Chemicals

1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) (LIPOID PC 18:1/18:1, batch 556890-216 01/073, CAS 4235-95-4) and 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) (LIPOID PC 14:0/14:0, batch 562236-01/028, CAS 18194-24-6) were purchased from Lipoid GmbH (Ludwigshafen, Germany). Latanoprost (LAT) (HY-B0577/CS-2758, batch 90,610, CAS 130203-82-4) was obtained from MedChemExpress (Monmouth Junction, New Jersey, United States of America). Cholesterol (CHOL) (C8667-5G, batch SLBW6939, CAS 57-88-5, ≥ 99%), α-tocopherol acetate (VE) (T3376-5G, batch MKCM6113, CAS 7695-91-2, ≥ 96%), L-leucine (LEU) (L8000-100G, batch BCBZ3028, CAS 61-90-5, ≥ 98%), trifluoroacetic acid (TFA) (302031-100ML, batch 102532920, CAS 76-05-1, ≥ 99%), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (475989, CAS 298-93-1, ≥ 98%), benzalkonium chloride (BAK) (B6295, batch 1298263, CAS 63449-41-2), boric acid (H3BO3) (B6768-1 KG, batch BCBS7652, CAS 10043-35-3, ≥ 99.5%) and sodium tetraborate decahydrate (Na2B4O7·10H2O) (S9640-2.5 KG, batch BCCB7490, CAS 1303-96-4, ≥ 99.5%) were supplied by Sigma-Aldrich (Madrid, Spain). Sodium chloride (NaCl) was obtained from Merck (1.06404.1000, Merck KGaA, Darmstadt, Germany). Betaine anhydrous (BET) (204241000, batch A0419439, CAS 107-43-7, ≥ 98%) and D-(+)-trehalose dihydrate (TREH) (BP2687-100, batch 215417, CAS 6138-23-4) were purchased from Fisher Scientific (Geel, Belgium). Acetonitrile (221881.1612) and dimethyl sulfoxide (DMSO) (A3672,0100) were acquired from PanReac AppliChem (Barcelona, Spain). Ophthalmic-grade sodium hyaluronate (HA) (F002503, batch 6/0001, molecular weight 400–800 kDa), was supplied by Abaran Materias Primas S.L. (Madrid, Spain). Single-dose Monoprost® (MF) (Laboratoires Théa, Madrid, Spain) and single-dose Lusan® (0.9% w/v NaCl) (Hartmann SA, Barcelona, Spain) were used as benchmarks for the in vivo studies. Water was purified using a Milli-Q® filtration system (Millipore Corporation, Billerica, MA, USA).

Liposome manufacture

Liposomes were prepared as per the lipid film hydration protocol developed by Bangham et al., albeit with some modifications, as shown in Fig. 1 [37, 38]. Components were selected considering the mechanical characteristics and in vitro tolerance results reported in previous studies [34, 38, 39]. The aqueous phase was composed of a borate buffer (0.84% w/v H3BO3, 0.14% w/v Na2B4O7·10H2O), 1.04% w/v TREH, 0.40% w/v BET and 0.90% w/v LEU. DOPC, DMPC, CHOL and VE were respectively incorporated in a weight ratio of 6:2:1:0.08 to constitute the lipid phase. First, the aforementioned lipid mixture was dissolved in chloroform and then placed in a rotary evaporator (Buchi R-205, Massó Analítica S.A., Spain) at reduced pressure (100 mPa and 50 mPa for 30 min, at 150 rpm) at 33 °C. Once this step was completed, organic solvent traces were removed with nitrogen flow for 30 s. Next, 10 mL of the aqueous phase were added to swell the lipid film at 185 rpm for 15 min. The liposomes obtained were left to stand at room temperature for 2 h and were then subjected to sonication in an ultrasonic bath for 15 min. (Bandelin® Sonorex Digiplus, DL 510 H, Berlin, Germany). Unilamellar vesicles were obtained with a high-pressure extruder (Lipex Biomembrane™, Vancouver, BC, Canada) by passing the liposomes through a 0.8 µm membrane (Nuclepore™ Track-Etch Membrane, 10417304, batch A30003737, Whatman™, Cytiva Europe GmbH, Freiburg, Germany) for 10 cycles, then by passing them through 0.2 µm polycarbonate filters (Nuclepore™ Track-Etch Membrane, 110606, batch 7084288, Whatman™, Cytiva Europe GmbH, Freiburg, Germany) for 10 cycles at 25 °C under a fume hood. At the end of the extrusion process the liposomes were diluted 1:1 with the aqueous phase (the same phase mentioned above) to achieve the desired final phospholipid concentration (1% w/v). To obtain liposomes dispersed in 0.2% w/v HA, 0.4% w/v HA was added to the aqueous phase beforehand and was then used to dilute (1:1) the liposomes after the extrusion cycles. After dilution, the liposomes were left overnight at 4 °C to allow full hydration.

