Biodegradable Intra-arterial Devices for Focal Drug Delivery to Targeted Organs

Abstract

This study presents the development of biodegradable intra-arterial drug delivery (IADD) devices for focal treatment of targeted organs, to enhance therapeutic efficacy while minimizing systemic toxicity. The IADD devices are fabricated using magnesium (Mg) and poly(glycerol sebacate) (PGS), leveraging their biocompatibility and tunable biodegradability, and are loaded with two model drugs, i.e., dexamethasone (DEX) or cisplatin (CIS). The IADD devices with helical and linear designs were fabricated for focal drug delivery to targeted organs and characterized for their microstructure and composition using scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDS), thermogravimetric analysis (TGA), and Fourier-transform infrared spectroscopy (FTIR). The results confirmed the successful incorporation and stability of the drugs within the device. The IADD devices demonstrated a sustained release of DEX and CIS over 30 days in vitro, with cumulative release of 373.11 ± 1.41 μg and 64.73 ± 0.06 μg, respectively. The IADD devices demonstrated cytocompatibility with endothelial cells and sustained pharmacological activity against glioma cells throughout the in vitro release period. We implanted DEX-loaded IADD devices into the artery upstream of a target organ in rat models. The devices implanted into the renal artery to target the kidney and the carotid artery to target the brain achieved 109-fold and 68-fold improvements, respectively, in organ vs systemic drug levels compared to oral drug administration. These results proved the safety and efficacy of the IADD devices for sustained, focal drug delivery of different drugs to the target organs, with reduced systemic drug exposure. Overall, the results demonstrated the potential of the IADD devices as a valuable platform technology to achieve focal drug delivery to targeted organs for a wide range of clinical applications, especially for delivering drugs with high efficacy, high systemic side-effects, and narrow therapeutic window.

1. Introduction

The development of focal drug delivery systems represents a significant advancement in treating diseases localized to specific tissues or organs. Compared with conventional methods, such as intravenous (IV) and oral administration, these systems allow for precise medication delivery directly to the targeted area while reducing systemic exposure and off-target side effects [1]. Several delivery techniques have been proven to achieve focal drug delivery, including intra-arterial (IA) injectables [2], pumps [3], and implants [4]. Studies show that IA chemotherapy achieves markedly higher local drug concentrations while reducing systemic exposure compared to conventional systemic routes. For example, IA cisplatin infusion in high-grade glioma patients yielded 8–12-fold higher tumor platinum levels and ~30% lower systemic levels (i.e., area under the curve, AUC) than intravenous dosing [5, 6]. Likewise, hepatic artery infusion of oxaliplatin or albumin-bound paclitaxel achieved 5–10-fold higher tumor/plasma ratios with 15–60% lower systemic exposure [7, 8]. In a recent single-center series of 70 glioblastoma patients undergoing 139 super-selective IA cerebral infusions, the completion rate of intended treatments was 95.7% and drug-related adverse events occurred in only 8.6% of cases [9]. Collectively, these findings underscore the pharmacokinetic and safety advantages of focal IA delivery of maximizing on-target effect while minimizing off-target exposure. However, current IA technology allows for acute delivery only, limited to the duration of endovascular catheterization.

The objective of this work is to develop and validate implantable intra-arterial devices as a platform technology for sustained focal drug delivery to targeted organs. In this study, we designed novel biodegradable intra-arterial drug delivery (IADD) devices, which could be placed into large- or middle-sized arteries via minimally invasive endovascular procedures and provide tunable release profiles for focal drug delivery to any organ perfused by a large- or middle-sized upstream artery. Compared with direct organ implants, IADD devices can be implanted endovascularly, offering a less invasive option while still achieving high local drug concentrations in the targeted downstream organs. Moreover, our IADD devices can be loaded with a wide range of therapeutics and fabricated into a variety of geometries such as helical structures, linear plank structures, half to full ring structures, or stent-like structures for controlling the amount of drug loading and release. Additionally, IADD devices has the potential for broad organ perfusion, compared to the diffusion-limited dispersal of direct organ implants [10].

For the first-generation prototypes of our novel IADD device, we selected magnesium (Mg) as the base material and poly(glycerol sebacate) (PGS) as the drug-loading polymer for the following reasons. Both Mg and PGS are biodegradable materials [11–13] eliminating the need for a subsequent procedure to remove the device, thereby minimizing potential damage to the implanted vessel [14]. Mg serves as the backbone of the device, providing the mechanical stiffness and strength required for endovascular delivery and for retaining the shape of the device in the artery. Mg is widely used in biomedical implants due to its biocompatibility and biodegradability, as well as mechanical properties [15–17] . Furthermore, Mg itself may improve intra-arterial drug delivery, as previous studies of Mg in vascular stents have shown that the released Mg²⁺ enhances blood flow, inhibits platelet activation, and prevents vasoconstriction [18]. Mg is fully bioresorbable, excreted through urine and feces as Mg²⁺ ions [19, 20]. PGS has gained prominence in tissue engineering and drug delivery due to its customizable and straightforward synthesis process [21]. PGS degrades via surface erosion in the body fluids, which affords stable drug release compared with polymers that degrade via bulk erosion such as poly-lactic acid (PLA) and poly-glycolic acid (PGA) [22]. The degradation products, glycerol and sebacic acid, are biocompatible and are naturally found and metabolized in the human body [23]. The synthesis conditions of PGS [24], such as reaction times and temperatures, can be fine-tuned to control the degree of esterification, allowing for precise tailoring of the chemical and mechanical properties of the synthesized polymer as well as its degradation behavior [25]. Notably, PGS is nonimmunogenic and has shown to be non-cytotoxic, causing minimal inflammatory response [26]. In addition, the mechanical properties of PGS, which are similar to blood vessels, can minimize the mechanical mismatch between the IADD device and arteries [13]. These characteristics make our IADD devices conducive for sustained intra-arterial drug delivery.

To demonstrate the potential of our novel IADD devices, we selected two model drugs—dexamethasone (DEX, C₂₂H₂₉FO₅) and cisplatin (CIS, Pt (NH₃)₂Cl₂)). Both DEX and CIS are acknowledged not only for their effectiveness, but also for their high systemic toxicity. DEX, a synthetic corticosteroid, is used to treat inflammatory and autoimmune diseases [27], allergic reactions [28], respiratory issues [29], endocrine disorders, and certain cancers [30]. However, its use is restricted due to severe side effects including metabolic disorders such as obesity and insulin resistance, hypertension, and susceptibility to infections [31]. CIS, a chemotherapeutic agent, is widely used against solid tumors by binding to DNA and inducing apoptosis, but its clinical application is limited by nephrotoxicity and other severe toxicities [32]. By delivering these drugs intra-arterially to the target organs, IADD devices could enhance target organ drug levels and reduce systemic exposure, thereby increasing drug efficacy and minimizing harmful off-target side effects.

To prove the concept of IADD as a platform technology, we fabricated and characterized DEX- and CIS-loaded IADD devices with two model geometries (i.e. helical shape and linear plank structures), and evaluated their drug release, cytocompatibility and pharmacological activity in vitro and in vivo. We implanted drug-loaded IADD devices in renal artery or carotid artery in vivo to target kidney or brain; and compared the drug distributions against oral dosing. Kidney and brain were chosen as representative organs to target because they have vastly different vascular permeabilities. In addition to DEX and CIS, many other drugs can be integrated into our IADD devices to broaden the therapeutic applications. The IADD devices can be particularly effective for treating localized cancers or disorders in a specific organ, including brain tumors, epilepsy and Parkinson’s disease, as well as chronic inflammatory conditions such as inflammatory bowel disease. By enabling organ-specific drug delivery, the IADD devices pave the way for more effective and safer therapeutic delivery strategies. The integrated safety and efficacy testing support these novel IADD devices as a promising platform technology for delivering focal on-target treatments to different organs and disease processes. In addition to validate safety and efficacy, this work provides a framework for further development of our IADD devices as a platform technology for intra-arterial drug delivery to many other targeted downstream organs.

