Gene Therapy (2009) 16, 894–904 & 2009 Macmillan Publishers Limited All rights reserved 0969-7128/09 $32.00 www.nature.com/gt ORIGINAL ARTICLE (Strept)avidin-displaying lentiviruses as versatile tools for targeting and dual imaging of gene delivery

MU Kaikkonen1,2,7, HP Lesch1,2,7, J Pikkarainen1,2,JKRa¨ty3, T Vuorio1, T Huhtala4, M Taavitsainen1,2, T Laitinen2, P Tuunanen5, O Gro¨hn5,ANa¨rva¨nen4, KJ Airenne1,7 and S Yla¨-Herttuala1,3,6,7 1Department of Biotechnology and Molecular Medicine, AI Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland; 2Ark Therapeutics Oyj, Microkatu, Kuopio, Finland; 3Gene Therapy Unit, Kuopio University Hospital, Kuopio, Finland; 4Biomedical Imaging Unit, AI Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland; 5Department of Neurology, AI Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland and 6Department of Medicine, AI Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland

Lentiviruses have shown great promise for human gene first-time dual imaging of virus biodistribution and transduction therapy. However, no optimal strategies are yet available for pattern by single-photon emission computed tomography and noninvasive imaging of virus biodistribution and subsequent magnetic resonance imaging after stereotactic injection into rat transduction in vivo. We have developed a dual-imaging brain. In addition, vector retargeting to cancer cells over- strategy based on avidin–biotin system allowing easy expressing CD46, epidermal growth factor and transferrin exchange of the surface ligand on HIV-derived lentivirus receptors using biotinylated ligands and antibodies was envelope. This was achieved by displaying avidin or strepta- demonstrated in vitro. In conclusion, we have generated novel vidin fused to the transmembrane anchor of vesicular lentivirus vectors for noninvasive imaging and targeting of stomatitis virus G protein on gp64-pseudotyped envelopes. lentivirus-mediated gene delivery. This study suggests that Avidin and streptavidin were efficiently incorporated on virus these novel vectors could be applicable for the treatment of particles, which consequently showed binding to biotin in central nervous system disorders and cancer. ELISA. These vectors, conjugated to biotinylated radionuclides Gene Therapy (2009) 16, 894–904; doi:10.1038/gt.2009.47; and engineered to express a ferritin transgene, enabled for the published online 14 May 2009