Fig. 1

Detailed diagram of liposome manufacture as per the lipid film hydration method. The image was created with BioRender (created on 13 November 2023)

In the case of the LAT-loaded liposomes, LAT dissolved in chloroform was included in the lipid phase to achieve a final concentration of 50 µg mL−1. The liposomal formulations produced, their lipid bilayer and their aqueous phase compositions are detailed in Table 1 below.

Table 1 Composition of the liposomal formulations: unloaded synthetic liposomes (B-LIP), unloaded synthetic liposomes with 0.2% w/v HA (B-HA-LIP), LAT-loaded synthetic liposomes (LAT-LIP) and LAT-loaded synthetic liposomes with 0.2% w/v HA (LAT-HA-LIP)Physicochemical characterisation of the liposomesMorphological evaluation

The liposomes were visualised using cryo-transmission electron microscopy (cryo-EM) to confirm their formation and explore their structure. The morphology and appearance of the four formulations were evaluated using a 200 kV FEI Talos Arctica device (FEI Company, Hillsboro, Oregon, USA). Regarding the cryo-EM analyses, B-LIP was previously diluted at a ratio of 1:10 and a small amount (3 µL) was subsequently placed on the upper side of Quantifoil® Lacey Carbon support (Cu/Rh lacey carbon grids) and then blotted. An FEI Vitrobot Mark IV device was employed to plunge the sample into liquid ethane. B-LIP was also processed in a Talos Arctica device using an X-field emission gun operating at 200 kV. EPU Software (ThermoFisher Scientific®) on a Falcon III device was used to capture the images. These were recorded under low-dose conditions at nominal magnifications of 22,000 (0.94 nm/pixel sampling rate) and 92,000 (0.11 nm/pixel sampling rate) for field and individual captures, respectively. Lastly, the ImageJ software (Fiji, version 1.54d) was used to process each image. The same procedure was repeated for B-HA-LIP, LAT-LIP and LAT-HA-LIP.

Vesicle size and zeta potential

Liposome sizes and zeta potential were determined at 25 °C by diluting the samples in Milli-Q® water at a ratio of 1:10. Both parameters were evaluated, respectively, in disposable transparent polystyrene cuvettes (batch 67.754, Sarstedt, Nümbrecht, Germany) and disposable folded capillary cells (DTS1070, Zetasizer Nano Series®, Malvern Panalytical Ltd, Malvern, United Kingdom) using the ZS Xplorer® software and Zetasizer Lab® (Malvern Panalytical Ltd, Malvern, United Kingdom). Material refractive index and absorption were correspondingly set at 1.35 and 0.001, whereas water refractive index, viscosity and dielectric constant were respectively set at 1.33, 1.05 and 78.5.

pH

The pH of the four different batches was measured in triplicate using a pH meter (model GLP-2, Crison Instruments SA, Barcelona, Spain) calibrated with standards at pH equal to 7.00 and 9.00 and equipped with a microelectrode (InLab, Mettler, Barcelona, Spain).

