A Single‐Administration Microneedle Skin Patch for Multi‐Burst Release of Vaccine against SARS‐CoV‐2

Abstract

The necessity for multiple injections and cold‐chain storage has contributed to suboptimal vaccine utilization, especially in pandemic situations. Thermally‐stable and single‐administration vaccines hold a great potential to revolutionize the global immunization process. Here, a new approach to thermally stabilize protein‐based antigens is presented and a new high‐throughput antigen‐loading process is devised to create a single‐administration, pulsatile‐release microneedle (MN) patch which can deliver a recombinant SARS‐CoV‐2 S1‐RBD protein—a model for the COVID‐19 vaccine. Nearly 100% of the protein antigen could be stabilized at temperatures up to 100 °C for at least 1 h and at an average human body temperature (37 °C) for up to 4 months. Arrays of the stabilized S1‐RBD formulations can be loaded into the MN shells via a single‐alignment assembly step. The fabricated MNs are administered at a single time into the skin of rats and induce antibody response which could neutralize authentic SARS‐CoV‐2 viruses, providing similar immunogenic effect to that induced by multiple bolus injections of the same antigen stored in conventional cold‐chain conditions. The MN system presented herein could offer the key solution to global immunization campaigns by avoiding low patient compliance, the requirement for cold‐chain storage, and the need for multiple booster injections.

The paper presents a new protein‐stabilization method and high‐throughput antigen‐loading process to create a single‐administration microneedle skin patch which could deliver multiple longitudinal doses of a recombinant S1‐RBD protein antigen derived from SARS‐CoV‐2 viruses. The single‐administration microneedle platform addresses the COVID‐19 crisis and prepares us for potential future outbreaks.

1. Introduction

During large scale outbreaks of highly infectious diseases, such as the current COVID‐19 pandemic caused by the virus SARS‐CoV‐2, widespread vaccination is often utilized to achieve community immunity at a global scale in order to prevent transmission and potential mutation of the pathogens, associated diseases, and loss of life. However, traditional vaccination methods relying on subcutaneous (s.c.) or intramuscular (IM) injections suffer from logistical burdens which hinder their rapid deployment on a massive scale. First, the injections are painful, and require trained professionals to administer the doses. Second, extensive research works have shown that vaccines often require booster shots to sustain an effective antibody level and long‐term immune protection.^{[ 1 ]} These booster shots are repeated over weeks or months which not only increase logistic burdens but also could be a challenge for patients to adhere to the schedules. Additionally, as patients need to travel to hospitals or medical centers for access to vaccinations, booster appointments during outbreaks can increase the risk of infection and the burden on healthcare facilities already stressed by increased caseloads. In developing countries, these hurdles can be even more problematic as many patients are living in remote areas and cannot afford repeated travels for multiple booster shots. These substantial problems of low‐patient compliance and adherence pose significant challenges to global vaccination campaign against SARS‐CoV‐2, poliovirus, measles virus, etc.^{[ 2 ]} Third, many vaccines have to be consistently stored under stringent conditions with very low temperatures. This cold‐chain requirement presents a significant obstacle to the dissemination of necessary vaccines in lower‐income or resource‐disadvantaged communities due to the need for extended storage of vaccines for priming and booster shots. Collectively, 1) the painful, and inconvenient nature of needle‐based injections which reduce patient compliance, 2) the need for booster injections, and 3) the requirement for cold‐chain storage are critical problems of current vaccines which need to be overcome in order to address the current COVID‐19 crisis, increase vaccine coverage worldwide, and prepare us for potential future outbreaks. A key solution to the aforementioned problems is to develop a new vaccine delivery system which is easy‐to‐use, carries thermally stable antigens, and can be efficiently distributed to people even for self‐administration at their homes in just a single time to create an effective and widespread community immune protection.

Transdermal microneedle (MN) skin patches offer an excellent candidate for such a delivery system. MNs can be self‐administered by patients at home, and may not require a skilled healthcare professional.^{[ 3 ]} In contrast to bolus injections by a hypodermic needle, the MNs are administered more superficially, avoid touching the nerve ends, and are, thus, almost painless^{[ 4} , ^{5 ]} Furthermore, the MNs target the dermal layer wherein there are many immune cells including dendritic and Langerhans cells which help to boost the immune reaction to vaccines.^{[ 3a ]} Multiple demonstrations have been performed to study MN patches for deliveries of diphtheria toxoid,^{[ 6 ]} recombinant anthrax,^{[ 7 ]} hepatitis B,^{[ 8 ]} rabies,^{[ 9 ]} and influenza^{[ 3a ]} vaccines. Currently, there is still no FDA‐approved MN patch for vaccination. However, the field has been rapidly evolving since the mid‐1990's and many studies are now transitioning from pre‐clinical to clinical trials in humans.^{[ 10 ]}

Although new mRNA or vector‐based COVID‐19 vaccines can be rapidly produced and have been successfully implemented, traditional vaccines using recombinant proteins have their own merits in the global vaccination effort due to their proven long‐term safety profile, the widespread availability of manufacturing facilities, and lower manufacturing costs.^{[ 11 ]} Indeed, at this moment, many vaccines using the spike protein (S‐protein) antigens (spike glycoprotein projecting externally from the shell of SARS‐CoV‐2) have emerged for clinical use and even shown to be effective against different SARS‐CoV‐2 strains including the highly contagious Delta variant (e.g. FDA‐approved Novavax with >90% efficacy against symptomatic SARS‐CoV‐2 infection as of December 2021).^{[ 12} , ^{13 ]} Among those S‐protein based antigens, the receptor binding domain (RBD) subunit could trigger broad and durable neutralizing antibodies against the SARS‐CoV‐2.^{[ 13} , ^{14 ]} Moreover, studies on MN patches with instantaneous release of the S‐protein have shown exciting success, demonstrating the potential benefits of the application of the technology for COVID‐19 vaccine delivery.^{[ 15 ]} For example, a high‐density microarray patch (HD‐MAP) coated with a stabilized recombinant SARS‐CoV‐2 spike glycoprotein and QS‐21 adjuvant has elicited both cellular and antibody immune responses and protection from a lethal virus challenge in an human angiotensin‐converting enzyme 2 (ACE2)‐transgenic mouse model.^{[ 16 ]}

