Liquid and Solid Self-Emulsifying Drug Delivery Systems (SEDDs) as Carriers for the Oral Delivery of Azithromycin: Optimization, In Vitro Characterization and Stability Assessment

Azithromycin (AZM) is a macrolide antibiotic used for the treatment of various bacterial infections. The drug is known to have low oral bioavailability (37%) which may be attributed to its relatively high molecular weight, low solubility, dissolution rate, and incomplete intestinal absorption. To overcome these drawbacks, liquid (L) and solid (S) self-emulsifying drug delivery systems (SEDDs) of AZM were developed and optimized. Eight different pseudo-ternary diagrams were constructed based on the drug solubility and the emulsification studies in various SEDDs excipients at different surfactant to co-surfactant (Smix) ratios. Droplet size (DS) < 150 nm, dispersity (Đ) ≤ 0.7, and transmittance (T)% > 85 in three diluents of distilled water (DW), 0.1 mM HCl, and simulated intestinal fluids (SIF) were considered as the selection criteria. The final formulations of L-SEDDs (L-F1(H)), and S-SEDDs (S-F1(H)) were able to meet the selection requirements. Both formulations were proven to be cytocompatible and able to open up the cellular epithelial tight junctions (TJ). The drug dissolution studies showed that after 5 min > 90% and 52.22% of the AZM was released from liquid and solid SEDDs formulations in DW, respectively, compared to 11.27% of the pure AZM, suggesting the developed SEDDs may enhance the oral delivery of the drug. The formulations were stable at refrigerator storage conditions.


Introduction
Oral drug delivery is the most widely used and the common route of drug administration because it is convenient, economical, comfortable, and requires no special training for use [1,2]. However, despite these advantages, problems such as poor solubility, low dissolution rates, and limited drug diffusion through the paracellular pathways before eventually entering the systemic circulation making it challenging for many drugs to reach the therapeutic levels via this route [3,4].
Azithromycin (AZM) is a semisynthetic 15-membered macrolide antibiotic [5], with a lipophilic nature (log P = 4), (pKa = 8.74), and a molecular weight of 749 g/mol [6,7]. The drug, which is included Excipients selection was mainly based on their ability to solubilize the highest amounts of the AZM. The impact of lipids classes as long carbon chain triglycerides (LCT) and medium carbon chain triglycerides (MCT) was observed, as such differences in carbon chain lengths were reported to have an influence on drug solubilization potential and the ability to facilitate the emulsification process [38,39]. Extra attention was given to include excipients with previously reported abilities to loosen the cellular TJ for the possibility of enhancing the drug paracellular route diffusion. Besides, certain compounds were reported to have other cellular activities (such as P-gp inhibition) along with the TJ loosening were also investigated. Other excipients which have not been reported for owning any such cellular activity were also included for their potentials in enhancing AZM solubility, as illustrated in Table 1. Medium-chain (C8) TJ opening [33,40,41] Castor Oil -Long-chain (C18) --Octanoic acid (caprylic acid or caprylate) -Medium-chain (C8) TJ opening [42,43] Palm Oil -Long-chain. Mixture of C16 and C18 --

Solubility study
The solubilization capacities of various excipients were investigated using the equilibrium method to ensure the selection of excipients that can solubilize the maximum amount of AZM. In brief, an excess amount of AZM (300 mg) was weighed in multiple 5 mL screw-capped bottles, then 3 mL of the different oils, surfactants, or co-surfactants were added separately into each bottle and vortexed for 2 min using ZX3 vortex mixer (VELP Scientifica, Italy). The filled bottles were then placed in a reciprocating shaker bath (Braun, Melsungen, Germany) at 500 oscillations/min for 72 h and maintained at 37 ± 2 • C. The bottles were centrifuged at 3000 rpm for 15 min to remove the undissolved AZM, and the resulted supernatant layer was filtered through a 0.45 µm syringe driven filter (Whatman, USA). Drug concentration in each vehicle was determined by HPLC.

•
Emulsification studies Based on the results of the AZM solubility study, the selected surfactants and co-surfactants were further screened for their emulsification abilities in the selected oils. Various surfactant to co-surfactant (Smix) ratios of (1:1), (2:1), (3:1), (4:1), (5:1), (6:1) and (1:2) (v/v) were mixed in separate vials and vortexed for few seconds until homogenous mixtures were obtained. A fixed oil ratio of 10% was added to each Smix ratio (90%); the mixtures were then gently vortexed and allowed to equilibrate at a reciprocating shaker bath for 2 h at room temperature. The produced mixtures had a range of hydrophilic-lipophilic balance (HLB) values within the range of 6.55 to 14.22. Although the characterization of SEDDs is mainly conducted using either distilled or deionized water as the dispersing medium [34,[55][56][57], yet, in this work, the self-emulsification efficiency of various mixtures was evaluated on the bases of the produced droplet size (DS), dispersity (Ð) and transmittance percentage (T%), in three different diluents namely: distilled water (DW), 0.1 mM HCl (pH = 4), and simulated intestinal fluids (SIF) (pH = 6.8) for mimicking the oral route of administration and select the most physically stable formulation(s) [58,59]. All the used diluents were freshly prepared and filtered through a 0.45 µm nylon membrane filter before use. Each formulation underwent a standard dilution of 1:1000 ratio (10 µL sample to 10 mL diluent) then gently stirred with a magnetic stirrer and allowed to equilibrate before loading into a cuvette in Pharmaceutics 2020, 12, 1052 5 of 29 a thermostatic chamber to measure DS and Ð through Photon Correlation Spectroscopy (PCS) using laser light scattering spectrometer Zetasizer 1000HSA (Malvern Instrument, UK). The measurement of T% was conducted at a dilution ratio of 10:1000 (100 µL sample to 10 mL diluent), then evaluated at 650 nm using a UV-Vis Spectrophotometer U-2000 (Hitachi, Japan) [60]. Only mixtures that produced DS < 150 nm, Ð ≤ 0.7, and T% > 85 in the three different diluents were selected for the construction of the pseudo-ternary diagram.

