ANTICANCER RESEARCH 33: 2525-2534 (2013)

Immunological Characterization of Glioblastoma Cells for Immunotherapy

TAE-YOUNG JUNG1,5,6, YOO-DUK CHOI2,6, YOUNG-HEE KIM5, JE-JUNG LEE3,6, HYUNG-SEOK KIM2,6, JU-SUN KIM7, SANG-KI KIM7, SHIN JUNG1,5,6 and DUCK CHO4,6

Departments of 1Neurosurgery, 2Pathology, 3Hematology, 4Laboratory Medicine, and 5Brain Tumor Research Laboratory, Chonnam National University Hwasun Hospital, Jeollanam-do, Republic of Korea; 6Research Institute of Medical Sciences, Chonnam National University, Gwangju, Republic of Korea; 7Department of Companion & Laboratory Animal Science, Kongju National University, Yesan, Republic of Korea

Abstract. The aim of this study was the immunological CD99 and ERBB2 in U251, U87 and U343 cell lines and characterization of glioblastoma cells. Glioblastoma cell lines tissues. These highly-expressed TAAs such as BIRC5, CD99 were cultured in serum and serum-free neurobasal (NBE) and ERBB2 in glioblastoma tissue could be the targets for medium conditions. These cell lines were characterized by immunotherapy. U87 and U343 cell lines could be useful for flow cytometry, reverse transcription-polymerase chain studying the efficacy of immunotherapy related to various reaction (RT-PCR), western blot and natural killer (NK) cell- TAAs and NK cell immunotherapy. cytotoxicity assays. A previously described NK cell expansion method that uses K562 cells expressing interleukin (IL)-15 Glioblastoma is a lethal malignant brain tumor and the and 4-1 BB Ligand (BBL) (K562-mb15-41BBL) was used. RT- current standard treatment of glioblastoma consists of surgical PCR and western blots for the expression of tumor-associated resection followed by concomitant chemoradiation therapy antigens (TAAs), were carried out in 32 glioblastoma and with temozolomide (1). Even with aggressive treatment, the seven normal brain tissues. U87 and U343 tumor cell lines median survival of patients is 14.6 months. Such failure leads showed increased expression for major histocompatibility many to believe that an aggressive combination of standard complex (MHC)-I and -II molecules. No significant therapies, along with other biologically-based therapies, could differences in the levels of CD133, MHC class I/II, MHC be necessary to improve the survival of patients with class I-related chain A (MICA), MICB, UL16 binding protein glioblastoma. Intra-tumoural immune response occurred 1-3 (ULBP 1-3) expression in these cell lines and in NK cell frequently in glioblastoma and there was a consistent cytotoxicity were observed between serum and NBE response, even after conventional treatment (2). conditions. Regardless of culture conditions, U87 and U343 One of the most promising strategies of immunotherapy cell lines were sensitive to expanded NK cells, with median was dendritic cells (DCs) pulsed with tumor-derived whole- cytotoxicities at 4:1 effector/target ratio of 43.2% and 46.5%, lysate/peptides to initiate T cell-mediated antitumor immunity respectively. In RT-PCR, U343 and U87 showed the (3-5). One study discussed the safety and immunogenicity of expression of most TAAs at a high ratio compared with U251. a novel vaccination with α-type 1 polarized DCs loaded with Western blots demonstrated positive expression for BIRC5, synthetic peptides for glioma-associated antigen epitopes with EPH receptor-A2 (EPHA2), interleukin-13 receptor α2 (IL13RA2), Chitinase 3-like 1 (CHI3L1, also known as human cartilage-glycoprotein-39 or YKL40) and premelanosome Correspondence to: Duck Cho, MD, Ph.D., Department of Laboratory Medicine, Chonnam National University Hwasun protein (PMEL, also known as glycoprotein 100) for recurrent Hospital, 160, Ilsim-ri, Hwasun-eup, Hwasun-gun, Jeollanam-do, malignant glioma, which achieved progression-free status 519-809, Republic of Korea. Tel: +82 613797951, Fax: +82 lasting at least 12 months (3). Human glial tumors express a 613797984, e-mail: [email protected] and Yoo-Duk Choi, MD, variety of tumor-associated antigens (TAAs), which can be the Ph.D., Department of Pathology, Chonnam National University targets for immunotherapy (6-8). Hwasun Hospital, 160, Ilsim-ri, Hwasun-eup, Hwasun-gun, In addition to DC-based immunotherapy, another one of Jeollanam-do, 519-809, Republic of Korea. Tel: +82 622205688, strategies was the use of activated NK cells for Fax: +82 613797099, e-mail: [email protected] immunotherapy of patients with glioblastoma. It was Key Words: Characterization, glioblastoma, immunotherapy, tumor- reported that NK cells recognized and killed human associated antigens, U251, U87, U343 cells. glioblastoma cells with stem cell-like properties and these

