Natural killer (NK) cells inhibit systemic metastasis of glioblastoma cells and have therapeutic effects against glioblastomas in the brain

Abstract

Background: Glioblastoma multiforme (GBM) is characterized by extensive local invasion, which is in contrast with extremely rare systemic metastasis of GBM. Molecular mechanisms inhibiting systemic metastasis of GBM would be a novel therapeutic candidate for GBM in the brain.

Methods: Patient-derived GBM cells were primarily cultured from surgical samples of GBM patients and were inoculated into the brains of immune deficient BALB/c-nude or NOD-SCID IL2Rgamma(null) (NSG) mice. Human NK cells were isolated from peripheral blood mononucleated cells and expanded in vitro.

Results: Patient-derived GBM cells in the brains of NSG mice unexpectedly induced spontaneous lung metastasis although no metastasis was detected in BALB/c-nude mice. Based on the difference of the innate immunity between two mouse strains, NK cell activities of orthotopic GBM xenograft models based on BALB/c-nude mice were inhibited. NK cell inactivation induced spontaneous lung metastasis of GBM cells, which indicated that NK cells inhibit the systemic metastasis. In vitro cytotoxic activities of human NK cells against GBM cells indicated that cytotoxic activity of NK cells against GBM cells prevents systemic metastasis of GBM and that NK cells could be effective cell therapeutics against GBM. Accordingly, NK cells transplanted into orthotopic GBM xenograft models intravenously or intratumorally induced apoptosis of GBM cells in the brain and showed significant therapeutic effects.

Conclusions: Our results suggest that innate NK immunity is responsible for rare systemic metastasis of GBM and that sufficient supplementation of NK cells could be a promising immunotherapeutic strategy for GBM in the brain.

Fig. 1

Brain and metastatic lung tumor formation in an orthotopic xenograft animal model using patient-derived GBM cells. a Pathologic validation of brain and metastatic lung tumors in various orthotopic xenograft animal models (n = 5 for each group). Tumor cells were transplanted into the mouse brain parenchyma. Immunohistochemistry against a cell proliferation marker (Ki-67) was performed. b Metastatic lung tumor formation rate was compared between BALB/c-nude and NSG recipient mouse strains

Fig. 2

NK cell inactivation provokes spontaneous lung metastasis of GBM cells in the brains of BALB/c-nude mice. a Experimental schedule illustrated. b Murine splenic NK cells were analyzed by flow cytometry. NKG2D-posive cell ratio was analyzed and compared. c, d Pathologic validation of brain (c) and metastatic lung tumors (d) in various orthotopic xenograft animal models. Immunohistochemistry against a cell proliferation marker (Ki-67) was performed. e Immunohistochemistry against human specific cytoplasmic antigen (STEM-121) and GBM specific markers (GFAP, Nestin and SOX2) in metastatic lung tumors. f Human-specific cytoplasmic antigen (STEM-121) was co-localized with GBM specific markers (GFAP or Nestin)

Fig. 3

Expression of NK1.1 and MHC class I molecules in orthotopic GBM xenograft tumors. a NK cells were immunostained with a specific antibody against NK1.1 in the orthotopic GBM xenograft tumors. b, c MHC class I molecule expression of patient-derived GBM cells were analyzed by western blotting (b) and immunohistochemistry in the orthotopic GBM xenograft tumors (c). GAPDH = loading control

Fig. 4

Large-scale in vitro expansion human NK cells and their phenotypic characteristics. a The expanded human NK cells were analyzed for various immune cell markers (n = 22). b The fold expansion of NK cells was determined by comparing the numbers of NK cells in the culture before (D0) and after (D14) the expansion (n = 20). c The viability of expanded NK cells was measured through propidium iodide staining (n = 20). Surface expression of activating (d) and inhibitory (e) receptors on expanded NK cells was analyzed by flow cytometry (n = 13). Data = mean + SD for (d) and (e)

Fig. 5

GBM cell lysis effects of expanded human NK cells. a Surface expression of HLA-ABC, HLA-E, NKG2D ligands (MIC-A/B, ULBP-1, and ULBP-2), and DNAM-1 ligands (CD112, CD155) on tumor cells was analyzed by flow cytometry. b In vitro cytotoxicity of expanded NK cells against K562 and U-87 MG cells was determined by ⁵¹Cr-release assay. Data = mean ± SD

Fig. 6

In vivo therapeutic effects of human NK cells against orthotopic GBM xenograft tumors. a Experimental schedule illustrated. b Tumor volumes were determined and compared one week after the last NK cell injection (n = 7 for each group). I.T. = intratumoral transplantation, I.V. = intravenous injection. Data = mean + SE. *p < 0.05, ***p < 0.001, vs. control. c Apoptotic tumor cells in the xenograft tumors were analyzed by the TUNEL assay. TUNEL-positive cells were calculated and compared with the control group. Arrow = TUNEL positive cell, Data = mean + SE. ***p < 0.001, vs. control. d Human NK cells were traced by immunohistochemistry using a CD56-specific antibody (n = 3 for each group). Number of NK cells were calculated and compared with the control group. Arrow = CD56 positive cell, Data = mean + SE. ***p < 0.001, vs. control

Fig. 7

Effects of treatment schedule on the treatment effects of intravenous supplementation of Human NK cells. a Experimental schedule illustrated. b Tumor volumes were determined and compared four weeks after the tumor cell transplantation (n = 7 for each group). Data = mean + SE. ***p < 0.001, vs. control, **p < 0.01, vs human NK cell treated once a week for three weeks