Expression Profile of Sonic Hedgehog Pathway Members in the Developing Human Fetal Brain.

The Sonic Hedgehog (SHH) pathway plays a central role in the developing mammalian CNS. In our study, we aimed to investigate the spatiotemporal SHH pathway expression pattern in human fetal brains. We analyzed 22 normal fetal brains for Shh, Patched, Smoothened, and Gli1-3 expression by immunohistochemistry. In the telencephalon, strongest expression of Shh, Smoothened, and Gli2 was found in the cortical plate (CP) and ventricular zone. Patched was strongly upregulated in the ventricular zone and Gli1 in the CP. In the cerebellum, SHH pathway members were strongly expressed in the external granular layer (EGL). SHH pathway members significantly decreased over time in the ventricular and subventricular zone and in the cerebellar EGL, while increasing levels were found in more superficial telencephalic layers. Our findings show that SHH pathway members are strongly expressed in areas important for proliferation and differentiation and indicate a temporal expression gradient in telencephalic and cerebellar layers probably due to decreased proliferation of progenitor cells and increased differentiation. Our data about the spatiotemporal expression of SHH pathway members in the developing human brain serves as a base for the understanding of both normal and pathological CNS development.

Participation in and support of clinical studies and other scientific investigations - Statement of the German Society for Pathology.

Clinical studies and preclinical investigations are essential in order to test new therapies and diagnostic techniques aimed at sustainable improvements in the treatment of patients. Fortunately, the number of clinical studies is continuously increasing and pathology and tissue-based research are more frequently involved. Pathologists are essential in this process and committed to support it by joining forces with our clinical partners. The investigative diagnostic technologies we apply to human cell and tissue samples and our specific expertise are essential contributions to the quality and success of preclinical investigations, clinical studies, and the implementation of results into clinical diagnostic pathology. In order to support this process, the German Society of Pathology has formulated a statement on the participation in and support of clinical studies and other scientific investigations with a special focus on tissue-based research.

Assessing CpG island methylator phenotype, 1p/19q codeletion, and MGMT promoter methylation from epigenome-wide data in the biomarker cohort of the NOA-04 trial.

BACKGROUND: Molecular biomarkers including isocitrate dehydrogenase 1 or 2 (IDH1/2) mutation, 1p/19q codeletion, and O(6)-methylguanine-DNA-methyltransferase (MGMT) promoter methylation may improve prognostication and guide treatment decisions for patients with World Health Organization (WHO) anaplastic gliomas. At present, each marker is individually tested by distinct assays. Illumina Infinium HumanMethylation450 BeadChip arrays (HM450) enable the determination of large-scale methylation profiles and genome-wide DNA copy number changes. Algorithms have been developed to detect the glioma CpG island methylator phenotype (G-CIMP) associated with IDH1/2 mutation, 1p/19q codeletion, and MGMT promoter methylation using a single assay. METHODS: Here, we retrospectively investigated the diagnostic and prognostic performance of these algorithms in comparison to individual marker testing and patient outcome in the biomarker cohort (n = 115 patients) of the NOA-04 trial. RESULTS: Concordance for IDH and 1p/19q status was very high: In 92% of samples, the HM450 and reference data agreed. In discordant samples, survival analysis by Kaplan-Meier and Cox regression analyses suggested a more accurate assessment of biological phenotype by the HM450 analysis. The HM450-derived MGMT-STP27 model to calculate MGMT promoter methylation probability revealed this aberration in a significantly higher fraction of samples than conventional methylation-specific PCR, with 87 of 91 G-CIMP tumors predicted as MGMT promoter-methylated. Pyrosequencing of discordant samples confirmed the HM450 assessment in 14 of 17 cases. CONCLUSIONS: G-CIMP and 1p/19q codeletion are reliably detectable by HM450 analysis and are associated with prognosis in the NOA-04 trial. For MGMT, HM450 suggests promoter methylation in the vast majority of G-CIMP tumors, which is supported by pyrosequencing.

Comparing the expression of integrins αvß3, αvß5, αvß6, αvß8, fibronectin and fibrinogen in human brain metastases and their corresponding primary tumors.

