HUMAN GENE THERAPY 13:1115–1125 (June 10, 2002) © Mary Ann Liebert, Inc. Adeno-Associated Viral Vectors Penetrate Human Solid Tumor Tissue In Vivo More Effectively than Adenoviral Vectors PER ØYVIND ENGER,1 FRITS THORSEN, 1 PER EYSTEIN LØNNING, 2,3 ROLF BJERKVIG,1–3 and FRANK HOOVER 3 ABSTRACT The transduction efficiencies of adeno-associated viral vectors (AAV, serotype 2) and adenovirus vectors (ADV, serotype 5) were examined in three different models of cancer. First, we used flow cytometry to quantitate AAV-GFP or ADV-GFP transduction in 13 cell lines derived from malignant tissue (6 gliomas, 6 mammary cancers, and 1 leukemia). These experiments showed variable transduction efficiency (0%–81%) between the cell lines, with ADV being more effective compared to AAV in 9 of 13 cell lines. Second, spheroids prepared from human glioblastomas were infected with ADV or AAV expressing GFP or lacZ cassettes, and after 2 weeks, uniform reporter gene expression was observed on the spheroid. Whereas AAV produced consistent transduction throughout the spheroids, ADV infection was mainly limited to the outer cell layers of the spheroids, suggesting that AAV were more efficient at penetrating solid tumor tissue. Third, human biopsies from glioblastoma multiforme patients were xenografted into nude rats and grown for 4 weeks followed by viral vector injection. Combined use of high-resolution magnetic resonance imaging (MRI) and histologic analysis allowed the identification of transduced cells and their spatial distribution within the tumors. AAV-mediated transgene expression was observed in cell clusters through the entire tumor, while ADV-mediated transduction was restricted to cells at the tumor periphery. Thus, while AAV and ADV vectors may infect tumor-derived cell lines to a similar degree, AAV penetrated glioblastoma spheroids and xenografts more efficiently compared to ADV vectors. These results suggest that AAV may be suitable for therapeutic gene delivery to malignant tumors. OVERVIEW SUMMARY Sustained therapeutic gene expression is a major limitation to cancer gene therapy. Thus, there is a need to identify efficient delivery systems. Here we demonstrate that both adeno-associated vectors (AAV, serotype 2) and adenovirus vectors (ADV, serotype 5) can transduce a broad range of glioma and mammary cancer cell lines grown as monolayers. We show that AAV effectively penetrated human glioma biopsy spheroids, as indicated by the presence of reporter gene positive cells throughout the tissue. In contrast, ADV-transduced cells were limited to the outer layers of the spheroids. To examine transduction in human material in vivo, we injected the vectors expressing GFP and lacZ into human glioblastomas xenografted into the nude rat brain. Fluores- cence microscopy and lacZ staining revealed clusters of AAVtransduced cells distributed within the tumors, while ADVmediated transduction was restricted to the tumor periphery. Taken together, these data suggest a positive potential for AAV vector systems in cancer gene therapy. INTRODUCTION G represents a novel and attractive alternative to conventional cancer therapies. Although in its infancy, gene therapy has already begun to provide encouraging results in human clinical trials (Somia and Verma, 2000). A major limitation to this technology has been sufficient and stable gene transfer to the target cells in solid tissue (Somia and Verma, 1Department ENE THERAPY of Anatomy and Cell Biology, University of Bergen, Bergen, Norway. of Oncology, Haukeland Hospital, Bergen, Norway. 3Gene Therapy Program, Bergen, Norway. 2Department 1115 1116 2000; Rainov and Kramm, 2001). Therefore, much of the current effort in the field is spent on identifying and developing suitable gene delivery systems of both viral and nonviral origin. Most clinical trials in cancer gene therapy have favored retroviral and adenoviral vectors for gene transfer (www. wiley.co.uk/genmed-2001). While these systems each have distinct advantages, current evidence suggests that there are significant limitations regarding their transduction efficiency in solid tumor tissue (Benedetti et al., 1997; Thorsen et al., 1997; Puumalainen et al., 1998; Enger et al., 1999; Sandmair et al., 1999; Kuriyama et al., 2001). Recombinant adeno-associated viruses (AAV) have been gaining favor as a reliable gene transfer system (Fisher et al., 1997; Monahan and Samulski, 2000a,b; Smith-Arica and Bartlett, 2001). These vectors are derived from parvoviruses and have small capsids (25 nm) containing single-stranded genomes, and have never been associated with any disease conditions in humans (Muzyczka et al., 1984; Snyder, 1999). Clinical trials in humans have demonstrated a safe toxicity profile, and additional studies are in progress (Kay et al., 2000; Flotte et al., 1996). Despite having many hallmarks necessary for cancer gene therapy, few studies have evaluated AAV for use in the oncologic setting. These studies have been limited to cell lines and cell lines transplanted into animals (Qazilbash et al., 1997; Kunke et al., 2000; Veldwijk et al., 2000). We have hypothesized that AAV’s small capsid size and innate antioncogenic properties (reviewed in Ponnazhagan, et al. ) could augment its potential for transducing solid tumor tissue and overcome some of the additional difficulties displayed by other vector systems (Somia and Verma, 2000). The elucidation of this information would be important to establish AAV’s future role in cancer therapy. To begin to address this issue, we have compared the transduction efficiencies of AAV (serotype 2) and ADV (serotype 5) in leukemia, glioma, and mammary cancers. These cancer subtypes were chosen because they display a variety of tumor biology and clinical behavior. Gliomas show a local invasive growth but rarely spread outside of the central nervous system (CNS), whereas mammary cancer is characterized by local invasive growth as well as metastasis. In contrast to these two diseases, leukemia is a nonsolid systemic disease from the outset. In this study, we designed experiments to examine transduction efficiency in three different oncologic models using increased levels of complexity that mimic the pathologic state in situ. First, we quantitatively assessed the transduction efficiency of ADV and AAV in 13 different cancer cell lines by flow cytometry to assess the targetable range of the cancers. Second, we compared the efficiency of these vector