Targeting Dendritic Cells with Antigen Via Dendritic Cell-Associated Promoters

ORIGINAL ARTICLE Targeting dendritic cells with antigen via dendritic cell-associated promoters

V Moulin1,3, ME Morgan1,3, D Eleveld-Trancikova1,3, JBAG Haanen2, E Wielders1, MWG Looman1, RAJ Janssen1, CG Figdor1, BJH Jansen1 and GJ Adema1

The induction of tumor-specific immune responses is largely dependent on the ability of dendritic cells (DCs) to present tumor-associated antigens to T lymphocytes. Therefore, we investigated the use of DC-associated promoter-driven genetic vaccines to specifically target DC in vivo. Restricted expression of vaccine-encoding genes in DC should enhance specificity and improves their safety for clinical applications. Hereto, 3--5 kb upstream sequences of the murine genes encoding CD11c, DC-SIGN, DC-STAMP and Langerin were isolated, characterized and subcloned into enhanced green fluorescent protein (EGFP) reporter constructs. Upon electroporation, EGFP was expressed in DC cell lines, but not in other cell lines, confirming DC- restricted promoter activity. When these promoters were cloned into a construct upstream of the gene for ovalbumin (OVA), it appeared that DC-STAMP promoter-driven expression of OVA (pDCSTAMP/OVA) in DC yielded the most efficient OVA-specific CD4 þ and CD8 þ T-cell responses in vitro. Administration of pDC-STAMP/OVA in vivo, using the tattoo gun vaccination system, evoked specific immune responses as evidenced in a mouse tumor model. Adoptively transferred pDC-STAMP/OVA-transfected DCs induced strong CD8 þ T-cell proliferation in vivo. These experiments demonstrate that our DC-directed promoter constructs are potential tools to restrict antigen expression in DC and could be implemented to modulate DC function by the introduction of relevant proteins.

Cancer Gene Therapy (2012) 19, 303--311; doi:10.1038/cgt.2012.2; published online 24 February 2012 Keywords: DNA vaccination; DC-speciﬁc promoter; dendritic cell; tattoo

INTRODUCTION cross-presenting antigen. Several ways of cutaneous DNA delivery Dendritic cells (DCs) are the most potent antigen presenting cells have been reported, ranging from the tape stripping, lipid- of the immune system. DCs reside in an immature stage in nearly mediated and nanoparticles-mediated transfer to needle-free all peripheral tissues where they capture, process and present electorporation, gene gun and tattoo.4--18 The availability of local antigens. Upon encountering pathogens, DCs mature, antigen produced by transfected keratinocytes, despite being migrate to draining lymph nodes and prime T lymphocytes.1 produced abundantly, was limited unless apoptosis of the Because of their essential role in catalyzing the adaptive immune keratinocytes was induced.2 Thus, it appears that sufficient response, DCs are currently exploited in tumor immunotherapies. targeting of DCs is crucial for the development of antigen-specific One method to direct DC responses for tumor immunotherapy is responses as a result of DNA vaccination. to use DNA vaccination combined with DC-specific promoters to To date, a few DC-specific promotors have been used to transcriptionally target protein expression to DCs. These proteins transcriptionally target DCs.17,19 --21 Despite the transfection of could be, for instance, cytokines, growth factors or signal very few cells, Ross and co-workers21 were able to detect strong transducers and they would ultimately have the ability to alter antigen-specific Th1 responses and these responses could under- the activation state of the DCs. This technology also allows mine the development of Th2 responses in an in vivo allergy expression of tumor antigens directly by DCs for the induction of model. Nevertheless, anti-tumor immunity, involving the T-cell antigen-specific immune responses, and limiting vaccine antigen responses or even tumor rejection, has yet to be demonstrated expression to DCs could also improve its safety for clinical after DNA vaccination with constructs carrying DC-specific applications. promotors. Recently, the 50 untranslated region of DC-STAMP Following DNA vaccination using a broadly and highly active gene was successfully used in lentiviral vectors yielding long-term promoter, such as cytomegalovirus (CMV) immediate early and cell-selective transgene expression in vivo.13 promoter (pCMV), antigens are delivered to DCs through either In order to develop additional vectors for permissive expression cross-presentation or direct antigen expression by DCs. Several in DCs in vitro and in vivo with protein, the promoters of the studies indicate that the latter method is mainly responsible for murine genes encoding cd11c, dc-sign,22 dc-stamp23 and langerin24 antigen presentation in the lymph nodes.2,3 It was shown that were cloned. We demonstrated that the DNA sequences that we during cutaneous DNA vaccination, in vivo transfected DCs that cloned were primarily active in the DC cell lines, D1, and directly express target antigens were primarily responsible for differentiated Raw 264.7 cells. All four promoters driving the activation of CD4 þ and CD8 þ T cells, as opposed to DCs ovalbumin (OVA) expression in DCs lead to specific CD8 þ T-cell

1Department of Tumor Immunology, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands and 2Division of Immunology, The Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands. Correspondence: Professor GJ Adema, Department of Tumor Immunology, TIL278, Nijmegen Centre for Molecular Life Sciences, Radboud University Nijmegen Medical Centre, POBOX 9101, 6500 HB Nijmegen, The Netherlands. E-mail: [email protected] 3These authors contributed equally to this work. Received 15 June 2011; revised 14 November 2011; accepted 22 December 2011; published online 24 February 2012 Transcriptional targeting of dendritic cells V Moulin et al 304 proliferation in vitro. Using the pDC-STAMP/OVA construct, it was phage cell line, Raw 264.7, was cultured in RPMI 1640 supplemented with shown that DC-STAMP promoter-driven OVA expression in DCs 10% FCS, glutamine, 50 mM b-Mercaptoethanol and antibiotics/antimyco- was exceptional in that it could induce CD4 þ as well as CD8 þ tics. To induce differentiation towards a DC phenotype, Raw 264.7 cells T-cell responses. Further, experiments with pDC-STAMP/OVA were cultured in the presence of murine recombinant GM-CSF (20 ng mlÀ1) demonstrated in vivo potential as pDC-STAMP/OVA-transfected and IL-4 (20 ng mlÀ1) for 4 days. The DC cell line, D1 (kindly provided by DCs were functional after adoptive transfer and tumor progression Dr Paola Ricciardi-Castagnoli) was maintained in culture as previously in the B16OVA melanoma model was hindered after tattoo described.25 The B3Z T-cell hybridoma26 and the OVA-transfected, murine vaccination. Our results conﬁrm that DC-associated promoters can melanoma cell line, B16OVA,27 (derived from B16F10, clone MO5) was successfully drive gene expression in DCs. Moreover, besides kindly provided by Dr Kenneth Rock. The medium used for the in vitro being potential tools to target tumor antigens to DCs, our T-cell stimulation was RPMI 1640 supplemented with 5% FCS, glutamine, promoters could be relevant tools to modify DC function by the 50 mM b-Mercaptoethanol and antibiotics/antimycotics. All transfections introduction of other effector proteins like cytokines and signal were performed using the Amaxa Nucleofector system (Amaxa, Cologne, transducers. Germany). Please see Table 1 for details regarding the amount of cells and the transfection conditions used.