2. Methods

2.1. Fabrication of Drug-loaded IADD Devices and Controls

Two structural variants of intra-arterial drug delivery (IADD) devices (helical or linear shaped structure) were fabricated using magnesium (Mg) wire cores and poly(glycerol sebacate) (PGS) coatings (Supplemental Figure 1). Full fabrication parameters including polymer synthesis conditions, curing durations, device dimensions, and sterilization are provided in Supplemental Methods. Briefly, PGS pre-polymer was synthesized from sebacic acid and glycerol following a modified polycondensation procedure [21]. The clean Mg wires of two model geometries (helical or linear shape) were embedded within the pre-PGS matrix and vacuum-cured at 120 °C to produce the IADD devices.

Two distinct drug-loading methods were employed: a two-step post-soaking method, in which pre-cured PGS-coated Mg devices were sequentially soaked in ethanol and drug solution (DEX in ethanol, CIS in DMF), and a one-step co-curing method, in which 10 % w/w of DEX was directly dispersed in the pre-PGS before curing. The helical devices were first developed using the two-step method to implant into renal arteries for focal drug delivery to kidney in rats (kidney cohort 1), but their extremely high amount of drug loading and release motivated the design of smaller, linear plank devices using both methods above. The linear devices fabricated by the one-step method were also implanted into renal arteries for focal drug delivery to kidney in rats (kidney cohort 2). The linear devices fabricated by the two-step method were used for physicochemical characterization, in vitro drug release, and in vivo drug delivery to the brain in rats.

2.2. Characterize Microstructure and Composition of Drug-loaded IADD Devices and Controls

The morphology of the devices was first examined using a 3D laser scanning microscope (Keyence VK-X150). Scanning electron microscopy (SEM) was then used to analyze the surface microstructure and the cross-sections of the drug-loaded IADD devices and non-drug-loaded controls. The devices were sputter-coated with a conductive layer of gold at 20 mA for 30 s (Sputter Model 108, Cressington Scientific Instruments Ltd., Watford, UK) before SEM. The cross-section morphology was examined using a Nova NanoSEM 450 (FEI Co., Hillsboro, OR, USA) equipped with an X-Max50 detector and AZtecEnergy software (Oxford Instruments, Abingdon, Oxfordshire, UK). Energy-dispersive X-ray spectroscopy (EDS) was used to analyze the elemental compositions of the drug-loaded IADD devices and controls. An accelerating voltage of 15 kV was used for SEM imaging and EDS analysis.

2.3. Thermogravimetric Analysis (TGA) of Drug-loaded IADD Devices and Controls

Thermogravimetric analysis (TGA) was conducted using a TG 209 F1 Libra instrument (Netzsch). First, the pre-cured PGS blocks with a dimension of 10 mm × 10 mm × 0.6 mm were cut out and loaded with DEX or CIS following the same two-step method as described above for the IADD device fabrication. Next, 10 mg of each sample was cut out from the block and loaded into alumina crucibles and heated from 200°C to 800°C at a heating rate of 10°C/min in an air flow rate of 20 mL/min for TGA. The TGA of DEX and CIS were performed in parallel for comparison.

2.4. Analyze Chemical Bonding of Drug-loaded IADD Devices and Controls

Fourier-transform infrared spectroscopy–attenuated total reflectance (FTIR-ATR) was utilized to analyze the chemical bonding of drug-loaded IADD devices, IADD control device without drugs, and drug-only controls. ThermoScientific Nicolet iS10 FTIR instrument was used to measure the transmittance of the samples and controls with the wavenumbers ranging from 4000 cm⁻¹ to 400 cm⁻¹. The drug-loaded PGS layer of 10 mg was scraped from drug-loaded IADD devices using a blade. Similarly, a 10 mg PGS layer without drug was scraped from the IADD control devices. DEX and CIS powders were also measured as drug-only controls. The FTIR-ATR spectra of the drug-loaded IADD devices were compared with the spectra of IADD controls and drug-only controls to determine chemical bonding and drug loading.

2.5. Measure Drug Release Profiles and Degradation of Drug-loaded IADD Devices and Controls In Vitro

The drug release profiles and degradation of the drug-loaded IADD devices were analyzed in vitro using the immersion method. Specifically, the linear drug-loaded IADD devices (n=5) fabricated by the two-step method and controls were placed in individual wells of a 48-well plate and immersed in physiologically relevant solutions. The DEX-loaded IADD devices were immersed in revised-simulated body fluid (rSBF) and the CIS-loaded IADD devices were immersed in artificial cerebrospinal fluid (aCSF). The rSBF and aCSF solutions were prepared according to the respective composition reported in literature [13, 33, 34]. The plates were placed inside a rotary shaker (Benchmark Incu-shaker Mini) at 120 rpm and incubated at 37 °C. The supernatants were collected from each well by complete media replacement (1 mL) every 24 h for 30 days.

To measure the DEX and CIS released into the media, liquid chromatography tandem mass spectrometry (LC-MS/MS) was used. DEX and CIS concentrations in daily release media were quantified using LC-MS/MS. For DEX, samples were extracted with ethyl acetate containing dexamethasone-d5 as an internal standard, dried, and reconstituted in methanol before analysis on a Waters I-class UPLC coupled to a Synapt G2-Si Q-TOF MS operated in ESI+ mode. CIS quantification followed a derivatization-based method (Shaik et al., 2017): samples were reacted with palladium acetate and diethyldithiocarbamate (DTCC), extracted with acetonitrile containing 8-Cyclopentyl-1,3-dipropylxanthine (DPCPX) as the internal standard, dried, and reconstituted prior to analysis on a Waters H-class UPLC-G2-XS Q-TOF MS in ESI+ mode. Both assays used standard curves for absolute quantification and lock-spray mass correction to ensure mass accuracy. Full extraction protocols, LC gradients, MS parameters, and MS/MS transitions are provided in the Supplementary Methods.

Mg²⁺ ion concentrations in the collected media were measured as a direct indicator of the IADD device degradation. Inductively coupled plasma optical emission spectrometry (ICP-OES; Optima 8000, PerkinElmer, Waltham, MA) was used to measure Mg²⁺ ion concentrations. First, 30 μL of media collected from each sample was diluted in deionized water (Millipore) to 3 mL. The diluted solutions were then fed into the ICP-OES using an autosampler. Mg²⁺ ion concentrations were calculated using the ICP-OES results multiplied by the dilution factor (100x).

2.6. Verify Cytocompatibility of Drug-loaded IADD Devices and Controls with HUVEC Cells In Vitro

Human umbilical vein endothelial cells (HUVEC) were used as the model cell to test the cytocompatibility of the drug-loaded IADD devices and controls in vitro. HUVECs were maintained in Endothelial Cell Growth Medium (Cell Applications Inc.) with 1% Penicillin-Streptomycin (P/S, Corning) solution. The cells were maintained in a standard incubator at 37°C with 5% CO₂.