Despite providing substantial immune protection, most available vaccine MN patches so far have been designed to only perform an immediate delivery of antigens, which thus might require repeated administrations for boosters. Patients are still responsible for keeping up with the vaccination schedule and need to store the MN patches appropriately (to avoid thermal degradation) for the follow‐up booster administrations. Some researchers have developed MNs for the sustained delivery of small‐molecule drugs or vaccines;^{[ 17} , ^{18 ]} however, there is a requirement for appropriately controlled release time in order to avoid the risk that the sustained release of vaccines could elicit immune tolerance responses.^{[ 19} , ^{20 ]} These shortcomings of current MN patches together make them suboptimal for use in global vaccination efforts. In this context, we have recently reported the use of single‐administration multi‐burst release core–shell MNs for the delivery of a clinically relevant vaccine against Streptococcus pneumoniae, Prevnar‐13.^{[ 21 ]} However, Prevnar‐13 is stable and can maintain structural integrity/immunogenicity over the course of the MN fabrication process and skin‐implantation. In contrast, many other important vaccines are highly susceptible to damage or denaturation by heat. Furthermore, our reported fabrication approach relies on a multiple‐step process to align and load vaccine cores into the MNs, which reduces the overall throughput of the manufacture, potentially causes loss of the loaded vaccine, and is prone to more fabrication errors. Thus, there remain significant challenges to 1) create such single‐administration MNs (often undergoing a high temperature fabrication process of 45–75 °C) for other thermally‐unstable antigens, 2) store the MNs for a long‐term use at a normal room temperature without the cold‐chain condition, 3) enhance the throughput for the process to load protein‐antigen into the MNs, and 4) embed the MNs inside the skin (at 37 °C) for a long period of time to perform the desired multiple burst release of the stabilized antigens, and trigger a long‐term high quality immune response similar to the effect of multiple primer‐booster injections.

Here, we present a new strategy to thermally stabilize protein antigens (using the S1‐RBD of SARS‐CoV‐2 as a model) and devise a one‐step antigen‐loading process to fabricate easy‐to‐use, painless, and single‐administration MNs which can be fully inserted into the skin and programmed to release the stabilized antigen at multiple desired times. We show that by using trehalose and sucrose as excipients we can stabilize the S1‐RBD protein antigen at an extreme temperature, up to 100 °C, for a short screening time (at least 1 h) and at the body temperature (37 °C) for at least 4 months. This is the first reported effort to stabilize S1‐RBD protein with trehalose and sucrose formulation. Furthermore, we also presented a new one‐step loading method to encapsulate the stabilized antigen into the core–shell MNs. This new approach eliminates the need for an impinging layer previously used in our published work, minimizing the potential errors caused by multiple alignment steps, and enhances the overall throughput of the manufacture. The MNs were then inserted into the skin of rats in a minimally‐invasive manner and exhibited no noticeable skin irritation even after long‐term implantation. These MNs successfully triggered antibody responses which were capable to neutralize authentic SARS‐CoV‐2 strains, including the Delta variant, at similar levels to those induced by multiple bolus injections of the same antigen stored in conventional cold‐chain conditions. The antigen‐stabilization approach and the antigen‐loading method for fabricating the single‐administration MN platform, presented herein, could provide the key solution to the global vaccination campaign (against the current/future pandemics) by avoiding 1) the pain, cost, and low patient compliance/adherence to the conventional vaccine injections, 2) the requirement of cold‐chain facilities, and 3) the need of multiple booster shots.

2. Experimental Section

2.1. Fabrication of the Core–Shell MNs

The MNs patch manufacturing follows the previously reported methods of micro‐molding and 3D layer‐by‐layer assembly with some additional modifications.^{[ 21 ]} The four main components of the core–shell MN are a shell, a core, a cap, and a supporting array. Each MN had a height of 600 µm and a base diameter of 300 µm, and encapsulates an antigen‐loaded core with a height of 400 µm and a base diameter of 200 µm. One 1 × 1 cm² MN patch contains a 15 × 15 MN array or 225 MNs. The shell and the cap were made by compression‐molding a Poly(D,L‐lactide‐co‐glycolide) (PLGA) (MilliporeSigma and Polysciences, Inc., USA) film into respective PDMS molds in the vacuum oven for 20 min at PLGA glass transition temperature. A Carver press (Model: 3850‐1011, Wabash, IN) was used to make PLGA films with defined thickness. The “scum layer” (or residual film) that resulted from the molding process was then removed by gently dropping and wiping acetone on top of the mold. The MN‐core preparation included dissolving or suspending S1‐RBD protein antigen in excipient solutions 0.5 m Trehalose (MilliporeSigma and Sucrose (MilliporeSigma) dissolved in MilliQ water. In the next step, the solution or suspension was filled into the prepared inverse PDMS mold. After solvent evaporation, the scum layer was then removed. The scum‐free protein antigen‐cores were next transferred onto a sacrificial PLGA film under mild heating. This whole structure was delaminated onto a solid Teflon‐coated substrate via heat‐assisted micro‐transfer molding. The supporting array was made by a two‐step process involving solution casting effervescent polymer solution^{[ 18a ]} and PLA molding to provide a flexible backing layer.

