Lentivirally transduced dendritic cells as a tool for cancer immunotherapy

THE JOURNAL OF GENE MEDICINE R E S E A R C H A R T I C L EJ Gene Med 2003; 5: 654–667.Published online 23 April 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jgm.400

Lentivirally transduced dendritic cells as a tool forcancer immunotherapy

Karine Breckpot1†

Melissa Dullaers1†

Aude Bonehill1

Sonja Van Meirvenne1

Carlo Heirman1

Catherine De Greef1

Pierre van der Bruggen2

Kris Thielemans1*

1Laboratory of Molecular andCellular Therapy, Department ofPhysiology and Immunology, MedicalSchool of the Vrije Universiteit Brussel(V.U.B.), Laarbeeklaan 103/E, 1090Brussels, Belgium2Ludwig Institute for CancerResearch, Brussels Branch, AvenueHippocrate 74, UCL 74.59, 1200Brussels, Belgium

*Correspondence to:Dr Kris Thielemans, Medical Schoolof the Vrije Universiteit Brussel(VUB), Laboratory of Molecular andCellular Therapy, Department ofPhysiology-Immunology,Laarbeeklaan 103/E, 1090 Brussels,Belgium.E-mail: [email protected]

†These authors contributed equallyto this work.

Received: 12 December 2002Revised: 13 February 2003Accepted: 24 February 2003

Abstract

Background Dendritic cells (DC) are the professional antigen-presentingcells of the immune system, fully equipped to prime naive T cells and thusessential components for cancer immunotherapy.

Methods We tested the influence of several elements (cPPT, trip, WPRE,SIN) on the transduction efficiency of human DC. Human and murine DCwere transduced with tNGFR-encoding lentiviruses to assess the effect oftransduction on phenotype and function. Human DC were transduced withlentiviruses encoding huIi80MAGE-A3 and murine DC with huIi80tOVA totest antigen presentation.

Results A self-inactivating (SIN) lentiviral vector containing the trip elementwas most efficient in transducing human DC. The transduction of DC withtrip/SIN tNGFR encoding lentiviral vectors at MOI 15 resulted in stable geneexpression in up to 94.6% (murine) and 88.2% (human) of the mature DC,without perturbing viability, phenotype and function. Human huIi80MAGE-A3-transduced DC were able to stimulate MAGE-A3-specific CD4+ and CD8+T cell clones and could prime both MAGE-A3-specific CD4+ and CD8+ Tcells in vitro. Murine huIi80tOVA-transduced DC were able to present OVApeptides in the context of MHC class I and class II in vitro and induceda strong OVA-specific cytotoxic T lymphocyte response in vivo, that wasprotective against subsequent challenge with OVA-expressing tumor cells.

Conclusions We show that, using lentiviral vectors, efficient gene transferin human and murine DC can be obtained and that these DC can elicit antigen-specific immune responses in vitro and in vivo. The composition of the transfervector has a major impact on the transduction efficiency. Copyright 2003John Wiley & Sons, Ltd.

Keywords dendritic cell; lentivirus; immunotherapy; T helper cell; cytotoxic Tcell

Introduction

Cancer immunotherapy is based on the fact that tumor cells expressantigens which can be recognised by the immune system and lead to tumorrejection. These tumor-associated antigens include tumor-specific sharedantigens, differentiation antigens, protein products of mutated genes andrearrangements unique to tumor cells, overexpressed tissue-specific antigensand exogenous viral proteins [1–3]. However, the development of effectivetherapeutic cancer vaccines has proven difficult, mainly because these tumor

Copyright 2003 John Wiley & Sons, Ltd.

Lentivirally Transduced Dendritic Cells 655

antigens are weak and generally self-derived antigens.Because of their unique capacities to activate naive Tcells [4–7], DC are extensively used as a tool for cancerimmunotherapy (as reviewed by Timmerman et al. andFong et al. [8,9]).

Dendritic cells loaded ex vivo with tumor antigen-derived peptides have been proven to be potent inducersof anti-tumor antigen immune responses in mice [10,11]and are extensively studied in clinical trials [12,13].Although peptide-loaded DC provide a safe method, thereare some drawbacks associated with the peptide-basedstrategy. There is only a transient presentation of theantigenic epitopes and it relies on the knowledge of theMHC haplotype of each patient and the correspondingclass I binding motifs of the antigenic peptide. Moreover,MHC class I peptide motifs are relatively well known,whereas the peptide sequences to load MHC class IImolecules are not as well defined. This is potentiallya major drawback, because an effective cancer vaccineshould not only activate CD8+ T cells, but also CD4+ Thelper cells. A strong T helper 1 (Th1) response is crucialfor inducing and maintaining the CD8+ T cell response[14–16]. Furthermore, CD4+ T cells not only play a rolein the maintenance of a CD8+ memory T cell response, butcan also contribute to the tumor rejection itself [17–20].

An approach that avoids the drawbacks related to theuse of peptides is to transduce the DC with the entiregene encoding the tumor antigen. Gene-modified DCoffer a potential advantage by providing a long-lastingexpression of the entire array of epitopes that fit in thepatient′s HLA haplotype. Another interesting feature ofgenetic modification is the possibility to fuse the gene ofinterest with a class II targeting signal, such as the first80 amino acids of the invariant chain (Ii80) [21]. This isimportant, since many of the tumor antigens are locatedin the cytosol and thus will be presented most efficientlyin MHC class I and only poorly in class II.

Recently, HIV-1 derived lentiviral vectors have emergedas a powerful tool for gene delivery into a variety ofdividing and non-dividing cells, including DC [22–25].These vectors are able to infect non-dividing cells becausethey have evolved a mitosis-independent nuclear importmachinery [26,27]. Since the first reports on lentiviralvectors, various modifications for the improvement ofnuclear import, stabilisation of transgene expression andvector safety have been made. Safety was improved bydeleting all accessory genes [28,29] and by engineeringself-inactivating (SIN) vectors [23,30]. The introductionof the deletion in the U3 region of the 3′ longterminal repeat (LTR) not only minimises the risk ofemergence of replication-competent recombinants, it alsoavoids problems linked to promoter interference. Toenhance nuclear import, the sequence from the polgene of HIV-1 encompassing the central polypurine tract(cPPT) or the cPPT flanked by its central terminationsequence (CTS), together forming a triple-stranded DNAflap (trip), was included in the vector design [27,31].To improve gene expression, the woodchuck hepatitis

virus posttranscriptional regulatory element (WPRE) wasadded [32].

We studied several variants of second-generation HIV-1-derived lentiviral vectors in terms of their ability totransduce human (huDC) and murine DC (muDC) andthe capacity of these lentivirally transduced DC to inducean immune response. We report that huDC and muDCcan be very efficiently and reproducibly transducedby HIV-1-derived lentiviral vectors, without harmfuleffects on viability, phenotype and function. We furthershow that lentivirally transduced huDC expressing thechimeric construct huIi80MAGE-A3 and muDC expressinghuIi80tOVA (a truncated form of ovalbumin) are able toefficiently process and present antigen in the contextof both MHC class I and MHC class II in vitro. Moreimportantly, the huIi80MAGE-A3 huDC were able toprime MAGE-A3-specific CD8+ and CD4+ T cells in vitro.Murine huIi80tOVA-transduced DC induced a potentOVA-specific CTL response in vivo that was shown tobe protective against subsequent challenge with OVA-expressing tumor cells.

Materials and methods

Mice and cell lines

Six to eight week-old female C57BL/6 and BALB/C micewere purchased from Harlan (Ad Horst, The Netherlands).Animals were maintained according to the institutionalguidelines.

293T and PhoenixAmpho cells were cultured in Dul-becco’s modified Eagle’s medium (DMEM; BioWhittaker,Verviers, Belgium) containing 10% fetal calf serum (FCS;Harlan), 100 U/ml penicillin, 100 µg/ml streptomycinand 2 mM L-glutamine (supplements from BioWhittaker).

The T-T hybridoma cell lines RF33.70 and MF2.2D9and the IL-2 sensitive cell line CTLL-2 have been describedpreviously [33].