MATERIALS AND METHODS Animals Vector constructs and plasmid isolation Female C57Bl/6N mice (Charles River Wiga, Sulzfeld, Germany) were The promoter sequences of murine cd11c, dc-sign, dc-stamp and langerin obtained at 6--8 weeks of age. OVA-specific TCR-transgenic OT-I and OT-II genes were determined using the Celera Database following an alignment mice were bred at the Central Animal Laboratory located at the Radboud with the corresponding cDNA. Using the primers sets described in Table 2 University Nijmegen Medical Centre (Nijmegen, The Netherlands). Animals and the Expand Long Template PCR system (Roche diagnostics, were maintained under specific pathogen-free barrier conditions at the Mannheim, Germany), 3--5 kb sequences upstream from the starting Central Animal Laboratory (Nijmegen, the Netherlands) and the experi- codon of the genes were amplified from C57BL/6N genomic DNA and ments were performed in accordance to the guidelines for animal care of subcloned into the multiple cloning site of the promoter-less enhanced the Nijmegen Animal Experiments Committee. green fluorescent protein (pEGFP)-1 vector (GenBank Accession: #U55761; Clontech, Saint-Germain-en-Laye, France). All constructs were checked by sequencing at the sequencing facility of the Department of Human Cell lines and transfections Genetics, Radboud University Nijmegen Medical Centre, and compared Murine embryonic fibroblast cells expressing H-2b class I molecules and with publicly available sequences. During PCR, restriction enzyme sites NIH3T3 cells (H-2d haplotype) were cultured in Dulbecco’s modified Eagle were introduced into the primers (see Table 2 for a list) and were used to medium supplemented with 10% fetal calf serum (FCS), glutamine and insert the fragments into the multiple cloning site. The CMV promoter was antibiotics/antimycotics (Invitrogen, Breda, The Netherlands). The macro- isolated from the pEGFP-N3 vector (GenBank Accession: #U57609; Clontech) and inserted in the multiple cloning site of pEGFP-1. The bovine keratin 5 promoter28 was amplified from a plasmid, kindly provided by Dr JL Jorcano, using primers introducing SalI and XmaI sites at the 50 and 30 Table 1. Transfection conditions for each cell type ends, respectively. The amplified fragment was further digested with SalI/ AgeI restriction enzymes and inserted into the multiple cloning site of the pEGFP-1 reporter system. The sequence coding for the full length of the Cell Amount of cells Programs; Amount of DNA type per transfection solutions chicken OVA was amplified from the pMFG/OVA plasmid, kindly provided by Dr Kris Thielemans, using the primer sets described in Table 2. pCMV (mg) Others (mg) Subsequently, the coding sequence of EGFP was replaced by the OVA- NIH3T3 2 Â 106 A-24; sol. R 5 10 coding sequence using AgeI/NotI sites. The plasmids were propagated in MEF 2 Â 106 T-20; sol. T 2.5 5 the Escherichia coli strain DH10 and purified from overnight cultures Raw 264,7 3 Â 106 U-24; sol. V 1.5 3 using either the Qiagen Plasmid or the EndoFree Plasmid kits (Qiagen, D1 4--8 Â 106 P-20; sol. T 3 6 Venlo, The Netherlands) according the manufacturer’s instructions. After DNA precipitatation, it was resuspended in pyrogen-free water and Abbreviation: pCMV, cytomegalovirus immediate early promoter. stored at À20 1C.

Table 2. Primers used to amplify the promoter regions

Oligos+RE sites Sequences Size (kb)

CD11c F---SacII 50-ACGACGCCGCGGACCAGGAGGGTTCGTTAGGGAA-30 3149 CD11c R---AgeI50-CGTCGTACCGGTAGAACAGAAGCAGGCTCTGAGCAAC-30 DC-SIGN F---SalI50-ACGACGGTCGACAAGGAGGTTCAGCAAGCAGCCAGA-30 4036 DC-SIGN R---AgeI50-CGTCGTACCGGTTTTCACAGCCACTTCTCAGTGCCAG-30 DC-STAMP F---SacII 50-ACGACGCCGCGGTGTATTCCAAGGTGGCTTGACCAG-30 4746 DC-STAMP R---AgeI50-CGTCGTACCGGTACAAGCACAAAGCACACCCTTCTGG-30 Langerin F---SalI50-ACGACGGTCGACTTCTCCTCACCAACACTGGCTGGCT-30 4047 Langerin R---ApaI50-CGTCGTGGGCCCCTCAATATCTGTCTCCCCAAGGAC-30 Keratinocyte 5 F---XmaI50-ACGACGCCCGGGGTTGAAACGCTGGGCAATATCG-30 5316 Keratinocyte 5 R---SalI50-CGTCGTGTCGACCCTGCAGGTCAACGGATCAA-30 Ovalbumin F---AgeI50-ACGACGACCGGTCGCCACCATGGGCTCCATCGGTGCAGCAAG-30 1160 Ovalbumin R---NotI50-TATCAAATCGCCGGCGTTAAGGGGAAACACATCTGCC-30 Abbreviation: RE, restriction enzyme.

