Hydrogel-Forming Microneedle Arrays for Enhanced Transdermal Drug Delivery

Unique microneedle arrays prepared from crosslinked polymers, which contain no drug themselves, are described. They rapidly take up skin interstitial fluid upon skin insertion to form continuous, unblockable, hydrogel conduits from attached patch-type drug reservoirs to the dermal microcirculation. Importantly, such microneedles, which can be fabricated in a wide range of patch sizes and microneedle geometries, can be easily sterilized, resist hole closure while in place, and are removed completely intact from the skin. Delivery of macromolecules is no longer limited to what can be loaded into the microneedles themselves and transdermal drug delivery is now controlled by the crosslink density of the hydrogel system rather than the stratum corneum, while electrically modulated delivery is also a unique feature. This technology has the potential to overcome the limitations of conventional microneedle designs and greatly increase the range of the type of drug that is deliverable transdermally, with ensuing benefits for industry, healthcare providers and, ultimately, patients.


Introduction
Microneedle arrays (MN) are minimally-invasive devices that painlessly by-pass the skin's stratum corneum, which is the principal barrier to topically-applied drugs. [ 1 ] MN (50-900 μ m in height, up to 2000 MN cm − 2 ) have been extensively investigated in recent years as a means to enhance transdermal drug and vaccine delivery. [2][3][4][5][6][7][8][9][10] No MN array-based drug delivery systems are presently marketed, due to the relatively recent developments that have made such devices a reality. Indeed, the subsequently rapid progression of the fi eld has inevitably led to some difficulties: silicon, the most commonly-employed MN material to date, has dubious biocompatibility and broken silicon or metal MN could cause skin problems; [ 5 ] solid, non-coated MN require a two-step application process, which is undesirable for patient use [ 5 , 6 ] ( Figure 1 a); accurately coating MN is diffi cult and these coated MN subsequently only deliver a very small amount of drug as a bolus [ 2-4 , 6 ] (Figure 1 b); MNinduced holes normally close over very quickly ( < 1 h). The rate of hole closure can be slowed by heavy occlusion, but cannot be prevented; [ 7 ] biomolecules can be signifi cantly degraded by the heating used to produce polymeric MN from molten polymers or carbohydrates [ 2-4 , 8 ] (Figure 1 c); and hollow MN have only one outlet and can become blocked by compressed dermal tissue. [2][3][4] (Figure 1 d).
Recently, MN loaded with drug have been prepared under ambient processing conditions from aqueous gels based on FDA-approved polymers. [ 9 ] Such systems, which dissolve in skin interstitial fl uid to release their drug payload, have overcome many of the above diffi culties. However, while transdermal delivery of small ( < 500 daltons) water-soluble molecules from both MN and baseplate is possible with such systems, only macromolecules (e.g., insulin) located in the MN themselves are deliverable. [ 10 ] In addition, while the polymers employed for fabricating these MN have long-established safety profi les, they have never before been administered intradermally. Regulatory authorities will inevitably require extensive studies investigating routes of metabolism and elimination, possibly delaying commercialisation. Since MN materials will encounter viable skin cells, unlike conventional transdermal and topical drug delivery systems, regulators may also require MN systems to be sterile. Here we describe unique MN arrays prepared from crosslinked polymers which contain no drug themselves, as shown in Figures 1 e-h. Instead, they rapidly take up skin interstitial fl uid upon skin insertion to form continuous, unblockable, hydrogel conduits from patch-type drug reservoirs to the dermal microcirculation (Figure 1 f,g). Importantly, such MN, which can be fabricated in a wide range of patch sizes and MN geometries, can be easily sterilised, resist hole closure while in place and are removed completely intact from the skin (Figure 1 h). Delivery of macromolecules is no longer limited to what can be loaded into the MN themselves and transdermal drug delivery is now controlled by the crosslink density of the hydrogel system rather than the stratum corneum, while electrically modulated delivery is also a unique feature.

In Vitro Permeation Studies
The ability of the novel hydrogel-forming MN in enhancing and controlling transdermal drug delivery was investigated by using six hydrophilic solute molecules of increasing molecular weight: fl uorescein-isothiocyanate labelled bovine serum albumin (FITC-BSA), insulin, methylene blue (MB), caffeine (CF), theophylline (TP), and metronidazole (MZT) with molecular weights of ≈ 67,000, ≈ 6000, 319. 85, 194.19, 180.17, 171.15 Daltons, respectively. For studies involving high-molecular weight molecules, a pore-forming agent (NaHCO 3 ) was added to the hydrogel formulation. Adhesive patches, as stated below, containing the above molecules at defi ned loadings, were then attached to the upper baseplates of hydrogel-forming MN, with the novel composite system termed as "integrated MN" ( Figure 2 a). Permeation was then investigated from this integrated MN system and compared with that of the adhesive patches alone.
Adhesive patches were prepared, by using a casting method, from aqueous blends of 10% w/w PMVE/MA and 5% w/w tripropyleneglycol methyl ether (TPM). [ 28 ] Required amounts of solute molecules were either added directly into the aqueous blends or dissolved in 0.01 M HCl prior to addition to the aqueous blends. An aliquot (2.7 g) of aqueous blend was cast

Materials and Methods
This study comprehensively investigated the utility of our novel hydrogel-forming MN arrays in transdermal drug delivery. Mechanical properties, in vitro/in vivo permeation of model drug substances, combined effect with iontophoresis (ITP), biocompatibility, and safety in human subjects were investigated. The methods for each study are detailed below.

