HLA-E protects glioma cells from NKG2D-mediated immune responses in vitro: implications for immune escape in vivo

The nonclassical MHC class I molecule HLA-E is the only known ligand for CD94/NKG2A and CD94/NKG2C expressed on NK and CD8+ alphabeta and gammadelta T cells. HLA-E may transmit either activating signals via CD94/NKG2C or inhibitory signals mediated by CD94/NKG2A. Here we show that HLA-E is expressed at mRNA and protein level in human long-term glioma cell lines, primary ex vivo polyclonal glioblastoma cell cultures and surgical glioblastoma specimens. Furthermore, immunohistochemistry revealed an enhanced in vivo expression of HLA-E in gliomas of lower grades and a massive overexpression in grade IV glioblastomas compared with normal CNS tissue. An immune-inhibitory effect of HLA-E on tumor-specific CTL has already been described. We show that siRNA-mediated silencing of HLA-E or blocking of CD94/NKG2A enables NKG2D-mediated lysis of 51Cr-labeled tumor cells by NK cells. Thus, our study provides the first evidence that expression and interaction of HLA-E on cancer cells with CD94/NKG2A expressed on lymphocytes compromises innate anti-tumor immune responses.     

Pitfalls in the assessment of MGMT expression and in its correlation with survival in diffuse astrocytomas: proposal of a feasible immunohistochemical approach

Immunohistochemical studies showed that O₆-methylguanine-DNA methyltransferase (MGMT) protein expression is negatively associated with survival in glioblastomas treated with alkylating agents in accordance with previous results of methylation-specific PCR. Implementation of this data in routine clinical diagnostics is limited due to often inappropriate study designs, e.g. pooling of tumor entities, WHO grades or primary and secondary glioblastomas, disregard concerning the infiltration zone or various epidemiological factors. The aim of our study was to evaluate MGMT expression and its prognostic value taking into consideration the aforementioned deficiencies. For this, 162 astrocytic tumors WHO II–IV (36 diffuse astrocytomas WHO II, 51 anaplastic astrocytomas, 75 primary glioblastomas) as well as 25 glioblastoma infiltration zones and 19 glioblastoma relapses were analyzed for immunohistochemical MGMT protein expression using tissue microarray technique. Expression of MGMT significantly decreased from WHO grade II (25.6%) to glioblastoma (16.8%, p = 0.01) with lowest levels in grade III tumors (10.2%, II/III p < 0.0001). Significant negative associations of MGMT and survival were detected for WHO grade II and IV (p = 0.003 and 0.013). The optimal cut-off value of MGMT positive nuclei in primary glioblastomas discriminating patients with significantly different survival rates was at 15% (Log–Rank p = 0.0002). Individual relapse tumors showed changes of MGMT expression to a varying degree. The infiltration zone demonstrated a significant increase of MGMT (p < 0.0001). We conclude that immunohistochemical MGMT assessment has potential as a powerful diagnostic tool but analysis should only be performed in a grade dependent manner, before radio-/chemotherapy and with special attention to the infiltration zone of diffuse astrocytomas.     

Expression pattern of the water channel aquaporin-4 in human gliomas is associated with blood-brain barrier disturbance but not with patient survival

Aquaporin-4 (AQP4), the most prominent CNS water channel, is restricted to the glia limitans and astrocytic endfeet. We previously showed the loss of spatial AQP4 expression in glioblastomas and a redistribution across the cell surface. However, opposing AQP4 functions have been described: protective in vasogenic but detrimental in cytotoxic brain edema. Thus, specific AQP4 induction to prevent or reduce vasogenic edema is suggested. To elucidate the AQP4 role in brain tumors, we investigated 189 WHO grade I-IV gliomas by immunohistochemistry and the prognostic significance for patients' survival. In gliomas, a remarkable de novo AQP4 redistribution was observed in comparison with normal CNS tissue. Surprisingly, the highest membraneous staining levels were seen in pilocytic astrocytomas WHO grade I and grade IV glioblastomas, both significantly higher than in WHO grade II astrocytomas. AQP4 up-regulation was associated with brain edema formation; however, no association between survival and WHO grade-dependent AQP4 expression was seen. Hence, AQP4 redistribution may go along with other tumor properties, such as vascular proliferation and resulting blood-brain barrier disturbance, features usually prominent in pilocytic astrocytomas WHO I and glioblastomas WHO grade IV. In summary, our findings question the protective role of AQP4 in vasogenic brain edema. Although AQP4 was associated with brain edema formation, one has to question the suitability of AQP4 induction as a promising approach in vasogenic brain edema prevention and treatment. In addition, our results provide unexpectedly high AQP4 levels in pilocytic astrocytomas and present AQP4 as tumor progression marker in WHO grade II-IV astrocytomas.     