systems in human tumor spheroids to examine penetration into solid tumor tissue. Third, we generated a unique in vivo model system in which the human glioblastoma tumor tissue was grown in nude rats followed by vector injection. To identify transduced cells and their spatial distribution within human tumor tissue in vivo, we performed highresolution magnetic resonance imaging (MRI) and histologic analyses. Collectively, we observed that AAV and ADV vectors could transduce a large range of malignant cells, however, AAV expressed distinct superiority in penetrating solid tumor tissue. These results clearly demonstrate that AAV have potential to be an effective delivery system for cancer gene therapy. ENGER ET AL. MATERIALS AND METHODS Viral vectors Recombinant replication-defective ADV shuttle vectors were obtained from Quantum Biotechnology Inc. (Montreal, Canada) and modified by oliogonucleotides to expand the multiple cloning site and to insert the EGFP gene (F.H. Hoover, unpublished data). Recombinant replication-defective ADV vectors (E1/E3-deficient) were generated by homologous recombination in a human embryonic kidney cell HEK 293. The ADV lacZ virus was propagated from stocks obtained from Quantum Biotechnologies. Virus was purified from cells according to published procedures (Graham et al., 1995). Viral titers were routinely 1 3 1012 virus particles per milliliter and were determined according to the QBI protocol and protein coat determination with UV adsorption at 260 nm. Recombinant AAV expressing the GFP gene and lacZ genes were gifts from Avigen, Inc. (Alameda, CA) that were prepared in an adenovirus-free system using a triple transfection technique with the pHLP19 and pLadeno5 helper vectors (Matsushita et al., 1998; Burton et al., 1999). AAV titers were 5.6 3 1011 and 8.2 3 1011 particles per milliliter for AAV-GFP and AAV-lacZ, respectively. Cell culture All cell cultures were kept at 37°C in a standard tissue-culture incubator (100% relative humidity, 5% CO2), except for the cell line MDA-MB-468, which was cultured in a CO2 2free incubator. The human glioma cell lines A172 and U373 were obtained from the American Type Culture Collection (ATCC, Rockville, MD). The human glioma cell line D-37 was kindly supplied by Dr. D.D. Bigner (Duke University Medical Center, Durham, NC). HF-66 is a cell line derived from a patient with glioblastoma multiforme (established at Henry Ford Midwest Neuro-Oncology Center, Detroit, MI). The GaMg cell line was obtained from a 42-year-old female, and histologically identified as a glioblastoma (Akslen et al., 1988). The BT4 C tumor cell line is an ethylnitrosurea-induced rat glioma, established in syngeneic BD-IX rats (Lærum et al., 1977). All glioma cell lines, as well as the biopsy tissue were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with penicillin (100 IU/ml), streptomycin (100 g/ml), 2% L -glutamine, 10% heat-inactivated newborn calf serum, and 4 times the prescribed concentration of nonessential amino acids, hereafter called complete medium (newborn calf serum: Sigma, Steinheim, Germany, all other reagents: Bio Whittaker, Verviers, Belgium). The human breast carcinoma cell lines MCF7, AU565, MDA-MB-468, HCC1395, HCC1569, and T-47D were all obtained from ATCC. The MCF7 cell line was cultured in DMEM supplemented with 10% fetal calf serum (BioWhittaker). The AU565, HCC1395, HCC1569 cells were grown in RPMI 1640 supplemented with 10% fetal calf serum (FCS). MDA-MB-468 cells were grown in L15 medium supplemented with 10% FCS. T-47D cells were grown in RPMI 1640 supplemented with 10% FCS and 200 U/ml of insulin (Sigma, St. Louis, MO). The leukemia cell line KG-1a was grown as a suspension culture in DMEM supplemented with 20% FCS and was obtained from ATCC. A brief description of these cell lines is provided in Table 1. 1117 AAV TRANSDUCTION IN HUMAN CARCINOMAS TABLE 1. ATTRIBUTES Cell line OF CANCER CELL LINES Histologic type A172 Human glioblastoma BT4C Rat glioblastoma D-37 Human glioblastoma GaMg Human glioblastoma HF-66 Human glioblastoma U373 Human glioblastoma Au565 Adenocarcinoma breast, human HCC1395 Ductal carcinoma, breast, human Meta-plastic, breast, human HCC1569 MCF7 MDA-MB-468 T-47D KG-1a Adenocarcinoma, breast, human Adenocarcinoma, breast, human Ductal carcinoma, breast, human Myeloblast cell line, human Patient 1 Glioblastoma Patient 2 Glioblastoma Patient 3 Glioblastoma Patient 4 Glioblastoma AND HUMAN GLIOBLASTOMA TUMOR SPHEROIDS Characteristics Nontumorogenic, hypertriploid, EGFR positive, and a3,vb1,4,5 integrin positive (Knott et al., 1998). Tumorogenic in BD IX rats, expresses N-CAM (Lærum et al., 1977). Forms spheroids, invades brain aggregates in vitro, EGFR, a2,3,4,5,6,v and b1,3,4,5 integrin positive (Haugland et al., 1997). Forms spheroids, invades brain aggregates in vitro, EGFR, a2,3,4,5,6,v and b1,3,4,5 integrin positive (Haugland et al., 1997). Forms spheroids, invades rat brain aggregates in vitro, a3,4,6,v and b1,4,5 integrin positive (Knott et al., 1998). Tumorogenic in nude mice, EGFR a3,4,6,v and b1,4,5 integrin positive (Knott et al., 1998). EGFR-positive, estrogen-receptor negative, overexpresses HER-2/neu and erbB-2 gene, HER-3, HER-4, and p53 positive (ATCC). Estrogen receptor positive, progesterone receptor negative, p53positive (ATCC). Estrogen and progesterone receptor negative, HER-2/neu positive, p53 negative (ATCC). Estrogen receptor positive, wnt7h1 (ATCC). EGFR positive, tumorogenic in nude mice (ATCC). Estrogen and progesterone receptor positive, wnt3, wnt7h positive (ATCC). Patient with erythroleukemia in myeloblastic relapse. Expresses CD 7, 29, 34, 44, 49d, 49e, 56 antigens (ATCC). Female, 35 years old, no previous treatment. Male, 71 years old, no previous treatment. Male, 69 years old, no previous treatment. Female, 47 years old, no previous treatment. EGFR, epidermal growth factor receptor. Source or relevant reference is provided in parentheses. Preparation of spheroids Spheroids composed of glioma cells were prepared as described previously (Bjerkvig et al., 1990). Briefly, tissue for the generation of spheroids and xenograft models were obtained from glioblastoma multiforme tumors at surgery. All patients gave their verbal informed consent, and the study was approved by the local ethical committee. Biopsies were minced into 0.5mm fragments and placed into agar-coated tissue flasks (Nunc, Roskilde, Denmark) with complete medium. After 5–10 