Helper Virus-Free Herpes Simplex Virus Type 1

0 downloads 0 Views 233KB Size Report
N.G. RAINOV,1,4 C. FRAEFEL,1,7 and X.O. BREAKEFIELD1 ... secretion into the medium during the first 24 hr after infection was 1026 ng/106 cells, whereas.
HUMAN GENE THERAPY 11:1429 –1438 (July 1, 2000) Mary Ann Liebert, Inc.

Helper Virus-Free Herpes Simplex Virus Type 1 Amplicon Vectors for Granulocyte-Macrophage Colony-Stimulating Factor-Enhanced Vaccination Therapy for Experimental Glioma U. HERRLINGER,1,2 A. JACOBS,1,3 A. QUINONES,4 C. WOICIECHOWSKY, 5,6 M. SENA-ESTEVES, 1 N.G. RAINOV,1,4 C. FRAEFEL, 1,7 and X.O. BREAKEFIELD 1

ABSTRACT Subcutaneous vaccination therapy with glioma cells, which are retrovirally transduced to secrete granulocytemacrophage colony-stimulating factor (GM-CSF), has previously proven effective in C57BL/6 mice harboring intracerebral GL261 gliomas. However, clinical ex vivo gene therapy for human gliomas would be difficult, as transgene delivery via retroviral vectors occurs only in dividing cells and ex vivo glioma cells have a low growth fraction. To circumvent this problem, a helper virus-free herpes simplex virus type 1 (HSV-1) amplicon vector was used. When primary cultures of human glioblastoma cells were infected with HSV-1 amplicon vectors at an MOI of 1, more than 90% of both dividing and nondividing cells were transduced. When cells were infected with an amplicon vector, HSVGM, bearing the GM-CSF cDNA in the presence of Polybrene, GM-CSF secretion into the medium during the first 24 hr after infection was 1026 ng/10 6 cells, whereas mock-infected cells did not secrete detectable GM-CSF. Subcutaneous vaccination of C57BL/6 mice with 5 3 10 5 irradiated HSVGM-transduced GL261 cells 7 days prior to intracerebral implantation of 10 6 wild-type GL261 cells yielded 60% long-term survivors (. 80 days), similar to the 50% long-term survivors obtained by vaccination with retrovirally GM-CSF-transduced GL261 cells. In contrast, animals vaccinated with the same number of nontranduced GL261 cells or with GL261 cells infected with helper virus-free packaged HSV1 amplicon vectors carrying no transgene showed only 10% long-term survivors. In conclusion, helper virusfree HSV-1 amplicon vectors appear to be effective for cytokine-enhanced vaccination therapy of glioma, with the advantages that both dividing and nondividing tumor cells can be infected, no viral proteins are expressed, and these vectors are safe and compatible with clinical use.

OVERVIEW SUMMARY Helper virus-free HSV-1 amplicon vectors provide a means for nonreplicative, nontoxic, efficient gene delivery to tumor cells for vaccination purposes. HSV-1 amplicon vectors were used to transduce GL261 glioma cells with GM-CSF. After infection, GL261 cells secreted high levels of GM-CSF into the supernatant. GM-CSF levels were sufficient to induce a potent

protective vaccination effect when irradiated, GM-CSF-transduced GL261 cells were injected subcutaneously into mice that were later challenged intracerebrally with wild-type GL261 cells. Since helper virus-free amplicon vectors infect more than 90% of both dividing and nondividing cells in primary cultures obtained from resected human glioblastoma tissue at an MOI of 1, these vectors appear to be promising tools for clinical ex vivo transduction of glioma cells for vaccination therapy.

1 Neurology

Service, Molecular Neurogenetics Unit, Massachusetts General Hospital and Harvard Medical School, Charlestown, MA 02129. of Neurology, University of Tuebingen, D-72076 Tuebingen, Germany. 3 Department of Neurology, University of Cologne, D-50931 Cologne, Germany. 4 Department of Neurosurgery, Martin-Luther University, D-06097 Halle am der Saale, Germany. 5 Neurosurgery Service, Molecular Neurogenetics Unit, Massachusetts General Hospital and Harvard Medical School, Charlestown, MA 02129. 6 Department of Neurosurgery, Humboldt University, D-1344 Berlin, Germany. 7 Institute of Virology, University of Zurich, CH-8057 Zurich, Switzerland. 2 Department

1429

1430

HERRLINGER ET AL.

INTRODUCTION

V

for glioma using subcutaneous injection of irradiated glioma cells, which have been genetically modified to secrete immunostimulating cytokines, has proven to be an effective treatment strategy in animal models of glioma. This strategy has been successful with syngeneic glioma cells that are transduced to express granulocyte macrophage colony-stimulat ing factor (GM-CSF) (Herrlinge r et al., 1997; Yu et al., 1997), or antisense to transforming growth factor b (Fahkrai et al., 1996), interleukin 2 (IL-2), and interferon g (IFN-g ) (Lichtor et al., 1995). In these previous experiments, transduction of glioma cells was carried out in culture with retrovirus vectors that can transduce only dividing cells. For clinical application of this ex vivo gene therapy, the growth fraction of glioblastoma tissue is too low (15–20%) to yield a significant transduction rate (Kleihues et al., 1997). Thus, for retrovirus infection, rapidly proliferating, short-ter m cell cultures need to be established from tumor resection samples. However, cell culture, even for several days, may lead to a selection of a subpopulation of glioma cells, from genetically heterogeneous tumors, and thus some tumor antigens may not be present in the cells used for vaccination. To avoid the need for short-term culture, vectors that can transduce both dividing and nondividing glioma cells need to be used. Fast and efficient transduction of dividing and nondividing glioma cells can be achieved by herpes simplex type 1 (HSV-1) vectors (Chiocca et al., 1990; Andreansky et al., 1996; Herrlinger et al., 1998, 2000). HSV-1 vectors have been used for vaccination purposes in two nonglioma animal tumor models: (1) a disabled infectious single-cycle herpes simplex virus (DISC-HSV) has been used in a lymphoma model (Dilloo et al., 1997). This viral is mutated in the gene for glycoprotein H (gH), but can be grown to high titers in complementing cell lines expressing gH. When this vector infects target cells, i.e., lymphoma cells, which do not express gH, virus replication proceeds, thereby killing the host cell, but progeny virions produced by these target cells lack gH in the envelope and thus are not able to infect other cells; (2) HSV-1 amplicon vectors have been used for vaccinatio n therapy in a hepatoma model (Tung et al., 1996). An HSV-1 amplicon vector carries DNA that, in its minimal version, bears noncoding HSV-1 sequences, including a DNA cleavage/packaging signal (, 1 kb) and an origin of DNA replication (, 1 kb), prokaryotic sequences for propagation of plasmid DNA in bacteria (Escherichia coli origin of DNA replication and antibiotic resistance gene), and a transgene cassette with the gene(s) of interest (Fraefel et al., 1998). For replication of amplicon DNA and packaging into HSV-1 virions, HSV-1 helper virus functions need to be expressed in trans. Tung et al. (1996) provided these functions through conventional use of a mutant helper virus, thus generating helper virus as well as amplicon vectors. For both HSV-1-based systems used so far for vaccination therapy of tumors, the gH 2 and the helper virus-con taining amplicon vectors, an important restriction applies: the vectors kill their target cells within days after infection due to either completion of the replicative cycle of HSV-1 (DISCHSV) or the expression of cytotoxic virus genes from the helper virus genome. This may limit the time available for antigen presentation in vivo. To overcome this problem, the present study used nontoxic helper virus-free amplicon vectors for transduc A C C IN A TIO N T H ER A P Y

