Introduction

DNA vaccine technologies offer considerable promise as effective immunization strategies.12 Cutaneous DC (epidermal Langerhans cells (LC) and dermal dendritic cells (DDC)) are potent antigen presenting cells (APC).34 In the presence of the appropriate stimuli (tumor necrosis factor-α (TNF-α), IL-1β, bacterial or viral components (lipopolysacharide, CpG DNA motifs, dsRNA)) cutaneous DC become activated and traffic to draining lymph nodes, where they can present antigenic peptides to naive T lymphocytes.567 Taken together, these features suggest that skin DC may be ideal targets for the delivery of transgenic antigen for the purpose of genetic immunization.

Using murine models our laboratory and others have demonstrated that after gene delivery to the skin, transfected DC are present in draining lymph nodes where they can induce specific immune responses.89 Similar studies addressing the mechanism of genetic immunization in human skin have not yet been reported. There are significant architectural differences between murine and human skin and considerable evidence that after skin transfection, the expression and distribution of transgenic proteins are different in human skin versus mouse skin. For example, in mice, following intradermal injection of naked DNA, the expression of transgenic proteins is evident in the epidermis, dermis, underlying fat tissue and muscle, while transgene expression in human skin is limited to the epidermis.10

To analyze gene expression and the mechanisms of immune stimulation following gene delivery to human skin, we utilized human skin organ cultures. This experimental model of human skin remains viable for several days in culture, maintaining the anatomy and physiology of normal skin.1112 In the present work, we employed the gene gun to transfect human skin organ cultures. Our studies demonstrate that epidermal LC are transfected directly in situ. Biolistic delivery of DNA did not affect the viability, morphology or the immune phenotype of transfected DCs. Interestingly we find that the use of the gene gun induced a rapid mobilization of resident epidermal LC from transfected human skin. Importantly, when skin samples were transduced with a plasmid encoding the human melanoma Ag MART-1, migrating LC efficiently presented the MART1/Melan-A27–35 peptide epitope to a human cytotoxic T cell (CTL) clone that recognizes the transgenic peptide sequence in the context of the MHC-class I HLA-A2 molecule.

Results

Gold particle delivery to human skin and epidermal dendritic cells in situ

To evaluate gold particle delivery to human skin, and particularly to epidermal LC (resident in the basal epidermis), we compared the distribution of gold particles in skin samples transfected with the gene gun using varying helium (He) pressures (400, 600, 800 and 900 psi). The anatomical localization of gold particles was evaluated immediately after delivery in vertical sections of skin and in epidermal sheets. With He pressures of 400 or 600 psi, gold beads reached only the stratum corneum (Figure 1a) whereas with 800 psi, gold particles were distributed uniformly throughout all epidermal strata (Figure 1b). Confocal microscopy analysis of epidermal sheets fixed immediately after shooting and stained with anti-CD1a mAb demonstrated the presence of gold particles in CD1a+ LC, indicating direct particle delivery into epidermal LC (Figure 1c–e). Rare gold particles were also found in the papillary dermis or in the epidermal/dermal junction. He pressures higher than 800 psi resulted in spontaneous detachment of the epidermis.

Figure 1
figure 1

Human skin and epidermal LCs contain gold particles after biolistic delivery of naked DNA. (a) After biolistic gene delivery to complete skin samples using a He discharge pressure of 600 psi, gold particles (arrow) are dectected exclusively in the stratum corneum (H&E, ×400). (b) Using a He pressure of 800 psi, gold particles can be found throughout all epidermal layers (arrow) (H&E, ×400). (c) Is a confocal microscopy image showing various 1 μm gold beads (arrows) in the basal side of an epidermal sheet which colocalize with a CD1a+ LC in d and e. Gold particles shown in c, (arrows) are present inside the cytoplasm of the LC shown in d (optical thicknes = 0.4 μm) (×600) (scale bar, 5 μm).

Full size image

We next evaluated the viability, morphology and phenotype of DC migrating from gene gun transfected epidermal/dermal explants. Migratory DC were collected 72 h after biolistic delivery of particles loaded with a plasmid encoding for firefly luciferase (pLuciferase) or plasmid backbone (pBackbone), and from untreated controls. The viability of migratory DC from transfected or control explants was ⩾85% according to PI exclusion detected by flow cytometry. Migratory DC from transfected explants contained gold particles when analyzed in cytospins (Figure 2a). Migratory DC from treated skin did not exhibit ultrastructural changes versus control DC. They showed typical membrane processes, abundant cytoplasm, numerous lysosomes and mitochondrias, and lack of Birbeck granules.11 Most of the gold particles were detected free in the cytoplasm of DC, consistent with direct particle cell delivery (Figure 2b). Twenty-five to 30% of the cells analyzed by TEM contained intracytoplasmic gold particles. Migratory DC from transfected explants exhibited the phenotype of mature skin DC (HLA-DRhigh, CD80high, CD86high and CD1alow) (data not shown).

