Predictive Toxicology in the Era of CAR-T and TCR Drugs
Regina Kelder

Predictive Toxicology in the Era of CAR-T and TCR Drugs

How labs are using in vitro human cells to de-risk powerful T cell therapies during early discovery. The first in our four-part series on predictive toxicology.

The human immune system is designed to repel foreign invaders, but when it comes to cancer the odds have not been in our favor. That is because tumors are supremely clever at evading T cells and B cells.

Fortunately we are getting better and better at recognizing our enemies, and we have technology to thank for that. One recent drug strategy involves extracting white blood cells from a cancer patient, and then genetically engineering the T cells to express receptors (called chimeric antigen receptors, or CARs) that seek out and kill cancer. A similar approach called T cell receptor or TCR therapy re-engineers the receptor itself.

Two CAR-T therapies have already been approved, one for children with acute lymphoblastic leukemia and the other for adults with advanced lymphomas. With the first cell based therapies targeting hematological malignances now approved, their next challenge is solid tumors and blood cancers. Solid tumors form an additional challenge due to the lack of tumor specific target antigens, which pose significant safety risks.

Because these products have the capacity to fill tremendous unmet medical needs, CAR-Ts are moving into the clinic much faster than, say, a new drug for diabetes or arthritis. This expedited timetable is certainly good news for patients with late stage disease, but it also means less time to conduct the rigorous safety tests that are required before the drug can be tested in humans.

Vetting T-cell therapies

This presents a dilemma for the pharmaceutical companies developing these products, and the CROs that test them. How can due diligence be performed in a compressed timeline on therapies as complex as cell-based products? To meet these challenges, drug developers have been departing from the usual step-wise fashion of discovery first, safety assessment later, and begun to incorporate safety and efficacy endpoints when candidate therapies haven’t even left the station.

Because conventional in vivo safety models are often not suitable for cell therapies, one novel way companies are vetting these therapies is by in vitro co-culturing a wide variety of healthy primary human cells with CARs or TCRs and assessing both T cell activation and reactivity against the healthy primary cells. This in vitro process, a type of predictive toxicology, helps establish whether the cell therapy will activate the proper target with laser-sharp precision or whether it unwantedly activates against healthy tissue along the way.

TCRs and CARs provide an excellent example of how predictive toxicology can be helpful in determining whether a drug is working properly or likely to go rogue. TCRs and CARs have their specific pros and cons but overall they both offer great promise in becoming a new wave of highly specific therapies against solid tumors.

TCRs and CARs involve re-engineering receptors onto T cells so that they better recognize cancer proteins and can spark an assault on tumors. There are subtle differences between CAR T-s and TCRs, though. CAR-Ts can only recognize abnormal proteins on the surface of tumor cells, while TCRs are also able to recognize tumor-specific proteins expressed inside of cells via human leukocyte antigens or HLA presentation.

Co-culturing strategies  

The big safety challenge for these cell therapies is that the tumor protein they are engineered to recognize is often also expressed on healthy organs in the human body. In an ideal world the TCRs and CARs would only program the tumor cells to die. Unfortunately, there is a risk they will also activate against healthy cells unleashing a storm of friendly fire that causes serious and possibly life-threatening side effects. By co-culturing the engineered T cells (TCRs or CARs) with a primary human cell we can look for signs of unwanted reactivity (when the TCR binds to both tumor and non-tumor targets) or specific reactivity where only tumor cells are triggered to die.  

“One of the major challenges with T cell therapies is that they are designed to bind to specific targets on tumor cells,” says Sanne Demandt, PhD, a Group Leader at Charles River’s Early Discovery site in Leiden. “The problem is these targets (antigens) are often also expressed on healthy cells. So you need to ‘de-risk’ the primary tissues to make sure the engineered T cells don’t attack the normal cells.”

A recent TCR safety co-culture study using human cells helped to de-risk a therapy. Demandt’s lab started by screening 10 healthy primary cell types, covering all the major organs in a human body, for the correct HLA type. “Upon co-culture with the TCR T cells, we were able to profile the TCR T cell activation and T cell mediated-cancer killing,” says Demandt. “The study included multiple lead TCRs for which we were able to compare their potency and toxicity against human healthy tissue, assisting in selecting the best lead candidate.”

Through these assays, essential safety indications are assessed for such therapies and lead candidates are identified for further characterization in more complex co-culture assays. By co-culturing primary human T cells and cancer cell lines one can then visualize and assess the T cell mediated toxicity in 3D cultures, dramatically improving prediction of T cell invasion and their efficacy against solid tumors.

A complementary approach

Using in vitro human cells to predict toxicity does not replace animal tests; in vivo study are still very useful to assess efficacy but their ability to assess the safety of cell therapies has shown to fall short in the past. As a scientific community we therefore need to invent novel ways to establish that these therapies are safe, says Demandt.  

“CARs and TCRs are one of the most exciting new therapies of the last decade in our fight against cancer,” says Demandt. “They have been shown to be highly efficacious, but we also need to make sure they are safe. The ability to mimic human organs and tissues in vitro and utilize them to test the safety of these therapies has been and will be essential in getting these novel drugs approved and available for patients."

Tune in tomorrow for a look at the benefits--and roadblocks--in using Big Data in predictive toxicology.