AIMS: To evaluate the expression of αv-series integrins in brain metastases. Inhibitors targeting these integrins are being tested for their therapeutic potential. MATERIAL AND METHOD: The extracellular regions of the αvß3, αvß5, αvß6, αvß8, the cytoplasmic domain of ß3, the αv-chain, and the ECM molecules fibronectin and fibrinogen were studied immunohistochemically in a series of 122 carcinoma and 60 melanomas metastatic to the central nervous system. In addition, 38 matched primary and metastatic tumors to the brain were compared directly. RESULTS: The αv-subunit was generally moderately to highly expressed in most tumors. αvß3 and cytoplasmic ß3 were weakly to moderately detectable in metastatic renal cell carcinomas and melanomas, αvß5 was prominently expressed in metastatic renal and colorectal carcinomas, αvß6 was most abundantly detectable in metastatic lung adenocarcinomas, but absent in melanomas. The tumor associated vessels in CNS metastases consistently expressed αvß3, αvß5, αv-, fibronectin and fibrinogen, however, mostly at low levels, while αvß6, αvß8 were lacking in vasculature. The comparative analysis of 38 matched primary tumors and brain metastases showed comparable levels of expression only for αvß3 and αvß8, while αvß6 and αvß5 were higher in primaries. CONCLUSION: We confirmed that integrin expression exhibits considerable heterogeneity according to tumor origin. αvß5 is the most promising target for integrin targeted treatment in brain metastases.

Differential expression of the tumor suppressor A-kinase anchor protein 12 in human diffuse and pilocytic astrocytomas is regulated by promoter methylation.

The scaffold protein A-kinase anchor protein 12 (AKAP12) exerts tumor suppressor activity and is downregulated in several tumor entities. We characterized AKAP12 expression and regulation in astrocytomas, including pilocytic and diffusely infiltrating astrocytomas. We examined 194 human gliomas and 23 normal brain white matter samples by immunohistochemistry or immunoblotting for AKAP12 expression. We further performed quantitative methylation analysis of the AKAP12 promoter by MassARRAY® of normal brain, World Health Organization (WHO) grade I to IV astrocytomas, and glioma cell lines. Our results show that AKAP12 is expressed in a perivascular distribution in normal CNS, strongly upregulated in tumor cells in pilocytic astrocytomas, and weakly expressed in diffuse astrocytomas of WHO grade II to IV. Methylation analyses revealed specific hypermethylation of AKAP12α promoter in WHO grade II to IV astrocytomas. Restoration experiments using 5-aza-2'-deoxycytidine in primary glioblastoma cells decreased AKAP12α promoter methylation and markedly increased AKAP12α mRNA levels. In summary, we demonstrate that AKAP12 is differentially expressed in human astrocytomas showing high expression in pilocytic but low expression in diffuse astrocytomas of all WHO-grades. Our results further indicate that epigenetic mechanisms are involved in silencing AKAP12 in diffuse astrocytomas; however, a tumor suppressive role of AKAP12 in distinct astrocytoma subtypes remains to be determined.

Longitudinal expression analysis of αv integrins in human gliomas reveals upregulation of integrin αvß3 as a negative prognostic factor.

Integrin inhibitors targeting αv series integrins are being tested for their therapeutic potential in patients with brain tumors, but pathologic studies have been limited by lack of antibodies suitable for immunohistochemistry (IHC) on formalin-fixed, paraffin-embedded specimens. We compared the expression of αv integrins by IHC in brain tumor and normal human brain samples with gene expression data in a public database using new rabbit monoclonal antibodies against αvß3, αvß5, αvß6, and αvß8 complexes using both manual and automated microscopy analyses. Glial tumors usually shared an αvß3-positive/αvß5-positive/αvß8-positive/αvß6-negative phenotype. In 94 WHO (World Health Organization) grade II astrocytomas, 85 anaplastic astrocytomas WHO grade III, and 324 glioblastomas from archival sources, expression of integrins generally increased with grade of malignancy. Integrins αvß3 and αvß5 were expressed in many glioma vessels; the intensity of vascular expression of αvß3 increased with grade of malignancy, whereas αvß8 was absent. Analysis of gene expression in an independent cohort showed a similar increase in integrin expression with tumor grade, particularly of ITGB3 and ITGB8; ITGB6 was not expressed, consistent with the IHC data. Parenchymal αvß3 expression and ITGB3 gene overexpression in glioblastomas were associated with a poor prognosis, as revealed by survival analysis (Kaplan-Meier logrank, p = 0.016). Together, these data strengthen the rationale for anti-integrin treatment of glial tumors.

Expression of integrins αvß3 and αvß5 and their ligands in primary and secondary central nervous system neoplasms.