0250-7005/2013 $2.00+.40 2525 ANTICANCER RESEARCH 33: 2525-2534 (2013) cells, although resistant to freshly-isolated NK cells, were antibodies, trastuzumab, or fluorescence-tagged isotype control highly susceptible to lysis mediated by both allogeneic and antibody for 15 min on ice. Trastuzumab-stained cells were washed autologous IL-2 (or IL-15)-activated NK cells (9). In recent and incubated with 1×PBS (1% BSA) containing fluorescein isothiocyanate (FITC) -conjugated or phycoerythrin (PE)-conjugated years there has been a remarkable improvement in a method antibody to human IgG1 for 15 min on ice. After washing all samples for highly efficient ex vivo NK cell expansion and activation. with 1xPBS (1% BSA), stained cells were then fixed in 2% Remarkable rates of highly cytotoxic NK cell expansion paraformaldehyde and analyzed by a FACSAria (BD Biosciences, San were reported using the protocol that employs irradiated, Jose, CA,USA). The experiment was repeated twice. Values are genetically engineered K562 cell line modified to express indicated as mean relative fluorescence intensity (MRFI), that is the membrane-bound IL-21 (mbIL21) or membrane bound IL- ratio of the mean of fluorescence intensity of cells stained with the 15 (mbIL15) and 4-1BB ligand (K562-mb15-41BBL) (10- selected mAb and that of the negative control; values of MRFI ≥2 were considered significant. 12). However, little has been reported on the immunological characterization of glioblastoma cell lines for the possible Flow cytometry-based NK cytotoxicity assay. Flow cytometric usefulness of expanded NK cells-based immunotherapy for cytotoxicity assay using Calcein AM-stained expanded NK cells glioblastoma. were performed as previously described (14). Target (T) cells Here, we investigated the expressions of TAAs in (U251, U343, U87) were cultured in serum and serum-free glioblastoma cell lines and human glioblastoma tissues to find neurobasal (NBE) medium conditions. To expand NK cells [effector appropriate targets for immunotherapy and the immunological (E) cells], peripheral blood mononuclear cells were isolated and co- characterization of glioblastoma cell lines for assessing the cultured with 100 Gy-irradiated K562 leukemia cells that had been modified to express 4-1BB ligand and membrane-bound interleukin efficacy of immunotherapy related with various TAAs and (IL)-15 (K562-mb15-41BBL cells), kindly provided by Dr. NK cell immunotherapy. Campana (National University of Singapore, Singapore), for 3 weeks in the presence of IL-2 and IL-15. The target cells and Materials and Methods effector cells were co-cultured at effector-to-target (E:T) ratios ranging from 4:1 to 1:1 for 4 h at 37˚C in a humidified incubator Cell lines and cell culture. Human glioma cell lines U87, U343 and containing 5% CO2. Cells were acquired and the date were analyzed U251 were obtained from the Korean Cell Line Bank, Seoul, Korea, by BD CellQuest™ Pro Software. and from the Brain Tumor Research Center, University of California, San Francisco, CA, USA, respectively. All cell lines Reverse transcription-polymerase chain reaction (RT-PCR) for were routinely grown in Dulbecco’s modified Eagle’s medium TAAs. Total RNA was isolated from U251, U87 and U343 cell lines (DMEM; GibcoBRL, Gaithersburg, MD., USA) supplemented with cultured under serum and NBE conditions, glioblastoma tissues 10% fetal bovine serum (FBS; GibcoBRL) at 37˚C in a humidified (n=32) and seven samples of brain from trauma patients using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). 95% air and 5% CO2 atmosphere, termed here ‘serum’ conditions. All cell lines were also cultured with serum-free neurobasal medium A total of 1μg of RNA was reverse transcribed to synthesize the (Invitrogen, Carlsbad, CA, USA) containing 50 ng/ml of both basic cDNA for RT-PCR. For the first-strand synthesis, 1μg of the purified fibroblast growth factor (FGF; Peprotech, Rocky Hill, NJ, USA) and total RNA was incubated with oligo (dT) (0.5 μg/μl; Promega, epidermal growth factor (EGF; Peprotech) supplemented with N2 Madison, WI, USA) for 5 min at 70˚C, followed by the addition of and B27 supplement without vitamin A and L-glutamine, termed buffer containing 4 μl of 5× Reaction Buffer (Promega), 1μl of dNTP ‘NBE’ conditions. (10 mM each), 3.5 μl of 25 mM MgCl2, 2 μl of RNase Inhibitor (40 Brain tissue specimens were obtained from patients who U/μl; Promega), and 1 μl of reverse transcriptase (200 U/μl; Promega). underwent surgery in the Department of Neurosurgery at Chonnam Reactions with differing numbers of PCR cycles were run for each National University Hwasun Hospital. Tissues were fixed in transcript and β-actin (ACTB) was used as an internal control. formalin and embedded in paraffin or immediately frozen in liquid RT-PCR was carried out for the expression of TAAs (EPHA2, nitrogen. Sections from all samples were evaluated histologically CHI3L1, AIM2, PMEL, MAGEA1, EGFR, ERBB2, TRNAP2, and graded according to the World Health Organization grading ABCC3, SOX2, CD99 type 1, CD99 type 2, IL13RA2 and BIRC5). system by two board-certified neuropathologists (13). The Primers and PCR conditions used for RT-PCR are summarized in Institutional Review Board of Chonnam National University Table I. Hwasun Hospital approved this study. Western blotting. Surgically-resected tumor samples were Flow cytometric analysis for the expression of CD133, Major homogenized with lysis buffer [50 mM Tris (pH 8.0), 5 mM EDTA, histocompatibility complex (MHC) class I and II, and NKG2D ligands 150 mM sodium chloride, 0.5% deoxycholic acid, 0.1% sodium (NKG2DLs). To evaluate cell surface expression of CD133 (Cell dodecyl sulfate (SDS), 1% NP-40, 1 mM phenylmethane- Signaling, Danvers, USA), MHC class I and II (BD Pharmingen, San sulfonylfluoride (PMSF), and 1 mg/ml protease inhibitor cocktail]. Diego, CA, USA), MHC class I-related chain A (MICA), MICB, Protein concentrations were determined using a Bio-Rad protein UL16 binding protein-1 (ULBP-1), ULBP-2 and ULBP-3 (R&D assay kit (Bio-Rad, Hercules, CA, USA). Subsequently, 50ug of the Systems, Minneapolis, MN, USA), glioblastoma cells were harvested whole- cell lysates were separated by 15% SDS- polyacrylamide gel by trypsanization and washed once with ice-cold 1x phosphate electrophoresis (PAGE) and transferred to a polyvinylidene buffered saline PBS [1% bovine serum albumin (BSA)]. Glioma cell difluoride membrane (Pall Corporation, Ann Arbor, MI, USA). The lines (1×105) were then stained with fluorescence-tagged target membrane was then incubated for 2 h at room temperature in a

