Ultrasound is one of the most widely used biomedical imaging techniques. In recent years, the decades-old technology has gotten even better due to the development of contrast-enhanced ultrasound imaging (CEUS), which uses microbubbles (gas-filled lipid structures) to help scatter sound in the body, allowing physicians to not only visualize organs and soft tissues, but also peer into vasculature to image blood flow, among other things.

Researchers at the California Institute of Technology (Caltech) have now taken the potential of CEUS one step further by producing gas vesicles in the form of genetically engineered, protein-shelled nanostructures, to be used in place of microbubbles. As described recently in ACS Nano, the tiny sound-scattering structures can be engineered to bestow ultrasound with various additional functionalities, such as harmonic responses, cell-targeting specificity, and multicolored images, simply by swapping out the gas vesicles’ outer proteins like Lego pieces.

“What we have here is the first acoustic biomolecule,” says study senior author Mikhail Shapiro, an assistant professor of chemical engineering at Caltech. “And we have gained the ability to engineer its properties at the genetic level in a way that could make it useful for a variety of applications.”

Approved by the US Food and Drug Administration (FDA) for limited uses, microbubble contrast agents work by resonating in an ultrasound beam, especially in the high frequencies used for diagnostic ultrasound imaging. Though the particles are far more reflective than normal body tissues and can greatly improve standard ultrasound images, the micron-sized bubbles are inherently limited by their size and physical instability. That is, they are too large to escape from blood vessels to image other tissues in the body.

“If I asked you to make a bubble that is nanometer in size, you wouldn’t be able to do that easily because the way that bubbles work is they trap gas and surface tension holds them together,” Shapiro says. The smaller the bubble, the greater the pressure difference between the inside and outside of the bubble, ultimately increasing its instability. Even micron-sized bubbles are unstable, he says, adding that “the gas inside the bubble wants to get out and dissolve in the surrounding media.”

In 2014, Shapiro and his colleagues introduced gas vesicles as a potential new contrast agent. These particles are naturally produced by some bacteria and archaea (single-celled microorganisms with no nucleus) to control buoyancy, helping the microorganisms access light and nutrients. Unlike microbubbles, gas vesicles are stable even at the nanometer scale because their protein shells exclude water but permit gas to diffuse in and out. At the time, the researchers found that gas vesicles produced by some microorganisms have excellent ultrasound back scatter properties, making them the first protein biomolecule that can be seen with ultrasound. And given that gas vesicles are based on a “genetically encodable structure,” the team hypothesized that they could produce variants with different properties by modifying them at the genetic level.

Gas vesicles are made up of two primary proteins: Gas vesicle protein A (GvpA), which provides the main structural backbone of the nanostructure’s shell, and gas vesicle protein C (GvpC), which is found on the surface of the vesicles and give them mechanical strength and structural reinforcement. In the new study, Shapiro, along with graduate student and study first author Anupama Lakshmanan and their colleagues, targeted GvpC for genetic modification.

The team began by harvesting gas vesicles produced by the cyanobacterium Anabaena flosaquae—a simple process involving the lysing of the bacteria (rupturing their cell membranes), which allows the buoyant particles to float up for collection. They then chemically treated the gas vesicles to remove their GvpC. Using standard genetic engineering techniques, they expressed several different GvpC variants in Escherichia coli, which they subsequently added to the GvpC-free gas vesicles harvested from Anabaena. “It’s like snapping on a new Lego piece,” Shapiro says.

The researchers conducted numerous tests on their genetically engineered gas vesicles. For one, they found that gas vesicle variants could be used to get cleaner ultrasound images by helping to distinguish the imaging agents from the background signal. Most tissues in the body scatter ultrasound signals linearly, making ultrasound images full of noise. But when tested in the bodies of mice, the gas vesicles variants—particularly those stripped completely of GvpC—produced strong nonlinear signals in harmonic frequencies, which easily stand out from the background noise.

In other experiments, they showed that GvpC could be fused with different proteins and peptides to yield new properties, such as altered surface charges and the ability to target specific cells. For instance, gas vesicles that have GvpC functionalized with a peptide called arginylglycylaspartic acid or RGD, which effectively recognizes and binds to integrin proteins that are overproduced in certain tumor cells, bound more readily to glioblastoma cells (a type of brain tumor cell) than other gas vesicles in vitro.

They also found that different gas vesicle variants collapse under different acoustic pressures. These collapse pressures could, in turn, be used to generate multicolored ultrasound images. To show this, the team made three gas vesicles, each with its own collapse pressure. By slowly subjecting the vesicles to increasing ultrasound pressures, they successively collapsed each variant population. When a gas vesicle population collapses, the overall ultrasound signal decreases and this change in signal can be mapped to a specific color. Importantly, if each variant is made to target specific cells, this technique could then be used to visualize different cells in different colors.

“I think this work is extremely useful, especially as a stepping stone toward [the] expression of these nanovesicles inside cells in experimental animals,” says Sasha Klibanov, a biomedical engineer at the University of Virginia, who was not involved in the study. “What is now done with fluorescence protein labels and requires confocal microscopy with rather shallow imaging depth (or infrared camera imaging with only modest spatial resolution) will be accomplished with ultrasound imaging, which is real-time, can go centimeters deep, and provide good spatial resolution (sub-mm or better, depending on frequency of ultrasound).”

Olivier Couture, an ultrasound researcher at France’s CNRS, says that the gas vesicles could help achieve ultrasound-based molecular imaging of small animals, allowing scientists to better understand various diseases, such as cancer or atherosclerosis. “Their size should allow them to penetrate tissues, such as tumors, which are difficult to attain for conventional microbubbles,” Couture says. However, he notes that the technology is still far from broad human applications and, like other contrast agents, will probably go through a long approval process.

Shapiro and his team are currently working on a number of gas vesicle projects, such as further fine-tuning their properties, expressing gas vesicle genes in cells that do not normally make them, and showing that the nanostructures can target tumor cells in vivo. “It’s a new biological material and we’re just starting to figure out all the ways we can make it useful,” Shapiro says.

Read the abstract in ACS Nano