Intradermal Delivery of Adenoviral Type-35 Vectors Leads to High Efficiency Transduction of Mature, CD8⁺ T Cell-Stimulating Skin-Emigrated Dendritic Cells

Dendritic cells (DC)³ are powerful APCs with the unique ability to prime naive T cells and thus initiate immune responses. They provide the link between the innate and the adaptive immune response, and as such regulate adaptive immunity at its most fundamental level (1). By way of vaccination, DC may be genetically modified to express any Ag of choice (2). By virtue of the intracellular expression of the Ag of interest, genetic vaccination ensures its processing via endogenous pathways for subsequent MHC class I-mediated presentation to CD8⁺ CTL, which are essential for an effective cell-mediated response to both viruses and tumors (3, 4, 5). The efficacy of in vivo genetic immunization has been shown to depend on DC-mediated immune activation (6, 7, 8). Massive uptake of the injected vectors by tissue-resident cells other than DC might interfere with DC transduction efficiency and subsequent effector T cell activation. It is therefore desirable to use vectors with the ability to transduce DC with high efficiency in the context of a tissue environment.

Recombinant adenovirus (Ad) type 35 (rAd35) is an attractive genetic vaccine carrier with the advantage of low pre-existing immunity in human populations, in contrast to the more commonly used rAd5 vector (9). In vitro-generated human monocyte-derived DC (MoDC) express the primary docking receptor for Ad35 (CD46), but not the receptor for Ad5. As a result, MoDC are more efficiently transduced by Ad35-based vectors than by rAd5, without interfering with the DC’s activation potential (10). Indeed, MoDC infected with Ad5F35 chimeric vectors have been used successfully to generate antitumor CD8⁺ T cell responses in vitro (11).

Unfortunately, studies in nonhuman primates could not demonstrate enhanced immunogenicity of Ad5F35 over Ad5 vectors in vivo, despite an improved in vitro transducibility of DC (12). Neither could the in vivo application of Ad5F35 circumvent pre-existent immunity to Ad5 in a mouse model (12). The latter observation can be remedied by the use of Ad35 rather than Ad5F35, thus abolishing all structural Ad5 capsid components to which immunity might exist (13, 14), while the first observation might be explained by the i.m. route of administration used, which is marked by a paucity of tissue-resident DC; the skin might be a more favorable vaccination site in this respect. Interpretation of in vivo human CD46-transgenic murine and nonhuman primate data is further complicated by the presence of CD46 on erythrocytes, which, unlike in humans, might seriously affect biodistribution of the rAd35 vector (15, 16). In conclusion, dissimilarities in the high-affinity CD46 receptor expression and anatomical barriers, which could result in differences in biodistribution characteristics, may render available preclinical in vivo models for rAd35 unrepresentative for the human situation. The exploration of alternative preclinical human models to test the dynamics of in vivo rAd35 infection is therefore warranted.

In view of these considerations, we have opted to introduce both Ad5F35 and Ad35 vectors in a near-physiological human skin explant model to explore their in vivo DC-targeting potential. Organotypic skin explant cultures allow for the study of human cutaneous DC and their functions in their natural complex tissue environment, closely maintaining the in vivo situation (17). We previously used this skin explant model successfully to assess the in situ DC infectivity potential of CD40-targeted rAd viruses and reported a high efficiency DC transduction and accompanying maturation induction (18). Our current data show that intracutaneous administration of rAd35 facilitates high efficiency transduction of mature skin-emigrated DC with an ability to specifically activate CTL, thus confirming the utility of Ad35 as an in vivo vaccine delivery vehicle.

