Jump to ContentJump to Main Navigation
Show Summary Details
More options …

Clinical Chemistry and Laboratory Medicine (CCLM)

Published in Association with the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM)

Editor-in-Chief: Plebani, Mario

Ed. by Gillery, Philippe / Greaves, Ronda / Lackner, Karl J. / Lippi, Giuseppe / Melichar, Bohuslav / Payne, Deborah A. / Schlattmann, Peter

IMPACT FACTOR 2017: 3.556

CiteScore 2017: 2.34

SCImago Journal Rank (SJR) 2017: 1.114
Source Normalized Impact per Paper (SNIP) 2017: 1.188

See all formats and pricing
More options …
Volume 56, Issue 6


Cerebrospinal fluid microRNAs as diagnostic biomarkers in brain tumors

Alena Kopkova / Jiri Sana
  • Central European Institute of Technology (CEITEC), Masaryk University, Brno, Czech Republic
  • Department of Comprehensive Cancer Care, Masaryk Memorial Cancer Institute, Faculty of Medicine, Masaryk University, Brno, Czech Republic
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Pavel Fadrus
  • Department of Neurosurgery, University Hospital Brno, Faculty of Medicine, Masaryk University, Brno, Czech Republic
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Ondrej Slaby
  • Corresponding author
  • Central European Institute of Technology (CEITEC), Masaryk University, Brno, Czech Republic
  • Department of Comprehensive Cancer Care, Masaryk Memorial Cancer Institute, Faculty of Medicine, Masaryk University, Brno, Czech Republic
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
Published Online: 2018-02-16 | DOI: https://doi.org/10.1515/cclm-2017-0958


Cerebrospinal fluid (CSF) is a body fluid that has many important functions and is in direct contact with the extracellular environment of the central nervous system (CNS). CSF serves as both the communication channel allowing the distribution of various substances among the CNS cells and the storage facility for the waste products these cells release. For these reasons, CSF is a potential source of diagnostic biomarkers of many CNS diseases, including brain tumors. Recent studies have revealed that CSF also contains circulating microRNAs (miRNAs), short non-coding RNAs that have been described as biomarkers in many cancers. However, CSF miRNAs are difficult to detect, which is why researchers face major challenges, including technological difficulties in its detection and its lack of standardization. Therefore, this review aims (i) to highlight the potential of CSF miRNAs as diagnostic, prognostic and predictive biomarkers in brain tumors, and (ii) to summarize technological approaches for detection of CSF miRNAs.

Keywords: biomarker; brain cancer; cerebrospinal fluid (CSF); diagnosis; microRNA; prognosis


With an incidence rate of around 22 cases per 100,000 people, primary brain tumors account for less than 2% of all malignancies. However, the malignant forms account for about 33% of primary brain tumors, with an estimated 5- and 10-year relative survival rates of 34.4% and 28.8%, respectively. A prognosis and the patient’s estimated survival depend on the tumor type. The most favorable prognosis is associated with pilocytic astrocytoma (with a 5-year survival rate of 94.2%), followed by malignant meningioma (65.2%) and lymphoma (29.2%). Patients with glioblastoma (GBM) have only a 5.1% 5-year survival rate [1]. Occurring in 20%–40% of adult cancer patients, metastases form another large group of brain tumors. Although the survival rate of brain metastasis patients is generally low, the prognosis varies among individual cases [2].

Diagnosis plays a crucial role in making the prognosis and choosing the best therapy for the diagnosed brain tumor. Despite significant recent advances in the diagnosis of brain tumors, such as various modifications of imaging methods followed by histopathological examinations, the diagnosis is still limited by a tumor’s size and localization as well as the heterogeneity of its tissue [3]. Hence, we need to develop new powerful diagnostic approaches that, together with the existing methods, would increase the accuracy of brain tumor diagnosis and thereby the survival of patients.

Promising diagnostic markers in many solid cancers seem to be circulating biomolecules that are altered in body fluids such as blood, urine or saliva. Nevertheless, the localization of brain tumors and the presence of the blood-brain barrier are deemed responsible for preventing the release of tumor-specific molecules into the above-mentioned body fluids [4]. Cerebrospinal fluid (CSF), which bathes the central nervous system (CNS) and is in direct contact with brain neoplasms, seems to be a suitable source of diagnostic biomarkers [5, 6].

MicroRNAs (miRNA) are short non-coding RNAs that generally bind to the 3′ untranslated regions of messenger RNAs (mRNAs) and repress protein translation [6, 7]. miRNAs may regulate over 50% of all human genes, and each miRNA can control hundreds of mRNA targets. The deregulated expression of miRNAs is closely associated with the pathology of many cancers, including brain tumors [8]. Moreover, circulating miRNAs (c-miRNAs) have been found in several types of human body fluids, CSF being one of them. These miRNAs might be both secreted by the cells in extracellular vesicles and be released by necrotic cells as naked oligonucleotides.

Many authors have also suggested that c-miRNAs participate in intercellular communication and may influence molecular cell processes, such as growth, invasion and drug resistance in recipient target cells [4, 9]. An increasing number of studies on the altered levels of specific miRNAs in CSF of brain tumor patients have clearly shown that CSF miRNAs are promising diagnostic biomarkers. These results, however, are far from conclusive, mainly because of their technological variations, but also due to other serious limitations which we discuss below.