AIMS: To study the expression of integrins αvß3 and αvß5 and their ligands in tumour, stroma and endothelial cells from human glioblastoma and CNS metastases from breast, lung and skin tumours. METHODS AND RESULTS: Integrin and integrin ligand expression was quantified in frozen tumour surgical specimens (15 glioblastomas and breast carcinoma metastases as well as 16 lung carcinoma and melanoma metastases) using immunohistochemistry. Gene expression profiles were evaluated in glioblastomas (n=424) and in normal brain (n=11). Overall, αvß3 expression was more common than αvß5 except in tumours derived from lung. αvß3 expression was most frequent in glioblastomas and melanoma metastases. Most lung-derived tumours expressed αvß5 but expression was less frequent in other tumours; about 20% of breast-derived tumours strongly expressed αvß5. Melanoma-derived tumours did not express αvß5. Expression of integrin ligands vitronectin, fibrinogen, fibronectin and osteopontin was variable between tumours, although most tumours expressed the ligands to some extent. Marked αvß3, but not αvß5, expression was common in stroma of CNS metastases. In blood vessels, αvß3 expression was more frequent than αvß5 and more pronounced in CNS metastases than in glioblastomas. Integrin ligand expression occurred in blood vessels in most tumours. In glioblastomas, mRNA expression of αvß3, αvß5, osteopontin and fibronectin were significantly upregulated over normal brain. CONCLUSIONS: Overall, we report distinct and heterogeneous patterns of integrin expression in primary and secondary brain tumours that may be relevant to the future development of integrin-targeting therapeutic approaches to brain tumours.

Activated leukocyte cell adhesion molecule is expressed in neuroepithelial neoplasms and decreases with tumor malignancy, matrix metalloproteinase 2 expression, and absence of IDH1R132H mutation.

Diffuse growth of gliomas is based on enhanced cell migration and remodeling of the extracellular matrix. Up-regulation of matrix metalloproteinases in gliomas is associated with a poor prognosis. The activated leukocyte adhesion molecule is considered to be indispensable for conversion of matrix metalloproteinase 2 into its active form. We therefore investigated the expression of activated leukocyte adhesion molecule in 9 malignant glial cell lines, 105 normal/reactive human brain specimens, 248 astrocytomas/glioblastomas, 98 ependymomas, 35 oligodendrogliomas, 10 neurocytomas, 10 primitive neuroectodermal tumors (PNET), and 36 medulloblastomas by immunohistochemistry and in selected cases by reverse transcriptase polymerase chain reaction. Correlation between activated leukocyte adhesion molecule expression and tumor grades and entities, proliferation activity, matrix metalloproteinase 2 expression, prognostic isocitrate dehydrogenase (IDH)1 mutation (R132H) status, O-6-methylguanine DNA-methyltransferase (MGMT) promoter status, or association with patient survival were analyzed. All oligodendrogliomas were strongly activated leukocyte adhesion molecule positive. Numbers of activated leukocyte adhesion molecule positive tumors were higher in glioblastomas (93%) than in diffuse astrocytomas (83%), but mean expression intensity was significantly reduced. Anaplastic ependymomas (68%) exhibited reduced numbers of activated leukocyte adhesion molecule-positive tumors and staining intensity compared with lower-grade ependymomas (85%). Activated leukocyte adhesion molecule expression in gliomas was independent of proliferative activity, MGMT status, patient survival, and age, whereas gliomas with IDH1 (R132H) mutation had significantly higher activated leukocyte adhesion molecule levels than their wild-type counterparts. Matrix metalloproteinase 2-negative glioblastomas exhibited significantly reduced activated leukocyte adhesion molecule expression levels compared with astrocytomas. In summary, our findings indicate that activated leukocyte adhesion molecule expression levels in gliomas are probably linked to other mechanisms than its supposed role as regulator of matrix metalloproteinase 2.

Parenchymal accumulation of CD163+ macrophages/microglia in multiple sclerosis brains.

