Since the beginning of the project, we were able to demonstrate the ability of gas vesicles to seed to seed cavitation activity in vitro, in cellulo, and in vivo. We first demonstrated the ability of purified GVs to nucleate cavitation activity using passive acoustic detection and high frame rate microscopy at 5 million frames per second. We showed that the collapse of GVs under ultrasound pressure releases the air from these vesicles. If the ultrasound pressure is high enough, over several cycles, the nanobubbles are converted into micron-scale bubbles, which can eventually undergo violent inertial cavitation. Passive acoustic measurement also enabled us to study the physical parameters that effects this activity, including the ultrasound pressure, pulse length, frequency, and concentration of GVs.
Then, we demonstrated that molecularly-targeted GVs could serve as ultrasound-triggered disruptors of tumor cells. The outer protein of GVs GvpC was edited to include an RGD peptide, targeting them to the overexpressed surface receptors of U87 glioblastoma cells. These GVs seeded cavitation activity that opened the membranes of nearby U87 tumor cells, enabling propidium iodide dye to enter into these cells. High frame rate microscopy directly showed the formation of bubbles during ultrasound application to cells treated with GVs.
After showing the GVs can nucleate cavitation in free medium and when attached to cells, we investigated them as genetically encoded seeds for cellular inertial cavitation and payload release. We showed that GVs expressed by bacteria and mammalian cells could be used as a remote kill switch for engineered cells. Moreover, we showed that by detonating these GVs, we release co-expressed molecular payload from these cells with spatiotemporal control.
Finally, we performed three proof of concept experiments demonstrating GV-seeded cavitation and tissue disruption in vivo. GVs were directly administered into MC26 hind limb tumors under ultrasound imaging. The acoustic contrast of these GVs disappeared after exposure to focused ultrasound, while passive acoustic measurements showed significantly higher cavitation activity compared to control tumors. The ability of systemically administered GVs to damage surrounding tissue was demonstrated using the natural accumulation of GVs in the liver. Insonation in the presence of GVs was shown to produce selective tissue damage with a high number of hemorrhagic foci surrounded by necrotic regions. Finally, we endeavored to demonstrate the use of genetically encoded cavitation nuclei in the context of in vivo cell-based therapy. Mechanotherapy with GV-expressing tumor-homing cells enhanced cancer immunotherapy with checkpoint inhibitors.