In this review, we summarize the diagnostic relevance of miRNAs in CSF for brain tumor patients and the technological approaches for their detection.

Methodical approaches used for miRNA detection in CSF

Detection of miRNA in CSF is a complex process that requires the following consecutive steps: CSF management and storage; isolation and expression analysis of miRNA; and final data normalization, analysis and interpretation. Unfortunately, the whole process can be affected by many factors and has not yet been standardized (Figure 1).

Summary of the most common techniques used for detection of CSF microRNAs.
Figure 1:

Summary of the most common techniques used for detection of CSF microRNAs.

CSF management and storage

In the first and critical step of CSF management, a sample of patients needs to be collected. This should be performed by experienced physicians using the standard lumbar puncture, external ventricular drain or cisternal aspiration at the time of craniotomy [10]. To avoid bias of the subsequent analyses due to the preanalytical factors, a standardized protocol for collection of CSF samples should be established. Such a protocol has already been proposed by Teunissen et al. in 2009 and 2011 [11, 12]. These authors highlighted three crucial points for the protocol to work. First, as the volume of CSF can influence the concentration of biomarkers, the volume of CSF withdrawal should be at least 12 mL. An inadequate total volume of the sampled CSF can lead to complications during the lumbar puncture, like brain herniation; the adequate CSF volume should then be assessed individually for each patient and the corresponding diagnosis [2]. Second, a lumbar puncture should be performed between the L3 and L5 vertebraes. Third, CSF samples containing blood-derived cells or even skin cells must be excluded. In an independent study, Müller et al. [13] confirmed the importance of this point by showing that even after centrifugation of samples before analysis, the number of blood-derived cells influenced the expression of miRNAs. In this context, Kirschner et al. [14] detected 136 hemolysis-susceptible miRNAs (e.g. miR-21, miR-106a, miR-92a, miR-17, miR-16 and miR-451). Similarly, Bache et al. [15] found that levels of miR-16-5p and miR-451a as well as the average levels for all miRNAs in each sample strongly correlated with the net absorbance of oxyhemoglobin and hemoglobin in CSF of subarachnoid hemorrhage (SAH) patients. Therefore, the collected CSF should be tested for the presence of hemoglobin or cells. The time from the CSF withdrawal to laboratory processing and storage of samples is also important and should thus be documented.

Next, the conditions of centrifugation for removing the debris, which should be done up to 60 min after the lumbar puncture, should be standardized at 400 g for 10 min at room temperature. However, in most of the existing studies, CSF samples are centrifuged at 500 g. More extensive and longer centrifugation is used to separate CSF extracellular vesicles and extract RNA from them. Akers et al. [10] used a different approach: they filtered the collected samples using a 0.8 μm pore size membrane filter. Finally, although miRNAs seem to be stable over a wide range of storage temperatures and resistant to repeated freeze-thaw [16], the samples should be protected against any biochemical changes and supernatants should be aliquoted and stored at −80 °C as soon as possible.

miRNA isolation

As the concentration of miRNAs in CSF is low, their extraction is challenging. A few studies have compared the efficiency of currently available commercial kits for the purification of RNA from CSF. This process still lacks standardization, and thus it is not easy to choose the most efficient method.

Burgos et al. tested four commercially available RNA isolation kits for CSF samples: mirVana miRNA isolation kit, mirVana PARIS kit (both Thermo Fisher Scientific), BioPure RNA isolation reagent (Bio Scientific), and miRNeasy mini kit (Qiagen). They followed the manufacturer-provided protocol with one modification, for the phenol/chloroform phase separation. The authors performed two consecutive RNA extraction steps from an organic phase to maximize RNA recovery and, in both these extractions, measured the total and miRNA concentrations. The tested kits gave different yields, with the highest obtained by the mirVana PARIS kit (±9 and 25 ng/mL of total RNA from first extraction; ±6 and 20 ng/mL of total RNA from second extraction) and the BioPure RNA isolation reagent (±10 and 10 ng/mL; ±1 and 40 ng/mL). In addition, the mirVana PARIS kit returned the greatest amount of spike-in cel-miR-39, measured by qPCR [17]. qPCR is commonly used for indirect quantification when RNA is in so small a concentration that RNA yield is immeasurable [18]. Thus, as the mirVana PARIS kit gave the greatest concentration of small RNAs and had a user-friendly protocol, it comes as no surprise that the authors selected this kit as the most suitable one for small RNA sequencing. They also found that in some cases the combination of the two extractions almost doubled the recovery of the total RNA yield. Finally, they also estimated correlations of miRNA yields isolated using the mirVana PARIS kit with the 0.5- to 1.25-mL starting CSF volumes. The next-generation sequencing (NGS) analysis confirmed high positive correlations (Spearman’s correlation over 0.95) between the starting volume of CSF and the levels of miRNAs with the coverage higher than five reads. The same experiment, but with RNAs extracted by the BioPure RNA isolation kit, showed similar associations as the experiment with the mirVana PARIS kit did. Thus, 0.5 mL of CSF should be a sufficient starting extraction volume to achieve reproducible results of miRNA analyses [17].