2.1.1. Drug Loading Process and 3D Layer‐by‐Layer Assembly

The core–shell MNs assembly process, which was based on the previously‐reported technology named StampEd Assembly of polymer Layers (SEAL), combines the technology used for computer chip manufacturing with soft lithography and an aligned sintering process as described in the main text.^{[ 22 ]} First, the core‐drugs previously prepared were aligned using the aligning device and loaded into the core–shell MNs. Second, the scum layer was them removed by the method described above, using acetone and water as solvents. The caps inside the PDMS mold were aligned and loaded on top of the PLA‐effervescent supporting array and sintered using heat. In the final step, the core–shell MNs were transferred onto the supporting array. The mold entrapping the core–shell MNs was then peeled off to yield free‐standing core–shell MNs on the supporting array which could be fully‐embedded into the skin. The optical images of the MNs were taken by the AmScope ME300TZC‐2L‐8M with a digital camera (AmScope Mu800).

2.2. S1‐RBD Protein Stability Study

S1‐RBD‐protein (Euprotein Inc., North Brunswick, NJ) was mixed with trehalose‐sucrose (TS) solution at different ratios (Table S1, Supporting Information) and vacuum dried overnight (mimicking the MN core fabrication). Samples were then exposed to heat for short‐term stability screenings (1 h) at 56, 70, and 100 °C. Long‐term stability study was conducted at 37 °C (body temperature) for 2, 4, 8, 12, and 16 weeks.

Enzyme‐linked immunosorbent assay (ELISA) was used to determine antigen stability after heat exposure. Incubated and non‐incubated S1‐RBD‐protein were diluted to a final concentration of 5 µg mL⁻¹ in 100 mm carbonate–bicarbonate buffer (pH 9.6). Bovine serum albumin (BSA) was diluted in carbonate–bicarbonate buffer to an equivalent concentration and used as a negative control. Immulon1B plates (Thermo Fisher Scientific) were coated with 100 µL per well with S1‐RBD in bicarbonate buffer and incubated overnight at 4 °C. Plates were then washed with 0.05% Tween‐20 in PBS (Thermo Fisher Scientific) and blocked for 1 h at room temperature with 2 µg mL⁻¹ BSA 0.05% Tween‐20 in PBS. As a primary antibody, 100 µL of SARS‐CoV‐2 Spike Protein (RBD) Recombinant Human Monoclonal Antibody (Thermo Fisher Scientific) diluted at 1:500 in PBST was loaded into each well and incubated for 1 h at room temperature. Wells were then washed three times with 50 µL of PBST before loading the secondary antibodies. Goat anti‐Human IgG (Gamma chain) cross‐adsorbed secondary antibody, HRP diluted at 1:1,000 in PBST (Thermo Fisher Scientific) was used as a secondary antibody (100 µL) and the samples were incubated for 1 h at room temperature. All of the wells were then washed three times with 50 µL of PBST before development. To develop each plate, 100 µL TMB substrate (SouthernBiotech) was added to each well. After 15 min, the reaction was stopped with TMB Stop Solution (SouthernBiotech). Plates were read at 450 nm on a BioTek Synergy HT plate reader using BioTek Gen5 (BioTek). Samples were run in duplicates. Data were representative of an average of three sample replicates, and normalized to initial S1‐RBD‐protein activity after vacuum drying. Statistical analysis was performed using two‐way analysis of variance (ANOVA) with repeated measures and Tukey test for multiple comparisons (GraphPad Prism 8.0 (GraphPad)).

SDS‐Page Coomassie Blue‐stained SDS‐PAGE gel was run on pure S1‐RBD protein or S1‐RBD:TS 1:4 v/v samples under varying conditions. All dried samples were provided as 20 µg dried protein and were resuspended in 20 µL Ultrapure Dnase and Rnase free water. 10 µL of each sample was mixed 1:1 with Laemmli+BME, incubated at 70 °C for 10 min, and then 20 µL of each sample was loaded per well, and run for 20 min at 200 V. Gel was Coomassie Blue stained over the weekend, destained using destain buffer with methanol, and imaged on a GelDoc light imager.

Angiotensin I converting enzyme 2 (ACE2) receptor complex binding assay was used to determine S1‐RBD protein binding affinity to ACE2 receptor after heat exposures. A commercial COVID‐19 Spike‐ACE2 Binding Assay Kit (RayBiotech, USA) with ACE2 on plate + Spike RBD query (CoV‐ACE2S2) format was used. Briefly, 5 µg of incubated and non‐incubated S1‐RBD protein was reconstituted in 125 µL assay buffer containing protein concentrated as instructed in the assay kit. Samples were then diluted 1:100 in assay diluent, added into ACE2 precoated microplate, and incubated at room temperature for 2.5 h with gentle shaking. Plate was washed four times with wash buffer. Next, 100 µL HRP‐Conjugated Anti‐IgG was added and incubated for 1 h at room temperature. To develop the plate, 100 µL TMB substrate was added to each well. The reaction was stopped with TMB Stop Solution after 30 min. Signals were read at 450 nm on a BioTek Synergy HT plate reader using BioTek Gen5 (BioTek). Samples were run in duplicates. Data were representative of an average of three sample replicates, and normalized to the positive control. Statistical analysis was performed using One‐way ANOVA with multiple comparisons (GraphPad Prism 8.0 (GraphPad, USA)).

Dynamic light scattering (DLS) was used to study the overall protein size/structure after heat exposure. Incubated and non‐incubated protein was dissolved in MilliQ‐water at 30 µg mL⁻¹ and filtered through a 0.45 µm PTFE syringe filter. For measurement, 0.45 mL of the sample was loaded into a UV‐transparent, disposable, semi‐micro cuvettes (Fisherbrand, Fisher Scientific, USA). DLS analysis was carried out using Zetasizer Nano ZS by Malvern Panalytical, UK. Data were plotted in Malvern Zetasizer software version 7.13.

