Fabrication of microneedle patches with lyophilized influenza vaccine suspended in organic solvent

Abstract

Skin vaccination by microneedle (MN) patch simplifies the immunization process to increase access to vaccines for global health. Lyophilization has been widely used to stabilize vaccines and other biologics during storage, but is generally not compatible with the MN patch manufacturing processes. In this study, our goal was to develop a method to incorporate lyophilized inactivated H1N1 influenza vaccine into MN patches during manufacturing by suspending freeze-dried vaccine in anhydrous organic solvent during the casting process. Using a casting formulation containing chloroform and polyvinylpyrrolidone, lyophilized influenza vaccine maintained activity during manufacturing and subsequent storage for 3 months at 40 C. Influenza vaccination using these MN patches generated strong immune responses in a murine model. This manufacturing process may enable vaccines and other biologics to be stabilized by lyophilization and administered via a MN patch.

Graphical Abstract

Introduction

Seasonal influenza viruses cause approximately 250,000 - 500,000 deaths worldwide each year [1], and influenza pandemics have the potential for much greater morbidity and mortality [2], Influenza vaccination remains the main strategy to control virus spread and prevent major epidemics and pandemics. Influenza vaccine is estimated to prevent up to a quarter of the predicted influenza-associated deaths [3], which means that effective distribution and storage of influenza vaccines is critical to expanding vaccination coverage [4]. However, current influenza vaccines are formulated as liquids with limited thermal stability, requiring storage within a narrow temperature range of 2 – 8 °C [5]. Deviation from these storage conditions can lead to vaccine degradation, reducing its potency [6]. For this reason, influenza vaccines must be maintained in a tightly controlled “cold chain” environment, resulting in increased costs and complexity in vaccine distribution and storage, especially in low-resource settings [7]. A thermostable influenza vaccine that can be partially or fully removed from the cold chain would reduce costs, facilitate logistics of storage and distribution and simplify vaccine administration [8].

Many vaccines have been previously stabilized through lyophilization or other drying processes to eliminate water molecules which reduce protein mobility; this process inhibits protein denaturation and contributes to long-term stability [9–11]. Lyophilization, or freeze-drying, has been extensively used to prepare dry, stable vaccines and other biologics. In lyophilization, a protein solution is frozen and then water is removed from the sample by sublimation, thereby locking the protein structure. Suitable excipients are included in the formulation to stabilize the protein during the freezing and dehydration processes. Significant work has been done to study and thereby minimize the stresses during these processes that can affect biomolecule structural integrity [12].

A limitation of lyophilized vaccine products is that they need to be reconstituted before use. This involves the addition of water to the lyophilized vaccine powder, which requires additional vials containing the water as well needles and syringes for the reconstitution before delivery. It also requires expertise, which had led to reconstitution errors, some of which have been fatal [13].

Vaccination by a microneedle (MN) patch (which is also sometimes called a microarray patch, or MAP) is an alternative to conventional vaccine administration by needle and syringe [14–18]. In this approach, the vaccine is prepared in a dry, solid state, that enables thermostabilization and administration without the need for reconstitution. The MN patch vaccination method involves skin patches that contain an array of solid microneedles measuring hundreds of microns in length. Upon application to skin, the MNs painlessly penetrate the skin’s upper layers, where they dissolve and release the encapsulated vaccine into the skin. Because the MNs dissolve in the skin, MN patches generate no biohazardous sharps waste. The simplicity of their use enables MN patches to be administered by personnel with minimal training, including self-administration by patients [19, 20]. Vaccination via MN patches has also been shown to improve immune responses compared to subcutaneous or intramuscular injections by targeting vaccine delivery to antigen-presenting cells in the skin, such as Langerhans cells and dermal dendritic cells [21, 22]. MN patch vaccination has been studied for many different vaccines in pre-clinical studies [15]. Specifically, influenza vaccination using a MN patch has been studied in human clinical trials, and was shown to be at least as immunogenic, well tolerated and strongly preferred compared to intramuscular injection [20, 23].

