lentiviral gene delivery to cns by spinal intrathecal administration to neonatal mice

THE JOURNAL OF GENE MEDICINE R E S E A R C H A R T I C L EJ Gene Med 2006; 8: 414–424.Published online 3 January 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jgm.861

Lentiviral gene delivery to CNS by spinal intrathecaladministration to neonatal mice

Elena Fedorova1

Lorenzo Battini1

Ainu Prakash-Cheng2

Daniele Marras1

G. Luca Gusella1*

1Department of Medicine, MountSinai School of Medicine, New York,NY 10029, USA2Department of Human Genetics,Mount Sinai School of Medicine, NewYork, NY 10029, USA

*Correspondence to:G. Luca Gusella, Mount Sinai Schoolof Medicine, One G. Levy Place, Box1243, New York, NY 10029, USA.E-mail: [email protected]

Received: 8 April 2005Revised: 23 August 2005Accepted: 30 September 2005

Abstract

Background Direct injection of lentivectors into the central nervoussystem (CNS) mostly results in localized parenchymal transgene expression.Intrathecal gene delivery into the spinal canal may produce a widerdissemination of the transgene and allow diffusion of secreted transgenicproteins throughout the cerebrospinal fluid (CSF). Herein, we analyze thedistribution and expression of LacZ and SEAP transgenes following theintrathecal delivery of lentivectors into the spinal canal.

Methods Four weeks after intrathecal injection into the spinal canal ofnewborn mice, the expression of the LacZ gene was assessed by histochemicalstaining and by in situ polymer chain reaction (PCR). Following the spinalinfusion of a lentivector carrying the SEAP gene, levels of enzymatically activeSEAP were measured in the CSF, blood serum, and in brain extracts.

Results Intrathecal spinal canal delivery of lentivectors to newborn miceresulted in patchy, widely scattered areas of β-gal expression mostly in themeninges. The transduction of the meningeal cells was confirmed by in situPCR. Following the spinal infusion of a lentivector carrying the SEAP gene,sustained presence of the reporter protein was detected in the CSF, as well asin blood serum, and brain extracts.

Conclusions These findings indicate that intrathecal injections of lentivec-tors can provide significant levels of transgene expression in the meninges.Unlike intracerebral injections of lentivectors, intrathecal gene deliverythrough the spinal canal appears to produce a wider diffusion of the trans-gene. This approach is less invasive and may be useful to address thoseneurological diseases that benefit from the ectopic expression of soluble fac-tors impermeable to the blood-brain barrier. Copyright 2006 John Wiley &Sons, Ltd.

Keywords lentiviral vectors; gene therapy; spinal injections; secreted alkalinephosphatase; beta-galactosidase; central nervous system

Introduction

Gene transduction of cells of the central nervous system (CNS) has greatpotential for the treatment of neurodegenerative disorders and for the devel-opment of biological models of disease pathogenesis. The complexity ofthe brain architecture and the high specialization of its many cell typesrequire that each approach be tailored to the particular pathologic condi-tion. Distinct transgene localization may be achieved by alternative routes ofadministration or different gene transduction approaches. These may lead to

Copyright 2006 John Wiley & Sons, Ltd.

Lentiviral Gene Delivery by Spinal Infusion 415

widespread gene delivery throughout the brain orrestricted targeting of the transgene to specific cellpopulations. Intracerebral gene delivery, which is widelyused in brain gene transfer, allows the prevalent targetingof some cell populations by directly inoculating the brainparenchyma, but results mostly in localized expression ofthe transgene around the site of injection. In contrast,administration of the vector into the cerebrospinalfluid (CSF) results in widespread transgene distribution.Intrathecally administered viral vectors have successfullyled to the production of significant amounts of bioactiveproteins such as IL-2 [1], IL-4 [2], β-glucoronidase[3], β-endorfin [4], Bcl-2 [5], NGF [6], α-1-iduronidase[7], IFN-γ [8], and oncostatin M [9]. CSF genedelivery can be performed either by intraventricularor by spinal canal intrathecal administration. Spinalintrathecal administration into the CSF is less invasivethan intraventricular delivery and has been shown toresult in significant transgene expression in the brain andspinal cord using plasmid DNA, liposome-complexed DNAor poliovirus-based vectors [6,10].

Studies in animal models suggest that intrathecaldelivery may be a promising approach to the treat-ment of neurodegenerative diseases [5], lysosomal stor-age diseases [3,11,12], and experimental autoimmuneencephalomyelitis [13–15]. Gene transfer to the CSF hasalso been applied to the treatment of several conditionssuch as leptomeningeal metastases [16,17] and chronicpain following traumatic injury [1,4] in mouse and ratmodels, respectively.

Different viral vectors have been employed for genetherapy of neurological conditions [18–20]. Among thevarious gene delivery systems, HIV-1-based lentiviralvectors were the first to be successfully utilized for thelong-term transduction of the brain [21–23]. Lentiviralvectors acquire broad tropism when pseudotyped withthe G-glycoprotein envelope from the vesicular stomatitisvirus (VSV-G) and combine the ability to stablyintegrate in the genome of terminally differentiated, non-proliferating neural cells with limited immunogenicity[24,25]. Lentiviral gene transfer to the rodent and primateCNS in vivo has been explored by direct intracerebral orspinal cord administration. By these routes, transgeneexpression has been achieved successfully with lentiviralvectors encoding both marker [21,23,26] and functionalgenes [27–31]. Recently, intraventricular administrationof an enhanced green fluorescent protein (EGFP)-expressing lentivector into the lumen of the mouse lateralventricle has been shown to transduce the ependymallayer throughout the ventricular system [32]. However,to date, no studies have evaluated the efficacy and theCNS distribution of lentiviral vectors delivered into theCSF through the spinal canal. Unlike the direct injectionof VSV-G-pseudotyped lentiviral vectors into the CNS,which results in the localized expression of the transgenein the parenchyma near the injection site, intrathecalgene delivery into the spinal canal may produce a widerdissemination of the transgene and diffusion of secretedtransgenic proteins throughout the CSF. Herein, we

analyze the distribution and the expression of transgenes,β-galactosidase (LacZ) and human secreted alkalinephosphatase (SEAP), following the intrathecal deliveryof lentiviral vectors into the spinal canal.

Materials and methods

Animals

Pregnant adult BALB/c mice, obtained from TaconicFarms (Germantown, NY, USA), and newborns weremaintained under pathogen-free conditions with freeaccess to regular chow in the animal facility at the MountSinai School of Medicine (New York, NY, USA). Adultmice were anesthetized by intraperitoneal injection oftribromoethanol (Avertin, 375 mg/kg).