days of culture, tumor cells rounded up, aggregated, and formed multicellular spheroids of various sizes. Spheroids of 250 mm in 1118 diameter were selected with a pasteur pipette under a Leica stereomicroscope (Leica, Heidelberg, Germany) and cultured separately in agar-coated multiwell dishes (Nunc) in complete medium. Generation of xenograft models Sixteen male and female nude rats (Rowett, Aberdeen, Scotland) weighing between 50 and 100 g were kept on a standard pellet diet, given unlimited access to water, and caged at constant temperature and humidity. The rats were anesthetized subcutaneously with midazolam, 0.2 g per 100 g of body weight; fentanyl citrate, 0.0126 g per 100 g of body weight; and fluanizone, 0.4 g per 100 g of body weight, before mounting the animals in a stereotactic frame (David Kopf Instruments, model 900, Tujunga, CA). After a midsagittal incision, a burrhole was made 1 mm posterior to the bregma suture and 3 mm to the right of the midline suture. Ten tumor spheroids (250 mm in diameter each) were implanted with a syringe in the brain 2.5 mm under the surface. The spheroids were derived from two of the human glioblastoma specimens, the burrhole was closed with bone wax, and the wound closed with polyamide thread suture. Animals were returned to their cages and observed daily. Routine monitoring by MRI (see below) 4 weeks after implantation in each case revealed that the tumor was growing. Typically, animals harboring tumors survived nonsymptomatic for up to 3 months (depending on the aggressive nature of the tumor). Animals were killed at the onset of tumor burden-induced symptoms such as overt weight loss or loss of appetite. MRI was performed in a Siemens (Erlangen, Germany) Magnetom Vision Plus 1.5T scanner and a small loop finger coil. Anesthetized rats were fixed in a polystyrene-immobilizing tube. Coronal T1 (TR 400 ms, TE 14 ms, slice thickness 2.0 mm, slice center distance 2.0 mm, total, 13 coronal slices covering the forebrain) presubcutaneous and postsubcutaneous injection of contrast agent (1.0 ml of gadolinium; 0.5 mmol/ml) and coronal T2 (TR 4000 ms, TE 96 ms, slice thickness 2.0 mm, slice center distance 2.0 mm, total, 19 coronal slices covering the forebrain) were obtained. After MRI, the brains were removed and prepared for histologic analyses (as described below). Viral transduction To determine if the culture medium affected the transduction efficiency, pilot experiments were performed assessing AAV transduction in spheroids and monolayers in three different media preparations: (1) complete medium with 10% FCS; (2) Optimem (serum-free media; BRL Life Technologies, Oslo, Norway), and (3) PBS. The different media were prewarmed and 500 ml of each was added to separate wells, containing 8 glioblastoma spheroids of 250 mm in diameter. To each well, 1 3 108 AAV-GFP particles were added and eight spheroids were kept as controls. After a 15-hr incubation, the medium containing virus was removed and replaced with complete medium, and incubated for 15 days. Monolayer cultures were processed similarly, but were only incubated for 4 days. These preliminary experiments were analyzed by confocal fluorescence laser microscopy (CLSM) and showed that different media preparations did not significantly affect transduction in solid tumor tissue or monolayer culture (data not shown). For further experiments, we therefore used complete medium. AAV-GFP and ADV-GFP particles were added at various multiplicity of ENGER ET AL. infections (100–500 MOI) in media to all the cell lines at subconfluency. After incubation for 4 days, we assayed for green fluorescent protein (GFP) activity using CLSM. Glioblastoma biopsy spheroids (250 mm in diameter) from four patients were kept individually in 1 ml of complete medium in 24 wells. ADV- and AAV-lacZ particles (1 3 108) were each added to eight spheroids and eight spheroids were used as controls for every patient. For two patients, we also added 1 3 107, 5 3 108, and 1 3 109 of AAV- and ADV-lacZ particles to the spheroids. To study transduction by confocal fluorescence microscopy, we also added 1 3 108 particles of AAV- or ADVGFP to spheroids from two patients. Xenografts were injected with viral vector 2–4 weeks after implantation using Hamilton syringes connected to the stereotactic frame. Vector was administered slowly using hand control. MRI was performed to verify the presence of tumor and to establish the location of the tumor for injection. ADV and AAV expressing lacZ or GFP was injected stereotactically in the same coordinates by reopening the original incision under anesthesia. To aid in transduction, 10 ml of 10% mannitol containing 4 3 108 particles were injected in the rats as described elsewhere (Mastakov et al., 2001). The rats were killed 1, 2, 3, and 6 weeks after viral injection, and the brains were removed, frozen, cryosectioned, and stained with lacZ. In rats injected with vectors carrying GFP, the whole brain was taken out and examined immediately using CLSM with a low-power objective. By using the transmission detector on the confocal microscope, the fluorescence distribution was related to the anatomic landmarks. This resulted in a detailed distribution map of GFP-transduced cells at the brain surface. All experiments were approved by the Norwegian Animal Research Authority (Oslo, Norway) in accordance with the Animal (Scientific Procedures) Act 1986. Flow cytometric analysis Quantification of viral transduction efficiency in monolayer cultures was determined by flow cytometric analysis of cells grown in 6-well plates (Nunc). After 4 days, the cultures were trypsinized with 3 ml of 0.025% trypsin (BioWhittaker), and resuspended in 4 ml of growth medium. The cells were then centrifuged at 100g for 4 min, and resuspended in Dulbecco’s phosphate-buffered saline (PBS; Sigma) with 0.5% D -Glucose (Merck, Darmstadt, Germany). The transduction efficiencies were determined using a FACSort flow cytometer (Becton Dickinson, Palo Alto, CA). The GFP fluorescence intensities were obtained by gating a two-parameter forward- and sidescatter cytogram to a one-parameter green fluorescence intensity plot. Both MilliQ-water and PBS were used as negative controls, while fluorescein isothiocyanate conjugated (FITC) fluorescent CaliBRITE beads (Becton Dickinson) were used as positive controls and to calibrate the instrument. Two regions M1 and M2 were chosen to determine the number of untransduced