Preparation and Characterization of AZM-Loaded L-SEDDs
Out of each identified self-emulsification area, three transparent formulations with DS < 150 nm were selected at three oil concentrations (low, medium, and high) for the maximum DC (or solubilization capacity) studies. There was a 5% difference between the three oil concentrations. A total of 24 L-SEDDs formulations were chosen for the DC studies. To perform the experiment, an excess amount of AZM (200 mg) was accurately weighed in multiple 5 mL screw-capped bottles, then 2 mL of each blank L-SEDDs formulation was added to each bottle. The samples were then treated as per the procedure of the AZM solubility studies, and the DC was measured. Both blank L-SEDDs and their corresponding AZM loaded L-SEDDs (AZM-L-SEDD) formulations were further characterized in terms of DS, Ð, T%, and zeta potential (ZP) in three diluents (DW, 0.1 mM HCl, and SIF). ZP was measured using Zetasizer nanoseries Nano-Z, (Malvern Instrument, UK). The AZM-L-SEDDs formulations with the highest AZM content, DS < 150 nm, Ð ≤ 0.7, T% > 85, and highest ZP value were selected for the solidification process.

Preparation and Characterization of AZM-Loaded S-SEDDs
Various solidification methods such as adsorption to solid carriers, spray drying, freeze-drying, rotary evaporation, melt extrusion-spheronization, and melt granulation are available for the preparation of S-SEDDs [63,64]. In this work, adsorption to solid carriers' method was adopted, as the technique is simple, which involves the addition of the L-SEDDs to the selected carriers with suitable mixing. Furthermore, the produced S-SEDDs by this method are stable and freely flowing [65,66]. Various water-soluble (mannitol and lactose) and water-insoluble (calcium carbonate and Aerosil 200 ® ) solidifying agents were screened to select the suitable one for the conversion of the optimized blank L-SEDDs and AZM-L-SEDDs into blank solid SEDDs (S-SEDDs) and AZM loaded solid SEDDs (AZM-S-SEDDs), respectively.
At first, the adsorption capacity (expressed as a weight ratio of L-SEDDs: solidifying agent) of each solidifying agent was studied through a drop-wise addition and mixing of a fixed portion of blank L-SEDDs (0.5 mL~535.6 mg) with an equivalent portion(s) of the solidifying agent (~535.6 mg) in a porcelain mortar until a non-sticky solid powder was produced. The mixing was done using a glass rod after each addition to ensure uniform distribution of the formulation. The obtained S-SEDDs were left for 24 h at room temperature to dry before further characterization for their DS, Ð, and ZP [67][68][69]. For DS and Ð measurements, 10 mg of the prepared S-SEDDs were dispersed in 10 mL of filtered diluent (DW, 0.1 mM HCl, or SIF). The S-SEDDs formulations solidified with a water-soluble solidifying agent were gently stirred with a magnetic stirrer for 10 min, then allowed to equilibrate and loaded into a cuvette in a thermostatic chamber. While those S-SEDDs solidified with a water-insoluble solidifying agent, the dispersions were first centrifuged using MiniSpin ® plus Eppendorf centrifuge (Eppendorf Ag, Hamburg, Germany) at 8000 rpm for 10 min to remove the water-insoluble solids then the samples were loaded into a cuvette in a thermostatic chamber [70].
The solidifying agent that produced S-SEDDs formulation with the lowest DS, Ð, and highest ZP was selected for solidifying AZM-L-SEDDs. Formulations were further examined for their DS, Ð, and ZP in the previously mentioned three diluents. AZM concentration in AZM-S-SEDDs formulation was quantified by dissolving the solid formulation (1.51 g) in 10 mL methanol and stirring it with a magnetic stirrer for 30 min. The solution was then sonicated in an ultrasonic bath (Branson 5510, Las Animas, CA, USA) for 2 min. Suitable aliquots were taken and diluted with the diluting solution and filtered with a 0.2 µm pore size nylon filter, then injected into HPLC.

Transmission Electron Microscope (TEM)
The morphology and shape of the final selected SEDD formulations in their liquid and solid forms were studied by TEM. A drop of the dispersed liquid or solid formulation in DW (as per the dispersing procedures for DS and Ð measurements) was placed onto 400 mesh carbon-coated copper grid, air dried, then negatively stained with 2% phosphotungstic acid for 5 min at room temperature. After that, the excess sample was removed using a filter paper and allowed to dry before observation. Images were taken under the transmission electron microscope TEM (LIBRA 120, Carl Zeiss, Oberkochen, Germany).

Cell Culture Studies
• Cell stock preparation Caco-2 cells with passage numbers of 25 and 26 were grown and maintained on culture flask in DMEM supplemented with 10% v/v of FBS and 1% v/v of 1% Pen-Strep (named as the complete medium) and were maintained at 37 • C in an atmosphere of 5% CO 2 and 90% relative humidity in a Binder ® constant climate chamber (Tuttlingen, Germany) [71]. Once 80% of the flask surface is covered by cell monolayer, cells were split and washed twice with DPBS, then the trypsin-EDTA solution was added and gently swirled, ensuring the complete covering of trypsin to all the cells. The flask was then incubated for 10 min in an atmosphere of 5% CO 2 and 90% relative humidity. Cells stock was prepared by aspirating out the trypsin-EDTA, and adding the complete medium, then centrifuging this cells stock in HERAEUS ® LABOFUGE 400R centrifuge ThermoFisher Scientific (Waltham, MA, USA) at 900 rpm for 5 min to create a pellet at the bottom of the centrifuge tube. The cell pellet was re-suspended in 3 mL of fresh complete medium for further use as a cell stock. All cell culture buffers and solutions were pre-warmed to 37 • C prior to contact with the cells, and all protocols involved the handling of Caco-2 cells in cultures were performed aseptically.

