CORDIS - Wyniki badań wspieranych przez UE

Natural Killer Cell-Based Anti-Cancer Immunotherapies:<br/>Research training in molecular medicine and biotechnology business

Final Report Summary - NATURIMMUN (Natural Killer Cell-Based Anti-Cancer Immunotherapies:Research training in molecular medicine and biotechnology business)

We have investigated basic principles how natural killer (NK) cells control tumors and viral infections and have developed NK cell-based therapies. In a coordinated research approach we gained novel data i) how tumors and viruses shape NK cells and escape NK cell control and ii) on novel ligands of activating NK receptors. iii) We further developed technologies for generation of therapeutic NK cells and for cross-linking NK cells to tumor cells and iv) a humanized mouse model for evaluation of NK cell-based therapies.

i) Shaping the NK cell compartment and immune evasion
NK cells are important components of the immune system involved in immunosurveillance of tumors and virally infected cells. Vice versa, human cancers and certain viruses can have profound effects on and shape the NK cell compartment. In regard of acute myeloid leukemia (AML) we investigated NK cells in patients receiving a novel maintenance therapy with histamine plus IL-2. In our study, AML patients displayed diminished and partly defective NK cells. The therapy strongly induced an immunomodulatory NK cell subtype and helped to restore the NK cell compartment (Cuapio et al., 2016). This provided additional support for the use of this therapy. For human cytomegalovirus (HCMV), a herpes family member, we studied the expansion of an NK cell subtype expressing the activating NKG2C receptor that is triggered by the virus. The results support that variable copy numbers of the NKG2C gene influence the development of different adaptive NK cell subsets in response to viral infection (Muntasell et al., 2016).
Given the importance of NK cells, it is not astonishing that tumors and viruses have further developed a wide array of mechanisms to avoid recognition by NK cells. One of these is the downregulation of human stress-induced ligands recognized by the activating NKG2D receptor. Normally, these ligands appear on the cell surface induced by oncogenic transformation or virus-infection. We identified a novel regulation of the stress ligand ULBP2 that is suppressed by an oncogenic RNA binding protein. Binding of the protein to ULBP2 mRNA decreases its stability and reduces ULBP2 levels on the cell surface. In consequence, tumor cells were protected from NK recognition (Schmiedel et al., 2016b). This suggests that modulation of stress ligands is an important escape mechanism to diminish NK cell recognition of cancer cells. For herpes virus infections we discovered another novel mechanism how the expression of stress ligands is suppressed, i.e. by proteasomal degradation. Consequently, HHV-6B infected cells could evade immune surveillance by NK cells (Schmiedel et al., 2016a). Another unexpected novel tumor evasion mechanism could be shown for colon cancer. NK cell killing was inhibited by the presence of fecal bacteria in the tumor environment. Bacterial proteins interacted with the inhibitory TIGIT receptor on NK cells leading to the inhibition of NK cell cytotoxicity (Gur et al. 2015).

ii) Novel ligands of activating receptors
NK cells are regulated in their function by a balance of signals transmitted via inhibitory and activating receptors. While ligands for inhibitory receptors are well established, ligands bound by certain important activating receptors are still not identified.
In this regard, we have searched for ligands of activating killer cell immunoglobulin-like receptors (KIR) receptors. We studied how HCMV stimulates NK cells via the activating KIR2DS1 receptor. It was necessary to use a clinical strain of HCMV to detect the recognition by KIR2DS1 of a specific class I molecule, HLA-C2. A conformational change of normal HLA-C2 triggered by HCMV was required for KIR2DS1-mediated NK cell activation (van der Ploeg et al., in prep. for Front. Immunol.). Similarly, we have searched for ligands of the activating NKG2C receptor that is involved in adaptive response to HCMV. Sensitive NKG2C+ reporter cells for screening of different HCMV strains, activating and not activating the receptor, have been developed. Potential candidates for NKG2C ligands have been revealed. Another approach to identify novel ligands has been undertaken for another class of activating receptors, the so-called natural cytotoxicity receptor (NCRs).