Construction of self-inactivatinglentiviral vectors

VVCW is a self-inactivating lentiviral vector derived bydeletion of 414 bp of the U3 region, between the integrasebinding site and the R region of the HIV-1 3′ long terminalrepeat (LTR).

The 3′ HIV-1 LTR from the digestion of thepBS-BAR plasmid [33] with BamHI and XhoI wasinserted between the BglII-EcoRI sites of the pcDNA3.0to give the pcD3/GRL plasmid. The XbaI site wasremoved from pcD3/GRL by digestion with XbaI andNcoI, subsequent fill-in with Klenow polymerase andself-ligation to generate pcD3/�GRL. The outwardprimers 5′GCTCTAGAATATCTTGTCTTCTTTGGGAGT3′and 5′GCTCTAGACTGGGTCTCTCTGGTTAGA3′, whichcontain the XbaI site (underlined), were used to amplifythe whole plasmid from positions 37 and 450 of the LTR.All amplification reactions were performed using the HighFidelity Expand kit from Roche (Indianapolis, IN, USA).The amplified product was digested with XbaI and self-circularized to produce pcD3/��GRL that contains theU3-deleted LTR. The proper deletion within the U3 regionof the LTR was confirmed by sequencing.

In parallel, the promoter and enhancer regions of the5′LTR of HIV-1 were substituted with the cytomegalovirus(CMV) early promoter amplified from pcDNA3.0 using theprimers: 5′GAAGATCTGCGATGTACGGGCCAGATATAC3′and 5′GACGATATCTTATATAGACCTCCCACCGTA3′ con-taining the BglII and EcoRV restriction sites (underlined),respectively. The amplified CMV promoter was digestedwith BglII and EcoRV and inserted into the correspondingsites of pcDNA3.1/neo to generate the pCMV/hf plasmid.The RNA starting point of the amplified CMV promotercoincides with the EcoRV site.

The R and U5 regions of the LTR, the primer bind-ing site, and the packaging signals along with the CMVpromoter were derived by digestion of pHR’CMV [22](from which the NotI site was removed by self-ligationfollowing NotI digestion and Klenow fill-in) with PvuII,

Copyright 2006 John Wiley & Sons, Ltd. J Gene Med 2006; 8: 414–424.

416 E. Fedorova et al.

which cuts 15 bases before R, and XhoI, which cutsafter the CMV. The fragment was inserted between theEcoRV and XhoI of pHR’CMV to generate the plas-mid pCLC, which contains the hybrid CMV/5′LTR andmost of the viral backbone. Finally, the modified 3′LTRwas excised from p��GRL with SalI and XhoI andused in a triple ligation with the SalI-SalI fragment ofpcDNA3.1/Hygro and SalI-XhoI fragment of pCLC to gen-erate the plasmid pCRC4. The central polypurine tractwas amplified from pNL4-3 using the oligonucleotides5′TGACATTCGAATTAATTAAAAGAAAAGGGGGGATTG-GGGG3′ and 5′CTATCGATGGCGCGCCAAAATTTTGAAT-TTTTGTAATTTGTTTTTG3′ containing the BstBI and ClaIsites (underlined), respectively, digested with BstBI andClaI and inserted in the ClaI site of pCRC4 to generatethe pVVC vector. The cis-acting posttranscriptional regu-latory element of the woodchuck hepatitis virus (WPRE),which has been shown to enhance transgene expres-sion [34], was amplified from the plasmid pKS/WPRE(kindly provided by Dr. Tom Hope) using the oligonucleo-tides 5′GACTCGAGAATCAACCTCTGGATTACAAAATT-TG3′ and 5′GAGTCGACAGGCGGGGAGGCGGCCCAAA-G3′, digested with XhoI and SalI and inserted in theXhoI site of pVVC to generate the VVCW lentiviral vector.

The LacZ gene from pHR’CMV/LacZ was digested withBamHI and inserted in VVCW under the control of theCMV promoter to obtain the VVCW/LacZ vector.

The VVEW vector was derived by substitution ofthe internal CMV promoter of VVCW, excised withClaI and NheI, with the EF-1 promoter from pCEF-6N,obtained following the partial digestion of pEF-6/V5-His(Invitrogen, Carlsbad, CA, USA) with HindIII, fill-in withKlenow and self-ligation, to create the NheI site, and bythe subsequent addition of a ClaI linker in place of thesecond HindIII site, 5′ of the promoter. The gene for thehuman secreted alkaline phosphatase from the pSEAP-2basic plasmid (Clontech, Palo Alto, CA, USA) was digestedwith NheI and XbaI and inserted in the NheI site of VVEW,under the control of the EF-1 promoter, to generateVVEW/SEAP3. The XbaI site cuts ten amino acids beforethe end of the SEAP gene. The stop codon is reconstitutedin the PmeI site of VVEW and the resulting SEAP maintainsits activity. The final VL/SPES self-inactivating lentiviralvector was generated by ligating the EcoRI-ClaI cassettecontaining the SV40 early promoter and the gene for theresistance to puromycin from the murine retroviral vectorpBABE/Puro [35] between AscI and ClaI of VVEW afterEcoRI and AscI on the respective fragments were bluntedby fill-in with Klenow (Figure 1a).

Production of viral supernatants

Infectious viral supernatants were produced by transienttransfection of 293T cells in suspension [36] usingEffectene (Qiagen, Carlsbad, CA, USA). A total of 4 µgof plasmid DNA/T75 flask was used for each transfectionin the following proportions: 2 µg of viral construct,0.6 µg of pMD.G and 1.4 µg of pCMV�R8.2 (pMD.G