Size exclusion‐high‐performance liquid chromatography (HPLC‐SEC) analysis was performed on Shimadzu LC‐10 system; Waters Acquity UPLC Protein BEH SEC column, 200 A, 1.7 µm, 4.6 mm × 150 mm (Waters Corporation, USA). Mobile phase contained 100 mm sodium phosphate (pH: 6.7) and was filtered with FisherScientific 0.2 µm aPES membrane vacuum filter unit (Fisherbrand, Fisher Scientific, USA) prior to use. 100% water seal wash (Optima) was used. Flow rate was 0.3 mL min⁻¹. Column temperature was held at 25 °C. Each 5 µL of sample was manually injected (≈10 µg protein column loading). Samples were prepared in mobile phase and filtered with Millipore Ultrafree low‐binding hydrophilic PTFE membrane filters (0.22 µm). Detection was at 220 and 280 nm wavelengths. Waters BEH200 SEC protein standard mix was utilized for calibration. Data collection was over 15 min and was plotted in GraphPad Prism 8.0 (GraphPad, USA).

2.3. S1‐RBD Protein Immunization

This study strictly followed the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the IACUC at University of Connecticut (#A18‐035). Animals were housed in the vivarium facility at the University of Connecticut, and received food and water ad libitum. The day before the insertion, the rats were shaved and Nair was applied for complete hair removal. The rats were placed under anesthesia using isoflurane (4% for induction and 3% for maintenance).

For the dosing study, female SAS SD rats (aged 8 weeks; Charles River Laboratories) were immunized with S1‐RBD‐protein at 10, 20, 50, and 100 µg dose⁻¹ via subcutaneous injections on days 0, 10, and 40. Saline was given as a negative control group. Blood collections were performed on days 0, 9, 14, 30, 44, 55, and 70. The serum was then collected from whole blood by centrifugation at 2,000 r.p.m. for 5 min and stored at −20 °C until use.

The MN patch was applied using a commercially available applicator (Micropoint Technologies Pte Ltd, Singapore) for 15 min. Then the supporting array was removed leaving the core–shell MNs embedded inside the skin. The subcutaneous injections were administered on days matching the MNs releases. Empty MNs were also given as a negative control group. Blood collections were performed accordingly. The serum was then collected from whole blood by centrifugation at 2000 rpm for 5 min and stored at −20 °C.

ELISA were performed on collected serum to determine the total binding antibody response. S1‐RBD‐protein (100 µL of 5 µg mL⁻¹) in 100 mm carbonate–bicarbonate buffer (pH 9.6) was added to Immulon 1B 96‐well plates and incubated overnight at 4 °C. Plates were then washed with 0.05% Tween‐20 in PBS and blocked for 1 h at room temperature with 2 µg mL⁻¹ bovine serum albumin (BSA) 0.05% Tween‐20 in PBS. Plates were then washed three times with 0.05% Tween‐20 in PBS. Serum diluted 1:10 in 0.05% Tween‐20 in PBS was then added at 90 µL per well and incubated for 1 h at room temperature. Wells were washed three times, and 100 µL of monoclonal goat anti‐rat IgG horseradish peroxidase (HRP)‐conjugated antibodies (1:1,000 in 0.05% Tween‐20 in PBS) were added. After incubation for 1 h at room temperature, wells were washed three times and 100 µL of TMB substrate solution was added to each well. The reaction was then stopped after 15 min by adding 100 µL TMB Stop Solution. Absorbance at 450 nm was measured using a Cytation 5 microplate reader (BioTek). The value was reported as optical density at 450 nm. Samples were run in duplicates. For end‐point binding titer, two‐fold serial dilutions were conducted. End‐point titers were determined as the final dilution at which signal in a well exceeded three times the background signal of the assay (as determined by averaging negative control values). Endpoint titers were reported as the reciprocal of that final dilution. Statistical analysis was performed in GraphPad Prism 8.0 (GraphPad).

2.3.1. Plaque Reduction Neutralization Test

Neutralization assays were performed in the BSL3 laboratory of the Ragon Institute of MGH, MIT, and Harvard. Vero‐E6 cells (ATCC) were seeded the day before the experiment in 48‐well plates at 40 000 cells per well in 250 uL of D10+ media (DMEM (Corning) supplemented with HEPES (Corning), 1× Penicillin 100IU/mL/Streptomycin 100 mg mL⁻¹ (Corning), 1× Glutamine (Glutamax, ThermoFisher Scientific), and 10% Fetal Bovine serum (FBS) (Sigma)) following harvest with Trypsin‐EDTA (Fisher Scientific). The serum samples were thawed on ice and diluted in D+ media (DMEM supplemented with HEPES, 1× Penicillin 100 IU/mL/Streptomycin 100 mg mL⁻¹, and 1× Glutamine) to the desired concentrations in 96‐well dilution blocks (Corning). Diluted stocks of SARS‐CoV‐2 variants (ancestral US‐WA1/2020 at 5e‐5 or delta B.1.617.2 at 5e‐4) were prepared in D+. An aliquot of viral stock was added to the serum dilution wells and the negative control wells (D+ only) in the dilution block to achieve the desired final concentration (5e‐6 for US‐WA1/2020 and 5e‐5 for delta) and incubated for 1 h at 37 °C at 5% CO₂. For each serum sample, 50 µL of virus‐serum dilutions (or no serum and no virus no serum controls) were added to triplicate wells in 48‐well plate. The viral inoculum was removed after 1 h and 0.5 mL of 1% methylcellulose (VWR) in DMEM supplemented with HEPES, 1× Penicillin 100 IU/mL/Streptomycin 100 mg mL⁻¹, 1× Glutamine, and 2% FBS was added to each well. After a 6‐day incubation at 37 °C and 5% CO₂, the methylcellulose media was carefully removed from each well, the cells were fixed in ice‐cold methanol at −20 °C for 15 min, and the plaques revealed with 1% aqueous Crystal violet solution (Fisher) and subsequent washes with water. The number of plaques in each well was counted manually.

