for the treatment of residual tumor tissue or local relapses in children with medulloblastoma. We showed that D341Med cells can be infected both in vitro and in vivo by the RVTG5360 retroviral vector which expresses the HSVtk gene. The viral titer on these medulloblastoma cells was 1 × 103 c.f.u./ml, lower than the titer determined on NIH3T3 cells or on the melanoma cell line A2058 (1 × 106 c.f.u./ml), and was comparable to other medulloblastoma cell lines. Although the susceptibility of D341Med cells to viral infection was lower than that previously reported for other experimental brain tumors,7 the vector RVTG5360 was capable of transferring the HSVtk gene into human medulloblastoma cells. In order to study the in vivo effects of retrovirus-


116

Figure 2 Effects of ganciclovir treatment on D341Med xenografts. Five million D341Med or D341Med-TK cells were injected subcutaneously in each inguinal region of athymic mice. When tumor xenografts reached a volume of at least 500 ml, 150 mg/kg GCV was administered intraperitoneally, twice a day, for 6 consecutive days. Tumor volume was determined based on two diameters. In the Figure, a representative experiment is reported.

a

b

Figure 3 Effects of in situ HSVtk gene transfer and GCV treatment on D341Med cells in vivo. Two million HSVtk producer cells were injected intratumorally in D341Med xenografts grown in athymic mice. In the same animal another D341Med tumor served as a control. Mice were treated with 150 mg/kg GCV intraperitoneally twice a day for 6 days. On day 19 mice were killed and tumors fixed for subsequent histologic examination. Areas of necrosis were visible in transduced D341Med tumors (a), but not in the control tumors (b).


117

Figure 4 Bystander effect in D341Med cells in vitro. 5 × 103 D341Med cells were co-cultured with 0–104 D341Med-TK cells in 96-well plates in the presence of 5 mg/ml GCV. Five days later the total number of living cells was determined by MTT assay. The results represent the mean ± s.d. of triplicates.

Figure 5 Connexin-43 expression in D341Med cells. Fifty micrograms of cell lysate from subconfluent prostate carcinoma cells PC3, A172 and U87 glioma cells, medulloblastoma cells D283Med, D341Med and one of its subclones (D341Med-B2), PDCNS1 and PDCNS3 ependimoma and PDCNS2 anaplastic astrocytoma cells, were electrophoresed through a 10% polyacrylamide gel, transferred to a nitrocellulose membrane and reacted with an anticonnexin-43-specific MoAb. Reaction was revealed by a chemiluminescence-based technique. Rat heart and brain lysates (30 mg) were used as positive controls. The band at 43 kDa represents the connexin-43 protein (Cx43), while the signal at 68 kDa, visible in most samples as well as in the protein molecular weight standards (MW), is probably due to a nonspecific binding of the antibody to bovine serum albumin.

mediated HSVtk gene transfer, we first determined whether the transfected D341Med cells might have acquired different growth properties and clonogenic ability in vitro. Both DNA synthesis and cell growth, determined by 3H-thymidine incorporation assay and MTT assay, respectively, were significantly increased in transduced D341Med cells compared with parental cells. On the other hand, clonogenicity in soft agar was higher for untransfected cells, but the ability of D341Med and

D341Med-TK cells to grow as xenografts in athymic mice, was not different. Thus, any antitumor effects of GCV treatment in vitro and in vivo, except for the aspects related to clonogenicity in semisolid medium, should be ascribed to the treatment itself and not to different biological characteristics of transfected versus untransfected D341Med cells. Ganciclovir treatment of cell cultures showed that expression of HSVtk was sufficient to induce cell death at


