Researchers have been building simple molecular machines out of DNA for some years now. This approach to molecular machinery is well suited to applications that involve conditional activation based on the proteins present in the surrounding environment; a lot of the necessary functional parts already exist in DNA and just have to be assembled in the right way. The Oisin Biotechnologies cell-killing technology is a smaller example of the type than the approach here, in which sizable DNA containers are constructed. They carry a cargo that will disrupt local blood flow, and are triggered into opening by cancerous cell surface proteins, thereby sabotaging the nutrient supply to tumors without harming other tissues. It is all quite clever and quite mechanical.
DNA nanorobots that travel the bloodstream, find tumors, and dispense a protein that causes blood clotting can trigger the death of cancer cells in mice. The researchers started with the goal of finding a path to design nanorobots that can be applied to treatment of cancer in humans. They first generated a self-assembling, rectangular, DNA-origami sheet to which they linked thrombin, an enzyme responsible for blood clotting. Then, they used DNA fasteners to join the long edges of the rectangle, resulting in a tubular nanorobot with thrombin on the inside. The authors designed the fasteners to dissociate when they bind nucleolin - a protein specific to the surface of tumor blood-vessel cells - at which point, the tube opens and exposes its cargo.
The scientists next injected the nanorobots intravenously into nude mice with human breast cancer tumors. The robots grabbed onto vascular cells at tumor sites and caused extensive blood clots in the tumors' vessels within 48 hours, but did not cause clotting elsewhere in the animals' bodies. These blood clots led to tumor-cell necrosis, resulting in smaller tumors and a better chance for survival compared to control mice. The team also found that nanorobot treatment increased survival and led to smaller tumors in a mouse model of melanoma, and in mice with xenografts of human ovarian cancer cells. The next step is to investigate any damage - such as undetected clots or immune-system responses - in the host organism, as well as to determine how much thrombin is actually delivered at the tumor sites. The authors showed in the study that the nanorobots didn't cause clotting in major tissues in miniature pigs, which satisfies some safety concerns, but more work is needed.