Chimeric antigen receptor T cell (CAR-T) therapy has transformed treatment options for many cancer patients. CAR-T infusion products are made by taking the patient’s own T cells, transferring the CAR construct into the donor T cells with an engineered virus, expanding them, and re-infusing them into the patient. CAR-T engraftment and expansion after infusion is an important indicator of treatment efficacy, however little is known about the clonal dynamics of CAR-Ts after infusion. Because the donor T cells each have their own unique endogenous T cell receptor (TCR) before CAR transfer, they can be tracked over time in patients receiving CAR-Ts. The Turtle lab (Clinical Research Division) examined the expansion of infusion products over time by tracking the endogenous TCR. Their work was recently published in Nature Communications.
By tracking TCR abundance over time, the authors could identify which clones expanded in the patient. They tracked the composition of CAR-T products in ten patients who received CD19-targeting therapy, determining the content of the product prior to infusion, at an early timepoint (7-14 days post infusion), and at a late timepoint (26-30 days post infusion). Interestingly, they found that many of the top ten clones at early and late timepoints were found in relatively low abundance in the infusion product, indicating that indeed some clones were favorably expanded over others. However, all products were polyclonal at each timepoint, indicating no preference for one clone specifically. The authors compared the dynamics of abundant clones over time and found three main groups: clones that steadily increased, clones that initially increased and then decreased, and clones that steadily decreased. Because the virus integrates the CAR construct randomly into the T cell, there is a possibility that site integration could influence clonal growth. Importantly, the authors determined that clonal outgrowth was not related to integration site.
In order to understand what did influence clonal expansion, the authors used single cell RNA sequencing of four patients with durable persistence of the CAR-T product. They found that CAR-Ts at early and late timepoints expressed higher levels of genes associated with T cell function compared to the infusion product, indicative of T cell activation in the patient. At the latest timepoints, cells expressed lower levels of cell proliferation genes which correlated with the timing of cancer regression, and therefore a reduction of CAR-T target antigen. Using principle component analysis, the authors were able to classify cells in the infusion product by their gene expression, leading to four clusters of starting cells. Cluster 2 was associated with higher levels of cytotoxicity genes necessary for T cell killing of cancer cells. Cluster 4 was associated with higher levels of proliferation genes and were determined to be actively dividing. The authors were able to link the single cell RNA sequencing to each cell’s unique TCR to track how the starting clusters of cells expanded over time. The majority of clones that expanded were derived from clusters 2 and 4, indicating that cells with higher markers of T cell proliferation and/or activation were strongly selected for in the patient. Dr. Alyssa Sheih, a postdoc in the lab who led the study, expanded on the findings: “This study is one of the first to link the transcriptional profile of CAR-T cells in the infusion product to their clonal kinetics after adoptive transfer. We wanted to understand why some CAR-T cell clones expanded and contributed to the anti-tumor response while other CAR-T cell clones disappeared after infusion. These findings are not enough to change the current manufacturing strategy but still an important step towards understanding mechanisms of therapeutic responses or failures to CAR-T cell immunotherapy.”
This work was supported by Juno Therapeutics/Celgene, Inc., an IIRC pilot grant, the Bezos Family, and the NHLBI funded National Gene Vector Biorepository at Indiana University.
Fred Hutch/UW Cancer Consortium members Stan Riddell, David Maloney, Hans-Peter Kiem, Jennifer Adair, Raphael Gottardo, and Cameron Turtle contributed to this work.
Alyssa Sheih, Valentin Voillet, Laïla-Aïcha Hanafi, Hannah A. DeBerg, Masanao Yajima, Reed Hawkins, Vivian Gersuk, Stanley R. Riddell, David G. Maloney, Martin E. Wohlfahrt, Dnyanada Pande, Mark R. Enstrom, Hans-Peter Kiem, Jennifer E. Adair, Raphaël Gottardo, Peter S. Linsley & Cameron J. Turtle. Clonal kinetics and single-cell transcriptional profiling of CAR-T cells in patients undergoing CD19 CAR-T immunotherapy. Nat Comm. 11, 219 (2020).