Dewetting: From Physics to the Biology of Intoxicated Cells

Dependence of endothelial drug delivery on monodisperse microbubble dynamics

Ultrasound insonification of microbubbles has been shown to increase vascular permeability to locally enhance drug delivery. For more effective and controllable therapeutic outcomes, uniform acoustic responses from microbubbles and a deeper understanding of the biophysical mechanisms of drug delivery are critical. In this study, we investigated the impact of monodisperse microbubble size and their dynamics on cellular responses and drug delivery outcomes in endothelial cells. Monodisperse microbubbles with radii of 1.5, 2.2, 2.7, and 2.9 μm were produced using a microfluidic flow-focusing device. Upon insonification (2 MHz, 220 kPa peak negative pressure, 10 cycles), the microbubble oscillation was captured in real-time at 10 million frames per second using ultra-high-speed imaging, while confocal microscopy was employed to observe cellular responses in both 2D and 3D. For the 65 microbubbles studied, the 2.2 μm microbubbles, i.e., corresponding to the resonant radius at 2 MHz ultrasound, exhibited the highest sonoporation rate (75 %), induced the largest membrane perforations (a median value of 78 μm2) and highest intracellular drug uptake. The 1.5 μm microbubbles achieved a comparable sonoporation rate (73 %), yet with significantly smaller membrane perforations (a median value of 20 μm2), lower intracellular drug uptake, and highest occurrence of transendothelial drug delivery pathways (64 %). Mechanistically, microbubble-generated shear stress was identified as the significant factor driving sonoporation, while normal stress did not show significance. In conclusion, our study highlights the importance of carefully selecting the microbubbles size to maximize microbubble-mediated drug delivery outcomes and facilitate safe translation of monodisperse microbubbles into clinical practice.

Characterizing Microbubble-Mediated Permeabilization in a Vessel-on-a-Chip Model

Drug transport from blood to extravascular tissue can locally be achieved by increasing the vascular permeability through ultrasound-activated microbubbles. However, the mechanism remains unknown, including whether short and long cycles of ultrasound induce the same onset rate, spatial distribution, and amount of vascular permeability increase. Accurate models are necessary for insights into the mechanism so a microvessel-on-a-chip is developed with a membrane-free extravascular space. Using these microvessels-on-a-chip, distinct differences between 2 MHz ultrasound treatments are shown with 10 or 1000 cycles. The onset rate is slower for 10 than 1000 cycles, while both cycle lengths increase the permeability in spot-wise patterns without affecting microvessel viability. Significantly less vascular permeability increase and sonoporation are induced for 10 versus 1000 cycles at 750 kPa (i.e., the highest studied peak negative acoustic pressure (PNP)). The PNP threshold for vascular permeability increases is 750 versus 550 kPa for 10 versus 1000 cycles, while this is 750 versus 220 kPa for sonoporation. Vascular permeability increases do not correlate with αvβ3-targeted microbubble behavior, while sonoporation correlates with αvβ3-targeted microbubble clustering. In conclusion, the further mechanistic unraveling of vascular permeability increase by ultrasound-activated microbubbles in a developed microvessel-on-a-chip model aids the safe and efficient development of microbubble-mediated drug transport.

Ultrasound-activated microbubbles mediate F-actin disruptions and endothelial gap formation during sonoporation

Locally opening up the endothelial barrier in a safe and controlled way is beneficial for drug delivery into the extravascular tissue. Although ultrasound-induced microbubble oscillations can affect the endothelial barrier integrity, the mechanism remains unknown. Here we uncover a new role for F-actin in microbubble-mediated endothelial gap formation. Unique simultaneous high-resolution confocal microscopy and ultra-high-speed camera imaging (10 million frames per second) reveal that single oscillating microbubbles (radius 1.3-3.8 μm; n = 48) induce sonoporation in all cells in which F-actin remodeling occurred. F-actin disruption only mainly resulted in tunnel formation (75 %), while F-actin stress fiber severing and recoil mainly resulted in cell-cell contact opening within 15 s upon treatment (54 %) and tunnel formation (15 %). F-actin stress fiber severing occurred when the fibers were within reach of the microbubble's maximum radius during oscillation, requiring normal forces of ≥230 nN. In the absence of F-actin stress fibers, oscillating microbubbles induced F-actin remodeling but no cell-cell contact opening. Together, these findings reveal a novel mechanism of microbubble-mediated transendothelial drug delivery, which associates with the underlying cytoskeletal F-actin organization.

Caveolin-1 protects endothelial cells from extensive expansion of transcellular tunnel by stiffening the plasma membrane

Large transcellular pores elicited by bacterial mono-ADP-ribosyltransferase (mART) exotoxins inhibiting the small RhoA GTPase compromise the endothelial barrier. Recent advances in biophysical modeling point toward membrane tension and bending rigidity as the minimal set of mechanical parameters determining the nucleation and maximal size of transendothelial cell macroaperture (TEM) tunnels induced by bacterial RhoA-targeting mART exotoxins. We report that cellular depletion of caveolin-1, the membrane-embedded building block of caveolae, and depletion of cavin-1, the master regulator of caveolae invaginations, increase the number of TEMs per cell. The enhanced occurrence of TEM nucleation events correlates with a reduction in cell height due to the increase in cell spreading and decrease in cell volume, which, together with the disruption of RhoA-driven F-actin meshwork, favor membrane apposition for TEM nucleation. Strikingly, caveolin-1 specifically controls the opening speed of TEMs, leading to their dramatic 5.4-fold larger widening. Consistent with the increase in TEM density and width in siCAV1 cells, we record a higher lethality in CAV1 KO mice subjected to a catalytically active mART exotoxin targeting RhoA during staphylococcal bloodstream infection. Combined theoretical modeling with independent biophysical measurements of plasma membrane bending rigidity points toward a specific contribution of caveolin-1 to membrane stiffening in addition to the role of cavin-1/caveolin-1-dependent caveolae in the control of membrane tension homeostasis.