Injectable Devices for Delivery of Liquid or Solid Protein Formulations

**Background:** Sustained delivery of protein therapeutics remains challenging, especially for localized administration in delicate anatomical sites like the eye. Miniaturized devices that can be injected via fine-gauge needles offer a minimally invasive approach. Prior work using thin-film PCL membranes achieved sustained release but faced limitations in device size, assembly complexity, and injectability. This study aimed to develop a tubular PCL device platform with improved payload capacity, uniform cylindrical geometry, and compatibility with 22-gauge needle injection.

**Methods:** Tubular PCL devices were fabricated by dip-casting a solution of PCL and poly(ethylene glycol) (PEG) in 2,2,2-trifluoroethanol onto stainless steel mandrels (353 ± 4 μm diameter). After drying, PEG was removed by sonication in water to create porous tubes. Wall thickness was controlled by mandrel draw speed (1–3 mm/s) and solution composition; tubes with wall thickness down to 30 μm were reliably produced. Devices were loaded with either a solid protein pellet or liquid formulation. Solid pellets were prepared by compressing lyophilized bovine serum albumin (BSA) in 300 μm inner diameter glass capillaries, achieving a pellet density of 1.05 ± 0.15 g/cm³. Devices were sealed either by self-fusing the tube ends or by inserting a necked PCL filament that expanded upon heating. In vitro release studies were conducted with BSA-loaded devices (payload 367 ± 133 μg BSA, total payload 400 ± 145 μg; porous length 7.9 ± 0.9 mm) incubated in PBS at 37°C with periodic media exchange. Protein concentration was quantified by micro-BCA assay. For in vivo safety, placebo devices (PCL rod core) were injected intravitreally into rabbit eyes using a 22-gauge thin-walled needle. Two devices were injected sequentially at 3-week intervals, and eyes were examined clinically and by histology at 6 weeks.

**Key Results:** Dip-casting produced uniform tubes with wall thickness as low as 30 μm, reducing the wall cross-sectional area to 27% of the device (compared to 65% in prior electrospun devices). Solid protein pellets achieved 76.5% of the theoretical maximum density (~1.37 g/cm³). In vitro release of BSA from solid-loaded devices was sustained for over 400 days (60 weeks). One device was excluded due to an abrupt release increase at day 94 attributed to handling damage. In the rabbit model, repeated intravitreal injection of two placebo devices was well tolerated. No significant intraocular inflammation was observed on clinical exams. Local irritation (conjunctival injection and chemosis) occurred after three of four injections but resolved by week 1. One injection caused a focal cataract due to needle misdirection. Histopathology showed peri-limbal inflammation consistent with mechanical trauma but no abnormalities in corneal endothelium, retina, or choroid.

**Clinical Implications:** This platform offers a promising approach for sustained delivery of protein therapeutics, particularly for ocular diseases requiring frequent intravitreal injections (e.g., anti-VEGF therapy). The ability to load solid or liquid formulations and the use of biocompatible, slowly degrading PCL may improve patient compliance and reduce treatment burden. However, the limited payload capacity (estimated ~520 μg for an 8 mm device) may restrict use to highly potent biologics. The authors note that sustained-release kinetics may allow lower total doses compared to bolus injections, but further studies are needed to define optimal release rates. The safety of repeated injections and multiple concurrent devices supports potential for long-term therapy. Future work should focus on formulation optimization for specific protein stability and advanced manufacturing for clinical translation.