Cancer Gene Therapy (2012), 303 --311 & 2012 Macmillan Publishers Limited Transcriptional targeting of dendritic cells V Moulin et al 305 Promoter analysis three times per week and mice were killed when the tumor reached a Upstream promoters of cd11c, dc-sign, dc-stamp and langerin were width of 13 mm. analyzed for the possible presence of transcription factor binding sites (TFBSs) using the program P-match (http://www.gene-regulation.com/ Detection of OVA-specific T cells cgi-bin/pub/programs/pmatch/bin/p-match.cgi) that scans input sequence Cell suspensions were stained in ice-cold phosphate-buffered saline with mononucleotide weight matrices as well as sets of aligned known supplemented with 2 mM EDTA, 1% bovine serum albumin and 0.02% TFBS from the TRANSFAC database (version 7.0 Public). Only high-quality sodium azide (fluorescent-activated cell sorting buffer). The monoclonal matrices were used in the screen and only those having minimal sums for antibody against CD11c (clone HL3), anti-CD4 and anti-CD8a were all false-positive and false-negative error rates were included. Additionally, b obtained from BD Biosciences (Erembodegem-Aalst, Belgium). H-2 K /OVA the promoter sequences were also analyzed for the presence of pre- Tetramers detecting CD8 þ T cells specific for OVA peptide (257--264; selected enhancer elements using the program MacVector (version 7.0, Pelimers, Sanquin, Amsterdam, The Netherlands) were used as previously MacVector, Cary, NC). These selected enhancer elements are consensus described.30 sequences rather than weight matrices and only exact matches were considered. RESULTS RNA isolation and reverse transcriptase PCR The promoters for cd11c, dc-sign, dc-stamp and langerin contain Total RNA was isolated from untreated or GM-CSF/IL-4-treated NIH3T3 and TFBSs involved in DC development and function Raw 264.7 cells and from immature or LPS-matured D1 cells using Trizol As DCs are pivotal to the development of immune responses reagent (Invitrogen) and was subsequently transcribed into cDNA using during DNA vaccination, we cloned a variety of promoters random hexamers and Moloney murine leukemia virus transcriptase preferentially targeting antigen production in DCs. To this end, (Invitrogen). Primers for DC-STAMP (forward 50-CCGCTGTGGACTATC we first characterized selected promoter sequences 3--5 kb TGCTG-30 and reverse 50-CTCAATGGCTGCTTTGATCG-30), DC-SIGN (forward upstream from the DC-associated genes, cd11c, dc-sign, dc-stamp 50-AAGGAAATGGGGAAGAGGCA-30 and reverse 50-ACTTGACTGTGGACCA and langerin, using bioinformatics. As shown in Figure 1, the GGCA-30), CD11c (forward 50-CTGAGAGCCCAGACGAAGACA-30 and reverse analysis revealed the presence of several TFBSs that may have a 50-TGAGCTGCCCACGATAAGAG-30) and Langerin (forward 50-GATCTGTC role in myeloid function and/or differentiation. Binding sequences TCCAGTTCTGAG-30 and reverse 50-TTGGCACTGAGACTGTCAAC-30) were for c-Rel and AP1 were found extensively throughout all the used for PCR analysis. As a control for RNA quality, b-actin was amplified promoters. Additional DC-associated TFBS, such as SPI1 and YY1, in parallel. were included only if they conformed to the criteria of high- quality matrices or exactly matched selected consensus sequences (see Materials and Methods for details). These analyses confirmed In vitro stimulations of B3Z, OT-I or OT-II T cells that 3--5 kb upstream of the start codon for our chosen DC- 4 5 Â 10 B3Z cells per well were cultured in flat-bottom 96-well plates with associated genes contained multiple promoter elements neces- varying amounts of D1 or murine embryonic fibroblast cells, which were sary to confer DC-directed activity. transfected with either pCMV/À, pK5/À, pDCSTAMP/À, pDC-SIGN/À, pCD11c/À or pLangerin/OVA constructs. After 24 h, b-Galactosidase The promoters for cd11c, dc-sign, dc-stamp and langerin are activity of the B3Z cells was determined by incubating the cells with preferentially active in DC cell lines 0.15 mM chlorophenolred- a-D-galactopyranoside (Calbiochem, Darmstadt, In order to examine the selected promoters for DC specificity, we Germany), 1 mM MgCl2, 0.125% Nonidet P-40 and 100 mM b-Mercaptoetha- nol in phosphate-buffered saline. After 3 h at 37 1C, the optical density was obtained several non-DC and DC cell lines. NIH3T3 fibroblasts and measured at 595 nm. the macrophage cell line Raw 264.7 were used as non-DC cell Naı¨ve OVA-specific CD4 þ (OT-II) and CD8 þ (OT-I) T cells were isolated lines. D1, the murine DC cell line, and Raw 264.7 cells treated with from spleen and inguinal lymph nodes using magnetic assisted cell sorting murine recombinant GM-CSF and IL-4 were used as example DC (Miltenyi Biotec, Auburn, CA). OT-I cells were first stained with anti-CD8b.2 cell lines. As shown in Figure 2a, untreated Raw cells did not FITC (BD PharMingen, Alphen aan den Rijn, The Netherlands) and then express CD11c, whereas an increased CD11c expression was ± incubated with anti-FITC magnetic assisted cell sorting beads. OT-II cells observed ( 60% of cells are CD11c-positive) upon their treatment were positively selected by use of directly-labeled anti-CD4 magnetic with recombinant GM-CSF and IL-4 for 4 days. The DC specificity of assisted cell sorting beads. Following isolation of the T-cell subsets, the cells our promoter sequences was first examined by determining the were stained with carboxyfluorescein diacetate succinimidyl ester (CFSE) expression of endogenous genes from which the promoters were (Molecular Probes Europe BV, Leiden, The Netherlands) at a concentration later isolated. No endogenous mRNA encoding cd11c, dc-sign, dc- stamp (full length and splice variant) and langerin was detectable of 0.5 mM in 1% FCS/phosphate-buffered saline for 10 min at 37 1Cinthe dark. Staining was stopped by the addition of pure FCS and the cells were in untreated or GM-CSF/IL-4-treated NIH3T3 cells (Figure 1b). Raw then washed with medium. In all, 2 Â 105 CFSE-stained OT-I or OT-II were 264.7 cells did express low levels of cd11c, dc-stamp and langerin. cultured in round-bottom 96-well plates with varying amounts of D1 cells, However, their expression was upregulated following treatment transfected 6 h earlier, with either irrelevant DNA (pCMV/EGFP) or pCMV/À, with GM-CSF and IL-4, confirming differentiation towards a DC-like pK5/À, pCD11c, pDC-SIGN/À, pDC-STAMP/À or pLangerin/OVA constructs. phenotype. No endogenous mRNA encoding dc-sign was detect- OT-I or OT-II cell proliferation was measured by CFSE dilution at 48 and 96 h, able in Raw 264.7 cells regardless of cytokine treatments. As respectively, after the beginning of the coculture. expected, D1 cells endogenously expressed all the four genes even after maturation with lipopolysaccharide. These results indicate that these genes are preferentially expressed within DC Immunization protocol and induction of melanoma cell types and are likely good candidates for transcriptional After shaving, mice were tattooed three times (day 0, 3 and 7) for 20 s on targeting of DCs. the left leg, using a Swiss Rotary Machine Mark II tattoo machine equipped Next, the DC-associated 3--5 kb upstream promoter sequences with Radical Clean Magnum 11-needle bars (B.S. Trading B.V., Amsterdam, were cloned into vectors encoding EGFP and analyzed following The Netherlands). This technique of dermal DNA vaccination has been transfection (Figure 2c). The bovine keratin 5 promoter (pK5), previously described.29 DNA vaccinations were performed with 15 mgof known to be active in epidermal murine cells,28 was also cloned in pCMV/EGFP as irrelevant DNA, 15 mg of pCMV/OVA or 30 mg of pDC- the same plasmid backbone for use as a negative control. Using STAMP/OVA. At day 8, 15 000 OVA-transfected B16OVA tumor cells were the Amaxa Nucleofector system, we electroporated DC cell lines injected subcutaneously into the left flank. Tumor growth was monitored (GM-CSF/IL-4-treated Raw 264.7 and D1 cells) and non-DC cell

