Using new machine learning techniques, researchers at UC San Francisco (UCSF), together with a team from IBM Research, have developed a virtual molecular library of thousands of “command phrases” for cells, based on combinations of “words” that have guided engineering immune cells to tirelessly seek out and kill cancer cells.
The book, published online December 8, 2022, in Sciencerepresents the first time that such sophisticated computational approaches have been applied to a field that, until now, has largely advanced through tinkering and ad hoc engineering cells with existing rather than synthesized molecules.
This breakthrough allows scientists to predict what elements – natural or synthesized – they need to include in a cell to give it the precise behaviors needed to respond effectively to complex diseases.
“This is a vital change for the field,” said Wendell Lim, Ph.D., Byers Professor Emeritus of Cellular and Molecular Pharmacology, who heads the UCSF Cell Design Institute and led the study. “Only by having this predictive power can we get to a place where we can rapidly design new cell therapies that achieve the desired activities.”
Meet the molecular words that make cellular command phrases
A large part of therapeutic cell engineering involves choosing or creating receptors which, when added to the cell, will allow it to perform a new function. Receptors are molecules that link the cell membrane to sense the external environment and provide the cell with instructions on how to respond to environmental conditions.
Putting the right receptor into a type of immune cell called a T cell can reprogram it to recognize and kill cancer cells. These so-called chimeric antigen receptors (CARs) have been effective against some cancers but not others.
Lim and lead author Kyle Daniels, Ph.D., a researcher in Lim’s lab, focused on the part of a receptor inside the cell that contains strings of amino acidscalled patterns. Each pattern acts as a command “word”, directing an action within the cell. How these words are joined together in a “sentence” determines the commands the cell will execute.
Many of today’s CAR-T cells are designed with receivers ordering them to kill the cancer, but also to pause after a short time, which is like saying, “Kill out a few rogue cells and then breathe.” As a result, cancers may continue to grow.
The team believed that by combining these “words” in different ways, they could generate a receptor that would allow CAR-T cells to complete the job without taking a break. They created a library of nearly 2,400 randomly combined command phrases and tested hundreds of them in T cells to see their effectiveness in fighting leukemia.
What the grammar of cellular commands can reveal about the treatment of diseases
Next, Daniels teamed up with computational biologist Simone Bianco, Ph.D., research director at the IBM Almaden Research Center at the time of the study and now director of computational biology at Altos Labs. Bianco and his team, researchers Sara Capponi, Ph.D., also at IBM Almeden, and Shangying Wang, Ph.D., who was then a postdoctoral fellow at IBM and is now at Altos Labs, applied new learning methods automatically to the data to generate entirely new receptor phrases that they believe would be more effective.
“We changed some of the words in the sentence and gave it new meaning,” Daniels said. “We predictively engineered cancer-killing T-cells without pausing because the new phrase told them, ‘Kill out those rogue tumor cells and keep going.’ “”
The combination of machine learning and cellular engineering creates a new synergistic research paradigm.
“The whole is definitely greater than the sum of its parts,” Bianco said. “It allows us to have a clearer picture not only of how to design cell therapies, but also to better understand the rules that underlie life itself and how living things do what they do. ”
Given the success of the work, Capponi added, “We will extend this approach to a diverse set of experimental data and hopefully redefine T-cell design.”
Researchers believe this approach will lead to cell therapies for autoimmunity, regenerative medicine and other applications. Daniels is interested in designing self-renewing stem cells to eliminate the need for blood donation.
He said the real power of the computational approach extends beyond creating command sentences, to understanding the grammar of molecular instructions.
“It’s the key to making cell therapies that do exactly what we want them to do,” Daniels said. “This approach facilitates the transition from understanding science to engineering to its application in real life.”
Kyle G. Daniels et al, Decoding the CAR T Cell Phenotype Using Combinatorial Signaling and Machine Learning Pattern Libraries, Science (2022). DOI: 10.1126/science.abq0225. www.science.org/doi/10.1126/science.abq0225
University of California, San Francisco
Quote: How AI Found the Words to Kill Cancer Cells (December 8, 2022) Retrieved December 8, 2022 from https://phys.org/news/2022-12-ai-words-cancer-cells.html
This document is subject to copyright. Except for fair use for purposes of private study or research, no part may be reproduced without written permission. The content is provided for information only.