Figure 2
figure 2

Biolistic delivery of DNA does not affect the morphology of migratory DCs from human epidermal/dermal explants and they contain intracytoplasmic gold particles. (a) Cytospin preparation shows skin migratory DCs containing intracellular gold particles (arrows) (May–Grünwald-Giemsa, ×400) (scale bar, 5 μm). (b) Ultrastructural analysis of skin migratory DCs (arrowheads) from gene gun transfected epidermal/dermal explants, one of them with one intracytosolic 1 μm gold particle (arrow and insert) (TEM, ×9000, insert ×19 000).

Full size image

Gene gun induces a rapid activation and migration of epidermal Langerhans cells

The kinetics of LC migration was evaluated 1, 3, 6, 12, 24 and 48 h after skin biolistic delivery of unloaded or loaded gold particles with either a plasmid encoding for EGFP (pEGFP), or pBackbone. Basal migration of LC from nontreated skin was included as control.

CD1a+ LC were quantified in epidermal sheets by fluorescence microscopy at each time-point and for each condition. The percentage of migrated LC was calculated according to the formula: 100% − (mean of LC number in 12 high power fields (HPF, ×400) at each time-point × 100/mean of LC number in 12 HPF at time 0). LC migration started earlier (1 h) following biolistic delivery, and after 3 h, areas lacking LC were evident in the epidermal sheets (Figure 3). At 6 h, 30 ± 5.6%, 48 ± 3.4% and 56 ± 1.1% of LC migrated out of the epidermis treated with unloaded gold particles, pBackbone, or pEGFP, respectively, whereas the spontaneous LC migration was 10.0 ± 0.4% (P < 0.0001) (Figure 3). Twelve hours after treatment, LC migration from skin treated with unloaded gold particles reached a plateau (49 ± 5.6%), but in those samples treated with pBackbone or with pEGFP the percentage of LC leaving epidermis increased (65.7 ± 3.4% and 69.5 ± 1.1%, respectively). After 24 h, 81 ± 2.3% and 85 ± 3.4% of LC had migrated from samples treated with pBackbone and pEGFP, respectively, whereas the spontaneous migration from nontreated samples remained low (22 ± 2%, P ⩽ 0.0001). At early time-points, (1 to 3 h) there was no significant difference in epidermal LC migration induced by the different treatments (P > 0.05). However, 6, 12, 24 and 48 h after treatment, LC migration induced by pBackbone and by pEGFP was significantly higher than that induced by unloaded gold particles (P < 0.05 and P < 0.001, respectively). To rule out the possibility that the reduction in LC numbers was not due to the down-modulation of the CD1a molecule by mature LC, we also determine the number of HLA-DR+ epidermal LC in epidermal sheets expressing EGFP, 24 h after transfection (Figure 4). A similar percentage of LC migration was detected when LC were quantified by their expression of HLA-DR versus CD1a (not shown).

Figure 3
figure 3

Quantitative analysis of epidermal LC migration from complete skin samples after biolistic delivery of gold particles or naked DNA. The number of CD1a+ LC remaining in the epidermis 1, 3, 6, 12, 24 and 48 h after biolistic delivery was quantified and compared with the total LC population at time 0 and with the basal migration from untreated skin samples. For each different condition, 12 representative fields were analyzed by fluorescence microscopy at ×400. Means ± 1 s.d. of different treatments from five independent experiments were compared by ANOVA and ad hoc Student–Newman Keules tests.

Full size image
Figure 4
figure 4

Most of LC migrated form the epidermis in response to gene gun transfection by the time that EGFP expression can be detected in nonmigratory epidermal cells. (a) Population of resident LC (HLA-DR+, red) in an epidermal sheet immediately after gene delivery (100% of LC at time 0). (b) Shows the population of HLA-DR+ LC remaining in the epidermis, of the same skin sample but 24 h after transfection. (c) The same epidermal sheet as in b shows the expression of EGFP in nonmigratory epidermal cells (mostly keratinocytes) 24 h after transfection (×200) (scale bar, 10 μm).