tion of glioma cells. Helper virus-free packaging is achieved by transfecting cells with an HSV-1 genome deleted for packag ing signals (Fraefel et al., 1996) contained either in overlap ping cosmids, which recombine to form an HSV-1 genome (Cunningham and Davison, 1993), or in a bacterial artificia l chromosome (Saeki et al., 1998). These packaging-defect ive helper genomes can provide all functions necessary for replication and packaging of amplicon DNA. Since intact packag ing sequences are present only in the amplicon DNA, only amplicons are packaged into HSV-1 virions, yielding stocks of highly infectious, nontoxic HSV-1 particles carrying the desired transgene and no viral genes. To facilitate clinical application s with a need for high titers of vectors, Sun et al. (1999) have described a method to produce helper virus-free vector stocks at titers up to 8 3 108 infectious particles/ml. On the basis of positive results in the GL261 mouse glioma model with GMCSF-enhanced vaccination therapy (Herrlinger et al., 1997; Yu et al. 1997), the feasibility and efficacy of vaccination therapy for experimental glioma were explored with helper-virus free amplicons expressing GM-CSF.

MATERIALS AND METHODS Cells and cell culture media The GL261 mouse glioma cell line was obtained from S. Saris (New England Medical Center, Boston, MA). These cells are matched for the major histocompatibili ty complex antigen H-2b with the C57BL/6 mouse host (Akbasak et al., 1991), in which they have been routinely passaged (Ausman et al., 1970). To obtain GL261 cells, which stably secrete mouse GM-CSF (RV-GM), they were infected with the supernatant of stably transfected c CRIP cells producing GM-CSF retrovirus vectors , as previously described (Herrlinger et al., 1997). GL261 cells and their derivatives, baby hamster kidney (BHK) cells (American Type Culture Collection [ATCC], Manassas, VA), African green monkey Vero cells (African green monkey kidney cells; ATCC), and 2-2 cells, derived from Vero cells by stable transfection with the ICP27-encoding gene of HSV-1 (Smith et al., 1993), were grown in Dulbecco’s modification of Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS) (Sigma, St. Louis, MO), penicillin (100 U/ml), and streptomycin (100 m g/ml) (1% P/S; GIBCO, Gaithersburg, MD).

HSV-1 amplicon plasmids, packaging, and titering HSV-1 amplicon plasmids included pHSVPrPUC (Bergold et al., 1993; kindly provided by H. Federoff, University of Rochester, Rochester, NY), which bears no transgene, and pHSVlacZ, which expresses E. coli lacZ from the HSV-1 immediate early 4/5 promoter (IE 4/5; Geller and Breakefield , 1988). pHSVGM was generated from pHSVPrPUC by digestion with BamHI, and dephosphorylation and insertion of the 1.0-kB BamHI–BamHI fragment from pGM-CSF-MFG (Dranoff et al., 1993) bearing the murine GM-CSF cDNA, thereby placing GM-CSF under control of the HSV-1 IE 4/5 promoter (Fig. 1). Helper virus-free stocks of amplicon vectors HSVGM, HSVlacZ, and HSVPrPUC were generated according to Fraefel et al. (1996). A set of five overlapping cosmids (0.4 m g

1431

HSV AMPLICONS FOR VACCINATION THERAPY each), which cover the whole HSV-1 genome (Cunningham and Davison, 1993) and are deleted for packaging signals, were cotransfected with 0.6 m g of amplicon DNA into 10 6 Vero 2-2 cells on 60-mm plates, using LipofectAMINE and Opti-MEM according to manufacturer instructions (GIBCO). On day 3 after transfection, amplicon stocks were harvested by scraping the cells, followed by three cycles of freezing and thawing and sonication for 16 sec. Helper virus-free amplicon stocks were checked for the presence of replicating helper virus by infecting 100,000 Vero cells in a 24-well plate with 500 m l of an amplicon stock and monitoring for plaque formation. Using this assay, contamination with replicating HSV-1 helper virus at 2 PFU/ml or more would have been detected. In all amplicon stocks produced and used for experiments described here, helper virus titers were less than 2 PFU/ml. Helper virus-free packaged stocks of HSVlacZ were titered on Vero cells (100,000 per well in 24-well plates) by infection with serial dilutions of HSVlacZ stock. Twenty-four hours later cells were washed with phosphate-buffer ed saline, pH 7.4 (PBS), fixed in ethanol, and stained for b -galactosidase for 4 hr at 37°C in 5-bromo-4-chl oro-3-indolyl -b - D -galactopyranoside (X-Gal) staining solution containing 5 mM potassium ferrocyanide, 5 mM potassium ferricyanide, 2 mM magnesium chloride, X-Gal (1 mg/ml; Fisher, Pittsburgh, PA), and 2.5% dimethyl sulfoxide (Sigma, St. Louis, MO) in PBS (Turner et al., 1990). Five different fields per well ( 3 100 magnification ) were counted to assess the number of LacZ 1 cells per well. Titers were expressed in transducing units (TU) per milliliter. All determinations were carried out at least in duplicate. Stocks of HSVGM virions were assessed for infectivity by infecting 10,000 GL261 cells with 500 m l of stock solution and determining GM-CSF secretion as described below.