Reactive macrophages/microglia exert both protective or damaging effects in multiple sclerosis (MS), which contribute to the relapsing-remitting nature of MS. CD163 is considered a marker of M2 (alternatively activated) macrophages. In the MS brain, CD163(+) perivascular macrophages express molecules for antigen recognition and presentation. Here we further investigated the accumulation of CD163(+) macrophages/microglia in the parenchyma of MS brains. CD163 expression pattern was investigated in different lesions of brain tissue specimens from five MS brains and five neuropathologically unaffected controls by immunohistochemistry. In the parenchyma of normal brain samples, immunoreactivity (IR) of CD163 was absent. In acute active lesions and at the rim of chronic active lesions of MS, strong accumulation of CD163(+) macrophages/microglia was seen. In chronic inactive lesions and in the center of chronic active lesion, CD163(+) macrophages/microglia were rare. Further, double-labeling showed that parenchymal and perivascular CD163(+) macrophages/microglia were myelin basic protein positive and HLA-DR(+), suggesting that CD163(+) macrophages/microglia could ingest and present antigen. In addition, in vitro incubating macrophage RAW264.7 cells with myelin turned LPS-induced inflammatory macrophages into an anti-inflammatory phenotype, indicating that myelin basic protein positive, CD163(+) macrophages/microglia in MS might have anti-inflammatory effects. The parenchymal CD163(+) macrophages/microglia, which had the capacity for antigen ingestion and presentation, might contribute to the resolution of inflammation in MS.

CD133 expression is associated with small round blue cell tumour morphology in human central nervous system neoplasms.

AIMS: CD133 is considered to be a marker for brain tumour-initiating cells. However, most data on CD133 are derived from animal or in-vitro studies. The aim of this study was to characterize CD133 expression, and the distribution and morphological features of CD133(+) cells, in primary and secondary human central nervous system (CNS) neoplasms. METHODS AND RESULTS: Tumours were analysed by real-time reverse transcription polymerase chain reaction, western blot, flow cytometry and immunohistochemistry. Our results show that only small round blue cell tumours (SRBCTs) exhibit strong and consistent CD133 expression. Interestingly, glioblastomas, large-cell carcinomas and sarcomas were negative for CD133. Only glioblastomas with a focal small-cell component exhibited CD133 immunoreactivity in the SRBCT component. In addition, CD133 expression did not correlate with the expression of other markers associated with stem cell differentiation, including CD15 and nestin. CONCLUSIONS: CD133 expression in human CNS neoplasms is independent of the grade of malignancy but strongly correlates with SRBCT morphology. Together with recent findings showing that CD133 is quickly upregulated upon hypoxia and that CD133(-) cells can also exhibit stem cell properties, our data strongly question the suitability of CD133 as a brain tumour stem cell marker in vivo.

Atypical teratoid/rhabdoid tumors may show morphological and immunohistochemical features seen in choroid plexus tumors.

Atypical teratoid/rhabdoid tumors (AT/RT) are aggressive embryonal tumors having a poor prognosis and are associated with mutations in the tumor suppressor gene hSNF5/SMARCB1/INI1. Differential diagnosis includes choroid plexus carcinoma which has occasionally been attributed as showing an inactivation of INI1/SMARCB1 nuclear staining in immunohistochemistry. However, these findings have been challenged by others. We therefore examined eight AT/RTs from six patients by immunohistochemistry for membranous expression of the inward rectifier potassium channel Kir7.1, which was in the central nervous system so far considered specific for choroid plexus tumors and normal choroid plexus epithelium. Two AT/RT cases exhibited membranous staining of Kir7.1, indicating a plexus epithelial differentiation of these tumors. The implications of these results on tumor diagnosis are discussed.

Immune surveillance of the normal human CNS takes place in dependence of the locoregional blood-brain barrier configuration and is mainly performed by CD3(+)/CD8(+) lymphocytes.

Despite the blood-brain barrier (BBB) the human CNS is continuously screened by blood-derived immunological cells. In certain brain areas the local BBB configuration grants passage of large molecules, whereas others are better shielded. We investigated whether these regional BBB compositions are paralleled by differences in the degree of cellular immunosurveillance by investigating tissue from 23 normal human brains for several CD markers, FoxP3, granzyme B, and perforin. Our results provide evidence that immunosurveillance is associated with locoregional BBB configuration and is mainly performed by CD3(+)/CD8(+)/granzyme B(-)/perforin(-) lymphocytes.

WT1 expression in normal and neoplastic cranial and peripheral nerves is independent of grade of malignancy.

