Hydrodynamic gene delivery in human skin using a hollow microneedle device

Abstract Microneedle devices have been proposed as a minimally invasive delivery system for the intradermal administration of nucleic acids, both plasmid DNA (pDNA) and siRNA, to treat localised disease or provide vaccination. Different microneedle types and application methods have been investigated in the laboratory, but limited and irreproducible levels of gene expression have proven to be significant challenges to pre‐clinical to clinical progression. This study is the first to explore the potential of a hollow microneedle device for the delivery and subsequent expression of pDNA in human skin. The regulatory approved MicronJet600® (MicronJet hereafter) device was used to deliver reporter plasmids (pCMV&bgr; and pEGFP‐N1) into viable excised human skin. Exogenous gene expression was subsequently detected at multiple locations that were distant from the injection site but within the confines of the bleb created by the intradermal bolus. The observed levels of gene expression in the tissue are at least comparable to that achieved by the most invasive microneedle application methods e.g. lateral application of a microneedle. Gene expression was predominantly located in the epidermis, although also evident in the papillary dermis. Optical coherence tomography permitted real time visualisation of the sub‐surface skin architecture and, unlike a conventional intradermal injection, MicronJet administration of a 50 &mgr;L bolus appears to create multiple superficial microdisruptions in the papillary dermis and epidermis. These were co‐localised with expression of the pCMV&bgr; reporter plasmid. We have therefore shown, for the first time, that a hollow microneedle device can facilitate efficient and reproducible gene expression of exogenous naked pDNA in human skin using volumes that are considered to be standard for intradermal administration, and postulate a hydrodynamic effect as the mechanism of gene delivery. Graphical abstract Figure. No Caption available.


