ACCEPTED ARTICLE PREVIEW
Accepted Article Preview: Published ahead of advance Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor
t
Kurt Schönfeld, Christiane Sahm, Congcong Zhang, Sonja Naundorf, Christian Brendel, Marcus Odendahl, Paulina Nowakowska, Halvard Bönig, Ulrike Köhl, Stephan Kloess, Sylvia Köhler, Heidi Holtgreve-Grez, Anna Jauch, Manfred Schmidt, Ralf Schubert, Klaus Kühlcke, Erhard Seifried, Hans G Klingemann, Michael A Rieger, Torsten Tonn, Manuel Grez and Winfried S Wels
m
an
us c
rip
Cite this article as: Kurt Schönfeld, Christiane Sahm, Congcong Zhang, Sonja Naundorf, Christian Brendel, Marcus Odendahl, Paulina Nowakowska, Halvard Bönig, Ulrike Köhl, Stephan Kloess, Sylvia Köhler, Heidi Holtgreve-Grez, Anna Jauch, Manfred Schmidt, Ralf Schubert, Klaus Kühlcke, Erhard Seifried, Hans G Klingemann, Michael A Rieger, Torsten Tonn, Manuel Grez and Winfried S Wels, Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor, Molecular Therapy accepted article preview online 06 November 2014; doi:10.1038/mt.2014.219
A cc
ep te
d
This is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication. NPG is providing this early version of the manuscript as a service to our customers. The manuscript will undergo copyediting, typesetting and a proof review before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers apply
Received 11 July 2014 ; accepted 29 October 2014; Accepted article preview online 06 November 2014
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor
Kurt Schönfeld1,§, Christiane Sahm1,§, Congcong Zhang1,§, Sonja Naundorf2, Christian Brendel1, Marcus Odendahl3, Paulina Nowakowska4, Halvard Bönig4, Ulrike Köhl5, Stephan Kloess5, Sylvia Köhler1, Heidi Holtgreve-Grez6, Anna Jauch6, Manfred Schmidt7, Ralf Schubert8, Klaus Kühlcke2, Erhard Seifried4, Hans G Klingemann9, Michael A Rieger1,10, Torsten Tonn3, Manuel Grez1
Georg-Speyer-Haus, Institute for Tumor Biology and Experimental Therapy, Frankfurt am Main,
us c
1
rip
t
and Winfried S Wels1,*
an
Germany; 2EUFETS GmbH, Idar-Oberstein, Germany; 3Institute for Transfusion Medicine, German Red Cross Blood Donation Service North-East, Dresden and Transfusion Medicine, Medical Faculty Carl
m
Gustav Carus, TU Dresden, Germany; 4Institute for Transfusion Medicine and Immunohematology,
ep te
d
Goethe University, and German Red Cross Blood Donation Service Baden-Württemberg - Hessen, Frankfurt am Main, Germany; 5Institute of Cellular Therapeutics, Hannover Medical School, Hannover,
A cc
Germany; 6Institute of Human Genetics, University of Heidelberg, Heidelberg, Germany; 7National Center for Tumor Diseases and German Cancer Research Center, Department of Translational Oncology, Heidelberg, Germany; 8Pediatric Pulmonology, Allergy and Cystic Fibrosis, University Hospital, Frankfurt am Main, Germany; 9Conkwest Inc., Cardiff-by-the-Sea, CA 92007, USA; 10
§
Department of Hematology/Oncology, University Hospital, Frankfurt am Main, Germany
These authors contributed equally to this work.
