Implantable device with magnetically rotating disk for needle‐free administrations of emergency drug

Abstract Prompt administration of first‐aid drugs can save lives during medical emergencies such as anaphylaxis and hypoglycemia. However, this is often performed by needle self‐injection, which is not easy for patients under emergency conditions. Therefore, we propose an implantable device capable of on‐demand administration of first‐aid drugs (i.e., the implantable device with a magnetically rotating disk [iMRD]), such as epinephrine and glucagon, via a noninvasive simple application of the magnet from the outside skin (i.e., the external magnet). The iMRD contained a disk embedded with a magnet, as well as multiple drug reservoirs that were sealed with a membrane, which was designed to rotate at a precise angle only when the external magnet was applied. During this rotation, the membrane on a designated single‐drug reservoir was aligned and torn to expose the drug to the outside. When implanted in living animals, the iMRD, actuated by an external magnet, delivers epinephrine and glucagon, similar to conventional subcutaneous needle injections.


| INTRODUCTION
Emergencies caused by life-threatening medical events, such as anaphylaxis or hypoglycemia, often occur in community settings where healthcare professionals are not present. [1][2][3][4] Anaphylaxis, a potentially fatal allergic reaction, has been reported to affect up to 5% of the US population, with a 1% mortality risk. 5,6 Hypoglycemia is defined as an abnormally low plasma glucose level, 7,8 which is more relevant in patients with type 1 diabetes, 9 accounting for a relevant death rate of more than 2%. [10][11][12] In such emergency situations, there is a golden time for prompt drug administration to prevent a fatal event of death and increase the survival rate. 13 Therefore, for the treatment of anaphylaxis and hypoglycemia, immediate administration of first-aid drugs, such as epinephrine [14][15][16] and glucagon, 10,17 is recommended when the patient first perceives the symptoms. For rapid systemic exposure, these drugs are prescribed to be administered by needle self-injections, 13,17 where the patients need to be equipped with a prefilled syringe or kit to be filled with drugs, 18,19 and properly trained for injections at all times. 20 However, the symptoms of anaphylaxis and hypoglycemia often include blocked breathing, dizziness, or trembling, 21,22 which makes self-injections difficult. 19 When involved with needle phobia, which is reported to interfere with 7%-22% of patients prescribed needle self-injections, 15,23,24 the chances of proper drug injections could be lower under such emergency conditions. As a result, an implantable device with on-demand drug delivery may be advantageous as a life-saving strategy in emergencies. The device is prefilled with drugs sufficient for multiple doses to be used with patients at all times after implantation. Moreover, when the symptoms are perceived, the patient can immediately respond to administering a first-aid drug through a modality without needles. 25,26 However, in many previous devices, such on-demand drug delivery was enabled with electronic actuators, circuits, and batteries embedded in the device, [27][28][29] making it bulky and inconvenient for implantation. Considering fatal but infrequent emergency occurrences, the limited lifetime of the battery due to the discharge through the implemented circuit can be a disadvantage, which may require replacement surgery even before complete drug consumption. [27][28][29][30][31] Therefore, we propose a battery-less implantable device actuated solely by magnetic force for the on-demand delivery of emergency drugs. Inside the device, there is a disk containing multiple drug reservoirs arranged at a precise angular spacing and hermetically sealed with a membrane. The disk contains a magnet that allows for the actuation of up and down movements by a magnetic force applied externally from the device. At each actuation, the disk can rotate at a precise angle along a guiding path, where the membrane of a specifically designated single reservoir can be aligned and torn by a pole formed in the device, thereby exposing and releasing the drug. This actuation is based on mechanical means enabled only when a magnet is applied from the outside skin, thereby on-demand drug delivery without electronic power.
In this study, an implantable device with a magnetically rotating disk (iMRD) was assessed with epinephrine and glucagon, representative first-aid drugs for anaphylaxis and hypoglycemia, respectively.
The drug-loaded iMRD was implanted subcutaneously in rats, and a magnet was applied from the outside skin at predetermined times to assess the capacity of on-demand drug delivery. We examined the pharmacokinetic (PK) profile and compared it to that of conventional subcutaneous needle injections. For glucagon, the pharmacodynamic (PD) profile (i.e., change in plasma blood glucose level) was also monitored in hypoglycemia-induced rats.