Prior studies have shown that MN patch formulations can be optimized to stabilize vaccines for long-term storage outside the cold chain [24, 25]. However, these formulations can be difficult to develop, because they involve optimization of vaccine stability and MN patch fabrication with sufficient mechanical strength, manufacturing processability and other critical attributes essential for effective needle penetration into skin. This is because typical MN patch manufacturing methods involve casting an aqueous solution containing the vaccine antigen as well as excipients onto a mold to form the MNs upon drying. Air drying, with the possible addition of vacuum and/or heat, is usually applied. Lyophilization is generally not suitable because it leads to mechanically weak, porous structures due to voids left by sublimation of water from a rigid, frozen structure. For these reasons, the MN patch manufacturing process involves extensive development of unique formulation and manufacturing methods for each type of vaccine to maintain stability during manufacturing and storage.

To reduce the formulation development burden for thermostabilized MN patches, and provide more formulation options, we sought to separate the process of vaccine lyophilization and MN patch fabrication. Specifically, we utilized the extensive prior knowledge of lyophilizing formulations, and conditions for stabilizing vaccines and other biomolecules as a first step, followed by direct encapsulation of lyophilized vaccine. without re-dissolving it into a solvent, into the MN patch matrix. We were guided by previous studies showing that lyophilized proteins in anhydrous organic solvent can maintain structural integrity and stability [26–28]. Motivated by this earlier work, we sought to investigate whether lyophilized vaccines can be cast into MN molds using an anhydrous organic solvent containing lyophilized vaccine and suitable excipients to form the MN patch matrix. We further investigated whether vaccine stabilized by lyophilization can be incorporated into MN patches that maintain thermostability without the use of additional stabilizers in the MN patch.

Materials and methods

Materials

Lyophilized FITC-labeled bovine serum albumin (FITC-BSA), lyophilized bovine immunoglobulin G (IgG), polyvinyl pyrrolidone (PVP, 40 kDa), and anhydrous chloroform were purchased from Sigma Aldrich (Milwaukee, WI). Lyophilized monovalent subunit influenza vaccine (A/Brisbane/10/10 (H1N1)) was generously provided by Seqirus (Maindenhead, United Kingdom). The lyophilized influenza vaccine provided by Seqirus was concentrated by them to an antigen concentration of 16.2 – 20.7 mg mL⁻¹ using tangential flow filtration and then lyophilizing it in the presence of 3% (w/v) sorbitol, as described previously [29]. Poly-dimethylsiloxane (PDMS) was purchased from Electron Microscopy Science (Hatfield, PA). Excised porcine skin was purchased from Pel-Freeze (Rogers, AR). Gentian violet dye was purchased from Rica Chemical (Arlington, TX). Depilatory cream (Nair®) was purchased from Church & Dwight Co. (Ewing, NJ). IgG ELISA (enzyme-linked immunosorbent assay) kits were purchased from R&D Systems (Minneapolis, MN). Lightning-Link conjugation Kit was purchased from Innova Biosciences (Cambridge, UK). Immulon 2HB 96-well microplate, TMB (3, 3′,5,5′-tetramethylbenzidine) substrate, TMB STOP Solution, 1X phosphate-buffered saline (PBS), and BSA were purchased from Thermo Fisher Scientific (Waltham, MA). Receptor-destroying enzyme was purchased from Denka Seiken (Tokyo, Japan). Aluminum pouches were purchased from Oliver-Tolas Healthcare (Grand Rapids, MI). Desiccant was purchased from Hammond Drierite (Xenia, OH). Animal lancets (4 mm Goldenrod) were purchased from Medipoints (Mineola, NY).

Animal welfare statement

All institutional and national guidelines for the care and use of laboratory animals were followed. All experimental protocols were approved by the Georgia Tech Institutional Animal Care and Use Committee (protocol A14030, originally approved in 2014). Female BALB/c mice (6 – 8 weeks old) were obtained from Charles River Laboratory (Wilmington, MA). Animals were anesthetized prior to euthanasia.