Short title:
ErbB2/HER2-specific CAR NK cells
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
*Corresponding author:
Winfried S. Wels Georg-Speyer-Haus, Institute for Tumor Biology and Experimental Therapy Paul-Ehrlich-Straße 42-44 D-60596 Frankfurt am Main, Germany Telephone:
+49-69-63395-188, Telefax: +49-69-63395-189
A cc
ep te
d
m
an
us c
rip
t
e-mail: [email protected]
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
ABSTRACT
Natural killer (NK) cells are an important effector cell type for adoptive cancer immunotherapy. Similar to T cells, NK cells can be modified to express chimeric antigen receptors (CARs) to enhance antitumor activity, but experience with CAR-engineered NK cells and their clinical development is still limited. Here, we redirected continuously expanding and clinically usable established human NK-92 cells to the tumor-associated ErbB2 (HER2) antigen. Following GMP-compliant procedures, we generated a stable
t
clonal cell line expressing a humanized CAR based on ErbB2-specific antibody FRP5 harboring CD28
rip
and CD3ζ signaling domains (CAR 5.28.z). These NK-92/5.28.z cells efficiently lysed ErbB2-
us c
expressing tumor cells in vitro and exhibited serial target cell killing. Specific recognition of tumor cells
an
and antitumor activity were retained in vivo, resulting in selective enrichment of NK-92/5.28.z cells in orthotopic breast carcinoma xenografts, and reduction of pulmonary metastasis in a renal cell carcinoma
m
model, respectively. γ-irradiation as a potential safety measure for clinical application prevented NK cell
ep te
d
replication, while antitumor activity was preserved. Our data demonstrate that it is feasible to engineer CAR-expressing NK cells as a clonal, molecularly and functionally well-defined and continuously
A cc
expandable cell therapeutic agent, and suggest NK-92/5.28.z cells as a promising candidate for use in adoptive cancer immunotherapy.
Key words:
Chimeric antigen receptor; ErbB2/HER2; natural killer cells; NK-92; adoptive cancer immunotherapy
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
INTRODUCTION
Successful application of chimeric antigen receptor (CAR)-modified T cells in patients with CD19positive malignancies has demonstrated the potency of this approach for adoptive cancer immunotherapy,1-4 and CAR T cells targeting a variety of different tumor antigens are under active clinical development.5 CAR-mediated retargeting of natural killer (NK) cells has been attempted less frequently, and so far no clinical data for such an approach are available. NK cells play an important role in cancer immunosurveillance,6-8 and represent an important effector cell type for adoptive cancer
rip
t
immunotherapy.9-11 In contrast to T cells, they do not require prior sensitization and recognition of
us c
peptide antigens presented in complex with MHC molecules. Instead, their cytotoxicity can be triggered
an
rapidly upon appropriate stimulation through germline-encoded cell surface receptors,12,13 that in part signal through CD3ζ. Hence, CD3ζ-containing CARs readily link to endogenous signaling pathways in
m
NK cells and trigger cytolytic activity, as demonstrated for CARs specific for differentiation antigens
ep te
d
expressed by hematologic malignancies,14-17 as well as antigens associated with solid tumors.18-21 Despite these advances, experience with CAR-engineered NK cells and their clinical development is still
A cc
limited. Due to efficient antiviral defense mechanisms, gene transfer into NK cells with retro- and lentiviral vectors as well as physical transfection methods are less efficient than in T cells, complicating the generation of large numbers of CAR-expressing cells.16,22 This restriction can be overcome by employing clinically applicable NK cell lines such as NK-92, which allow isolation and expansion of CAR-expressing cells from a bulk of untransduced cells.18 Phase I studies in cancer patients demonstrated the safety of infusion of unmodified NK-92 cells, which were irradiated prior to application to prevent permanent engraftment. Clinical responses were achieved in a subset of patients.23,24 Similarly, CAR-engineered NK-92 cells may be developed as a targeted allogeneic cell therapeutic agent. Here, we describe the generation and the molecular and functional characterization of a clonal ErbB2-specific NK-92 cell line suitable for clinical applications.
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
These NK-92/5.28.z cells were derived from a single cell clone after lentiviral transduction with a vector encoding a second generation CAR that targets the ErbB2 (HER2) receptor tyrosine kinase, a tumorassociated self-antigen expressed at elevated levels by many human cancers of epithelial origin.25 ErbB2-specific NK-92/5.28.z cells efficiently lysed ErbB2-expressing tumor cells in vitro that were resistant to parental NK-92 cells, and exhibited serial target cell killing. Importantly, specific recognition of ErbB2-positive tumor cells and antitumoral activity were retained in vivo, resulting in selective enrichment of NK-92/5.28.z cells in orthotopic breast carcinoma xenografts, and reduction of pulmonary
rip
t
metastasis of renal cell carcinoma cells in murine models.