Preparation and Mechanical Testing of MN
Aqueous blends containing 15% w/w poly(methylvinylether/ maelic acid) (PMVE/MA) and 7.5% w/w poly(ethyleneglycol) 10,000 (PEG) were utilized to fabricate MN by using laser-engineered silicone micromould templates. [ 10 , 15 , 16 ] Optimum polymeric composition was determined in our previous hydrogel work. [11][12][13] MN were crosslinked (esterifi cation reaction) by heating at 80 ° C for 24 h. [11][12][13] Mechanical properties of these array. (f) shows application of the integrated hydrogel MN patch to the skin surface. This is followed by diffusion of water into the crosslinked integrated MN patch. (g) shows diffusion of water, which causes controlled swelling of the MN arrays, forming an in situ hydrogel conduit. This further results in liberation and diffusion of drug molecules from the adhesive patch through the hydrogel MN into the skin. (h) shows that the hydrogel MN arrays remain intact, even after removal from the skin, thereby leaving no polymeric material in the skin following drug delivery. previously. [ 10 , 15 , 30 ] FITC-BSA was quantifi ed using a gradient HPLC method (Agilent Technologies 1200 Series, Stockport, UK) in which the separation was performed on a C 4 (4.6 mm × 50 mm, 5 μ m) analytical column (Symmetry 300, Waters Ireland, Dublin). The mobile phase gradient consisted of 0.1% triforoacetic acid (TFA) in acetonitrile (ACN) and 0.1% TFA aqueous solution. The gradient linearity changed from 20:80 (ACN: TFA) to 90:10 in 8 min and continued for 1.0 min at the ratio of 90:10. The reversal to the initial conditions was attained within 1.0 min, followed by a 3.0 min re-equilibration period. The total analysis run time was 13.0 min. The injection volume was 20 μ l and elution fl ow rate was 1.0 ml min − 1 . FITC-BSA was detected using a fl uorescence detector with excitation and emission wavelengths set at 490 nm and 520 nm, respectively. Data were processed using Agilent Rapid Res software. CAF was analyzed following modifi cation of a literature method. [ 31 ] Briefl y, an isocratic HPLC method was used with a reverse phase C-18 (Waters Spherisorb 5 μ m ODS 4.6 × 150 mm, Waters Ireland, Dublin) analytical column. The column was thermostated at 35 ° C. The mobile phase consisted of 90% of 0.52% glacial acetic acid and 10% of 50% ACN: 50% tetrahydrofuran. The fl ow rate was 1.0 ml min − 1 . Sample injection volume was 50 μ l. UV detection was performed at 273 nm. Data were processed using Agilent Rapid Res software. MB was analysed into silicone moulds (30 × 30 mm 2 ) and dried under a gentle air fl ow for 48 h. Attaching the formed adhesive patches with gentle pressure to hydrogel-forming MN yielded the integrated MN system.
Transdermal permeation was investigated in vitro across dermatomed neonatal porcine skin (300-400 μ m thickness, previously shown by us to be the optimum skin model for prediction of in vivo performance of dissolving polymeric MN) [ 29 ] by using a modifi ed Franz-cell setup, as described previously. [ 10 , 15 , 16 ] Integrated MN were applied to the dermatomed skin by using a custom spring-activated applicator [ 16 ] at a force of 11.0 N per array. Adhesive patches were applied using gentle pressure. Sample volumes of 300 μ L were withdrawn, in triplicates, and analysed by the methods below. The release medium for FTIC-BSA was 0.05% w/w SDS in phosphate buffered saline (PBS) at pH 7.4, for insulin 0.1 M Tris buffer at pH 10 (due to stability issues) and for the other drug molecules PBS at pH 7.4 was used.

Analysis of Drug Permeation
Receptor medium samples for insulin, TP and MZT were analysed using validated analytical methods reported

In Vitro ITP Studies
We further investigated the combined effect of integrated MN and ITP on in vitro permeation of TP, MB, fl uorescein sodium (FS), insulin and FITC-BSA. In this case, TP, MB and FS were loaded at a concentration of 3 mg cm − 2 in the adhesive patches. FITC-BSA and insulin were loaded at concentration of 2.5 and 5 mg cm − 2 , respectively. The methodology for this study is similar to that stated above. For studies involving application of an electrical current, an Ag electrode was used as the anode and an Ag-AgCl electrode used as the cathode. The delivery electrode was placed directly on top of the integrated MN, whilst the return electrode was placed into the receiver medium via the side arm of the Franz cell. A commercially available power supply (Phoresor II, Iomed, Lake City, FL, USA) was used to deliver a current of 0.5 mA for a period of 6 h for in vitro studies. A sample volume of 300 μ l was collected at predetermined time intervals and was assayed by the analytical methods described above. For samples containing FS, UV-spectroscopy was used at a λ max of 497 nm.

