**4.2. Cerebrospinal fluid**

**4.1. Serum and plasma**

50 Neurooncology - Newer Developments

elevated after tumor removal.

The first analysis of exosomes in serum identified the presence of 11 miRNAs in the samples from two different patients with primary GBM in 2008 (153). However, the levels were generally lower in exosomes but correlated well with the tumor profile (153). In 2011, Roth et al. analyzed miRNA profiles from the blood of 20 patients with GBM and 20 matched healthy controls (154). Among 1158 tested miRNAs, 52 were significantly deregulated, and of these, two candidates, miR-128 (upregulated) and miR-342-3p (downregulated), remained the most significant miRNAs. The altered expression of these two miRNAs was confirmed in a valida‐ tion cohort by RT-PCR. In this model, the discrimination between blood samples of patients with GBM and healthy controls reached an accuracy of 81%, specificity of 79%, and sensitivity of 83%. In 2012, Wang et al. determined the plasma miRNA levels of 50 patients with glioma and 10 healthy donors using RT-PCR (155). The plasma level of miR-21 was increased and the levels of miR-128 and miR-342-3p were significantly decreased in the patients with glioma compared with those in normal controls. These miRNAs were able to discriminate patients with glioma from healthy controls with high specificity and sensitivity. However, there were not significant differences between patients with glioma and other brain tumors such as meningioma or pituitary adenoma. Yang et al. performed genomewide serum miRNA analysis using serum samples of 122 untreated astrocytoma patients and 123 normal controls (156). The seven-miRNA panel (miR-15b\*, miR-23a, miR-133a, miR-150\*, miR-197, miR-497, and miR-548b-5p) demonstrated a high sensitivity (88.00%) and specificity (97.87%) for malignant astrocytoma prediction (156). These identified miRNAs also exhibited a global decrease in tumor tissues relative to normal tissues. Interestingly, these miRNAs in serum were markedly

In miRNA microarray analysis of the serum of patients with GBM and normal controls, 115 miRNAs were upregulated in the GBM group and 24 miRNAs were downregulated (157). In these microRNAs, a six-membered serum miRNA expression profile (upregulated miRs; miR-576-5p, miR-340, and miR-626, downregulated miRs; miR-320, let-7g-5p, and miR-7-5p) could serve as a noninvasive biomarker for GBM diagnosis (157). Manterola et al. found that the serum expression levels of miR-320 and miR-574-3p were significantly altered in the patients with GBM (158). In addition, small noncoding RNA (RNU6-1) was an independent predictor of a diagnosis of GBMs. Shao et al. compared the expression levels of miR-454-3p between preoperative plasmas from 70 patients with glioma and 70 healthy controls and between these preoperative and postoperative plasmas (159). The expression levels of miR-454-3p in plasma in patients with glioma were significantly higher, and the area under receiver operating characteristic (ROC) curve (AUC) of the expression of miR-454-3p for glioma diagnosis was 0.9063. In addition, the expression levels of miR-454-3p in the postop‐ erative plasmas were significantly downregulated relative to the preoperative plasmas. Wu et al. analyzed serum from 83 patients with glioma and 69 healthy controls and evaluated the availability of the serum miR-29 family in the screening of glioma (160). The predictive value of the serum miR-29 family for glioma was moderate (AUC = 0.74), but that in high-grade glioma detection was sufficient (AUC = 0.81). Another study also showed that the expressions of miR-29b in blood were significantly different compared with those of a healthy control (161).

Baraniskin et al. (170) reported the first investigation of microRNA expression in CSF from patients with glioma in 2012. The results demonstrated that miR-15b and miR-21 were differentially expressed in CSF samples from patients with glioma compared with control subjects with various neurologic disorders, including CNS lymphoma and carcinomatous brain metastases. The combination of miR-15b and miR-21 resulted in increased diagnostic accuracy, with 90% sensitivity and 100% specificity in distinguishing patients with glioma from control subjects. Interestingly, miR-15 levels were significantly higher in glioma than in CNS lymphoma or metastatic tumor; however, miR-21 levels were significantly lower in glioma than in CNS lymphoma or metastatic tumor.

Teplyuk et al. (171) determined the CSF levels of several cancer-associated miRNAs for 118 patients diagnosed with different types of brain cancers. The levels of miR-10b and miR-21 were significantly increased in the CSF of patients with GBM and brain metastasis of breast and lung cancer compared with non-neoplastic conditions. Members of the miR-200 family were useful markers for discriminating between GBM and metastatic brain tumors. In addition, quantification of as few as seven microRNAs in CSF (miR-10b, miR-21, miR-125b, miR-141, miR-200a, miR-200b, and miR-200c) enabled differential recognition of GBM and metastatic brain tumor with high accuracy (91%–99%) (171).

Akers et al. analyzed extracellular vesicles (EVs) containing miRNAs in the CSF from patients with GBMs. The EV fraction was isolated by differential centrifugation. Although the analytic algorithm for quantitatively assessing EV miRNA remains underdeveloped, the authors showed that the CSF miR-21 levels of patients with GBM were 10-fold higher than those in the CSF of nononcologic patients (172).

Recently, another study reported results to those of Akers and coworkers. Exosomal miR-21 levels in the CSF of patients with glioma were found to be significantly higher than in controls, although no difference was detected in serum-derived exosomal miR-21 expression. Interest‐ ingly, the CSF-derived exosomal miR-21 levels correlated with a level of dissemination and a location of recurrence. Therefore, exosomal miR-21 in CSF may be a predictive marker for the early stages of tumor recurrence or metastasis (173).