Introduction
Intradermal delivery of plasmid DNA (pDNA) has been proposed for a number of therapeutic applications (Abdul-Wahab et al. 2014, Hart andVile, 1994), including DNA vaccines, which are targeted to the skin to exploit the highly evolved and competent local immune system (Robinson, 1999, Kutzler andWeiner, 2008).
However, although detectable expression of exogenous genes in human and animal skin was achieved over 20 years ago (Hengge et al. 1995, Hengge et al. 1996, Hengge et al. 1998, Yu et al. 1999), clinical translation has been hindered by extraand intra-cellular barriers to efficient and reproducible cell transfection.
Cell barriers to transfection have been well characterised in numerous cell types, including skin cells.Viral vectors enable efficient and long-term expression of exogenous genes (Howarth et al. 2010) but they have a number of acknowleged constraints (Boulaiz et al. 2005, Lv et al. 2006, Mingozzi and High 2013, Thomas et al. 2003).Non-viral vectors provide an alternative, but whilst these vectors are comparably safe, not immunogenic, low cost and easy to manufacture (Li andHuang, 2000, Wang et al. 2013), their transfection efficiency is often significantly reduced compared to their viral counterparts (He et al. 2010, Nayerossadat et al. 2012).Physical enhancement techniques have also been used promote cell transfection.These include electroporation (Broderick et al. 2011, Drabick et al. 2001, Guo et al. 2011, Roos et al. 2009), iontophoresis (Brand andIversen, 1996, Vlassov et al. 1996), particle bombardment (Klein et al. 1987, Yang et al. 1990), tattooing (Bins et al. 2005, Oosterhuis et al. 2012), ultrasound (Newman and Bettinger, 2007) and radiofrequency ablation (Birchall et al. 2006).However these approaches are limited by the complexity and cost of the equipment, cell death and restricted access to the tissue or organ of interest (Meacham et al. 2014).The challenge of gene delivery at the cellular level is compounded in the skin, where a diverse range of cell populations exist in a complex three dimensional architecture.
The most attractive target for localised gene expression in this stratified organ is the region with the greatest density of viable proliferating cells, i.e. the basal layer of the viable epidermis.This target layer exists 100 -200 m below the skin surface and is sandwiched by the terminally differentiating keratinocytes of the upper epidermis and the underlying connective tissue of the dermis.Potential targets for nucleic acid therapies also exist in the elastin and collagen packed layer of the dermis but cells in this region are not constrained to a membranous organisation and are often motile.
Overcoming the physical and biological barriers of the skin, which has evolved to act as a barrier to exogenous materials and insults, to target exogenous nucleic acids to specific cell types is therefore challenging.Despite these challenges, numerous studies have demonstrated that vectorless (naked) pDNA can transfect cells in both human and animal skin (Coulman et al. 2006, Hengge et al. 1995, Hengee et al. 1996, Yan et al. 2014).This has been traditionally achieved via conventional intradermal injection, whereby a single hypodermic needle with millimetre dimensions is used to overcome the stratum corneum barrier and transfer a bolus of pDNA into the skin.However the resulting gene expression levels are low and often restricted to the viable epidermal cells that surround the injection site (Hengge et al. 1996).This relatively inefficient method of delivery has therefore been augmented by physical transfection methods, such as electroporation (Broderick et al. 2011, Wells, 2004), which temporarily permeabilises cell membranes to enhance transfection levels in the tissue.Recent advances in this area, including Phase 2 clinical trial data (Trimble et al. 2015), are very encouraging.
However conventional injections combined with electroporation have acknowleged limitations e.g.pain, bleeding and the inherent costs and complexity associated with electroporation technology.The ideal delivery system would therefore be a single and simple minimally invasive device (Miller and Pisani, 1999) that can achieve vectorless delivery of nucleic acids across both the tissue barrier, the stratum corneum, and the major cellular barrier, the cell membrane (Ibraheem et al. 2014).
Microneedle (MN) technology uses single or multiple microscopic needles to deliver therapeutic materials through the stratum corneum (SC) barrier into the underlying viable epidermis, or dermis, in a minimally invasive manner.The primary advantage of MNs, compared to other 'physical' delivery methods, is the simplicity of the device.
Studies have exemplified the potential of MNs for delivery of nucleic acids, both pDNA and siRNA, for the treatment of genetic skin diseases, cutaneous cancers, hyperproliferative diseases and vaccinations and these are summarised in a recent review (McCaffrey et al. 2015).These studies have employed solid MNs to deliver pDNA into the skin using either the "soak and poke" (Birchall et al. 2005, Coulman et al. 2006, Pearton et al. 2008, Yan et al. 2014) or the "coat and poke" approach (DeMuth et al. 2013, Gill et al. 2010, González-González et al. 2011, Kim et al. 2013, Pearton et al. 2012, Song et al. 2012).The former involves topical application of a liquid pDNA formulation prior to application of a MN device (Pearton et al. 2008).The "soak and poke" method results in reasonable levels of gene expression (Birchall et al. 2005, Chabri et al. 2004, Coulman et al. 2006, Pearton et al. 2008, Yan et al. 2014) however it is impossible to control the delivered dose using this method.The alternative strategy is to "coat and poke" i.e. a MN device is dry coated with a pDNA formulation and then applied to the skin (DeMuth et al. 2013, González-González et al. 2011, Pearton et al. 2012).MN coating methods are able to control the spatial distribution and mass of a therapeutic that is loaded onto MNs and thus dosing is more reproducible for the coat and poke approach (Kim et al. 2013).However the transfection efficiency of this method is significantly reduced (Pearton et al. 2012).
Both of these approaches suffer from low to moderate levels of gene expression that are restricted to cells adjacent to the site of microdisruption.Transfection can be heightened by increasing tissue disruption using repeated MN applications to the skin surface (Yan et al. 2014) or lateral application of MNs (Mikszta et al. 2002).
However, these approaches cause significant disruption of the tissue barrier that are unlikley to be clinically acceptable.
The hollow MN approach administers the therapeutic cargo in a liquid formulation through one or more needles of micron dimensions.The advantages of hollow MN systems are conferred by its ability to inject a finite dose using a standard syringe and with minimal training.Re-formulation of commercially available medicines and vaccines is often not required and therefore the potential of hollow MNs has been established both pre-clinically and clinically (Daugimont et al. 2010, Gupta et al. 2009, Gupta et al. 2012, Kim and Prausnitz, 2011, Levin et al. 2014, Levin et al. 2015, Norman et al. 2013, Roxhed et al. 2008, Van Damme et al. 2009, Wang et al. 2006).To the best of our knowledge, the ability of hollow MNs to deliver nucleic acids to human skin has not yet been reported.The aim of this study was to investigate the potential of a regulatory approved hollow MN device, MicronJet (Levin et al. 2015), for pDNA delivery to human skin in an effort to combine the favourable transfection efficiencies previously witnessed with liquid delivery (soak and poke approach) with the finite dosing that can be achieved using a hollow MN delivery system.

Materials and methods
All reagents and laboratory consumables were purchased from Fisher Scientific, UK, (Loughborough, UK), unless otherwise indicated. 2

Preparation of pDNA
The reporter plasmids pCMVβ and pEGFP-N1 were amplified and purified from previously transformed stocks of DH5α Escherichia coli using antibiotic selective conditions and a Qiagen Mega Kit (Qiagen Ltd, Crawley, UK), (Birchall et al. 2005).

Collection, processing and maintenance of an ex vivo human skin
Excised human breast skin was collected immediately after surgery and transported to the laboratory in culture media, which consisted of Dulbecco's Modified Eagle Medium (DMEM), supplemented with 100 unit mL -1 of penicillin and 100 μg mL -1 of streptomycin (all Life Technologies, UK) at 4°C.Collected skin samples were used within 2 h of excision.Subcutaneous adipose tissue was removed by blunt dissection and skin was pinned, dermis side down, onto a cork dissection board to facilitate injection.An 8 mm biopsy punch was used to isolate areas of treated skin, which were further maintained in the organ culture by incubating at the air-liquid interface at 37°C in an atmosphere of 5% (v/v) CO 2 /95% (v/v) air (Ng et al. 2009).Data in this study was obtained from replicate samples obtained from ten independent skin donors (aged 51-80 years old).
2.5 Gene expression in primary and continuous human keratinocyte cell lines 2.5.1 HaCaT cell culture Immortalised human keratinocyte cells (HaCaT cells), (Boukamp et al. 1988) were received as a gift from Professor Mark Gumbleton (School of Pharmacy and Pharmaceutical Sciences, Cardiff University).Cells were cultured in a growth medium, which consisted of DMEM, supplemented with 10% fetal bovine serum (FBS), 100 unit mL -1 of penicillin and 100 μg mL -1 of streptomycin (all Life Technologies, UK) at 37°C in a humidified atmosphere containing 5% CO 2 .

