DEVELOPMENT OF NANOEMULSION TO IMPROVE THE OCULAR BIOAVAILABILITY AND PATIENT COMPLIANCE IN POSTOPERATIVE TREATMENT USING INDOMETHACIN

Objective: To develop a new cationic nanoemulsion (NE) for ophthalmic delivery of indomethacin (IND) to improve the permeability and retention time of formulations, thereby improving the drug's ocular bioavailability. Methods: Based on the solubility profile of indomethacin in various solvents, captex 8000 was selected as oil phase, span 20 as a surfactant and tween 20 as co-surfactant to construct pseudo ternary phase diagrams and nanoemulsion region was recognized. Sonication was used as the method of NE preparation. Optimization was done using 3 2 factorial designs by considering the oil and the ratio of surfactant to co-surfactant (Smix) quantities as independent variables and evaluated for different physicochemical properties. Ex vivo transcorneal permeability was studied using bovine cornea, the In vivo drug pharmacokinetics of optimized NE and marketed formulation were assessed in rabbit aqueous humor and also in plasma. Results: The mean globule size, zeta potential, viscosity, refractive index, pH, surface tension and the osmolarity values for the prepared indomethacin nanoemulsions (IND-NEs) were found between 129.8±1.1 to 191.4±1.6 nm, +13.20±4.6 to+23.45±4.82, 15.3±0.1 to 32.7±0.0 mPas, 1.346±0.007 to 1.386±0.005, 5.5±0.4 to 6.9±0.9, 32.0±2.6 to 52.3±3.4 mN/m and 303-395 mOsm/l respectively and all these values found to be falling under the recommended values for ophthalmic use. From the In vitro release studies, it was found that the IND-NEs exhibited sustained drug release with 67.91±2.01 to 95.90±1.93 % drug release at 24h when compared to the drug solution which showed 99.81±5.21 % drug release within 2h. The Ex vivo drug permeation through the corneal membrane at 4h from the optimized NE and drug solution was found to be 524±1.5 µg/cm 2 and 175±2.6 µg/cm 2 respectively. Further, the optimized NE was found to be nonirritant with the lowest ocular irritation potential (Iirr) of 1 towards the rabbit's eyes. The area under the drug concentration vs. time curve for 24h (AUC (0–24h) ) for optimized NE and the marketed formulation was found to be 1514.99 ng/ml/h and 974.14 ng/ml/h in aqueous humour; 2266.83 ng/ml/h and 778.15 ng/ml/h in plasma respectively. Conclusion: Due to its improved corneal absorption and prolonged drug release along with less systemic absorption, the optimized NE offers an effective postoperative treatment with increased ocular bioavailability and improved patient compliance with a decrease in the number of installations per day and a decrease or disappearance of systemic side effects of IND.


INTRODUCTION
Postoperative pain and irritation symptoms are relatively common during the first hours after surgery [1], which is associated with a breakdown of the blood-aqueous barrier as a result of surgical trauma induced prostaglandin production. In general, corticosteroids are prescribed to control postoperative inflammation; however, they found increasing intraocular pressure, delaying in corneal epithelial and stromal wound healing, and more susceptible to microbial infections. Hence, the use of non-steroidal anti-inflammatory drugs becomes an alternative treatment [2].
The topical route is preferred to administer ophthalmic dosage forms due to benefits like the ease of application, targetability, reduced side effects and also cost-effectiveness [3]. Most of the conventional ophthalmic dosage forms are in an aqueous solution of water-soluble drugs and either as an ointment or suspension of water-insoluble drugs. Frequent administration is required for aqueous solutions due to poor ocular bioavailability and drug loss through the nasolacrimal drainage which may lead to extremely unwanted pharmacokinetics and systemic toxicity suggesting the necessity of new alternatives to ocular drug delivery. Coming to the water-insoluble drugs, ointments suffer from poor patient acceptance because of blurred vision and matted eyelids, and suspensions lead to particle irritation, poor bioavailability, and changes in polymorphism and particle size upon storage [4]. Besides, the complex structure of the human eye can limit the bioavailability of ocular drugs. Due to poor ocular bioavailability, higher concentrations of the conventional form of analgesic used during surgery may result in ocular and systemic side effects.
NEs are advantageous for topical ocular drug delivery due to their ability to solubilize drugs in a large quantity, enhance absorption, achieve sustained drug release and target the affected area of the eye [5]. NEs are kinetically stable colloidal dispersions with droplet sizes on the order of 100 nm. Because of their small size, they are optically transparent in appearance, possess high surface area per unit volume, have excellent stability and variable viscosity [6]. Apart from their excellent physicochemical properties and stability, the NE systems for ocular delivery are easy to fabricate and characterize. NEs can deliver both hydrophilic and lipophilic drug moieties both to the anterior and posterior segments of the eye, in a safe and reproducible manner with improved patient compliance [7]. Because of the increased bioavailability and reduced drug toxicity, the NEs may serve as the potential delivery systems for ocular administration [8].
The objective of the present study was to develop a new cationic NE for ophthalmic delivery of indomethacin IND with improved permeability and retention time thereby improving drug's ocular bioavailability.