Surface tension

Liposomal formulations surface tension was measured as per the Wilhelmy plate method. Briefly, all measurements were taken in triplicate using a K-11 digital tensiometer (Kruss GmbH, Hamburg, Germany) previously calibrated with Milli-Q® water (70.0 ± 2.0 mN/m) at 33 °C to simulate ocular surface temperature [40]. Before each analysis, the liposomal formulations were pre-heated to 33 °C and equilibrated for 3 min.

Osmolarity

Osmolarity was measured by the freezing point depression technique using a Fiske® single-sample micro osmometer (model 210, Fiske® Associates, Norwood, Massachusetts, United States). Three standard solutions of 50, 290 and 850 mOsm/L were employed to calibrate the equipment.

Rheological studies

Viscosity was measured by a parallel plate system linked to a Discovery HR1 hybrid rheometer (TA Instruments, New Castle, Delaware, United States). Viscosity was determined by increasing shear rates from 0–1000 s−1 in 30 steps. The study was performed at room temperature.

Quantitation of the latanoprost in the liposomal formulation

High-performance liquid chromatography coupled with ultraviolet detection (HPLC-UV) was employed to quantify the LAT in the formulations produced. The apparatus used to perform the analyses was a Waters® Acquity Arc Bio UHPLC device (Waters, Madrid, Spain) paired with a Waters® Photodiode Array 2998 detector; samples were analysed with the HPLC line of the instrument. The Empower 3® software was used to collect and process the chromatographical results. Analyses were based on a pre-existing method employing an Ascentis® C18 HPLC Column (10 cm × 4.6 mm, 3 µm) (265458-04, Supelco®, Sigma-Aldrich, Madrid, Spain) as stationary phase [41]. The reversed-phase HPLC column was kept at 30.0 ± 0.2 °C throughout the analysis. The injection volume was 10 µL and the flow rate was set at 1 mL min−1. The composition of the mobile phase was acetonitrile – 0.1% v/v TFA in Milli-Q® water (70:30% v/v). The isocratic elution was examined for 3.00 min and LAT was detected at 210.0 nm [36]. Several standard dilutions were prepared from a 5 mg mL−1 LAT stock solution in acetonitrile (10, 25, 40, 50, 60, 75, 100 µg mL−1) to calculate the regression line. The HPLC-UV analyses were only performed in triplicate on the LAT-LIP formulation. Precision, expressed as relative standard deviation percentage, remained lower than 1.55% for all the concentrations of the regression line, whereas the accuracy, expressed as the relative error, was in any case below 1.05%. Furthermore, limit of detection (LOD) and limit of quantitation (LOQ) were considered when LAT was not detectable in the chromatograms in order to determine the maximum range of quantitation. As per IUPAC principles, LOD and LOQ were calculated using the standard error of the intercept (σ) obtained from the calculation of the regression line divided by its slope (m) [42].

$$LOD=\left(\frac\right)\cdot 3.3$$

LOD and LOQ resulted respectively equal to 0.27 µg mL−1 and 0.83 µg mL−1, whereas linearity parameters, σ and m were correspondingly calculated as 899.40 and 10892.12.

Total drug loading

To quantify the total LAT loading present (LATtot), the liposomal formulation (1 mL) was lyophilised (Telstar Lyoquest® 30 benchtop freeze dryer, Telstar, Terrassa, Spain) once elaborated (freezing: − 60 °C/60 min, drying: − 60 °C/12 h/0.1 mBar). Afterwards, the same volume of acetonitrile (1 mL) was added to solubilise the pellet. Following centrifugation (5000 rpm, 5 min, 20 °C, Micro 220R, Hettich®, Aizarnazabal, Guipuzcoa, Spain) and filtration (0.22 µm-pore nylon syringe filters, JNY022013N, Filter-Lab®, Barcelona, Spain), the solution was placed in vials and injected into the chromatographic system as described in 2.4. LATtot was calculated as the ratio between the LAT concentration present in the analysed samples and the theoretical concentration (50 µg mL−1) included in preparation.

