A little more than a decade ago, Lawrence Livermore researchers embarked on an ambitious endeavor to combine biology and engineering into the iChip (in vitro chip-based human investigational platform)—an experimental device for evaluating the effects of potentially harmful chemicals, viruses, and drugs in a human-based system without relying on animal or human test subjects. Since then, Laboratory teams have refined and leveraged these platforms to better understand human response mechanisms under stress, such as with infection or biothreat exposure, and to drive advances in therapeutic and medical countermeasure development.
The Laboratory’s brain-on-a-chip has become a hallmark of this bioengineering innovation. Over the years, internal investments through the Laboratory Directed Research and Development (LDRD) Program have supported multidisciplinary teams in honing the technology’s capabilities and applications. The platform contains an engineered device for culturing and maintaining brain cells, including neurons and other cells such as astrocytes, microglia, and oligodendrocytes. These cells are grown atop microelectrode arrays (MEAs), which measure the electrical signals generated by neurons as they relay information to one another, yielding intricate neuronal information that can be analyzed to evaluate cell and tissue responses to external stimuli.
Initially developed as a 2D system using primary rat cell cultures, the brain-on-a-chip has evolved to include multiple cell types, including human-induced pluripotent stem cells, within a 3D MEA structure. (See S&TR, May 2019, Small Brain-on-a-Chip Promises Big Payoffs.) In addition, the 3D MEA technology has been integrated into a novel Laboratory-developed neurovascular unit (NVU) that also incorporates a blood–brain barrier (BBB) for more precise recapitulation of the brain’s neurovascular environment. (See the box below.) These advances have enabled three brain-on-a-chip configurations—2D, 3D, and NVU—each of which provides key data across the application space.
In correlated studies, project teams have demonstrated the device’s utility, while others have focused on enhancing data acquisition and incorporating computational methods for improved data analytics and interpretation. With each progressive step, researchers have increased the device’s functionality, bringing attention to its potential as a next-generation testing platform for mission-relevant and human health applications. Livermore biologist Nick Fischer, the co-principal investigator for the 3D MEA strategic initiative, says, “Internal investments have springboarded this technology to where we are today, and now we are collaborating with external sponsors to drive key aspects of the technology to address critical scientific questions.”
Proof in Practice
Staff scientist and biologist Heather Enright and Livermore neuroscientist Doris Lam have both led brain-on-a-chip LDRD initiatives, including those to test the device’s context of use. Enright says, “In 2D, we can quickly screen different drugs, pathogens, and insults of interest, then down-select conditions for testing in our advanced 3D systems (3D and NVU), which are lower throughput but more physiologically relevant. We leverage these different formats in our experimental pipeline across our projects.”
In one study, Enright’s team used the 2D platform to noninvasively evaluate changes in neuronal function over time for both rat and human cells subjected to varying doses of the highly potent opioid, fentanyl. The researchers then compared results derived from the system to those obtained through in vivo rodent studies. Through this work, the team successfully demonstrated the effectiveness of the device to recapitulate dose-dependent outcomes, which can be integrated into a broader approach for evaluating other compounds of national interest and for identifying new potential treatments.
Lam’s project illustrated how the device can serve as a test bed for therapeutics. Military operations can place personnel at significant risk for traumatic brain injury (TBI), and although medical interventions can help mitigate the injury, therapeutics that can repair damaged tissue would be a medical breakthrough. Lam and colleagues chemically reprogrammed astrocytes—support cells grown on the platform—into working nerve cells, monitored their growth and development into neuronal networks, and tracked long-term changes to their function over months. In addition to leveraging the device to experimentally validate the reprogrammed cells, the project also advanced computational methods for improved data analyses.
“Similar to how online social networks can be analyzed to understand how people make friends or build relationships, we can apply the same computational models to neuroscience,” says computer scientist Jose Cadena, whose expertise in graph-based modeling allowed Lam’s research team to analyze how the brain creates different network “communities” and which ones are related to different functions or behaviors. “We can analyze the types of neurons talking to one another over time, and whether different functional communities within the network are connecting to each other, and how those connections change under different experimental conditions.” Each electrode on the device becomes one entity, or node, in the graph, and then connections, called edges, are made based on highly correlated activity. In the case of the TBI study, graph-based models helped identify the types of neurons involved: whether a neuron is excitatory (tends to activate other neurons) or inhibitory (tends to reduce activity). More recently, Cadena led an LDRD project to further develop this computational approach, using graph-based models to identify and monitor functional changes in cell cultures after exposure to a sarin surrogate and to evaluate recovery after administration of a standard treatment.
Although further research is needed, this computational method can be applied across a range of applications. Says Lam, “Our 2D and 3D MEAs contain thousands of neurons. Computational models can help us extract more information from the recorded data to better understand neurological effects and outcomes.” Fischer adds, “The next level of computation will include analyses of different data streams to synergistically deepen insight into responses from a variety of insults and help predict how infections, chemical agents, or countermeasures will realistically affect humans.”