Quantitative analysis of IND
Stock solution and working sample preparation 100 mg of IND was dissolved in 50 ml of HPLC grade methanol and the final volume was adjusted up to 100 ml to prepare a stock solution of 1000μg/ml concentration. For the calibration curve, the stock solution was again diluted with the help of diluents to get final working concentrations of 10-500ng/ml.

Quantitative analysis conditions
The quantitative analysis of IND was performed using ultrafast liquid chromatography (UFLC, Shimadzu L220) containing LC-20AD isocratic pump, SPD-20A UV/Vis detector, and rheodyne injector was used. The UFLC process was carried out on a C8 column (100 mm x 4.6 mm and the particle size of 5 µm). The mobile phase comprised of 0.5% v/v O-Phosphoric acid: methanol: acetonitrile in the ratio of 40:20:40, respectively. The flow rate was set at 1.0 ml/min and about 20 µl of sample solutions were injected into the UFLC in triplicate using Hamilton microsyringe and measured at a wavelength of 270 nm [9].
The chromatographic data analysis was done by using the LC solution software. A standard plot of concentration of the drug (ng/ml) vs. peak area was plotted. The linearity of the calibration curve was established by the correlation coefficient value obtained from the graph.

Solubility study
The solubility of IND in quite a few solvents was determined by adding an excess amount of drug in 2 ml of the solvent (captex 100, captex 200, captex 355, captex 8000, capryol 90, capmul MCM C8, capmul MCM L8, labrafil M 1994, labrafil M 2125, acconol E, tween 20, tween 40, tween 80, PEG 200, PG and span 20) individually in 5 ml stopper vials and mixed with a vortex mixer for 10 min. To attain an equilibrium, the vials were stored at 25 ˚C in an isothermal shaker for 72h. The samples were centrifuged at 3000 rpm for 30 min. The supernatant was collected and filtered using a 0.22 mm membrane filter. The concentration of IND in the filtrate was analyzed by the UFLC [10].

Ternary phase diagrams
The most stable emulsion systems usually consist of blends of two or more emulsifiers, one portion having lipophilic tendencies, the other hydrophilic [11]. To find out the suitable blend that can better solubilize IND and can form emulsions of nano-size, the aqueous titration method was used to make pseudo ternary systems and the pseudo ternary diagrams were plotted using CHEMIX School software [12]. Surfactant and co-surfactants were mixed in altered weight proportions of 1:1, 1:2, 1:3and 2:1 and were designated as Smix ratios. Then, altered weight ratios of oil and Smix (1:9, 2:8, 3:7, 4:6, 5:5, 6:4, 7:3, 8:2 and 9:1) were prepared [13,14] and the aqueous phase was added dropwise to each altered weight ratios of oil and Smix with continuous stirring till solution converts turbid. For every phase diagram, the systems were inspected for the transparency in appearance visually.

Optimization by 3 2 factorial designs
A 2-factor, 3-level factorial design using Design-Expert (Version 12, Stat-Ease Inc., Minneapolis, MN) software was applied to optimize the formulation compositions. The selected independent variables and experimental responses were the percentage of oil (X1) and Smix (X2); and mean globule size (Y1) and percentage drug release (Y2), respectively. Table 1 illustrates the selected factor levels from the phase diagram for the factorial design. To recognize the significant effect of independent variables on the experimental responses, response surface analyses were carried out [15].

Preparation of IND-NEs
IND-NEs were prepared by the ultrasonic emulsification method by using probe sonicator. A total of 9 NEs were prepared in which the oil phase contained drug, oil and Smix, and aqueous phase contained PG dissolved in water. IND (0.1% w/v) was added by vortexing into the optimized quantity of oil and Smix until clear solutions were obtained. The hot aqueous phase (50˚C) was dropwise added to the oil phase with continuous stirring at 250 rpm using a magnetic stirrer to produce a primary emulsion, which is then sonicated at an amplitude of 50, pulse 5s on and 3s off for 10 min to get nanoemulsion [16].