| iMRD design and working principle
The iMRD was prepared by assembling and bonding the constituent units, each of which was prepared by 3D printing, as shown in  Through a punctured membrane, the drug in the reservoir is released and exposed to the outside through the perforated ceiling and floor of the iMRD (⑨) while the disk is back to the stationary state. This process is repeated at every actuation to rotate the disk and puncture the membrane of the next designated reservoir.

| In vitro performance tests
To examine the performance of on-demand drug delivery, the iMRD, each loaded with epinephrine or glucagon, was fully submerged in PBS (pH 7.4) and actuated six times consecutively using an external magnet applied on the top of the iMRD at a 1 mm spacing to simulate the skin over the implanted device. 32,33 For both epinephrine and glucagon, almost all drug in a reservoir was dissolved and exposed to the outside of the iMRD quite rapidly, within minutes. As shown in Figure 1a,b, the amount of drug released per actuation was highly reproducible, which was measured as 980.13 ± 10.45 and 19.20 ± 0.85 μg per actuation for epinephrine and glucagon, respectively.
After six actuations, the iMRD was disassembled to observe the rotating disk, where the six distinct holes made in the membrane over the designated reservoirs were clearly observed, while the rest remained intact ( Figure S2 and Movie S2).
When the iMRD was immersed in PBS for a long-term period of 30 days, the drug was released only when the external magnet approached to actuate the iMRD, as shown in Figure 1c,d. No added drug was detected immediately before actuation, indicating that there was no leak or unexpected membrane breakage when the iMRD was in the stationary state. Considering the implantation, the stability of epinephrine and glucagon loaded in the iMRD was also tested under simulated biological conditions, that is, incubation at body temperature. As shown in Figure 2e,f, almost all the drugs retained their stability for up to 90 days. When glucagon after 90-day incubation was injected into hypoglycemic animals, the increase in blood glucose level was not different from that of the animals injected with fresh glucagon; hence, there was retained biological activity ( Figure S3).

| In vivo performance tests
To assess the capacity of needle-free delivery and prompt systemic exposure to drugs, an iMRD loaded with epinephrine, a representative emergency drug, was first tested under in vivo experimental conditions. For this, the iMRD loaded with epinephrine was subcutaneously implanted (i.e., Epi-iMRD) for 45 days and actuated by a magnet at scheduled times, where the PK profile was compared with the animals subcutaneously injected with epinephrine (i.e., Epi-inj). As shown in Figure 3 (Table S1) was also similar between the two groups (p > 0.05). In this study, the drug was administered at a relatively wide time interval of 15 or more days for a total of 60 days, and it should be noted that the PK profile was reproducible at all administration times with the iMRD.
When loaded with glucagon, another emergency drug, iMRD delivered the drug in a manner similar to subcutaneous injections. As shown in Figure 4a-d (Table S2), the PK profiles of Glu-iMRD and Glu-inj were quite similar, showing the same T max (15 min) and similar C max and AUC (p > 0.05). This PK profile was reproducible at all administration times using iMRD. For glucagon, we also examined the PD profile, that is, the profile of blood glucose levels, in hypoglycemia-induced animals. As shown in Figure 4e-h, after actuation of the iMRD, the blood glucose level increased, indicating viable bioactivity of glucagon. Notably, the PD profile of Glu-iMRD was quite similar to that of Glu-inj at all administration times for 60 days, suggesting an effective hypoglycemic treatment similar to conventional needle injections.
F I G U R E 2 In vitro performance test of the device, which was performed while the implantable device with a magnetically rotating disk (iMRD) was fully immersed in pH 7.4 PBS at 37 C. The device was actuated six consecutive times, and the amount of newly released (a) epinephrine and (b) glucagon was measured after each actuation. Cumulative release amounts of (c) epinephrine and (d) glucagon were measured before and after the iMRD actuation at scheduled times of 1, 3, 5, 10, 20, and 30 days. For stability evaluation, (e) epinephrine and (f) glucagon were stored in the reservoir of the device at 37 C for 30, 60, and 90 days. After each incubation time, the content was measured using highperformance liquid chromatography (HPLC). For each drug kind, and measurement and incubation time, the experiment was performed in triplicate. Data are presented in the form of mean ± standard deviation (SD).