The US‐WA1/2020 ancestral variant was obtained from BEI Resources. The delta variant isolate was obtained from the MassCPR variant repository. In brief, the variant was isolated at the Ragon BSL3 by rescue on Vero‐E6 cells from primary clinical specimens. The whole genome of subsequent viral stocks was sequenced to confirm that no additional mutation arose during virus expansion. Statistical analysis was performed in GraphPad Prism 8.0 (GraphPad).

2.3.2. Mechanical Testing

An ADMET eXpert 5952F fatigue tester (ADMET, USA) was equipped with two machined aluminum grips in its upper and lower clamp. MN patches were adhered to the lower aluminum grips with a thin layer of Loctite 411 adhesive (Henkel, Germany). The upper piston was lowered to apply compressive force to the MN patch at a rate of 6 mm min⁻¹ and data was sampled at 100 samples per second. The compressive force was triggered to stop at a ceiling value of 444.8 N. The compression testing data was plotted as compression force per needle as a function of the displacement.

2.3.3. Skin Irritation Assessment

The optical images of the rats’ skin were taken right after the MNs insertion and throughout the vaccination study. Three people who were blinded to the experiment setup scored the severity of erythema, eschar, and edema. Four (4) was the maximum possible score which indicates the severe skin irritation and (0) was the minimum score for normal, no irritation skin. The scores were displayed as the average and standard error of the mean (SEM).

3. Results and Discussion

3.1. Stabilization of S1‐RBD‐Protein Antigen for Long‐Term Skin‐Embedded MNs

The need to overcome the cold‐chain storage requirements and maintain vaccine thermostability is a major barrier to most vaccination campaigns. Incorporating vaccines into MN patches and eliminating the need for a reconstitution process can enable enhanced thermostability as compared to liquid form.^{[ 23 ]} Still, developing stabilized formulations that are specific to each vaccine and MN manufacturing process is essential for patch production, storage, and usage. Throughout the process of drying and storage, proteins can undergo critical stresses owing to the removal of their hydration shell which makes them susceptible to mis‐folding and denaturation.^{[ 24 ]} We sought to identify combinations of excipients which could successfully stabilize a protein antigen derived from the S1 protein of SARS‐CoV‐2, and containing the RBD of the virus, for implementing the single‐administration MN platform. We selected TS as the stabilizing excipients. This is because Trehalose and Sucrose could serve as water substitutions by direct interactions with proteins through hydrogen bonding. The sugars additionally help to minimize protein exposure to the environment and fill in void volumes. Highly viscous trehalose/sugar glass also entraps protein conformations, slows molecular dynamics, and eventually prevents protein misfolding.^{[ 25 ]} In the experiment, we performed formulation screenings for S1‐RBD protein antigen to identify the best formulation that would properly stabilize the antigen in varying temperatures. The S1‐RBD protein antigen was mixed with 0.5 m TS solution at 3 different ratios of 1:1 v/v, 1:2 v/v, and 1:4 v/v (Table S1, Supporting Information) and then vacuum dried overnight with desiccant at room temperature.

We exposed samples to heat for short‐term stability screenings (1 h) at 56, 70, and 100 °C to mimic temperatures that may be observed during the MN manufacturing process. A long‐term stability analysis was conducted at 37 °C (approximate human body temperature) for 2, 4, 8, 12, and 16 weeks. ELISAs were used to assess antigenicity of S1‐RBD protein after heat exposure. Figure 1 and Figure S1, Supporting Information illustrate the stabilization of S1‐RBD protein using the various TS formulations. While the pure S1‐RBD protein is stable and maintains ≈100% activity at 56 °C for a short period of time (≈60 min), the protein quickly degrades at 75–100 °C and exhibits a significant degradation for a long‐term heat exposure at 37 °C. In contrast, formulations of S1‐RBD which were mixed with TS retained 100% S1‐RBD protein activity even at high temperatures of 75–100 °C after 1‐h incubation. SDS‐Page analysis further confirmed that the S1‐RBD protein formulated with TS maintained its structural integrity after heat exposure (Figure S2, Supporting Information). Moreover, long‐term incubation at 37 ^°C indicated that the formulations with greater volumes of TS provided improved protein stability and outperformed those with lower concentrations of TS. Specifically, the S1‐RBD:TS at 1:4 v/v ratio formulation retained 100% antigen activity and showed a statistically significant difference from other groups. The 1:1 v/v and 1:2 v/v S1‐RBD:TS ratios maintained 70% and 80% activities, respectively in long‐term stability study. In comparison, pure S1‐RBD‐protein without excipients demonstrated a 60% reduction in activity. In general, the presence of TS as excipients has provided a crucial protection to S1‐RBD at high temperatures for both short‐term and long‐term incubation intervals. The S1‐RBD:TS at 1:4 v/v ratio has demonstrated the best performance among other formulations and thus was chosen for the continuation of our study. Additionally, we performed the ACE2 binding assay to investigate whether the S1‐RBD formulated with TS at 1:4 v/v ratio could maintain its functionality after heat exposure at 100 °C for 1 h and 37 °C for 16 weeks (Figure S3, Supporting Information). Results have indicated that the S1‐RBD:TS after heat incubations was able to maintain approximately the same level of binding efficacy to the ACE2 receptor as compared to the dried protein without heating (no statistically significant difference). Meanwhile, pure S1‐RBD protein significantly lost its activity after being exposed to heat. HPLC‐SEC and DLS data could further confirm that the S1‐RBD:TS 1:4 v/v has maintained the overall protein structure (Figure S4, Supporting Information). To sum up, we have identified suitable excipient formulations using various TS ratios to stabilize S1‐RBD for short‐term (up to 1 h) under high heat exposure (75–100 °C) and long‐term (up to 16 weeks) at average human body temperature of 37 °C, which is important for our next step of loading the protein into MNs.