Full size image

Expression of transgenic proteins in human skin and in Langerhans cells

The efficiency of transfection by gene gun was evaluated in epidermal and dermal sheets obtained 24 h after biolistic delivery of pLuciferase, or pBackbone as a control. As expected, according to the distribution of gold particles present primarily in the epidermis when using a delivery He pressure of 800 psi (Figure 1a), higher levels of luciferase activity were found in epidermal sheets when compared with the expression in the underlying dermis (P < 0.0001) (Figure 5a and b).

Figure 5
figure 5

Analysis of luciferase expression in skin samples and in epidermal LC. (a) Most of the luciferase expression was detected in epidermal sheets (75 000 ± 7200 RLU/μg of protein), while as shown in (b) little expression was present in the dermis (220 ± 29 RLU/μg of protein) (P < 0.0001). (c and d) The expression of two different epidermal cell suspension fractions. (c) Epidermal cells depleted from LC, express up to 65 000 RLU/μg of protein in (d), CD1a+ LC fraction expressed up to 1450 ± 200 RLU/μg of protein, statistically significant when compared with the luciferase expression of LC from control samples shot with the plasmid backbone (P < 0.0001). Means ± 1 s.d. of five independent experiments are represented. Comparisons between luciferase expression by samples transfected with pLuciferase or pBackbone are included as a P value in each individual diagram.

Full size image

To investigate in situ transfection of epidermal LC and to circumvent the problem of their rapid migration out of epidermis after gene gun treatment, gene expression was analyzed in epidermal cell suspensions prepared immediately after gene delivery and cultured for 24 h. LC were enriched through paramagnetic columns by positive selection according to their CD1a expression. Luciferase activity was analyzed in CD1a+ and CD1a− cell fractions. Most of the luciferase expression was found in the CD1a− cell fraction (P < 0.0001). However, CD1a+ LC expressed significantly higher levels of luciferase when compared with control CD1a+ LC obtained from skin transfected with pBackbone (P < 0.0001) (Figure 5c and d). We were unable to demonstrate luciferase activity in migratory DCs from epidermal/dermal explants collected 48 or 72 h following transfection (data not shown).

To rule out the possibility that the luciferase activity detected in CD1a+ cell fraction of epidermal cell suspensions was caused by contaminating keratinocytes, CD1a+ enriched epidermal cells from EGFP transfected skin explants were stained with anti-HLA-DR and evaluated in cytopin preparations by fluorescence microscopy (Figure 6a–d). Cells with dendritic morphology, co-expressing EGFP and HLA-DR were found in cytospins prepared from purified CD1a+ DC, but were few in number (30–50 out of 3 × 104) (four independent experiments, data not shown).

Figure 6
figure 6

Biolistic delivery of naked DNA to human skin results in the expression of EGFP by isolated LC. (a) HLA-DR (red) and (b) EGFP (green), co-localize in a cell with dendritic morphology (d) isolated from transfected human epidermis. (c) Dendrites (arrows) are identified using differential interference contrast microscopy (scale bar, 5 μm) (×1000).

Full size image

Skin dendritic cells can be directly transfected in situ by gene gun

After genetic immunization, DC may be directly transfected and/or they may acquire Ag by internalizing transgenic proteins released from neighboring cells (mostly keratinocytes). While it is unlikely that the luciferase synthesized by keratinocytes and subsequently released and internalized by DC retains enzymatic activity, we directly investigated if under our experimental conditions DC were transfected directly. Epidermal/dermal explants were shot with gold beads loaded with pΔOVA (a plasmid DNA encoding for an intracytoplasmic form of chicken ovoalbumin under the control of the hIE-CMV promoter) or with pK14-ΔOVA (a plasmid where ΔOVA is driven by the keratinocyte specific promoter K-14).13 Seventy-two hours after transfection, the presence of mRNA for ΔOVA was analyzed by RT-PCR in epidermal sheets and in migratory DC (depleted from skin migratory CD3+ T cells by immunomagnetic separation). As expected, transcripts for OVA were found in epidermal sheets after delivery of either pΔOVA or pK14-ΔOVA. By contrast, transcripts for ΔOVA were found only in migratory DC from skin transfected with pCMV-ΔOVA, but not in migratory DCs from skin transfected with pK14-ΔOVA (Figure 7). These results further demonstrate that cutaneous DCs are directly transfected using the gene gun and they are able to transcribe the transgenic DNA. False positive bands, due to the presence of contaminating plasmid DNA in wells from where migratory DC were harvested, were excluded by treating mRNA samples with DNase I previously to RT-PCR (Figure 7).