2526 Jung et al: Immunological Characterization of Glioblastoma Cell Lines

Table I. Primers and polymerase chain reaction (PCR) conditions in reverse transcription PCR for tumor-associated antigens (TAAs).

Target Forward Primers Reverse Primers Initiation Denaturation Annealing Extension

SOX2 5’-AACGGCTCGCCCAC 5’-CAGGGGCAGTGTGCC 54˚C (30 sec) CTACAGC-3’ GTATTTGG-3’ EGFR 5’-ACCTGCGTGAAGAAG 5’-CACATCTCCATCACT 55˚C (30 sec) TGTCC-3’ TATCTCC-3’ ACTB 5’-AATCTGGCACCACACC 5’-CTTCTCCTTAATGTCA TTCTAC-3’ CGCACG-3’ IL13RA2 5’-AGTATGGCTTTCGTTT 5’-TACCGAGCTCGGATCCA GCTTGGC-3’ CTAGT-3’ ABCC3 5’-GCCATCGCACTGGAG 5’-GACCCTGGTGTAGTCCA 56˚C (30 sec) ACTGA-3’ TGATAGTG-3’ ERBB2 5’-TCTGACGTCCATCGT 5’-AGGGCATAAGCTGTGT 57˚C (30 sec) CTCTG-3’ CACC-3’ MAGEA1 5’-GCCTGCTGCCCCTGA 5’-AGGAGAGACCTAGGCA CGAGAG-3’ GGTG-3’ AIM2 5’-GTGTATGCCCC 5’-CCTGGTGACCTAGTC 95˚C 95˚C 72˚C ACCAACC-3’ ATTGG-3’ (5 min) (30 sec) (1 min) CD99 type 1 5’-GTGCGGCTAGCACCATG 5’-TAGTCTCGAGCTATTTC 58˚C (30 sec) GCCCGCGGGGCTG-3’ TCTAAAAGAGTACG-3’ CD99 type 2 5’-GTGCGGCTAGCACCATG 5’-GCTCTAGACCCTAGGT GCCCGCGGGGCTG-3’ CTTCAGCCAT-3’ PMEL 5’-TGGCTCTTGGTCTCAGA 5’-AGGTGCAGTGCTTATGA 59˚C (30 sec) AGA-3’ CTT-3’ TRNAP2 5’-GAGGTGCGAGCCGACA 5’-TCTGTACACACATCACA CAAG-3’ CTC-3’ BIRC5 5’-GGACCACCGCATCTCTA 5’-CTTTGCAATTTTGTTCT 59.8˚C (30 sec) CATT-3’ TGGCT-3’ CHI3L1 5’-CCTGCTCAGCGCAGC 5’-GCTTTTGACGCTTTCCT 60˚C (1 min) ACTGT-3’ GGTC-3’ EPHA2 5’-TCAGCAGCAGCGACTT 5’-CAGTGGCCAGGGAA 62˚C (30 sec) CGAGGCA-3’ GGTGCA-3’