& 2012 Macmillan Publishers Limited Cancer Gene Therapy (2012), 303 --311 Transcriptional targeting of dendritic cells V Moulin et al 306

Figure 1. TFBSs involved in DC development and function are present in all promoters. Promoters were analyzed as explained in Materials and methods. Each promoter sequence is represented by a single horizontal line. TFBS identiﬁed using P-match are indicated above the promoter sequence, whereas those identiﬁed by MacVector are indicated below the sequence.

lines (NIH3T3 fibroblasts and untreated Raw 264.7 cells) with OVA production via pDC-STAMP activates both OVA-specific constructs containing the DC promoters driving expression of the CD8 þ and CD4 þ T cells EGFP reporter. As a positive control, a construct containing the Next, we examined whether antigens, expressed under the control ubiquitous CMV promoter was also used. Figure 2d shows that of promoters that are preferentially active in DCs, are processed 55.39% of pCMV/EGFP-transfected NIH3T3 were EGFP þ , whereas and presented to T cells and are capable of initiating specific only very few EGFP þ NIH3T3 were detected after their transfection immune responses. For this purpose, the coding sequence for the with the negative control (pK5/EGFP construct) or the pDC- protein EGFP was replaced by a sequence coding the foreign SIGN/À, pDC-STAMP/À or pLangerin/EGFP plasmids. Surprisingly, antigen OVA (full length). Using these new constructs, D1 cells pCD11c/EGFP was active within NIH3T3 cells (approximately 14%). were transfected and tested for their capacity to activate the OVA- The treatment of NIH3T3 cells with GM-CSF and IL-4 did not affect specific hybridoma B3Z. The B3Z T-cell hybridoma contains a lacZ the EGFP expression driven by most of the promoters, except reporter under the control of the il-2 promoter. Upon engagement pCD11c, which showed a slight increase. A similar EGFP expression of its TCR, which recognizes the OVA peptide (257--264) in the pattern was observed in untreated Raw 264.7 cells (Figure 2e), context of H-2kb IL-2 is produced along with b-Galactosidase.26 As however, after the differentiation of Raw 264.7 cells towards a DC a negative control, the same constructs were also transfected into phenotype using IL-4 and GM-CSF, a significantly increased murine embryonic fibroblasts expressing H-2kb The pCMV/OVA- activity of pCD11c (1.7-fold), pDC-SIGN (3.6-fold) and pDC-STAMP transfected murine embryonic fibroblast cells were able to induce (4.2-fold) was seen. As displayed in Figure 2f, all DC promoters IL-2 production in the B3Z cells (Figure 3a). However, none of the were able, to a varying degree, to drive EFGP expression in D1 OVA constructs were capable of activating IL-2 production in this cells. Furthermore, the mean fluorescence intensity of the EGFP non-DC cell line. In contrast, all of the constructs transfected into expression, which gives an indication of the amount of the D1 cells were able to activate B3Z cells (Figure 3b). The pDC- antigen produced per cell, indicated that pDC-STAMP was STAMP/OVA plasmid was the most efficient of all the DC the most active of all the DC-associated promoters (Figure 2g). promoters in inducing IL-2 production. Taken together, these data demonstrate that the We then tested the capacity of D1 cells, transfected with distinct four isolated DC-associated promoters contain the essential DC-promoter/OVA constructs, to induce the in vitro proliferation of elements to confer DC-specific expression and are active within naı¨ve CD8 þ (OT-I) and CD4 þ (OT-II) OVA-specific T cells. OT-I and our DC cell lines. OT-II T cells, isolated from spleen and inguinal LN, were labeled

Cancer Gene Therapy (2012), 303 --311 & 2012 Macmillan Publishers Limited Transcriptional targeting of dendritic cells V Moulin et al 307

