Time matters when moving and growing fast: predicting glioblastoma prognosis and clinical outcome in a patient-specific manner

We developed a microfluidic assay to enumerate and isolate highly proliferative and motile cells from intratumorally heterogenous population of patient-derived glioblastoma cells. In our Nature BME article, we show that the relative abundance of these cells is predictive of clinical outcomes.
Time matters when moving and growing fast: predicting glioblastoma prognosis and clinical outcome in a patient-specific manner

For patients with malignant cancer, early advance planning of treatment and management regimen is important to guide patient-specific needs and palliative care. In the case of glioblastoma (GBM), the most common and devastating primary brain cancer with a median survival of 14.6 months, it is crucial to develop an individualized clinical plan. Despite thorough state-of-the art surgical resection, followed by concurrent chemo- and radio-therapy, GBMs are incurable and recur frequently. Notwithstanding our growing understanding of the clinical and surgical parameters, and the molecular and cellular tumor characteristics that influence GBM patient survival, no standardized and reproducible point-of-care methodology is currently available to predict patient prognosis. The lack of such a patient-oriented prognostic tool for GBM is largely due to the inter- and intra-tumoral heterogeneity of molecular markers, which adds complexity and cost to single cell analysis. Although patient-derived xenografts recapitulate key aspects of the tumor biology and microenvironment, they are costly, laborious, and successful only in a small percentage of cases. Thus, our groups merged forces to develop a tool to accurately predict patient-specific mortality and disease trajectory

It is well appreciated that highly metastatic subpopulations of cancer cells have enhanced motility that is intimately linked to the aggressiveness of the disease. However, motility alone tells only part of the story. To colonize distant sites, motile cells also need to squeeze through confining spaces and proliferate rapidly. Recently, the Konstantopoulos lab developed a novel Microfluidic Assay for quantification of Cell Invasion (MAqCI) assay, which showed remarkable capability in predicting the metastatic potential of breast cancer cell lines by assessing their migratory and proliferative capacities. While accumulating evidence indicates that the migratory behavior of GBM cells is qualitatively informative in delineating tumor aggressiveness, no effective quantitative approach has been developed. We posited that prior work failed to account for the proliferative potential of highly motile cells. Moreover, previous studies failed to model key features of the complex human brain topography. Naturally, we wondered whether a microfluidic platform that mimics the confined architecture of perivascular conduits and white-matter tracts of the human brain parenchyma could be exploited to better phenocopy the native microenvironment of GBMs. If so, can we study GBM cell decision-making in real-time during its migratory journey and residency in a confined space model of its native microenvironment? Would the behavioral pattern of the GBM cells instruct us about the aggressiveness of the disease? Can we ultimately develop a method to distinguish a subpopulation of migratory and proliferative cells within a patient-derived GBM biopsy as a metric for predicting individual patient clinical prognosis? 

Making use of patient-derived clinical specimens available to us directly from the operating room (Quiñones-Hinojosa Lab), we sought to answer these questions in double-blinded retrospective and prospective studies. Using therapy-naive cells isolated from primary GBM patients, we measured the relative abundance of highly motile cells that are capable of navigating and squeezing through the microchannels, as well as the proliferative capacity of this highly motile cell subpopulation. By combining these phenotypic features into a single composite score, we retrospectively categorized each patient by their progression-free survival time into short or long-term survival outcomes and predicted the time to recurrence with high sensitivity, specificity, and accuracy. Our retrospective-based findings provided an impetus to further test the efficacy of MAqCI in a pilot prospective study, which remarkably, predicted the prognosis of all patients in the cohort.

Primary GBM cells harvested from patients following surgical resection are allowed to migrate in MAqCI, which recapitulates key aspects of the complex topography and the confining microenvironment of the brain parenchyma. Using the migratory and proliferative potentials of the cells to generate a composite MAqCI score, which is then used to predict patient-specific prognosis. Higher composite MAqCI score correlates with shorter progression-free survival prognosis and lower time to recurrence.

For GBM patients, time matters when the cells are moving and growing fast. Given that our prognostication is currently restricted to binary survival times, in the future, we hope to fine tune our methodology to precisely quantitate clinical outcomes. In addition, we envision to utilize this platform to screen potential therapeutics to stop these migratory cancer cells on their tracks.  Likewise, at a molecular level, can we interrogate the “go” or “grow” mechanisms of cancer cells to put the brakes on these devastating cancer cells? Can this technology be adapted for predicting aggressiveness of other invasive cancers? Moving forward, we aim to address these fundamental and clinical questions using MAqCI.

Given the uncertainty of this disease, the time-to-death until now has remained remarkably unpredictable for each GBM patient. Our work now offers a promising and robust approach to predict individualized clinical outcomes for GBM patients, which may equip patients and clinicians to strategize effectively in fighting a disease where time is of the essence. 

Link to our manuscript: A microfluidic cell-migration assay for the prediction of progression-free survival and recurrence time of patients with glioblastoma

Please sign in or register for FREE

If you are a registered user on Research Communities by Springer Nature, please sign in