To summarize our discussion of the methods of extraction of small RNA from CSF, several modifications of the manufacturer-provided protocols have been suggested. In McAlexander et al’s. [19] study, which used the miRCURY: Biofluids kit (Exiqon), the increasing input volumes of CSF from Macaca nemestrina showed a quantitative recovery of four endogenous miRNAs. But as kits vary in the input of biofluid, it seems reasonable to choose the optimal input volume for a particular combination of biofluid and method [19]. Some studies have also supplemented the RNA extraction kit with glycogen or exogenous RNA (east tRNA or MS2 phage RNA) as the carrier. McAlexander et al. used pure glycogen, which – as an inert carrier – does not affect the downstream assays. Unlike pure glycogen, RNA carriers might cause non-specific hybridization or amplification in the quantification assays. The authors also studied how glycogen affects RNA yield in blood plasma samples. The results were inconsistent because adding glycogen increased RNA yields only in the case of some extraction kits [19].

Bache et al. [15] successfully used the MS2 carrier during the RNA isolation process handled by the miRCURY RNA isolation kit – Biofluids (Exiqon). The RNA obtained was then analyzed using high-throughput real-time PCR. However, for RNA isolation in the explorative phase of the project, the authors used the total RNA purification kit supplied by the Norgen Biotek Corp. [15]. In our experiments, the urine microRNA purification kit, also supplied by the Norgen Biotek Corp., was the most sufficient method for isolating miRNA from CSF, but only after reducing the elution volume from 50 μL to 20 μL and extending elution time to 20 min before centrifugation. However, these preliminary results need further studies and thus have not yet been published.

Finally, low concentrations of total and small RNA greatly decrease the accuracy of the standard methods used for the quantification and quality control of RNA, such as the Nanodrop and Qubit technologies. Hence, alternative methods might work better, e.g. the automated capillary electrophoresis, i.e. Bioanalyzer (Agilent). One can even use a special chip for a small RNA analysis, able to detect RNA at as low concentration as 50 pg/μL. Unfortunately, the price-achievement ratio of this chip is far from outstanding. This problem of inaccurate measurement of RNA quality and quantity might be alleviated by using the optimum volumes for the procedure chosen [4, 20].

miRNA expression analysis

Final quantification of miRNAs in the extracted RNA samples is very important. Simply put, imprecise miRNA levels detected in the experiment can greatly affect the diagnostic power of these molecules. Although the techniques of CSF miRNA quantification do not differ from the commonly used approaches, the low RNA concentrations of the samples examined should always be taken into account.

The miRNA expression analyses are divided into two major groups: single miRNA analyses and global (or whole-genome) miRNA profiling. Currently, levels of individual miRNAs in CSF are measured exclusively using modified real-time PCR technologies. Among them, the most common approach is based on the specific reverse transcription with the stem-loop primers followed by the real-time PCR using the TaqMan detection system (Thermo Fisher Scientific) [16, 21, 22]. Baraniskin et al. [16] confirmed the reproducibility of this method. It is not the only suitable approach, however; another one is based on PCR with the universal reverse transcription followed by SYBR Green quantitative PCR with miRNA specific forward and reverse locked nucleic acid (LNA)-modified primers (Exiqon) [15]. Another study used digital PCR handled according to the TaqMan chemistry protocol for miRNA levels analyses in CSF. The results, consistent with the simultaneously performed NGS analysis, proved that the digital PCR experiments worked correctly [23]. This PCR technology seems suitable only for low-concentrated RNA samples and samples with undefined normalizer molecules, such as CSF.

For the global miRNA profiling, four high-throughput platforms are currently available, namely, hybridization arrays, technologies based on real-time PCR, nCounter platform and NGS. These platforms differ in the number of miRNAs they can detect, sensitivity, specificity, dynamic range, quantity of input material and total cost. Thus, all the circumstances of the profiling need to be considered before selecting the adequate technology.

Several studies have compared the high-throughput technologies for the CSF miRNA analysis. Sørensen et al. performed small RNAseq of RNA samples from CSF and detected 246 miRNAs, of which 71 were found in over 90% of the patients examined. Simultaneously, the authors used the Exiqon qPCR arrays, which enabled them to analyze 372 miRNAs in one experimental run, in which they detected 210 miRNAs. Nevertheless, only 21 molecules occurred in over 90% of the investigated samples. Both qPCR and small RNAseq detected all but one differentially expressed miRNAs, but in general the qPCR-based technology showed a lower detection rate [24].

Bache et al. analyzed the global miRNA expression in CSF using the TaqMan Low Density Array (TLDA; Thermo Fisher Scientific) with 754 miRNAs on two independent cards. They detected 168 miRNAs in over 60% of healthy patients (128 miRNAs in all healthy patients) and 216 in over 60% of SAH patients (SAH; 155 miRNAs in all SAH patients). These 155 miRNAs were then validated using the Exiqon technology, with which only 74 miRNAs were detected in over 60% of the healthy patients (19 miRNAs in all healthy patients) and 100 in over 60% of the SAH patients (60 miRNAs in all SAH patients) [15]. These different numbers of the detected miRNAs by the two technologies may result from the preamplification step that precedes the TLDA analysis.

Table 1 summarizes available studies that analyze CSF miRNAs using high-throughput approaches.

Table 1:

A summary of the studies analyzing global microRNA expression profiles in CSF samples.