Figure 7
figure 7

RT-PCR analysis of transgenic transcripts in epidermis and skin migratory DCs demonstrates that skin dendritic cells can be directly transfected using biolistic gene delivery. (a) Transcription products of the non-restricted hIE-CMV promoter driven ΔOVA transgene are present in migratory DCs from epidermal/dermal explants. Transcripts from K14 promoter driving the expression of the ΔOVA transgene can be detected in transfected epidermal sheets (Ep). However, they were not be detected in the population of migratory DC harvested from the same epidermal dermal explant (Mig. DC) ruling out the possibility of contaminating keratinocytes among DCs. (b) The transcription of the human β-actin is shown as positive control of the RT-PCR reactions. Controls, no template in PCR (−T), no RNA in RT reaction (−RNA RT) and the direct PCR of DnaseI treated samples (−RT/PCR) did not show any band in either case when ΔOVA or β-actin-specific primers were used.

Full size image

Langerhans cells transfected in situ with a plasmid encoding for the human melanoma antigen MART-1/ Melan-A stimulate specific CTLs

To evaluate the ability of LC to present transgenic Ag derived peptides to CD8+ T cells, the pCI-MART-1 construct encoding for the HLA-A2 restricted melanoma Ag MART-1/Melan-A was delivered into HLA-A2 positive skin by gene gun. Epidermal cell suspensions were cultured for 24 h and enriched CD1a+ LC were co-cultured with the HLA-A2 restricted CTL line M9.2 specific for the MART1-Melan-A27–35 peptide epitope (LC/T cell ratio = 1/20). Twenty-four hours later, the capability of transfected CD1a+ LC to induce specific T cell activation was assessed by measuring the levels of INF-γ released by the MART-1/Melan-A specific CTLs. The pCI-MART1 transfected LC stimulated significantly higher levels of INF-γ release than pCI backbone transfected controls (P = 0.0001) (Figure 8). Neither transfected LC nor T cells alone produced significant amounts of INF-γ (Figure 8). Addition of the HLA-A2 blocking mAb BB7.2 significantly reduced INF-γ production (P = 0.0001), supporting the HLA-A2-restricted nature of CTL peptide recognition from transfected DC (Figure 8). Together, these results demonstrate that LC isolated from transfected human skin are fully capable of processing and presenting transgenic Ag to MHC class I restricted T cells.

Figure 8
figure 8

LC isolated from pMART-1 transfected human epidermis present antigen to MART-1/Melan A27–35 specific CTLs. CD1a+ cells obtained from human skin after delivery of pCI-MART1, or an identical quantity of pBackbone, were co-cultured with an HLA-A2-restricted CTL. The capability of transfected LC to induce specific T cell activation assessed by measuring the levels of INF-γ released by the CTL is shown. Additional controls included incubation of CD1a+ cells, CTL alone or transfected LC incubated with the blocking HLA-A2 mAb BB7.1. The pCI-MART1 transfected LC stimulated significantly higher levels of INF-γ release than pCI backbone transfected controls (P = 0.0001). INF-γ release was significantly reduced by the addition of BB7.1 mAb (P = 0.0001). Means ± 1 s.d. from five independent studies are shown.

Full size image

Discussion

The skin immune system offers a unique potential as a target for the purpose of genetic immunization.14151617 The development of strategies for direct delivery of Ag to DC in vivo, and particularly to accessible DC resident in human skin, is a rational and attractive approach for vaccine design.

Biolistic gene delivery has been shown to combine the safety advantage of using naked plasmid DNA with the possibility of direct physical delivery of DNA into the cytoplasm or the nuclei of living cells. Such intracellular delivery of cDNA could circumvent the potential obstacles of in vivo gene delivery, such as DNA degradation, or the requirement for cellular internalization of exogenous DNA.18 Recently the gene gun has been successfully used for the treatment of tumors or infectious diseases in mice and in nonhuman primate models.891920

To evaluate gene delivery to dendritic cells in human skin, we adapted a human skin organ culture system that has been shown to be a representative model for the expression of exogenous genes delivered to human skin in vivo.10 Importantly these complete human skin cultures remain viable and maintain the anatomical architecture of living skin for several days.1112 This is particularly relevant for genetic immunization studies, as the mechanisms of immunization are likely to depend not only on the type of cell transfected and the quantity of gene expressed, but also on the unique anatomical relationship of interacting cells in the cutaneous immune system present in intact skin.1617

