These man-made cancer-fighting molecules are infinitely variable, like a set of Legos, each one constructed to allow a patient’s immune cells to kill cancer cells. But, unlike selecting a red versus a blue block while building a colorful plastic tower, the consequences of your design choices when building what’s known as a CAR could have life-or-death consequences.
In the first comprehensive study of its kind, scientists have mapped out how a critical design choice affects how CARs, or chimeric antigen receptors, signal immune attack and how well cells carrying those CARs can eradicate cancer in mice. Their findings will inform the next generations of this form of cancer immunotherapy, called CAR T-cell therapy, the researchers say.
The study was published Tuesday in the journal Science Signaling by a cross-disciplinary research team at Fred Hutchinson Cancer Research Center and was supported by government and philanthropic funding.
“The breakthrough in CAR T cells is really how well they work in patients. But we shouldn’t rest on those successes,” said Dr. Stanley Riddell, the senior scientist on the project. “There are many patients who still don’t respond to CAR T cells, and we need to understand whether there are things we can do to overcome the barriers that prevent more patients from responding. I think this work gives us a direction to do that.”
CAR T-cell therapy sounds like science fiction: A patient with a deadly blood cancer and few other options for treatment has some immune cells, called T cells, removed from her bloodstream, genetically reprogrammed to sprout the tiny CAR, multiplied in the lab and put back into her body. There, the CARs recognize her cancer cells and activate her T cells to kill them.
But this is science fact: Two CAR T-cell therapies are now on the market in the U.S. to treat patients with advanced leukemias and lymphomas, and researchers around the world are developing new CAR T-cell therapies they hope will work better and help more patients.
This enthusiasm for CAR T-cell therapy, combined with technological advances that make it easier than ever to engineer these molecules and program them into cells, have resulted in a bit of a cart-ahead-of-the-horse situation, Riddell said.
“In part because the field exploded, the rush was to test things in the clinic. And maybe not enough attention has been paid to the basic biology: When you make a molecule like this, how does it really work?” said Riddell, who leads the Hutch’s Immunotherapy Integrated Research Center.
“No one’s really gone back and said, ‘Hey, are we really using the best foundation to build our therapy?’” added Alex Salter, a graduate student in the Riddell Lab who led the experiments.
So, for his Ph.D. project, Salter decided to help his colleagues worldwide figure that out.
CARs stretch from the outside of the T cell to the inside. The outside part recognizes a specific cancer target, and the inside part activates the T cell once that target is detected.
One of the main components of that inside part of the CAR is called a costimulatory domain. Scientists who create these engineered T-cell therapies have for years often incorporated one of two costimulatory domains — known as CD28 and 4-1BB — into their CARs as T-cell activating components. CD28, for example, is part of the Food and Drug Administration-approved CAR T-cell product Yescarta, and 4-1BB is part of the other currently approved CAR T-cell product, Kymriah.
The research team conducted laboratory tests of the two costimulatory domains within experimental CARs created in the Riddell Lab, each of which is currently being tested in Fred Hutch clinical trials.
Within minutes of a CAR detecting its target, thousands of molecular changes take place in the T cell as signals are sent to and fro and the T cell begins to activate its killing programs.
“What we wanted to know is: What are the instructions that are being delivered to the T cells, and how are those instructions informing what’s happening at the functional level?” Salter explained.
The team learned that the conventional wisdom about how these costimulatory domains work is incorrect. Whether the CAR is built with 4-1BB or CD28, they found, it largely uses the same sets of signals to trigger the T cell to act against cancer, not different sets of signals as scientists had assumed.
An important caveat
According to the project's senior scientist, Dr. Stanley Riddell, it’s not possible to use the results of this study to draw conclusions about how well other CAR T-cell products that use one component or another might work to treat a patient’s cancer.
There are many other differences in CAR design and T-cell manufacture that vary from product to product and have effects on CAR T-cell function, and each patient has a unique medical situation that affects how a therapy works for them.