Figure 2. The promoters for cd11c, dc-sign, dc-stamp and langerin are endogenously active within DC cell lines. (a) NIH3T3 or Raw 264.7 cells were treated in the presence of GM-CSF and IL-4 or left untreated. After 4 days, cells were collected and the CD11c expression was analyzed. The thin line represents untreated cells and the thicker line indicates cells treated with GM-CSF/IL-4. (b) Reverse transcriptase PCR was used to determine mRNA expression of untreated- and cytokine-treated NIH3T3 (UT 3T3 and GM/IL-4 3T3, respectively) and untreated- and cytokine- treated Raw 264.7 cells (UT RAW and GM/IL-4 RAW) for DC-associated genes. Immature (D1) or mature D1 cells (D1 þ LPS) were also examined. The household protein, b-actin, was transcribed in parallel to confirm the DNA quality. These data are representative of three independent experiments. (c) 3--5 kb before the starting codon of the genes encoding cd11c, dc-sign, dc-stamp and langerin were isolated and cloned into vectors containing either OVA or EGFP as reporter genes. pCMV, pK5 and different DC promoter constructs driving EGFP expression were transfected into NIH3T3 cells (d), Raw 264.7 cells (e) and D1 cells (f and g). Both NIH3T3 and Raw 264.7 cells were treated for 4 days with GM- CSF/IL-4 (black bars) or left untreated (white bars). At 24 h after transfection, the percentages of EGFP þ cells (d-- f) and the mean fluorescence intensity (g) were analyzed by fluorescent-activated cell sorting (FACS). The results are representative of four independent experiments. with CFSE and cultured with D1 cells electroporated with the OVA and pCMV/EGFP. Together, these results reveal that all the various DC promoter/OVA constructs. As a positive control, D1 promoters preferentially active in DCs were able to drive antigen cells were transfected with pCMV/OVA, whereas pK5/OVA and expression and thereby activate in vitro the proliferation of naı¨ve pCMV/EGFP were used as negative controls. The DC promoter/ antigen-specific CD8 þ T cells. However, only pDC-STAMP was OVA constructs induced OT-I division to the same degree as the capable of inducing detectable levels of OVA-specific CD4 þ T-cell positive control pCMV/OVA (Figure 4a). The negative control activation. constructs, pK5/OVA and pCMV/EGFP, only induced background levels of OT-I cell proliferation. The activation of the naı¨ve OVA TCR-transgenic CD4 þ T cells Adoptive transfer of pDC-STAMP/OVA-transfected D1 leads to (OT-II) was detectable at 96 h and required more electroporated CD8 þ T-cell expansion in vivo D1 cells as compared with the stimulation required for the OT-I. Of As pDC-STAMP/OVA induced proliferation of both CD4 þ and the DC promoters, only pDC-STAMP was able to induce OT-II T-cell CD8 þ T cells in vitro, we set out to examine the activity of this proliferation (Figure 4b). However, the amount of dividing cells construct in vivo. To achieve this, we first transfected D1 cells induced by pDC-STAMP/OVA (67.91%) was lower than that of in vitro with pDC-STAMP/OVA and then transferred these cells to dividing OT-II cells following their stimulation with pCMV/OVA- mice containing 5 (and 6)-CFSE-labeled OT-I cells to examine the transfected D1 (80.05%). The other DC promoter constructs, ability of these transfected D1 cells to stimulate proliferation pCD11c/-, pDC-SIGN/- and pLangerin/OVA, only induced OT-II in vivo. As positive controls, mature D1 cells loaded with the OVA proliferation comparable to the negative control constructs pK5/ peptide (257--264) and pCMV-OVA-transfected D1 were used.

& 2012 Macmillan Publishers Limited Cancer Gene Therapy (2012), 303 --311 Transcriptional targeting of dendritic cells V Moulin et al 308 Following a similar protocol of tattooing a total of three times every 3 days, we tested our constructs directly in an in vivo tumor model. One day after the last tattoo treatment, OVA-transfected B16 melanoma cells were injected subcutaneously into the flank. As shown in Figure 6a, the pCMV/OVA-vaccinated mice were successfully protected against B16OVA tumor outgrowth confirming the effectiveness of this vaccination technique. Mice tattooed with pCMV/EGFP, however, quickly succumbed to tumor outgrowth, highlighting the necessity of OVA-encoding constructs to confer tumor protection to mice. The pDC-STAMP/OVA vaccination, though less effective than pCMV/OVA, significantly protected several mice from tumors, demonstrating that tattoo vaccination of DC-directed promoter constructs can generate a weak antigen- specific immunity. Because DC-STAMP/OVA vaccination lead to the delay of tumor outgrowth, we investigated whether endogenous-specific CD8 þ T-cell formation could be detected using H-2Kb tetramers complexed with the OVA peptide (257--264). Fifteen days after the tattoo treatments, the lymph nodes draining the site of the tattoo were isolated and examined using fluorescent-activated cell sorting. pCMV/OVA vaccination lead to development of OVA peptide-specific CD8 þ T cells. However, as shown in Figure 6b, pDC-STAMP/OVA vaccination did not induce the production of detectible amounts of OVA-specific CD8 þ T cells. Although, cellular responses were not detected with our assays, tattoo vaccination with pDC-STAMP/OVA significantly protected some mice from tumor outgrowth, indicating the induction of a weak but nonetheless detectable OVA-specific immune response. Figure 3. The B3Z hybridoma is preferentially activated by pDC- STAMP/OVA-transfected D1 cells. Murine embryonic fibroblast cells (a) or D1 cells (b) were transfected for 24 and 6 h, respectively, in the DISCUSSION absence of DNA (No DNA) or with the pCMV/À, pK5/À, pCD11c/À, DCs have a prominent role in initiating tumor-directed immune pDC-SIGN/À, pDC-STAMP/À or pLangerin/OVA constructs. Varying responses necessary for tumor immunotherapy. Currently, the amounts of transfected D1 cells were cultured with 5 Â 104 OVA- emphasis in the clinic is on producing DCs ex vivo, loading them specific B3Z T cells. The b-Galactosidase activity of B3Z cells was with relevant antigen and then administering them to cancer measured as described in the Material and methods. Results are patients. However, this process is costly and labor intensive, thus ± presented as the mean s.d. of duplicates and similar results were alternative methods are being explored to target DCs in vivo with obtained in two (a) and four (b) independent experiments. not only relevant tumor antigens but also with molecules capable of improving their pro-inflammatory qualities. Transcriptional targeting of DCs is one possible method to direct proteins to The pK5/OVA- and pCMV/EGFP-transfected D1 were used as DCs in vivo. To achieve this, we have isolated the promoters of the negative controls. The transfection procedure using electropora- DC-associated genes encoding CD11c, DC-SIGN, DC-STAMP and tion resulted in maturation of the D1 cells as evidenced by the Langerin. Using several in vitro assays, we have confirmed that our upregulation of maturation markers (data not shown). pDC- promoter sequences are primarily active in DC cell lines and that STAMP/OVA-transfected D1 cells induced the strong expansion of the antigen produced can be processed and presented to T cells. OT-I cells as evidenced by the dilution of the CFSE label (Figure 5). The DC-STAMP promoter proved to be the most potent, as it had Both positive controls also induced a large expansion of the highest activity and led to both CD8 þ and CD4 þ T-cell transferred OT-I and the negative control, D1 cells transfected responses in vitro and demonstrated in vivo potential. Thus, we with pCMV/EGFP, did not stimulate any detectable amounts of have developed a set of tools that can be used to both target DCs proliferation. pK5/OVA-transfected D1 cells showed a slight to generate antigen-specific immunity and to, on a wider scale, background proliferation. In conclusion, these results indicate modulate DC function when proteins, such as signal transducers that pDC-STAMP/OVA-transfected DCs have the capacity to or cytokines, are encoded. process and present transfected antigens in vivo. Developing sufficient immunity against tumor-related antigens is essential for tumor immunotherapy. However, typical tumor antigens are also found on normal tissues and are considered self- Detectable tumor protection after tattoo vaccination with antigens and it is difficult to develop immune responses against pDC-STAMP/OVA them. In models using cutaneous DNA vaccinations, combining Several methods exist to administer DNA in vivo via the skin. pCMV with endogenous tumor antigen genes, protective immune The skin is an ideal location for DNA vaccination because of the responses have not been detected.32 However, DC vaccinations abundance of DCs.15 The epidermis is rich in Langerhans cells with the same endogenous antigen are quite successful.32 These and the dermal region contains dermal DCs. Both of these data suggest that widespread expression of antigen is detrimental antigen presenting cells were putative targets for our DC-directed to the development of immunity against self-antigens. Thus, the DNA constructs. Another method that targets cutaneous DCs transcriptional targeting of DCs could be advantageous in this is the tattooing of DNA.18,29,31 Using a relatively low amount setting, especially in the priming phase of an immune response,30 of DNA (30 mg), tattooing induces strong cellular and humoral where one likes to avoid antigen expression by nonprofessional responses against antigens driven by ubiquitous promoters antigen presenting cells. Moreover, expression of vaccines limited like pCMV. to DC also improves their safety for clinical applications.