In the present work, using gene gun as a transfection method, we studied the expression of transgenic proteins in human skin organ cultures. For the purpose of genetic immunization, we were interested in the possibility of directly transfecting cutaneous DC. To effectively deliver gold particles to the cytoplasm of epidermal LC it was necessary to use a He discharge pressure of 800 psi, which is four times higher than the pressure commonly used to transfect mouse or nonhuman primate skin.821 Even with the application of a high pressure, gold particles were mostly found in the epidermis and only a few beads were detected in the epidermal/dermal junction. Accordingly, transgenic proteins were expressed mainly in the epidermis. These features are distinct from previous results in mice or nonhuman primate models in which, using a He pressure of 200–300 psi, gold particles were localized in the epidermis, as well and in the underlying dermis.1821 Similarly, using intradermal injection of naked DNA to transfect human, pig, or murine skin, Hengge et al10 have demonstrated that the expression of transgenic proteins varied among species. Whereas in mouse the distribution of expression of the reporter protein included epidermis, dermis, hypodermis and underlying muscle tissue, in human and pig skin the expression of transgenes was limited to the epidermis. It can be speculated that differences in the anatomical and histological skin architecture among species, such as thickness of keratin layer, hair distribution or different number of epidermal cell strata, might result in such differential patterns of gene expression.

The viability of the migratory DC from transfected skin samples was as high as those migrating from untreated controls. The expression of molecules characteristic of DC activation (HLA-DR, CD80 and CD86) was upregulated even in those samples shot with gold beads loaded with pBackbone (not shown). This indicates that the conditions used to transfect human skin augments the costimulatory function of the cutaneous APCs, supporting an adjuvant effect inherent in the transfection method.

A key function of LC is to migrate from the periphery to secondary lymphoid organs where they present Ag to T cells. In the skin, different proinflammatory mediators can initiate the process of mobilization of LC.5722 Spontaneous LC migration from mouse or human skin explants has also been described.23 Our data suggest that the delivery of either gold coated with DNA, or gold particles alone, may provide a signal for LC activation and mobilization. This migration could be detected as early as 1 h after treatment. Since no significant difference in LC mobilization was evident following different treatments at early time-points, we can conclude that delivery of gold particles produces a mechanical injury sufficient to trigger the migration of LC. Our observations using human skin are consistent with studies in mouse showing that an increased number of overlapped shots delivered to skin results in proportional augmentation in the number of DC present in draining lymph nodes regardless of gene expression.9 The observation that either pBackbone or pEGFP induced increasing LC migration at later time-points demonstrates that this event is independent from the presence or expression of transgenes. Moreover, it is tempting to speculate that the resulting migration, which was significantly higher than that after delivery of gold particles alone, was a result of a combination of both mechanical injury and non-specific activation of LC by the presence of DNA, probably induced by immunostimulatory sequences contained in the plasmid backbone.242526 Although in the epidermis, luciferase activity and EGFP expression were found mainly in keratinocytes, the analysis of transgenic protein expression in epidermal cell suspensions clearly demonstrated luciferase expression in CD1a+ LC. In contrast, we were unable to demonstrate luciferase activity in the migratory DC collected 72 h after transfection. This was consistent with previous murine studies where despite the presence of Ag-specific mRNA, transgenic proteins could not been detected in skin migratory DC.2728 The lack of expression of intact transgenic protein might be attributed to low transfection efficiency however; it is interesting that those studies collected migrated DC 3–4 days after transfection. There is evidence that luciferase expression in gene gun transfected monocyte-derived DC reaches its highest expression 12–24 h after transfection and decreases to undetectable levels after 48 h.29 Similarly, we demonstrated luciferase expression in highly purified epidermal the CD1a+ LC 24 h after transfection, but we were unable to detect transgenic protein in migratory DC that require at least 48 h to be collected. However, the fact that transgenic mRNA was still present 72 h after transfection, together with the ability of transfected DC to load Ag transgenic peptides into MHC class I molecules, confirms efficient transfection of skin DC.

After gene delivery to the skin, DC might be directly transfected and/or they may acquire Ag by internalizing transgenic proteins released from neighboring cells (mostly keratinocytes). Murine studies suggest that directly transfected DC are the primary APC in draining lymph nodes. However, recent works suggested that DC can internalize and process Ag produced by other cells, potentially through the uptake of cell components or debris from dying or apoptotic cells.303132 Under our experimental conditions, both mechanisms for Ag presentation may occur. However, the rapid migration of LC out of the epidermis after transfection reduced the probability of transgenic protein uptake by LC from neighboring cells. Moreover, the presence of gold particles inside the cytoplasm of DC and synthesis of transgenic OVA mRNA further demonstrated that DC were indeed directly transduced. Since both, DC and T cells are able to migrate out of skin explants,123334 we performed RT-PCR experiments using DC purified from CD3+ T cells by magnetic sorting. To rule out the presence of contaminating keratinocytes, we used migratory DC from epidermal/dermal explants transfected with the plasmid encoding for ΔOVA under the control of the keratinocyte specific promoter K14. Whereas the presence of K14-ΔOVA, transcripts were found in transfected epidermal cells, they were not detected in the migratory cell population, thus confirming the purity of the DC population analyzed in those samples.

