Chimeric Antigen Receptor-Engineered NK-92 Cells: An Off-the-Shelf Cellular Therapeutic for Targeted Elimination of Cancer Cells and Induction of Protective Antitumor Immunity

Abstract

Significant progress has been made in recent years toward realizing the potential of natural killer (NK) cells for cancer immunotherapy. NK cells can respond rapidly to transformed and stressed cells and have the intrinsic potential to extravasate and reach their targets in almost all body tissues. In addition to donor-derived primary NK cells, also the established NK cell line NK-92 is being developed for adoptive immunotherapy, and general safety of infusion of irradiated NK-92 cells has been established in phase I clinical trials with clinical responses observed in some of the cancer patients treated. To enhance their therapeutic utility, NK-92 cells have been modified to express chimeric antigen receptors (CARs) composed of a tumor-specific single chain fragment variable antibody fragment fused via hinge and transmembrane regions to intracellular signaling moieties such as CD3ζ or composite signaling domains containing a costimulatory protein together with CD3ζ. CAR-mediated activation of NK cells then bypasses inhibitory signals and overcomes NK resistance of tumor cells. In contrast to primary NK cells, CAR-engineered NK-92 cell lines suitable for clinical development can be established from molecularly and functionally well-characterized single cell clones following good manufacturing practice-compliant procedures. In preclinical in vitro and in vivo models, potent antitumor activity of NK-92 variants targeted to differentiation antigens expressed by hematologic malignancies, and overexpressed or mutated self-antigens associated with solid tumors has been found, encouraging further development of CAR-engineered NK-92 cells. Importantly, in syngeneic mouse tumor models, induction of endogenous antitumor immunity after treatment with CAR-expressing NK-92 cells has been demonstrated, resulting in cures and long-lasting immunological memory protecting against tumor rechallenge at distant sites. Here, we summarize the current status and future prospects of CAR-engineered NK-92 cells as off-the-shelf cellular therapeutics, with special emphasis on ErbB2 (HER2)-specific NK-92 cells that are approaching clinical application.

Keywords: NK-92; adoptive cancer immunotherapy; chimeric antigen receptor; leukemia; lymphoma; natural killer cells; solid tumors.

Figure 1

Expression of first- and second-generation chimeric antigen receptors (CARs) in NK-92 cells. (A) Schematic representation of first- and second-generation CARs for expression in NK-92 cells that consist of an extracellular single chain fragment variable (scFv) antibody domain for target recognition fused via a hinge region derived from CD8α (hinge) to transmembrane and intracellular domains of CD3ζ (left), transmembrane and intracellular domains of CD28, and the intracellular domain of CD3ζ (middle), or transmembrane and intracellular domains of CD137 (4-1BB) and the intracellular domain of CD3ζ (right). (B) To assess endogenous expression of CD28 and CD137, lysates of NK-92 cells were subjected to SDS-PAGE and subsequent immunoblotting with CD28- and CD137-specific antibodies as indicated. Lysates of peripheral blood mononuclear cells (PBMCs) from healthy donors were included for comparison. (C) For analysis of CAR expression, lysates of NK-92 cells transduced with lentiviral vectors that encode CD19-specific CARs containing CD3ζ, composite CD28-CD3ζ, or CD137-CD3ζ signaling domains as represented in (A) were subjected to SDS-PAGE under reducing (R, left panel) or non-reducing conditions (NR, right panel) and subsequent immunoblotting with CD8α-specific antibody, which detects the hinge domain. The positions of CAR monomers, homodimers of the CD3ζ CAR, and heterodimers of the CD3ζ CAR with endogenous CD3ζ are indicated. Data in panel (C) are from Oelsner et al. (74).

Figure 2

Activity of NK-92/5.28.z against ErbB2-expressing breast carcinoma cells. Cytotoxicity of CAR-engineered ErbB2-specific NK-92/5.28.z cells (filled circles) against ErbB2-overexpressing and trastuzumab-sensitive MDA-MB453 (left), or ErbB2-overexpressing and trastuzumab-resistant JIMT-1 (middle) and CAL-51 (right) breast carcinoma cells was investigated in flow cytometry-based cytotoxicity assays after coincubation of NK cells and tumor cells at different effector to target ratios (E/T) for 2 h. Parental NK-92 cells were included for comparison (open circles). Mean values ± SEM are shown; n = 3.