Keywords: avidin; streptavidin; lentivirus; targeting; MRI; SPECT

Introduction vectors can be biotinylated chemically3–6 or meta- bolically7–10 while bringing the biotinylated targeting Lentiviruses are able to efficiently transduce both molecule in conjunction with avidin. The second dividing and nondividing cells leading to long-term approach consists of displaying avidin on the virus gene expression.1 This ability has made them especially surface. Our group has previously described a baculo- attractive for gene transfer into differentiated tissues virus displaying avidin that could be efficiently targeted giving rise to numerous clinical trials based on the ex vivo in vitro.11 transduction of cells. However, for many applications As important as choosing the targeting strategy itself is targeted gene transduction to specific tissues through the use of noninvasive imaging techniques to elucidate the local or intravenous injection would be the ultimate specificity of new targeted vectors. In vivo imaging preferred method of gene delivery. Unfortunately, the techniques have the potential to provide critical informa- specificity of the various vector pseudotypes is not tion about the safety and kinetics of viral administration. always sufficient for the therapeutic need. Therefore, Compared to the traditional immunological and histologi- many attempts have been made to develop targetable cal methods, noninvasive imaging methods can follow vectors based on modification of the virus surface virus localization with anatomical accuracy in real time. genetically, chemically or by using bispecific targeting Current imaging methods in gene therapy can be divided molecules.2 One of the most versatile approaches is into biodistribution and reporter gene imaging (transduc- offered by the avidin–biotin system, which has been tion). Reporter gene imaging visualizes the transgene adapted for vector targeting by two ways. First, the expression whereas biodistribution demonstrates the vector location throughout the body. Single-photon emission Correspondence: Professor S Yla¨-Herttuala, Department of Biotech- tomography (SPECT) and magnetic resonance imaging nology and Molecular Medicine, AI Virtanen Institute for Molecular (MRI) have been previously utilized to image the biodis- Sciences, University of Kuopio, PO Box 1627, FIN-70211, Kuopio, tribution of herpes simplex viruses12,13 and baculo- Finland. viruses14,15 but studies using other viral vectors have not E-mail: Seppo.Ylaherttuala@uku.fi 7These authors contributed equally to this work. been described. To date, the only imaging modality Received 31 January 2009; revised 31 March 2009; accepted 1 April described for lentiviruses are based on optical transduction 2009; published online 14 May 2009 imaging of the luciferase reporter gene.16,17 Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 895 In the present work, we created recombinant lentivirus to attach to biotin-coated ELISA plates and the bound vectors displaying avidin or streptavidin fused to the virus was detected by gp64 antibody. Results showed transmembrane anchor of vesicular stomatitis virus G that AVD/GP64 and SA/GP64 viruses bound to biotin- protein (VSV-G) (VSV-G EctoDomain, VSV-GED)18 pseu- coated plates in a dose-dependent manner but the dotyped with the baculovirus major envelope glycopro- extent of binding was 3–6 times higher for SA/GP64 tein gp64. We present data on targeting of these (Figure 1d). On the basis of these results, we chose lentivirus vectors to transferrin and epidermal growth SA/GP64 lentivirus for in vivo studies. factor (EGF) and CD46 receptors overexpressed on tumor cells. Moreover, we demonstrated the utility of Targeting in vitro (strept)avidin display for noninvasive dual imaging of To analyze lentivirus targeting in vitro, we tested biotinyl- lentivirus biodistribution and transgene expression in ated ligands and antibodies selectively targeting receptors vivo while evaluating for the first time the transduction expressed at high levels on tumor cells (Figure 2). pattern of gp64-pseudotyped lentivirus vectors in the Transferrin receptors mediate iron transport into the central nervous system (CNS). cells and are upregulated on rapidly dividing cells such as gliomas.19 We bound ligand and/or antibody against the transferrin receptor to the rat malignant glioma cells Results (BT4C) and human glioblastoma cells (D54) followed by lentivirus transduction (Figure 2a). On rat glioma cells, a Characterization of (strept)avidin-displaying 20–30% increase in transduction efficiency was seen with lentiviruses AVD/GP64 when targeted with biotinylated transferrin We generated gp64-pseudotyped lentivirus vectors or transferrin antibody whereas with SA/GP64 an displaying avidin or streptavidin fused to the trans- increase of 50–60% was achieved. Targeting the human membrane anchor of VSV-G (Figure 1a). The optimal D54 cells with biotinylated transferrin led to a 20 and ratio of gp64 to SA/AVD-VSV-GED plasmid was found 40% increase in the number of green fluorescent protein to be 3:1 and higher amount of SA/AVD-VSV-GED (GFP)-expressing cells with AVD/GP64 and SA/GP64, plasmid resulted in decreased titers (data not shown). respectively. No change in the percentage of GFP- The viral supernatant was harvested, clarified using a expressing cells was seen with GP64 control virus after 0.22 mm filter and concentrated by ultracentrifugation treatment of BT4C and D54 cells with the targeting (19 500 r.p.m., 2 h 20 min). The incorporation of gp64, molecules. streptavidin and avidin into the lentiviral particles was Besides transferrin, overexpression of epidermal determined by immunoblot analysis from equalized growth factor receptor (EGFR) is frequently found in amounts of concentrated virus particles (p24gag TU per several human cancers and is associated with an ng) (Figure 1b). Gp64 was detected with gp64 antibody aggressive tumor behavior.20 Cancer cell lines with high in all lentivirus preparations, suggesting that codisplay EGFR expression (SKOV3, HepG2, A549) were targeted of (strept)avidin fusion proteins did not affect its with biotinylated EGFR antibody (Cetuximab) whereas incorporation. Probing with anti-avidin confirmed the the non-biotinylated antibody served as a control (Figure proper incorporation of the 23 kDa AVD-VSV-GED 2b). 100% increase in transduction was seen with the SA/ protein on the surface of AVD/GP64 viruses. Similarly, GP64 when targeted with the biotinylated anti-EGFR. the SA-VSV-GED with molecular weight of 27 kDa was Interestingly, the transduction efficiency of AVD/GP64 detected in the immunoblots with anti-streptavidin. was only enhanced by 10–40% with the same antibody- Recombinant proteins of the same size were recognized mediated targeting. The targeting effect was abolished by VSV-G antibody against the 15 C-terminal amino when the non-biotinylated anti-EGFR was used confirm- acids (497-511) of the VSV-G showing significantly ing that the unbound antibodies are not responsible for higher incorporation of SA-VSV-GED compared to the increased transduction and avidin–biotin bond is AVD-VSV-GED. mediating virus binding. No significant increase in the The efficacy of lentivirus production was further transduction was achieved with GP64 vector in all the determined by measuring the p24gag concentration from cell lines studied. To test the cognate ligand of EGFR, we concentrated virus preparations (Figure 1c). No signifi- pretreated SKOV3 cells with biotinylated EGF before cant differences in particle concentrations (p24gag transduction with the lentivirus vectors. In line with the pg ml1) were detected between the gp64-pseudotyped anti-EGFR results, EGF was able to increase the SA/ lentivirus compared to SA/GP64 or AVD/GP64. GP64 transduction by 50% whereas no significant effect To determine if (strept)avidin display had an effect on was seen with AVD/GP64 or GP64 vectors. the transduction efficiency of viruses, we assayed the CD46 has also been shown to be overexpressed on titers on four cell lines in parallel (HeLa, BT4C, SKOV3, some human neoplasms protecting these cells from HepG2) (Figure 1c). Titers of 107–108 transducing units destruction by the complement system.21 A human per ml (TU per ml) were reached in BT4C, SKOV3 and glioblastoma cell line, U118MG, was chosen to examine HepG2 cells whereas 106 TU per ml was obtained in the efficacy of CD46 antibody-mediated viral targeting HeLa cells. The titers of SA/GP64 or AVD/GP64 (Figure 2c). Biotinylated anti-CD46 was shown to lentiviruses were close to the titers of GP64-only viruses, increase the transduction efficiency of SA/GP64 by suggesting that simultaneous expression of two different 300% and of AVD/GP64 by 200% compared to un- envelope proteins had no major influence on the virus targeted virus and control virus. production. We used ELISA to determine if (strept)avidin fusion In vitro MRI proteins displayed on the lentivirus surface were For noninvasive imaging of virus transduction, we accessible for biotin binding. Lentiviruses were allowed utilized ferritin as endogenous MRI reporter gene. The

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 896

Figure 1 Schematic presentation of the display constructs SA-VSV-GED and AVD-VSV-GED with signal sequences (SS) and transmembrane domains (TM) indicated (a). Characterization of (strept)avidin-displaying lentiviruses in comparison to control virus GP64 (b–d). (b) Detection of envelope glycoproteins by immunoblot analysis. The concentrated virus preparations were detected with anti-gp64, anti- avidin, anti-streptavidin and anti-VSV-G. (c) Comparison of viral titers based on capsid protein p24 amount (pg ml1) and infectivity (TU per ml) in different cell lines. (d) Binding of increasing amounts of lentiviruses on biotin-coated ELISA plates. Error bars correspond to ±s.d. of three independent transductions. The data from the same viral dilutions of SA/GP64 and AVD/GP64 were compared pair wise to the GP64 control virus using a two-way ANOVA with Bonferroni’s post-test; **Po0.01, ***Po0.001.

overexpression of ferritin has been shown to be sufficient duced HepG2 cells 48 h post-transduction (p.t.) to augment the iron uptake of cells giving rise to (Figure 3). Transduced cells grown in their normal significant contrast in situ in the absence of exogenously growth medium, having 10% fetal bovine serum as the administered MRI substrates.22–24 To evaluate the ability only iron source, showed little increase in the R2 of human heavy chain of ferritin (hHF) expressing compared to the control cells. Supplementation of the lentivirus to increase iron uptake in vitro, we measured medium with iron resulted in twofold increase in the R2 relaxation rate R2(R2 ¼ 1/T2) in AVD/GP64-hHF-trans- compared to the controls. These results confirmed that

