Background CAR-T cell therapies have demonstrated remarkable clinical benefit in liquid tumors, but significant obstacles such as antigen loss, tumor heterogeneity, T cell exhaustion, and poor persistence limit the frequency and durability of complete responses. To overcome these challenges and expand this class of therapeutics to solid tumors, T cells will need to be multi-specific, resistant to exhaustion, and maintain their fitness in the tumor microenvironment. Engineering NextGen therapies with these enhanced capabilities will require the introduction of additional genetic modules that often do not fit within the payload capacity of a single vector, necessitating the use of multiple vectors. However, engineering cells with multiple vectors results in a heterogeneous cell product that requires purification for each vector, which is not currently possible in the clinical setting. Here, we present STASH Select, a simple method for enriching cells containing multiple genetic modifications using a single selection marker.

Methods Human CAR-T cells were generated from healthy donors and characterized by flow cytometry, ELISA, tumor killing assays, bulk RNAseq, and MACS.

Results The core principle of STASH-Select is the conditional cellular localization of a selectable marker. We integrate a protease cleavage site and an ER-Tag onto the C-terminus of the EGFRt selection marker, which “STASHes” it inside the cell. Co-expression of the cognate TEV protease results in cleavage of the ER Tag, which releases the EGFRt from the ER and allows it to translocate to the cell surface. Encoding the STASHed EGFRt on one vector and the protease on a second vector creates an AND gate system, where only cells transduced with both vectors express high levels of surface EGFRt that can be used as a selection handle for isolating double+ cells. We screened 22 ER retention domains and identified four novel, compact (<45aa) human ER tags that yield high performance STASH-Select systems. We tested the two-vector STASH-select in primary human T cells and were able to enrich an initial double+ population of 12.3% to 96.3% purity via EGFRt MACS. We expanded this approach to a three-vector STASH-Select using Split TEV protease. Configuring the STASHed EGFRt and the two TEV fragments onto three separate vectors produces a 3-vector AND gate which allowed for purification of three clinically relevant cargos with a single-step MACS selection, enriching an initial population of cells from 16.7% to >90% triple+.

Conclusions STASH-Select is a novel selection platform designed to propel NextGen smart therapies containing multiple enhancing modules into the clinic.