Particle size measurements
The mean globule size of the IND-NEs was determined by using photon correlation spectroscopy (Zeta seizer, ZS Nano, Malvern Instruments Ltd., UK), which analyzes the fluctuations in the light scattering of the particle. Light scattering was checked at 25˚C at a 90˚ angle. While zeta potential measurements were performed using a disposable zeta cuvette. For each sample, the mean diameter and zeta potential±standard deviation of three determinations was calculated applying multimodal analysis [17].

Refractive index and pH measurement
The refractive index of the IND-NEs was measured by using an Abbes refractometer by placing one drop of formulation on the slide (Nirmal International, New Delhi, India) at 25±0.5˚C in triplicate. The pH of the NEs was measured by a digital pH meter (DPH 504) in triplicate at 25˚C [20].

Surface tension measurements
Surface tension measurements were carried out by the Du Nouy ring method using an electronic Tensiometer (model-K100, KRÜSS, Germany) [21]. The force referred to the wetted length acting on a ring as a result of the tension of the withdrawn liquid lamella when moving the ring from one phase to another is measured in this method and the mathematical formula involved is as followed.

Osmolarity determination
Osmolarity is an important parameter by which one can predict the irritability of the formulation. It was calculated using the following equation [22].

Stability testing
To evaluate the stability, IND-NEs were subjected to thermodynamic stability testing, which comprises of a heating-cooling cycle, freezethaw cycle and centrifugation. The nanoemulsions were examined for stability at temperatures 45 ˚C (heating) and 4 ˚C (cooling) subsequently for not less than 48h. These alternating cycles of heating and cooling were repeated at least six times. Selected nanoemulsions were kept in a deep freezer at-20 ˚C for about 48h. After that, the nanoemulsions were removed from the freezer and kept at room temperature (25 ˚C) to return to the original form. This step of freeze-thaw was repeated 2-3 times. After freeze-thaw cycle nanoemulsions were subjected to centrifugation where they were centrifuged at 3,500 rpm for 30 min, and checked for any phase separation, change in mean globule size and drug content [23].

Drug content
A small fixed volume of IND-NEs was taken and diluted with methanol. Then the samples were analyzed by using UFLC and the concentrations were calculated from the calibration curve [24,25].

In vitro release study
In vitro release of IND-NEs was evaluated and compared with that of control by the diffusion technique using Franz diffusion cells and dialysis membrane. 1 ml of simulated tear fluid (STF) with a dose equivalent amount of IND was served as control. The setup was kept under continuous stirring with a Teflon-coated magnetic bar on a magnetic stirrer at 100 rpm, maintaining a constant temperature of 32±0.5 ˚C. 1 ml sample was withdrawn at every predetermined time points i.e. 0.5, 1, 2, 4, 6, 8, 12 and 24h. The same amount of fresh STF was replaced into the receiver compartment. Withdrawn samples were estimated for the drug content by UFLC 270 nm [26].

Surface morphology
Transmission electron microscopy (TEM) was used to study the morphology of the optimized NE observed under different magnifications by following the standard protocol practicing at RUSKA Labs, PV Narshimha Rao Telangana Veterinary University, Telangana, India. For the TEM study, a drop of the NE was appropriately diluted with water and applied on a carbon-coated grid, and then treated with a drop of 2% phosphotungstic acid and kept aside for the 30s. The carbon-coated grid was then dried and taken onto a slide and observed under the microscope. A combination of bright field imaging at cumulative magnification and diffraction modes was used to evaluate the size and morphology of the NE [27].

Ex vivo corneal permeation studies
The Ex vivo permeation studies were carried out using bovine corneas, which is a reliable method for the prediction of drug transport across the corneal membrane. The eyeballs of bovine were obtained from a slaughterhouse and transported to the laboratory in normal saline maintained at 4 ˚C. The corneas were carefully removed along with a 5-6 mm of surrounding sclera and washed with cold saline. The washed corneas were kept in cold, freshly prepared STF of pH 7.4. The study was carried out by using Franz-diffusion cells in such a way that the epidermal side was in intimate contact with the formulation in the donor compartment. The receptor compartment was filled with STF at 34±0.5 ˚C. The receptor medium was stirred at 50 rpm. The samples were withdrawn at different time intervals and replaced with an equal volume of STF. The permeation study was carried out for 4h, and samples were analyzed by UFLC at 270 nm. Results were expressed as the mean of three experiments±SD. The amount of drug permeated per unit area through the excised cornea (µg/cm2) versus time (h) was plotted and the permeation parameters of drug in the different formulae were calculated [28][29][30].

