The age of molecular-scale computing is entering a new era, thanks to the development of a system that uses synthetic proteins and Boolean logic to identify cancer cells.
The proteins can lock onto chemical markers on the surface of cells in predetermined combinations, performing the roles of logical AND, OR and NOT gates. It’s similar to the way binary computers do their thing, but with biochemistry rather than electronic bits.
“We were trying to solve a key problem in medicine, which is how to target specific cells in a complex environment,” Marc Lajoie, one of the lead authors of a study published today in the journal Science, explained in a news release.
“Unfortunately, most cells lack a single surface marker that is unique to just them. So, to improve cell targeting, we created a way to direct almost any biological function to any cell by going after combinations of cell surface markers,” Lajoie said.
Lajoie worked on the effort during his stint as a postdoctoral scholar at the University of Washington’s Institute for Protein Design. He’s now co-director for protein and cell engineering at Lyell Immunopharma, a California-based startup aiming to commercialize the technique.
Co-authors of the study include researchers from UW and Seattle’s Fred Hutchinson Cancer Research Institute, some of whom are also affiliated with Lyell.
The technique takes advantage of protein engineering, a field that’s been pioneered at the Institute for Protein Design. Proteins interact with cells based on their molecular structures. A protein can lock onto a cell if its structure meshes with a corresponding structure on the cell’s surface, much as a key meshes with a lock. The virus that causes COVID-19, for example, uses a spiky molecular structure to latch onto the cells that are its victims.
Lots of cells share similar structures, however. A specific type of cancer cell may not have one unique and accessible target structure to lock onto. That can be a problem for the targeted therapies that make use of the body’s own killer T cells to attack cancer cells.
The newly developed approach makes use of a molecular tool known as Co-LOCKR to get around the problem. “Co-LOCKR” is a mouthful of an acronym, standing for “Colocalization-dependent Latching Orthogonal Cage/Key pRoteins.” In simpler terms, the tool is a multiple-key locking system or a two-factor authentication system that targets the intended cells while avoiding off-target effects.
Co-LOCKR proteins can be programmed to execute logical functions. For example, they can serve as an AND gate, setting off a molecular beacon only if two surface markers are present. A different configuration can act as an OR gate, lighting up the beacon if either one of two markers is present. And yet another configuration works like a NOT gate, ensuring that the beacon isn’t activated if a given protein is present.
Just as in a computer, the logical gates can be combined — for example, to light up the beacon if marker A is present in combination with either B or C, but never if D is also present.
When the beacon (actually, a type of molecule known as a peptide) is activated, that would signal killer T cells to go after the cancer cell that’s targeted. Meanwhile, the killer cells ignore healthy cells that aren’t flashing the beacon of doom.
In a series of experiments, Lajoie and his colleagues mixed Co-LOCKR proteins and genetically engineered T cells (also known as chimeric antigen receptor T cells, or CAR-T cells) with a soup of potential target cells. Only the cells with the desired combinations of surface markers ended up being killed, demonstrating that the technique works in the lab.
There are other protein-based techniques that use if-then logic to distinguish bad cells from good cells. But in an email to GeekWire, Lajoie said Co-LOCKR works on the scale of minutes rather than hours, with single-cell resolution.
And although cell-based cancer treatment is an obvious application, it’s not the only one. “Co-LOCKR can be used for any application that would benefit from proximity-dependent detection, including intracellular targets such as DNA sequences or protein complexes,” Lajoie told GeekWire.
Lajoie’s fellow co-director at Lyell, Scott Boyken, said the potential applications even extend beyond medicine. “The ability to turn on activity based on proximity is a substantial advance for de novo protein design and opens multiple doors for programming biological functions,” Boyken said via email.
Co-LOCKR has its limitations. For example, the NOT gate won’t work as intended if there aren’t enough of the cellular markers to set it off. “You need to have information ahead of time about which ‘healthy’ antigens are highly expressed on the off-target tissue and not on the target tissue,” Lajoie said.
But assuming the limitations can be overcome, the technique seems likely to take a prominent place in the standard biomedical toolbox.
“We believe Co-LOCKR will be useful in many areas where precise cell targeting is needed, including immunotherapy and gene therapy,” said UW biochemist David Baker, director of the Institute for Protein Design.
About the study: Lajoie, Boyken and Fred Hutch researcher Alexander Salter are principal authors of the study published by Science, “Designed Protein Logic to Target Cells With Precise Combinations of Surface Antigens.” Co-authors include Jilliane Bruffey, Anusha Rajan, Robert Langan, Audrey Olshefsky, Vishaka Muhunthan, Matthew Bick, Mesfin Gewe, Alfredo Quijano-Rubio, JayLee Johnson, Garreck Lenz, Alisha Nguyen, Suzie Pun, Colin Correnti, Stanley Riddell and David Baker.
About the Co-LOCKR illustration: Here’s Lajoie’s technical explanation for what’s going on: “There are three types of cells depicted: orange has antigen A, blue has antigen B, and green has both antigens A and B. The ‘Cage’ can bind to antigen A on the orange and green cells. The ‘Key’ can bind to antigen B on the blue and green cells. Only the green cells result in Co-LOCKR activation as a result of co-localizing the Cage and Key components of the system.
“The ‘molecular beacon’ is the activated Co-LOCKR system: Cage and Key co-localization results in activation of the Bim peptide (part of the Cage protein), which acts as the beacon. Then the Bcl2 Effector protein is able to bind to the activated Bim peptide. We showed in the paper that the Effector protein could be a small-molecule-tagged protein or a chimeric antigen receptor (CAR) on a T cell, but there are many other possible options as well.”