Infection protocol and determination of GM-CSF secretion Infectability of GL261 cells with HSV-1 amplicons was determined with HSVlacZ amplicon stocks and BHK cells as a positive control. Cells in log-phase growth were harvested with trypsin, washed in Hanks’ balanced salt solution (HBSS; GIBCO), and resuspended at 1 3 10 8 cells/ml HBSS. Cultures were irradiated with 35 Gy from a 137 Cs source. Cells were then diluted in growth medium and plated into 24-well plates at a density of 10,000 cells per well in a total volume of 500 m l. Various volumes of amplicon stocks were added for a 4-hr incubation time immediately after plating, and then cells were rinsed and placed in growth medium. One day postinfection , cells were stained for LacZ as described above. In some experiments, infections were carried out with various concentra tions of Polybrene (0–40 m g/ml). To analyze the influence of irradiation and/or infection with HSVlacZ stocks on cell survival, 10,000 cells were harvested

postirradiation and/or postinfection, an aliquot was stained with trypan blue (Sigma) to determine percent viability, and another aliquot was counted with a Coulter Counter (Coulter Electronics, Hialeah, FL). Cells were counted directly after and 24, 48, and 96 hr after radiation/infection. All experiments were done in duplicate. To assess the influence of timing of irradiation on transduc tion efficiency, cells were trypsinized, harvested, and split into two samples. One sample was immediately irradiated with 35 Gy to induce growth arrest, and then plated and infected 1 day later with HSVlacZ for a 4-hr period. The other sample was plated, infected for 4 hr, and irradiated 1 day later with 35 Gy. Two days after the start of the experiment the percentage of LacZ-positive cells per well was determined by LacZ staining . To analyze the secretion of GM-CSF into the medium after infection with amplicon vectors encoding GM-CSF, GL261 glioma cells in log-phase growth were irradiated with 35 Gy, plated at a density of 10,000 cells per well in 24-well plates, and infected for 4 hr with various volumes of HSVGM, as described above for infection with HSVlacZ. The medium was changed every 24 hr and conditioned medium was stored at 2 70°C. GM-CSF secretion levels was assayed by enzymelinked immunosorben t assay (ELISA) according to the manufacturer instructions (Endogen, Cambridge, MA).

In vivo experiments For in vivo experiments, 5 3 10 5 wild-type GL261 cells or retrovirally GM-CSF-transduced GL261 cells (RV-GM) were plated into 100-mm plates. After 24 hr, cells were irradiated with 35 Gy to induce growth arrest, as described above, and washed with PBS. Subsequently, GL261 cells were incubated in 8 ml of medium per plate with helper virus-free packaged amplicon stock (HSVGM, HSVlacZ, or HSVPrPUC) supplemented with Polybrene (40 m g/ml) for 4 hr. Uninfected plates with wild-type GL261 cells (negative control) and RV-GM cells (positive control) were incubated in growth medium 1 Polybrene (40 m g/ml) without amplicon virus. Plates were washed three times with PBS, trypsinized, and resuspended in HBSS to a concentration of 5 3 10 5 cells/100 m l. To determine the GM-CSF secretion during the first 24 hr after infection, a sample of 10,000 cells of each solution was plated into 24-well plates in triplicate and GM-CSF secretion was determined in the supernatant 24 hr later by ELISA. For subcutaneous vaccination, 100 m l of a cell suspension containing 5 3 105 cells was injected subcutaneously into the right flank of 6-week-old female C57BL/6 mice (Charles River Laboratories, Wilmington, MA) immediately after infection with amplicon vectors. The various treatment groups were vaccinated with (1) control GL261 cells, (2) HSVlacZ-infected cells, (3) HSVGM-infected cells, (4) HSVPrPUC-infected cells, or (5) RV-GM cells. For intracranial tumor cell implantation ,

FIG. 1. DNA structure of the GM-CSF-expressing HSV-1 amplicon vector (HSVGM). OriS, the HSV-1 origin of DNA replication and pac, the HSV-1 packaging signal, enables the DNA to be packaged into HSV-1 virions. The GM-CSF gene is driven by the HSV-1 IE4/5 promoter.

1432

HERRLINGER ET AL.

7 days after subcutaneous vaccination, mice were anesthetize d by intraperitoneal injection with 60–100 m l of a solution made with 2 parts of bacteriostatic 0.9% NaCl (Abbott, Abbott Park, IL) and 1 part each of xylazine (Rompun, 20 mg/ml; Miles, Kankakee, IL) and ketamine (Ketalar, 100 mg/ml; Parke-Davis, Detroit, MI). Control GL261 cells (106 ) were injected intracra nially 2 mm rostral and 2 mm to the right of the bregma, at a depth of 2.5 mm from the dura. Animals were observed daily and killed by cervical dislocation at terminal disease, when movement and feeding became difficult. Long-term surviving animals were killed on day 81 after intracranial tumor cell implantation. Differences in survival between treatment groups were analyzed for significance by the log-rank test. Animal studies were done in accordance with the guidelines issued by the Massachusetts General Hospital Subcommittee of Animal Care.