Data normalization

The last issue to address about miRNA profiling of CSF concerns how to normalize raw data. This is a challenging but important problem to solve, especially in the context of circulating miRNAs. Various technical and biological factors affect miRNA levels in body fluids. Examples of the technical factors are variability induced during collection and storage of the clinical specimens, the approach to miRNA extraction and the methods of final miRNA quantification. Examples of the biological factors are the amount of miRNAs in tissues, the intensity of their secretion into body fluids, the form of circulating miRNAs affecting their ability to cross various barriers and the stability of miRNA. It is because of these various factors that miRNA raw data need to be normalized. This normalization has been broadly recognized as one of the key factors in efficient miRNA analysis in CSF [35].

Several methods of CSF miRNA normalization have already been proposed. The first one is normalization by spike-in miRNA, which is added to the samples before the RNA extraction and quantitative analysis. This exo-genous miRNA is normally absent in the organism studied (i.e. Caenorhabditis elegans [C. elegans] or plant miRNA), so its level reflects the technological variability only. The miRNA spike-in normalization has worked well in several studies comparing and optimizing the methods of CSF miRNAs detection. Burgos et al. [17] used quantitative analysis of spiked-in C. elegans miRNA cel-miR-238 to compare yields of small RNAs obtained in two consecutive extractions from CSF. Analyzing the effect of the starting CSF volume on the quantitative recovery of RNA yield, McAlexander et al. [19] spiked cel-miR-39 into the lysis/denaturant buffer before adding it to CSF. According to the authors, synthetic RNA should not be added directly to CSF because endogenous RNases would degrade it immediately. For normalization of sample-to-sample variation, Bergman et al. [30] used exogenous miRNA spike-in from Arabidopsis thaliana ath-miR-159a and C. elegans cel-miR-54, cel-miR-39 and cel-miR-238; the authors added synthetic miRNAs to each denatured CSF sample.

This approach, however, does not reflect the biological factors and possible variations in the CSF management prior to the RNA extraction. To take account of them, data can be normalized using the most stable endogenous miRNAs or another short RNA that is not associated with the pathogenic condition. Recent works have identified several such endogenous controls in CSF, for instance, Baraniskin et al. (who used miR-24) [16], Teplyuk et al. (miR-24 and miR-125) [22] , Denk et al. (let-7c, miR-21, miR-24, miR-99b, miR-328 and miR-1274B) [25], Sørensen et al. (miR-15a-5p, miR-21-5p, miR-23a-3p, miR-23b-3p, miR-99a-5p, miR-125b-5p, miR-145-5p, miR-204-5p and miR-320a) [24], and Gui et al. (RNU44 and RNU6B) [29]. Nevertheless, it is not easy to choose the best normalizer, and no RNA molecule shows stable levels in all pathological conditions and CSF components, a conclusion drawn by Akers et al. [36]. In their study, miR-125 and miR-24 were inappropriate references for the quantitative miRNA analyses of extracellular vesicles in CSF.

A third approach, possible in high-throughput analyses, is to normalize levels of individual miRNAs according to the selected parameters based on comprehensive characterization of the sample. In other words, small RNA-seq raw data are generally expressed in reads per million (RPM), i.e. specifically mapped reads to individual miRNA divided by the total number of aligned reads and multiplied by one million. This approach was successfully used by Yagi et al. [23]. To estimate the relative miRNA levels from small RNA-seq data, the trimmed mean of M-values normalization method (TMM) can also be used. It estimates scale factors between the analyzed samples; these scale factors can then be used in the statistical methods for differential expression analysis [37]. This approach was used by Sørensen et al. [24] for the CSF miRNA analysis.

The simplest approach to normalizing raw data obtained by hybridization high-throughput platforms, like Affymetrix, is to re-scale each chip in an experiment by its total intensity. Many computational approaches can be used for such normalization, such as using selected genes instead of the entire gene set or scaling the individual intensities so that the median or mean value of intensities are identical within a single array or across all arrays.

Many normalization strategies are available and new ones will be developed in the future because choosing the most suitable approach to normalize the miRNA data for the method in use is crucial to draw unbiased conclusions.

CSF miRNAs as biomarkers in brain tumors

During the last decade, several studies have shown that deregulated levels of CSF miRNAs might be associated with malignant tumors of the CNS and might thus represent a novel group of possible diagnostic biomarkers. Many researchers have also focused on the use of CSF miRNAs as prognostic and predictive biomarkers of therapy response in patients with the CNS malignancies.

Diagnostic markers

In 2011, Baraniskin et al. [16] analyzed levels of six candidate miRNAs (miR-15b, miR-19b, miR-21, miR-92a, miR-106b and miR-204) in CSF samples obtained from 23 patients with primary CNS B-cell lymphoma (PCNSL) and 30 control patients with various neurologic disorders. These miRNAs were selected based on the two following criteria: (i) they are expressed by lymphoma cells at moderate or high levels and (ii) they show low or undetectable concentrations in CSF derived from the control patients. Only miR-19b, miR-21 and miR-92a showed significantly higher levels in the CSF samples from the patients with PCNSL than those from the control patients. Moreover, 22 of 23 PCNSL patients (95.7%) were correctly identified according to the levels of the above three miRNAs [16]. Several months later, the same authors published another study, in which miR-15b and miR-21 were higher in CSF samples from patients with glioma than in those from the control patients. On the other hand, CSF samples from patients with glioma had higher levels of miR-15b and lower levels of miR-21 than the CSF sample from the PCNSL patients [21].