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 897

Figure 2 Flow cytometric analysis of green fluorescent protein (GFP) expression in cancer cells transduced with lentivirus vectors targeted by biotinylated ligands or antibodies. (a) Targeting the transferrin receptor on human glioblastoma (D54) and rat glioma (BT4C) cells with biotinylated transferrin ligands and antibodies. (b) Targeting the epidermal growth factor receptor (EGFR) on lung carcinoma (A549), hepatocellular carcinoma (HepG2) and human ovarian carcinoma (SKOV3) cells with biotinylated EGF and Cetuximab. (c) Targeting the CD46 on human glioblastoma cells (U118MG) with biotinylated anti-CD46. Results represent means of three independent experiments±s.d. *Po0.05, **Po0.01, ***Po0.001. One-way ANOVA, with Dunnett’s post hoc test. lentivirus transduction leads to ferritin expression and MRI of viral transduction in vivo thus to iron storage inside the cells at levels detectable To monitor the transgene expression after virus injection with MRI. to corpus callosum, we performed scans on animals at 4.7 T MRI at days 3, 14, 28, 94, 141 and 211. Using T2*- weighted MRI, we detected the increase in intracellular Imaging virus biodistribution by SPECT/CT in vivo iron caused by ferritin expression as a negative contrast Virus biodistribution was studied using SPECT/CT by along the needle track and in the injection site after attaching 111In-labeled biotin-poly-Lys-DTPA on (strept) inoculation of SA/GP64-hHF (Figure 5a). Control ani- avidin on the viral surface. Rats were monitored during mals receiving SA/GP64-LacZ lentivirus or 111In-labeled 3 days after injection of the radiolabeled lentivirus or biotin-poly-Lys-DTPA alone showed minor signal loss at radiolabeled chelate alone into the right corpus callosum. the injection site at day 3 (Figures 5b and c). This On the first and third days only planar images were resolved, with one exception, before later measurements taken. On day 2 both planar and three-dimensional (3D) and was attributed to small hemorrhage caused by images were taken (Figure 4). Clear accumulation of the injection. One animal receiving SA/GP64-LacZ, how- signal was seen in the brain area in all rats up to the third ever, showed sustained negative contrast on T2*- day (Figures 4a–c). Interestingly, one of the rats injected weighted sequence up to day 35 that was attributed to with radiolabeled virus showed leakage to the circula- extensive chronic hemorrhage as confirmed visually tion during the injection leading to a clear signal in the from the histological sections (data not shown). peritoneal area on day 2 in 3D images (Figure 4e). No To verify the source of MRI contrast, we performed accumulation of indium was seen in the abdomen of the histological evaluation of randomly selected animals by rats with successful injections (Figure 4d). Three rats Perls’ Prussian Blue, anti-ferritin and X-gal stainings 5 or injected with 111In-labeled biotin-poly-Lys-DTPA alone 211 (anti-ferritin only) days after viral inoculation. The showed signal accumulation to the kidney area on the results showed ferritin expression and iron accumulation first day, which shifted to the bladder on day 2 (Figure in the cells of the corpus callosum at the site of injection 4f). This is explained by the small molecular size of the (Figures 6a–d). Ferritin expression was still detected chelate (below 10 000 Da) that leads to its transfer to the from transduced tissues at 211 days time point (Figure circulation and subsequent elimination by the kidneys 6e). The X-gal staining pattern of SA/GP64-LacZ and bladder. transduced rats was also detected along the needle track,

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 898 protein gp64 has proven especially attractive as its expression is not cytotoxic and allows the production of vectors with high particle stability.29 Numerous studies have shown that gp64-pseudotyped HIV and FIV vectors efficiently transduce various cell types in vitro and in vivo.29–32 The specificities of viral attachment proteins do not always coincide with those that are required for targeted gene delivery, and incorporation of additional ligands is needed. Ideally, different targeting molecules would be readily coupled to a validated vector and this vector– ligand complex would retain its integrity under physio- B 13 logical conditions. The very high affinity (Kd 10 – 1015 M) of the binding of streptavidin and avidin to biotin could offer such an approach. Indeed, a recent study by Pereboeva et al.10 provided evidence of the applicability of this system for the targeting of adeno- virus in vivo. To create an efficient and versatile tool for targeting and imaging purposes, we constructed lenti- viral vectors pseudotyped with gp64 and displaying avidin and streptavidin. This two-step system is easier to control than the three-step process that relies on the avidin bridge between the biotinylated vector and targeting molecule while offering better valence for coverage because all four biotin-binding sites are avail- able on the lentivirus envelope. The results from the targeting studies in vitro proved Figure 3 Measurement of R2 of the pelleted HepG2 cells 48 h post- transduction (p.t.) in vitro.(a) The results for ferritin-expressing that retargeting to EGFR, CD46 and transferrin receptors lentivirus (black bars) and nontransduced control cells (white increased the transduction of SA/GP64 and in the case of bars) with or without iron supplementation. Means±s.d. of three CD46 also that of AVD/GP64, thus providing potential independent experiments. (b) R2 maps of the tubes from different targets for ex vivo and in vivo work. The targeting was experimental groups containing HepG2 pellets. The region of able to increase the transduction efficiency up to interest (ROI) is represented by a white box. threefold even though a short virus incubation time (15 min) was used. This effect was more significant with SA/GP64 probably due to its higher incorporation on the in the corpus callosum and in the choroid plexus, thus viral membranes translating into more efficient biotin- mimicking the ferritin MRI contrast (Figures 6f–j). No binding capacity. Optimization of virus production hHF expression was detected by quantitative RT-PCR might be needed to attain higher avidin levels on from the livers or the spleens of the experimental animals AVD/GP64 surface. On the other hand, each avidin at 28 days p.t. (data not shown). monomer has a high isoelectric point (pI) of 10.5 and a single oligosaccharide moiety whereas streptavidin has a mildly acidic pIofB6 and is devoid of sugars.33 Owing Discussion to these dissimilarities, streptavidin has the advantage of lower nonspecific binding to lectin-like and negatively Modification of a lentivirus surface for every new gene charged molecules than avidin, which might contribute therapy application can be laborious and time –consum- to the differences seen in targeting efficiency. To some ing, requiring cloning, production and safety evaluation extent, the differential biotin-binding efficacy of these of every new virus. Ideally, one virus platform could vectors might be due to the slight structural deviations be used for many different purposes. To solve this between avidin and streptavidin; biotin-binding sites challenge, we developed a novel versatile lentivirus being more accessible on streptavidin.34 Even though based on avidin–biotin system allowing easy exchange of avidin binds free biotin more efficiently, the biotin the ligand on lentivirus envelope. This strategy opens conjugates have been shown to be more stable with new fields for lentivirus purification, targeting and streptavidin than with avidin.35 imaging applications. In this work, the two latter In line with previous studies by Morizono et al.36 and approaches were studied. Lin et al.37, the separation of targeting moiety and fusion Targeting lentiviral entry is one of the major goals in protein circumvents the problems resulting from the the development of vectors for gene therapy. Selective inability of ligand binding to induce the conformational gene delivery would reduce the potential side effects and changes required for fusion and/or sequestration of ease clinical applications.2 First method described to alter vectors into endosomes and their subsequent degrada- lentiviral tropism was pseudotyping.25 The most widely tion. Consequently, the ligand or antibody binding can used pseudotyping tool is the VSV-G that stabilizes induce receptor activation and lead to endocytosis the otherwise fragile lentivirus and offers a remarkably through clathrin-coated pits, which is the one of the broad tropism.26 However, the cytotoxicity of VSV-G possible pathways for gp64-mediated entry.38,39 All the limits its use and therefore other alternatives have receptors retargeted in this study are known to exploit been sought.26–28 The baculovirus major envelope the clathrin-mediated uptake route.