E1/E3-deleted, replication-incompetent rAd5, rAd5F35, or rAd35 vectors were generated in PER.C6 or PER.C6/55K cells using pBR322-based adaptor plasmids pAdApt or pAdApt535 together with cosmids pWE.Ad.AflII-rITR.dE3, pWE.Ad35.pIX-rITR.dE3, or pWE.Ad.AflII-rITR/Fib35, respectively, as previously described (19, 20). GFP or the circumsporozoite (CS) protein of Plasmodium falciparum (Pf CS) was cloned into the adaptor plasmids under control of an immediate/early CMV promoter and an SV40 polyadenylation signal. These plasmids were linearized and transfected into PER.C6 or PER.C6/55K cells together with the linearized cosmids using lipofectamine (Invitrogen Life Technologies). Homologous recombination led to the generation of rAd5, rAd5Fib35, or rAd35 vectors containing the enhanced GFP or CS transgene. Generated vectors were analyzed for transgene expression, amplified in 24–48 triple-layer T175-cm² flasks, purified by double CsCl gradient ultracentrifugation, and dialyzed into PBS containing 5% sucrose. Purified rAd vectors were stored at −80°C. Virus particle (vp) titers were determined by HPLC, whereas IU were assessed by PFU. Equal amounts of vp/IU ratios (<30) were found with assays meeting criteria defined by Food and Drug Administration guidelines. All viruses were injected into skin explants in the same volume of 10 μl.

A detailed description of Ad infection and DC culture in the skin explant system has been published previously (18). Human skin specimens were obtained after informed consent from patients undergoing corrective breast or abdominal plastic surgery, following hosptital guidelines. Skin biopsies (6 mm) were intradermally (i.d.) injected with plain IMDM (Invitrogen Life Technologies) or with a mixture of 100 ng of GM-CSF (Schering-Plough) and 1000 IU of IL-4 (Centraal Laboratorium van de Bloedtransfusiedienst), cultured on rafts according to previously described methods (18, 21), and subsequently i.d. injected with 10⁹ vp per biopsy, or with plain medium as a negative control. Ad vectors were injected into the dermis in a total volume of 10 μl. Following injection, the explants (12–20 samples/condition) were cultured, floating freely on medium containing 5% HPS (Centraal Laboratorium van de Bloedtransfusiedienst), with their epidermal side up, before their removal 2 days later. The explants were discarded, and the medium, containing migrated cells, was harvested and pooled per test condition. Samples of the skin explant-conditioned medium were stored at −20°C for cytokine analysis. Cytokine content of the explant-conditioned medium was analyzed using the cytokine bead array (BD Biosciences) for the simultaneous flow cytometric detection of IL-10, IL-12p70, IL-6, IL-8, TNF-α, and IL-1β, according to the manufacturer’s instructions and using cytokine bead array analysis software (BD Biosciences). Absolute numbers of migrated DC were counted in hemocytometers using trypan blue exclusion, after which they were used for flow cytometric or functional analyses.

Fresh 6-mm biopsies (10 per condition) were immediately processed and placed in 10-cm-diameter culture dishes containing 15 ml of 0.05% trypsin (Invitrogen Life Technologies) for 4–5 h at 37°C, 5% CO₂. The epidermis and dermis were separated with tweezers and washed with IMDM 10% FCS, and single-cell suspensions were made of each by pushing through 100-μm-pore nylon cell strainers (Falcon; BD Biosciences) with the plunger of a 2-ml syringe. The cell suspensions were resuspended in 5 ml of IMDM and counted before flow cytometric analysis.

Cells were incubated on ice for 30 min in PBS with 0.1% BSA and 0.01% NaN₃, in the presence of appropriate dilutions of FITC- or PE-labeled mouse mAbs to CD83 (Beckman-Coulter), CD1a, CD86 (BD Pharmingen), CD46, or CD80 (BD Biosciences). A second incubation step was performed for the unconjugated mAb against coxsackievirus and Ad receptor (CAR) (RmcB (22)) with FITC-labeled goat anti-mouse Abs (Centraal Laboratorium van de Bloedtransfusiedienst). The cells were subsequently analyzed, using a FACSCalibur and CellQuest FACS analysis software (BD Biosciences).