and pCMV�R8.2 were kindly provided by Dr. DiderTrono) and 40 µl of Effectene. Seven million 293T cells incomplete Dulbecco’s modified Eagle’s medium (DMEM)(DMEM supplemented with 10% fetal bovine serumand 2 mM L-glutamine) were mixed with the lipid-DNAcomplexes, and immediately transferred to a poly-L-lysine-coated T75 flask for an overnight incubation at37 ◦C, 5% CO2. The medium was then replaced only forthe first 24 h following the transfection with completemedium supplemented with 4 mM sodium butyrate.Three supernatants were collected at 48, 72 and 96 h,filtered through a 0.45 µm pore-size filter and stored inpools at 4 ◦C until concentration. Viral titer was increasedby centrifugation of supernatants for 5 h at 4000 gand reconstitution of the pellet in phosphate-bufferedsaline (PBS) (0.5% of the initial volume). Viral titration(transducing units (TU)/ml) was performed on 293Tcells immediately after resuspension of the pelleted virususing the endpoint dilution method. Target cells wereseeded at 3 × 104/well in poly-L-lysine-coated 24-wellplates the day before the infection. Concentrated viruswas serially diluted in complete DMEM supplementedwith 10 µg/ml polybrene and added to the target cells at37 ◦C, 5% CO2 for 72 h. The medium was then replacedfor 24 h with complete DMEM supplemented with 4 mMsodium butyrate. The TU of VVCW/LacZ or VL/SPES wasdetermined 96–120 h after the infection by colorimetricdetection (Roche Diagnostics, Indianapolis, IN, USA) ofthe β-galactosidase (β-gal)-positive cells or by measuringthe SEAP activity in the supernatants of infected cells(Clontech). Titration of viral preparations and scoringof titer were performed independently by three people.Using this protocol, the viruses described in this paperachieved titers ranging from 1010 to 1012 TU/ml. Thesetiters are in agreement with those previously shownby Nanmoku et al., who used a similar methodology[37], and correspond to a p24 concentration of 1 µg/ml(107 to 109 viral particles per ng of p24). By usingthe fluorescence-activated cell sorting (FACS) titrationmethod, in which transduction levels between 5 and20% are generally used to deduce the final number ofvirions/ml, the same preparations are titered between108 and 1010 TU/ml. Throughout the paper, we refer tothe titers as measured using the endpoint dilution method.

Intrathecal injections

Surface injections to the spinal canal at the level ofribs were made transcutaneously similarly to the methodreported by Elliger et al. [12]. To track the injectedsolution, lissamine green (Sigma, St. Louis, MO, USA) wasadded to the lentiviral vectors VVCW/LacZ or VL/SPES toa final concentration of 0.1% in PBS and a final viraltiter of 106 to 1012 TU/ml. At this concentration oflissamine green, no inhibition of lentiviral transductionwas observed in vitro. For injections, 10 µl of VVEW/LacZor 20 µl of the VL/SPES lentiviral vectors in PBS or an



equal volume of PBS alone were administered to 2–3-day-old BALB/c newborn mice using an ultra-fine needle(Thiebaum Biomedical Instruments, Thonon-les-Bains,France) attached to a microsyringe over a 2-min period.The appearance of the marker dye in the fontanelles,visible through the skin, indicated the successful deliveryof the lentiviral vector into the spinal canal (Figure 1b).

Tissue processing

Four or six weeks after intrathecal injections, theanimals were anesthetized by intraperitoneal injectionof tribromoethanol (Avertin, 375 mg/kg) and killed bydecapitation. Brains obtained from experimental animalswere bisected and one half was processed for histologicalanalysis, while the other half was used either for thepreparation of protein extracts or genomic DNA. Forhistological studies, the brain halves were fixed in 4%paraformaldehyde in PBS for 4 h at 4 ◦C, transferred into15% sucrose for 12 h, and snap-frozen using the GentleJane snap-freezing system (Instrumedics, Hackensack, NJ,USA) before 5-µm sections were cut with a microtomecryostat (Polysciences, Warrington, PA, USA). Randomlyselected, unfixed brain halves from treated or untreatedanimals were washed in PBS before protein extractswere prepared by homogenization in cell lysis buffer(20 mM Tris-HCl, pH7.5, 150 mM NaCl, 1 mM EDTA,10% glycerol, 1% Triton X-100). Protein concentrationswere determined using the Bio-Rad protein assay (Bio-Rad, Hercules, CA, USA) and equal amounts of brainprotein extract were subjected to determination of SEAPactivity.

Histochemical detection of β-gal

The expression of β-gal was determined in 5-µm sectionsor in tissue samples with a histochemical staining kit(Roche Diagnostics), according to the manufacturer’sprotocol. The sections or tissue samples were first fixedfor 10 min in 2% formaldehyde, washed extensivelywith PBS, and incubated in 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-Gal) solution for 4 h at 37 ◦C.Sections from the brains injected with PBS were usedas negative controls. The sections were then mountedin Aqua-Poly/Mount medium (Polysciences) and β-galexpression was assessed by light microscopy.

CSF collection

Four weeks after administration of lentiviral vectorVL/SPES, mice were anesthetized by intraperitoneal injec-tion of tribromoethanol (Avertin, 375 mg/kg). Muscletissue was separated away from the cisterna magna andthe surface around the cisterna magna was cleaned toremove traces of blood. CSF was obtained by cisternalpuncture with a 30-G needle, and collected via a capillary

tube (pulled into a pipette with a fine point). Sampleswere immediately used for detection of SEAP activity.

Measurement of SEAP

Levels of enzymatically active SEAP in mouse serumand CSF were measured using the Tropix Phospha-Light Chemiluminometric Reporter Assay kit (AppliedBiosystems, Foster City, CA, USA). Assays were performedaccording to the manufacturer’s protocol. Following20 min of incubation in the reaction buffer, luminescencewas measured using a MiniLumat LB 9506 tubeluminometer (EG&G Berthold, Germany).

In situ polymerase chain reaction(PCR)

Amplification parameters, including MgCl2 concentra-tion, pH, and annealing temperature, were optimizedon standard PCR, before in situ PCR. Using frozen tis-sue blocks mounted on charged slides, in situ PCRwas performed according to a three-step methoddescribed by Nuovo [38]. The tissue sections wereincubated for 15 min at 37 ◦C in the presence of pro-teinase K (10 mg/ml; Sigma) to facilitate the penetra-tion of reagents into the cells. In situ PCR was per-formed with the following reaction mixture: 2.5 mMMgCl2, 200 mM dNTPs, 100 mM digoxigenin-11-2′-deoxyuridine-5′-triphosphate (Roche Diagnostics), andprimers (1 ng/ml) specific for the lacZ gene (LacZ5,5′CAAGCCGTTGCTGATTCGAGGCGTTAACCG3′; LacZ3,5′GATCACACTCGGGTGATTACGATCGCGCTG3′). A vol-ume of 100 µl of the PCR mix was applied to each slide,and the slides were covered with frames and placedin a thermocycler (Hybaid, Teddington, Middlesex, UK).Amplification was performed for 20 cycles, which pro-vided a definitive signal and limited the nonspecificbackground staining. After DNA amplification, the tis-sues were washed twice in 0.1X saline–sodium citrate(SSC) at 45 ◦C to eliminate the unbound nucleotides.Digoxigenin-tagged PCR products were detected with acommercially available kit (Roche Diagnostics), accordingto the manufacturer’s instructions. Slides were blockedovernight with 1% bovine serum albumin (BSA) andthen incubated for 2 h with an anti-digoxigenin antibodybound to alkaline phosphatase. After a thorough rinse,the detection reaction was performed according to thenitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phos-phate (NBT/BCIP) method (Roche Diagnostics), whichgenerates a brown precipitate.

