FROM THE AACR ANNUAL MEETING
A differently engineered chimeric antigen receptor (CAR) T cell promises to overcome major limitations of current CAR T cell therapies. Rather than engineer the CAR T cells to have a receptor that recognizes specific tumor antigens one at a time and requiring different CAR T cells for every antigen, this technique engineers a T cell receptor that can bind one invariant end of a bifunctional molecule. The molecule is constructed such that the other end can bind to whatever tumor cell surface marker is of interest. In this way, the CAR T cells can be constructed once and be directed to various tumor markers.
Standard CAR T cells are engineered to express on their surfaces receptors that recognize a specific antigen. These cells have been used up to now to recognize and kill tumor cells – for example, B cell leukemias carrying the pan-B cell marker CD19. The CAR T cells and their progeny, including memory T cells, remain in the body and continue to carry out their functions, potentially providing immune surveillance in case cancer cells arise again. But they uniquely recognize just CD19 – a problem, in that they kill even normal B cells, so-called off-target toxicity.
Beyond the unique specificity of standard CAR T cells, Philip Low, Ph.D., director of the Center for Drug Discovery at Purdue University in West Lafayette, Indiana, said these cells have three major limitations. First, they may lyse tumor cells so rapidly that a systemic tumor lysis syndrome or “cytokine storm” occurs. Second, the persisting CAR T cells can kill normal cells – for example, ones directed against CD19 killing normal B cells. Third, tumor cells have unstable genomes, leading to tumor heterogeneity, with some cells potentially losing the targeted antigens and therefore becoming “invisible” to the CAR T cells.
“So what we have done is basically designed a solution to all three, and we call it a universal CAR T cell because of its ability, with the help of an adapter molecule, to recognize all of these mutated tumor cells within a heterogeneous tumor,” he said at the annual meeting of the American Association for Cancer Research. The key was to make a CAR T cell with a surface receptor that binds to the dye fluorescein. Then fluorescein is coupled through a short linker to a molecule that binds specifically to a molecule expressed on tumor cells. In this way the CAR T cell can be made to interact with any tumor cell, depending on what is coupled to the fluorescein. The technique is analogous to a socket wrench. Every socket has the same size hole that the ratchet handle fits into regardless of the size of the “business end” of the sockets, which recognize different size nuts.
Dr. Low gave an example of folic acid, for which he says a receptor is overexpressed on about 40% of human tumors but almost never on normal cells. “We link fluorescein to the vitamin folic acid,” he said. CAR T cells are injected into an animal, and nothing happens unless a folate-fluorescein conjugate is also injected. “As soon as we inject folate-fluorescein, the folate binds to the tumor cell surface, the fluorescein part of the folate-fluorescein binds to the CAR T cell, this forces a very tight interaction between the engineered T cell and the cancer cell, and we found it leads to melting away of the tumor,” he said.
This technique addresses the three major problems with standard CAR T cell therapy. By titrating the binding affinity, concentration, and rate of administration of the fluorescein conjugate, the rate of tumor killing can be controlled, mitigating tumor lysis syndrome. Plus, normal cells may be spared if the parameters are adjusted so that the conjugate binds only to cells with high levels of the target molecule, such as tumor cells.
Because its low molecular weight, the bi-specific conjugate rapidly disappears from the circulation, and the cell killing can be terminated, allowing normal cells to regenerate – for example, in the case of normal B cells that carry CD19. Since CAR T cells generate progeny that stay in the body, the progeny remain “dormant” but are ready to be activated again by addition of the conjugate to attack tumor cells if they arise.
A major issue is dealing with tumor heterogeneity; Dr. Low’s method seems to address that, as well. “We have tumor-specific ligands for over 90% of all human cancers,” he said. “Within another couple of months we’ll have them for 100%.”