Peptide delivery with poly(ethylene glycol) diacrylate microneedles through swelling effect

Abstract Transdermal delivery of therapeutic biomolecules (including peptides) can avoid enzymatic digestion that occurs in the oral route. (Polyethylene glycol) diacrylate (PEGDA)‐based microneedles, with good biocompatibility, are easily fabricated through photo‐polymerization with a precisely controlled structure. It has successfully been used for the transdermal delivery of small molecule drugs such as 5‐fluorouracil. However, the delivery of peptide‐based therapeutics using this platform is seldom reported. This is because of the potential damage to the peptide during the photo‐polymerization process of PEGDA. Herein, we introduce a method to load PEGDA microneedles with peptides without compromising peptide potency. Using gap junction inhibitor (Gap 26) as an example, the peptide was loaded into PEGDA microneedles through the swelling effect of PEGDA in the aqueous solution. The peptide‐loaded microneedles were applied to a keloid scar model and exhibited inhibition expression of collagen I, a predominant marker of keloid scar, demonstrating its potential therapeutic effects.


| I N TR ODU C TI ON
Microneedles are emerging as a proficient transdermal delivery system.
It allows a variety of molecules to be transported into skin in a minimally invasive way, which overcomes the limitations of conventional needle injection. [1][2][3] Polymer-based microneedles are specifically attractive due to their excellent biocompatibility, biodegradability, and nontoxicity. [4][5][6] Drugs can be loaded into the microneedle tips and/or base, or coated on the microneedle tips. 1,4,5 One example is poly(ethylene glycol) diacrylate (PEGDA)-based microneedles. [7][8][9] PEGDA is biocompatible and can be cross-linked in a short time under UV exposure (a few seconds), which facilitates the control of morphology and dimension of microneedles. 7,10 Small molecule drugs like 5-fluororacil (5-FU) and curcumin have been integrated within microneedles during the fabrication process, 11 in which drugs are mixed with the monomer before the mixture is placed in a mold and exposed to UV light. For example, our group has utilized PEGDA-made microneedles to deliver the 5-FU for treating keloid scar cells. 12  such as hydrophobic campothecin (CPT)-that targets DNA topoisomerase I (Topo I) and inhibit collagen synthesis 13 can also be encapsulated and delivered in a similar manner.
In recent years, peptide-based therapeutics have played a major role in new drug development. [14][15][16][17] However, this process has not been without difficulties. Peptides face hurdles including instability under high temperature, light, and high/low pHs, and digestion by gastrointestinal enzymes during oral delivery. Microneedles have been suggested as a means to overcome this delivery hurdle since it can bypass oral uptake routes. The easiest way to load the peptide is to coat them on the surface of needles, however, it often requires optimization and adjusting the coating formulations by adding excipients to increase viscosity and decrease surface tension in order to achieve uniform and sufficient coating. 18 Preloading drugs into the needles during the fabrication is an alternative way. However, preloading processes should not disrupt or significantly compromise peptide activity. PEGDA microneedles have been used to deliver several small molecular drugs that are preloaded into the polymer matrix during the fabrication. However, this procedure is unsuitable for peptides since high energy UV rays can denature peptides thus compromise the biological activity of peptide-based therapeutics. 19,20 This report introduces a gentle strategy to load peptides into the PEGDA microneedles without the above complications based on the swelling ability of PEGDA in aqueous solution. We discovered that substances with molecular mass less than 4 kDa readily enter the PEGDA matrix during swelling in aqueous solutions. Furthermore, this process is dependent on the period of UV exposure to crosslink PEGDA monomers (part of the microneedle fabrication process). As a proof of concept, the model peptide (Gap 26, a connexin mimetic peptide that inhibits cellular gap-junction), [21][22][23][24][25][26][27][28] is loaded into the PEGDA microneedles using this PEGDA swelling strategy ( Figure 1). After examining the loading capacity and release profile of the peptide-loaded microneedles, the potential therapeutic effect is explored on a keloid scar model comprising of cells and ex vivo skin. Research Corporation (Singapore). Dulbecco's modified eagle medium (DMEM) (w/4.5 g/L D-glucose, w/phenol red), fetal bovine serum (FBS), penicillin-streptomycin (10,000 U/ml), Phosphate buffered saline (1x, PBS), and trypsin-EDTA (0.25%, 10x) were obtained from Gibco Life Technologies (USA). All reagents were of analytical reagent grade and used without further purification.

| Fabrication of blank and CPT loaded PEGDA microneedles
First, a negative polydimethylsiloxane (PDMS) mold of microneedles was prepared by curing PDMS (base: curing agent 5 10:1) on the stainless microneedle template for 3 hr at 708C. In order to make blank PEGDA microneedles, the mixed solution of PEGDA (MW 5 250) and photoinitiator 2-hydroxy-2-methylpropiophenone (0.5% v/v) was added to the negative PDMS mold pretreated with plasma. The gas bubbles in the solution were removed by vacuum before the whole sample was exposed to UV (Intensity: 10 mW/cm 2 ) for crosslinking.
Finally, PEGDA microneedles patch was peeled off from the PDMS mold.

| Swelling ratio and mesh size determination
The microneedle patches were placed in distilled water for 22 hr. Then the samples were taken out and gently dried on the paper, and the weight in the swollen state (W swell ) was obtained. Later, the patches were transferred into an oven preheated to 508C and dehydrated until a stable dry weight (W dry ). The swelling ratio can be calculated by Equation 1.
To calculate the mesh size, 29 the polymer volume fraction of the swollen state t 2s was first obtained by Equation 2.

