Polymorphic HLAs form the primary immune barrier to cell therapy. In addition, innate immune surveillance impacts cell engraftment, yet a strategy to control both, adaptive and innate immunity, is lacking. Here we employed multiplex genome editing to specifically ablate the expression of the highly polymorphic HLA-A/-B/-C and HLA class II in human pluripotent stem cells. Furthermore, to prevent innate immune rejection and further suppress adaptive immune responses, we expressed the immunomodulatory factors PD-L1, HLA-G, and the macrophage “don’t-eat me” signal CD47 from the AAVS1 safe harbor locus. Utilizing in vitro and in vivo immunoassays, we found that T cell responses were blunted. Moreover, NK cell killing and macrophage engulfment of our engineered cells were minimal. Our results describe an approach that effectively targets adaptive as well as innate immune responses and may therefore enable cell therapy on a broader scale. Significance To enable cell therapy on a broader scale, the development of universal donor stem cell products that can be administered to multiple patients in need has been proposed, yet a strategy controlling both adaptive and innate immune rejection has not been reported. In this study, we employed multiplex genome editing to selectively ablate the highly polymorphic HLA class Ia and class II molecules and introduced the immunoregulatory factors PD-L1, HLA-G, and CD47 to control T cell- and NK cell-mediated immune responses and macrophage engulfment in vitro and in vivo. Our strategy demonstrates the power of cell engineering and informs future studies aiming to generate “off-the-shelf” universal cell products that may make cell therapy available to a larger pool of patients.

Anti-CD19 chimeric antigen receptor (CD19-CAR)-engineered T cells are approved therapeutics for malignancies. The impact of the hinge domain (HD) and the transmembrane domain (TMD) between the extracellular antigen-targeting CARs and the intracellular signaling modalities of CARs has not been systemically studied. In this study, a series of 19-CARs differing only by their HD (CD8, CD28, or IgG₄) and TMD (CD8 or CD28) was generated. CARs containing a CD28-TMD, but not a CD8-TMD, formed heterodimers with the endogenous CD28 in human T cells, as shown by co-immunoprecipitation and CAR-dependent proliferation of anti-CD28 stimulation. This dimerization was dependent on polar amino acids in the CD28-TMD and was more efficient with CARs containing CD28 or CD8 HD than IgG₄-HD. The CD28-CAR heterodimers did not respond to CD80 and CD86 stimulation but had a significantly reduced CD28 cell-surface expression. These data unveiled a fundamental difference between CD28-TMD and CD8-TMD and indicated that CD28-TMD can modulate CAR T-cell activities by engaging endogenous partners.