Boosting Precision and Minimizing Toxicity of CAR T-Cell Therapies
Developing a predictive assay to improve CAR T-cell risk stratification among patients
Technology Overview
Dr. Stephen E.P. Smith
Chimeric antigen receptor (CAR) T-cell therapy has been a transformative treatment for many pediatric patients with refractory B-cell malignancies and other cancers. The bioengineered therapy’s potential remains largely unfulfilled, however, due in part to unpredictable side effects such as cytokine release syndrome (CRS) and neurotoxicity.
Standard preclinical assays in animal models often do not accurately predict the activity of custom-made CARs within human patients, contributing to the failures of costly clinical trials. Neuroimmunologist Stephen E.P. Smith, PhD, has developed a new assay based on quantitative multiplex co-immunoprecipitation (QMI) that can more accurately measure the multitude of signaling interactions associated with stimulation of bioengineered CAR T cells. In QMI, complexed proteins are co-immunoprecipitated from CAR T cells and quantified using labeled antibody probes detected by flow cytometry. By detecting even small changes or disruptions to the protein complexes, this assay can predict whether those interactions will yield good or bad clinical responses.
To test the QMI assay’s potential, Dr. Smith and team compared CAR products made with the same DNA instructions. For unknown reasons, however, some of the clinical products elicited a complete remission free of side effects, while others caused CRS or immune effector cell-associated neurotoxicity syndrome (ICANS). The team measured a QMI-based biosignature from more than 40 clinical products and correlated the biosignature variations with patient outcomes. The team simultaneously measured 400 protein-protein interactions from each sample, and identified a group of interactions that differed between the optimal response-associated samples and those that caused CRS. A retrospective analysis using a machine-learning classifier trained on QMI biosignatures used this difference to identify CRS-associated CAR samples with high accuracy.
The predictive algorithm created by Dr. Smith’s company, Q-Immune, could provide an important new tool for two critical points in clinical trials. First, it could supply pharmaceutical companies with preclinical predictions about the likelihood of success in patients to help weed out poor CAR candidates. Second, given the large patient-to-patient variability in drug activity, the tool could aid medical centers by predicting if a particular CAR T-cell therapy is likely to cause CRS in individual patients and help risk stratify patients pre-infusion. For low-risk patients, the therapy could be conducted as a more time- and cost-efficient outpatient procedure.
Dr. Smith has extensive experience in neuroimmunotherapy, proteomics, QMI, and other technology focused on deciphering the language of the cell. His approach is to determine the mechanistic basis of how CARs redirect endogenous signaling pathways and to develop approaches that optimize the cellular response through optimization of protein interaction networks.
He is interested in industry partnerships that could provide CAR T-cell products and clinical data to further refine the QMI experimental assay and increase its utility in mapping and quantifying signal transduction responses for more rational CAR T-cell design. As the QMI and predictive modeling continue to improve, the testing framework could be scaled up relatively quickly to increase capacity.
Stage of Development
- Preclinical in vitro
- Preclinical in vivo
- Preclinical ex vivo
- Licensing agreements
Partnering Opportunities
- Clinical trials
- Collaborative research and development
- Sponsored research agreement
- Consultation agreement
- Investigator-initiated clinical trials
- Drug development opportunity
- Tissue sample access
- Cell line access
Publications
Ritmeester-Loy SA, Draper IH, Bueter EC, … Smith SEP. Differential protein-protein interactions underlie signaling mediated by the TCR and a 4-1BB domain–containing CAR. Sci Signal. 2024;17(826):eadd4671.
Neier SC, Ferrer A, Wilton KM, Smith SEP, et al. The early proximal αβ TCR signalosome specifies thymic selection outcome through a quantitative protein interaction network. Sci Immunol. 2019;4(32):eaal2201.
Wehle DT, Bass CS, Sulc J, Mirzaa G, Smith SEP. Protein interaction network analysis of mTOR signaling reveals modular organization. J Biol Chem. 2023;299(11):105271.
Learn More