•
Cytotoxicity assay The cytocompatibility of the optimized blank L-SEDDS and S-SEDDs formulas was investigated using the microtiter tetrazolium assay (MTT). The assay was done according to previously described methods [72,73], with slight modifications. Proper dilutions of the cell stock with the complete medium were made to get the required seeding density of 51,000 cells/cm 2 on Biofil ® tissue culture 96 well plates (Guangzhou, China).
After 24 h of seeding on the 96 well plates at 37 • C in an incubator with 5% CO 2 and 90% relative humidity, the attached cells were washed twice with DPBS, then concentrations of blank L-SEDDS (0.25, 0.5, 1, and 2% v/v) and blank S-SEEDs (0.39, 0.77, 1.54, and 3.08% w/v) were selected for the study. The SEDDs were dispersed in DMEM solution, gently vortexed, filtered using a 0.2 µm sterile filter Pharmaceutics 2020, 12, 1052 7 of 29 membrane, and added to the 96 wells plates then incubated for 4 h. DMEM alone was employed as blank. After incubation, the supernatant layer was discarded, and all cells were washed twice with DPBS. To observe cells' vibrant reductive activity, a final concentration of 1 mg/mL of thiazolyl blue tetrazolium bromide dye solution (MTT solution) was dissolved in DPBS within a tube wrapped in aluminum foil to protect it from light, and the cells in the well plates were loaded with 100 µL of sterile and filtered MTT solution and incubated for another 4 h. Then the supernatant layers were removed, the precipitated formazan crystals were dissolved in DMSO, and the plates were gently rotated on an orbital shaker for 5 min. The color of the resulted formazan solutions in DMSO was measured at 570 nm with background subtraction at 690 nm by a Multiskan™ FC microplate photometer (Thermo Scientific, Waltham, MA, USA) with Skanit software version 3.2. The percentage of cells viability was calculated with respect to the control [74], as follows: Viability % = Abs sample − Abs blank Abs control − Abs blank * 100 (1) where, Abs sample : Absorbance of cells treated with formulations. Abs control : Absorbance of cells not treated with formulations. Abs blank : Absorbance of blank. •

Transepithelial electrical resistance evaluation
Transepithelial electrical resistance (TEER) is a widely accepted quantitative technique to measure the integrity of tight junction dynamics in cell culture models of epithelial monolayers where the TJ proteins in the paracellular route contribute to an ohmic resistance (RTEER) in the equivalent circuit [75]. The TJ opening was investigated as per the previous protocol [76]. Cells suspension corresponding to 51,000 cells/cm 2 was prepared from the cells' stock and added to the apical transwell compartments of all plates. The plates were maintained at 37 • C in an incubator with 5% CO 2 and 90% relative humidity until they formed a polarized/differentiated monolayer, which took approximately a period of 21-23 days. During that period, a regular replacement of cells medium in both chambers (every 48 h), and monolayer's integrity monitoring were conducted by measuring the TEER using an EVOM2 ® epithelial volt ohmmeter with STX100 electrodes (World Precision Instruments Ltd., Stevenage, UK). Figure 1 shows a schematic diagram of the measuring procedure of the TEER. The Caco-2 cell monolayer with average TEER values ≥ 300 Ω.cm 2 indicated that cells are intact [69], and accordingly, they were used in this study. Raw data were measured in Ohm (Ω) and converted to Ω × cm 2 based on the area of transwell plate inserts, which is equal to 0.4 cm 2 for the used 24-well plate [77]. and filtered MTT solution and incubated for another 4 h. Then the supernatant layers were removed, the precipitated formazan crystals were dissolved in DMSO, and the plates were gently rotated on an orbital shaker for 5 min. The color of the resulted formazan solutions in DMSO was measured at 570 nm with background subtraction at 690 nm by a Multiskan™ FC microplate photometer (Thermo Scientific, Waltham, MA, USA) with Skanit software version 3.2. The percentage of cells viability was calculated with respect to the control [74], as follows: where, Abs sample: Absorbance of cells treated with formulations. Abs control: Absorbance of cells not treated with formulations. Abs blank: Absorbance of blank. •