an
us c
RESULTS
ErbB2-targeted NK-92 cells display antigen-specific cell killing activity
m
We first compared the activity of NK-92 cells carrying codon-optimized CARs which harbor the ErbB2-
ep te
d
specific scFv(FRP5) antibody fragment26,27 linked to human CD3ζ, or composite CD28-CD3ζ or CD137-CD3ζ signaling domains via a CD8α hinge region (Figure 1a). In all constructs an unpaired
A cc
cysteine within the hinge region was replaced by serine, which increased overall CAR expression in NK-92 cells (Supplementary Figure S1). Upon lentiviral transduction and flow cytometric sorting of NK-92 cells we found similar surface expression of CARs 5.z and 5.28.z, with CAR 5.137.z expressed at lower levels (Figure 1b). All three CARs mediated selective killing of ErbB2-expressing human MDA-MB453 breast carcinoma cells, while more prominent cytotoxicity was observed for NK-92 cells carrying the second generation CARs 5.28.z or 5.137.z (65%, 62% and 51% specific lysis at an E/T ratio of 10:1 for NK-92 cells expressing CARs 5.28.z, 5.137.z or 5.z, respectively) (Figure 1c). ErbB2negative MDA-MB468 cells were not lysed by any of the NK-92 derivatives.
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
NK-92/5.28.z cells derived from a single cell clone retain ErbB2-specific cytotoxicity Due to its stable surface expression and high cytotoxic activity, CAR 5.28.z was chosen as a candidate receptor to generate a clinically applicable ErbB2-specific NK-92 cell line. VSV-G pseudotyped lentiviral CAR vector particles were produced and NK-92 cells from a certified NK-92 master cell bank23 were transduced following GMP-compliant procedures. Single cell clones were derived by limiting dilution, and CAR-expressing cells were identified by flow cytometric analysis with ErbB2-Fc fusion protein. A total of 15 CAR-expressing single cell clones were functionally and molecularly
t
characterized, which harbored between one and four vector copies. One cell clone termed NK-92/5.28.z
rip
which displayed high and stable CAR-expression during continuous culture in a setting reflecting large-
us c
scale expansion under GMP conditions was selected for further analysis (Figure 2a). Linear
an
amplification-mediated PCR (LAM-PCR), DNA sequencing and fluorescence in situ hybridization
on chromosome 9 (Figure 2b).
m
revealed one vector integration each in an intergenic region on chromosome 2, and in the TRAF2 gene
ep te
d
Next, cytotoxic activity of the retargeted cells was evaluated. Clonal NK-92/5.28.z cells displayed high cytotoxicity towards ErbB2-expressing MDA-MB453 cells (86% specific lysis at an E/T
A cc
ratio of 10:1), which were resistant to parental NK-92 (Figure 2c). As observed before, NK-92/5.28.z cells like parental NK-92 failed to lyse ErbB2-negative MDA-MB468 cells included as a control. Nevertheless, MDA-MB468 cells which express the pancarcinoma antigen EpCAM were readily killed by EpCAM-specific NK-92/31.28.z cells,21 demonstrating that enhanced activity of the CAR NK cells against otherwise NK-resistant tumor cells is strictly determined by CAR specificity. Likewise, RencalacZ/ErbB2 murine renal cell carcinoma cells stably expressing human ErbB2 were selectively killed by NK-92/5.28.z cells, while otherwise isogenic Renca-lacZ/EGFR cells expressing epidermal growth factor receptor displayed no enhanced sensitivity to the effector cells (Figure 2d). This indicates that cell killing was indeed mediated by interaction of CAR 5.28.z with its target antigen. In addition to breast carcinoma cells, NK-92/5.28.z also effectively lysed ErbB2-positive ovarian carcinoma and
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
melanoma cells that were resistant to parental NK-92 (Supplementary Figure S2). Coculture of NK92/5.28.z with ErbB2-positive targets induced secretion of IFN-γ, TNF-α, IL-10 and the chemokine MIP-1α, while no measurable amounts of IL-4 and IL-6 were produced by the NK cells (Supplementary Figure S3 and data not shown). Potential reactivity against normal tissues was investigated using primary cells derived from different human tissues as targets. At a relatively high E/T ratio of 10:1, we only observed minimal cytotoxicity of NK-92/5.28.z cells towards lung epithelial cells but no cytotoxicity above background values towards cardiomyocytes, lung fibroblasts and peripheral
rip
t
blood mononuclear cells (Figure 2e).
us c
NK-92/5.28.z cells specifically recognize ErbB2-expressing targets in mixed cultures and are
an
capable of serial target cell killing
Next we investigated selectivity of NK-92/5.28.z cells and kinetics of target cell killing in more detail.
