**4. Discussion**

The results of the present study showed massive spontaneous in situ formation of liposomes from a porous inorganic drug carrier combined with phospholipids. This effect was observed in media simulating the conditions of the different regions of the GI tract. As a consequence, this effect led to dissolution rate enhancement of a model low soluble drug.

A successful drug loading in the internal pores of the carrier is essential for effective porous structure exploitation. In this study, the loading efficiency was visually controlled. Absence of external deposition of drug crystals and phospholipid material was confirmed by visual image analysis for every formulation (excluding physical mixtures), as shown in Figure 4. The results suggest that the phospholipid and nifedipine were loaded in the internal structure of the porous FCC particles. This result was corroborated by FIB cross-sectioning of the consolidated and loaded FCC, as shown in Figure 5. By visual analysis of internal porosity reduction, as clearly seen on the FIB-SEM images (Figure 5d), we could confirm the loading of drug and phospholipid into internal carrier structures. The results of BET measurements (Figure 6) showed a reduction in specific surface area, which further confirmed the hypothesis of internal (i.e., intraparticulate) deposition of the loaded material.

Both calcium carbonate and calcium phosphate, the main constituents of FCC, readily dissolve in the low pH of the gastric medium, which triggers an immediate release of liposomes (Video S1). On the other hand, in neutral pH media, the formation of liposomes from FCC particle followed a different mechanism. Since the carrier was not immediately dissolved in the neutral pH, the liquid was taken up into the particles by the carrier's micro-capillaries. This observation suggests a capillary pump action for the porous microstructure of FCC. The capillary pump effect supposedly drives the hydrated phospholipids out of the pores and promotes self-assembly of liposomes. This process can be seen in the confocal microscopy images, revealing the liposomes emerging (budding) out of the pores of FCC (Figure 1).

In SGF (i.e., medium with a low pH), a significantly large population of the liposomes had a size of 121 ± 36 nm and were observed as uni-lamellar structures in cryo-EM images. This size range was possibly due to fast release of liposomes from FCC, aided by the eddy currents created as a result of CO<sup>2</sup> formation from calcium carbonate dissolution in the acidic pH. The liposomes formed in this process were able to encapsulate up to 25% of the nifedipine, resulting in a faster dissolution of the drug compared to the drug release rate from the formulation without phospholipids. The different liposome formation mechanism in neutral pH produced large multi-lamellar liposomes (490 ± 112 nm) as seen in the cryo-EM images. These liposomes had a lower encapsulation efficiency, as compared to the uni-lamellar liposomes in SGF. This finding was surprising since the multi-lamellar liposomes were expected to have a higher encapsulation efficiency for hydrophobic drugs compared to uni-lamellar liposomes. The low encapsulation efficiency in neutral media was explained by the slow rate of formation and release of the liposomes from FCC particles, which allowed enough time for the drug to escape from the liposomes before their detachment from the carrier. This assumption was further confirmed by the results that showed a similar rate of dissolution of nifedipine from FCC phospholipid formulations and the Nif-FCC (the formulation without phospholipid) in neutral pH. In other words, in neutral pH, nifedipine dissolution takes place independently of the liposome formation, and the dissolution rate is enhanced solely by the large specific surface area of the carrier.

Due to the surface-active properties of phospholipids, faster dissolution rates are expected from formulations with higher DMPC content. Despite the latter, an increased phospholipid content resulted in slower dissolution rates in SGF (Figures 2 and 3). The higher phospholipid amounts in FCC reduce the overall surface area available for drug release by blocking out the carrier's porous meshwork, which leads to a reduction in the surface available for dissolution. This effect was confirmed by the results of BET surface area experiments, where a reduction of specific surface area with an increase in the phospholipid content was measured. For example, the D5N20 sample with 5% phospholipid content had a threefold larger specific surface area and an almost sevenfold faster dissolution as compared to D20N20.

The importance of the available medium-to-solid contact surface is emphasized since FCC itself dissolves in the low pH dissolution medium. This is forcing single particles to disintegrate, which leads to a faster release of the liposomes that can better encapsulate the drug (i.e., not letting sufficient time for the drug to escape the forming liposomes) and results in a faster rate of dissolution. Additions of more phospholipid lead to a slower wetting, confirmed by the water sorption studies, and therefore slower reaction of FCC with the medium. This makes the drug release dependent on the diffusion of the drug from the carrier's pores. A complete dissolution of calcium carbonate in SGF results in disappearance of the carrier particles; therefore, the only microclimate pH change may be expected within the formed liposomes. In theory, a change in the dissolution kinetics of the drug is expected to be negligible due to close proximity of undissolved drug crystals to the lipophilic layer of the phospholipids.

The high surface area of FCC itself effectively exposes the deposited drug to the medium and, even without phospholipids, leads to a faster dissolution of nifedipine from the FCC formulations compared to the physical mixture (Figure 3, Nif-FCC vs. physical mixture).

The X-ray diffraction and DSC results did not exclude the fact that the nifedipine might be partially deposited in a non-crystalline form in the carrier. The DSC thermograms of FCC formulations showed a reduced melting enthalpy, which was possibly a result of partial amorphization of the drug (Figure S2). Moreover, the X-ray diffractograms clearly indicated the crystalline state of nifedipine in all formulations (Figure 7); therefore, an enhanced dissolution cannot be solely attributed to the drug amorphization by phospholipids.

The wetting constant and the specific surface area depend on each other according to Equation (5), which represents the power law function (Figure 6). The reduction of the individual area of a pore's cross section depends on squared diameter of remaining pore opening as follows from simple geometric considerations. Thus, under an assumption of constant rate of capillary transport, there is a quadratic dependency of the porous cross-section area on the rate of material deposition in a material with idealized equisized pore geometries. However, for different pore size distributions in real materials, the power exponent is expected to be significantly different from two. Therefore, due to loading of materials into the FCC particles, the smaller pores are filled first and the total liquid throughput (i.e., water uptake) through the particle meshwork changes according to power law with exponents greater than two. The applicability of the power law correlation can be extended to the dissolution rate and liposomal encapsulation efficiency as a function of the phospholipid content in the FCC-based formulations. The latter suggests a guidance for the design or section of porous carriers for such formulations.

On the basis of the results, we could show that the combination of porous FCC microparticles and phospholipids can be used to produce solid formulations oral liposomal delivery. This method is suitable for classical scale-up. We demonstrated that the fast water uptake and dissolution of the carrier is essential for achieving the highest encapsulation efficiency and the fastest drug dissolution rate.
