NK cells have therapeutic potential for a wide variety of human malignancies. However, because NK cells expand poorly in vitro, have limited life spans in vivo, and represent a small fraction of peripheral white blood cells, obtaining sufficient cell numbers is the major obstacle for NK-cell immunotherapy. Genetically-engineered artificial antigen-presenting cells (aAPCs) expressing membrane-bound IL-15 (mbIL15) have been used to propagate clinical-grade NK cells for human trials of adoptive immunotherapy, but ex vivo proliferation has been limited by telomere shortening. We developed K562-based aAPCs with membrane-bound IL-21 (mbIL21) and assessed their ability to support human NK-cell proliferation. In contrast to mbIL15, mbIL21-expressing aAPCs promoted log-phase NK cell expansion without evidence of senescence for up to 6 weeks of culture. By day 21, parallel expansion of NK cells from 22 donors demonstrated a mean 47,967-fold expansion (median 31,747) when co-cultured with aAPCs expressing mbIL21 compared to 825-fold expansion (median 325) with mbIL15. Despite the significant increase in proliferation, mbIL21-expanded NK cells also showed a significant increase in telomere length compared to freshly obtained NK cells, suggesting a possible mechanism for their sustained proliferation. NK cells expanded with mbIL21 were similar in phenotype and cytotoxicity to those expanded with mbIL15, with retained donor KIR repertoires and high expression of NCRs, CD16, and NKG2D, but had superior cytokine secretion. The mbIL21-expanded NK cells showed increased transcription of the activating receptor CD160, but otherwise had remarkably similar mRNA expression profiles of the 96 genes assessed. mbIL21-expanded NK cells had significant cytotoxicity against all tumor cell lines tested, retained responsiveness to inhibitory KIR ligands, and demonstrated enhanced killing via antibody-dependent cell cytotoxicity. Thus, aAPCs expressing mbIL21 promote improved proliferation of human NK cells with longer telomeres and less senescence, supporting their clinical use in propagating NK cells for adoptive immunotherapy.

Chimeric antigen receptors (CAR) combine an antigenbinding domain with a CD3-Z signaling motif to redirect Tcell specificity to clinically important targets. First-generation CAR, such as the CD19-specific CAR (designated CD19R), may fail to fully engage genetically modified T cells because activation is initiated by antigen-dependent signaling through chimeric CD3-Z, independent of costimulation through accessory molecules. We show that enforced expression of the fulllength costimulatory molecule CD28 in CD8 + CD19R + CD28À T cells can restore fully competent antigen-dependent T-cell activation upon binding CD19 + targets expressing CD80/CD86. Thus, to provide costimulation to T cells through a CD19-specific CAR, independent of binding to CD80/CD86, we developed a second-generation CAR (designated CD19RCD28), which includes a modified chimeric CD28 signaling domain fused to chimeric CD3-Z. CD19R + and CD19RCD28 + CD8 + T cells specifically lyse CD19 + tumor cells. However, the CD19RCD28 + CD8 + T cells proliferate in absence of exogenous recombinant human interleukin-2, produce interleukin-2, propagate, and up-regulate antiapoptotic Bcl-X L after stimulation by CD19 + tumor cells. For the first time, we show in vivo that adoptively transferred CD19RCD28 + T cells show an improved persistence and antitumor effect compared with CD19R + T cells. These data imply that modifications to the CAR can result in improved therapeutic potential of CD19-specific T cells expressing this second-generation CAR. (Cancer Res 2006; 66(22): 10995-1004)