and transduced cells, respectively. The percentage of gated cells into these regions was determined, as well as the mean fluorescence intensity in each region. Three parallels were performed for every flow cytometric experiment. Confocal Fluorescence Laser Microscopy (CLSM) GFP-mediated transduction of cells in multicellular tumor spheroids was assessed by a Leica NT CLSM with an argon- 1119 AAV TRANSDUCTION IN HUMAN CARCINOMAS FIG. 1. Increasing multiplicity of infection (MOI) results in increased number of transduced cells. Adeno-associated viral vectors (AAV) or adenovirus vectors (ADV) mediated transduction in AU565 and MCF7 breast cancer and A172 and HF-66 glioma cell lines at MOI 100, 500, and 1000. Cells were infected with recombinant virus and cultured for 4 days prior to performing flow cytometry. ADV displayed higher transduction rates than AAV at similar MOIs. krypton laser using FITC filter optics (Leica). After 8–21 days, infected spheroids were transferred to an object glass in 1 drop of PBS, and examined under the microscope. Sixteen optical sections covering a total of 160 mm were recorded from each spheroid. All sections were recorded with identical gain settings on the CLSM. The sections were superimposed into a single image, and the mean fluorescence intensity was then determined within a central region of each spheroid. Nontransduced (no virus medium) spheroids were used as negative controls. In rats harboring xenografts and injected with vectors carrying GFP, after being sacrificed (1, 2, 3, and 6 weeks), the whole brain was taken out and examined immediately using CLSM with a low power objective. By using the transmission detector on the confocal microscope, the fluorescence distribution was related to the anatomic landmarks. b-Galactosidase staining Sections from xenografts, spheroids, or monolayer cultures were fixed for 20 min in 0.2% glutaraldehyde and 2% paraformaldehyde in PBS. Thereafter, they were washed 3 3 5 min in PBS and stained for b-galactosidase activity with 5-bromo-4-chloro-3-indolyl-b-D -galactopyranoside (X-Gal, Sigma) as previously described (Thorsen et al., 1997). Stained spheroids were fixed in 2% glutaraldehyde in 0.1 M sucroseadjusted cacodylate buffer for 24 hr, prior to dehydration in ethanol and embedding in a mixture of 1:1 of Epon (epoxy resin, Agar 100) (Agar Scientific, Stensteed, Essex, UK) and propylenoxide. Polymerization was carried out at 60°C for 48 hr. Xenografts and spheroids were cut using a microtome (Reichert, Wetzler, Germany) in 20-mm thick sections collected onto Super Frost Slides (Kebo Labs, Bergen, Norway) and processed as above. Sections were examined under a light microscope, and pictures were taken at 203 magnification using Nikon digital camera (Nikon Coolpix 990, Nikon Corporation, Tokyo, Japan). RESULTS Quantitative analysis of ADV and AAV transduction efficiency in monolayer cell cultures We infected six glioma and six mammary cancer cell lines in monolayer cultures (Table 1) with ADV-GFP or AAV-GFP at an MOI (virus particles per cell) between 300 and 500 to identify any qualitative differences in transduction. Four days postinfection, we confirmed by fluorescence microscopy that both AAV and ADV vectors were able to transduce cells of different morphology within the same cell line (data not shown). To quantitate differences in transduction efficiency, ADVGFP and AAV-GFP vectors were added to all cell lines including a leukemia suspension cell line, KG-1a, at an MOI of 100. Flow cytometry was performed on day 4. All cell lines, with the exception of the leukemia cell line, displayed positive but varying levels of GFP. Table 2 shows that transduction rates of the different cell lines ranged between 1.5% and 50%, but with the majority having a transduction rate over 20% at an MOI of 100. In general, the cell lines were more permissive to transduction by the ADV vectors compared to the AAV. The experiments using ADV and AAV were performed on the cell lines simultaneously. To examine the relationship between dose and transduction, AAV-GFP and ADV-GFP were added at MOIs of 100, 500, and 1000 to two glioma and two mammary cancer cell lines (Fig. 1). The results were quantitated by flow cytometry 4 days after infection. For all cell lines, the transduction rates increased with increasing MOI. At all MOIs, transduction rates were higher for ADV, and were above 80% for all cell lines at 1000 MOI. Penetration of ADV and AAV in human biopsy spheroids Several studies have concluded that gene transfer to solid tumors is a limiting factor (Benedetti et al., 1997; Thorsen et al., 1120 ENGER ET AL. TABLE 2. CANCER CELL LINE PERMISSIBILITY Glioma cell line A172 BT4C D-37 GaMG HF-66 U373 AAV % transduction ADV % transduction Breast cancer cell line 13 (12.5–14) 1.5 (1.5–2.0) 22 (20–23) 44.5 (44–45.5) 16 (15.5–16.5) 38 (37.5–38) 18 (16.5–19) 1.5 (1.5–1.5) 51.5 (50–52.5) 81 (80–82) 20 (19–20.5) 81 (78–84.5) AU565 MCF7 MDA-MB-468 HCC1395 HCC1569 T-47D AAV % transduction 43 26 27 51 26 7 (41–45.5) (25–27) (26–27.5) (49–52.5) (25.5–27) (6–7.5) ADV % transduction 55 (54.5–5.6) 59 (57.5–60) 14 (13.5–14) 23.5 (23–24.5) 46 (43.5–47.5) 26 (24.5–27) AAV-GFP or ADV-GFP were added to malignant cell lines at a multiplicity of infection of 100. After 4 days, cells were analyzed for GFP expression using flow cytometry. The results represent the mean transduction efficiency from 3 parallel experiments, each based on 5000 cells analyzed by flow cytometry. Range of transduction efficiency is given in parenthesis. AAV, adeno-associated vector; ADV, adenovirus vector; GFP, green fluorescent protein. 