m
Mixtures of tdTOMATO-expressing ErbB2-positive MDA-MB453 and EGFP-expressing ErbB2-
ep te
d
negative MDA-MB468 breast carcinoma cells were incubated with NK-92/5.28.z cells. Cultures were followed by live cell imaging for approximately 7 hours, with phase-contrast and fluorescent images
A cc
taken every 4 min 45 s. Evaluation of serial images of individual microscopic fields revealed multiple brief contacts of single NK-92/5.28.z cells with MDA-MB468 cells (green cells), which remained unaffected by the NK cells and continued to replicate (Figure 3a and Supplementary Movie S1). In contrast, NK-92/5.28.z cells made prolonged contacts with MDA-MB453 cells (red cells) interspersed with the ErbB2-negative targets, followed by cell lysis. Thereby single NK-92/5.28.z cells sequentially attacked and killed multiple ErbB2-positive targets, with cell death indicated by massive membrane blebbing, the appearance of apoptotic bodies and loss of the marker gene signal typically occurring between 1 and 3 hours after initial contact (Figure 3b and Supplementary Movie S2).
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
NK-92/5.28.z cells maintain specific target cell killing upon irradiation In phase I clinical trials with untargeted NK-92, irradiation of cells with 10 Gy prior to infusion had been included as a safety measure to prevent permanent engraftment.23,24 Similar safety measures may be important for clinical use of retargeted NK-92 cells. Hence, we tested the effects of γ-irradiation on growth and cytotoxic activity of clonal NK-92/5.28.z cells. After exposure to 5 or 10 Gy, further replication was prevented and the number of viable NK-92/5.28.z cells declined gradually, with living cells no longer detectable at day 5 (10 Gy) and day 7 (5 Gy), respectively (Figure 4a). To assess effects
t
on cytotoxic activity, NK-92/5.28.z cells irradiated with 10 Gy were cultured for 24 hours and then co-
rip
incubated for two hours with target cells. Similar to untreated NK cells, irradiated NK-92/5.28.z retained
us c
high and specific cytotoxicity towards ErbB2-expressing MDA-MB453 targets (86% specific lysis at an
an
E/T ratio of 10:1) (Figure 4b).
m
Targeted NK-92 cells accumulate in ErbB2-positive breast carcinomas
ep te
d
The potential of NK-92/5.28.z cells to reach established tumors was investigated in an orthotopic breast carcinoma model. MDA-MB453 cells transduced with an EGFP-encoding lentiviral vector were
A cc
implanted into the mammary fat pad of female NSG mice, and allowed to grow until tumors were palpable. Then NK-92/5.28.z and unmodified parental NK-92 cells were labeled with fluorescent DiD labeling reagent, and intravenously injected into the tumor-bearing animals. Twenty-four hours later, tumors were excised, single cell suspensions were prepared, and analyzed for the presence of EGFPexpressing tumor cells and DiD-labeled NK cells. In mice injected with untargeted NK-92, only a few of the NK cells were found in the tumors (Figure 5, upper panel). In contrast, NK-92/5.28.z cells were strongly enriched in MDA-MB453/EGFP xenografts (Figure 5, lower panel). Importantly, we also found conjugates of NK-92/5.28.z and MDA-MB453/EGFP cells in the cell suspensions prepared from the tumors. These data demonstrate that NK-92/5.28.z cells retain target cell specificity in vivo, and are capable of penetrating tissues and reach distant tumor sites.