Transduction of human ex vivo glioblastoma cells Primary cell cultures from resected tumor specimens of three consecutive patients with histologically proven glioma (WHO grade IV) were prepared by placing small tumor pieces into cell culture flasks containing DMEM (as described above) but with 15% FCS and allowing the glioblastoma cells to migrate out. After 2 or 3 passages in culture and after verifying descendenc e from the glial lineage by glial fibrillary acidic protein (GFAP) immunocytochem istry, 15,000 cells were plated on slide chambers in growth medium. Cells were incubated for 24 hr at 37°C and then 10 m M bromodeoxyurid ine (BrdU) was added to the cultures. Twenty-four hours later, cells were infected with HSVlacZ vectors at a multiplicity of infection (MOI) of 1 for 6 hr. Cells were washed and resuspended in growth medium with 10% FCS and incubated for an additional 24 hr to allow time for lacZ expression. Cells were then washed in PBS and stained for b -galactosidase, as described above (Turner et al., 1990), and for BrdU incorporation into dividing cells as follows. Cells were fixed with methanol and air dried; DNA was denatured in 2 M HCl for 60 min at 37°C followed by neutralization in 0.1 M borate buffer, pH 8.5. After thorough washing with PBS, slides were incubated in a humidified chamber for 60 min at room temperature with 400 m l of a solution contain ing anti-bromodeox yuridine–fluorescein-conju gated antibody (50 pg/ml; Boehringer Ingelheim, Pearl River, NY) in PBS with 0.1% bovine serum albumin (BSA). The percentage of fluorescent cells was determined with a UV microscope.

RESULTS Infection of GL261 glioma cells with HSVlacZ GL261 glioma cells are moderately infectable with HSV-1 vectors. Infection with a helper-free preparation of HSVlacZ amplicon for 4 hr at an MOI of 1 (1 transducing unit [TU]/cell) yielded 10% transduced GL261 cells (Fig. 2A). To infect the same percentage of 2-2 or BHK cells, a 10-fold lower MOI (0.1) is sufficient (data not shown). With MOIs of 5 and higher, more than 75% of GL261 cells could be infected with this vector (Fig. 2A). To increase the percentage of infected cells at a given MOI, Polybrene was added during infection and washed off 1 day later. At an MOI of 2.5, the presence of Polybrene at 40 m g/ml during the 4-hr incubation with HSVlacZ stocks en-

hanced the percentage of LacZ-positive cells by 2.5-fold, as compared with infection without Polybrene (Fig. 2B). In comparison, 25% LacZ-positive cells could be achieved by infection at an MOI of 2.5 without Polybrene, or at an MOI of 1 in the presence of Polybrene. Polybrene at concentrations up to 40 m g/ml did not affect cell viability as measured by trypan blue exclusion assay.

Effect of irradiation on infection of GL261 cells with HSVlacZ Infection with helper virus-free HSVlacZ amplicon stocks (MOI 5 or 50) for 4 hr without Polybrene was not cytotoxic to GL261 cells. Cell counts and trypan blue staining 2 and 4 days after infection did not differ significantly from viable cell counts of uninfected cultures (Fig. 3A). Irradiation with 15 Gy prior to infection did not alter the growth rate of cells, as compared with nonirradiated cultures (data not shown). Irradiation with 35 Gy arrested the growth of GL261 cells (Fig. 3B), and subsequent infection with HSVlacZ at an MOI of 5 or 50 did not cause significant cytotoxicity . The timing of irradiation was also evaluated in relation to infection efficiency. Some cultures were first iradiated and them infected 1 day later; other cultures were first infected, and then irradiated 1 day later. At all MOIs tested (5, 10, and 20), there was no significant difference between these two types of cultures regarding the transduction efficiency, as measured by the percentage of LacZ-positive cells 2 days after infection (Fig. 3C). Also, transduction efficiency was not different from that of GL261 cells that had not been irradiated when they were infected with HSVlacZ (Fig. 2A).

Infection of GL261 with HSVGM and GM-CSF expression over time Mouse GL261 glioma cells (10,000 cells/well in 24-well plates, previously irradiated with 35 Gy) were infected with various doses of helper virus-free HSVGM vector over 4 hr. After infection the cells were washed and fresh medium was added. GM-CSF was measured in the medium 24 hr after infection. Cumulative GM-CSF secretion increased proportion ally to the vector dose (Fig. 4A). Highest GM-CSF secretion was found for 500- m l vector stocks (about 5 3 10 5 virion DNA equivalents per 104 cells), with 1026 ng/106 cells/24 hr. GMCSF secretion by nonirradiated cells was not significantly different as compared with irradiated cells (data not shown), and irradiated and/or HSVlacZ-infected GL261 cells did not secrete detectable GM-CSF (level of detection by ELISA: 1 ng/10 6 cells/24 hr). Time course experiments with sampling and replacement of the medium of infected GL216 cells every 24 hr revealed that GM-CSF values of the supernatant decreased rapidly within a few days after infection. GM-CSF secretion by day 3 after infection had fallen to 1% of the GM-CSF secretion during the first 24 hr after infection (Fig. 4B). In supernatants sampled on day 7 after infection, GM-CSF levels were undetectable .

In vivo vaccination GL261 glioma cells infected with GM-CSF-encoding , helper virus-free amplicon vector, HSVGM, were used for vac-

1433

HSV AMPLICONS FOR VACCINATION THERAPY

B

A

FIG. 2. Infection of GL261 glioma cells with lacZ amplicon vector (HSVlacZ). GL261 cells were infected with various MOIs of helper virus-free HSVlacZ amplicon vector over a 4-hr period, with LacZ staining carried out 1 day after infection. (A) Effect of increasing MOI, in transducing units per cell; (B) effect on infection efficiency of Polybrene at various concentration s during the 4-hr infection period at an MOI of 2.5.

cination studies. For infection, 5 3 10 5 GL261 cells per animal were plated, irradiated 24 hr later, and then infected with 8 ml of amplicon vector stock for 4 hr. The MOI was equivalent to 160 m l (estimated at 16 3 10 4 virion DNA equivalents) per 10 4 cells in Fig. 4A. Immediately after infection , cells were thoroughly washed, trypsinized, and subsequentl y injected subcutaneous ly (500,000 cells/animal) into the left flank of C57BL/6 mice (n 5 10 animals). A parallel aliquot of HSVGM-infected cells placed in culture produced 303 6 16 ng of GM-CSF/10 6 cells during the first 24 hr after infection . As negative controls, animals were injected subcutaneous ly with uninfected GL261 cells (n 5 11 animals), cells infected with 8 ml of HSVlacZ stock (MOI of 16; n 5 9), or cells infected with 8 ml of HSVPrPUC stock (n 5 4). Aliquots of uninfected or HSVlacZ-infected cells placed in culture did not show any GM-CSF expression. As a positive control, animals were injected with the same number of retrovirally transduced , GM-CSF-secreting GL261 cells (RV-GM, n 5 10), which produce about 190 ng of GM-CSF/10 6 cells over 24 hr (Herrlinger et al., 1997). Median survival time (MST) of animals vaccinated with uninfected GL261 cells was 25 days (range, 18 to . 80 days (Fig. 5). Median survival was not significantly prolonged in control animals receiving cells infected with HSPVrPUC (MST, 28 days; range, 23 to 39 days) or HSVlacZ (MST, 31 days; range, 24 to . 80 days) as compared with animals vaccinated with wild-type GL261 cells (p 5 0.4 and 0.24, respectively). However, MST was markedly prolonged in animals vaccinated with GM-CSF-transduced cells as compared with animals vaccinated with wild-type cells, with 50% longterm survivors on day 80 with retrovirally infected cells (p 5 0.03) and 60% long-term survivors with HSVGM ampliconinfected cells (p 5 0.015).