Partnering for Success
In 2024, Enright and Fischer led a one-year seed project with the Defense Threat Reduction Agency (DTRA) to demonstrate the efficacy of the 2D system for recapitulating infection and its potential for countermeasure screening. The success of this project has led to a five-year follow-on effort funded by DTRA to expand scope to include both the 2D and 3D NVU models for pathogen infection and countermeasure screening and includes engagement with the United States Army Medical Research Institute of Infectious Disease for exposures requiring high containment. This work falls under DTRA’s CAMO (comparing animal models to organ tissue equivalents) program, which aims to leverage organ-on-a-chip models to generate human-relevant data earlier in the experimental pipeline and to facilitate reduction of animal testing.
“We are currently the only central nervous system–based platform that DTRA is funding as part of this larger effort,” says Enright. “In addition to the expanded scope for new pathogens and countermeasures of interest, the project has specific milestones for ruggedizing our systems for their use in high-containment facilities. These milestones are critical to enable our work with pathogens and will facilitate transfer of our technology to our collaborators,” says Enright.
A continued focus for these studies will include species comparisons. Indeed, part of the one-year seed project was to evaluate human versus rat responses to infection to identify species differences and to evaluate if both could be used for countermeasure screening. Noah Goshi, a Livermore staff scientist who was integral to the development of the NVU and who leads aspects of the DTRA project, is developing a mouse cell–based platform that can be compared to large existing data repositories for in vivo responses to pathogens of interest, allowing the team to validate data off the brain-on-a-chip models for the CAMO program.
Complexity Mirrors Reality
In 2022, Livermore staff scientist and biologist Heather Enright spearheaded a disruptive research project funded through the Laboratory Directed Research and Development (LDRD) Program to understand the effects of SARS-CoV-2 infection and prolonged exposure on brain function. Although the project utilized both the 2D and 3D brain-on-a-chip platforms, a primary focus of the research was further developing the 3D system to create a neurovascular unit (NVU) that incorporates a blood–brain barrier (BBB) in addition to brain cells. “Having one platform with both blood–brain barrier and brain components allows us to more accurately recapitulate the brain and mirror a more relevant exposure (through the blood–brain barrier), which is critical for both threat and countermeasure assessment,” says Enright.
Within the 3D platform, designed and fabricated by Livermore engineer Michael Triplett, a coated steel pin is emplaced through the hydrogel-extracellular matrix material that surrounds the cultured cells and the corresponding microelectrode arrays. As the cells grow and mature, the pin serves as a placeholder for the BBB. “When the time comes, we remove the molding pin, which leaves a void within the gel. We then use a micro syringe needle to inject endothelial cells into the cavity,” says Livermore staff scientist Noah Goshi. “The cells mature for several days to form the intact blood-brain barrier, after which we can test the system.” Follow-on LDRD studies have demonstrated how substances can be flowed through the endothelial channel across the device and what effects these substances have on the integrity of the BBB boundary. The NVU is also being further developed for high containment use as part of the study funded by the Defense Threat Reduction Agency.
Looking ahead, this research will be augmented by work done on the computational side. “We are starting to engage with Battelle as part of a program for specific context of use cases,” says Enright. The broader program integrates a computational component with the expectation that a database will be made available so that teams can benchmark results in their systems and evaluate their relevance to animal data. These studies will help researchers draw similarities and differences between species to inform next steps and streamline the selection of potential countermeasures.
Goshi is also working on a version of the platform that incorporates olfactory cells to model that particular route of infection. He says, “Inhaled pathogens or toxins can travel through olfactory receptor neurons, which form a direct connection between the air one breathes and the brain.” The model will have brain cells on one side and olfactory receptor neurons on the other with small single direction pathways to mimic the anatomical connection between the nose and the brain. The olfactory side can be flooded with the virus and its progression tracked into the brain side. Goshi says, “We are trying to understand how quickly that process occurs and identify specific time points in which medical countermeasures might be particularly effective.”
With its many potential applications and growing capabilities, the Laboratory’s brain-on-a-chip serves as another example of the transformative breakthroughs that are possible with the support of internal investments, foresight into emerging needs, and a whole-of-Laboratory approach that takes advantage of multidisciplinary expertise. “At Livermore, we are good at building technologies, and bringing what we build into an application space is absolutely critical to what we are trying to achieve,” says Fischer. “Over the past few years, we have been applying our brain-on-a-chip models to critical mission areas. Now, we are excited to expand our applications with new sponsors and collaborators.”
—Caryn Meissner
For further information contact Heather Enright (925) 423-8232 (enright3 [at] llnl.gov (enright3[at]llnl[dot]gov)).