The used HLA-A2-restricted CD8⁺ CTL clone, specifically recognizing the aa 327–335 epitope of Pf CS, has been described previously (20), and was provided by G. Corradin (University of Lausanne, Lausanne, Switzerland) and J. Lopez (Queensland Institute of Medical Research, Brisbane, Australia). For the readout of T cell activation, 96-well nitrocellulose plates (Multiscreen-HA; Millipore) were precoated with the anti-human IFN-γ mAb 1-D1K (15 μg/ml in filtered PBS; Mabtech). After overnight incubation at 4°C, the plates were washed and blocked for 1 h at 37°C with culture medium supplemented with 10% FBS. In situ transduced HLA-A2⁺ skin-emigrated DC were cocultured in the presence of the Pf CS 327–335-specific CTL clone (tested ratios 1:2 and 1:1) for 16 h at 37°C/10% CO₂. Plates were washed with PBS/0.05% Tween 20, followed by a 2- to 4-h incubation at room temperature with the biotinylated anti-human IFN-γ mAb 7-B6-1 (1 μg/ml in filtered PBS; Mabtech). After washing with PBS/0.05% Tween 20, plates were incubated for 1 h at room temperature with 1/2000 diluted extravidin-alkaline phosphatase conjugate (Sigma-Aldrich). Spots were developed with 5-bromo-4-chloro-3-indolyl phosphate/NBT substrate (Sigma-Aldrich) and quantitated with the aid of an Automated ELISA-Spot Assay Video Analysis System (A.EL.VIS).

Transduction efficiencies, transgene expression levels, absolute numbers of (transduced) DC, and cytokine release levels were compared after infection with Ad5 or Ad35-based vectors, using paired Student’s t test (two-sided); differences were considered significant when p < 0.05.

A first prerequisite for successful infection of DC by Ad vectors is the expression of the relevant Ad receptor on the DC surface. CD46 has been identified as a primary docking receptor for Ad35. We therefore determined the expression of CD46 on CD1a⁺ cutaneous DC by flow cytometric analysis, both premigration (intracutaneous DC at day 0) and postmigration (skin-emigrated DC at day 2); see Fig. 1,A. In single-cell suspensions derived from uncultured skin explants (Fig. 1,A, day 0), CD1a⁺ DC were found to be immature (evidenced by a lack of CD83 expression) and to display barely detectable membrane expression of the Ad35 receptor CD46. In contrast, CD1a⁺ DC that had emigrated over the course of a 2-day culture period from i.d. GM-CSF- and IL-4-injected skin explants displayed a mature CD83⁺ phenotype and high membrane expression levels of CD46 (Fig. 1,A, day 2). Skin-emigrated DC were gated by their characteristic side scatter and forward scatter properties, respectively, as described (see Fig. 1,A). We previously reported the maturing effect of premigration i.d. injection of GM-CSF and/or IL-4 on cutaneous DC postmigration, resulting in higher levels of the classic DC maturation marker CD83 and costimulatory molecules such as CD80 and CD86 (18). Although DC migrated from explants i.d. injected with control medium already expressed CD46 (in contrast to skin DC in skin explants before culture; Fig. 1,A), CD46 was further up-regulated through GM-CSF- and IL-4-induced maturation of skin-emigrated DC (see Fig. 1,B). Thus, the level of membrane expression of CD46 correlated to the DC maturation state. In contrast, the Ad5 receptor CAR was not expressed on skin-emigrated DC, independent of their maturation state (Fig. 1 B).

When either Ad5, Ad35, or Ad5F35 (i.e., an Ad5 vector equipped with an Ad35-derived fiber) vectors encoding the GFP reporter gene were i.d. injected into skin explants (at 10⁹ vp per explant) at day 0 of culture and emigrated DC were harvested 2 days later (at which time migration was complete) and tested for GFP expression by FACS, low transduction efficiencies (<20%) were observed without any significant differences between the vectors used (data not shown). This may be explained by absent or low expression of CAR and CD46 on resting CD1a⁺ DC in fresh skin explants (Fig. 1,A). In combination with the relatively high expression of both receptors on other skin-resident cells such as fibroblasts and keratinocytes (shown for CD46 on the CD1a-negative population in Fig. 1,A), this could hamper efficient DC transduction. Coinjection of GM-CSF and IL-4 at day 0, to increase DC maturation and concomitant CD46 expression, could not overcome this. In contrast, when explants were either injected with medium or GM-CSF and IL-4 and then cultured on nitrocellulose filter rafts at the medium/air interface (as described previously (18)) for 24 h before injection of the Ad vectors (at day 1), higher transduction efficiencies of explant-emigrated DC were observed 2 days later (Fig. 2,A). Importantly, the latter regimen resulted in significantly higher DC transduction efficiencies by Ad35 and Ad5F35 as compared with Ad5 (a 2- to 3-fold increase) with particularly high transduction efficiencies of mature CD83⁺ DC (∼70%; see Fig. 2,B). These data indicate that the expression of CD46 on migrating mature DC at the time of Ad35 or Ad5F35 encounter is essential for high efficiency transduction to occur in the context of the skin microenvironment. Ad injections 24 h subsequent to the start of explant culture were therefore used throughout the remainder of this study. GFP transgene expression levels detected by the use of this methodology were also significantly higher after Ad35-mediated transduction as compared with Ad5-mediated transduction (see Fig. 2,B). These results demonstrate that transduction with Ad35 is equivalent to Ad5F35 (Fig. 2), which, in light of the low seroprevalence of Ad35 (9), makes Ad35 the preferred vector for in vivo use. Consequently, all subsequent experiments were performed with Ad35 only.