Figure 1.

Stabilization of S1‐RBD‐protein antigen using Trehalose‐Sucrose dry formulation at different temperatures a) 100 °C for short‐term screening, and b) long‐term incubation for 3 months at 37 °C. (n = 3, independent experiments). Data are mean ± s.d. Each sample was run in duplicates. Statistical analysis was performed using two‐way analysis of variance (ANOVA) with repeated measures and Tukey test for multiple comparisons. The statistics denote the comparison of final time points.

3.2. New MN Fabrication Process with a High‐Throughput One‐Step Loading of Stabilized S1‐RBD Antigen into MN Patch

The MN patch manufacturing, presented herein, is technically based on our previously reported work^{[ 21 ]} and yet, we devised a new vaccine‐core loading process to further improve the manufacturing throughput and avoid multiple complex layer‐by‐layer alignment steps which are prone to errors and potentially cause loss of protein antigens. Using micro‐molding, 3D assembling, and heat sintering processes, we create a core–shell MN structure with three main components including a polymeric shell, a cap, and a dried vaccine core, positioned on top of a dissolvable supporting structure (Figure 2 ). Each MN has a height of 600 µm and a base diameter of 300 µm, and encapsulates a vaccine‐loaded core with a height of 400 µm and a diameter of 200 µm. Briefly, a biodegradable polymer film of poly(D,L‐lactide‐co‐glycolide) (PLGA) is compression‐molded into a silicone or poly(dimethylsiloxane) (PDMS) mold to form the base for the MN shell. The excessive PLGA scum layer on top of PDMS mold is then removed (Figure 2a: i,ii). The S1‐RBD protein cores are made separately by casting S1‐RBD protein mixed in TS solutions (1:4 v/v ration) onto another PDMS mold, which has structure and relative spacing similar to the MN shell mold. After a drying period, we then transfer arrays of the micro‐molded vaccine cores onto a PLGA film. The key novelty of this fabrication process is that; we align arrays of the vaccine‐cores, and load them directly into arrays of the heated MN bases/shells previously prepared (Figure 2a: iii,iv). The whole MN part is then placed in a heated vacuum oven for a short period of time, which allows the PLGA polymer to fill in the PDMS mold and encapsulate the loaded vaccine cores. To further elaborate on the novelty of the antigen‐loading process, here we only perform one alignment to load arrays of the cores into the MN shells while our previous work^{[ 21 ]} needs two separate alignment steps to 1) impinging the PLGA shell for creation of empty wells and 2) inserting the MN core into the wells (Figure 2b). Thus, the fabrication method herein significantly enhances the throughput and reduces the complexity and potential errors caused by multiple alignment steps in our previous work.^{[ 21 ]} This one‐step loading procedure also offers a high consistency, efficiency, and maximizes the amount of loaded cargos. In the next step, the core–shell MNs are capped and transferred onto a two‐layered Polylactic acid (PLA)‐ effervescent supporting array (Figure 2a: v). The effervescent layer provides a rapid detachment of the MNs from the supporting array upon contact with the local bodily fluid at the site of insertion,^{[ 26 ]} which minimizes the patch wear time (only 5–10 min before patch removal) and optimizes the MN application process. The MNs are released from the PDMS mold to yield the final complete vaccine‐loaded core–shell MN patch (Figure 2a: vi, b,c). From our previous work, the MNs have demonstrated their capabilities to penetrate and stay embedded under the skin (like an invisible tattoo) following insertion and rapid healing of the skin. The patch can be safely removed and disposed after administration (≈5–10 min) and produces no biohazardous sharp wastes. Mechanical strength measurement (Figure 2d) indicated a failure force of 0.15N per MN (under compression) which is sufficient for MNs penetration into the human skin without damages to the MN structures.^{[ 27 ]} The release times of the MNs are highly tunable by tailoring the molecular weight and compositions of the PLGA shell to accommodate different antigen release paradigms.^{[ 21 ]} Using PLGA shells with high molecular weights (MW) and more lactide component, we could obtain the delayed burst release (to mimic the booster dose) up to 4 months as seen in Figure S5, Supporting Information. Moreover, we conducted an in vitro release study of S1‐RBD protein loaded in the core–shell MN to investigate whether different encapsulated cargo could lead to changes in release timing of the PLGA MN. Results in Figure S6, Supporting Information demonstrate that the burst release time of S1‐RBD loaded MN was similar with our prior work.^{[ 21 ]} This is in agreement with published research in PLGA core–shell structures, which highlights that the release time points are influenced by polymer molecular weight and composition of PLGA shell, and not size, morphology, or loaded cargo.^{[ 28 ]} In principle, we could combine multiple sets of MNs made of different PLGA shells into one patch for a single‐time administration of antigen with built‐in booster doses. The embedded MNs can provide multiple burst releases of the encapsulated payload after predictable periods of time, offering an effect similar to that obtained from conventional multiple bolus injections.

Figure 2.

A new fabrication process with a high‐throughput one‐step MN core loading and testing of the S1‐RBD protein loaded core–shell microneedles. a) Assembly process illustrations of the MNs patch i,ii) PLGA polymer is compression molded into a negative PDMS MN mold under heating. iii,iv) Arrays of micro‐molded prepared antigen cores are loaded into arrays of MNs shell via a one‐step alignment process. v) After removing the scum layer, a supporting array of PLA‐effervescent with the cap‐layer, is aligned and assembled with the core–shell MNs through a heat‐sintering process. vi) The PDMS mold is peeled off to obtain the free‐standing core–shell MNs on the supporting array. b) The previously reported impinging and loading method to create the core–shell MN structure.^{[ 21 ]} c) The optical image of the MNs patch containing S1‐RBD protein. (Scale bar: 1 cm). d) The optical image of fabricated core–shell MNs (Scale bar: 300 µm). e) The mechanical behavior of S1‐RBD loaded core–shell microneedles made of PLGA 50:50 30 kDa under compressive force. Dash line represents insertion force to human skin of average 0.058N/needle. Data are mean ± s.e.m. n = 3 different microneedle patches.