Cancer Gene Therapy (2012), 303 --311 & 2012 Macmillan Publishers Limited Transcriptional targeting of dendritic cells V Moulin et al 309

Figure 4. Proliferation of naı¨ve CD8 þ and CD4 þ T cells induced by pDCPromoter/OVA-transfected D1. (a)2Â 105 naı¨ve OVA-specific CD8 þ T cells (OT-I) were labeled with CFSE and cultured with 2 Â 103 D1 cells, which were, 6 h earlier, electroporated with constructs containing the ova gene under the control of pCMV, pK5, pCD11c, pDC-SIGN, pDC-STAMP, pLangerin or with pCMV/EGFP as irrelevant DNA. (b)2Â 105 CFSE- labeled, OVA-specific CD4 þ T cells (OT-II) were cultured with 3 Â 104 D1 cells transfected with the same constructs as mentioned previously (see above). After 48 h (a) and 96 h (b) of culture, CFSE dilution profiles were analyzed by FACS analysis. The bar graphs represent percentages of dividing cells for multiple experiments. Percentages of dividing cells were normalized to pCMV/OVA using the following formula: relative percent of dividing cells ¼ (% pDCPromoter/OVA construct/% pCMV/OVA) Â 100. The results are presented as mean of three (a) and four (b) experiments±s.d.

One downside of transcriptional targeting of DCs via DC-specific nodes after tattooing with our pDC-STAMP/OVA construct (data promoters is the low transfection efficiency of DCs following not shown). This could be the result of many factors, including poor in vivo vaccination. Furthermore, it has been shown previously that in vivo transfection and especially low expression, which has been the strength of the antigen expression is of high importance in previously reported for cutaneous DNA vaccination.2,3 However, driving the specific immune response,1 indicating that next to DC- this appears not to be a huge obstacle to stimulating specific specificity also expression levels are important. We were con- immune responses. It was shown before that very few transfected sistently unable to detect EGFP þ cells within the skin or lymph DCs are able to generate sufficient immune responses.2,3

& 2012 Macmillan Publishers Limited Cancer Gene Therapy (2012), 303 --311 Transcriptional targeting of dendritic cells V Moulin et al 310 Promoter activity is also essential for the development of strong immune responses to transfected antigens. Of the other DC- associated promoters that have been described in the literature (fascin,20 dc-stamp,23 cd11c,33 dectin-219 and the promoter for murine MHC class II34), the fascin promoter appears the most promising because of its high activity. Fascin promoter constructs are capable of inducing humoral and cellular responses after in vivo gene-gun vaccination to the same extent as those caused by CMV-driven constructs.21 This is largely a consequence of its activity in matured DCs. Like the pCMV promoter, the fascin promoter increases its activity as a result of DC maturation,35,36 resulting in large amounts of antigen production. Of our isolated DC-associated promoters, the DC-STAMP promoter was the most efficient in driving antigen expression in D1 cells as measured by EGFP fluorescence, which is likely the reason that it was able to induce the strongest in vitro proliferation of OVA-specific T cells. However, pDC-STAMP transcription appears to be lower after maturation (Figure 1c), and electroporation itself induce only partial maturation of D1 cells (data not shown). In contrast to pDC- STAMP, pCMV produced at least three times the amount of protein as determined by the mean fluorescence intensity of EGFP, allowing it to consistently outperform pDC-STAMP in both in vitro and in vivo assays. That pDC-STAMP is functional in vivo and is supported by the ability of pDC-STAMP/OVA-transfected D1 to stimulate CD8 þ Figure 5. Adoptively transferred pDC-STAMP/OVA-transfected T-cell proliferation after adoptive transfer and by the fact that we D1cells are functional in vivo. D1 cells were electroporated with detected a significant delay in tumor outgrowth after in vivo pCMV/OVA, pK5/OVA, pDC-STAMP/OVA or pCMV/EGFP as a negative control. As a positive control, LPS-matured D1 were loaded with the vaccination. Despite multiple attempts, we were unable to detect OVA peptide (257--264). Subsequently, the D1 cells were adoptively specific CD8 þ T-cell development or specific CD8 þ or CD4 þ transferred to mice harboring previously-injected CFSE þ OT-I cells. T-cell proliferation (data not shown) following pDC-STAMP/OVA Three days after the injection of D1, draining lymph nodes were vaccination. However, the detection of specific cellular responses isolated and examined for proliferation of OT-I cells by CFSE dilution. is not always detected concomitantly with tumor regression.37 The Similar results were found in three independent experiments. limited capacity of the DC-specific promoters used in this study to induce immunity may be related to the relatively low expression levels they induce in DC. Increasing their strength by adding enhancer-type elements like those recognized by CIITA (class II, major histocompatibility complex, transactivator) might be important to improve their activity. Vaccination by tattooing was employed as it has the potential to be an inexpensive and easily portable method to administer DNA. As it is a relatively novel method of dermal DNA vaccination, it could be argued that more traditional and widely tested methods will be more successful in combination with our constructs. Moreover, pDC-STAMP/OVA vaccination with several alternative methods of DNA vaccination (intramuscular, subcuta- neous injection and gene gun) yielded similar or worse results in the tumor setting (data not shown). Therefore, it is essential to find ways to boost the immune response of pDC-STAMP/OVA after tattoo vaccination. To this end, we plan to examine the use of a prime-boost strategy, which is known to greatly increase the potency of DNA vaccinations.38,39 Previously Dresch et al.13 have showed that lentiviral vector containing the 50 untranslated region of DC-STAMP yields long-term, DC-specific OVA expression in vivo. Moreover, their work confirms our data with respect to DC-STAMP promoter.13 To transcriptionally target DCs, we have isolated several novel promoters that have a DC-specific activity and are capable of driving antigen expression. Although we have focused on the Figure 6. pDC-STAMP/OVA-vaccinated mice have delayed tumor induction of immunity by permissive expression of relevant growth. (a) The left leg of C57Bl/6 mice (n ¼ 5--6) was tattooed at antigen in DC, the additional introduction of signal transducers day 0, 3 and 7 with pCMV/EGFP, pCMV/OVA or pDC-STAMP/OVA. On or cytokines using the same techniques may favorably influence day 8, OVA-expressing B16 tumor cells were injected sub- the behavior of DCs. Using these strategies, it may be possible to cutaneously into the left flank. Tumor growth was monitored and develop methods to use DNA vaccination in the treatment of mice were killed upon obtainment of large tumors (see Materials cancer patients. and Methods). Data is representative of two independent experiments. (b) Fifteen days after DNA vaccination, lymph nodes were isolated and examined for OVA peptide-specific (257--264) CD8 þ cells using tetramer staining. Numbers in the upper-right corners represent percentages of OVA peptide-specific CD8 þ T cells from CONFLICT OF INTEREST the total CD8 þ cell population. The authors declare no conflict of interest.