Ocular irritation studies
Ocular irritancy of optimized NE was studied on healthy adult New Zealand albino rabbits free from visible ocular abnormalities, weighing about 2-3 kg of either sex. 30 μl of test formulation previously sterilized using 0.2 µm syringe filter was instilled into the right eye of each rabbit (n=3) and observed for ocular irritation reactions like redness, conjunctival chemosis and discharge for 24h keeping the untreated eye as a control. The scoring was given from 0 to 4 for the absence to highest observed abnormality and an on the whole Iirr was calculated by summing up the total scores for each category. The Iirr of more than 4 was considered the presence of clinically significant irritation [31].

Drug pharmacokinetics in plasma
6 Rabbits were divided into 2 groups of three animals each. Gruop1 received control (Indocollyre 0.1% ophthalmic solution, Bosch and Lomb) and Group 2 received test (optimized NE) samples of 30 µl each. 200 µl of blood samples were collected into heparinized tubes at 0.5, 0.75, 1, 2, 4, 8, 12 and 24h time-points through the marginal vein. After collection, samples were immediately stored at temperature (-20 ˚C) until the analysis [32,33].

Drug pharmacokinetics in aqueous humour
To study and compare the pharmacokinetics of optimized NE in aqueous humour with that of marketed formulation, 24 animals (three animals corresponding to each sampling point) were used and the time points include 0.5, 0.75, 1, 2, 4, 8, 12 and 24h. Each animal was placed in an individual restraining box, 30 µl of the test sample (optimized NE) and control (marketed product) were topically applied on to the right eye and left eye, respectively. 100 µl of aqueous humor was collected from three animals at each time point by inserting 22 G needle of an insulin syringe into the anterior segment of the eye through the cornea without causing any injury to iris and lens. Before collecting the samples, a combination of 35 mg/kg ketamine hydrochloride and 5 mg/kg xylazine was used to anesthetize the rabbits by injecting intramuscularly [34]. After collection, samples were immediately stored at temperature (-20 ˚C) until the analysis.

Analysis of aqueous humour and blood samples
Different calibration standards ranging from 50-500 ng/ml were prepared by adding 10 µl of known working solution of drug to 90 µl of drug-free rabbit aqueous humor/blood. All samples were vortexed to ensure complete mixing and 20 µl was used for UFLC analysis. The aqueous humor/blood samples collected were mixed with 200 µl of acetonitrile, vortexed and centrifuged for 30 min at 3000rpm. 20 µl of the organic phase was used for UFLC analysis [35,36].

Calibration curve
The calibration curve was plotted between peak area and concentration (ng/ml) as shown in fig. 1. The linearity was seen at a concentration range of 2.5 to 100 ng/ml with a regression coefficient (R 2 ) value of 0.998. The retention time was found to be 5.4±0.34 min as shown in fig. 2.

Solubility study
The solubility of IND in various oils, surfactants and co-surfactants were depicted in fig. 3. Among oils, captex 8000 showed very good solubility. Therefore, it was selected as the oil phase. Among surfactants and co-surfactants, IND had the highest solubility in span 20 and tween 20; hence they were selected as surfactant and cosurfactant respectively for the phase study.    [37]. The zeta potential of IND-NEs ranged from+13.20±4.6 to+23.45±4.82. The positive charge helps to improve the retention time and also helps in improving corneal absorption due to electrostatic interactions with the corneal membrane [38]. The viscosity ranged from 15.3±0.1 to 32.7±0.0 mPas. In general, 15-150 mPas of viscosity is recommended as this low viscosity will ensure the transparency and hence no visual impairment upon eye drop instillation into the eyes without compromising the prolonged retention time [39].

Stability
IND-NEs remained clear with no phase separation or drug precipitation, indicating their excellent physical stability. All the formulations were found to be consistent concerning their mean globule size, drug content, phase separation and transparency during the stability study. The physical data comparing their mean globule size and drug content before and after stability studies were given in table 4.