Figure 3

Reciprocal natural killer (NK)—dendritic cell (DC) cross talk. Upon activation by target tumor cells or cytokines, NK cells produce IFN-γ and tumor necrosis factor (TNF)-α that can promote DC maturation. DC maturation is also strongly dependent on the engagement of activating receptors on NK cells such as NKp30 and NKG2D. Mature DCs (mDCs) will in turn produce interleukin (IL)-12, IL-15, and IL-18, which enhance cytotoxicity and IFN-γ secretion of NK cells. NK cells can also distinguish immature (iDC) and mDCs through activating NKp30 and inhibitory killer cell immunoglobulin-like receptors and NKG2A/CD94 and eliminate immature DCs (iDCs), thereby maintaining the quality of the mDC population (DC editing). NK-cell cytotoxicity can be further augmented by IFN-α secreted by plasmacytoid DCs (pDCs). NK-induced tumor cell lysis provides antigens, which can be taken up by DCs for antigen presentation. Once maturated, antigen-loaded mDCs will migrate into tumor-draining lymph nodes, cross-present tumor antigens to naïve T cells, and induce their differentiation toward tumor-specific CD8⁺ cytotoxic T cells and CD4⁺ T helper 1 (Th1) cells.

Figure 4

Induction of protective antitumor immunity by CAR-engineered NK-92 cells. (A) Murine GL261/ErbB2 glioblastoma cells (5 × 10³) stably expressing human ErbB2 were stereotactically injected into the right striatum of syngeneic C57BL/6 mice. Seven days later, animals were treated once per week for 3 weeks by intratumoral injection of 2 × 10⁶ parental NK-92 (n = 6) or NK-92/5.28.z cells (n = 8), which express an ErbB2-specific CAR with CD28 and CD3ζ signaling domains. Animals that were cured upon NK-92/5.28.z treatment (n = 5) were rechallenged at day 126 by injection of 5 × 10³ GL261/ErbB2 cells into the left brain hemisphere without further NK-cell therapy and symptom-free survival was followed. Naïve C57BL/6 mice injected into the brain with GL261/ErbB2 cells at day 126 served as a control (n = 5) (69). (B) Induction of IgG serum antibodies against glioblastoma cells in NK-92/5.28.z-treated animals (n = 4) from the experiment summarized in (A) was investigated by flow cytometry with GL261/ErbB2 (upper panel) and ErbB2-negative parental GL261 cells (lower panel) using sera collected at day 210. Sera from naïve C57BL/6 mice (n = 2) served as controls. MFI, mean fluorescence intensity (geometric mean). Data for GL261/ErbB2 cells are from Zhang et al. (69). (C) T cells in the NK-92/5.28.z-pretreated and rechallenged mice were depleted by intravenous injection of CD4- and CD8-specific antibodies at days 568, 573, 578, 582, 587, and 592 (left and middle panels). At day 575, these animals were rechallenged a second time with 5 × 10³ GL261/ErbB2 cells injected into the left brain hemisphere without further NK-cell therapy. Tumor development was assessed by MRI at day 595 (right panels).

Figure 5

Inducible caspase-9 (iCasp9) as a safety switch for CAR-engineered NK-92 cells. (A) NK-92 cells transduced with a lentiviral vector that encodes iCasp9 and ErbB2-specific CAR 5.28.z separated by a Thosea asigna virus self-cleaving peptide (T2A) (NK-92/iCasp9_T2A_5.28.z) were incubated in the presence of 10 nM of the homodimerizer AP20187 for iCasp9 activation. Lysates of cells collected after 10, 20, 30, or 60 min of exposure to AP20187 were subjected to SDS-PAGE and subsequent immunoblotting with a caspase-9-specific antibody. Lysates of NK-92/iCasp9_T2A_5.28.z cells kept without dimerizer and parental NK-92 cells incubated in the absence or presence of AP20187 served as controls. (B) Cytotoxicity of NK-92/iCasp9_T2A_5.28.z cells against ErbB2-overexpressing MDA-MB453 breast carcinoma cells was investigated in flow cytometry-based cytotoxicity assays after co-incubation of NK cells and tumor cells at different effector to target ratios (E/T) for 2 h in the absence (filled circles) or presence of AP20187 (open circles). Parental NK-92 cells were included for comparison (gray boxes). Mean values ± SEM are shown; n = 2.