Fabrication of microneedle patches

MN patches were fabricated based on methods described previously [30]. Molds were made by etching PDMS sheets using a CO₂ laser (VLS 3.50, Universal Laser Systems, Scottsdale, AZ) to make 10 by 10 arrays of cavities with the desired MN dimensions. PVP was dissolved in chloroform at a concentration of 20% (w/v), and mixed with (i) lyophilized IgG at the indicated concentration, (ii) lyophilized FITC-labeled BSA at a concentration of 3 mg mL⁻¹, or (iii) lyophilized influenza vaccine at a concentration of 3 mg mL⁻¹. For placebo MN patches, 20% (w/v) PVP in chloroform was used without antigen. Then, 100 L of the solution was cast onto the MN patch mold. Vacuum (−30 mmHg) was then applied for 4 h. During the vacuum process, the mold was covered with a glass slide to prevent evaporation of chloroform, and dried overnight. Then, a solution of 20% (w/v) PVP in chloroform was cast onto the mold under vacuum to form the patch backing layer. MN patches were peeled off the mold within 3 h. Residual chloroform was removed by vacuum over a period of 48 h. Efficient chloroform removal was expected due to its relatively high vapor pressure at room temperature (~20 kPa), but residual solvent content was not measured. Each MN was measured to be ~650 μm tall with a base diameter of ~300 μm, and tapering to a sharp tip. MNs were photographed using a fluorescence stereomicroscope (Olympus SZX16, Tokyo, Japan).

Microneedle insertion into porcine skin

Hair from excised porcine skin was removed by shaving using a razor blade to expose smooth skin without damage. The skin was then cleaned using 70% isopropanol. The fat tissue of the subcutaneous layer was removed, and the skin was then laid flat, with the stratum corneum side up on a hard surface at room temperature (20 – 25 °C). The surface of the skin was dried with a paper towel, and then stretched and fixed using needles at the edges of the excised skin to mimic natural skin tension. A MN patch was applied to the skin and pressed by thumb for 30 s to insert the MNs into the skin. After that, the MN patches were left on the skin for 15 min to ensure dissolution of the inserted MNs in the skin.

After removing the MN patch from the skin, either gentian violet dye was applied to insertion sites to confirm MN penetration into the skin, or the skin was embedded in cryomolds to obtain histological sections. Histological sections of the skin were performed using a cryostat (CM3050, Leica, Buffalo Grove, IL) and imaged by fluorescence microscopy (IX70, Olympus). In addition, sectioned skin was stained with hematoxylin-eosin (H&E) and imaged by brightfield microscopy (IX70, Olympus). To facilitate calculation of delivery efficiency of the MN patches loaded with either IgG or influenza vaccine, the MNs were cut from the patch backing using a scalpel blade. The collected MNs were then dissolved in PBS to quantify IgG or vaccine content, which was considered as the full dose in the patch.

ELISA assay for IgG and influenza vaccine

IgG concentration was quantified using an IgG ELISA kit according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN). Influenza vaccine samples from MN patches reconstituted in deionized water were assayed for binding activity by a sandwich ELISA [31]. Briefly, polyclonal, strain-specific antibodies were received from the Center for Biologics Evaluation and Research of the Food and Drug Administration (Silver Spring, MD). Antibodies were conjugated to horseradish peroxidase with a Lightning-Link conjugation kit. The lyophilized vaccine was used to generate a reference standard curve. After the indicated storage time, samples were dissolved in PBS and run parallel to the reference standard curve on an Immulon 2HB 96-well microplate. The horseradish peroxidase (HRP) substrate reaction involved TMB chromogen (tetramethylbenzidine), and 2 N sulfuric acid as a stop solution. The microplate absorbance at 620 nm was read using an iMark plate reader (Bio-Rad, Hercules, CA).

