Materials and methods

Materials

Daunorubicin hydrochloride (European Pharmacopoeia reference standard) was supplied by Fagron BV (The Netherlands). The hydrophobized fumed silica particles AEROSIL® R972 and AEROSIL® R816 were supplied by Evonik Gulf FZE (Dubai, UAE), with AEROSIL® R816 being the least hydrophobic, as indicated by its carbon content. Glucidex® 9 (potato maltodextrin, DE: 9) was manufactured by Roquette (France) and provided by Azelis (Belgium). Cyclohexane (≥99.5%), d-mannitol (≥98%), sodium chloride (≥99.5%), pepsin (from porcine gastric mucosa, 3200–4500 units/mg), pancreatin (from the porcine pancreas), sodium taurocholate hydrate (≥97%), sodium phosphate monobasic (≥99.5%), Bile extract (≥97%), and Tween® 20 (polysorbate 20) were supplied by Sigma-Aldrich (Germany). Certipur® buffer solutions: pH 2 (citric acid/sodium hydroxide/hydrogen chloride), pH 4 (citric acid/sodium hydroxide/hydrogen chloride), pH 7 (di‑sodium hydrogen phosphate/potassium dihydrogen phosphate), and pH 9 (boric acid/potassium chloride/sodium hydroxide) were provided by Merck (Germany).

Preparation of daunorubicin-containing antibubbles

The inner water phase (W1) for all four primary emulsions constituted 1 mL of pH 4 buffer containing 10% of maltodextrin (2 DE) and 2% daunorubicin mixed by manual shaking in a 2 mL Eppendorf tubes (Eppendorf, Germany). The oil phase consisted of 4 mL of cyclohexane containing hydrophobized fumed silica AEROSIL® R972 particles at concentrations of 1.5%, 2.5%, or 5% (Table 1). The dispersion in the oil phase was homogenized using sonication (BANDELIN SONOPLUS UV 2200, Germany) at 30% power for 30 s to get an even mixture while also preventing evaporation of cyclohexane at the same time. The inner water phase was added to the oil phase and homogenized using ULTRA-TURRAX® at 15,000 rpm for 1 min to prepare the W1/O primary emulsions. The outer water phase consisted of 16 mL of distilled water containing 5% maltodextrin (9 DE), 15% d-mannitol, 0.5% fumed silica AEROSIL® R816, and 0.5% AEROSIL® R972 for three formulations, while one formulation included only the relatively mildly hydrophobic particles, i.e., 1% AEROSIL® R816 instead (Table 1). The dispersions of the outer aqueous phases (W2) were homogenized with ULTRA-TURRAX® at 5000 rpm and sonication at 30% power for 40 s. The process was repeated thrice until all foam disappeared, all particles were dispersed in the aqueous phase, and a slightly opaque dispersion was obtained. The double emulsions (W1/O/W2) were prepared by adding the primary emulsions into the external water phases and homogenizing using ULTRA-TURRAX® at 7000 rpm for 40 s.

Double emulsion and antibubble variants | R972 particle conc. in the oil phase | Particle type and conc. in the outer water phase
V1 | 1.5% | R816 (0.5%) + R972 (0.5%)
V2 | 2.5% | R816 (0.5%) + R972 (0.5%)
V3 | 5.0% | R816 (0.5%) + R972 (0.5%)
V4 | 2.5% | R816 (1.0%)

The emulsion samples were placed instantly in a − 80 °C ultra-freezer (BINDER GmbH, Germany). An ultra-freezer was required to freeze samples quickly to minimize creaming and preserve the internal structure of double emulsions. The frozen product was then subjected to freeze drying (Lyovapor™ L-300, BUCHI, Switzerland) for 24–30 h at −80 °C with a vacuum of 0.01 mbar. This removes the volatile oil from the middle layer and water from both the water phases and converts the structure into a glassy state. The antibubbles were obtained by rehydrating the freeze-dried mixture with a suitable solution. The overall formation process and resulting antibubble structure are schematically illustrated in Fig. 1.

