Wyss researchers developed a technology framework to isolate and characterize brain-derived extracellular vesicles and signal the brain diagnostics field to take new directions
By Benjamin Boettner
(BOSTON) – A new collaborative study by David Walt, Ph.D.’s and George Church, Ph.D.’s teams at the Wyss Institute shows that a commonly targeted protein that researchers hoped would allow them to obtain liquid biopsies of neuronal tissue in the brain from patients’ blood does not live up to the task. Their findings, published in Nature Methods, signal that alternative approaches need to be developed to move the field forward.
The diagnosis and monitoring of brain diseases – including neurodegenerative diseases like Parkinson’s and Alzheimer’s disease, a variety of mental disorders, as well as brain tumors – remains a Herculean challenge. Brain biopsies cannot be easily performed, and much of the understanding of what causes and advances brain diseases is based on postmortem tissue analysis, which obviously does not help living patients.
A vision of liquid brain biopsies
One potential solution to this problem are neuron-derived extracellular vesicles (NDEVs), tiny membrane-enclosed particles that are released by neuronal cells in the brain into the surrounding cerebrospinal fluid (CSF) at a constant rate, and contain many of the molecules that neurons produce, including nucleic acids, proteins, and lipids. These intriguing vesicles, besides allowing cells in the brain to communicate with each other, importantly, also cross the blood-brain barrier, which functions as a security system that tightly regulates the exchange of molecules between the blood stream and brain.
Researchers speculate that, once circulating in the blood stream, NDEVs could be isolated in a minimally-invasive and repetitive way, and investigated for their neuron-specific contents, and potential biomarker signatures that emerge during disease progression.
To advance this idea, the groups of Wyss Core Faculty members David Walt, and George Church, and neurodegenerative disease expert Alice Chen-Plotkin, M.D., in 2018, set out to lay the ground work for isolating and characterizing NDEVs in a collaborative project that is funded by the Chan Zuckerberg Initiative’s Neurodegenerative Challenge Network. The ultimate aim is to get a much better handle on diagnosing and monitoring Parkinson’s disease.
Walt is also the Hansjörg Wyss Professor of Biologically Inspired Engineering at Harvard Medical School (HMS), Professor of Pathology at Brigham and Women’s Hospital, and a Howard Hughes Medical Institute (HHMI) Professor. Church is also Professor of Genetics at HMS, and Professor of Health Sciences and Technology at Harvard and MIT. Alice Chen-Plotkin is the Parker Family Associate Professor of Neurology at Penn Medicine.
Up a blind alley and a way out
The vision of using NDEVs as a surrogate for brain biopsies stands and falls with the ability to effectively isolate them from blood plasma. Until now, researchers in the field have been strongly relying on a transmembrane protein called L1CAM that was thought to be enriched in NDEVs and an antibody that specifically latches on to it, allowing NDEVs to be pulled out of blood plasma and other fluids. “The concept of non-invasively reading out the molecular contents of neurons from EVs in CSF or plasma is very exciting, but as the EV field is still quite new, there are many technical challenges to getting this to really work,” said co-first author Dima Ter-Ovanesyan, Ph.D., a Wyss Research Scientist working with Church who is spearheading the project together with Maia Norman, an M.D.-Ph.D. graduate student on Walt’s team. “Although there are many papers that have used L1CAM for EV immuno-isolation, some initially unexpected results by our team led us to step back and investigate L1CAM more thoroughly. In the process, I think the tools we developed will be helpful for validating other candidate NDEV markers.”
Norman and Ter-Ovanesyan developed a framework of technologies that enabled them to specifically isolate NDEVs from other components of the fluids they can be found in, and analyze with extreme sensitivity candidate surface proteins such as the purported L1CAM for their potential to allow their capture with specific antibodies.
The first part of their approach uses a traditional method that separates molecules and larger molecular structures based on their different sizes. By fine-tuning this fractionation method, also known as “size exclusion chromatography” (SEC), they were able to concentrate NDEVs in only a few fractions, and thus separate them from larger structures on one end of the size spectrum and from free molecules on the other.
To then further investigate whether NDEVs can be isolated from individual CSF or blood plasma fractions using the L1CAM-specific antibody, they made use of a second assay, called a Simoa assay, that was developed in Walt’s group. Simoa, a digital version of the often-used versatile ELISA assay, allows researchers to capture a protein (and its associated vesicle) of interest with a specific antibody molecule that is linked to a magnetic bead. It then detects the bound protein with a second antibody that produces a fluorescent signal. Since “sandwiching” the target protein by the two antibodies can be performed with one bead-bound antibody in an individual well of a multi-well plate at a single-molecule level, Simoa enables researchers to count individual capture events, allowing much higher sensitivity than a conventional ELISA assay does.
To be able to evaluate whether the L1CAM-specific antibody would indeed preferentially bind to L1CAM located on NDEVs, the team developed additional reference Simoa assays for three different membrane proteins that are generally present on EVs, and another one that detects a soluble protein called albumin that never can be found in EVs. “Much to our surprise we saw that the commonly used immunocapture target L1CAM eluted in the albumin-containing fractions and not in fractions that contained EVs,” said Norman. “This finding clearly shows that the vast majority of L1CAM in CSF and blood is not associated with EVs, and serves to redirect the field away from the use of L1CAM and towards the use of novel targets for isolating NDEVs.”
The reason that L1CAM/L1CAM antibody system failed as a useful NDEV isolation approach likely is due to the fact that L1CAM is also produced by cells as a version that does not insert in the membranes of neuronal cells and their derived EVs, and instead is secreted into cerebrospinal fluid. Alternatively, the portion of the L1CAM protein that extrudes out of NDEVs and is recognized by the L1CAM-specific antibody can be cleaved off by specific proteinase enzymes and released into cell-surrounding fluids.
Refocusing through discussion
Walt recently presented the team’s study at the 2021 conference held by the International Society for Extracellular Vesicles in a session titled “The L1CAM Controversy: Pulling Down Consensus: ISEV and Michael J. Fox Foundation Session” and attended by more than 400 experts, including neurologists and diagnostics developers in the field. “Our report has stimulated robust discussion in this multi-disciplinary community, which we hope will rapidly shift directions to develop alternative methods for reliably isolating NDEVs,” said Walt. “The framework we built provides an excellent basis for validating future candidate reagents that are so urgently needed to bring liquid biopsies closer to their clinical application.”
Additional authors on the study are Wyss Institute researchers Wendy Trieu, Roey Lazarovits, and Ju Hyun Lee; Emma Kowal at MIT; and Aviv Regev, Ph.D., Professor of Biology at MIT, Core Member of the Broad Institute, and Investigator at the HHMI. The study was funded by the Chan Zuckerberg Initiative (CZI), Good Ventures, the National Institutes of Health Center for Excellence in Genomic Science, the HHMI, and the Klarman Cell Observatory.