Expression and activity of l-isoaspartyl methyltransferase decrease in stage progression of human astrocytic tumors

Protein l-isoaspartyl methyltransferase (PIMT) functions as a repair enzyme that acts upon damaged proteins bearing abnormal aspartyl residues. We previously reported that PIMT expression and activity are reduced by half in human epileptic hippocampus. Here we investigated PIMT regulation in astrocytic tumors, which are the most common human brain tumors. PIMT expression and enzyme activity were significantly decreased in all grades of human astrocytic tumors. More precisely, PIMT levels were significantly lower by 76% in pilocytic astrocytomas (grade I), 46% in astrocytomas (grade II), 69% in anaplastic astrocytomas (grade III), and a marked 80% in glioblastomas (grade IV) as compared to normal brains. RT-PCR analysis showed that levels of type I PIMT mRNA were up-regulated while those of type II PIMT mRNA were down-regulated in glioblastomas. Furthermore, the reduced PIMT levels correlated closely with a decrease in the number of neuron cells in astrocytic tumors as assessed by measuring the neuron-specific enolase level. Many proteins with abnormal aspartyl residues accumulated in brain tumors and some were specific to individual grades of astrocytic tumors. Similar results were obtained, either by measuring the reduction in PIMT activity and expression or by measuring the formation of abnormal proteins, in an orthotopic rat brain tumor model implanted with invasive CNS-1 glioma cells. The novelty of these findings was to provide the first evidence for a marked reduction of PIMT expression and activity during stage progression of astrocytic tumors in humans.     

Increased expression of podoplanin in malignant astrocytic tumors as a novel molecular marker of malignant progression

Podoplanin (aggrus) is a mucin-like transmembrane sialoglycoprotein that is expressed on lymphatic endothelial cells. Podoplanin is putatively involved in cancer cell migration, invasion, metastasis, and malignant progression and may be involved in platelet aggregation. Previously, we showed upregulated expression of podoplanin in central nervous system (CNS) germinomas, but not in non-germinomatous germ cell tumors, except for parts of immature teratomas in limited numbers. However, little information exists about its role in CNS astrocytic tumors. In this study, 188 astrocytic tumors (30 diffuse astrocytomas, 43 anaplastic astrocytomas, and 115 glioblastomas) were investigated using immunohistochemistry with an anti-podoplanin antibody, YM-1. In 11 of 43 anaplastic astrocytomas (25.6%) and in 54 of 115 glioblastomas (47.0%), podoplanin was expressed on the surface of anaplastic astrocytoma cells and glioblastoma cells, especially around necrotic areas and proliferating endothelial cells. However, the surrounding brain parenchyma was not stained by YM-1. On the other hand, podoplanin expression was not observed in diffuse astrocytoma (0/30: 0%). Furthermore, we investigated the expression of podoplanin using quantitative real-time PCR and Western blot analysis in 54 frozen astrocytic tumors (6 diffuse astrocytomas, 14 anaplastic astrocytomas, and 34 glioblastomas). Podoplanin mRNA and protein expression were markedly higher in glioblastomas than in anaplastic astrocytomas. These data suggest that podoplanin expression might be associated with malignancy of astrocytic tumors.     