118

drug concentrations which were comparable with those reported for most brain tumor cell lines and which can be obtained in vivo.23 Moreover, no toxicity was observed at the same GCV concentrations in untransfected D341Med cells. Some decline of cell viability was seen when GCV was added to control cell cultures at concentrations 20 times higher than the 100% lethal dose for D341Med-TK cells. Under these conditions, due to the suboptimal solubility of GCV and the high pH of the GCV solution (around pH 11.0), we believe that some degree of cell death could be the result of the final pH of the culture medium rather than a consequence of drug concentration. Based on the effect of GCV in vitro, we should expect that in vivo treatment of D341Med-TK tumors could eradicate the tumors grown as xenografts in athymic mice. Indeed this was observed in all animals, although we treated animals harboring tumors which had already reached a size of at least 500 mm3. This result demonstrates that in vivo, as well as in vitro, TK expression was sufficiently sustained to mediate cell suicide. The efficacy of our HSVtk approach was less pronounced when we performed in situ HSVtk gene transfer by injecting producer cells into the tumor xenografts. Tumors of 4–5 mm in diameter, injected with 2 × 106 producer cells showed areas of necrosis, whereas no necrosis was found in the control tumors (tumors injected with buffer alone). Nevertheless, no significant reduction in tumor growth of treated tumors compared with control D341Med tumors grown in the same animals, was observed. Whether this was due, in part, to the low efficiency of infection in vivo remains to be determined. With regard to this issue we are conducting a comparative study of medulloblastoma infectability by using retroviral as well as other viral vectors, such as adenoviruses. Another important aspect to be considered for the in vivo results of our suicide gene approach is the so-called ‘bystander effect’. Different authors have demonstrated that cell death upon HSVtk gene transfer is in part the consequence of the direct effect of GCV in transduced cells, but that the antitumor effect is amplified by transmission of toxic metabolites to neighboring cells through gap junctions.15,16 Furthermore, it is known that connexin-43 is a major component of gap junctions and that its expression is related to the magnitude of the bystander effect in human CNS tumor cell lines, such as the glioma cell line U87.24 Based on these observations we studied the bystander effect in D341Med cells in vitro. Ganciclovir treatment of D341Med and D341Med-TK cell cocultures showed that the bystander effect has a role for the antitumor activity in vitro in this medulloblastoma cell line, but it appears to be less efficient than in other experimental CNS tumor cell lines, since it is not appreciable at a ratio of tk−/tk+ cells lower than 1:2−2, compared with ratios of 1:2−5 reported for other cell lines.13,14 In order to assess whether there was any relationship with connexin-43 expression at the protein level, we assessed connexin-43 expression by Western blot analysis. D341Med cells expressed very low levels of connexin-43, compared with rat brain lysate (used as a positive control) and other human CNS tumor cell lines. These results suggest that in D341Med medulloblastoma cells the suicide effect of HSVtk gene transfer might be partially impaired by a low

bystander effect and that this might be a consequence of low connexin-43 expression. These observations lead to the speculation that the transfer of the connexin-43 gene into D341Med cells could increase the bystander effect, as previously demonstrated in the U87 glioma cell line.24 Moreover, in a potential clinical application of the HSVtk/GCV system to the gene therapy of children affected by medulloblastoma, the therapeutic efficacy could be increased by the antitumor immune response elicited by the tk product in an immunocompetent host.29 In conclusion, we were able to inhibit tumor growth of the medulloblastoma cell line D341Med both in vitro and in vivo following retrovirus-mediated transfer of HSVtk gene, but the overall efficacy of the in situ gene transfer in our in vivo model was lower than expected, probably due in part to a low connexin-43 expression. Further studies have been undertaken in order to determine the efficacy of other vectors for gene transfer, such as adenoviruses, and to better understand the relationship between connexin-43 expression and the bystander effect in human medulloblastoma.

Materials and methods Cell cultures The human medulloblastoma cell line D341Med and the NIH3T3 cell line were purchased from ATCC (Rockville, MD, USA), as were the producer cells GP+E8630 and GP+EnvAm12.31 PDCNS1, PDCNS2 and PDCNS3 are short-term cell cultures obtained from pediatric patients affected by ependimoma (PDCNS1 and PDCNS2) and by anaplastic astrocytoma, who had surgery in our Department. All cell lines were cultured in Dulbecco’s modified Eagle’s medium (Seromed, Milan, Italy) supplemented with 10% fetal calf serum, 2 mM L-glutamine (Seromed), 50 units/ml penicillin (Sigma, Milan, Italy), 50 mg/ml streptomycin (Sigma). Cultures were maintained in a humidified atmosphere with 5% CO2 at 37°C. When used for in situ gene transfer, the medium was removed and the cells were rinsed with saline. The monolayer was then incubated in 0.05% trypsin-EDTA (Seromed) for 5–10 min at 37°C, cells collected in HBSS (Seromed), washed twice and resuspended at the appropriate concentration for further use. Cell proliferation studies DNA synthesis was determined by 3H-thymidine incorporation assay. Briefly, cells were seeded in 96-well plates and cultured in complete culture medium for 24 h. Two days later, triplicate wells were pulsed for 4 h with 0.5 mCi of 3H-thymidine. Following the pulse, wells were harvested on to glass fiber filters by using a Skatron (Dasit, Milan, Italy) cell harvester and acid-precipitable radioactivity retained on the filters was determined by scintillation counting. Cell growth was indirectly determined by MTT assay as previously described.32 Briefly, cells were plated in a 96-well plate in complete culture medium. On the day of the assay, 20 ml of 5 mg/ml solution of MTT (Sigma) in phosphate-buffered saline was added to triplicate samples and the plate was incubated in the dark for a further 4 h. The plate was then centrifuged, the medium removed, and 150 ml dimethyl-sulfoxide (Sigma) added to each well. After 10 min incubation in the dark with