MHC class I restricted CTL are a critical component of immune responses against tumors or virus-infected cells.3536 Induction of CTL responses requires presentation of Ag through the MHC class I-restricted processing pathway of professional APC.26273536 Through genetic immunization, a gene encoding a target Ag may be introduced into the cytoplasm of an APC, resulting in endogenous production of the Ag and MHC class I access.3738 In the present study, we demonstrated that purified LC from MART-1 transfected skin induced secretion of INF-γ from a MART-1-specific CTL clone, and MART-1 presentation was significantly diminished after blocking MHC class I molecules, demonstrating that LC from transfected human skin processed and presented transgenic Ags T cells in an MHC class I-restricted fashion.

Our results extend those reported by Hengge et al10 regarding gene expression in human skin. Additionally, we can conclude that complete skin organ cultures, as well and epidermal/dermal explants, constitute a suitable model to analyze gene expression and also the mechanism of genetic immunization relevant to subsequent applications in clinical trials. To our knowledge, this is the first demonstration that human skin LC can be transfected directly in situ using the gene gun. Importantly, these studies show that this transfection method also provides an activation signal, stimulating a massive mobilization of LC immediately after treatment. This may obviate the need to include additional adjuvants commonly used in murine models, few of which have been approved for human clinical trials. Finally, our studies demonstrate functional significance. Activated LC from transfected human skin efficiently present transgenic Ag to MHC class I-restricted T cells. These observations clearly support further development of DNA-based vaccines designed to express antigenic proteins in DC, and demonstrate the feasibility of genetically engineer human LC in vivo.

Materials and methods

Skin samples

Human skin samples (from abdominal plastic surgery and foreskins) were obtained according with the guidelines of the University of Pittsburgh Medical Center and used immediately after excision.

Antibodies and T cell line

The following mouse anti-human monoclonal antibodies (mAb) were used: CD1a (clone Na 1/34, DAKO, Glostrup, Denmark), HLA-DR β chain (clone DK22, DAKO), CD80 (clone BB1, PharMingen, San Diego, CA, USA), CD86 (clone B70/B7.2, PharMingen), and CD3 (clone UCHT1, Sigma-Aldrich, St Louis, MO, USA). Two mAb, both recognizing the human MHC class I HLA-A2 molecule (MA2.1 and HB54), were obtained from the American Type Culture Collection (Rockville, MD, USA).

A long-term HLA-A2 restricted CTL line (M9.2) recognizing the peptide sequence MART-1/Melan-A27–35 was used as T cell responder.29

Plasmid DNA

Complete skin samples and epidermal/dermal explants were transfected with one of the following plasmids: (1) pLuciferase, encoding for the reporter firefly luciferase gene in the pcDNA3 backbone (kindly provided by Dr L Wang, University of Pittsburgh, PA, USA); (2) pEGFP-C2 encoding for the enhanced version of the green fluorescent protein from the Aequorea victoria jellyfish (Clontech Laboratories, Palo Alto, CA, USA); (3) pcDNA-ΔOVA, encoding the intracytoplasmic truncated form of the chicken ovoalbumin (ΔOVA) in the pcDNA3 backbone; (4) pCI-MART1, encoding the full length of the human melanoma peptide MART-1 in the pCI-backbone;29 or (5) pK14-ΔOVA encoding the ΔOVA cDNA under the control of the human keratin 14 cell type specific promoter (K14).13 Transgene expression in pLuciferase, pEGFP, p-ΔOVA and pCI-MART-1 was under the control of the human immediate–early cytomegalovirus promoter (hIE-CMV). The p-K14-ΔOVA was constructed from the pcDNA 3.1 vector by replacing the hIE-CMV promoter with the PCR amplified 2.01 kb genomic DNA fragment of the human K14 promoter (5′primer: GACCCGGGCTCCGGAGCTT and 3′primer GATCGGG TAAATTGCAAAG).