1997; Puumalainen et al., 1998; Enger et al., 1999; Sandmair et al., 1999; Kuriyama et al., 2001; Rainov and Kramm, 2001). We generated a three-dimensional model comprising multicellular tumor spheroids established from primary glioblastoma specimens, and infected with 1 3 108 viral vector particles. CLSM showed that spheroids transduced with AAV-GFP or ADV-GFP were evenly fluorescent at the spheroid surface (Fig. 2). Negative controls were largely devoid of positive signal, although some cells displayed autofluorescence. We observed only minor differences in the fluorescence intensity between groups of spheroids from different patients (data not shown). Spheroids were sectioned to compare viral penetration and transduction of AAV-lacZ and ADV-lacZ within the central areas. Because biopsy spheroids are heterogenous with regard to cellular composition, we examined at least eight spheroids for each vector system for all four patients (Fig. 3). Microscopy showed that transduction varied little between different spheroids from the same patient, or between groups of spheroids obtained from different patients. However, a remarkably different transduction profile was observed between the two vector systems. While AAV-transduced cells were evenly distributed throughout the spheroids, ADV-transduced cells formed a halo at the periphery with few cells transduced in the center of the spheroid (Fig. 3). Immunohistochemical experiments showed that these spheroids consisted mostly of glial-derived tumor cells and endothelial cells; homogenous cellularity through the nontransduced regions was confirmed using toluidine blue staining (data not shown). Additional transduction experiments FIG. 2. Human biopsy spheroids are effectively transduced by adeno-associated viral vectors (AAV) and adenovirus vectors (ADV). AAV-GFP or ADV-GFP vectors were added to media containing human glioblastoma spheroids. Confocal laser scanning microscopy was performed 8 days after adding vector. Examination revealed an even distribution of green fluorescent protein (GFP) reporter gene expression across the surface of the spheroid (A, shown for AAV) compared to control spheroids (B). Spheroids from several patients were examined and all were similar. ADV yielded similar results (not shown). Scale bar 5 100 mm. 1121 AAV TRANSDUCTION IN HUMAN CARCINOMAS on smaller spheroids (100 mm) gave similar results. Furthermore, comparable observations were recorded when the number of ADV and AAV particles was increased and decreased by a factor of 10 (data not shown). ADV and AAV transduction in vivo using human biopsy xenografts Genotypic heterogeneity and phenotypic diversity are wellestablished characteristics of human tumor tissue. In order to examine the efficiency of vector systems for cancer gene transfer in vivo, these parameters need to be considered, thus cell lines xenotransplanted into animals in vivo are limited by this contention. Therefore we used animal models that in every way examined recapitulated human tumor tissue in situ. Nude rats received tumor spheroids sterotactically placed intracerebrally. During tumor growth, examination of the brain sections with MRI indicated that the spheroids developed into diffuse and invading tumors reminiscent and characteristic of the glioma phenotype in situ. After 4 weeks of growth, ADV and AAV (4 3 108 particles) containing GFP or lacZ transgenes were injected slowly and directly into the center of the growing tumor as assessed by MRI. Following sacrifice (1–6 weeks) and using CLSM on the fresh rat brain, we identified transduced cells by GFP expression from both vectors. However, we observed a striking difference between the two vectors with respect to the transduction profile. Whereas AAV transduction produced confluent areas with GFP expression up to 2–3 mm from the injection site, ADV-transduced cells were more scattered. In fact, after injection of ADV-GFP, single cells expressing GFP were detected around vessels even in the contralateral hemisphere (Fig. 4A and 4B). To determine more precisely the domains of transduced cells, we used AAV-lacZ and ADV lacZ. AAV- lacZ transduced cells were observed centrally in the tumor, and at distances up to 3 mm from the injection site. In general, clusters of positive cells (10–15) were separated by a distance of 1–2 mm (Fig. 4E and 4F). After injection of ADV-lacZ, positive cells were consistently observed at the periphery of the tumor where they formed a characteristic rim at the margin. In contrast, no positive cells were observed in the central regions of the tumors (Fig. 4G and 4H), consistent with the observations from the spheroids. The transduction profile was similar in rats sacrificed at different time points, and control sections of rat brain stained for the lacZ gene product showed no positive cells (data not shown). from human tissue in contrast to ADV, which only demarcate the periphery; and (3) in animal tumor models prepared from human biopsy material, AAV-mediated transduction is observed in the central regions of the growing tumor, while ADVmediated transduction is restricted to the tumor periphery. Cellular tropism of ADV and AAV Human tissue is a natural host for ADV5 and AAV2 vectors. Cellular entry is mediated by several combinations of integrins and the CAR protein for ADV (Bergelson et al., 1997; Nemerow et al., 1999; Li et al., 2001), whereas heparan sulfate DISCUSSION Sufficient and stable gene delivery is a critical prerequisite for the development of successful gene transfer strategies. Here, we have assessed the ability of recombinant ADV5 and AAV2 vectors to transduce and penetrate various cancer cell lines and human biopsy material in vitro. Furthermore, we have evaluated these vectors in a unique in vivo model system using human biopsy material displaying biological characteristics resembling the mother tumor in situ. Our salient findings are: (1) ADV and AAV are capable of transducing a wide range of breast cancer and glioma cell lines; (2) AAV efficiently and completely transduce entire glioblastoma spheroids prepared FIG. 3. Sections from the center of human spheroids after transduction with adeno-associated viral vectors (AAV)-lacZ or adenovirus vector (ADV)-lacZ vectors. Spheroids were selected from several specimens from four patients and were similar in diameter. After a period of up to 2 weeks postinfection, the spheroids were processed for lacZ histochemistry and sectioned. Spheroids transduced with AAV (A, C, E, G) showed even transduction through mid-sections, while ADV vectors (B, D, F, H) transduced few cells in the interior of the spheroid. Increased particles of ADV did not result in any major differences in transduction efficiency. These were the same batches of ADV and AAV as used in flow cytometry experiments where ADV yielded higher transduction rates. Scale bar 5 100 mm. 1122 proteoglycan, fibroblast growth factor receptor I, and the avb5 integrin appear to be responsible for AAV infection (Summerford and Samulski, 1998; Qiu et al., 1999, 2000; Qing et al., 1999; Qiu and Brown, 1999; Summerford et al., 1999). These cell surface markers are widely expressed, which is consistent with our findings that both the ADV5 and AAV2 vectors in this study were capable of transducing a wide