shaking, the absorbance at 540 nm was determined, as a measure of the formazan concentration generated in each well, using a ELX808 Elisa reader (Bio-Tek Instruments, Winooski, VT, USA).

Plasmid Plasmid pTG5360 was derived from pLXSP33 by insertion of an expression cassette containing the HSV-1tk gene isolated from ptk134 under the control of the murine phosphoglycerolkinase (pGK) gene promoter.35 The puromycin acetyltransferase36 gene was inserted downstream of the viral long terminal repeats. Isolation of RVTG5360-12 producer clone and D341Med-tk+ cells Ecotropic retroviral supernatant recovered from GP+E−86 transfected with pTG5360 by calcium-phosphate coprecipitation technique37 was used to transduce the amphotropic GP+EnvAm12 cell line. Stable producer clones were isolated by culture in the presence of puromycin (5 mg/ml; Sigma) and clonal dilution. One hundred producer cell clones were titrated on NIH3T3 cells in the presence of polybrene (8 mg/ml; Sigma) and the best clone RVTG5360-12 titer (9 × 106 c.f.u./ml) was selected for further studies. To generate D341Med-tk+ cells, 105 cells were seeded and infected with 5 ml supernatant recovered from the best RVTG5360-12 clone in the presence of polybrene (4 mg/ml). Next day puromycin (5 mg/ml) was added and selection continued for 4 days. D341Med-tk+ cells were expanded in complete culture medium supplemented with puromycin. In vitro sensitivity to ganciclovir Control NIH3T3 cells and cells transduced with RVTG5360-12 were plated in triplicate at a density of 103 cells per well in a six-well plate (Costar, Badhoevedorp, The Netherlands). Twenty-four hours later, ganciclovir (GCV; Syntex Laboratories, Palo Alto, CA, USA) was added at increasing concentration (from 0 to 100 mm) and the cultures continued for 7–8 days. The medium was subsequently removed, cells were trypsinized, counted and viability was determined by the trypan blue (Sigma) exclusion test, following the manufacturer’s instructions. Determination of the bystander effect Five thousand D341Med cells were mixed in complete culture medium with RVTG5360-12 transduced D341Med cells (D341Med-tk+) at a ratio of 1:2 to 1:2−6 and then plated in a 96-well plate (Costar), in triplicate. Next day, GCV was added at a final concentration of 5 mg/ml. Medium was changed every 3 days and on day 5 or 9 the number of living cells was determined by MTT assay and by the trypan blue exclusion test. Medulloblastoma xenografts and ganciclovir treatment in vivo Human medulloblastoma xenografts were established in both inguinal regions of athymic mice by injecting 5 × 106 D341Med cells subcutaneously. Approximately 3 weeks later, when the tumor masses reached a diameter of at least 5 mm, 2 × 106 HSVtk vector-producer cells in 20 ml of Hank’s balanced salt solution were injected into one of the tumors over a period of 5 min. The other tumor was injected with the same volume of Hank’s solution. After 72 h, GCV was administered i.p., 150 mg/kg twice

a day for 6 days and tumor diameters measured daily with calipers to determine the volume. We also tried to establish human medulloblastoma xenografts in the CNS of six athymic mice by injecting 1 × 107 D341Med cells into the brain parenchyma through a stereotactic apparatus. The mice were observed for 80 days after the injection, killed and autopsy examination of the brain was carried out.