Transepithelial electrical resistance evaluation
Transepithelial electrical resistance (TEER) is a widely accepted quantitative technique to measure the integrity of tight junction dynamics in cell culture models of epithelial monolayers where the TJ proteins in the paracellular route contribute to an ohmic resistance (RTEER) in the equivalent circuit [75]. The TJ opening was investigated as per the previous protocol [76]. Cells suspension corresponding to 51,000 cells/cm 2 was prepared from the cells' stock and added to the apical transwell compartments of all plates. The plates were maintained at 37 °C in an incubator with 5% CO2 and 90% relative humidity until they formed a polarized/differentiated monolayer, which took approximately a period of 21-23 days. During that period, a regular replacement of cells medium in both chambers (every 48 h), and monolayer's integrity monitoring were conducted by measuring the TEER using an EVOM2 ® epithelial volt ohmmeter with STX100 electrodes (World Precision Instruments Ltd., Stevenage, UK). Figure 1 shows a schematic diagram of the measuring procedure of the TEER. The Caco-2 cell monolayer with average TEER values ≥ 300 Ω.cm 2 indicated that cells are intact [69], and accordingly, they were used in this study. Raw data were measured in Ohm (Ω) and converted to Ω × cm 2 based on the area of transwell plate inserts, which is equal to 0.4 cm 2 for the used 24-well plate [77]. On the day of the experiment, the cell cultures medium was removed, the monolayers were washed with DPBS, and fresh medium was placed in both chambers; cell cultures were left for 15  On the day of the experiment, the cell cultures medium was removed, the monolayers were washed with DPBS, and fresh medium was placed in both chambers; cell cultures were left for 15 min. to equilibrate before starting the measurements. The monolayers TEER values were measured in all the studied plates for the first 10 min of the test before adding any samples, including DMSO as control, DMEM as blank, pure AZM, blank and AZM incorporated L-SEDDs, as well as blank and AZM incorporated S-SEDDs formulations.
The selected studied concentration was based on the MTT assay results. The blank and AZM incorporated L-SEDDs at the concentration of 0.5%, and an equivalent concentration of blank and AZM incorporated S-SEDDs formulations, as well as pure AZM, were dispersed in DMEM. All samples were filtered using a 0.2 µm sterile filter membrane immediately before use. At the apical side, 10 µL of the medium was replaced with 10 µL of each of the seven studied solutions, and once the addition is made, TEER values were measured at different time intervals including, every minute for 30 min, and then every hour for a period of 4 h. The resistance of the monolayer was calculated using the following equation: where, R cell layer: Resistance of the Caco-2 monolayer. R sample: Resistance reading of the studied sample. R blank: Resistance reading of the blank.

The In Vitro AZM Release Studies
The in vitro release of the optimized L-SEDDs and S-SEDDs formulations was performed in comparison to the pure AZM powder using Varian ® 7000 USP dissolution apparatus II (Santa Clara, CA, USA), paddle method, as recommended by USFDA [78]. Three different dissolution media were employed, namely: DW, 0.1 mM HCl (pH = 4), and SIF (pH = 6.8). The SIF was prepared by mixing 250 mL of 0.2 M potassium phosphate monobasic solution with 118 mL of 0.2 M sodium hydroxide solution then diluted with DW to 1000 mL [79]. A volume of 900 mL of the selected dissolution medium was placed in each vessel of the dissolution apparatus and was kept at 37 ± 2 • C with a paddle rotating speed of 100 rpm. An amount of 100 mg of pure AZM powder, and equivalent volume and weight of AZM-loaded L-SEDDs and S-SEDDs formulations that are containing 100 mg of the drug were placed in the vessels [15,80]. From each vessel, 5 mL aliquot was drawn at selected time intervals of 5, 15, 30, 60, 90, 120, 180, and 240 min, respectively, and were filtered through 0.45 µm polytetrafluoroethylene (PTEF) filters (Titan ® , West Springfield, MA, USA). The drawn aliquots were directly replaced with an equal volume of the same fresh dissolution media, and AZM was quantified by HPLC.

Stability Studies
Stability studies of the optimized AZM incorporated L-SEDDS and S-SEDDS formulas were conducted for three months under three different temperatures and relative humidity conditions, namely refrigerator (4 ± 2 • C), room condition (25 ± 2 • C and 60 ± 5 relative humidity %), and humidity chamber (40 ± 2 • C and 75 ± 5 relative humidity %). The samples were analyzed for their DS, Ð, ZP, DC %, and the in vitro release profiles at different time intervals of 0, 0.5, 1, 2, and 3 months.

Statistical Analysis
All experiments were carried out for at least in triplicate. To test the statistical significance, a one-way analysis of variance (ANOVA) with Tukey's HSD (honest significant difference) tests were used. The difference in all the analyses was considered statistically significant when p < 0.05. All the statistical tests were done using (Minitab ® statistical software, version 17.2.1.0, Minitab Inc., USA).
Based on the literature review, many reports showed that drugs' solubilities in various oils and surfactants might vary based on the carbon chain lengths of the oils and surfactants [81], including MCT [82][83][84], and LCT [85,86]. This attitude is probably linked to the employed drug physicochemical properties [82,87]. Among the studied excipients, AZM showed higher solubility in the MCT rather than the LCT. This may be linked to the high log p-value of AZM (4.02), where such drugs are shown to be more soluble in MCT than LCT [88]. Furthermore, MCT are reported to have a higher solvent capacity [89,90]. In contrast, AZM showed poor solubility in Cremophor RH40 ® and Cremophor EL ® despite being medium-chain surfactants; this is likely because these surfactants are derived from edible oil (castor oil), where some lipophilic drugs reported to have very low solubility in such oils during SEDDs formulation studies [91][92][93].

Emulsification Efficiency Studies
Based on the solubility study results, all possible combinations of the various excipients were investigated for emulsification efficiency, as illustrated in Table 2. As more than two types of chemicals were incorporated in the SEDDs formulations, then one of the components was used at a fixed ratio (oil) [94], while others (Smix) were studied at the previously mentioned ratios in the method section. Precisely, a fixed ratio of 10% of the oil was employed as recommended to produce spontaneous L-SEDDs formulations with small DS upon dispersion [28]. Among the studied excipients, the highest solubility of AZM was observed in the oils of Capryol 90 ® (80.16 ± 0.83 mg/mL), and Octanoic acid (74.5 ± 0.035 mg/mL), in the surfactants of Labrasol ® (97.44 ± 0.7 mg/mL), and Tween 20 ® (69.43 ± 0.81 mg/mL) and in the co-surfactants of Transcutol ® HP (85.41 ± 0.27 mg/mL), and Span 20 ® (53.48 ± 0.75 mg/mL).
Based on the literature review, many reports showed that drugs' solubilities in various oils and surfactants might vary based on the carbon chain lengths of the oils and surfactants [81], including MCT [82][83][84], and LCT [85,86]. This attitude is probably linked to the employed drug physicochemical properties [82,87]. Among the studied excipients, AZM showed higher solubility in the MCT rather than the LCT. This may be linked to the high log p-value of AZM (4.02), where such drugs are shown to be more soluble in MCT than LCT [88]. Furthermore, MCT are reported to have a higher solvent capacity [89,90]. In contrast, AZM showed poor solubility in Cremophor RH40 ® and Cremophor EL ® despite being medium-chain surfactants; this is likely because these surfactants are derived from edible oil (castor oil), where some lipophilic drugs reported to have very low solubility in such oils during SEDDs formulation studies [91][92][93].