In Vivo Studies
Delivery of insulin, FITC-BSA and MZT from the integrated MN (with 600 μ m height, 300 μ m width at base, 50 μ m interspacing, & 19 × 19 arrays) was investigated in a rat model. For studies involving insulin, diabetic Sprague-Dawley rats were selected. The methodology was similar to that in our previous publication, which involved soluble MN. [ 10 ] Positive controls were performed by subcutaneously injecting bovine insulin solution in PBS pH 7.4 at a dose of 0.2 IU per animal. Integrated MN based on adhesive patches containing 5 mg cm − 2 insulin were applied to the shaved skin (Gillette Mach 3 Power, Procter & Gamble, Hampshire, UK) on the rats' backs by using the spring-activated applicator at a force of 11.0 N per array. Skin integrity was confi rmed using transepidermal water loss measurements made using a Defl in Vapometer (Defl in Technologies Kuopio, Finland) Following application, the integrated MN were held in place for 12 h with the help of a self-adhesive silicone sheet. Blood samples were collected at pre-defi ned time intervals over a 12 h period by lateral tail vein prick and blood glucose levels (BGL) were measured using a glucometer (Accu-Check Aviva, Roche Ltd., Mannheim, Germany). Blood glucose levels were expressed as the percentage of initial BGL and calculated values were plotted against time to obtain blood glucose leveltime profi les. C max , denoting maximum% decrease in BGL, was calculated by subtracting the lowest% BGL from 100, as described elsewhere. [ 32 ] T max , denoting time required to achieve C max , was also reported.
Integrated MN containing FITC-BSA at 2.5 mg cm − 2 loading in patches, were applied to healthy rats in an analogous way to insulin-loaded patches. A 0.25 ml aliquot of blood was drawn from the tail vein and collected into heparinised tubes (Microvette CB300, Sarstedt, Leicester, UK) at pre-defi ned time intervals over a period of 24 h. Plasma samples were obtained by centrifuging the collected blood samples at 10 000 rpm for 10 min. All samples were analysed, within 2.0 h of collection using spectrofl uorimetry at an excitation wavelength of 490 nm and emission wavelength of 520 nm.
In vivo delivery of MZT from the integrated MNs was also investigated in rat models. The concentration of MZT versus time data obtained using the dried blood spot (DBS) method reported previously [ 30 ] was subjected to non-compartmental analysis (WinNonlin version 2.1, Pharsight, USA) to determine pharmacokinetic parameters, namely, area under curve (AUC), C max and T max . The steady state blood concentration (C ss ) of MZT after application of integrated MNs was calculated by using the equation: For in vivo studies, involving application of an electrical current, a circular PVC ring (diameter 2.0 cm) was secured onto the rats back at the site of integrated hydrogel MN or adhesive patch application, using cyanoacrylate glue (Loctite Ltd, Dublin Ireland). The Ag electrode anode was then placed on top of the integrated MN or adhesive patch and held in place using a layer of double-sided adhesive tape (Henkel Ltd, Cheshire, UK) and an additional layer of Micropore tape (3M Ireland, Dublin). A second PVC ring (diameter 2.0 cm) was then secured at an adjacent site, ≈ 2 cm away from the site of MN application. An aliquot of PBS pH 7.4 (1.0 ml) was then placed into the centre of this second PVC ring and an Ag-AgCl electrode was attached at the site using Micropore tape. The Phoresor™ II was used to deliver a current of 0.5 mA for a 2 h period after which the integrated MN and/or adhesive insulin/FITC-BSA patch were removed.
All animal experiments throughout this study were conducted according to the policy of the Federation of European Laboratory Animal Science Associations (FELASA) and The European Convention for the protection of vertebrate Animals used for Experimental and Other Scientifi c Purposes, with implementation of the principle of the 3Rs (replacement, reduction, refi nement).
To depict conditions pertaining to actual in vivo application of MN, a Franz-cell setup was used to expose only the MN and not the baseplate to the extraction medium. A Silescol membrane was sandwiched between the receptor and donor compartment of the Franz-cells and the hydrogel-forming MN were applied in a similar manner as stated above. The receptor compartment was fi lled with 12.0 ml of DMEM (Dulbecco's modifi ed Eagle's medium) containing 4.5 g L − 1 glucose, 2.0 mM www.afm-journal.de www.MaterialsViews.com Hydrogel-forming MN arrays were attached to waterproof plasters (Elastoplast, Beiresdorf, Hamburg, Germany) using double-sided adhesive tape. The spring loaded applicator was used to apply the MN-containing plasters Each volunteer was subjected to four interventions, a control plaster (without MN), a plaster containing MN baseplate only, a plaster containing hydrogel-forming MN (11 × 11 array) and a plaster containing hydrogel MN (19 × 19 array), as shown in Figure 6 c.
The study was conducted at a controlled room temperature of 20 ° C and a relative humidity of 45 ± 5%. The subjects were acclimatized in this room for 15 min prior to the start of the measurements. This study involved the application of waterproof plasters, as described above, to the ventral forearm on three different occasions; 1) Patches applied and removed immediately (i.e., 0 h treatment group); 2) Patches applied and removed after 2 h (i.e., 2 h treatment group), and 3) Patches applied and removed after 24 h (i.e., 24 h treatment group).
Before application of the patches the ventral forearm were cleansed with a sterile swab (Boots Pharmaceuticals, Nottingham, UK) in each case. Four square areas ( ≈ 1.5 cm 2 ) were marked on the ventral forearm of each subject using a ballpoint pen. The square areas were located at equidistance on each forearm (Figure 6 c). Before the application of the plasters, TEWL and clinical photographs of skin were recorded, as described below. Following application of the plasters, the applicator was activated against each plaster to ensure consistent MN penetration in each volunteer (confi rmed using OCT, as described previously). [ 16 ] Immediately upon application, and after removal of the plasters (i.e., in 0, 2 and 24 h treatment groups), the VAS, TEWL and any change in skin colour was recorded in all the four interventions, as follows;