Primary keratinocytes isolation and culture
Superficial skin tissue was mechanically isolated from a full thickness skin sample using a surgical blade (approx.1x1 cm 2 ) and transferred to a 0.25% trypsin solution which was incubated overnight at 4°C.The resulting suspension of cells was filtered through a 70 μm cell strainer and centrifuged at 300 x g for 7 mins.The cell pellet was gently re-suspended in media (EpiLife medium supplemented with 10% human keratinocyte growth supplement (HGKS) and 100 unit mL -1 penicillin, 100 μg mL -1 of streptomycin and 2.5 μg mL -1 of amphotericin B (all Life Technologies, UK) before seeding in a 25cm 2 flask.Growth media was changed every 2-3 days until the primary keratinocytes reached 70-80% confluence.

HaCaT cell transfection
Twenty four hours before treatment, cells were seeded into 24-well plates at a density of 7.5x10 4 cells cm -2 in 0.5 mL growth media.Cell populations were then treated using either (i) naked plasmid (pEGFP-N1 or pIRES-TdTomato) or (ii) pDNA lipoplex formed with Lipofectamine2000 transfection reagent (positive control).
Naked plasmid (0.6 μg in 100 μL PBS) was added to a well containing 500 μL of DMEM.Transfection complexes were prepared according to the manufacturer's protocol (Life Technologies, UK).Briefly, 2 μL of Lipofectamine2000 was diluted in 50 μL of DMEM and pDNA (0.6 μg /per well) was diluted in 50 μL of DMEM.After 5 min of incubation, pDNA and Lipofectamine2000 dilutions were mixed gently and allowed to equilibrate for 20 min at room temperature.Cells were washed once with PBS, and then lipoplexes (100 μL) were added to each well containing 500 μL of DMEM and mixed gently by rocking the plate back and forth.After 6 h, the cells were washed three times with PBS and the treatment media was replaced with DMEM supplemented with 10% FBS and the cells were incubated for a further 42 h.All experiments were performed at least in triplicate.

Primary keratinocyte cell transfection
Primary keratinocyte transfection was performed in a comparable manner to HaCaT transfection (Section 2.5.3), with the following exceptions.Twenty four hours before treatment cells were seeded into 24-well plates at a density of 5x10 4 cells cm -2 in 0.5 mL growth media.pDNA lipoplexes were prepared in EpiLife media and cells were treated as detailed in 2.5.3.After 6 h the treatment was removed and the cells were washed three times with PBS.EpiLife media, supplemented with 10% HGKS, was then added to the cell populations which were incubated for a further 42 h.All experiments were performed at least in triplicate.

Examining gene expression in vitro
Following a 42 h incubation period (Section 2.5.3 and 2.5.4), cell populations were washed three times with PBS before addition of EpiLife imaging media (Life Technologies, UK) to permit live cell imaging.The presence of GFP and TdTomato in cell monolayers was visualised using the Leica DM IRB epifluorescence microscope and imaging system (Leica Microsystems Ltd, UK).The transfection efficiency of pEGFP-N1 and pIRES-TdTomato was determined by flow cytometry using the FACSVerse flow cytometry system (BD Biosciences, UK).To prepare for FACS analysis cells were washed with PBS and then treated with trypsin-EDTA at 37°C for 10 min.The cell suspension was then re-suspended in growth medium, centrifuged (Thermo Fisher Scientific, UK) at 300 × g for 5 min and washed with PBS.Pelleted cells were fixed with 100 μL of fixation buffer (BD Biosciences, UK) for 15 min at 4°C, washed and re-suspended in 200 μL of staining buffer (BSA), (BD Biosciences, UK) before flow cytometry analysis.GFP fluorescence was detected using a 527/32 filter, while TdTomato used a 568/42 filter (PE channel).On each sample, a minimum of 10,000 events was collected.Analysis of collected data was performed with the FlowJo Flow Cytometry Analysis Software for Mac Version 10. (Tree Star Inc., USA).
The percentage of gated cells in each sample indicates the percentage of fluorescent cells.