SOX2: SRY (sex determining region Y)-box 2, EGFR: epidermal growth factor receptor, ACTB: actin, beta, IL13RA2: interleukin-13 receptor, alpha 2, ABCC3: ATP-binding cassette, sub-family c (CFTR/MRP), member 3, ERBB2: v-erb-b2 erythroblastic leukemia viral oncogene homolog-2, neuro/glioblastoma-derived oncogene homolog (avian), MAGEA1: melanoma antigen family A,1 (directs expression of antigen MZ2-E), AIM2: absent in melanoma-2, CD99: CD99 molecule, PMEL: premelanosome protein, TRNAP2: transfer RNA proline-2 (anticodon AGG), BIRC5: baculoviral IAP repeat containing-5, CHI3L1: chitinase 3-like-1 (cartilage glycoprotein-39), EPHA2: EPH receptor A2.

solution containing TBST [10 mM TrisCl (pH 8.0), 150 mM NaCl, Results and 0.05% Tween 20] supplemented with 5% non-fat dry milk, and probed overnight at 4˚C with antibodies to BIRC5 (Santa Cruz Expression of CD133, MHC class I and II, and NKG2DL Biotechnology, Santa Cruz, CA, USA), CD99 (Abcam, Abcam, molecules. The expression of CD133, MHC class I, class II Cambridge, United Kingdom), ERBB2 (Cell signaling). The bound and NKG2DL molecules was evaluated by antibodies were visualized with goat anti-rabbit (1:40000; Jackson immunofluorescence analysis in a panel of cells exposed to Immunoresearch, West Grove, PA, USA) or anti-mouse (1:40000, BD Biosciences) conjugated with horseradish peroxidase using serum and NBE conditions. Table II shows the results of enhanced chemiluminescence reagents (ECL, Amersham MRFI values. The NBE conditions induced expression of Bioscience). ACTB was used as an internal control. We used CD133 at slightly higher levels than serum conditions for MultiGauge 3.0 software (LAS-4000) to estimate the expression U251, U87 and U343 cell lines. However, no significant level of BIRC5, CD99 and ERBB2. differences in the levels of CD133, MHC class I/II, MICA/B, ULBP 1-3 expression in these cell lines were Data analysis. Statistical analyses of cytotoxicity of expanded NK observed between serum and NBE conditions. Low or lack cells, depending on the culture media, were carried out to evaluate the statistical significance of the differences between groups using of expression of MHC class I and II was observed in U251 repeated measurements of ANOVA, with a p-value of ≤0.05 cells. U343 cells had significant expression of MHC class I considered to be significant (SPSS version 20.0 for Windows; SPSS, and low or lack of expression of MHC class II. NKG2DLs Inc., Chicago, IL, USA). (MICA/B, ULBP 1-3) were heterogeneously-expressed in

2527 ANTICANCER RESEARCH 33: 2525-2534 (2013)

Table II. The expression of CD133, MHC class I, II and NKG2DL molecules in glioblastoma cell lines.

Cell line Medium Molecule

CD133 MHC I MHC II MICA MICB ULBP-1 ULBP-2 ULBP-3

U251 Serum 1.1 1.0 1.0 1.3 2.8 2.7 1.3 2.1 NBE 1.5 1.5 1.0 1.4 2.8 3.7 1.3 3.1 U87 Serum 1.1 11.4 7.0 1.2 1.0 4.3 1.4 3.2 NBE 1.2 8.0 3.5 3.3 1.7 5.9 2.3 4.2 U343 Serum 1.1 112.7 1.0 13.0 2.0 2.5 7.3 3.2 NBE 1.4 38.8 1.1 16.4 1.7 1.9 5.9 2.8

Values are indicated as mean relative fluorescence intensity (MRFI) that is the ratio between the mean of fluorescence intensity of cells stained with the selected monoclonal antibody and that of the negative control; bold values indicate significant MRFI (≥2). CD133: cluster of differentiation 133; MHC: major histocompatibility complex; MICA: MHC class I-related chain A; MICB: MHC class I-related chain B; ULBP: UL16 binding protein.

U251, U87 and U343 cell lines, and ULBP-1 and ULBP-3 Table III. Different expression of tumor-associated antigens in tumor were more commonly detected. cell lines according to different culture medium.

