Glioblastoma multiforme (GBM), the most aggressive brain cancer, recurs due to the resistance of glioblastoma stem cells (GSCs) to all standard therapies. This dismal therapy outcome has sparked intense interest in immunotherapy as a means of overcoming one or more of the factors that have limited the impact of available treatments: (a) the aggressive tumor’s rapid growth rate, (b) their molecular heterogeneity and natural tendency to invade critical brain structures, and (c) the tumor regenerative power of a small subset of glioblastoma stem cells (GSCs).​ Using in vitro studies, researchers found that GSCs, but not normal astrocytes, responded to lyzing by healthy allogeneic natural killer cells (NK). However, the vast majority of tumor cells studied to date have powerful immune defenses that allow them to avoid NK cell-mediated cytotoxicity. These include the disruption of receptor-ligand interactions between NK and tumor cells, as well as the release of immunosuppressive cytokines such as transforming growth factor (TGF-) into the microenvironment. It may not be possible to eradicate enough self-renewing GSCs to sustain complete responses, even though NK cells could be protected from evasive GBM tumors. The susceptibility of GSCs to NK cell surveillance in vivo is not well understood.​ Mass cytometry and single-cell RNA sequencing of primary tumor samples revealed that GBM tumor-infiltrating NK cells had an altered phenotype associated with impaired lytic function when compared to the matched peripheral blood NK cells from GBM patients or healthy donors. This immune evasion strategy was attributed to direct cell-to-cell contact between GSCs and NK cells via v integrin-mediated TGF- activation. GSC-engrafted mice were treated with allogeneic NK cells in combination with integrin or TGF-signalling inhibitors, or with TGFBR2-edited allogeneic NK cells, which prevented GSC-induced NK cell dysfunction and tumor growth.​ A preclinical study was conducted in which single-cell analysis used primary GBM tissue from patients who underwent surgery to determine how much NK cells penetrate the active tumor sites and how effectively they eliminate patient-derived GSCs. NK cells have an altered phenotype that correlates with reduced NK cell cytolytic function using an experimental approach that allowed head-to-head comparison of NK cell markers at the single-cell level in peripheral blood and primary tumor specimens from patients with GBM. GSC-induced NK cell dysfunction was completely prevented in a patient-derived xenograft (PDX) orthotopic mouse model of GBM by direct blockade of v integrin or TGF- or by CRISPR gene editing of the TGF- receptor 2 (TGFBR2) on allogeneic NK cells, resulting in effective tumor control. GSCs, which cause the majority of GBM tumor recurrences after therapy, were highly susceptible to NK cell-mediated killing in vitro but evaded NK cell recognition via a mechanism requiring direct v integrin-mediated cell-to-cell contact, leading to TGF- release, and activation by the GCSs. These findings shed light on an important mechanism of NK cell immune evasion by GSCs and point to the v integrin/TGF- axis as a potential therapeutic target in GBM. Methodology Mass cytometry and single-cell RNA sequencing is used to determine if NK cells eliminate GSCs. For flow cytometry, freshly isolated TI-NK, GP-NK, and HC-NK cells were incubated for 20 minutes at room temperature with live/dead aqua and surface markers. As described in that paper, EGFR and SOX9 first identify neoplastic GBM cells, and then SOX2, POU3F2, OLIG2, and SALL2 identified GSCs, while the remaining were defined as mature cells. The cells were separated into single-cell suspensions using accurate for GSCs and trypsin for the attached astrocytes. The cells were then stained for 20 minutes before washing and acquiring by flow cytometry. After 48 hours, supernatants were collected and the secretion of TGF- and MMP2/3/9 was assessed in the supernatant by TGF- 1 ELISA kit or MMP2/3/9 Luminex kit as per the manufacturer's protocol. CRISPR gene editing of primary NK, crRNAs to target CD9, CD103, and CD51 were designed using the integrated DNA technologies predesigned data set. Mice were treated with either cilengitide or galunisertib in the presence or absence of intracranial NK cell injection for the release of TGF- Beta which inhibit the NK cell-mediated killing. All studies were performed in accordance with the Declaration of Helsinki. Results & discussion The K562 cells were used as a positive control because of their marked sensitivity to NK cell-mediated killing due to the lack of expression of HLA class I. Healthy donor NK cells killed GSCs and K562 cells with equal efficiency and much more readily than healthy human astrocytes, which appeared to be less susceptible to NK cell-mediated killing. Flow cytometry was used to analyze the expression of NK cell activating and inhibitory receptor ligands on GSCs. Furthermore, by using single-cell RNA sequencing data, it was found that NK cell activating ligands are also abundantly expressed on non-GSC neoplastic cells, which supports the finding that these cells can be killed by NK cells. The purpose of this study was to determine the contribution of both activating and inhibitory receptors to the NK cell-dependent cytotoxicity against GSCs by using receptor-specific blocking antibodies. Activated HC-NK cells were cocultured with GSCs in the presence or absence of blocking antibodies: anti-NKG2D, anti-DNAM, anti-NKp30, or anti-HLA class I. HC-NK, GP-NK, and TI-NK cells were expressed of NK cell markers in a comparative mass cytometry showing cytotoxicity against GSCs with or without anti-DNAM. These markers are associated with NK cell activation and cytotoxicity, inhibition, and the TGF- pathway. Impact of the research Immunotherapy is a promising treatment for patients with glioblastoma, but it has had limited success so far. Genetically modifying NK cells allowed us to overcome the immunosuppressive environment in the brain, which allowed us to eliminate tumor-regenerating GSCs. NK cell therapies are showing promising results in early trials, and hence similar strategies can be applied to other types of solid tumors as well.