Visual Analogue Scale (VAS) Pain Scores
Pain sensation associated with plasters was determined on a VAS. The pain intensity rating (i.e., VAS) was measured immediately after plaster application and immediately upon removal. The VAS was determined by asking the volunteers to mark on a 10 cm line, anchored by word descriptors at each end ("No Pain = 0 cm" and the other end the "Worst Pain Imaginable = 10 cm). The volunteers mark on the line the point that he or she feels represents their perception of their current state. The VAS score is determined by measuring in centimeters from the left hand end of the line to the point that the volunteer marks. [ 33 ]

Transepidermal Water Loss
TEWL was measured to determine the level of disruption to skin barrier function following application of the plasters. The Vapometer was used to measure TEWL at a control site (previously marked with squares of ≈ 1.5 cm 2 areas) pre-and post-application of plasters. TEWL measurements have been established as a routine method for evaluating the integrity of skin, which has been subjected to either physical or chemical treatment. [ 34 , 35 ] TEWL measurements were taken by carefully resting the TEWL probe horizontally on the application site, with the probe head vertical and perpendicular to the skin. L-glutamine, 10% foetal calf serum, 1.0 IU ml − 1 penicillin and 1.0 μ g ml − 1 streptomycin. After 24 h of application, the DMEM solution was collected and sterile fi ltered. The medium (100%, 50% and 10% diluted in fresh DMEM) was exposed to cell monolayers of Balb3T3 and NRERT-1 keratinocytes at 1 × 10 4 cells ml − 1 in the 96-well culture plate for 24 hrs. Following 24 h incubation at 37 ° C, the test extracts were assessed for cell viability by means of the quantitative MTT assay at 540 nm (Biotrak II Plate Reader, Amersham Biosciences, Cambridge, UK). A total of six replicates were studied for each extract, including control (sterile DMEM media) in the same plate. In 3D keratinocyte organotypic raft type culture studies, the 3D keratinocyte tissues were pre-incubated (at 37 ° C and 5% CO 2 ) for 1.0 h in assay medium before starting the experiment. Prior to dosing with MN extract the assay medium was renewed. An aliquot (50 μ L) MN extract (undiluted) was applied atop the tissues in triplicate and incubated for 24.0 h. At the end of the exposure period, the tissue cultures were gently rinsed with PBS, incubated in fresh medium for a further 24.0 h and fi nally placed into plates containing 0.5 mg ml − 1 of MTT solution at 1.0 ml cm − 2 cultured tissue (4.0 ml). Following 3.0 h incubation and a rinse with PBS, the reduced MTT was extracted by submerging the tissues in 2.0 ml of isopropanol and shaking for 2.0 h. The absorbance of the extraction solutions was again measured at 540 nm. The mean absorbance was calculated for each tested sample (undiluted MN extract). The mean result of the negative control (H 2 O) tissues was set to represent 100% viability.