Preparation of IgG or influenza vaccine stability samples

Lyophilized IgG or influenza vaccine were suspended at a concentration of 3 mg mL⁻¹ in a 20% (w/v) PVP in chloroform and cast as 333 μl droplets in glass vials. The solution was dried overnight at room temperature and then exposed to a vacuum at −30 mmHg for 48 h to remove residual chloroform from the samples. Samples were then packaged in aluminum pouches and sealed with an impulse heat sealer (AIE-300, American International Electric, City of Industry, CA) with desiccant. Samples were then stored in an environmental test chamber maintained at 40°C (model 6020, Caron, Marietta, OH). At specified time points, samples were reconstituted in PBS, and active IgG or influenza vaccine concentration was measured by ELISA. To calculate the IgG or influenza vaccine percent activity, the sample concentration was quantified at specified time points and divided by the untreated samples’ measurement at the beginning of the experiment. The obtained value was multiplied by 100% to obtain percent activity. Both treated and untreated samples were prepared in the same manner. All the measurements were done using ELISA.

Immunization of mice

BALB/c mice (8–10 weeks old, female) were anesthetized with isoflurane. The caudal dorsal skin was depilated with depilatory cream, rinsed with water and patted dry. Each animal was treated either with a MN patch vaccine, containing 3.3 1.7 μg of influenza vaccine in the MN portion of the patch, or a placebo MN patch without influenza vaccine but otherwise identical to the MN patches with influenza vaccine. MN patches were applied to the skin and left for 15 min to allow dissolution of the encapsulated influenza vaccine in the skin.

At specified time points (14 and 21 days) after immunization, blood (150 – 200 μl) was collected from the animals submandibularly, and centrifuged at 4,500 g for 3 min for serum separation. Serum samples were collected and stored at −80 °C until analysis. At the final time point, mice were euthanized.

Hemagglutination inhibition titers

Immunogenicity of influenza MN patch vaccination was tested by measuring hemagglutination inhibition (HAI) titers, as previously described [31]. We used HAI titers as the method to quantify immune response because they are generally considered the best correlate with protective efficacy of influenza vaccines [32]. Serum was treated with the receptor-destroying enzyme (Denka Seiken, Tokyo, Japan) overnight at 37°C.

Samples were then incubated at 56°C for 30 min for complement inactivation, followed by overnight incubation with packed red blood cells at 4°C. After centrifugation, the supernatant was collected, serially diluted in PBS, and mixed with separate strain-specific influenza viruses for 30 min. The mixture was incubated with 0.5% chicken red blood cells for 30 min. The reciprocal of the highest serum dilution that prevented hemagglutination was read as the HAI titer.

Statistical Analysis

A two-tailed Student’s t-test was performed when comparing two different conditions. When comparing three or more conditions, a one-way analysis of variance (ANOVA) with multiple comparisons was performed. All statistical analysis was performed using GraphPad Prism 8. Differences were considered to be statistically significant at a level of p < 0.05.

Results

Design and fabrication of microneedle patches, and delivery into the skin

MN patches were designed to (i) encapsulate lyophilized influenza vaccine and (ii) dissolve in the skin in a relatively short time to release the vaccine. We used chloroform as a solvent for MN fabrication because of its strong hydrophobic character that prevents dissolution of lyophilized vaccine during the manufacturing process. In addition, chloroform is very volatile, which means it can be quickly removed from the MN matrix during the drying process. We selected PVP as the MN matrix material because it is soluble in both chloroform (for manufacturing) and water (for dissolution in the skin). PVP has been used to fabricate MN patches in prior studies because it has strong mechanical properties and also rapidly dissolves in the skin [33–35].

To visualize the encapsulation of lyophilized proteins inside the MN matrix, we used FITC-BSA (Fig. 1a–b). We then tested if the MN patch had enough mechanical strength to penetrate the superficial layer of the skin, and deliver encapsulated FITC-BSA into the skin. To test this, we applied MN patches to excised pig skin ex vivo and observed the fluorescence of FITC-BSA located in the skin, exhibiting the pattern of the 10 × 10 array of the MN patch (Fig. 1c). Successful penetration of MNs into the pig skin and delivery of FITC-BSA was confirmed with histological sections of the insertion site (Fig. 1d).