Semi-quantitative PCR analysis fromgenomic DNA

Genomic DNA from randomly selected, unfixed brainhalves, spleen and liver of treated and untreated micewas isolated using the Purigene DNA isolation kit (Gentra



Systems, Minneapolis, MN, USA) according to the manu-facturer’s protocol. Genomic DNA (200 ng) in PCR buffer(2.5 mM MgCl2, 200 µM dNTPs, 5U Taq polymerase) wasdivided into two tubes and primers specific for SEAP orβ-actin were added before amplification was performedfor 35 or 30 cycles, respectively. The following primers ata final concentration of 100 nM were used for the specificamplification of the 299 bp SEAP and 104bp β-actinfragments: (forward) 5′CAACGAGGTCATCTCCGTGA3′and (reverse) 5′ACCACAGTCCATGCCATCAC3′; (for-ward) 5′CTGAACCCTAAGGCCAACCGTG3′ and (reverse)5′GGCATACAGGGACAGCACAGCC3′, respectively. Todetermine the sensitivity of the assay, SEAP was ampli-fied at the same time from the VL/SPES plasmid that wasserially diluted to 103, 102, 101, and −100 copies perreaction tube.

Statistical analysis

Serum SEAP levels were expressed in ng/ml usingthe standard curve generated from the positive controlpurified human placental alkaline phosphatase suppliedwith the assay kit. The statistical analysis of controland lentivirus-injected groups was performed using thesoftware program InStat 3.0 applying the two-tailed, non-parametric Wilcoxon test.

Results

Gene delivery via spinal intrathecalinjections

Increasing amounts of the β-gal-expressing VVCW/LacZlentiviral vector (Figure 1a) in a total volume of 10 µlwere administered through the spinal canal into the CSF of2–3-day-old BALB/c newborns, as described in Materialsand Methods (Table 1). The successful delivery into theCSF was monitored by visually following the distributionof the tracking die, lissamine green, that was addedto the viral preparation (Figures 1b and 1c). Littermatecontrol mice were infused with an equal volume ofPBS only. The procedure had a good survival rate asonly one animal died out of 18 animals injected withthe virus and 6 animals injected with PBS. Expressionof the transgene was measured 4 weeks after infectionby sectioning the brains and staining for β-gal. Whenfewer than 107 viral particles were injected, no β-galexpression could be observed (Table 1). The injection of107 lentiviral particles into the CSF resulted in patchy,disseminated small areas of intense β-gal staining onthe ventrolateral surface of the brain that were visiblewith the naked eye, indicating widespread transductionand expression of the lacZ gene (Figure 1d). The controlPBS-injected brains showed no transduction (Figures 1eand 1g). In all the animals infused with the higherdose of lentivector, transgene expression was detectablemostly in non-parenchymal cells of the brain (Figure 1f).

Microscopic analysis showed that the cells of pia materwere preferentially transduced. With only one exception,in which one parenchymal cell stained positive for β-gal, no β-gal expression was detected in the neuralparenchymal elements of the CNS or in nerves regardlessof the amount of injected viral particles (data not shown).Longitudinal sections from virus-injected mouse brains aswell as brains from PBS-treated and untreated mice wereexamined morphologically. No morphological differenceswere observed between the treated and untreated mousebrains 4 weeks after the inoculation (Figures 1h and 1i).We did not observe inflammatory response in the mouseCNS either 1 (data not shown) or 4 weeks after thetreatment (Figures 1h and 1i). Perivascular aggregatesof inflammatory cells, which are a common findingsecondary to viral infection of meninges, were not seen inany of the slides.

Detection of the transgene by in situPCR

To determine the presence of the transgene in transducedbrain cells, sections prepared from brains harvested4 weeks after virus or PBS treatment were subjected toin situ PCR, using primers specific for the lacZ gene.The lacZ gene was specifically amplified from the virus-infused brains whereas no amplification was observed inPBS-injected brains (Figure 1j). In situ PCR revealed thepresence of the lacZ gene on the surface of the brainconfirming the prevalent distribution of β-gal in the cellsof the meninges. The in situ PCR also showed that theneural parenchymal elements of the CNS were negativefor lacZ.

Expression of a soluble reporterfollowing intrathecal delivery oflentiviral vector

The pattern of transgene distribution indicated that thisroute of administration may be particularly suitable forthe expression of soluble factors that may diffuse acrossthe brain and be internalized in regions of the braindistant from the transduced cells. To determine thelevel of transgene released into the CSF following thisapproach, we delivered the VL/SPES lentiviral vector(Figure 1a) carrying the SEAP gene under the control ofthe EF-1 promoter (Table 2). As no adverse effects wereobserved with the infusion of 10 µl of VVCW/LacZ, andexpecting that the dilution of the soluble factor in the CSFwould hamper its detection, we increased the volume ofVL/SPES infusion to 20 µl, a volume still lower than thatpreviously used by others [12]. The survival rate in thisset of experiments was similar to that observed in theprevious animal group (Table 1) with no correlation tothe dose of administered virus. Initially, we focused onthe detection of SEAP in the CSF, using the VVCW/LacZ-or PBS-infused animals as control baseline of reporter



Figure 1. Lentiviral vector delivery and transgene expression following intrathecal injection of VVCW/LacZ into spinal canal.(a) Schematic of the self-inactivating lentiviral vectors VVCW/LacZ and VL/SPES. CMV: early cytomegalovirus promoter; SD:splicing donor; SA: splicing acceptor; RRE: REV responsive element; CPP: central polypurine tract; EF-1: human EF-1 (elongationfactor-1) promoter; SV40 ori: SV40 origin of replication/promoter; puro: puromycin acetyltransferase gene; LacZ: β-galactosidasegene; SEAP: human secreted placental alkaline phosphatase gene; WPRE: woodchuck hepatitis virus posttranscriptional regulatoryelement; � deletion in U3 LTR. (b, c) Lentiviral vector injected through the spinal canal diffuses in the CSF. A tracing dye, lissaminegreen, was added to the lentiviral vector preparation to monitor the localization of the injected virus. The presence of the markerdye in the fontanelles of the 3-day-old mouse, visible through the skin, indicated the successful delivery of the lentiviral vectorinto the spinal canal (b). The rapid diffusion of the lissamine green after the injection accounts for the staining in the tissuessurrounding the brain. An untreated newborn mouse is shown as control (c). (d, e) Expression of β-gal on the surface of adultmouse brain after administration of VVCW/LacZ lentivector into the spinal canal of newborn mice. Brains from VVCW/LacZ- andvehicle (PBS)-injected mice were harvested 4 weeks after infusion and stained for β-gal expression. Diffuse areas of transgeneexpression (some of which are indicated by the arrows) were observed on the ventrolateral surface of the brain in all mice injectedwith lentiviral vector VVCW/LacZ (d, original magnification 20×). No positive staining was present in brains of control untreatedmice (data not shown) or mice injected intrathecally with PBS (e, original magnification 20×). (f–i) Histological analysis of mousebrains 4 weeks after spinal intrathecal administration of VVCW/LacZ. Positive β-gal staining was detectable in the meninges of thebrain from the VVCW/LacZ-treated animals (f, original magnification 1000×), but not from the PBS-treated control (g, originalmagnification 1000×). Morphological analysis by hematoxylin and eosin staining of serial longitudinal frozen sections indicated nodifferences between the brains of the lentiviral vector-treated (h, original magnification 100×) and untreated animals (i, originalmagnification 100×). Insert magnification: 1000×. Expression of the transgene following viral delivery was not associated withincreased cellularity. (j) Detection of the transgene by in situ PCR. In situ PCR was performed on murine brains untreated or4 weeks following the spinal intrathecal infusion of the VVCW/LacZ lentiviral vector. The presence of the LacZ gene was confirmedin the meningeal cells on the brain surface. Some of the positive cells are indicated by the arrows. Original magnification 400×