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
NK-92/5.28.z cells exhibit specific antitumor activity in vivo For evaluation of in vivo antitumor activity we chose an experimental lung metastasis model. NSG mice received intravenous injections of Renca-lacZ/ErbB2 cells, followed by i.v. injections of parental NK-92 or retargeted NK-92/5.28.z cells at days 1 and 3 after tumor cell inoculation. Control mice received PBS. Four weeks after tumor challenge, lungs were excised and tumor nodules on the lung surface were counted. While treatment with parental NK-92 cells did not affect metastasis formation in comparison to
t
PBS-treated controls, retargeted NK-92/5.28.z cells reduced the number of pulmonary tumor nodules in
rip
this experiment by approximately 50% (mean number of lung surface metastases: PBS: 37.7 ± 5.4; NK-
us c
92: 36 ± 5.1; NK-92/5.28.z: 19 ± 2.6; p0.05; *, p
Accepted Article Preview: Published ahead of advance Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor
t
Kurt Schönfeld, Christiane Sahm, Congcong Zhang, Sonja Naundorf, Christian Brendel, Marcus Odendahl, Paulina Nowakowska, Halvard Bönig, Ulrike Köhl, Stephan Kloess, Sylvia Köhler, Heidi Holtgreve-Grez, Anna Jauch, Manfred Schmidt, Ralf Schubert, Klaus Kühlcke, Erhard Seifried, Hans G Klingemann, Michael A Rieger, Torsten Tonn, Manuel Grez and Winfried S Wels
m
an
us c
rip
Cite this article as: Kurt Schönfeld, Christiane Sahm, Congcong Zhang, Sonja Naundorf, Christian Brendel, Marcus Odendahl, Paulina Nowakowska, Halvard Bönig, Ulrike Köhl, Stephan Kloess, Sylvia Köhler, Heidi Holtgreve-Grez, Anna Jauch, Manfred Schmidt, Ralf Schubert, Klaus Kühlcke, Erhard Seifried, Hans G Klingemann, Michael A Rieger, Torsten Tonn, Manuel Grez and Winfried S Wels, Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor, Molecular Therapy accepted article preview online 06 November 2014; doi:10.1038/mt.2014.219
A cc
ep te
d
This is a PDF file of an unedited peer-reviewed manuscript that has been accepted for publication. NPG is providing this early version of the manuscript as a service to our customers. The manuscript will undergo copyediting, typesetting and a proof review before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers apply
Received 11 July 2014 ; accepted 29 October 2014; Accepted article preview online 06 November 2014
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
Selective Inhibition of Tumor Growth by Clonal NK Cells Expressing an ErbB2/HER2-Specific Chimeric Antigen Receptor
Kurt Schönfeld1,§, Christiane Sahm1,§, Congcong Zhang1,§, Sonja Naundorf2, Christian Brendel1, Marcus Odendahl3, Paulina Nowakowska4, Halvard Bönig4, Ulrike Köhl5, Stephan Kloess5, Sylvia Köhler1, Heidi Holtgreve-Grez6, Anna Jauch6, Manfred Schmidt7, Ralf Schubert8, Klaus Kühlcke2, Erhard Seifried4, Hans G Klingemann9, Michael A Rieger1,10, Torsten Tonn3, Manuel Grez1
Georg-Speyer-Haus, Institute for Tumor Biology and Experimental Therapy, Frankfurt am Main,
us c
1
rip
t
and Winfried S Wels1,*
an
Germany; 2EUFETS GmbH, Idar-Oberstein, Germany; 3Institute for Transfusion Medicine, German Red Cross Blood Donation Service North-East, Dresden and Transfusion Medicine, Medical Faculty Carl
m
Gustav Carus, TU Dresden, Germany; 4Institute for Transfusion Medicine and Immunohematology,
ep te
d
Goethe University, and German Red Cross Blood Donation Service Baden-Württemberg - Hessen, Frankfurt am Main, Germany; 5Institute of Cellular Therapeutics, Hannover Medical School, Hannover,
A cc
Germany; 6Institute of Human Genetics, University of Heidelberg, Heidelberg, Germany; 7National Center for Tumor Diseases and German Cancer Research Center, Department of Translational Oncology, Heidelberg, Germany; 8Pediatric Pulmonology, Allergy and Cystic Fibrosis, University Hospital, Frankfurt am Main, Germany; 9Conkwest Inc., Cardiff-by-the-Sea, CA 92007, USA; 10
§
Department of Hematology/Oncology, University Hospital, Frankfurt am Main, Germany
These authors contributed equally to this work.