The tumor cell lines EL4 thymoma (C57BL/6, H-2b)and B16-F0 melanoma (C57BL/6, H-2b) were culturedin DMEM containing 5% FCS and all supplements(complete medium). Their derivatives E.G7-OVA (EL4cells transfected with OVA cDNA) and MO4 (B16 cellstransfected with OVA cDNA) were maintained in completemedium with addition of 400 µg/ml and 2 mg/ml G418,respectively (Eurobiochem, Bierges, Belgium).

The EBV-transformed B cells were cultured in Iscove’smodified Eagle’s medium (IMEM; BioWhittaker) supple-mented with 10% FCS and all supplements.

The human MAGE-A3-specific T cell clones, ESBI684and R12-C9, which recognise MAGE-A3-derived peptides,respectively EVDPIGHLY in the context of HLA-A1 andTQHFVQENYLEY in the context of HLA-DP0401, havebeen described by Schultz et al. [34,35].

3T6 and human CD40 ligand (huCD40L) transfected3T6 fibroblasts were cultured in RPMI 1640 (Invitrogen,Merelbeke, Belgium) containing all supplements.

Copyright 2003 John Wiley & Sons, Ltd. J Gene Med 2003; 5: 654–667.

656 K. Breckpot et al.

Lentivirus production andcharacterisation

PlasmidsThe multiple attenuated packaging plasmid pCMV�R8.9and the vesicular stomatitis virus glycoprotein (VSV.G)encoding plasmid pMD.G were a kind gift from Dr. D.Trono (University of Geneva, Switzerland). The transfervectors pHR′CMVeGFP and pHR′CMVeGFP WPRE were akind gift from Dr. L. Naldini (University of Torino MedicalSchool, Torino, Italy) and encode the enhanced greenfluorescent protein (eGFP). pRRL-cPPT PGKeGFP WPRESIN was a kind gift from Dr. A. Follenzi (University ofTorino Medical School, Torino, Italy).

Elements of the plasmid pBH-10, a plasmid containingthe entire HIV genome and the above-mentioned plas-mids, were used to construct the following eGFP-encodingtransfer vectors: pHR′cPPT CMVeGFP WPRE, pHR′tripCMVeGFP WPRE, pHR′trip CMVeGFP WPRE SIN andpHR′trip CMVeGFP SIN (Figure 1). The latter plasmid wasused to construct a trip/SIN-containing transfer vectorencoding the fusion gene huIi80MAGE-A3 or huIi80tOVAfollowed by an internal ribosomal entry site (Ires)and a truncated nerve growth factor receptor (tNGFR)gene, resulting in the pHR′trip CMVhuIi80MAGE-A3-Ires-tNGFR SIN and pHR′trip CMVhuIi80tOVA-Ires-tNGFR SIN transfer vectors (further referred to as

pHR′huIi80MAGE-A3 and pHR′huIi80tOVA). The frag-ments encoding the huIi80MAGE-A3 and tOVA have beendescribed respectively by Schultz et al. [35] and VanMeirvenne et al. [14,33].

Virus stock preparationsLentiviral vector particles were generated in humanembryonal kidney 293T cells by the transient cotrans-fection method, previously described by Naldini et al.[36], with minor adjustments for production in 175-cm2

bottles. The virus stock was stored at −80 ◦C in X-VIVO15 (BioWhittaker) containing 1% heat inactivated humanAB (huAB) serum (PAA Laboratories, Linz, Austria) forhuDC transduction and in serum-free OptiMEM medium(Invitrogen) for muDC transduction.

Viral titers were determined by infection of 293T cellswith serial dilutions of the vector stock. Seventy-twohours after infection the number of eGFP- or tNGFR-positive cells was scored by fluorescence-activated cellsorting (FACS) analysis to determine the titer.

Generation and transduction ofdendritic cells

For the generation of huDC, peripheral blood mono-nuclear cells (PBMC) were isolated from buffy coat

pHR’trip CMVhuIi80tOVA-Ires-tNGFR SIN

pHR’CMVeGFP

5’LTR 3’LTR

ψ

∆gag RRE CMV eGFP

pHR’CMVeGFP WPRE

pHR’CPPT CMVeGFPWPRE

pHR’trip CMVeGFPWPRE

pHR’trip CMVeGFPWPRE SIN

pHR’trip CMVeGFPSIN

pHR’trip CMVhuIi80MAGE-A3-Ires-tNGFR SIN

5’LTR 3’LTR

ψ

ψ

ψ

ψ

ψ

ψ

ψ

∆gag RRE CMV eGFP

WPRE5’LTR 3’LTR∆gag RRE CMV eGFP

WPRECPPT5’LTR 3’LTR∆gag RRE CMV eGFP

WPREtrip

5’LTR ∆3’LTR∆gag RRE CMV eGFP

WPREtrip

∆3’LTR5’LTR ∆gag RRE CMV eGFP

trip

5’LTR ∆gag RRE CMV huIi80 MAGE-A3 Ires-tNGFR ∆3’LTR

trip

5’LTR ∆gag RRE CMV huIi80 tOVA Ires-tNGFR ∆3’LTR

trip

Figure 1. Schematic representation of the modified lentiviral transfer vectors. Abbreviations: LTR: long terminal repeat; �:packaging signal; �gag: frame-shifted gag gene; RRE: rev-responsive element; CMV: cytomegalovirus promotor; eGFP: enhancedgreen fluorescent protein; cPPT: central polypurine tract; WPRE: woodchuck hepatitis posttranscriptional regulatory element;trip: central polypurine tract + termination sequence; huIi80: first 80 amino acids of the human invariant chain; MAGE-A3;melanoma-associated gene A3; tOVA: truncated form of the model antigen ovalbumin; Ires: internal ribosomal entry site; tNGFR:truncated form of the nerve growth factor receptor



preparations of normal donors or hemochromato-sis patients by gradient centrifugation (Lymphoprep;Nycomed Pharma AS, Brussels, Belgium). Subsequently,CD14+ cells were separated by magnetic sorting usingthe VarioMACS technique (Miltenyi Biotech GmbH, Ber-gisch Gladbach, Germany) following the manufacturer’sinstructions. Monocytes were differentiated to huDC inX-VIVO 15 supplemented with 1% heat-inactivated huABserum at a cell density of 106 huDC/ml. Recombinantgranulocyte-macrophage colony stimulating factor (GM-CSF; Novartis, Brussels, Belgium) and home-made IL-4(described by Tuyaerts et al. [37]) were added, respec-tively, at 1000 and 100 U/ml. Cells were incubated in ahumidified atmosphere containing 5% CO2 at 37 ◦C. Onday 3, huDC were infected with recombinant lentivirusesas follows: 105 huDC were resuspended in 100 µl X-VIVO 15 containing 1% huAB serum, the lentiviruses ata multiplicity of infection (MOI) of 15, protamine sul-phate (10 µg/ml; LeoPharma, Thornhill, ON, Canada),IL-4 (100 U/ml) and GM-CSF (1000 U/ml). The imma-ture transduced huDC were cultured at 106 cells/ml untilday 6. Subsequently, huDC were matured at a cell den-sity of 5 × 105 huDC/ml X-VIVO 15 containing 1% huABserum and the following cytokines: IL-1β (100 U/ml), IL-6(1000 U/ml), tumor necrosis factor α (TNFα, 100 U/ml)and prostaglandin E2 (PGE2, 1 µg/ml). The cytokinesIL-1β, IL-6, and TNF α were purchased from PeprotechInc. (Rocky Hill, NJ, USA). PGE2 was obtained fromSigma-Aldrich (Bornem, Belgium).