Emulsification Efficiency Studies
Based on the solubility study results, all possible combinations of the various excipients were investigated for emulsification efficiency, as illustrated in Table 2. As more than two types of chemicals were incorporated in the SEDDs formulations, then one of the components was used at a fixed ratio (oil) [94], while others (Smix) were studied at the previously mentioned ratios in the method section. Precisely, a fixed ratio of 10% of the oil was employed as recommended to produce spontaneous L-SEDDs formulations with small DS upon dispersion [28]. As the present SEDDs formulations are designed for oral delivery, the potential effects of GIT conditions on SEDDs emulsification were considered by performing the emulsification studies in different diluents, namely DW, 0.1 mM HCl (pH = 4), and SIF (pH = 6.8). The formulations that were able to maintain their DS < 150 nm, Ð ≤ 0.7, and T% > 85 in all the three diluents were selected for the next steps.
The selection criteria are based on the fact that the smaller DS is associated with greater absorption, and faster release, facilitated hydrophobic drug solubilization, and quick transport from the stomach and distribution along the GIT [38,[95][96][97]. Furthermore, the L-SEDDS formulations with DS < 150 nm showed higher size robustness upon dilutions in the different media in the emulsification study. On the other hand, smaller Ð values were required to get a better system homogeneity [98,99]. While T% > 85 was considered, as it indicates good emulsification in terms of rapid and reproducible equilibrium [100][101][102].

Construction of Pseudo-Ternary Phase Diagram
Ternary phase diagrams are constructed to determine self-emulsification areas at which various concentrations of the excipients (oil, surfactant, and co-surfactant) would possess a transparent appearance and DS < 150 nm. These parameters were considered for selecting the optimum L-SEDDs formulations. Based on the emulsification study, eight diagrams were constructed (Figure 3).
Different blank L-SEDDs samples with three levels of oil: Smix ratios including (8:1), (5:1), and (3.5:1) that meet the set selection criteria were further investigated. These ratios represent high (H), medium (M), and low (L) concentrations on the pseudo-ternary diagram as described in Table 3. This approach was considered to avoid the use of a high concentration of surfactants (>60%), which is associated with GIT irritation when used in L-SEDDS formulations [105,106]. In the pseudo ternary phase diagram, each blue dot represents a formulation with various combination ratios of (oil: surfactant: co-surfactant) where all set criteria are met.
Octanoic acid) (46.9 ± 0.674 mg/mL); this might be attributed to the higher solubility of AZM in Capryol 90 ® than in octanoic acid oil. AZM addition to the blank L-SEDDs formulations also caused some changes in their DS, Đ, ZP, and T% characteristics (Tables 4 and 5). These changes might be linked to the entering of the drug molecule into the interfacial surface where surfactant molecules exist [98]. Unlike formulations constructed with octanoic acid, formulations with Capryol 90 ® oil showed characteristics robustness upon drug incorporation. Such observations were in line with previously reported data using high oil concentrations of 25% Capmul 808G EP/NF [107], as well as 30% of oil mixture (Maisine ® 35-1 and Labrafac ® CC (1:1)) [108,109].
Increasing oil concentration was associated with significant increases in DS and Đ, as well as a decrease in T% in all formulations. These effects could be attributed to the presence of an insufficient amount of surfactant and co-surfactant in the mixture to reduce the DS, and Đ values.
The ZP measurement results showed that the charge of the prepared L-SEDDs formulations was negative when dispersed in DW and SIF. This may be linked to the presence of free fatty acids in the oil phase and/or surfactant [110]. However, upon the dilution of L-SEDDs formulations with 0.1 mM HCl, the ZP charge was shifted to be positive. This could be due to the neutralization of fatty acids and their negatively charged hydroxyl groups (OH − ) by the available positively charged hydrogens (H + ) in such acidic medium.
Furthermore, increasing oil(s) concentration was associated with a significant increase in the ZP values of the formulations, yet, relatively lower ZP values were obtained for formulations constructed with octanoic acid oil compared with those constructed with Capryol 90 ® oil. This might be related to the presence of a carboxyl group in the structure of octanoic acid [104]. The blank L-F1(H) formulation as well as its drug-loaded form AZM-L-F1(H) were shown to have the highest ZP values of (-23.03 ± 1.1 mV)) and (−26.47 ± 0.65 mV) respectively.