Fig. 1.

Microneedle (MN) patch encapsulation and delivery of lyophilized FITC-labeled bovine serum albumin (BSA) prepared by casting in anhydrous chloroform. Representative brightfield (a) and fluorescence (b) imaging of MNs encapsulating the lyophilized FITC-labeled BSA. Scale bars: 200 m. (c) Representative fluorescence imaging of pig skin after delivery of FITC-labeled BSA by MN patch demonstrating delivery of the BSA into the skin ex vivo. Scale bar: 200 mm. (d) Histological section of pig skin after delivery of FITC-labeled BSA using a MN patch ex vivo. The site of MN penetration into the skin is indicated by the white arrow and localized fluorescence of FITC-labeled BSA. Scale bar: 200 m

Lyophilized IgG and influenza vaccine stability after casting in anhydrous chloroform formulation

This project is motivated by the hypothesis that lyophilized biomolecules can be suspended in chloroform solution without damaging the biomolecules. To assess this hypothesis, we compared the activity of lyophilized bovine IgG (as determined by ELISA binding activity) to lyophilized bovine IgG suspended in chloroform, or suspended in chloroform containing PVP (i.e., the casting solution used to make MN patches). These solutions were cast in glass vials, allowed to dry, and reconstituted to determine influenza vaccine potency. This comparison showed no statistically significant difference in vaccine activity before or after suspension, casting and drying in chloroform with or without PVP (Fig. 2a). This indicates no loss in vaccine binding activity in vitro. Experiments below address immunogenicity in vivo.

Fig. 2.

Stability of lyophilized IgG after casting using anhydrous chloroform and polyvinylpyrollidone (PVP). (a) Stability of IgG measured in dry form (1^st bar), compared to IgG cast in chloroform alone or with PVP (2^nd and 3^rd bars, respectively), dried overnight and reconstituted in phosphate-buffered saline. Binding activity was measured by ELISA. Dotted line indicates 100% stability. Data are shown as mean SEM (n = 6 - 10). (b) IgG dose in microneedles cut from microneedle patches as a function of the initial IgG concentration of the casting solution used. Data are shown as mean SEM (n = 3)

To further characterize the MN patches, we varied the IgG concentration in the chloroform solution in order to encapsulate varying doses in the MN patches. We found that the IgG dose in the MNs decreased as the IgG concentration in the MN casting solution decreased, as expected (Fig. 2b). The IgG dose in the MNs ranged from 1.7 to 7.8 g per MN patch.

Guided by these findings, we evaluated the activity of lyophilized influenza vaccine after suspension in chloroform, with or without PVP, followed by drying in glass vials. We similarly found that lyophilized influenza vaccine binding activity was not significantly impacted compared to untreated lyophilized influenza vaccine, thereby indicating that influenza vaccine MN patches could be fabricated using this process (Fig. 3).

Fig. 3.

Stability of lyophilized influenza vaccine after casting using anhydrous chloroform and polyvinylpyrollidone (PVP). Stability of influenza vaccine was measured in dry form (1^st bar), compared to vaccine cast in chloroform alone or with PVP (2^nd and 3^rd bars, respectively), dried overnight and reconstituted in phosphate-buffered saline. Binding activity was measured by ELISA. Dotted line indicates 100% stability. Data are shown as mean SEM (n = 5 - 6)

Lyophilized influenza vaccine: encapsulation in microneedle patch matrix, stability during storage and delivery into porcine skin

Using the chloroform-based casting solution, we fabricated MN patches encapsulating lyophilized influenza vaccine. We produced arrays of MNs measuring ~650 μm long, with a base diameter of 300 μm, and tapering to a sharp tip (Fig. 4a). The MNs were mounted atop wide base structures to serve as pedestals that facilitate MN insertion into skin by succeeding surface deformation during the application process. Influenza vaccine antigen content was measured to be 3.3 ± 1.7 μg per MN patch.