3.3. In Vivo Immunogenicity of S1‐RBD Protein Loaded Core–Shell MNs

Prior to the MN construction, we first sought to identify an appropriate dose of S1‐RBD for use in the platform. We conducted a dose‐finding study by vaccinating SAS SD rats with the pure S1‐RBD protein antigen at 10, 20, 50, and 100 µg dose⁻¹ via subcutaneous (s.c.) injections on days 0, 10, and 40. Serum from vaccinated rats was collected and analyzed using ELISA to track changes in S1‐RBD‐specific IgG antibody responses over the course of the immunization schedule (Figure S7, Supporting Information). The negative‐control group that received saline injections on the same schedule did not demonstrate a detectable S1‐RBD‐specific antibody response. Animals receiving the 10 µg dose⁻¹ (Figure S7b, Supporting Information) and 20 µg dose⁻¹ (Figure S7c, Supporting Information) developed antibodies specific to S1‐RBD protein, though these responses were not statistically significantly higher than negative control animals (Figure S7a, Supporting Information). In contrast, the 50 µg dose⁻¹ (Figure S7d, Supporting Information) and 100 µg dose⁻¹ (Figure S7e, Supporting Information) groups exhibited statistically significant S1‐RBD‐specific IgG responses at day 55 in comparison to the saline control animals. As there was no statistically significant increase in antigen‐specific IgG induced by the 100 µg dose over the 50 µg dose, the 50 µg dose was selected for continuation of the study (Figure 7f).

To demonstrate the immunogenicity of the S1‐RBD loaded single‐administration core–shell MNs, similar to the multiple vaccine‐injection regimes, we first fabricated the MN patches each containing 50 µg S1‐RBD protein and applied the patches on rats. One 1 cm × 1 cm MN patch could contain the same amount of active protein as one dose of s.c. injection. We confirmed this dose matching and the stability of the S1‐RBD protein after the MN fabrication process using ELISA (Figure S8, Supporting Information). The rats then received S1‐RBD vaccinations at a single time with the MN administration onto the skins (S1‐RBD MN) or multiple‐time s.c bolus injections (S1‐RBD s.c). In the S1‐RBD MN group, rats received three MN patches (50 µg of stabilized S1‐RBD protein per patch for the total amount of 150 µg protein) at the initial time for the releases at day 0, 10, and 40. Similarly, S1‐RBD s.c group was injected with 50 µg of pure S1‐RBD stored at −80 °C per dose at time points matching the release time of the MN. Subcutaneous, intradermal, and intramuscular routes have been shown to elicit similar immune responses in several vaccines.^{[ 29 ]} Many preclinical studies on MN for vaccination have utilized s.c administration as a reference to compare with MN.^{[ 30} , ^{31 ]} And therefore, s.c. injection was chosen in our work. We used a vaccination schedule that had a priming dosage starting on day 0 followed by two boosts on day 10 and on day 40 (Figure 3a). The MNs were constructed accordingly (with the PLGA shells, adapted from our prior published work for the burst releases at days 0, 10, and 40) to replicate this administration schedule.^{[ 21 ]} Furthermore, since PLGA degradation releases acidic by‐products that decrease the local environmental pH, we assessed the stability of R1‐RBD to ensure the protein maintained its stability at low pH (Figure S9, Supporting Information).^{[ 32 ]} We also performed investigation of the MN insertion process. Results from Figure S10, Supporting Information demonstrate that the MN could be gently and efficiently inserted into the skin. And the skin was then quickly healed to entrap the MNs. The implantation of MNs inside the skin did not cause any noticeable skin erythema, edema, or any irritation and inflammation in all groups for the entire duration of the studies, which suggests that the core–shell MNs are safe and well‐tolerated (Figure 3b and Figure S11, Supporting Information). Histological analysis of the rat skin further showed no detectable traces of tissue damages or inflammatory cells in the group immunized via MN (Figure S12, Supporting Information).

Figure 3.

Immunogenicity of S1‐RBD protein loaded core–shell microneedle patch. a) Vaccinations and serum collection schedule. b) Erythema and eschar formation, and Edema formation on the rat skins. Data are mean ± s.e.m. n = 5 rats per group, independent experiments. Scores range from 0 (normal) to 4 (severe). c)The immune responses over the course of time from the rats receiving the S1‐RBD:TS 1:4 v/v protein loaded core–shell MNs, multiple bolus subcutaneous injections of pure S1‐RBD protein stored at −80 °C, and empty MNs. Bars represent mean ± S.E.M. (n = 4–5 rats per each group, independent experiments). d) End‐point IgG titers of serum collected on day 70 were analyzed by ELISAs. Statistical analysis was determined via One‐way ANOVA with Tukey's multiple comparisons post‐hoc test. e,f) Plaque reduction neutralization test (PRNT) results of serum collected on day 70 against US‐WA1/2020 and Delta B.1.617.2 variants. Bars represent mean ± s.e.m (n = 4–5 rats per each group, independent experiments). Statistical analysis was determined via Kruskal–Wallis ANOVA with Dunn's multiple comparisons test.