The DEX-loaded IADD devices, CIS-loaded IADD devices, and non-drug-loaded control devices were prepared using the two-step method as described in Section 2.1 and disinfected prior to in vitro cell study. Cytocompatibility of the devices was studied using the exposure culture method [35]. Specifically, the cells were first seeded into a 12-well plate at a density of 8,000 cells/cm² and cultured for 24 h. Afterward, the cells were exposed to the drug-loaded IADD devices (n=3), ethanol-IADD control devices (n=3), and DMF-IADD control devices (n=3) for another 24 h. Each device of interest was placed in a Transwell insert (Corning) and introduced to each well. The control wells (n=3) were the cells only exposed to the inserts without any devices. After exposure culture, Sulforhodamine B (SRB) assay was used to assess cell viability. The cells in each well were fixed with 1 mL of ice-cold trichloroacetic acid (10% w/v) and incubated at 4 °C for 1 h. The culture wells were washed with water and air-dried. SRB solution (0.057% w/v, 2 mL) was added to each well and incubated for 30 min, followed by washing with 1% v/v acetic acid. The wells were dried, and 3 mL of 10 mM TRIS base (pH 10.5) was added to each well. Absorbance was measured at 510 nm using a microplate reader [36]. The percentage of cell viability was calculated as follows:

$$\mathrm{\%}\mathrm{o}\mathrm{f}\mathrm{c}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{v}\mathrm{i}\mathrm{a}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}=\left(\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}{\mathrm{O}\mathrm{D}}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}/\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}\left.{\mathrm{O}\mathrm{D}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}\right)\times 100\right.$$ Equation (1):

The mean $\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}{\mathrm{O}\mathrm{D}}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}$ is the mean absorbance values of the triplicates samples; and, the mean $\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{n}{\mathrm{O}\mathrm{D}}_{\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{o}\mathrm{l}}$ is the mean absorbance values obtained from triplicate control measurements.

2.7. Evaluate Chemotherapeutic Potential of CIS-loaded IADD Devices and Controls In Vitro

F98 glioma cells (CRL-2397, ATCC) were used to assess the anti-cancer activity of CIS released from the CIS-loaded IADD devices using SRB assay [36]. F98 rat glioma cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Sigma-Aldrich) with 10% Fetal Bovine Serum (FBS, Gibco) and 1% Penicillin-Streptomycin (P/S, Corning) solution. The cells were seeded into a 96-well plate at a density of 5,000 cells/cm² and cultured in 150 μL of media. After 24h, 50 μL of CIS-containing aCSF solution collected from the drug release study (as described in section 2.5) was added to each well. The immersion solution collected from the non-drug-loaded IADD devices were used as controls. After 24 h, the media from each well was collected, and 30 μL of media from each sample was diluted (1:100) with deionized water to 3 mL for ICP-OES analysis of Mg²⁺ and Ca²⁺ ion concentrations. The cells were stained using the same method described in section 2.7. The experiment was performed in triplicate, and results are reported as Mean ± SD. The percentage of cell viability was calculated the same as shown in Equation 1.

2.8. Assess Focal Delivery of DEX to the Targeted Kidney or Brain in Vivo using Rat Models

Male Sprague Dawley rats (180–200 g) and male Fischer rats (180–200 g) procured from Charles River Labs were used to assess the focal drug delivery to the kidney and brain, respectively. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of California, Riverside. Animals were housed in groups of three under controlled conditions with a 12 h light/dark cycle and free access to food and water.

2.8.1. Focal delivery of DEX to the kidney in rats

Sprague Dawley rats were divided into two experimental cohorts. Both cohorts were used to evaluate focal drug delivery of DEX from the IADD devices to kidney versus oral delivery. Cohort 1 consisted of the rats that were implanted with helical-shaped DEX-loaded IADD devices (fabricated by the two-step method) in the left renal artery (n=2) and the corresponding rats that received DEX (1 mg/mL) in drinking water as oral controls (n=3). Cohort 2 consisted of the rats that were implanted with linear-shaped DEX-loaded IADD devices (fabricated by the one-step method) in the left renal artery (n=3) and their oral control group (n=3).

All surgical instruments were sterilized using an autoclave at 121°C and 15 psi for 30 min and handled aseptically thereafter. Animals were weighed before the procedure, and pre-operative analgesics, buprenorphine extended release (0.65 mg/kg s.c.) and meloxicam (2 mg/kg s.c.), were administered 30 min prior to anesthesia. Rats were anesthetized with 4% isoflurane in 1 L/min O₂, followed by ketamine (80 mg/kg i.p.) and xylazine (10 mg/kg i.p.). They were then transferred to a controlled heating pad with isoflurane maintained at 1% in 1 L/min O₂, and veterinary ophthalmic lubricant was applied to their eyes. An i.p. injection of 5 mL/kg normal saline was given pre-operatively. Animals were positioned supine with limbs fixed using low-tack adhesive. After confirming adequate anesthesia by monitoring reflexes, abdominal hair was shaved, and the skin was cleaned with chlorhexidine followed by povidone-iodine solution. A midline laparotomy incision of 2–3 cm was made to access the peritoneal cavity, and the intestine and colon were manipulated to expose the left renal artery. Vessels were separated from connective tissue and fat using blunt dissecting curved forceps and wet cotton swabs. The aorta was clamped above and below the renal artery bifurcation with microvascular clips. Next, both ends of the left renal artery were temporarily occluded using silk suture knot. A 1 mm incision was made on the occluded renal artery, a sterile device was inserted, and the incision was closed with tissue adhesive. The knot was opened first from the lateral, followed by the medial end. Clamps were released to allow kidney reperfusion, ensuring that the arterial occlusion did not exceed 15–20 min. The area was checked for active bleeding and observed for an additional 10–15 min. The abdominal wall was closed in two layers with 4–0 absorbable sutures using a simple continuous pattern for the muscle and a simple interrupted pattern for the skin. Topical ointment was applied to the suture site, and the animal was returned to its cage on a heating pad until fully recovered. The total surgery time ranged from 30 to 45 min. The internal diameter of the renal artery in rats was approximately 1.1 mm with a cross-sectional area of 0.95 mm². For cohort 1, the diameter of the helical-shaped device was approximately 1 mm with a cross-sectional area of approximately 0.62 mm² and a lumen diameter of 0.45 mm. For cohort 2, the dimensions of the linear-shaped devices were approximately 0.4 mm × 0.25 mm with a cross-sectional area of 0.1 mm². Post-surgery, animals received meloxicam (2 mg/kg s.c.) every 24 h for 72 h. Blood collection for serum drug level estimation was performed via retro-orbital puncture on day 7 under anesthesia with isoflurane. The serum was separated by centrifugation (6000 rpm, 10 min, 4 °C). The animals were sacrificed using pentobarbital and phenytoin cocktail (Euthasol) (200 mg/kg) and the left and right kidneys were isolated, homogenized in ice-cold phosphate buffer (10% w/v), centrifuged (10000 rpm, 10 min, 4 °C), and the supernatants were collected for DEX quantification using LC-MS/MS as described previously. Liver, heart, lungs, and spleen were also collected from rats in the Kidney Cohort 2 with linear IADD devices. Implanted and control renal arteries were isolated and stored in 4% paraformaldehyde in phosphate-buffered saline (PBS) for further analyses.

2.8.2. Focal delivery of DEX to the brain in rats

Fischer rats were divided into two groups: Brain-IADD (n=5) and Brain-Control (n=3). The Brain-IADD group had linear-shaped DEX-loaded IADD devices (fabricated by the two-step method) surgically implanted in the right carotid artery for focal drug delivery to the brain. The Brain-Control group received DEX (5 mg/mL) in drinking water.

