(M1430-01-02) Impact of Changes in Ophthalmic Emulsion Globule Size Distribution and Viscosity on Tear Film Thickness and Menisci Characteristics

Purpose: The purpose of this study was to obtain tear film thickness and menisci measurements on rabbit ocular surface after instillation of cyclosporine ophthalmic emulsion formulations with differences in viscosity and globule size distribution (GSD). This information will support the development of a mechanistic absorption model to better understand and predict formulation effects on the bioavailability of cyclosporine to the cornea and the conjunctiva in humans.

Methods: Five compositionally identical cyclosporine emulsion formulations (F1, F3, F4A, F4B and F5 with varying GSD and viscosity) and three placebos (P1- glycerin and carbomer; P2- glycerin, carbomer and Tween 80 and P3- glycerin and excess carbomer) were prepared and characterized in-house (Table 1). In addition, three lots of commercially available cyclosporine emulsion drug product (Restasis), one lot of over-the-counter artificial tear product (Refresh Liquigel) and sterile water were tested. In total, sixteen female Dutch Belted (DB) rabbits were enrolled. After instillation, each formulation was tested for tear variables like central tear film thickness (TFT), upper tear film meniscus cross-sectional area (UMA) and lower tear film meniscus cross-sectional area (LMA). Tear variable measurements were taken at baseline (just before test article instillation), immediately after instillation, and at 5, 10, 20, and 40 minutes’ post instillation for each formulation. A total of 416 experiments were conducted and measurements for all tear variables were done using the Heidelberg Spectralis^® Eye Explorer (HEYEX) software on the scans captured with anterior segment optical coherence tomography (AS-OCT).

Results: Instillation of all 13 formulations resulted in immediate increase in the levels of tear variables like TFT, UMA and LMA. TFT and LMA levels were followed by a sharp decrease within 10 minutes and then a more gradual decrease afterwards. UMA levels were followed by a sharp decrease within 5 minutes post instillation. For most of the formulations, tear variable levels returned to the baseline within 20 minutes post instillation. Formulation F1, which had the highest viscosity and similar GSD as compared with F4A and F4B, produced higher TFT levels post 5 minutes’ instillation than the F4A and F4B. Similarly, F5 formulation, which had the highest GSD and similar viscosity to F3 and F4A, produced the highest TFT levels among all the cyclosporine emulsion formulations post 5 minutes instillation. Formulations F4A and F4B, similar in GSD but different with respect to viscosity, produced similar but lower levels of TFT compared to F1, F3 and F5 at post instillation time points. Formulation F3, which had the lowest GSD and medium viscosity, produced TFT levels similar to the F1 and F5 formulations post 5 minutes’ instillation. Post 20 minutes installation, formulations F3 with lower GSD/medium viscosity and F5 with higher GSD/medium viscosity had higher TFT levels compared to other cyclosporine emulsion formulations. Placebo formulations with low to high viscosity established a correlation between TFT levels and viscosity. Refresh Liquigel^® and P3 formulations appeared to produce the greatest increase in TFT among all 13 formulations immediately post instillation; however, by 5 minutes post instillation, TFT levels decreased to levels that were comparable to the other formulations. All 3 lots of Restasis^® exhibited comparable TFT levels at all time points. Water exhibited the lowest increase in TFT levels immediately post instillation, returning to the baseline within 5 minutes. Restasis^® (all 3 lots), F1, F3, F4A, F4B formulations and Refresh Liquigel^® exhibited comparable UMA and LMA levels at all time points. Water instillation produced only a minimal increase in UMA and LMA immediately after instillation and returned to baseline within 5 minutes post instillation. It was observed that the F5 formulation, which had the highest value for GSD, gave rise to the largest change in the tear variables. Formulation P3, which had the highest viscosity ( >2-fold compared to other formulations), appeared to have the greatest changes to TFT and LMA. Overall, TFT data revealed that earlier time points within 10 minutes post instillation correlate well with formulations that vary in GSD and viscosity. Similar tear variables of different lots of Restasis^® further supports the hypothesis that observed differences in tear variables can be attributed to differences in the GSD and viscosity of the formulations.

Conclusion: Data showed that GSD and viscosity are critical quality attributes (CQAs) of the formulation that impact the tear variables, albeit in different ways. Among all tear variables, change in TFT appears to be the most sensitive measure to capture differences in GSD and viscosity. These data will be utilized for the validation of a previously developed in silico model along with rabbit pharmacokinetics data, to examine the influence of certain physicochemical properties on the predicted local bioavailability of cyclosporine ophthalmic emulsion and thereby help inform CQA limits that are expected to be more clinically relevant.

TFT measurements (µm) post instillation at different time points (immediately post-dose; 5, 10, 20, and 40 minutes post-dose). TFT was calculated = post-dose cornea thickness (µm) – baseline (pre-dose) cornea thickness (µm). Data expressed as Mean ± SD.



UMA measurements (mm2) at pre (baseline) and post instillation (immediately post-dose; 5, 10, 20, and 40 minutes). Data expressed as Mean ± SD.



LMA measurements (mm2) at pre (baseline) and post instillation (immediately post-dose; 5, 10, 20, and 40 minutes). Data expressed as Mean ± SD.