Summary Accurate pathological diagnosis is crucial for optimal management of cancer patients. For the ~100 known central nervous system (CNS) tumour entities, standardization of the diagnostic process has been shown to be particularly challenging - with substantial inter-observer variability in the histopathological diagnosis of many tumour types. We herein present the development of a comprehensive approach for DNA methylation-based CNS tumour classification across all entities and age groups, and demonstrate its application in a routine diagnostic setting. We show that availability of this method may have substantial impact on diagnostic precision compared with standard methods, resulting in a change of diagnosis in up to 12% of prospective cases. For broader accessibility we have designed a free online classifier tool (www.molecularneuropathology.org) requiring no additional onsite data processing. Our results provide a blueprint for the generation of machine learning-based tumour classifiers across other cancer entities, with the potential to fundamentally transform tumour pathology.

USP8 is frequently mutated in adenomas causing Cushing's disease, especially in those from female adult patients diagnosed at a younger age.

The migration of polymorphonuclear granulocytes (PMN) into the brain parenchyma and release of their abundant proteases are considered the main causes of neuronal cell death and reperfusion injury following ischemia. Yet, therapies targeting PMN egress have been largely ineffective. To address this discrepancy we investigated the temporo-spatial localization of PMNs early after transient ischemia in a murine transient middle cerebral artery occlusion (tMCAO) model and human stroke specimens. Using specific markers that distinguish PMN (Ly6G) from monocytes/macrophages (Ly6C) and that define the cellular and basement membrane boundaries of the neurovascular unit (NVU), histology and confocal microscopy revealed that virtually no PMNs entered the infarcted CNS parenchyma. Regardless of tMCAO duration, PMNs were mainly restricted to luminal surfaces or perivascular spaces of cerebral vessels. Vascular PMN accumulation showed no spatial correlation with increased vessel permeability, enhanced expression of endothelial cell adhesion molecules, platelet aggregation or release of neutrophil extracellular traps. Live cell imaging studies confirmed that oxygen and glucose deprivation followed by reoxygenation fail to induce PMN migration across a brain endothelial monolayer under flow conditions in vitro. The absence of PMN infiltration in infarcted brain tissues was corroborated in 25 human stroke specimens collected at early time points after infarction. Our observations identify the NVU rather than the brain parenchyma as the site of PMN action after CNS ischemia and suggest reappraisal of targets for therapies to reduce reperfusion injury after stroke.Electronic supplementary materialThe online version of this article (doi:10.1007/s00401-012-1076-3) contains supplementary material, which is available to authorized users.

The activation of immune cells by targeting checkpoint inhibitors showed promising results with increased patient survival in distinct primary cancers. Since only limited data exist for human brain metastases, we aimed at characterizing tumor infiltrating lymphocytes (TILs) and expression of immune checkpoints in the respective tumors.Two brain metastases cohorts, a mixed entity cohort (n = 252) and a breast carcinoma validation cohort (n = 96) were analyzed for CD3+, CD8+, FOXP3+, PD-1+ lymphocytes and PD-L1+ tumor cells by immunohistochemistry. Analyses for association with clinico-epidemiological and neuroradiological parameters such as patient survival or tumor size were performed.TILs infiltrated brain metastases in three different patterns (stromal, peritumoral, diffuse). While carcinomas often show a strong stromal infiltration, TILs in melanomas often diffusely infiltrate the tumors. Highest levels of CD3+ and CD8+ lymphocytes were seen in renal cell carcinomas (RCC) and strongest PD-1 levels on RCCs and melanomas. High amounts of TILs, high ratios of PD-1+/CD8+ cells and high levels of PD-L1 were negatively correlated with brain metastases size, indicating that in smaller brain metastases CD8+ immune response might get blocked. PD-L1 expression strongly correlated with TILs and FOXP3 expression. No significant association of patient survival with TILs was observed, while high levels of PD-L1 showed a strong trend towards better survival in melanoma brain metastases (Log-Rank p = 0.0537).In summary, melanomas and RCCs seem to be the most immunogenic entities. Differences in immunotherapeutic response between tumor entities regarding brain metastases might be attributable to this finding and need further investigation in larger patient cohorts.

The immune response in the central nervous system (CNS) is under tight control of regulatory mechanisms, resulting in the establishment of immune privilege. CNS injury induces an acute inflammatory reaction, composed mainly of invading leukocytes and activated microglial cells/macrophages. The generation of this robust immune response requires binding of receptors such as CD14, a pattern recognition receptor of the immune system. CD14, a surface molecule of monocytic cells, is up-regulated after monocyte stimulation and is involved in cellular activation. To examine CD14 expression in human brain lesions we investigated sections of brains obtained at autopsy from 25 cases following closed traumatic brain injury (TBI) and 5 control brains by immunohistochemistry. Detection of CD14 in controls demonstrated constitutive expression by perivascular cells, but not in parenchymal microglial cells, equivalent to known expression pattern of ED2 in rats. Following TBI, numbers of CD14(+) cells in perivascular spaces and in the brain parenchyma increased in parallel within 1-2 days, both at the lesion and in adjacent perilesional areas. The number of CD14(+) cells in perivascular spaces and in the brain parenchyma reached maximum levels within 4-8 days and remained elevated until weeks after trauma. In contrast to activated parenchymal microglia/macrophages, resting parenchymal microglial cells lacked CD14. Thus, early CD14 expression constitutes an essential part of the acute inflammatory CNS response following trauma.