Gene delivery to human skin

Genetic immunization was performed by biolistic bombardment using an Accell helium (He) pulse gun (PowderJect, Madison, USA) on human complete skin samples or on epidermal/dermal explants attached to a cork plate. One μm gold particles (BioRad Laboratories, Hercules, CA, USA) were conjugated with naked plasmid DNA as previously described.8 Biolistic genetic immunization was performed by delivering two overlapping shots (∼2 mg gold beads and 4 μg of plasmid DNA). The optimization of gold particle delivery in human skin was determined as described in Results. He discharge pressures ranging from 400 psi to 900 psi were tested on skin samples and the distribution of gold particles was analyzed in cryosections stained with hematoxylin and eosin (H&E).

Preparation of epidermal and dermal sheets

Skin samples (composed of epidermis and complete dermis) or epidermal/dermal explants (composed of epidermis and a very thin layer of dermis), were immunized by biolistic bombardment. After transfection, they were cultured on sterile steel meshes with the epidermal side up in RPMI 1640 (Irvine Scientific, Santa Ana, CA, USA) supplemented with 10% heat inactivated FCS, 20 mM Hepes, 2 mM L-glutamine, 200 U/ml penicillin/ streptomycin (Gibco, Grand Island, NY, USA) and 20 μg/ml gentamicin (Sigma) (complete medium), at 37°C in 5% CO2. Twenty-four hours later, the skin was rinsed in PBS and incubated in a solution of 20 mM EDTA in PBS (pH 7.3–7.5) 2 h at 37°C, with continuous shaking. The epidermis was separated from the dermis with fine forceps under stereoscopic microscope and epidermal and dermal sheets were fixed in cold acetone and immunostained.

Preparation of epidermal cell suspensions

Epidermal cell suspensions were prepared immediately after transfection as previously described.39 Single cell suspensions were cultured for 24 h in complete medium. For CTL cytokine secretion assays, only heat-inactivated human AB serum was used throughout the procedure. To obtain highly purified LC, 24 h cultured epidermal cell suspensions were incubated with mouse anti-human CD1a (DAKO) mAb followed by bead conjugated, goat anti-mouse Igs (Miltenyi, Auburn, CA, USA). LC were isolated by positive selection using a magnetic bead separation kit (Miltenyi) according to the manufacturer's protocol. Enriched LC populations obtained in this manner reproducibly consisted of >95% of CD1a+ cells as determined by flow cytometry. LC viability assessed by propidium iodine (PI) exclusion was always >85%.

Preparation of dendritic cells migrated from human skin explants (skin migratory DCs)

Skin explants composed of epidermis and a thin layer of dermis (epidermal/dermal explants) (∼0.3 mm thickness) were obtained from abdominal skin samples by using skin graft knife (Padgett Instruments, Kansas City, MO, USA), and biolistic transfected in sterile conditions. Transfected or nontransfected control epidermal/dermal explants were cultured in complete medium, at 37°C in 5% CO2, with the epidermal side up over sterile steel meshes. Skin migratory DCs were collected after 48 or 72 h of culture.

Cytospin preparation

Skin migratory DCs or cultured LC enriched from epidermal cell suspensions were spun on to glass slides using a Shandon cytocentrifuge (at 230 g), air dried and stained with May–Grünwald-Giemsa, or fixed in cold acetone for immunostaining.

Transmission electron microscopy (TEM)

Skin migratory DCs from transfected epidermal/dermal explants were washed in PBS, spun down and the resulting pellet was fixed in 2.5% glutaraldehyde, and processed as previously described.8 Ultrathin sections were analyzed using a JEOL 1210 transmission electron microscope (JEOL, Chicago, IL, USA).

Quantification of Langerhans cell migration after genetic immunization

Complete skin samples were immunized by biolistic gene delivery and cultured over sterile steel meshes in the air liquid interface in complete medium supplemented with 10% human AB serum. Epidermal sheets dissected at different time-points (0, 1, 3, 6, 12, 24, and 48 h) after transfection were fixed in cold acetone, and immunostained with anti-HLA-DR mAb (DAKO), followed by Cy3- conjugated goat anti-mouse Igs (Jackson Laboratories).

The number of LCs in the epidermis was assessed in 12 fields (×400) using a Nikon FXL microscope (Chicago, IL, USA). Comparisons among basal LC migration from nontransfected samples; shot with unloaded gold particles, or with pBackbone, or with pEGFP were evaluated.