range of human breast and glioma cell lines. Flow cytometry experiments showed that the quantities of cells transduced by AAV or ADV varied substantially, and this may be accounted for by a differential expression of the receptor components on these cells. Our experiments included cell lines expressing a variety of different phenotypes including wild-type as well as mutant p53 (Table 1), indicating little correlation between phenotype and permissibility. We observed that the human leukemia cell line, KG1a, and the rat glioma cell line , BT4C, were completely or relatively nonpermissive for AAV2 and ADV5 vectors. In other studies, leukemia (Itou et al., 1998), rat brain tissue and glioma cell lines have been transduced (Okada et al., 1996; Rosenfeld et al., 1997; Mizuno and Yoshida, 1998; Mizuno et al., 1998; Cunningham et al., 2000; Tenenbaum et al., 2000). Accordingly, the explanation for the low receptivity of AAV and ADV transduction does not seem to be a general feature of cell lines derived from particular tissues. An alternative explanation is that the cytomegalovirus (CMV) promoter common to these vectors or other cis-acting elements were downregulated by certain factors in these cell lines. An increase in MOI resulted in a increase in the number of transduced cells. This indicates that cell surface receptors were not a limiting component of transduction. Penetration of AAV and ADV particles in human glioblastoma spheroids In order to study transduction efficiency in solid tumor tissue, we generated a three-dimensional model comprising multicellular tumor spheroids established from primary glioblastoma specimens. Previous studies have shown that in vitro, these spheroids contain glioma cells, astrocytes, neurons, and blood vessels, and express cell ploidy and growth kinetics similar to the mother tumor (Sutherland, 1988; Bjerkvig et al., 1990). Using this model system, we observed an even transduction over the surface of the spheroids with ADV- or AAVGFP, whereas only AAV were able to transduce the lacZ re- ENGER ET AL. porter gene into cells in the central regions of the spheroids. This is in contrast to and could not have been predicted based on the results observed in the cell lines. These findings were consistently observed in spheroids from different tumors and with use of increased and decreased viral particle numbers. Notably, the same AAV and ADV preparations were used in the monolayer cultures and spheroid experiments, and typically more cells were transduced by ADV vectors in the cell line model. Two explanations can account for these observations that are not mutually exclusive. First, it is possible that the low or differential levels of CAR protein in glioma spheroids could restrict transduction (Grill et al., 2001). Second, the smaller size of the AAV-capsid (20 nm, compared to adenovirus [100 nm]), may facilitate AAV movement through solid tumor tissue. Therefore, based on the available evidence (Benedetti et al., 1997; Thorsen et al., 1997; Puumalainen et al., 1998; Enger et al., 1999; Sandmair et al., 1999; Kuriyama et al., 2001), it can be concluded that AAV are able to penetrate solid tumor tissue more effectively than either adenoviral or retroviral vectors. Comparison of AAV and ADV in animal models Transplantation of human biopsy material into animal models represents a model different from xenografts derived from immortalized cell lines (Engebraaten et al., 1990, 1999). This model displays diverse phenotypes consistent with the biologic heterogeneity of glioblastoma multiforme in situ (Horten et al., 1981; Stromblad et al., 1982; Engebraaten et al., 1990, 1999; Peterson et al., 1994). Furthermore, this model has the physical barriers to gene delivery associated with the extracellular matrix. Comparing AAV and ADV vectors in this in vivo model revealed several important observations and differences not recorded in cell culture experiments. Most notably, AAV transduction was seen scattered throughout central regions of the tumor, while ADV transduction demarcated cells in a halo configuration. These observations are consistent with those observed in the spheroid model, and suggest that AAV can diffuse more effectively compared to ADV. Additionally, the AAV-transduced cells in vivo had a more clustered appearance than the vast spread of lacZ-positive cells observed in the spheroids. This may be explained by the dynamic microenvironment of the tumor in vivo where spread of the vector is mediated by flow of cerebrospinal fluid and pressure gradients within the tu- FIG. 4. Adeno-associated viral vectors (AAV) and adenovirus vectors (ADV) transduction in human glioblastoma xenografts. Human biopsy spheroids were sterotactically injected into the hemispheres of young nude rats. After 4 weeks of tumor development, AAV and ADV vectors carrying lacZ or green fluorescent protein (GFP) were injected into the center of the tumor. After a period of up to 6 weeks, the brains were analyzed for reporter gene expression. Shown in (A) is a confluent area with fluorescence representing clusters of GFP-positive cells in the tumor after AAV injection. In the lower right corner at higher magnification, is a picture of the area where the tumor infiltrates the surface illustrating a close correlation with the area of fluorescence. In contrast, cells transduced with ADV are distributed in a more scattered pattern (B, arrows indicating transduced cells). Fluorescence was also detected in the contralateral hemisphere suggesting convection based spread of the vector (blue arrow). The pictures were obtained by imposing a fluorescence picture of the brain surface on a digital photo image of the same region (See materials and methods for details). (C) and (D) are MRI scans of rat brains harboring tumors that were injected with AAV and ADV, respectively. Blue spots indicates the area of transduction in a region of 3 mm anterior and posterior to the plane of the needle tract (red line). After injection of AAV-lacZ, transduced cells are observed in the central regions of the tumor (E). Transduced cells are also seen at the periphery of the tumor, although these are few in number (F). After injection with ADV, transduced cells were virtually absent from the central regions of all tumors (G), wherease a strong band of positive cells was observed at the periphery of the tumor (H). AAV TRANSDUCTION IN HUMAN CARCINOMAS 1123 1124 ENGER ET AL. mor. Also, the in vivo