AZM Incorporation and Its Impact on L-SEDDs Properties
The optimal drug incorporation into the L-SEDDs formulations depends on the compatibility between the added drug and the physicochemical properties of the formulation. The results of AZM incorporation and its related impacts on the L-SEDDs characteristics are shown in Tables 4 and 5.
The results of exploring DC or solubilization capacity of various formulations showed that it is directly proportional to the oil concentration increment where L-F1 (L) , L-F1 (M) , and L-F1 (H) formulations were found to have a DC of 12.3 ± 0.05, 32.23 ± 0.07, and 60.42 ± 0.4 mg/mL, respectively. While the DC of L-F2 (L) , L-F2 (M) , and L-F2 (H) formulations were 18.68 ± 0.16, 29.97 ± 0.14 and 46.9 ± 0.67 mg/mL, respectively. L-F1 (H) and L-F2 (H) formulations of both oils with Smix of Tween 20 ® and Transcutol HP ® at 2:1 ratio showed the highest DC, and their characteristics were within the selection criteria. The higher was the oil concentration, the more was the amount of AZM incorporated in the formulations. This finding is probably related to the higher solubility of AZM in the oil phase than to its solubility in the used surfactant ( Figure 2). However, L-F1 (H) formulation (prepared with Capryol 90 ® oil) showed higher drug content (60.42 ± 0.4 mg/mL) than L-F2 (H) formulation (prepared with Octanoic acid) (46.9 ± 0.674 mg/mL); this might be attributed to the higher solubility of AZM in Capryol 90 ® than in octanoic acid oil.
AZM addition to the blank L-SEDDs formulations also caused some changes in their DS, Ð, ZP, and T% characteristics (Tables 4 and 5). These changes might be linked to the entering of the drug molecule into the interfacial surface where surfactant molecules exist [98]. Unlike formulations constructed with octanoic acid, formulations with Capryol 90 ® oil showed characteristics robustness upon drug incorporation. Such observations were in line with previously reported data using high oil concentrations of 25% Capmul 808G EP/NF [107], as well as 30% of oil mixture (Maisine ® 35-1 and Labrafac ® CC (1:1)) [108,109].
Increasing oil concentration was associated with significant increases in DS and Ð, as well as a decrease in T% in all formulations. These effects could be attributed to the presence of an insufficient amount of surfactant and co-surfactant in the mixture to reduce the DS, and Ð values.  The ZP measurement results showed that the charge of the prepared L-SEDDs formulations was negative when dispersed in DW and SIF. This may be linked to the presence of free fatty acids in the oil phase and/or surfactant [110]. However, upon the dilution of L-SEDDs formulations with 0.1 mM HCl, the ZP charge was shifted to be positive. This could be due to the neutralization of fatty acids and their negatively charged hydroxyl groups (OH − ) by the available positively charged hydrogens (H + ) in such acidic medium.
Furthermore, increasing oil(s) concentration was associated with a significant increase in the ZP values of the formulations, yet, relatively lower ZP values were obtained for formulations constructed with octanoic acid oil compared with those constructed with Capryol 90 ® oil. This might be related to the presence of a carboxyl group in the structure of octanoic acid [104]. The blank L-F1 (H) formulation as well as its drug-loaded form AZM-L-F1 (H) were shown to have the highest ZP values of (−23.03 ± 1.1 mV)) and (−26.47 ± 0.65 mV) respectively.
Based on all the previously described characterizations, L-F1 (H) and its drug-loaded AZM-L-F1 (H) formulations were found to meet the selection criteria and were selected for further study. The formulation was composed of 22.22%, 51.85%, and 25.93% (v/v) of Capryol 90 ® , Tween 20 ® , and Transcutol HP ® respectively.

Preparation of Solid Self-Emulsifying Drug Delivery System (S-SEDDs)
The water-insoluble solidifying agents such as Aerosil 200 ® and calcium carbonate as well as the water-soluble solidifying agents such as mannitol and lactose are the most common solidifiers used for the production of S-SEDDs formulations [111][112][113][114][115][116]. In this study, the adsorption capacity (expressed as a weight ratio of L-SEDDs (L-F1 (H) ): Solidifying agent) of these solidifying agents were investigated. It was observed that each solidifying agent had a different adsorption capacity to yield a non-sticky solid powder of S-SEDDs formulations, as illustrated in Table 6. Table 6. Adsorption capacity and characterization of blank S-SEDDs (S-F1 (H) ) formulation. Mean ± SD, n = 3.
The results of the solidification process showed that S-F1 (H) formulation had bigger DS, higher Ð values, and significantly lower AZM DC (38.79 ± 0.52 mg/g) in comparison to the L-F1 (H) formulation which had smaller DS, lower Ð values and higher AZM DC (60.42 ± 0.4 mg/mL). Such effects are linked to the need of using large amounts of Aerosil 200 ® (~535.6 mg) to produce solid S-SEDDs formulation. The same pattern of characteristic change was previously reported upon solidifying L-SEDDs formulation [117]. Table 7. Characterization of the optimized blank and AZM loaded S-F1(H) formulations in DW, 0.1 mM HCl, and SIF diluents. Mean ± SD, n = 3.

Transmission Electron Microscope (TEM)
The TEM images of the optimized AZM-loaded L-F1 (H) and S-F1 (H) formulations have revealed the formation of emulsion upon their dispersion ( Figure 4A,B). The dispersed droplets of the prepared liquid and solid SEDDs formulations were spherical in shape with no signs of aggregation ( Figure 4C,D). These findings are in line with previous TEM studies of liquid and solid SEDDS formulations [118][119][120][121].

Cytotoxicity Assay
Despite the fact that AZM is known to be cytocompatible [122,123] and the used excipients are considered relatively safe where they are used in the food, cosmetics or pharmaceutical industries