Fig. 1: Schematic illustration of the formation and structure of particle-stabilized antibubbles. (a) The drug-containing inner aqueous phase (W1) is first emulsified in cyclohexane (oil phase) to form a W1/O emulsion, which is then dispersed in an external aqueous phase (W2) to obtain a W1/O/W2 double emulsion. Freeze-drying removes water and cyclohexane, yielding a dry porous matrix that forms antibubbles (W1/A/W2) upon rehydration. (b) Schematic structure of an antibubble with a hollow air-filled interior containing drug-loaded aqueous cores stabilized by silica nanoparticles.

Characterization of antibubbles

Entrapment efficiency

Entrapment efficiency (EE) of daunorubicin in the antibubbles was determined by quantifying the drug present in the external aqueous phase after rehydration and comparing it to the total drug content. The total daunorubicin concentration was measured after complete disruption of the antibubble structure induced by a surface-active agent (i.e., Tween 20). To prepare the sample, 0.25 g of freeze-dried powder was rehydrated in 10 mL distilled water that contained 10% maltodextrin solution to decrease the chance of structural breakdown due to osmotic pressure imbalances. The rehydrated sample was allowed to stand for a few minutes to enable the low-density antibubbles to rise, after which the lower aqueous phase was carefully sampled without disturbing the floating antibubble layer. An aliquot (0.5 mL) of the clear solution was withdrawn using a sterile syringe from below the antibubble foam and centrifuged at 2500 rpm for 5 min to allow any suspended particles to settle. Subsequently, 300 μL of the clear supernatant was carefully removed and mixed with 300 μL of pH 2 buffer (to ensure daunorubicin remains in protonated form). The mixture was vortexed for 10 s, and the absorbance was measured in a 96-well plate (Greiner Bio-One GmbH, Austria) at 480 nm using a spectrophotometer (Epoch microplate spectrophotometer, BioTek, USA).

To determine the maximum possible drug release, Tween 20 was added to the original antibubble suspension at a concentration of 1% of mixture, vortexed, and then allowed to stand for additional 5 to 10 min to completely destabilize the particle-stabilized interfaces. A sample from this suspension was centrifuged to obtain a clear supernatant, after which absorbance measurements were performed (as described in the above paragraph). The entrapment efficiency was calculated using the following equation:where, Cmax is the daunorubicin concentration obtained after completely disrupting the antibubbles with Tween 20. Ci is the initial daunorubicin concentration in the external aqueous phase after rehydration. The concentration of daunorubicin was quantified using a calibration curve in the range of 5–100 μg/mL, where absorbance exhibited a linear relationship with drug concentration in pH 2 buffer.

(1)EE%=Cmax−CiCmax×100

Microscopic observation and image analysis

The microscopic images of double emulsions were recorded using an optical microscope equipped with a digital camera (Olympus, Japan). The antibubbles were also observed after rehydration in a 10% maltodextrin solution. The double emulsions and antibubble suspensions were placed on a large cavity glass slide and then visualized under the microscope at 10× and 20× magnifications.

The size of the antibubbles was recorded using images taken by the optical microscope. ImageJ software (Schneider et al., 2012) was used to calculate the antibubbles' size. The size of 500 antibubbles was measured for each measurement, and then the average size and standard deviation were calculated.

Factors controlling drug release from antibubbles

Effect of pH

Freeze-dried V3 antibubble powder (0.27 g) was dispersed in pH 2, 5, and 7 buffers (each 10 mL) by gently manually shaking test tubes. Samples were collected at 0, 0.5, 1, and 2 h after dispersions were allowed to stand for 2 min and then analyzed in the same manner as described in Section 2.3.1. Since daunorubicin solubility and extinction coefficient in spectrophotometric analysis vary with pH, all samples were analyzed after adjusting pH to 2 to ensure consistent detection.

Effect of bile salts

Freeze-dried V3 antibubble powder (0.27 g) was dispersed in 10 mL of pH 7 buffer containing either 0.5% (w/v) bile extract or sodium taurocholate (NaTC) at 5 or 20 mM. These concentrations were selected to reflect physiologically relevant intestinal bile salt levels, with 5 mM representing a lower physiological level and 20 mM representing an upper fed-state level (Santos et al., 2025). The dispersions were gently mixed and allowed to settle before analysis, while samples were taken at 0, 0.5, 1, and 2 h. Daunorubicin was quantified in the samples following the detection procedure described in Section 2.3.1, ensuring consistency across all measurements.

