Expression of a CD20-specific chimeric antigen receptor enhances cytotoxic activity of NK cells and overcomes NK-resistance of lymphoma and leukemia cells

Abstract

Despite the clinical success of CD20-specific antibody rituximab, malignancies of B-cell origin continue to present a major clinical challenge, in part due to an inability of the antibody to activate antibody-dependent cell-mediated cytotoxicity (ADCC) in some patients, and development of resistance in others. Expression of chimeric antigen receptors in effector cells operative in ADCC might allow to bypass insufficient activation via FcgammaRIII and other resistance mechanisms that limit natural killer (NK)-cell activity. Here we have generated genetically modified NK cells carrying a chimeric antigen receptor that consists of a CD20-specific scFv antibody fragment, via a flexible hinge region connected to the CD3zeta chain as a signaling moiety. As effector cells we employed continuously growing, clinically applicable human NK-92 cells. While activity of the retargeted NK-92 against CD20-negative targets remained unchanged, the gene modified NK cells displayed markedly enhanced cytotoxicity toward NK-sensitive CD20 expressing cells. Importantly, in contrast to parental NK-92, CD20-specific NK cells efficiently lysed CD20 expressing but otherwise NK-resistant established and primary lymphoma and leukemia cells, demonstrating that this strategy can overcome NK-cell resistance and might be suitable for the development of effective cell-based therapeutics for the treatment of B-cell malignancies.

Fig. 1

Transduction of NK-92 cells with retroviral vector pL-scFv(Leu-16)-ζ. a Schematic representation of the pL-scFv(Leu-16)-ζ-SN construct. The Moloney murine leukemia virus 5′ long terminal repeat (LTR) controls the expression of the chimeric antigen receptor scFv(Leu-16)-ζ which consists of an N-terminal immunoglobulin heavy-chain leader peptide (SP), the CD20-specific single-chain antibody scFv(Leu-16), a Myc-tag, the hinge region of murine CD8α, and the murine CD3ζ chain. The neomycin-resistance gene for G418 selection of transduced cells is driven by the SV40 early promoter. b Left panel: RT-PCR analysis with primers specific for the chimeric receptor sequence of mRNA from parental NK-92 cells (lane 1), NK-92 cells transduced with empty pLXSN vector (lane 2), and NK-92-scFv(Leu-16)-ζ cells (lane 3). The position of the scFv(Leu-16)-ζ DNA fragment is indicated. Right panel: Immunoblot analysis of scFv(Leu-16)-ζ expression. Lysates of NK-92 (lanes 1 and 3) and transduced NK-92-scFv(Leu-16)-ζ cells (lanes 2 and 4) were separated by SDS-PAGE under non-reducing (NR) or reducing (R) conditions. Immunoblot analysis was performed with a CD3ζ chain specific Mab followed by HRP-conjugated anti-mouse antibody and chemiluminescent detection. The positions of endogenous and chimeric CD3ζ proteins and of molecular weight standards (kDa) are indicated. c Surface expression of chimeric scFv(Leu-16)-ζ antigen receptor. After G418 selection of transduced cells, NK-92-scFv(Leu-16)-ζ cells expressing homogeneous levels of the chimeric antigen receptor on their surface were enriched by sorting with Mab 9E10 and goat anti-mouse IgG-coated magnetic beads, and subsequent sorting by FACS with Mab 9E10 and FITC-conjugated secondary antibody as indicated. After each selection step surface expression of scFv(Leu-16)-ζ was determined by flow cytometry using Mab 9E10 and FITC-labeled goat anti-mouse secondary antibody. NK-92 cells transduced with empty pLXSN served as a control

Fig. 2

NK-92-scFv(Leu-16)-ζ cells display enhanced cell killing toward target cells expressing high levels of CD20. a Flow cytometric analysis of CD20 surface expression on Raji, DOHH-2, WSU-NHL, NALM-6, BV173, and primary human B-cells with CD20-specific Mab L27 and PE-conjugated secondary antibody. b Raji, DOHH-2, WSU-NHL, NALM-6, and BV173 target cells expressing different levels of CD20, CD20-negative K562 human erythroleukemia cells and primary human B-cells were stained with PKH67-GL and incubated at effector to target ratios of 1:1 or 10:1 with parental NK-92 (filled squares) or NK-92-scFv(Leu-16)-ζ cells (open squares) for 2 h. For comparison, Raji cells were also incubated with IL-2 activated primary human NK cells (filled circles). Cells were stained with propidium iodide, and dead target cells were quantified as double positive cells by flow cytometry. Mean values of triplicate samples are shown. The standard deviation is indicated by error bars