To ascertain the in situ DC-targeting properties of Ad35 relative to Ad5, skin explants were harvested 24 h after Ad-GFP injection and single-cell suspensions were made of the separated epidermal and dermal layers. GFP expression was subsequently determined in relation to CD1a to quantitate the in situ transduction efficiency of CD1a⁺ DC relative to the total dermal and epidermal populations (shown for a dermal fraction in Fig. 3,A). GFP expression was mostly detected in the dermal fractions (at the site of Ad injection), consistent with a barrier function of the basal membrane of the epidermis interfering with Ad passage into the epidermal layer (data not shown). From these experiments, it became clear that Ad35 infected both the dermal population as a whole and CD1a⁺ DC at higher efficiencies than Ad5 (Fig. 3, A and B). Clearly, the presence of other skin-resident cells that were infected by Ad35 did not interfere with the improved in situ infection of DC by Ad35 over Ad5. Of note, the Ad35-mediated i.d. transduction efficiency of as yet unmigrated CD1a⁺ DC was much lower (up to 15%) than the transduction efficiency of the mature CD83⁺ DC that eventually ended up migrating from the skin 24 h later (up to 70%; see Fig. 3 C).

Transduction efficiencies among DC, migrated over the course of 2 days from Ad-injected skin explants, remained constant for at least another 5 days up to day 7 after the i.d. administration of the Ad vectors (with maintained higher efficiencies for Ad35 as compared with Ad5; see Fig. 4,A). This held true both for medium-preinjected explants and for explants injected with the DC-stimulatory cytokines GM-CSF and IL-4, 24 h before Ad injection. Although this continued transgene expression might be expected to result in long-term in vivo stimulation of specific T cells, the phenotypic development of the skin-emigrated DC differed notably between these two conditions with possible functional consequences. Whereas at day 2 a majority of the migrated DC were in a mature state irrespective of medium or cytokine injection (determined by CD83 expression; see Fig. 3,C), most DC migrated from medium-conditioned explants had lost CD83 expression by day 7 and also displayed lower levels of the costimulatory molecules CD80 and CD86 (see Fig. 4,B). This is in keeping with our previous observations in this model (18). Administration of Ad5 or Ad35 caused only slight and nonsignificant increases in the expression of these DC activation markers at day 7 (Fig. 4 B). In contrast, a majority of DC from GM-CSF- and IL-4-conditioned explants showed stable maturation, evidenced by a conserved expression of CD83, CD80, and CD86 at day 7, irrespective of Ad5 or Ad35 infection. In all conditions, the DC activation markers were uniformly expressed, irrespective of GFP expression. In conclusion, conditioning of Ad vaccination sites with the DC-stimulating cytokines GM-CSF and IL-4 may ensure a stable mature T cell-stimulatory phenotype to accompany long-term transgene expression in Ad35-transduced DC after their migration from the skin. Ad35 infection does not interfere with this intracutaneous cytokine-induced prolongation of DC maturation.