Murine DC were generated according to the protocoldescribed by Lutz et al. [38] with minor adjustments.Briefly, bone marrow cells were collected from the tibiaand femurs of 6–8-week-old C57BL/6 mice and the redblood cells were lysed. The cells were seeded in 6-wellplates at 106 cells/well in 2 ml DMEM containing 5%FCS, all supplements, 50 µM β-ME and 20 ng/ml muGM-CSF (home-made). On day 3, 2 ml of fresh mediumcontaining 20 ng/ml muGM-CSF were added. On day6, 2 ml medium/well were collected and the cells werepelleted and resuspended in fresh medium with 20 ng/mlmuGM-CSF. On day 7 of the culture, the muDC wereharvested, washed and resuspended at 106 cells/100 µlin OptiMEM containing recombinant lentiviral particlesat an MOI of 15 and 10 µg/ml protamine sulphate. Thecells were plated into a non-tissue culture treated 24-wellplate and incubated at 37 ◦C. After 2 h, DMEM containing5% FCS, all supplements, 50 µM β-ME and 20 ng/mlmuGM-CSF was added to reach a final concentration of106 cells/ml. In order to mature the cells, 100 ng/ml LPS(E. coli serotype O55:B5; Sigma-Aldrich) was added.

Flow cytometry

Flow cytometry was performed to analyse the phenotypeand transgene expression of the lentivirally transducedDC. All stainings were performed for 30 min on icein phosphate-buffered saline containing 1% BSA and0.02% sodium azide. The first step of each staining

was performed in the presence of 10% goat serum(Sigma-Aldrich) to reduce non-specific antibody (Ab)binding. FITC or PE conjugated monoclonal Abs againsthuCD80, huCD83, huCD86, huCD14 and muI-Ab (cloneAFG-120.1) were purchased from BD Pharmingen(Erembodegem, Belgium). The following Abs wereaffinity-purified and biotinylated in our laboratory: anti-hutNGFR (clone HB8737), anti-muCD11c (clone N418),anti-muB7.1 (clone 16-10A1), anti-muB7.2 (clone GL-1),anti-muCD40 (clone FGK45), and anti-HLA-DR (cloneL243). Biotinylated Abs were detected with streptavidin-PE (BD Pharmingen).

The stained cells were analysed on a FACSCaliburflow cytometer (Beckton Dickinson, San Jose, CA, USA)using CellQuest software. All stainings were comparedwith irrelevant isotype control Abs purchased from BDPharmingen.

Mixed leukocyte reaction

In order to obtain human allogeneic T cells, PBMC wereplated at 5 × 106 cells/ml in a final volume of 44 ml in a175-cm2 culture flask (Falcon, Beckton Dickinson). After2 h of plastic adherence the non-adherent fraction (T cellenriched) was collected. Murine allogeneic T cells werederived from the spleens of female BALB/c mice using anylon wool column.

Mature huDC and muDC, non- or lentivirally trans-duced, were cocultured in graded numbers with 2 × 105

allogeneic T cells in round-bottomed 96-well platesin a final volume of 200 µl. After 4 days of incuba-tion, 1 µCi 3H-thymidine (Amersham Pharmacia Biotech,Roosendaal, The Netherlands) was added and 3H-thymidine uptake was measured 18 h later using liquidscintillation counting (Microbeta; Wallac, Turku, Fin-land).

IL-12 dosage

Human DC were cocultured at a ratio 1 : 1 with 105 3T6or 3T6-huCD40L cells in 1 ml RPMI 1640 in a 24 well.Murine, LPS matured DC were plated at 1 × 106 cells/mlmedium. The supernatant of the matured DC wascollected 24 h after transduction. Fifty microliters of thesupernatant were used to measure the IL-12 content in asandwich ELISA following the manufacturer’s instructions(Endogen, Woburg, MA, USA, for detection of murineIL-12 p40 + p70, and Medsystems Diagnostics GmbH,Vienna, Austria, for detection of human IL-12 p70).

Semiquantitive RT-PCR

Total RNA was extracted from the huDC using the SVtotal RNA isolation system (Promega, Madison, USA).Total RNA (1 µg) was converted into first-strand cDNA



using random hexamers and SuperScript II reverse tran-scriptase according to the manufacturer’s recommenda-tions (Superscript first-strand synthesis system for RT-PCR, Invitrogen). cDNA was amplified using BIOTAQDNA polymerase (Bioline, London, UK) following themanufacturer’s instructions in a GeneAmp PCR system(Perkin Elmer, Wellesley, MA, USA). In order to deter-mine the level of OVA expression in tumor samples,RNA was extracted using Trizol (Invitrogen) accordingto the manufacturer’s instructions. RT-PCR was per-formed using the Access RT-PCR system (Promega). TheMAGE-A3 cDNA was amplified using the MAGE-A3-sense:5′CCCAGATCTGGGGAGTGTCGTCGGA 3′ and the MAGE-A3-antisense: 5′ CCCCTCGAGTCACTCTTCCCCCTCTCTC3′ primers. The ovalbumin sequence was amplified withOVA235-5′-sense: 5′GGGGGATCCATTTGCCAGTGGGAC-AAT 3′ and OVA3′-antisense: 5′GGGGATCCGGGGAAACA-CATCTGCCAA 3′ primers. To assess the RNA and cDNAquality and quantity, RT-PCR for the house-keeping geneβ-actin was performed.

In vitro antigen presentation assay

HLA-A1+ and HLA-DP0401+ huDC, transduced withpHR′huIi80MAGE-A3 at a MOI of 15, were coculturedwith, respectively, the MAGE-A3-specific CD4+ T cellclone R12-C9 or the CD8+ T cell clone ESBI684. Peptide-loaded huDC were used as a positive control (MAGE3.A1and MAGE3.DP0401). Dendritic cells were loaded ata density of 2 × 106 DC/ml in DMEM with 10 µg/mlpeptide for 2 h at 37 ◦C. Prior to use, the DC were washedtwice in DMEM. Cocultures were performed in triplicatewith 2 × 104 DC at a DC/T ratio of 4 : 1 in a round-bottomed 96-well plate in IMEM supplemented with 10%huAB serum containing 25 U/ml huIL-2. After 24 h ofcoculture the supernatant was tested for the presence ofinterferon γ (IFNγ ) by means of ELISA (Endogen).

The antigen-presenting capacity of muDC was tested inan in vitro antigen presentation assay using OVA-specific Tcell hybridoma cells, as described by Van Meirvenne et al.[33]. The DC were transduced on day 7 and subsequentlymatured. The test was performed on day 9.

Retroviral transduction ofEBV-transformed B cells

The retroviral vector pMFG encoding huIi80MAGE-A3-Ires-tNGFR used to transduce EBV-transformed B cellsand the procedure to transduce the EBV transformed Bcells have been described previously by Chaux et al. [39]and Schultz et al. [34].

In vitro priming of MAGE-A3-specificCD4+ and CD8+ T cells

Autologous, lentivirally transduced, mature huDCexpressing the fusion protein huIi80MAGE-A3 were

washed and added at a ratio of 1 : 10 for CD4+ and1 : 3 for CD8+ VarioMACS sorted T cells to 1 × 105 T cellsin 200 µl X-VIVO 15 containing 1% huAB serum in around-bottomed 96-well plate. The first stimulation wasperformed in the presence of IL-6 (1000 U/ml) and IL-12(10 ng/ml). Both CD4+ and CD8+ T cells were restimu-lated with the same number of transduced huDC on days7, 14 and 21 in the presence of IL-2 (10 U/ml) and IL-7(5 ng/ml).

The microcultures were screened 10 days after thefourth stimulation for their capacity to produce IFNγ

when stimulated with autologous non- or huIi80MAGE-A3-transduced EBV B cells. Therefore, 2.5 × 103 T cellswere cocultured with autologous 5 × 103 EBV B cells in200 µl medium containing IL-2 (25 U/ml). After 24 h,the supernatant was collected and the IFNγ contentwas determined by ELISA (Endogen). Each well ofthe T cell culture was tested in duplicate. MAGE-A3specificity of the microcultures that were MAGE-A3responsive in the first screening was confirmed in ELISAfor the CD4+ T cell microcultures and in a standard51chromium (51Cr) release assay [35] for the CD8+ T cellmicrocultures.