Cytotoxicity Assay
Despite the fact that AZM is known to be cytocompatible [122,123] and the used excipients are considered relatively safe where they are used in the food, cosmetics or pharmaceutical industries (such as Transcutol ® HP [124], Tween 20 ® [125], and Capryol 90 ® [126]), However, the cytotoxicity study in this work was conducted before proceeding to the TEER assay to make sure that the SEDDs effects on TEER must not be linked to the cytotoxicity of the excipients or their combination in the developed SEDDs formulations. Caco-2 cell monolayer model was used in this study because it is a reliable in vitro model, an approved standard by USFDA and pharmaceutical companies to investigate the cytotoxicity and behavior of developed formulations and/or drug in the intestinal tract [127,128]. The MTT assay is one of the most popular in vitro cytotoxicity assays. Its mechanism is based on the conversion of the water-soluble thiazolyl blue tetrazolium bromide dye to an insoluble purple-colored formazan, which is quantified by a microplate photometer [129]. The amount of formed crystals and the obtained readings represent the metabolic activities, and thus the number of the present viable cells [129,130]. The selected concentrations of the finally optimized blank formulations were chosen so that both formulations would have the same drug concentration when later preforming the TEER evaluation.
The results of the MTT assay indicated that the blank L-F1 (H) and S-F1 (H) formulations had low cytotoxicity with no induced toxic effects during the 4 h incubation period and at the different studied concentrations. As shown in Figure 5, the cells viabilities were above~90% and~86% for the tested blank L-F1 (H) and S-F1 (H) formulations, respectively. based on the conversion of the water-soluble thiazolyl blue tetrazolium bromide dye to an insoluble purple-colored formazan, which is quantified by a microplate photometer [129]. The amount of formed crystals and the obtained readings represent the metabolic activities, and thus the number of the present viable cells [129,130]. The selected concentrations of the finally optimized blank formulations were chosen so that both formulations would have the same drug concentration when later preforming the TEER evaluation.
The results of the MTT assay indicated that the blank L-F1(H) and S-F1(H) formulations had low cytotoxicity with no induced toxic effects during the 4 h incubation period and at the different studied concentrations. As shown in Figure 5, the cells viabilities were above ~90% and ~86% for the tested blank L-F1(H) and S-F1(H) formulations, respectively.

Transepithelial Electrical Resistance Evaluation
The modulation of TJ is a potent strategy to improve hydrophobic drug delivery [131]; such modulations were related to the used formulations' excipients, mostly the surfactants [132][133][134] and some fatty acids [42]. In general, the opening of the TJ is associated with a decrease of transepithelial electrical resistance (TEER) [135,136]. The TEER evaluation study results are shown in Figure 6.

Transepithelial Electrical Resistance Evaluation
The modulation of TJ is a potent strategy to improve hydrophobic drug delivery [131]; such modulations were related to the used formulations' excipients, mostly the surfactants [132][133][134] and some fatty acids [42]. In general, the opening of the TJ is associated with a decrease of transepithelial electrical resistance (TEER) [135,136]. The TEER evaluation study results are shown in Figure 6.  It was revealed that pure AZM had no significant impact on TJ in the Caco-2 cell line model and within the studied period of 4 h. Such results are in contrast with previous studies that showed pure AZM could increase the TEER values in airway epithelial cells [21]. Such a difference in AZM behavior might be linked to the variations in cell lines' type, source, passage number, and culturing conditions. These results are suggesting the need for more studies to understand the exact cellular mechanisms of AZM permeation and its relation to its bioavailability using different cell lines and in vivo levels. Upon treating the cells with the set samples, and during the first 10 min of the measurement period, all samples (AZM, controls, blank and drug-loaded formulations) showed almost no significant changes in the TEER values. After that, only blank and drug-loaded formulations of L-SEDDs and S-SEDDs were able to decrease the TEER values sharply. These declines were significant (p < 0.05) in comparison to the TEER values of pure AZM, and could be linked to the ability of the optimized formulations to modulate the intestinal tight junction, possibly due to the presence of Capryol 90 ® , which has a proven ability to produce an appreciable reversible opening of TJ [33], which in turn might offer better delivery of AZM via paracellular pathways [137]

The In Vitro Release Studies
The in vitro release of AZM is usually conducted in a medium that mimics the intestinal pH such as phosphate buffer (pH = 6). This is because AZM is not sufficiently stable in acidic pH mediums such as simulated gastric fluids (SGF, pH = 1.2) [138]. Accordingly, three different release solutions were employed in this study including 0.1 mM HCl (pH = 4), DW (pH = 5.5), and SIF medium (pH = 6.8) [103,104,139]. Figure 7 represents the in vitro release profiles of AZM from AZM-L-F1 (H) and AZM-S-F1 (H) formulations, compared to the release of the pure AZM in the media of DW, 0.1 mM HCl, and SIF, respectively.
The liquid and solid SEDDs formulations were able to significantly (p < 0.05) increase the percentage of AZM release by 2.2-fold in DW, 1.9 and 1.8 folds in HCl, 1.7 and 1.6 folds in SIF, respectively in comparison with the pure AZM during the 4 h dissolution period.
The cumulative release percentages from the pure AZM were only 43.68 ± 0.84%, 50.57 ± 1.24% and 58.87 ± 2.23% in DW, HCl and SIF media respectively. In contrast, the AZM-L-F1 (H) formulation released > 90% of the AZM within the first 5 min of the dissolution in all the studied dissolution media including 96.65 ± 1.77%, 96.21 ± 1.06%, 98.49 ± 0.77% in DW, HCl, and SIF, respectively. While the AZM-S-F1 (H) formulation showed a slower release pattern whereby in 5 min, only 52.22%, 74.74%, and 70.01% were released in DW, HCL, and SIF, respectively. Unlike the L-SEDDs, S-SEDDs formulation needed 60 min, 30 min, and 15 min to reach > 90% of AZM in DW, HCL, and SIF. Such delay of AZM release from S-SEDDs formulation was in line with previous studies [34,140]. The researchers linked it to that the S-SEDDs formulation needed more step such as desorption of the adsorbed SEDDs from the Aerosol 200 ® during dissolution process [141]. Based on these results, the developed formulations were able to significantly enhance the dissolution rate of AZM in comparison to the pure drug powder.
Pharmaceutics 2020, 12, x 19 of 29 It was revealed that pure AZM had no significant impact on TJ in the Caco-2 cell line model and within the studied period of 4 h. Such results are in contrast with previous studies that showed pure AZM could increase the TEER values in airway epithelial cells [21]. Such a difference in AZM behavior might be linked to the variations in cell lines' type, source, passage number, and culturing conditions. These results are suggesting the need for more studies to understand the exact cellular mechanisms of AZM permeation and its relation to its bioavailability using different cell lines and in vivo levels. Upon treating the cells with the set samples, and during the first 10 min of the measurement period, all samples (AZM, controls, blank and drug-loaded formulations) showed almost no significant changes in the TEER values. After that, only blank and drug-loaded formulations of L-SEDDs and S-SEDDs were able to decrease the TEER values sharply. These declines were significant (p < 0.05) in comparison to the TEER values of pure AZM, and could be linked to the ability of the optimized formulations to modulate the intestinal tight junction, possibly due to the presence of Capryol 90 ® , which has a proven ability to produce an appreciable reversible opening of TJ [33], which in turn might offer better delivery of AZM via paracellular pathways [137] 3. 8