MN mediated delivery of pDNA in vitro
Twenty four hours before transfection HaCaT cells were seeded into 24-well plates at density of 7.5x10 5 cells cm -2 in 0.5 mL growth media.Cell populations were then treated using: (i) naked plasmid (pEGFP-N1; 5 μg per well) injected into the media by pipette, (ii) naked plasmid (pEGFP-N1; 5 μg per well) injected into the media by MNs and (iii) pDNA lipoplex (pEGFP-N1; 5 μg per well) formed with Lipofectamine2000 transfection reagent (prepared according supplier's recommended protocol) as a positive control.Transfection was performed using a comparable protocol to that described in Section 2.5.3.The transfection efficiency of pEGFP-N1 in cells was determined by fluorescent microscopy and flow cytometry (Section 2.5.5).
2.7 Gene expression in ex vivo human skin 2.7.1 Conventional and MN injection of pDNA into human skin Following collection and preparation of excised skin (Section 2.4) 50 μL of pDNA (pCMVβ and pEGFP-N1) diluted in PBS, at a range of concentrations (detailed in results), was injected into the tissue using MicronJet MNs or a traditional hypodermic needle (BD Microlance 3, 26G x 3/8").50 μL of PBS was injected as a negative control.Following MN injection, the treatment sites were excised using 8 mm biopsy punches and incubated for 24 h at the air-liquid interface at 37°C and 5% (v/v) CO 2 /95% (v/v) air.Injections were performed at least in triplicate.

Lateral MN disruption of human skin
Split-thickness human skin was pinned onto a cork dissection board (Section 2.4).20 μl of pDNA solution (pCMVβ) at concentration of 3.5 mg/ml was applied to the skin surface.A 4x4 array of MN-like structures, manufactured using silicon wet etching techniques (Wilke et al. 2005), was provided by the Tyndall National Institute.In this instance the wafer was removed from the etchant early in order to leave a flat uppermost surface, Figure 8D (insert), instead of the nm-sharp tip associated with conventional microneedles (O'Mahony, 2014).The flat tip of this truncated conical structure (more accurately termed a frustum tip) helps to enhance the skin disruption and maximise reliability when scraped laterally across the skin surface.This array was then applied in a downward motion to the edge of the treated area, under constant pressure, for a period of approximately 2 secs and subsequently moved laterally across the skin surface.(Olympus, Middlesex, UK).Selected slides were subsequently stained using haematoxylin and eosin (H&E) to visualize the skin architecture (Fischer et al. 2008).

Intravital fluorescence imaging
Skin treated with pEGFP-N1 plasmid was incubated for 24 h at 37°C, rinsed in PBS and fixed in 4% paraformaldehyde/PBS at 4°C for 24 h.Skin samples were then rinsed in PBS and imaged intravitally using the Kodak FX-Pro imaging system (Bruker Corporation), with wavelengths set to 470 nm (excitation) and 535 nm (emission).

Fluorescence microscopy of skin sections
Following 24 h incubation, skin samples were rinsed in PBS and fixed in 4% paraformaldehyde/PBS at 4°C for 24 h.Skin was then cryosectioned (2.7.3.2) and imaged using the Leica DM IRB epifluorescence microscope and imaging system (Leica Microsystems Ltd., UK).

Preparation of epidermal sheets and fluorescent microscopy
Skin treated with pEGFP-N1 was incubated for 24 h at 37°C before epidermal sheets were isolated by heat separation.Skin biopsy punches were incubated in water at 60°C for 55 seconds after which the epidermis was removed from the dermis using forceps under a dissection microscope.Isolated epidermal sheets were then washed in PBS, fixed in acetone for 20 min at -20°C, washed again and mounted on slides using Dako fluorescent mounting medium (Dako, UK).The tissue was then visualised by a Leica DMI6000B confocal microscope (Leica Microsystems (UK) Ltd, UK) with LAS AF software (Leica Microsystems (UK) Ltd, UK).

Fluorescence analysis of skin cells by flow cytometry
Excised skin was maintained in organ culture for 24 h prior to removal of the lower dermis by blunt dissection (Ng et al. 2009).The remaining tissue was immersed in an enzyme solution containing Dispase II (2 units/ml) (Roche, UK), Collagenase I (197units/ml) and Deoxyribonuclease I (DNA-se I) from bovine pancreas (20 units/ml) in RPMI 1640 (all Invitrogen, UK) for 1 h at 37°C.Epidermal sheets were then mechanically separated from the dermis using forceps and placed in 0.25% (w/v) solution of trypsin for 30 min at 37°C.Following incubation, the epidermal sheets were pipetted up and down to increase cell yield and the suspension of the cells was passed through a 70 μm cell strainer.Single cell suspensions were subsequently centrifuged at 400 x g for 10 min and washed in PBS.Pelleted cells were fixed with 100 μL of fixation buffer (BD Biosciences, UK) for 15 min at 4°C, washed and re-suspended in 200 μL of stain buffer (BSA), (BD Biosciences, UK) before flow cytometry analysis (Section 2.5.5).