Distinct differences in binding capacity to saccharide epitopes in supratentorial pilocytic astrocytomas, astrocytomas, anaplastic astrocytomas, and glioblastomas

We monitored the expression of glycan-binding sites on a panel of 10 biotinylated neoglycoconjugates by means of quantitative computer-assisted microscopy to further study the molecular mechanisms in the extensive infiltration of the surrounding brain parenchyma by most astrocytic tumors. Three distinct histological compartments were analyzed for each of the 108 astrocytic tumors (15 pilocytic astrocytomas (WHO grade I), 25 astrocytomas (WHO grade II), 30 anaplastic astrocytomas (WHO grade III), and 38 glioblastomas (WHO grade IV) included in our series. These compartments were tumors (nonperivascular tumor astrocytes), perivascular tumor astrocytes, and blood vessel walls. Clear differences were observed between the pilocytic and the diffuse astrocytic tumors. Furthermore, malignant progression in the latter category was paralleled by a decrease in cells' ability to bind distinct sugar epitopes, especially the D-GalNAc(alpha1-3)-D-GalNAc-beta1-R determinant of the Forssman pentasaccharide in tumors, the alpha-L-fucose in perivascular tumor areas, and the beta-D-glucose in tumor vessel walls. Markedly, the level of binding site expression for alpha-D-mannose decreased in the tumors, the perivascular tumor areas, and the vessel walls. These glycohistochemical results imply the functional relevance of protein-carbohydrate interactions in this tumor system.     

p53 protein alterations in adult astrocytic tumors and oligodendrogliomas

BACKGROUND: p53 is a tumor suppressor gene implicated in the genesis of a variety of malignancies including brain tumors. Overexpression of the p53 protein is often used as a surrogate indicator of alterations in the p53 gene. AIMS: In this study, data is presented on p53 protein expression in adult cases (>15 years of age) of astrocytic (n=152) and oligodendroglial (n=28) tumors of all grades. Of the astrocytic tumors, 86% were supratentorial in location while remaining 14% were located infratentorially - 8 in the the cerebellum and 13 in the brainstem. All the oligodendrogliomas were supratentorial. MATERIALS AND METHODS: p53 protein expression was evaluated on formalin-fixed paraffin-embedded sections using streptavidin biotin immunoperoxidase technique after high temperature antigen retrieval. RESULTS: Overall 52% of supratentorial astrocytic tumors showed p53 immunopositivity with no correlation to the histological grade. Thus, 58.8% of diffuse astrocytomas (WHO Grade II), 53.8% of anaplastic astrocytomas (WHO Grade III) and 50% of glioblastomas (WHO Grade IV) were p53 protein positive. In contrast, all the infratentorial tumors were p53 negative except for one brainstem glioblastoma. Similarly, pilocytic astrocytomas were uniformly p53 negative irrespective of the location. Among oligodendroglial tumors, the overall frequency of p53 immunopositivity was lower (only 28%), though a trend of positive correlation with the tumor grade was noted - 25% in Grade II and 31.5% in grade III (anaplastic oligodendroglioma). Interestingly, p53 labeling index (p53 LI) did not correlate with the histopathological grade in both astrocytic and oligodendroglial tumors. CONCLUSIONS: Thus, this study gives an insight into the genetic and hence biological heterogeneity of gliomas, not only between astrocytic tumors vs. oligodendrogliomas but also within astrocytic tumors with regard to their grade and location. With p53 gene therapy trials in progress, this will possibly have future therapeutic implications.     

Macrophage migration inhibitory factor (MIF) expression in human malignant gliomas contributes to immune escape and tumour progression

Macrophage migration inhibitory factor (MIF), which inhibits apoptosis and promotes angiogenesis, is expressed in cancers suppressing immune surveillance. Its biological role in human glioblastoma is, however, only poorly understood. We examined in-vivo expression of MIF in 166 gliomas and 23 normal control brains by immunohistochemistry. MIF immunoreactivity was enhanced in neoplastic astrocytes in WHO grade II glioma and increased significantly in higher tumour grades (III–IV). MIF expression was further assessed in 12 glioma cell lines in vitro. Quantitative RT-PCR showed that MIF mRNA expression was elevated up to 800-fold in malignant glioma cells compared with normal brain. This translated into high protein levels as assessed by immunoblotting of total cell lysates and by ELISA-based measurement of secreted MIF. Wild-type p53-retaining glioma cell lines expressed higher levels of MIF, which may be connected with the previously described role of MIF as a negative regulator of wild-type p53 signalling in tumour cells. Stable knockdown of MIF by shRNA in glioma cells significantly increased tumour cell susceptibility towards NK cell-mediated cytotoxicity. Furthermore, supernatant from mock-transfected cells, but not from MIF knockdown cells, induced downregulation of the activating immune receptor NKG2D on NK and CD8⁺ T cells. We thus propose that human glioma cell-derived MIF contributes to the immune escape of malignant gliomas by counteracting NK and cytotoxic T-cell-mediated tumour immune surveillance. Considering its further cell-intrinsic and extrinsic tumour-promoting effects and the availability of small molecule inhibitors, MIF seems to be a promising candidate for future glioma therapy.     