No. (%) of positivity Expanded NK cell- mediated cytotoxic activity. To determine whether U251, U87 and U343 cell lines affect NK cell Serum NBE cytotoxic activity, assays were performed in duplicate under serum and NBE conditions with E:T ratios from 4:1 to 1:1 for U251 3/14 (21.4) 4/14 (28.6) expanded NK cells (Figure 1). U87 and U343 cell lines were U87 12/14 (85.7) 11/14 (78.6) U343 11/14 (78.6) 12/14 (85.7) sensitive to expanded NK cells, with median cytotoxicities at 4:1 effector/target ratio of 43.2% and 46.5%, respectively. Cytotoxicity was not related to levels of expression of NK receptor ligands such as human leukocyte antigen (HLA) class I/II, NKG2DLs. U87 and U343 cell lines exhibited greater Table IV. Different expression of tumor-associated antigens (TAAs) in glioblastoma tissue and normal brain tissue. expanded NK cell-mediated cytotoxicity than U251 cell lines. Expanded NK cell-mediated cytotoxicity against U87 and No. (%) of positivity U343 cell lines cultured in NBE was higher than that in serum. However, there were no statistical differences between TAA Glioblastoma tissue Normal brain tissue U251, U87 and U343 cell lines (p=0.343, 0.186 and 0.291, IL13RA2 2/32 (6.3) 0/7 (0.0) respectively). CD99 type 2 7/32 (21.9) 0/7 (0.0) PMEL 8/32 (25.0) 0/7 (0.0) Expression of TAAs by RT-PCR. U343 and U87 cell lines TRNAP2 13/32 (40.6) 0/7 (0.0) exhibited expression of most of TAAs studied compared to EPHA2 13/32 (40.6) 6/7 (85.7) U251 cells (Table III). Depending on the culture media, EGFR 17/32 (53.1) 0/7 (0.0) ERBB2 17/32 (53.1) 0/7 (0.0) IL13RA2, CD99 type 2, PMEL, TRNAP2, ERBB2 and MAGEA1 22/32 (68.8) 4/7 (0.0) BIRC5 showed the different expression (Figure 2). Different CD99 type 1 23/32 (71.9) 0/7 (0.0) expression of TAAs in glioblastoma tissues (n=32) and BIRC5 26/32 (81.3) 0/7 (0.0) normal brain tissues (n=7) was compared (Figure 3) and CHI3L1 26/32 (81.3) 3/7 (42.9) described in Table IV. SOX2 29/32 (90.6) 4/7 (57.1) AIM2 29/32 (90.6) 7/7 (100.0) ABCC3 32/32 (100.0) 4/7 (57.1) Western blot assays for ERBB2, CD99 and BIRC5. Western blots for ERBB2, CD99 and BIRC5 were performed to validate the protein expression in U251, U87 and U343 glioblastoma cell lines and glioblastoma tissues. All glioblastoma cell lines showed the positive protein expression positive expression in serum condition of U251 cell line, in for BIRC5 (Figure 4A). CD99 showed the positive expression serum and neurobasal media cultured U87 cell line, and in in serum cultured U251 and U87 cell lines, and in serum and neurobasal media cultured U343 cell line. In 32 glioblastoma neurobasal media cultured U343 cell line. ERBB2 showed the tissues, BIRC5 showed the positive expression in 96.9% (31

2528 Jung et al: Immunological Characterization of Glioblastoma Cell Lines

Figure 2. The expression of tumor-associated antigens (TAAs) in glioblastoma cell lines. Lane 1: U251 in serum, 2: U251 in NBE, 3: U87 in serum, 4: U87 in NBE, 5: U343 in serum, 6: U343 in NBE. For full TAA names see Table I.

out of 32), CD99 in 93.8% (30 out of 32) and ERBB2 in 100% (32 out of 32) (Figure 4B).

Discussion

Once naïve cytotoxic T-cells have differentiated into effector cytotoxic T-lymphocytes (CTLs) after vaccination against tumors, they are able to kill tumor cells expressing the relevant antigens without a requirement for co-stimulation or T-cell help (15, 16). The expression of class I and II MHC molecules has an important role in exploiting CTL responses. In this study, U87 and U343 glioblastoma cell lines showed the increased expression for class I and II MCH Figure 1. Natural killer cell cytotoxic activity for glioblastoma cell lines molecules compared to U251 cells. under both serum and serum-free neurobasal medium (NBE) condition. It is well-known that NK cell cytotoxicity is tightly A: U251, B: U87, C: U343. Effector/Target rato: 4:1, 2:1, 1:1. regulated by a balance between activating and inhibitory

2529 ANTICANCER RESEARCH 33: 2525-2534 (2013)

Figure 3. The expression of tumor-associated antigens (TAAs) in human glioblastoma tissue (A) and normal brain (B). Values in parenthesis indicate the proportion of positive cases. For full TAA names see Table I.