Short title:
ErbB2/HER2-specific CAR NK cells
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
*Corresponding author:
Winfried S. Wels Georg-Speyer-Haus, Institute for Tumor Biology and Experimental Therapy Paul-Ehrlich-Straße 42-44 D-60596 Frankfurt am Main, Germany Telephone:
+49-69-63395-188, Telefax: +49-69-63395-189
A cc
ep te
d
m
an
us c
rip
t
e-mail: [email protected]
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
ABSTRACT
Natural killer (NK) cells are an important effector cell type for adoptive cancer immunotherapy. Similar to T cells, NK cells can be modified to express chimeric antigen receptors (CARs) to enhance antitumor activity, but experience with CAR-engineered NK cells and their clinical development is still limited. Here, we redirected continuously expanding and clinically usable established human NK-92 cells to the tumor-associated ErbB2 (HER2) antigen. Following GMP-compliant procedures, we generated a stable
t
clonal cell line expressing a humanized CAR based on ErbB2-specific antibody FRP5 harboring CD28
rip
and CD3ζ signaling domains (CAR 5.28.z). These NK-92/5.28.z cells efficiently lysed ErbB2-
us c
expressing tumor cells in vitro and exhibited serial target cell killing. Specific recognition of tumor cells
an
and antitumor activity were retained in vivo, resulting in selective enrichment of NK-92/5.28.z cells in orthotopic breast carcinoma xenografts, and reduction of pulmonary metastasis in a renal cell carcinoma
m
model, respectively. γ-irradiation as a potential safety measure for clinical application prevented NK cell
ep te
d
replication, while antitumor activity was preserved. Our data demonstrate that it is feasible to engineer CAR-expressing NK cells as a clonal, molecularly and functionally well-defined and continuously
A cc
expandable cell therapeutic agent, and suggest NK-92/5.28.z cells as a promising candidate for use in adoptive cancer immunotherapy.
Key words:
Chimeric antigen receptor; ErbB2/HER2; natural killer cells; NK-92; adoptive cancer immunotherapy
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
INTRODUCTION
Successful application of chimeric antigen receptor (CAR)-modified T cells in patients with CD19positive malignancies has demonstrated the potency of this approach for adoptive cancer immunotherapy,1-4 and CAR T cells targeting a variety of different tumor antigens are under active clinical development.5 CAR-mediated retargeting of natural killer (NK) cells has been attempted less frequently, and so far no clinical data for such an approach are available. NK cells play an important role in cancer immunosurveillance,6-8 and represent an important effector cell type for adoptive cancer
rip
t
immunotherapy.9-11 In contrast to T cells, they do not require prior sensitization and recognition of
us c
peptide antigens presented in complex with MHC molecules. Instead, their cytotoxicity can be triggered
an
rapidly upon appropriate stimulation through germline-encoded cell surface receptors,12,13 that in part signal through CD3ζ. Hence, CD3ζ-containing CARs readily link to endogenous signaling pathways in
m
NK cells and trigger cytolytic activity, as demonstrated for CARs specific for differentiation antigens
ep te
d
expressed by hematologic malignancies,14-17 as well as antigens associated with solid tumors.18-21 Despite these advances, experience with CAR-engineered NK cells and their clinical development is still
A cc
limited. Due to efficient antiviral defense mechanisms, gene transfer into NK cells with retro- and lentiviral vectors as well as physical transfection methods are less efficient than in T cells, complicating the generation of large numbers of CAR-expressing cells.16,22 This restriction can be overcome by employing clinically applicable NK cell lines such as NK-92, which allow isolation and expansion of CAR-expressing cells from a bulk of untransduced cells.18 Phase I studies in cancer patients demonstrated the safety of infusion of unmodified NK-92 cells, which were irradiated prior to application to prevent permanent engraftment. Clinical responses were achieved in a subset of patients.23,24 Similarly, CAR-engineered NK-92 cells may be developed as a targeted allogeneic cell therapeutic agent. Here, we describe the generation and the molecular and functional characterization of a clonal ErbB2-specific NK-92 cell line suitable for clinical applications.
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
These NK-92/5.28.z cells were derived from a single cell clone after lentiviral transduction with a vector encoding a second generation CAR that targets the ErbB2 (HER2) receptor tyrosine kinase, a tumorassociated self-antigen expressed at elevated levels by many human cancers of epithelial origin.25 ErbB2-specific NK-92/5.28.z cells efficiently lysed ErbB2-expressing tumor cells in vitro that were resistant to parental NK-92 cells, and exhibited serial target cell killing. Importantly, specific recognition of ErbB2-positive tumor cells and antitumoral activity were retained in vivo, resulting in selective enrichment of NK-92/5.28.z cells in orthotopic breast carcinoma xenografts, and reduction of pulmonary
rip
t
metastasis of renal cell carcinoma cells in murine models.