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 899

Figure 4 Sagittal and axial three-dimensional (3D) images of the virus distribution in the head (a–c) and abdomen (d–f) area at day 2 after intracerebral administration. (a and d) 111In-labeled SA/GP64-hHF lentivirus showed a clear accumulation in the brain but no signal was detected in abdomen area. (b and e) 111In-labeled SA/GP64-hHF lentivirus partially escaping to the circulation resulted in accumulation of the signal in the brain and in the peritoneum area. (c and f) Injection of 111In-labeled multi-Lys-DTPA chelate alone led to signal accumulation in the brain and in the bladder.

Pseudotyping of our vectors with gp64 can be an analyze the subsequent transgene expression in the advantage for applications requiring a broad vector target organs. Imaging of CNS, tumor and cardiovascu- tropism, but it can limit the selective targeting of specific lar tissues could enable evaluation of the most efficient cell types if (strept)avidin-binding molecules are not solely therapeutic approach. Imaging methods can be divided responsible for virus binding. To achieve more stringent into biodistribution and transduction imaging. Transduc- targeting in vivo, cell-binding-defective envelope protein tion imaging visualizes the transgene expression mutants should be considered as described for hemagglu- whereas biodistribution demonstrates the vector location tinin of influenza A,36 measles virus40 and Sindbis in the body. The virus biodistribution may yield an virus16,37,41 envelope proteins, thus allowing the targeting inaccurate picture of transduction pattern, as a virus protein alone direct viral binding. It might also be feasible may be able to enter into many cells but unable to to create similar mutants of gp64 as the motifs responsible express its transgenes in all cells. Therefore, it is desirable for receptor binding have been elucidated.42 We are to obtain images with both imaging modalities to obtain currently investigating the utility of these approaches for conclusive data on the safety and efficacy of gene targeting of (strept)avidin-displaying viruses in vivo. therapy. Noninvasive imaging would be essential to follow In our study, noninvasive SPECT and MRI were the fate of lentivirus vectors in tissues, to quantify combined for the first time to simultaneously monitor relative injected dose in various organs and to lentiviral particle distribution and gene expression in rat

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 900

Figure 5 Longitudinal results of T2*-weighted gradient echo magnetic resonance images in vivo. Rat brain was imaged 3–211 days after injection of (a) 111In-labeled SA/GP64-hHF lentivirus, (b) SA/GP64-LacZ or (c) 111In-labeled multi-Lys-DTPA chelate alone. Lentivirus injection led to signal void area (arrows), corresponding to ferritin expression and subsequent accumulation of iron. Injection of the SA/GP64-LacZ and chelate alone resulted in minor loss of signal at day 3 (arrows) along the needle track caused by bleeding that was no longer detectable at later time points.

brain. SPECT results showed an accumulation of the In conclusion, this work gives proof of principle of signal from 111In-labeled virus at the injection site, the utility of (strept)avidin display as a versatile tool indicating that the radiolabeled virus particles remained for targeting of lentivirus vectors. This conjugate- in the rat brain. We recognize that indirect radiolabeling based strategy is readily adaptable for different targets of the virus might underestimate the particle number in to increase the gene delivery for ex vivo applications. In case not all vector particles are labeled thus requiring a this study, we also extended the use of (strept)avidin second evaluation method to verify the presence of display to image lentivirus biodistribution and using vector particles in tissue. In this respect, we performed a human ferritin as a marker to follow the duration of reverse transcriptase (RT)-PCR from the liver and spleen transgene expression. These new vectors could be used of the animals, confirming that no signal was seen for designing new gene therapy strategies for a variety of elsewhere in the body. Interestingly, accidental injection diseases, such as CNS disorders and cancer. of the virus to the circulation in one rat resulted in clear accumulation to the peritoneal area. There was no signal in the liver, kidneys or bladder, suggesting that the virus stayed intact in the circulation and had tropism to the Materials and methods peritoneum. These results provide evidence of the applicability of SPECT imaging in verifying the success Construction of plasmids for lentiviruses production of virus administration. Envelope plasmids encoding gp64 or (strept)avidin SPECT imaging enables to monitor the success of fused to membrane anchor of VSV-G were constructed vector delivery, but it does not provide evidence of by removing VSV-G from plasmid pCMV-VSV-G and the efficiency of gene transfer. Ferritin was chosen replacing it with a linker (50-AATTCCCTAGGATTTAA as a transgene to study the success of lentiviral gene ATGCTAGCGTTTAAACGATATCG-30 and 50-AATTCG delivery to the rat brain, as it requires no exogenous ATATCGTTTAAACGCTAGCATTTAAATCCTAGGG-30). contrast agents for detection by MRI.22–24 Hemorrhage Gp64 was cut from pBACsurf-1 (Novagen, Darmstadt, caused by injection appeared on T2*-weighted gradient- Germany) with SnaBI and SpeI and cloned to the linker echo images of all animals at day 3 but as acute digested with PmeI and AvrII (New England Biolabs Inc., hematoma resolved within a few days,43 only the specific Ipswich, MA, USA). Construction of avidin- and strep- signal from ferritin was detected. The MRI contrast tavidin-displaying vectors began by the removal of gp64 was seen during the entire 7-month time course gene from the Baavi vector.11 The VSV-G TM/CTD (VSV- investigated, thus confirming stable expression of ferritin GED)18 domain was amplified by PCR using the follow- in the brain. The transgene expression was detected ing primers: forward 50-GGGGTGATACTGGGCTATCC along the needle track, in the corpus callosum and in the AA-30, reverse 50-AGATCTTTACTTTCCAAGTCGGT cuboidal epithelial cells of the choroid plexus. Transduc- TCA-30 and ligated to SmaI (New England Biolabs Inc.) tion of the choroid plexus cells and corpus callosum digested vector giving rise to AVD-VSV-GED baculo- suggests that gp64-pseudotyped lentiviruses exhibit virus vector. Tetrameric avidin was replaced with somewhat similar tropism to baculoviruses.44,45 streptavidin sequence resulting in SA-VSV-GED vector. Altogether, these results indicate that gp64-pseudotyped AVD-VSV-GED and SA-VSV-GED were digested from lentivirus vectors may provide an efficient tool for gene these vectors with SnaBI and SpeI enzymes and cloned to therapy of CNS disorders. Furthermore, the technology pCMV-linker vector digested with PmeI and AvrII. presented here can be adapted to monitor the therapeutic hHF gene (pENTR 221-Ferritin; RZPD German Re- gene delivery to other tissues. By coexpression of a source Center for Genome Research, Berlin, Germany) therapeutic gene with ferritin from a dual expression and LacZ11 were cloned to the third-generation lentivirus cassette, the ferritin expression can be linked with transfer plasmid pBOB-CAG under CAG (chicken beta- treatment efficacy. actin promoter with CMV-IE enhancer) promoter, using