In order for Ad-transduced DC to generate effective T cell-mediated immunity against viruses or tumors, it is important that sufficient numbers of mature transgene-expressing DC leave the vaccination site and migrate to draining lymph nodes to activate specific CTL. As shown in Fig. 5,A for GM-CSF- and IL-4-conditioned skin explants, DC migrated in smaller absolute numbers subsequent to i.d. administration of Ad35 vectors as compared with Ad5 or medium controls. However, due to the high Ad35-mediated transduction efficienciy of CD83⁺ mature DC, the total number of DC with a mature CD83⁺ T cell-stimulatory phenotype that also expressed the GFP transgene was significantly higher after Ad35 administration as compared with Ad5 administration (p < 0.01; Fig. 5 A). This mature CD83⁺ phenotype of skin-emigrated DC was further typified by the expression of CCR7, indicative of the ability of these DC to home to the paracortical T cell areas in the draining lymph nodes (21).

The observed decrease in absolute numbers of migrated DC did not appear to be due to an Ad35-induced decrease in the release levels of the migration-mediating cytokines IL-1β or TNF-α in the skin environment, as detected by flow cytometric cytokine bead array in explant-conditioned medium, collected 24 h after Ad injection (data not shown). Although rAd35 did not influence the (generally low) release levels of the type-1 T cell-stimulatory cytokine IL-12p70 in explant-conditioned medium, it did lead to a significant decrease in the release of the immunosuppressive cytokine IL-10 (Fig. 5 B). The resulting rise in IL-12:IL-10 ratio suggests a conditioning of the microenvironment that might be more conducive to the generation of cell-mediated immunity.

Using a CTL clone that specifically recognized an HLA-A2-restricted epitope derived from Pf CS aa 327–335 (20), we determined the ability of skin-emigrated DC from HLA-A2⁺ skin explants, i.d. injected with either Pf CS-expressing Ad5 or Ad35 vectors, to present the relevant epitope and activate CD8⁺ T cells (readout by IFN-γ ELISPOT assay). As shown in Fig. 5 C for two separate HLA-A2⁺ skin donors, equivalent or slightly higher numbers of Pf CS-specific CD8⁺ T cells were activated in response to relatively low numbers of DC emigrated from explants injected with Pf CS-encoding Ad35 as compared with an equivalent number of DC migrated from Ad5-injected explants. Importantly, these data demonstrate the ability of skin-emigrated DC, transduced in situ by Ad35, to properly process and present epitopes from the transgene product for subsequent T cell activation.

Their central role in the generation of both cellular and humoral immunity has made DC the major target of most novel vaccination strategies (5). DC-based genetic vaccination may allow for high-level intracellular expression of the Ag target gene of interest over long periods of time, thus optimizing the chances of in vivo effector CTL activation by the transduced DC (3). Particularly attractive in this regard would be the use of vectors with the ability to target and transduce DC with high efficiency in vivo. Such vectors would obviate the need for the costly and laborious ex vivo generation and Ag loading of autologous DC. Their ability to infect DC at high efficiency through binding to CD46 on the DC surface make subgroup B Ad vectors, including rAd35, interesting candidate vehicles for in vivo DC-targeted vaccination (10). The more commonly used rAd5 vectors, belonging to the subgroup C, are relatively poor human DC transducers due to the absence of their primary docking receptor CAR on the DC surface (10, 23). In contrast, chimeric rAd5F35 vectors, consisting of the Ad5 capsid and the Ad35 fiber shaft and knob, were shown to transduce MoDC with high efficiency (10).

Three previously reported observations have suggested a particular suitability of the rAd35-based vectors for the in vivo transduction of DC for vaccination purposes: 1) Extensive serological screening has demonstrated the absence of pre-existent immunity to several members of the Ad subgroup B family (including Ad35) in the general human population (9). This is in stark contrast to Ad5 and may allow repeated in vivo use of Ad35-based vaccines before limiting immunity arises, which might interfere with infection efficiency and longevity of the transduced DC (13). 2) rAd5F35 vectors transduce DC in vitro with conserved high efficiency in the presence of primary skin-derived fibroblasts (10). 3) rAd5F35-transduced DC are more potent in vitro CTL activators than rAd5-transduced DC, as shown for melanoma-associated transgene products by the activation of a gp100-specific CTL clone and the induction of CTL-recognized Ag on melanoma-specific and tumor-reactive CD8⁺ T cells (11).