Subsequently, Teplyuk et al. reported that miR-10b was detected in most of CSF of GBM patients (89%) and patients with leptomeningeal and brain metastases of both lung and breast carcinomas (81%). This miRNA was, however, not detected in patients with various non-neoplastic neurological diseases. Furthermore, miR-21 level was higher in most of the CSF samples taken from GBM and metastatic patients than in the control CSFs. MiR-10b was expressed in most extracranial tissues and in peripheral blood serum. This miRNA was absent in both brain tissue and CSF of patients without pathologically diagnosed malignancy. This observation might suggest that, in physiological conditions, miR-10b and other miRNAs cannot pass the blood-brain barrier and, thus, CSF miRNAs might reflect a specific brain miRNA signature. Metastatic cells might bring their own miRNA signature to CSF, as miR-10b is abundant in lung and breast tissues and is present in the CSF of breast and lung cancer patients with the CNS metastases. To confirm this presumption, the authors evaluated the levels of miR-200a, miR-200b, miR-200c and miR-141 in the CSF samples from GBM and metastatic brain cancer patients and from healthy donors. Comparing these two patient groups is relevant because the miR-200 family is highly expressed in a primary carcinoma but not in normal brain or GBM and the other primary brain tumors. The levels of the miRNAs examined were higher in most of the CSF samples from the patients with leptomeningeal and brain metastases, but not from the GBMs or healthy donors. These results show that the miR-200 family could be a promising biomarker able to classify brain metastases and primary brain cancers [22].

Using NanoString nCounter assay, Drusco et al. looked for possible diagnostic biomarkers of all CNS tumors across CSF miRNAs. They divided CSF samples from 34 neoplastic and 14 healthy patients into seven groups (GBM, medulloblastoma, lung metastasis, breast metastasis, primary CNS lymphoma, benign and normal) and performed data analysis as well as comparisons among all groups. Results were further validated using qRT-PCR and hybridization in situ. The CSF levels of miR-125b, miR-223, miR-451, miR-711 and miR-935 exhibited different patterns between the examined groups, which suggests that these miRNAs might be efficient diagnostic biomarkers of the CNS malignancies [38].

Grotzer et al. analyzed CSF miRNAs in medulloblastoma and control patients. They detected 1254 miRNAs in the tumor CSF samples, and 86 of these miRNAs were differentially expressed between the groups. The authors also used culture medium in vitro to analyze miRNAs excreted by the medulloblastoma cells. Fifty-seven detected miRNAs were specific to the metastasis-related cell lines which represent the most aggressive medulloblastoma subtypes. Interestingly, the three miRNAs associated with metastases (miR-125a, miR-125b and miR-1290) which were over-represented in the culture medium of cancer metastasis-related cell lines were also detected in the CSF samples from the medulloblastoma patients [39].

Finally, Akers et al. analyzed miR-21 levels in extracellular vesicles (EV) extracted from CSF samples from 13 GBM patients and 14 non-tumor control patients. miR-21 ranged from 0.14 to 1.04 copies per EV in the GBM samples and from 5.26·10−4 to 1.48·10−1 copies per EV in the control patients. The cut-off (<0.25) copy per EV classified the GBM patients with 87% sensitivity and 93% specificity. The authors also measured the levels of miR-21 in EV depleted CSF of the GBM patients, but miR-21 was undetectable [36]. Yagi et al. [23] confirmed, by NGS, that miRNA in CSF of the healthy donors is more abundant in the exosomal fraction than in the supernatant.

Table 2 summarizes studies comparing CSF microRNA levels among brain tumors.

Table 2:

CSF microRNAs with significantly different levels between brain tumors and non-tumor brain tissue.

Prognostic and predictive markers

Teplyuk et al. were the first to study whether levels of CSF miRNAs reflects the activity of the GBM and metastatic brain cancers. Members of neither the miR-10b nor miR-200 famillies were detected in the CSF samples from cancer patients in remission. Similarly, miR-21 was detected at significantly lower levels in cancer patients in remission than in patients with progressing GBM or metastatic brain cancer. To examine whether levels of specific CSF miRNAs are associated with disease status and activity as well as tumor response to the therapy, miRNAs were quantified in the CSF of GBM and lung cancer patients under erlotinib treatment. During the treatment, the CSF levels of members of both the miR-10b and miR-200 families increased in patients with relapsing non-small cell lung carcinoma. However, the miRNA levels decreased after increasing the treatment dosage. The miR-21 levels in the CSF samples from the GBM patients with pseudo-progression were similar to the levels in the CSF samples from the non-neoplastic patients, whereas the miR-10b levels were higher in the first set of patients than in the second one. On the other hand, disease progression was accompanied by a notable increase in the miRNA CSF levels. In summary, Teplyuk et al’s. [22] study indicates that CSF miRNAs may serve as biomarkers of response to the therapy and brain cancer progression.

Shi et al. examined CSF samples from recurrent glioma patients to detect miRNAs associated with this cancer, aiming to test whether these miRNAs can serve as prognostic biomarkers. For this purpose, they compared the miR-21 levels in CSF, blood serum and exosome. While blood serum-derived exosomal miR-21 levels of the glioma patients did not differ from the non-tumor control group, CSF miR-21 levels of exosomal origin were higher in the glioma patients. The CSF exosomal miR-21 levels reflected the presence of both tumor recurrence with anatomical site preference and spinal/ventricle metastasis. Therefore, the authors suggested that exosomal miR-21 might be a promising diagnostic and prognostic biomarker of glioma [40]. Finally, Tumilson et al. [41] observed that the increased levels of miR-21 in CSF were associated with a poor prognosis. Moreover, they suggested that CSF miR-21 has potential as a predictive biomarker of temozolomid resistance.


More often than not, prognoses for patients with brain malignancies are unfavorable. Not only does the prognosis depend on a tumor, it can also vary among patients with the same tumor type. Like in other cancers, early and accurate diagnosis is crucial for effective treatment but is seldom easy, mainly due to the tumor’s localization or heterogeneity, common in gliomas. For these reasons, there is a clinical need for new biomarkers of brain tumors that would allow for precise diagnosis, prognosis and prediction of a response to therapy. These biomarkers should be stable, and their analysis should use common methods (like qRT-PCR) that provide simple, reproducible and fast analysis. From this perspective, brain tumors associated with miRNAs released and detected in CSF seem promising molecules, so it is not surprising that more and more research on this topic is being published. Despite the promising published data indicating CSF miRNAs as specific and sensitive brain tumor biomarkers, technical issues of the analysis are hurdles still to overcome. This is exactly why further research should focus on these technical issues, before even thinking of clinical implementation of CSF miRNAs.

Given that the CSF miRNAs are so promising as diagnostic, prognostic and predictive biomarkers of therapy response in brain tumor patients, we should continue our efforts to solve the technical hurdles and better understand the molecular mechanisms leading to the cellular secretion of miRNAs into body fluids.


  • 1.

    Ostrom QT, Gittleman H, Fulop J, Liu M, Blanda R, Kromer C, et al. CBTRUS Statistical Report: Primary brain and central nervous system tumors diagnosed in the United States in 2008–2012. Neuro-Oncol 2015;17:1–62. CrossrefGoogle Scholar

  • 2.

    Khuntia D, Khuntia D. Contemporary review of the management of brain metastasis with radiation. Adv Neurosci 2015;2015:e372856. Google Scholar

  • 3.

    Ahmed R, Oborski MJ, Hwang M, Lieberman FS, Mountz JM. Malignant gliomas: current perspectives in diagnosis, treatment, and early response assessment using advanced quantitative imaging methods. Cancer Manag Res 2014;6:149–70. PubMedWeb of ScienceGoogle Scholar

  • 4.

    Shalaby T, Grotzer MA. Tumor-associated CSF microRNAs for the prediction and evaluation of CNS malignancies. Int J Mol Sci 2015;16:29103–19. PubMedWeb of ScienceCrossrefGoogle Scholar

  • 5.

    Weston CL, Glantz MJ, Connor JR. Detection of cancer cells in the cerebrospinal fluid: current methods and future directions. Fluids Barriers CNS 2011;8:14. PubMedCrossrefGoogle Scholar

  • 6.

    Shalaby T, Achini F, Grotzer MA. Targeting cerebrospinal fluid for discovery of brain cancer biomarkers. J Cancer Metastasis Treat 2016;2:176–87. CrossrefGoogle Scholar

  • 7.

    Gomes HR. Cerebrospinal fluid approach on neuro-oncology. Arq Neuropsiquiatr 2013;71:677–80. CrossrefPubMedGoogle Scholar

  • 8.

    Sana J, Hajduch M, Michalek J, Vyzula R, Slaby O. MicroRNAs and glioblastoma: roles in core signalling pathways and potential clinical implications. J Cell Mol Med 2011;15:1636–44. Web of SciencePubMedCrossrefGoogle Scholar

  • 9.

    Mittelbrunn M, Sánchez-Madrid F. Intercellular communication: diverse structures for exchange of genetic information. Nat Rev Mol Cell Biol 2012;13:328–35. Web of ScienceCrossrefPubMedGoogle Scholar

  • 10.

    Akers JC, Ramakrishnan V, Yang I, Hua W, Mao Y, Carter BS, et al. Optimizing preservation of extracellular vesicular miRNAs derived from clinical cerebrospinal fluid. Cancer Biomark Sect Dis Markers 2016;17:125–32. Google Scholar

  • 11.

    Teunissen CE, Petzold A, Bennett JL, Berven FS, Brundin L, Comabella M, et al. A consensus protocol for the standardization of cerebrospinal fluid collection and biobanking. Neurology 2009;73:1914–22. PubMedCrossrefWeb of ScienceGoogle Scholar

  • 12.

    Teunissen CE, Tumani H, Bennett JL, Berven FS, Brundin L, Comabella M, et al. Consensus guidelines for CSF and blood biobanking for CNS biomarker studies. Mult Scler Int 2011;2011:e246412. Web of ScienceGoogle Scholar

  • 13.

    Müller M, Kuiperij HB, Claassen JA, Küsters B, Verbeek MM. MicroRNAs in Alzheimer’s disease: differential expression in hippocampus and cell-free cerebrospinal fluid. Neurobiol Aging 2014;35:152–8. PubMedCrossrefWeb of ScienceGoogle Scholar

  • 14.

    Kirschner MB, Kao SC, Edelman JJ, Armstrong NJ, Vallely MP, van Zandwijk N, et al. Haemolysis during sample preparation alters microRNA content of plasma. PLoS One 2011;6:e24145. Web of ScienceCrossrefPubMedGoogle Scholar

  • 15.

    Bache S, Rasmussen R, Rossing M, Laigaard FP, Nielsen FC, Møller K. MicroRNA changes in cerebrospinal fluid after subarachnoid hemorrhage. Stroke 2017;4:2391–8. Web of ScienceGoogle Scholar

  • 16.

    Baraniskin A, Kuhnhenn J, Schlegel U, Chan A, Deckert M, Gold R, et al. Identification of microRNAs in the cerebrospinal fluid as marker for primary diffuse large B-cell lymphoma of the central nervous system. Blood 2011;117:3140–6. Web of SciencePubMedCrossrefGoogle Scholar

  • 17.

    Burgos KL, Javaherian A, Bomprezzi R, Ghaffari L, Rhodes S, Courtright A, et al. Identification of extracellular miRNA in human cerebrospinal fluid by next-generation sequencing. RNA 2013;19:712–22. Web of ScienceCrossrefPubMedGoogle Scholar

  • 18.

    Gallego JA, Gordon ML, Claycomb K, Bhatt M, Lencz T, Malhotra AK. In vivo microRNA detection and quantitation in cerebrospinal fluid. J Mol Neurosci 2012;47:243–8. CrossrefWeb of SciencePubMedGoogle Scholar

  • 19.

    McAlexander MA, Phillips MJ, Witwer KW. Comparison of methods for miRNA extraction from plasma and quantitative recovery of RNA from cerebrospinal fluid. Front Genet 2013;4:83. PubMedGoogle Scholar

  • 20.

    Pritchard CC, Cheng HH, Tewari M. MicroRNA profiling: approaches and considerations. Nat Rev Genet 2012;13: 358–69. CrossrefPubMedWeb of ScienceGoogle Scholar

  • 21.

    Baraniskin A, Kuhnhenn J, Schlegel U, Maghnouj A, Zöllner H, Schmiegel W, et al. Identification of microRNAs in the cerebrospinal fluid as biomarker for the diagnosis of glioma. Neuro-Oncol 2012;14:29–33. CrossrefGoogle Scholar

  • 22.

    Teplyuk NM, Mollenhauer B, Gabriely G, Giese A, Kim E, Smolsky M, et al. MicroRNAs in cerebrospinal fluid identify glioblastoma and metastatic brain cancers and reflect disease activity. Neuro-Oncol 2012;14:689–700. PubMedCrossrefGoogle Scholar

  • 23.

    Yagi Y, Ohkubo T, Kawaji H, Machida A, Miyata H, Goda S, et al. Next-generation sequencing-based small RNA profiling of cerebrospinal fluid exosomes. Neurosci Lett 2017;636:48–57. PubMedWeb of ScienceCrossrefGoogle Scholar

  • 24.

    Sørensen SS, Nygaard A-B, Carlsen AL, Heegaard NH, Bak M, Christensen T. Elevation of brain-enriched miRNAs in cerebrospinal fluid of patients with acute ischemic stroke. Biomark Res 2017;5:24. CrossrefPubMedWeb of ScienceGoogle Scholar

  • 25.

    Denk J, Boelmans K, Siegismund C, Lassner D, Arlt S, Jahn H. MicroRNA profiling of CSF reveals potential biomarkers to detect Alzheimer’s disease. PLoS One 2015;10:e0126423. CrossrefPubMedWeb of ScienceGoogle Scholar

  • 26.

    van Harten AC, Mulders J, Scheltens P, van der Flier WM, Oudejans CB. Differential expression of microRNA in cerebrospinal fluid as a potential novel biomarker for Alzheimer’s disease. J Alzheimers Dis 2015;47:243–52. CrossrefWeb of SciencePubMedGoogle Scholar

  • 27.

    Bhomia M, Balakathiresan NS, Wang KK, Papa L, Maheshwari RK. A panel of serum miRNA biomarkers for the diagnosis of severe to mild traumatic brain injury in humans. Sci Rep 2016;6:28148. CrossrefWeb of ScienceGoogle Scholar

  • 28.

    Quintana E, Ortega FJ, Robles-Cedeño R, Villar ML, Buxó M, Mercader JM, et al. MiRNAs in cerebrospinal fluid identify patients with MS and specifically those with lipid-specific oligoclonal IgM bands. Mult Scler 2017;23:1716–26. Web of ScienceCrossrefPubMedGoogle Scholar

  • 29.

    Gui Y, Liu H, Zhang L, Lv W, Hu X. Altered microRNA profiles in cerebrospinal fluid exosome in Parkinson disease and Alzheimer disease. Oncotarget 2015;6:37043–53. Web of SciencePubMedGoogle Scholar

  • 30.

    Bergman P, Piket E, Khademi M, James T, Brundin L, Olsson T, et al. Circulating miR-150 in CSF is a novel candidate biomarker for multiple sclerosis. Neurol Neuroimmunol Neuroinflammation 2016;3:e219. CrossrefWeb of ScienceGoogle Scholar

  • 31.