2530 Jung et al: Immunological Characterization of Glioblastoma Cell Lines

signals. NK cells express killer immunoglobulin-like receptors (KIRs), most of which recognize specific corresponding human leukocyte antigen (HLA) class I molecules on target cells, and deliver inhibitory signals (17, 18). NKG2D is an activating receptor of NK cells, mediating cytolytic responses to tumor cells (19). To achieve specific targeting, NKG2D interacts with different such as MICA and MICB, ULBP 1-3, RAET1E and RAET1G (19-21). In this study, we examined whether expanded NK cell cytotoxicity was significantly related to the expression of NK inhibitory receptor ligand (i.e., MHC class I, -II) and NK activating receptor ligand (i.e., NKG2DL) in glioblastoma cell lines. U87 and U343 cell lines were more sensitive to expanded NK cells than U251, whereas U87 and U343 express higher level of MHC class I than U251. The expression of NKG2DLs (MICA/B, ULBP 1-3) in U251, U87 and U343 cell lines was heterogeneous. This result indicates that expanded NK cell cytotoxicity toward glioblastoma cell lines was not related to levels of expression of NK receptor ligands. This result corresponds to previous report that there was little correlation between expression of MHC class I in pediatric solid tumor and breast cancer cells and their susceptibility to expanded NK cell cytotoxicity (22, 23). The tumor sphere is composed of free-floating cluster of tumor cells in a medium lacking adherent substrates but containing necessary growth factors, such as epidermal growth factor and fibroblast growth factor which allow the neural stem cells to form into the characteristic 3-D clusters (24). However, these spheres are not identical to stem cells and only contain a small percent of neural stem cell. In serum- cultured media, U251 cells were found to contain 3.2% CD133+ cells, U87 cells contained 0.9% CD133+ cells and A172 glioma cells 2.1% CD133+ cells (25). A 0.05-3% population possessing cancer stem cell properties was identified in glioma cell line (26, 27). Tumor spheres of U87 cells cultured in serum-free medium exhibited at least a 10% increase of CD133 expression (28). In this study, CD133 expression in the glioma cell lines was a minor fraction of total cells and cells under serum-free conditions expressed CD133 molecules at slightly higher levels than under serum conditions. However, no significant differences in the levels of CD133 were observed between serum and NBE conditions. Glioblastoma stem-like cells are more sensitive to NK cell-mediated lysis relatively to their corresponding serum- cultured glioblastoma cells from the same tumor specimen (29). This study showed the increased sensitivity of serum- free (NBE) cultured U87 and U343 cell lines to expanded NK cells, however no statistical significance was observed. Human glial tumors have been shown to express a wide variety of TAAs (6-8). Many TAAs expressed in glioma are Western blot for v-erb-b2 erythroblastic leukemia viral Figure 4. expressed in other types of cancer: melanoma, breast and oncogene homolog-2 (ERBB2), CD99 molecule (CD99) and baculoviral IAP repeat containing-5 (BIRC5) in glioblastoma cell lines (A) and ovarian cancer and multiple cancer types (7, 30). All of these human glioblastoma tissue (B). antigens elicit CTL responses. Most 20 human glioma cell