In vitro release studies
To facilitate comparison between release behaviors of different NE formulae, the In vitro drug release profiles of the IND-NEs were graphically illustrated in fig. 5. The release results of NE revealed that the prepared formulations were found to exhibit sustained release with 67.91±2.01 to 95.90±1.93 % drug release at 24h in comparison to control which showed 99.81±5.21 % drug release within 2 h. The variation in drug release between IND-NEs can be emphasized from the formulation composition variables and the droplet size of the formulations. *Values are expressed as mean±SD (n=3), MGS-mean globule size, PDI-poly dispersity index, ZP-zeta potential, RI-refractive index. From the response surface analysis, it was evident that the oil and Smix compositions showed a significant effect on mean globule size and % drug release. The effects were represented as 3D surface plots in fig. 6. It is prominently noticed that an increase in oil content provided a decrease in globule size to some extent and a further increase in oil content lead to a slight increase in globule size and an opposite effect was seen for percentage drug release. However, by increasing the Smix content, the globule size decreased and the percentage drug release increased. Zainol S et al. (2012) reported an increase in the mean globule size when the oil composition of levodopa lipid emulsions was increased [42]. Reddy et al. (2011) developed felodipine nanoemulsions and reported that the mean globule size reduced from 231.8 to 162.7 nm by increasing the surfactant concentration from 1 to 1.5% [43]. In the present study, the drug release was found to be affected by their droplet size. The decrease in droplet size increased the total number of oil globules and a subsequent increase in their surface area, which led to an increase in drug release. Thus the formulation IND-NE7 with the medium oil and high Smix was found to have the lowest mean globule size and significantly high drug release; hence it was chosen as the optimized formulation and was subjected to further study. Similar results were reported by Yadav et al. (2020) where low oil and high surfactant levels of emulsion composition for ezetimibe showed the lowest globule size of 24.4±2.07 nm and high % drug release [44].

Surface morphology
The droplets in the IND-NE7 appeared dark, and spherical with size more or less similar and less than 100 nm. The image can be seen in fig. 7.

Ex vivo corneal permeation studies using bovine corneas
The drug permeated through the corneal membrane at 4 h from the IND-NE7 and drug solution was found to be 524±1.5 µg/cm 2 and 175±2.6 µg/cm 2, respectively. The results for the drug solution were found quite the opposite to the in vitro drug diffusion rate. This can be explained by the structural complexity of the corneal membrane that opposes the aqueous solution because of the outer lipophilic layer. In the case of IND-NE7, the nano-sized globules can be easily infiltrated by endocytosis and the permeation enhancing the ability of the formulation ingredients will cause temporary changes in the tight junctions of the corneal membrane that enhances the permeation by transcellular pathway. Similar results were reported by Sayed et al. (2017), where the high in vitro drug diffusion rate from the drug solution was reversed in case of corneal permeation and the nanovesicles showed more corneal permeation compared to the drug solution due to the above said reasons [45]. The graphical representation of the comparative ex vivo drug permeation profiles of test and control were shown in fig. 8 and the drug permeation parameters were depicted in table 5.

Ocular irritation test
IND-NE7 found nonirritant with a minimal score of 1 in the Draize eye test. The rabbits showed slight redness of the conjunctiva, which disappeared completely within 15 min, but no lachrymation or chemosis was observed throughout the study indicating the optimized formulation is nonirritant and could be tolerated by the rabbit eye.

In vivo pharmacokinetics
The graphical representation of the comparative In vivo drug concentration profiles of test and control was shown in fig. 9 and the pharmacokinetic parameters were depicted in    The obtained results were found to be statistically significant when analyzed by a one-sample t-test using GraphPad Prism software version, 8.0.2 and were depicted in table 7.

CONCLUSION
Novel cationic IND-NEs for topical ophthalmic delivery were prepared successfully using captex 8000, span 20, tween 20, PG and glycerol. Almost all prepared IND-NEs showed acceptable physicochemical properties and thermodynamic stability. However, IND-NE7 showed the highest drug release among all and was nonirritant to the rabbits' eyes. The drug release rate from NEs was found to be dependent on the oil and Smix content used in NE preparation. IND-NE7 improved IND permeation across the bovine cornea and showed improved corneal absorption with prolonged drug release compared to marketed eye drops. Thus, IND-NE7 offers an effective postoperative treatment with increased ocular bioavailability and improved patient compliance with a decrease in the number of installations per day and a decrease or disappearance of systemic side effects of IND.