$$_\, \left[\%\right]=\frac_ \;in \;LAT-LIP}_ \;in\; LAT-LIP}\cdot100$$

Encapsulation efficiency

Encapsulation efficiency (EE) refers to the LAT percentage trapped inside the vesicles’ lipid bilayer. Ultrafiltration was employed to determine the LAT concentration present in the aqueous core and in the buffered vehicle, which was then deducted from total LAT concentration [36]. LAT-LIP was diluted with Milli-Q® water at a ratio of 1:10 and a suitable volume (0.5 mL) was pipetted into centrifugal tubes fitted with appropriate filters (Amicon® Ultra 0.5 mL with Ultracel® 50 kDa regenerated cellulose 50000 NMWL, UFC505024, Millipore®, Sigma-Aldrich, Madrid, Spain). This was then centrifuged (14000 rpm, 5 min, 20 °C) and filtered. The ultrafiltered solutions were freeze-dried (freezing: − 60 °C/60 min, drying: − 60 °C/12 h/0.1 mBar), incubated in acetonitrile (1 mL) to dissolve the residual LAT present and vortex-mixed (Vortex D-051, Dinko®, Barcelona, Spain) for 3 min. Samples were then filtered, put in vials and subjected to HPLC-UV analyses as described in 2.4. Encapsulation efficiency was expressed considering the LOD and LOQ. Considering the total LAT concentration present in the LAT-LIP and the LAT concentration detected after ultrafiltration (LATfree), EE was calculated by the following equation:

$$EE \,\left[\%\right]= \frac_- _}_}\cdot 100$$

In vitro studies in human corneal and conjunctival cells

Human immortalised corneal epithelial cells (hTERT-HCECs) (Evercyte GmbH, Vienna, Austria) and human immortalised conjunctival epithelial cells (IM-HConEpiC) (Innoprot, Bizkaia, Spain) were used to conduct cytotoxicity studies of the four liposomal formulations, which were previously filtered through 0.22 µm-pore cellulose acetate syringe filters (JCAS022025K, Filter-Lab®, Barcelona, Spain). MF was used as the benchmark. This research group has previously investigated the in vitro viability of these liposomal formulations’ components (e.g. synthetic phospholipids and osmoprotective substances, alone and in combination) [38, 43]. Culture cells were maintained under appropriate conditions (37 °C, 5% v/v CO2, 95% v/v humidity) in T75 tissue culture flasks (Sarstedt, Madrid, Spain) and their supplemented media were changed every 2 days. The hTERT-HCECs were cultured with EpiLife™ cell culture medium (Life Technologies Corporation, Madrid, Spain) and supplemented with EpiLife™ Defined Growth Supplement (EDGS® 100X, Life Technologies Corporation, Madrid, Spain) and 1% w/v penicillin (10000 units mL−1) and streptomycin (10000 µg mL−1) (Pen Strep®, Life Technologies Corporation, Madrid, Spain). To ensure correct cell attachment and preservation, the flasks were coated with 2% w/v gelatine before adding the cell suspension. Conjunctival cells were cultured with IM-Ocular Epithelial Cell Medium (Innoprot, Bizkaia, Spain) and collagen (1 mg mL−1) was employed as flask coating (Collagen I Coating Kit, Innoprot, Bizkaia, Spain).