2531 ANTICANCER RESEARCH 33: 2525-2534 (2013) lines including U251 cells expressed various TAAs: AIM2, Holtzman MP, Reinhart TA, Whiteside TL, Butterfield LH, CCNB, EPHA2, PMEL, MGAT5, IL13RA2, ERBB2, hTERT, Hamilton RL, Potter DM, Pollack IF, Salazar AM and + MAGEA1, MART1, SART1 and BIRC5, and glioblastoma Lieberman FS: Induction of CD8 T-cell responses against novel glioma-associated antigen peptides and clinical activity by specimens expressed most TAAs as well (7). Other studies vaccinations with {alpha}-type 1 polarized dendritic cells and showed the glioblastoma tissue showed the strong expression polyinosinic-polycytidylic acid stabilized by lysine and of EGFRvIII, PMEL, CHI3L1, HNRPL, SOX11, Ezh2, carboxymethylcellulose in patients with recurrent malignant ABCC3, NELFA, GALT3, ART4, FOSL1, UBE2V and so on glioma. J Clin Oncol 29: 330-336, 2011. (6, 8). When the characteristics of tumor cells are studied, the 4 Prins RM, Soto H, Konkankit V, Odesa SK, Eskin A, Yong WH, different media can influence different genotypic/phenotypic Nelson SF and Liau LM: Gene expression profile correlates with profile properties (31-34). The differences in mRNA T-cell infiltration and relative survival in glioblastoma patients expression appear to depend on the culture medium which vaccinated with dendritic cell immunotherapy. Clin Cancer Res 17: 1603-1615, 2011. impacts expressional and functional characteristic. 5 Wheeler CJ and Black KL: DCVax-Brain and DC vaccines in In this study, U343 and U87 cells showed the most TAAs the treatment of GBM. Expert Opin Investig Drugs 18: 509-519, included IL13LA2, CD99 type 2, PMEL, TRNAP2, EPHA2, 2009. EGFR, ERBB2, MAGEA1, CD99 type1, BIRC5, CHI3L1, 6 Saikali S, Avril T, Collet B, Hamlat A, Bansard JY, Drenou B, SOX2, AIM2 and ABCC3 compared to U251 cells which Guegan Y and Quillien V: Expression of nine tumour antigens showed the expression for EPHA2, BIRC5, SOX2 and in a series of human glioblastoma multiforme: interest of ABCC3. Depending on the culture media, IL13RA2, CD99 EGFRvIII, IL-13Ralpha2, gp100 and TRP-2 for immunotherapy. J Neurooncol 81: 139-148, 2007. type 2, PMEL, TRNAP2, ERBB2 and BIRC5 showed the 7 Zhang JG, Eguchi J, Kruse CA, Gomez GG, Fakhrai H, Schroter different expression within the same cell line. Out of these S, Ma W, Hoa N, Minev B, Delgado C, Wepsic HT, Okada H TAAs, ERBB2, BIRC5 and CD99 showed the positive and Jadus MR: Antigenic profiling of glioma cells to generate expression in most glioblastoma tissues and the negative allogeneic vaccines or dendritic cell-based therapeutics. Clin expression in normal brains. The glycoprotein CD99 is Cancer Res 13: 566-575, 2007. implicated in cell adhesion and migration in human glioma 8 Zhang JG, Kruse CA, Driggers L, Hoa N, Wisoff J, Allen JC, cell and tissue (35). This study could suggest the new Zagzag D, Newcomb EW and Jadus MR: Tumor antigen precursor protein profiles of adult and pediatric brain tumors possibility of CD99 target for future immunotherapy. identify potential targets for immunotherapy. J Neurooncol 88: These highly expressed TAAs such as BIRC5, CD99 and 65-76, 2008. ERBB2 in glioblastoma tissues could be the targets of 9 Castriconi R, Daga A, Dondero A, Zona G, Poliani PL, Melotti synthetic multi-peptides vaccine or dendritic cell A, Griffero F, Marubbi D, Spaziante R, Bellora F, Moretta L, immunotherapy. U87 and U343 cell lines could be useful for Moretta A, Corte G and Bottino C: NK cells recognize and kill studying the efficacy of immunotherapy related with various human glioblastoma cells with stem cell-like properties. J TAAs and NK-cell immunotherapy. Immunol 182: 3530-3539, 2009. 10 Imai C, Iwamoto S and Campana D: Campana, Genetic Acknowledgements modification of primary natural killer cells overcomes inhibitory signals and induces specific killing of leukemic cells. Blood 106: This work was supported by a research grant from the Research 376-383, 2005. Institute of Medical Sciences, Chonnam National University (2012- 11 Cho D and Campana D: Expansion and activation of natural CURIMS-DR003) and by Basic Science Research Program through killer cells for cancer immunotherapy. Korean J Lab Med 29: 89- the National Research Foundation of Korea (NRF) funded by the 96, 2009. Ministry of Education, Science and Technology (2010-0023757). 12 Denman CJ, Senyukov VV, Somanchi SS, Phatarpekar PV, Kopp LM, Johnson JL, Singh H, Hurton L, Maiti SN, Huls MH, References Champlin RE, Cooper LJ and Lee DA: Membrane-bound IL-21 promotes sustained ex vivo proliferation of human natural killer 1 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, cells. PLoS One 7: e30264, 2012. Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, 13. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Jouvet A, Scheithauer BW and Kleihues P: The 2007 WHO Lacombe D, Cairncross JG, Eisenhauer E and Mirimanoff RO: Classification of Tumours of the central nervous system. Acta Radiotherapy plus concomitant and adjuvant temozolomide for Neuropathol 114: 97-109, 2007. glioblastoma. N Engl J Med 352: 987-996, 2005. 14 Jang YY, Cho D, Kim SK, Shin DJ, Park MH, Lee JJ, Shin MG, 2 Kim YH, Jung TY, Jung S, Jang WY, Moon KS, Kim IY, Lee Shin JH, Suh SP and Ryang DW: An improved flow cytometry- MC, and Lee JJ: Tumour-infiltrating T-cell subpopulations in based natural killer cytotoxicity assay involving calcein AM glioblastomas. Kim YH, Jung TY, Jung S, Jang WY, Moon KS, staining of effector cells. Ann Clin Lab Sci 42: 42-49, 2012. Kim IY, Lee MC and Lee JJ: Tumour-infiltrating T-cell 15 Izumoto S: Peptide vaccine. Adv Exp Med Biol 746: 166-177, subpopulations in glioblastomas. Br J Neurosurg 26: 21-27, 2012. 2012. 3 Okada H, Kalinski P, Ueda R, Hoji A, Kohanbash G, Donegan 16 Yamanaka R and Kajiwara K: Dendritic cell vaccines. Adv Exp TE, Mintz AH, Engh JA, Bartlett DL, Brown CK, Zeh H, Med Biol 746: 187-200, 2012.