Diagnostic value of WT1 in neuroepithelial tumours

Aims: Currently, clinical trials using WT1 (Wilms tumour gene) peptide vaccines are conducted in haematopoietic malignancies and solid cancers. Single reports showed that the Wilms tumour gene product WT1 is also expressed in astrocytic neoplasms. Our aim was to investigate WT1 expression in a large cohort of various neuroepithelial tumours of different World Health Organization (WHO) grades and in normal central nervous system (CNS) tissue specimens to test its potential value as a diagnostic marker. Methods: Specimens were assessed by RT-PCR, Western blotting and immunohistochemistry. The samples investigated in our study consisted of 334 human neuroepithelial tumours, among those 33 oligodendrogliomas, 219 astrocytomas (including 105 glioblastomas) and 47 ependymomas. Results: Our results showed a de novo WT1 expression in neuroepithelial tumours. In diffuse astrocytomas and ependymomas, WT1 expression increased significantly with the grade of malignancy. In contrast, no significant difference was seen between WHO grade-II and -III oligodendrogliomas. Controlling for WHO grade, the comparison of oligodendrogliomas with ependymal and astrocytic tumours showed higher expression values for the latter. Conclusions: Our study shows that WT1 is expressed de novo in numerous neuroepithelial tumours and increases with the grade of malignancy. These results suggest an important role of WT1 in tumourigenesis and progression in human brain tumours.     

Trk A, B, and C are commonly expressed in human astrocytes and astrocytic gliomas but not by human oligodendrocytes and oligodendroglioma

Neurotrophins regulate the proliferation and differentiation of neurons in the central nervous system via a family of specialized receptors, including TrkA, TrkB, and TrkC. As little is known about their expression or potential role in human glial tissues and glial tumors, we undertook an immunohistochemical analysis of human glia, glioma tissues and cell cultures of glial tumors to characterize the expression of Trk family members (full-length TrkA, TrkB, the truncated form of TrkB, and TrkC). In normal human brain Trk A, B, and C immunoreactivity was found in neurons and some weak staining was also seen astrocytes. No Trk expression was seen on oligodendrocytes. Strong reactivity was seen in reactive astrocytes in a glial scar. In a total of 34 glioma tissue specimens, which included 16 astrocytic tumors (4 low-grade astrocytomas and 12 glioblastomas multiforme) and 15 oligodendrogliomas (8 low-grade and 7 anaplastic) as well as 3 oligoastrocytomas (WHO grade II), TrkA, B, and C immunoreactivity was observed exclusively in specimens from astrocytic gliomas (16/16), but not in any of the oligodendrocytic gliomas (0/15). In the oligoastrocytomas, staining was restricted to the astrocytic component. In the astrocytoma and oligodendroglioma specimens, Trk A, B, and C immunoreactivity was also seen in the surrounding reactive astrocytes. Trk expression was independent of age, sex or histological grade of the investigated tumors. In six primary cell cultures, one derived from human astrocytes and five established from malignant astrocytomas, only TrkA immunoreactivity could be detected, while TrkB (both full-length and truncated isoforms) and TrkC were absent. The TrkA expression in primary cell cultures decreased with continuous cell passaging, and no Trk could be detected in established cell lines derived from glioblastoma. In conclusion, our data suggest that in human glial tissues Trk A, B, and C may be expressed in a lineage-restricted manner, thereby distinguishing between astrocytes and oligodendrocytes in a marker-like fashion. Trk expression, like GFAP expression appears to be increased in activated (reactive)/ neoplastic astrocytes. Received: 15 January 1998 / Revised, accepted: 27 March 1998     