Analysis of the expression of transgenic proteins

Luciferase detection:

Expression of luciferase after transfection was assessed in (1) epidermal or dermal sheets (prepared 24 h after transfection and culture); (2) in epidermal cell suspensions (prepared immediately after transfection and cultured for 24h); and (3) in skin migratory DCs (obtained from epidermal/dermal explants). Epidermal or dermal sheets were obtained as described above and resuspended in 500 μl of lysis buffer (0.05% Triton ×100, 0.1 M Tris-HCl and 2 mM EDTA) and homogenized using a Tissue Tearor (Fisher Scientific, Pittsburgh, PA, USA) at 4°C. Complete epidermal cell suspensions, CD1a+ and CD1a− epidermal cell fractions, or HLA-DR+ skin migratory DC were resuspended in 200 μl of lysis buffer at 4°C. After pelleting the cell debris, 10 μl of supernatants were assayed by triplicate using the Luciferase Assay System (Promega, Madison, WI, USA) and luminescence was analyzed by an Autolumat LB 953 luminometer (EG&G Berthold, Bundora, Victoria, Australia). The level of the sample luminescence was recorded as relative light units (RLU)/μg of protein after calculating the amount of protein present in the sample assayed, against a bovine serum albumin standard curve.

EGFP detection:

The expression of EGFP in mature LC was evaluated in cytospins prepared from enriched LC suspensions as described above, fixed in cold acetone, and immunostained with anti-HLA-DR mAb overnight at 4°C, followed by a secondary Cy3 conjugated goat anti-mouse Igs (Jackson Laboratories, West Grove, PA, USA). After mounting, cytospins were analyzed using a photomicroscope Nikon FXL (Chicago, IL, USA). Images were collected by means of Sony 3 chip color digital camera.

RT-PCR detection of ΔOVA gene transcripts

Skin epidermal/dermal explants were biolistically transfected with pΔOVA or pK14-ΔOVA and cultured during 72 h in complete medium. After culture, skin migratory cells were harvested and incubated with anti-HLA-DR mAb followed by bead-conjugated goat anti-mouse Igs. Skin migratory DCs were highly enriched by immunomagnetic sorting of the CD1a+ cell population (>95% of CD1a+) cells. Total RNA was isolated from transfected epidermal sheets, and from migratory DCs (∼5 × 105 cells/sample) using 200 μl TRIAZOL Reagent (GibcoBRL, Life Technologies, Rockville, MD, USA). The resulting RNA was DNase treated (DNase I, Amp Grade, GibcoBRL, Life Technologies), and subsequently used in RT-PCR reactions. The first strand of cDNA was synthesized by SuperScript II enzyme (GibcoBRL, Life Technologies) using oligo-dT primer and 2 μg DNase I treated total RNA in a 20 μl reaction volume, at 42°C for 30 min. Two μl of the RT reaction was directly used in PCR amplification for the ΔOVA transgene expression products and for the human β-actin as control (primers used: ΔOVA-5′ TCAGAGTGACTGAGCAAGAAAGC, and ΔOVA 3′ACCCATACCCATTAAGACAGATGTG, results in a 311 bp fragment, human β-actin-5′ GTGGGGCG CCCCAGGCACCA, and human β-actin-3′ CTCCTT AATGTCACGCACGATTTC, results in a 539 bp fragment). The conditions of the PCR were: 94°C for 1 min, one cycle pre-denaturation, followed by 35 cycles of amplification at 94°C for 30 s, 60°C for 30 s and with 72°C for 1 min extension segments. As a control of RNA purity, total RNA (∼0.5 μg per reaction) was also used in PCR reactions (RT-PCR). The resulted PCR products were detected by gel electrophoresis in a 2% agarose gel.

Cytokine release assays

For cytokine release assays, gold beads loaded with pCI-MART1 or pCI-backbone were biolistically delivered in HLA-A2+ skin samples. Epidermal cell suspensions were prepared immediately after treatment and cultured 24 h in complete medium supplemented with 10% human AB serum. Enriched CD1a+ LC (purity >90%) were obtained by positive selection through paramagnetic columns as described. Enriched CD1a+ LC were co-cultured with responder HLA-A2 restricted CTL clone (M9.2) (ratio LC/CTL = 1/20), in 96 round-bottom well plates (Costar, Corning, NY, USA), in complete medium. After 24 h, supernatants were harvested and human IFN-γ content was measured by ELISA, implementing anti-human IFN-γ capture and detection antibodies, purchased from PharMingen (San Diego, CA, USA). The lower limit of detection of IFNγ was 1 U/ml.

Statistic analysis

Means ± 1 s.d. obtained from migration experiments and cytokine release assays were compared by ANOVA and ad hoc Student–Newman Keuls tests. Comparisons between two different means ± 1 s.d. from luciferase expression assays were performed by Student t test. A P value <0.05 was considered significant.