Back

A Singular Base Editing Platform for Polyfunctional Multiplex Engineering of Immune Cells

Skeate, J. G.; Slipek, N. J.; Lahr, W. S.; Roy, S.; Wick, B. J.; Stelljes, E. M.; Gilkey, A. K.; Thenge, P. P.; Diers, M. D.; Kar, B.; Krueger, J. B.; Niemeyer, E. M.; Lonetree, C.-l.; Kluesner, M. G.; Bell, J. B.; Clement, K.; Provenzano, P.; Moriarity, B. S.; Webber, B. R.

2025-07-16 bioengineering
10.1101/2025.07.11.664404 bioRxiv
Show abstract

Current methods to engineer antigen-specific receptors rely on randomly integrating vectors or double-strand break induced targeted integration, both of which pose safety risks. To implement an all-in-one tool for multiplex knockout (KO) and knock in (KI), we expand the use of cytosine and adenine base editor (ABE) nickase activity to stimulate homology-directed repair (HDR) and insert clinically relevant chimeric antigen receptors (CARs) into specific loci. Through a novel sgRNA design strategy and a recombinant adeno-associated virus (rAAV) delivered DNA template, we enhanced the efficiency of ABE8e-stimulated HDR in human T cells. By combining KI of CD19-, CD33-, or mesothelin-targeting CARs with >95% quadplex gene KO (B2M/CD3{varepsilon}/PDCD1/CISH), we achieve single-step generation of highly functional off-the-shelf CAR T cell products with enhanced function. Importantly, we found no detectable translocations or significant off-target edits and demonstrated efficacy against multiple cancer lines, and a suppressive 3D spheroid culture model. This efficient engineering process of Iterative Nicking for Synchronous Engineered Reprogramming of T cells (INSERT) establishes a safe, simplified platform for advanced therapeutic CAR T engineering.

Published in Molecular Therapy (predicted rank #1) · training set

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.