In vivo CTL induction

On day 7, muDC were transduced with either emptylentivirus particles (pHR′empty) or pHR′huIi80tOVA atan MOI of 15 and matured in the presence of 100 ng/mlLPS for 4 h. Six to eight week old C57BL/6 mice receivedone subcutaneous injection of 1 × 105 transduced muDCin the interscapular area (4 animals/group). Sevendays after immunisation, the spleen cells were isolatedand restimulated in vitro for 5 days with mitomycinC treated E.G7-OVA cells at an effector/stimulator(E/S) ratio of 10 : 1. After restimulation, the CTLwere tested for their cytolytic activity against EL4and E.G7-OVA cells in a standard 51Cr release assay[14].

Tumor protection

On day 7, muDC were transduced with either pHR′emptyor with pHR′huIi80tOVA at an MOI of 15 and maturedin the presence of 100 ng/ml LPS for 4 h. Six to eightweek old C57BL/6 mice received one subcutaneousinjection of 1 × 105 muDC in the interscapular area (10animals/group). Seven days after immunisation, 2 × 105

MO4 tumor cells were grafted subcutaneously in theinterscapular area. Tumor growth and survival wereevaluated. Mice were sacrificed when the tumor reached adiameter of more than 15 mm. The study was terminated75 days after tumor challenge and all surviving animalswere sacrificed. The data were recorded in a Kaplan-Meier plot and the log-rank test was used to determine pvalues.



Results

Enhanced transduction efficiencies andincreased fluorescence intensities areobserved with a trip/SIN lentiviralvector

In a series of preliminary experiments, we determined theoptimal timing and MOI for the transduction of huDC.Lentiviral transduction at a MOI of 15, early duringthe GM-CSF and IL-4 driven differentiation of bloodmonocytes, gave the best results.

The influence of several elements on the transductionefficiency and on the level of transgene expression, asmeasured by the mean fluorescence intensity (MFI) inflow cytometry, was assessed. The different elements(Figure 1) include a safety improving element (SIN),elements to improve nuclear import of the viral genome(cPPT, trip) and an element to improve gene expression(WPRE). Immature day 3 huDC were transduced at MOI15 and matured on day 6. Analysis was performed 24 hpost-maturation. No cytotoxicity, as assessed via trypanblue exclusion, was observed using the different typesof lentiviruses in three independent experiments. Asshown in Figure 2, we obtained the highest transductionefficiency with the trip/SIN-containing lentiviral vector.Transduction efficiencies up to 88% with MFI up to 2600were obtained. Compared with the pHR′CMVeGFP vector,the transduction efficiency with pHR′trip CMVeGFP SIN isincreased 2 times and MFI 12 times. Based on these resultswe performed all further experiments with trip/SINlentiviral vectors.

Lentiviral transduction of murine andhuman dendritic cells has no harmfuleffects on the mature phenotype andfunction of the dendritic cells

To evaluate transduction of muDC and to confirm theefficient lentiviral transduction of huDC, we used VSV.Gpseudotyped, tNGFR encoding trip/SIN lentiviruses.Murine DC and human DC were transduced withpHR′tNGFR at a MOI of 15 and subsequently matured.Flow cytometry was performed 48 h post-transduction todetect tNGFR expression. An average of 92% (range ±3%, n = 8) of the muDC and an average of 80%(range ± 9%, n = 3) of the huDC stained positive fortNGFR (Figure 3). Expression of tNGFR could still bedetected 1 week after transduction (data not shown).Trypan blue exclusion and propidium iodide stainingindicated that there was no deleterious effect on cellviability compared with mock-transduced DC (data notshown).

Flow cytometry was performed to compare theexpression of several immunologically important DCsurface markers in the mock- and pHR′tNGFR-transducedmature huDC and muDC. As shown in Figure 4,transduction with pHR′tNGFR of huDC and muDC

had no effect on the expression of MHC class II,important adhesion molecules, costimulatory moleculesand maturation markers.

Furthermore, we evaluated the ability of transducedDC to stimulate allogeneic T cells in allo-MLR. Weobserved that human and murine lentivirally transducedDC display an allostimulatory capacity comparable withthat of non-transduced DC (Figure 5). Finally, weassessed the ability of lentivirally transduced DC toproduce IL-12, a major determinant of DC functionand mediator of Th1 development. We could notdetect major differences in IL-12 secretion betweennon- or lentivirally transduced huDC (IL-12 p70) andmuDC (IL-12 p40 + p70) (data not shown). The averagesecretion levels obtained were 340 pg/ml for huDC(105 cells/1 ml/24 hours) and 120 pg/ml for muDC(1 × 106 cells/1 ml/24 hours).

Considering the normal upregulation of phenotypicmarkers following maturation, the potent allostimulatorycapacity and the ability of lentivirally transduced DC toproduce IL-12, we conclude that lentiviral transductiondoes not alter DC maturation or function.

In vitro antigen presentation bylentivirally transduced murine andhuman dendritic cells

In order to study antigen presentation, pHR′tNGFR- orpHR′huIi80tOVA-transduced muDC were cocultured withT-T hybridomas that recognise OVA-specific peptides inMHC class I (RF33.70) and MHC class II (MF2.2D9). Asa reference we used muDC loaded with the OVA-derivedpeptides SIINFEKL or TEWTSSNVMEERKIKV, presentedin respectively H2-Kb and I-Ab. The pHR′huIi80tOVA-transduced muDC showed a potent antigen presentationin MHC class I and MHC class II that increased with theDC/T cell ratio. pHR′tNGFR-transduced muDC inducedno activation of the T-T hybridomas, demonstrating thespecificity of the response (Figure 6B).

We also assessed the capacity of pHR′huIi80MAGE-A3-transduced HLA-A1+ and HLA-DP0401+ huDC to presentMAGE-A3-derived epitopes to the CD8+ T cell cloneESBI684 and the CD4+ T cell clone R12-C9, respectively.As a reference for IFNγ production we used huDCloaded with the MAGE-A3-derived peptides EVDPIGHLYor TQHFVQENYLEY, presented respectively in the contextof HLA-A1 and HLA-DP0401. Mock-transduced huDCinduced no activation of the MAGE-A3-specific Tcell clones. Figure 6A shows that pHR′huIi80MAGE-A3-transduced huDC were able to present MAGE-A3-derivedantigenic epitopes in the context of HLA class I and HLAclass II. We further demonstrated that the mRNA contentof the lentivirally transduced huDC is comparable to thatof an endogenously MAGE-A3-expressing myeloma cellline, U266 (data not shown).

Taken together these results indicate that both muDCand huDC are able to efficiently present antigenicepitopes in the context of both MHC class I and MHCclass II.



R1

R1

R1

R1

R1

R1

M1

FL1-Height

M1

M1

M1

M1

M1

28.5%MFI=112

CMVeGFP WPRE

cppt CMVeGFP WPRE

trip CMVeGFP WPRE

44.1%MFI=399

trip CMVeGFP SIN

78.6%MFI =1500

CMVeGFP

trip CMVeGFP WPRE SIN

47.5%MFI=160

13.9%MFI=130

15.1%MFI=111

0100 101 102 103 104

20

40

60

80

100

00 200 400 600 800

FSC-Height1000

200

400

600

800

SS

C-H

eigh

t

1000

00 200 400 600 800

FSC-Height1000

200

400

600

800

SS

C-H

eigh

t

1000

00 200 400 600 800

FSC-Height1000

200

400

600

800

SS

C-H

eigh

t

1000

00 200 400 600 800

FSC-Height1000

200

400

600

800

SS

C-H

eigh

t

1000

00 200 400 600 800

FSC-Height1000

200

400

600

800

SS

C-H

eigh

t

1000

00 200 400 600 800

FSC-Height1000

200

400

600

800

SS

C-H

eigh

t

1000

Cou

nts

FL1-Height

0100 101 102 103 104

20

40

60

80

100

Cou

nts

FL1-Height

0100 101 102 103 104

20

40

60

80

100

Cou

nts

FL1-Height

0

100 101 102 103 104

20

40

60

80

100

Cou

nts

FL1-Height

0

100 101 102 103 104

20

40

60

80

100

Cou

nts

FL1-Height

0

100 101 102 103 104

20

40

60

80

100

Cou

nts

Figure 2. Comparison of the transduction efficiencies and mean fluorescence intensities (MFI) of huDC transduced with the sixmodified lentiviral vectors encoding eGFP. Immature huDC were transduced at a MOI of 15 and matured on day 6. Flow cytometricanalysis was performed 24 h post-maturation and 4 days after the lentiviral transduction. The percentage of positive cells in thegate (M1) and their mean fluorescence intensity are shown. The data are representative for three independent experiments