Cancer Gene Therapy (2012), 303 --311 & 2012 Macmillan Publishers Limited Transcriptional targeting of dendritic cells V Moulin et al 311 ACKNOWLEDGEMENTS 20 Ross R, Sudowe S, Beisner J, Ross XL, Ludwig-Portugall I, Steitz J et al. We would like to thank the animal caretakers of the Central Animal Laboratory of the Transcriptional targeting of dendritic cells for gene therapy using the promoter Radboud University Medical Centre Nijmegen for their conscientious work and of the cytoskeletal protein fascin. Gene Ther 2003; 10: 1035 --1040. dedication. This research was also made possible by grants from the Dutch Cancer 21 Sudowe S, Ludwig-Portugall I, Montermann E, Ross R, Reske-Kunz AB. Prophylactic Society (KWF 2004-3137 and KWF 2000-2283), the Netherlands Organization for and therapeutic intervention in IgE responses by biolistic DNA vaccination Scientific Research (NWO Vici 918.66.615 to GJA and NWO 912-02-34). primarily targeting dendritic cells. J Allergy Clin Immunol 2006; 117: 196 --203. 22 Geijtenbeek TB, Torensma R, van Vliet SJ, van Duijnhoven GC, Adema GJ, van Kooyk Y et al. Identification of DC-SIGN, a novel dendritic cell-specific ICAM-3 receptor that supports primary immune responses. Cell 2000; 100: REFERENCES 575 --585. 1 Steinman RM, Hemmi H. Dendritic cells: translating innate to adaptive immunity. 23 Hartgers FC, Vissers JL, Looman MW, van Zoelen C, Huffine C, Figdor CG et al. DC- Curr Top Microbiol Immunol 2006; 311:17--58. STAMP, a novel multimembrane-spanning molecule preferentially expressed by 2 Akbari O, Panjwani N, Garcia S, Tascon R, Lowrie D, Stockinger B. DNA vaccination: dendritic cells. Eur J Immunol 2000; 30: 3585 --3590. transfection and activation of dendritic cells as key events for immunity. J Exp Med 24 Valladeau J, Duvert-Frances V, Pin JJ, Dezutter-Dambuyant C, Vincent C, 1999; 189: 169 --178. Massacrier C et al. The monoclonal antibody DCGM4 recognizes Langerin, a 3 Porgador A, Irvine KR, Iwasaki A, Barber BH, Restifo NP, Germain RN. Predominant protein specific of Langerhans cells, and is rapidly internalized from the cell role for directly transfected dendritic cells in antigen presentation to CD8+ T cells surface. Eur J Immunol 1999; 29: 2695 --2704. after gene gun immunization. J Exp Med 1998; 188: 1075 --1082. 25 Winzler C, Rovere P, Rescigno M, Granucci F, Penna G, Adorini L et al. Maturation 4 Donnelly RF, Raj Singh TR, Woolfson AD. Microneedle-based drug delivery stages of mouse dendritic cells in growth factor-dependent long-term cultures. systems: microfabrication, drug delivery, and safety. Drug Deliv 2010; 17: J Exp Med 1997; 185: 317 --328. 187 --207. 26 Karttunen J, Sanderson S, Shastri N. Detection of rare antigen-presenting cells by 5 Mahe B, Vogt A, Liard C, Duffy D, Abadie V, Bonduelle O et al. the lacZ T-cell activation assay suggests an expression cloning strategy for T-cell Nanoparticle-based targeting of vaccine compounds to skin antigen-presenting antigens. Proc Natl AcadSci USA 1992; 89: 6020 --6024. cells by hair follicles and their transport in mice. J Invest Dermatol 2009; 129: 27 Falo Jr LD, Kovacsovics-Bankowski M, Thompson K, Rock KL. Targeting antigen 1156 --1164. into the phagocytic pathway in vivo induces protective tumour immunity. Nat 6 Mir LM. Nucleic acids electrotransfer-based gene therapy (electrogenetherapy): Med 1995; 1: 649 --653. past, current, and future. Mol Biotechnol 2009; 43: 167 --176. 28 Casatorres J, Navarro JM, Blessing M, Jorcano JL. Analysis of the control of 7 Prow TW, Chen X, Prow NA, Fernando GJ, Tan CS, Raphael AP et al. Nanopatch- expression and tissue specificity of the keratin 5 gene, characteristic of basal targeted skin vaccination against West Nile Virus and Chikungunya virus in mice. keratinocytes. Fundamental role of an AP-1 element. J Biol Chem 1994; 269: Small (Weinheim an der Bergstrasse, Germany) 2010; 6: 1776 --1784. 20489 --20496. 8 Singh TR, Garland MJ, Cassidy CM, Migalska K, Demir YK, Abdelghany S et al. 29 Bins AD, Jorritsma A, Wolkers MC, Hung CF, Wu TC, Schumacher TN et al. A rapid Microporation techniques for enhanced delivery of therapeutic agents. Recent Pat and potent DNA vaccination strategy defined by in vivo monitoring of antigen Drug Deliv Formul 2010; 4: 1 --17. expression. Nat Med 2005; 11: 899 --904. 