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 901

Figure 6 Immunohistochemistry results showing transgene expression after (a–d) SA/GP64-hHF or (e–i) SA/GP64-LacZ lentivirus inoculation into the rat corpus callosum. (a) Ferritin expression was detected by Perls’ Prussian Blue staining in the (b) corpus callosum and (c) along the needle track 5 days post-transduction (p.t.). (d) Same area as in b detected with an antibody against human heavy chain of ferritin. (e) Ferritin expression in the cuboidal epithelial cells of the choroid plexus 211 days p.t. (f and g) LacZ expression was detected by X-gal staining (h) along the needle track, (i) in the corpus callosum and (j) in the choroid plexus 5 days p.t. c.c., corpus callosum; LV, lateral ventricle. gateway technology (Invitrogen, Carlsbad, CA, USA). All dish were 12.5, 6.25, 13.0 and 25.0 mg, respectively. constructs were verified by sequencing. The transfer plasmids were pBOB-CAG-hHF, pBOB- CAG-LacZ and LV1-GFP.47 The amount of AVD-VSV- Preparation and titering of lentivirus vectors GED or SA-VSV-GED was optimized using a range Third-generation self-inactivating lentiviruses were of 1–9 mg. The virus containing supernatant was prepared by four or five plasmid transient transfection harvested, clarified using a 0.22 mm filter and concen- in 293T cells as described.46 The amounts of pMDLg/ trated by ultracentrifugation (1 95 000 r.p.m., 2 h and pRRE, pRSV-Rev, pGp64 and transfer plasmid per 15 cm 20 min).

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 902 Transforming units of lentiviruses (TU per ml) were 4 1C and washed three times followed by the addition of determined by analyzing the number of virus particles complete growth medium. After 72 h, GFP expression able to transduce cells. BT4C,48 HeLa, HepG2 or SKOV3 was analyzed by flow cytometry (FACSCanto II; BD (ATCC) cells were seeded in 96-well plates at 5000 cells Biosciences). per well. The lentivirus transduction was carried out next day with serial dilutions (1:1000–1:10 00 000) of the Stereotactic injections into rat brain concentrated virus preparation. After 3 days, the cells Male inbred Wistar rats (200–250 g) were anesthetized were analyzed by flow cytometry (FACSCanto II; BD intraperitoneally with a solution containing ketamine Biosciences, San Diego, CA, USA) to reveal the percen- (Ketalar; Pfizer Oy Animal Health, Espoo, Finland) and tage of the GFP-positive cells. Titers were calculated as medetomidine (Domitor; Orion Pharma, Espoo, Finland) described previously.46 The amount of lentiviral capsid and placed into stereotactic apparatus (Kopf Instru- protein p24 (pg ml1) was determined by p24 ELISA kit ments, Tujunga, CA, USA). For SPECT/CT, a total (PerkinElmer, Waltham, MA, USA). volume of 15 ml of the 111In-labeled solution was injected into the right corpus callosum (coordinates: 1.0 mm Immunoblotting caudal to bregma, 1.5 mm right to sutura sagittalis and The display of gp64, streptavidin-VSV-GED and avidin- depth of 3.0–3.5 mm) using a Hamilton syringe with a VSV-GED on the lentivirus surface was detected by using 27-gauge needle. Control rats (n ¼ 3) received chelate mouse anti-gp64 (eBioscience, San Diego, CA, USA), only whereas virus-injected animals received SA/GP64- rabbit anti-avidin, anti-streptavidin and anti-VSV-G hHF (n ¼ 15) or SA/GP64-LacZ lentivirus (n ¼ 7). (Bethyl Laboratories, Montgomery, TX, USA) antibodies. Primary antibodies were detected with secondary alka- In vitro and in vivo MRI line-phosphatase-conjugated goat anti-mouse or anti- To confirm the iron uptake of ferritin-expressing lenti- rabbit antibodies (Bio-Rad, Hercules, CA, USA) followed virus, we measured R2 relaxation times from the by a color reaction (NBT/BCIP; Roche, Basel, Switzer- lentivirus-transduced (MOI 10) HepG2 cells incubated land). in medium with or without 2 mM ferric ammonium citrate supplement (Sigma-Aldrich) 48 h p.t. Samples ELISA were imaged using 4.7 T MRI system (Magnex, Abing- Biotin-coated ELISA plates (Reacti-Bind Biotin-Coated ton, UK) interfaced to Varian Unity INOVA console Polystyrene Strip Plates; Pierce, Rockford, IL, USA) were (Varian Inc., Palo Alto, CA, USA). Standard spin-echo used to analyze the binding capacity of avidin and pulse sequence with time to repetition (TR) of 2 s and streptavidin on lentivirus surface. Plates were washed, time to echo (TE) of 15, 35 and 55 ms was used to obtain blocked with bovine serum albumin (BSA; 40 min, data for T2 maps. In the measurements slice thickness +37 1C; Sigma-Aldrich, St Louis, MO, USA) and incu- was 1.5 mm, field of view (FOV) of 50 50 mm was bated with diluted virus samples (1 h, +37 1C). Bound covered with 256 128 points. Quantitative relaxation virus was detected with mouse anti-gp64 antibody rate maps were computed using standard two-parameter (eBioscience) followed by horseradish peroxidase fit and central region of the pellet was chosen for the (HRP)-conjugated anti-mouse antibody (Silenus Labora- region-of-interest (ROI) analysis of R2 value. tories, Hawthorn, Australia). Finally, the plates were The experimental animals underwent MRI 3, 14, 28, washed and substrate (0.5% TMB diluted 1:50 in 0.1 M 52, 63, 94, 141 and 211 days after the stereotactic injec-