100 101 102 103 104

0

20

40

60

80

100

Cou

nts

SS

C-H

eigh

t

0 200 400 600 800 10000

200

400

600

800

1000

FSC-Height

100 101 102 103 104 100 101 102 103 104 100 101 102 103104

0

20

40

60

80

100

Cou

nts

Cou

nts

SS

C-H

eigh

t

0 200 400 600 800 10000

200

400

600

800

1000

FSC-Height

0 1000

FSC-Height

0 1000

FSC-Height

0

80

SS

C-H

eigh

t

0

1000

Cou

nts

0

80

SS

C-H

eigh

t

0

1000

A

Non-transduced pHR’tNGFR

Human DC B

Non-transduced pHR’tNGFR

Murine DC

tNGFR

DC DC DC DC

0.5% 88.2% 0.8% 93.1%

Figure 3. Transduction efficiency of lentivirally transduced huDC and muDC. Both huDC and muDC were transduced with trip/SINtNGFR encoding lentiviral particles at a MOI of 15 and subsequently matured. The expression of tNGFR was analysed via flowcytometry. The forward/side scatter profile and the tNGFR expression of the huDC are shown in (A) and of the muDC in (B). Thedata shown are representative for three (huDC) and eight (muDC) experiments

In vitro priming of MAGE-A3-specificCD4+ and CD8+ T cells

Human DC of donor 1 (HLA-A2/A3, B7, Cw7) and ofdonor 2 (HLA-DRB103/07, DQB102, DPB10401/1011)were transduced with pHR′huIi80MAGE-A3, subsequentlymatured and used to stimulate, respectively, autologousCD8+ T cells (DC/T ratio 1 : 3) and CD4+ T cells(DC/T ratio 1 : 10). Two 96-well plates, each wellcontaining 105 T cells, were stimulated four times atweekly intervals. The first stimulation was performed inthe presence of the cytokines IL-6 and IL-12, whereasthe following three stimulations were performed inthe presence of IL-2 and IL-7. After a resting periodof 10 days, the CD4+ and CD8+ T cells were testedfor their MAGE-A3 specificity in a coculture withautologous, retrovirally transduced (pMFGhuIi80MAGE-A3-Ires-tNGFR) or non-transduced EBV B cells. The IFNγ

secretion after stimulation with non-transduced EBVB cells was plotted against the IFNγ secretion afterstimulation with MAGE-A3-transduced EBV B cells toidentify the MAGE-A3-specific microcultures. We obtainedtwo CD4+ and four CD8+ MAGE-A3-specific T cellmicrocultures (Figure 7A). These T cells were stimulatedan additional two times with autologous huIi80MAGE-A3-transduced EBV B cells. After the last stimulationwith EBV B cells, the microcultures were tested again toconfirm their MAGE-A3 specificity. Only one out of the twoCD4+ T cell microcultures tested could be maintained.These CD4+ T cells showed specific IFNγ secretionupon restimulation with autologous huIi80MAGE-A3-transduced huDC.

The MAGE-A3 specificity of the 4 CD8+ T cellmicrocultures was confirmed in a 51Cr release assaywith autologous huDC as targets at an E/T ratio of 1 : 1.We obtained approximately 30% lysis of the MAGE-A3-expressing huDC whereas only 4–14% lysis was obtainedwith the non-transduced huDC as targets (Figure 7B).

We show here that lentivirally transduced huDCcan prime both CD4+ and CD8+ MAGE-A3-specificT cells in vitro. Currently, these T cell clones havebeen subcloned and experiments to identify the pre-sented peptide have been performed (manuscript inpreparation).

In vivo CTL induction and tumorprotection

We wanted to investigate whether lentivirally transducedmuDC are also capable of inducing an anti-OVA immuneresponse in vivo. Therefore, mice were immunised with asingle injection of muDC transduced with pHR′emptyor pHR′huIi80tOVA at an MOI of 15. Seven dayspost-immunisation, the spleen cells were collected,restimulated with E.G7-OVA cells and tested for OVA-specific lysis in a standard 51Cr release assay. Figure 6shows that pHR′empty-transduced muDC induced no CTLagainst EL4 or E.G7-OVA. Murine DC transduced withpHR′huIi80tOVA, however, induced OVA-specific CTL in75% of the immunised mice, while no lysis was inducedagainst EL4 cells (Figure 8).

Since we were able to induce OVA-specific CD8+T cells in vivo with a single immunisation, wewanted to investigate if this immune response was



M2

88.8%MFI=111

0

20

10

30

40

50

60

70

80

Cou

nts

101 102 103 104100

M2

82.3%MFI=60

0

101 102 103 104

20

40

60

100

80

Cou

nts

100

M2

86.8%MFI=976

0100 101 102 103 104

20

40

60

80

100

Cou

nts

M2

9.8%MFI = 45

0100 101 102 103 104

20

40

60

80

100

Cou

nts

M2

97.7%MFI=936

0100 101 102 103 104

20

40

60

80

100

Cou

nts

Lenti-tNGFR

Class II Class II

CD11c

CD80

CD86

CD40

CD83

CD80

CD86

CD14

M1

90.4%MFI=1548

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

B Mock

M1

80.3%MFI=97

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M1

78.8%MFI=148

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M1

85.0%MFI=2700

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M1

69.5%MFI=352

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M2

97.4%MFI=964

0100 101 102 103 104

20

40

60

80

100

Cou

nts

A Mock

M2

94.5%MFI=163

0

20

10

30

40

50

60

70

80

Cou

nts

101 102 103 104100

M2

81.7%MFI=76

0101 102 103 104

20

10

30

40

50

60

70

80

Cou

nts

100

M2

94.7%MFI=623

0100 101 102 103 104

20

40

60

80

100

Cou

nts

M2

10.4%MFI = 45

0100 101 102 103 104

20

40

60

80

100

Cou

nts

M1

90.0%MFI=1533

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

Lenti-tNGFR

M1

84.8%MFI=240

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M1

85.5%MFI=268

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M1

85.8%MFI=2680

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

M1

68.8%MFI=424

0

10

20

30

40

Cou

nts

50

100 101 102 103 104

Figure 4. Phenotypic analysis of mature, mock or lentivirally transduced huDC and muDC. Flow cytometry was performed toanalyse the DC phenotype. (A) Phenotype of the huDC and (B) phenotype of the muDC. The percentage of positive cells in the gate(M1, M2) and their mean fluorescence intensity are shown. The data are representative for three independent experiments

protective against subsequent challenge with OVA-expressing tumor cells. We immunised mice withpHR′empty- or pHR′huIi80tOVA-transduced muDC. Themice were challenged with 2 × 105 MO4 cells 7 dayspost-immunisation. Animals immunised with pHR′empty

DC started to develop tumors from day 7 post tumorchallenge. These tumors developed quickly and reachedan exclusion diameter of 15 mm in approximately10 days. A number of mice that were immunisedwith the huIi80tOVA muDC also developed tumors,



A B

0

20000

40000

60000

80000

100000

120000

140000

120 240 480

ratio T/muDC

Non TD muDCMock muDCpHR'tNGFR muDC

T c

ell p

rolif

erat

ion

(cpm

)

0

5000

10000

15000

20000

25000

30000

100 300 900

ratio T/huDC

Non TD huDCMock huDCpHR'tNGFR huDC

T c

ell p

rolif

erat

ion

(cpm

)

Figure 5. Allostimulatory capacity of lentivirally transduced huDC and muDC. Non-, mock and lentivirally transduced DC werecocultured for 4 days with allogeneic T cells. Proliferation was measured using 3H-thymidine incorporation (represented incounts/min). (A) Results with huDC and (B) with muDC, representative for three independent experiments

A

B

Peptide

pHR'tNGFRpHR'huIi80tOVA

MHC Class II

0

20000

40000

60000

80000

100000

120000

140000

5 10 20 40

T/DC ratio

CT

LL-2

pro

lifer

atio

n (c

pm)