Fig. 4.

Microneedle (MN) patch insertion and delivery of lyophilized influenza vaccine to pig skin ex vivo, (a) Representative brightfield image of MNs encapsulating lyophilized influenza vaccine. Scale bar: 200 m. (b) Representative brightfield image of pig skin after delivery of influenza vaccine by MN patch ex vivo. The skin was subsequently stained with propidium iodide to selectively stain sites of MN penetration. (c) Histological section of pig skin after delivery of influenza vaccine by MN patch showing a site of MN penetration into the skin (arrow). Scale bar: 200 m. MN patch before (d) and after (e) insertion into pig skin ex vivo, showing dissolution of the MNs, leaving only the base pedestals. Scale bar: 1 mm

Similar to the MN patches containing FITC-BSA, we applied MN patches containing influenza vaccine to pig skin ex vivo and found that they were sufficiently strong to penetrate the skin, as shown by selective staining of pores formed after insertion (Fig. 4b) as well as histological evidence of MN penetration across the epidermis and into the upper dermis layers (Fig. 4c). A comparison between MNs before and after the skin insertion showed efficient dissolution of the MNs, leaving only the supporting base structures (Fig. 4d–e).

To assess long-term stability of the lyophilized influenza vaccine in chloroform/PVP formulation, we cast and dried influenza vaccine casting solution in glass vials, and then stored them sealed with desiccant in a stability chamber at 40°C. Over the 12-week study, there was no statistically significant loss of binding activity (Fig. 5).

Fig. 5.

Stability of lyophilized influenza vaccine in microneedle (MN) patches stored at 40 °C for up to 12 weeks. Lyophilized influenza vaccine was encapsulated in MN patches, reconstituted in phosphate-buffered saline and activity was measured by ELISA. Patches were assayed either immediately after manufacturing (1^st bar) or after storage with desiccant at 40 °C for 1, 2, 4 or 12 weeks. Dotted line indicates 100% stability. Data are shown as mean SEM (n = 5 - 13). **p < 0.005 by one-way ANOVA with Holm-Sidak’s multiple comparison

Immune responses to lyophilized influenza vaccine delivered by microneedle patch

Immunogenicity of the influenza vaccination by MN patches was evaluated in a murine model. MN patches were administered to two mouse groups: (i) MN patch loaded with 3.3 ± 1.7 μg influenza vaccine and (ii) placebo MN patches. Serological analysis showed no significant difference in HAI titers between the groups 14 days post-vaccination (Fig. 6). However, 21 days after vaccination, mice immunized with influenza vaccine developed vaccine-specific humoral responses, whereas mice given placebo MN patches did not. The mean HAI titer in immunized mice was 51, which is greater than the threshold level of 40 that is generally considered to be seroprotective in humans [36]. [36].

Fig. 6.

Hemagglutination inhibition (HAI) titers in mice after influenza vaccination by microneedle (MN) patch. HAI titer was measured 14 and 21 days after application of a MN patch encapsulating lyophilized influenza vaccine (MN patch vaccine) or a MN patch placebo containing no vaccine (MN patch placebo) Data are shown as geometric mean SEM (n = 4, placebo, n = 10 immunized). ***p < 0.0005 by Student’s t-test

Discussion

Through a combination of controlled freezing and dehydration conditions, aided with optimized formulations, lyophilization has become a widely used and highly effective method to preserve activity of vaccines, biologics and some small-molecule drugs, enabling storage outside the cold chain [5, 12]. While formulations used to stabilize drugs and vaccines in MN patches have been guided by lessons from lyophilization, it has been difficult to directly use lyophilized formulations in MN patches because the freeze drying process is not easily compatible with MN patch manufacturing methods to fabricate MNs with robust mechanical strength that is critical for efficient insertion into skin.