Both hTERT-HCECs and IM-HConEpiC were exposed to the four liposomal formulations for 1 h and 4 h to respectively simulate short-term treatment and chronic topical exposure [44]. The cytotoxicity of the four formulations was evaluated by the MTT assay [38]. hTERT-HCECs and IM-HConEpiC were seeded in 96-well plates (Sarstedt, Madrid, Spain) at a cell density equal to 20000 and 25000 cells per well, respectively, and incubated overnight (16 h). After that, supernatants were discarded and cells were exposed for 1 h and 4 h to B-LIP, B-HA-LIP, LAT-LIP, LAT-HA-LIP ([DOPC-DMPC] = 0.5% w/v) and MF using a volume of 100 µL of formulation per well ([LAT] = 0.0025% w/v) and 100 µL of the supplemented medium per well. Afterwards, supernatants were carefully removed and MTT solution (0.33 mg mL−1) was added to the plates and incubated for 4 h at 37 °C. Following careful aspiration of the MTT solution, 100 µL of DMSO were added to each well to dissolve the formazan crystals. The extent to which MTT was reduced to its formazan salt by hTERT-HCECs was assessed by SPECTROStar Nano® absorbance microplate reader (BMG LABTECH, Ortenberg, Germany) measurements at 550 nm with prior shaking for 5 min. The negative control was a 0.9% w/v NaCl solution, and the positive control was an aqueous solution of 0.005% w/v BAK, both diluted 1:1 with supplemented cell culture medium. Although it is a component often included as a preservative in topical ophthalmic medications, BAK was selected as it is linked to inflammatory cascades in corneal and conjunctival cells [45, 46].

In vivo studies

Normotensive albino male New Zealand rabbits (n = 11) were purchased from Granja San Bernardo (Tulebras, Spain); at the end of the study they had a mean weight of 3.24 ± 0.16 kg. The animals were housed in individual cages with free access to food and water and were maintained under 12 h light-dark cycles (lights on from 8:00 am to 8:00 pm) at a room temperature of 18 °C and at approximately 50% relative humidity in a controlled atmosphere. In vivo studies fulfilled the 3R principles and followed the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic Vision Research, European Council Directive (86/609/EEC) and the Spanish Regulation on Experimental Studies with Animals (RD 53/2013 of 1 February 2013, modified by RD 118/2021 of 23 February 2021). The Animal Experimentation Ethics Committee of the Complutense University of Madrid approved the study protocol under the code ES280790000086.

In vivo tolerance

B-HA-LIP and LAT-HA-LIP tolerance were assessed by topically administering 25 µL of each formulation in both animal eyes (n = 6 animals). 0.9% w/v NaCl and MF were used as controls following the same procedure (n = 6 animals). Evaluation was performed by a masked observer both before the instillations and 4 h later. Ocular signs were classified according to the Draize test (ISO 10993-10:2010). This test includes a scoring system that encompasses six distinct components of observable alteration, such as toxicity and inflammation, in the anterior segment of the eye; these elements include the density and area of corneal opacification, the severity of eventual iritis, conjunctival redness, oedema, and any discharge that may be produced (Table 2). Before beginning the study, the authors selected a total score equal to 10 as the cut-off for establishing good tolerance of the tested formulations [47]. Pictures of the rabbits’ eyes before and 4 h after each instillation were taken using a VX75 slit lamp (Luneau Technology, Chartres, France).

Table 2 Scoring system, according to the Draize test, employed to determine the in vivo tolerance of the liposomal formulationsHypotensive studies in rabbits

The efficacy of the proposed liposomal formulations (LAT-HA-LIP) in reducing IOP was assessed in a total of six rabbits (n = 12 eyes). The same protocol was followed for both MF and 0.9% w/v NaCl (n = 12 eyes each). The latter was used as a negative control to determine IOP fluctuations throughout the day due to circadian rhythms [48, 49]. MF was used as the benchmark for comparing changes in IOP over time as its effect is already described in clinical practice. Unloaded liposomes (B-HA-LIP) were also evaluated (n = 12 eyes). IOP measurements were taken with an iCare® TonoVet TV01 rebound tonometer fitted with original iCare® tonometer probes (V1089342/05/17, iCare Finland Oy, Vantaa, Finland). Each IOP measurement corresponded to the average of six correct single measurements and was taken by placing the tip of the probe at 4–8 mm and guaranteeing its contact with the centre of the cornea. To establish baseline IOP, two consecutive tonometric readings were performed at 30 min and immediately before instillation in each eye. After that, 25 µL of each tested formulation was gently instilled in both eyes of each rabbit. All the experiments began at the same time every day (9 am) to avoid the bias induced in IOP by circadian rhythms. Subsequently, IOP readings were logged every hour for a total of 11 h the first day, at 24, 28 and 32 h after instillation the following day, and then once per day until baseline IOP values were completely restored. Several parameters were set and calculated to analyse the results of these hypotensive studies: total IOP reduction percentage at each timepoint (ΔIOP), maximum IOP reduction percentage (IOPmax), area under the IOP curve from the beginning of the study (t0) until the time of the measurement recorded prior to restoration of baseline IOP values (t’) (AUCt0-t’), time when each formulation started to be effective (tonset) and total IOP reduction time (teffective). AUCt0-t’ was estimated using the lineal trapezoidal rule. To determine whether use of the formulations enhanced bioavailability, a 95–105% interval was determined as per bioequivalence research guidelines [50, 51].