Optical coherence tomography (OCT) of human skin following MN delivery
Excised human skin was removed from -20°C storage and allowed to equilibrate to room temperature for approximately 2 h.Sub-cutaneous adipose tissue was removed by blunt dissection and skin was pinned, dermis side down, onto a cork dissection board to facilitate injection.Real time injections of methylene blue were visualised using the VivoSight System (Michelson Diagnostics, UK).The system is a Swept Source Fourier Domain OCT with clinical CE and FDA approval.This MultiBeam OCT system with <7.5 µm lateral resolution and <5 µm axial resolution in tissue is capable of achieving between 1.2 and 2.0 mm imaging depth and a 6×6 mm lateral field of view.Each OCT scan consists of 60 B-scans (in the x-y plane) with 4 µm pixels, separated by 100 µm in the y direction.

Data processing
Image processing including scale bar inclusion was performed ImageJ software (National Institute of Health, USA).Statistical analysis of differences between transfection was determined by t-test analysis using GraphPad Prism software version 6.0c.

Characterisation of MN devices
MicronJet is a CE marked and FDA cleared device consisting of three MNs, each 600 μm in length with a lumen that is approximately 60 μm in diameter, bonded to a plastic adapter that permits attachment to a Luer fit syringe (Fig 1).

Discussion
There is a growing body of research that describes delivery of pDNA to murine (Gill et al. 2010, González-González et al. 2011, Kumar et al. 2012, Kim et al. 2014, Song et al. 2012) and other animal (DeMuth et al. 2013, Yan et al. 2014) models using solid MNs, particularly in the context of vaccination.Whilst these studies provide valuable information related to biological end-points and also exemplify the potential of solid MNs for pDNA delivery, the gross architecture and biology of animal and human skin have acknowledged differences, including the density, locations, proportions and phenotypes of both epidermal and dermal cells.Studies have also evaluated pDNA delivery in a viable ex vivo human skin model (Ng et al. 2009), but whilst studies have demonstrated proof-of-concept (Birchall et al. 2005, Coulman et al. 2006, Daugimont et al. 2010, McCaffrey et al. 2015, Pearton et al. 2012, Song et al. 2012), the 'soak and poke' and 'coat and poke' strategies are associated with limitations related to dosing reproducibility and transfection efficiency.
In this study hollow MNs (Figure 1) were used, for the first time, for the minimally invasive delivery of a pDNA solution into human skin.Experiments using two different reporter plasmids indicate that following MicroJet delivery of a known bolus dose of pDNA into human skin, transfection is more extensive and reproducible (Figure 3, 4) than that previously observed using the "soak and poke" (Birchall et al. 2005, Coulman et al. 2006, Pearton et al. 2008, Yan et al. 2014) or "coat and poke" MN approaches (DeMuth et al. 2013, Gill et al. 2010, González-González et al. 2011, Kim et al. 2013, Pearton et al. 2012, Song et al. 2012).Most notably, the punctate areas of intense gene expression are remote from the injection site.This suggests that the MicronJet device facilitates transfection by a different mechanism than other published MN delivery strategies.A series of experiments were therefore conducted to elucidate the mechanism by which gene expression is enhanced upon MicronJet delivery of reporter pDNA to ex vivo human skin.
Initial studies used a well-characterised immortalised human keratinocyte cell line (HaCaT cells) to exemplify the functionality of the reporter pDNA.Cell delivery using a commercial non-viral transfection reagent produced transfection rates of over 40% (Figure 2), however this did not translate to primary keratinocytes (Figure 2).The nuclear envelope is the major intracellular barrier to nuclear delivery of pDNA (Dean et al. 2005) and therefore transfection efficiencies are greatest in rapidly dividing cells, where the nuclear envelope is briefly but repeatedly compromised during mitosis (Escriou et al. 2001, Mortimer et al. 1999, Nicolau and Sene 1982, Wilke et al. 1996).Primary cells proliferate slowly and therefore transfection efficiencies in these cells are reduced.In terminally differentiating primary keratinocytes, where nuclear entry relies on passive translocation via the nuclear pore complex (Mattaj and Englmeier, 1998, Stoffler et al. 1999, Wilson et al. 1999), this problem is exacerbated.Therefore whilst immortalised keratinocytes and non-viral vectors are useful laboratory tools, the more representative primary keratinocyte cells used in this study exemplify the cellular barriers to transfection that exist in vivo.
Treatment of the skin with a solid MN temporarily disrupts the stratum corneum and the plasma membranes of underlying viable cells.It has therefore been hypothesised that pre-treatment of the skin surface with a topical pDNA solution prior to application of a MN device ("soak and poke") can facilitate almost simultaneous passage of pDNA across the skin barrier and into the cytoplasm of the cells in the viable epidermis (Pearton et al. 2012, Yan et al. 2014).If the cell membrane recovers after this mechanical trauma, and cell viability is maintained, intracellular delivery of pDNA will translate to expression of the exogenous gene product.The drawback of this mechanism of delivery is the exclusive proximity of positively-expressing cells to the MN insertions, which necessitates extensive or repeated MN applications to facilitate reasonable levels of gene expression."Coat and poke" (Pearton et al. 2012) and "poke and soak" (Chabri et al. 2004, Mikszta et al. 2002, Pearton et al. 2008, Yan et al. 2014) strategies produce lower levels of gene expression due to the time delay between MN induced disruption of the cell membranes and exposure of the compromised cells to the pDNA.This delay is dependent on either the dissolution rate of the relatively insoluble macromolecular nucleic acid within the local environment of the skin tissue ("coat and poke") or the interval between treatment of the skin with the MN device and topical application of a pDNA solution ("poke and soak").Therefore a significant observable enhancement in both the level and distribution of gene expression following the use of hollow MNs, compared to their solid counterparts, suggests that this method of MN delivery improves cellular delivery as well as providing more reproducible and controlled dosing into the tissue.
Delivery of a liquid bolus into human skin using a hollow MN may be considered to be more comparable to a conventional intradermal injection than solid MNs.
However, the level and location of gene expression observed in ex vivo human skin following delivery of pDNA by hollow MNs is also significantly greater than that achieved using conventional intradermal injection (Figure 8B).Conventional intradermal injections use a 260 m needle bore and expel the liquid bolus in a single stream, compared to Micronjet, which uses three 60 m bores.The smaller bore sizes of the MicronJet increase the exit velocity of the liquid when a comparable downward pressure is applied to the syringe.The increased velocity is likely to facilitate greater direct cell transfection.In vitro experiments partially exemplified this point (Figure 6), with increased transfection evident when the naked pDNA was applied to cell monolayers using the MicronJet device compared to a pipette tip, although absolute gene expression levels remain low.Direct cell transfection by a high velocity liquid is also supported by a significant body of literature related to jet injections (Cartier et al. 2000, Fargnoli et al. Fargnoli et al. 2016, Furth et al. 1995, Sawamura et al. 1999, Walther et al. 2001, Walther et al. 2002).However, although a jet injection mechanism may contribute to the heightened gene expression observed in this study, the limited number of jets and the extensive distribution of gene expression suggest that it is unlikely to be the sole or major reason for the enhancement.