. The In Vitro Release Studies
The in vitro release of AZM is usually conducted in a medium that mimics the intestinal pH such as phosphate buffer (pH = 6). This is because AZM is not sufficiently stable in acidic pH mediums such as simulated gastric fluids (SGF, pH = 1.2) [138]. Accordingly, three different release solutions were employed in this study including 0.1 mM HCl (pH = 4), DW (pH = 5.5), and SIF medium (pH = 6.8) [103,104,139]. Figure 7 represents the in vitro release profiles of AZM from AZM-L-F1(H) and AZM-S-F1(H) formulations, compared to the release of the pure AZM in the media of DW, 0.1 mM HCl, and SIF, respectively.
The liquid and solid SEDDs formulations were able to significantly (p < 0.05) increase the percentage of AZM release by 2.2-fold in DW, 1.9 and 1.8 folds in HCl, 1.7 and 1.6 folds in SIF, respectively in comparison with the pure AZM during the 4 h dissolution period.
The cumulative release percentages from the pure AZM were only 43.68 ± 0.84%, 50.57 ± 1.24% and 58.87 ± 2.23% in DW, HCl and SIF media respectively. In contrast, the AZM-L-F1(H) formulation released > 90% of the AZM within the first 5 min of the dissolution in all the studied dissolution media including 96.65 ± 1.77%, 96.21 ± 1.06%, 98.49 ± 0.77% in DW, HCl, and SIF, respectively. While the AZM-S-F1(H) formulation showed a slower release pattern whereby in 5 min, only 52.22%, 74.74%, and 70.01% were released in DW, HCL, and SIF, respectively. Unlike the L-SEDDs, S-SEDDs formulation needed 60 min, 30 min, and 15 min to reach > 90% of AZM in DW, HCL, and SIF. Such delay of AZM release from S-SEDDs formulation was in line with previous studies [34,140]. The researchers linked it to that the S-SEDDs formulation needed more step such as desorption of the adsorbed SEDDs from the Aerosol 200 ® during dissolution process [141]. Based on these results, the developed formulations were able to significantly enhance the dissolution rate of AZM in comparison to the pure drug powder.
(A)  Tables 8 and 9. Both formulations showed better stability in the refrigerator condition (4 ± 2 °C) when compared with the other studied storage conditions (i.e., room and humidity chamber conditions). For instance, after three months of storage in the refrigerator, the ZP values of both formulations were maintained unchanged, while there were slight but significant increases in DS of ~8 and 5 nm, respectively. The same increment patterns were seen in the Đ values of the AZM-L-F1(H) formulation at the three storage conditions. In contrast, there were no changes in the Đ values of the AZM-S-F1(H) formula under all the storage conditions during the three months period.

Chemical Stability
All formulations showed no statistically significant changes in DC% when stored in the refrigerator (p < 0.05). However, under the room storage condition (25 ± 2 °C/60 ± 5% RH), only AZM-S-F1(H) formulation was stable with no significant decreases in the DC% for the period of three   Tables 8 and 9. Both formulations showed better stability in the refrigerator condition (4 ± 2 • C) when compared with the other studied storage conditions (i.e., room and humidity chamber conditions). For instance, after three months of storage in the refrigerator, the ZP values of both formulations were maintained unchanged, while there were slight but significant increases in DS of~8 and 5 nm, respectively. The same increment patterns were seen in the Ð values of the AZM-L-F1 (H) formulation at the three storage conditions. In contrast, there were no changes in the Ð values of the AZM-S-F1 (H) formula under all the storage conditions during the three months period.

Chemical Stability
All formulations showed no statistically significant changes in DC% when stored in the refrigerator (p < 0.05). However, under the room storage condition (25 ± 2 • C/60 ± 5% RH), only AZM-S-F1 (H) formulation was stable with no significant decreases in the DC% for the period of three months. On the other hand, decreases in DC% were observed in both formulations when stored at the humidity chamber (40 ± 2 • C/75 ± 5% RH).

Conclusions
Liquid and solid emulsifying SEDDS were successfully developed and optimized. The medium carbon chain triglycerides (MCT) were better than the long carbon chain triglycerides (LCT) as solubilizing agents for AZM. The hydrophobic solidifying agent (Aerosil 200 ® ) was found to have the best solidification capacity among the studied solidifying agents to produce solid SEDDs. In this study, the main suggested reasons behind the low oral bioavailability of AZM were addressed, whereby both the liquid and solid SEDDs formulations improved the solubility and dissolution rate of the drug. However, more studies are required to specify the exact reasons that are responsible for the low oral bioavailability of this drug. The cytotoxicity study using MTT assay revealed that the AZM-SEDDs formulations had a low toxicity profile. Furthermore, the transepithelial electrical resistance (TEER) evaluations showed that both formulations had the abilities to open the tight epithelial junction. The formulations were relatively stable at the refrigerator storage condition for 3 months. Therefore, it is suggested that the developed liquid and solid self-emulsifying drug delivery systems are potential nanocarriers for the oral delivery of AZM.