    Burgos K, Malenica I, Metpally R, Courtright A, Rakela B, Beach T, et al. Profiles of extracellular miRNA in cerebrospinal fluid and serum from patients with Alzheimer’s and Parkinson’s diseases correlate with disease status and features of pathology. PLoS One 2014;9:e94839. CrossrefPubMedWeb of ScienceGoogle Scholar

  • 32.

    You W-D, Tang Q-L, Wang L, Lei J, Feng J-F, Mao Q, et al. Alteration of microRNA expression in cerebrospinal fluid of unconscious patients after traumatic brain injury and a bioinformatic analysis of related single nucleotide polymorphisms. Chin J Traumatol 2016;19:11–5. CrossrefGoogle Scholar

  • 33.

    Tietje A, Maron KN, Wei Y, Feliciano DM. Cerebrospinal fluid extracellular vesicles undergo age dependent declines and contain known and novel non-coding RNAs. PLoS One 2014;9:e113116. PubMedWeb of ScienceCrossrefGoogle Scholar

  • 34.

    Akers JC, Ramakrishnan V, Kim R, Phillips S, Kaimal V, Mao Y, et al. miRNA contents of cerebrospinal fluid extracellular vesicles in glioblastoma patients. J Neurooncol 2015;123: 205–16. PubMedCrossrefGoogle Scholar

  • 35.

    Sheinerman KS, Umansky SR. Circulating cell-free microRNA as biomarkers for screening, diagnosis and monitoring of neurodegenerative diseases and other neurologic pathologies. Front Cell Neurosci 2013;7:150. PubMedWeb of ScienceGoogle Scholar

  • 36.

    Akers JC, Ramakrishnan V, Kim R, Skog J, Nakano I, Pingle S, et al. MiR-21 in the extracellular vesicles (EVs) of cerebrospinal fluid (CSF): a platform for glioblastoma biomarker development. PLoS One 2013;8:e78115. Web of ScienceCrossrefPubMedGoogle Scholar

  • 37.

    Robinson MD, Oshlack A. A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biol 2010;11:R25. PubMedWeb of ScienceCrossrefGoogle Scholar

  • 38.

    Drusco A, Bottoni A, Laganà A, Acunzo M, Fassan M, Cascione L, et al. A differentially expressed set of microRNAs in cerebro-spinal fluid (CSF) can diagnose CNS malignancies. Oncotarget 2015;6:20829–39. PubMedWeb of ScienceGoogle Scholar

  • 39.

    Grotzer M, Shalaby T, Fiaschetti G, Baulande S, Gerber N, Baumgartner M. Detection and quantification of extracellular microRNAs in medulloblastoma. J Cancer Metastasis Treat 2015;1:67–75. CrossrefGoogle Scholar

  • 40.

    Shi R, Wang P-Y, Li X-Y, Chen J-X, Li Y, Zhang X-Z, et al. Exosomal levels of miRNA-21 from cerebrospinal fluids associated with poor prognosis and tumor recurrence of glioma patients. Oncotarget 2015;6:26971–81. PubMedWeb of ScienceGoogle Scholar

  • 41.

    Tumilson CA, Lea RW, Alder JE, Shaw L. Circulating microRNA biomarkers for glioma and predicting response to therapy. Mol Neurobiol 2014;50:545–58. PubMedWeb of ScienceCrossrefGoogle Scholar

About the article

Corresponding author: Assoc. Prof. Ondrej Slaby, PhD, Central European Institute of Technology (CEITEC), Masaryk University, University Campus Bohunice, Building A35, Kamenice 753/5, 625 00 Brno, Czech Republic, Phone: +420549497574

aAlena Kopkova and Jiri Sana contributed equally to this work.

Received: 2017-10-17

Accepted: 2017-12-28

Published Online: 2018-02-16

Published in Print: 2018-05-24

Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.

Research funding: This work was financially supported by the Czech Ministry of Health grants no. 15-34553A and 15-33158A; by the project MZ CR – RVO (MOU, 00209805); and by the Ministry of Education, Youth and Sports of the Czech Republic under the project CEITEC 2020 (LQ1601).

Employment or leadership: None declared.

Honorarium: None declared.

Competing interests: The funding organization(s) played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.

Citation Information: Clinical Chemistry and Laboratory Medicine (CCLM), Volume 56, Issue 6, Pages 869–879, ISSN (Online) 1437-4331, ISSN (Print) 1434-6621, DOI: https://doi.org/10.1515/cclm-2017-0958.

Export Citation

©2018 Walter de Gruyter GmbH, Berlin/Boston.Get Permission

Citing Articles

Here you can find all Crossref-listed publications in which this article is cited. If you would like to receive automatic email messages as soon as this article is cited in other publications, simply activate the “Citation Alert” on the top of this page.

Alena Kopkova, Jiri Sana, Pavel Fadrus, Tana Machackova, Marek Vecera, Vaclav Vybihal, Jaroslav Juracek, Petra Vychytilova-Faltejskova, Martin Smrcka, Ondrej Slaby, and George Calin
PLOS ONE, 2018, Volume 13, Number 12, Page e0208580

Comments (0)

Please log in or register to comment.
Log in