an
us c
RESULTS
ErbB2-targeted NK-92 cells display antigen-specific cell killing activity
m
We first compared the activity of NK-92 cells carrying codon-optimized CARs which harbor the ErbB2-
ep te
d
specific scFv(FRP5) antibody fragment26,27 linked to human CD3ζ, or composite CD28-CD3ζ or CD137-CD3ζ signaling domains via a CD8α hinge region (Figure 1a). In all constructs an unpaired
A cc
cysteine within the hinge region was replaced by serine, which increased overall CAR expression in NK-92 cells (Supplementary Figure S1). Upon lentiviral transduction and flow cytometric sorting of NK-92 cells we found similar surface expression of CARs 5.z and 5.28.z, with CAR 5.137.z expressed at lower levels (Figure 1b). All three CARs mediated selective killing of ErbB2-expressing human MDA-MB453 breast carcinoma cells, while more prominent cytotoxicity was observed for NK-92 cells carrying the second generation CARs 5.28.z or 5.137.z (65%, 62% and 51% specific lysis at an E/T ratio of 10:1 for NK-92 cells expressing CARs 5.28.z, 5.137.z or 5.z, respectively) (Figure 1c). ErbB2negative MDA-MB468 cells were not lysed by any of the NK-92 derivatives.
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
NK-92/5.28.z cells derived from a single cell clone retain ErbB2-specific cytotoxicity Due to its stable surface expression and high cytotoxic activity, CAR 5.28.z was chosen as a candidate receptor to generate a clinically applicable ErbB2-specific NK-92 cell line. VSV-G pseudotyped lentiviral CAR vector particles were produced and NK-92 cells from a certified NK-92 master cell bank23 were transduced following GMP-compliant procedures. Single cell clones were derived by limiting dilution, and CAR-expressing cells were identified by flow cytometric analysis with ErbB2-Fc fusion protein. A total of 15 CAR-expressing single cell clones were functionally and molecularly
t
characterized, which harbored between one and four vector copies. One cell clone termed NK-92/5.28.z
rip
which displayed high and stable CAR-expression during continuous culture in a setting reflecting large-
us c
scale expansion under GMP conditions was selected for further analysis (Figure 2a). Linear
an
amplification-mediated PCR (LAM-PCR), DNA sequencing and fluorescence in situ hybridization
on chromosome 9 (Figure 2b).
m
revealed one vector integration each in an intergenic region on chromosome 2, and in the TRAF2 gene
ep te
d
Next, cytotoxic activity of the retargeted cells was evaluated. Clonal NK-92/5.28.z cells displayed high cytotoxicity towards ErbB2-expressing MDA-MB453 cells (86% specific lysis at an E/T
A cc
ratio of 10:1), which were resistant to parental NK-92 (Figure 2c). As observed before, NK-92/5.28.z cells like parental NK-92 failed to lyse ErbB2-negative MDA-MB468 cells included as a control. Nevertheless, MDA-MB468 cells which express the pancarcinoma antigen EpCAM were readily killed by EpCAM-specific NK-92/31.28.z cells,21 demonstrating that enhanced activity of the CAR NK cells against otherwise NK-resistant tumor cells is strictly determined by CAR specificity. Likewise, RencalacZ/ErbB2 murine renal cell carcinoma cells stably expressing human ErbB2 were selectively killed by NK-92/5.28.z cells, while otherwise isogenic Renca-lacZ/EGFR cells expressing epidermal growth factor receptor displayed no enhanced sensitivity to the effector cells (Figure 2d). This indicates that cell killing was indeed mediated by interaction of CAR 5.28.z with its target antigen. In addition to breast carcinoma cells, NK-92/5.28.z also effectively lysed ErbB2-positive ovarian carcinoma and
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
melanoma cells that were resistant to parental NK-92 (Supplementary Figure S2). Coculture of NK92/5.28.z with ErbB2-positive targets induced secretion of IFN-γ, TNF-α, IL-10 and the chemokine MIP-1α, while no measurable amounts of IL-4 and IL-6 were produced by the NK cells (Supplementary Figure S3 and data not shown). Potential reactivity against normal tissues was investigated using primary cells derived from different human tissues as targets. At a relatively high E/T ratio of 10:1, we only observed minimal cytotoxicity of NK-92/5.28.z cells towards lung epithelial cells but no cytotoxicity above background values towards cardiomyocytes, lung fibroblasts and peripheral
rip
t
blood mononuclear cells (Figure 2e).
us c
NK-92/5.28.z cells specifically recognize ErbB2-expressing targets in mixed cultures and are
an
capable of serial target cell killing
Next we investigated selectivity of NK-92/5.28.z cells and kinetics of target cell killing in more detail.
m
Mixtures of tdTOMATO-expressing ErbB2-positive MDA-MB453 and EGFP-expressing ErbB2-
ep te
d
negative MDA-MB468 breast carcinoma cells were incubated with NK-92/5.28.z cells. Cultures were followed by live cell imaging for approximately 7 hours, with phase-contrast and fluorescent images
A cc
taken every 4 min 45 s. Evaluation of serial images of individual microscopic fields revealed multiple brief contacts of single NK-92/5.28.z cells with MDA-MB468 cells (green cells), which remained unaffected by the NK cells and continued to replicate (Figure 3a and Supplementary Movie S1). In contrast, NK-92/5.28.z cells made prolonged contacts with MDA-MB453 cells (red cells) interspersed with the ErbB2-negative targets, followed by cell lysis. Thereby single NK-92/5.28.z cells sequentially attacked and killed multiple ErbB2-positive targets, with cell death indicated by massive membrane blebbing, the appearance of apoptotic bodies and loss of the marker gene signal typically occurring between 1 and 3 hours after initial contact (Figure 3b and Supplementary Movie S2).
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
NK-92/5.28.z cells maintain specific target cell killing upon irradiation In phase I clinical trials with untargeted NK-92, irradiation of cells with 10 Gy prior to infusion had been included as a safety measure to prevent permanent engraftment.23,24 Similar safety measures may be important for clinical use of retargeted NK-92 cells. Hence, we tested the effects of γ-irradiation on growth and cytotoxic activity of clonal NK-92/5.28.z cells. After exposure to 5 or 10 Gy, further replication was prevented and the number of viable NK-92/5.28.z cells declined gradually, with living cells no longer detectable at day 5 (10 Gy) and day 7 (5 Gy), respectively (Figure 4a). To assess effects
t
on cytotoxic activity, NK-92/5.28.z cells irradiated with 10 Gy were cultured for 24 hours and then co-
rip
incubated for two hours with target cells. Similar to untreated NK cells, irradiated NK-92/5.28.z retained
us c
high and specific cytotoxicity towards ErbB2-expressing MDA-MB453 targets (86% specific lysis at an
an
E/T ratio of 10:1) (Figure 4b).
m
Targeted NK-92 cells accumulate in ErbB2-positive breast carcinomas
ep te
d
The potential of NK-92/5.28.z cells to reach established tumors was investigated in an orthotopic breast carcinoma model. MDA-MB453 cells transduced with an EGFP-encoding lentiviral vector were
A cc
implanted into the mammary fat pad of female NSG mice, and allowed to grow until tumors were palpable. Then NK-92/5.28.z and unmodified parental NK-92 cells were labeled with fluorescent DiD labeling reagent, and intravenously injected into the tumor-bearing animals. Twenty-four hours later, tumors were excised, single cell suspensions were prepared, and analyzed for the presence of EGFPexpressing tumor cells and DiD-labeled NK cells. In mice injected with untargeted NK-92, only a few of the NK cells were found in the tumors (Figure 5, upper panel). In contrast, NK-92/5.28.z cells were strongly enriched in MDA-MB453/EGFP xenografts (Figure 5, lower panel). Importantly, we also found conjugates of NK-92/5.28.z and MDA-MB453/EGFP cells in the cell suspensions prepared from the tumors. These data demonstrate that NK-92/5.28.z cells retain target cell specificity in vivo, and are capable of penetrating tissues and reach distant tumor sites.
© 2014 The American Society of Gene & Cell Therapy. All rights reserved
ACCEPTED ARTICLE PREVIEW
NK-92/5.28.z cells exhibit specific antitumor activity in vivo For evaluation of in vivo antitumor activity we chose an experimental lung metastasis model. NSG mice received intravenous injections of Renca-lacZ/ErbB2 cells, followed by i.v. injections of parental NK-92 or retargeted NK-92/5.28.z cells at days 1 and 3 after tumor cell inoculation. Control mice received PBS. Four weeks after tumor challenge, lungs were excised and tumor nodules on the lung surface were counted. While treatment with parental NK-92 cells did not affect metastasis formation in comparison to
t
PBS-treated controls, retargeted NK-92/5.28.z cells reduced the number of pulmonary tumor nodules in
rip
this experiment by approximately 50% (mean number of lung surface metastases: PBS: 37.7 ± 5.4; NK-
us c
92: 36 ± 5.1; NK-92/5.28.z: 19 ± 2.6; p0.05; *, p