Transduction of human ex vivo glioblastoma cells All three cultures of ex vivo human glioblastoma cells contained 65.7 6 4.7% dividing cells as determined by BrdU stain-

ing. Human glioblastoma cells were highly infectable with helper-free HSV-1 amplicon preparations since, on infection with helper-free HSVlacZ preparations at an MOI of 1, 97 6 1.4% of cells stained positive for b -galactosidase. b -Galactosidase-positive cells comprised both dividing, i.e., BrdU-positive, cells (67.5 6 10.6%) and nondividing, BrdU-negative cells (29.5 6 9.2%; Fig. 6) in the cultures .

DISCUSSION This article describes helper virus-free HSV-1 amplicon vectors as a new, safe, and efficient vector for vaccination therapy of glioma, and potentially other cancers. Helper virus-free amplicon vectors efficiently infected glioma cells and led to highlevel GM-CSF expression for the first few days after infection . Levels of GM-CSF secretion from amplicon-infect ed GL261 cells and the vaccination effect were similar to those observed with retrovirally transfected GL261 cells. For clinical immunotherapy, helper-virus free amplicon vectors are attractive since, in contrast to retroviral vectors, they should allow for immediate, efficient transduction of glioma cells removed from patients independently of cell proliferation . Helper virus-free HSV-1 amplicon vectors were chosen for these experiments because, for clinical cytokine-enhance d vaccination therapy of glioma, immediate ex vivo transduction of glioma cells independent of their cell cycle will be important. Using helper virus-free amplicon vectors, the expression level of GM-CSF during the first days postinfection was markedly beyond 30 ng/10 6 cells/24 hr, a level that has been reported to be necessary for a potent vaccination effect (Jaffee et al., 1996). For a future clinical application of amplicon vectors, helper virus-free amplicon preparations appear to be attractive since both dividing and nondividing glioma cells can be infected (Fig. 6). Infection of more than 90% of cells occurred at a low MOI of 1 (titered on highly permissive Vero cells), allowing the use of relatively low titers for infection in a clinical paradigm. The

1434

HERRLINGER ET AL.

A

B

C

FIG. 3. Influence of infection with HSVlacZ amplicon vectors and/or irradiation with 35 Gy on viability and transduction of GL261 cells. (A) Cell count and trypan blue staining for viable cells at various time points without infection and after infectio n on day 0 with HSVlacZ supernatant (MOI of 5 or 50). (B) Viable cell counts at various time points after irradiation with 35 Gy and infection with HSVlacZ supernatant (MOI of 5 or 50). (C) Percentage of LacZ-positive cells on day 2 after infection with HSVlacZ (MOI of 5, 10, or 20) on day 0 and irradiation with 35 Gy on day 1 (infection before radiation), or irradiation with 35 Gy on day 0 and infection with HSVlacZ on day 1 (MOI 5, 10, or 20) (radiation before infection) .

infectability of human primary glioblastoma cultures appears to be much greater than the infectability of some mouse (e.g., GL261; Fig. 2) and rat glioma cell lines (Herrlinger et al., 1998). In contrast to other vectors suitable for transduction of nondi-

viding tumor cells, such as vaccinia virus (Qin and Chatterjee , 1996a,b) and adenovirus vectors, which are both generated by homologous recombination into complicated virus genomes, plasmid-based HSV-1 amplicon vectors can be easily con-

1435

HSV AMPLICONS FOR VACCINATION THERAPY

A

B

FIG. 4. GM-CSF-secretion of irradiated (35 Gy) GL261 cells after infection with HSVGM stocks. (A) GM-CSF (ng/106 cells/24 hr) as determined by ELISA in the conditioned medium of 10,000 GL261 cells 24 hr after infection with various volumes of HSVGM for 4 hr in the presence of Polybrene (40 m g/ml). (B) GM-CSF (ng/106 cells/24 hr) as determined by ELISA in the medium of 10,000 GL261 cells at various time points after infection with 500 m l of virus stock. Medium was replaced every 24 hr. structed and modified by recombinant DNA technology. However, it is still easier to generate recombinant adenovirus vector stocks through infection than amplicon vector stocks through transfection. Also, adenovirus vectors can be produced to higher

titers than HSV-1 amplicon vectors. However, helper virus-free amplicon vectors can now be produced at titers close to 10 9 TU/ml by optimizing the packaging protocol and by the addition of previral particles (Sun et al., 1999).

FIG. 5. Survival of animals vaccinated 7 days prior to intracranial implantation of 10 6 wild-type glioma cells. Animals were subcutaneously vaccinated with 5 3 10 5 uninfected, wild-type (wt) GL261 cells (GL261 wt), HSVPrPUC-infected GL261 cells (HSVPrPUC), HSVlacZ-infected GL261 cells (HSVlacZ), HSVGM-infected GL261 cells (HSVGM), or GL261 cells that were retrovirally transduced to secrete GM-CSF (RV-GM). Intracranial GL261 gliomas were generated 7 days after vaccination by injection of 10 6 wild-type GL261 cells into the right frontal lobe. Results were pooled from two similar experiments. Survival was not significantly different after vaccination with HSVPrPUC (empty vector)- or HSVlacZ-infected cells as compared with vaccination with GL261 wt (*p 5 0.4 or 0.24, respectively). Survival was significantly prolonged in animals vaccinated with RVGM cells or HSVGM-infected cells as compared with vaccination with GL261 wt (*p 5 0.03 or 0.015, respectively) or HSVPrPUC-infected cells (*p 5 0.02 or 0.027, respectively). *Log rank test.