NaAc, 1.5 mM citric acid, 0.005% 30% H2O2) was added. tion. For imaging, animals were anesthetized with

Color development was stopped with 2 M H2SO4 and 1% halothane in 74%:25% N2O/O2. MRI was performed absorbances were measured at 450 nm. using the 4.7 T magnet. Animals were imaged by T2*-weighted gradient-echo sequence (TR/TE ¼ 1.5 s/ Targeting in vitro 0.015 s. The parameters for T2* scans were TR/TE ¼ Cells were seeded at 10 000 cells per 96-well plate in their 1.5 s/0.015 s. Total of 11 slices were imaged from each recommended medium. After 24 h medium was re- brain with a slice thickness of 1 mm, gap between the moved and cells were washed twice with cold 1% BSA- slices being 0.8 mm. FOV of 4 4 cm was covered with PBS (phosphate-buffered saline) and placed on ice for 256 128 points. Image reconstruction and analyses 5 min. Targeting molecules were as follows: biotinylated were performed using MatLab 7.1. (The MathWorks transferrin (5 mg ml1; Molecular Probes/Invitrogen, Inc., Natick, MA, USA). Carlsbad, CA, USA), biotinylated EGF (40 mgml1; Molecular Probes), biotinylated mAb to human transfer- In vivo SPECT/CT imaging rin receptor (1 mg ml1; Ancell Corporation, Bayport, Biotinylated multi-Lys-DTPA conjugate was synthesized MN, USA), biotinylated mAb to human CD46 as described15 and labeled with 111indium chloride (1 mg ml1; Exbio, Praha, Czech Republic) and anti- (Mallinckrodt Medical, Petten, Holland). After incuba- EGFR (2 mg ml1; Cetuximab/Erbitux, IMC-C225; Merck tion the radiochemical purity of the multi-Lys-DTPA was KGaA, Darmstadt, Germany). Cetuximab was biotinyl- checked using instant thin-layer chromatography (SG; ated according to manufacturer’s instruction by Pall Corporation, OEM Healthcare, Ann Arbor, MI, N-hydroxysuccinimid biotin (EZ-Link NHS-Biotin; USA). Virus was added to the labeled multi-Lys-DTPA Pierce). These molecules were diluted 1:50 (ligands) solution and a total volume of 15ml having an activity 1:200 (antibodies) in 1% BSA-PBS before addition to the B2 MBq was injected. In chelate controls, the virus cells (30 min, 4 1C). Cells were washed three times before volume was replaced with labeling buffer. The injections the addition of lentivirus vectors (20 ml) at multiplicity of were given as described above. infection (MOI) of 0.4 (SKOV3, HepG2), 0.6 (BT4C, D54) All animals were imaged during 3-day period. The or 1 (A549, U118MG). Cells were incubated 15 min at animals were anesthetized using 1.5–2.0% isoflurane in

