(T1430-01-05) Formulation Development of Novel PLGA-Based Long-Acting Monolithic Implant

Purpose: PLGA-based long-acting implants (LAIs) have enjoyed enormous clinical and economic success over the last 30 years. There are currently dozens of approved products on the market that cumulatively yield billions of dollars in annual revenue.[1] These products include treatments for bacterial infections, cancer, and opioid dependence, demonstrating the range of applications of this platform. The growing popularity of parenteral drug delivery ensures that this platform will continue to attract the interest of drug developers. Despite this history of success, PLGA-based implants remain difficult to formulate due to the complexity of drug release, among other issues. Drug release from PLGA-based LAIs can be controlled by several simultaneous physicochemical phenomena.[2] This includes water uptake and swelling, dissolution and diffusion of the drug, degradation of PLGA, and microclimate pH changes. The relative importance of each phenomenon depends on the site of treatment, drug substance, drug loading, and other factors. The complexity of this situation makes it difficult to predict the pharmacokinetics, pharmacodynamics, and treatment outcomes of PLGA-based LAIs. Improved understanding of the physicochemical mechanisms of drug release will accelerate the development of effective products. In this study, we developed a novel antifungal intravitreal implant according to design requirements we developed through collaboration with ophthalmologists and pharmacologists. This involved building a pharmacokinetic (PK) model to estimate the drug release kinetics that would be required to replace multiple bolus injections (the standard of care) with a single long-acting implant. We then developed a formulation that demonstrated the required release kinetics in-vitro. Importantly, in-vitro drug release was substantially linear and had the desired duration due to addition of a rate-controlling hygroscopic pore-former. This formulation strategy may be broadly applicable for modifying the drug release kinetics from PLGA-based implants.

Methods: Pharmacokinetic (PK) modeling was performed with MATLAB Simbiology Model Builder (MathWorks, USA). PK parameters for intravitreal delivery of voriconazole were extracted from literature. This included physiological, disease-specific (fungal  endophthalmitis), and drug-specific parameters.[3] Resomer 502H (PLGA 50:50, acid-terminated, inherent viscosity: ~0.2 dL/g) was the selected PLGA grade due to its high hygroscopicity and fast degradation rates, which allowed complete drug release within the targeted period of 21 days. Voriconazole (basic pKa: 1.9, LogP: 1.8, solubility at pH 7: 500 ug/mL) was the selected antifungal drug substance due to its favorable potency, tolerability, and stability. PEG 1000 and PVP K12 were selected as hygroscopic pore-forming additives due to their favorable chemistry for this purpose and their demonstrated tolerability in the eye. Intravitreal implants were manufactured by extrusion with a Haake MiniLab twin-screw extruder (Thermo Electron, Germany) equipped with a circular die and down-stream puller assembly. Implant diameter and length were small enough to allow injection into the eye with a tolerable needle gauge. In-vitro release testing was performed in 30 mL phosphate-buffered saline (PBS) at 37 °C and 100 rpm in a shaker.

Results: Fig. 1a and 1b summarize the pharmacokinetic model for treating fungal endophthalmitis with an intravitreal long-acting implant. Fig 1a shows the target in-vivo drug release curve needed to generate intravitreal drug concentrations exceeding the minimum inhibitory concentration (MIC) of the two fungus species most commonly implicated in fungal endophthalmitis. Fig 1b compares the intravitreal drug concentrations generated by this model implant with those generated by repeated bolus injections. This model suggests that a single LAI may generate therapeutically effective drug levels in the eye for up to 20 days, whereas bolus injections generate effective drug levels for only about 1 day per injection. Fig. 1c shows the measured in-vitro drug release of three implant formulations. The 40/42/18 VCZ/PLGA/PVP formulation matched the target drug release rate and duration. Fig. 2 explains the dominant mechanism of drug release from the PVP formulation. As shown in the SEM cross-sectional images (Fig. 2b), small pores initially appeared on the outer layers of the implant, then progressively appeared nearer the center. We attribute this to progressive dissolution of PVP. This created an interconnected network of channels within the implant that allowed the drug to dissolve and escape. Fig. 3 demonstrates our hypothesis for the different release behavior of the PEG and PVP formulations. While both additives are water-soluble pore-formers, PVP may form a viscous gel at high solid fractions, whereas PEG dissolves more readily. Initial water ingress into the PVP implant may effectively create a concentrated polymer solution, resulting in a gel. This may slow both water and drug transfer within the implant, extending drug release.

Conclusion: Drug release kinetics from PLGA-based implants may be modulated by the addition of excipients. In addition to the soluble polymeric excipients discussed here, other substances such as hygroscopic insoluble salts and pH modifiers could be investigated during formulation of new products.

References: [1] Nkanga, C., et. al. 2020. Clinically established biodegradable long acting injectables: An industry perspective. Adv. Drug Deliv. Rev. 167, 19.
[2] Fredenberg, M., et. al. 2011. The mechanisms of drug release in poly(lactic-co-glycolic acid)-based drug delivery systems–a review, Int. J. Pharm. 415, 34.
[3] Shen, Y., et. al. 2007. Clearance of intravitreal voriconazole. Inv. Ophth. Vis. Sci. 48, 2238.


Fig. 1. (a) Target in-vivo release profile for long-acting implant. (b) Predicted pharmacokinetics of the target long-acting implant compared to bolus injections and an estimated MIC 90. (c) In-vitro drug release of 3 implant formulations.


Fig 2: (a) Drug release, water content, and dry mass loss of PVP formulation. (b) SEM cross-section images of PVP formulation.


Fig. 3: Photos comparing water uptake and swelling behavior of PEG and PVP.