The development of antibodies against the S1‐RBD protein over time was recorded from serum samples of vaccinated rats using ELISA. Antigen‐specific IgG responses in S1‐RBD‐MN rats were highly similar to those of S1‐RBD‐s.c rats receiving multiple bolus s.c. injections, with no statistically significant difference identified (Figure 3c,d and Figure S13, Supporting Information). The end‐point binding antibody titer increased to 10⁴–10⁵, which is even comparable to the reported range for those values obtained after vaccination of rodents with commercial COVID‐19 vaccines (e.g. Moderna).^{[ 33 ]} Meanwhile, the negative‐control group receiving empty PLGA MNs exhibited no measurable immune response. Furthermore, to investigate the functionality of the induced antibodies, serum samples from vaccinated rats were analyzed for neutralizing activities against two authentic SARS‐CoV‐2 strains (ancestral US‐WA1/2020 and delta B.1.617.2) using the plaque reduction neutralization test. The rats immunized with S1‐RBD loaded single‐administration MNs developed robust neutralizing antibodies against both SARS‐CoV‐2 strains relative to background activity in unvaccinated animals. Neutralizing titers trended higher in S1‐RBD‐MN vaccinated animals in comparison to animals receiving multiple s.c. injections of the same antigen, though these differences were not statistically significant (Figure 3e,f). The MNs may have enhanced immunogenicity through stimulation of professional antigen presenting immune cells in the dermal layer of the skin via mechanical stress^{[ 34} , ^{35 ]} though this mechanism may require further investigations with our MN system. On the other hand, serum from the unimmunized control group displayed modest neutralization capacity (although some certain level of neutralizing activity is still naturally present). Note that we did not use adjuvants in this study so the antibody titers were not extremely high and exhibited some variations. The inclusion of an adjuvant may provide a dose‐sparing and consistent effect on the MN‐delivered protein antigen formulation or increase the long‐term durability of the immune response, though this will need to be determined in future studies.^{[ 36,b ]} However, the results presented herein clearly demonstrate the immunogenic efficacy of our single‐administration stabilized‐antigen MN platform, comparable to that obtained by multiple injections of the fresh antigens. Our MNs again offer unique significant advantages of being painless and easy‐to‐use while avoiding the need of cold chain requirements and repeated administrations.

4. Conclusion

We have devised a new approach to thermally stabilize a protein antigen model S1‐RBD which is a candidate for SARS‐COV‐2 vaccines development, and incorporate the antigen into a recently‐developed single‐administration multi‐burst release MN platform via a novel high‐throughput single‐step antigen‐loading process. The MNs are capable of releasing stabilized protein antigens at specific predetermined time points with only a single‐time skin application, effectively mimicking the immunogenic effect of multiple bolus vaccine injections. We successfully utilized this platform to deliver a highly relevant protein antigen, the S1‐RBD protein of SARS‐CoV‐2, which triggers a significant neutralizing antibody response in a rat model. The inclusion of stabilizing agents allowed the S1‐RBD MN patch to retain immunogenicity, even at an extreme temperature of 100 °C for at least 1 h and after long‐term storage (at least 4 months at 37 °C) outside of the normal cold‐chain conditions typically required for vaccines, further demonstrating the utility of this platform in geographically isolated and austere environments. This is the first demonstration of using common excipients to stabilize S1‐RBD protein antigen—a candidate for COVID‐19 vaccine development. Loading of the MNs with recombinant SARS‐CoV‐2 S1‐RBD protein induced both binding and functional neutralizing antibody responses in rats against different strains of SARS‐CoV‐2 virus, including the variant of concern B.1.617.2 (Delta), thus demonstrating the potential of this platform to being a viable needle‐free alternative to the traditional vaccination paradigms. The single‐administration burst‐release MNs with stabilized protein antigens, presented herein, could revolutionize the vaccination process by avoiding painful, inconvenient injections, trained medical personnel, cold‐chain facilities for vaccine storage, and the need for booster shots.

For the future development and clinical translation of this technology, there are limitations of the current work which need to be addressed. The addition of suitable adjuvants to the formulations may help to boost and provide consistence for the observed immune response^{[ 16} , ^{37 ]} and thus, there is a need for studying the effect of adjuvants on this MN system. Additionally, future studies need to assess the activity of the MNs in large animal models (e.g. pigs or monkeys with immune systems and skin properties more similar to those of humans) with larger sample sizes. Furthermore, continued improvements on the fabrication processes, especially product streamlining and automation, could enable the mass production of the MN patch in response to surge events and expedite the immunization of populations against SARS‐CoV‐2 variants or potential future viruses in pandemics. Regardless of these future works, the single‐administration MNs with stabilized protein antigens and a high‐throughput method for antigen loading, presented herein, could be the key solution to the success of global immunization campaigns against highly contagious viruses and infectious diseases by enabling an easy‐to‐use, easy‐to‐distribute, and painless vaccine delivery platform with built‐in programmable booster release.

Conflict of Interest

T.D.N. has a conflict of interest with the companies PiezoBioMembrane Inc. and SingleTimeMicroneedles Inc.

Author Contributions

K.T.M.T. and T.D.G. contributed equally to this work. K.T.M.T., T.D.G., E.R.T., S.M.S., and T.D.N. designed the project and experiments. K.T.M.T., T.D.G., S.M.S., and T.D.N. wrote the manuscript. K.T.M.T., T.D.G., T.T.L., F.L., and Y.L. performed the research and experiments. K.L. and A.G. contributed to the experimental design and aided in the drafting of the manuscript.

Supporting information

Supporting information

Tran K. T. M., Gavitt T. D., Le T. T., Graichen A., Lin F., Liu Y., Tulman E. R., Szczepanek S. M., Nguyen T. D., A Single‐Administration Microneedle Skin Patch for Multi‐Burst Release of Vaccine against SARS‐CoV‐2. Adv. Mater. Technol. 2022, 2200905. 10.1002/admt.202200905

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.