Effect of particles' concentration on drug release from antibubbles

Different formulations with various particle concentrations at the inner interface and particle types at the outer interface were evaluated to optimize the antibubble formulation for maximum protection against drug leakage in the gastric environment. Four antibubble formulations (as already presented in Table 1) were tested in pH 2 buffer (simulating gastric pH conditions). The samples were gently mixed and allowed to settle before analysis, and the drug release was recorded at 0, 0.5, 1, and 2 h. Daunorubicin was quantified following the detection procedure described in Section 2.3.1. Since the pH of all samples was already 2, no additional pH adjustment was required before spectrophotometric analysis.

Drug release from optimized antibubble formulation under simulated gastrointestinal conditions

The release of daunorubicin from the antibubbles was determined in simulated gastric fluid (SGF) and simulated intestinal fluid (SIF). SGF was an aqueous mixture of pepsin (2 mg/mL) and NaCl (2 mg/mL), whose pH was adjusted to 1.5 using 5 M HCl. SIF was an aqueous mixture of monobasic potassium phosphate (6.8 mg/mL) and pancreatin (5 mg/mL), whose pH was adjusted to 6.8 using 1 M NaOH. Additionally, SIF was supplemented with bile extract 0.5 (wt%) (Santos et al., 2025).

The daunorubicin-loaded antibubble powder (0.33 g) was added to a conical tube (50 mL) containing 10 mL of SGF (preheated to 37 °C). The samples were capped and placed in a shaking incubator (Binder, USA), rotating at 100 rpm, and maintained at 37 °C. Aliquots of samples were withdrawn at different time points (5 min, 1, and 2 h), and their absorbance was recorded at 480 nm to quantify the released daunorubicin (Section 2.3.1). After completion of gastric phase (i.e., 2 h), the SGF mixture containing the antibubble suspension was transferred to an equal volume of SIF. After mixing SGF with SIF, the pH of the medium was readjusted to ∼6.8 using 1 M NaOH. The release of daunorubicin was quantified by measuring the absorbance at 0, 2, and 4 h after transition to SIF, where 0 h corresponds to immediately after addition to SIF. All absorbance measurements were blank-corrected using the corresponding release medium without daunorubicin, including the bile extract-containing medium.

After completion of the release experiment, the mixture was treated with Tween 20 (at a concentration of 1% of the mixture) to attain the maximum drug release. The total drug release was then determined from the absorbance, as already described. Cumulative release (CR) of the drug was calculated using Eq. 2:where, Cs and Cmax are the cumulative drug release and total drug concentration at a specific time.

(2)CR%=CsCmax×100

Preservation of encapsulated daunorubicin by antibubbles

To evaluate the protective effect of drug-loaded antibubbles, the stability of encapsulated daunorubicin was assessed at two storage conditions, pH 9 at 4 °C and pH 9.5 at 23 °C, where daunorubicin is known to be unstable (Beijnen et al., 1986; Respaud et al., 2013). Pure daunorubicin solution (10 mL each with a concentration of 100 μg/mL) at both conditions was also kept alongside to monitor the degradation of pure drug. Daunorubicin-loaded antibubbles (0.31 g each) were added to water (whose pH was adjusted to 9 or 9.5 with 1 M NaOH). Samples from free daunorubicin solutions and drug-loaded antibubble solutions (from below antibubbles cake) were collected at 0 and 24 h and analyzed following the procedure outlined in Section 2.3.1. Samples examined at 4 °C were stored in the refrigerator throughout the experiment (except during sampling) to prevent additional degradation caused by light or temperature.

Data analysis

All experiments were conducted in triplicate. Statistical analysis and graph preparation were performed using GraphPad Prism 10.0.3 (GraphPad Software, Inc., USA). Where applicable, the effects were analyzed using analysis of variance (ANOVA), and means were compared using Tukey's test. All evaluations were conducted at a 5% significance level.