Electroporation as a vaccine delivery system and a natural adjuvant to intradermal administration of plasmid DNA in macaques

In vivo electroporation (EP) is used to enhance the uptake of nucleic acids and its association with DNA vaccination greatly stimulates immune responses to vaccine antigens delivered through the skin. However, the effect of EP on cutaneous cell behavior, the dynamics of immune cell recruitment and local inflammatory factors, have not been fully described. Here, we show that intradermal DNA vaccination combined with EP extends antigen expression to the epidermis and the subcutaneous skin muscle in non-human primates. In vivo fibered confocal microscopy and dynamic ex vivo imaging revealed that EP promotes the mobility of Langerhans cells (LC) and their interactions with transfected cells prior to their migration from the epidermis. At the peak of vaccine expression, we detected antigen in damaged keratinocyte areas in the epidermis and we characterized recruited immune cells in the skin, the hypodermis and the subcutaneous muscle. EP alone was sufficient to induce the production of pro-inflammatory cytokines in the skin and significantly increased local concentrations of Transforming Growth Factor (TGF)-alpha and IL-12. Our results show the kinetics of inflammatory processes in response to EP of the skin, and reveal its potential as a vaccine adjuvant.

Among the various vaccination approaches against infectious diseases such as human immunodeficiency virus (HIV), deoxyribonucleic acid (DNA) vaccines have several advantages: they are easily produced, provide opportunities for molecular engineering, lack anti-vector immunity, and have the potential to promote both cellular and humoral immune responses 1 . However, despite their high immunogenicity in murine models, DNA vaccines have shown poor efficacy in large animal models and humans 2 . New strategies to improve DNA vaccines include the optimization of transcriptional control elements and codons [3][4][5] , the use of adjuvants, such as Toll-like receptor (TLR) ligands 6 , cytokine expressing plasmids [7][8][9] or apoptosis-based adjuvants [10][11][12] and the choice of an appropriate delivery system such as local electroporation (EP) [13][14][15] .
In particular, EP has been largely used to enhance plasmid DNA uptake and increase the number of antigen-producing cells 16,17 . In addition, EP modifies blood vessel permeability and facilitates leukocyte extravasation in the exposed area 18 . However, the effects of EP on cutaneous antigen presenting cells (APCs) and on the dynamics of cell recruitment at the vaccine site have not been fully described.

Results
Electroporation stimulates local antigen expression, especially in the epidermis. We used in vivo bioluminescence and fibered confocal fluorescence microscopy to monitor the expression of antigens in the skin after the id injection of DNA vaccine with or without EP. When EP was applied, luciferase was expressed for up to two weeks after the injection of auxo-GTU ® -Luc-EGFP plasmid with a peak at 24 h, but was poorly expressed (p = 0.014) after immunization without EP (Fig. 1a,b). This observation was confirmed by the direct visualization of cells expressing EGFP: there were significantly more EGFP + cells at 4 h (p = 0.0183), 24 h (p = 0.0002) and 48 h (p = 0.0002) after id vaccination with EP than after id vaccination with DNA alone (Fig. 1c,d). Furthermore, antigen expression at 24 h seemed to be mainly located in the epidermis (Fig. 1c,e).
Langerhans cells are highly mobile, interact with transfected cells and leave the epidermis after DNA injection with EP. Fluorescent stained epidermal cells, after id injection of HLA-DR antibody (Ab), were organized into a homogeneous network with dendritic morphology and were considered as Langerhans cells (LCs). These cells were imaged in vivo for 5 days after DNA ± EP in three independent experiments (Fig. 2a). Their number remained stable between 24 h and 96 h after the injection of plasmid DNA or PBS without EP (Fig. 2a,b). Interestingly, LC density was significantly decreased from 48 h after EP (p < 0.0001), suggesting their migration from the epidermis (Fig. 2b). LCs were imaged continuously for 18 h in skin explants taken 24 h after vaccination, to assess their behavior and motility. EP was associated with morphological changes of LCs, including dendritic retraction and rounding up (Fig. 2c), which corresponds to an activated LC profile 20 . EP also led to LC migration from the epidermis because their number decreased significantly (p = 0.017) over time (Fig. 2d). Furthermore, the velocity and displacement of LCs were significantly higher (p < 0.0001) after EP than in non-electroporated skin, and EP was associated with a high confinement ratio, illustrating the directional movement of LCs after vaccination with EP ( Fig. 2e). In addition, some LCs interacted for several hours with antigen producing cells (EGFP + cells), resulting in double stained cells, which suggests antigen capture by LCs (Fig. 2f).

