Will Programmable Immune “Cellbots” Be the Next Weapon Against Cancer?

Xconomy San Francisco — 

[Corrected, 9/29/16, 4:24 p.m. See below.] Cancer immunotherapy has captured the medical world’s attention, knocking out a few types of cancer in a small portion of the most desperate patients. But it has bumped into many limitations.

Researcher Wendell Lim (pictured) and his lab mates at the University of California, San Francisco, believe they can improve upon early immunotherapies by rewiring a key component of the immune system: the roving attack dogs called T cells that sniff out and kill intruders. They also think their idea could work beyond cancer on other diseases, as well.

The Lim Lab’s latest paper, published Thursday in the journal Cell, describes how they reprogrammed T cells to do two things: sense when they’re near a tumor with particular markings on its surface, and produce drugs on the spot to kill only the tumor cells. The work is based in part on the way cells communicate with each other in group settings to carry out complex tasks, like forming a limb or an organ.

Lim calls programmable T cells “cellbots.” Several devices used to simulate cells in communication sit on his office table like high-tech tea candles, blinking at each other. Lim, wearing a collared sweater, shorts, and worn leather sandals, fiddles with them as he talks.

“Cancer is a recognition problem on one level. Recognition by our immune system, or by drugs. How do we solve problems in voice recognition, or face recognition? We have algorithms that deconstruct the information, whether it’s frequencies or facial features. That’s actually how immune cells work in their native way,” he says—integrating information from many sources to decide, for example, when to move from the blood stream into a different part of the body. “If we can break down how to use different sensors in cells to achieve more sophisticated recognition and response and algorithms, I think it’s going to be a big difference.”

Changing a patient’s own T cells to become better cancer fighters is not entirely radical. Several publicly traded companies, including the healthcare giant Novartis (NYSE: NVS), hope to have such treatments approved in the next couple years. This first wave of T cell therapies (called CAR-T for “chimeric antigen receptor T cells”) have had great results in a small subset of blood-borne cancers. The latest results came Monday, from Kite Pharma (NASDAQ: KITE) of Santa Monica, CA, which recently became the first industry partner of Cell Design Labs, the San Francisco-based company Lim has co-founded to turn his lab’s work into medicine.

Lim and company want to go further than current CAR-T therapies, not just tinkering with a part of a gene here and there, but making genetic manipulations that alter fundamental principles of T cells: how they gather information about their surroundings and react to what’s around them. Lim says an effort akin to building a computer program is needed—what’s often called synthetic biology. The field of T cell therapy to date has “stumbled into a bunch of ad-hoc solutions,” Lim says. “We’d like to have a more uniform, systematic kind of programming logic and language.”

Earlier this year, Lim and colleagues published their first work on what they call “synNotch” cells—named so because the researchers have redesigned the Notch receptor, which extends outside the cell like a powerful nose that sniffs out its surroundings. The Notch redesign also includes parts of the cell’s internal machinery that responds to signals by, for example, producing a toxin to kill invaders.

In the earlier papers, the Lim Lab introduced the synNotch cells and showed that they can get to a tumor site in a mouse and, like current T cell therapies, latch onto a protein and kill tumor cells—without latching onto the same protein in other tissues. It’s intriguing because current T cell therapies will kill any cells, healthy or cancerous, that display the protein they’re trained to go after. Even the successful therapies, to treat types of leukemia and lymphoma, also kill healthy immune cells that share the tumor target, though the side effects of that are generally manageable. The danger is far worse going after solid tumors, like breast, lung, and colon cancer, which share a lot of target proteins with healthy tissue. T cell therapies have not had much success there. Her2, a protein abundant in many cancers, is also present in healthy lung tissue, and a National Cancer Institute test several years ago ended in a patient death when her modified T cells attacked her lungs. SynNotch cells are designed to avoid such problems by recognizing two targets instead of one.

In the paper published today, Lim and colleagues have shown the possibility of pinpoint targeting, but also using the synNotch cells as on-the-spot drug factories that might be able to make all kinds of drugs, not just the stuff that T cells naturally produce. The paper describes an experiment with a mouse that has one kind of tumor on its left side, another kind on its right side. The synNotch cells are injected into the mouse’s tail. Twenty five days later, the tumor the cells have been programmed to recognize has disappeared. The other tumor, similar but missing a key marker on its surface, continued to grow. Lim’s lab have induced the cells to produce well-known cancer drugs like pembrolizumab (Keytruda), blinatumumab (Blincyto), as well as molecules that have been tried as cancer drugs but are too poisonous when injected into the bloodstream. Delivering them at short range via synNotch cells, Lim proposes, could make them viable drugs.

Yvonne Chen, who runs a synthetic biology lab at UCLA working on similar problems as Lim, calls synNotch “a powerful platform” that could have many uses.

“I think this potential was evident from the initial studies, and what the current study adds is confirmation that the general strategy is indeed applicable to a variety of outputs,” says Chen, who was not involved in the study.

The idea might face a stiff test in relatively short order. Lim says Cell Design Labs could start a … Next Page »

Single PageCurrently on Page: 1 2