tumors grow and expand to be larger than the spheroids, such that AAV may not be available to tumor cells born later. In contrast, cells transduced by ADV were observed at the tumor periphery and not the injection site. One explanation could be that ADV initially transduced cells at the injection site, but as the tumor expanded cells remained in the same spatial configuration, such that at the time of analysis cells were positioned at the tumor periphery. Furthermore, the presence of ADV-mediated GFP-positive cells in the perivascular space of capillaries outside the tumor region, suggests that the spread of this vector to some degree is convection-based. In summary, we have generated a rational and graded comparison between AAV and ADV using models with increasing complexity: (1) cell line data from highly standardized material; (2) three-dimensional models in vitro from human biopsy spheroids that are less standardized, and (3) in vivo models in which the experimental conditions are not standardized and thus are more representative of the pathologic state in situ. Our interpretation of the current data, is that even though ADV and AAV are comparable in their tropism for cancer cell types, AAV are better gene transfer vectors for cancer gene therapy because they penetrate solid tumor tissue in vivo more effectively. Taken together, the evidence provided here suggests that AAV could play a role in achieving therapeutic efficacy for cancer gene therapy. ACKNOWLEDGMENTS This work was supported by grants from the Norwegian Health Ministry and the Norwegian Cancer Society. We thank Dr. Kenneth Chahine and the team at Avigen Inc. (Alameda, CA) for their kind gift of high-titer AAV-EGFP and AAV-lacZ particles. We are grateful to Iren Sefland (University of Oslo), Tove Johansen, and Beatrice Probst for technical assistance. Tore-Jacob Raa and Aina Johannessen are gratefully acknowledged for their animal husbandry. We thank Drs. Karl-Henning Kalland, Stig-Ove Bøe, and Kjersti Lønning for critical discussions. REFERENCES AKSLEN, L.A., ANDERSEN, K.J., and BJERKVIG, R. (1988). Characteristics of human and rat glioma cells grown in a defined medium. Anticancer Res. 8, 797–803. BERGELSON, J.M., CUNNINGHAM, J.A., DROGUETT, G., KURTJONES, E.A., KRITHIVAS, A., HONG, J.S., HORWITZ, M.S., CROWELL, R.L., and FINBERG, R.W. (1997). Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320–1323. BENEDETTI, S., DIMECO, F., POLLO, B., CIRENEI, N., COLOMBO, B.M., BRUZZONE, M.G., CATTANEO. E,, VESCOVI A, DIDONATO S., COLOMBO, M.P., and FINOCCHIARO, G. (1997). Limited efficacy of the HSV-TK/GCV system for gene therapy of malignant gliomas and perspectives for the combined transduction of the interleukin-4 gene. Hum. Gene Ther. 20, 1345–1353. BJERKVIG, R., TØNNESEN, A., LÆRUM, O.D., and BACKLUND, E.O. (1990). Multicellular tumor spheroids maintained in organ culture. J. Neurosurg. 72, 463–475. BURTON, M., NAKAI, H., COLOSI, P., CUNNINGHAM, J., MITCHELL, R., and COUTO, L. (1999). Coexpression of factor VIII heavy and light chain adeno-associated viral vectors produces biologically active protein. Proc. Natl. Acad. Sci. U.S.A. 96, 12725– 12730. CUNNINGHAM, J., OIWA, Y., NAGY, D., PODSAKOFF, G., COLOSI, P., and BANKIEWICZ, K.S. (2000). Distribution of AAVTK following intracranial convection-enhanced delivery into rats. Cell Transplant. 9, 585–594. ENGEBRAATEN, O., BJERKVIG, R., LUND-JOHANSEN, M., WESTER, K., PEDERSEN, P.H., MORK, S., BACKLUND, E.O., and LÆRUM, O.D. (1990). Interaction between human brain tumour biopsies and fetal rat brain tissue in vitro. Acta Neuropathol. 81, 130–140. ENGEBRAATEN, O., HJORTLAND, G.O., HIRSCHBERG, H., and FODSTAD, O. (1999). Growth of precultured human glioma specimens in nude rat brain. J Neurosurg. 90, 125–132. ENGER, P.Ø., VISTED, T., THORSEN, F., BJERKVIG, R., and LUND-JOHANSEN, M. (1999). Retroviral transfection of the lacZ gene from Liz-9 packaging cells to glioma spheroids. Int. J. Dev. Neurosci. 17, 665–672. FISHER, K.J., JOOSS, K., ALSTON, J., YANG, Y., HAECKER, S.E., HIGH, K., PATHAK, R., RAPER, S.E., and WILSON, J.M. (1997). Recombinant adeno-associated virus for muscle directed gene therapy. Nat. Med. 3, 306–312. FLOTTE, T.R., CARTER, B.J., CONRAD, C.K., GUGGINO, W.J., REYNOLDS, T.C., ROSENSTEIN, B.J., TAYLOR, G., WALDEN S., and WETZEL, R.A. (1996). A phase I study of an adeno-associated virus-CFTR gene vector in adult CF patients with mild lung disease. Hum. Gene Ther. 7, 1145–1159. GRAHAM, F.L., and PREVEC, L. (1995). Methods for construction of adenovirus vectors. Mol. Biototechnol. 3, 207–220. GRILL, J., VAN BEUSECHEM, V.W., VAN DER VALK, P., DIRVEN, C.M., LEONHART, A., PHERAI, D.S., HAISMA, H.J., PINEDO, H.M., CURIEL, D.T., and GERRITSEN, W.R. (2001). Combined targeting of adenoviruses to integrins and epidermal growth factor receptors increases gene transfer into primary glioma cells and spheroids. Clin. Cancer Res. 7, 641–650. HAUGLAND, K.H., TYSNES, B.B., and TYSNES, O.B. (1997). Adhesion and migration of human glioma cells are differently dependent on extracellular matrix molecules. Anticancer Res. 17, 1035–1043. HORTEN, B.C., BASLER, G.A., and SHAPIRO, W.R. (1981). Xenograft of human malignant glial tumors into brains of nude mice. A histopatholgical study. J. Neuropathol. Exp. Neurol. 40, 493–511. ITOU, T., MIYAMURA, K., ABE, A., EMI, N., TANIMOTO, M., TERASAKI, H., SHIMADZU, M., and SAITO, H. (1998). Recombinant adeno-associated virus-mediated gene transfer into human leukemia cell lines. Int. J. Hematol. 67, 27–35. KAY, M.A., MANNO, C.S., RAGNI, M. V., LARSON, P.J., COUTO, L.B., MCCLELLAND, A., GLADER, B., CHEW, A.J., TAI, S.J., HERZOG, R.W., ARRUDA, V., JOHNSON, F., SCALLAN, C., SKARSGARD, E., FLAKE, A.W., and HIGH, K.A. (2000). Evidence for gene transfer and expression of factor IX in haemophilia B patients treated with an AAV vector. Nat. Genet. 24, 257–261. KNOTT, J.C., MAHESPARAN, R., GARCIA-CABRERA, I., TYSNES, B.B., EDVARDSEN, K., NESS, G.O., MORK, S., LUNDJOHANSEN, M., and BJERKVIG, R. (1998). Stimulation of extracellular matrix components in the normal brain by invading glioma cells. Int. J. Cancer 75, 864–872. KUNKE, D., GRIMM, D., DENGER, S., KREUZER, J., DELIUS, H., KOMITOWSKI, D., and KLEINSCHMIDT, J.A. (2000). Preclinical study on gene therapy of cervical carcinoma using adeno-associated virus vectors. Cancer Gene Ther. 7, 766–777. KURIYAMA, N., KURIYAMA, H., JULIN, C.M., LAMBORN, K.R., and ISRAEL, M.A. (2001). Protease pretreatment increases the efficacy of adenovirus-mediated gene therapy for the treatment of an experimental glioblastoma model. Cancer Res. 61, 1805–1809. LI, E., BROWN, S.L., STUPACK, D.G., PUENTE, X.S., CHERESH, 1125 AAV TRANSDUCTION IN HUMAN CARCINOMAS D.A., and NEMEROW, G.R. (2001). Integrin alphavbeta1 is an adenovirus coreceptor. J. Virol. 