pDNA delivery using hollow MNs results in gene expression in cells remote from the point of injection, but which are restricted to the 'bleb' that is created by the intradermal liquid bolus (Figures 3 & 4).This 'bleb', as it is commonly referred to, is a discrete area of skin tissue that temporarily expands in response to the intradermal administration of a liquid volume.This type of event is also witnessed during the hydrodynamic delivery of nucleic acids to animal models (Liu et al. 1999, Zhang et al. 1999, Bonamassa et al. 2011, Suda et al. 2007a), where a high speed and large volume tail vein injection of pDNA causes the liver to more than double in size (Suda et al. 2007b).Delivery of a large volume of liquid into the confined tissue space of the liver results in hydrostatic pressures that create temporary pores in the plasma membranes of cells, thus increasing permeability and heightening transfection.In the same human skin model using the hollow MN device results in a significant enhancement to the level of exogenous gene expression (Figure 3).It is therefore the specific delivery system i.e. the hollow MN that is responsible for the increase in gene expression witnessed in this study.
Real time OCT videos and supportive microscopy images (Figure 7 & Suppl.Figure 1A-C) indicate that hollow MN delivery of the liquid bolus is more superficial than that of conventional intradermal injection.This results in multiple microdisruptions to the skin architecture in the papillary dermis and at the epidermal-dermal junction (Figure 7 and Suppl.Figurures 1A-C).These disruptions were created during the MN injection process (Figure 7B and Suppl.Figure 1B) but remain following completion of bolus delivery (Figure 7C and Suppl.Figure 1C).The disruptions were located at numerous points that are distant from the point of MN injection but within the confines of the bleb.Some of these changes were also co-localised with points of transfection (Figure 7D).These disruptions were not observed upon conventional intradermal injection (Figure 7A and Suppl.Figure 1A).We therefore postulate that, unlike conventional intradermal injections, hollow MNs are able to force large volumes of liquid (relative to the skin volume) at high speed (typically less than few seconds to administer 50 l) into the papillary dermis, where elastin and collagen fibres are tightly packed and are in direct contact with the viable cells of the epidermal basement membrane.This creates hydrostatic pressures that physically disrupt the local tissue architecture and cause a transient increase in the permeability of cells in this region, which results in notable gene expression located exclusively with the skin 'bleb' (Figure 3 & 4).This proposed mechanism of transfection is akin to the hydrodynamic delivery witnessed in animal models, the main difference being that we have used a hollow MN device to replicate this in human skin tissue and have achieved it using relatively low injection volumes (50 l).
The pattern of gene expression following hollow MN mediated delivery of pDNA was not uniform.Discrete areas of gene expression were observed, predominantly in the viable epidermis but also in the dermis (Figure 3-5).The disparate pattern of gene expression is unsurprising given the proposed mechanism of transfection, i.e.
hydrodynamic delivery is unlikely to create a uniform pattern of disruption in a heterogeneous tissue architecture.Indeed, OCT images and histology (Figure 7) support this, highlighting discrete areas of mechanical tissue disruption in the papillary dermis and epidermis, rather than uniform tissue disruption.Lateral application of a MN device (Mikszta et al. 2002) was used in the human skin model as an alternative method to create extensive superficial physical disruptions in the viable epidermis and papillary dermis (Figure 8 D, G-H).Delivery using this method also produced a random pattern of discrete areas of gene expression proximal to those points of mechanical disruption.Whilst lateral microneedle application is not clinically useful, as it causes significant mechanical damage to the stratum corneum barrier, it provides an interesting comparator and supports the hypothesis that intradermal delivery of a pDNA solution through the MicronJet system also physically disrupts cell membranes in the papillary dermis and viable epidermis by a hydrodynamic effect, in this case without compromising the protective stratum corneum barrier.
Transfection in the excised human tissue is therefore considerable following pDNA delivery by the MicronJet, but when skin cells are subsequently extracted from the bleb the relative number of transfected cells is low (Figure 5).This is likely a compound factor of the inherent difficulty in transfecting the dominant cell type of the epidermis i.e. primary terminally differentiating cells (Figure 2), restricted access of pDNA to all cells within the three dimensional skin architecture and, potentially, damage to transfected cells during extraction from the tissue.Successful transfection also relies on transient disruption of the plasma membrane, but in a proportion of cells the physical disruption is likely to be permanent and thus cause cell death.
Therefore whilst the data provides an insight into the numbers of cells transfected it is likely to be an under-estimate of the transfection efficiency i.e. of the number of cells expressing the exogenous gene as a proportion of those cells that are able to be transfected.Additionally, levels of localised gene expression are significantly greater than that achieved by conventional intradermal injections or other MN delivery technologies and therefore may be suitable where high transfection efficiencies are not required e.g.nucleic acid vaccines or immunotherapy approaches.
The observable enhancement in performance, compared to conventional intradermal injection and solid MNs, encourages development of hollow MN delivery systems for nucleic acid therapies.Whilst electroporation is now being investigated as an adjunct to enhance intracellular delivery of pDNA following the application of solid MNs (Wei et al. 2014), the hydrodynamic mechanism of cellular delivery associated with use of the hollow MN device may provide a simpler alternative.To the best of our knowledge, this is the first time that hydrodynamic gene delivery has been achieved in human skin and therefore further work is now required to optimise delivery and translate these results from reporter genes to therapeutically relevant pDNA.The extent and reproducibility of exogenous gene expression in human skin indicates that the CE marked MicronJet device may therefore provide a valuable platform for nucleic acid vaccination strategies or immunotherapy approaches.It is also important to note that whilst multiple sub-surface microdisruptions created by the MicronJet in the skin are able to facilitate cellular uptake and subsequent expression of pDNA, there have been thousands of MicronJet injections with no reports of associated infection and therefore the enhancement in transfection is not at the expense of the minimally invasive principles of the MN delivery system.

