(W1230-09-61) A Mutli-Mechanistic Three Polymer In Situ Gelling Ocular Drug Delivery Vehicle: Optimization under a DOE Strategy

Purpose: Rapid clearance from the eye through mechanical and physiologic processes continues to be a main determinant of poor ocular pharmacokinetics. In situ gelling eye drops have been explored for their promise to extend residence time of topical therapies on the eye.^1-4 The ocular surface, rich in mono- and divalent cations along with mucins both, membrane associated and secreted, allows for activation of polymers to form a gel network. Cross-linking ion activation with Ca²⁺ and K⁺ ions and binding to gel forming MUC5AC present ideal delivery targets for the polymer components.⁵ The previously demonstrated synergistic combinations of gellan gum (GG) plus carrageenan (CG) and gellan gum (GG) plus hyaluronan (HA) served as blueprints for the development of a three polymer in situ gelling ocular drug delivery vehicle (ODDV). The purpose of this work was to combine the ion-activation from GG and CG with the mucoadhesive capabilities of HA and optimize a formulation for a strong and highly mucoadhesive gel using a design of experiments approach. A space filling mixture design (SFMD) was chosen because it fills in the gaps typically left by traditional mixture designs.

Methods: Sample Preparation: % (w/v) solutions of GG, CG, and HA were made by adding polymer in 12 mL deionized water and heating between 70 – 90 ˚C for 10 minutes with magnetic stirring. Polymer solutions were then left to cool and set at room temperature overnight. Simulated Tear Fluid (STF) was prepared with 6.78 g/L sodium chloride, 2.18 g/L sodium bicarbonate, 0.0084 g/L calcium chloride, and 1.38 g/L potassium chloride. Sodium hydroxide was used to adjust STF to pH 7.4.
Design of Experiments SFMD: A prescreening study determined a range of concentrations to set values for the SFMD based on previous concentration vs gel strength experiments. The SFMD was built using JMP Pro 18 (JMP, Cary, NC) in which three factors, % GG (0.5 – 0.8), %CG (0.1 – 0.5), %HA (0.1 – 0.2) were set. 18 distinct formulations were generated and the rheology was run as described below. After values were measured the data was entered into JMP Pro 18 and analyzed using a generalized regression model for the most optimum formulation.
Rheology: 1 mL polymer was combined with 1 mL deionized water, water with mucin (muc(-)), STF, or STF with mucin (muc(+)) and rotated on a Rotator Genie (Scientific Industries, Bohemia, NY) at low speed for 5 revolutions. 1.1 mL of sample were added to a Kinexus rotational rheometer (Malvern Panalytical, Westborough, MA) and an amplitude oscillation sweep was performed over a shear strain range [0.100 – 100.00 Pa] at 34 ˚C.

Results: Never-before-seen synergistic interactions previously demonstrated by our lab identified three polymer candidates for formulation: GG, CG, and HA. Prescreening rheological studies conducted determined experimental concentration ranges of GG at (0.5 – 1), CG at (0.1 – 0.6), and HA at (0.1 – 0.6) % (w/v). The SFMD model generated 18 distinct formulations that expressed nonlinear blending or concentration-independent increases in gel strength and viscosity. A generalized regression was used to analyze the data and using the maximize desirability function, it was determined the optimum concentrations for the ternary polymer combination to be 0.560% GG, 0.340% CG, and 0.100% HA. Desirability was optimized with first priority given to gel strength, followed by viscosity.

Conclusion: Addition of a third polymer to a previously synergistic two-polymer combination showed stronger synergy when optimized using a design of experiments optimization. The mixture design generated 18 formulations in a targeted window that, when tested, produced concentration-independent effects in both gel strength and viscosity. Applying a self-validating ensemble model to the data followed by processing using the prediction profiler in JMP18 Pro an optimized formulation composed of 0.560% GG, 0.340% CG, and 0.100% HA was determined. Addition of the mucoadhesive element, HA, gives the potential to target gelling mucins in tear film and in the conjunctiva on the ocular surface, allowing for additional support in retention of the vehicle on the eye. Combination of ion-actived gelation and mucoadhesion in a three polymer formulation shows promise as an ODDV for prolonged delivery based on the synergistic interactions observed.

References: 1. Dubashynskaya et al. (2020), Pharmaceutics 12, 22
2. Wu et al. (2019), Asian J Pharm Sci. 14, 1-15
3. Li P et al. (2018), Drug Dev Ind Pharm. 44, 829-836
4. Cassano et al. (2021) Gels, 7, 130
5. Cher I. (2012), Clin. Exp. Ophthamol, 6, 634 - 43

Acknowledgements: This research was supported by Countermeasures Against Chemical Threats, NIH Grant AR055073.

Figure: Binary combinations tested for concentration dependent synergy. A. Amplitude oscillation was performed over a range of concentrations of gellan gum (GG) (0.5% - 1.0% (w/v)) and carrageenan (CG) (0.1% - 0.6% (w/v)) with and without STF as singular polymer samples and binary polymer combinations (n = 3). (GG) + STF served as the positive control to indicate the basal level of gel strength. Data shown as mean ± SD B. Amplitude oscillation was performed over a range of concentrations of gellan gum (GG) (0.5% - 1.0% (w/v)) and hyaluronan (HA) (0.1% - 0.6% (w/v)) with and without STF as singular polymer samples and binary polymer combinations (n = 3). (GG) + STF served as the positive control to indicate the basal level of gel strength. Data shown as mean ± SD


Figure: Space filling mixture design for three component mixture formulation. A. SFMD generates 18 distinct formulations for testing. B. Ternary plot defines the experimental space. Colored dots indicate distinct formulations and fill the concentration gaps typical of classical boundary point mixture designs. C. Amplitude oscillation was performed on 18 formulations composed of three polymer mixtures containing gellan gum %(GG) (0.5 - 0.8 (w/v)), carrageenan %(CG) (0.1 - 0.5 (w/v)), and hyaluronan %(HA) (0.1 - 0.2 (w/v)) with and without STF as singular component samples and Mixture (M1-M18) samples (n = 3). Components plotted as the sum of G’ of individual polymer samples. Data shown as mean ± SD. D. Amplitude oscillation was performed on formulations from A as described in C. Components plotted as the sum of dynamic viscosity of individual polymer samples. Data shown as mean ± SD.


Figure: DOE method optimizes three component mixture formulation. Optimized formulation using analyzed data via a generalized regression model and prediction profiler in JMP Pro 18.