EP causes immune cell infiltration in the skin and promotes the mobility of recruited APCs in the dermis post-vaccination.
Flow cytometry revealed the presence of four leukocyte populations (CD45 + cells) in the skin (both epidermis and dermis) (Fig. 3a). EP without DNA injection was able to induce lymphocytes T CD3+ infiltration in the skin 24 h after treatment (Fig. 3b). Furthermore, when EP was associated to the administration of DNA vaccine, the number of dendritic cells and T lymphocytes was significantly higher in electroporated skin 24 h after immunization than in skin injected with PBS alone (p = 0.0226 and p = 0.0055). There was a similar trend for monocytes/macrophages (p = 0.15) and DR − CD66 + neutrophils (p = 0.20) (Fig. 3b). We also observed this cell recruitment in the dermis by confocal microscopy, which revealed the presence of many HLA-DR + cells 24 h after DNA injection with EP (Fig. 3c). In contrast to the epidermis, the dermal APC confinement ratio was not modified by EP whereas cell velocity (p < 0.0001) and displacement (p = 0.0022) were still significantly higher in electroporated skin than in non-electroporated skin (Fig. 3d).

DNA injection combined with EP affects deep skin layers.
Because the hypodermis and the subcutaneous muscle are interconnected and localized under the skin (Fig. 4a), we investigated the effect of EP on these skin layers. We detected antigen production in hypodermal cells after the id injection of plasmid DNA with or without EP (Fig. 4b, upper panel), and found that both DR + CD163 + and DR − CD66 + cells infiltrated this tissue (Fig. 4c). By contrast, we could detect antigen expression in subcutaneous muscle only after DNA administration with EP (Fig. 4b, lower panel). In this condition we observed a tendency of DR + CD163 + and DR − CD66 + cells infiltration at the site of immunization (p = 0.0625) (Fig. 4c).
Vaccine antigen expression co-localizes with areas of damaged keratinocytes. We observed damaged keratinocytes on the epidermal surface 24 h after PBS + EP (Fig. 5a). Interestingly, when EP was associated with DNA injection, vaccine antigen was detected in these dying keratinocytes (Fig. 5b). Furthermore, the number of apoptotic cells was significantly higher after vaccination with EP than after DNA alone (p = 0.0313) (Fig. 5c,d). These cells were mostly CD45 − (Fig. 5e) and may represent an additional stimulus for immune activation 12 .
EP creates an early inflammatory environment. Finally, we analyzed the cytokine microenvironment in supernatants of skin biopsies post-EP. The concentrations of TGF-α and IL-12 were significantly higher in skin 24 h after EP than in non-electroporated skin (p = 0.0274 and p = 0.0117) (Fig. 6). At this time point, the concentration of other pro-inflammatory (GM-CSF, IL-8, TNF-alpha, IL-18) and anti-inflammatory (IL-1ra, IL-10) cytokines, tended to be higher in electroporated skin than in non-electroporated skin.

