(T1330-09-49) In Vivo and Biopredictive In Vitro Dissolution Testing of Dexamethasone Intravitreal Implants in Rabbits

Purpose: OZURDEX is a long-acting intravitreal implant comprising dexamethasone and a biodegradable polymer, poly(lactic-co-glycolic acid) (PLGA). In previous research, we developed multiple OZURDEX-like formulations that were compositionally and structurally equivalent but varied in the PLGA blockiness and acid number [1]. These variables were correlated with changes to in-vitro drug release, but this varied with methodology and the biopredictiveness of these results was unknown. Biopredictive in-vitro release testing (IVRT) methods and in-vitro-in-vivo correlations (IVIVCs) can aid formulation development, improve quality control specifications, and eliminate the need for clinical bioequivalence studies in some cases [2]. The principal challenge of developing biopredictive IVRTs is mechanistically simulating the in-vivo drug release process in vitro. This requires identification and characterization of the relevant physiological factors in the compartment of interest. Further, one must appreciate the impact of these factors on the physicochemical processes of drug release. In this study, we measured drug release from two formulations of OZURDEX-like intravitreal implants with two IVRT methods. These methods differed only in the dissolution medium composition; one medium was unbuffered isotonic saline, and the other was 12 mM phosphate-buffered saline (PBS), pH 7.4. We then administered these formulations to the vitreous of rabbits and measured in-vivo drug release and ocular pharmacokinetics. The saline-based IVRT showed excellent correlation with in-vivo release, both in the total duration of drug release and shape of the release profile. This method may allow IVIVC development if the good biopredictiveness extends to similar formulations with different release rates.

Methods: Formulation design and manufacturing of OZURDEX-like implants were described in previous work [1]. Herein, Formulation Ev-O refers to implants comprising PLGA from Evonik Industries AG (Essen, Germany). Ev-O was optimized to match commercial OZURDEX in all critical quality attributes (CQAs) we have identified. Formulation Ak-A refers to implants comprising PLGA from Akina Inc. (West Lafayette, IN), which had a higher acid number than the Evonik PLGA but was identical otherwise.Implants were individually weighed and loaded into OZURDEX injectors acquired from an ophthalmologist’s clinic, then sterilized by gamma radiation. IVRT was performed in 30 mL of medium at 37 °C with agitation on a shaker table. In-vitro drug release was measured by reverse-phase high pressure liquid chromatography (HPLC). New Zealand White Rabbits were selected for this study due to the anatomical similarity of the rabbit and human eye. In-vivo drug release was determined by sacrificing six rabbits at six timepoints after injection (using a total of 36 rabbits), recovering the remaining implant by dissection, and assaying drug content by HPLC. In-vivo pharmacokinetics in the aqueous and vitreous compartments were determined by collecting samples of aqueous and vitreous fluid during the same dissection as above and measuring drug concentrations by liquid chromatography/triple quadrupole mass spectrometry (LC/MS-MS). HPLC and LC/MS-MS methods were evaluated for specificity, linearity, accuracy, and precision, and were highly satisfactory relative to typical specifications.

Results: Fig. 1 compares the in-vitro drug release of the Ev-O and Ak-B formulations in saline and PBS dissolution media. Release profiles were identical in saline but significantly different in PBS. Additionally, drug release lasted nearly 4 times longer in PBS than in saline. The latter result demonstrates the significance of buffering in the release medium. Drug release from these formulations was dominated by PLGA hydrolytic erosion, which is accelerated by acid catalysis in unbuffered saline. PLGA erosion was slower for both formulations in PBS, but the Ak-B formulation showed a significant burst release, possibly due to osmosis generated by the greater hygroscopicity of the Akina PLGA. Fig. 2 compares these two formulations in vivo. As in the saline IVRT, drug release was complete at 28 days and the two formulations were indistinguishable by all measures. This demonstrates that saline is a superior IVRT with respect to biopredictiveness for this rabbit model and the performance difference detected by the PBS method is not biologically relevant. Fig. 3 compares saline in-vitro drug release and rabbit in-vivo release for both formulations. An in-vitro-in-vivo correlation chart demonstrates the predictivity of the saline IVRT.

Conclusion: This study determined that: The saline IVRT was biopredictive for OZURDEX-like implants in the rabbit vitreous with respect to the drug release profile and duration, despite the fact that the rabbit vitreous fluid is buffered and the saline IVRT is not. Other physiological factors, such as molecular diffusivity in the vitreous, may be more important than vitreous buffering in this context. The difference in acid number between these two formulations was not significant in vivo in rabbits, as predicted by the saline IVRT but not by the PBS IVRT. This indicates the advantage of the saline IVRT for formulation development and quality control as compared to the PBS IVRT.

References: [1] Costello, M. A., et. al. Role of PLGA Variability in Controlled Drug Release from Dexamethasone Intravitreal Implants. Mol. Pharmaceutics, 2023, 20 (12), 6330-6344.
[2] Guidance for the Industry: Extended Release Solid Oral Dosage Forms: Development, Evaluation, and Application of In Vitro/In Vivo Correlations. U.S. Department of Health, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), 1997.
[3] Bhagat, R. Comparison of the Release Profile and Pharmacokinetics of Intact and Fragmented Dexamethasone Intravitreal Implants in Rabbit Eyes. J. Ocul. Pharmacol. Ther., 2014, 30 (10), 854-858.

Acknowledgements: This work was supported by the Broad Agency Announcement (BAA) Contract # 75F40120C00198 from the U.S. Food and Drug Administration (FDA). The content reflects the views of the authors and should not be construed to present the FDA's views or policies.
All animal procedures were approved by UT Austin’s Institutional Animal Care and Use Committee.

Fig. 1. (a) In vitro drug release of two OZURDEX-like formulations in saline dissolution medium. (b) In vitro drug release of these formulations in PBS.

Fig 2: (a) Drug release in the rabbit vitreous of two OZURDEX-like formulations and commercial OZURDEX (literature data [3]). (b) Dexamethasone concentrations of our two formulations in the rabbit aqueous. (c) Dexamethasone concentrations in the rabbit vitreous.

Fig. 3: (a) Comparison of saline in-vitro and rabbit in-vivo drug release for the Ev-O formulation. (b) The same comparison for the Ak-A formulation. (c) In-vitro-in-vivo correlation using both formulations.