A single microfluidic chip, loaded with human brain cells grown from stem cells, now mimics the oxygen starvation and cell death patterns of a stroke. This “stroke-on-a-chip” replaces the rodent neuronal cultures currently used in the lab’s bespoke model, and aims to become the standard tool for certain stroke studies that still require live animals. The problem is stark: thousands of drugs that protect neurons in rodent stroke models have failed in human clinical trials. Rodent cells do not faithfully replicate human stroke pathophysiology, so promising animal results often do not translate. This chip reproduces the spatial and temporal gradients of oxygen-glucose deprivation—from the lethal core to the salvageable penumbra—that existing in vitro systems cannot capture. It also enables real-time, quantitative measurements of how cells respond at different distances from the injury. If validated, the model could replace roughly 100 animals per year in the applicants’ labs and up to 1,000 per year across the UK, according to letters of support. More importantly, a human-based system should yield more clinically relevant data on stroke mechanisms and potential therapies, sparing animal lives while improving the odds that a neuroprotective drug actually works in patients.
View original technical description
Aim: This PhD project aims to achieve a significant Replacement of animals, in 2 distinct ways: by replacing the use of rodent neuronal cells with human iPS-derived neuronal cells in our bespoke microfluidics stroke model we have developed in-house, and by making the enhanced, “humanised” microfluidics model the “go-to” model for some areas of stroke research that currently require the use of animals. Context: This work is targeting the replacement of rodents used in in vitro stroke research and more gradually in vivo work on pathomechanisms and prospective therapeutics. This model is expected to replace per year ~100 animals in Carswell and Gibson labs, and ~1000 animals in UK according to enclosed letters of support (detailed metrics in Section 4). Thousands of agents are neuroprotective in animal stroke models but have not translated to the clinic. Therefore, apart from sparing animal lives, our human-based model has additional health benefits of better recapitulating human pathophysiology, providing greater translational value. Objectives: We have developed a ‘stroke-on-a-chip’ microfluidic model that currently uses mouse primary neuronal cultures. The objectives of this work are to develop, validate, and apply this model using human induced pluripotent stem cell (iPS)-derived neuronal cells instead of rodent neuronal cultures to transform it into a human-based stroke model. The PhD student has access to support for all methodologies and established protocols we are proposing. This work will succeed because we have established protocols for access by the student and because two main features make our model stand out from other human stroke models: 1) Our model replicates the spatial and temporal gradients of oxygen glucose deprivation spreading from ‘core’ to ‘penumbra’ found in vivo which cannot be achieved using existing in vitro systems. This will provide crucially important information on critical thresholds of OGD that induce: loss of neuronal viability, connectivity, electrical function and protein synthesis in core and penumbra, all of which could be salvaged by interventional therapeutics. 2) Our multi-chamber microfluidics system allows quantitative, real-time measurements of cell-cell responses in remote and proximal regions to the ischaemic insult. This crucially important knowledge on spatial definition of cellular reactivity and their modulation by therapeutics cannot be acquired in conventional systems. In summary, we are uniquely poised to replace the use of animals to investigate stroke pathomechanisms and prospective therapeutics, which together with our strong dissemination plan, will achieve a lasting 3Rs legacy beyond the PhD project.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know