1436

FIG. 6. Percentage of LacZ 1 and BrdU 1 cells in primary glioblastoma cultures infected with HSVlacZ at an MOI of 1 and stained for b -galactosidase expression and BrdU incorporation. Cells were infected with HSVlacZ at an MOI of 1 and stained for b -galactosidase expression by histochemistry and for BrdU incorporation using a fluorescein-conju gated antibody. Values are given as percentages of total cells. Averages and standard errors of the mean are calculated from two independent experiments . The rapid decrease in GM-CSF secretion within a few days after infection, observed with the helper virus-free amplicon vectors, does not appear to pose a limitation to the applicabil ity of these vectors for vaccination purposes, since a strong vaccination effect was nevertheless observed in animal studies. The rapid decline in GM-CSF release may reflect the loss of amplicon DNA over time and/or the downregulation of the HSV1 IE 4/5 promoter driving GM-CSF expression in HSVGM due to degradation of the VP16 virion protein, which trans-activates this promoter. The length of transgene expression might be extended by use of HSV/EBV (Wang and Vos. 1996) or HSV/AAV (Johnston et al.., 1997) hybrid vectors and/or other strong viral or mammalian promoters. Another reason for the apparent marked drop in GM-CSF release from amplicon-in fected GL261 cells might be that GM-CSF secretion on day 1 is overestimated by the portion of GM-CSF brought in with the infection solution . To increase the level of transgene expression with a given dose of helper virus-free amplicon vector, Polybrene was added during infection. With Polybrene at 40 m g/ml, infection efficiency was increased by 2.5-fold (Fig. 2B). Polybrene is known to increase the transduction efficiency by other viruses that also rely on receptor binding for entry into the target cell, such as retroviruses (Andreadis and Palsson, 1997) and adenoviruse s (Arcasoy et al., 1997), as well as by lipofection (Abe et al., 1998). Irradiation of target cells before or after infection with helper virus-free amplicon vectors did not prove to increase transduction efficiency. Irradiation has been described as a means to enhance the transduction efficiency of adeno-associ ated virus (Alexander et al., 1994 and 1996) and adenoviru s

HERRLINGER ET AL. (Zeng et al., 1997) vectors. Also, the transduction of fibroblasts by retroviral plasmids was greatly enhanced by ionizing radiation, with the most favorable timing of infection being a few hours after irradiation at the peak of the DNA damage repair process (Stevens et al., 1996). GM-CSF secretion from GL261 glioma cells infected with helper-virus free amplicon vectors was high enough (303 ng/10 6 cells/24 hr) to induce as strong a vaccination effect in the GL261 mouse model as conventional retrovirus-transf ected GM-CSFsecreting GL261 cells (190 ng/10 6 cells/24 hr [Herrlinger et al., 1997]; Fig. 5). Most of the vaccination effect appears to be due to the enhancing effect of GM-CSF on tumor antigen presentation (Banchereau and Steinman, 1998). Infection with HSV1 amplicon vectors that do not contain any transgene (HSVPrPUC) did not enhance the immunogenic ity of GL261 cells in this vaccination paradigm, whereas the expression of a foreign antigen, E. coli b -galactosidase, had a slight although nonsignificant immunostimulatory effect (Fig. 5). This finding correlates well with reports from other groups showing high immunogenicity of HSV-1-mediated (Brubaker et al., 1996) and adenovirus- mediated (Tripathy et al., 1996) expression of b -galactosidase, as well as immunogenicity of other transgene products such as that of the neomycin resistance gene (Tapscott et al., 1994). Successful vaccination therapy with HSV-1 amplicon vectors has been corroborated with IL-2 as the transgene in a hepatoma model (Tung et al., 1996). However, in this helper viruscontaining system, the helper virus may have cytotoxic effects on the cells used for vaccination. Also, de novo synthesis of HSV-1 proteins by the helper virus could be immunogenic in its own right as demonstrated in vaccination experiments using a conditionally replicating HSV-1 virus and a colon cancer cell line (Toda et al., 1999). This argument also applies to the successful vaccination experiments using a lytic gH 2 HSV-1 mutant expressing GM-CSF in a leukemia model (Dilloo et al., 1997). In contrast to helper virus-containing amplicon vectors and gH 2 HSV-1 vectors, nontoxic infection of target cells without de novo synthesis of immunogenic HSV-1 proteins can be provided by using helper virus-free HSV-1 amplicon vectors . In conclusion, helper-free amplicon vectors appear to be promising tools for an effective immunogene therapy of cancer, with a transgene capacity up to 20 kb in the standard amplicon. These vectors can provide efficient, nontoxic transduc tion of dividing and nondividing target cells. Thus, helper virus-free amplicon vectors may be able to infect resected tumor tissue without culture of cells, thus retaining the full antigenic profile of the tumor.

ACKNOWLEDGMENTS The authors thank Dr. Glenn Dranoff (Dana Farber Cancer Center) for helpful discussions and Ms. Suzanne McDavitt for assistance in preparing the manuscript. This work was supported by a grant from the Deutsche Forschungsge meinschaf t (U.H.), the Gertrud Reemtsma Stiftung (A.H.J.), the Swiss National Research Foundation (C.F.), and the Wilhelm-TönnisScholarship from the Deutsche Gesellschaft für Neurochirurgi e (C.W.), and by grants from the National Cancer Institute, CA69246 (X.O.B.), Junta Nacional de Investiga ção Científica e

HSV AMPLICONS FOR VACCINATION THERAPY Tecnológica (Portugal) Programa Praxis XXI/BD/3248/94 (M.S.E.), the Federal Ministry of Education and Research of Germany, BioRegio 0311661/1402 (N.G.R.), the State of Saxony-Anhalt (2794A /0087H; N.G.R.), Novartis Pharma (N.G.R.), and the E.-Kröner-Fresenius Foundation, Bad Homburg, Germany (N.G.R.).