In aggregate, the above listed data clearly indicate the possible utility of Ad35-based vectors for in vivo DC-targeted vaccination approaches. However, despite an improved DC tropism in vitro, in nonhuman primates rAd5F35 vectors were recently shown to be less immunogenic in vivo than rAd5 vectors (12). This may be explained by the used route of administration in that particular instance, because i.m. genetic vaccination is characterized by the preferential transduction of stromal cells rather than DC, which are in short supply in muscle tissues (24). In this respect, i.d. injection might be more effective, as is indeed borne out by our present findings. A relatively dense network of DC lines the skin (17) and, as shown in the current study, i.d. injection of rAd35, preceded by DC maturation induction (and concomitant CD46 up-regulation), results in high efficiency in situ infection of these cutaneous DC. Importantly, in comparison with rAd5, a higher rAd35-mediated transduction efficiency was observed among phenotypically mature DC that subsequently migrated from the skin explants. Although we did not formally demonstrate whether these DC were originally derived from epidermal Langerhans cells (LC) or from dermal DC, expression of CD1a by most of the DC and Langerin by a subpopulation with high CD1a expresson levels (data not shown) suggests most of them to be of LC origins (17, 25). This is of particular interest in view of a recent study by Rozis et al. (26), who showed that subgroup B Ad viruses preferentially infect LC and suggested that the skin might be an appropriate site for the administration of Ad35-based vaccines that target LC. Besides CD1a⁺ DC, CD1a⁻ cells also migrated from the skin explants. We recently reported a subpopulation of CD1a⁻CD83⁻ cells within the DC gate to exhibit macrophage-like characteristics with both CD14 and strong diffuse CD68 expression (27). These cells expressed low levels of costimulatory molecules and failed to stimulate T cells. Of note, we found their number to increase over time subsequent to migration from the skin, unless a strong DC-maturation signal was provided before migration, such as injection of GM-CSF and IL-4. This finding was confirmed in the present study, with maintained expression of CD83, CD80, and CD86 levels following coinjection of GM-CSF and IL-4 together with the used Ad vectors, up to 7 days subsequent to skin emigration. Of note, we previously demonstrated this maintained CD83⁺ mature state of the skin-emigrated DC to be accompanied by the expression of CCR7, indicative of their ability to migrate to the T cell areas in the skin-draining lymph nodes (18, 21).

In essence, previous observations pointing to the suitability of Ad35-based vectors for in vivo DC-targeted vaccination approaches were confirmed in our present study of rAd35-mediated DC transduction in a dermal substrate (see below).

1) An important issue in the application of Ad-based vaccines is the presence of pre-existent neutralizing Abs. Whereas pre-existent neutralizing Abs to Ad5 may necessitate high initial rAd5 vaccine doses and may limit the applicability of repeated rAd5 booster vaccinations, the absence of neutralizing Abs to Ad35 in the general population allows for accurate dose control, i.e., one single effective dose of Ad35-based vectors for all vaccinees. Moreover, Ad35-based vaccines may be used for multiple vaccinations before limiting immunity arises. It was demonstrated recently that the most effective neutralizing Ab responses against rAd5 were directed at the hexon proteins on the viral capsid, also contained within the rAd5F35 vector, rather than to the Ad5 fiber knob or penton base, as was originally assumed (14). In this study, we report that rAd35 and rAd5F35 vectors are equally effective in the transduction of cutaneous DC. Besides demonstrating the applicability of rAd35 in i.d. vaccination to be equal to that of rAd5F35, this observation also indicates that the rAd35-associated improved DC transduction efficiency was not due to lower levels of Ad35-neutralizing Abs as compared with Ad5-neutralizing Abs in the pooled human serum that was used in the skin explant cultures. Rather, the improved transduction efficiency of skin-emigrated DC was the result of an intrinsically superior in situ DC infectivity of Ad35-based vectors over rAd5. This was confirmed by a maintained improved transduction efficiency with rAd35 over rAd5 in serum-free skin explant cultures or cultures containing FCS rather than human serum (data not shown).