Western blot analysis Subconfluent cells were washed twice in phosphate-buffered saline (PBS) and then 1 ml of lysis solution (1% sodium dodecylsulfate (SDS), 1.0 mM sodium vanadate, 10 mM Tris pH 7.4) was added per 10 cm diameter Petri dish. Cells were subsequently scraped from the dish, transferred to a microcentrifuge tube and boiled for 5 min. The viscosity of the sample was reduced by several passages through a 26-gauge needle and insoluble material removed by centrifugation. A small aliquot of the supernatant was diluted to approximately 0.1% SDS and used to measure the protein concentration in the sample using the BCA method (Pierce, Beijerland, The Netherlands). Fifty micrograms of protein were elecrophoresed through a 10% SDS-polyacrylamide gel, and transferred to a nitrocellulose membrane by use of a semi-dry transfer cell (Bio-Rad, Milan, Italy) following the manufacturer’s instructions. Membrane was then placed into blocking buffer (5% non-fat dry milk, 1% bovine serum albumin, 10 mM Tris, 100 mM NaCl, 0.1% Tween 20) for 1 h at room temperature and then incubated with an anticonnexin-43 monoclonal antibody (Transduction Laboratories, Mamhead, UK) diluted 1:250 in blocking solution. After 3 × 10 minutes washing in 10 mM Tris, 100 mM NaCl, 0.1% Tween 20 to remove the primary antibody, the membrane was reacted with a rabbit anti-mouse horseradish peroxidase conjugated IgG (Transduction Laboratories) at room temperature for 1 h. Reaction was then revealed using the ECL Western blotting analysis kit (Amersham, Milan, Italy).

Acknowledgements We are grateful to Dr F Giangaspero, Servizio di Anatomia Patologica, Cesena Hospital, for performing histologic analysis of medulloblastoma xenografts, Y Poitevin for excellent technical assistance and Dr S Benini and Dr A Piccin for helpful assistance in cell culture and animal studies, respectively. This work was supported by a grant from AIRC (Associazione Italiana per la Ricerca sul Cancro), Progetto Terapia Genica.

References 1 Heideman RL et al. Tumors of the central nervous system. In: Pizzo PA, Poplack DG (eds). Principle and Practice of Pediatric Oncology. Lippincott-Raven: Philadelphia 1997, pp 633–697. 2 Borrelli E, Heyman R, Hsi M, Evans RM. Targeting of an inducible toxic phenotype in animal cells. Proc Natl Acad Sci USA 1988; 85: 7572–7576. 3 Culver KW et al. In vivo gene transfer with retroviral-producer cells for treatment of experimental brain tumors. Science 1992; 256: 1550–1552. 4 Ezzeddine ZD et al. Selective killing of glioma cells in culture and in vivo by retrovirus gene transfer of the herpes simplex virus thymidine kinase gene. New Biol 1991; 3: 608–614.

119


120

5 Moolten FL. Tumor chemosensitivity conferred by inserted herpes thymidine kinase genes: paradigm for a prospective cancer control strategy. Cancer Res 1986; 46: 5276–5281. 6 Moolten FL, Wells JM. Curability of tumors bearing herpes thymidine kinase genes transferred by retroviral vectors. J Natl Cancer Inst 1990; 82: 297–300. 7 Ram Z et al. In situ retroviral gene transfer for the treatment of brain tumors in rats. Cancer Res 1992; 53: 83–88. 8 Field AK et al. 9-(2-Hydroxyl-1-hydroxymethyl, ethoxy)guanine: a selective inhibitor of herpes group virus replication. Proc Natl Acad Sci USA 1983; 80: 4139–4143. 9 Matthews T, Boehme R. Antiviral activity and mechanisms of action of ganciclovir. Rev Draft Dis 1988; 10: S490–S494. 10 Freeman SM et al. The ‘Bystander effect’: tumor regression when a fraction of tumor mass is genetically modified. Cancer Res 1993; 53: 5274–5283. 11 Beck C et al. The thymidine kinase/ganciclovir-mediated ‘suicide’ effect is variable in different tumor cells. Hum Gene Ther 1995; 6: 1525–1530. 12 Chen CY et al. Effect of herpes simplex virus thymidine kinase expression levels on ganciclovir-mediated cytotoxicity and bystander effect. Hum Gene Ther 1995; 6: 1467–1476. 13 Samejima Y, Meruelo D. Bystander killing induces apoptosis and is inhibited by forskolin. Gene Therapy 1995; 2: 50–58. 14 Elshami AA et al. Gap junctions play a role in the ‘bystander effect’ of the herpes simplex virus thymidine kinase/ganciclovir system in vitro. Gene Therapy 1996; 3: 85–92. 15 Li Bi W, Parysek LM, Warnick R, Stambrook PJ. In vivo evidence that metabolic cooperation is responsible for the bystander effect observed with HSVtk retroviral gene therapy. Hum Gene Ther 1993; 4: 725–731. 16 Fick J et al. The extent of heterocellular communication mediated by gap junctions is predictive of bystander tumor cytotoxicity in vitro. Proc Natl Acad Sci USA 1995; 92: 11071–11075. 17 Loewenstein WR. Junctional intercellular communication: the cell-to-cell membrane channel. Physiol Rev 1981; 61: 829–913. 18 Beyer EC, Paul DL, Goodenough DA. A protein from rat heart homologous to a gap junction protein from liver. J Cell Biol 1987; 105: 2621–2629. 19 Kumar NM, Gilula NB. The gap junction communication channels. Cell 1996; 84: 381–388. 20 Wu GY, Wu CH. Delivery systems for gene therapy. Biotherapy 1991; 3: 87–95. 21 Miller DG, Mohammed AA, Miller AD. Gene transfer by retrovirus vectors occurs only in cells that are actively replicating at the time of infection. Mol Cell Biol 1990; 10: 4239–4242. 22 Plautz G, Nabel EG, Nabel GJ. Selective elimination of recombinant genes in vivo with a suicide retroviral vector. New Biol 1991; 3: 709–715. 23 Fletcher C et al. Human pharmacokinetics of the antiviral drug DHPG. Clin Pharmacol Ther 1986; 40: 281–286.

