Skip to main content
Premium Trial:

Request an Annual Quote

Extracellular Vesicles Provide Biomarkers for Glioblastoma Tumor Subtyping


NEW YORK – A group led by researchers at the University of Sussex in the UK has used extracellular vesicles (EVs) to discover RNA biomarkers that might be linked to glioblastoma (GBM).

The team expects to publish a follow-up validation study on the EV-based RNA biomarkers on blood samples from a GBM patient cohort in early 2020.  

Researchers and commercial groups have recently begun to use EVs derived from tumor cells to find biomarkers linked to GBM. Specifically, they believe that RNA biomarkers in the EVs could be linked to tumor resistance to anti-cancer blood vessel growth drugs. In addition, some RNA biomarkers may indicate different levels of tumor aggression.

Current methods to diagnose glioblastoma require a tissue biopsy and the tumor is often not detected until it begins to affect other parts of the brain. Because of the ability of tumor-derived EVs to cross the blood-brain barrier, however, researchers believe they could help detect the disease in earlier stages.

"The new idea is to have a dynamic way to diagnose glioblastoma way earlier using biomarkers associated with [the disease] in a patient's blood," University of Sussex junior postdoc Thomas Simon explained. 

In a study published last week in Communications Biology, Simon, University of Sussex cancer cell signaling professor Georgios Giamas, and their team examined the protein content of EVs produced by different GBM cell lines and patient-derived stem cells. 

The researchers first performed cell invasion assays to determine the colony-forming abilities of astrocytes (as control samples), GBM subtype cells, and patient-derived stem cells in hydrogels. The team then aimed to understand the cells' migration, proliferation, and invasion capabilities, later classifying them based on the phenotypes. 

They then assessed the expression of different markers related to classical and mesenchymal GBM subtypes, finding distinctive expression within the panel of GBM cell lines and stem cells.

After combining and comparing the invasiveness and expression data of protein markers through clustering analysis, the researchers identified seven distinctive signatures. Based on these signatures, they then compared GBM and stem cells, clustering them together according to their similarities. 

"We came up with a cell clustering that was partly reminiscent of GBM subtyping; for example, two of the cell lines we clustered together show a substantial number of mesenchymal characteristics," Simon said. "Following this, we concentrated the EVs from the same cells and described their proteomic content."

The team also saw that the vesicles had a size range of 50 nm to 150 nm, confirming EV isolation from the different GBM cell cultures. Using mass spectrometry, the researchers then deciphered the proteomic content of EVs derived from various GBM subtypes. They found that the EV protein content, as well as specific biomarkers like CD44, could partly mirror the team's initial cell cluster. 

"Our results suggest TGFBI and SERPINE1 as potential [small] EV-associated biomarkers for the aggressive mesenchymal subtype," the study authors noted. "Both our data and recent publications suggest that changes in the EV-specific marker expression patterns could help identify highly invasive [and] aggressive tumors." 

At the same time, Simon acknowledged that his team ran into several technical challenges when developing the EV-based biomarkers. While the researchers initially struggled to grow GBM cells to perform methods including Western blotting and mass spectrometry, they eventually were able to develop a classification that represents or mirrors glioblastoma subtype, he said. 

One of the study's limitations involved using immortalized tumor cell lines for studying GBM subtypes, which Simon noted is not representative of patient tumors. He said that the team will need to perform additional validation work in an in vivo setting and consider the role that patients' blood microenvironment may play in analyzing EVs. 

Massachusetts General Hospital Chief of Neurosurgery Bob Carter, who was not involved in the study, also highlighted that one of the major challenges of liquid biopsy-based approaches is the admixture of tumor EVs with normal physiologic EVs. While EVs may have strong potential as biomarkers for abundant tumor-specific signatures in GBM, he also stressed the importance of validating the signatures in patient samples in specific biofluids such as plasma and cerebrospinal fluid (CSF). 

Simon and his team are currently collaborating with the Brighton and Sussex University Hospital on a recently launched validation study that involves collecting blood samples from patients who have undergone glioblastoma resection surgery. The group is collecting about 5 ml of blood from each patient and aims to examine samples from about 30 patients, in addition to control samples. 