PeptidepHR'tNGFRpHR'huIi80tOVA

MHC Class I

020000400006000080000

100000120000140000160000180000

5 10 20 40

T/DC ratio

CT

LL-2

pro

lifer

atio

n (c

pm)

HLA-A1

0

500

1000

1500

2000

2500

Mock huDCpeptide huDCpHR'huIi80MAGE-A3 huDC

#pg/

ml I

FN

γ

HLA-DP0401

0

500

1000

1500

2000

2500

Mock huDCpeptide huDCpHR'huIi80MAGE-A3 hu DC

#pg/

ml I

FN

γ

Figure 6. In vitro antigen presentation by lentivirally transduced muDC and huDC. (A) pHR′huIi80MAGE-A3-transduced huDC werecocultured with the MAGE-A3-specific CD4+ (R12-C9) and CD8+ (ESBI684) T cell clones for 24 h. The IFNγ content was measuredin ELISA. (B) pHR′tNGFR- and pHR′huIitOVA-transduced muDC were cocultured with T-T hybridomas that recognise OVA-specificpeptides in MHC class I (RF33.70) and class II (MF2.2D9). The IL-2 secretion of these hybridomas was measured as proliferationof the IL-2-dependent CTLL-2 cell line (represented in counts/min). The data shown are representative for three independentexperiments

but with a delay starting from day 25 on. Weobserved long-term survival rates of 20–70% withthe pHR′huIi80tOVA muDC immunised mice in threeindependent experiments (Figure 9A). The differencein survival we observed between pHR′empty andpHR′huIi80tOVA muDC immunised mice was significant

with p ≤ 0.0001 (Kaplan-Meier cumulative survival, log-rank test) in all three experiments.

We investigated if the OVA gene was still expressedin these later developing tumors in the pHR′huIi80tOVAmuDC immunised mice and performed a RT-PCR on RNAextracted from several tumor samples. We observed a



B

A CD8 T cell stimulation

0

50

100

150

200

250

300

0 50 100 150 200 250 300

2E72E8

1D101E9

EBV B stimulation: # pg/ml IFNγ

EB

V B

huI

i80M

AG

E-A

3st

imul

atio

n: #

pg/

ml I

FN

γCD4 T cell stimulation

0500

1000150020002500300035004000

0 1000 2000 3000 4000

1D9

1C10

EBV B stimulation: # pg/ml IFNγ

EB

V B

huI

i80M

AG

E-A

3st

imul

atio

n: #

pg/

ml I

FN

γ

CD4 T cells

0200400600800

100012001400

1C10

pHR'empty pHR'huIi80MAGE-A3

#pg/

ml I

FN

γ

CD8 T cells

05

101520253035

1D10 2E7 2E8 1E9

pHR'empty pHR'huIi80MAGE-A3

% s

peci

fic ly

sis

Figure 7. In vitro priming of MAGE-A3-specific CD4+ and CD8+ T cells. Lentivirally transduced huDC (pHR′huIi80MAGE-A3) wereused to prime autologous CD4+ and CD8+ T cells. After four stimulations at weekly intervals, the CD4+ and CD8+ T cells werescreened for MAGE-A3 specificity. The CD4+ and CD8+ T cells were therefore cocultured in duplicate with autologous EBV B cellsor huIi80MAGE-A3-transduced autologous EBV B cells. (A) IFNγ production after stimulation with huIi80MAGE-A3-transducedEBV B cells is plotted against the IFNγ production after stimulation with non-transduced EBV B. The marked dots representT cell microcultures that are considered MAGE-A3 reactive. The MAGE-A3-responsive CD4+ and CD8+ T cells were stimulatedfor an additional 2 weeks with huIi80MAGE-A3-expressing EBV B cells and subsequently retested. (B) IFNγ production of theCD4+ T cells after restimulation with autologous pHR′huIi80MAGE-A3-transduced huDC and the percentage lysis of autologouspHR′huIi80MAGE-A3-transduced huDC by the CD8+ T cells after restimulation

0102030405060708090

100

100 50 25 12.5 6 3

A B

% S

peci

fic ly

sis

0102030405060708090

100

% S

peci

fic ly

sis

EL-4 lysis EG7.OVA lysis

E/T ratio

100 50 25 12.5 6 3

E/T ratio

Figure 8. In vivo CTL induction after immunisation with lentivirally transduced muDC. Mice were subcutaneously immunised with105 pHR′empty- or pHR′huIitOVA-transduced DC. After 7 days, the spleen cells were isolated and restimulated for 5 days withE.G7-OVA cells. Chromium release assay was performed to evaluate the cytotoxic activity against EL4 and E.G7-OVA cells. (A) Lysisinduced by CTL generated from mice immunised with pHR′empty-transduced DC and (B) lysis of EL4 and E.G7-OVA cells by CTLgenerated from mice immunised with pHR′huIi80tOVA-transduced DC. One out of three experiments is shown

clear OVA cDNA expression in a tumor sample of a mouseimmunised with pHR′empty-transduced DC. However, wecould not detect OVA cDNA expression in any of the testedtumor samples of the OVA-immunised mice (Figure 9B).This suggests that these mice developed tumors afterthe OVA-expressing tumor cells were destroyed by the

immune system, thus resulting in selection of antigen lossvariants.

The results from these experiments show that immu-nisation with pHR′huIi80tOVA-transduced DC induces animmune response that is protective against subsequentchallenge with OVA-bearing tumor cells.



Cum

ulat

ive

Sur

viva

l (%

)

pHR’huIi80tOVA muDC

p ≤ 0,0001

80706050403020100

100

80

60

40

20

0 pHR’empty muDC

Days post tumor challenge

A

pHR’OVA 2

pHR’OVA 1

pHR’Empty

Ovalbuminβ-actinB

5 ng 1 ng100 pg10 pg 1 pg 5 ng 1 ng100 pg10 pg 1 pg

Figure 9. Tumor protection induced by immunisation withlentivirally transduced muDC. Mice were subcutaneouslyimmunised with 105 pHR′empty- or pHR′huIitOVA-transducedDC. One week later the mice received 2 × 105 MO4 cellssubcutaneously in the interscapular region. Tumor growth wasevaluated every 3–4 days until 75 days after tumor cell injection.The data are recorded in a Kaplan-Meier survival plot (A) and thelog-rank test was used to determine p values. One out of threeexperiments is shown. (B) OVA mRNA content of tumor samplestaken from two pHR′huIitOVA DC immunised mice that diddevelop tumors, in comparison with a tumor sample taken froma mouse immunised with pHR′empty DC. RNA and cDNA qualityand quantity were assessed in RT-PCR for the house-keepinggene β-actin

Discussion

Dendritic cells are potent antigen-presenting cells respon-sible for initiating primary T cell responses and mayprovide an effective tool for immunotherapeutic interven-tions. In this regard, the delivery of genes encoding tumorantigens is widely studied.

Non-viral methods have been used for gene transfer intonon-dividing DC; however, their use is limited, due to poortransfection efficiencies and transient gene expression[40,41].

Adenoviral methods have been applied successfullyto transduce DC but their utility for vaccination maybe limited due to expression of immunodominant viralepitopes [42,43]. Lentiviruses, on the other hand, donot express viral components and have proven to behighly efficient in infecting various non-dividing celltypes [26,44,45]. Since the first reports on lentiviralvectors, various modifications for the improvement ofnuclear import, stabilisation of transgene expression andvector safety have been made. Safety was improved bydeleting all accessory genes [46] and by engineering self-inactivating vectors [23]. Attempts to improve nuclear

import were undertaken by including cPPT or trip in thelentiviral vector design [27,31]. The woodchuck hepatitisvirus posttranscriptional regulatory element was insertedin the transfer vector to improve gene expression [32].We tested the effect of these elements on the transductionefficiency of huDC. We observed that the combinationof trip (cPPT-CTS) with the �U3-3′LTR, resulting inself-inactivation, gave the best transduction efficiencies,obtaining up to 88% transduced huDC, with high eGFPexpression levels (MFI = 2600). Since the cPPT incombination with the CTS (trip) had a more beneficialeffect on transduction efficiency than the cPPT alone weconclude that the CTS plays an important role in the activenuclear import of the viral pre-integration complex. Theseresults are consistent with those of Firat et al. [47] onhuman DC and Manganini et al. on human monocytes andT lymphocytes [48]. It was quite unexpected that insertionof WPRE in the vector design had only a beneficial effecton MFI whereas transduction efficiency was unaltered oreven less, dependent on the other vector elements. This isin contrast with the first report on the effect of WPRE ontransgene expression of Zufferey et al. [32]. The higherMFI obtained with the �U3-3′LTR vector is probably dueto the decrease in LTR-mediated transcriptional silencingor promoter interference.

