Charles River: Understanding and mitigating underlying mechanisms behind unwanted toxicities in CAR T cell therapies
Teams from Charles River and Distributed Bio presented research at AACR 2021 in April detailing how their technology and expertise benefits CAR T cell therapy developers.
Distributed Bio’s phage-display libraries (SuperHuman 2.0 and Tungsten) have been designed to be biophysically stable and non-immunogenic. Their size and clonal diversity, coupled with the use of streamlined work-flows, enable the rapid identification of large panels of sequence-unique binders which can then be assessed for CAR suitability by exploiting the integrated approach offered by Distrbuted Bio’s partnership with Charles River, they explained.
Inherent risks in CAR-T therapy development
Outlining how the global market for chimeric antigen receptors (CAR) therapy is continuing to grow rapidly, with many clinical success stories for B-cell malignancies, the experts noted how the path to clinical success has not been smooth, however, with early reports of organ damage and even deaths following CAR T cell therapy, highlighting the significant risks involved in the development of such treatment approaches.
There is a clear requirement to understand and mitigate the underlying mechanisms that result in unwanted toxicities, said the Charles River and Distributed Bio teams.
Off-target binding and tonic signaling are indicated in many pre-clinical and clinical cell therapy failures, and it is becoming increasingly apparent that CAR design – particularly the selection of the tumor targeting moiety with exquisite specificity, desirable biophysical attributes and appropriate affinity – is key to developing a successful therapy, said Katherine Vousden, Science Director, large molecules, at Charles River Laboratories.
Navigating unwanted toxicities
She weighed in on the mechanism behind some of those CAR T toxicities for us:
“There are general toxicities that are related to T-cell activation and the subsequent cytokine release syndromes that can result from over-activation of those T-cells. Over-activation of T-cells can result from the level of antigen expressed on the tumor cells or the number of tumor cells in the patient – known as the tumor burden – as well as elements more specific to the CAR and its design – including the affinity of the single chain (scFv), or antigen binding domain on the CAR T for its target."
It can be challenging for therapy developers then to ensure clinical efficacy while avoiding these systemic cytokine toxicities. “There is a real balance to be struck between a threshold level of activation and cytokine secretion – you need a certain amount of cytokine secretion to actually activate and destroy the tumor cells but you don’t want so much that you get to a level that results in this vicious cycle of cytokine release.”
There are also other inherent risks, toxicities than can result from specific interactions between the CAR and its target antigen if it is expressed also by non-malignant cells, which is termed on-target, off tumor, said the CRO's science director. “Not all tumor associated antigens are only expressed on the tumor and an example of that was where a high affinity HER2 binder was used in a CAR T construct which unfortunately resulted in a clinical death due to respiratory failure – assumed to be caused by on-target, but off tumor binding - meaning healthy tissue was targeted by the CAR T."
And then there is also tonic signaling, which is target independent T-cell activation, even in the absence of its target antigen, she added. “T-cells can become spent in this way and they don’t persist, they are not present in sufficient quantities to then actually target the tumor.”
Stability
Tonic signaling is a complex area and the CRO believes there to be a number of different root causes for this, continued Vousden.
“Specifically, as it relates to Distributed Bio’s portfolio, some of the self-activation by T-cells can be caused by an instability of the binding partner, so the single chain. If a single chain is derived from an IgG, which would be typical if it had been isolated through an immunization strategy for antibody discovery, what you can often see is an instability between the heavy and the light chain interface - it doesn’t fold as well as it did in the context of the whole IgG anymore. And that instability and that unfolding that you get between the heavy and light chain, particularly if the CAR construct is expressed at quite a high density on the T-cell, can enable different binding partners from adjacent CARs to interact with each other, which can cause and drive the stimulation of the T-cell in the absence of antigen.
“One of the biggest advantages of Distributed Bio’s platform is the fact that it is a single chain discovery platform that has that stability component hardwired into its design, so only very well behaved heavy and light chain germlines are chosen in the construct of the library. Additionally, it is thermally challenged during its construction so that any misfolding of single chains that occurs results in their removal from the population for selection. That stability component is one of principal aspects of Distributed Bio’s platform, to help mitigate [that target independent T-cell activation].”
Having a panel of sequence diverse, epitope diverse, and kinetically diverse single chains so the client or Discovery teams elsewhere within Charles River can take that panel and test for what those differences mean and how that translates to the in vitro or in vivo model is also massively powerful, Vousden stressed.