In Vitro 3D Skin Irritancy Test
In vitro skin irritancy tests were performed by using a 3D reconstructed human skin model (EpiSkin, Skin Ethic Laboratories, Lyon, France). The assay endpoint for irritancy measurements was IL-1 α (Interleukin-1 alpha). An aliquot (10 μ l) of sterile-fi ltered hydrogel extract, obtained as described above, was exposed to the reconstructed epidermis of the EpiSkin skin model. A negative (PBS) and positive (5% w/v sodium dodecyl sulphate, SDS) control were also included in the protocol. In each case, the epidermis of EpiSkin skin model was exposed at room temperature, according to the manufacturer's recommendations. Following exposure for 24 h, EpiSkin™ constructs were washed with PBS, and transferred to wells that contained fresh maintenance medium, and allowed to recover for 42 h at 37 ° C, 5% CO 2 . The IL-1 α content of the recovery medium was assessed using a human IL-1 α specifi c ELISA (Thermo Scientifi c/Pierce, IL, USA).

Human Volunteer Studies
In this study, gamma-sterilised hydrogel-forming MN were applied to the ventral forearm skin of six healthy human volunteers (3 men and 3 women, aged between 23 and 31 years), with no pre-existing skin conditions. They were asked not to wash or apply any cosmetic formulations on the ventral forearm during the study period. The School of Pharmacy's Ethical Committee, Queen's University Belfast, approved this study. The camera was fi xed on a tripod and was adjusted to know height so that the distance between the ventral forearm and the camera lens were consistent throughout the study. Images were then printed on high quality photographic paper and were evaluated and scored blindly by two-experienced dermatologists.
A clinical score for each test site was assigned, using the ICD scores. ICD scores were rated as 0, 0.5, 1, 2 or 3, which represent either: a negative, a barely perceptible macular erythema, a mild erythema, a moderate-intense uniform erythema or an intense erythema, vesiculation or erosion, respectively. [ 36 , 37 ] Optical coherence tomographic (OCT) studies, which allowed measurement of depth of MN insertion and permitted study of in-skin swelling, were conducted as described previously. [ 16 ]

Statistical Analysis
Physical characterisation data and in vitro drug permeation experiments were analysed using a One-Way Analysis of Variance (ANOVA), with Tukey's HSD post-hoc test. Comparisons for animal studies, cell culture work and human volunteer investigations were made using the Kruskal-Wallis tests and the Mann-Whitney U-test. In all cases p < 0.05 denoted signifi cance.

Mechanical Characterisation
We studied the mechanical properties of MN prepared from a hydrogel-forming formulation selected as a result of our previous work with such materials. [11][12][13] Here we report on fi ve mechanical tests; axial and transverse deformation, skin penetration, MN base plate deformation and break. Increase in axial forces caused a progressive reduction in MN height, shown in Figure 2 Figure 2 f,g shows representative digital images of MN after transverse force application. It was observed that insertion forces as low as 0.03 N per needle resulted in 100% of our MN penetrating the stratum corneum of neonatal porcine skin (thickness 400 μ m) in vitro, as evidenced by methylene blue staining in Figure 2 i. In contrast, Figure 2 j shows the force required to break the hydrogel MN base-plates was approximately 23.55 N. At break point, the MN baseplate had bent to an angel of approximately 79.28º. This refl ects not only the considerable strength of the material, but also its conformability, important since, on a micron-scale, the skin is not perfectly fl at. Hence, MN arrays must ideally have a degree of fl exibility in their baseplate to ensure all MN penetrate the skin when applied. While this is true for the system described here, the same cannot be said for silicon, where we have seen that hard brittle baseplates can fracture upon application.

