Genetic modification of T cells with an artificial tumor-targeting chimeric antigen receptor (CAR) is a new approach for adoptive cell therapy for cancer. Defining cell surface molecules that are both selectively expressed on cancer cells and can be safely targeted with T cells or NK cells is a significant challenge in this research field.

NKp44 is a member of the natural cytotoxicity receptor (NCR) families and also known as NCR2. Expression of NKp44 is limited to activating NK cells, which leads to a marked increase in cytotoxicity against tumors. The receptor contains one extracellular immunoglobulin domain, type I transmembrane (TM) domain, and intracellular (IC) domain, and its surface expression seems to require binding of the TM domain to adaptor molecules of DAP12 accessory protein that contains ITAMs. The ligand for NKp44 is considered damage-associated molecular pattern molecules, which have been reported to be expressed by various types of cancer cells but not by healthy cells. Therefore, a wide range of cancer cells may be safely targeted if the ligand-binding domain of this receptor is used in a construction of a chimeric antigen receptor (CAR) as an antigen recognition site, instead of using single chain variable domains derived from monoclonal antibody.

We created several NKp44-based CAR constructs, which shares the extracellular NKp44 IG domain as a ligand-binding domain. Surface expression levels and subsequent functional properties can differ among T cells or NK cells transduced with novel CARs with different structural characteristics. We thus tested whether swapping the domains other than the antigen-binding domain affected expression and function. The CAR genes were retrovirally transduced into human primary T cells according to a standard method. We also transduced human primary NK cells with NKp44-based CARs, by a previously reported method (Imai C, et al. Blood 2005), to compare the expression pattern of the CAR in NK cells with that in human T cells.

Retroviral transfer of wild type NKp44 gene and a construct harboring IC(p44) both did not induce NKp44 surface expression (Fig 1A,B). By sharp contrast, primary NK cells were able to express the CAR protein on the cell surface after transfer of these two genes. Removal of the IC(p44) [EH(p44)-TM(p44)-IC(CD3z)] allowed slight surface expression in T cells (Fig1C). The replacement of TM(p44) with TM(CD8a) resulted in higher surface expression in T cells (Fig 1D). These observations indicated the presence of IC(p44) as well as TM(p44) in the CAR constructs hampered surface expression in T cells most likely due to the lack of DAP12 expression. In addition to TM replacement, replacement of EH(p44) with EH(CD8a) markedly increased surface expression of the CAR (Fig 1E). Similarly, we tested use of CD28 domains instead of CD8a. Surprisingly, as different from the case of CD8a, the construct EH(p44)-TM(CD28)-IC(CD3z) yielded highest surface expression among the all CAR constructs created in this study in T cells as well as in NK cells (Fig1F), while the replacement of EH(p44) of the abovementioned CAR with EH(CD28) resulted in marked reduction of the CAR expression (Fig 1G).

We confirmed surface expression of NKp44 ligand with flow cytometric analysis using recombinant human NKp44 Fc chimera protein (R&D Systems, McKinley Place, Minneapolis, USA) on various tumor cell lines including myeloid leukemia (K562, THP-1, U937, KY821, HL60), T-cell leukemia (PEER, MOLT4, HSB2), Burkitt lymphoma (Raji), BCR-ABL-positive B-ALL (OP-1), osteosarcoma (MG63, NOS1, NOS2, NOS10, U2OS, SaOS2), rhabdomyosarcoma (Rh28, RMS-YM), neuroblastoma (SK-N-SH, NB1, NB16, IMR32), and cervical carcinoma (Hela). Function of the best construct [EH(p44)-TM(CD28)-IC(CD3z)] was further evaluated. Primary T cells transduced with this NKp44-based CAR exerted powerful cytotoxicity against tumor cell lines tested and produced interferon-g and granzyme B, while GFP-transduced T cells and control T cells transduced with truncated NKp44-based CAR did not.

In conclusion, we have created a novel CAR based on the antigen-binding property derived from NKp44 receptor immunoglobulin domain. This CAR should be effective to redirect T cells as well as NK cells against various types of cancer including hematological malignancies.

Figure 1

Schematic representation of gene constructs and their surface expression of NKp44-based CARs in human T cells and NK cells.

Figure 1

Schematic representation of gene constructs and their surface expression of NKp44-based CARs in human T cells and NK cells.

Close modal

Imai:Juno Therapeutics: Patents & Royalties.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution