NK cells – Versatile tools for viral defense and cancer treatment


Correspondence: Dr. Adelheid Cerwenka, Innate Immunity Research Group, German Cancer Research Center, Im Neuenheimer Feld 280, 69120 Heidelberg, Germany

e-mail: a.cerwenka@dkfz.de

Fax: +49-6221-403755


About 150 NK cell researchers met at the German Cancer Research Center (DKFZ) in Heidelberg, Germany from 26–28 September 2012 for the Natural Killer Cell Symposium which was organized by the NK cell study group of the German Society for Immunology (DGfI) and sponsored by the European Journal of Immunology (EJI), the European Federation of Immunological Societies (EFIS) and the DGfI. The meeting was a forum for the discussion of the function and regulation of these fascinating innate immune cells and the opportunities for the transfer of this knowledge to cancer immunotherapy.


NK cells belong to the innate immune system and exert multiple functions among which the rapid elimination of virus-infected or transformed cells is most important. This is achieved by the secretion of cytolytic granules and the production of effector cytokines such as IFN-γ or TNF-α. The activity of NK cells is regulated by a complex interplay of a multitude of activating and inhibitory signals. A comprehensive understanding of the functions of the increasing number of NK subsets, their receptors and ligands is a prerequisite for the utilization of these cells for cancer immunotherapy of cancer. In this meeting report we present recent findings in NK-cell biology and the application of NK cells in clinical settings that were discussed during the 2012 symposium. Since 2008, there has been a tradition to hold these relatively small annual meetings on NK cells at different locations in Germany. This year, the meeting also attracted internationally renowned speakers who were essential for the good spirit and open discussions of unpublished results. The Keynote lecture was delivered by Günter Hämmerling from the DKFZ, Heidelberg who focused on the control of tumor immunity by the microenvironment. For young scientists this meeting was a great opportunity to present their posters in short oral presentations and to discuss their results with the invited speakers at a “Meet the Speakers” lunch.

NK cells in inflammation and infection

Innate lymphoid cells (ILCs) of the ILC22 type protect the intestinal mucosa from infection by secreting IL-22. Marco Colonna (St. Louis, MO, USA) reported that the aryl hydrocarbon receptor (AHR) was essential for the development of ILC22 cells [1]. Ahr−/− mice succumbed to intestinal bacterial infections and showed impaired development of parts of the gut-associated lymphoid tissue. AHR-ligands triggered the Notch signaling pathway which was required for the development and function of NKp46+ ILC22. Colonna also presented evidence for a previously unappreciated heterogeneity among human NKp44+ cells, a population comprised mainly of activated innate lymphoid cells, and he described a novel NKp44+/CD103+ subset of intraepithelial lymphocytes. Chiara Romagnani (Berlin, Germany) identified physiological stimuli, inducing divergent cytokine profiles, in RORγt+ ILCs (ILC22)—while triggering of activating receptors such as NKp44 selectively induced TNF-α, cytokine stimulation preferentially induced IL-22 and GM-CSF expression (unpublished results). Andreas Diefenbach (Freiburg, Germany) reported that instructive signals from the commensal microbiota were essential for the priming of NK cells by non-mucosal mononuclear phagocytes. In germ-free mice, such phagocytes failed to produce type I interferons resulting in a severely compromised NK-cell mediated immunity to viral infections [2]. Dominique Knoll from Christian Bogdan's laboratory in Erlangen, Germany, provided novel insight into the activation of NK cells by neutrophils in a Leishmania infantum infection model [3]. After depletion of neutrophils by antibodies, NK-cell cytotoxicity and IFN-γ production was reduced by about 50%. In the spleens of infected mice NK cells were found in direct contact with neutrophils suggesting that physical conjugation of these cell types may facilitate NK-cell activation. The NK-cell activating effect of neutrophils was independent of IL-12 or IL-18.

The role of NK cells in viral infection mouse models was discussed during subsequent sessions. Ofer Mandelboim (Jerusalem, Israel) elaborated on an elegant way to enhance NK-cell cytotoxicity during influenza virus infection. Ncr1−/− mice succumbed to influenza virus infection indicating the critical role of the activating NK receptor NKp46 for the innate immune response against this virus. Since sialylation of NKp46 is required for its binding to influenza viral hemagglutinin on infected cells and sialylation is counteracted by the viral neuraminidase, the use of neuraminidase inhibitors such as zanamivir enhanced NK cell function in influenza-infected mice [4]. Cytomegalovirus is known to employ a multitude of mechanisms to thwart NK cells as well as cytotoxic T cells. Marina Babić Čač from Stipan Jonjić's group in Rijeka (Croatia) reported on the involvement of a highly abundant transcript (MAT) of mouse cytomegalovirus in the recognition of MHC class I molecules in conjunction with the cytomegalovirus protein m04 by Ly49 receptors [5]. A so far unknown feature of the MAT transcript regulated the differential recognition of MHC-I/m04 complexes by activating and inhibitory Ly49 receptors. Frank Momburg (Heidelberg, Germany) reported on a recently discovered poxvirus ligand for the NK-cell receptors NKp46 and NKp30 [6]. Hemagglutinin molecules encoded by vaccinia or ectromelia virus bind to these receptors but exert differential effects. Hemagglutinin binding to NKp30 inhibited NK-cell activation in a dominant fashion while engagement of NKp46 had an activating effect. Mario Mondelli (Pavia, Italy) studied the modulation of soluble and NK cell-expressed Siglec-7 molecules in patients with chronic hepatitis C virus infection. Siglec-7 is an inhibitory receptor highly expressed on NK cells and monocytes. His findings suggested that Siglec-7 can bind to the major HCV envelope protein E2 and that the NK-cell interaction with E2 leads to an increased shedding of membrane-bound Siglec-7 into the serum. Siglec-7-deficient NK cells occurring in chronic HCV patients displayed a decreased degranulation and IFN-γ secretion in vitro which could contribute to the inability to clear HCV infection (unpublished results).

Figure 1.
Figure 2.
Figure 3.
Figure 4.

NK-cell education and NK-cell receptors

The complex mechanisms of NK-cell education were elaborated upon by Petter Höglund (Stockholm, Sweden). Using statistical models, he compared expression patterns of inhibitory NK-cell receptors in mice and humans. The stochastic expression of mainly 1–3 Ly49 receptors by clonal NK cells creates a complex NK-cell repertoire in mice, which is formed primarily without the influence of MHC I molecules but is fined-tuned by the presence of MHC class I allotypic molecules [7]. MHC I also influences the human KIR repertoire but requires CMV infection to do so (data provided by Beziat and Malmberg), suggesting potential species differences in repertoire formation.

Stefan Leibelt from Alexander Steinle's group (Frankfurt, Germany) reported that a previously uncharacterized mouse C-type lectin-like receptor Clr-f exhibits tissue-restricted expression by gut epithelial cells and inhibits effector responses by NKR-P1G-expressing cells. The presence of NKR-P1G+ lymphocytes coincided with Clr-f expression by the gut epithelial cells suggesting Clr-f's involvement in mucosal immunosurveillance (unpublished results). Doris Urlaub from Carsten Watzl's group (Dortmund, Germany) investigated the crosstalk between activating NK-cell receptors controlling adhesion and cytotoxicity. The conformational activation of the NK-cell adhesion molecule LFA-1 was shown to be regulated by inside-out signaling originating from various activating (e.g., NKG2D), co-activating (e.g., 2B4) or inhibitory (e.g. KIR2DL2/3) receptors. 2B4 triggering induced the high-affinity state of LFA-1 suggesting that co-activation is able to qualitatively modulate NK activity [8]. Jan Spanholtz (Nijmegen, The Netherlands) compared developmental stages of human NK cells in vivo and after ex vivo maturation from CD34+ progenitor NK cells using a stromal cell-free culture system [9]. Ex vivo NK-cell differentiation in this system was able to closely reproduce known in vivo developmental stages and resulted in the generation of highly cytotoxic CD56dim NK cells that can be used for therapeutic purposes. Debanjana Chatterjee from the group of Roland Jacobs (Hanover, Germany) analyzed the interaction of NK cells with umbilical cord-derived mesenchymal stem cells (UCMSCs). She found that UCMSCs suppressed NK-cell cytotoxicity and cytokine production. Notch signaling and the cyclooxygenase-2 enzyme upregulated by NK cells after engagement with UCSMCs were identified as the mediators of NK-cell suppression [10].

Francesco Colucci (Cambridge, UK) studied the interaction of NK cells with invasive fetal placental cells, which might be considered paradigmatic for invasive malignant cells and for allogeneic cells. He reported that maternal NK cells contribute to reproductive success by responding to paternal MHC class I molecules expressed on fetal placental cells. Allorecognition of H-2Kb molecules expressed by fetal cells in BALB/c × BALB.B pregnancies resulted in an appropriate remodeling of the uterine vessels, whereas transgenic H-2Dd inhibited NK-cell activation and vascularization in C57BL/6 mice [11]. In both cases the interaction between maternal NK cells and allogeneic MHC class I molecules affected fetal weight.

NK cells and cancer

The presentation by Mark Smyth (Melbourne, Australia) kicked off the session on NK cells and cancer that represented a major thematic focus of this symposium. Since NK-cell activation is balanced by a variety of receptor-mediated signals that are either activating or inhibitory, it is important to understand how tumor cells can activate NK cells or escape from NK-cell recognition. Smyth described an important role of the NK-cell activating receptor DNAM-1 and its ligand CD155 (PVR) for the suppression of early mouse B lymphoma development. Smyth additionally reported that another receptor for the CD155 ligand, CD96 (Tactile), had the opposite effect to DNAM-1 and promoted carcinogen-induced fibrosarcoma development. Using NLRP3-deficient mice, Smyth also found that the NLRP3 inflammasome was a critical regulator of tumor metastasis and the cross-talk between NK cells and myeloid cells. A unique myeloid population (CD11b+/Gr-1intermediate) accumulated in the absence of NLPR3 in the lung tumor microenvironment and attracted NK cells to the environment by way of the CCL5 and CXCL9 chemokines which, in turn, where associated with rejection of B16 melanoma metastases [12]. Veronika Sexl from Vienna (Austria) reported on studies into the JAK-STAT pathway in mouse NK-cell development and effector functions. She reported that activated STAT1 and STAT3 proteins were regulators of the NK-cell mediated elimination of tumor cells and cytokine secretion, whereas STAT5 was essential for NK-cell development and survival. Stat5 deletion in NK cells resulted in a severe reduction of NK-cell numbers and an increase in melanoma metastases [13].

Eric Vivier from Marseille (France) provided novel insights into the cell biology of the recently described NKp30 ligand, B7-H6 [14]. Using novel anti-B7-H6 monoclonal antibodies he found that B7-H6 is not only expressed on tumor cells but can also be expressed on monocytes and neutrophils after Toll-like receptor triggering (LPS) or by IL-1β. B7-H6 could be found on CD14+/CD16+ monocytes and in a soluble form in the serum of sepsis patients. Vivier also provided evidence of the anti-tumor efficacy of an anti-KIR antibody in a pre-clinical human KIR transgenic mouse model. Nathalie Fiegler from Adelheid Cerwenka's group in Heidelberg (Germany) reported on mechanisms controlling the expression of the B7-H6 molecule, using anti-B7-H6 mAbs generated by the Cerwenka group. She found that histone deacetylase inhibitors modulated B7-H6 expression on tumor cells (unpublished results). Laurence Zitvogel from Villejuif (France) examined the prognostic impact of NK-cell infiltrates in gastrointestinal sarcomas (GISTs). In contrast to soft tissue sarcomas, this rare tumor entity was enriched in tumor-infiltrating NK cells. GIST NK-TILs displayed a homogeneous and fully activated functional phenotype. The author described how the treatment with the c-kit tyrosine kinase inhibitor imatinib modulates the NK-cell topography and activity in GIST. A high density of tumor-infiltrating NK cells correlated with a longer event-free survival and lower metastases occurrence in GIST patients [15]. In addition, the amount of soluble ligands of NKG2D and NKp30 correlated with the prognosis of GIST patients.

Julia Hilpert from Helmut Salih's group in Tübingen (Germany) presented a novel strategy to therapeutically target GITR ligand expression on leukemia cells in a high proportion of AML and CLL patients. Given that GITRL impairs both the direct and Rituximab-induced reactivity of GITR-expressing NK cells [16], the authors used GITR-Ig fusion proteins containing Fc modified to possess an enhanced capacity to stimulate ADCC and therefore to effectively neutralize the harmful GITRL-GITR interaction and target GITRL-expressing leukemia cells for NK cyto-toxicity. Sven Borchmann from the group of Elke Pogge von Strandmann (Cologne, Germany) reported on studies into the anti-tumor activity of the bispecific immunoligand ULBP2-scFvCEA which engages the activating NK receptor NKG2D on one side and carcinoembryonic antigen on the other [17]. CEA is frequently overexpressed in colorectal and other cancers, and the bispecific construct enhanced the NK cell-mediated lysis of CEA+ tumor cells in vitro and suppressed the outgrowth of luciferase-transfected MC38-huCEA cells in vivo.

Andrea Velardi (Perugia, Italy) summarized the clinical outcomes of allogeneic hematopoietic transplantations performed as means to cure leukemia and other hematological malignancies [18]. Velardi's group has established that in HLA-haplotype mismatched (“haploidentical”) transplants the presence of donor-versus-recipient alloreactive NK cells significantly improved event-free survival without causing graft-versus-host disease. Velardi, Ruggeri and coworkers have shown that the early infusion of donor regulatory T (Treg) cells could mitigate an otherwise lethal GvHD effect of high numbers of donor conventional T (Tcon) cells and thus support the recovery of pathogen-specific T cells. Potent NK cell-mediated graft-versus-leukemia effects were maintained in the Treg/Tcon cell-based conditioning regimens. Daniela Pende (Genoa, Italy) reported on in vitro assays and diagnostic algorithms to differentiate rare inherited diseases associated with cytolytic defects and hemophagocytic lymphohistiocytosis (HLH). Pende elaborated on her experience with pediatric leukemia patients transplanted with cells from haploidentical donors. The depletion of TCRα/β+/CD19+ cells from the graft represents a novel and successful transplantation strategy that provides for a more complete pool of hematopoietic stem cells in addition to mature myeloid cells, alloreactive NK cells and pathogen-reactive immune effector cells (NK, TCRγ/δ+ T cells). Using this approach, mature, alloreactive NK cells were available much earlier than when using CD34+ stem cell grafts as in this instance the alloreactive NK cells need to be newly generated [19]. Lena Oevermann from Rupert Handgretinger's group in Tübingen (Germany) emphasized the importance of including KIR genotyping in the donor selection algorithm and, for transplant pediatric ALL patients, using grafts from donors having a high KIR haplotype B-content score since each person possesses an individual pattern of KIR genes assigned to either haplotype A or B, the latter harboring at least one activating KIR receptor [20]. Ulrike Koehl (Hanover, Germany) reported on clinical trials conducted by a German/Swiss consortium in which allogeneic, unstimulated or IL-2-stimulated NK cells were employed for the treatment of patients with high-risk tumours or leukaemia [21]. Unfortunately, high levels of soluble MICA (NKG2D ligand) in neuroblastoma impaired the cytotoxicity of the donor NK cell infusion, but this could be reversed by both a high-dose infusion of activated NK cells or retargeted NK cells.

Concluding remarks

The panel discussion on “Perspectives of NK-cell based immunotherapies” hosted by Adelheid Cerwenka concluded the sessions on cancer. At the beginning of the discussion, Laurence Zitvogel summarized the different strategies of NK cell-based immunotherapies. Mark Smyth then discussed the importance of antibody-based approaches engaging NK cells to exert antibody-dependent cytotoxicity against leukemia and cancer cells. Rather than attempting to treat large established tumors, the premetastatic niche seems to be an attractive target for NK cell-based immunotherapy. Andrea Velardi advocated NK cell-based immunotherapy for various types of leukemia, except acute lymphoblastoid leukemia where NK cells have so far been found to be ineffective. However, it was noted that the severe side effects of GvHD need to be controlled much better. Laurence Zitvogel and Eric Vivier emphasized the need for improved immunoprofiling of different tumor entities by immunohistochemistry or serum biomarkers sensing the activity of NK cells. The participants pointed out that, on a larger scale, it is still unclear which tumor types are significantly infiltrated by NK cells and whether the infiltrated NK cells are functional and able to influence the prognosis of cancer patients. For adoptive transfer regimens, it was concluded that the numbers of infused NK cells, the modes of NK-cell activation ex vivo and the counteraction of NK cell inhibition by tumor-derived soluble factors or through MHC-I/KIR interactions need to be understood better.

To sum up, this symposium continued the great tradition of relatively small meetings on different aspects of NK cells and we are all looking forward to the next meeting of the NK cell study group of the German Society for Immunology (DGfI) that will be organized by Roland Jacobs and colleagues in Hanover and also to the next meeting to be held in Heidelberg in September this year, this time organized by the Society for Natural Immunity.


The authors would like to thank EJI-EFIS and the German Society for Immunology for their generous financial support of this symposium.

URL: www.nk2012-heidelberg.de

NK2013 – 14th Meeting of the Society for Natural Immunity

18–22 September 2013

Heidelberg, Germany

Invited Speakers

Marcus Altfeld, Ragon Institute of MGH, MIT and Harvard, USA

Bibi Bielekova, NINDS, National Institutes of Health, USA

Jack Bui, University of California, San Diego, USA

Michael Caligiuri, The Ohio State Unversity, Columbus, USA

Marco Colonna, Washington University, St. Louis, USA

Sarah Cooley, Masonic Cancer Center, USA

Mariapia Degli-Eposti, Lions Eye Institute, Australia

Andreas Diefenbach, University of Freiburg, Germany

Katharine Hsu, Memorial Sloan-Kettering Cancer Center, NY, USA

Stipan Jonjic, University of Rijeka, Croatia

Rolf Kiessling, Karolinska University Hospital Solna, Sweden

Lewis Lanier, University of California, San Francisco, USA

Hans-Gustaf Ljunggren, Karolinska Institutet, Sweden

Eric Long, NIAID, NIH, USA

Ofer Mandelboim, Hebrew University, Jerusalem, Israel

Alessandro Moretta, University of Genoa, Italy

Christian Münz, University of Zurich, Switzerland

Jordan Orange, Baylor College of Medicine, USA

David Raulet, University of California Berkeley, USA

Hergen Spits, University of Amsterdam, Nederland

Markus Uhrberg, University Düsseldorf, Germany

Andre Veillette, IRCM, Montreal, Canada

Eric Vivier, CNRS-INSERM, Marseille, France

Ray Welsh, University of Massachusetts Medical School, USA

Wayne Yokoyama, Washington University, St. Louis, USA

Laurence Zitvogel, INSERM, U805, Institut Gustave Roussy, France

For further information, visit http://www.nk2013.com