REFERENCES ABE, A., MIYANOHARA, A., and FRIEDMANN, T. (1998). Polybrene increases the efficiency of gene transfer by lipofection. Gene Ther. 5, 708–711. AKBASAK, A., OLDFIELD, E.H., and SARIS, S.C. (1991). Expression and modulation of major histocompatibility antigens on murine primary brain tumor in vitro. J. Neurosurg. 75, 922–929. ALEXANDER, I.E., RUSSELL, D.W., and MILLER, A.D. (1994). DNA-damaging agents greatly increase the transduction of nondividing cells by adeno-associated virus vectors. J. Virol. 68, 8282– 8287. ALEXANDER I.E., RUSSELL, D.W., SPENCE, A.M., and MILLER, A.D. (1996). Effects of gamma irradiation on the transduction of dividing and nondividing cells in brain and muscle of rats by adenoassociated virus vectors. Hum. Gene Ther. 7, 841–850. ANDREADIS, S., and PALSSON, B.O. (1997). Coupled effects of Polybrene and calf serum on the efficiency of retroviral transduction and the stability of retroviral vectors. Hum. Gene Ther. 8, 285–291. ANDREANSKY, S.S., HE, B., GILLESPIE, G.Y., SOROCEANU, L., MARKERT, J., CHOU J., ROIZMAN, B., and WHITLEY, R.J. (1996). The application of genetically engineered herpes simplex viruses to the treatment of experimental brain tumors. Proc. Natl. Acad. Sci. U.S.A. 93, 11313–11318. ARCASOY, S.M., LATOCHE, J.D., GONDOR, M., PITT, B.R., and PILEWSKI, J.M. (1997). Polycations increase the efficiency of adenovirus-mediated gene transfer to epithelial and endothelial cells in vitro. Gene Ther. 4, 32–38. AUSMAN, J.I., SHAPIRO, W.R., and RALL, D.P. (1970). Studies on the chemotherapy of experimental brain tumors: development of an experimental model. Cancer Res. 30, 2394– 2400. BANCHEREAU, J., and STEINMAN, R.M. (1998). Dendritic cells and the control of immunity. Nature (London) 392, 245–252. BERGOLD, P., CASACCIA-BONNEFIL, P., XIU-LIU, Z., and FEDEROFF, H.J. (1993). Transsynaptic neuronal loss induced in hippocampal slice cultures by a herpes simplex virus vector expressing the GluR6 subunit of the kainate receptor. Proc. Natl. Acad. Sci. U.S.A. 90, 6165– 6169. BRUBAKER, J.O., THOMPSON, C.M., MORRISON, L.A., KNIPE, D.M., SIBER, G.R., and FINBERG, R.W. (1996). Th1-associated immune responses to beta-galactosida se expressed by a replication defective herpes simplex virus. J. Immunol. 157, 1598–1604. CHIOCCA, E.A., CHOI, B.B., CAI, W.Z., DELUCA, N.A., SCHAFFER, P.A., DIFIGLIA, M., BREAKEFIELD, X.O., and MARTUZA, R.L. (1990). Transfer and expression of the lacZ gene in rat brain neurons mediated by herpes simplex virus mutants. New Biol. 2, 739–746. CUNNINGHAM, C., and DAVISON, A.J. (1993). A cosmid-based system for constructing mutants of herpes simplex virus type 1. Virology 197, 116–124. DILLOO, D., RILL, D., ENTWISLE, C., BOURSNELL, M., ZHONG, W., HODLEN, W., HOLLADAY, M., INGLIS, S., and BRENNER, M. (1997). A novel herpes vector for the high-efficiency transduction of normal and malignant human hematopoietic cells. Blood 89, 119–127. DRANOFF, G., JAFFEE, E., LAZENBY, A., GOLUMBEK, P., LEVITSKY, H., BROSE, K., JACKSON, V., HAMADA, H., PARDOLL,

1437 D., and MULLIGAN, R.C. (1993). Vaccination with irradiated tumor cells engineered to secrete murine granulocyte-ma crophage colony-stimulat ing factor stimulates potent, specific, and long-lasting antitumor immunity. Proc. Natl. Acad. Sci. U.S.A. 90, 3539– 3543. FAHKRAI, H., DORIGO, O., SHAWLER, D.L., LIN, H., MERCOLA, D., BLACK, K.L., ROYSTON, I., and SOBOL, R.E. (1996). Eradication of established intracranial rat gliomas by transforming growth factor beta antisense gene therapy. Proc. Natl. Acad. Sci. U.S.A. 93, 2909– 2914. FRAEFEL, C., SONG, S., LIM, F., LANG, P., YU, L., WANG, Y., WILD, P., and GELLER, A.I. (1996). Helper virus-free transfer of herpes simplex virus type 1 plasmid vectors into neural cells. J. Virol. 70, 7190– 7197. FRAEFEL, C., JACOBY, D.R., LAGE, C., HILDERBRAND, H., CHOU, J.Y., ALT, F.W., BREAKEFIELD, X.O., and MAJZOUB, J.A. (1997). Gene transfer into hepatocytes mediated by helper virus free HSV/AAV hybrid vectors. Mol. Med. 3, 813–825. FRAEFEL, C., BREAKEFIELD, X.O., and JACOBY, D.R., (1998). HSV-1 amplicon. In Gene Therapy for Neurological Disorders and Brain Tumors. E.A. Chiocca, X.O. Breakefield, eds. (Humana Press, Totowa, NJ) pp. 63–82. GELLER, A.I., and BREAKEFIELD, X.O. (1988). A defective HSV1 vector expresses Escherichia coli b -galactosidase in cultured peripheral neurons. Science 241, 1667– 1669. HERRLINGER, U., KRAMM, C.M., JOHNSTON, K.M., LOUIS, D.N., FINKELSTEIN, D., REZNIKOFF, G., DRANOFF, G., BREAKEFIELD, X.O., and YU, J.S. (1997). Vaccination for experimental gliomas using GM-CSF transduced glioma cells. Cancer Gene Ther. 4, 345–352. HERRLINGER, U., KRAMM, C.M., ABOODY-GUTERMAN, K.S., SILVER, J.S., IKEDA, K., JOHNSTON, K.M., PECHAN, P.A., BARTH, R.F., FINKELSTEIN, D., CHIOCCA, E.A., LOUIS, D.N., and BREAKEFIELD, X.O. (1998). Pre-existing herpes simplex virus 1 (HSV-1) immunity decreases, but does not abolish, gene transfer to experimental brain tumors by a HSV-1 vector. Gene Ther. 5, 809–819. HERRLINGER, U., JACOBS, A., AGHI, M., SCHUBACK, D.E., and BREAKEFIELD, X.O. (2000). HSV-1 vectors for gene therapy of experimental CNS tumors. In Methods in Molecular Medicine: Gene Therapy Protocols. P. Robbins, ed. (Humana Press, Totowa, NJ) pp. 287–312 JAFFEE, E.M., THOMAS, M.C., HUANG, A.Y.C., HAUDA, K.M., LEVITSKY, H.I., and PARDOLL, D.M. (1996). Enhanced immune priming with spatial distribution of paracrine cytokine vaccines. J. Immunother. 19, 176–183. JOHNSTON, K.M., JACOBY, D., PECHAN, P.A., FRAEFEL, C., BORGHESANI, P., SCHUBACK, D., DUNN, R.J., SMITH, F.I., and BREAKEFIELD, X.O. (1997). HSV/AAV hybrid amplicon vectors extend transgene expression in human glioma cells. Hum. Gene Ther. 8, 359–370. KLEIHUES P., BURGER, P.C., PLATE, K.H., OHGAKI, H., and CAVENEE, W.K. (1997). Glioblastoma. In Pathology and Genetics of Tumours of the Nervous System. P. Kleihues and W.K. Cavenee, eds. (International Agency for Research on Cancer, Lyon, France) pp. 16–24. LICHTOR, T., GLICK, R.P., KIM, T.S., HAND, R., and COHEN, E.P. (1995). Prolonged survival of mice with glioma injected intracerebrally with double cytokine-secre ting cells. J. Neurosurg. 83, 1038– 1044. QIN, H.X., and CHATTERJEE, S.K. (1996a). Cancer gene therapy using tumor cells infected with recombinant vaccinia virus expressing GM-CSF. Hum. Gene Ther. 7, 1853– 1860. QIN, H.X., and CHATTERJEE, S.K. (1996b). Construction of recombinant vaccinia virus expressing GM-CSF and its use as tumor vaccine. Gene Ther. 3, 59–66.