Table 1. Mouse groups treated by intrathecal injections oflentiviral vector VVCW/LacZ

InjectionInjected

transducing unitsTreated

miceSurviving

miceβ-gal

expression

PBS 0 6 5 –VVCW/LacZ: 1 × 104 5 5 –

1 × 105 6 6 –1 × 107 7 7 + (n = 7)

Table 2. Mouse groups treated by intrathecal injections oflentiviral vector VL/SPES

InjectionInjected

transducing unitsTreated

miceSurviving

miceSEAP

expression

PBS 0 11 11 –VL/SPES: 2 × 108 17 15 + (n = 15)

2 × 109 5 5 + (n = 4)2 × 1010 5 4 + (n = 4)

activity. CSF of all VL/SPES-infused but not controlmice contained significant levels of SEAP 4 weeks afterthe treatment (p < 0.0002; Figure 2a). Correspondingly,SEAP was measurable in the total protein extractsobtained from brain hemispheres inclusive of pia layers ofanimals transduced with VL/SPES but not in those infusedwith PBS (Figure 2b). The collection of CSF from micerequired the sacrifice of the animals, thereby makingstudies of the kinetics of transgene expression in CSFimpossible to follow in the same subject.

Since proteins of the spinal fluid may enter thecirculation through the intercellular space of thechoriomeningeal cells of the sagittal sinus, we testedthe blood of lentivirus- and PBS-treated animals for thepresence of SEAP immediately before the collection ofthe CSF. Indeed, the levels of SEAP were significantlyhigher in the sera of VL/SPES injected mice than incontrol animals, with a mean of 12 ng/ml (p < 0.05;Figure 2c).

We used the measurement of serum SEAP as aproxy of the kinetics of expression of the reporter inthe CSF. Blood from VL/SPES- or PBS-infused animalswas collected at different time points and assessedfor SEAP levels (Figure 3). Expression of SEAP couldbe observed as early as day 8 post-treatment in thelentivector- but not in PBS-treated mice. We found aprogressively higher SEAP expression in the blood withlevels of SEAP increasing within 3 weeks after injection(Figure 3). This more likely reflects the accumulationof the enzyme in serum, rather than expansion oftransduced neural precursors, which would have shownparenchymal staining [32]. The differences in SEAP serumlevels observed between the lentivirus-treated and thecontrol groups remained highly statistically significantduring the entire length of the experiment (day 42).Despite the seemingly uneven SEAP expression in thetransduced animals, the variability was not statisticallysignificant and within the range observed in other in vivo

Figure 2. Expression of SEAP after spinal intrathecal deliveryof the VL/SPES lentivector into newborn mice. Analysis ofSEAP expression was performed on CSF (a), brain extractsinclusive of pial layers (b), and blood serum (c) 4 weeks afterinfusion. Mice injected with PBS were used as control. In allcases, VL/SPES infusion of 2 × 108 TU resulted in statisticallysignificant increase in SEAP activity (p < 0.0001 for CSF andserum; p < 0.05 for brain extracts)

systems for the transduction of different soluble factors[39–41].

The expression of SEAP activity paralleled the sustainedexpression of β-gal that we observed following intrathecalinfusion and was confirmed by semi-quantitative PCRusing specific primers for SEAP in brains from treatedanimals (Figure 4). Amplification of the β-actin gene wasperformed as a comparative analysis from one half of thesame sample that was amplified for SEAP. The sensitivityof PCR detection was estimated amplifying SEAP fromthe serially diluted (from 103 to 100 copies) VL/SPEScontrol plasmid. To rule out that SEAP expression in bloodresulted from the diffusion of the virus in the circulationduring the intrathecal delivery, we also looked for thepresence of the reporter gene in liver and spleen, whichhave been shown to be the major targets of similarly



Figure 3. Sustained expression of SEAP delivered by a lentiviralvector. The serum SEAP level was measured after spinalintrathecal injections of lentiviral vector VL/SPES or vehicle(PBS) into newborn mice. Different time points were measuredfrom two groups of animals that were followed up to 21 (fourmice injected with 2 × 108 TU of VL/SPES and two PBS-treatedmice; solid line) or from 21 to 42 days (five mice injected with2 × 108 TU of VL/SPES and two PBS-treated mice; broken line).Symbols represent individual animals. Values of SEAP are givenas the amount of enzyme after subtraction of the backgroundbasal level

Figure 4. Detection of SEAP transgene in lentivirus-transducedmice by semi-quantitative PCR. Animals were sacrificed 4 weeksafter the infusion with VL/SPES and spleen, liver and brain werecollected for the isolation of genomic DNA. A 299 bp fragmentfrom the SEAP transgene or 104 bp fragment from the β-actinwas amplified by PCR as described in Materials and Methods.As positive control and to gauge the sensitivity of the assay,VL/SPES plasmid DNA serially diluted to 103, 102, 101, and100 copies per reaction was amplified in the same PCR. Theexperiment shows one animal (serum SEAP = 12 ng/ml) out ofsix representative samples. Control indicates the reaction inwhich DNA was omitted. MW: DNA molecular weight marker(bp)

pseudotyped lentivectors upon systemic administration[42,43]. Under the same PCR conditions and sensitivity,no amplification of the SEAP gene was detectable inliver and spleen from the VL/SPES-infused animals(Figure 4), indicating that the reporter protein measuredin blood indeed derives from the expression of thetransgene in the pia. Furthermore, these findings suggestthat the stability of the transgene delivered by spinalintrathecal injection may be similar to that reportedfollowing the injection of lentiviral vector into the brainparenchyma [23]. Despite the relatively high levels of

SEAP released in serum of treated animals, no specificantibodies against the heterologous protein were detectedby enzyme-linked immunosorbent assay (ELISA) in serum(dilution 1 : 10-1 : 10 000) of immunocompetent mice atdays 14, 21, and 35 following VL/SPES injection (datanot shown).