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 903 N2/O2 with ratios 70:30, respectively, and imaged using conditions: 2 min at 50 1C, 10 min at 95 1C, followed the small animal SPECT/CT device (Gamma Medica; by 40 cycles of 15 s at 95 1C and 1 min at 60 1C. Input Northridge, CA, USA). Imaging was performed in two amounts of cDNAs were corrected by amplification of partially overlapping sections covering the head and the 18S ribosomal RNA as an endogenous control using abdomen regions. For planar imaging, 60–90 s acquisi- TaqMan Ribosomal RNA Control Reagents (Applied tion time with energy windows 140 keV±10% and Biosystems). 247 keV±10% for 111In was used. 3D imaging was performed using 64 projections and 60 s per projection. Statistical analyses Matrix size was 80 80 in 125 125 m FOV. CT imaging Statistical analyses were performed by GraphPadPrism was performed using the same coordinates as SPECT 5 (GraphPad Software Inc., San Diego, CA, USA). with 512 projections, 1024 1024 projection matrix size and a voltage of 70 kV. The planar SPECT images were combined to X-ray image with Matlab 7.1. (The Conflict of interest MathWorks Inc.). 3D SPECT reconstruction was carried out by LumGem program (Gamma Medica) using The authors M Kaikkonen, H Lesch, J Pikkarainen, Butterworth six-order low-pass postfiltering with cutoff JRa¨ty, T Vuorio and M Taavitsainen are employees of frequency of 0.3. 3D images with dimensions of Ark Therapeutics Ltd. No competing financial interests 512 512 512 matrix size in 87 87 87 mm FOV were exist for T Huhtala, T Laitinen, P Tuunanen, O Gro¨hn, obtained. Both CT and SPECT images were interpolated ANa¨rva¨nen, K Airenne and S Yla¨-Herttuala. into final 256 256 256 size by using vendor software (Gamma Medica). The images are combined using IDL software (ITT Visual Information Solutions, Washington, Acknowledgements VA, USA) and monitored with Amira software (Mercury, We thank Siiri Va¨isto¨, Juha Ruuskanen, Anne Martikai- Carlsbad, CA, USA). nen, Katja Salminen, Joonas Malinen and Riikka Eisto for excellent technical assistance. We thank Ann-Marie Histology and immunohistochemistry Ma¨a¨tta¨ and Haritha Samaranayake for their help with Randomly selected rats were killed and perfused with the animal work. We are grateful to Olli Laitinen for PBS intracardially 5, 28 and 211 days after injection. fruitful conversations during the course of the project. Brains were removed and frozen within Tissue-Tek We thank Markku Kulomaa for providing streptavidin- OCT Compound (Electron Microscopy Sciences, Forth encoding plasmid and anti-avidin antibody and Edward m Washington, PA, USA) and cryosections of 12 m were A Bayer and Meir Wilchek for the generous gift of anti- cut. Samples from liver, spleen, kidney, hearth, lung and streptavidin. We also thank Paolo Arosio for the rHO2 muscle were collected and frozen in liquid nitrogen antibody. This study was supported by Ark Therapeutics for mRNA analysis. Group Plc., Finnish Academy, Finnish Foundation for Ferritin expression was detected using a modified Cardiovascular Research and Finnish Cultural Founda- Perl’s Prussian Blue staining and antibody against hHF tion, North Savo Regional fund. rHO2.49 In the Perl’s Prussian Blue staining, the tissues sections were incubated in a solution of 7% ferrocyanide and 3% hydrochloric acid (Sigma-Aldrich) and subse- References quently in DAB solution (Zymed Laboratories Inc., San Francisco, CA, USA). For antibody stainings, the 1 Naldini L, Blomer U, Gage FH, Trono D, Verma IM. Efficient primary antibody rHO2 was probed with an HRP- transfer, integration, and sustained long-term expression of the conjugated mouse antibody (Vectastain ABC Kit; Vector transgene in adult rat brains injected with a lentiviral vector. Labs, Burlingame, CA, USA). Sections were counter- Proc Natl Acad Sci USA 1996; 93: 11382–11388. stained in Harris hematoxylin (Merck KGaA) and 2 Waehler R, Russell SJ, Curiel DT. Engineering targeted viral mounted with Permount (Fisher Chemicals, Fair Lawn, vectors for gene therapy. Nat Rev Genet 2007; 8: 573–587. NJ, USA). The LacZ expression of was analyzed as 3 Smith JS, Keller JR, Lohrey NC, McCauslin CS, Ortiz M, Cowan described.46 K et al. Redirected infection of directly biotinylated recombinant adenovirus vectors through cell surface receptors and antigens. Quantitative RT-PCR Proc Natl Acad Sci USA 1999; 96: 8855–8860. The tissue samples were dissolved in TRI Reagent 4 Zhong Q, Kolls JK, Schwarzenberger P. Retrovirus molecular (Sigma-Aldrich). Total RNA was isolated by phenol- conjugates. A novel, high transduction efficiency, potentially chloroform extraction, treated with DNaseI (Promega, safety-improved, gene transfer system. J Biol Chem 2001; 276: Madison, WI, USA) and reverse transcribed. Similar 24601–24607. treatment was carried out for HepG2 cells to obtain a 5 Ponnazhagan S, Mahendra G, Kumar S, Thompson JA, Castillas M. Conjugate-based targeting of recombinant adeno-associated positive control for human ferritin transcription. SYBR virus type 2 vectors by using avidin-linked ligands. J Virol 2002; Green (Applied Biosystems, Foster City, CA, USA) 76: 12900–12907. RT-PCR was used to quantify the amount of human 6 Purow B, Staveley-O’Carroll K. Targeting of vaccinia virus using and rat ferritin mRNA using specific primers: forward biotin-avidin viral coating and biotinylated antibodies. J Surg Res 0 for human and rat ferritin 5 -GGAACATGCTGAGA 2005; 123: 49–54. 0 AACTGATGAAG, reverse for human ferritin 5 -CACA 7ParrottMB,MokH,CamposSK,AdamsKE,MercierGT,Barry GTCTGGTTTCTTGATATCCTGA and reverse for rat MA. Metabolically biotinylated adenovirus for cell targeting, 0 ferritin 5 -CACGGTCAGGTTTCTTTATATCCTGC. Samples ligand screening, and vector purification. Mol Ther 2003; 8: 688–700. were amplified using ABI PRISM 7700 Sequence Detec- 8 Arnold GS, Sasser AK, Stachler MD, Bartlett JS. Metabolic tion System (Applied Biosystems) under the following biotinylation provides a unique platform for the purification and