| Histology
To evaluate in vivo biocompatibility, the tissues surrounding the iMRD were biopsied at 15 and 60 days after implantation. In this study, we assessed H&E-stained tissues near two different locations, that is, the top and side walls of the iMRD, which represent the tissues near and far from where an external magnet was applied, respectively ( Figure 5a). Overall, the results of the histological analysis were not different between Epi-iMRD and Glu-iMRD.
As shown in Figure 5c, a mild inflammatory response was observed in the periphery, which was not further upregulated as time elapsed after implantation. Figure 5d shows fibrotic capsules formed at both tissue locations, as reported for many other nondegradable implants. 28,34,35 However, the thickness was not significantly different between the two locations (p > 0.05), suggesting no significant mechanical stress due to magnetic actuation. From 15 days after implantation, the capsule thickness stabilized without an increase, as observed in our previous studies. 36,37

| DISCUSSION
In life-threatening medical emergencies in community settings, such as anaphylaxis and hypoglycemia, prompt drug administration can actually save lives. 3,15,17 These events require rapid systemic drug exposure, which is often achieved using needle injections. For this purpose, the needle-based auto-injectors are already commercially available, 38,39 where the individual needs to be well established and experienced with such a first-aid injection plan without the presence of medical professionals. [40][41][42] However, medical emergencies are often accompanied by symptoms that interfere with delicate movements, and needle injections are difficult to perform. 43,44 Therefore, we proposed an implantable device, that is, the iMRD, that could deliver a first-aid drug of interest easily and noninvasively by a simple magnet application from the outside skin. The structure of the iMRD was carefully designed to provide on-demand drug delivery capacity (Figure 1), and the use of magnets in an implantable device has already been adapted clinically, such as in cochlear implants. 45 Multiple drug reservoirs were hermetically sealed with a biocompatible titanium membrane 46,47 ; therefore, no drug was released in the absence of magnetic applications (Figure 2c,d). These reservoirs were located at an accurate angular spacing on the disk, and the movement path of the disk was precisely guided through the grooves in the iMRD. Thus, only at the time of magnet application, the disk could rotate to accurately align and tear the membrane of a single designated reservoir ( Figure S2 and Movie S2).
We pursued to obtain a similar PK profile between the groups of iMRD and subcutaneous injections. For this, we intentionally prepared the ceiling and floor of the iMRD to be perforated so that a bodily fluid could be readily in contact with the membrane. The drug in each reservoir was formulated to be a highly porous tablet of pure drug itself by freeze-drying. Thus, right after the breakage of the membrane, the drug could be rapidly dissolved and exposed to the outside of the iMRD (<10 min), which allowed for prompt systemic exposure of the drug, as observed with subcutaneous needle injections (Figures 3 and 4).
The actuation of the IMRD was based solely on magnetic force without any other power sources, thereby eliminating batteries or electronics. This would allow for smaller dimensions and long-term operation of the device after implantation compared to other implantable devices for on-demand drug delivery. 48,49 The device size can be further reduced when fabricated in a more sophisticated manufacturing facility. 50 The attraction force between the disk and bottom magnets (c.a. 0.12 N) was strong enough to prevent an accidental movement of the disk (weighing c.a. 2 g). Upon this setting, the iMRD herein was designed to actuate with a relatively strong magnetic field (>3000 G) applied at a distance greater than the thickness of human skins (Table S3) Considering translational applications, the actuation device housing may need to be made of harder materials, such as metals, to better protect the drug in the device after implantation. 48,52 In our prototype, 12 separate drug reservoirs were used. The average incidence of anaphylaxis and severe hypoglycemia is reported to be once a year or once every 2 years, 53,54 implying that the iMRD in its current form can be envisioned for more than a 10-year use after a one-time implantation. In this aspect, an iMRD with a smaller number of reservoirs would still be useful; hence, a smaller device dimension. Unlike implantable drug delivery devices based on liquid pumping, 37,55-57 the iMRD releases the drug by opening a single reservoir, each filled with an accurate dose of drugs in dry form, which is expected to generally improve the drug stability for prolonged periods compared with that in liquids, especially for epinephrine and glucagon. 58