24 Colombo BM et al. The ‘bystander effect’: association of U-87 cell death with ganciclovir-mediated apoptosis of nearby cells and lack of effect in athymic mice. Hum Gene Ther 1995; 6: 763–772. 25 Culver KW et al. Gene therapy for the treatment of malignant brain tumors with in vivo tumor transduction with the herpes simplex thymidine kinase gene/ganciclovir system. Hum Gene Ther 1994; 5: 343–379. 26 Eck SL et al. Treatment of advanced CNS malignancies with the recombinant adenovirus H5.010RSVTK: a phase I trial. Hum Gene Ther 1996; 7: 1465–1482. 27 Raffel C et al. Gene therapy for the treatment of recurrent pediatric malignant astrocytomas with in vivo tumor transduction with the herpes simplex thymidine kinase gene/ganciclovir system. Hum Gene Ther 1994; 5: 863–890. 28 Kun LE et al. Stereotactic injection of herpes simplex thymidine kinase vector producer cells (PA317-G1Tk1SvNa.7) and intravenous ganciclovir for the treatment of progressive or recurrent primary supratentorial pediatric malignant brain tumors. Hum Gene Ther 1995; 6: 1231–1255. 29 Barba D, Hardin J, Sadelain M, Gage FH. Development of antitumor immunity following thymidine kinase-mediated killing of experimental brain tumors. Proc Natl Acad Sci USA 1994; 91: 4348–4352. 30 Markowitz D, Goff S, Bank A. A safe packaging line for gene transfer: separating viral genes on two different plasmids. J Virol 1988; 62: 1120–1124. 31 Markowitz D, Goff S, Bank A. Construction and use of a safe and efficient amphotropic packaging cell line. Virology 1988; 167: 400–404. 32 Twentyman PR, Fox NE, Rees JKH. Chemosensitivity testing of fresh leukemia cells using the MTT colorimetric assay. Br J Haematol 1989; 71: 19–24. 33 Miller D, Rosman G. Improved retroviral vectors for gene transfer and expression. Biotechniques 1989; 7: 980–988. 34 Spandidos DA, Harrison PR, Paul J. Replication and amplification of recombinant plasmid molecules as extra chromosomal elements in transformed mammalian cells. Exp Cell Res 1982; 141: 149–158. 35 Adra C, Boehr P, McBurney M. Cloning and expression of the mouse pgk-1 gene and nucleotide sequence of its promoter. Gene 1987; 60: 65–74. 36 Morgenstern J, Land H. Advanced mammalian gene transfer: high titer retroviral vectors with multiple drug selection markers and a complementary helper free packaging cell line. Nucleic Acids Res 1990; 8: 3587–3596. 37 Graham FL, van der Erb AJ. A new technique for the assay of infectivity of human adenovirus type 5. J Gen Virol 1973; 52: 456–467.