The rats were prepared and anesthetized using the similar procedures as described above for the kidney cohorts. A pre-operative subcutaneous injection of 10 mL/kg normal saline was given before placing the animal in a supine position with its limbs fixed to the surface using low-tack adhesive, ensuring the upper extremities remained in a normal position to prevent lung compression. Once the depth of anesthesia was confirmed by monitoring reflexes such as withdrawal from a toe pinch, the hair over the incision site was shaved with a sterile razor blade, and the skin was cleaned with chlorhexidine followed by povidone-iodine solution. The site was then infused with bupivacaine. For common carotid artery implantation, an incision was made on the neck to expose the common carotid artery by dissecting the submandibular glands and sternohyoid muscles. The artery was occluded using vascular clamps. A 1-mm incision was made on the artery, a sterile device was inserted, and the incision was closed using tissue adhesive. The carotid artery was occluded for approximately 15 mins. The device self-adhered to the wall of the artery owing to its property of tissue adhesion. After removing the clamps and ensuring there was no internal bleeding along with no visual displacement of the device within the artery, the skin incision was closed using simple interrupted sutures. A topical ointment was applied to the suture site, and the animal was returned to its cage on a heating pad until fully recovered. The total surgery time, from incision to suture, was approximately 30–45 min. The internal diameter of the carotid artery in rat was approximately 1.1 mm with a cross-sectional area of 0.95 mm². The dimensions of our linear devices were approximately 0.4 mm × 0.25 mm with a cross-sectional area of 0.1mm². Post-surgery, the animals received meloxicam (2 mg/kg s.c.) every 12–24 h for 72 h. To determine serum drug level, blood collection was performed via retro-orbital puncture on day 7 under isoflurane anesthesia. The serum was separated by centrifugation (6000 rpm, 10 min, 4 °C). Animals were sacrificed using pentobarbital and phenytoin cocktail (Euthasol) (200 mg/kg) and the brain and liver were isolated, homogenized in ice-cold phosphate buffer (10% w/v), centrifuged (10000 rpm, 10 min, 4 °C), and the supernatants collected for DEX quantification using LC-MS/MS as described previously. The liver was included in this experiment as an indicator for off-target delivery to another organ. Implanted and control carotid arteries were isolated and stored in 4% paraformaldehyde in PBS for further analyses.

To quantify the efficacy of focal drug delivery, we determined the normalized organ drug levels (NODL), using the equation below:

$$\mathrm{N}\mathrm{O}\mathrm{D}\mathrm{L}=\left([\text{Organ}\text{DEX}{]}_{\text{device}}/[\text{Serum}\text{DEX}{]}_{\text{device}}\right)/\left([\text{Organ}\text{DEX}{]}_{\text{oral}}/[\text{Serum}\text{DEX}{]}_{\text{oral}}\right)$$ Equation (2)

NODL was calculated based on average values pooled from each cohort. For the linear IADD devices in rat kidney cohort 2, $[\text{Serum}\text{DEX}{]}_{\text{device}}$ values were below the lower limits of detection (LLOD) of our LC-MS/MS quantification. For NODL calculations only, we assumed this LLOD for undetected values in order to perform the NODL calculation. Importantly, this assumption should underestimate the actual NODL for this cohort.

2.8.3. Histopathology of implanted arteries:

Staining for histopathology was performed as described briefly with slight modifications [37]. Briefly, arteries (renal and carotid) were fixed in 10% neutral buffered formalin, processed, and embedded in optimum cutting temperature (OCT) medium. Sections of 10-μm thick were stained in Mayer’s hematoxylin (3–5 min), rinsed, and “blued” in a mildly alkaline solution (10–15 s). Sections were counterstained with eosin Y (30 s) and cover-slipped with a resinous medium.

2.9. Statistical Analysis

All sets of data were expressed as Mean±SD and analyzed using GraphPad Prism v10.2.3 (GraphPad Prism software, San Diego, CA, USA). IC₅₀ values were calculated with non-linear regression (curve fit) method as the drug concentration at the 50% cell viability. Comparison of in vivo drug release data was performed using one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc comparisons. Normality was verified using the Shapiro-Wilk test, and the datasets met the assumptions for parametric analysis. The comparison of DEX levels between the brain and serum levels was performed using the paired t-test. The difference in the means for all the tests was considered statistically significant at p<0.05.

3. Results

The results of this work are to demonstrate proof-of-concept for our novel intra-arterial drug delivery (IADD) devices by evaluating the design and fabrication processes, in vitro release kinetics, cytocompatibility, and in vivo safety and efficacy of organ-specific delivery in rat models.

3.1. Microstructure and Composition of Drug-loaded IADD Devices and Controls

We designed and fabricated IADD devices and systematically characterized their microstructure and composition to ensure reproducibility and consistency across batches (Fig. 1). Figure 1A shows the viscosity of pre-PGS from 0°C to 120°C. The viscosity of pre-PGS used for IADD device fabrication was similar to previously reported viscosity profiles [24]. For this study, the PGS was synthesized with reduced time to have a lower viscosity, thus achieving a uniform PGS (or drug-loaded PGS) coating onto Mg substrates in the IADD devices. Figure 1B and 1C show the images of two model IADD devices with different geometries, i.e. a helical shape (Fig. 1B) and a linear plank structure (Fig. 1C). For the helical design, the gaps between the Mg coils were completely filled with drug-loaded PGS coating, resulting in a solid tubular structure. For the linear design, the Mg wire was centrally positioned and surrounded with drug-loaded PGS coating. Further examination of the device cross-section in SEM (Figure 1D) confirmed that the Mg microwire was fully embedded within the PGS matrix. EDS analyses of the DEX-loaded, CIS-Loaded, and no-drug control IADD devices in Figure 1E showed that the coating layer primarily consisted of carbon (C > 70 at.%) and oxygen (O > 20 at.%), consistent with the composition of poly(glycerol sebacate). All the other elements detected in the coating layer, such as magnesium (Mg), fluorine (F), platinum (Pt), and chlorine (C), had less than 1 at.%. Trace amounts of Mg (~ 0.1 – 0.4 at.%) were detected, reflecting minimal diffusion from the metallic wire core. The presence of fluorine (0.6 at.%) in DEX-loaded IADD device indicated the successful incorporation of DEX, while platinum (0.1 at.%) and chlorine (0.2 at.%) signals in CIS-loaded IADD devices confirmed CIS loading, with a Pt:Cl ratio of 1:2 matching the stoichiometric composition of the drug. Notably, Pt and Cl signals were not detected in the EDS of DEX-IADD devices and no-drug control IADD, while F was not found in the EDS of CIS-IADD devices and no-drug control IADD, confirming the expected drug loading in the as-fabricated devices.

Figure 1.

Characterization of IADD devices. (A) Viscosity of PGS pre-polymer. (B, C) Two different geometries of IADD devices (B - helical shape and C - linear structure). (D) SEM image with (E) elemental composition of drug-loaded IADD devices. In [D], the dashed white circle depicts the central Mg wire, the dashed yellow box depicts the surrounding PGS, and the solid green box depicts the region of the device used for elemental composition analysis via EDS.

We further analyzed the drug-loaded and control devices using TGA and FTIR-ATR (Fig. 2). Fig. 2A shows the weight loss of drug-loaded IADD devices, IADD controls, DEX, and CIS, when the temperature increase caused the decomposition of PGS and drugs. PGS began decomposing at 394 °C and fully decomposed at 504 °C, leaving no solid residue. The DEX-only sample began decomposing at 285 °C and decomposed almost completely by 495 °C, leaving less than 2% solid residue, likely carbon black from incomplete combustion. The CIS-only sample began decomposing at 315 °C and continued to 364 °C; and, the weight loss then reached a plateau after decomposition, with approximately 65% mass remaining, because Pt in CIS has a high melting temperature of 270 °C. The DEX-loaded and the control IADD devices showed similar heat decomposition since the DEX amount is small relative to the PGS matrix. In the CIS-loaded IADD devices, 3% residue remained, consistent with Pt as a solid remnant after the CIS decomposition. Fig. 2B shows the FTIR-ATR spectra of drug-loaded IADD devices in comparison with non-drug-loaded IADD controls and drug controls. The FTIR spectra of drug-loaded IADD devices in Figure 2B were highly similar to IADD controls, and revealed evidence of drug incorporation without chemical degradation. The PGS in the IADD device exhibits its characteristic ester carbonyl stretch (~1730–1740 cm⁻¹), along with broad O-H stretching and C-O bands in the 1000–1300 cm⁻¹ region, similar to the FTIR-ATR spectra of PGS reported in the literature [38]. Upon loading DEX or CIS, the composite spectra of DEX-IADD and CIS-IADD retained the core polymer peaks, with attenuated and slightly shifted drug-related features. The FTIR-ATR spectra of DEX-loaded IADD devices and control devices are similar, and no significant difference was observed, considering the low amount of drug relative to the PGS matrix and overlap of peaks between IADD controls and DEX. In the CIS-IADD spectrum, weak N-H stretching (~3200–3400 cm⁻¹) and amide bending (~1600–1650 cm⁻¹) bands remained discernible, although diminished relative to pure CIS. The absence of new peaks and the preservation of polymer backbone bands confirmed the drug encapsulation in PGS without covalent modification or decomposition.