Discussion
Electroporation has long been used to facilitate the DNA access to the cell nucleus and its association with DNA vaccination clearly enhances the production of vaccine antigen in mice and large animal models [21][22][23] . However, it appears that high antigen expression is not the only factor that determines vaccine immunogenicity because electrical pulses applied prior to or simultaneously to DNA vaccination induce similar immune responses 24 . We used the auxo-GTU ® -multiHIV plasmid to investigate the cellular effects of vaccination with EP. This plasmid is known for its ability to stimulate strong cellular responses which are significantly larger and more persistent when EP is used 19 . Consistent with other studies 17,21,23,25,26 we confirmed that vaccine antigen was more strongly expressed when injection was combined with EP. Interestingly, EP seemed to affect the localization of transfected cells and they were predominantly present in the epidermis. High antigen availability in the epidermis may be beneficial due to the LC population localized in this skin layer. LCs are essential for T-cell responses, as shown by the induction of modified vaccinia virus Ankara (MVA)-specific CD8+ T cells 27 , and for antitumor cellular immunity 28 . In vitro experiments have revealed that LC are particularly effective at priming and cross-priming naïve CD8 + T cells 29 . However, there has been disagreement about LC functions in vivo and it is now accepted (d) Quantification of EGFP + cells from 4 to 96 h after vaccination, from 10 different frames. Paired and unpaired t-test. (e) Distribution of EGFP + cells in the epidermis and the dermis after DNA vaccination with electroporation. Data are presented as mean ± SD; *p < 0.05; **p < 0.01; ***p < 0.001; EP, electroporation.
that their role depends on the quantity and the quality of immune signals 30 . EP in vivo promotes the migration of LCs out of murine skin explants 31 . In our experiments, DNA vaccination with EP caused morphological changes to LCs that have been previously associated with their migratory behavior 20 . In addition, the high motility of LCs, their interaction with DNA transfected cells, and their migration from the epidermis, shows that LCs are involved in the mechanism of vaccination with EP. Moreover, after EP, we found vaccine antigen in damaged keratinocytes on the skin surface. Thus, vaccination combined with EP provides an additional source of cell-associated antigen, which is easily accessible to surrounding cells for cross presentation 32 . Moreover, we showed that simple EP of macaque skin led to immune cell infiltration at the site of injection with a peak at 24 h (data not shown). These immune cell infiltrates have also been observed after intramuscular vaccination with EP in mice 33 . Recently, Markelc et al. demonstrated that electrical pulses increase blood vessel permeability, which is accompanied by leukocyte extravasation 34 . In our study, the overlap between the peak of antigen expression and APC recruitment may lead to robust antigen processing and presentation at the site of vaccination. Furthermore, the recruitment of lymphocytes to the skin (although non-specific to the antigen) may promote potent local responses after vaccine boost, as occurs during the recall of peripheral infection 35 .
We also identified several cytokines produced locally in response to electrical pulses without vaccination. IL-12 and GM-CSF have already been used as adjuvants to DNA vaccines 36,37 . In particular, IL-12 promotes the accumulation of CD8 + T cells 38 and the differentiation of naïve T cells to Th1 39 , whereas GM-CSF plays a central role in the development and maturation of dendritic cells and its overexpression in the epidermis induces both keratinocyte proliferation and apoptosis 40 . Interestingly, we found that the TGF-α concentrations were significantly higher in electroporated skin than in non-electroporated skin. This cytokine up-regulates TLR5 and TLR9 expression on keratinocytes and enhances the cell response to CpG DNA 41 (CpG for cytosine-phosphate-guanine). Other pro-inflammatory cytokines induced by EP, participate in immune cell recruitment (IL-8) 42 or in cell maturation and migration (TNFα, IL-18) 43,44 . We also observed delayed production of anti-inflammatory cytokine IL-10 in electroporated skin, which may be important to limit inflammation caused by EP.
The analysis of subcutaneous tissues revealed that the effects of EP penetrate deep into the skin. We showed that when vaccination was performed with EP, the vector reached subcutaneous tissues, such as the subcutaneous muscle. In mouse skin, most transfected cells following EP are observed in the hypodermis and around the panniculus carnosus 16 . This muscle is involved in skin movements and is almost absent in humans, which may explain several interspecies differences regarding vaccine delivery through the skin, as occurs for some adeno-associated viral vectors which show skin muscle tropism after id injection 45 . Furthermore, recruited inflammatory cells in the panniculus carnosus may play an additional role in vaccine responses by creating a favorable environment for antigen processing and presentation.
Nevertheless, epidermal injury resulting from EP may induce the production of damage-associated molecular patterns and activate immune actors independent of vaccination. Indeed, uric acid and high mobility group protein B1 (HMGB1) have been detected in electroporated muscle with or without vaccination and have been associated with robust T cell responses to plasmid DNA antigens 46 . Tissue damage and inflammation depend on EP parameters such as the intensity, the duration of electrical pulses and the type of electrodes 47,48 . The immune enhancer potency of cell death has been reported after vaccine delivery by micro-needles 49 and this property is also being used to develop laser-based vaccine adjuvants 50 .
However, for clinical use, this immune stimulation due to electroporation should be counterbalanced by the discomfort which may occur with this procedure. In our study electroporation performed with needleless tweezers electrodes 51 , resulted in very slight and transitory superficial tissue injury. The same electroporation protocol has been previously tested on pig skin 19 approved as biomedical model for human skin studies 52 .
Our findings demonstrate that EP is an effective vaccine delivery system and adjuvant. The application of EP to the skin broadens immune responses to vaccine in all skin layers and subcutaneous tissues by extending the delivery of antigen. Thus, vaccine protocols that promote local immune actors may abolish the need for chemical-based adjuvants.