Latanoprost determination in rabbit tear fluid

LAT concentration in rabbit tear fluid (RTF) was evaluated for LAT-HA-LIP at three different timepoints: 10, 30 and 60 min (n = 4 eyes). A volume of 25 µL of LAT-HA-LIP was topically administered in the lower conjunctival sac. RTF was collected after each timepoint using sterile diagnostic Schirmer’s test strips (T213, Contacare Ophthalmics and Diagnostics, Gujarat, India), which were positioned in the inferior eyelid for a total of 60 s and then protected from light and stored in plastic tubes at − 80 °C until processing. The same procedure was performed to determine LAT concentration in RTF after topical MF administration (n = 4 eyes per timepoint). As a negative control, RTF was collected from untreated animals (n = 6) following the same protocol. Subsequently, LAT in RTF was determined after extraction of the test strips with acetonitrile-Milli-Q® water in a 70:30% v/v ratio (500 µL in 1.5 mL plastic tubes). Afterwards, each sample was subjected to sonication (Sonorex Digiplus DL 510 H, Bandelin, Berlin, Germany) for 5 min and then centrifuged (5000 rpm, 5 min, 20 °C). Supernatants were analysed with the ultra-performance liquid chromatography (UPLC) line of a Waters® Acquity Arc Bio UHPLC device. The chromatographic conditions were the same as described in 2.4. except for injection volume, which was set at 50 µL. Seven standard dilutions were prepared from a 5 mg mL−1 LAT stock solution in acetonitrile (0.18, 0.35, 0.70, 0.88, 1.16, 1.75, 3.50 µg mL−1) to obtain the regression line (y = 28,486.45·x + 2614.28, R2 = 0.9950). LOD and LOQ were respectively established at 0.09 µg mL−1 and 0.26 µg mL−1. To estimate LAT concentration, samples volume was set at 7.5 µL, considered equal to the total RTF [52].

Statistical analysis

Each liposomal formulation was prepared in three separate batches and each batch was analysed in triplicate. Cytotoxicity data were collected from three separate experiments on three different days (biological replicates), and seven wells were tested for each sample (technical replicates) to ensure reproducibility. The results were expressed as a decrease in cell viability [%] relative to the negative control. Physicochemical, in vitro and hypotensive results were expressed as the mean ± standard deviation of the means (SD). In vivo tolerance determinations were reported as the mean ± standard error of the means (SEM). Unpaired t-tests were selected to determine statistical differences during physicochemical, in vivo tolerance, AUCt0-t’ comparisons and LAT determination in RTF assays. Two-way multivariate analysis of variance (ANOVA) using Šídák's multiple comparisons test was employed to compare results for the formulations tested in the cell viability studies. Differences between LAT-HA-LIP and MF in the hypotensive efficacy treatments were considered significant when the two-sided 95–105% confidence interval for the difference between the means of the selected parameters excluded zero [53]. GraphPad Prism® software (version 9.5.0, GraphPad Software LLC) was used for the statistical determinations. A probability value lower than 0.05 (p-value < 0.05) was considered statistically significant.