Figure 2.

(A) TGA curves of DEX- and CIS-loaded IADD devices in comparison with non-drug-loaded IADD device and standard model drugs of DEX and CIS. (B) FTIR-ATR spectra of DEX- and CIS-loaded IADD devices in comparison with non-drug-loaded IADD device and standard model drugs of DEX and CIS.

3.2. Sustained Drug Release and Degradation of DEX-Loaded and CIS-Loaded IADD Devices In Vitro

Building on the successful fabrication of the IADD devices, we next quantified their drug release profiles in physiological solutions and device degradation in vitro over 30 days (Fig. 3). The amounts of absolute and cumulative DEX and CIS released from DEX-loaded and CIS-loaded IADD devices over 30 days, with daily medium replacement, are presented (Fig 3A and 3B). The DEX-loaded IADD devices released cumulative 373.11 ± 1.41 μg of DEX over the 30 days, with a burst release on day 1 (274.65 ± 91.90 μg), equivalent to 73% of total release. By day 7, DEX release stabilized between 0.39 – 4.12 μg/day with an average release of 1.78 ± 1.12 μg/day. CIS-loaded IADD devices did not demonstrate a burst release phase and released a cumulative 64.73 ± 0.06 μg of CIS over 30 days. Similar to DEX, CIS release stabilized by day 7 between 0.97 – 7.07 μg/day with minor fluctuations, releasing an average of 1.72 ± 1.2 μg/day. For comparison, the DEX-loaded and CIS-loaded IADD devices with the same linear structure were fabricated by the same two-step method for this in vitro study. Collectively, these results demonstrated that the IADD devices achieved reliable, sustained drug release for at least 30 days, with distinct release profiles likely reflecting the different polymer-drug interactions, drug solubilities, and immersion media.

Figure 3.

The drug release and magnesium ion release profiles over 30 days of immersion in physiologically relevant solutions in vitro. (A) Cumulative amount of DEX and CIS released from DEX-loaded and CIS-loaded IADD devices per day for 30 days. (B) Absolute amount of DEX and CIS released from DEX-loaded and CIS-loaded IADD devices per day for 30 days. (C, D) Magnesium ion concentrations in rSBF containing the DEX-loaded devices and in aCSF containing the CIS-loaded devices, as well as the respective non-drug-loaded IADD control devices. Dashed lines reflect the baseline magnesium ion concentrations in the immersion media of rSBF or aCSF, that is, 1.3 mM (31.6 mg/L) in (C) and 1.5 mM (36 mg/L) in (D).

To evaluate the device degradation over time, Mg²⁺ ion concentrations in the collected media were quantified (Fig 3C and 3D). Degradation of Mg wire substrate releases Mg²⁺ ions. The measured Mg²⁺ ion concentrations demonstrated stable ion release throughout the 30-day period, without evidence of burst release that would indicate rapid or uncontrolled degradation. For DEX-loaded and non-drug-loaded IADD devices incubated in rSBF, Mg²⁺ ion concentrations remained near the baseline of rSBF controls at 31.6 mg/L (1.3 mM) for the first 23 days. Slight increases were observed after day 24, with maximal concentrations remaining below 45 mg/L (Fig. 3C). Similarly, for CIS-loaded and non-drug-loaded IADD devices incubated in aCSF, Mg²⁺ ion concentrations remained near aCSF controls of 36 mg/L (1.5 mM) (Fig 3D). These results indicated that the drug-loaded PGS or PGS coating layer effectively maintaining the structural integrity of the devices during the 30-day in vitro drug release under physiologically relevant conditions.

3.3. Cytocompatibility of Drug-loaded IADD Devices with HUVECs

Given the intended intra-arterial application of the IADD devices, we next evaluated their cytocompatibility with vascular endothelial cells (HUVEC) (Fig. 4). We determined the inhibitory concentration (IC₅₀) profiles of the two drugs (DEX and CIS) and their respective solvents of ethanol and DMF (Figs. 4A and 4B). As expected, the drugs and solvents exhibited reduced cell viability at high concentrations. Importantly, these concentrations were substantially higher than those expected to be released from our IADD devices under physiological conditions. We assessed the cytocompatibility of the drug-loaded IADD and control devices with HUVECs using the exposure culture method (Fig. 4C). No significant differences in viability were observed across the groups of interest, including DEX-loaded IADD (100.5 ± 1.57%), CIS-loaded IADD (94.94 ± 0.37%), ethanol-IADD control (98.38 ± 2.11%), DMF-IADD control (101.0 ± 1.33%), and non-treated cell only control (100.0 ± 1.14%) (Fig. 4C). Collectively, these in vitro results suggested that the IADD devices are cytocompatible with endothelial cells for potential intra-arterial implantation in vivo, while also confirming effective removal of residual solvents from the IADD devices during fabrication.

Figure 4.

Cytocompatibility of drug-loaded IADD devices with HUVEC cells in vitro along with drugs and solvents controls. (A-B) IC₅₀ curves of (A) DEX and CIS drugs and (B) ethanol and DMF solvents. (C) Percentages of HUVEC viability when exposed to the DEX-loaded, CIS-loaded, ethanol-soaked, and DMF-soaked IADD devices in comparison with the cells only controls.

3.4. CIS-loaded IADD Devices Reduced Viability of Glioma Cells In Vitro

To evaluate the pharmacological activity of the released drug, we examined the viability of F98 glioma cells exposed to the media collected daily from the CIS-loaded IADD devices during the 30-day in vitro drug release. The inhibitory concentration (IC₅₀) for CIS was determined as 2.21 μg/mL (Fig. 5A), confirming effective cytotoxic activity against F98 cells. When the cells were treated with the daily release media from CIS-loaded IADD device, an average reduction in cell viability of 38.21 ± 10.56% was observed (Fig. 5B). These findings validated that the CIS released from the IADD devices remained pharmacologically active during prolonged release. Interestingly, F98 cells cultured with the daily release media from non-drug-loaded IADD controls exhibited higher cell densities and elevated viability compared to non-treated cell only controls (Fig. 5B). We speculate that the enhanced viability may be because the degrading PGS would release glycerol and the cells could use glycerol as an energy source [39, 40].

Figure 5.

In vitro drug activity of CIS released from the CIS-loaded IADD devices against F98 glioma cells. (A) IC₅₀ curve of CIS. (B) Percentage of glioma cell viability when exposed to the release media collected from CIS-loaded IADD device (blue) versus non-drug-loaded control device (black). ‘Medium collection day’ refers to the day that the release media were collected. (C) Mg²⁺ and Ca²⁺ ion concentrations in the culture media of the F98 cells doped with CIS-loaded IADD release media (solid lines). Dashed lines reflect ion concentrations from culture media controls: 0.8 mM (19.4 mg/L) of Mg²⁺ (bottom right y-axis) and 1.8 mM (72.1 mg/L) of Ca²⁺ (top right y-axis).