11, 5405–5409. LAERUM, O.D., RAJEWSKY, M.F., SCHACHNEC, M., STAVROU, D., HAGLID, K.H., and HANGEN, Å. (1977). Phenotypic properties of neoplastic cell lines developed from fetal rat brain cells in culture after exposure to ethylnitrosurea in vivo. Z. Krebsforsch. 89, 273–295. MASTAKOV, M.Y., BAER, K., XU, R., FITZSIMONS, H., and DURING, M.J. (2001). Combined injection of rAAV with mannitol enhances gene expression in the rat brain. Mol. Ther. 3, 225–232. MATSUSHITA, T., ELLIGER, S., ELLIGER, C., PODSAKOFF, G., VILLARREAL, L., KURTZMAN, G.J., IWAKI, Y., and COLOSI, P. (1998). Adeno-associated virus vectors can be efficiently produced without helper virus. Gene Ther. 5, 938–945. MIZUNO, M., and YOSHIDA, J. (1998). Improvement of transduction efficiency of recombinant adeno-associated virus vector by entrapment in multilamellar liposomes. Jpn. J. Cancer Res. 89, 352–354. MIZUNO, M., YOSHIDA, J., COLOSI, P., and KURTZMAN, G. (1998). Adeno-associated virus vector containing the herpes simplex virus thymidine kinase gene causes complete regression of intracerebrally implanted human gliomas in mice, in conjunction with ganciclovir administration. Jpn. J. Cancer Res. 89, 76–80. MONAHAN, P.E., and SAMULSKI, R.J. (2000a). Adeno-associated virus vectors for gene therapy: more pros than cons? Mol. Med. Today 6, 433–440. MONAHAN, P.E., and SAMULSKI, R.J. (2000b). AAV vectors: Is clinical success on the horizon? Gene Ther. 7, 24–30. MUZYCZKA, N., SAMULSKI, R.J., HERMONAT, P., SRIVASTAVA, A., and BERNS, K.I. (1984). The genetics of adeno-associated virus. Adv. Exp. Med. Biol. 179, 151–161. NEMEROW, G.R., and STEWART, P.L. (1999). Role of alpha(v) integrins in adenovirus cell entry and gene delivery. Microbiol. Mol. Biol. Rev. 63, 725–734. OKADA, H., MIYAMURA, K., ITOH, T., HAGIWARA, M., WAKABAYASHI, T., MIZUNO, M., COLOSI, P., KURTZMAN, G., and YOSHIDA, J. (1996). Gene therapy against an experimental glioma using adeno-associated virus vectors. Gene Ther. 3, 957–964. PETERSON, D. L., SHERIDAN, P. J., and BROWN, W. E., Jr. (1994). Animal models for brain tumors: Historical perspectives and future directions. J. Neurosurg. 80, 865–876. PONNAZHAGAN, S., CURIEL, D.T., SHAW, D.R., ALVAREZ, R.D., and SIEGAL, G.P. (2001). Adeno-associated virus for cancer gene therapy. Cancer Res. 61, 6313–6321. PUUMALAINEN, A.M., VAPALAHTI, M., AGRAWAL, R.S., KOSSILA, M., LAUKKANEN, J., LEHTOLAINEN, P., VIITA, H., PALARVI, L., VANNINEN, R., and YLA-HERTTUALA S. (1998). Beta-galactosidase gene transfer to human malignant glioma in vivo using replication-deficient retroviruses and adenoviruses. Hum. Gene Ther. 9, 1769–1774. QAZILBASH, M.H., XIAO, X., SETH, P., COWAN, K.H., and WALSH, C.E. (1997). Cancer gene therapy using a novel adeno-associated virus vector expressing human wild-type p53. Gene Ther. 4, 675–682. QING, K., MAH, C., HANSEN, J., ZHOU, S., DWARKI, V., and SRIVASTAVA, A. (1999). Human fibroblast growth factor receptor 1 is a co-receptor for infection by adeno-associated virus 2. Nat. Med. 5, 71–77. QIU, J., and BROWN, K.E. (1999). Integrin alphaVbeta5 is not involved in adeno-associated virus type 2 (AAV2) infection. Virology 264, 436–440. QIU, J., HANDA, A., KIRBY, M., and BROWN, K.E. (2000). The interaction of heparin sulfate and adeno-associated virus 2. Virology 269, 137–147. QIU, J., MIZUKAMI, H., and BROWN, K.E. (1999). Adeno-associated virus 2 co-receptors? Nat. Med. 5, 467–468. RAINOV, N.G., and KRAMM, C.M. (2001). Vector delivery methods and targeting strategies for gene therapy of brain tumors. Curr. Gen. Ther. 1, 367–383. ROSENFELD, M.R., BERGMAN, I., SCHRAMM, L., GRIFFIN, J.A., KAPLITT, M.G., and MENESES, P.I. (1997). Adeno-associated viral vector gene transfer into leptomeningeal xenografts. J. Neurooncol. 34, 139–144. SANDMAIR, A.M., LOIMAS, S., POPTANI. H,, VAINIO, P., VANNINEN R., TURUNEN, M., TYYNELA, K., VAPALAHTI, M., and YLA-HERTTUALA, S. (1999). Low efficacy of gene therapy for rat BT4C malignant glioma using intra-tumoural transduction with thymidine kinase retrovirus packaging cell injections and ganciclovir treatment. Acta Neurochir. (Wien). 141, 867–872. SMITH-ARICA, J.R., and BARTLETT, J.S. (2001). Gene therapy: Recombinant adeno-associated virus vectors. Curr. Cardiol. Rep. 3, 43–49. SOMIA, N., and VERMA, I.M. (2000). Gene therapy: trials and tribulations. Nat. Rev. Genet. 1, 91–99. STROMBLAD, L.G., BRUN, A., SALFORD, L. G., and STENEVI, U. (1982). A model for xenotransplantation of human malignant astrocytomas into the brain of normal adult rats. Acta Neurochir. 65, 217–26. SNYDER, R.O. (1999). Adeno-associated virus-mediated gene delivery. J. Gene Med. 1, 166–175. SUMMERFORD, C., BARTLETT, J.S., and SAMULSKI, R.J. (1999). AlphaVbeta5 integrin: A co-receptor for adeno-associated virus type 2 infection. Nat. Med. 5, 78–82. SUMMERFORD, C., and SAMULSKI, R.J. (1998). Membrane-associated heparan sulfate proteoglycan is a receptor for adeno-associated virus type 2 virions. J. Virol. 72, 1438–1445. SUTHERLAND, R.M. (1988). Cell and environment interactions in tumor microregions: The multicellular spheroid model. Science 240, 177–184. TENENBAUM, L., JURYSTA, F., STATHOPOULOS, A., PUSCHBAN, Z., MELAS, C., HERMENS, W.T., VERHAAGEN, J., PICHON, B., VELU, T., and LEVIVIER, M. (2000). Tropism of AAV-2 vectors for neurons of the globus pallidus. Neuroreport 11, 2277–2283. THORSEN, F., VISTED, T., LEHTOLAINEN, P., YLÄ-HERTTUALA , S., and BJERKVIG, R. (1997). Release of replication-deficient retroviruses from a packaging cell line: Interaction with glioma tumor spheroids in vitro. Int. J. Cancer 71, 874–880. VELDWIJK, M.R., FRUEHAUF, S., SCHIEDLMEIER, B., KLEINSCHMIDT, J.A., and ZELLER, W.J. (2000). Differential expression of a recombinant adeno-associated virus 2 vector in human CD341 cells and breast cancer cells. Cancer Gene Ther. 7, 597–604. Address reprint requests to: Frank Hoover Haukeland Hospital Department of Oncology Oncology Research Laboratory 5021 Bergen Norway E-mail: [email protected] Received for publication January 9, 2002; accepted after revision April 22, 2002. Published online: May 30, 2002.
© Copyright 2019