9 Vandermeulen G, Daugimont L, Richiardi H, Vanderhaeghen ML, Lecouturier N, 30 Toes RE, Schoenberger SP, van der Voort EI, Offringa R, Melief CJ. CD40- Ucakar B et al. Effect of tape stripping and adjuvants on immune response after CD40Ligand interactions and their role in cytotoxic T lymphocyte priming and intradermal DNA electroporation. Pharm Res 2009; 26: 1745 --1751. anti-tumor immunity. Semin Immunol 1998; 10: 443 --448. 10 Babiuk S, Baca-Estrada ME, Foldvari M, Baizer L, Stout R, Storms M et al. Needle- 31 van den Berg JH, Nujien B, Beijnen JH, Vincent A, van Tinteren H, Kluge J et al. free topical electroporation improves gene expression from plasmids adminis- Optimization of intradermal vaccination by DNA tattooing in human skin. Hum tered in porcine skin. Mol Ther 2003; 8: 992 --998. Gene Ther 2009; 20: 181 --189. 11 Bennett AM, Phillpotts RJ, Perkins SD, Jacobs SC, Williamson ED. Gene gun 32 Schreurs MW, Eggert AA, de Boer AJ, Vissers JL, van Hall T, Offringa R et al. mediated vaccination is superior to manual delivery for immunisation with DNA Dendritic cells break tolerance and induce protective immunity against a vaccines expressing protective antigens from Yersinia pestis or Venezuelan melanocyte differentiation antigen in an autologous melanoma model. Cancer Equine Encephalitis virus. Vaccine 1999; 18: 588 --596. Res 2000; 60: 6995 --7001. 12 Dincer Z, Jones S, Haworth R. Preclinical safety assessment of a DNA vaccine using 33 Lauterbach H, Gruber A, Ried C, Cheminay C, Brocker T. Insufficient APC capacities particle-mediated epidermal delivery in domestic pig, minipig and mouse. Exp of dendritic cells in gene gun-mediated DNA vaccination. J Immunol 2006; 176: Toxicol Pathol 2006; 57: 351 --357. 4600 --4607. 13 Dresch C, Edelmann SL, Marconi P, Brocker T. Lentiviral-mediated transcriptional 34 Cho JH, Youn JW, Sung YC. Cross-priming as a predominant mechanism for targeting of dendritic cells for induction of T cell tolerance in vivo. J Immunol inducing CD8(+) T cell responses in gene gun DNA immunization. J Immunol 2008; 181: 4495 --4506. 2001; 167: 5549 --5557. 14 Geusens B, Strobbe T, Bracke S, Dynoodt P, Sanders N, Gele MV et al. Lipid- 35 Bros M, Ross XL, Pautz A, Reske-Kunz AB, Ross R. The human fascin gene promoter mediated gene delivery to the skin. Eur J Pharm Sci 2011; 43: 199 --211. is highly active in mature dendritic cells due to a stage-specific enhancer. 15 Peachman KK, Rao M, Alving CR. Immunization with DNA through the skin. J Immunol 2003; 171: 1825 --1834. Methods (San Diego, Calif) 2003; 31: 232 --242. 36 Rea D, Havenga MJ, van Den Assem M, Sutmuller RP, Lemckert A, Hoeben RC et al. 16 Steele KE, Stabler K, VanderZanden L. Cutaneous DNA vaccination against Ebola Highly efficient transduction of human monocyte-derived dendritic cells with virus by particle bombardment: histopathology and alteration of CD3-positive subgroup B fiber-modified adenovirus vectors enhances transgene-encoded dendritic epidermal cells. Vet Pathol 2001; 38: 203 --215. antigen presentation to cytotoxic T cells. J Immunol 2001; 166: 5236 --5244. 17 Sudowe S, Ludwig-Portugall I, Montermann E, Ross R, Reske-Kunz AB. Transcrip- 37 Coulie PG, Karanikas V, Colau D, Lurquin C, Landry C, Marchand M et al. A tional targeting of dendritic cells in gene gun-mediated DNA immunization favors monoclonal cytolytic T-lymphocyte response observed in a melanoma patient the induction of type 1 immune responses. Mol Ther 2003; 8: 567 --575. vaccinated with a tumor-specific antigenic peptide encoded by gene MAGE-3. 18 van den Berg JH, Nuijen B, Beijnen JH, Vincent A, van Tinteren H, Kluge J et al. Proc Natl Acad Sci USA 2001; 98: 10290 --10295. Optimization of intradermal vaccination by DNA tattooing in human skin. Hum 38 Pokorna D, Rubio I, Muller M. DNA-vaccination via tattooing induces stronger Gene Ther 2009; 20: 181 --189. humoral and cellular immune responses than intramuscular delivery supported 19 Bonkobara M, Zukas PK, Shikano S, Nakamura S, Cruz Jr PD, Ariizumi K. Epidermal by molecular adjuvants. Genet Vaccines Ther 2008; 6:4. Langerhans cell-targeted gene expression by a dectin-2 promoter. J Immunol 39 Ramshaw IA, Ramsay AJ. The prime-boost strategy: exciting prospects for 2001; 167: 6893 --6900. improved vaccination. Immunol Today 2000; 21: 163 --165.

& 2012 Macmillan Publishers Limited Cancer Gene Therapy (2012), 303 --311