Analysis of the IDH1 codon 132 mutation in brain tumors

A recent study reported on mutations in the active site of the isocitrate dehydrogenase (IDH1) gene in 12% of glioblastomas. All mutations detected resulted in an amino acid exchange in position 132. We analyzed the genomic region spanning wild type R132 of IDH1 by direct sequencing in 685 brain tumors including 41 pilocytic astrocytomas, 12 subependymal giant cell astrocytomas, 7 pleomorphic xanthoastrocytomas, 93 diffuse astrocytomas, 120 adult glioblastomas, 14 pediatric glioblastomas, 105 oligodendrogliomas, 83 oligoastrocytomas, 31 ependymomas, 58 medulloblastomas, 9 supratentorial primitive neuroectodermal tumors, 17 schwannomas, 72 meningiomas and 23 pituitary adenomas. A total of 221 somatic IDH1 mutations were detected and the highest frequencies occurred in diffuse astrocytomas (68%), oligodendrogliomas (69%), oligoastrocytomas (78%) and secondary glioblastomas (88%). Primary glioblastomas and other entities were characterized by a low frequency or absence of mutations in amino acid position 132 of IDH1. The very high frequency of IDH1 mutations in WHO grade II astrocytic and oligodendroglial gliomas suggests a role in early tumor development.     

Differential expression of S100 calcium-binding proteins characterizes distinct clinical entities in both WHO grade II and III astrocytic tumours

The computer-assisted microscopic analysis of Feulgen-stained nuclei enabled us to identify two subgroups of astrocytomas (WHO grade II) and two subgroups of anaplastic astrocytomas (WHO grade III) with significantly distinct clinical outcomes (Decaestecker et al. Brain Pathol 1998; 8: 29-38). The astrocytomas labelled in the present study as typical (TYP-ASTs) behaved clinically like real astrocytomas while atypical astrocytomas (ATYP-ASTs) behaved similarly to anaplastic astrocytomas. The anaplastic astrocytomas that we labelled as typical (TYP-ANAs) behaved clinically like anaplastic astrocytomas while atypical ones (ATYP-ANAs) behaved like glioblastomas. In the present study, we investigate whether some biological characteristics could be evidenced across these four groups of TYP- and ATYP-ASTs and TYP- and ATYP-ANAs. The data show that the levels of expression (immunohistochemically assayed and quantitatively determined by means of computer-assisted microscopy) of vimentin, the glial fibrillary acidic protein and the platelet-derived growth factor-alpha did not differ significantly across these four groups of astrocytic tumours. The level of cell proliferation (determined by means of both the anti-proliferating cell nuclear antigen and the anti-MIB-1 antibodies; P < 0.001 to P < 0.0001) differed very significantly between the astrocytomas and anaplastic astrocytomas, but not between the typical and atypical variants identified in each group. In sharp contrast, the levels of expression of the S100A3 and S100A5 proteins differed markedly in the solid tumour tissue in relation to the astrocytic tumour types and grades. In addition, while the levels of expression of S100A6 did not change in the astrocytic tumour tissue in relation to histopathological grade, the levels of expression of this S100 protein (but not those of S100A3 and S100A5) differed markedly in the blood vessel walls according to whether these vessels originated from low- or high-grade astrocytic tumours.     

Immunohistochemical expression of P-glycoprotein and glutathione S-transferases in cerebral gliomas and response to chemotherapy

The expression of the drug resistance-related proteins glutathione S-transferases (GST) and P-glycoprotein (Pgp) was analyzed quantitatively in samples of 53 astrocytic gliomas (eight WHO grade 1, 11 WHO grade 2, 9 WHO grade 3 and 25 glioblastomas, WHO grade 4). Sections of these tumors were immunohistochemically stained with antibodies to Pgp (MDR1-gene product) and to GST subclasses alpha, mu and pi. Pgp expression was not detected in tumor cells of the majority of low-grade astrocytomas (69%) and the percentage of Pgp stained cells generally increased with tumor grade. However, 4 of the 34 malignant gliomas were negative. Many neoplastic cells of most tumors were dominantly stained for GST-pi. The other two subclasses were expressed in a less consistent fashion with no linear correlation to grading. Grade 2 astrocytomas exhibited the highest percentage of cells with GST expression. GST-alpha was absent in 9 tumors, GST-mu in 8 and GST-pi in 4. Four tumors showed no expression of any GST subclass or Pgp in neoplastic cells. Of 13 patients 5 with a more favorable clinical course after radiation and chemotherapy had a lower percentage of neoplastic cells immunostained for Pgp and the three GST subclasses than 8 patients with a worse clinical course. These results suggest a relationship between expression of drug resistance-related proteins in gliomas and response to chemotherapy with ACNU/VM26. Received: 18 November 1996 / Revised: 5 May 1997 / Revised, accepted: 16 June 1997     