Instead, the strength and speed of the signal is much more intense when the CAR is built using CD28, they discovered. This is due to previously unknown interactions between the CD28 costimulatory domain in the Fred Hutch CARs and an important T-cell signaling molecule that is not part of the CAR.
These changes in signal intensity resulted in the T cell activating more of its attack functions when it was armed with a CAR containing CD28, the team found. Pitted against cancer cells growing in a dish, the CAR T cells with CD28 multiplied and attacked more intensely than did the T cells with the 4-1BB CAR.
When the scientists gave the engineered cells at high doses to mice with lymphoma, both CAR designs eliminated the mice’s cancers. However, when the scientists gave the cells at lower doses, the T cells built with the CD28 design quickly tired out and could not hold back the tumors. Instead, it was the slower-burning CAR T cells designed with 4-1BB which, like the famously slow-and-steady tortoise, were able to keep the mice’s cancer in check for longer.
The work would not have been possible without the combined expertise of an interdisciplinary group of Fred Hutch scientists, particularly the team of Dr. Amanda Paulovich, an expert in the study of proteins, and computational biologist and biostatistician Dr. Raphael Gottardo.
The complex protein network through which a CAR activates a T cell is only fully visible with a high-powered method called mass spectrometry. Paulovich credited her lab members Richard Ivey and Jacob Kennedy for their critical role in the development, implementation and interpretation of the mass-spectrometry methods used in the study. Gottardo and team member Dr. Valentin Voillet developed computational methods to sift through the overwhelming amount of data generated by this methodology to reveal what the cells were doing.
These findings are consistent with a growing consensus in the field about how these costimulatory domain components likely affect the way CAR T cells work in patients, the researchers said.
“Our findings may help explain why the CAR T cells behave differently in patients. And I think that data is now fairly well accepted that the CD28 CAR T cells tend to expand rapidly and then don’t persist as long. The 4-1BB CARs tend to expand more slowly and persist longer,” Riddell said.
Riddell laid out one very important caveat: It’s not possible to use the results of this study to draw conclusions about how well other CAR T-cell products that use one component or another might work to treat a patient’s cancer. There are many other differences in CAR design and T-cell manufacture that vary from product to product and have effects on CAR T-cell function, and each patient has a unique medical situation that affects how a therapy works for them.
The researchers think that future studies may show that both the more-intense CD28 and the more mellow 4-1BB should each be used in specific circumstances.
A CAR built with CD28 “may be better for other types of cancers where we do want a stronger signal coming out of the receptor,” Salter explained. “So, what we’re learning from the signaling work will hopefully inform what type of CAR structures and CAR signaling domains we want in future receptors as they’re applied to different types of cancer.”
And the team is learning how to further fine-tune these precision components to dial in a particular effect. For example, Salter has created some CD28 CARs with tiny changes that calm down their signaling a bit, and he’s now testing in the lab how these changes affect the engineered cells’ function against cancers.
They hope their work will be built upon by researchers everywhere to learn more about how T cells work and to design better therapies. Collaborator Paulovich’s team, for example, is currently refining their methods for studying the T cells’ signals, crafting standardized laboratory tests that could be made widely available to scientists.
Riddell and Salter envision, one day, an entire library of engineered cell therapies, each one optimized for treating different groups of patients.
“The more we know about the receptors and how they work the more likely we are to be able to say in the future [to a patient], ‘This [therapy] will give you an 80 percent chance of responding and this other therapy may be less than that,’” Riddell said. “We’re not there yet. But I think that this is the type of thing that we have to do to get there.”
Grants from the National Institutes of Health and a gift from the Bezos family funded this research.
Susan Keown is an associate editor at Fred Hutchinson Cancer Research Center. She has written about health and research topics for a variety of research institutions, including the National Institutes of Health and the Centers for Disease Control and Prevention. Reach her at email@example.com or on Twitter @sejkeown.