| iMRD Fabrication
The major constituent units of the iMRD were designed using 3D CAD SolidWorks (Dassault Systemes, Vélizy-Villacoublay, France) and fabricated using a 3D printer (Objet 30 Pro, Stratasys, Rehovot, Israel). As shown in Figure 1a (see Figure S1), the iMRD was composed of three main parts: a top case, rotating disk, and bottom case.
In the top case, a groove for a guiding path was formed in the inner wall to allow for the rotation of the disk at a precise angle during its upward and downward movement. A pole was placed on the ceiling to puncture the membrane on the drug reservoir when the disk moved upward. The ceiling was perforated so that the drug exposed after membrane breakage could be released outside the iMRD.
Three guide balls were made at the side wall of the rotating disk ( Figure S1) with an angular spacing of 120 , allowing the disk to slide and rotate through the tilted grooves made in the top case during the upward and downward movements. In the body of the disk, 12 individual drug reservoirs were prepared at an angular spacing of 30 , each of which was filled with 1000 μg of epinephrine or 20 μg of glucagon. To load the drug, 10 μl of an aqueous solution of epinephrine (100 mg/ml) or glucagon (2 mg/ml) prepared in sterile water was added to each reservoir, and the whole disk was exposed to liquid nitrogen and lyophilized for more than 6 h (FreeZone 6 dryer system; Labconco, USA).
Immediately thereafter, the disk was seamlessly sealed with a titanium membrane using medical epoxy, which was then covered with a cap. A donut-shaped magnet (0.16 T) was bonded to the bottom of the disk (i.e., the disk magnet). In the bottom case, a donut-shaped magnet (0.14 T) was inserted into the shaft and bonded to the floor using medical epoxy (i.e., the bottom magnet). The floor was also prepared to be perforated to allow drug release to the outside. For the final assembly, the rotating disk was inserted into the shaft of the bottom case. Then, the top and bottom cases were aligned using male and female connectors and bonded using medical epoxy.

| iMRD characterizations
To measure the actual drug-loading amount, a drug-loaded rotating disk without a membrane was fully immersed in 10 ml of phosphate- To evaluate the in vitro performance of on-demand drug delivery, the iMRD was fully immersed in 10 ml of PBS (pH 7.4) at 37 C, and actuated with a magnet (0.3 T) externally approaching the ceiling of the iMRD at a gap of 1 mm to mimic the presence of the skin layer. 32,33 Actuation was performed at scheduled times of 1, 3, 5, 10, 20, and 30 days. Immediately before and after each actuation, 5 ml of medium was extracted and replaced with fresh PBS. The collected medium was analyzed by HPLC as described above to measure the concentrations of epinephrine and glucagon.