In Vitro Release Studies
For successful passive transdermal transport, molecules should ideally have intermediate Log P (1-3) values and molecular weights less than 500 Daltons. [ 1 , 14 ] A wide range of techniques aimed at bypassing the stratum corneum barrier have successfully overcome these limitations. However, tape-stripping, sonophoresis, electroporation and heat/laser ablation are all much less practical than MN technology, which has been widely acclaimed for its potential to realise considerable patient benefi t in the near future. Recently, we have demonstrated a signifi cant enhancement in transdermal insulin transport in vitro using our dissolving poly(methylvinylether-co-maleic acid) (PMVE/MA) MN. [ 10 ] While percutaneous administration of these insulin-loaded MN arrays to diabetic rats resulted in a dose-dependent hypoglycaemic effect, sustained administration was limited by the fact that only the insulin contained in the needles themselves was released. This is an acknowledged problem for macromolecule-loaded dissolving polymeric MN. Since this is the type of MN system that currently shows the most promise and due to the ever-increasing number of macromolecular drugs with delivery problems, solutions are urgently required. In the present study, we have for the fi rst time, demonstrated MN-mediated sustained percutaneous delivery of both low and high molecular weight compounds. When drug-loaded patches were attached to drug-free hydrogel-forming MN and the integrated devices applied to dermatomed neonatal porcine skin in vitro, controlled drug administration was achieved, with rates and extents of permeation independent of molecular weight and Log P values ( Figure 3 ). The demonstrated ability to deliver relatively small water soluble molecules in milligram per square centimeter amounts in a sustained fashion is likely to be of great value to industry and, ultimately, patients. Most conventional drugs are small water soluble molecules and many treatment regimes would benefi t from the enhanced bioavailability and reduced dosing frequency that transdermal delivery offers. The biotechnology www.afm-journal.de www.MaterialsViews.com wileyonlinelibrary.com © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim using MN coupled with iontophoresis (ITP) may allow rapid delivery of vaccines or hormones (e.g., insulin), as well as enabling delivery of conventional drug substances to be precisely controlled. [ 17 , 18 ] Ultimately, this may enable bolus, pulsatile or responsive drug administration. Combination of MN and ITP has shown to lead to a synergistic enhancement in transdermal delivery of a range of molecules. [17][18][19][20][21] However, the necessity for a two-stage application process (MN applied and removed, then ITP applied) when using this approach may be cumbersome for clinicians and patients, particularly given the importance of ensuring that the drug-loaded formulation is placed correctly over the microporated skin area. Secondly, it is known that the process of skin closure begins almost immediately upon MN removal, unless the skin area is kept under heavy occlusion. [ 7 ] This effectively limits the time available for enhanced drug delivery. Such problems may be overcome by the amalgamation of ITP with in situ polymeric MN delivery systems. As such, the present study evaluated the potential for the combination of our novel hydrogel MN devices and ITP. Figure 4 a-e illustrates the combined effect of ITP and integrated hydrogel MN on permeation of fi ve different solutes across dermatomed neonatal porcine skin. In general, the combination of ITP with the novel hydrogel MN array led to a boom continues to produce peptide and protein drugs with great potential. However, most of these exciting new drugs must be dosed parenterally, due to their high molecular weights, extreme hydrophilicities and propensity for gastrointestinal destruction. Here we have shown sustained transdermal transport of a peptide and protein in milligram per square centimeter amounts over a 24 hour period. Importantly, we have previously demonstrated that the nature of this hydrogel formulation can easily be altered on a case-by-case basis to improve delivery, [11][12][13] suggesting utility in delivery of a range of macromolecules.

Combined Effect of ITP and Integrated MN
MN have been widely hailed as exciting alternatives to parenteral injection, complete with enhanced ease of use and reduced potential for transmission of infection. However, MNbased delivery typically relies on relatively slow drug diffusion through MN-induced holes in the stratum corneum or dissolution of drug-loaded polymeric MN or drug coatings on silicon or metal MN. This is in stark contrast to the rapid delivery possible using conventional needles and syringes. It has recently been proposed that the combination of skin barrier impairment

In Vivo Studies
In order to demonstrate the utility and advantages of our novel hydrogel MN technology, we carried out a range of in vivo rat experiments ( Figure 5 a- Sustained transdermal delivery of the high-molecular weight protein FTIC-BSA was clearly observed following application of greater rate and overall extent of transdermal delivery, in comparison to integrated MN alone. However, signifi cant enhancements were only observed for the biomolecules insulin and FITC-BSA. This was particularly notable for FITC-BSA and is understandable, given the approximate 12,000 Dalton cut-off associated with conventional ITP. [ 1 , 19 ] In an additional ex vivo experiment, optical coherence tomographic (OCT) analysis revealed that application of an electrical current leads to a dramatic increase in the swelling of the hydrogel-forming MN. It can be seen in Figure 4 f-i that, 1 h after MN insertion into rat skin, MN height increased from ≈ 610 to 650 μ m, whilst the MN base width increased from ≈ 307 μ m to 360 μ m. However, following application of an electrical current for 1 h, MN height increased to ≈ 1076 μ m (p < 0.001), and the MN base width to approximately 802 μ m (p < 0.001). This equates to an approximate 5.75-fold increase in MN surface area, in comparison to integrated hydrogel MN swelling under passive conditions. Such pronounced dimensional changes, most likely due to enhanced water uptake by electoosmosis as we have previously postulated for the PEGcrosslinked PMVE/MA system, [ 13 ] may partially explain the enhancements in delivery observed. The ability of these MN to imbibe skin interstitial fl uid is likely to fi nd use in minimallyinvasive patient monitoring, since drug concentrations in skin interstitial fl uid frequently refl ect those in plasma. [ 22 ] This would overcome many of the problems associated with direct blood sampling. www.afm-journal.de www.MaterialsViews.com wileyonlinelibrary.com literature, the minimum effective concentration for MZT is 6 μ g ml − 1 . [ 23 ] If MZT loading in the integrated MN was to be increased, the minimum effective concentration could easily be achieved over a prolonged period of time, negating the need for invasive intravenous infusion. One of the major advantages of the current hydrogel-based MN delivery system could be negating the need for hospitalisation, as required in infusion dosing. In the present study, MZT was used as a model drug. However, considering the potential of this new delivery system, drugs with narrow therapeutic indices, having a range of physicochemical properties (log P/log D) and/or exhibiting stability problems in infusion media could be effi ciently delivered with this approach.
It was again observed that application of an electrical current led to a marked increase in the rate and extent of in-skin swelling of the hydrogel-forming MN arrays. Importantly, as with in vitro studies, the integrated hydrogel MN arrays were removed fully intact following the complete period of application in all animals, regardless of whether ITP was employed or not ( Figure 5 h,i). This refl ects the retention of good mechanical strength by the swollen hydrogel and is likely to be of great importance as MN technology moves forwards towards commercialization, given that regulatory authorities, healthcare professionals and ultimately patients may have concerns about integrated MN, with peak plasma concentrations reaching 8.86 ± 1.49 μ g ml − 1 at 24 h. The combination of integrated hydrogel MN and ITP led to a signifi cantly accelerated FTIC-BSA permeation, with detectable FTIC-BSA levels found in plasma after only 30 min. Following termination of the electric current and complete removal of the MN device after 2 h of application, FTIC-BSA plasma levels were found to be 3.89 ± 0.96 μ g ml − 1 . This is ≈ 4.8 fold greater than the 0.81 ± 0.32 μ g ml − 1 detected at the same period through the use of integrated hydrogel MN alone ( p < 0.001). It was found that the FTIC-BSA plasma levels continued to rise to peak levels of 4.92 ± 1.15 μ g ml − 1 at 4 h. In contrast, application of an adhesive FTIC-BSA-loaded patch, or the use of ITP alone, did not result detection of FTIC-BSA in plasma.
The low molecular weight hydrophilic drug metronidazole (MZT) was found to rapidly appear in plasma and maintained quite constant levels over the 48 h study period. The mean pharmacokinetic parameters of MZT, administered through the integrated MN, was determined by a non-compartmental analysis of the whole blood, where the C max was 4.64 ± 1.05 μ g ml − 1 at a T max of 4.67 ± 1.15 h (Figure 5 g). The mean area under the curve (AUC 0-48 h ) was found to be 172.24 ± 24 μ g h ml − 1 . The C ss was calculated by dividing AUC 0-48 h with total study time (48 h) was found to be 3.59 ± 0.73 μ g ml − 1 . Based on the

