“CAR-T cells are already used today to treat a number of blood cancers, but solid tumors continue to pose significant challenges for this mode of therapy in terms of both safety and efficacy,” said Irving. “We have potentially addressed both these issues by engineering, directly into the CAR design, on and off switches that are engaged by drugs that have been approved by regulatory agencies and are already in use in the clinic. This should hasten the advancement of these remotely controlled CAR-T cells into clinical trials.”

CAR-T cells express synthetic receptors that detect specific molecular markers, or antigens, on cancer cells using an antibody fragment as an external sensor. When the CAR binds its antigen on a cancer cell, its signaling modules are activated and trigger the T cell’s innate cytotoxic weaponry to destroy tumor cells.

The trouble is that many solid tumor antigens are found on healthy cells as well, raising the risks of so-called “off-tumor, on-target” effects. These can provoke destructive immune responses that are difficult to control and potentially lethal to the patient. Conversely — and perhaps more often — the immunosuppressive conditions of the solid tumor microenvironment can push anti-tumor T cells, including those equipped with CARs, into a state of dysfunction known as “exhaustion.”

“Having the ability to remotely switch CAR-T cells on to varying degrees using different doses of an activating drug — and then off on demand, as needed — would improve the safety of this therapy,” said Giordano Attianese. “Further, the remote control of CAR-T cell activity could also be used to mitigate T cell exhaustion, improving the durability of patient responses to the therapy.”

The latter possibility stems from studies — led by Ludwig Stanford’s Crystal Mackall — showing that giving CAR-T cells periods of rest between bouts of active tumor targeting rewrites their gene-expression programs, reverses exhaustion and broadly boosts their functional efficacy.

Classical CARs are single chain receptors that directly link binding to the tumor antigen to internal signaling components derived from the functional modules, or “domains,” of a few key proteins from T cells. The antigen-sensor of the CAR, which sticks out of the engineered T cell like a wand, is usually derived from the antigen-binding fragment of an antibody molecule, which can be developed to detect virtually any target on tumor cells with exquisite specificity. The internal signaling components usually come from a protein called CD3-ζ, which is absolutely required to activate the T cell upon antigen binding, and another borrowed from a “co-stimulatory” protein (like 4-1BB or CD28) that boosts the function and persistence of T cells after activation.

To enable control of CAR activity, Irving, Giordano Attianese and colleagues separated the antigen-sensing moiety (the antibody fragment) and the activation domain (CD3-ζ) on two separate chains, the ‘receptor chain’ and the ‘signaling chain’. Tapping the expertise of Bruno Correia of the École Polytechnique Fédérale de Lausanne (EPFL), they also included an extra module able to dimerize the two chains upon application of a cancer drug called venetoclax.

When it binds those external modules, the venetoclax molecule acts like a bridge, bringing the two chains together to create an active CAR complex — and the intensity of the subsequent CAR-T cell response depends on how much of the drug is used. The researchers named this CAR construct the “inducible-ON” (iON) CAR.

To be truly safe, however, the CAR-T cells also need to be switched off promptly if they pose a danger to patients. To that end, the researchers added an additional druggable component onto the CD3- ζ signaling chain that is responsive to another approved cancer drug named lenalidomide. This binding, however, marks the receptor for degradation by the cell’s waste-disposal machinery. The researchers show that the all-in-one iON/OFF CAR (iONØ-CAR) T cells can be switched on by venetoclax and quickly deactivated — within 4-6 hours — by lenalidomide.

The researchers next plan to better characterize the performance of their iON and iONØ-CARs against different tumor models. They will also test whether remote control of the cells indeed prevents toxicity of overactive CAR-T responses and if periodic bouts of rest can improve the long-term control of tumors.

Melita Irving leads the T cell engineering group in Ludwig Lausanne’s Human integrated tumor immunology discovery engine (Hi-TIDe).

The Irving lab research was supported by Ludwig Cancer Research, the Swiss National Science Foundation, the ISREC Foundation, the Prostate Cancer Foundation and the Fondazione Teofilo Rossi di Montelera e di Premuda.