In this study, we developed a method for incorporating lyophilized formulations into MN patches by separating the lyophilization process from the MN patch manufacturing process. More specifically, we encapsulated lyophilized influenza vaccine and IgG (as a model antigen) in MN patches by suspending the lyophilized compounds in anhydrous chloroform as the casting solution used during manufacturing. This approach was guided by prior literature informing us that proteins can be stable if incorporated into organic solvent without the presence of any water, which has been hypothesized to be due to restriction of protein conformational mobility in anhydrous environments compared to mixed organic-aqueous solvents [26, 27]. The protein conformation may have also been protected because the protein was in a solid state as a lyophilized particle, that further limited protein mobility.

Prior studies have demonstrated that long-term stability of vaccines can be achieved using optimized formulations during the MN encapsulation process but these studies involved extensive excipient screening and formulation development [31, 37]. In contrast, the solvent-based encapsulation process shown here leverages pre-optimized lyophilized vaccine formulations for maintaining long-term stability. This allows independent optimization of lyophilization methods for vaccine stabilization that is separate from optimization of the MN patch manufacturing process to achieve MNs with good mechanical strength for tissue penetration and rapid dissolution kinetics in the skin. The use of PVP as the MN patch matrix material was important to this process, as this polymer exhibits solubility in both chloroform for casting and water for dissolution in skin. This segregated, two-step development approach can simplify and expedite development of stabilized MN patch vaccines, especially since stable lyophilized vaccines already exist, such as the lyophilized influenza vaccine we obtained from Seqirus for use in this study.

Stability of proteins and influenza vaccine post fabrication was evaluated by ELISA, which measured the ability of antigen-specific binding activity. As a more definitive measure of vaccine potency, we also administered influenza vaccine via MN patches to mice, which showed that this method of vaccination produced strong HAI titers that exceeded the threshold levels required for seroprotection in humans. Future studies should more broadly assess the utility of the proposed MN patch fabrication method using other vaccines, proteins and biologics.

MN patch-based immunization can be an attractive vaccination modality with numerous advantages over traditional hypodermic needle-based administration. Although not assessed in this study, MN patches generally provide stronger immune responses compared to subcutaneous or intramuscular injections because MN patches target antigen delivery to skin-resident dendritic cells and direct draining to lymph nodes [38]. The small size of MN patches means smaller packaging, simplified distribution, and less-costly storage. In addition to eliminating biohazardous sharps waste, the simplified vaccination methods, possibility enabling self-administration, makes MN patch immunization particularly attractive for low-resource settings and mass vaccination campaigns for routine or pandemic situations [39, 40]. For industrialized countries, MN patch vaccination can also be attractive because they are painless and convenient to access, which has been shown to be strongly preferred compared by hypodermic needle injections [19, 41] and offer cost-effective solutions [42, 43]. The versatility of encapsulating a pre-lyophilized vaccines, proteins or other biologic formulations in a MN patch without loss of activity, as shown in this study, could significantly accelerate clinical translation of MN patch-based immunization.

Conclusion

This study sought to simplify the complex process of MN patch formulation development that must not only enable fabrication of MNs that are strong, water-soluble, manufacturable and biocompatible, but must also stabilize fragile biomolecules like vaccines during manufacturing and storage. Our approach was to separate the process of vaccine stabilization into a first step of lyophilization, that is already well-established for many pharmaceuticals. This allows the second step of MN patch fabrication to focus on other MN patch properties. This study showed that a MN patch manufacturing process using a casting solution comprised of PVP and chloroform enabled incorporation of lyophilized BSA, IgG or influenza vaccine into the MNs. IgG and influenza vaccine were shown to remain stable during the formulation and casting process, and influenza vaccine retained binding activity for at least 12 weeks during storage at 40°C. Finally, vaccination of mice with MN patches prepared with lyophilized influenza vaccine were immunogenic, achieving HAI titers that exceeded levels typically required for seroprotection in humans. We conclude that manufacturing MN patches using lyophilized biologics cast in an anhydrous organic solvent can provide a useful means to produce thermostable MN patches that benefit from the stabilizing effects of lyophilization.