Biocompatability
Biocompatibility of a given material can be tested by ISO and USP standards, [ 24 , 25 ] which principally depend upon exposure of material to specifi c cells in vitro, using both direct and indirect contact approaches. Alternatively, 3D human cells cultures can also be used. [ 26 ] In these methods, leachable products from the material will, if toxic, induce membrane damage and impairment of metabolic processes. The biocompatibility of our MN materials was evaluated in vitro using cell monolayers (Balb/3T3 and NRERT-1 keratinocytes) and a 3D skin model. Figure 6 a,b shows no signifi cant reduction in cell viability in all three tests. Skin irritancy studies ( Figure 6 c), showed that exposure to SDS elicited a signifi cant increase in production of the irritancy biomarker cytokine interleukin-1 α (IL-1 α ) in a 3D skin model when compared local or systemic reactions that may occur if portions of the needle were to break within the skin. It is our assertion that these integrated hydrogel MN are well-suited for sustained delivery of both small water soluble drugs and also higher molecular weight peptide and protein molecules. Pulsatile delivery can be achieved through application of ITP to meet ondemand requirements (e.g., delivery of insulin after a meal) or for rapid vaccine delivery.

Safety and Scalable Production
All drug delivery systems must not only be effective, they must also be completely safe for patient use. Accordingly, much current effort is focussed on MN prepared from FDAapproved polymeric materials, rather than silicon or metal. Such systems will also need to be shown to be safe, since administration of these polymers by the intradermal route is completely new. Sterility, or low bioburden, is likely to be a prerequisite for regulatory approval, while MN arrays must be  of the skin site were MN arrays had been applied returned to normal (i.e., 0) within 1 h of MN removal. Furthermore, MN arrays remained intact and no damage to the MN was noted during application, wear (as evidenced by OCT, Figure 6 l), or removal of MN-containing plasters.
Following removal of MN-containing plasters from the volunteers, MN dimensional changes were measured. A greater increase in MN dimensions was noted for the higher density 361 MN array in comparison to the 121 MN array ( p < 0.001). MN height increased from 597.2 ± 7.4 to 731.2 ± 6.8 μ m and MN width at base from 306.9 ± 4.5 to 379.3 ± 5.2 μ m after 24 h wear for the 361 MN array. In comparison, the MN height of the 121 array increased from 599.6 ± 6.9 to 686.8 ± 7.8 μ m, whilst the MN width at base increased from 305.8 ± 4.6 to 359.4 ± 5.6 μ m, after residing in the skin of a human volunteer for a 24 h period. Overall, these values equate to an increase in MN surface area of 35.0%, and 51.5% for the 121 and 361 MN array designs, respectively.
Transepidermal water loss (TEWL) measurements, used worldwide to assess skin barrier function showed that the skin barrier returned to normal within 30 minutes of MN removal, regardless of the treatment protocol. Figure 6 k show the results of TEWL values for the 24 h treatment group. Similar trends in TEWL values were observed in the 0 h and 2 h treatment groups.
Interestingly, we found that MN removed from skin, even after 1 min insertion, were incapable of re-insertion, due to softening by rapid imbibing of skin interstitial fl uid. This is a further important safety advantage over conventional MN designs.