Discussion

This study analyzed the feasibility of gene deliverythrough the spinal canal using replication-defective self-inactivating lentiviral vectors. We showed that followingintrathecal injection into the spinal canal of newbornmice, the lentivectors reached different areas of the CNSvia the CSF in the subarachnoid space. The vectorsappeared not to gain access to the parenchymal cellsof the brain and spinal cord, but transduced the piamater and leptomeningeal cells lining the CSF space onthe ventrolateral surface of the brain. This transductionpattern resembles the distribution of the transgeneobserved upon administration of HSV-1 and adenoviralvectors into the CSF [2,4,44,45]. Intracisternal injectionsof HSV vectors efficiently transduced the ependymal orleptomeningeal cells in the ventricular and subarachnoidspaces rostrally or caudally to the site of injection,but not parenchymal cells [2,8]. Transgene expressionwas observed in cells lining the CSF space includingthose of the Virchow-Robin spaces and the ependymal,choroidal, and leptomeningeal cells forming the blood-CSF barrier surrounding both the brain and the spinalcord. Administration of adenoviral vectors into theCSF of nonhuman primates led to the transductionof the meninges covering the brain and spinal cord,with transgene expression found predominantly in thearachnoid cells and to a lesser extent in the cellsof the pia mater [44]. Ependymal cells lining thecerebral ventricles could also be targeted followingintracerebroventricular injection of adenoviral vectors[4]. Differently from what was observed with adenoviralvectors, in which a clear immune cell infiltrate occurredfollowing intrathecal infusion, we did not detect anymorphological abnormality, cellularity or side effects inthe mouse CNS 1 (data not shown) or 4 weeks followingthe administration of the lentivector through the sameroute. More sensitive staining with anti-CD3 or anti-F4/80 antibodies excluded the presence of T lymphocytesor macrophages in the areas surrounding the transducedcells (data not shown). Our findings support the evidenceof the lower immunogenicity of the lentiviral vectors[24,25]. An advantage offered by this delivery route is thatthe infusion of the lentivector into the CSF bypasses thetransitory cell damage observed following direct injectionof the CNS, which is likely due to the toxicity of the VSV-Genvelope or to the activation of the cells of the microglia[26].

AAV vectors demonstrated quantitative and qualitativedifferences in the transduction patterns according to theserotype. AAV1 and AAV2 vectors transduced mostly



neurons after intraventricular injections of neonatalmice brain. In particular, transduction with AAV1vectors resulted in widespread distribution and long-term expression of the β-glucoronidase gene and reversalof the diseased phenotype in a mouse model formucopolysaccharidosis type VII [3]. In contrast, AAV5vectors transduced poorly following intraventricularinjection whereas AAV4 showed a strong preference forthe transduction of ependymal rather than parenchymalcells [46]. These differences are likely due to the variabledistribution of attachment receptors for the recombinantviruses, to the natural viral tropism as well as to theroute of administration that provide physical access tocertain cell types. Similarly, it is likely that delivery intothe CSF of lentiviral vectors pseudotyped with differentenvelopes [7,47,48] may achieve variable distribution andtransduction efficiency.

Various laboratories have shown that the expressionin the meninges of a diffusible transgene can affect thecells of the brain parenchyma through the circulation ofthe spinal fluid. For example, Furlan et al. reported theinduction of MHC class II in parenchyma cells followingthe intrathecal delivery of HSV-1 vector carrying theinterferon-γ gene [8]. Here, we showed that followingthe lentivector delivery via spinal intrathecal infusion, asecreted reporter protein can be released into the CSF.The use of SEAP does not allow the determination ofthe penetration of the transgenic protein into the brainparenchyma, and the enzyme measured in the brainextracts may solely derive from transduced cells in the piamater. However, our data clearly indicate that the releaseof a soluble factor from transduced meningeal cells candiffuse into the systemic circulation, possibly through thearachnoid granulation in the superior sagittal sinus. Thesustained expression of the heterologous human SEAP inmice was unexpected since the homology of murine andhuman SEAP is less than 80%. Despite this heterogeneity,no antibody response to SEAP could be detected in thetreated animals for up to 6 weeks. The early expressionof heterologous SEAP in the newborn mice, which arenot fully immunologically competent, may explain thelack of anti-SEAP response. Indeed, intrathymic injectionof foreign antigens has been successfully used to inducetolerance in neonatal mice [49,50]. Our findings suggestthat early systemic expression, rather than intrathymicdelivery, of a heterologous antigen may be sufficient toachieve tolerance in mice. Further work, beyond the scopeof the present study, will be necessary to address thesequestions.

With a survival rate higher than 93%, intrathecalinfusion via the spinal canal can be very useful forthe generation of small experimental models to studythe pathogenesis of neurological diseases and CNSbiology. This route of administration also offers severaladvantages for functional gene therapy. The feasibilityof the meningeal gene therapy approach for secretedproteins is based on the accessibility of the pia matercells following intrathecal delivery into the spinal canaland on the ability of the lentiviral vectors to stably

transduce terminally differentiated cells. This minimallyinvasive procedure allows the wide distribution of thelentivector around the CNS. The spatial distribution ofthe transgene on the meninges facilitates the diffusionof the secreted enzyme to nontransduced areas and theuptake by distal cells [51]. As no inflammatory responsewas observed following the lentivector delivery, multipleadministrations may be feasible to increase the numberof transduced cells. Finally, the inherent characteristicof lentiviral vectors to integrate into genomic DNAprovides the long-term expression of the transgene inbrain [21,52], while the terminal differentiation of themeningeal cells reduces the concerns about possibleactivation of proto-oncogenes [19]. It should be notedthat the lack of common endpoint measurements in theexperiments reported in literature, which involve differentvector systems across the various routes of administration,prevents any direct quantitative comparison. Followingspinal canal delivery, the restriction of the transgeneexpression to the meninges may appear low withoutconsidering the large surface of the brain. However, thesignificance of this expression can be better appreciatedconsidering the robust systemic accumulation of SEAP,which reaches levels over 30 ng/ml. As these expressionlevels are sustained, they may reach the therapeuticranges for the treatment of the severe neurodegenerativecomplications associated with the deficiency of solubleenzymes in genetic diseases such as Niemann PickA, Gaucher, mucopolysaccharidosis IIIB, and Tay-Sachs[53–55]. The enzyme replacement therapies currentlyemployed, or under development to treat these diseases,present some limitations. In fact, they rely on bolusintravenous injection of large amounts of recombinantfactors that have a very short serum half-life and thatcannot cross the blood-brain barrier. Gene delivery bylentivectors into the spinal canal can be performedearly after birth to provide enzyme access to the CNS.The continuous sustained expression of the normalgene within the first days of life could significantlyreduce the neurological damage observed in patientsaffected by these diseases. As this delivery approachis simple and lacks side effects, intrathecal spinalcanal delivery may also be useful to address CNSconditions such as amyotrophic lateral sclerosis, spinalcord injuries, and chronic pain, in which large numbersof neurons in the brain or the spinal cord are affected[18].

