*4.1. Physicochemical Characterization*

The first part of this study revolved around the production and physicochemical characterization of DSPC-based microbubbles with cholesterol. Results indicated that the mean size of the microbubbles decreased with increasing concentrations of cholesterol. In contrast, Kaur et al. found that microbubbles with DSPC and cholesterol (1:1 molar ratio) were not significantly different in size from microbubbles with DSPC only [25]. However, those microbubbles were air-filled and did not contain any emulsifier such as PEG40 stearate or DSPE-PEG2000 like the microbubbles investigated in the present study. In our study, the span value increased with increasing concentrations of cholesterol, indicating that microbubbles with cholesterol were more polydisperse than those without cholesterol. Furthermore, the variability in polydispersity was larger between batches of microbubbles with cholesterol than those without cholesterol.

The addition of cholesterol to the indirect DSPC-based microbubble coating affected both the ligand and the lipid phase distribution. Indirect DSPC microbubbles without cholesterol had a mostly homogeneous ligand distribution as shown by fluorescence microscopy imaging, which is in agreement with results from Langeveld et al. [16]. However, all types of microbubbles with cholesterol had significantly more heterogeneous and variable ligand distribution than those without cholesterol. While the ligand distribution of microbubbles with 12 mol% cholesterol was the most homogeneous and comparable to that of the indirect DSPC microbubbles without cholesterol, indirect DSPC microbubbles with 14 mol% cholesterol had a more heterogeneous ligand distribution. The increased number of buckles in the coating is likely the reason for this increase in heterogeneity.

The indirect DSPC microbubbles without cholesterol had a lipid phase distribution similar to previous reports, with dark LC domains and a bright interdomain LE region [14,16]. All types of microbubbles with cholesterol had a significantly smaller LC phase area than those without cholesterol, indicating that cholesterol molecules modified the lateral molecular packing of the microbubble coating. The impact of cholesterol on the lipid phase distribution was most evident in microbubbles with 12 mol% cholesterol, where all components appeared to be miscible and in a single homogeneous phase. With a higher concentration of cholesterol, specifically 14 mol%, the quantified LC phase area was larger than in microbubbles with 12 mol% cholesterol. A previously reported analysis of the lipid phase behavior in binary monolayers of DPPC or DSPC with cholesterol suggested a three-state phase model [23], where cholesterol either reduced or increased the lateral molecular packing. According to that study, the lateral molecular packing of a lipid monolayer is expected to decrease with low concentrations of cholesterol and increase with higher concentrations of cholesterol. This is in agreement with our results of the lateral molecular packing, quantified here as LC phase area, decreasing up to 12 mol% and then increasing at 14 mol% cholesterol. Other work focused on lipid phase behavior in monolayers includes atomic force microscopy images of monolayers with DPPC and 33 mol% cholesterol, showing a homogeneous phase distribution [22]. While we found microbubbles with 32 mol% cholesterol to be highly unstable, those with 12 mol% had a homogeneous phase distribution. This suggests that the phase behavior of phospholipids in a monolayer cannot be directly translated to the phase behavior of phospholipids in a microbubble coating, which is supported by a direct comparison of lipid phase behavior in monolayers and microbubble coatings with the same ternary mixture of DPPC or DSPC with DSPE-PEG2000 and PEG40-stearate [16].

Interestingly, cholesterol (10–50 mol%) has been used for many years to stabilize liposomes with DPPC or DSPC by increasing the lateral molecular packing [34], emphasizing the difference in lamellar structures, i.e., bilayers, of a liposome compared to the

phospholipid monolayer coating of a microbubble. DSPC forms lamellar structures when suspended in water at room temperature [35]. However, during microbubble production by probe sonication, the lamellar structures are disrupted, and the molecules self-assemble as a monolayer of phospholipids at the gas–liquid interface [36]. In a model membrane system with monolayer-bilayer junctions, cholesterol was shown to be involved in lipiddriven budding of the membrane, with higher concentrations of cholesterol resulting in increased budding [37]. These findings are in agreement with the increased budding and formation of buckles we found in microbubbles with higher concentrations of cholesterol in the coating. In this context, budding refers to the formation of lipid bilayer-coated vesicles, while buckle formation refers to bilayers that are still attached to the lipid monolayer coating of the microbubble.

The present study includes a normalization factor in the analysis of the 4Pi microscopy data to compensate for a difference in fluorescence intensity between the middle and the top or bottom of the microbubbles. The normalization factor did not affect the proper quantification of the LC phase area in microbubbles without cholesterol. Since the difference in fluorescence signal between LC and LE phase in those microbubbles was much larger than the difference in signal between the middle and top or bottom of the microbubble, the LC phase area could easily be quantified in microbubbles with clear separation of the lipids into LC and LE phase. The imaging artifact only became evident during the analysis of microbubbles with a homogeneous lipid phase distribution, i.e., containing cholesterol. All experiments in this study were performed at a room temperature of 19–21 ◦C. Since the 4Pi confocal microscope operates at a limited range of temperature, this practice facilitated comparison of the data obtained from the 4Pi confocal microscopy and the acoustic characterization with the combined confocal and Brandaris 128 system and was in accordance with previous microscopy studies on lipid and ligand distribution in microbubble coatings [14,16]. Slight fluctuations in the temperature of the sample due to, for instance, the light or ultrasound are not expected to affect the lipid phase distribution, since the transition temperature for DSPC is 55 ◦C [38]. Furthermore, it was previously reported that in lipid bilayers of DPPC and cholesterol (10 or 20 mol%), the lipid phase distribution was only affected by temperatures above 40 ◦C [39]. Processing of lipid films in the sonicator bath and with the probe sonicator at power 3 did not affect the temperature of the samples.