Materials and Methods
Animals. Adult female cynomolgus macaques (Macaca fascicularis), imported from Mauritius and weighing 3-6 kg, were housed in CEA facilities (accreditation: C 92-032-02) and handled in accordance with European guidelines for nonhuman primates (NHP) care (EU Directive N 63/2010). This study was approved by the regional committee for animal care and use (Comité d'Ethique en Expérimentation Animale n° 44, AP N° 11_008). Animals were handled under sedation with an intramuscular injection of 10 mg/kg ketamine hydrochloride (Merial, France).
Plasmids. The vaccine vector auxo-GTU ® -multiHIV encodes a fusion protein composed of the full-length sequences of Rev, Nef, Tat, p17 and p24 proteins and a stretch of selected T cell epitopes from Pol and Env

DNA vaccine injection. Skin was shaved at the site of injection. Each animal received id injections of 100 µl
(1 mg/ml) of auxo-GTU ® -multiHIV plasmid with or without EP. Phosphate-buffered saline (PBS) (100 µL) was injected as a control. EP was performed with a portable pulse generator (CUY21 EDIT; Nepa Gene, Ichikawa, Japan) and tweezer electrodes (6 pulses of 10 msec with output current 300-600 mA) 51  Technologies, France). For epidermal APC imaging, the monoclonal antibody (Ab) anti-HLA-DR (clone L243, Ozyme, France) was amine-labeled with the Fluoprobe 490 kit (Interchim, France). Five µg of fluorescent Ab was injected id into the middle of each vaccination site, 5 min and then 49 h after vaccination. Epidermis imaging was performed with the probe S-1500, which covers a 300 µm circular diameter in a focal plane and provides images from the skin surface, with a slice thickness of 15 µm and a lateral resolution of 3.3 µm. The Z-1800 probe was used for dermal imaging. This probe is 1.8 mm in diameter and acquires fluorescent signals emitted from 100 µm in depth, with a lateral resolution of 3.5 µm. Images were collected from at least 10 different areas per injection site from three independent experiments (n = 30). Images were analyzed with ImageJ 1.6 software (National Institute of Mental Health, Bethesda, USA).

Ex vivo imaging on skin explants.
Microscopic studies of tissue explants were performed as described previously 53 . Briefly, skin biopsies were performed 24 hours after the injection auxo-GTU ® -Luc-EGFP plasmid and 2 hours after the in vivo id injection of anti-HLA-DR-AF700 Ab (Ozyme, France). Fat tissue was removed and each biopsy was cut into two equal parts and placed in a 6-well plate (MatTek Corporation, Ashland, USA) in contact with complete culture medium to visualize dermal and epidermal skin layers. Images were acquired with a Plan Fluor 20x DIC objective (NA: 0.45) on a Nikon A1R confocal fast laser scanning system (Nikon Corporation, Japan) equipped with a thermostatic chamber (37 °C; 5% CO 2 ). Images were recorded with a high-speed resonant scanner every 10 minutes for 22 hours. Volocity software (Perkin Elmer, Waltham, USA) was used to reconstruct three-dimensional images and to assess cell motility (velocity, displacement and confinement ratio).
Immunohistofluorescence of skin sections. Skin biopsies were performed 24 hours post-injection and embedded in optimal cutting temperature compound (OCT) and frozen in liquid nitrogen. Ten micrometer sections were fixed in 4% of Paraformaldehyde (PFA) for 15 min at room temperature. Tissue was permeabilized with Triton X-100 (0.3%), Bovine serum albumin (BSA) (1%) and goat serum (1%). Skin sections were incubated overnight at 4 °C with 10 µg/ml of anti-GFP-AF488 Ab (Invitrogen, France) or anti-p24 Ab (kind gift from B. Verrier, UMR 5086 CNRS/UCBL, Lyon, France) to detect the expression of vaccine antigens in animals injected with the auxo-GTU ® -Luc-EGFP or the auxo-GTU ® -multiHIV plasmid, respectively. For keratinocyte staining, Figure 6. Cytokine production in the skin after EP. The cytokine concentrations were measured in the supernatants of skin biopsies after 6 h of incubation at 37 °C and 5% CO 2 . The values of the change fold were expressed as the ratio between the cytokine concentrations in the supernatant in each condition and the ones measured in the non-electroporated condition. Skin biopsies were performed 24, 48 and 72 hours after intradermal injection of PBS followed by EP. Kruskal-Wallis test (n = 5). Data are presented as mean ± SD; *p < 0.05, EP: electroporation.

Flow cytometry.
To characterize skin immune cells, 8 mm-biopsies were taken from the site of vaccination.