To further assess the metabolic impact of the released drug, Mg²⁺ and Ca²⁺ ion concentrations were measured in the culture media of the F98 glioma cells exposed to CIS-loaded IADD release media (Fig. 5C). These ions served as indirect indicators of cellular metabolic stability. Mg²⁺ and Ca²⁺ concentrations remained stable throughout the 30-day sampling period, with average levels comparable to those observed in the media control group. These findings suggest that the release media from CIS-loaded IADD devices did not significantly alter Mg²⁺ or Ca²⁺ homeostasis in F98 cells, indicating preserved cellular metabolic balance during exposure.

3.5. DEX-loaded IADD Devices Demonstrated Focal Drug Delivery in Targeted Kidney and Brain In Vivo

The in vivo studies of the DEX-loaded IADD devices demonstrated their safety and efficacy for focal drug delivery to two model organs in rats. Specifically, the DEX-loaded IADD devices with two different geometries were implanted in the left renal artery of rats, upstream of the left kidney; or in the right carotid artery of rats, upstream of the brain. The devices with helical design (fabricated by the two-step method) were implanted into rat Kidney Cohort 1 (Figure 6); the devices with linear design (fabricated by the one-step method) were implanted into rat Kidney Cohort 2 (Figure 7); and, the devices with linear design (fabricated by the two-step method) were implanted into rat brain cohort (Figure 8). All the animals survived during the whole 7-day period post implantation with the IADD devices. For comparison, oral control rats received DEX through their drinking water (1 mg/L or 5 mg/L). These concentrations are the respective therapeutic level dosages in rats, and sufficient to induce adrenal suppression in short-term studies [41–43]. The organ-specific drug accumulation and distribution in all three cohorts of animals showed significantly higher drug concentrations in the respective target organs of kidney or brain than those of the controls. In addition, histopathological analyses of explanted arteries showed no thrombus or necrosis, indicating in vivo safety of the IADD devices.

Figure 6.

DEX-loaded IADD devices enhanced focal drug delivery in targeted kidney in vivo in the rat Kidney Cohort 1. (A) Schematic representation of the in vivo drug administration via helical-shaped IADD device (left) or orally (right). (B) DEX concentrations in the kidneys and the serum, after IADD device implantation in the left renal artery for 7 days (left) or after 7 days of oral DEX administration (right).

Figure 7.

DEX-loaded IADD devices enhanced focal drug delivery in targeted kidney in vivo in the rat Kidney Cohort 2. (A) Schematic representation of the in vivo drug administration via the linear-shaped IADD device (left) or orally (right). (B) DEX concentrations in the left kidney (target), right kidney (contralateral), serum, and off-target organs (liver, spleen, heart, lungs), after IADD device implantation in the left renal artery for 7 days (left) or after 7 days of oral DEX administration (right). “UD” indicates values below the LC-MS/MS LLOD. *p < 0.05 compared to all other compartments in the IADD-treated animals.

Figure 8.

DEX-loaded IADD devices enhanced focal drug delivery in targeted brain in vivo in the rat Brain Cohort. (A) Schematic representation of the in vivo drug administration via the linear-shaped IADD device (left) or orally (right). (B) DEX concentrations in the brain, serum, and the liver, after IADD implantation in the right carotid artery for 7 days (left) or after 7 days of oral DEX administration (right). * p<0.05 when comparing the DEX levels in Brain vs Serum in the IADD-treated animals. # p<0.05 when comparing the DEX levels Brain vs Serum in the animals treated with oral administration.

3.5.1. DEX-loaded IADD devices increased DEX levels in targeted kidney in vivo

In the Kidney Cohort 1, the DEX-loaded IADD devices (Fig 6A) contained a larger volume of drug-loaded PGS per device because of the helical-shaped design and two-step methods. In rats implanted with the devices in the left renal artery, DEX concentrations in the left kidney (606.40 ± 81.74 ng/mL) were markedly higher than in the serum (14.40 ± 8.88 ng/mL) and the right kidney (29.75 ± 5.62 ng/mL) (Fig. 6B). This 42-fold increase between the left kidney and serum confirmed the ability of the IADD device platform to achieve focal drug delivery. In contrast, oral DEX administration (1 mg/mL in drinking water) resulted in comparable concentrations across all compartments, including the left kidney (11.88 ± 3.60 ng/mL), right kidney (7.88 ± 5.99 ng/mL), and serum (8.08 ± 5.54 ng/mL), as expected for systemic drug distribution of oral administration (Fig. 6B). Notably, the DEX concentration in the left kidney from the IADD-treated group was approximately 51-fold higher than that achieved by oral administration, thereby demonstrating the ability to achieve more therapeutic-like drug levels in the target organ. The calculated normalized organ drug level (NODL) was 29, reflecting a 29-fold enhancement in kidney-to-serum drug concentration using the IADD devices versus using oral dosing. Given the very high drug levels observed in the left kidney, we hypothesized that partial saturation of renal uptake might have contributed to the off-target exposure in the right kidney. To address this, in the Kidney Cohort 2, smaller, linear-shaped IADD devices fabricated by the one-step method were implanted in the left renal artery.

In the Kidney Cohort 2, the smaller linear-shaped IADD devices reduced lumen occlusion (Fig. 7A). In this cohort, DEX concentrations in the left kidney reached 154.4 ± 38.54 ng/mL, approximately 2.25-fold higher than levels in the left kidney following oral administration (68.39 ± 2.92 ng/mL) (Fig. 7B). Within the IADD-treated animals, serum DEX levels were below the lower limit of detection (LLOD = 0.97 ng/mL), indicating minimal systemic exposure. Additionally, DEX concentrations in the left kidney were significantly greater than those in the right kidney (36.95 ± 8.77 ng/mL) and liver (1.59 ± 0.61 ng/mL), while no detectable DEX was found in other off-target organs including the heart, lungs, and spleen (Fig. 7B). In contrast, the animals with oral DEX administration showed uniform DEX distributions across the left kidney (68.39 ± 2.92 ng/mL), right kidney (70.74 ± 10.07 ng/mL), and serum (47.02 ± 8.49 ng/mL), suggesting systemic drug exposure (Fig. 7B). For statistical analysis, undetectable concentrations were conservatively treated as LLOD. The calculated NODL for the Kidney Cohort 2 was 109, representing a 109-fold enhancement in focal drug targeting using the IADD device with linear design versus oral delivery. This improvement, substantially greater than the 29-fold enhancement achieved using the IADD device with helical design, indicates reduced drug loading and saturation in renal update for better local dosing efficiency.

3.5.2. DEX-loaded IADD devices increased DEX levels in targeted brain in vivo

DEX-loaded IADD devices with the linear geometry (fabricated by the two-step method) significantly increased drug levels in the brain when implanted in the common carotid artery of rats (Fig. 8). In the device-treated animals, the DEX concentrations in the brain (83.46 ± 29.04 ng/mL) were significantly higher than in the serum (7.34 ± 1.24 ng/mL) (Fig. 8B). This 11-fold increase in brain versus serum concentration is particularly noteworthy given the restrictive nature of the blood-brain barrier. This is illustrated in the rats treated with oral administration, in which DEX levels in the brain (4.14 ± 1.05 ng/mL) were significantly lower than in the serum (24.95 ± 5.68 ng/mL). Furthermore, DEX concentrations in the brain from the IADD-treated animals were approximately 20-fold higher than those treated with oral dosing. To further validate the efficacy of focal drug delivery, the DEX concentrations were quantified in the liver as a potential site of off-target accumulation (Fig. 8B). The DEX levels in the liver of IADD treated rats (25.74 ± 10.93 ng/mL) were comparable to that in rats with oral administration (38.93 ± 4.86 ng/mL), despite that the IADD device resulted in 20-fold increase in brain DEX concentration. Based on these data, a normalized organ drug level (NODL) of 68 was calculated, indicating a 68-fold improvement in brain-specific targeting using IADD-mediated focal drug delivery versus oral administration. These findings are qualitatively consistent with the results from the renal artery implantation to target kidney and confirmed that the IADD devices can achieve both focal and therapeutic-like drug profiles for targeted organs such as kidney and brain.