Detection of IDH1 mutations in gliomatosis cerebri, but only in tumors with additional solid component: evidence for molecular subtypes

The current WHO classification of brain tumors defines gliomatosis cerebri (GC) as an extensively infiltrating astrocytic glioma involving at least three cerebral lobes. The relation of GC to diffuse astrocytomas and glioblastoma is uncertain. Due to malignant biological behavior, GC is allotted to WHO grade III. Recent reports showed IDH1 mutations in astrocytic and oligodendroglial tumors WHO grades II and III and in secondary glioblastomas with a frequency of up to 90%, whereas IDH1 mutations occurred in only 5% of primary glioblastomas. Here, we examined the frequency of IDH1 mutations in 35 GC samples by direct sequencing, derived cleaved amplified polymorphic sequence analysis and immunohistochemistry. We identified IDH1 mutations in 10/24 (42%) cases, which also included a solid tumor portion (type 2 GC), but not in 11 “classical” cases without solid tumor mass (type 1 GC). TP53 mutations were revealed in two type 2 GC, but not in any type 1 GC, while combined chromosomal losses of 1p and 19q were not found at all. Our data suggest that GC consists of two histological/molecular subtypes, type 1 being clearly distinct from diffuse astrocytoma, and type 2 sharing features with diffuse astrocytoma.     

Fascin expresion in glial tumors and its prognostic significance in glioblastomas

Fascin is a ‐55 kDa‐actin binding protein. Actin bundles rearranged by fascin proteins are concentrated in cell membrane protrusions and these protrusions provide motility of the cell. In this study, we evaluated fascin expression in glial tumors and its relation with histologic grade. Its prognostic value in glioblastomas (GBs) was also investigated. Seventy‐six glial tumors including 44 glioblastomas with known survival time, 18 anaplastic astocytomas (AAs), six diffuse astrocytomas (DAs), and eight pilocytic astrocytomas (PAs) were examined immunohistochemically for fascin expression. Fascin was observed in the neurons of normal brain tissue and endothelium of vascular spaces in the glial tumors. Fascin expression was correlated with histologic grade in DAs. PAs expressed low levels of fascin. Half of the GBs showed high levels of fascin expression. In the GB group, overall survival was poor for cases with percentage of stained cells >50% having moderate or strong staining intensity. In GBs, overall survival was also poor for >50‐year‐old cases and cases that refused radiotherapy. Multivariate Cox regression analysis revealed that age (>50 years, P = 0.021) and higher level of fascin expression (immunohistochemical score >8, P = 0.040) were independent poor prognostic factors. In conclusion, fascin expression levels are correlated with histologic grade and fascin overexpression may play an important role in the biologic behavior of glial astrocytic tumors and in the prognosis of GBs.     

HLA-E contributes to an immune-inhibitory phenotype of glioblastoma stem-like cells

Cancer stem cells are an attractive target for immunotherapeutic approaches to glioblastoma. However, an immune inhibitory phenotype of cells currently classified as "glioma-initiating cells" (GIC) might counteract recognition by immune effector cells. Here, we investigate the contribution of the non-classical MHC molecule HLA-E to the immunosuppressive phenotype of GIC. HLA-E is expressed in GIC lines and its expression is reduced upon differentiation of GIC in serum-containing culture conditions. Constitutive HLA-E inhibits natural killer (NK) cell-mediated lysis of GIC since small-interfering RNA-mediated HLA-E gene silencing enhances the immunogenicity of GIC. Increased GIC lysis was observed both in the CD133+ and in the CD133- compartment. Furthermore, the use of interferon-γ as a possible agent to boost an immune response against glioblastoma cells might be limited by the concurrent upregulation of HLA-E.