| Drug stability evaluation
Considering the drug stability after implantation, a reservoir was prepared as in the iMRD, which was filled with epinephrine or glucagon and sealed with the membrane. The drug-containing reservoir was stored in an incubator at 37 C for 30, 60, and 90 days. After each incubation, the membrane was torn to expose the drug in the reservoir, which was fully immersed in 1 ml of PBS (pH 7.4) to dissolve the drug. The medium was collected and assessed by HPLC, as described above, and compared with that of a fresh drug without incubation.
Experiments were performed in triplicate for each incubation time for epinephrine and glucagon.

| Animal study
All in vivo experimental procedures were performed using Wistar rats (Charles River, Massachusetts, USA), weighing 180 ± 30 g, following institutional guidelines approved by the Institutional Animal Care and Use Committee at Seoul National University Hospital Biomedical Research Institute (approval no. 21-0234-S1A0). The animals were housed in a pathogen-free facility with controlled temperature, humidity, and a 12:12 h light/dark cycle. Standard chow and water were provided ad libitum. Prior to implantation, all iMRDs were sterilized with ethylene oxide gas. 61 To implant the iMRD, the animal was anesthetized by isoflurane inhalation, and the flank area was shaved and sterilized with betadine. A 33-mm incision was made and the iMRD was implanted into the subcutaneous pocket. The incision was closed with surgical sutures and disinfected with betadine. Gentamicin (10 mg/kg) and acetaminophen (50 mg/kg) were then injected subcutaneously and intraperitoneally, respectively. for Epi-inj and Epi-iMRD, respectively. At 0, 15, 30, 60, 120, and 180 min after each drug administration, 250 μl of blood was extracted from the tail vein and centrifuged at 3000 g for 10 min to collect 100 μl of plasma. The drug concentration in blood plasma was analyzed using an epinephrine/adrenaline ELISA kit.

| PK and PD evaluation of glucagon
For the PK test of glucagon, the animals were divided into two groups: (1) Glu-inj (n = 4): animals treated with subcutaneous injections of 10 μg glucagon; and (2) Glu-iMRD (n = 4): animals implanted with glucagonloaded iMRD containing 20 μg glucagon in each reservoir. Drug administration and blood extraction were performed as previously described. A rat glucagon EIA kit was used to measure glucagon concentration in the blood plasma. For the PD test of glucagon, we measured the blood glucose level using a glucometer (Accu-check Performa, Basel, Switzerland).
Prior to each administration day, the animals were fasted overnight to induce hypoglycemia, which was confirmed by a blood glucose level of less than 70 mg/dl. 36,62 At 0, 15, 30, 60, 120, and 180 min after glucose administration, 1-2 μl of blood was collected from the tail vein and measured using a glucometer.

| Histology
At 15 and 60 days post-implantation, the animals in the Epi-iMRD (n = 4) and Glu-iMRD (n = 4) groups were sacrificed by carbon dioxide inhalation, and the tissue surrounding the iMRD was harvested from two distinct locations, that is, the top and side walls of the iMRD ( Figure 5a). The tissue was then fixed in 4% paraformaldehyde and embedded in paraffin wax. The paraffin block was cut into slices of 4 μm thickness, among which three slides were randomly selected from each day and location in each animal to prepare tissue slides.
The tissue was then stained with hematoxylin and eosin (H&E) to assess the degree of inflammation and fibrous capsule thickness around the iMRD, as described in our previous studies. 34,63 The analysis was performed by a professional pathologist (C.L.) in a blinded manner using an optical microscope at Â40 and Â100 magnification (Nikon, Eclipse Ci-L, Tokyo, Japan).

| Statistical analysis
Data are presented as mean ± standard deviation, and differences between groups were determined by two-way analysis of variance (ANOVA) followed by post hoc Tukey's test for multiple comparisons.
Student's t-test was used to compare the means of the two groups.
Statistical analysis was performed using GraphPad Prism 7 (GraphPad Software, CA, USA). Differences were considered statistically significant when p < 0.05.

DATA AVAILABILITY STATEMENT
Data are available from the authors upon reasonable request.