3.5.3. Histopathology of explanted arteries and characterization of IADD Devices Post Implantation

After 7 days of implantation, histopathological changes in the explanted arteries were compared with non-implanted control arteries to assess the safety of the IADD devices in vivo (Fig. 9). Hematoxylin and eosin (H&E) staining of arterial sections provided a qualitative assessment of local tissue response. Representative sections from the left renal artery and right carotid artery harvested at day 7 revealed patent lumens with preserved intima, media, and adventitia architecture at the implantation sites (Fig. 9B,D). Limited peri-arterial inflammatory infiltrates and focal adventitial remodeling were observed near the arteriotomy and adhesive interfaces, consistent with mild early post-surgical changes. No occlusive thrombus formation or transmural necrosis was identified in any of the examined fields, supporting the local vascular safety of the IADD devices and implantation procedure.

Figure 9.

Histopathological analyses of explanted arteries. Representative H&E images of renal (A, non-implanted; B, explanted) and carotid (C, non-implanted; D, explanted) arteries. A fragment of the IADD device is observable in [B], showing the magnesium core (dashed circle) and surrounding PGS. The shape of the PGS protrusion is an artifact of sectioning. For the carotid artery in [D], the device dislodged during staining.

Explanted IADD devices were examined and compared to non-implanted control devices to assess their structural integrity following in vivo exposure (Fig. 10). In explanted devices, SEM-EDS analyses revealed partial surface degradation and interfacial modifications at the Mg-PGS boundary (Fig. 10D–F). The polymer layer remained largely intact but exhibited localized thinning and minor surface cracking, suggesting gradual surface erosion rather than bulk degradation. Elemental mapping confirmed that Mg remained in the device core (Fig. 10I), while an increased oxygen signal near the interface indicated the formation of magnesium oxide or hydroxide during physiological exposure (Fig. 10E,F,H). The polymer-rich regions contained predominantly carbon (C, 76.7–76.9 at.%) and oxygen (O, 22.4–22.5 at.%), which were similar in explanted and non-implanted devices (Fig. 10G,H). Fluorine peaks (F, ~0.4 At%) were present in the polymer-rich regions of the explanted devices (Fig. 10J), indicating retention of DEX. Overall, the post-explantation profiles indicate limited surface oxidation and degradation of the Mg substrate and preservation of the polymer coatings, confirming the structural stability of the IADD devices after 7 days in vivo.

Figure 10.

Representative SEM-EDS images of control and explanted devices. (A,D) 50x gross SEM-EDS images of non-implanted control [A] vs explanted [D] IADD devices. The white squares indicate regions for higher magnification examination. (B,C,E,F) 350x magnification SEM-EDS overlay images of non-implanted control (B,C) and explanted (E,F) devices. The regions of interest (white squares) indicate regions used to analyze the elemental distributions. (G-J) Elemental distributions of carbon (G), oxygen (H), magnesium (I), and fluorine (J). The pseudo-colors in the SEM-EDS overlay [A-F] match the color conventions in [G-J]. The orange color in [E,F] reflects the colocalization of magnesium (red) and oxygen (yellow).

4. Discussion

4.1. Design Considerations of the IADD Devices for Focal Drug Delivery

The IADD devices are versatile for a wide range of therapeutic agents and can be tailored for focal drug delivery to different target organs, thus potentially addressing a variety of disease processes. When determining the suitability of a candidate therapeutic agent to be delivered by the IADD devices (e.g. small molecule compounds or large molecule biologics such as proteins, vaccines, peptides, or nucleotides), multiple factors must be considered, including but not limited to, suitability of candidate agents, physicochemical properties of device materials and drugs, the device-drug interactions, the processing conditions and the scalability of manufacturing processes. First, the amounts of drugs that can be loaded to the IADD devices are inherently limited by the sizes of arteries; and, thus, the selected therapeutic agents must have a therapeutic window at relatively low dosage levels, to achieve the intended clinical responses. Second, the selected candidate agents are expected to have a high systemic toxicity if distributed to non-target organs, which could justify the benefits over the risks in using the IADD devices; even though the IADD devices can be placed in arteries via minimally-invasive, fluoroscopy-guided endovascular procedures. Third, the candidate therapeutic agents should be compatible with the device materials in the device fabrication and drug-loading processes, while the device materials provide tunable mechanisms for controlling in vivo drug release. Last but not the least, from an absorption, distribution, metabolism, and excretion (ADME) perspective, the candidate agents chosen for IADD should exhibit absorption and distribution profiles that favor localized action and do not require hepatic metabolism for conversion to active form. Collectively considering all these factors, the IADD devices can be designed and optimized to deliver a broad spectrum of medications effectively to targeted organs. As an example, in this work, the IADD devices consisting of bioresorbable Mg microwire substrates and biodegradable PGS polymer coatings were fabricated with helical or linear geometries, both of which demonstrated the safety and efficacy in sustained in vivo focal drug delivery. We believe that other biodegradable biocompatible metals (e.g., Zn, Mn, Mo, Fe, and their alloys) and other biodegradable generally-recognized-as-safe polyesters (or hydrogels) can be used to fabricate the IADD devices with other geometries (such as half-to-full ring or tubular structures or other stent-like mesh structures) to control the device degradation, modify drug loading, and achieve tunable drug release profiles for different therapeutic agents, including but not limited to small molecule compounds and large molecule biologics.

In this work, the selected drugs (DEX and CIS) are compatible with the IADD device materials, i.e., magnesium (Mg) substrates (core backbone) and poly (glycerol sebacate) (PGS) coatings (drug-loading matrix), and processing parameters. Specifically, the drugs had at least moderate solubility in the solvents used for fabrication, such as ethanol or N, N-Dimethylformamide (DMF), to facilitate efficient loading without compromising the structural integrity of the device and pharmacological activity.

Other considerations include the chemical stability of the drug during the fabrication and sterilization processes, as well as potential interactions with PGS and Mg that could affect drug efficacy or release kinetics. Hydrophobic drugs may benefit from the hydrophobic nature of PGS, enabling sustained release, whereas hydrophilic agents might require modifications to the metallic core and drug-loading polymer matrix or the incorporation of encapsulating carriers to achieve desired release profiles. Furthermore, the biodegradation rate of polymers (PGS in this study) should align with the therapeutic window of the drug to ensure consistent and controlled delivery over the intended treatment period. The physicochemical properties of the drugs, such as molecular weight, lipophilicity, and charge, also play crucial roles in determining the mechanism and rate of drug release from the IADD devices.

4.2. Drug Loading, Release and Cellular Responses In Vitro

Elemental and thermal analyses verified successful drug incorporation and stability of both DEX and CIS within the PGS matrix. The observed cumulative release over 30 days shows that the polymer matrix can sustain extended elution while preserving the mechanical integrity of the Mg core for the duration of drug release from the IADD devices. Although DEX exhibited an initial burst release, the subsequent sustained release phase indicates steady uniform surface erosion of PGS for drug delivery. The absence of abrupt Mg²⁺ ion spikes throughout 30-day incubation confirms that Mg core did not undergo rapid degradation and the IADD device retained the structural integrity.