2) In the skin microenvironment, the docking receptor of Ad35 is not exclusively expressed on DC. In fact, before activation, quiescent CD1a⁺ DC are either negative for CD46 or express it at lower levels than other (stromal) cells in the skin (see Fig. 1,A). This may, at least in part, explain why immediate injection of rAd35 into the dermis did not result in a higher transduction efficiency of subsequently migrated DC as compared with rAd5 injection. In contrast, an increased transduction efficiency of rAd35 (and rAd5F35) over rAd5 became apparent upon i.d. injection of the rAd vectors subsequent to a 24-h culture of the skin explants (i.d. injected with either medium or GM-CSF and IL-4). This is most likely due to an up-regulation of the high-affinity CD46 receptor on the cutaneous DC, which accompanied their culture-induced maturation. Other contributing factors to the improved transduction efficiency may be the increased frequency of migrating and accumulated DC in the dermis at the time of i.d. rAd injection and possibly their more preferential localization in draining lymph vessels at that time, which may render the DC more easily physically accessible to the injected rAd viruses. A preferential infection of mature (CD46⁺) migrating DC would explain the considerably higher transduction efficiency among skin-emigrated DC as compared with dermal CD1a⁺ DC in situ, an observation we also previously made for CD40-targeted rAd viruses (18). Our in situ findings are in keeping with previously published in vitro observations by Rea et al. (10), revealing a preserved high efficiency transduction of DC by Ad35-based vectors, even in the presence of CD46⁺ stromal skin fibroblasts that, like the DC, were also more efficiently transduced by rAd35 than by rAd5 (see Fig. 3). Thus, rAd35 vectors do not truly target DC in the sense that they uniquely infect DC, but also other cell types. Nevertheless, they infect DC with high efficiency in the context of dermal tissue.

3) DC that were i.d. transduced by rAd35 and had subsequently migrated from the skin were able to activate CD8⁺ T cells, specifically recognizing an HLA-A2-restricted epitope contained within the transgene product (in this case the Pf CS protein). This demonstrated that the followed sequence of cutaneous DC maturation, rAd35-mediated infection, and migration allowed for the correct expression, processing, and presentation of the transgene product for subsequent CD8⁺ T cell activation: an absolute requirement for successful vaccination.

In keeping with previous reports, we did not observe rAd35-induced DC maturation (10), nor did we find an up-regulation of the T cell-stimulatory cytokine IL-12p70. In contrast, neither did we find any adverse effects of Ad35 on DC maturation or on the release of such immunostimulatory cytokines as IL-1β, IL-6, or IL-12p70, as was recently reported by Iacobelli-Martinez et al. (28), who found evidence for immunosuppressive effects of rAd35 (but not rAd5) through CD46 binding and subsequent interference with IFN-γ-induced C/EBPβ protein expression. In fact, we observed a significant decrease in the levels of the immunosuppressive cytokine IL-10 in skin explant-conditioned medium upon intracutaneous delivery of rAd35, suggestive of immunopotentiation rather than immunosuppression. Although we did observe a decrease in overall numbers of skin-emigrating DC upon rAd35 injection, the high rAd35-mediated transduction efficiency, as compared with rAd5, ultimately resulted in significantly higher numbers of migrated DC that were both mature and transduced. Indeed, by conditioning of the skin with GM-CSF and IL-4 before rAd35 injection, these numbers were further up-regulated, and the skin-emigrated DC were characterized both by a stable mature phenotype and a maintained transgene expression for up to 7 days after i.d. rAd35 delivery.

In conclusion, our data clearly demonstrate the feasibility of i.d. vaccination with rAd35 and warrant the further development of rAd35 vectors for DC-targeted vaccine delivery in vivo. From the results obtained in our human skin explant model, we propose the following rAd35-based i.d. vaccination scheme: 1) prior conditioning of the vaccination site with DC-stimulatory cytokines, which will ensure sufficient numbers of migrating DC in the dermis as well as sufficient expression levels of CD46 on the DC to facilitate high efficiency in situ DC transduction, followed by 2) the i.d. injection of rAd35 encoding the Ag of interest, which will result in the migration of DC with high transduction efficiencies, a stable mature T cell-stimulatory phenotype, and the ability to activate specific CD8⁺ effector T cells.

The authors have no financial conflict of interest.