According to Simon, the study's main goal is to see if each GBM subtype produces different patterns of biomarkers in EVs than the ones previously identified in the immortalized cell lines. The team also aims to see if the pathways it observes in the clinical cohort could be linked to subtypes of GBM. 

"From using patient samples, we want to see if we can find any clinical value from the biomarkers," Simon said. "We want to compare what occurs in [controls] to what we observe in GBM patients."

However, Simon acknowledged that his team will have a much harder time examining patient samples than in vitro GBM cancer cell lines due to the high heterogeneity in patients' EVs. While the researchers are currently collecting GBM patient samples and data, Simon expects to complete the study by early 2020. 

Clinical applications

Simon believes that the new study's results may eventually lead to GBM-specific subtype diagnoses, based on the different biomarkers and pathways they identified. In addition, he envisions the biomarkers guiding targeted, personalized treatment and helping clinicians predict GBM relapse in cancer patients.  

Memorial Sloan Kettering Cancer Center researcher Ingo Mellinghoff, who is not involved in the UK-based research, also believes EVs may be able to cross the blood brain barrier and potentially provide a molecular snapshot of the GBM tumor. He sees that EVs — compared to cell-free DNA — could also provide additional information about a broad set of analytes, including proteins, RNA, lipids, and metabolites.

As Simon's group seeks to develop a diagnostic tool based on EVs for brain cancer, other researchers are using analytes such as circulating tumor DNA (ctDNA) in order to detect or monitor brain cancer in patients.

Mellinghoff's own team at MSKCC has begun examining ctDNA in cerebral spinal fluid (CSF) in order to detect cases of glioblastoma. The researchers anticipate using the technique to eventually monitor the evolution and status of gliomas in patients who have undergone brain tumor surgeries and subsequent therapy. 

Researchers at the University of California, San Francisco are also testing CSF, as well as blood samples, from pediatric patients diagnosed with diffuse midline glioma. The team developed a droplet digital PCR assay targeting driver mutations, finding that changes in circulating DNA correlated with radiation response.

Rather than viewing EVs as a better method for analyzing biomarkers linked to GBM than using ctDNA or circulating tumor cells (CTCs), Simon believes that researchers should use a combined approach, applying all three tools for biomarker detection. 

"By looking at CTCs or [cell-free tumor DNA] and combining them into a pattern, you could identify patterns in terms of specific pathways," Simon said. "Now, it depends on what methods and particles you'd want to use to concentrate on, since contamination is an issue."

Mellinghoff agreed that researchers could eventually use both EVs and ctDNA in CSF, and potentially blood, from GBM patients to begin correlating measurements in liquid samples with results from solid tumor tissue samples.  However, he argued that the UK researchers will need to demonstrate that they can reliably detect EVs in blood and CSF in future validation studies. Mellinghoff emphasized that most of the current results regarding EVs — including their biological role and cargos— are "only descriptive" and that researchers have limited information as to whether EVs can provide enough data concerning clinical specificity and sensitivity. 

"Since there are currently no therapies that are specific for any of the molecular GBM subtypes, it would also be important to determine whether EVs can provide quantitative information regarding specific biomarkers, such as mutant oncoproteins," Mellinghoff said. 

Like Mellinghoff, Carter thinks that EVs have a special advantage as a source for RNA biomarkers, which rapidly degrade when outside the cell. However, he also believes that a combination of ctDNA and EVs hold "particular promise" to increase the sensitivity of liquid biopsy assays by capturing both RNA and DNA signals. 

In addition, MGH neurosurgery professor Leonora Balaj noted that each analyte type could be informative at different stages of GBM.

"For example, CTCs may be released in advanced stages of [GBM], while exosomes may be informative even in the early stages of the disease," Balaj explained. 

While Simon highlighted that EVs contain a lot of actionable data for cancer detection, he acknowledged that EV purity is typically weak and that his team lacks optimized methods to guarantee the biomarkers it might identify derive from the tumors. In addition, he noted that vesicles contain a wide variety of elements — including lipids, proteins, and genetic content — that do not relate to the biomarkers.