Production
In order to be commercially viable, polymeric MN must be capable of large scale production. Manufacture of arrays with areas greater than 0.5-1 cm 2 will also be required in order to provide fl exibility of dosing for drugs with different therapeutic windows. Clearly, centrifuge-based mould fi lling is unsuitable for production of multiple arrays or arrays with large surface areas. In addition, there are no similar dosage forms currently produced by the pharmaceutical or medical devices industries. It is our assertion that MN manufacture should resemble production of conventional transdermal patches as closely as possible. Accordingly, we used knife casting to spread aqueous PEG-containing PMVE/MA gels on laser engineered silicone moulds without sidewalls. The moulds were then placed in a vacuum chamber. This resulted in fully-formed MN arrays. The maximum patch size produced by us to date is 25 cm 2 , containing approximately 18,000 individual MN (Figure 6 m).

Conclusion
We have shown for the fi rst time that MN can be successfully prepared from hydrogel-forming materials. Such MN swell in skin to produce continuous, unblockable conduits from patch-type drug reservoirs to the dermal microcirculation, thus allowing prolonged transdermal drug administration. Pulsatile or bolus delivery can be achieved under electrical control, while to both control and hydrogel MN extract. These results suggest that the hydrogel-based MN material is likely to be both biocompatible and non-irritant.

Sterilization and Human Volunteer Studies
Both aseptic manufacture and post-manufacture gamma sterilization rendered MN sterile, as evidenced by the BP Sterility Test. [ 27 ] Gamma-sterilised ( 60 Co, 10 kGy) MN were applied to the ventral forearm skin of six healthy human volunteers. Application of a conventional sticking plaster only and the plaster containing the MN baseplate received a VAS score of 0 cm (i.e., no pain was felt at all) in all volunteers. However, application of plasters containing MN arrays elicited higher VAS scores, which were dependent upon the density of the MN array applied. For example, the mean VAS score associated with application of the hydrogel MN of density 361 MN per 0.5 cm 2 (0.33 ± 0.10 cm) was greater than that for a hydrogel MN of density 121 MN per 0.5 cm 2 (0.14 ± 0.08) ( p < 0.001). Similarly, both hydrogel MN systems had greater recorded VAS score during wear for both the 2 h and the 24 h application groups. VAS scores were increased for both MN array designs during the 24 h treatment protocol, in comparison to the 2 h treatment protocol. Comments from volunteers regarding sensations included: "A prickly feeling noticed at times during the presence of 361 array"; "Slight pressure noticed for 361 array about 10 h after application, but not classed as pain"; "Defi nitely feel a sensation during 361 array application, like a scratch"; "Could feel pressure during 361 array presence, slight itchy feeling" and "For 24 h period, slight itchy feeling at night for the 361 array". Figure 6 l shows representative cross-sectional OCT images of 121 MN arrays inserted in human skin in vivo. The images confi rm that the MN punctured the stratum corneum barrier and the MN extended ≈ 460 μ m into the skin. The widths of the MN induced holes in the stratum corneum were ≈ 265 μ m in diameter. However, it is obvious that there was a clear space of ≈ 136 μ m between the bottom of the MN baseplate and the upper surface of the stratum corneum , indicating that the entirety of the MN lengths were not inserted into the skin. When the 361 MN array were inserted into the skin of human volunteers, it was found that the MN extended approximately 440 μ m into the skin and that the holes in the stratum corneum were ≈ 258 μ m in diameter. Once again, a clear space of approximately 156 μ m between the bottom of the MN baseplates and the upper surface of the stratum corneum was observed. Figure 6 d-i show sample clinical photographs for a 24 h treatment protocol. The irritant contact dermatitis (ICD) scores indicated no reaction with the plaster only treatment, whilst only one individual had a response graded as 0.5 for the plaster containing the MN baseplate only. The skin response following application of the MN-containing plasters was graded as only barely perceptible (0.5) or mild erythema (1), regardless of the length of time the MN array remained within the skin. Interestingly, for the 0 h and 24 h treatment protocols, the grading of the skin response following removal of the patch containing a 361 MN array was lower than that for a patch containing a 121 MN array, as shown in Figure 6 j. In all cases, ICD scores MN are removed intact from skin. The ability of MN to imbibe skin interstitial fl uid may fi nd further use in minimally-invasive patient monitoring. MN materials were shown to be biocompatible and non-irritant in vitro and raised no safety concerns in human volunteers. This technology has the potential to overcome the limitations of conventional MN designs and greatly increase the range of type of drug deliverable transdermally, with ensuing benefi ts for industry, healthcare providers and, ultimately, patients. We are actively pursuing commercialisation of the technology with two major companies.