To evaluate cytotoxicity and pharmaceutic activity, two distinct in-vitro cell models were used. HUVECs represent the endothelial cells at the implantation site of IADD devices, while F98 glioma cells represent the cancer cells at the targeted organ site that need focal drug treatment. In the HUVEC culture, cell viability remained above 90% for all device and control groups, indicating the desired cytocompatibility of IADD device with inner arterial wall. In contrast, in the F98 assay, the release media from the CIS-loaded IADD device caused a 38% reduction in glioma cell viability, indicating inhibitory effect against cancer cells. The differences in the outcomes between HUVEC and F98 assays can be interpreted in the light of their respective in vitro culture methods that mimic the distinct in vivo conditions. In the HUVEC assay, the cells were exposed to the intact device initially and then its dynamic degradation products in the same culture well gradually, where drug availability is governed by the slow degradation of the PGS matrix and localized diffusion from the device surface. This creates a concentration gradient, with HUVECs slowly getting exposed to CIS that is released into the immediate environment over 24 h. Thus, the amount of drug exposure at a given moment is well below the cytotoxic level, similar to the in vivo condition in an artery where continuous blood flow moderates the dynamic exposure level. By contrast, F98 glioma cells were exposed directly to the accumulated 24-hour release media, representing the total amount of CIS liberated from the IADD device in 24 hours. This results in higher effective drug concentrations than that in the exposure culture of HUVECs, and, thus, lowered cancer cell viability, similar to the higher drug levels at the targeted organ site in vivo. Overall, these in vitro results suggest the feasibility of IADD devices for focal drug delivery to targeted organ.

4.3. Localized Focal Drug Delivery and in Vivo Performance in a Rat Model

Our in vivo results confirmed the successful focal drug delivery to both kidney and brain—organs with distinct perfusion and barrier properties. For the smaller linear IADD devices, kidney and brain targeting achieved 109-fold and 68-fold (NODL) improvements in focal drug delivery, respectively. The brain targeting data demonstrates an intra-arterial concentration advantage even across a restrictive barrier. These outcomes align with pharmacokinetic theory for regional intra-arterial delivery, that is, high extraction/clearance favoring organ deposition.

Collectively, these localized gains can be leveraged in two complementary ways: (1) maintain standard target organ drug levels while substantially reducing systemic drug levels, and (2) substantially increase target organ drug levels while maintaining standard systemic drug levels. The optimal balance of each affordance will depend on the specific clinical target. For example, treating brain cancers may benefit more from (2) given the restrictions of the blood-brain barrier, whereas treating kidney cancers may benefit more from (1).

Additional considerations can even further improve IADD-mediated focal drug delivery. As described by Dedrick’s kinetic framework, IADD would provide significant pharmacokinetic benefit when the drug exhibits high systemic clearance, high regional extraction, and/or low regional blood flow. According to this model, the concentration advantage of regional over systemic administration is most apparent in these conditions [44]. Our findings confirmed this, as DEX levels in the left kidney and the brain were substantially higher than those in the serum and the off-target organs, highlighting the ability of IADD device to deliver higher doses locally while minimizing systemic exposure.

The computational models by Cooke et al. further emphasize that rapid drug extraction and sustained retention are essential for maximizing regional deposition [45]. This principle is evident in our study, where we observed a marked difference between target and non-target kidney drug concentrations. However, detectable DEX levels in the contralateral kidney and the liver indicate that optimization may be necessary to further reduce systemic leakage and enhance local retention. Modifying drug formulation and delivery parameters could improve these outcomes, similar to the strategies suggested by Cooke et al. for enhancing regional deposition.

IADD-mediated focal drug delivery may be particularly advantageous for the disease processes that benefit from localized drug treatment, such as tumors, infection, inflammation, and pain. Previous studies have shown that intra-arterial delivery can restrict the drug distribution to the target region and thus reduce systemic toxicity [46, 47]. By enhancing the pharmacokinetic control of our IADD devices, we can leverage the advantages of IADD for sustained and precise therapeutic effects.

4.4. Limitations and Future Directions

Our study demonstrates several significant advantages of our IADD technology, including effective drug loading and controlled release, robust pharmacological efficacy, precise localization of drug delivery, and safe biocompatible degradation. These strengths underscore the potential of IADD device as a promising platform for localized treatment across various medical applications, leveraging the biocompatibility and biodegradability of metallic core and polymer matrix to enhance therapeutic outcomes while reducing the need for invasive removal procedures.

However, several limitations must be addressed to fully realize the clinical potential of IADD device. The current drug loading capacity is constrained by the solubility of therapeutic agents and processing steps inherent to PGS as the polymer matrix, limiting the maximum dosage that can be delivered effectively. Thus, ideal candidate drugs should have high potencies and low therapeutic dosage windows. Additionally, the observed initial burst release of DEX may be useful in cases where an initial bolus of a drug is needed to quickly reach the highest therapeutic concentration (C_max). Nevertheless, extended studies to characterize long-term release kinetics and mitigate early dosage spikes are still needed. Although the linear IADD devices increased focal drug delivery to the target kidney and brain by 109-fold and 68-fold respectively, the DEX levels in the other compartments were still slightly elevated, indicating possible systemic leakage and associated off-target side effects. Lastly, vascularization of certain tissues, such as the cornea or striatum, poses challenges for the specificity of intra-arterial drug delivery to these areas, because there are no major feeding arteries specific to these tissues. Intra-arterial implants carry established risks including thrombosis, embolism, and vessel occlusion. In our current study, histological examination of the implanted arteries revealed preservation of lumen and intact intima architecture, with no evidence of thrombus formation or luminal occlusion in the sections harvested at Day 7. Nonetheless, we recognize that the limited time-point and sample size restrict the ability to assess longer-term vascular risks, and that extended safety studies with I-ADD devices are warranted.

To overcome these limitations, future research should focus on enhancing drug loading efficiency and optimizing drug release profiles for different clinical applications. For example, tailoring polymer design by incorporating other polymers such as polycaprolactone (PCL) [48], polylactic acid (PLA) [49], poly(glycolic acid) (PGA), biodegradable hydrogels [50–53], silk, hybrid co-polymers, or their combinations could improve drug loading and release profiles. Moreover, incorporating 3D structural designs such as interconnected porous structures or meshes and multi-layer systems can also increase drug loading and sustain longer release. In addition, our current estimation methods for drug-loading (based on EDS atomic percentages and TGA residual mass) underestimate actual loaded mass, so development of more accurate loading quantification (e.g., direct extraction, imaging-based distribution mapping) is needed to better define therapeutic payloads. Additionally, combining PGS with angiogenic factors can improve tissue perfusion and enable effective drug delivery to poorly vascularized tissues. To reduce systemic leakage, optimizing drug formulations and delivery parameters are essential, potentially through advanced encapsulation techniques or refined device architectures.

Future studies will include dedicated histopathologic and functional evaluations in disease models, such as primary brain tumors. An important future direction is to assess organ-specific toxicities from our intra-arterial intervention. This will be particularly important in tumor models, which are less resilient to chemical and ischemic stressors [54]. By implementing these approaches, future iterations of IADD devices may achieve enhanced therapeutic efficacy, safety, and broad clinical applicability.

5. Conclusion

This study reports on the design and fabrication of novel drug-loaded IADD devices for focal drug delivery to targeted organs. The comprehensive characterization of IADD devices verified successful drug loading of two different drugs (DEX and CIS) into prototype devices with two distinct geometries (linear and helical) with two fabrication methods (one-step and two-step), used to focally target two organs (kidney or brain). The drug release profiles demonstrated that IADD devices could sustain drug release for at least 30 days. We confirmed cytocompatibility of drug-loaded IADD devices with endothelial cells, and pharmacological efficacy of the released CIS against glioma cells in vitro. Implantation of DEX-loaded IADD devices In vivo in a rat model demonstrated >60-fold improvements in focal targeting to the kidney and brain. Moreover, the IADD devices can be modified for different therapeutic agents toward different applications, to achieve optimal drug loading efficiency and drug release profiles. Overall, this study introduces IADD devices as a potential platform technology applicable to an array of clinical applications in which focal administration may improve treatment outcomes.