Acknowledgements

We thank Michael Lipkowitz (Mount Sinai School of Medicine)and Sujana Chandrasekhar (Mount Sinai School of Medicine)for their helpful suggestions, and Denise McAloose (MountSinai School of Medicine) for the assistance with neu-ropathological examination of histological sections. This workwas supported by the National Institutes of Health grantRO1DK63611.



References

1. Yao MZ, Wang JH, Gu JF, et al. Interleukin-2 gene has superiorantinociceptive effects when delivered intrathecally. Neuroreport2002; 13: 791–794.

2. Furlan R, Poliani PL, Galbiati F, et al. Central nervous systemdelivery of interleukin 4 by a nonreplicative herpes simplex type1 viral vector ameliorates autoimmune demyelination. HumGene Ther 1998; 9: 2605–2617.

3. Passini MA, Watson DJ, Vite CH, et al. Intraventricular braininjection of adeno-associated virus type 1 (AAV1) in neonatalmice results in complementary patterns of neuronal transductionto AAV2 and total long-term correction of storage lesions in thebrains of beta-glucuronidase-deficient mice. J Virol 2003; 77:7034–7040.

4. Finegold AA, Mannes AJ, Iadarola MJ. A paracrine paradigm forin vivo gene therapy in the central nervous system: treatment ofchronic pain. Hum Gene Ther 1999; 10: 1251–1257.

5. Azzouz M, Hottinger A, Paterna JC, et al. Increased motoneuronsurvival and improved neuromuscular function in transgenicALS mice after intraspinal injection of an adeno-associated virusencoding Bcl-2. Hum Mol Genet 2000; 9: 803–811.

6. Meuli-Simmen C, Liu Y, Yeo TT, et al. Gene expression along thecerebral-spinal axis after regional gene delivery. Hum Gene Ther1999; 10: 2689–2700.

7. Watson DJ, Kobinger GP, Passini MA, et al. Targetedtransduction patterns in the mouse brain by lentivirus vectorspseudotyped with VSV, Ebola, Mokola, LCMV, or MuLV envelopeproteins. Mol Ther 2002; 5: 528–537.

8. Furlan R, Brambilla E, Ruffini F, et al. Intrathecal delivery ofIFN-gamma protects C57BL/6 mice from chronic-progressiveexperimental autoimmune encephalomyelitis by increasingapoptosis of central nervous system-infiltrating lymphocytes.J Immunol 2001; 167: 1821–1829.

9. Kerr C, Langdon C, Graham F, et al. Adenovirus vectorexpressing mouse oncostatin M induces acute-phase proteinsand TIMP-1 expression in vivo in mice. J Interferon Cytokine Res1999; 19: 1195–1205.

10. Jackson CA, Cobbs C, Peduzzi JD, et al. Repetitive intrathecalinjections of poliovirus replicons result in gene expression inneurons of the central nervous system without pathogenesis.Hum Gene Ther 2001; 12: 1827–1841.

11. Elliger SS, Elliger CA, Lang C, et al. Enhanced secretion anduptake of beta-glucuronidase improves adeno-associated viral-mediated gene therapy of mucopolysaccharidosis type VII mice.Mol Ther 2002; 5: 617–626.

12. Elliger SS, Elliger CA, Aguilar CP, et al. Elimination of lysosomalstorage in brains of MPS VII mice treated by intrathecaladministration of an adeno-associated virus vector. Gene Ther1999; 6: 1175–1178.

13. Furlan R, Pluchino S, Marconi PC, et al. Cytokine gene deliveryinto the central nervous system using intrathecally injectednonreplicative viral vectors. Methods Mol Biol 2003; 215:279–289.

14. Furlan R, Poliani PL, Marconi PC, et al. Central nervous systemgene therapy with interleukin-4 inhibits progression of ongoingrelapsing-remitting autoimmune encephalomyelitis in BiozziAB/H mice. Gene Ther 2001; 8: 13–19.

15. Martino G, Poliani PL, Furlan R, et al. Cytokine therapy inimmune-mediated demyelinating diseases of the central nervoussystem: a novel gene therapy approach. J Neuroimmunol 2000;107: 184–190.

16. Reijneveld JC, Taphoorn MJ, Voest EE. A simple mouse modelfor leptomeningeal metastases and repeated intrathecal therapy.J Neurooncol 1999; 42: 137–142.

17. Vincent AJ, Vogels R, Someren GV, et al. Herpes simplex virusthymidine kinase gene therapy for rat malignant brain tumors.Hum Gene Ther 1996; 7: 197–205.

18. Costantini LC, Bakowska JC, Breakefield XO, et al. Gene therapyin the CNS. Gene Ther 2000; 7: 93–109.

19. Davidson BL, Breakefield XO. Viral vectors for gene delivery tothe nervous system. Nat Rev Neurosci 2003; 4: 353–364.

20. Baker D, Hankey DJ. Gene therapy in autoimmune,demyelinating disease of the central nervous system. Gene Ther2003; 10: 844–853.

21. Naldini L, Blomer U, Gage FH, et al. Efficient transfer,integration, and sustained long-term expression of the transgene

in adult rat brains injected with a lentiviral vector. Proc Natl AcadSci U S A 1996; 93: 11382–11388.

22. Naldini L, Blomer U, Gallay P, et al. In vivo gene delivery andstable transduction of nondividing cells by a lentiviral vector.Science 1996; 272: 263–267.

23. Blomer U, Naldini L, Kafri T, et al. Highly efficient and sustainedgene transfer in adult neurons with a lentivirus vector. J Virol1997; 71: 6641–6649.

24. Kafri T. Gene delivery by lentivirus vectors an overview. MethodsMol Biol 2004; 246: 367–390.

25. Baekelandt V, Eggermont K, Michiels M, et al. Optimizedlentiviral vector production and purification procedure preventsimmune response after transduction of mouse brain. Gene Ther2003; 10: 1933–1940.