1438 SAEKI, Y., ICHIKAWA, T., SAEKI, A., CHIOCCA, E.A., TOBLER, K., ACKERMANN, M., BREAKEFIELD, X.O., and FRAEFEL, C. (1998). Herpes simplex virus type 1 DNA amplified as bacterial artificial chromosome in Escherichia coli: Rescue of replication-com petent virus progeny and packaging of amplicon vectors. Hum. Gene Ther. 10, 2787–2794. SMITH, C.A., BATES, P., RIVERA-GONZALEZ, R., GU, B., and DELUCA, N.A. (1993). ICP4, the major transcriptional regulatory protein of herpes simplex virus forms a tripartite complex with TATA-binding protein and TFIIB. J. Virol. 67, 4676–4687. STEVENS, C.W., ZENG, M., and CERNIGLIA, G.J. (1996). Ionizing radiation greatly improves gene transfer efficiency in mammalian cells. Hum. Gene Ther. 7, 1727– 1734. SUN, M., ZHANG, G.R., YANG, T., YU, L., and GELLER, A.I. (1999). Improved titers for helper virus-free herpes simplex virus type 1 plasmid vectors by optimization of the packaging protocol and addition of noninfectious herpes simplex virus-related particles (previral DNA replication enveloped particles) to the packaging procedure. Hum. Gene Ther. 10, 2005– 2011. TAPSCOTT, S.J., MILLER, A.D., OLSON, J.M., BERGER, M.S., GROUDINE, M., and SPENCE, A.M. (1994). Gene therapy of rat 9L gliosarcoma tumors by transduction with selectable genes does not require drug selection. Proc. Natl. Acad. Sci. U.S.A. 91, 8185– 8189. TODA, M., RABKIN, S.D., KOJIMA, H., and MARTUZA, R.L. (1999). Herpes simplex virus as an in situ cancer vaccine for the induction of specific anti-tumor immunity. Hum. Gene Ther. 10, 385–393. TRIPATHY S.K., BLACK, H.B., GOLDWASSER, E., and LEIDEN, J.M. (1996). Immune responses to transgene-enco ded proteins limit the stability of gene expression after infection of replication-defe ctive adenovirus vectors. Nature Med. 2, 545–550. TUNG, C., FEDEROFF, H.J., BROWNLEE, M., KARPOFF, H., WEIGEL, T., BRENNAN, M.F., and FONG, Y. (1996). Rapid pro-

HERRLINGER ET AL. duction of interleukin-2-se creting tumor cells by herpes simplex virus-mediated gene transfer: Implications for autologous vaccine production. Hum. Gene Ther. 7, 2217– 2224. TURNER, D.L., SNYDER, E.Y., and CEPKO, C.L. (1990). Lineageindependent determination of cell type in the embryonic mouse retina. Neuron 4, 833–845. WANG, S., and VOS, J.M. (1996). A hybrid herpesvirus infectious vector based on Epstein–Barr virus and herpes simplex virus type 1 for gene transfer into human cells in vitro and in vivo. J. Virol. 70, 8422– 8430. YU, J.S., BURWICK, J.A., DRANOFF, G., and BREAKEFIELD, X.O. (1997). Gene therapy for metastatic brain tumors by vaccination with granulocyte-m acrophage colony-stimulati ng factor-transduced tumor cells. Hum. Gene Ther. 8, 1065– 1072. ZENG M., CERNIGLIA, G.J., ECK, S.L., and STEVENS, C.W. (1997). High-efficiency stable transfer of adenovirus into mammalian cells using ionizing radiation. Hum. Gen. Ther. 8, 1025–1032.

Address reprint requests to: Dr. Xandra O. Breakefield Massachusetts General Hospital East Department of Molecular Neurogenetic s 13th St., Bldg. 149 Charlestown, MA 02129 E-mail: breakefield @helix.mgh.harva rd.edu Received for publication August 16, 1999; accepted after revision April 20, 2000.