iii) Generation of therapeutic NK cells and novel reagents to enhance cytotoxicity
Strengthening the NK cell response would be highly desirable for future immunotherapies of cancer. This could be achieved by infusion of ex vivo expanded NK cells, by genetic modification, by reagents cross-linking NK cells to tumor cells, or by a combination of these methods.
In regard of ex vivo expansion of therapeutic NK cells, we have developed a proprietary protocol to expand NK cells using irradiated autologous peripheral blood (PB) mononuclear cells as feeder cells. The system was fully automated for clinical applications (Granzin et al., 2015 and 2016; Delso-Vallejo et al., in prep. for Front. Immunol.). Combined with automated cell isolation this will provide a basis for adoptive immunotherapy. To harmonize the manufacturing of therapeutic NK cell products in future clinical studies we have further collaborated with international partners on quality control and GMP-conform manufacturing protocols (Koehl et al., 2015). In a second approach, we have improved and characterized a proprietary system to generate NK cells from umbilical cord blood (UCB) stem cells (Lehmann et al., 2014). We showed an important role of the transcription factor ZNF683/HOBIT for NK cell differentiation (Post et al., in prep. for Front. Immunol.). In addition, cytotoxicity of UCB-NK cells was investigated against a variety of solid tumors and displayed high cytotoxicity (Veluchamy et al., manuscript in revision).
To enhance NK cell cytotoxicity we have conducted work to retarget NK cells by multispecific antibody reagents as well as by transduction with chimeric antigen receptor (CAR) constructs. We have developed a novel trispecific antibody platform for therapeutic antibodies with the capacity to bind to two different tumor antigens, thus increasing specificity, and cross-linking them via the third arm to NK cells. Improved cytotoxicity in vitro against leukemia cell lines has been shown (Valverde da Silva et al., in prep. for Front. Immunol.).
We have further investigated NK reactivity towards colon carcinoma cells enhanced by clinically approved EGFR antibodies (cetuximab). The cytotoxicity of NK cells for EGFR+ tumor cells was significantly enhanced. This provides a rationale to strengthen NK cell immunotherapy through a combination with cetuximab for metastatic colorectal cancer patients (Veluchamy et al., 2016).

iv) Improvement of preclinical models and evaluation of NK cell-based therapies
The preclinical evaluation of NK cell-based therapies in mouse models is hampered by the fact that human cells and reagents do not fully react with murine immune cells. To circumvent this problem, mouse models with humanized immune system (HIS mice) can be used. Here we aimed to improve an available HIS mouse models for evaluation of NK cell-based therapies.
In this regard, we developed a novel method to boost the inefficient human NK cell development in mice observed after engraftment of human hematopoietic stem cells. The differentiation of NK cells depends on the interplay with myeloid cells and human myeloid cells are poorly reconstituted in HIS mice due to competition with the murine cells. Therefore, we evaluated mice that lack the Flt3 receptor and display reduced murine myeloid differentiation. In these mice, human dendritic cells and consequently human NK cells could be successfully boosted by human Flt3 ligand providing a novel mouse model with increased NK cell numbers (Lopez-Lastra et al., manuscript in prep.). This will be valuable for future evaluations of reagents/cancer therapies involving human NK cells. In the context of DCs we also evaluated immune reconstitution of DCs post stem cell transplantation in patients (Ciocarlie et al., 2013; Elze et al., 2015).
Finally, we evaluated as a therapeutic example the UCB-NK cells in combination with the clinically approved cetuximab in a human solid cancer model. A human colorectal carcinoma cell line was engrafted in HIS mice and the tumor load and survival rate of mice treated with cetuximab and UCB-NK cells monitored. A significant inhibition of tumor growth and improvement of survival rates was observed. These results provide a rational for NK infusion therapies not only for leukemia, but also for solid cancer treatment (Veluchamy, Lopez-Lastra, et al., manuscript in prep.).

Taken together we have progressed in this project according the work plan. Our data will provide the basis for further improvements in the use of NK cells for cancer therapy. Corresponding improved protocols for future clinical NK cell-based tumor therapies can now be developed.
We want to emphasize that we have a joint publication in preparation that will comprise reviews and original articles describing research results from this project. These will be published during the first half of 2017 in Frontiers of Immunology, section Alloimmunity and Transplantation, under the topic "Tailoring NK Cell Receptor-Ligand Interactions: an Art in Evolution" which is hosted by Drs. Gianfranco Pittari, Antoine Toubert, and Ulrike Koehl.