26. Zhao C, Strappe PM, Lever AM, et al. Lentiviral vectors for genedelivery to normal and demyelinated white matter. Glia 2003;42: 59–67.

27. Bensadoun JC, Deglon N, Tseng JL, et al. Lentiviral vectors asa gene delivery system in the mouse midbrain: cellular andbehavioral improvements in a 6-OHDA model of Parkinson’sdisease using GDNF. Exp Neurol 2000; 164: 15–24.

28. Blomer U, Kafri T, Randolph-Moore L, et al. Bcl-xL protects adultseptal cholinergic neurons from axotomized cell death. Proc NatlAcad Sci U S A 1998; 95: 2603–2608.

29. Bosch A, Perret E, Desmaris N, et al. Reversal of pathology inthe entire brain of mucopolysaccharidosis type VII mice afterlentivirus-mediated gene transfer. Hum Gene Ther 2000; 11:1139–1150.

30. Kordower JH, Emborg ME, Bloch J, et al. Neurodegenerationprevented by lentiviral vector delivery of GDNF in primatemodels of Parkinson’s disease. Science 2000; 290: 767–773.

31. Tanase K, Teng Q, Krishnaney AA, et al. Cervical spinal corddelivery of a rabies G protein pseudotyped lentiviral vectorin the SOD-1 transgenic mouse. Invited submission from theJoint Section Meeting on Disorders of the Spine and PeripheralNerves, March 2004. J Neurosurg Spine 2004; 1: 128–136.

32. Consiglio A, Gritti A, Dolcetta D, et al. Robust in vivo genetransfer into adult mammalian neural stem cells by lentiviralvectors. Proc Natl Acad Sci U S A 2004; 101: 14835–14840.

33. Cara A, Rybak SM, Newton DL, et al. Inhibition of HIV-1replication by combined expression of gag dominant negativemutant and a human ribonuclease in a tightly controlled HIV-1inducible vector. Gene Ther 1998; 5: 65–75.

34. Zufferey R, Donello JE, Trono D, et al. Woodchuck hepatitisvirus posttranscriptional regulatory element enhancesexpression of transgenes delivered by retroviral vectors. J Virol1999; 73: 2886–2892.

35. Morgenstern JP, Land H. Advanced mammalian gene transfer:high titre retroviral vectors with multiple drug selection markersand a complementary helper-free packaging cell line. NucleicAcids Res 1990; 18: 3587–3596.

36. Escobedo J, Koh TJ. Improved transfection technique foradherent cells using a commercial lipid reagent. Biotechniques2003; 35: 936–938, 940.

37. Nanmoku K, Kawano M, Iwasaki Y, et al. Highly efficient genetransduction into the brain using high-titer retroviral vectors.Dev Neurosci 2003; 25: 152–161.

38. Nuovo GJ. PCR in situ hybridization. Methods Mol Biol 1994;33: 223–241.

39. Bloquel C, Bessis N, Boissier MC, et al. Gene therapy of collagen-induced arthritis by electrotransfer of human tumor necrosisfactor-alpha soluble receptor I variants. Hum Gene Ther 2004;15: 189–201.

40. Park F, Ohashi K, Kay MA. The effect of age on hepatic genetransfer with self-inactivating lentiviral vectors in vivo. Mol Ther2003; 8: 314–323.

41. Zinn KR, Szalai AJ, Stargel A, et al. Bioluminescence imagingreveals a significant role for complement in liver transductionfollowing intravenous delivery of adenovirus. Gene Ther 2004;11: 1482–1486.

42. Follenzi A, Battaglia M, Lombardo A, et al. Targeting lentiviralvector expression to hepatocytes limits transgene-specificimmune response and establishes long-term expression ofhuman antihemophilic factor IX in mice. Blood 2004; 103:3700–3709.

43. Peng KW, Pham L, Ye H, et al. Organ distribution of geneexpression after intravenous infusion of targeted and untargetedlentiviral vectors. Gene Ther 2001; 8: 1456–1463.



44. Driesse MJ, Kros JM, Avezaat CJ, et al. Distribution ofrecombinant adenovirus in the cerebrospinal fluid of nonhumanprimates. Hum Gene Ther 1999; 10: 2347–2354.

45. Martino G, Poliani PL, Marconi PC, et al. Cytokine gene therapyof autoimmune demyelination revisited using herpes simplexvirus type-1-derived vectors. Gene Ther 2000; 7: 1087–1093.

46. Davidson BL, Stein CS, Heth JA, et al. Recombinant adeno-associated virus type 2, 4, and 5 vectors: transduction of variantcell types and regions in the mammalian central nervous system.Proc Natl Acad Sci U S A 2000; 97: 3428–3432.

47. Mazarakis ND, Azzouz M, Rohll JB, et al. Rabies virusglycoprotein pseudotyping of lentiviral vectors enablesretrograde axonal transport and access to the nervous systemafter peripheral delivery. Hum Mol Genet 2001; 10: 2109–2121.

48. Watson DJ, Wolfe JH. Lentiviral vectors for gene transfer tothe central nervous system. Applications in lysosomal storagedisease animal models. Methods Mol Med 2003; 76: 383–403.

49. DeMatteo RP, Chu G, Ahn M, et al. Long-lasting adenovirustransgene expression in mice through neonatal intrathymictolerance induction without the use of immunosuppression.J Virol 1997; 71: 5330–5335.

50. Hamilton T, DeMatteo R, McClane S, et al. Adenoviral-mediated gene transfer to murine small intestine is moreefficient in neonates than adults. J Pediatr Surg 1997; 32:373–377.

51. Skorupa AF, Fisher KJ, Wilson JM, et al. Sustained productionof beta-glucuronidase from localized sites after AAV vectorgene transfer results in widespread distribution of enzymeand reversal of lysosomal storage lesions in a large volumeof brain in mucopolysaccharidosis VII mice. Exp Neurol 1999;160: 17–27.

52. Jakobsson J, Ericson C, Jansson M, et al. Targeted transgeneexpression in rat brain using lentiviral vectors. J Neurosci Res2003; 73: 876–885.

53. Futerman AH, van Meer G. The cell biology of lysosomal storagedisorders. Nat Rev Mol Cell Biol 2004; 5: 554–565.

54. Cheng SH, Smith AE. Gene therapy progress and prospects:gene therapy of lysosomal storage disorders. Gene Ther 2003;10: 1275–1281.

55. Scriver CR, Stanbury JB. The Metabolic and Molecular Bases ofInherited Disease (7th edn). McGraw-Hill Health ProfessionsDivision: New York, 1995; 3 (v–xxxvi), 4605, I-4693.


lentiviral gene delivery to cns by spinal intrathecal administration to neonatal mice

Documents