11
(2) where Q is the swelling ratio, q is polymer density, and q H2O is water density.
The average molecular weight between two consecutive crosslinks Finally, mesh size can be calculated by the Equation 4.
where l is the bond length (1.5 A o ), M r is the molecular weight of the PEG repeating unit (44 g/mol), and C n is the characteristic ratio for PEG, equal to 4.

| Immunostaining for collagen I
The sliced skin samples were treated by 0.2% Triton-X 100 in PBS for

| PEGDA microneedles loaded with hydrophobic camptothecin
About 10 lM CPT was shown to inhibit the proliferation of keloid fibroblasts (Supporting Information Figure S1A) and suppress the collagen I expression (Supporting Information Figure S1B). Even only after 24 hr, the proliferative activity of CPT of fibroblasts was reduced by 18.21%. Simultaneously, the cellular collagen I expression per cell decreased 55% under the treatment.
As a hydrophobic molecule, CPT can be loaded into PEGDA microneedles during microneedle fabrication by dissolving it with PEGDA monomer. The PEGDA microneedles exhibited homogeneous pyramidlike structures with a base diameter of 300 lm and a depth of 850 lm following photo-polymerization (Supporting Information Figure S2A,B).
The fluorescence from microneedles confirmed the presence of CPT (Supporting Information Figure S3A). The release profile of CPT was

| Swelling effect of PEGDA microneedles in aqueous solution
PEGDA exhibits swelling behavior in aqueous solutions. 29 As shown in

| Therapeutic effect of Gap26 loaded microneedles in ex vivo keloid model
A keloid disease model was next used to explore the potential efficacy of Gap26-loaded microneedles. Briefly, the model was built by injecting keloid fibroblasts into the dermis region of an ex vivo human skin sample (Figure 4a). Keloid fibroblasts were labeled with the lipophilic dye DiI (Supporting Information Figure S6) to enable identification (Figure 4b).
Before the study commenced, keloid fibroblasts were injected into the dermis region of the skin. At day 14, the microneedle patch was applied on the skin for 1 week. The patch was then replaced with a

| D I SCUSSION
In recent years, there has been increasing interest to apply biomolecules including peptide/protein as targeted therapeutics. Delivery these molecules by microneedles can avoid their digestions by gastrointestinal enzymes during oral delivery. Previously reported methods to load these biomolecules in the microneedles have been limited. In this paper, we developed a facile, effective, and gentle strategy to load peptide into biocompatible PEGDA microneedles. A model peptide, Gap26 (connexin mimetic peptide) was chosen as a candidate therapeutic and applied on an ex vivo keloid scar model to prove this concept.
To date, PEGDA microneedles have mainly been used to deliver small molecular drugs (e.g., 5-FU and curcumin) that are chemically stable to high temperature, light, and extreme pH conditions. These stable molecules can be mixed with the monomer before UV exposure or heating to generate microneedles. For example, CPT can reduce collagen biosynthesis, a key step in scar formation. 13,37 When this drug was encapsulated in PEGDA microneedles through photolithography, no noticeable change was observed in its efficacy to inhibit cell proliferation and collagen I expression (Supporting Information Figures S1, S3, S7).
Subsequently, we loaded hydrophilic sugar and peptide molecules by swelling PEGDA microneedles in an aqueous solution (Figures 1, 2b, and 3c). On hydration, the PEGDA matrix allows solutes to diffuse into the hydrogel matrix. As discussed above, the mesh size of PEG hydrogel, 38,39 related to its crosslinking density determines whether molecules can readily diffuse into the matrix. Thus, molecules larger than the mesh size cannot be efficiently loaded into PEGDA. Here, we controlled the mesh size of PEGDA microneedles by tuning the UV exposure time during the fabrication and ended up with a PEGDA matrix with a mesh size of 2.0 nm (Supporting Information Figure S4). Thereafter, we tested the loading of various molecules with distinct molecular weights (FITC, FITC-dextran 4, 10, 20 kDa). Due to the difference in molecular weight, their hydrodynamic sizes vary from angstrom scale to 1.3 to 2.0 to 2.8 nm. 40 In this case, FITC-dextran molecules bigger than 10 kDa are larger than the mesh size of PEGDA. Thus, it is expected that minimal quantities of larger molecules were observed in microneedles (Figure 2b,c)