We then used trip/SIN lentiviral vectors encodingthe tNGFR to transduce huDC and muDC in order tostudy the influence of the lentiviral transduction oncell viability, phenotype and function. We used tNGFRas a transgene rather then eGFP, to avoid possiblemis-interpretation of the transduction efficiency due topseudotransduction, which can be observed using eGFP-encoding vectors [49]. Both huDC and muDC were veryefficiently transduced. We used a MOI of 15 and obtainedtransduction efficiencies up to 89 and 95%, respectively,for huDC and muDC, which is consistent with previousreports [25,47,50,51]. The lentiviral transduction hadno adverse effect on cell viability or on the phenotypeof the mature DC. The allostimulatory capacity and thecapacity to produce IL-12, a key factor in the initiation ofcell-mediated immunity, was not hampered by lentiviraltransduction of the DC.

Lentiviruses encoding the truncated variant of themodel antigen OVA fused to the first 80 amino acidsof the invariant chain were used to transduce muDC.We showed that these DC are almost as potent as OVApeptide-loaded DC in presenting OVA peptides in thecontext of MHC class I and class II in vitro. Given that thelentivirally transduced DC present OVA peptides in thecontext of MHC class I and class II, we argued that thesehuIi80tOVA-transduced muDC should be potent inducersof OVA-specific CTL in vivo. Indeed, when we immunisedmice with huIi80tOVA-transduced muDC, we detected astrong and specific CTL activity against OVA-expressingtumor cells with lysis of 75–100%. Firat et al. [47]obtained specific lysis of up to 50% against melanoma-associated antigens. However, these antigens are weakerthan ovalbumin. Esslinger et al. [52] also induced antigen-specific CTL in vivo as they showed via tetramer staining



and quantification of the IFNγ production by theseT cells. However, they did not perform a cytotoxicityassay. We further showed that the immune responseinitiated after injection of the huIi80tOVA-transducedmuDC was protective against subsequent challenge withOVA-expressing tumor cells. Despite this protection, anumber of huIi80tOVA-immunised mice developed lethaltumors. We showed by RT-PCR that these tumors hadlost OVA expression. These results are consistent with theresults of Metharom et al. [50], who performed similarexperiments with murine tyrosinase related protein 2,which is endogenously expressed by B16 melanoma.Epitope loss through downregulation of the tumor antigenor molecules related to antigen presentation (MHC classI, TAP) is a major problem in cancer immunotherapy.This problem can be partially addressed by optimising thevaccination strategy through immunisation with differentantigens in a serial vaccination regimen.

We investigated whether huDC transduced with thecancer-germline antigen, MAGE-A3, fused to the huIi80were able to present MAGE-A3 peptides in the contextof HLA class I and class II. Indeed, the mature huDCtransduced at a MOI of 15 with huIi80MAGE-A3 wereable to present MAGE-A3 peptides to a CD4+ T cellclone which recognises MAGE-A3 in the context of HLA-DP0401 and to a CD8+ T cell clone which recognisesMAGE-A3 in the context of HLA-A1 in vitro. Moreimportantly, the huIi80MAGE-A3-transduced DC wereable to prime MAGE-A3-specific CD4+ and CD8+ T cellsin vitro, as tested by IFNγ production and cytotoxicityassay, respectively. Dyall et al. [53] showed earlier thatlentivirally transduced huDC can activate flu-specific CTLin vitro; however, these are not naive T cells. Firat et al.[47] used tetramer staining to show that lentivirallytransduced huDC were able to induce proliferation ofCTL against a melanoma poly-epitope in vitro. Herewe report, for the first time, the induction of antigen-specific CD4+ T cells by lentivirally transduced DC. Theinduction of T helper cells is very important, since therequirement for both CD4+ and CD8+ T cells for theinduction of an effective immune response has beendocumented [14,54]. The role of CD4+ T cells haslargely been attributed to providing regulatory signalsfor the induction and maintenance of the CD8+ effectorT cells [18]. However, the CD4+ T cells also play a moredirect role in the effector phase of the tumor rejection[19,20].

We conclude that the use of lentivirally transduced DCexpressing tumor antigens fused to MHC class II targetingsignals, such as Ii80, is an effective approach for DC-basedcancer immunotherapy.

Acknowledgements

This work was supported by grants from the Institute of Scienceand Technology Flanders (IWT) to K.B., from the Fund forScientific Research-Flanders (FWO) to M.D., and from theMinistry of Science (IUAP/PAI IV), the Fortis Bank, and ‘DeBelgische Federatie voor Kankerbestrijding’ to K.T.

References1. Van Der Bruggen P, Zhang Y, Chaux P, et al. Tumor-specific

shared antigenic peptides recognized by human T cells. ImmunolRev 2002; 188: 51–64.

2. Pardoll DM. Cancer vaccines. Trends Pharmacol Sci 1993; 14:202–208.

3. Pardoll DM. Spinning molecular immunology into successfulimmunotherapy. Nat Rev Immunol 2002; 2: 227–238.

4. Banchereau J, Briere F, Caux C, et al. Immunobiology ofdendritic cells. Annu Rev Immunol 2000; 18: 767–811.

5. Steinman RM. The dendritic cell system and its role inimmunogenicity. Annu Rev Immunol 1991; 9: 271–296.

6. Steinman RM, Inaba K, Turley S, et al. Antigen capture,processing, and presentation by dendritic cells: recent cellbiological studies. Hum Immunol 1999; 60: 562–567.

7. Cella M, Sallusto F, Lanzavecchia A. Origin, maturation andantigen presenting function of dendritic cells. Curr Opin Immunol1997; 9: 10–16.

8. Fong L, Engleman EG. Dendritic cells in cancer immunotherapy.Annu Rev Immunol 2000; 18: 245–273.

9. Timmerman JM, Levy R. Dendritic cell vaccines for cancerimmunotherapy. Annu Rev Med 1999; 50: 507–529.

10. Mayordomo JI, Zorina T, Storkus WJ, et al. Bone marrow-derived dendritic cells pulsed with synthetic tumour peptideselicit protective and therapeutic antitumour immunity. Nat Med1995; 1: 1297–1302.

11. Zitvogel L, Mayordomo JI, Tjandrawan T, et al. Therapy ofmurine tumors with tumor peptide-pulsed dendritic cells:dependence on T cells, B7 costimulation, and T helper cell1-associated cytokines. J Exp Med 1996; 183: 87–97.

12. Nestle FO, Alijagic S, Gilliet M, et al. Vaccination of melanomapatients with peptide- or tumor lysate-pulsed dendritic cells. NatMed 1998; 4: 328–332.

13. Schuler-Thurner B, Schultz ES, Berger TG, et al. Rapid inductionof tumor-specific type 1 T helper cells in metastatic melanomapatients by vaccination with mature, cryopreserved, peptide-loaded monocyte-derived dendritic cells. J Exp Med 2002; 195:1279–1288.

14. De Veerman M, Heirman C, Van Meirvenne S, et al. Retrovirallytransduced bone marrow-derived dendritic cells require CD4+ Tcell help to elicit protective and therapeutic antitumor immunity.J Immunol 1999; 162: 144–151.

15. Kalams SA, Walker BD. The critical need for CD4 help inmaintaining effective cytotoxic T lymphocyte responses. J ExpMed 1998; 188: 2199–2204.

