AUTHORS
John Michael Warrington, BS, Jeffrey Lotthammer, PhD2, Henry Humble3, Lyra Morina, BS4, Ryan J. Emenecker5, Alex S. Holehouse, PhD6, Nathan Singh, MD, MS7
AFFILIATIONS
1 Washington University School of Medicine, St Louis, Missouri, United States
2 Washington University School of Medicine, St Louis, Missouri, United States
3 University of Tennesee-Knoxville, Knoxville, Tennessee, United States
4 Washington University School of Medicine, St Louis, Missouri, United States
5 Washington University School of Medicine, St Louis, Missouri, United States
6 Washington University School of Medicine, St Louis, Missouri, United States
7 Washington University School of Medicine, St Louis, Missouri, United States
ABSTRACT
Background: Durable responses to chimeric antigen receptor (CAR) T-cell therapy require rapid cytotoxicity followed by sustained function during recurrent antigen exposure. The CAR is a single-chain protein with an extracellular target recognition domain and an intracellular signaling domain consisting of 4-1BB (or CD28) and CD3ζ linked end-to-end. CAR intracellular domains are generally studied as modular elements, limiting identification of the precise sequence features responsible for these temporally distinct activities. Unlike folded proteins, the 4-1BB–CD3ζ (BBζ) intracellular domain is entirely disordered, precluding conventional structure-informed mutational analysis. We therefore deployed an engineering framework to independently evaluate the contributions of amino-acid composition and residue order to CAR T-cell function.
Methods: We generated 166 unique CD19 CAR signaling domains, including tiled deletions, glycine-serine replacements, residue shuffles, and variants selectively depleted of physicochemical residue classes. CAR T cells from at least two independent donors were cocultured with CD19-positive NALM6 cells at a 1:4 effector-to-target ratio, with targets replenished every 2 days. Acute cytotoxicity, proliferation, and sustained tumor control were measured by flow cytometry. Functional failure was defined as the first day on which the target-to-effector ratio exceeded 10:1. CAR surface expression was measured to exclude differences in receptor abundance or baseline fitness.
Results: Progressive sequence reconstruction revealed distinct design requirements for acute and sustained CAR activity. A minimal construct retaining only the consensus residues from all 3 CD3ζ immunoreceptor tyrosine-based activation motifs (ITAMs) positionally conserved within a length-matched glycine-serine scaffold failed to restore acute activity. In contrast, preserving each complete ITAM at its native position while replacing the remainder of the intracellular domain with glycine-serine linkers restored acute NALM6 control despite only 39% sequence identity to BBζ. This ITAM-only CAR nevertheless failed during repeated challenge, establishing that complete, positionally conserved ITAMs are sufficient for initial activity but not durable function.
We next reconstructed sustained activity within this ITAM-only background. Addition of either (or both) reported 4-1BB tumor necrosis factor receptor–associated factor (TRAF)-binding motifs failed to extend tumor control. Reintroduction of amino acid classes, strikingly revealed that only when a contiguous 21-amino acid 4-1BB segment spanning intracellular residues 18–38 (PVQTTQEEDGCSCRFPEEEEG) was present were CAR T-cells capable of persisting to the same degree as wild-type BBζ. Thus, this element was sufficient to confer sustained function outside the remainder of native 4-1BB.
Next we sought to determine if this sequence was a bona fide motif or if its function was driven through non-specific chemical complementarity (e.g. electrostatic interactions). In an otherwise wild-type BBζ CAR, selective depletion of cysteine, aliphatic, charged, or polar residues from this element abolished sustained activity despite preserved CAR surface expression and acute cytotoxicity, revealing an independent requirement for amino-acid composition. Likewise, deletion, glycine-serine linker substitution, or shuffling of this region alone shortened tumor control to the duration observed with the ITAM-only CAR. In contrast, size-matched shuffling of every other intracellular region tested did not detectably impair sustained control. Because shuffling preserved amino-acid composition but disrupted order, together, these findings identified that both native residue composition and order are specific functional requirements within residues 18–38.
Conclusion: Systematic engineering of 166 CAR signaling domains uncovered a hierarchical sequence code for CAR T-cell function. Conserved ITAM residues alone were insufficient for initial activity but complete ITAMs in their native positions enabled acute tumor control. However, sustained activity required an additional 21–amino acid 4-1BB element governed by both its chemical composition and native residue order. These findings separate the sequence determinants of immediate and durable CAR activity and identify a compact 4-1BB element that can serve as a mechanistic building block for rationally engineering longer-lasting CAR T cells.
DISCLOSURES
John Michael Warrington, BS: No relevant conflicts of interest to declare.