Conclusions
This study exemplifies, for the first time, the use of hollow MNs for efficient vectorless delivery and expression of pDNA in human skin and proposes hydrodynamic delivery as the mechanism for enhanced gene expression, even with a 'standard' injection volume.Hydrodynamic gene delivery has not previously been observed in human skin tissue.The approved MicronJet hollow MN system ensures accurate delivery of a finite dose of exogenous DNA into the skin and provides more efficient and reproducible gene expression compared to solid MN delivery strategies.The delivery method provides potential advantages over other chemical and physical gene delivery methods with respect to cost, simplicity and flexibility of dosing and therefore could be the basis for future nucleic acid based immunotherapies and vaccines.
Pl e a s e n o t e: C h a n g e s m a d e a s a r e s ul t of p u blis hi n g p r o c e s s e s s u c h a s c o py-e di ti n g, fo r m a t ti n g a n d p a g e n u m b e r s m a y n o t b e r efl e c t e d in t his ve r sio n.Fo r t h e d efi nitiv e ve r sio n of t hi s p u blic a tio n, pl e a s e r ef e r t o t h e p u blis h e d s o u r c e.You a r e a d vis e d t o c o n s ul t t h e p u blis h e r's v e r sio n if yo u wi s h t o cit e t hi s p a p er. Thi s v e r sio n is b ei n g m a d e a v ail a bl e in a c c o r d a n c e wit h p u blis h e r p olici e s. S e e h t t p://o r c a .cf. a c. u k/ p olici e s. h t ml fo r u s a g e p olici e s.Co py ri g h t a n d m o r al ri g h t s fo r p u blic a tio n s m a d e a v ail a bl e in ORCA a r e r e t ai n e d by t h e c o py ri g h t h ol d e r s .
transfection of immortalised and primary keratinocytes Immortalised keratinocytes (HaCaT) were used to exemplify the functionality of the pEGFP-N1 and pIRES-TdTomato reporter plasmids.Freshly isolated human keratinocytes were subsequently used to determine transfection efficiency as these cells are more representative of the in vivo environment.A significant proportion (44.1 ± 0.2%) of HaCaT cells were positive for gene expression following treatment with the pEGFP-N1:Lipofectamine2000 lipoplex (Fig 2A).The detection of gene expression was reduced (10.1 ± 0.4%) in cells transfected with pIRES-TdTomato:Lipofectamine2000 (Fig 2D).Gene expression efficiency in freshly isolated primary skin cells was reduced by an order of magnitude, with 3.8 ± 0.4% of the cell population transfected by pEGFP-N1 lipoplexes, and 0.8 ± 0.1% for pIRES-TdTomato lipoplexes.Positive transfection by GFP and TdTomato was confirmed by fluorescence micrographs of HaCaT cells (Figure 2B and 2E) and primary keratinocytes (Figure 2C and 2F).