Gene Therapy Targeting and imaging of (strept)avidin lentivirus MU Kaikkonen et al 904 targeting of multiple AAV vector serotypes. Mol Ther 2006; 14: 30 Schauber CA, Tuerk MJ, Pacheco CD, Escarpe PA, Veres G. 97–106. Lentiviral vectors pseudotyped with baculovirus gp64 efficiently 9 Stachler MD, Bartlett JS. Mosaic vectors comprised of modified transduce mouse cells in vivo and show tropism restriction against AAV1 capsid proteins for efficient vector purification and targeting hematopoietic cell types in vitro. Gene Therapy 2004; 11: 266–275. to vascular endothelial cells. Gene Therapy 2006; 13: 926–931. 31 Kang Y, Xie L, Tran DT, Stein CS, Hickey M, Davidson BL 10 Pereboeva L, Komarova S, Roth J, Ponnazhagan S, Curiel DT. et al. Persistent expression of factor VIII in vivo following Targeting EGFR with metabolically biotinylated fiber-mosaic nonprimate lentiviral gene transfer. Blood 2005; 106: 1552–1558. adenovirus. Gene Therapy 2007; 14: 627–637. 32 Sinn PL, Burnight ER, Hickey MA, Blissard GW, McCray Jr PB. 11 Raty JK, Airenne KJ, Marttila AT, Marjomaki V, Hytonen VP, Persistent gene expression in mouse nasal epithelia following Lehtolainen P et al. Enhanced gene delivery by avidin-displaying feline immunodeficiency virus-based vector gene transfer. J Virol baculovirus. Mol Ther 2004; 9: 282–291. 2005; 79: 12818–12827. 12 Schellingerhout D, Bogdanov A, Marecos E, Spear M, Breake- 33 Green NM. Avidin and streptavidin. Methods Enzymol 1990; 184: field X, Weissleder R. Mapping the in vivo distribution of herpes 51–67. simplex virions. Hum Gene Ther 1998; 9: 1543–1549. 34 Livnah O, Bayer EA, Wilchek M, Sussman JL. Three-dimensional 13 Schellingerhout D, Rainov NG, Breakefield XO, Weissleder R. structures of avidin and the avidin-biotin complex. Proc Natl Quantitation of HSV mass distribution in a rodent brain tumor Acad Sci USA 1993; 90: 5076–5080. model. Gene Therapy 2000; 7: 1648–1655. 35 Pazy Y, Kulik T, Bayer EA, Wilchek M, Livnah O. Ligand exchange 14 Raty JK, Liimatainen T, Wirth T, Airenne KJ, Ihalainen TO, between proteins: exchange of biotin and biotin derivatives between Huhtala T et al. Magnetic resonance imaging of viral particle avidin and streptavidin. JBiolChem2002; 277: 30892–30900. biodistribution in vivo. Gene Therapy 2006; 13: 1440–1446. 36 Lin AH, Kasahara N, Wu W, Stripecke R, Empig CL, Anderson 15 Raty JK, Liimatainen T, Huhtala T, Kaikkonen MU, Airenne KJ, WF et al. Receptor-specific targeting mediated by the coexpres- Hakumaki JM et al. SPECT/CT imaging of baculovirus bio- sion of a targeted murine leukemia virus envelope protein and distribution in rat. Gene Therapy 2007; 14: 930–938. a binding-defective influenza hemagglutinin protein. Hum Gene 16 Morizono K, Xie Y, Ringpis GE, Johnson M, Nassanian H, Lee B et al. Ther 2001; 12: 323–332. Lentiviral vector retargeting to P-glycoprotein on metastatic 37 Morizono K, Bristol G, Xie YM, Kung SK, Chen IS. Antibody- melanoma through intravenous injection. Nat Med 2005; 11: 346–352. directed targeting of retroviral vectors via cell surface antigens. 17 Pariente N, Morizono K, Virk MS, Petrigliano FA, Reiter RE, J Virol 2001; 75: 8016–8020. Lieberman JR et al. A novel dual-targeted lentiviral vector leads 38 Long G, Pan X, Kormelink R, Vlak JM. Functional entry of to specific transduction of prostate cancer bone metastases baculovirus into insect and mammalian cells is dependent on in vivo after systemic administration. Mol Ther 2007; 15: 1973–1981. clathrin-mediated endocytosis. J Virol 2006; 80: 8830–8833. 18 Kaikkonen MU, Raty JK, Airenne KJ, Wirth T, Heikura T, Yla- 39 Matilainen H, Rinne J, Gilbert L, Marjomaki V, Reunanen H, Herttuala S. Truncated vesicular stomatitis virus G protein Oker-Blom C. Baculovirus entry into human hepatoma cells. improves baculovirus transduction efficiency in vitro and in vivo. J Virol 2005; 79: 15452–15459. Gene Therapy 2006; 13: 304–312. 40 Funke S, Maisner A, Muhlebach MD, Koehl U, Grez M, Cattaneo 19 Recht L, Torres CO, Smith TW, Raso V, Griffin TW. Transferrin R et al. Targeted cell entry of lentiviral vectors. Mol Ther 2008; 16: receptor in normal and neoplastic brain tissue: implications for 1427–1436. brain-tumor immunotherapy. J Neurosurg 1990; 72: 941–945. 41 Yang L, Bailey L, Baltimore D, Wang P. Targeting lentiviral 20 Herbst RS. Review of epidermal growth factor receptor biology. vectors to specific cell types in vivo. Proc Natl Acad Sci USA 2006; Int J Radiat Oncol Biol Phys 2004; 59: 21–26. 103: 11479–11484. 21 Wojnicz D, Bar J, Jankowski S. The role of membrane glyco- 42 Zhou J, Blissard GW. Identification of a GP64 subdomain proteins CD46, CD55 and CD59 in protection of tumor cells involved in receptor binding by budded virions of the against complement lysis. Postepy Hig Med Dosw 2002; 56: 603–616. baculovirus Autographica californica multicapsid nucleopoly- 22 Genove G, DeMarco U, Xu H, Goins WF, Ahrens ET. A new hedrovirus. J Virol 2008; 82: 4449–4460. transgene reporter for in vivo magnetic resonance imaging. 43 Bradley Jr WG. Hemorrhage and hemorrhagic infections in the Nat Med 2005; 11: 450–454. brain. Neuroimaging Clin N Am 1994; 4: 707–732. 23 Cohen B, Dafni H, Meir G, Harmelin A, Neeman M. Ferritin as 44 Sarkis C, Serguera C, Petres S, Buchet D, Ridet JL, Edelman L et an endogenous MRI reporter for noninvasive imaging of gene al. Efficient transduction of neural cells in vitro and in vivo by a expression in C6 glioma tumors. Neoplasia 2005; 7: 109–117. baculovirus-derived vector. Proc Natl Acad Sci USA 2000; 97: 24 Cohen B, Ziv K, Plaks V, Israely T, Kalchenko V, Harmelin A et al. 14638–14643. MRI detection of transcriptional regulation of gene expression in 45 Lehtolainen P, Tyynela K, Kannasto J, Airenne KJ, Yla-Herttuala S. transgenic mice. Nat Med 2007; 13: 498–503. Baculoviruses exhibit restricted cell type specificity in rat brain: a 25 Cronin J, Zhang XY, Reiser J. Altering the tropism of lentiviral comparison of baculovirus- and adenovirus-mediated intracer- vectors through pseudotyping. Curr Gene Ther 2005; 5: 387–398. ebral gene transfer in vivo. Gene Therapy 2002; 9: 1693–1699. 26 Burns JC, Friedmann T, Driever W, Burrascano M, Yee JK. 46 Follenzi A, Naldini L. Generation of HIV-1 derived lentiviral Vesicular stomatitis virus G glycoprotein pseudotyped retroviral vectors. Methods Enzymol 2002; 346: 454–465. vectors: concentration to very high titer and efficient gene 47 Lesch HP, Turpeinen S, Niskanen EA, Mahonen AJ, Airenne KJ, transfer into mammalian and nonmammalian cells. Proc Natl Yla-Herttuala S. Generation of lentivirus vectors using recombi- Acad Sci USA 1993; 90: 8033–8037. nant baculoviruses. Gene Therapy 2008; 15: 1280–1286. 27 Park F, Ohashi K, Kay MA. Therapeutic levels of human factor 48 Sandmair AM, Turunen M, Tyynela K, Loimas S, Vainio P, VIII and IX using HIV-1-based lentiviral vectors in mouse liver. Vanninen R et al. Herpes simplex virus thymidine kinase gene Blood 2000; 96: 1173–1176. therapy in experimental rat BT4C glioma model: effect of the 28 Watson DJ, Kobinger GP, Passini MA, Wilson JM, Wolfe JH. percentage of thymidine kinase-positive glioma cells on treat- Targeted transduction patterns in the mouse brain by lentivirus ment effect, survival time, and tissue reactions. Cancer Gene Ther vectors pseudotyped with VSV, Ebola, Mokola, LCMV, or MuLV 2000; 7: 413–421. envelope proteins. Mol Ther 2002; 5: 528–537. 49 Cozzi A, Corsi B, Levi S, Santambrogio P, Albertini A, Arosio P. 29 Kumar M, Bradow BP, Zimmerberg J. Large-scale production of Overexpression of wild type and mutated human ferritin H- pseudotyped lentiviral vectors using baculovirus GP64. Hum chain in HeLa cells: in vivo role of ferritin ferroxidase activity. Gene Ther 2003; 14: 67–77. J Biol Chem 2000; 275: 25122–25129.

Gene Therapy