16. Toes RE, Ossendorp F, Offringa R, et al. CD4 T cells and theirrole in antitumor immune responses. J Exp Med 1999; 189:753–756.

17. Li Q, Yu B, Grover AC, et al. Therapeutic effects of tumorreactive CD4+ cells generated from tumor-primed lymph nodesusing anti-CD3/anti-CD28 monoclonal antibodies. J Immunother2002; 25: 304–313.

18. Dorothee G, Vergnon I, Menez J, et al. Tumor-infiltrating CD4+T lymphocytes express APO2 ligand (APO2L)/TRAIL uponspecific stimulation with autologous lung carcinoma cells:role of IFN-alpha on APO2L/TRAIL expression and -mediatedcytotoxicity. J Immunol 2002; 169: 809–817.

19. Qin Z, Blankenstein T. CD4+ T cell-mediated tumor rejectioninvolves inhibition of angiogenesis that is dependent onIFN gamma receptor expression by nonhematopoietic cells.Immunity 2000; 12: 677–686.

20. Schnell S, Young JW, Houghton AN, et al. Retrovirallytransduced mouse dendritic cells require CD4+ T cell helpto elicit antitumor immunity: implications for the clinical use ofdendritic cells. J Immunol 2000; 164: 1243–1250.

21. Pieters J, Bakke O, Dobberstein B. The MHC class II-associatedinvariant chain contains two endosomal targeting signals withinits cytoplasmic tail. J Cell Sci 1993; 106: 831–846.

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. Zufferey R, Dull T, Mandel RJ, et al. Self-inactivating lentivirusvector for safe and efficient in vivo gene delivery. J Virol 1998;72: 9873–9880.

24. Vigna E, Naldini L. Lentiviral vectors: excellent tools forexperimental gene transfer and promising candidates for genetherapy. J Gene Med 2000; 2: 308–316.



25. Chinnasamy N, Chinnasamy D, Toso JF, et al. Efficient genetransfer to human peripheral blood monocyte-derived dendriticcells using human immunodeficiency virus type 1-basedlentiviral vectors. Hum Gene Ther 2000; 11: 1901–1909.

26. Lewis PF, Emerman M. Passage through mitosis is required foroncoretroviruses but not for the human immunodeficiency virus.J Virol 1994; 68: 510–516.

27. Follenzi A, Ailles LE, Bakovic S, et al. Gene transfer by lentiviralvectors is limited by nuclear translocation and rescued by HIV-1pol sequences. Nat Genet 2000; 25: 217–222.

28. Zufferey R, Nagy D, Mandel RJ, et al. Multiply attenuatedlentiviral vector achieves efficient gene delivery in vivo. NatBiotechnol 1997; 15: 871–875.

29. Chinnasamy D, Chinnasamy N, Enriquez MJ, et al. Lentiviral-mediated gene transfer into human lymphocytes: role of HIV-1accessory proteins. Blood 2000; 96: 1309–1316.

30. Miyoshi H, Blomer U, Takahashi M, et al. Development of a self-inactivating lentivirus vector. J Virol 1998; 72: 8150–8157.

31. Zennou V, Petit C, Guetard D, et al. HIV-1 genome nuclearimport is mediated by a central DNA flap. Cell 2000; 101:173–185.

32. 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.

33. Van Meirvenne S, Straetman L, Heirman C, et al. Efficientgenetic modification of murine dendritic cells by electroporationwith mRNA. Cancer Gene Ther 2002; 9: 787–797.

34. Schultz ES, Lethe B, Cambiaso CL, et al. A MAGE-A3 peptidepresented by HLA-DP4 is recognized on tumor cells by CD4+cytolytic T lymphocytes. Cancer Res 2000; 60: 6272–6275.

35. Schultz ES, Zhang Y, Knowles R, et al. A MAGE-3 peptiderecognized on HLA-B35 and HLA-A1 by cytolytic T lymphocytes.Tissue Antigens 2001; 57: 103–109.

36. Naldini L, Blomer U, Gage FH, et al. Efficient transfer,integration, and sustained long-term expression of the transgenein adult rat brains injected with a lentiviral vector. Proc Natl AcadSci U S A 1996; 93: 11382–11388.

37. Tuyaerts S, Noppe SM, Corthals J, et al. Generation of largenumbers of dendritic cells in a closed system using cell factories.J Immunol Methods 2002; 264: 135–151.

38. Lutz MB, Kukutsch N, Ogilvie AL, et al. An advanced culturemethod for generating large quantities of highly pure dendriticcells from mouse bone marrow. J Immunol Methods 1999; 223:77–92.

39. Chaux P, Vantomme V, Stroobant V, et al. Identification ofMAGE-3 epitopes presented by HLA-DR molecules to CD4(+) Tlymphocytes. J Exp Med 1999; 189: 767–778.

40. Van Tendeloo VF, Snoeck HW, Lardon F, et al. Nonviraltransfection of distinct types of human dendritic cells: high-efficiency gene transfer by electroporation into hematopoietic

progenitor- but not monocyte-derived dendritic cells. Gene Ther1998; 5: 700–707.

41. Zhong L, Granelli-Piperno A, Pope M, et al. Presentation ofSIVgag to monkey T cells using dendritic cells transfected with arecombinant adenovirus. Eur J Immunol 2000; 30: 3281–3290.

42. Dietz AB, Vuk-Pavlovic S. High efficiency adenovirus-mediatedgene transfer to human dendritic cells. Blood 1998; 91:392–398.

43. Kaplan JM, Yu Q, Piraino ST, et al. Induction of antitumorimmunity with dendritic cells transduced with adenovirusvector-encoding endogenous tumor-associated antigens. JImmunol 1999; 163: 699–707.

44. Amado RG, Chen IS. Lentiviral vectors – the promise of genetherapy within reach? Science 1999; 285: 674–676.

45. Follenzi A, Sabatino G, Lombardo A, et al. Efficient gene deliveryand targeted expression to hepatocytes in vivo by improvedlentiviral vectors. Hum Gene Ther 2002; 13: 243–260.

46. Kim VN, Mitrophanous K, Kingsman SM, et al. Minimalrequirement for a lentivirus vector based on humanimmunodeficiency virus type 1. J Virol 1998; 72: 811–816.

47. Firat H, Zennou V, Garcia-Pons F, et al. Use of a lentiviralflap vector for induction of CTL immunity against melanoma.Perspectives for immunotherapy. J Gene Med 2002; 4: 38–45.

48. Manganini M, Serafini M, Bambacioni F, et al. A humanimmunodeficiency virus type 1 pol gene-derived sequence(cPPT/CTS) increases the efficiency of transduction of humannondividing monocytes and T lymphocytes by lentiviral vectors.Hum Gene Ther 2002; 13: 1793–1807.

49. Haas DL, Case SS, Crooks GM, et al. Critical factors influencingstable transduction of human CD34(+) cells with HIV-1-derivedlentiviral vectors. Mol Ther 2000; 2: 71–80.

50. Metharom P, Ellem KA, Schmidt C, et al. Lentiviral vector-mediated tyrosinase-related protein 2 gene transfer to dendriticcells for the therapy of melanoma. Hum Gene Ther 2001; 12:2203–2213.

51. Rouas R, Uch R, Cleuter Y, et al. Lentiviral-mediated genedelivery in human monocyte-derived dendritic cells: optimizeddesign and procedures for highly efficient transductioncompatible with clinical constraints. Cancer Gene Ther 2002;9: 715–724.

52. Esslinger C, Romero P, MacDonald HR. Efficient transductionof dendritic cells and induction of a T-cell response bythird-generation lentivectors. Hum Gene Ther 2002; 13:1091–1100.

53. Dyall J, Latouche JB, Schnell S, et al. Lentivirus-transducedhuman monocyte-derived dendritic cells efficiently stimulateantigen-specific cytotoxic T lymphocytes. Blood 2001; 97:114–121.

54. Bennett SR, Carbone FR, Karamalis F, et al. Induction of a CD8+cytotoxic T lymphocyte response by cross-priming requirescognate CD4+ T cell help. J Exp Med 1997; 186: 65–70.


Lentivirally transduced dendritic cells as a tool for cancer immunotherapy

Documents