Fig. 3 .
Fig. 3. Images of ex vivo human skin following MicronJet 600µm injection of (A) PBS as a negative control or (B-H) naked pCMVß plasmid (250 μg in 50 μl), followed by incubation in X-gal solution.(A, B, C) provide en face images of the skin surface and (D) is the underside of a skin sample i.e. viewed from the dermis.Transverse cryosections (10µm) illustrate the (E) microneedle insertion site and successful transfection of pCMVß plasmid in the (F) epidermis, (G) epidermis and dermis and (H) dermis of the skin.Areas of transfection are highlighted by arrows.Donor skin in these exemplar images was obtained from 71 (A-C, E-H) and 54 years old females.
experiment ex vivo human skin was treated with pEGFP-N1 delivered via MicronJet MNs 24 hours prior to epidermal cells being extracted from the tissue and analysed by fluorescence activated cell sorting (FACS).FACS plots indicate that a small population of the extracted epidermal cells were GFP positive (Figure5B).Positive expression of the reporter protein in these skin-derived cells is further supported by fluorescence microscopy 5D).

Fig. 5 .
Fig. 5. Detectable GFP in epidermal cells that were isolated from freshly excised human skin following MicronJet 600 µm injection of naked pEGFP-N1 (250 μg).Control skin samples received an injection of PBS.Flow cytometry plots of (A) control epidermal cells and (B) epidermal cells transfected with pEGFP-N1 plasmid.Fluorescence micrographs of (C) a control epidermal cell suspension and (D) positive GFP expression within the epidermal cell suspension that has been isolated from human skin post-treatment.Donor skin was obtained from a 67 years old female.

Fig. 6 .
Fig. 6.Representative flow cytometry plots and accompanying micrographs to illustrate pEGFP-N1 expression in HaCaT cells following treatment of the cell monolayer by different delivery strategies.Gene delivery was facilitated by (A) topical naked plasmid (5 μg) delivery from a pipette, (B, D) topical lipoplex (pEGFP-N1 plasmid with Lipofectamine2000) delivery from a pipette (positive control) or (C, E) naked plasmid (5 μg) extruded on to the cell monolayer through a MicronJet 600µm.
Figure 7B and 7C).These microdisruptions appeared in a region of the tissue that is approximately 200-500μm below the skin surface and are a feature that were not Figure 1A).Subsequent histological examination of the tissue following injection of pCMVβ supported OCT images, showing in some cases positive gene expression localised to areas of significant tissue disruption (Figure 7D, 7E and 7F).

Fig
Fig. 7. (A, B,C) OCT images highlight (A) untreated and (B) microneedle treated ex vivo human skin, both (B) during and (C) after a 50μl microneedle injection of methylene blue using MicronJet 600µm needles.In (B) the microneedle can be identified adjacent to the bleb that is being created.The dashed orange line highlights the

Fig. 8 .
Fig. 8. En face images of ex vivo human skin treated with naked pCMVß plasmid and stained with X-gal solution.Skin samples were intradermally injected using a hypodermic needle to deliver (A) PBS (negative control), (B) 250µg of pCMVß in 50µl animal models this facilitates gene expression in large numbers of cells in the liver and detectable levels of gene expression in other internal organs.This technique has also been transferred to animal skin, where rapid intradermal injection of a large volume (up to 100µl) of pDNA solution into a mouse foot pad facilitates significant gene expression(González-González et al. 2010).González-González et al. suggest    that the presence of nucleic acid in the epidermis is not enough to facilitate keratinocyte uptake and propose that the hydrodynamic pressure created by intradermal injection of a relatively large volume of liquid into the confined space of the mouse foot pad causes transient membrane disruption of skin cells and facilitates cell uptake and subsequent plasmid expression.To date, hydrodynamic delivery has not been successfully used to facilitate transfection in human skin.In human skin conventional intradermal injection of a pDNA solution also results in observable expansion of the skin tissue.However there is limited gene expression and this is restricted to the point of injection and/or the needle track (Figure8A-C; E-F).The extent and pattern of transfection following conventional injection is therefore not characteristic of a hydrodynamic effect and is more likely caused by direct jet injection at the needle orifice and/or temporary mechanical damage to the cells at the point of needle insertion.The pressure created in the human tissue by conventional intradermal injection of 50 µl of a pDNA solution is therefore not great enough to facilitate extensive transfection in the cellular region of the skin.However delivery of the same injection volume (50 µl) containing the same reporter plasmid (pCMVB) to