2532 Jung et al: Immunological Characterization of Glioblastoma Cell Lines

17 Lanier LL: Up on the tightrope: natural killer cell activation and 28 Ajeawung N and Kamnasaran D: An efficient approach to Enrich inhibition. Nat Immunol 9: 495-502, 2008. Glioma Stem Cells from Glioma Cell Lines in Culture. 18 Cho D, Kim SK and Carson WE 3rd: NK cell-based WebmedCentral Oncology 1(9): WMC00557, 2010. immunotherapy for treating cancer: will it be promising? Korean 29 Avril T, Vauleon E, Hamlat A, Saikali S, Etcheverry A, Delmas J Hematol 46: 3-5, 2011. C, Diabira S, Mosser J and Quillien V: Human glioblastoma 19 Bauer S, Groh V, Wu J, Steinle A, Phillips JH, Lanier LL and stem-like cells are more sensitive to allogeneic NK and T cell- Spies T: Activation of NK cells and T cells by NKG2D, a mediated killing compared with serum-cultured glioblastoma receptor for stress-inducible MICA. Science 285: 727-729, 1999. cells. Brain Pathol 22: 159-174, 2012. 20 Cosman D, Mullberg J, Sutherland CL, Chin W, Armitage R, 30 Renkvist N, Castelli C, Robbins PF and Parmiani G: A listing Fanslow W, Kubin M and Chalupny NJ: ULBPs, novel MHC of human tumor antigens recognized by T cells. Cancer Immunol class I-related molecules, bind to CMV glycoprotein UL16 and Immunother 50: 3-15, 2001. stimulate NK cytotoxicity through the NKG2D receptor. 31 Lindley DJ, Roth WJ, Carl SM and Knipp GT: The effects of Immunity 14: 123-133, 2001. media on pharmaceutically relevant transporters in the human 21 Li P, Morris DL, Willcox BE, Steinle A, Spies T and Strong RK: HT-29 adenocarcinoma cell line: does culture media need to be Complex structure of the activating immunoreceptor NKG2D controlled? J Pharm Sci 101: 1616-1630, 2012. and its MHC class I-like ligand MICA. Nat Immunol 2: 443- 32 Pfeifer N, Baston-Bust DM, Hirchenhain J, Friebe-Hoffmann U, 451, 2001. Rein DT, Krussel JS and Hess AP: Selection of the In Vitro 22 Cho D, Shook DR, Shimasaki N, Chang YH, Fujisaki H and Culture Media Influences mRNA Expression of Hedgehog Campana D: Cytotoxicity of activated natural killer cells against Genes, Il-6, and Important Genes regarding Reactive Oxygen pediatric solid tumors. Clin Cancer Res 16: 3901-3909, 2010. Species in Single Murine Preimplantation Embryos. 23 Park YK, Shin DJ, Cho D, Kim SK, Lee JJ, Shin MG, Ryang ScientificWorldJournal 2012: 479315, 2012. DW, Lee JS, Park MH, Yoon JH and Jegal YJ: Interleukin-21 33 Roth WJ, Lindley DJ, Carl SM and Knipp GT: The effects of increases direct cytotoxicity and IFN-gamma production of ex intralaboratory modifications to media composition and cell vivo expanded NK cells towards breast cancer cells. Anticancer source on the expression of pharmaceutically relevant Res 32: 839-846, 2012. transporters and metabolizing genes in the Caco-2 cell line. J 24 Lee J, Kotliarova S, Kotliarov Y, Li A, Su Q, Donin NM, Pharm Sci 101: 3962-3978, 2012. Pastorino S, Purow BW, Christopher N, Zhang W, Park JK and 34 von Kleist S, Chany E, Burtin P, King M and Fogh J: Fine HA: Tumor stem cells derived from glioblastomas cultured Immunohistology of the antigenic pattern of a continuous cell in bFGF and EGF more closely mirror the phenotype and line from a human colon tumor. J Natl Cancer Inst 55: 555-560, genotype of primary tumors than do serum-cultured cell lines. 1975. Cancer Cell 9: 391-403, 2006. 35 Seol HJ, Chang JH, Yamamoto J, Romaqnuolo R, Suh Y, Weeks 25 Qiang L, Yang Y, Ma YJ, Chen FH, Zhang LB, Liu W, Qi Q, Lu A, Aqnihotri S, Smith CA and Rutka JT: Overexpression of N, Tao L, Wang XT, You QD and Guo QL: Isolation and CD99 Increases the Migration and Invasiveness of Human characterization of cancer stem like cells in human glioblastoma Malignant Glioma Cells. Genes Cancer 3: 535-549, 2012. cell lines. Cancer Lett 279: 13-21, 2009. 26 Kondo T, Setoguchi T and Taga T: Persistence of a small subpopulation of cancer stem-like cells in the C6 glioma cell line. Proc Natl Acad Sci USA 101: 781-786, 2004. 27 Patrawala L, Calhoun T, Schneider-Broussard R, Zhou J, Claypool K and Tang DG: Side population is enriched in tumorigenic, stem-like cancer cells, whereas ABCG2+ and Received March 20, 2013 ABCG2- cancer cells are similarly tumorigenic. Cancer Res 65: Revised April 23, 2013 6207-6219, 2005. Accepted April 24, 2013

2533