OBJECTIVE: Wilms' tumor protein (WT1) expression is usually absent in normal glial cells of the CNS but is highly upregulated in brain tumor cells and its expression correlates with tumor grade. However, knowledge on WT1 expression in tumors of the peripheral nerve system (PNS) is limited. As WT1 antibodies not only serve as biomarker for cancerous tissue but also are considered for cancer immunotherapy, knowledge of WT1 expression in tumorous and normal peripheral nerve tissue is important for therapeutical purposes. METHODS: We analyze the immunohistochemical expression of WT1 in 101 samples consisting of 13 normal nerves, 10 neurofibromas, 69 schwannomas and 9 malignant peripheral nerve sheath tumors (MPNST). Tumor samples included 14 specimen from patients with a proven neurocutaneous disorder (neurofibromatosis type 1 and 2) and 3 cases of schwannomatosis. In 50 vestibular schwannomas tumor growth extension was correlated to WT1 expression. RESULTS: WT1 expression is present in Schwann cells of the majority of normal human nerves (11/13). In peripheral nerve sheath tumors, cytoplasmic WT1 protein is expressed in the cytoplasm of the neoplastic cells in all tumors, including MPNST, neurofibromas and schwannomas. The WT1 expression is independent of tumor malignancy or tumor growth extension and is not associated with a neurocutaneous disorder. CONCLUSION: WT1 expression in normal and neoplastic tissue differs in the peripheral and the central nervous system. These findings may point to a different functional role of WT1 in the PNS and the CNS.

Patients with IDH1 wild type anaplastic astrocytomas exhibit worse prognosis than IDH1-mutated glioblastomas, and IDH1 mutation status accounts for the unfavorable prognostic effect of higher age: implications for classification of gliomas.

WHO grading of human brain tumors extends beyond a strictly histological grading system by providing a basis predictive for the clinical behavior of the respective neoplasm. For example, patients with glioblastoma WHO grade IV usually show a less favorable clinical course and receive more aggressive first-line treatment than patients with anaplastic astrocytoma WHO grade III. Here we provide evidence that the IDH1 status is more prognostic for overall survival than standard histological criteria that differentiate high-grade astrocytomas. We sequenced the isocitrate dehydrogenase 1 gene (IDH1) at codon 132 in 382 patients with anaplastic astrocytoma and glioblastoma from the NOA-04 trial and from a prospective translational cohort study of the German Glioma Network. Patients with anaplastic astrocytomas carried IDH1 mutations in 60%, and patients with glioblastomas in 7.2%. IDH1 was the most prominent single prognostic factor (RR 2.7; 95% CI 1.6-4.5) followed by age, diagnosis and MGMT. The sequence from more favorable to poorer outcome was (1) anaplastic astrocytoma with IDH1 mutation, (2) glioblastoma with IDH1 mutation, (3) anaplastic astrocytoma without IDH1 mutation and (4) glioblastoma without IDH1 mutation (p < 0.0001). In this combined set of anaplastic astrocytomas and glioblastomas both, IDH1 mutation and IDH1 expression status were of greater prognostic relevance than histological diagnosis according to the current WHO classification system. Our data indicate that much of the prognostic significance of patient age is due to the predominant occurrence of IDH1 mutations in younger patients. Immunohistochemistry using a mutation-specific antibody recognizing the R132H mutation yielded similar results. We propose to complement the current WHO classification and grading of high-grade astrocytic gliomas by the IDH1 mutation status and to use this combined histological and molecular classification in future clinical trials.

Application of mutant IDH1 antibody to differentiate diffuse glioma from nonneoplastic central nervous system lesions and therapy-induced changes.

Differentiation of gliomas and reactive gliosis may be challenging both at primary tumor occurrence and at posttherapy biopsy. The most frequent IDH1 mutation found in the majority of WHO grade II and III gliomas can be visualized with an antibody specifically detecting mutant IDH1 protein. In this study, mIDH1R132H immunoreactivity in 120 reactive gliosis specimens of various etiologies is compared with Wilms Tumor 1 (WT1) and p53 expression, both markers applied for the differentiation of reactive gliosis and glioma. Although WT1 and p53 positive glial cells were found in 17% and 63% of cases respectively, all samples were negative for mIDH1R132H. Furthermore, we investigated 19 posttherapy gliomas (6 WHO II, 13 WHO III) with extensive reactive changes and detected mIDH1R132H positive cells in 13 specimens. In 5 of these cases, tumor cells were missed by conventional staining, showing the improved sensitivity of mIDH1R132H. Thus, mIDH1R132H is a tumor-specific marker that is superior to other established markers to differentiate reactive from neoplastic cells in grade II and III gliomas and allows identifying tumor cells in posttherapy specimens with extensive reactive changes. As IDH mutations are not characteristic of grade IV primary glioblastomas, this antibody cannot differentiate primary glioblastoma from reactive gliosis. Thus, caution has to be taken and a combined panel with other markers is needed.