Fig. 3

Specificity of target cell recognition and cell killing. Raji cells stained with PKH67-GL were incubated with parental NK-92 cells (open bar) or NK-92-scFv(Leu-16)-ζ cells (filled and hatched bars) for 2 h at an effector to target ratio of 10:1 either in the absence of competitor, or in the presence of 15 μg/ml rituximab or an irrelevant control antibody as indicated. Killing of target cells was determined as described in the legend to Fig. 2. Mean values of triplicate samples are shown. The standard deviation is indicated by error bars

Fig. 4

Target-cell killing is mediated via interaction of the chimeric scFv(Leu-16)-ζ antigen receptor with CD20. a Surface expression of human CD20 on NIH3T3 murine fibroblasts stably transfected with a CD20 cDNA construct (NIH3T3-CD20 cells) was determined by flow cytometry using CD20-specific Mab L27 and a secondary PE-labeled antibody. Parental NIH3T3 cells served as a control. b Cytotoxic activity of NK-92 and NK-92-scFv(Leu-16)-ζ toward CD20 expressing NIH3T3-CD20 and CD20-negative NIH3T3 cells at the indicated effector to target ratios was analyzed in a 3 h MTT cytotoxicity assay as described in the methods section. The relative number of viable target cells is expressed in % of untreated controls (set to 100%). Mean values of triplicate samples are shown. The standard deviation is indicated by error bars. c Cytotoxic activity of NK-92 and NK-92-scFv(Leu-16)-ζ toward NIH3T3-CD20 cells was also investigated by microscopical analysis after 16 h. Control cells were incubated in the absence of NK cells, or were treated with 8 μM of the apoptosis-inducing drug staurosporine as indicated. Representative fields are shown

Fig. 5

Selectivity and kinetics of target cell killing. NIH3T3-CD20 cells transduced with a retroviral vector encoding enhanced green fluorescent protein (eGFP) were mixed at a 1:1 ratio with parental NIH3T3 cells and grown overnight. Then NK-92-scFv(Leu-16)-ζ cells were added at an effector to target ratio of 1:1, and microscopic images were taken at 1.5-min intervals for 6.4 h. a Fluorescence microscopic image of eGFP- and CD20-positive NIH3T3-CD20(eGFP), and eGFP- and CD20-negative NIH3T3 cells before addition of NK cells. b–f Microscopic images of the same field taken at the indicated time points after addition of NK cells. As examples, some NIH3T3-CD20(eGFP) and NIH3T3 cells are indicated by black and white arrows, respectively. An NK-92-scFv(Leu-16)-ζ cell is indicated by a red circle in b. The complete image sequence is provided as a QuickTime movie in the supplementary material

Fig. 6

In vivo anti-tumor activity of NK-92-scFv(Leu-16)-ζ cells. About 4 × 10⁶ Raji cells were mixed with PBS, or NK-92-scFv(Leu-16)-ζ or parental NK-92 cells at an effector to target ratio of 5:1 as indicated and immediately injected subcutaneously into the flanks of NOD/SCID γ_c^−/− mice. Tumor growth was followed by caliper measurements. Mice were sacrificed when tumors reached 1.5 cm in diameter or when animals appeared to be in distress

Fig. 7

Cytotoxic activity of NK cells against primary chronic lymphocytic leukemia cells. CLL cells were obtained from routine peripheral blood samples of untreated patients. Mononuclear cells were enriched by density gradient centrifugation, analyzed for expression of CD20 by flow cytometry, and directly used for cytotoxicity experiments. Cytotoxic activity of CD20-specific NK-92-scFv(Leu-16)-ζ and parental NK-92 cells was analyzed as described in the legend of Fig. 2. Mean values of triplicate samples are shown. The standard deviation is indicated by error bars