Pituitary adenomas and craniopharyngiomas are CDX2 negative neoplasms.

OBJECTIVES: Previous studies have shown an inverse correlation between the expression of CDX2 (also known as CDX3) and tumour grade, stage and lymph node dissemination in colorectal adenomas and adenocarcinomas. Although less frequent, expression of CDX2 has also been reported in various other epithelial tissues and carcinomas. While many neoplasms have been studied, to date, no data is available on CDX2 expression in craniopharyngiomas. Furthermore, only very few data are available on CDX2 expression in normal pituitary gland tissue and/or pituitary adenomas. MATERIAL AND METHODS: We investigated CDX2 expression in 28 normal pituitary glands, 75 pituitary adenomas of varying hormonal activity (including 7 invasive adenomas and 7 atypical adenomas) and 23 craniopharyngiomas (17 adamantinous and 6 papillary) in tissue microarrays. RESULTS: None of the pituitary adenomas, craniopharyngiomas and normal pituitary glands showed expression of CDX2. CONCLUSIONS: There is no evidence for that CDX2 might play a role in tumourigenesis, invasive growth or tumour recurrence of pituitary adenomas or in tumourigenesis of craniopharyngiomas. But, presence of CDX2 expression might be useful in distinguishing intrasellar metastases from primary tumours of the sellar region.

Chemoradiotherapy of newly diagnosed glioblastoma with intensified temozolomide.

PURPOSE: To evaluate the toxicity and efficacy of chemoradiotherapy with temozolomide (TMZ) administered in an intensified 1-week on/1-week off schedule plus indomethacin in patients with newly diagnosed glioblastoma. PATIENTS AND METHODS: A total of 41 adult patients (median Karnofsky performance status, 90%; median age, 56 years) were treated with preirradiation TMZ at 150 mg/m(2) (1 week on/1 week off), involved-field radiotherapy combined with concomitant low-dose TMZ (50 mg/m(2)), maintenance TMZ starting at 150 mg/m(2) using a 1-week on/1-week off schedule, plus maintenance indomethacin (25 mg twice daily). RESULTS: The median follow-up interval was 21.7 months. Grade 4 hematologic toxicity was observed in 15 patients (36.6%). Treatment-related nonhematologic Grade 4-5 toxicity was reported for 2 patients (4.9%). The median progression-free survival was 7.6 months (95% confidence interval, 6.2-10.4). The 1-year survival rate was 73.2% (95% confidence interval, 56.8-84.2%). The presence of O(6)-methylguanine-DNA methyltransferase (MGMT) gene promoter methylation in the tumor tissue was associated with significantly superior progression-free survival. CONCLUSION: The dose-dense regimen of TMZ administered in a 1-week on/1-week off schedule resulted in acceptable nonhematologic toxicity. Compared with data from the European Organization for Research and Treatment of Cancer/National Cancer Institute of Canada trial 26981-22981/CE.3, patients with an unmethylated MGMT gene promoter appeared not to benefit from intensifying the TMZ schedule regarding the median progression-free survival and overall survival. In contrast, data are promising for patients with a methylated MGMT promoter.

CD8(+)/perforin/granzyme B(+) effector cells infiltrating cerebellum and inferior olives in gluten ataxia.

Up to 8% of patients with gluten sensitivity (GS) develop neurological symptoms such as ataxia, dementia, seizures or peripheral neuropathy. The underlying immunological mechanisms still remain to be elucidated. We here report the case of a 68-year-old male patient suffering from progressive ataxia and dementia associated with chronic diarrhea and both elevated IgG and IgA antigliadin-antibodies. At autopsy, frequent argyrophilic glial and neuronal inclusions within the basal nucleus of Meynert were considered as the structural correlative for the cognitive decline. Significant neuronal loss in the cerebellar cortex and the inferior olives was accompanied by infiltrating CD8(+)/perforin(+)/granzyme B(+) cells as well as reactive astrogliosis and microglial activation. These CD8(+) cytotoxic T and NK cells are likely to act as effector cells responsible for neuronal cell death in patients with gluten sensitivity and neurological disease and might therefore at least partly be responsible for cerebellar symptoms in gluten ataxia. In conclusion, our results, showing an absence of B- or plasma cells but multiple CD8(+) as well as granzyme B and perforin expressing cells in ataxia-associated brain areas, suggest that there are also prominent cytotoxic effects in neuropathogenesis of GS.